Archive for the ‘Na-K-ATPase’ Category

The Reconstruction of Life Processes requires both Genomics and Metabolomics to explain Phenotypes and Phylogenetics

Writer and Curator: Larry H. Bernstein, MD, FCAP 





This discussion that completes and is an epicrisis (summary and critical evaluation) of the series of discussions that preceded it.

  1. Innervation of Heart and Heart Rate
  2. Action of hormones on the circulation
  3. Allogeneic Transfusion Reactions
  4. Graft-versus Host reaction
  5. Unique problems of perinatal period
  6. High altitude sickness
  7. Deep water adaptation
  8. Heart-Lung-and Kidney
  9. Acute Lung Injury

The concept inherent in this series is that the genetic code is an imprint that is translated into a message.  It is much the same as a blueprint, or a darkroom photographic image that has to be converted to a print. It is biologically an innovation of evolutionary nature because it establishes a simple and reproducible standard for the transcription of the message through the transcription of the message using strings of nucleotides (oligonucleotides) that systematically transfer the message through ribonucleotides that communicate in the cytoplasm with the cytoskeleton based endoplasmic reticulum (ER), composing a primary amino acid sequence.  This process is a quite simple and convenient method of biological activity.  However, the simplicity ends at this step.  The metabolic components of the cell are organelles consisting of lipoprotein membranes and a cytosol which have particularly aligned active proteins, as in the inner membrane of the mitochondrion, or as in the liposome or phagosome, or the structure of the  ER, each of which is critical for energy transduction and respiration, in particular, for the mitochondria, cellular remodeling or cell death, with respect to the phagosome, and construction of proteins with respect to the ER, and anaerobic glycolysis and the hexose monophosphate shunt in the cytoplasmic domain.  All of this refers to structure and function, not to leave out the membrane assigned transport of inorganic, and organic ions (electrolytes and metabolites).

I have identified a specific role of the ER, the organelles, and cellular transactions within and between cells that is orchestrated.  But what I have outlined is a somewhat limited and rigid model that does not reach into the dynamics of cellular transactions.  The DNA has expression that may be old, no longer used messages, and this is perhaps only part of a significant portion of “dark matter”.  There is also nuclear DNA that is enmeshed with protein, mRNA that is a copy of DNA, and mDNA  is copied to ribosomal RNA (rRNA).  There is also rDNA. The classic model is DNA to RNA to protein.  However, there is also noncoding RNA, which plays an important role in regulation of transcription.

This has been discussed in other articles.  But the important point is that proteins have secondary structure through disulfide bonds, which is determined by position of sulfur amino acids, and by van der Waal forces, attraction and repulsion. They have tertiary structure, which is critical for 3-D structure.  When like subunits associate, or dissimilar oligomers, then you have heterodimers and oligomers.  These constructs that have emerged over time interact with metabolites within the cell, and also have an important interaction with the extracellular environment.

When you take this into consideration then a more complete picture emerges. The primitive cell or the multicellular organism lives in an environment that has the following characteristics – air composition, water and salinity, natural habitat, temperature, exposure to radiation, availability of nutrients, and exposure to chemical toxins or to predators.  In addition, there is a time dimension that proceeds from embryonic stage to birth in mammals, a rapid growth phase, a tapering, and a decline.  The time span is determined by body size, fluidity of adaptation, and environmental factors.  This is covered in great detail in this work.  The last two pieces are in the writing stage that completes the series. Much content has already be presented in previous articles.

The function of the heart, kidneys and metabolism of stressful conditions have already been extensively covered in  in the following and more:

The Amazing Structure and Adaptive Functioning of the Kidneys: Nitric Oxide – Part I

Nitric Oxide and iNOS have Key Roles in Kidney Diseases – Part II

The pathological role of IL-18Rα in renal ischemia/reperfusion injury –

Summary, Metabolic Pathways



Read Full Post »

Depth Underwater and Underground

Writer and Curator: Larry H. Bernstein, MD, FCAP 



Deep diving for mammals is dangerous for humans and land based animals for too long, and it has dangerous consequences, most notable in nitrogen emboli  with very deep underwater diving. Other mammals live in water and have adapted to a water habitat.  This is another topic that needs further exploration.

Deep diving has different meanings depending on the context. Even in recreational diving the meaning may vary:

In recreational diving, a depth below about 30 metres (98 ft), where nitrogen narcosis becomes a significant hazard for most divers, may be considered a “deep dive”

In technical diving, a depth below about 60 metres (200 ft) where hypoxic breathing gas becomes necessary to avoid oxygen toxicity may be considered a “deep dive”.

Early experiments carried out by Comex S.A. (Compagnie maritime d’expertises) using hydrox and trimix attained far greater depths than any recreational technical diving. One example being the Comex Janus IV open-sea dive to 501 metres (1,644 ft) in 1977. The open-sea diving depth record was achieved in 1988 by a team of Comex divers who performed pipe line connection exercises at a depth of 534 metres (1,752 ft) in the Mediterranean Sea as part of the Hydra 8 program. These divers needed to breathe special gas mixtures because they were exposed to very high ambient pressure (more than 50 times atmospheric pressure).

Then there is the adaptation to the water habitat as a living environment. The two main types of aquatic ecosystems are marine ecosystems and freshwater ecosystems.

Marine ecosystems are part of the earth’s aquatic ecosystem. The habitats that make up this vast system range from the productive nearshore regions to the barren ocean floor. The marine waters may be fully saline, brackish or nearly fresh. The saline waters have a salinity of 35-50 ppt (= parts per thousand). The freshwater has a salinity of less than 0.5 ppt. The brackish water lies in between these 2. Marine habitats are situated from the coasts, over the continental shelf to the open ocean and deep sea. The ecosystems are sometimes linked with each other and are sometimes replacing each other in other geographical regions. The reason why habitats differ from another is because of the physical factors that influence the functioning and diversity of the habitats. These factors are temperature, salinity, tides, currents, wind, wave action, light and substrate.

Marine ecosystems are home to a host of different species ranging from planktonic organisms that form the base of the marine food web to large marine mammals. Many species rely on marine ecosystems for both food and shelter from predators. They are very important to the overall health of both marine and terrestrial environments. Coastal habitats are those above the spring high tide limit or above the mean water level in non-tidal waters.  They are close to the sea and include habitats such as coastal dunes and sandy shores, beaches , cliffs and supralittoral habitats. Coastal habitats alone account for approximately 30% of all marine biological productivity.

All plant and animal life forms are included from the microscopic picoplankton all the way to the majestic blue whale, the largest creature in the sea—and for that matter in the world. It wasn’t until the writings of Aristotle from 384-322 BC that specific references to marine life were recorded. Aristotle identified a variety of species including crustaceans, echinoderms, mollusks, and fish.
Today’s classification system was developed by Carl Linnaeus external link as an important tool for use in the study of biology and for use in the protection of biodiversity. Without very specific classification information and a naming system to identify species’ relationships, scientists would be limited in attempts to accurately describe the relationships among species. Understanding these relationships helps predict how ecosystems can be altered by human or natural factors.

Preserving biodiversity is facilitated by taxonomy. Species data can be better analyzed to determine the number of different species in a community and to determine how they might be affected by environmental stresses. Family, or phylogenetic, trees for species help predict environmental impacts on individual species and their relatives.

For generations, whales and other marine mammals have intrigued humans. 2,400 years ago, Aristotle, a Greek scientist and philosopher, recognized that whales are mammals, not fish, because they nurse their young and breathe air like other mammals. There are numerous myths and legends surrounding marine mammals. The Greeks believed that killing a dolphin was as bad as murdering a human. An Amazon legend said that river dolphins came to shore dressed as men to woo pretty girls during fiestas. During the Middle Ages, there were numerous legends surrounding the narwhals’ amazing tusk, which was thought to have come from the unicorn.

Insert movie

Marine mammals evolved from their land dwelling ancestors over time by developing adaptations to life in the water. To aid swimming, the body has become streamlined and the number of body projections has been reduced. The ears have shrunk to small holes in size and shape. Mammary glands and sex organs are not part of the external physiology, and posterior (hind) limbs are no longer present.

Mechanisms to prevent heat loss have also been developed. The cylindrical body shape with small appendages reduces the surface area to volume ratio of the body, which reduces heat loss. Marine mammals also have a counter current heat exchange mechanism created by convergent evolution external link where the heat from the arteries is transferred to the veins as they pass each other before getting to extremities, thus reducing heat loss. Some marine mammals also have a thick layer of fur with a water repellent undercoat and/or a thick layer of blubber that can’t be compressed. The blubber provides insulation, a food reserve, and aids with buoyancy. These heat loss adaptations can also lead to overheating for animals that spend time out of the water. To prevent overheating, seals or sea lions will swim close to the surface with their front flippers waving in the air. They also flick sand onto themselves to keep the sun from directly hitting their skin. Blood vessels can also be expanded to act as a sort of radiator.

One of the major behavioral adaptations of marine mammals is their ability to swim and dive. Pinnipeds swim by paddling their flippers while sirenians and cetaceans move their tails or flukes up and down.

Some marine mammals can swim at relatively high speeds. Sea lions swim up to 35 kph and orcas can reach 50 kph. The fastest marine mammal, however, is the common dolphin, which reaches speeds up to 64 kph. While swimming, these animals take very quick breaths. For example, fin whales can empty and refill their huge lungs in less than 2 seconds. Marine mammals’ larynx and esophagus close automatically when they open their mouths to catch prey during dives. Oxygen is stored in hemoglobin in the blood and in myoglobin in the muscles. The lungs are also collapsible so that air is pushed into the windpipe preventing excess nitrogen from being absorbed into the tissues. Decreasing pressure can cause excess nitrogen to expand in the tissues as animals ascend to shallower depths, which can lead to decompression sickness,  aka “the bends.” Bradycardia, the reduction of heart rate by 10 to 20%, also takes place to aid with slowing respiration during dives and the blood flow to non-essential body parts. These adaptations allow sea otters to stay submerged for 4 to 5 minutes and dive to depths up to 55 m. Pinnipeds can often stay down for 30 minutes and reach average depths of 150-250 m. One marine mammal with exceptional diving skills is the Weddell seal, which can stay submerged for at least 73 minutes at a time at depths up to 600 m. The length and depth of whale dives depends on the species. Baleen whales feed on plankton near the surface of the water and have no need to dive deeply so they are rarely seen diving deeper than 100 m external link. Toothed whales seek larger prey at deeper depths and some can stay down for hours at depths of up to 2,250 m external link.

Human Experience

Albert Behnke: Nitrogen Narcosis

Casey A. Grover and David H. Grover
The Journal of Emergency Medicine, 2014; 46(2):225–227

As early as 1826, divers diving to great depths noted that descent often resulted in a phenomenon of intoxication and euphoria. In 1935, Albert Behnke discovered nitrogen as the cause of this clinical syndrome, a condition now known as nitrogen narcosis. Nitrogen narcosis consists of the development of euphoria, a false sense of security, and impaired judgment upon underwater descent using compressed air below 34 atmospheres (99 to 132 feet). At greater depths, symptoms can progress to loss of consciousness. The syndrome remains relatively unchanged in modern diving when compressed air is used. Behnke’s use of non-nitrogencontaining gas mixtures subsequent to his discovery during the 1939 rescue of the wrecked submarine USS Squalus pioneered the use of non-nitrogencontaining gas mixtures, which are used by modern divers when working at great depth to avoid the effects of nitrogen narcosis.

Behnke’s first duty station as a licensed physician was as assistant medical officer for Submarine Division 20 in San Diego, which was then commanded by one of the Navy’s rising stars, Captain Chester W. Nimitz of World War II fame.
In this setting, Dr. Behnke spent his free time constructively by learning to dive, using the traditional ‘‘hard-hat’’ gear aboard the USS Ortalon, a submarine rescue vessel to which he also rotated. Diving was not a notable specialty of the Navy at the time, and the service was slow in developing the infrastructure for it. Dr. Behnke devoted his efforts to research on the topic of diving medicine, as well as developing a more sound understanding of the biophysics of diving. In 1932, he wrote a letter to the Surgeon General describing some of his observations on arterial gas embolism, which earned him some accolades from the Navy and resulted in his transfer to Harvard’s School of Public Health as a graduate fellow. After 2 years at Harvard, the Navy assigned duty to Dr. Behnke at the Navy’s submarine escape training tower at Pearl Harbor. He worked extensively here on developing techniques for rescuing personnel from disabled submarines on the sea floor. In 1937, he was one of three Navy physicians assigned to the Navy’s Experimental Diving Unit. This team worked on improving the rescue system, plus updating the diving recompression tables originally developed by the British in 1908.

The intoxicating effects of diving were first described by a French physician named Colladon in 1826, who reported that descent in a diving bell resulted in his feeling a ‘‘state of excitement as though I had drunk some alcoholic liquor’’.
The etiology of this phenomenon remained largely unknown until the 1930s, when the British military researcher Damant again highlighted the issue, and reported very unpredictable behavior in his divers during descents as deep as 320 feet during the British Admiralty Deep Sea diving trials. Two initial theories arose as to the etiology for this effect, the first being from psychological causes by Hill and Phillip in 1932, and the second being from oxygen toxicity by Haldane in 1935.

Dr. Behnke and his colleagues at the Harvard School of Public Health had another idea as to the etiology of this phenomenon. In 1935, based on observation of individuals in experiments with a pressure chamber, Dr. Behnke published an article in the American Journal of Physiology in which he posited that nitrogen was the etiology of the intoxicating effects of diving.

Nitrogen narcosis, described as ‘‘rapture of the deep’’ by Jacques Cousteau, still remains a relatively common occurrence in modern diving, despite major advances in diving technology since Behnke’s initial description of the pathophysiologic cause of the condition in 1935. The development of symptoms of this condition varies from diver to diver, but usually begins when a depth of 4 atmospheres (132 feet) is reached in divers using compressed air. More sensitive divers can develop symptoms at only 3 atmospheres (99 feet), and other divers may not be affected up to depths as high as 6 atmospheres (198 feet). Interestingly, tolerance to nitrogen narcosis can be developed by frequent diving and exposure to the effects of compressed air at depth.

  1. Acott C. A brief history of diving and decompression illness. SPUMS J 1999;29:98–109.
    2. Bornmann R. Dr. Behnke, founder of UHMS, dies. Pressure 1992; 21:14.
    3. Behnke AR, Thomson RM, Motley P. The psychologic effects from breathing air at 4 atmospheric pressures. Am J Physiol 1935; 112:554–8.
    4. Behnke AR, Johnson FS, Poppen JR, Motley P. The effect of oxygen on man at pressures from 1 to 4 atmospheres. AmJ Physiol 1934; 110:565–72.

Exhaled nitric oxide concentration and decompression-induced bubble formation: An index of decompression severity in humans?

J.-M. Pontier, Buzzacott, J. Nastorg, A.T. Dinh-Xuan, K. Lambrechts
Nitric Oxide 39 (2014) 29–34

Introduction: Previous studies have highlighted a decreased exhaled nitric oxide concentration (FE NO) in divers after hyperbaric exposure in a dry chamber or following a wet dive. The underlying mechanisms of this decrease remain however unknown. The aim of this study was to quantify the separate effects of submersion, hyperbaric hyperoxia exposure and decompression-induced bubble formation on FE NO after a wet dive.
Methods: Healthy experienced divers (n = 31) were assigned to either

  • a group making a scuba-air dive (Air dive),
  • a group with a shallow oxygen dive protocol (Oxygen dive) or

a group making a deep dive breathing a trimix gas mixture (deep-dive).
Bubble signals were graded with the KISS score. Before and after each dive FE NO values were measured using a hand-held electrochemical analyzer.
Results: There was no change in post-dive values of FE NO values (expressed in ppb = parts per billion) in the Air dive group (15.1 ± 3.6 ppb vs. 14.3 ± 4.7 ppb, n = 9, p = 0.32). There was a significant decrease in post-dive values of FE NO in the Oxygen dive group (15.6 ± 6 ppb vs. 11.7 ± 4.7 ppb, n = 9, p = 0.009). There was an even more pronounced decrease in the deep dive group (16.4 ± 6.6 ppb vs. 9.4 ± 3.5 ppb, n = 13, p < 0.001) and a significant correlation between KISS bubble score >0 (n = 13) and percentage decrease in post-dive FE NO values (r = -0.53, p = 0.03). Discussion: Submersion and hyperbaric hyperoxia exposure cannot account entirely for these results suggesting the possibility that, in combination, one effect magnifies the other. A main finding of the present study is a significant relationship between reduction in exhaled NO concentration and dive-induced bubble formation. We postulate that exhaled NO concentration could be a useful index of decompression severity in healthy human divers.

Brain Damage in Commercial Breath-Hold Divers

Kiyotaka Kohshi, H Tamaki, F Lemaıtre, T Okudera, T Ishitake, PJ Denoble
PLoS ONE 9(8): e105006

Background: Acute decompression illness (DCI) involving the brain (Cerebral DCI) is one of the most serious forms of diving related injuries which may leave residual brain damage. Cerebral DCI occurs in compressed air and in breath-hold divers, likewise. We conducted this study to investigate whether long-term breath-hold divers who may be exposed to repeated symptomatic and asymptomatic brain injuries, show brain damage on magnetic resonance imaging (MRI).
Subjects and Methods: Our study subjects were 12 commercial breath-hold divers (Ama) with long histories of diving work in a district of Japan. We obtained information on their diving practices and the presence or absence of medical problems, especially DCI events. All participants were examined with MRI to determine the prevalence of brain lesions.
Results: Out of 12 Ama divers (mean age: 54.965.1 years), four had histories of cerebral DCI events, and 11 divers demonstrated ischemic lesions of the brain on MRI studies. The lesions were situated in the cortical and/or subcortical area (9 cases), white matters (4 cases), the basal ganglia (4 cases), and the thalamus (1 case). Subdural fluid collections were seen in 2 cases. Conclusion: These results suggest that commercial breath-hold divers are at a risk of clinical or subclinical brain injury which may affect the long-term neuropsychological health of divers.

Decompression illness

Richard D Vann, Frank K Butler, Simon J Mitchell, Richard E Moon
Lancet 2010; 377: 153–64

Decompression illness is caused by intravascular or extravascular bubbles that are formed as a result of reduction in environmental pressure (decompression). The term covers both arterial gas embolism, in which alveolar gas or venous gas emboli (via cardiac shunts or via pulmonary vessels) are introduced into the arterial circulation, and decompression sickness, which is caused by in-situ bubble formation from dissolved inert gas. Both syndromes can occur in divers, compressed air workers, aviators, and astronauts, but arterial gas embolism also arises from iatrogenic causes unrelated to decompression. Risk of decompression illness is
affected by immersion, exercise, and heat or cold. Manifestations range from itching and minor pain to neurological symptoms, cardiac collapse, and death. First aid treatment is 100% oxygen and definitive treatment is recompression to increased pressure, breathing 100% oxygen. Adjunctive treatment, including fluid administration and prophylaxis against venous thromboembolism in paralyzed patients, is also recommended. Treatment is, in most cases, effective although residual deficits can remain in serious cases, even after several recompressions.

Bubbles can have mechanical, embolic, and biochemical effects with manifestations ranging from trivial to fatal. Clinical manifestations can be caused by direct effects from extravascular (autochthonous) bubbles such as mechanical distortion of tissues causing pain, or vascular obstruction causing stroke-like signs and symptoms. Secondary effects can cause delayed symptom onset up to 24 h after surfacing. Endothelial damage by intravascular bubbles can cause capillary leak, extravasation of plasma, and haemoconcentration. Impaired endothelial function, as measured by decreased effects of vasoactive compounds, has been reported in animals and might occur in man. Hypotension can occur in severe cases. Other effects include platelet activation and deposition, leucocyte-endothelial adhesion, and possibly consequences of vascular occlusion believed to occur in thromboembolic stroke such as ischaemia-reperfusion injury, and apoptosis.

Classification of initial and of all eventual manifestations of decompression illness in 2346 recreational diving accidents reported to the Divers Alert Network from 1998 to 2004 For all instances of pain, 58% consisted of joint pain, 35% muscle pain, and 7% girdle pain. Girdle pain often portends spinal cord involvement. Constitutional symptoms included headache, lightheadedness, inappropriate fatigue, malaise, nausea or vomiting, and anorexia. Muscular discomfort included stiffness, pressure, cramps, and spasm but excluded pain. Pulmonary manifestations included dyspnoea and cough.

Other than depth and time, risk of decompression sickness is affected by other factors that affect inert gas exchange and bubble formation, such as immersion (vs dry hyperbaric chamber exposure), exercise, and temperature. Immersion decreases venous pooling and increases venous return and cardiac output. Warm environments improve peripheral perfusion by promoting vasodilation, whereas cool temperatures decrease perfusion through vasoconstriction. Exercise increases both peripheral perfusion and temperature. The effect of environmental conditions on risk of decompression sickness is dependent on the phase of the pressure exposure. Pressure, exercise, immersion, or a hot environment increase inert gas uptake and risk of decompression sickness. During decom-pression these factors increase inert gas elimination and therefore decrease the risk of decompression sickness. Conversely, uptake is reduced during rest or in a cold environment, hence a diver resting in a cold environment on the bottom has decreased risk of decompression sickness. Rest or low temperatures during decompression increase the risk. If exercise occurs after decompression when super-saturation is present, bubble formation increases and risk of decompression sickness rises.

Exercise at specific times before a dive can decrease the risk of serious decompression sickness in animals and incidence of venous gas emboli in both animals and man. The mechanisms of these effects are unknown but might involve modulation of nitric oxide production and effects on endothelium. Venous gas emboli and risk of decompression sickness increase slightly with age and body-mass index.

Arterial gas embolism should be suspected if a diver has a new onset of altered consciousness, confusion, focal cortical signs, or seizure during ascent or within a few minutes after surfacing from a compressed gas dive.

If the diver spends much time at depth and might have absorbed substantial inert gas before surfacing, arterial gas embolism and serious decompression sickness can coexist, and in such cases, spinal cord manifestations can predominate. Other organ systems, such as the heart, can also be affected, but the clinical diagnosis of gas embolism is not reliable without CNS manifestations. Arterial gas embolism is rare in altitude exposure; if cerebral symptoms occur after altitude exposure, the cause is usually decompression sickness.

Nondermatomal hypoaesthesia and truncal ataxia are common in neurological decompression sickness and can be missed by cursory examination. Pertinent information includes level of consciousness and mental status, cranial nerve function, and motor strength. Coordination can be affected disproportionately, and abnormalities can be detected by assessment of finger-nose movement, and, with eyes open and closed, ability to stand and walk and do heel-toe walking backwards and forwards. Many of these simple tests can be done on the scene by untrained companions.

Panel: Differential diagnosis of decompression illness
Inner-ear barotrauma
Middle-ear or maxillary sinus overinfl ation
Contaminated diving gas and oxygen toxic effects
Musculoskeletal strains or trauma sustained before, during, or after diving
Seafood toxin ingestion (ciguatera, pufferfish, paralytic shellfish poisoning)
Immersion pulmonary edema
Water aspiration
Decompression chamber

Decompression chamber

Decompression chamber. fluidic or pneumatic ventilator is shown at the left. The infusion pump is contained within a plastic cover, in which 100% nitrogen is used to decrease the fi re risk in the event of an electrical problem. The monitor screen is outside the chamber and can be seen through the viewing port. Photo from Duke University Medical Center, with permission.

Long-term outcomes of 69 divers with spinal cord decompressionsickness, by manifestation
n %
No residual symptoms 34 49·3
Any residual symptom 35 50·7
Mild paraesthesias, weakness, or pain 14 20·3
Some impairment of daily activities 21 30·4
Difficulty walking 11 15·9
Impaired micturition 13 18·8
Impaired defecation 15 21·7
Impaired sexual function 15 21·7

Decompression illness occurs in a small population but is an international problem that few physicians are trained to recognise or manage. Although its manifestations are often mild, the potential for permanent injury exists in severe cases, especially if unrecognised or inadequately treated. Emergency medical personnel should be aware of manifestations of decompression illness in the setting of a patient with a history of recent diving or other exposure to substantial pressure change, and should contact an appropriate consultation service for advice.

Diving Medicine: Contemporary Topics and Their Controversies

Michael B. Strauss and Robert C. Borer, Jr
Am J Emerg Med 2001; 19:232-238

SCUBA diving is a popular recreational sport. Although serious injuries occur infrequently, when they do knowledge of diving medicine and/or where to obtain appropriate consultation is essential. The emergency physician is likely to be the first physician contact the injured diver has. We discuss 8 subjects
in diving medicine which are contemporary, yet may have controversies associated with them. From this information the physician dealing primarily with the injured diver will have a basis for understanding and managing, as
well as where to find additional help, for his/her patients’ diving injuries.

Over the past 10 years, new knowledge and equipment improvements have made diving safer and more enjoyable. Estimates of actively participating sports divers show a striking increase over this time interval while the number of SCUBA diving deaths annually has remained nearly level at approximately 100. A further indicator of recreational diving safety is that reflected in the nearly constant number of diving injuries (1000 per annum) over the most recent 5 reported years, or approximately 0.53 to 3.4 incidents/10,000 dives.

Divers Alert Network.
The Divers Alert Network (DAN) is a nonprofit organization directed and staffed by experts in the specialty of diving medicine.6 DAN provides immediate consultation for both divers and physicians in the diagnosis and initial management of diving injuries. This 24-hour service is available free world-wide through a dedicated emergency telephone line: 1-919-684-4326. The DAN staff will also identify the nearest appropriate recompression treatment facility and knowledgeable physicians for an expedient referral. General diving medical inquiries can be answered during normal weekday hours either through an information telephone line: 1-919-684-2948 or through an interactive web site

Use of 100% Oxygen for Initial, on the Scene, Management of Diving Accidents
The breathing of pure oxygen is crucial for the initial management of the diving related problems of arterial gas embolism (AGE), decompression sickness (DCS), pulmonary barotrauma (thoracic squeeze), aspiration pneumonitis, and hypoxic encephalopathy associated with near drowning. In 1985, Dick reported that in many cases the neurologic symptoms of AGE and DCS were resolved with the immediate breathing of pure oxygen on the surface. The breathing of pure oxygen reduces bubble size by increasing the differential pressure for the inert gas to diffuse out of the bubble and it also speeds the washout of inert gas from body tissues. The early elimination of the bubble prevents hypoxia and the interaction of the bubble with the blood vessel lining. This interaction leads to secondary problems of capillary leak, bleeding, inflammation, ischemia, and cell death. These secondary problems are the reasons not all DCS symptoms resolve with recompression chamber treatment. The immediate use of pure oxygen for the medical management of these diving problems is analogous to the use of cardiopulmonary resuscitation for the witnessed cardiac arrest; the sooner initiated the better the results.

Diving Education

Medical Fitness for Diving

Asthma has the potential risk for AGE. Neuman reviewed the subject of asthma and diving. He and his coauthors recommend that asthmatics who are asymptomatic, not on medications and have no exercised induced abnormality on pulmonary function studies be allowed to dive.

Conditions leading to loss of consciousness, such as insulin dependent diabetes and epilepsy, can result in drowning. Carefully controlled diving studies in diabetics, who are free from complications, are now defining the safe requirements for diving. Epilepsy remains as a disqualification except in individuals with a history of febrile seizures ending prior to 5 years of age.

Availability of Hyperbaric Oxygen Treatment Facilities

The availability of these chambers makes it possible for divers who become symptomatic after SCUBA diving to readily receive recompression treatment. This is important because the closer the initiation of recompression treatment to the onset of DCS (and AGE) signs and symptoms, the greater the likelihood of full recovery.

Improved Diving Equipment

Mixed and Rebreather Gas Diving
Mixed gas diving involves changing the breathing gas from air which has 20% oxygen to higher oxygen percentages (nitrox). As the amount of oxygen is increased in the gas mixture, the amount of the inert gas (nitrogen) is reduced. With oxygen enriched air there is less tissue deposition of inert gas per unit of time under water for any given depth. However, because of increased oxygen partial pressures, the seizure threshold for oxygen toxicity is lowered. For normal sports diving activities, oxygen toxicity with mixed gas diving is only a theoretical concern.

Decompression Illness is More Than Bubbles

When AGE occurs, DCS symptoms may be concurrent or appear during or after recompression treatment even though the decompression tables were not violated on the dive. When DCS occurs in this situation it appears resistant to recompression treatment (Neuman) perhaps because of the inflammatory reaction generated by the bubble-blood vessel interaction from the AGE. In cases of DCI where components of both DCS and AGE are suspected, the diver should be observed for a minimum of 24 hours after the recompression treatment is completed for the delayed onset of DCS.

No theory of DCS discounts the primary role of bubbles in this condition. However, new information suggests that there are precursors to bubble formation and post-bubbling events that occur as a consequence of the bubbles. As mentioned earlier, venous gas emboli are a common occurrence diving ascent and ordinarily are filtered out harmlessly by the lungs. Precursors to DCS include stasis, dehydration and too rapid of ascents. These conditions allow the ubiquitous VGE to enlarge, coalesce and occlude the venous side of the circulation. Massive venous bubbling to the lungs can cause pulmonary vessel obstruction described as the chokes. If right to left shunts occur in the heart, VGE can become AGE to the brain. If the arterial flow is slow enough and/or the gradients large enough, autochthonus (ie, spontaneous) bubbles can form in the arterial circulation and lead to any of the consequences of AGE. In such situations it could be difficult to determine whether the DCI event was from AGE or DCS even after careful analysis of the dive profile. Hollenbeck’s model for diving paraplegia includes the setting of venous stasis (Batson’s plexus of veins) in the spinal canal, bubble formation, bubble enlargement possibly from off gassing of the spinal cord, blood vessel occlusion, and venous side infarctions of the spinal cord.
Contemporary Management of DCS

Problem Intervention Effect
Bubble Recompression
with HBO
Reduce bubble size
1. Washout inert gas.
2. Change bubble composition by diffusion.
Stasis and dehydration Hydration: oral fluids if alert, IV fluids otherwise. Improve blood flow.
InflammationCell Ischemia ? Anti-inflammatory medicationsHBO Reduce interaction between bubble and blood vessel endothelium.
Improve oxygen availability to hypoxic tissues, reduce edema and also reduces the interaction between bubble and blood vessel endothelium.



We anticipate that in the future there will be further improvements for the safety and enjoyment of the recreational SCUBA diver. For example, the dive computer of the future will be able to individualize dive profiles for different personal medical parameters such as age, body composition and fitness level. Diver locators could quickly target a missing diver and save time and gas consumption as well as prevent serious diving mishaps. Drugs may be developed that would minimize the effect of bubbles interacting with body tissues and prevent DCS and AGE.

Extracorporeal membrane oxygenation therapy for pulmonary decompression illness

Yutaka Kondo, Masataka Fukami and Ichiro Kukita
Kondo et al. Critical Care 2014; 18:438

Pulmonary decompression illness is rarely observed in clinical settings, and most patients die prior to hospitalization. We administered ECMO therapy to rescue a patient, even though this therapy has rarely been reported with good outcome in patients with decompression illness. In addition, we had to select venovenous ECMO even with the patient showing right ventricular failure. A lot of physicians may select venoarterial ECMO if the patient shows right ventricular failure, but the important physiological mechanism of pulmonary decompression illness is massive air embolism in the pulmonary arteries, and the bubbles diminish within the first 24 hours. The management of decompression illness therefore differs substantially from the usual right-sided heart failure.

Extremes of barometric pressure

Jane E Risdall, David P Gradwell
Anaesthesia and Intensive Care Medicine 16:2
Ascent to elevated altitude, commonly achieved through flight, by climbing or by residence in highland regions, exposes the individual to reduced ambient pressure. Although there are physical manifestations of this exposure as a consequence of Boyle’s law, the primary physiological challenge is of hypobaric hypoxia. The acute physiological and longer-term adaptive responses of the cardiovascular, respiratory, hematological and neurological systems to altitude are described, together with an outline of the presentation and management of acute mountain sickness, high-altitude pulmonary edema and high-altitude cerebral edema. While many millions experience modest exposure to altitude as a result of flight in pressurized aircraft, fewer individuals are exposed to increased ambient pressure. The pressure changes during diving and hyperbaric exposures result in greater changes in gas load and gas toxicity. Physiological effects include the consequences of increased work of breathing and redistribution of circulating volume. Neurological manifestations may be the direct result of pressure or a consequence of gas toxicity at depth. Increased tissue gas loads may result in decompression illness on return to surface or subsequent ascent in flight.

  • understand the physical effects of changes in ambient pressure and the physiological consequences on the cardiovascular respiratory and neurological systems
  • gain an awareness that exposure to reduced ambient pressure produces both acute and more chronic effects, with differing signs, symptoms and time to onset at various altitudes
  • develop an awareness of the toxic effects of ‘inert’ gases at increased ambient pressures and the pathogenesis and management of decompression illness

Decompression illness According to Henry’s law, at a constant temperature the amount of gas which dissolves in a liquid is proportional to the pressure of that gas or its partial pressure, if it is part of a mixture of gases. Breathing gases at increased ambient pressure will increase the amount of each gas dissolved in the fluid phases of body tissues. On ascent this excess gas has to be given up. If the ascent is controlled at a sufficiently slow rate, elimination will be via the respiratory system. If the ascent is too fast, excess gas may come out of solution and form free bubbles in the tissues or circulation. Bubbles may contain any of the gases in the breathing mixture, but it is the presence of inert gas bubbles (nitrogen or helium) that are thought most likely to give rise to problems, since the elimination of excess oxygen is achieved by metabolism as well as ventilation. These bubbles may act as venous emboli or may trigger inflammatory tissue responses giving rise to symptoms of decompression illness (DCI). Signs and symptoms of DCI may appear up to 48 hours after exposure to increased ambient pressure and include joint pains, motor and sensory deficits, dyspnoea, cough and skin rashes.

Neurological effects of deep diving

Marit Grønning, Johan A. Aarli
Journal of the Neurological Sciences 304 (2011) 17–21

Deep diving is defined as diving to depths more than 50 m of seawater (msw), and is mainly used for occupational and military purposes. A deep dive is characterized by the compression phase, the bottom time and the decompression phase. Neurological and neurophysiologic effects are demonstrated in divers during the compression phase and the bottom time. Immediate and transient neurological effects after deep dives have been shown in some divers. However, the results from the epidemiological studies regarding long term neurological effects from deep diving are conflicting and still not conclusive.

Possible immediate neurological effects of deep diving
Syndrome Pressure
Hyperoxia/oxygen seizures >152 kPa (5 msw)
Nitrogen narcosis >354 kPa (25 msw)
High pressure nervous syndrome >1.6 MPa (150 msw)
Neurological decompression sickness

Neurological effects have been demonstrated, both clinically and neurophysiologically in divers during the compression phase and the bottom time. Studies of divers before and after deep dives have shown immediate and transient neurological effects in some divers. However, the results from the epidemiological and clinical studies regarding long term neurological effects from deep diving are conflicting and still not conclusive. Prospective clinical studies with sufficient power and sensitivity are needed to solve this important issue.

Today deep diving to more than 100 msw is routinely performed globally in the oil- and gas industry. In the North Sea remote underwater intervention and maintenance is performed by the use of remotely operated vehicles (ROV), both in conjunction to and as an alternative to manned underwater operations. There will, however, always be a need for human divers in the technically more advanced underwater operations and for contingency repair operations.

P300 latency indexes nitrogen narcosis

Barry Fowler, Janice Pogue and Gerry Porlier
Electroencephalography, and clinical Neurophysiology, 1990, 75:221-229

This experiment investigated the effects of nitrogen narcosis on reaction time (RT) and P300 latency and amplitude, Ten subjects breathed either air or a non-narcotic 20% oxygen-80% helium (heliox) mixture in a hyperbaric chamber at 6.5, 8.3 and 10 atmospheres absolute (ATA), The subjects responded under controlled accuracy conditions to visually presented male or female names in an oddball paradigm. Single-trial analysis revealed a strong relationship between RT and P300 latency, both of which were slowed in a dose-related manner by hyperbaric air but not by heliox. A clear-cut dose-response relationship could not be established for P300 amplitude. These results indicate that P300 latency indexes nitrogen narcosis and are interpreted as support for the slowed processing model of inert gas narcosis.

Adaptation to Deep Water Habitat

Effects of hypoxia on ionic regulation, glycogen utilization and antioxidative ability in the gills and liver of the aquatic air-breathing fish Trichogaster microlepis

Chun-Yen Huang, Hui-Chen Lina, Cheng-Huang Lin
Comparative Biochemistry and Physiology, Part A 179 (2015) 25–34

We examined the hypothesis that Trichogaster microlepis, a fish with an accessory air-breathing organ, uses a compensatory strategy involving changes in both behavior and protein levels to enhance its gas exchange ability. This compensatory strategy enables the gill ion-regulatory metabolism to maintain homeostasis during exposure to hypoxia. The present study aimed to determinewhether ionic regulation, glycogen utilization and antioxidant activity differ in terms of expression under hypoxic stresses; fish were sampled after being subjected to 3 or 12 h of hypoxia and 12 h of recovery under normoxia. The air-breathing behavior of the fish increased under hypoxia. No morphological modification of the gills was observed. The expression of carbonic anhydrase II did not vary among the treatments. The Na+/K+-ATPase enzyme activity did not decrease, but increases in Na+/K+-ATPase protein expression and ionocyte levels were observed. The glycogen utilization increased under hypoxia as measured by glycogen phosphorylase protein expression and blood glucose level, whereas the glycogen content decreased. The enzyme activity of several components of the antioxidant system in the gills, including catalase, glutathione peroxidase, and superoxidase dismutase, increased in enzyme activity. Based on the above data, we concluded that T. microlepis is a hypoxia-tolerant species that does not exhibit ion-regulatory suppression but uses glycogen to maintain energy utilization in the gills under hypoxic stress. Components of the antioxidant system showed increased expression under the applied experimental treatments.

Divergence date estimation and a comprehensive molecular tree of extant cetaceans

Michael R. McGowen , Michelle Spaulding, John Gatesy
Molecular Phylogenetics and Evolution 53 (2009) 891–906

Cetaceans are remarkable among mammals for their numerous adaptations to an entirely aquatic existence, yet many aspects of their phylogeny remain unresolved. Here we merged 37 new sequences from the nuclear genes RAG1 and PRM1 with most published molecular data for the group (45 nuclear loci, transposons, mitochondrial genomes), and generated a supermatrix consisting of 42,335 characters. The great majority of these data have never been combined. Model-based analyses of the supermatrix produced a solid, consistent phylogenetic hypothesis for 87 cetacean species. Bayesian analyses corroborated odontocete (toothed whale) monophyly, stabilized basal odontocete relationships, and completely resolved branching events within Mysticeti (baleen whales) as well as the problematic speciose clade Delphinidae (oceanic dolphins). Only limited conflicts relative to maximum likelihood results were recorded, and discrepancies found in parsimony trees were very weakly supported. We utilized the Bayesian supermatrix tree to estimate divergence dates among lineages using relaxed-clock methods. Divergence estimates revealed rapid branching of basal odontocete lineages near the Eocene–Oligocene boundary, the antiquity of river dolphin lineages, a Late Miocene radiation of balaenopteroid mysticetes, and a recent rapid radiation of Delphinidae beginning [1]10 million years ago. Our comprehensive,  time calibrated tree provides a powerful evolutionary tool for broad-scale comparative studies of Cetacea.

Mitogenomic analyses provide new insights into cetacean origin and evolution

Ulfur Arnason, Anette Gullberg, Axel Janke
Gene 333 (2004) 27–34

The evolution of the order Cetacea (whales, dolphins, porpoises) has, for a long time, attracted the attention of evolutionary biologists. Here we examine cetacean phylogenetic relationships on the basis of analyses of complete mitochondrial genomes that represent all extant cetacean families. The results suggest that the ancestors of recent cetaceans had an explosive evolutionary radiation 30–35 million years before present. During this period, extant cetaceans divided into the two primary groups, Mysticeti (baleen whales) and Odontoceti (toothed whales). Soon after this basal split, the Odontoceti diverged into the four extant lineages, sperm whales, beaked whales, Indian river dolphins and delphinoids (iniid river dolphins, narwhals/belugas, porpoises and true dolphins). The current data set has allowed test of two recent morphological hypotheses on cetacean origin. One of these hypotheses posits that Artiodactyla and Cetacea originated from the extinct group Mesonychia, and the other that Mesonychia/Cetacea constitutes a sister group to Artiodactyla. The current results are inconsistent with both these hypotheses. The findings suggest that the claimed morphological similarities between Mesonychia and Cetacea are the result of evolutionary convergence rather than common ancestry.

The order Cetacea traditionally includes three suborders: the extinct Archaeoceti and the recent Odontoceti and Mysticeti. It is commonly believed that the evolution of ancestral cetaceans from terrestrial to marine (aquatic) life was accompanied by a fast and radical morphological adaptation. Such a scenario may explain why it was, for a long time, difficult to morphologically establish the position of Cetacea in the mammalian tree and even to settle whether Cetacea constituted a monophyletic group.

Biochemical analyses in the 1950s  and 1960s had shown a closer relationship between cetaceans and artiodactyls (even-toed hoofed mammals) than between cetaceans and any other eutherian order and karyological studies in the late 1960s and early 1970s unequivocally supported cetacean monophyly (Arnason, 1969, 1974). The nature of the relationship between cetaceans and artiodactyls was resolved in phylogenetic studies of mitochondrial (mt) cytochrome b (cytb) genes (Irwin and Arnason, 1994; Arnason and Gullberg, 1996) that placed Cetacea within the order Artiodactyla itself as the sister group of the Hippopotamidae (see also Sarich, 1993). The Hippopotamidae/ Cetacea relationship was subsequently supported in studies of nuclear data (Gatesy et al., 1996; Gatesy, 1997) and statistically established in analysis of complete mt genomes (Ursing and Arnason, 1998). The relationship has also been confirmed in analyses of combined nuclear and mt sequences (Gatesy et al., 1999; Cassens et al., 2000) and in studies of short interspersed repetitive elements (SINEs). Artiodactyla and Cetacea are now commonly referred to as Cetartiodactyla.

Previous analyses of the complete cytb gene of more than 30 cetacean species (Arnason and Gullberg, 1996) identified five primary lineages of recent cetaceans, viz., Mysticeti and the four odontocete lineages Physeteridae (sperm whales), Platanistidae (Indian river dolphins), Ziphiidae (beaked whales) and Delphinoidea (iniid river dolphins, porpoises, narwhals and dolphins). However, these studies left unresolved the relationships of the five lineages as well as those between the three delphinoid families Monodontidae (narwhals, belugas), Phocoenidae (porpoises) and Delphinidae (dolphins). Similarly, the relationships between the four mysticete families Balaenidae (right whales), Neobalaenidae (pygmy right whales), Eschrichtiidae (gray whales) and Balaenopteridae (rorquals) were not conclusively resolved in analyses of cytb genes.

Fig. (not shown). Cetartiodactyl relationships and the estimated times of their divergences. The tree was established on the basis of maximum likelihood analysis of the concatenated amino acid (aa) sequences of 12 mt protein-coding genes. Length of alignment 3610 aa. Support values for branches A–H are shown in the insert.
Cetruminantia (branch A) receives moderate support and Cetancodonta (B) strong support. Cetacea (C) splits into monophyletic Mysticeti (baleen whales) and monophyletic Odontoceti (toothed whales). Odontoceti has four basal lineages, Physeteridae (sperm whales: represented by the sperm and pygmy sperm whales), Ziphiidae (beaked whales: bottlenose and Baird’s beaked whales), Platanistidae (Indian river dolphins: Indian river dolphin) and Delphinoidea. Delphinoidea encompasses the families Iniidae (iniid river dolphins: Amazon river dolphin, La Plata dolphin), Monodontidae (narwhals/belugas: narwhal), Phocoenidae (porpoises: harbour porpoise) and Delphinidae (dolphins: white-beaked dolphin). The common odontocete branch and the branches separating the four cetacean lineages are short. These relationships are therefore somewhat unstable (cf. Section 3.1 and Table 1). Iniid river dolphins (F) are solidly nested within the Delphinoidea (E). Thus, traditional river dolphins (Platanistidae + Iniidae) do not form a monophyletic unit. Molecular estimates of divergence times (Sanderson 2002) were based on two calibration points, A/C-60 and O/M-35 (cf. Section 3.2). Due to the short lengths of internal branches, some estimates for these divergences overlap. NJ: neighbor joining; MP: maximum parsimony; LBP: local bootstrap probability; QP: quartet puzzling. The bar shows the number of aa substitutions per site.

The limited molecular resolution among basal cetacean lineages has been known for some time. Studies of hemoglobin and myoglobin (Goodman, 1989; Czelusniak et al., 1990) have either joined Physeteridae and Mysticeti to the exclusion of Delphinoidea (myoglobin data) or Mysticeti and Delphinoidea to the exclusion of Physeteridae (hemoglobin data). Thus, neither of the data sets identified monophyletic Odontoceti by joining the two odontocete lineages (Physeteridae and Delphinoidea) to the exclusion of Mysticeti. A similar instability was recognized and cautioned against in analyses of some mt data, notably, sequences of rRNA genes (Arnason et al., 1993b). The suggestion (Milinkovitch et al., 1993) of a sister group relationship between Physeteridae and the mysticete family Balaenopteridae (rorquals) was based on a myoglobin data set (which joins Physeteridae and Mysticeti to the exclusion of Delphinoidea) that was complemented with partial data of the mt 16S rRNA gene.

The cetancodont divergence times calculated using A/C-60 and O/M-35 as references have been included in Fig. 1. As a result of the short branches separating several cetacean lineages, the estimates of these divergences overlap. The same observation has been made in calculations based on SINE flanking sequences (Nikaido et al., 2001). There is a general consistency between the current and the flanking sequence datings, except for those involving the Balaenopteridae, which are somewhat younger in our analysis than in the SINEs study. The currently estimated age of the divergence between Hippopotamus and Cetacea (c53.5 MYBP) is consistent with the age (>50 MY) of the oldest archaeocete fossils identified so far (Bajpai and Gingerich, 1998). This suggests that the ages allocated to the two references, A/C-60 (the divergence between ruminant artiodactyls and cetancodonts) and O/M-35 (the divergence between odontocetes and mysticetes) are reasonably accurate.

The dating of the divergence between the blue and fin whales is of interest regarding hybridization between closely related mammalian species. Previous molecular analyses (Arnason et al., 1991b; Spilliaert et al., 1991) demonstrated the occurrence of hybridization between these two species. These studies, which were based on three hybrids (one female and two males), showed that either species could be the mother or father in these hybridizations. The two male hybrids had rudimentary testes, whereas the female hybrid was in her second pregnancy. This suggests that the blue and fin whales may be close to the limit for permissible species hybridization among mammals.

The current data set has allowed examination of the coherence between the molecular results and two prevalent morphological hypotheses related to cetacean evolution. The first hypothesis, which in essence originates from Van Valen (1966, 1968), postulates that monophyletic Artiodactyla and monophyletic Cetacea evolved separately from the extinct Palaeocene group Mesonychia. This hypothesis was recently reinforced in a morphological study (Thewissen et al., 2001) that included mesonychians, two archaeocete taxa (Ambuloocetus and Pakicetus) and some extant and fossil artiodactyls. The study of Thewissen et al. (2001) showed a sister group relationship between monophyletic Artiodactyla and monophyletic Cetacea, with Mesonychia as the basal sister group of Artiodactyla/Cetacea, a conclusion consistent with the palaeontological age of Mesonychia relative to that of Artiodactyla and Cetacea. The second hypothesis favours a sister group relationship between Mesonychia and Cetacea with the Mesonychia/Cetacea clade as the sister group of monophyletic Artiodactyla (O’Leary and Geisler, 1999; see also Gatesy and O’Leary, 2001).

Although the position of Mesonychia differs in the two morphological hypotheses, both correspond to a sister group relationship between Cetacea and monophyletic Artiodactyla among extant cetartiodactyls. Thus, both hypotheses can be tested against the current data set. The result of such a test has been included in Table 1, topology (m)(not shown). As evident, both these morphological hypotheses are incongruent with the mitogenomic findings.

Morphological studies have not provided an answer to the question whether mysticetes and odontocetes had separate origins among the archaeocetes (Fordyce and de Muizon, 2001). However, the long common cetacean branch and the short branches separating the five extant cetacean lineages strongly suggest an origin of modern cetaceans from the same archaeocete group (probably the Dorudontidae).

The limbs of Ambulocetus constitute somewhat of an evolutionary enigma. As evident in Thewissen et al.’s (1994) paper, Ambulocetus has very large hind limbs compared to its forelimbs, a difference that is less pronounced in later silhouette drawings of the animal. It is nevertheless evident that evolution from the powerful hindlimbs of Ambulocetus to their rudimentation in archaeocetes constitutes a remarkable morphological reversal if Ambulocetus is connected to the cetacean branch after the separation of the hippopotamid and cetacean lineages.

For natural reasons, systematic schemes have traditionally been based on external morphological characteristics. The rates of morphological and molecular evolution are rarely (if ever) strictly correlated, however, and this may give rise to inconsistency between traditional systematics and molecular findings. The emerging consensus that the order Cetacea resides within another traditional order, Artiodactyla, makes apparent the incongruity in cetartiodactyl nomenclature (Graur and Higgins, 1994). In this instance, a possible solution for maintaining reasonable consistency between nomenclature and phylogeny would be to recognize Cetartiodactyla as an order with three suborders: Suina, Tylopoda and Cetruminantia. According to such a scheme, Cetacea would (together with the Hippopotamidae) constitute a parvorder within the infraorder Cetancodonta.

Cytochrome b and Bayesian inference of whale phylogeny

Laura May-Collado, Ingi Agnarsson
Molecular Phylogenetics and Evolution 38 (2006) 344–354

In the mid 1990s cytochrome b and other mitochondrial DNA data reinvigorated cetacean phylogenetics by proposing many novel

and provocative hypotheses of cetacean relationships. These results sparked a revision and reanalysis of morphological datasets, and the collection of new nuclear DNA data from numerous loci. Some of the most controversial mitochondrial hypotheses have now become benchmark clades, corroborated with nuclear DNA and morphological data; others have been resolved in favor of more traditional views. That major conflicts in cetacean phylogeny are disappearing is encouraging. However, most recent papers aim specifically to resolve higher-level conflicts by adding characters, at the cost of densely sampling taxa to resolve lower-level relationships. No molecular study to date has included more than 33 cetaceans. More detailed molecular phylogenies will provide better tools for evolutionary studies. Until more genes are available for a high number of taxa, can we rely on readily available single gene mitochondrial data? Here, we estimate the phylogeny of 66 cetacean taxa and 24 outgroups based on Cytb sequences. We judge the reliability of our phylogeny based on the recovery of several deep-level benchmark clades. A Bayesian phylogenetic analysis recovered all benchmark clades and for the Wrst time supported Odontoceti monophyly based exclusively on analysis of a single mitochondrial gene. The results recover the monophyly, with the exception of only one taxa within Cetacea, and the most recently proposed super- and subfamilies. In contrast, parsimony never recovered all benchmark clades and was sensitive to a priori weighting decisions. These results provide the most detailed phylogeny of Cetacea to date and highlight the utility of both Bayesian methodology in general, and of Cytb in cetacean phylogenetics. They furthermore suggest that dense taxon sampling, like dense character sampling, can overcome problems in phylogenetic reconstruction.

Some long standing debates are all but resolved: our understanding of deeper level cetacean phylogeny has grown strong. However, the strong focus of most recent studies, aiming specifically to resolve these higher level conflicts by adding mostly characters rather than taxa, has left our understanding of lower level relationships among whale species lagging behind. Mitogenomic data, for example, is available only for 16 cetacean species, and no molecular study to date has included more than 33 cetaceans. It seems timely to focus on more detailed (genus, and species level) molecular phylogenies. These will provide better tools for detailed evolutionary studies, and are necessary to test existing morphological phylogenetic hypotheses, and current cetacean classification.

We judge the reliability of our phylogeny based on the recovery of the previously mentioned benchmark clades, in addition to the less controversial clades Perissodactyla, Euungulata (sensu Waddell et al., 2001; Perissodactyla+ Cetartiodactyla), Cetacea, and Mysticeti. Because Cytb is thought to be most reliable at lower taxonomic levels (due to high substitution rates), recovering ‘known’ deeper clades gives credibility to these new findings which have not been addressed by studies using few taxa. We compare the performance of Bayesian analyses versus parsimony under four different models, and briefly examine the sensitivity of the results to taxon sampling. We use our results to discuss agreement and remaining conflict in cetacean phylogenetics, and provide comments on current classification.

The Bayesian analysis recovered all seven benchmark clades. Support for five of the benchmark clades is high (100 posterior probabilities) but rather low for Cetancodonta (79) and marginal for the monophyly of Odontoceti. The analysis also recovered all but one family level, and most sub- and superfamily level cetacean taxa. The results broadly corroborate current cetacean classiffcation, while also pointing to some lower-level groups that may need redefinition.

Many recent cetacean phylogenetic studies include relatively few taxa, in part due to a focus on generating more characters to resolve higher level phylogenetics. While addressing crucial questions and providing the backbone for lower level phylogenies, such studies have limited utility for classification, and for comparative evolutionary studies. In some cases sparse taxon sampling may also confound the results. Of course, taxon sampling is usually simply constrained by the availability of character data, but for some reason many studies have opted to include only one, or a few outgroup taxa, even if many are available.

We find that as long as outgroup taxon sampling was extensive, Bayesian analyses of Cytb recovered all the a priori identified benchmark clades. When only a few outgroups were chosen, however, the Bayesian analysis negated Odontoceti monophyly, as have many previous parsimony analyses of mitochondrial DNA. Furthermore, in almost every detailed comparison possible our results mirror the findings O’Leary et al. (2004), the most ‘character-complete’ (but including relatively few cetacean taxa) analysis to date (37,000 characters from morphology, SINE, and 51 gene fragments). This result gives credibility to our findings, including previously untested lower level clades.

  • Monophyly and placement of Mysticeti (baleen whales).
  • Monophyly of Odontoceti (toothed whales)
  • Delphinoids
  • River Dolphins
  • Beaked and sperm whales

A major goal of phylogenetics is a phylogeny of life (i.e., many taxa), based on multiple lines of evidence (many characters of many types). However, when phylogenies based on relatively few characters can be judged reliable based on external evidence (taxonomic congruence with other phylogenies using many characters, but few taxa), they seem like very promising and useful ‘first guess’ hypotheses. The evolution of sexual dimorphism, echolocation, social behavior, and whistles and other communicative signals, and major ecological shifts (e.g., transition to fresh water) are among the numerous interesting questions in cetacean biology that this phylogeny can help answer.

Deep-diving sea lions exhibit extreme bradycardia in long duration dives

Birgitte I. McDonald1, and Paul J. Ponganis
The Journal of Experimental Biology (2014) 217, 1525-1534

Heart rate and peripheral blood flow distribution are the primary determinants of the rate and pattern of oxygen store utilization and ultimately breath-hold duration in marine endotherms. Despite this, little is known about how otariids (sea lions and fur seals) regulate heart rate (fH) while diving. We investigated dive fH in five adult female California sea lions (Zalophus californianus) during foraging trips by instrumenting them with digital electrocardiogram (ECG) loggers and time depth recorders. In all dives, dive fH (number of beats/duration; 50±9 beats min−1) decreased compared with surface rates (113±5 beats min−1), with all dives exhibiting an instantaneous fH below resting (<54 beats min−1) at some point during the dive. Both dive fH and minimum instantaneous fH significantly decreased with increasing dive duration. Typical instantaneous fH profiles of deep dives (>100 m) consisted of:

(1) an initial rapid decline in fH resulting in the lowest instantaneous fH of the dive at the end of descent, often below 10 beats min−1 in dives longer than 6 min in duration;
(2) a slight increase in fH to ~10–40 beats min−1 during the bottom portion of the dive; and
(3) a gradual increase in fH during ascent with a rapid increase prior to surfacing.

Thus, fH regulation in deep-diving sea lions is not simply a progressive bradycardia. Extreme bradycardia and the presumed associated reductions in pulmonary and peripheral blood flow during late descent of deep dives should

(a) contribute to preservation of the lung oxygen store,
(b) increase dependence of muscle on the myoglobin-bound oxygen store,
(c) conserve the blood oxygen store and
(d) help limit the absorption of nitrogen at depth.

This fH profile during deep dives of sea lions may be characteristic of deep-diving marine endotherms that dive on inspiration as similar fH profiles have been recently documented in the emperor penguin, another deep diver that dives on inspiration.

The resting ƒH measured in this study (54±6 beats min−1) was lower than predicted for an animal of similar size (~80 beats min−1 for an 80 kg mammal). In part, this may be due to the fact that the sea lions were probably sleeping. The resting ƒH in our study was also lower than previous measurements in captive juvenile California sea lions (87±17 beats min−1, average mass 30 kg)  and wild Antarctic fur seals (78±5 beats min−1, body mass 30–50 kg). However, we found a significant negative relationship between mass and resting ƒH even with our small sample size of five sea lions (resting ƒH = –0.58 Mb +100.26, r2=0.81, F1,3=12.37, P=0.039). For a 30 kg sea lion, this equation predicts a resting ƒH of 83 beats min−1, which is similar to what was measured previously in juvenile sea lions, suggesting this equation may be useful in estimating resting ƒH in sea lions.

The sea lions exhibited a distinct sinus arrhythmia fluctuating between a minimum of 42±9 and a maximum of 87±12 beats min−1, comparable to the sinus arrhythmias described in other diving birds and mammals, including sea lions. The minimum instantaneous ƒH during the sinus arrhythmia was similar to the mean minimum ƒH in dives less than 3 min (37±7 beats min−1), indicating that in dives less than 3 min (estimated cADL), ƒH only decreased to levels observed during exhalation at rest. This is consistent with observations in emperor penguins and elephant seals, where it was proposed that in dives shorter than the aerobic dive limit (ADL) the reduction in ƒH is regulated by a mechanism of cardiorespiratory control similar to that governing the respiratory sinus arrhythmia, with a further reduction only occurring in dives longer than the ADL.

Fig. 3. (not shown) Instantaneous fH and dive depth profiles of a California sea lion (CSL12_2). Data are from (A) a short, shallow dive (1.3 min, 45 m), (B) a mid-duration dive (4.8 min, 239 m) and (C) a long-duration dive (8.5 min, 305 m). Minimum instantaneous fH reached 37 beats min−1 in the short dive
(A) 19 beats min−1 in the mid-duration dive
(B) and 7 beats min−1 in the long duration dive
(C) Prominent features typical of mid- and long-duration dives include

  • a surface interval tachycardia (pre- and post-dive);
  • a steady rapid decrease in fH during initial descent;
  • a gradual decline in fH towards the end of descent with the lowest fH of the dive at the end of descent;
  • a slight increase and sometimes variable fH during the bottom portion of the dive; and
  • a slow increase in fH during ascent,
  • often ending in a rapid increase just before surfacing.

We obtained the first diving ƒH data from wild sea lions on natural foraging trips, demonstrating how they regulate ƒH over a range of dive durations. Sea lions always decreased dive ƒH from surface ƒH values; however, individual sea lions exhibited different dive ƒH, accounting for a significant amount of the variation in the relationship between dive duration and ƒH (intra-individual correlation: 75–81%)). The individual differences in dive ƒH exhibited in this study suggest that different dive capacities of individual sea lions may partially account for the range of dive strategies exhibited in a previous study (Villegas-Amtmann et al., 2011). Despite the individual differences in ƒH, the pattern of the dive ƒH response was similar in all the sea lions. As predicted, sea lions only consistently displayed a true bradycardia on mid- to long- duration dives (>4 min) (Fig. 5A). Additionally, as seen in freely diving phocids, dive ƒH and minimum ƒH were negatively related to dive duration, with the longest duration dives having the lowest dive ƒH and displaying the most intense bradycardia, often below 10 beats min−1 (Fig. 5A,B).

Profiles of mean fH at 10 s intervals of dives

Profiles of mean fH at 10 s intervals of dives

Fig 4.  Profiles of mean fH at 10 s intervals of dives for (A) six duration categories and (B) five depth categories. Standard error bars are shown. Data were pooled from 461 dives performed by five sea lions. The number of dives in each category and the number of sea lions performing the dives in each category are provided in the keys.

The mild bradycardia and the dive ƒH profiles observed in the shorter duration dives (<3 min) were similar to those observed in trained juvenile California sea lions and adult Stellar sea lions, but much more intense than ƒH observed in freely diving Antarctic fur seals. Surprisingly, although dive ƒH of trained Steller sea lions was similar, Steller sea lions regularly exhibited lower minimum ƒH, with minimum ƒH almost always less than 20 beats min−1 in dives less than 2 min in duration. In the wild, California sea lions rarely exhibited a minimum ƒH less than 20 beats min−1 in similar duration dives (Fig. 5B), suggesting greater blood oxygen transport during these natural short-duration dives.

Fig. 5. (not shown)  fH decreases with increasing dive duration. Dive duration versus (A) dive fH (total number of beats/dive duration), (B) minimum instantaneous fH and (C) bottom fH (total beats at bottom of dive/bottom time) for California sea lions (461 dives from five sea lions).

Although California sea lions are not usually considered exceptional divers, they exhibited extreme bradycardia, comparable to that of the best diving phocids, during their deep dives. In dives greater than 6 min in duration, minimum ƒH was usually less than 10 beats min−1 and sometimes as low as 6 beats mins−1 (Fig. 5B), which is similar to extreme divers such as emperor penguins (3 beats min−1), elephant seals (3 beats min−1), grey seals (2 beats min−1) and Weddell seals (<10 beats min−1), and even as low as what was observed in forced submersion studies. Thus, similar to phocids, the extreme bradycardia exhibited during forced submersions is also a routine component of the sea lion’s physiological repertoire, allowing them to perform long-duration dives.

While the degree of bradycardia observed in long dives of California sea lions was similar to the extreme bradycardia observed in phocids, the ƒH profiles were quite different. In general, phocid ƒH decreases abruptly upon submergence. The intensity of the initial phocid bradycardia either remains relatively stable or intensifies as the dive progresses, and does not start to increase until the seal begins its ascent. In contrast, the ƒH profiles of sea lions were more complex, showing a more gradual decrease during descent, with the minimum ƒH of the dive usually towards the end of descent (Figs 3, 6). There was often a slight increase in ƒH during the bottom portion of the dive, and as soon as the sea lions started to ascend, the ƒH slowly started to increase, often becoming irregular during the middle of ascent, before increasing rapidly as the sea lion approached the surface.

Fig. 6. (not shown) Instantaneous fH and dive depth profiles of the longest dive (10.0 min, 385 m) from a California sea lion (CSL12_1). During this dive, instantaneous fH reached 7 beats min−1 and was less than 20 beats min−1 for over 5.5 min. Post-dive fH was high in the first 0.5–1 min after surfacing, but then declined to ~100 beats min−1 towards the end of the surface interval.

Implications for pulmonary gas exchange

The moderate dive ƒH in short, shallow dives compared with the much slower ƒH of deep long-duration dives suggests more pulmonary blood flow and greater potential for reliance on lung O2. Most of these dives were to depths of less than 100 m (well below the estimated depth of lung collapse near 200 m), so maintenance of a moderate ƒH during these dives may allow sea lions to maximise use of the potentially significant lung O2 stores (~16% of total body O2 stores) throughout the dive. This is supported by venous blood O2 profiles, where, occasionally, there was no decrease in venous blood O2 between the beginning and end of the dive; this can only occur if pulmonary gas exchange continues throughout the dive. Greater utilization of the lung O2 store in sea lions is consistent with higher dive ƒH in other species that both dive on inspiration and typically perform shallow dives (dolphins, porpoises, some penguin species), and in deeper diving species when they perform shallow dives (emperor penguins).

In deeper dives of sea lions, although ƒH was lower and bradycardia more extreme, the diving ƒH profiles suggest that pulmonary gas exchange is also important. In long-duration dives, even though ƒH started to decrease upon or shortly after submergence, the decrease was not as abrupt as in phocids. Additionally, in long deep dives, despite having overall low dive ƒH, there were more heart beats before resting ƒH was reached compared with short, shallow dives. In dives less than 3 min in duration, there were ~10–15 beats until instantaneous ƒH reached resting values. In longer duration dives (>3 min), there were usually ~30–40 beats before instantaneous ƒH reached resting values. We suggest the greater number of heart beats early in these deeper dives enables more gas exchange and blood O2 uptake at shallow depths, thus allowing utilisation of the postulated larger respiratory O2 stores in deeper dives The less abrupt decline in ƒH we observed in sea lions is similar to the more gradual declines documented in emperor penguins and porpoises, where it has also been proposed that the gradual decrease in ƒH allows them to maximise pulmonary gas exchange at shallower depths. However, as sea lions swam deeper, ƒH decreased further (Figs 3, 6), and by 200 m depth (the approximate depth of lung collapse, instantaneous ƒH was 14 beats min−1. Such an extreme decline in ƒH in conjunction with increased pulmonary shunting due to lung compression at greater depths will result in minimization of both O2 and N2 uptake by blood, even before the depth of full lung collapse (100% pulmonary shunt) is reached.

Implications for blood flow

ƒH is often used as a proxy to estimate blood flow and perfusion during diving because of the relative ease of its measurement. This is based on the assumption that stroke volume does not change during diving in sea lions, and, hence, changes in ƒH directly reflect changes in cardiac output. As breath-hold divers maintain arterial pressure while diving, changes in cardiac output should be associated with changes in peripheral vascular resistance and changes in blood flow to tissues. In Weddell seals, a decrease in cardiac output of ~85% during forced submersions resulted in an 80–100% decrease in tissue perfusion in all tissues excluding the brain, adrenal glands and lung. Sea lions exhibited extremely low instantaneous ƒH values that often remained low for significant portions of the dive (Figs 4, 6), suggesting severe decreases in tissue perfusion in dives greater than 5 min in duration. In almost all dives greater than 6 min in duration, instantaneous ƒH reached 10 beats min−1, and stayed below 20 beats min−1 for more than a minute. At a ƒH of 20 beats min−1, cardiac output will be ~36% of resting cardiac output and only about 18% of average surface cardiac output. At these levels of cardiac suppression, most of this flow should be directed towards the brain and heart.


We successfully obtained diving ƒH profiles from a deep-diving otariid during natural foraging trips. We found that

(1) ƒH decreases during all dives, but true and more intense bradycardia only occurred in longer duration dives and
(2) in the longest duration dives, ƒH and presumed cardiac output were as low as 20% of resting values.

We conclude that, although initial high ƒH promotes gas exchange early in deep dives, the extremely low ƒH in late descent of deep dives (a) preserves lung O2, (b) conserves blood O2, (c) increases the dependence of muscle on myoglobin-bound O2 and (d) limits N2 absorption at depth. This ƒH profile, especially during the late descent/early bottom phase of deep dives is similar to that of deep-diving emperor penguins, and may be characteristic of deep diving endotherms that dive on inspiration.

Dive duration was the fixed effect in all models, and to account for the lack of independence caused by having many dives from the same individual, individual (sea lion ID) was included as a random effect. Covariance and random effect structures of the full models were evaluated using Akaike’s information criterion (AIC) and examination of residual plots. AICs from all the tested models are presented with the best model in bold.

Additionally, dives were classified as short-duration (less than 3 min, minimum cADL), mid-duration (3–5 min, range of cADLs) or long-duration (>5 min) dives. Differences in pre-dive ƒH, dive ƒH, minimum ƒH, post-dive ƒH, and heart beats to resting between the categories were investigated using mixed effects ANOVA, followed by post hoc Tukey tests. In all models, dive duration category was the fixed effect and individual (sea lion ID) was included as a random effect. Model fit was accessed by examination of the residuals. All means are expressed ±s.d. and results of the Tukey tests were considered significant at P<0.05. Statistical analysis was performed in R.

Investigating Annual Diving Behaviour by Hooded Seals (Cystophora cristata) within the Northwest Atlantic Ocean

Julie M. Andersen, Mette Skern-Mauritzen, Lars Boehme
PLoS ONE 8(11): e80438.

With the exception of relatively brief periods when they reproduce and molt, hooded seals, Cystophora cristata, spend most of the year in the open ocean where they undergo feeding migrations to either recover or prepare for the next fasting period. Valuable insights into habitat use and diving behavior during these periods have been obtained by attaching Satellite Relay Data Loggers (SRDLs) to 51 Northwest (NW) Atlantic hooded seals (33 females and 18 males) during icebound fasting periods (200422008). Using General Additive Models (GAMs) we describe habitat use in terms of First Passage Time (FPT) and analyze how bathymetry, seasonality and FPT influence the hooded seals’ diving behavior described by maximum dive depth, dive duration and surface duration. Adult NW Atlantic hooded seals exhibit a change in diving activity in areas where they spend .20 h by increasing maximum dive depth, dive duration and surface duration, indicating a restricted search behavior. We found that male and female hooded seals are spatially segregated and that diving behavior varies between sexes in relation to habitat properties and seasonality. Migration periods are described by increased dive duration for both sexes with a peak in May, October and January. Males demonstrated an increase in dive depth and dive duration towards May (post-breeding/pre-molt) and August–October (post-molt/pre-breeding) but did not show any pronounced increase in surface duration. Females dived deepest and had the highest surface duration between December and January (post-molt/pre-breeding). Our results suggest that the smaller females may have a greater need to recover from dives than that of the larger males. Horizontal segregation could have evolved as a result of a resource partitioning strategy to avoid sexual competition or that the energy requirements of males and females are different due to different energy expenditure during fasting periods.

Novel locomotor muscle design in extreme deep-diving whales

P. Velten, R. M. Dillaman, S. T. Kinsey, W. A. McLellan and D. A. Pabst
The Journal of Experimental Biology 216, 1862-1871

Most marine mammals are hypothesized to routinely dive within their aerobic dive limit (ADL). Mammals that regularly perform deep, long-duration dives have locomotor muscles with elevated myoglobin concentrations that are composed of predominantly large, slow-twitch (Type I) fibers with low mitochondrial volume densities (Vmt). These features contribute to extending ADL by increasing oxygen stores and decreasing metabolic rate. Recent tagging studies, however, have challenged the view that two groups of extreme deep-diving cetaceans dive within their ADLs. Beaked whales (including Ziphius cavirostris and Mesoplodon densirostris) routinely perform the deepest and longest average dives of any air-breathing vertebrate, and short-finned pilot whales (Globicephala macrorhynchus) perform high-speed sprints at depth. We investigated the locomotor muscle morphology and estimated total body oxygen stores of several species within these two groups of cetaceans to determine whether they

(1) shared muscle design features with other deep divers and
(2) performed dives within their calculated ADLs.

Muscle of both cetaceans displayed high myoglobin concentrations and large fibers, as predicted, but novel fiber profiles for diving mammals. Beaked whales possessed a sprinterʼs fiber-type profile, composed of ~80% fast-twitch (Type II) fibers with low Vmt. Approximately one-third of the muscle fibers of short-finned pilot whales were slow-twitch, oxidative, glycolytic fibers, a rare fiber type for any mammal. The muscle morphology of beaked whales likely decreases the energetic cost of diving, while that of short-finned pilot whales supports high activity events. Calculated ADLs indicate that, at low metabolic rates, both beaked and short-finned pilot whales carry sufficient onboard oxygen to aerobically support their dives.

Serial cross-sections of the m. longissimus dorsi of Mesoplodon densirostris

Serial cross-sections of the m. longissimus dorsi of Mesoplodon densirostris

Fig. Serial cross-sections of the m. longissimus dorsi of Mesoplodon densirostris (A–D) and Globicephala macrorhynchus (E–H). Scale bars, 50μm. Muscle sections stained for the alkaline (A,E) and acidic (B,F) preincubations of myosin ATPase were used to distinguish Type I and II fibers. Muscle sections stained for succinate dehydrogenase (C,G) and α-glycerophosphate dehydrogenase (D,H) were used to distinguish glycolytic (gl), oxidative (o) and intermediate (i) fibers.

Previous studies of the locomotor muscles of deep-diving marine mammals have demonstrated that these species share a suite of adaptations that increase onboard oxygen stores while slowing the rate at which these stores are utilized, thus extending ADL. Their locomotor muscles display elevated myoglobin concentrations and are composed predominantly of large Type I fibers. Vmt are also lower in deep divers than in shallow divers or athletic terrestrial species. The results of this study indicate that beaked whales and short-finned pilot whales do not uniformly display these characteristics and that each possesses a novel fiber profile compared with those of other deep divers.

The phylogeny of Cetartiodactyla: The importance of dense taxon sampling, missing data, and the remarkable promise of cytochrome b to provide reliable species-level phylogenies

Ingi Agnarsson, Laura J. May-Collado
Molecular Phylogenetics and Evolution 48 (2008) 964–985

We perform Bayesian phylogenetic analyses on cytochrome b sequences from 264 of the 290 extant cetartiodactyl mammals (whales plus even-toed ungulates) and two recently extinct species, the ‘Mouse Goat’ and the ‘Irish Elk’. Previous primary analyses have included only a small portion of the species diversity within Cetartiodactyla, while a complete supertree analysis lacks resolution and branch lengths limiting its utility for comparative studies. The benefits of using a single-gene approach include rapid phylogenetic estimates for a large number of species. However, single-gene phylogenies often differ dramatically from studies involving multiple datasets suggesting that they often are unreliable. However, based on recovery of benchmark clades—clades supported in prior studies based on multiple independent datasets—and recovery of undisputed traditional taxonomic groups, Cytb performs extraordinarily well in resolving cetartiodactyl phylogeny when taxon sampling is dense. Missing data, however, (taxa with partial sequences) can compromise phylogenetic accuracy, suggesting a tradeoff between the benefits of adding taxa and introducing question marks. In the full data, a few species with a short sequences appear misplaced, however, sequence length alone seems a poor predictor of this phenomenon as other taxa.

The mammalian superorder Cetartiodactyla (whales and eventoed ungulates) contains nearly 300 species including many of immense commercial importance (cow, pig, and sheep) and of conservation interest and aesthetic value (antelopes, deer, giraffe, dolphins, and whales) (MacDonald, 2006). Certain members of this superorder count among the best studied organisms on earth, whether speaking morphologically, behaviorally, physiologically or genetically. Understanding the interrelationships among cetartiodactyl species, therefore, is of obvious importance with equally short sequences were not conspicuously misplaced. Although we recommend awaiting a better supported phylogeny based on more character data to reconsider classification and taxonomy within Cetartiodactyla, the new phylogenetic hypotheses provided here represent the currently best available tool for comparative species-level studies within this group. Cytb has been sequenced for a large percentage of mammals and appears to be a reliable phylogenetic marker as long as taxon sampling is dense. Therefore, an opportunity exists now to reconstruct detailed phylogenies of most of the major mammalian clades to rapidly provide much needed tools for species-level comparative studies.

Our results support the following relationship among the four major cetartiodactylan lineages (((Tylopoda ((Cetancodonta (Ruminantia + Suina))), with variable support. This arrangement has not been suggested previously, to our knowledge (see review in O’Leary and Gatesy, 2008 and discussion).

Relationships among clades within Cetancodonta are identical to those found by May-Collado and Agnarsson (2006).

Within Ruminantia all our analyzes suggest the following relationships among families: (((((Tragulidae((((Antilocapridae(((Giraffidae(( Cervidae(Moschidae + Bovidae))))) with relatively high support, supporting the subdivision of Ruminantia into Tragulina and Pecora.
In the rare cases where our results are inconsistent with benchmark clades, ad hoc explanations seem reasonable. The placement of M. meminna (Tragulidae) within Bovidae is likely an artifact of missing data, although remarkably it is the only conspicuous misplacement of a species across the whole phylogeny at the family level (while three species appear to be misplaced at the subfamily level within Cervidae in the full analysis, see Fig. 5a). This is supported by the fact that the placement of Moschiola receives low support, and the removal of Moschiola prior to analysis increases dramatically the support for clades close to where it nested (not shown, analysis available from authors), suggesting it had a tendency to ‘jump around’. Two other possibilities cannot be ruled out, however. One, that possibly the available sequence in Genbank may be mislabeled. And second, it should be kept in mind that the validity of Tragulidae has never been tested with molecular data including more than two species.

Oxygen and carbon dioxide fluctuations in burrows of subterranean blind mole rats indicate tolerance to hypoxic–hypercapnic stresses

Imad Shams, Aaron Avivi, Eviatar Nevo
Comparative Biochemistry and Physiology, Part A 142 (2005) 376 – 382

The composition of oxygen (O2), carbon dioxide (CO2), and soil humidity in the underground burrows from three species of the Israeli subterranean mole rat Spalax ehrenbergi superspecies were studied in their natural habitat. Two geographically close populations of each species from contrasting soil types were probed. Maximal CO2 levels (6.1%) and minimal O2 levels (7.2%) were recorded in northern Israel in the breeding mounds of S. carmeli in a flooded, poor drained field of heavy clay soil with very high volumetric water content. The patterns of gas fluctuations during the measurement period among the different Spalax species studied were similar. The more significant differentiation in gas levels was not among species, but between neighboring populations inhabiting heavy soils or light soils: O2 was lower and CO2 was higher in the heavy soils (clay and basaltic) compared to the relatively light soils (terra rossa and rendzina). The extreme values of gas concentration, which occurred during the rainy season, seemed to fluctuate with partial flooding of the tunnels, animal digging activity, and over-crowded breeding mounds inhabited by a nursing female and her offspring. The gas composition and soil water content in neighboring sites with different soil types indicated large differences in the levels of hypoxic–hypercapnic stress in different populations of the same species. A growing number of genes associated with hypoxic stress have been shown to exhibit structural and functional differences between the subterranean Spalax and the aboveground rat (Rattus norvegicus), probably reflecting the molecular adaptations that Spalax went through during 40 million years of evolution to survive efficiently in the severe fluctuations in gas composition in the underground habitat.

map of the studied sites

map of the studied sites

Schematic map of the studied sites: S. galili (2n =52): 1— Rehania (chalk); 2— Dalton (basaltic); S. golani (2n =54): 3— Majdal Shams (terra tossa); 4—Masa’ada (basaltic soils); S. carmeli (2n =58): 5— Al-Maker (heavy clay); 6— Muhraqa (terra rossa).

Comparison of gas composition (O2 and CO2) and water content between light and heavy soils inhabited by S. carmeli

Comparison of gas composition (O2 and CO2) and water content between light and heavy soils inhabited by S. carmeli

Comparison of gas composition (O2 and CO2) and water content between light and heavy soils inhabited by S. carmeli, Al-Maker (heavy soil) and Muhraqa (light soil). AverageTSD of measurements in the burrows of approximately 10 animals at a given date is presented. **p <0.01, T-test and Mann– Whitney test).

Subterranean mammals, which live in closed underground burrow systems, experience an atmosphere that is different from the atmosphere above-ground. Gas exchange between these two atmospheres depends on diffusion through the soil, which in turn, depends on soil particle size, water content, and burrow depth. Heavy soils (clay and basaltic), hold water and have little air space for gas diffusion. A large deviation from external gas composition is found in the burrows of Spalax living in these soil types. The maximal measured concentration of CO2 was 6.1% in Spalax breeding mounds, which is one of the highest concentrations among studied mammals in natural conditions. At the same time 7.2% O2 was measured in water saturated heavy clay soil

seasonal variation from August to March in mean O2, CO2, and soil water content

seasonal variation from August to March in mean O2, CO2, and soil water content

Example of seasonal variation from August to March in mean O2, CO2, and soil water content (VWC) in the Al-Maker population (2n =58, heavy soil). Values are presented as mean TSD.

In this study new data were presented for a wild mammal that survives in an extreme hypoxic–hypercapnic environment. Interestingly, the very low concentrations of O2 experienced by Spalax are correlated with the expression pattern of hypoxia related genes.  So far, we have shown higher and longer-term mRNA expression of erythropoietin, the main factor that regulates the level of circulating red blood cells, in subterranean Spalax compared to the above-ground rat in response to hypoxic stress, as well as differences in the response of erythropoietin to hypoxia in different populations of Spalax experiencing different hypoxic stress in nature. We also demonstrated that erythropoietin pattern of expression is different in Spalax than in Rattus throughout development, a pattern suggesting more efficient hypoxic tolerance in Spalax starting as early as in the embryonic stages. Furthermore, vascular endothelial growth factor (VEGF), which is a critical angiogenic factor that responds to hypoxia, is constitutively expressed at maximal levels in Spalax muscles, the most energy consuming tissue during digging. This level is 1.6-fold higher than in Rattus muscles and is correlated with significantly higher blood vessel concentration in the Spalax muscles compared to the Rattus muscles. Likewise, myoglobin the globin involved in oxygen homeostasis in skeletal muscles, exhibits different expression pattern under normoxia and in response to hypoxia in Spalax muscles compared to rat muscles as well as between different populations of Spalax exposed to different hypoxic stress in nature (unpublished results). Similarly, neuroglobin, a brain-specific globin involved in reversible oxygen binding, i.e., presumably in cellular homeostasis, is expressed differently in the Spalax brain compared to Rattus brain. Like erythropoietin and myoglobin also neuroglobin is expressed differently in Spalax populations experiencing different oxygen supply (unpublished results). Furthermore, Spalax p53 harbors two amino acid substitutions in its binding domain, which are identical to mutations found in p53 of human cancer cells. These substitutions endow Spalax p53 with several-fold higher activation of cell arrest and DNA repair genes compared to human p53 and favor activation of DNA repair genes over apoptotic genes. The study of specific tumoral variants indicates that such preference of growth arrest over apoptosis possibly results as a response to the hypoxic environmental stress known in tumors. Differences in the structure of other molecules related to homeostasis, namely, hemoglobin, haptoglobin (Nevo, 1999), and cytoglobin (unpublished) were also observed in Spalax.

Stress, adaptation, and speciation in the evolution of the blind mole rat, Spalax, in Israel

Eviatar Nevo
Molecular Phylogenetics and Evolution 66 (2013) 515–525

Environmental stress played a major role in the evolution of the blind mole rat superspecies Spalax ehrenbergi, affecting its adaptive evolution and ecological speciation underground. Spalax is safeguarded all of its life underground from aboveground climatic fluctuations and predators. However, it encounters multiple stresses in its underground burrows including darkness, energetics, hypoxia, hypercapnia, food scarcity, and pathogenicity. Consequently, it evolved adaptive genomic, proteomic, and phenomic complexes to cope with those stresses. Here I describe some of these adaptive complexes, and their theoretical and applied perspectives. Spalax mosaic molecular and organismal evolution involves reductions or regressions coupled with expansions or progressions caused by evolutionary tinkering and natural genetic engineering. Speciation of Spalax in Israel occurred in the Pleistocene, during the last 2.00–2.35 Mya, generating four species associated intimately with four climatic regimes with increasing aridity stress southwards and eastwards representing an ecological speciational adaptive trend: (Spalax golani, 2n = 54?S. galili, 2n = 52?S. carmeli, 2n = 58?S. judaei, 2n = 60). Darwinian ecological speciation occurred gradually with relatively little genetic change by Robertsonian chromosomal and genic mutations. Spalax genome sequencing has just been completed. It involves multiple adaptive complexes to life underground and is an evolutionary model to a few hundred underground mammals. It involves great promise in the future for medicine, space flight, and deep-sea diving.

Stress is a major driving force of evolution (Parsons, 2005; Nevo, 2011). Parsons defined stress as the ‘‘environmental factor causing potential injurious changes to biological systems with a potential for impacts on evolutionary processes’’. The global climatic transition from the middle Eocene to the early Oligocene (45–35 Ma = Million years ago) led to extensive convergent evolution underground of small subterranean mammals across the planet (Nevo, 1999; Lacey et al., 2000; Bennett and Faulkes, 2000; Begall et al., 2007). The subterranean ecotope provided small mammals with shelter from predators and extreme aboveground climatic stressful fluctuations of temperature and humidity. However, they had to evolve genomic adaptive complexes for the immense underground stresses of darkness, energy for burrowing in solid soil, low productivity and food scarcity, hypoxia, hypercapnia, and high infectivity. These stresses have been described in Nevo (1999, 2011) and Nevo et al. (2001); and Nevo list of Spalax publication at with many cited references relevant to these stresses).

blind subterranean mole rat of the Spalax ehrenbergi superspecies

blind subterranean mole rat of the Spalax ehrenbergi superspecies

The blind subterranean mole rat of the Spalax ehrenbergi superspecies in Israel. An extreme example of adaptation to life underground

Circadian rhythm and genes

adaptive circadian genes. We identified the circadian rhythm of Spalax
(Nevo et al., 1982) and described, cloned, sequenced, and expressed several circadian genes in Spalax. These include Clock, MOP3, three Period (Per), and cryptochromes (Avivi et al., 2001, 2002, 2003). The Spalax circadian genes are differentially conserved, yet characterized by a significant number of amino acid substitutions. The glutamine-rich area of Clock, which is assumed to function in circadian rhythmicity, is expanded in Spalax compared with that of mice and humans and is different in amino acid composition from that of rats. All three Per genes of Spalax oscillate with a periodicity of 24 h in the suprachaismatic nucleus, eye, and Harderian gland and are expressed in peripheral organs. Per genes are involved in clock resetting. Spalax Per 3 is unique in mammals though its function is still unresolved. The Spalax Per genes contribute to the unique adaptive circadian rhythm to life underground. The cryptochrome (Cry) genes, found in animals and plants, act both as photoreceptors and as ingredients of the negative feedback mechanism of the biological Clock. The CRY 1 protein is significantly closer to the human homolog than to that of mice, as was also shown in parts of the immunogenetic system. Both Cry 1 and Cry 2 mRNAs were found in the SCN, eye, harderian gland, and in peripheral tissues. Remarkably, the distinctly hypertrophied harderian gland is central in Spalax’s unique underground circadian rhythmicity (Pevet et al., 1984).

  • Spalax eye mosaic evolution
  • Gene expression in the eye of Spalax
  • Brain evolution in Spalax to underground stresses
  • Spalax: four species in Israel

The morphological, physiological, and behavioral Spalax eye patterns are underlain by gene expression representing regressive and progressive associated transcripts. Regressive transcripts involve B-2 microglobulin, transketolase, four keratins, alpha enolase, and different heat shock proteins. Several proteins may be involved in eye degeneration. These include heat shock protein 90alpha (hsp90alpha), found also in the blind fish Astyanax mexicanus, two transcripts of programmed cell death proteins, oculospanin, and peripherin 2, both belonging to the Tetraspanin family, in which 60 different mutations cause eye degeneration in humans. Several progressive transcripts in the Spalax eye are found in the retina of many mammals involving gluthatione, peroxidase 4, B spectrin, and Ankyrin; the last two characterize rod cells in the retina. Some transcripts are involved in metabolic processing of retinal, a vertebrate key component in phototransduction, and a relative of vitamin A.

cross section of the developing eye of the mole rat

cross section of the developing eye of the mole rat

Light micrographs showing cross section of the developing eye of the mole rat Spalax ehrenbergi. (A) Optic cup and lens vesicle initially develop normally (x100). (B) Eye at a later embryonic stage. Note appearance of iris-ciliary body rudiment (arrows), and development of the lens nucleus (L). ON, optic nerve (x100). (C) Eye at a still later fetal stage. Note massive growth of the iris-ciliary body complex colobomatous opening (arrow) (x100). (D) Early postnatal stage. The iris-ciliary body complex completely fills the chamber. The lens is vascularized and vacuolated (x100). (E) Adult eye. Eyelids are completely closed and pupil is absent. Note atrophic appearance of the optic disc region (arrow) (x65). (F) Higher magnification of the adult retina. The different retinal layers are retained: PE, pigment epithelium: RE, receptor layer; ON, outer nuclear layer: IN, inner nuclear layer; GC, ganglion cell layer (x500) (from Sanyal et al., 1990, Fig. 1).

The brains of subterranean mammals underwent dramatic evolution in accordance with underground stresses for digging and photoperiodic perception associated with vibrational, tactile, vocal, olfactory, and magnetic communication systems replacing sight, as is seen in Spalax. The brain of Spalax is twice as large as that of the laboratory rat of the same body size. The somatosensory region in the isocortex of Spalax is 1.7 times, the thalamic nuclei 1.3 times, and the motor cortex 3.1 times larger than in the sighted laboratory rat Rattus norvegicus matched to body size.

The ecological stress determinant in Spalax brain evolution is highlighted by the four species of the Spalax ehrenbergi superspecies in Israel. They differentiated chromosomally (by means of Robertsonian mutations and fission), allopatrically, and clinally southwards into four species associated with different climatic regimes, following the gradient of increasing aridity stress and decreasing predictability southwards towards the desert: Spalax galili (2n = 52) ->S. golani (2n = 54)->S. carmeli (2n = 58)->S. judaei (2n = 60), and eastwards S. galili ->S. golani (2n = 52–>54) (Fig. 2). This chromosomal speciation trend southwards is associated with the regional aridity stress southwards (and eastwards) in Israel, budding new species adapted genomically, proteomically, and phenomically (i.e., in morphology, physiology, and behavior) to increasing stresses of higher solar radiation, temperature, and drought southwards (Nevo, 1999; Nevo et al., 2001; Nevo
list of Spalax at A uniquely recent discovery of incipient sympatric ecological speciation at a microscale in Spalax triggered by local stresses occurs within Spalax galili.

retinal input to primary visual structures in Spalax

retinal input to primary visual structures in Spalax

Relative degree of retinal input to primary visual structures in Spalax, hamster, rat, and Spalacopus cyanus (South American Octodontidae, ‘‘coruro’’). These rodents are of similar body size (120–140 g). B. Relative degree of change in the proportions of retinal input to different primary visual structures in Spalax compared with measures obtained in other rodents. A relative progressive development in Spalax is seen in structures involved in photoperiodic and neuroendocrine functions (SCN, BNST).The main regressive feature is the drastic relative reduction of retinal input to the superior colliculus. The main regressive feature is the drastic reduction of retinal input to the superior colliculus. The relative size of other visual structures in Spalax is modified compared to that of the other species. c. Comparison of the absolute size (volume, mm3 x 10-4) of visual structures in Spalax and other rodents. The size of the SCN is equivalent in all species. The vLGN and dLGN are reduced by 87–93% in Spalax. The retino-recipient layers of the superior colliculus are reduced by 97%. Abbreviations: SCN: suprachiasmatic nucleus; BNST: bed nucleus of the stria terminalis; dLGN: dorsal lateral geniculate nucleus; SC: superior colliculus [From Cooper et al., 1993 (Fig 3)].

Subterranean life has a high energetic cost if an animal has to burrow in order to obtain its food. For a 150 g Thomomys bottae, burrowing 1 m may be 360–3400 times more expensive energetically than moving the same distance on the surface (Vleck, 1979). Mean rates of oxygen consumption during burrowing at 22 oC are from 2.8 to 7.1 times the RMR. Vleck developed a model examining the energetics of foraging by burrowing and found that, in the desert, Thomomys adjusts the burrow segment length to minimize the cost of burrowing. Since burrowing becomes less economic as body size increases, Vleck (1981) predicted that the maximum possible body size that a subterranean mammal can attain depends on a balance between habitat productivity and the cost of burrowing in local soils. Vleck’s cost of burrowing hypothesis has been verified in multiple cases. Heth (1989) demonstrated longer burrows in the rendzina soil and shorter ones in the terra rossa soil, associating lower productivity in the former for Spalax.

Food is a limiting factor for subterranean mammals. The abundance and distribution of food explain some of the ecological, physiological, and behavioral characteristics of subterranean mammals. In a field test of Spalax foraging strategy, we concluded that Spalax was a generalist due to the constraints of the subterranean ecotope. Restricted foraging time primarily during the winter when soil is wet, and the high energetic investment of tunneling to get to food items is significantly reduced than in summertime.
We also identified a decrease in the basic metabolic rate towards the desert, i.e., economizing energetics. The maintenance of adequate O2 transport in a subterranean mammal confronting hypoxia requires adaptation along the O2 transport system, achieved by increasing the flow of O2 in the convection systems (ventilation and perfusion) and by reduction of oxygen pressure (PO2) gradients at the diffusion barriers (lung blood, blood-tissue (Arieli, 1990). The PO2 gradient between blood capillaries and respiring mitochondria capillaries is large, and any adaptation at this level could be significant for O2 transport. Reduction of diffusion distance in a muscle can be achieved, like in Spalax, by increasing the number of capillaries that surround muscle fiber or by reducing fiber areas.

Geographic distribution in Israel of the four chromosomal species belonging to the S. ehrenbergi superspecies

Geographic distribution in Israel of the four chromosomal species belonging to the S. ehrenbergi superspecies

Geographic distribution in Israel of the four chromosomal species belonging to the S. ehrenbergi superspecies that are separated by narrow hybrid zones (2n = 52, 54, 58, and 60, now named as S. galili, S. golani, S. carmeli, and S. judaei, respectively; see Nevo et al., 2001).

Spalacid evolution, based on mtDNA, is driven by climatic oscillations and stresses. The underground ecotope provided subterranean mammals with shelter from extreme climate (temperature and humidity) fluctuations, and predators. However, they had to extensively and intensively adapt to the multiple underground stresses (darkness, energetic, low productivity and
food scarcity, hypoxia, hypercapnia, and high infectivity). All subterranean mammals, including spalacids as an extreme case, share convergent molecular and organismal adaptations to their shared unique underground ecotope. Evolution underground, as exemplified here in spalacids, led to mosaic molecular and organismal evolutionary syndromes to cope with multiple stresses.

Speciation involves all rates – from gradual to rapid. Subterranean mammals, with the spalacid example discussed above, provide uniquely rich evolutionary global tests of speciation and adaptation, convergence, regression, progression, and mosaic evolutionary processes. Adaptation and speciation underground was one of the most dramatic natural experiments verifying Darwinian evolution.

The Spalax genome sequencing has just been completed. It is being analyzed and will soon be published in 2012. This will be a milestone in understanding how numerous mammals across the globe, who found underground shelter from climatic fluctuations and stresses above ground, cope with the new suite of stresses they encountered underground, demanding a new engineering overhaul on all organizational levels, selecting for adaptive complexes to cope with the new underground stresses. The main current and future challenges are to compare and contrast genome sequences and identify the genomic basis of adaptation and speciation.

This global Cenozoic experiment could answer the following open questions: How heterozygous is the whole genome? How prevalent are retrotransposons and what is their functional role? How many genes are involved in the Spalax genome and how are they regulated? What are the genic and regulatory networks resisting the multiple stresses underground? How much of the Spalax genome is conserved and how much is reorganized to cope with the underground stresses? How is the solitary blind mole rat, Spalax, different from the social naked mole rat Heterocephalus? How are the processes of reduction, expansion, and genetic tinkering and engineering reflected across the genome? How effective is copy number variation in regulation? Is there similarity in the transcriptomes of subterranean mammals? How could we harness the rich genome repertoire of Spalax to revolutionize medicine, especially in the realm of hypoxia tolerance and the related major diseases of the western world, e.g., cancer, stroke, and cardiovascular diseases? What is the phylogenetic origin of Spalax? How much of the Spalax genome represents its phylogenetic roots and how much of coding and noncoding genomic regions are shared with other subterranean mammals across the globe in adapting to life underground?

The Atmospheric Environment of the Fossorial Mole Rat (Spalax Ehrenbergi): Effects of Season, Soil Texture, Rain, Temperature and Activity

  1. Arieli
    Comp Biochen Physiol. 1978; 63A:569-5151. The fossorial mole rat (Spalax ehrenbergi) may inhabit heavy soil with low gas permeability.
  2. Air composition in burrows in heavy soil deviates from atmospheric air more than that of burrows in light soil.
  3. In winter and spring O2 and CO2 concentrations in breeding mounds were 16.5% O2 and 2.5-3x CO2 and the extreme values measured were 14.0% O2 and 4.8% Cot.
  4. Hypoxia and hypercapnia in the burrow develop shortly after rain and when ambient temperature drops.
  5. Composition of the burrows air is influenced by the solubility of CO2 in soil water and by faster penetration of oxygen than outflowing of CO2.

Hypo-osmotic stress-induced physiological and ion-osmoregulatory responses in European sea bass (Dicentrarchus labrax) are modulated differentially by nutritional status

Amit Kumar Sinha, AF Dasan, R Rasoloniriana, N Pipralia, R Blust, G De Boeck
Comparative Biochemistry and Physiology, Part A 181 (2015) 87–99

We investigated the impact of nutritional status on the physiological, metabolic and ion-osmoregulatory performance of European sea bass (Dicentrarchus labrax)when acclimated to seawater (32 ppt), brackishwater (20 and 10 ppt) and hyposaline water (2.5 ppt) for 2 weeks. Following acclimation to different salinities, fish were either fed or fasted (unfed for 14 days). Plasma osmolality, [Na+], [Cl−] and muscle water contentwere severely altered in fasted fish acclimated to 10 and 2.5 ppt in comparison to normal seawater-acclimated fish, suggesting ion regulation and acid–base balance disturbances. In contrast to feed-deprived fish, fed fish were able to avoid osmotic perturbation more effectively. This was accompanied by an increase in Na+/K+-ATPase expression and activity, transitory activation of H+-ATPase (only at 2.5 ppt) and down-regulation of Na+/K+/2Cl− gene expression. Ammonia excretion rate was inhibited to a larger extent in fasted fish acclimated to low salinities while fed fish were able to excrete efficiently. Consequently, the build-up of ammonia in the plasma of fed fish was relatively lower. Energy stores, especially glycogen and lipid, dropped in the fasted fish at low salinities and progression towards the anaerobic metabolic pathway became evident by an increase in plasma lactate level. Overall, the results indicate no osmotic stress in both feeding treatments within the salinity range of 32 to 20 ppt. However, at lower salinities (10–2.5 ppt) feed deprivation tends to reduce physiological, metabolic, ion-osmo-regulatory and molecular compensatory mechanisms and thus limits the fish’s abilities to adapt to a hypo-osmotic environment.

The absence of ion-regulatory suppression in the gills of the aquatic air-breathing fish Trichogaster lalius during oxygen stress

Chun-Yen Huang, Hsueh-Hsi Lin, Cheng-Huang Lin, Hui-Chen Lin
Comparative Biochemistry and Physiology, Part A 179 (2015) 7–16

The strategy for most teleost to survive in hypoxic or anoxic conditions is to conserve energy expenditure, which can be achieved by suppressing energy-consuming activities such as ion regulation. However, an air-breathing fish can cope with hypoxic stress using a similar adjustment or by enhancing gas exchange ability, both behaviorally and physiologically. This study examined Trichogaster lalius, an air-breathing fish without apparent gill modification, for their gill ion-regulatory abilities and glycogen utilization under a hypoxic  treatment. We recorded air-breathing frequency, branchial morphology, and the expression of ion-regulatory proteins (Na+/K+-ATPase and vacuolar-type H+-ATPase) in the 1st and 4th gills and labyrinth organ (LO), and the expression of glycogen utilization (GP, glycogen phosphorylase protein expression and glycogen content) and other protein responses (catalase, CAT; carbonic anhydrase II, CAII; heat shock protein 70, HSP70; hypoxia-inducible factor-1α, HIF-1α; proliferating cell nuclear antigen, PCNA; superoxidase dismutase, SOD) in the gills of T. lalius after 3 days in hypoxic and restricted conditions. No morphological modification of the 1st and 4th gills was observed. The air breathing behavior of the fish and CAII protein expression both increased under hypoxia. Ion-regulatory abilities were not suppressed in the hypoxic or restricted groups, but glycogen utilization was enhanced within the groups. The expression of HIF-1α, HSP70 and PCNA did not vary among the treatments. Regarding the antioxidant system, decreased CAT enzyme activity was observed among the groups. In conclusion, during hypoxic stress, T. lalius did not significantly reduce energy consumption but enhanced gas exchange ability and glycogen expenditure.

The combined effect of hypoxia and nutritional status on metabolic and ionoregulatory responses of common carp (Cyprinus carpio)

Sofie Moyson, HJ Liew, M Diricx, AK Sinha, R Blusta, G De Boeck
Comparative Biochemistry and Physiology, Part A 179 (2015) 133–143

In the present study, the combined effects of hypoxia and nutritional status were examined in common carp (Cyprinus carpio), a relatively hypoxia tolerant cyprinid. Fish were either fed or fasted and were exposed to hypoxia (1.5–1.8mgO2 L−1) at or slightly above their critical oxygen concentration during 1, 3 or 7 days followed by a 7 day recovery period. Ventilation initially increased during hypoxia, but fasted fish had lower ventilation frequencies than fed fish. In fed fish, ventilation returned to control levels during hypoxia, while in fasted fish recovery only occurred after reoxygenation. Due to this, C. carpio managed, at least in part, to maintain aerobic metabolism during hypoxia: muscle and plasma lactate levels remained relatively stable although they tended to be higher in fed fish (despite higher ventilation rates). However, during recovery, compensatory responses differed greatly between both feeding regimes: plasma lactate in fed fish increased with a simultaneous breakdown of liver glycogen indicating increased energy use, while fasted fish seemed to economize energy and recycle decreasing plasma lactate levels into increasing liver glycogen levels. Protein was used under both feeding regimes during hypoxia and subsequent recovery: protein levels reduced mainly in liver for fed fish and in muscle for fasted fish. Overall, nutritional status had a greater impact on energy reserves than the lack of oxygen with a lower hepatosomatic index and lower glycogen stores in fasted fish. Fasted fish transiently increased Na+/K+-ATPase activity under hypoxia, but in general ionoregulatory balance proved to be only slightly disturbed, showing that sufficient energy was left for ion regulation.

The effect of temperature and body size on metabolic scope of activity in juvenile Atlantic cod Gadus morhua L.

Bjørn Tirsgaard, Jane W. Behrens, John F. Steffensen
Comparative Biochemistry and Physiology, Part A 179 (2015) 89–94

Changes in ambient temperature affect the physiology and metabolism and thus the distribution of fish. In this study we used intermittent flow respirometry to determine the effect of temperature (2, 5, 10, 15 and 20 °C) and wet body mass (BM) (~30–460 g) on standard metabolic rate (SMR, mg O2 h−1), maximum metabolic rate (MMR, mg O2 h−1) and metabolic scope (MS, mg O2 h−1) of juvenile Atlantic cod. SMR increased with BM irrespectively of temperature, resulting in an average scaling exponent of 0.87 (0.82–0.92). Q10 values were 1.8–2.1 at temperatures between 5 and 15 °C but higher (2.6–4.3) between 2 and 5 °C and lower (1.6–1.4) between 15 and 20 °C in 200 and 450 g cod. MMR increased with temperature in the smallest cod (50 g) but in the larger cod MMR plateaued between 10, 15 and 20 °C. This resulted in a negative correlation between the optimal temperature for MS (Topt) and BM, Topt being respectively 14.5, 11.8 and 10.9 °C in a 50, 200 and 450 g cod. Irrespective of BM cold water temperatures resulted in a reduction (30–35%) of MS whereas the reduction of MS at warm temperatures was only evident for larger fish (200 and 450 g), caused by plateauing of MMR at 10 °C and above. Warm temperatures thus seem favorable for smaller (50 g) juvenile cod, but not for larger conspecifics (200 and 450 g).

Read Full Post »

Introduction to Proteomics

Author and Curator: Larry H. Bernstein, MD, FCAP  


We have had a considerable extended discussion of preoteins and peptides, protein sinthesis, amino acid incorporation into protein, and metabolism of carbohydrates and lipids.  It is also clear that the historic practice of medicine, and the classification of biological systems has been highly dependent on the observations related to the observed phenotypical traits and disturbances of normal function that could be measured by traditional metabolic pathways for over a century.

What did we gain from the genomic revolution?

  1. Traceability of protein expression to a basic coded message
  2. The possibility of tracing disturbed cellular function to mutation related loss-of-function
  3. The ability to trace generational traits over long periods of time
  4. The promise of regenerating the enterprise of pharmacology and pharmaceutical intervention based on the silencing of or readjustment of regulated metabolic pathways to bring an adaptive rebalancing favoring extended life

What can we expect as we progress further as a result of the last two decades?

  1. There is a huge amount of information, as well as missing information that is necessary for adequately tackling the mastery of the life processes.
  2. There is a complex web of knowledge that goes beyond the genome and the one-gene one-enzyme, and the DNA-RNA-protein hypotheses that can only be realized by more full disclosure of the many metabolic control circuits involved in cellular homeostasis and adaptive control.
  3. The ability to come to disclosure and understanding of this cellular balancing will require the comprehensive exploration of the proteome and the active role of proteins and peptides in the functioning of all cells, and the organism.
  4. Proteomics will open up the discovery of new approaches to diagnostics and pharmaceutical discovery.

What about proteins?  What can proteins do? What can’t they do!

  • Enzymes are proteins that make sure that chemical reactions in your body take place up to a million times faster than they would without enzymes.
  • Antibodies are proteins that help your immune system to fight disease.
  • When you get an injury, the bleeding stops because of blood clots, thanks to the proteins fibrinogen and thrombin.
  • Transport! Some proteins carry vitamins ot hormones from one place to another, or form tunnels (pores) in cell membranes that will let only specific molecules (or ions) through. Hemoglobin, a protein in your blood, carries oxygen from your lungs to your cells.
  • Strength and support! Other proteins like collagen and keratin are strong and tough and make up your skin, hair, and fingernails. Collagen also supports your cells and organs so they don’t slosh around.
  • Motion! The proteins myosin and actin make up much of your muscle tissue. They work together so your muscles can move you around. Some bacteria have cilia and flagella made out of proteins. The bacteria can whip these around to move from place to place.

Proteins (/ˈprˌtnz/ or /ˈprti.ɨnz/) are large biological molecules, or macromolecules,

Proteins perform a vast array of functions within living organisms, including

  1. catalyzing metabolic reactions,
  2. replicating DNA,
  3. responding to stimuli, and
  4. transporting molecules from one location to another.

Proteins differ from one another primarily in

  1. their sequence of amino acids,
  2. which is dictated by the nucleotide sequence of their genes, and
  3. which usually results in folding of the protein into

A linear chain of amino acid residues is called a polypeptide. A protein contains at least one long polypeptide. Short polypeptides, containing less than about 20-30 residues, are rarely considered to be proteins and are commonly called peptides, or sometimes oligopeptides. The individual amino acid residues are bonded together by peptide bonds and adjacent amino acid residues. The sequence of amino acid residues in a protein is defined by

In general, the genetic code specifies 20 standard amino acids; however, in certain organisms the genetic code can include selenocysteine and—in certain archaeapyrrolysine. Shortly after or even during synthesis,

  • the residues in a protein are often chemically modified by posttranslational modification,
  • which alters the physical and chemical properties, folding, stability, activity, and ultimately, the function of the proteins.

Posttranslational modification (PTM) is a step in protein biosynthesis. Proteins created by ribosomes translating mRNA into polypeptide chains may undergo PTM (such as folding, cutting and other processes) before becoming the mature protein product.  After translation, the posttranslational modification of amino acids extends the range of functions of the protein by attaching it to other biochemical functional groups (such as acetate, phosphate, various lipids and carbohydrates), changing the chemical nature of an amino acid (e.g. citrullination), or making structural changes (e.g. formation of disulfide bridges).

Also, enzymes may remove amino acids from the amino end of the protein, or cut the peptide chain in the middle. For instance, the peptide hormone insulin is cut twice after disulfide bonds are formed, and a propeptide is removed from the middle of the chain; the resulting protein consists of two polypeptide chains connected by disulfide bonds. Also, most nascent polypeptides start with the amino acid methionine because the “start” n mRNA also codes for this amino acid. This amino acid is usually taken off during post-translational modification. Other modifications, like phosphorylation, are part of common mechanisms for controlling the behavior of a protein, for instance activating or inactivating an enzyme.

posttranslational modification of insulin

posttranslational modification of insulin

Posttranslational modification of insulin. At the top, the ribosome translates a mRNA sequence into a protein, insulin, and passes the protein through the endoplasmic reticulum, where it is cut, folded and held in shape by disulfide (-S-S-) bonds. Then the protein passes through the golgi apparatus, where it is packaged into a vesicle. In the vesicle, more parts are cut off, and it turns into mature insulin.

Genetic Code mapped

Genetic Code mapped

The genetic code diagram showing the amino acid residues as target of modification.

PTMs involving addition of cofactors for enhanced enzymatic activity

Sometimes proteins have non-peptide groups attached, which can be called prosthetic groups or cofactors.  Examples of cofactors include metal ions like iron and zinc. Proteins can also work together to achieve a particular function, and they often associate to form stable protein complexes.



Coenzymes are molecules that work at the active site of an enzyme and aid in recognizing, attracting, or repulsing a substrate or product. Many are derived from vitamins. The substrate is the molecule upon which an enzyme catalyzes a reaction transforming A to B by removal or addition of a hydrogen, or a hydroxyl group, or a methyl group, and so forth. This is  how an alcohol or an aldehyde is produced. Such a reaction is critical is carbohydrate metabolism for producing two 3-carbon sugars from a 6-carbon sugar. Coenzymes shuttle chemical groups from one enzyme to another enzyme. They may bind loosely to enzymes, while another group of cofactors do not.

Prosthetic groups are cofactors that bind tightly to proteins or enzymes. As if holding on for dear life, they are not easily removed. They can be organic or metal ions and are often attached to proteins by a covalent bond. The same cofactors can bind multiple different types of enzymes and may bind some enzymes loosely, as a coenzyme, and others tightly, as a prosthetic group. Some cofactors may always tightly bind their enzymes. It’s important to note, though, that these prosthetic groups can also bind to proteins other than enzymes.  A holoenzyme is an enzyme with any metal ions or coenzymes attached to it that is now ready to catalyze a reaction.



Around the world, millions of people don’t get enough protein. Protein malnutrition leads to the condition known as kwashiorkor. Lack of protein can cause growth failure, loss of muscle mass, decreased immunity, weakening of the heart and respiratory system, and death.

All Protein Isn’t Alike

Protein is built from building blocks called amino acids. Our bodies make amino acids in two different ways: Either from scratch, or by modifying others. A few amino acids (known as the essential amino acids) must come from food.

  • Animal sources of protein tend to deliver all the amino acids we need.
  • Other protein sources, such as fruits, vegetables, grains, nuts and seeds, lack one or more essential amino acids.

Vegetarians need to be aware of this. People who don’t eat meat, fish, poultry, eggs, or dairy products need to eat a variety of protein-containing foods each day in order to get all the amino acids needed to make new protein.
Molecular Biologists Guide to Proteomics

PR. Graves and TA.J. Haystead*
Microbiol Mol Biol Rev. Mar 2002; 66(1): 39–63  PMC120780

The emergence of proteomics, the large-scale analysis of proteins, has been inspired by the realization that

  • the final product of a gene is inherently more complex and
  • closer to function than the gene itself.

Shortfalls in the ability of bioinformatics to predict

  • both the existence and function of genes have also illustrated
  • the need for protein analysis.

Moreover, only through the study of proteins can posttranslational modifications be determined,

  • which can profoundly affect protein function.

Proteomics has been enabled by

  • the accumulation of both DNA and protein sequence databases,
  • improvements in mass spectrometry, and
  • the development of computer algorithms for database searching.

In this review, we describe why proteomics is important,

  • how it is conducted, and
  • how it can be applied to complement other existing technologies.

We conclude that currently, the most practical application of proteomics is

  • the analysis of target proteins as opposed to entire proteomes.

This type of proteomics, referred to as functional proteomics, is always

  • driven by a specific biological question.

In this way, protein identification and characterization has a meaningful outcome. We discuss some of the advantages

  • of a functional proteomics approach and

provide examples of how different methodologies can be utilized to address a wide variety of biological problems.

Entry of our laboratory into proteomics 5 years ago was driven by a need to define a complex mixture of proteins (∼36 proteins) we had affinity isolated that bound specifically to the catalytic subunit of protein phosphatase 1 (PP-1, a serine/threonine protein phosphatase that regulates multiple dephosphorylation events in cells). We were faced with the task of trying to understand the significance of these proteins, and the only obvious way to begin to do this was to identify them by sequencing. Since the majority of intact eukaryotic proteins are not immediately accessible to Edman sequencing

  • due to posttranslational N-terminal modifications,
  • we invented mixed-peptide sequencing.

This method enables internal peptide sequence information to be derived from proteins

  • electroblotted onto hydrophobic membranes.

Using the mixed-peptide sequencing strategy, we identified all 36 proteins in about a week. The mixture contained at least two known PP-1 regulatory subunits, but most were novel proteins of unknown function. Herein lies the lesson of proteomics. Identifying long lists of potentially interesting proteins often generates more questions than it seeks to answer.

Despite learning this obvious lesson, our early sequencing experiences were an epiphany that has subsequently altered our whole scientific strategy for probing protein function in cells. The sequencing of the 36 proteins has opened new avenues to further explore the functions of PP-1 in intact cells. Because of increased sensitivity, our approaches now routinely use state-of-the-art mass spectrometry (MS) techniques. However, rather than using proteomics to simply characterize large numbers of proteins in complex mixtures, we see the real application of this technology as a tool to enhance the power of existing approaches currently used by the modern molecular biologist such as classical yeast and mouse genetics, tissue culture, protein expression systems, and site-directed mutagenesis.

Importantly, the one message we would want the reader to take away from reading this review is that one should always let the biological question in mind drive the application of proteomics rather than simply engaging in an orgy of protein sequencing. From our experiences, we believe that if the appropriate controls are performed, proteomics is an extremely powerful approach for addressing important physiological questions. One should always design experiments to define a selected number of relevant proteins in the mixture of interest. Examples of such experiments that we routinely perform include defining early phosphorylation events in complex protein mixtures after hormone treatment of intact cells or comparing patterns of protein derived from a stimulated versus nonstimulated cell in an affinity pull-down experiment. Only the proteins that were specifically phosphorylated or bound in response to the stimulus are sequenced in the complex mixtures. Sequencing proteins that are regulated then has a meaningful outcome and directs all subsequent biological investigation.

The term “proteomics” was first coined in 1995 and was defined as the large-scale characterization of the entire protein complement of a cell line, tissue, or organism. Today, two definitions of proteomics are encountered. The first is the more classical definition, restricting the large-scale analysis of gene products to studies involving only proteins. The second and more inclusive definition combines protein studies with analyses that have a genetic readout such as mRNA analysis, genomics, and the yeast two-hybrid analysis. However, the goal of proteomics remains the same, i.e., to obtain a more global and integrated view of biology by studying all the proteins of a cell rather than each one individually.

Using the more inclusive definition of proteomics, many different areas of study are now grouped under the rubric of proteomics (Fig. (Fig.1).1). These include protein-protein interaction studies, protein modifications, protein function, and protein localization studies to name a few. The aim of proteomics is not only to identify all the proteins in a cell but also to create a complete three-dimensional (3-D) map of the cell indicating where proteins are located. These ambitious goals will certainly require the involvement of a large number of different disciplines such as molecular biology, biochemistry, and bioinformatics. It is likely that in bioinformatics alone, more powerful computers will have to be devised to organize the immense amount of information generated from these endeavors.

Types of proteomics and their applications to biology

Types of proteomics and their applications to biology

In the quest to characterize the proteome of a given cell or organism, it should be remembered that the proteome is dynamic. The proteome of a cell will reflect the immediate environment in which it is studied. In response to internal or external cues, proteins can be modified by posttranslational modifications, undergo translocations within the cell, or be synthesized or degraded. Thus, examination of the proteome of a cell is like taking a “snapshot” of the protein environment at any given time. Considering all the possibilities, it is likely that any given genome can potentially give rise to an infinite number of proteomes.

The first major technology to emerge for the identification of proteins was the sequencing of proteins by Edman degradation. A major breakthrough was the development of microsequencing techniques for electroblotted proteins. This technique was used for the identification of proteins from 2-D gels to create the first 2-D databases.  One of the most important developments in protein identification has been the development of MS technology. In the last decade, the sensitivity of analysis and accuracy of results for protein identification by MS have increased by several orders of magnitude. It is now estimated that proteins in the femtomolar range can be identified in gels. Because MS is more sensitive, can tolerate protein mixtures, and is amenable to high-throughput operations, it has essentially replaced Edman sequencing as the protein identification tool of choice.

The growth of proteomics is a direct result of advances made in large-scale nucleotide sequencing of expressed sequence tags and genomic DNA. Without this information, proteins could not be identified even with the improvements made in MS. Protein identification (by MS or Edman sequencing) relies on the presence of some form of database for the given organism. The majority of DNA and protein sequence information has accumulated within the last 5 to 10 years. In 1995, the first complete genome of an organism was sequenced, that of Haemophilus influenzae. At the time of this writing, the sequencing of the genomes of 45 microorganisms has been completed and that of 170 more is under way ( To date, five eukaryotic genomes have been completed: Arabidopsis thaliana, Saccharomyces cerevisiae, Schizosaccharomyces pombe, Caenorhabditis elegans, and Drosophila melanogaster. In addition, the rice, mouse, and human genomes are near completion.

One of the first applications of proteomics will be to identify the total number of genes in a given genome. This “functional annotation” of a genome is necessary because

  • it is still difficult to predict genes accurately from genomic data. One problem is that
  • the exon-intron structure of most genes cannot be accurately predicted by bioinformatics.

To achieve this goal, genomic information will have to be integrated with

  • data obtained from protein studies to confirm the existence of a particular gene.

The analysis of mRNA is

  • not a direct reflection of the protein content in the cell.

Many studies have shown a poor correlation

  • between mRNA and protein expression levels.

The formation of mRNA is only the first step in a long sequence of events resulting in the synthesis of a protein (Fig. (Fig.2).2).

  1. mRNA is subject to posttranscriptional control in the form of alternative splicing, polyadenylation, and mRNA editing. Many different protein isoforms can be generated from a single gene at this step.
  2. mRNA then can be subject to regulation at the level of protein translation. Proteins, having been formed, are subject to posttranslational modification. It is estimated that up to 200 different types of posttranslational protein modification exist. Proteins can also be regulated by proteolysis and compartmentalization. It is clear that the tenet of “one gene, one protein” is an oversimplification.
Mechanisms by which a single gene can give rise to multiple gene products

Mechanisms by which a single gene can give rise to multiple gene products

Mechanisms by which a single gene can give rise to multiple gene products. Multiple protein isoforms can be generated by RNA processing when RNA is alternatively spliced or edited to form mature mRNA. mRNA, in turn, can be regulated by stability and efficiency
One of the most important applications of proteomics will be the characterization of posttranslational protein modifications. Proteins are known to be modified posttranslationally in response to a variety of intracellular and extracellular signals. For example, protein phosphorylation is an important signaling mechanism and disregulation of protein kinases or phosphatases can result in oncogenesis. By using a proteomics approach, changes in the modifications of many proteins expressed by a cell can be analyzed simultaneously.
Of fundamental importance in biology is the understanding of protein-protein interactions. The process of cell growth, programmed cell death, and the decision to proceed through the cell cycle are all regulated by signal transduction through protein complexes. Proteomics aims to develop a complete 3-D map of all protein interactions in the cell. One step toward this goal was recently completed for the microorganism Helicobacter pylori. Using the yeast two-hybrid method to detect protein interactions, 1,200 connections were identified between H. pylori proteins covering 46.6% of the genome. A comprehensive two-hybrid analysis has also been performed on all the proteins from the yeast S. cerevisiae.
mixed peptide sequencing with MS

mixed peptide sequencing with MS

The process of mixed-peptide sequencing involves separation of a complex protein mixture by polyacrylamide gel electrophoresis (1-D or 2-D) and then transfer of the proteins to an inert membrane by electroblotting (Fig. (Fig.4).4). The proteins of interest are visualized on the membrane surface, excised, and fragmented chemically at methionine (by CNBr) or tryptophan (by skatole) into several large peptide fragments.
FASTF and FASTS search programs

FASTF and FASTS search programs

The mixed-sequence data are fed into the FASTF or TFASTF algorithms, which sort and match the data against protein (FASTF) and DNA (TFASTF) databases to unambiguously identify the protein. The FASTF and TFASTF programs were written in collaboration with William Pearson (Department of Biochemistry, University of Virginia). Because minimal sample handling is involved, mixed-peptide sequencing can be a sensitive approach for identifying proteins in polyacrylamide gels at the 0.1- to 1-pmol level.  A recent variation of T/FASTF has been devised for MS (101) (Fig. (Fig.5B).5B). The T/FASTF/S programs are available at (Table (Table11).

triple quadrupole MS

triple quadrupole MS

Triple-quadrupole mass spectrometers are most commonly used to obtain amino acid sequences. In the first stage of analysis, the machine is operated in MS scan mode and all ions above a certain m/z ratio are transmitted to the third quadrupole for mass analysis (Fig. (Fig.6)6) (82, 173). In the second stage, the mass spectrometer is operated in MS/MS mode and a particular peptide ion is selectively passed into the collision chamber. Inside the collision chamber, peptide ions are fragmented by interactions with an inert gas by a process known as collision-induced dissociation or collisionally activated dissociation. The peptide ion fragments are then resolved on the basis of their m/z ratio by the third quadrupole (Fig. (Fig.6).6). Since two different mass spectra are obtained in this analysis, it is referred to as tandem mass spectrometry (MS/MS). MS/MS is used to obtain the amino acid sequence of peptides by generating a series of peptides that differ in mass by a single amino acid.

The largest application of proteomics continues to be protein expression profiling. Through the use of two-dimensional gels or novel techniques such as ICAT, the expression levels of proteins or changes in their level of modification between two different samples can be compared and the proteins can be identified. This approach can facilitate the dissection of signaling mechanisms or identify disease-specific proteins.

Cancer cells are good candidates for proteomics studies because they can be compared to their non-transformed counterparts. Analysis of differentially expressed proteins in normal versus cancer cells can

(i) identify novel tumor cell biomarkers that can be used for diagnosis,

(ii) provide clues to mechanisms of cancer development, and

(iii) identify novel targets for therapeutic intervention. Protein expression profiling has been used in the study of breast, esophageal, bladder and prostate cancer. From these studies, tumor-specific proteins were identified and 2-D protein expression databases were generated. Many of these 2-D protein databases are now available on the World Wide Web.

Read Full Post »

Integrins, Cadherins, Signaling and the Cytoskeleton

Curator: Larry H. Bernstein, MD, FCAP 


We have reviewed the cytoskeleton, cytoskeleton pores and ionic translocation under lipids. We shall now look at this again, with specific attention to proteins, transporters and signaling.

Integrins and extracellular matrix in mechanotransduction

Lindsay Ramage
Queen’s Medical Research Institute, University of Edinburgh,

Edinburgh, UK
Cell Health and Cytoskeleton 2012; 4: 1–9

Integrins are a family of cell surface receptors which

  • mediate cell–matrix and cell–cell adhesions.

Among other functions they provide an important

  • mechanical link between the cells external and intracellular environments while
  • the adhesions that they form also have critical roles in cellular signal-transduction.

Cell–matrix contacts occur at zones in the cell surface where

  • adhesion receptors cluster and when activated
  • the receptors bind to ligands in the extracellular matrix.

The extracellular matrix surrounds the cells of tissues and forms the

  • structural support of tissue which is particularly important in connective tissues.

Cells attach to the extracellular matrix through

  • specific cell-surface receptors and molecules
  • including integrins and transmembrane proteoglycans.

Integrins work alongside other proteins such as

  • cadherins,
  • immunoglobulin superfamily
  • cell adhesion molecules,
  • selectins, and
  • syndecans

to mediate

  • cell–cell and
  • cell–matrix interactions and communication.

Activation of adhesion receptors triggers the formation of matrix contacts in which

  • bound matrix components,
  • adhesion receptors,
  • and associated intracellular cytoskeletal and signaling molecules

form large functional, localized multiprotein complexes.

Cell–matrix contacts are important in a variety of different cell and

tissue properties including

  1. embryonic development,
  2. inflammatory responses,
  3. wound healing,
  4. and adult tissue homeostasis.

This review summarizes the roles and functions of integrins and extracellular matrix proteins in mechanotransduction.

Integrins are a family of αβ heterodimeric receptors which act as

  • cell adhesion molecules
  • connecting the ECM to the actin cytoskeleton.

The actin cytoskeleton is involved in the regulation of

  1. cell motility,
  2. cell polarity,
  3. cell growth, and
  4. cell survival.

The integrin family consists of around 25 members which are composed of differing

  • combinations of α and β subunits.

The combination of αβ subunits determines

  • binding specificity and
  • signaling properties.

In mammals around 19 α and eight β subunits have been characterized.

Both α and β integrin subunits contain two separate tails, which

  • penetrate the plasma membrane and possess small cytoplasmic domains which facilitate
  • the signaling functions of the receptor.

There is some evidence that the β subunit is the principal

site for

  • binding of cytoskeletal and signaling molecules,

whereas the α subunit has a regulatory role. The integrin


  • link the ECM to the actin cytoskeleton within the cell and with cytoplasmic proteins,

such as talin, tensin, and filamin. The extracellular domains of integrin receptors bind the ECM ligands.

The ECM is a complex mixture of matrix molecules, including -glycoproteins, collagens, laminins, glycosaminoglycans, proteoglycans,
and nonmatrix proteins, – including growth factors.
These can be categorized as insoluble molecules within the ECM, soluble molecules, and/or matrix-associated biochemicals, such as systemic hormones or growth factors and cytokines that act locally.

The integrin receptor formed from the binding of α and β subunits is shaped like a globular head supported by two rod-like legs (Figure 1). Most of the contact between the two subunits occurs in the head region, with the intracellular tails of the subunits forming the legs of the receptor.6 Integrin recognition of ligands is not constitutive but is regulated by alteration of integrin affinity for ligand binding. For integrin binding to ligands to occur the integrin must be primed and activated, both of which involve conformational changes to the receptor.

The integrins are composed of well-defined domains used for protein–protein interactions. The α-I domains of α integrin subunits comprise the ligand binding sites. X-ray crystallography has identified an α-I domain within the β subunit and a β propeller domain within the α subunit which complex to form the ligand-binding head of the integrin.

The use of activating and conformation-specific antibodies also suggests that the β chain is extended in the active integrin. It has since been identified that the hybrid domain in the β chain is critical for integrin activation, and a swing-out movement of this leg activates integrins.

DBP6: Integrin





Linking integrin conformation to function

Figure  Integrin binding to extracellular matrix (ECM). Conformational changes to integrin structure and clustering of subunits which allow enhanced function of the receptor.

integrin coupled to F-actin via linker

integrin coupled to F-actin via linker

Integrin extracellular binding activity is regulated from inside the cell and binding to the ECM induces signals that are transmitted into the cell.15 This bidirectional signaling requires

  • dynamic,
  • spatially, and
  • temporally regulated formation and
  • disassembly of multiprotein complexes that
    form around the short cytoplasmic tails of integrins.

Ligand binding to integrin family members leads to clustering of integrin molecules in the plasma membrane and recruitment of actin filaments and intracellular signaling molecules to the cytoplasmic domain of the integrins. This forms focal adhesion complexes which are able to maintain

  • not only adhesion to the ECM
  • but are involved in complex signaling pathways

which include establishing

  1. cell polarity,
  2. directed cell migration, and
  3. maintaining cell growth and survival.

Initial activation through integrin adhesion to matrix recruits up to around 50 diverse signaling molecules

  • to assemble the focal adhesion complex
  • which is capable of responding to environmental stimuli efficiently.

Mapping of the integrin

  • adhesome binding and signaling interactions

identified a network of 156 components linked together which can be modified by 690 interactions.

The binding of the adaptor protein talin to the β subunit cytoplasmic tail is known to have a key role in integrin activation. This is thought to occur through the disruption of

  • inhibitory interactions between α and β subunit cytoplasmic tails.

Talin also binds

  • to actin and to cytoskeletal and signaling proteins.

This allows talin to directly link activated integrins

to signaling events and the cytoskeleton.
Genetic programming occurs with the binding of integrins to the ECM

Signal transduction pathway activation arising from integrin-

ECM binding results in changes in gene expression of cells

and leads to alterations in cell and tissue function. Various

different effects can arise depending on the

  1. cell type,
  2. matrix composition, and
  3. integrins activated.

One way in which integrin expression is important in genetic programming is in the fate and differentiation of stem cells.
Osteoblast differentiation occurs through ECM interactions

with specific integrins

  • to initiate intracellular signaling pathways leading to osteoblast-specific gene expression
  • disruption of interactions between integrins and collagen;
  • fibronectin blocks osteoblast differentiation and

Disruption of α2 integrin prevents osteoblast differentiation, and activation of the transcription factor

  • osteoblast-specific factor 2/core-binding factor α1.

It was found that the ECM-integrin interaction induces osteoblast-specific factor 2/core-binding factor α1 to

  • increase its activity as a transcriptional enhancer
  • rather than increasing protein levels.

It was also found that modification of α2 integrin alters

  • induction of the osteocalcin promoter;
  • inhibition of α2 prevents activation of the osteocalcin promoter,
  • overexpression enhanced osteocalcin promoter activity.

It has been suggested that integrin-type I collagen interaction is necessary for the phosphorylation and activation of osteoblast-specific transcription factors present in committed osteoprogenitor cells.

A variety of growth factors and cytokines have been shown to be important in the regulation of integrin expression and function in chondrocytes. Mechanotransduction in chondrocytes occurs through several different receptors and ion channels including integrins. During osteoarthritis the expression of integrins by chondrocytes is altered, resulting in different cellular transduction pathways which contribute to tissue pathology.

In normal adult cartilage, chondrocytes express α1β1, α10β1 (collagen receptors), α5β1, and αvβ5 (fibronectin) receptors. During mechanical loading/stimulation of chondrocytes there is an influx of ions across the cell membrane resulting from activation of mechanosensitive ion channels which can be inhibited by subunit-specific anti-integrin blocking antibodies or RGD peptides. Using these strategies it was identified that α5β1 integrin is a major mechanoreceptor in articular chondrocyte responses to mechanical loading/stimulation.

Osteoarthritic chondrocytes show a depolarization response to 0.33 Hz stimulation in contrast to the hyperpolarization response of normal chondrocytes. The mechanotransduction pathway in chondrocytes derived from normal and osteoarthritic cartilage both involve recognition of the mechanical stimulus by integrin receptors resulting in the activation of integrin signaling pathways leading to the generation of a cytokine loop. Normal and osteoarthritic chondrocytes show differences at multiple stages of the mechanotransduction cascade (Figure 3). Early events are similar; α5β1 integrin and stretch activated ion channels are activated and result in rapid tyrosine phosphorylation events. The actin cytoskeleton is required for the integrin-dependent Mechanotransduction leading to changes in membrane potential in normal but not osteoarthritic chondrocytes.

Cell–matrix interactions are essential for maintaining the integrity of tissues. An intact matrix is essential for cell survival and proliferation and to allow efficient mechanotransduction and tissue homeostasis. Cell–matrix interactions have been extensively studied in many tissues and this knowledge is being used to develop strategies to treat pathology. This is particularly important in tissues subject to abnormal mechanical loading, such as musculoskeletal tissues. Integrin-ECM interactions are being used to enhance tissue repair mechanisms in these tissues through differentiation of progenitor cells for in vitro and in vivo use. Knowledge of how signaling cascades are differentially regulated in response to physiological and pathological external stimuli (including ECM availability and mechanical loading/stimulation) will enable future strategies to be developed to prevent and treat the progression of pathology associated with integrin-ECM interactions.

Cellular adaptation to mechanical stress: role of integrins, Rho, cytoskeletal tension and mechanosensitive ion channels

  1. Matthews, DR. Overby, R Mannix and DE. Ingber
    1Vascular Biology Program, Departments of Pathology and Surgery, Children’s Hospital, and 2Department of Pediatrics, Massachusetts General Hospital, Harvard Medical School, Boston, MA J Cell Sci 2006; 119: 508-518.

To understand how cells sense and adapt to mechanical stress, we applied tensional forces to magnetic microbeads bound to cell-surface integrin receptors and measured changes in bead isplacement with sub-micrometer resolution using optical microscopy. Cells exhibited four types of mechanical responses: (1) an immediate viscoelastic response;

(2) early adaptive behavior characterized by pulse-to-pulse attenuation in response to oscillatory forces;

(3) later adaptive cell stiffening with sustained (>15 second) static stresses; and

(4) a large-scale repositioning response with prolonged (>1 minute) stress.

Importantly, these adaptation responses differed biochemically. The immediate and early responses were affected by

  • chemically dissipating cytoskeletal prestress (isometric tension), whereas
  • the later adaptive response was not.

The repositioning response was prevented by

  • inhibiting tension through interference with Rho signaling,

similar to the case of the immediate and early responses, but it was also prevented by

  • blocking mechanosensitive ion channels or
  • by inhibiting Src tyrosine kinases.

All adaptive responses were suppressed by cooling cells to 4°C to slow biochemical remodeling. Thus, cells use multiple mechanisms to sense and respond to static and dynamic changes in the level of mechanical stress applied to integrins.

Microtubule-Stimulated ADP Release, ATP Binding, and Force Generation In Transport Kinesins

J Atherton, I Farabella, I-Mei Yu, SS Rosenfeld, A Houdusse, M Topf, CA Moores

1Institute of Structural and Molecular Biology, Department of Biological Sciences, Birkbeck College, University of London, London, United Kingdom; 2Structural Motility, Institut Curie, Centre National de la Recherche Scientifique, Paris, France; 3Department of Cancer Biology, Lerner Research Institute, Cleveland Clinic, Cleveland, United States
eLife 2014;3:e03680.

Kinesins are a large family of microtubule (MT)-based motors that play important roles in many cellular activities including

  • mitosis,
  • motility, and
  • intracellular transport

Their involvement in a range of pathological processes also highlights their significance as therapeutic targets and the importance of understanding the molecular basis of their function They are defined by their motor domains that contain both the microtubule (MT) and ATP binding sites. Three ATP binding motifs—the P-loop, switch I, switch II–are highly conserved among kinesins, myosin motors, and small GTPases. They share a conserved mode of MT binding such that MT binding, ATP binding, and hydrolysis are functionally coupled for efficient MT-based work.

The interior of a cell is a hive of activity, filled with proteins and other items moving from one location to another. A network of filaments called microtubules forms tracks along which so-called motor proteins carry these items. Kinesins are one group of motor proteins, and a typical kinesin protein has one end (called the ‘motor domain’) that can attach itself to the microtubules.

The other end links to the cargo being carried, and a ‘neck’ connects the two. When two of these proteins work together, flexible regions of the neck allow the two motor domains to move past one another, which enable the kinesin to essentially walk along a microtubule in a stepwise manner.

Atherton et al. use a technique called cryo-electron microscopy to study—in more detail than previously seen—the structure of the motor domains of two types of kinesin called kinesin-1 and kinesin-3. Images were taken at different stages of the cycle used by the motor domains to extract the energy from ATP molecules. Although the two kinesins have been thought to move along the microtubule tracks in different ways, Atherton et al. find that the core mechanism used by their motor domains is the same.

When a motor domain binds to the microtubule, its shape changes, first stimulating release of the breakdown products of ATP from the previous cycle. This release makes room for a new ATP molecule to bind. The structural changes caused by ATP binding are relatively small but produce larger changes in the flexible neck region that enable individual motor domains within a kinesin pair to co-ordinate their movement and move in a consistent direction. This mechanism involves tight coupling between track binding and fuel usage and makes kinesins highly efficient motors.

A number of kinesins drive long distance transport of cellular cargo with dimerisation allowing them to take multiple 8 nm ATP-driven steps toward MT plus ends. Their processivity depends on communication between the two motor domains, which is achieved via the neck linker that connects each motor domain to the dimer-forming coiled-coil

Kinesins are a superfamily of microtubule-based

  • ATP-powered motors, important for multiple, essential cellular functions.

How microtubule binding stimulates their ATPase and controls force generation is not understood. To address this fundamental question, we visualized microtubule-bound kinesin-1 and kinesin-3 motor domains at multiple steps in their ATPase cycles—including their nucleotide-free states—at ∼7 Å resolution using cryo-electron microscopy.

All our reconstructions have, as their asymmetric unit, a triangle-shaped motor domain bound to an αβ-tubulin dimer within the MT lattice (Figure 1). The structural comparisons below are made with respect to the MT surface, which, at the resolution of our structures (∼7 Å, Table 1), is the same (CCC > 0.98 for all). As is well established across the superfamily, the major and largely invariant point of contact between kinesin motor domains and the MT is helix-α4, which lies at the tubulin intradimer interface (Figure 1C, Kikkawa et al., 2001).

However, multiple conformational changes are seen throughout the rest of each domain in response to bound nucleotide (Figure 1D). Below, we describe the conformational changes in functionally important regions of each motor domain starting with the nucleotide-binding site, from which all other conformational changes emanate.

The nucleotide-binding site (Figure 2) has three major elements: (1) the P-loop (brown) is visible in all our reconstructions;

(2) loop9 (yellow, contains switch I) undergoes major conformational changes through the ATPase cycle; and

(3) loop11 (red, contains switch II) that connects strand-β7 to helix-α4,

the conformation and flexibility of which is determined by MT binding and motor nucleotide state.

Movement and extension of helix-α6 controls neck linker docking

the N-terminus of helix-α6 is closely associated with elements of the nucleotide binding site suggesting that its conformation alters in response to different nucleotide states. In addition, because the orientation of helix-α6 with respect to helix-α4 controls neck linker docking and because helix-α4 is held against the MT during the ATPase cycle,

  • conformational changes in helix-α6 control movement of the neck linker.

Mechanical amplification and force generation involves conformational changes across the motor domain

A key conformational change in the motor domain following Mg-ATP binding is peeling of the central β-sheet from the C-terminus of helix-α4 increasing their separation (Figure 3—figure supplement 2); this is required to accommodate rotation of helix-α6 and consequent neck linker docking (Figure 3B–E).

Peeling of the central β-sheet has previously been proposed to arise from tilting of the entire motor domain relative to static MT contacts, pivoting around helix-α4 (the so-called ‘seesaw’ model; Sindelar, 2011). Specifically, this model predicts that the major difference in the motor before and after Mg-ATP binding would be the orientation of the motor domain with respect to helix-α4.

Kinesin mechanochemistry and the extent of mechanistic conservation within the motor superfamily are open questions, critical to explain how MT binding, and ATP binding and hydrolysis drive motor activity. Our structural characterisation of two transport motors now allows us to propose a model that describes the roles of mechanochemical elements that together drive conserved MT-based motor function.

Model of conserved MT-bound kinesin mechanochemistry. Loop11/N-terminus of helix-α4 is flexible in ADP-bound kinesin in solution, the neck linker is also flexible while loop9 chelates ADP. MT binding is sensed by loop11/helix-α4 N-terminus, biasing them towards more ordered conformations.

We propose that this favours crosstalk between loop11 and loop9, stimulating ADP release. In the NN conformation, both loop11 and loop9 are well ordered and primed to favour ATP binding, while helix-α6—which is required for mechanical amplification–is closely associated with the MT on the other side of the motor domain. ATP binding draws loop11 and loop9 closer together; causing

(1) tilting of most of the motor domain not contacting the MT towards the nucleotide-binding site,

(2) rotation, translation, and extension of helix-α6 which we propose contributes to force generation, and

(3) allows neck linker docking and biases movement of the 2nd head towards the MT plus end.

In both motors, microtubule binding promotes

  • ordered conformations of conserved loops that
  • stimulate ADP release,
  • enhance microtubule affinity and
  • prime the catalytic site for ATP binding.

ATP binding causes only small shifts of these nucleotide-coordinating loops but induces

  • large conformational changes elsewhere that
  • allow force generation and
  • neck linker docking towards the microtubule plus end.

Family-specific differences across the kinesin–microtubule interface account for the

  • distinctive properties of each motor.

Our data thus provide evidence for a

conserved ATP-driven

  • mechanism for kinesins and
  • reveal the critical mechanistic contribution of the microtubule interface.

Phosphorylation at endothelial cell–cell junctions: Implications for VE-cadherin function

I Timmerman, PL Hordijk, JD van Buul

Cell Health and Cytoskeleton 2010; 2: 23–31
Endothelial cell–cell junctions are strictly regulated in order to

  • control the barrier function of endothelium.

Vascular endothelial (VE)-cadherin is one of the proteins that is crucial in this process. It has been reported that

  • phosphorylation events control the function of VE-cadherin.

This review summarizes the role of VE-cadherin phosphorylation in the regulation of endothelial cell–cell junctions and highlights how this affects vascular permeability and leukocyte extravasation.

The vascular endothelium is the inner lining of blood vessels and

  • forms a physical barrier between the vessel lumen and surrounding tissue;
  • controlling the extravasation of fluids,
  • plasma proteins and leukocytes.

Changes in the permeability of the endothelium are tightly regulated. Under basal physiological conditions, there is a continuous transfer of substances across the capillary beds. In addition the endothelium can mediate inducible,

  • transient hyperpermeability
  • in response to stimulation with inflammatory mediators,
  • which takes place primarily in postcapillary venules.

However, when severe, inflammation may result in dysfunction of the endothelial barrier in various parts of the vascular tree, including large veins, arterioles and capillaries. Dysregulated permeability is observed in various pathological conditions, such as tumor-induced angiogenesis, cerebrovascular accident and atherosclerosis.

Two fundamentally different pathways regulate endothelial permeability,

  • the transcellular and paracellular pathways.

Solutes and cells can pass through the body of endothelial cells via the transcellular pathway, which includes

  • vesicular transport systems, fenestrae, and biochemical transporters.

The paracellular route is controlled by

  • the coordinated opening and closing of endothelial junctions and
  • thereby regulates traffic across the intercellular spaces between endothelial cells.

Endothelial cells are connected by

  • tight, gap and
  • adherens junctions,

of which the latter, and particularly the adherens junction component,

  • vascular endothelial (VE)-cadherin,
  • are of central importance for the initiation and stabilization of cell–cell contacts.

Although multiple adhesion molecules are localized at endothelial junctions, blocking the adhesive function of VE-cadherin using antibodies is sufficient to disrupt endothelial junctions and to increase endothelial monolayer permeability both in vitro and in vivo. Like other cadherins, VE-cadherin mediates adhesion via homophilic, calcium-dependent interactions.

This cell–cell adhesion

  • is strengthened by binding of cytoplasmic proteins, the catenins,
  • to the C-terminus of VE-cadherin.

VE-cadherin can directly bind β-catenin and plakoglobin, which

  • both associate with the actin binding protein α-catenin.

Initially, α-catenin was thought to directly anchor cadherins to the actin cytoskeleton, but recently it became clear that

  • α-catenin cannot bind to both β-catenin and actin simultaneously.

Data using purified proteins show that

  • monomeric α-catenin binds strongly to cadherin-bound β-catenin;
  • in contrast to the dimer which has a higher affinity for actin filaments,
  • indicating that α-catenin might function as a molecular switch regulating cadherin-mediated cell–cell adhesion and actin assembly.

Thus, interactions between the cadherin complex and the actin cytoskeleton are more complex than previously thought. Recently, Takeichi and colleagues reported that

  • the actin binding protein EPLIN (epithelial protein lost in neoplasm)
  • can associate with α-catenin and thereby
  • link the E-cadherin–catenin complex to the actin cytoskeleton.

Although this study was performed in epithelial cells,

  • an EPLIN-like molecule might serve as
  • a bridge between the cadherin–catenin complex and
  • the actin cytoskeleton in endothelial cells.

Next to β-catenin and plakoglobin, p120-catenin also binds directly to the intracellular tail of VE-cadherin.

Numerous lines of evidence indicate that

  • p120-catenin promotes VE-cadherin surface expression and stability at the plasma membrane.

Different models are proposed that describe how p120-catenin regulates cadherin membrane dynamics, including the hypothesis

  • that p120-catenin functions as a ‘cap’ that prevents the interaction of VE-cadherin
  • with the endocytic membrane trafficking machinery.

In addition, p120-catenin might regulate VE-cadherin internalization through interactions with small GTPases. Cytoplasmic p120-catenin, which is not bound to VE-cadherin, has been shown to

  • decrease RhoA activity,
  • elevate active Rac1 and Cdc42, and thereby is thought
  • to regulate actin cytoskeleton organization and membrane trafficking.

The intact cadherin-catenin complex is required for proper functioning of the adherens junction. Mutant forms of VE-cadherin which

  • lack either the β-catenin, plakoglobin or p120 binding regions reduce the strength of cell–cell adhesion.

Moreover, our own results showed that

  • interfering with the interaction between α-catenin and β-catenin,
  • using a cell-permeable peptide which encodes the binding site in α-catenin for β-catenin,
  • resulted in an increased permeability of the endothelial monolayer.

Several mechanisms may be involved in the regulation of the organization and function of the cadherin–catenin complex, including endocytosis of the complex, VE-cadherin cleavage and actin cytoskeleton reorganization. The remainder of this review primarily focuses on the

  • role of tyrosine phosphorylation in the control of VE-cadherin-mediated cell–cell adhesion.

Regulation of the adhesive function of VE-cadherin by tyrosine phosphorylation

It is a widely accepted concept that tyrosine phosphorylation of components of the VE–cadherin-catenin complex

  • Correlates with the weakening of cell–cell adhesion.

One of the first reports that supported this idea showed that the level of phosphorylation of VE-cadherin was

  • high in loosely confluent endothelial cells, but
  • low in tightly confluent monolayers,

when intercellular junctions are stabilized.

In addition, several conditions that induce tyrosine phosphorylation

of adherens junction components, like

  • v-Src transformation
  • and inhibition of phosphatase activity by pervanadate,

have been shown to shift cell–cell adhesion from a strong to a weak state. More physiologically relevant;

permeability-increasing agents such as

  • histamine,
  • tumor necrosis factor-α (TNF-α),
  • thrombin,
  • platelet-activating factor (PAF) and
  • vascular endothelial growth factor (VEGF)

increase tyrosine phosphorylation of various components of the cadherin–catenin complex.

A general idea has emerged that

  • tyrosine phosphorylation of the VE-cadherin complex
  • leads to the uncoupling of VE-cadherin from the actin cytoskeleton
  • through dissociation of catenins from the cadherin.

However, tyrosine phosphorylation of VE-cadherin is required for efficient transmigration of leukocytes.

This suggests that VE-cadherin-mediated cell–cell contacts

  1. are not just pushed open by the migrating leukocytes, but play
  2. a more active role in the transmigration process.

A schematic overview of leukocyte adhesion-induced signals leading to VE-cadherin phosphorylation

Regulation of the integrity of endothelial cell–cell contacts by phosphorylation of VE-cadherin

Regulation of the integrity of endothelial cell–cell contacts by phosphorylation of VE-cadherin

Regulation of the integrity of endothelial cell–cell contacts by phosphorylation of VE-cadherin.

Notes: A) Permeability-inducing agents such as thrombin, histamine and VEGF, induce tyrosine phosphorylation (pY) of VE-cadherin and the associated catenins. Although the specific consequences of catenin tyrosine phosphorylation in endothelial cells are still unknown, VE-cadherin tyrosine phosphorylation results in opening of the cell–cell junctions (indicated by arrows) and enhanced vascular permeability. How tyrosine phosphorylation affects VE-cadherin adhesiveness is not yet well understood; disrupted binding of catenins, which link the cadherin to the actin cytoskeleton, may be involved. VEGF induces phosphorylation of VE-cadherin at specific residues, Y658 and Y731, which have been reported to regulate p120-catenin and β-catenin binding, respectively. Moreover, VEGF stimulation results in serine phosphorylation (pSer) of VE-cadherin, specifically at residue S665, which leads to its endocytosis. B) Adhesion of leukocytes to endothelial cells via ICAM-1 increases endothelial permeability by inducing phosphorylation of VE-cadherin on tyrosine residues. Essential mediators, such as the kinases Pyk2 and Src, and signaling routes involving reactive oxygen species (ROS) and Rho, have been shown to act downstream of ICAM-1. Different tyrosine residues within the cytoplasmic domain of VE-cadherin are involved in the extravasation of neutrophils and lymphocytes, including Y658 and Y731. (β: β-catenin, α: α-catenin, γ: γ-catenin/plakoglobin).

N-glycosylation status of E-cadherin controls cytoskeletal dynamics through the organization of distinct β-catenin- and γ-catenin-containing AJs

BT Jamal, MN Nita-Lazar, Z Gao, B Amin, J Walker, MA Kukuruzinska
Cell Health and Cytoskeleton 2009; 1: 67–80

N-glycosylation of E-cadherin has been shown to inhibit cell–cell adhesion. Specifically, our recent studies have provided evidence that the reduction of E-cadherin N-glycosylation promoted the recruitment of stabilizing components, vinculin and serine/ threonine protein phosphatase 2A (PP2A), to adherens junctions (AJs) and enhanced the association of AJs with the actin cytoskeleton. Here, we examined the details of how

  • N-glycosylation of E-cadherin affected the molecular organization of AJs and their cytoskeletal interactions.

Using the hypoglycosylated E-cadherin variant, V13, we show that

  • V13/β-catenin complexes preferentially interacted with PP2A and with the microtubule motor protein dynein.

This correlated with dephosphorylation of the microtubule-associated protein tau, suggesting that

  • increased association of PP2A with V13-containing AJs promoted their tethering to microtubules.

On the other hand, V13/γ-catenin complexes associated more with vinculin, suggesting that they

  • mediated the interaction of AJs with the actin cytoskeleton.
  • N-glycosylation driven changes in the molecular organization of AJs were physiologically significant because transfection of V13 into A253 cancer cells, lacking both mature AJs and tight junctions (TJs), promoted the formation of stable AJs and enhanced the function of TJs to a greater extent than wild-type E-cadherin.

These studies provide the first mechanistic insights into how N-glycosylation of E-cadherin drives changes in AJ composition through

  • the assembly of distinct β-catenin- and γ-catenin-containing scaffolds that impact the interaction with different cytoskeletal components.

Cytoskeletal Basis of Ion Channel Function in Cardiac Muscle

Matteo Vatta, and Georgine Faulkner,

1 Departments of Pediatrics (Cardiology), Baylor College of Medicine, Houston, TX 2 Department of Reproductive and Developmental Sciences, University of Trieste, Trieste, Italy
3 Muscular Molecular Biology Unit, International Centre for Genetic Engineering and Biotechnology, Padriciano, Trieste, Italy

Future Cardiol. 2006 July 1; 2(4): 467–476.

The heart is a force-generating organ that responds to

  • self-generated electrical stimuli from specialized cardiomyocytes.

This function is modulated

  • by sympathetic and parasympathetic activity.

In order to contract and accommodate the repetitive morphological changes induced by the cardiac cycle, cardiomyocytes

  • depend on their highly evolved and specialized cytoskeletal apparatus.

Defects in components of the cytoskeleton, in the long term,

  • affect the ability of the cell to compensate at both functional and structural levels.

In addition to the structural remodeling,

  • the myocardium becomes increasingly susceptible to altered electrical activity leading to arrhythmogenesis.

The development of arrhythmias secondary to structural remodeling defects has been noted, although the detailed molecular mechanisms are still elusive. Here I will review

  • the current knowledge of the molecular and functional relationships between the cytoskeleton and ion channels

and, I will discuss the future impact of new data on molecular cardiology research and clinical practice.

Myocardial dysfunction in the end-stage failing heart is very often associated with increasing

  • susceptibility to ventricular tachycardia (VT) and ventricular fibrillation (VF),

both of which are common causes of sudden cardiac death (SCD).

Among the various forms of HF,

myocardial remodeling due to ischemic cardiomyopathy (ICM) or dilated cardiomyopathy (DCM)

  • is characterized by alterations in baseline ECG,

which includes the

  • prolongation of the QT interval,
  • as well as QT dispersion,
  • ST-segment elevation, and
  • T-wave abnormalities,

especially during exercise. In particular, subjects with

severe left ventricular chamber dilation such as in DCM can have left bundle branch block (LBBB), while right bundle branch block (RBBB) is more characteristic of right ventricular failure.  LBBB and RBBB have both been repeatedly associated with AV block in heart failure.

The impact of volume overload on structural and electro-cardiographic alterations has been noted in cardiomyopathy patients treated with left ventricular assist device (LVAD) therapy, which puts the heart at mechanical rest. In LVAD-treated subjects,

  • QRS- and both QT- and QTc duration decreased,
  • suggesting that QRS- and QT-duration are significantly influenced by mechanical load and
  • that the shortening of the action potential duration contributes to the improved contractile performance after LVAD support.

Despite the increasing use of LVAD supporting either continuous or pulsatile blood flow in patients with severe HF, the benefit of this treatment in dealing with the risk of arrhythmias is still controversial.

Large epidemiological studies, such as the REMATCH study, demonstrated that the

  • employment of LVAD significantly improved survival rate and the quality of life, in comparison to optimal medical management.

An early postoperative period study after cardiac unloading therapy in 17 HF patients showed that in the first two weeks after LVAD implantation,

  • HF was associated with a relatively high incidence of ventricular arrhythmias associated with QTc interval prolongation.

In addition, a recent retrospective study of 100 adult patients with advanced HF, treated with an axial-flow HeartMate LVAD suggested that

  • the rate of new-onset monomorphic ventricular tachycardia (MVT) was increased in LVAD treated patients compared to patients given only medical treatment,

while no effect was observed on the development of polymorphic ventricular tachycardia (PVT)/ventricular fibrillation (VF).

The sarcomere

The myocardium is exposed to severe and continuous biomechanical stress during each contraction-relaxation cycle. When fiber tension remains uncompensated or simply unbalanced,

  • it may represent a trigger for arrhythmogenesis caused by cytoskeletal stretching,
  • which ultimately leads to altered ion channel localization, and subsequent action potential and conduction alterations.

Cytoskeletal proteins not only provide the backbone of the cellular structure, but they also

  • maintain the shape and flexibility of the different sub-cellular compartments, including the
  1. plasma membrane,
  2. the double lipid layer, which defines the boundaries of the cell and where
  • ion channels are mainly localized.

The interaction between the sarcomere, which is the basic for the passive force during diastole and for the restoring force during systole. Titin connects

  • the Z-line to the M-line of the sarcomeric structure
    (Figure 1).

In addition to the strategic

  • localization and mechanical spring function,
  • titin is a length-dependent sensor during
  • stretch and promotes actin-myosin interaction

Titin is stabilized by the cross-linking protein

  • telethonin (T-Cap), which localizes at the Z-line and is also part of titin sensor machinery (Figure 1).

The complex protein interactions in the sarcomere entwine telethonin to other

  • Z-line components through the family of the telethonin-binding proteins of the Z-disc, FATZ, also known as calsarcin and myozenin.

FATZ binds to

  1. calcineurin,
  2. γ-filamin as well as the
  3. spectrin-like repeats (R3–R4) of α-actinin-2,

the major component of the Z-line and a pivotal

  • F-actin cross-linker (Figure 1).contractile unit of striated muscles, and
  • the sarcolemma,

the plasma membrane surrendering the muscle fibers in skeletal muscle and the muscle cell of the cardiomyocyte,

  • determines the mechanical plasticity of the cell, enabling it to complete and re-initiate each contraction-relaxation cycle.

At the level of the sarcomere,

  • actin (thin) and myosin (thick) filaments generate the contractile force,

while other components such as titin, the largest protein known to date, are responsible for

  • the passive force during diastole and for the restoring force during systole, and (titin).
  • the Z-line to the M-line of the sarcomeric structure
    (Figure 1).

In addition to the strategic

  • localization and mechanical spring function,
  • it acts as a length-dependent sensor during stretch and
  • promotes actin-myosin interaction.

Stabilized by the cross-linking protein telethonin (T-Cap),

  • titin localizes at the Z-line and is
  • part of titin sensor machinery

Another cross-linker of α-actinin-2 in the complex Z-line scaffold is

  • the Z-band alternatively spliced PDZ motif protein (ZASP),
  • which has an important role in maintaining Z-disc stability

in skeletal and cardiac muscle (Figure 1).

ZASP contains a PDZ motif at its N-terminus,

  • which interacts with C-terminus of α-actinin-2,
  • and a conserved sequence called the ZASP like motif (ZM)
  • found in the alternatively spliced exons 4 and 6.

It has also been reported

  • to bind to the FATZ (calsarcin) family of Z-disc proteins (Figure 1).

The complex protein interactions in the sarcomere entwine telethonin to other Z-line components through the family of the telethonin-binding proteins of the

  1. Z-disc,
  2. FATZ, also known as calsarcin and
  3. myozenin

FATZ binds to calcineurin,

  1. γ-filamin as well as the
  2. spectrin-like repeats (R3–R4) of α-actinin-2, the major component of the Z-line and a pivotal F-actin cross-linker (Figure 1).
sarcomere structure

sarcomere structure

Figure 1. Sarcomere structure

The diagram illustrates the sarcomeric structure. The Z-line determines the boundaries of the contractile unit, while Titin connects the Z-line to the M-line and acts as a functional spring during contraction/relaxation cycles.

Sarcomeric Proteins and Ion Channels

In addition to systolic dysfunction characteristic of dilated cardiomyopathy (DCM) and diastolic dysfunction featuring hypertrophic cardiomyopathy (HCM), the clinical phenotype of patients with severe cardiomyopathy is very often associated with a high incidence of cardiac arrhythmias. Therefore, besides fiber stretch associated with mechanical and hemodynamic impairment, cytoskeletal alterations due to primary genetic defects or indirectly to alterations in response to cellular injury can potentially

  1. affect ion channel anchoring, and trafficking, as well as
  2. functional regulation by second messenger pathways,
  3. causing an imbalance in cardiac ionic homeostasis that will trigger arrhythmogenesis.

Intense investigation of

  • the sarcomeric actin network,
  • the Z-line structure, and
  • chaperone molecules docking in the plasma membrane,

has shed new light on the molecular basis of

  • cytoskeletal interactions in regulating ion channels.

In 1991, Cantiello et al., demonstrated that

  • although the epithelial sodium channel and F-actin are in close proximity,
  • they do not co-localize.

Actin disruption using cytochalasin D, an agent that interferes with actin polymerization, increased Na+ channel activity in 90% of excised patches tested within 2 min, which indicated that

  • the integrity of the filamentous actin (F-actin) network was essential
  • for the maintenance of normal Na+ channel function.

Later, the group of Dr. Jonathan Makielski demonstrated that

  • actin disruption induced a dramatic reduction in Na+ peak current and
  • slowed current decay without affecting steady-state voltage-dependent availability or recovery from inactivation.

These data were the first to support a role for the cytoskeleton in cardiac arrhythmias.

F-actin is intertwined in a multi-protein complex that includes

  • the composite Z-line structure.

Further, there is a direct binding between

  • the major protein of the Z-line, α-actinin-2 and
  • the voltage-gated K+ channel 1.5 (Kv1.5), (Figure 2).

The latter is expressed in human cardiomyocytes and localizes to

  • the intercalated disk of the cardiomyocyte
  • in association with connexin and N-cadherin.

Maruoka et al. treated HEK293 cells stably expressing Kv1.5 with cytochalasin D, which led to

  • a massive increase in ionic and gating IK+ currents.

This was prevented by pre-incubation with phalloidin, an F-actin stabilizing agent. In addition, the Z-line protein telethonin binds to the cytoplasmic domain of minK, the beta subunit of the potassium channel KCNQ1 (Figure 2).

Molecular interactions between the cytoskeleton and ion channels

Molecular interactions between the cytoskeleton and ion channels

Figure 2. Molecular interactions between the cytoskeleton and ion channels

The figure illustrates the interactions between the ion channels on the sarcolemma, and the sarcomere in cardiac myocytes. Note that the Z-line is connected to the cardiac T-tubules. The diagram illustrates the complex protein-protein interactions that occur between structural components of the cytoskeleton and ion channels. The cytoskeleton is involved in regulating the metabolism of ion channels, modifying their expression, localization, and electrical properties. The cardiac sodium channel Nav1.5 associates with the DGC, while potassium channels such as Kv1.5, associate with the Z-line.

Ion Channel Subunits and Trafficking

Correct localization is essential for ion channel function and this is dependent upon the ability of auxiliary proteins to

  • shuttle ion channels from the cytoplasm to their final destination such as
  • the plasma membrane or other sub-cellular compartments.

In this regard, Kvβ-subunits are

  • cytoplasmic components known to assemble with the α-subunits of voltage-dependent K+ (Kv) channels
  • at their N-terminus to form stable Kvα/β hetero-oligomeric channels.

When Kvβ is co-expressed with Kv1.4 or Kv1.5, it enhances Kv1.x channel trafficking to the cell membrane without changing the overall protein channel content. The regulatory Kvβ subunits, which are also expressed in cardiomyocytes, directly decrease K+ current by

  • accelerating Kv1.x channel inactivation.

Therefore, altered expression or mutations in Kvβ subunits could cause abnormal ion channel transport to the cell surface, thereby increasing the risk of cardiac arrhythmias.

Ion Channel Protein Motifs and Trafficking

Cell membrane trafficking in the Kv1.x family may occur in a Kvβ subunit-independent manner through specific motifs in their C-terminus. Mutagenesis of the final asparagine (N) in the Kv1.2 motif restores the leucine (L) of the Kv1.4 motif

  • re-establishing high expression levels at the plasma membrane in a Kvβ-independent manner

Cytoskeletal Proteins and Ion Channel Trafficking

Until recently, primary arrhythmias such as LQTS have been almost exclusively regarded as ion channelopathies. Other mutations have been identified with regard to channelopathies. However, the conviction that primary mutations in ion channels were solely responsible for

  • the electrical defects associated with arrhythmias

has been shaken by the identification of mutations in the

  • ANK2 gene encoding the cytoskeletal protein ankyrin-B

that is associated with LQTS in animal models and humans.

Ankyrin-B acts as a chaperone protein, which shuttles the cardiac sodium channel from the cytoplasm to the membrane. Immunohistochemical analysis has localized ankyrin-B to the Zlines/T-tubules on the plasma membrane in the myocardium. Mutations in ankyrin-B associated with LQTS

  • alter sodium channel trafficking due to loss of ankyrin-B localization at the Z-line/transverse (T)-tubules.

Reduced levels of ankyrin-B at cardiac Z-lines/T-tubules were associated with the deficiency of ankyrin-B-associated proteins such as Na/K-ATPase, Na/Ca exchanger (NCX) and inositol-1, 4, 5-trisphosphate receptors (InsP3R).

Dystrophin component of the Dystrophin Glycoprotien Complex (DGC)

Synchronized contraction is essential for cardiomyocytes, which are connected to each other via the extracellular matrix (ECM) through the DGC. The N-terminus domain of dystrophin

  • binds F-actin, and connects it to the sarcomere, while
  • the cysteine-rich (CR) C-terminus domain ensures its connection to the sarcolemma (Figure 2).

The central portion of dystrophin, the rod domain, is composed of

  • rigid spectrin-like repeats and four hinge portions (H1–H4) that determine the flexibility of the protein.

Dystrophin possesses another F-actin binding domain in the Rod domain region, between the basic repeats 11- 17 (DysN-R17).

Dystrophin, originally identified as the gene responsible for Duchenne and Becker muscular dystrophies (DMD/BMD), and later for the X-linked form of dilated cardiomyopathy (XLCM), exerts a major function in physical force transmission in striated muscle. In addition to its structural significance, dystrophin and other DGC proteins such as syntrophins are required for the

  • correct localization,
  • clustering and
  • regulation of ion channel function.

Syntrophins have been implicated in ion channel regulation.  Syntrophins contain two pleckstrin homology (PH) domains, a PDZ domain, and a syntrophin-unique (SU) C-terminal region. The interaction between syntrophins and dystrophin occurs at the PH domain distal to the syntrophin N-terminus and through the highly conserved SU domain. Conversely, the PH domain proximal to the N-terminal portion of the protein and the PDZ domain interact with other membrane components such as

  1. phosphatidyl inositol-4, 5-bisphosphate,
  2. neuronal NOS (nNOS),
  3. aquaporin-4,
  4. stress-activated protein kinase-3, and
  5. 5,

thereby linking all these molecules to the dystrophin complex (Figure 2).

Among the five known isoforms of syntrophin, the 59 KDa α1-syntrophin isoform is the most highly represented in human heart, whereas in skeletal muscle it is only present on the

  • sarcolemma of fast type II fibers.

In addition, the skeletal muscle γ2-syntrophin was found at high levels only at the

  • postsynaptic membrane of the neuromuscular junctions.

In addition to syntrophin, other scaffolding proteins such as caveolin-3 (CAV3), which is present in the caveolae, flask-shaped plasma membrane microdomains, are involved

  • in signal transduction and vesicle trafficking in myocytes,
  • modulating cardiac remodeling during heart failure.

CAV3 and α1-syntrophin, localizes at the T-tubule and are part of the DGC. In addition, α1-syntrophin binds Nav1.5, while

  • caveolin-3 binds the Na+/Ca2+ exchanger, Nav1.5 and the L-type Ca2+ channel as well as nNOS and the DGC (Figure 2).

Although ankyrin-B is the only protein found mutated in patients with primary arrhythmias, other proteins such as caveolin-3 and the syntrophins if mutated may alter ion channel function.


It is important to be aware of the enormous variety of clinical presentations that derive from distinct variants in the same pool of genetic factors. Knowledge of these variants could facilitate tailoring the therapy of choice for each patient. In particular, the recent findings of structural and functional links between

  • the cytoskeleton and ion channels

could expand the therapeutic interventions in

  • arrhythmia management in structurally abnormal myocardium, where aberrant binding
  • between cytoskeletal proteins can directly or indirectly alter ion channel function.

Executive Summary

Arrhythmogenesis and myocardial structure

  • Rhythm alterations can develop as a secondary consequence of myocardial structural abnormalities or as a result of a primary defect in the cardiac electric machinery.
  • Until recently, no molecular mechanism has been able to fully explain the occurrence of arrhythmogenesis in heart failure, however genetic defects that are found almost exclusively in ion channel genes account for the majority of primary arrhythmias such as long QT syndromes and Brugada syndrome. The contractile apparatus is linked to ion channels
  • The sarcomere, which represents the contractile unit of the myocardium not only generates the mechanical force necessary to exert the pump function, but also provides localization and anchorage to ion channels.
  • Alpha-actinin-2, and telethonin, two members of the Z-line scaffolding protein complex in the striated muscle associate with the potassium voltage-gated channel alpha subunit Kv1.5 and the beta subunit KCNE1 respectively.
  • Mutations in KCNE1 have previously been associated with the development of arrhythmias in LQTS subjects.
  • Mutations in both alpha-actinin-2, and telethonin were identified in individuals with cardiomyopathy. The primary defect is structural leading to ventricular dysfunction, but the secondary consequence is arrhythmia.

Ion channel trafficking and sub-cellular compartments

  • Ion channel trafficking from the endoplasmic reticulum (ER) to the Golgi complex is an important check-point for regulating the functional channel molecules on the plasma membrane. Several molecules acting as chaperones bind to and shuttle the channel proteins to their final localization on the cell surface
  • Ion channel subunits such as Kvβ enhance Kv1.x ion channel presentation on the sarcolemma. The α subunits of the Kv1.x potassium channels can be shuttled in a Kvβ-independent manner through specific sequence motif at Kv1.x protein level.
  • In addition, cytoskeletal proteins such as ankyrin-G bind Nav1.5 and are involved in the sodium channel trafficking. Another member of the ankyrin family, ankyrin-B was found mutated in patients with LQTS but the pathological mechanism of ankyrin-B mutations is still obscure, although the sodium current intensity is dramatically reduced.

The sarcolemma and ion channels

  • The sarcolemma contains a wide range of ion channels, which are responsible for the electrical propagating force in the myocardium.
  • The DGC is a protein complex, which forms a scaffold for cytoskeletal components and ion channels.
  • Dystrophin is the major component of the DGC and mutations in dystrophin and DGC cause muscular dystrophies and X-linked cardiomyopathies (XLCM) in humans. Cardiomyopathies are associated with arrhythmias
  • Caveolin-3 and syntrophins associate with Nav1.5, and are part of the DGC. Syntrophins can directly modulate Nav1.5 channel function.


  • The role of the cytoskeleton in ion channel function has been hypothesized in the past, but only recently the mechanism underlying the development of arrhythmias in structurally impaired myocardium has become clearer.
  • The recently acknowledged role of the cytoskeleton in ion channel function suggests that genes encoding cytoskeletal proteins should be regarded as potential candidates for variants involved in the susceptibility to arrhythmias, as well as the primary target of genetic mutations in patients with arrhythmogenic syndromes such as LQTS and Brugada syndrome.
  • Studies of genotype-phenotype correlation and and patient risk stratification for mutations in cytoskeletal proteins will help to tailor the therapy and management of patients with arrhythmias.

Read Full Post »

Introduction to Signaling

Curator: Larry H. Bernstein, MD, FCAP


We have laid down a basic structure and foundation for the remaining presentations.  It was essential to begin with the genome, which changed the course of teaching of biology and medicine in the 20th century, and introduced a central dogma of translation by transcription.  Nevertheless, there were significant inconsistencies and unanswered questions entering the twenty first century, accompanied by vast improvements in technical advances to clarify these issues. We have covered carbohydrate, protein, and lipid metabolism, which function in concert with the development of cellular structure, organ system development, and physiology.  To be sure, the progress in the study of the microscopic and particulate can’t be divorced from the observation of the whole.  We were left in the not so distant past with the impression of the Sufi story of the elephant and the three blind men, who one at a time held the tail, the trunk, and the ear, each proclaiming that it was the elephant.

I introduce here a story from the Brazilian biochemist, Jose

Eduardo des Salles Rosalino, on a formativr experience he had with the Nobelist, Luis Leloir.

Just at the beginning, when phosphorylation of proteins is presented, I assume you must mention that some proteins are activated by phosphorylation. This is fundamental in order to present self –organization reflex upon fast regulatory mechanisms. Even from an historical point of view. The first observation arrived from a sample due to be studied on the following day of glycogen synthetase. It was unintended left overnight out of the refrigerator. The result was it has changed from active form of the previous day to a non-active form. The story could have being finished here, if the researcher did not decide to spent this day increasing substrate levels (it could be a simple case of denaturation of proteins that changes its conformation despite the same order of amino acids). He kept on trying and found restoration of maximal activity. This assay was repeated with glycogen phosphorylase and the result was the opposite – it increases its activity. This led to the discovery

  • of cAMP activated protein kinase and
  • the assembly of a very complex system in the glycogen granule
  • that is not a simple carbohydrate polymer.

Instead, it has several proteins assembled and

  • preserves the capacity to receive from a single event (rise in cAMP)
  • two opposing signals with maximal efficiency,
  • stops glycogen synthesis,
  • as long as levels of glucose 6 phosphate are low
  • and increases glycogen phosphorylation as long as AMP levels are high).

I did everything I was able to do by the end of 1970 in order to repeat the assays with PK I, PKII and PKIII of M. Rouxii and using the Sutherland route to cAMP failed in this case. I then asked Leloir to suggest to my chief (SP) the idea of AA, AB, BB subunits as was observed in lactic dehydrogenase (tetramer) indicating this as his idea. The reason was my “chief”(SP) more than once, had said to me: “Leave these great ideas for the Houssay, Leloir etc…We must do our career with small things.” However, as she also had a faulty ability for recollection she also used to arrive some time later, with the very same idea but in that case, as her idea.
Leloir, said to me: I will not offer your interpretation to her as mine. I think it is not phosphorylation, however I think it is glycosylation that explains the changes in the isoenzymes with the same molecular weight preserved. This dialogue explains why during the reading and discussing “What is life” with him he asked me if as a biochemist in exile, talking to another biochemist, I expressed myself fully. I had considered that Schrödinger would not have confronted Darlington & Haldane because he was in U.K. in exile. This might explain why Leloir could have answered a bad telephone call from P. Boyer, Editor of The Enzymes, in a way that suggested that the pattern could be of covalent changes over a protein. Our FEBS and Eur J. Biochemistry papers on pyruvate kinase of M. Rouxii is wrongly quoted in this way on his review about pyruvate kinase of that year (1971).


Another aspect I think you must call attention to the following. Show in detail with different colors what carbons belongs to CoA, a huge molecule in comparison with the single two carbons of acetate that will produce the enormous jump in energy yield

  • in comparison with anaerobic glycolysis.

The idea is

  • how much must have been spent in DNA sequences to build that molecule in order to use only two atoms of carbon.

Very limited aspects of biology could be explained in this way. In case we follow an alternative way of thinking, it becomes clearer that proteins were made more stable by interaction with other molecules (great and small). Afterwards, it’s rather easy to understand how the stability of protein-RNA complexes where transmitted to RNA (vibrational +solvational reactivity stability pair of conformational energy).

Millions of years later, or as soon as, the information of interaction leading to activity and regulation could be found in RNA, proteins like reverse transcriptase move this information to a more stable form (DNA). In this way it is easier to understand the use of CoA to make two carbon molecules more reactive.

The discussions that follow are concerned with protein interactions and signaling.

Read Full Post »

Extracellular evaluation of intracellular flux in yeast cells

Larry H. Bernstein, MD, FCAP, Reviewer and Curator

Leaders in Pharmaceutical Intelligence

This is the fourth article in a series on metabolomics, which is a major development in -omics, integrating transcriptomics, proteomics,  genomics, metabolic pathways analysis, metabolic and genomic regulatory control using computational mapping.  In the previous two part presentation, flux analysis was not a topic for evaluation, but here it is the major focus.  It is a study of yeast cells, and bears some relationship to the comparison of glycemia, oxidative phosphorylation, TCA cycle, and ETC in leukemia cell lines.  In the previous study – system flux was beyond the scope of analysis, and explicitly stated.  The inferences made in comparing the two lymphocytic leukemia cells was of intracellular metabolism from extracellular measurements.  The study of yeast cells is aimed at looking at cellular effluxes, which is also an important method for studying pharmacological effects and drug resistance.

Metabolomic series

1.  Metabolomics, Metabonomics and Functional Nutrition: the next step in nutritional metabolism and biotherapeutics

2.  Metabolomic analysis of two leukemia cell lines. I

3.  Metabolomic analysis of two leukemia cell lines. II.

4.  Extracellular evaluation of intracellular flux in yeast cells

Q1. What is efflux?

Q2. What measurements were excluded from the previous study that would not allow inference about fluxes?

Q3. Would this study bear any relationship to the Pasteur effect?

Q4 What is a genome scale network reconstruction?

Q5 What type of information is required for a network prediction model?

Q6. Is there a difference between the metabolites profiles for yeast grown under aerobic and anaerobuc conditions – under the constrainsts?

Q7.  If there is a difference in the S metabolism, would there be an effect on ATP production?



Connecting extracellular metabolomic measurements to intracellular flux
states in yeast

Monica L Mo1Bernhard Ø Palsson1 and Markus J Herrgård12*

Author Affiliations

1 Department of Bioengineering, University of California, San Diego, La Jolla, CA 92093, USA

2 Current address: Synthetic Genomics, Inc, 11149 N Torrey Pines Rd, La Jolla, CA 92037, USA

For all author emails, please log on.

BMC Systems Biology 2009, 3:37  doi:10.1186/1752-0509-3-37


The electronic version of this article is the complete one and can be found online at:


Received: 15 December 2008
Accepted: 25 March 2009
Published: 25 March 2009

© 2009 Mo et al; licensee BioMed Central Ltd.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.



Metabolomics has emerged as a powerful tool in the

  • quantitative identification of physiological and disease-induced biological states.

Extracellular metabolome or metabolic profiling data, in particular,

  • can provide an insightful view of intracellular physiological states in a noninvasive manner.


We used an updated genome-scale

  • metabolic network model of Saccharomyces cerevisiae, iMM904, to investigate
  1. how changes in the extracellular metabolome can be used
  2. to study systemic changes in intracellular metabolic states.

The iMM904 metabolic network was reconstructed based on

  • an existing genome-scale network, iND750,
  • and includes 904 genes and 1,412 reactions.

The network model was first validated by

  • comparing 2,888 in silico single-gene deletion strain growth phenotype predictions
  • to published experimental data.

Extracellular metabolome data measured

  • of ammonium assimilation pathways 
  • in response to environmental and genetic perturbations

was then integrated with the iMM904 network

  • in the form of relative overflow secretion constraints and
  • a flux sampling approach was used to characterize candidate flux distributions allowed by these constraints.

Predicted intracellular flux changes were

  • consistent with published measurements
  • on intracellular metabolite levels and fluxes.

Patterns of predicted intracellular flux changes

  • could also be used to correctly identify the regions of
  • the metabolic network that were perturbed.


Our results indicate that

  • integrating quantitative extracellular metabolomic profiles
  • in a constraint-based framework
  • enables inferring changes in intracellular metabolic flux states.

Similar methods could potentially be applied

  • towards analyzing biofluid metabolome variations
  • related to human physiological and disease states.


“Omics” technologies are rapidly generating high amounts of data

  • at varying levels of biological detail.

In addition, there is a rapidly growing literature and

  • accompanying databases that compile this information.

This has provided the basis for the assembly of

  • genome-scale metabolic networks for various microbial and eukaryotic organisms [111].

These network reconstructions serve

  • as manually curated knowledge bases of
  • biological information as well as
  • mathematical representations of biochemical components and
  • interactions specific to each organism.

genome-scale network reconstruction is

  • structured collection of genes, proteins, biochemical reactions, and metabolites
  • determined to exist and operate within a particular organism.

This network can be converted into a predictive model

  • that enables in silico simulations of allowable network states based on
  • governing physico-chemical and genetic constraints [12,13].

A wide range of constraint-based methods have been developed and applied

  • to analyze network metabolic capabilities under
  • different environmental and genetic conditions [13].

These methods have been extensively used to

  • study genome-scale metabolic networks and have successfully predicted, for example,
  1. optimal metabolic states,
  2. gene deletion lethality, and
  3. adaptive evolutionary endpoints [1416].

Most of these applications utilize

  • optimization-based methods such as flux balance analysis (FBA)
  • to explore the metabolic flux space.

However, the behavior of genome-scale metabolic networks can also be studied

  • using unbiased approaches such as
  • uniform random sampling of steady-state flux distributions [17].

Instead of identifying a single optimal flux distribution based on

  • a given optimization criterion (e.g. biomass production),

these methods allow statistical analysis of

  • a large range of possible alternative flux solutions determined by
  • constraints imposed on the network.

Sampling methods have been previously used to study

  1. global organization of E. coli metabolism [18] as well as
  2. to identify candidate disease states in the cardiomyocyte mitochondria [19].

Network reconstructions provide a structured framework

  • to systematically integrate and analyze disparate datasets
  • including transcriptomic, proteomic, metabolomic, and fluxomic data.

Metabolomic data is one of the more relevant data types for this type of analysis as

  1. network reconstructions define the biochemical links between metabolites, and
  2. recent advancements in analytical technologies have allowed increasingly comprehensive
  • intracellular and extracellular metabolite level measurements [20,21].

The metabolome is

  1. the set of metabolites present under a given physiological condition
  2. at a particular time and is the culminating phenotype resulting from
  • various “upstream” control mechanisms of metabolic processes.

Of particular interest to this present study are

  • the quantitative profiles of metabolites that are secreted into the extracellular environment
  • by cells under different conditions.

Recent advances in profiling the extracellular metabolome (EM) have allowed

  • obtaining insightful biological information on cellular metabolism
  • without disrupting the cell itself.

This information can be obtained through various

  • analytical detection,
  • identification, and
  • quantization techniques

for a variety of systems ranging from

  • unicellular model organisms to human biofluids [2023].

Metabolite secretion by a cell reflects its internal metabolic state, and

  • its composition varies in response to
  • genetic or experimental perturbations
  • due to changes in intracellular pathway activities
  • involved in the production and utilization of extracellular metabolites [21].

Variations in metabolic fluxes can be reflected in EM changes which can

  • provide insight into the intracellular pathway activities related to metabolite secretion.

The extracellular metabolomic approach has already shown promise

  • in a variety of applications, including
  1. capturing detailed metabolite biomarker variations related to disease and
  2. drug-induced states and
  3. characterizing gene functions in yeast [2427].

However, interpreting changes in the extracellular metabolome can be challenging

  • due to the indirect relationship between the proximal cause of the change
    (e.g. a mutation)
  • and metabolite secretion.

Since metabolic networks describe

  • mechanistic,
  • biochemical links between metabolites,

integrating such data can allow a systematic approach

  • to identifying altered pathways linked to
  • quantitative changes in secretion profiles.

Measured secretion rates of major byproduct metabolites

  • can be applied as additional exchange flux constraints
  • that define observed metabolic behavior.

For example, a recent study integrating small-scale EM data

  • with a genome-scale yeast model
  • correctly predicted oxygen consumption and ethanol production capacities
  • in mutant strains with respiratory deficiencies [28].

The respiratory deficient mutant study

  • used high accuracy measurements for a small number of
  • major byproduct secretion rates
  • together with an optimization-based method well suited for such data.

Here, we expand the application range of the model-based method used in [28]

  • to extracellular metabolome profiles,
  • which represent a temporal snapshot of the relative abundance
  • for a larger number of secreted metabolites.

Our approach is complementary to

  • statistical (i.e. “top-down”) approaches to metabolome analysis [29]
  • and can potentially be used in applications such as biofluid-based diagnostics or
  • large-scale characterization of mutants strains using metabolite profiles.

This study implements a constraint-based sampling approach on

  • an updated genome-scale network of yeast metabolism
  • to systematically determine how EM level variations

are linked to global changes in intracellular metabolic flux states.

By using a sampling-based network approach and statistical methods (Figure 1),

  • EM changes were linked to systemic intracellular flux perturbations
    in an unbiased manner
  • without relying on defining single optimal flux distributions
  • used in the previously mentioned study [28].

The inferred perturbations in intracellular reaction fluxes were further analyzed

  • using reporter metabolite and subsystem (i.e., metabolic pathway) approaches [30]
  • in order to identify dominant metabolic features that are collectively perturbed (Figure 2).

The sampling-based approach also has the additional benefit of

  • being less sensitive to inaccuracies in metabolite secretion profiles than
  • optimization-based methods and can effectively be used – in biofluid metabolome analysis.
integration of exometabolomic (EM) data

integration of exometabolomic (EM) data

Figure 1. Schematic illustrating the integration of exometabolomic (EM) data with the constraint-based framework.

(A) Cells are subjected to genetic and/or environmental perturbations to secrete metabolite patterns unique to that condition.
(B) EM is detected, identified, and quantified.
(C) EM data is integrated as required secretion flux constraints to define allowable solution space.
(D) Random sampling of solution space yields the range of feasible flux distributions for intracellular reactions.
(E) Sampled fluxes were compared to sampled fluxes of another condition to determine

  • which metabolic regions were altered between the two conditions (see Figure 2).

(F) Significantly altered metabolic regions were identified.


sampling and scoring analysis to determine intracellular flux changes

sampling and scoring analysis to determine intracellular flux changes

Figure 2. Schematic of sampling and scoring analysis to determine intracellular flux changes.

(A) Reaction fluxes are sampled for two conditions.
(B & C) Sample of flux differences is calculated by selecting random flux values from each condition

  • to obtain a distribution of flux differences for each reaction.

(D) Standardized reaction Z-scores are determined, which represent

  • how far the sampled flux differences deviates from a zero flux change.

Reaction scores can be used in

  1. visualizing perturbation subnetworks and
  2. analyzing reporter metabolites and subsystems.

This study was divided into two parts and describes:

(i) the reconstruction and validation of an expanded S. cerevisiae metabolic network, iMM904; and
(ii) the systematic inference of intracellular metabolic states from

  • two yeast EM data sets using a constraint-based sampling approach.

The first EM data set compares wild type yeast to the gdh1/GDH2 (glutamate dehydrogenase) strain [31],

  • which indicated good agreement between predicted metabolic changes
  • of intracellular metabolite levels and fluxes [31,32].

The second EM data set focused on secreted amino acid measurements

  • from a separate study of yeast cultured in different
    ammonium and potassium concentrations [33].

We analyzed the EM data to gain further insight into

  • perturbed ammonium assimilation processes as well as
  1. metabolic states relating potassium limitation and
  2. ammonium excess conditions to one another.

The model-based analysis of both

  • separately published extracellular metabolome datasets
  • suggests a relationship between
  1. glutamate,
  2. threonine and
  3. folate metabolism,
  • which are collectively perturbed when
    ammonium assimilation processes are broadly disrupted
  1. either by environmental (excess ammonia) or
  2. genetic (gene deletion/overexpression) perturbations.

The methods herein present an approach to

  • interpreting extracellular metabolome data and
  • associating these measured secreted metabolite variations
  • to changes in intracellular metabolic network states.

Additional file 1. iMM904 network content.

The data provided represent the content description of the iMM904 metabolic network and
detailed information on the expanded content.

Format: XLS Size: 2.7MB Download file

This file can be viewed with: Microsoft Excel Viewer

Additional file 2. iMM904 model files.

The data provided are the model text files of the iMM904 metabolic network
that is compatible with the available COBRA Toolbox [13]. The model structure
can be loaded into Matlab using the ‘SimPhenyPlus’ format with GPR and compound information.

Format: ZIP Size: 163KB Download file

Conversion of the network to a predictive model

The network reconstruction was converted to a constraint-based model using established procedures [13].

Network reactions and metabolites were assembled into a stoichiometric matrix 

  • containing the stoichiometric coefficients of the reactions in the network.

The steady-state solution space containing possible flux distributions

  • is determined by calculating the null space of S= 0,

where is the reaction flux vector.

Minimal media conditions were set through constraints on exchange fluxes

  • corresponding to the experimental measured substrate uptake rates.

All the model-based calculations were done using the Matlab COBRA Toolbox [13]

  • utilizing the glpk or Tomlab/CPLEX (Tomopt, Inc.) optimization solvers.

Chemostat growth simulations

The iMM904 model was initially validated by

  1. simulating wild type yeast growth in aerobic and anaerobic
    carbon-limited chemostat conditions
  2. and comparing the simulation results to published experimental data

on substrate uptake and byproduct secretion in these conditions [34].

The study was performed following the approach taken to validate the iFF708 model in a previous study [35].

The predicted glucose uptake rates were determined

  1. by setting the in silico growth rate to the measured dilution rate,
    – equivalent under continuous culture growth,
  2. and minimizing the glucose uptake rate.

The accuracy of in silico predictions of

  • substrate uptake and byproduct secretion by the iMM904 model
  • was similar to the accuracy obtained using the iFF708 model
  • and results are shown in Figure S1 [see Additional file 3].

Additional file 3. Supplemental figures. 

The file provides the supplemental figures and descriptions of S1, S2, S3, and S4.

Format: PDF Size: 513KB Download file

This file can be viewed with: Adobe Acrobat Reader

Genome-scale gene deletion phenotype predictions

The iMM904 network was further validated by

  • performing genome-scale gene lethality computations
  • following established procedures to determine growth phenotypes
  1. under minimal medium conditions and
  2. compared to published data.

A modified version of the biomass function used in previous iND750 studies

  1. was set as the objective to be maximized and
  2. gene deletions were simulated by

setting the flux through the corresponding reaction(s) to zero.

The biomass function was based on the experimentally measured

  1. composition of major cellular constituents
  2. during exponential growth of yeast cells and
  3. was reformulated to include trace amounts of
  4. additional cofactors and metabolites
  5. with the assumed fractional contribution of 10-.

These additional biomass compounds were included

according to the biomass formulation used in the iLL672 study

  • to improve lethality predictions through
  • the inclusion of additional essential biomass components [3].

The model was constrained by limiting

  1. the carbon source uptake to 10 mmol/h/gDW
  2. and oxygen uptake to 2 mmol/h/gDW.

Ammonia, phosphate, and sulfate were assumed to be non-limiting.

The experimental phenotyping data was obtained

  • using strains that were auxotrophic for
  1. methionine,
  2. leucine,
  3. histidine, and
  4. uracil.

These auxotrophies were simulated

  1. by deleting the appropriate genes from the model and
  2. supplementing the in silico strain with the appropriate supplements
  3. at non-limiting, but low levels.

Furthermore, trace amounts of essential nutrients that are present

  • in the experimental minimal media formulation
  1. 4-aminobenzoate,
  2. biotin,
  3. inositol,
  4. nicotinate,
  5. panthothenate,
  6. thiamin)
  • were supplied in the simulations [3].

Three distinct methods to simulate the outcome of gene deletions were utilized:

  1. Flux-balance analysis (FBA) [36-38],
  2. Minimization of Metabolic Adjustment (MoMA) [39], and
  3. a linear version of MoMA (linearMoMA).

In the linearMoMA method, minimization of the quadratic objective function
of the original MoMA algorithm

  • was replaced by minimization of the corresponding 1-norm objective function
    (i.e. sum of the absolute values of the differences of wild type FBA solution
    and the knockout strain flux solution).

The computed results were then compared to growth phenotype data
(viable/lethal) from a previously published experimental gene deletion study [3].

The comparison between experimental and in silico deletion phenotypes involved

  • choosing a threshold for the predicted relative growth rate of
  • a deletion strain that is considered to be viable.

We used standard ROC curve analysis

  • to assess the accuracy of different prediction methods and models
  • across the full range of the viability threshold parameter,
    results shown in Figure S2 [see Additional file 3].

The ROC curve plots the true viable rate against the false viable rate

  • allowing comparison of different models and methods
  • without requiring arbitrarily choosing this parameter a priori [40].

The optimal prediction performance corresponds to

  • the point closest to the top left corner of the ROC plot
    (i.e. 100% true viable rate, 0% false viable rate).

Table 1

Table 1 Comparison of iMM904 and iLL672 gene deletion predictions and experimental data under minimal media conditions
Media Model Method True viable False viable False lethal True lethal True viable % False viable % MCC
Glucose iMM904 full FBA 647 10 32 33 95.29 23.26 0.6
iMM904 full linMOMA 644 10 35 33 94.85 23.26 0.58
iMM904 full MOMA 644 10 35 33 94.85 23.26 0.58
iMM904 red FBA 440 9 28 33 94.02 21.43 0.61
iMM904 red linMOMA 437 9 31 33 93.38 21.43 0.6
iMM904 red MOMA 437 9 31 33 93.38 21.43 0.6
iLL672 full MOMA 433 9 35 33 92.52 21.43 0.57
Galactose iMM904 full FBA 595 32 36 59 94.29 35.16 0.58
iMM904 full linMOMA 595 32 36 59 94.29 35.16 0.58
iMM904 full MOMA 595 32 36 59 94.29 35.16 0.58
iMM904 red FBA 409 12 33 56 92.53 17.65 0.67
iMM904 red linMOMA 409 12 33 56 92.53 17.65 0.67
iMM904 red MOMA 409 12 33 56 92.53 17.65 0.67
iLL672 full MOMA 411 19 31 49 92.99 27.94 0.61
Glycerol iMM904 full FBA 596 43 36 47 94.3 47.78 0.48
iMM904 full linMOMA 595 44 37 46 94.15 48.89 0.47
iMM904 full MOMA 598 44 34 46 94.62 48.89 0.48
iMM904 red FBA 410 20 34 46 92.34 30.3 0.57
iMM904 red linMOMA 409 21 35 45 92.12 31.82 0.56
iMM904 red MOMA 412 21 32 45 92.79 31.82 0.57
iLL672 full MOMA 406 20 38 46 91.44 30.3 0.55
Ethanol iMM904 full FBA 593 45 29 55 95.34 45 0.54
iMM904 full linMOMA 592 45 30 55 95.18 45 0.54
iMM904 full MOMA 592 44 30 56 95.18 44 0.55
iMM904 red FBA 408 21 27 54 93.79 28 0.64
iMM904 red linMOMA 407 21 28 54 93.56 28 0.63
iMM904 red MOMA 407 20 28 55 93.56 26.67 0.64
iLL672 full MOMA 401 13 34 62 92.18 17.33 0.68
MCC, Matthews correlation coefficient (see Methods). Note that the iLL672 predictions were obtained directly from [3] and thus the viability threshold was not optimized using the maximum MCC approach.
Mo et al. BMC Systems Biology 2009 3:37


The values reported in Table 1 correspond to selecting

  • the optimal viability threshold based on this criterion.

We summarized the overall prediction accuracy of a model and method

  • using the Matthews Correlation Coefficient (MCC) [40].

The MCC ranges from -1 (all predictions incorrect) to +1 (all predictions correct) and

  • is suitable for summarizing overall prediction performance

in our case where there are substantially more viable than lethal gene deletions.

ROC plots were produced in Matlab (Mathworks, Inc.).


Table 1. Comparison of iMM904 and iLL672

  • gene deletion predictions and
  • experimental data

Inferring perturbed metabolic regions based on EM profiles

The method implemented in this study is shown schematically in Figures 1 and 2

Constraining the iMM904 network 

Relative levels of quantitative EM data were incorporated into the constraint-based framework

  • as overflow secretion exchange fluxes to simulate the required low-level production of
  • experimentally observed excreted metabolites.

The primary objective of this study is to associate

  • relative metabolite levels that are generally measured for metabonomic or biofluid analyses
  • to the quantitative ranges of intracellular reaction fluxes required to produce them.

However, without detailed kinetic information or dynamic metabolite measurements available,

  • we approximated EM datasets of relative quantitative metabolite levels
  • to be proportional to the rate in which they are secreted and detected
  • (at a steady state) – into the extracellular media.

This approach is analogous to approximating uptake rates based

  • on metabolite concentrations from a previous study performing sampling analysis
  • on a cardiomyocyte mitochondrial network
  • to identify differential flux distribution ranges

for various environmental (i.e. substrate uptake) conditions [19].

The raw data was normalized by the raw maximum value of the dataset
(thus the maximum secretion flux was 1 mmol/hr/gDW) with

  • an assumed error of 10%
  • to set the lower and upper bounds and thus
  • inherently accounting for sampling calculation sensitivity.

The gdh1/GDH2 strains were flask cultured under minimal glucose media conditions; thus,

  • glucose and oxygen uptake rates were set at 15 and 2 mmol/hr/gDW, respectively,
  • for the gdh1/GDH2 strain study.

In the anaerobic case the oxygen uptake rate was set to zero, and

  • sterols and fatty acids were provided as in silico supplements as described in [35].

For the potassium limitation/ammonium toxicity study

  • the growth rate was set at 0.17 1/h, and
  • the glucose uptake rate was minimized
  • to mimic experimental chemostat cultivation conditions.

These input constraints were constant for each perturbation and comparative wild-type condition

  • such that the calculated solution spaces between the conditions
  • differed based only on variations in the output secretion constraints.

FBA optimization of EM-constrained networks

A modified FBA method with minimization of the 1-norm objective function

  • between two optimal flux distributions was used
  • to determine optimal intracellular fluxes
  • based on the EM-constrained metabolic models.

This method determines two optimal flux distributions simultaneously

  • for two differently constrained models (e.g. wild type vs. mutant) –
  • these flux distributions maximize biomass production in each case and
  • the 1-norm distance between the distributions is as small as possible
  • given the two sets of constraints.

This approach avoids problems with

  • alternative optimal solutions when comparing two FBA-computed flux distributions
  • by assuming minimal rerouting of flux distibution between a perturbed network and its reference network.

Reaction flux changes from the FBA optimization results were determined

  • by computing the relative percentage fold change for each reaction
  • between the mutant and wild-type flux distributions.

Random sampling of the steady-state solution space

We utilized artificial centering hit-and-run (ACHR) Monte Carlo sampling [19,41]

  • to uniformly sample the metabolic flux solution space
  • defined by the constraints described above.

Reactions, and their participating metabolites, found to participate in intracellular loops [42]

  • were discarded from further analysis as these reactions can have arbitrary flux values.

The following sections describe the approaches used for the analysis of the different datasets.

Sampling approach used in the gdh1/GDH2 study

Due to the overall shape of the metabolic flux solution space,

  • most of the sampled flux distributions resided close to the minimally allowed growth rate
    (i.e. biomass production) and
  • corresponded to various futile cycles that utilized substrates but
  • did not produce significant biomass.

In order to study more physiologically relevant portions of the flux space

  • we restricted the sampling to the part of the solution space
  • where the growth rate was at least 50% of the maximum growth rate
  • for the condition as determined by FBA.

This assumes that cellular growth remains an important overall objective by the yeast cells

  • even in batch cultivation conditions, but
  • that the intracellular flux distributions
  • may not correspond to maximum biomass production [43].

To test the sensitivity of the results to the minimum growth rate threshold,

  • separate Monte Carlo samples were created for each minimum threshold
  • ranging from 50% to 100% at 5% increments.

We also tested the sensitivity of the results

  • to the relative magnitude of the extracellular metabolite secretion rates
  • by performing the sampling at three different relative levels

(0 corresponding to no extracellular metabolite secretion, maximum rate of 0.5 mmol/hr/gDW,
and maximum rate of 1.0 mmol/hr/gDW).

For each minimum growth rate threshold and extracellular metabolite secretion rate,

  • the ACHR sampler was run for 5 million steps and
  • a flux distribution was stored every 5000 steps.

The sensitivity analysis results are presented in Figures S3 and S4 [see Additional File 3], and

  • the results indicate that the reaction Z-scores (see below) are not significantly affected by
  1. either the portion of the solution space sampled or
  2. the exact scaling of secretion rates.

The final overall sample used was created by combining the samples for all minimum growth rate thresholds

  • for the highest extracellular metabolite secretion rate (maximum 1 mmol/hr/gDW).

This approach allowed biasing the sampling towards

  • physiologically relevant parts of the solution space
  • without imposing the requirement of strictly maximizing a predetermined objective function.

The samples obtained with no EM data were used as control samples

  • to filter reporter metabolites/subsystems whose scores were significantly high
  • due to only random differences between sampling runs.

Sampling approach used in the potassium limitation/ammonium toxicity study

Since the experimental data used in this study was generated in chemostat conditions, and

  • previous studies have indicated that chemostat flux patterns predicted by FBA are
  • close to the experimentally measured ones [43],
  • we assumed that sampling of the optimal solution space was appropriate for this study.

In order to sample a physiologically reasonable range of flux distributions,

  • samples for four different oxygen uptake rates
    (1, 2, 3, and 4 mmol/hr/gDW with 5 million steps each)
  • were combined in the final analysis.

Standardized scoring of flux differences between perturbation and control conditions

Z-score based approach was implemented to quantify differences in flux samples between two conditions (Figure 2).
First, two flux vectors were chosen randomly,

  • one from each of the two samples to be compared and
  • the difference between the flux vectors was computed.

This approach was repeated to create a sample of 10,000 (n) flux difference vectors

  • for each pair of conditions considered (e.g. mutant or perturbed environment vs. wild type).

Based on this flux difference sample, the sample mean (μdiff,i) and standard deviation (σdiff,i)

  • between the two conditions was calculated for each reaction i. The reaction Z-score was calculated as:


reaction Z-score

reaction Z-score

which describes the sampled mean difference deviation

  • from a population mean change of zero (i.e. no flux difference
    between perturbation and wild type).

Note that this approach allows accounting for uncertainty in the

  • flux distributions inferred based on the extracellular metabolite secretion constraints.

This is in contrast to approaches such as FBA or MoMA that would predict

  • a single flux distribution for each condition and thus potentially
  • overestimate differences between conditions.

The reaction Z-scores can then be further used in analysis

  • to identify significantly perturbed regions of the metabolic network
  • based on reporter metabolite [44] or subsystem [30] Z-scores.

These reporter regions indicate, or “report”, dominant perturbation features

  • at the metabolite and pathway levels for a particular condition.

The reporter metabolite Z-score for any metabolite can be derived from the reaction Z-scores

  • of the reactions consuming or producing j (set of reactions denoted as Rj) as:


reporter z-score for any metabolite j

reporter z-score for any metabolite j

where Nis the number of reactions in Rand mmet,is calculated as


distributional correction for m_met,j SQRT

distributional correction for m_met,j SQRT

To account and correct for background distribution, the metabolite Z-score was normalized

  • by computing μmet,Nj and σmet,,Nj corresponding to the mean mmet and
  • its standard deviation for 1,000 randomly generated reaction sets of size Nj.

Z-scores for subsystems were calculated similarly by considering the set of reactions R

  • that belongs to each subsystem k.

Hence, positive metabolite and subsystem scores indicate a significantly perturbed metabolic region

  • relative to other regions, whereas
  • a negative score indicate regions that are not perturbed
  • more significantly than what is expected by random chance.

Perturbation subnetworks of reactions and connecting metabolites were visualized using Cytoscape [45].

Results and discussion

  1. Reconstruction and validation of iMM904 network iMM904 network content 

A previously reconstructed S. cerevisiae network, iND750,

  • was used as the basis for the construction of the expanded iMM904 network.
  • Prior to its presentation here, the
    iMM904 network content was the basis for a consensus jamboree network that was recently published
  • but has not yet been adapted for FBA calculations [46].

The majority of iND750 content was carried over and

  • further expanded on to construct iMM904, which accounts for
  1. 904 genes,
  2. 1,228 individual metabolites, and
  3. 1,412 reactions of which
  •                       395 are transport reactions.

Both the number of gene-associated reactions and the number of metabolites

  • increased in iMM904 compared with the iND750 network.

Additional genes and reactions included in the network primarily expanded the

  • lipid,
  • transport, and
  • carbohydrate subsystems.

The lipid subsystem includes

  • new genes and
  • reactions involving the degradation of sphingolipids and glycerolipids.

Sterol metabolism was also expanded to include

  • the formation and degradation of steryl esters, the
  •                      storage form of sterols.

The majority of the new transport reactions were added

  • to connect network gaps between intracellular compartments
  • to enable the completion of known physiological functions.

We also added a number of new secretion pathways

  • based on experimentally observed secreted metabolites [31].

A number of gene-protein-reaction (GPR) relationships were modified

  • to include additional gene products that are required to catalyze a reaction.

For example, the protein compounds

  • thioredoxin and
  • ferricytochrome C

were explicitly represented as compounds in iND750 reactions, but

  • the genes encoding these proteins were not associated with their corresponding GPRs.

Other examples include glycogenin and NADPH cytochrome p450 reductases (CPRs),

  1. which are required in the assembly of glycogen and
  2. to sustain catalytic activity in cytochromes p450, respectively.

These additional proteins were included in iMM904 as

  • part of protein complexes to provide a more complete
  • representation of the genes and
  • their corresponding products necessary for a catalytic activity to occur.

Major modifications to existing reactions were in cofactor biosynthesis, namely in

  • quinone,
  • beta-alanine, and
  • riboflavin biosynthetic pathways.

Reactions from previous S. cerevisiae networks associated with

  • quinone,
  • beta-alanine, and
  • riboflavin biosynthetic pathways

were essentially inferred from known reaction mechanisms based on

  • reactions in previous network reconstructions of E. coli [2,47].

These pathways were manually reviewed

  • based on current literature and subsequently replaced by
  • reactions and metabolites specific to yeast.

Additional changes in other subsystems were also made, such as

  1. changes to the compartmental location of a gene and
  2. its corresponding reaction(s),
  3. changes in reaction reversibility and cofactor specificity, and
  4. the elucidation of particular transport mechanisms.

A comprehensive listing of iMM904 network contents as well as

  • a detailed list of changes between iND750 and iMM904 is included
    [see Additional file 1].

Predicting deletion growth phenotypes

The updated genome-scale iMM904 metabolic network was validated

  • by comparing in silico single-gene deletion predictions to
  • in vivo results from a previous study used
  • to analyze another S. cerevisiae metabolic model, iLL672 [3].

This network was constructed based on the iFF708 network [22],

  • which was also the starting point for
  • reconstructing the iND750 network [2].

The experimental data used to validate the iLL672 model consisted of

3,360 single-gene knockout strain phenotypes evaluated

  • under minimal media growth conditions with
  1. glucose,
  2. galactose,
  3. glycerol, and
  4. ethanol

as sole carbon sources. Growth phenotypes for the iMM904 network were predictedusing

  1. FBA [3234],
  2. MoMA [35], and
  3. linear MoMA methods

as described in Methods and subsequently compared to the experimental data (Table 1).

Each deleted gene growth prediction comparison was classified as

  1. true lethal,
  2. true viable,
  3. false lethal, or
  4. false viable.

The growth rate threshold for considering a prediction viable was chosen

  • for each condition and method separately
  • to optimize the tradeoff between true viable and false viable predictions
    (maximum Matthews correlation coefficient, see Methods).

Since iMM904 has 212 more genes than iLL672 with experimental data, we also present results

  • for the subset of iMM904 predictions with genes included in iLL672 (reduced iMM904 set).

When the same gene sets are compared, iMM904 improves gene lethality predictions under

  • glucose,
  • galactose, and
  • glycerol conditions

over iLL672 somewhat, but is less accurate

  • at predicting growth phenotypes under the ethanol condition.

It should be noted that the iLL672 predictions were obtained directly from [3]

  • thus the growth rate threshold was not optimized similarly to iMM904 predictions.

Overall, when viability cutoff is chosen

  • as indicated above for each method separately,
  • the three prediction methods perform similarly
  1. FBA,
  2. MOMA, and
  3. linear MOMA) .

While the full gene complement in iMM904 greatly increased

  • the number of true viable predictions,
  • the full model also made significantly more false viable predictions
  • compared with reduced iMM904 and iLL672 predictions.

However, it is important to note that 143 reactions involved in dead-end biosynthetic pathways were actually

  • removed from iFF708 to build the iLL672 reconstruction [3].

These dead-ends are considered “knowledge gaps” in pathways

  • that have not been fully characterized and, as a result,
  • lead to false viable predictions when determining gene essentiality
  • if the pathway is in fact required for growth under a certain condition [2,26].

As more of these pathways are elucidated and

  • included in the model to
  • fill in existing network gaps,
  • we can expect false viable prediction rates to consequently decrease.

Thus, while a larger network has a temporarily reduced capacity to accurately predict gene deletion phenotypes,

  • it captures a more complete picture of currently known metabolic functions and
  • provides a framework for network expansion as new pathways are elucidated [48].


Inferring intracellular perturbation states from metabolic profiles – Aerobic and anaerobic gdh1/GDH2 mutant behavior

The gdh1/GDH2 mutant strain was previously developed [49,50]

  • to lower NADPH consumption in ammonia assimilation, which would
  • favor the NADPH-dependent fermentation of xylose.

In this strain, the NADPH-dependent glutamate dehydrogenase, Gdh1, was

  • deleted and the NADH-dependent form of the enzyme, Gdh2,
  •                     was overexpressed.

The net effect is to allow efficient assimilation of ammonia

  • into glutamate using NADH instead of NADPH as a cofactor.

While growth characteristics remained unaffected,

  • relative quantities of secreted metabolites differed between the wild-type and mutant strain
  • under aerobic and anaerobic conditions.

We analyzed EM data for the gdh1/GDH2 and wild-type strains reported

  • in [31] under aerobic and anaerobic conditions separately using
  • both FBA optimization and
  • sampling-based approaches as described in Methods.

43 measured extracellular and intracellular metabolites from the original dataset [31],

  • primarily of central carbon and amino acid metabolism,
  • were explicitly represented in the iMM904 network [see Additional file 4].

Extracellular metabolite levels were used

  • to formulate secretion constraints and
  • differential intracellular metabolites were used
  • to compare and validate the intracellular flux predictions.

Perturbed reactions from the FBA results were

  • determined by calculating relative flux changes, and
  • reaction Z-scores were calculated from the sampling analysis
  • to quantify flux changes between the mutant and wild-type strains,
  • with Z reaction > 1.96 corresponding to a two-tailed p-value < 0.05 and
  • considered to be significantly perturbed [see Additional file 4].

Additional file 4. Gdh mutant aerobic and anaerobic analysis results. 

The data provided are the full results for the exometabolomic analysis of aerobic and anerobic gdh1/GDH2 mutant.

Format: XLS Size: 669KB Download file

This file can be viewed with: Microsoft Excel Viewer

To validate the predicted results, reaction flux changes from both FBA and sampling methods were compared to differential intracellular metabolite level data measured from the same study. Intracellular metabolites involved in highly perturbed reactions (i.e. reactants and products) predicted from FBA and sampling analyses were identified and
compared to metabolites that were experimentally identified as significantly changed (< 0.05) between mutant and wild-type. Statistical measures of recall, accuracy, and
precision were calculated and represent the predictive sensitivity, exactness, and reproducibility respectively. From the sampling analysis, a considerably larger number of
significantly perturbed reactions are predicted in the anaerobic case (505 reactions, or 70.7% of active reactions) than in aerobic (394 reactions, or 49.8% of active reactions). The top percentile of FBA flux changes equivalent to the percentage of significantly perturbed sampling reactions were compared to the intracellular data. Results from both analyses are summarized in Table 2. Sampling predictions were considerably higher in recall than FBA predictions for both conditions, with respective ranges of 0.83–1
compared to 0.48–0.96. Accuracy was also higher in sampling predictions; however, precision was slightly better in the FBA predictions as expected due to the smaller
number of predicted changes. Overall, the sampling predictions of perturbed intracellular metabolites are strongly consistent with the experimental data and significantly
outperforms that of FBA optimization predictions in accurately predicting differential metabolites involved in perturbed intracellular fluxes.

Table 2. Statistical comparison of the differential intracellular metabolite data set (< 0.05) with metabolites involved in perturbed reactions predicted by FBA optimization and sampling analyses for aerobic and anaerobic gdh1/GDH2 mutant.


Table 2 Statistical comparison of the differential intracellular metabolite data set (p < 0.05)
with metabolites involved in perturbed reactions predicted by FBA optimization and
sampling analyses for aerobic and anaerobic gdh1/GDH2 mutant.
                           Aerobic                         Anaerobic                             Overall
FBA Sampling FBA Sampling FBA
Recall 0.48 0.83 0.96 1 0.71 0.91
Accuracy 0.55 0.62 0.64 0.64 0.6 0.63
Precision 0.78 0.69 0.64 0.63 0.68 0.66
Overall statistics indicate combined results of both conditions.
Mo et al. BMC Systems Biology 2009 3:37

Figure 3.
 Perturbation reaction subnetwork of gdh1/GDH2 mutant under aerobic conditions.

The network illustrates a simplified subset of highly perturbedPerturbation subnetworks can be drawn to visualize predicted significantly perturbed intracellular reactions and illustrate their connection to the observed secreted metabolites in the aerobic and anaerobic gdh1/GDH2 mutants.

Perturbation reaction subnetwork of gdh1.GDH2 mutant under aerobic conditions.

Perturbation reaction subnetwork of gdh1.GDH2 mutant under aerobic conditions.

Figure 3 shows an example of a simplified aerobic perturbation subnetwork consisting primarily of proximal pathways connected directly to a subset of major secreted

  • glutamate,
  • proline,
  • D-lactate, and
  • 2-hydroxybuturate.

Figure 4 displays anaerobic reactions with Z-scores of similar magnitude to the perturbed reactions in Figure 3. The same subset of metabolites is also present in the
larger anaerobic perturbation network and indicates that the NADPH/NADH balance perturbation induced by the gdh1/GDH2 manipulation has widespread effects
beyond just altering glutamate metabolism anaerobically.

Interestingly, it is clear that the majority of the secreted metabolite pathways involve connected perturbed reactions that broadly converge on glutamate.

Note that Figures 3 and 4 only show the subnetworks that consisted of two or more connected reactions  for a number of secreted metabolites no contiguous perturbed pathway could be identified by the sampling approach. This indicates that the secreted metabolite pattern alone is not sufficient to determine which specific
production and secretion pathways are used by the cell for these metabolites.

Reactions connected to aerobically-secreted metabolites predicted from the sampling analysis of the gdh1/GDH2 mutant strain.
The major secreted metabolites

  • glutamate,
  • proline,
  • D-lactate, and
  • 2-hydroxybuturate

were also detected in the anaerobic condition. Metabolite abbreviations are found in Additional file 1.

Figure 4.

Perturbation reaction subnetwork of gdh1/GDH2 mutant under anaerobic conditions.

Perturbation reaction subnetwork of gdh1.GDH2 mutant under anaerobic conditions

Perturbation reaction subnetwork of gdh1.GDH2 mutant under anaerobic conditions

Subnetwork illustrates the highly perturbed anaerobic reactions of similar Z-reaction magnitude to the reactions in Figure 3.

A significantly larger number of reactions indicates mutant metabolic effects are more widespread in the anaerobic environment.
The network shows that perturbed pathways converge on glutamate, the main site in which the gdh1/GDH2 modification was introduced, which
suggests that the direct genetic perturbation effects are amplified under this environment. Metabolite abbreviations are found in Additional file 1.

To further highlight metabolic regions that have been systemically affected by the gdh1/GDH2 modification, reporter metabolite and subsystem methods [30] were used to
summarize reaction scores around specific metabolites and in specific metabolic subsystems. The top ten significant scores for metabolites/subsystems associated with more
than three reactions are summarized in Tables 3 (aerobic) and 4 (anaerobic), with Z > 1.64 corresponding to < 0.05 for a one-tailed distribution. Full data for all reactions,
reporter metabolites, and reporter subsystems is included [see Additional file 4].

Table 3. List of the top ten significant reporter metabolite and subsystem scores for the gdh1/GDH2 vs. wild type comparison in aerobic conditions.

Table 3
List of the top ten significant reporter metabolite and subsystem scores for the gdh1/GDH2 vs. wild type comparison in aerobic conditions.
Reporter metabolite Z-score No of reactions*
L-proline [c] 2.71 4
Carbon dioxide [m] 2.51 15
Proton [m] 2.19 51
Glyceraldehyde 3-phosphate [c] 1.93 7
Ubiquinone-6 [m] 1.82 5
Ubiquinol-6 [m] 1.82 5
Ribulose-5-phosphate [c] 1.8 4
Uracil [c] 1.74 4
L-homoserine [c] 1.72 4
Alpha-ketoglutarate [m] 1.71 8
Reporter subsystem Z-score No of reactions
Citric Acid Cycle 4.58 7
Pentose Phosphate Pathway 3.29 12
Glycine and Serine Metabolism 2.69 17
Alanine and Aspartate Metabolism 2.65 6
Oxidative Phosphorylation 1.79 8
Thiamine Metabolism 1.54 8
Arginine and Proline Metabolism 1.44 20
Other Amino Acid Metabolism 1.28 5
Glycolysis/Gluconeogenesis 0.58 14
Anaplerotic reactions 0.19 9
*Number of reactions categorized in a subsystem or found to be neighboring each metabolite
Mo et al. BMC Systems Biology 2009 3:37

Table 4. List of top ten significant reporter metabolite and subsystem scores for the gdh1/GDH2 vs. wild type comparison in anaerobic conditions.


Table 4
List of top ten significant reporter metabolite and subsystem scores for the gdh1/GDH2 vs. wild type comparison in anaerobic conditions.
Reporter metabolite Z-score No of reactions
Glutamate [c] 4.52 35
Aspartate [c] 3.21 11
Alpha-ketoglutarate [c] 2.66 17
Glycine [c] 2.65 7
Pyruvate [m] 2.56 7
Ribulose-5-phosphate [c] 2.43 4
Threonine [c] 2.28 6
10-formyltetrahydrofolate [c] 2.27 5
Fumarate [c] 2.27 5
L-proline [c] 2.04 4
Reporter subsystem Z-score No of reactions
Valine, Leucine, and Isoleucine Metabolism 3.97 15
Tyrosine, Tryptophan, and Phenylalanine Metabolism 3.39 23
Pentose Phosphate Pathway 3.29 11
Purine and Pyrimidine Biosynthesis 3.08 40
Arginine and Proline Metabolism 2.96 19
Threonine and Lysine Metabolism 2.74 14
NAD Biosynthesis 2.66 7
Alanine and Aspartate Metabolism 2.65 6
Histidine Metabolism 2.24 10
Cysteine Metabolism 1.85 10
Mo et al. BMC Systems Biology 2009 3:37
Open Data

Perturbations under aerobic conditions largely consisted of pathways involved in mediating the NADH and NADPH balance. Among the highest scoring aerobic subsystems
are TCA cycle and pentose phosphate pathway – key pathways directly involved in the generation of NADH and NADPH. Reporter metabolites involved in these
subsystems –

  • glyceraldehyde-3-phosphate,
  • ribulose-5-phosphate, and
  • alpha-ketoglutarate – were also identified.

These results are consistent with flux and enzyme activity measurements

  • of the gdh1/GDH2 strain under aerobic conditions [32],
  1. which reported significant reduction in the pentose phosphate pathway flux
  2. with concomitant changes in other central metabolic pathways.

Levels of several TCA cycle intermediates (e.g. fumarate, succinate, malate) were also elevated

  • in the gdh1/GDH2 mutant according to the differential intracellular metabolite data.

Altered energy metabolism, as indicated by

  • reporter metabolites (i.e. ubiquinone- , ubiquinol, mitochondrial proton)
  • and subsystem (oxidative phosphorylation),

is certainly feasible as NADH is a primary reducing agent for ATP production.

Pentose phosphate pathway and NAD biosynthesis also appears

  • among the most perturbed anaerobic subsystems, further suggesting
  • perturbed cofactor balance as a common, dominant effect under both conditions.

Glutamate dehydrogenase is a critical enzyme of amino acid biosynthesis as it acts as

  • the entry point for ammonium assimilation via glutamate.

Consequently, metabolic subsystems involved in amino acid biosynthesis were broadly perturbed

  • as a result of the gdh1/GDH2 modification in both aerobic and anaerobic conditions.

For example, the proline biosynthesis pathway that uses glutamate as a precursor

  • was significantly perturbed in both conditions,
  • with significantly changed intracellular and extracellular levels.

There were differences, however, in that more amino acid related subsystems were

  • significantly affected in the anaerobic case (Table 4),
  • further highlighting that altered ammonium assimilation in the mutant
  • has a more widespread effect under anaerobic conditions.

This effect is especially pronounced for

  • threonine and nucleotide metabolism,
  • which were predicted to be significantly perturbed only in anaerobic conditions.

Intracellular threonine levels were amongst the most significantly reduced

  • relative to other differential intracellular metabolites in the anaerobically grown gdh1/GDH2 strain
    (see [31] and Additional file 4), and
  • the relationship between threonine and nucleotide biosynthesis is further supported

by threonine’s recently discovered role as a key precursor in yeast nucleotide biosynthesis [51].

Other key anaerobic reporter metabolites are

  • glycine and 10-formyltetrahydrofolate,
  • both of which are involved in the cytosolic folate cycle (one-carbon metabolism).

Folate is intimately linked to biosynthetic pathways of

  • glycine (with threonine as its precursor) and purines
  • by mediating one-carbon reaction transfers necessary in their metabolism and
  • is a key cofactor in cellular growth [52].

Thus, the anaerobic perturbations identified in the analysis emphasize the close relationship

  • between threonine, folate, and nucleotide metabolic pathways as well as
  • their potential connection to perturbed ammonium assimilation processes.

Interestingly, this association has been previously demonstrated at the transcriptional level

  • as yeast ammonium assimilation (via glutamine synthesis) was found to be
  • co-regulated with genes involved in glycine, folate, and purine synthesis [53].

In summary, the overall differences in predicted gdh1/GDH2 mutant behavior

  • under aerobic and anaerobic conditions show that changes in flux states
  • directly related to modified ammonium assimilation pathway
  1. are amplified anaerobically whereas the
  2. indirect effects through NADH/NADPH balance are more significant aerobically.

Perturbed metabolic regions under aerobic conditions were predominantly

  • in central metabolic pathways involved in responding to the changed NADH/NADPH demand
  • and did not necessarily emphasize that glutamate dehydrogenase was the site of the genetic modification.

The majority of affected anaerobic pathways were involved directly

  • in modified ammonium assimilation as evidenced by

1) significantly perturbed amino acid subsystems,

2) a broad perturbation subnetwork converging on glutamate (Figure 4), and

3) glutamate as the most significant reporter metabolite (Table 4).

Potassium-limited and excess ammonium environments

A recent study reported that potassium limitation resulted in significant

  • growth retardation effect in yeast due to excess ammonium uptake
  • when ammonium was provided as the sole nitrogen source [33].

The proposed mechanism for this effect was that ammonium

  • could to be freely transported through potassium channels
  • when potassium concentrations were low in the media environment, thereby
  • resulting in excess ammonium uptake [33].

As a result, yeast incurred a significant metabolic cost

  • in assimilating ammonia to glutamate and
  • secreting significant amounts of glutamate and other amino acids
  • in potassium-limited conditions as a means to detoxify the excess ammonium.

A similar effect was observed when yeast was grown

  • with no potassium limitation,
  • but with excess ammonia in the environment.

While the observed effect of both environments (low potassium or excess ammonia) was similar,

  • quantitatively unique amino acid secretion profiles suggested that
  • internal metabolic states in these conditions are potentially different.

In order to elucidate the differences in internal metabolic states, we utilized

  • the iMM904 model and the EM profile analysis method to analyze amino acid secretion profiles
  • for a range of low potassium and high ammonia conditions reported in [33].

As before, we utilized amino acid secretion patterns as constraints to the iMM904 model,

  1. sampled the allowable solution space,
  2. computed reaction Z-scores for changes from a reference condition (normal potassium and ammonia), and
  3. finally summarized the resulting changes using reporter metabolites.

Figure 5 shows a clustering of the most significant reporter metabolites (Z ≥ 1.96 in any of the four conditions studied)

  • obtained from this analysis across the four conditions studied.

Interestingly, the potassium-limited environment perturbed only a subset of

  • the significant reporter metabolites identified in the high ammonia environments.

Both low potassium environments shared a consistent pattern of

  • highly perturbed amino acids and related precursor biosynthesis metabolites
    (e.g. pyruvate, PRPP, alpha-ketoglutarate)
  • with high ammonium environments.

The amino acid perturbation pattern (indicated by red labels in Figure 5) was present in

  • the ammonium-toxic environments, although the pattern was
  • slightly weaker for the lower ammonium concentration.

Nevertheless, the results clearly indicate that a similar

  • ammonium detoxifying mechanism that primarily perturbs pathways
  • directly related to amino acid metabolism
  • exists under both types of media conditions.

Figure 5.

Clustergram of top reporter metabolites - y in ammonium-toxic and potassium-limited conditions

Clustergram of top reporter metabolites – y in ammonium-toxic and potassium-limited conditions

Clustergram of top reporter metabolites (i.e. in yellow) in ammonium-toxic and potassium-limited conditions.

Amino acid perturbation patterns (shown in red) were shown to be consistently scored across conditions, indicating that potassium-limited environments K1 (lowest
concentration) and K2 (low concentration) elicited a similar ammonium detoxification response as ammonium-toxic environments N1 (high concentration) and N2
(highest concentration). Metabolites associated with folate metabolism (highlighted in green) are also highly perturbed in ammonium-toxic conditions. Metabolite
abbreviations are found in Additional file 1.

In addition to perturbed amino acids, a secondary effect notably appears at high ammonia levels in which metabolic regions related to folate metabolism are significantly affected. As highlighted in green in Figure 3, we predicted significantly perturbed key metabolites involved in the cytosolic folate cycle. These include tetrahydrofolate derivatives and other metabolites connected to the folate pathway, namely glycine and the methionine-derived methylation cofactors S-adenosylmethionine and S-adenosyl-homocysteine. Additionally, threonine was identified to be a key perturbed metabolite in excess ammonium conditions. These results further illustrate the close
connection between threonine biosynthesis, folate metabolism involving glycine derived from its threonine precursor, and nucleotide biosynthesis [51] that was discussed in
conjunction with the gdh1/GDH2 strain data. Taken together with the anaerobic gdh1/GDH2 data, the results consistently suggest highly perturbed threonine and folate
metabolism when amino acid-related pathways are broadly affected.

In both ammonium-toxic and potassium-limited environments, impaired cellular growth was observed, which can be attributed to high energetic costs of increased
ammonium assimilation to synthesize and excrete amino acids. However, under high ammonium environments, reporter metabolites related to threonine and folate
metabolism indicated that their perturbation, and thus purine supply, may be an additional factor in decreasing cellular viability as there is a direct relationship between
intracellular folate levels and growth rate [54]. Based on these results, we concluded that while potassiumlimited growth in yeast indeed shares physiological features with
growth in ammonium excess, its effects are not as detrimental as actual ammonium excess. The effects on proximal amino acid metabolic pathways are similar in both
environments as indicated by the secretion of the majority of amino acids. However, when our method was applied to analyze the physiological basis behind differences in
secretion profiles between low potassium and high ammonium conditions, ammonium excess was predicted to likely disrupt physiological ammonium assimilation processes,
which in turn potentially impacts folate metabolism and associated cellular growth.


The method presented in this study presents an approach to connecting intracellular flux states to metabolites that are excreted under various physiological conditions. We
showed that well-curated genome-scale metabolic networks can be used to integrate and analyze quantitative EM data by systematically identifying altered intracellular
pathways related to measured changes in the extracellular metabolome. We were able to identify statistically significant metabolic regions that were altered as a result of
genetic (gdh1/GD2 mutant) and environmental (excess ammonium and limited potassium) perturbations, and the predicted intracellular metabolic changes were consistent
with previously published experimental data including measurements of intracellular metabolite levels and metabolic fluxes. Our reanalysis of previously published EM data
on ammonium assimilation-related genetic and environmental perturbations also resulted in testable hypotheses about the role of threonine and folate pathways in mediating
broad responses to changes in ammonium utilization. These studies also demonstrated that the samplingbased method can be readily applied when only partial secreted
metabolite profiles (e.g. only amino acids) are available.

With the emergence of metabolite biofluid biomarkers as a diagnostic tool in human disease [55,56] and the availability of genome-scale human metabolic networks [1],
extensions of the present method would allow identifying potential pathway changes linked to these biomarkers. Employing such a method for studying yeast metabolism was possible as the metabolomic data was measured under controllable environmental conditions where the inputs and outputs of the system were defined. Measured metabolite biomarkers in a clinical setting, however, is far from a controlled environment with significant variations in genetic, nutritional, and environmental factors between different
patients. While there are certainly limitations for clinical applications, the method introduced here is a progressive step towards applying genome-scale metabolic networks
towards analyzing biofluid metabolome data as it 1) avoids the need to only study optimal metabolic states based on a predetermined objective function, 2) allows dealing with noisy experimental data through the sampling approach, and 3) enables analysis even with limited identification of metabolites in the data. The ability to establish potential
connections between extracellular markers and intracellular pathways would be valuable in delineating the genetic and environmental factors associated with a particular

Authors’ contributions

Conceived and designed the experiments: MLM MJH BOP. Performed experiments: MLM MJH. Analyzed the data: MLM MJH. Wrote the paper: MLM MJH BOP. All authors have read and approved the final manuscript.


We thank Jens Nielsen for providing the raw metabolome data for the mutant strain, and Jan Schellenberger and Ines Thiele for valuable discussions. This work was supported by NIH grant R01 GM071808. BOP serves on the scientific advisory board of Genomatica Inc.



  1. Duarte NC, Becker SA, Jamshidi N, Thiele I, Mo ML, Vo TD, Srivas R, Palsson BO: Global reconstruction of the human metabolic network based on genomic and bibliomic data. 

Proc Natl Acad Sci USA 2007, 104(6):1777-1782. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Duarte NC, Herrgard MJ, Palsson B: Reconstruction and Validation of Saccharomyces cerevisiae iND750, a Fully Compartmentalized Genome-Scale Metabolic Model. 

Genome Res 2004, 14(7):1298-1309. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Kuepfer L, Sauer U, Blank LM: Metabolic functions of duplicate genes in Saccharomyces cerevisiae. 

Genome Res 2005, 15(10):1421-1430. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Nookaew I, Jewett MC, Meechai A, Thammarongtham C, Laoteng K, Cheevadhanarak S, Nielsen J, Bhumiratana S: The genome-scale metabolic model iIN800 of Saccharomyces cerevisiae and its validation: a scaffold to query lipid metabolism. 

BMC Syst Biol 2008, 2:71. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text

  1. Edwards JS, Palsson BO: Systems properties of the Haemophilus influenzae Rd metabolic genotype. 

J biol chem 1999, 274(25):17410-17416. PubMed Abstract | Publisher Full Text

  1. Edwards JS, Palsson BO: The Escherichia coli MG1655 in silico metabolic genotype: Its definition, characteristics, and capabilities. 

Proc Natl Acad Sci USA 2000, 97(10):5528-5533. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Thiele I, Vo TD, Price ND, Palsson B: An Expanded Metabolic Reconstruction of Helicobacter pylori (IT341 GSM/GPR): An in silico genome-scale characterization of single and double deletion mutants. 

J Bacteriol 2005, 187(16):5818-5830. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Vo TD, Greenberg HJ, Palsson BO: Reconstruction and functional characterization of the human mitochondrial metabolic network based on proteomic and biochemical data. 

J Biol Chem 2004, 279(38):39532-39540. PubMed Abstract | Publisher Full Text

  1. Reed JL, Vo TD, Schilling CH, Palsson BO: An expanded genome-scale model of Escherichia coli K-12 (iJR904 GSM/GPR). 

Genome Biology 2003, 4(9):R54.51-R54.12. BioMed Central Full Text

  1. Feist AM, Henry CS, Reed JL, Krummenacker M, Joyce AR, Karp PD, V H, Palsson BO: A genome-scale metabolic reconstruction for Escherichia coli K-12 MG1655 that accounts for 1261 ORFs and thermodynamic information. 

Molecular Systems Biology 2007, 3:121. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Suthers PF, Dasika MS, Kumar VS, Denisov G, Glass JI, Maranas CD: A genome-scale metabolic reconstruction of Mycoplasma genitalium, iPS189. 

PLos Comp Biol 2009, 5(2):e1000285. Publisher Full Text

  1. Price ND, Reed JL, Palsson BO: Genome-scale models of microbial cells: evaluating the consequences of constraints. 

Nat Rev Microbiol 2004, 2(11):886-897. PubMed Abstract | Publisher Full Text

  1. Becker SA, Feist AM, Mo ML, Hannum G, Palsson BO, Herrgard MJ: Quantitative Prediction of Cellular Metabolism with Constraint-based Models: The COBRA Toolbox. 

Nature protocols 2007, 2(3):727-738. PubMed Abstract | Publisher Full Text

  1. Reed JL, Palsson BO: Genome-Scale In Silico Models of E. coli Have Multiple Equivalent Phenotypic States: Assessment of Correlated Reaction Subsets That Comprise Network States. 

Genome Res 2004, 14(9):1797-1805. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Fong SS, Palsson BO: Metabolic gene deletion strains of Escherichia coli evolve to computationally predicted growth phenotypes. 

Nature Genetics 2004, 36(10):1056-1058. PubMed Abstract | Publisher Full Text

  1. Ibarra RU, Edwards JS, Palsson BO: Escherichia coli K-12 undergoes adaptive evolution to achieve in silico predicted optimal growth. 

Nature 2002, 420(6912):186-189. PubMed Abstract | Publisher Full Text

  1. Schellenberger J, Palsson BØ: Use of randomized sampling for analysis of metabolic networks. 

J Biol Chem 2009, 284(9):5457-5461. PubMed Abstract | Publisher Full Text

  1. Almaas E, Kovács B, Vicsek T, Oltvai ZN, Barabási AL: Global organization of metabolic fluxes in the bacterium Escherichia coli. 

Nature 2004, 427(6977):839-843. PubMed Abstract | Publisher Full Text

  1. Thiele I, Price ND, Vo TD, Palsson BO: Candidate metabolic network states in human mitochondria: Impact of diabetes, ischemia, and diet. 

J Biol Chem 2005, 280(12):11683-11695. PubMed Abstract | Publisher Full Text

  1. Kell DB: Metabolomics and systems biology: making sense of the soup. 

Curr Opin Microbiol 2004, 7(3):296-307. PubMed Abstract | Publisher Full Text

  1. Kell DB, Brown M, Davey HM, Dunn WB, Spasic I, Oliver SG: Metabolic footprinting and systems biology: the medium is the message. 

Nat Rev Microbiol 2005, 3(7):557-565. PubMed Abstract | Publisher Full Text

  1. Goodacre R, Vaidyanathan S, Dunn WB, Harrigan GG, Kell DB: Metabolomics by numbers: acquiring and understanding global metabolite data. 

Trends Biotechnol 2004, 22(5):245-252. PubMed Abstract | Publisher Full Text

  1. Lenz EM, Bright J, Wilson ID, Morgan SR, Nash AF: A 1H NMR-based metabonomic study of urine and plasma samples obtained from healthy human subjects. 

J Pharm Biomed Anal 2003, 33(5):1103-1115. PubMed Abstract | Publisher Full Text

  1. Allen J, Davey HM, Broadhurst D, Heald JK, Rowland JJ, Oliver SG, Kell DB: High-throughput classification of yeast mutants for functional genomics using metabolic footprinting. 

Nat Biotech 2003, 21(6):692-696. Publisher Full Text

  1. Nicholson JK, Connelly J, Lindon JC, Holmes E: Metabonomics: a platform for studying drug toxicity and gene function. 

Nat Rev Drug Discov 2002, 1(2):153-161. PubMed Abstract | Publisher Full Text

  1. Mortishire-Smith RJ, Skiles GL, Lawrence JW, Spence S, Nicholls AW, Johnson BA, Nicholson JK: Use of metabonomics to identify impaired fatty acid metabolism as the mechanism of a drug-induced toxicity. 

Chem Res Toxicol 2004, 17(2):165-173. PubMed Abstract | Publisher Full Text

  1. Sabatine MS, Liu E, Morrow DA, Heller E, McCarroll R, Wiegand R, Berriz GF, Roth FP, Gerszten RE: Metabolomic identification of novel biomarkers of myocardial ischemia. 

Circulation 2005, 112(25):3868-3875. PubMed Abstract | Publisher Full Text

  1. Cakir T, Efe C, Dikicioglu D, Hortaçsu AKB, Oliver SG: Flux balance analysis of a genome-scale yeast model constrained by exometabolomic data allows metabolic system identification of genetically different strains. 

Biotechnol Prog 2007, 23(2):320-326. PubMed Abstract | Publisher Full Text

  1. Bang JW, Crockford DJ, Holmes E, Pazos F, Sternberg MJ, Muggleton SH, Nicholson JK: Integrative top-down system metabolic modeling in experimental disease states via data-driven Bayesian methods. 

J Proteome Res 2008, 7(2):497-503. PubMed Abstract | Publisher Full Text

  1. Oliveira AP, Patil KR, Nielsen J: Architecture of transcriptional regulatory circuits is knitted over the topology of bio-molecular interaction networks. 

BMC Syst Biol 2008, 2:17. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text

  1. Villas-Boas SG, Moxley JF, Akesson M, Stephanopoulos G, Nielsen J: High-throughput metabolic state analysis: the missing link in integrated functional genomics of yeasts. 

Biochem J 2005, 388(Pt 2):669-677. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Moreira dos Santos M, Thygesen G, Kötter P, Olsson L, Nielsen J: Aerobic physiology of redox-engineered Saccharomyces cerevisiae strains modified in the ammonium assimilation for increased NADPH availability. 

FEMS Yeast Res 2003, 4(1):59-68. PubMed Abstract | Publisher Full Text

  1. Hess DC, Lu W, Rabinowitz JD, Botstein D: Ammonium toxicity and potassium limitation in yeast. 

PLoS Biol 2006, 4(11):e351. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Nissen TL, Schulze U, Nielsen J, Villadsen J: Flux distributions in anaerobic, glucose-limited continuous cultures of Saccharomyces cerevisiae. 

Microbiology 1997, 143(Pt 1):203-218. PubMed Abstract | Publisher Full Text

  1. Famili I, Forster J, Nielsen J, Palsson BO: Saccharomyces cerevisiae phenotypes can be predicted by using constraint-based analysis of a genome-scale reconstructed metabolic network. 

Proc Natl Acad Sci USA 2003, 100(23):13134-13139. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Bonarius HPJ, Schmid G, Tramper J: Flux analysis of underdetermined metabolic networks: The quest for the missing constraints. 

Trends in Biotechnology 1997, 15(8):308-314. Publisher Full Text

  1. Edwards JS, Palsson BO: Metabolic flux balance analysis and the in silico analysis of Escherichia coli K-12 gene deletions. 

BMC Bioinformatics 2000, 1:1. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text

  1. Varma A, Palsson BO: Metabolic Flux Balancing: Basic concepts, Scientific and Practical Use. 

Nat Biotechnol 1994, 12:994-998. Publisher Full Text

  1. Segre D, Vitkup D, Church GM: Analysis of optimality in natural and perturbed metabolic networks. 

Proc Natl Acad Sci USA 2002, 99(23):15112-15117. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Baldi P, Brunak S, Chauvin Y, Andersen CA, Nielsen H: Assessing the accuracy of prediction algorithms for classification: an overview. 

Bioinformatics 2000, 16(5):412-424. PubMed Abstract | Publisher Full Text

  1. Price ND, Schellenberger J, Palsson BO: Uniform Sampling of Steady State Flux Spaces: Means to Design Experiments and to Interpret Enzymopathies. 

Biophysical Journal 2004, 87(4):2172-2186. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Price ND, Thiele I, Palsson BO: Candidate states of Helicobacter pylori’s genome-scale metabolic network upon application of “loop law” thermodynamic constraints. 

Biophysical J 2006, 90(11):3919-3928. Publisher Full Text

  1. Schuetz R, Kuepfer L, Sauer U: Systematic evaluation of objective functions for predicting intracellular fluxes in Escherichia coli. 

Mol Syst Biol 2007, 3:119. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Patil KR, Nielsen J: Uncovering transcriptional regulation of metabolism by using metabolic network topology. 

Proc Natl Acad Sci USA 2005, 102(8):2685-2689. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, Ideker T: Cytoscape: a software environment for integrated models of biomolecular interaction networks. 

Genome Res 2003, 13(11):2498-2504. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Herrgard MJ, Swainston N, Dobson P, Dunn WB, Arga KY, Arvas M, Bluthgen N, Borger S, Costenoble R, Heinemann M, et al.: A consensus yeast metabolic network reconstruction obtained from a community approach to systems biology. 

Nat Biotech 2008, 26:1155-1160. Publisher Full Text

  1. Forster J, Famili I, Fu PC, Palsson BO, Nielsen J: Genome-Scale Reconstruction of the Saccharomyces cerevisiae Metabolic Network. 

Genome Research 2003, 13(2):244-253. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Reed JL, Patel TR, Chen KH, Joyce AR, Applebee MK, Herring CD, Bui OT, Knight EM, Fong SS, Palsson BO: Systems Approach to Genome Annotation: Prediction and Validation of Metabolic Functions. 

Proc Natl Acad Sci USA 2006, 103(46):17480-17484. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Nissen TL, Kielland-Brandt MC, Nielsen J, Villadsen J: Optimization of ethanol production in Saccharomyces cerevisiae by metabolic engineering of the ammonium assimilation. 

Metab Eng 2000, 2(1):69-77. PubMed Abstract | Publisher Full Text

  1. Roca C, Nielsen J, Olsson L: Metabolic engineering of ammonium assimilation in xylose-fermenting Saccharomyces cerevisiae improves ethanol production. 

Appl Environ Microbiol 2003, 69(8):4732-4736. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Hartman JL IV: Buffering of deoxyribonucleotide pool homeostasis by threonine metabolism. 

Proc Natl Acad Sci USA 2007, 104(28):11700-11705. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

  1. Gelling CL, Piper MD, Hong SP, Kornfeld GD, Dawes IW: Identification of a novel one-carbon metabolism regulon in Saccharomyces cerevisiae. 

J Biol Chem 2004, 279(8):7072-7081. PubMed Abstract | Publisher Full Text

  1. Denis V, Daignan-Fornier B: Synthesis of glutamine, glycine and 10-formyl tetrahydrofolate is coregulated with purine biosynthesis in Saccharomyces cerevisiae. 

Mol Gen Genet 1998, 259(3):246-255. PubMed Abstract | Publisher Full Text

  1. Hjortmo S, Patring J, Andlid T: Growth rate and medium composition strongly affect folate content in Saccharomyces cerevisiae. 

Int J Food Microbiol 2008, 123(1–2):93-100. PubMed Abstract | Publisher Full Text

  1. Kussmann MRF, Affolter M: OMICS-driven biomarker discovery in nutrition and health. 

J Biotechnol 2006, 124(4):758-787. PubMed Abstract | Publisher Full Text

  1. Serkova NJ, Niemann CU: Pattern recognition and biomarker validation using quantitative 1H-NMR-based metabolomics. 

Expert Rev Mol Diagn 2006, 6(5):717-731. PubMed Abstract | Publisher Full Text


Read Full Post »

The multi-step transfer of phosphate bond and hydrogen exchange energy

Curator: Larry H. Bernstein, MD, FCAP, Leaders in Pharmaceutical Intelligence

In this subtext of the series we expand on a tie between respiration and glycolysis, and the functioning of the mitochondrion to discover the key role played by oxidative phosphorylation, “acetyl coenzyme A, and electron transport.  This was crucial to understanding cellular energetics, which explains the high energy of fatty acid catabolism from stored adipose tissue, and the criticality of the multi-step sequence of reactions in energy transfer.

This portion considerably provides a response to the TWO points made by Jose EDS Rosallis:

  1. Just at the beginning, when phosphorylation of proteins is presented, I assume you must mention that some proteins are activated by phosphorylation. This is fundamental in order to present self –organization reflex upon fast regulatory mechanisms. This poiny needs further clarification, but he makes important observations here.
  • Even from an historical point of view. The first observation arrived from a sample due to be studied on the following day of glycogen synthetase. It was unintended left overnight out of the refrigerator. The result was it had changed from active form of the previous day to a non-active form.

The story could have being finished here, if the researcher did not decide to spent this day increasing substrate levels (it could be a simple case of denaturation of proteins that changes its conformation despite the same order of amino acids). He kept on trying and found restoration of maximal activity.

  • This assay was repeated with glycogen phosphorylase and the result was the opposite it increases its activity.

This led to the discovery of cAMP activated protein kinase and the assembly of a very complex system in the glycogen granule that is not a simple carbohydrate polymer. Instead

  • it has several proteins assembled and preserves the capacity to receive from a single event (rise in cAMP) two opposing signals with maximal efficiency,
  • stops glycogen synthesis, as long as levels of glucose 6 phosphate are low and
  • increases glycogen phosphorylation as long as AMP levels are high).

I did everything I was able to do by the end of 1970 in order to repeat these assays with

  • PK I, PKII and PKIII of M. Rouxii and Sutherland route to cAMP failed in this case.

I ask Leloir to suggest to my chief (SP) the idea of AA, AB, BB subunits as was observed in lactic dehydrogenase (tetramer)
(Nathan O. Kaplan discovery) indicating this as his idea. The reason was my “chief” (SP) more than once,  said to me: “Leave these great ideas for the Houssay, Leloir etc…We must do our career with small things. ” However, as she also had a faulty ability for recollection she also used to arrive some time later, with the very same idea but in that case, as her idea.

[This reminds me of when I was studying the emergence of lactic dehysrogenase isoenzyme patterns in the developing eye lens of cattle, I raised reservations about Elliott Vessells challenge to Nathan Kaplan, but that not being my primary problem, my brilliant mentor (H.M.), a very young full professor of anatomy said – leave that to NOK.}

Leloir, said to me: I will not offer your interpretation to her as mine. I think it is not phosphorylation, however I think it is

  • glycosylation that explains the changes in the isoenzymes with the same molecular weight preserved.

This dialogue explains why during the Schroedinger’s “What is life?” reading with him he asked me if from biochemist in exile, to biochemist I expressed all of my thoughts to him. Since I had considered that Schrödinger did not confront Darlington & Haldane for being in exile. This may explain why Leloir could have answered a bad telephone call from P. Boyer, Editor of The Enzymes in a way that suggests the the pattern could be of covalent changes over a protein. Our FEBS and Eur J. Biochemistry papers on pyruvate kinase of M. Rouxii is wrongly quoted in this way on his review about pyruvate kinase of
that year(1971).

  1. show in detail with different colors what carbons belongs to CoA a huge molecule, in comparison with the single two carbons of acetate that will produce the enormous jump in energy yield in comparison with anaerobic glycolysis. The idea is how much must have being spent in DNA sequences to build that molecule in order to use only two atoms of carbon. Very limited aspects of biology could be explained in this way. In case we follow an alternative way of thinking, it becomes clearer that proteins were made more stable by interaction with other molecules (great and small). Afterwards, it rather easy to understand how the stability of protein-RNA complexes where transmitted to RNA (vibrational +solvational reactivity stability pair of conformational energy). Latter, millions of years, or as soon as, the information of interaction leading to activity and regulation could be found in RNA, proteins like reverse transcriptase move this information to a more stable form (DNA). In this way it is easier to understand the use of CoA to make two carbon molecules more reactive.

The outline of what I am presenting in series is as follows:

  1. Signaling and Signaling Pathways
  1. Signaling transduction tutorial.
  1. Carbohydrate metabolism

3.1  Selected References to Signaling and Metabolic Pathways in Leaders in Pharmaceutical Intelligence

  1. Lipid metabolism

4.1  Studies of respiration lead to Acetyl CoA

4.2 The multi-step transfer of phosphate bond and hydrogen exchange energy

  1. Protein synthesis and degradation
  2. Subcellular structure
  3. Impairments in pathological states: endocrine disorders; stress hypermetabolism; cancer.

Oxidation-Reduction Reactions

Rachel Casiday, Carolyn Herman, and Regina Frey
Department of Chemistry, Washington University
St. Louis, MO 63130


OX-Phos steps

OX-Phos steps


Key Concepts:

  • ATP as Free-Energy Currency in the Body
  • Coupled Reactions
    • Standard Free-Energy Change for Coupled Reactions
    • ATP Dephosphorylation Coupled to Nonspontaneous Reactions
    • Coupled Reactions to Generate ATP
  • Structure and Function of the Mitochondria
  • Oxidation-Reduction Reactions in the Electron-Transport Chain
    • Electron-Carrier Proteins (NOTE: This section includes a separate link and an animation.)
    • Relationship Between Reduction Potentials and Free Energy
  • Proton Gradient as Means of Coupling Oxidative and Phosphorylation Components of Oxidative Phosphorylation
  • ATP Synthetase Uses Energy From Proton Gradient to Generate ATP

Every day, we build bones, move muscles, eat food, think, and perform many other activities with our bodies. All of these activities are based upon chemical reactions. However, most of these reactions are not spontaneous (i.e., they are accompanied by a positive change in free energy, DG>0) and do not occur without some other source of free energy. Hence, the body needs some sort of “free-energy currency,” (Figure 1) a molecule that can store and release free energy when it is needed to power a given biochemical reaction.

The four questions:

  1. How does the body “spend” free-energy currency to make a nonspontaneous reaction spontaneous? The answer, which is based on thermodynamics, is to use coupled reactions.
  2. How is food used to produce the reducing agents (NADH and FADH2) that can regenerate the free-energy currency? The answer, from biology, is found in glycolysis and the citric-acid cycle.
  3. How are the reducing agents (NADH and FADH2) able to generate the free-energy currency molecule (ATP)? Once again, coupled reactions are key.
  4. What mechanism does the body use to couple the reducing agent reactions and the generation of ATP? ATP is synthesized primarily by a two-step process consisting of an electron-transport chain and a proton gradient.  This process is based on electrochemistry and equilibrium, as well as thermodynamics.

The free-energy change (DG) for the net reaction is given by the sum of the free-energy changes for the individual reactions.  The phospholipids that form cell membranes are formed from glycerol with a phosphate group and two fatty-acid chains attached.This step actually consists of two reactions:

(1) the phosphorylation of glycerol, and

(2) the dephosphorylation of ATP (the free-energy-currency molecule). The reactions may be added as shown in Equations 2-4, below:

      Glycerol + HPO42- –>  (Glycerol-3-Phosphate)2- + H2O DGo= +9.2 kJ
+      ATP4- + H2O –>       ADP3- + HPO42- + H+ DGo30.5 kJ
     Glycerol + ATP4- –> (Glycerol-3-Phosphate)2- +ADP3- + H+ DGo21.3 kJ

ATP is the most important “free-energy-currency” molecule in living organisms (see Figure 2, below). Adenosine triphosphate (ATP) is a useful free-energy currency because the dephosphorylation reaction is very spontaneous; i.e., it releases a large amount of free energy (30.5 kJ/mol). Thus, the dephosphorylation reaction of ATP to ADP and inorganic phosphate (Equation 3) is often coupled with nonspontaneous reactions (e.g., Equation 2) to drive them forward. The body’s use of ATP as a free-energy currency is a very effective strategy to cause vital nonspontaneous reactions to occur.

structure of ATP

structure of ATP

This is the two-dimensional (ChemDraw) structure of ATP, adenosine triphosphate. The removal of one phosphate group (green) from ATP requires the breaking of a bond (blue) and results in a large release of free energy. Removal of this phosphate group (green) results in ADP, adenosine diphosphate.

flowchart of food energy

flowchart of food energy

This flowchart shows that the energy used by the body for its many activities ultimately comes from the chemical energy in our food. The chemical energy in our food is converted to reducing agents (NADH and FADH2). These reducing agents are then used to make ATP. ATP stores chemical energy, so that it is available to the body in a readily accessible form.

Glycolysis   Glucose + 2 HPO42- + 2 ADP3- + 2 NAD+ –>
2 Pyruvate + 2 ATP4- + 2 NADH + 2 H+ + 2 H2O
Intermediate Step   2(Pyruvate + Coenzyme A + NAD+ –>
Acetyl CoA + CO2 + NADH)
Citric-Acid Cycle 2(Acetyl CoA + 3 NAD++ FAD + GDP3-
+ HPO42- + 2H2O –> 2 CO2 + 3 NADH + FADH2
+ GTP4- + 2H+ + Coenzyme A)

The structures of the important molecules in Equations 5-7 are shown in Table 1, below.

How is Food Used to Make the Reducing Agents Needed for the Production of ATP?

To make ATP, energy must be absorbed. This energy is supplied by the food we eat, and then used to synthsize two reducing agents, NADH and FADH2 that are needed to produce ATP. One of the principal energy-yielding nutrients in our diet is glucose (see structure in Table 1 in the blue box below), a simple six-carbon sugar that can be broken down by the body. When the chemical bonds in glucose are broken, free energy is released. The complete breakdown of glucose into CO2 occurs in two processes: glycolysis and the citric-acid cycle. The reactions for these two processes are shown in the blue box below.






Acetyl CoA








two-dimensional representations of several important molecules in Equations 5-7.

As seen in Equations 5-7 in the blue box, glycolysis and the citric-acid cycle produce a net total of only four ATP or GTP molecules (GTP is an energy-currency molecule similar to ATP) per glucose molecule. This yield isfar below the amount needed by the body for normal functioning, and in fact is far below the actual ATP yield for glucose in aerobic organisms (organisms that use molecular oxygen). For each glucose molecule the body processes, the body actually gains approximately 30 ATP molecules! (See Figure 4, below.)  So, how does the body generate ATP?

The process that accounts for the high ATP yield is known as oxidative phosphorylation. A quick examination of Equations 5-7 shows that glycolysis and the citric-acid cycle generate other products besides ATP and GTP, namely NADH and FADH2 (blue). These products are molecules that are oxidized (i.e., give up electrons) spontaneously. The body uses these reducing agents (NADH and FADH2) in an oxidation-reduction reaction .  As you will see later in this tutorial, it is the free energy from these redox reactions that is used to drive the production of ATP.

flowchart - making of ATP

flowchart – making of ATP

This flowchart shows the major steps involved in breaking down glucose from the diet and converting its chemical energy to the chemical energy in the phosphate bonds of ATP, in aerobic (oxygen-using) organisms. Note: In this flowchart, red denotes a source of carbon atoms (originally from glucose),green denotes energy-currency molecules, and blue denotes the reducing agents that can be oxidized spontaneously.

In the discussion above, we see that glucose by itself generates only a tiny amount of ATP. However, during the breakdown of glucose, a large amount of NADH and FADHis produced; it is these reducing agents that dramatically increase the amount of ATP produced. How does this work?

How are the reducing agents (NADH and FADH2) able to generate the free-energy currency molecule (ATP)?

As discussed in an earlier section about coupling reactions, ATP is used as free-energy currency by coupling its (spontaneous) dephosphorylation (Equation 3) with a (nonspontaneous) biochemical reaction to give a net release of free energy (i.e., a net spontaneous reaction). Coupled reactions are also used to generate ATP by phosphorylating ADP. The nonspontaneous reaction of joining ADP to inorganic phosphate to make ATP (Equation 8, below, and Figure 2, above) is coupled to the oxidation reaction of NADH or FADH(Equation 9, below). (Recall, NADH and FADH2 are produced in glycolysis and the citric-acid cycle as described in the blue box). For simplicity, we shall henceforth discuss only the oxidation of NADH; FADH2 follows a very similar oxidation pathway.

The oxidation reaction for NADH has a larger, but negative, DG than the positive DG required for the formation of ATP from ADP and phosphate. This set of coupled reactions is so important that it has been given a special name: oxidative phosphorylation. This name emphasizes the fact that an oxidation (of NADH) reaction (Equation 9 and Figure 5, below) is being coupled to a phosphorylation (of ADP) reaction (Equation 8, below, and Figure 2, above). In addition, we must consider the reduction reaction (gaining of electrons) that accompanies the oxidation of NADH. (Oxidation reactions are always accompanied by reduction reactions, because an electron given up by one group must be accepted by another group.) In this case, molecular oxygen (O2) is the electron acceptor, and the oxygen is reduced to water (Equation 10, below) .

The individual reactions of interest for oxidative phosphorylation are:


ADP3- + HPO42- + H+ –>
ATP4- + H2O

DGo= +30.5 kJ

NADH –> NAD+ + H+ +  2e

DGo158.2 Kj

1/2 O2 + 2H+ + 2e –> H2O

DGo61.9 kJ


The net reaction is obtained by summing the coupled reactions, as shown in Equation 11, below.

ADP3- + HPO42- + NADH + 1/2 O2 + 2H+ –>
ATP4- + NAD+ + 2 H2O
DGo= -189.6 kJ

The molecular changes that occur upon oxidation of NADH are shown:



This is a two-dimensional (ChemDraw) representation showing the change that occurs when NADH is oxidized to NAD+. “R” represents the part of the structure that is shown in black in the drawing of NADH in Table 1, and does not change during the oxidation half-reaction. The molecular changes that occur upon oxidation are shown in red.

In this tutorial, we have seen that nonspontaneous reactions in the body occur by coupling them with a very spontaneous reaction (usually the ATP reaction shown in Equation 3). We have just seen that ATP is produced by coupling the phosphorylation reaction with NADH oxidation (a very spontaneous reaction). But we have not yet answered the question: by what mechanism are these reactions coupled?

Coupling Reactions in Biological Systems

Every day your body carries out many nonspontaneous reactions. As discussed earlier, if a nonspontaneous reaction is coupled to a spontaneous reaction, as long as the sum of the free energies for the two reactions is negative, the coupled reactions will occur spontaneously. How is this coupling achieved in the body? Living systems couple reactions in several ways, but the most common method of coupling reactions is to carry out both reactions on the same enzyme. Consider again the phosphorylation of glycerol (Equations 2-4). Glycerol is phosphorylated by the enzyme glycerol kinase, which is found in your liver. The product of glycerol phosporylation, glycerol-3-phosphate (Equation 2), is used in the synthesis of phospholipids.

Glycerol kinase is a large protein comprised of about 500 amino acids. X-ray crystallography of the protein shows us that there is a deep groove or cleft in the protein where glycerol and ATP attach (see Figure 6, below). Because the enzyme holds the ATP and the glycerol in place, the phosphate can be transferred directly from the ATP to glycerol. Instead of two separate reactions where ATP loses a phosphate (Equation 3) and glycerol picks up a phosphate (Equation 2), the enzyme allows the phosphate to move directly from ATP to glycerol (Equation 4).

The coupling in oxidative phosphorylation uses a more complicated (and amazing!) mechanism, but the end result is the same: the reactions are linked together, the net free energy for the linked reactions is negative, and, therefore, the linked reactions are spontaneous.



This is a schematic representation of ATP and glycerol bound (attached) to glycerol kinase. The enzyme glycerol kinase is a dimer (consists of two identical subuits). There is a deep cleft between the subunits where ATP and glycerol bind. Since the ATP and phosphate are physically so close together when they are bound to the enzyme, the phosphate can be transferred directly from ATP to glycerol. Hence, the processes of ATP losing a phosphate (spontaneous) and glycerol gaining a phosphate (nonspontaneous) are linked together as one spontaneous process

Questions on ATP: The Body’s Free-Energy Currency (How Free-Energy Currency Works)

  • Biological systems involve many molecules containing phosphate groups, such as ATP. Although ATP is the most commonly used free-energy currency, any of these phosphorylated molecules could, in theory, be used as free-energy currency. The standard free-energy change (DGo) for the dephosphorylation (removal of a phosphate group) of several biological compounds is given below:
Acetyl phosphate DGo = -47.3 kJ/mol
Adenosine triphosphate (ATP) DGo = -30.5 kJ/mol
Glucose-6-phosphate DGo = -13.8 kJ/mol
Phosphoenolpyruvate (PEP) DGo = -61.9 kJ/mol
Phosphocreatine DGo = -43.1 kJ/mol

Neglecting any differences in difficulty synthesizing or accessing these molecules by biological systems, rank the molecules in order of their efficiency as a free-energy currency (i.e., the amount of nonspontaneous reactions enabled per phosphate removed from a molecule of free-energy currency) from the most efficient to the least efficient.

  • What, if any, changes are there in the shape of the ring as NADH is oxidized to NAD+(see Figure 5)? (Hint: Consider which atoms lie in the same plane in each structure.)

Mechanism of Coupling the Oxidative-Phosphorylation Reactions

In order to couple the redox and phosphorylation reactions needed for ATP synthesis in the body, there must be some mechanism linking the reactions together. In cells, this is accomplished through an elegant proton-pumping system that occurs inside special double-membrane-bound organelles (specialized cellular components) known as mitochondria. A number of proteins are required to maintain this proton-pumping system and catalyze the oxidative and phosphorylation reactions.

Synthesis of ATP (Equation 8) is coupled with the oxidation of NADH (Equation 9) and the reduction of O2 (Equation 10). There are three key steps in this process:

  1. Electrons are transferred from NADH, through a series of electron carriers, to O2. The electron carriers are proteins embedded in the inner mitochondrial membrane. (More detail about the structure of the mitochondria is presented in the next section.) (See Figure 7a.)
  2. Transfer of electrons by these carriers generates a proton (H+) gradient across the inner mitochondrial membrane. (See Figure 7b.)
  3. When Hspontaneously diffuses back across the inner mitochondrial membrane, ATP is synthesized. The large positive free energy of ATP synthesis is overcome by the even larger negative free energy associated with proton flow down the concentration gradient. (See Figure 7c.)

These steps are outlined below.

  1. Electron Transport (Oxidation-Reduction Reactions) Through a Series of Proteins in the Inner Membrane of the Mitochondria


Generation of H+(Proton) Concentration Gradient Across the Inner Mitochondrial Membrane During the Electron-Transport Process (via a Proton Pump)

. Generation of H+ (Proton) Concentration Gradient Across the Inner Mitochondrial Membrane

. Generation of H+ (Proton) Concentration Gradient Across the Inner Mitochondrial Membrane

Synthesis of ATP Using Free Energy Released From Spontaneous Diffusion of H+Back to the Matrix Inside the Inner Mitochondrial Membrane

. Synthesis of ATP Using Free Energy Released From Spontaneous Diffusion of H+

. Synthesis of ATP Using Free Energy Released From Spontaneous Diffusion of H+

The three major steps in oxidative phosphorylation are

(a) oxidation-reduction reactions involving electron transfers between specialized proteins embedded in the inner mitochondrial membrane; 

(b) the generation of a proton (H+) gradient across the inner mitochondrial membrane (which occurs simultaneously with step (a)); and 

(c) the synthesis of ATP using energy from the spontaneous diffusion of electrons down the proton gradient generated in step (b).

Note: Steps (a) and (b) show cytochrome oxidase, the final electron-carrier protein in the electron-transport chain described above. When this protein accepts an electron (green) from another protein in the electron-transport chain, an Fe(III) ion in the center of a heme group (purple) embedded in the protein is reduced to Fe(II). The coordinates for the protein were determined using x-ray crystallography, and the image was rendered using SwissPDB Viewer and POV-Ray (see References).

Cells use a proton-pumping system made up of proteins inside the mitochondria to generate ATP. Before we examine the details of ATP synthesis, we shall step back and look at the big picture by exploring the structure and function of the mitochondria, where oxidative phosphorylation occurs.

Structure and Function of the Mitochondria



This is a schematic diagram showing the membranes of the mitochondrion. The purple shapes on the inner membrane represent proteins, which are described in the section below. An enlargement of the boxed portion of the inner membrane in this figure is shown in Figure.

The mitochondrial membranes are crucial for this organelle’s role in oxidative phosphorylation. As shown in Figure 8, mitochondria have two membranes, an inner and an outer membrane. The outer membrane ispermeable to most small molecules and ions, because it contains large protein channels called porins. The inner membrane is impermeable to most ions and polar molecules. The inner membrane is the site of oxidative phosphorylation. Although the membrane is mostly impermeable, it contains special H+ (proton) channels and pumps that enable the coupling of the redox reaction involving NADH and O2 (Equations 9-10) to the phosphorylation reaction of ADP (Equation 8), as described below (“Oxidation-Reduction Reactions and Proton Pumping in Oxidative Phosphorylation”). (Recall the discussion of protein channels in the “Maintaining the Body’s Chemistry: Dialysis in the Kidneys” Tutorial .)

As shown in Figure 8, inside the inner membrane is a space known as the matrix; the space between the two membranes is known as the intermembrane space. The matrix side of the inner membrane has a negative electrical charge relative to the intermembrane space due to an H+ gradient set up by the redox reaction (Equations 9 and 10). This charge difference is used to provide free energy (G) for the phosphorylation reaction (Equation 8).

Oxidation-Reduction Reactions and Proton Pumping in Oxidative Phosphorylation

Phosphorylation of ADP (Equation 8) is coupled to the oxidation-reduction reaction of NADH and O2 (Equations 9 and 10). Electrons are not transferred directly from NADH to O2, but rather pass through a series of intermediate electron carriers in the inner membrane of the mitochondrion. Why? This allows something very important to occur: the pumping of protons across the inner membrane of the mitochondrion. As we shall see, it is this proton pumping that is ultimately responsible for coupling the oxidation-reduction reaction to ATP synthesis.

Two major types of mitochondrial proteins (see Figure 9, below) are required for oxidative phosphorylation to occur. Both classes of proteins are located in the inner mitochondrial membrane.

  1. The electron carriers (NADH-Q reductase, ubiquinone (Q), cytochrome reductase, cytochrome c, and cytochrome oxidase shown in shades of purple in Figure 9 below) transport electrons in a stepwise fashion from NADH to O2.  Three of these carriers (NADH-Q reductase, cytochrome reductase, and cytochrome oxidase) are also proton pumps, and simultaneously pump H+ ions (protons) from the matrix to the intermembrane space. (Proton movement from one side of the membrane to the other is shown as blue arrows in Figure 9, below.) The protons that are pumped across the membrane complete the redox reaction (Equations 9 and 10). The creation of a proton gradient across the membrane is one way of storing free energy.
  2. ATP synthetase (shown in red in Figure 9 below) allows H+ ions to diffuse back into the matrix and uses the free energy released to synthesize ATP from ADP and HPO42-. The ATP synthetase is essential for the phosphorylation to occur (Equation 8). (Proton movement from one side of the membrane to the other is shown as blue arrows in Figure 9, below.)

The electron carriers can be divided into three protein complexes (NADH-Q reductase (1), cytochrome reductase (3), and cytochrome oxidase (5)) that pump protons from the matrix to the intermembrane space, and two mobile carriers (ubiquinone (2) and cytochrome c (4)) that transfer electrons between the three proton-pumping complexes. (Gold numbers refer to the labels on each protein in Figure 9, below.) Because electrons move from one carrier to another until they are finally transferred to O2, the electron carriers (shown in Figure 9,below) are said to form an electron-transport chain.

Figure  below, is a schematic representation of the proteins involved in oxidative phosphorylation. To see an animation of oxidative phosphorylation, click on “View the Movie.”

Proteins of inner space

Proteins of inner space

This is a schematic diagram illustrating the transfer of electrons from NADH, through the electron carriers in the electron transport chain, to molecular oxygen. Please click on the pink button below to view a QuickTime animation of the functions of the proteins embedded in the inner mitochondrial membrane that are necessary for oxidative phosphorylation. Click the blue button below to download QuickTime 4.0 to view the movie.

NADH-Q reductase (1), cytochrome reductase (3) , and cytochrome oxidase (5) are electron carriers as well as proton pumps, using the energy gained from each electron-transfer step to move protons (H+) against a concentration gradient, from the matrix to the intermembrane space.Ubiquinone (Q) (2) and cytochrome c (Cyt C) (4) are mobile electron carriers. (Ubiquinone is not actually a protein.) All of the electron carriers are shown in purple, with lighter shades representing increasingly higher reduction potentials. Together, these electron carriers form a “chain” to transport electrons from NADH to O2. The path of the electrons is shown with the green dotted line.

ATP synthetase (red) has two components: a proton channel (allowing diffusion of protons down a concentration gradient, from the intermembrane space to the matrix), and a catalytic component to catalyze the formation of ATP.

For a more complete description of each step in oxidative phosphorylation (indicated by the gold numbers), click here.

view movie

view movie

Click here for a brief description of each of the electron carriers in the electron-transport chain. It is important to note that, although NADH donates two electrons and O2 ultimately accepts four electrons, each of the carriers can only transfer one electron at a time. Hence, there are several points along the chain where electrons can be collected and dispersed. For the sake of simplicity, these points are not described in this tutorial.

In the section above, we see that the oxidation-reduction process is a series of electron transfers that occurs spontaneously and produces a proton gradient. Why are the electron tranfers from one electron carrier to the next spontaneous?

What causes electrons to be transferred down the electron-transport chain?

As seen in Table 2, below, and Figure 7a, in these carriers, the species being oxidized or reduced is Fe, which is found either in a iron-sulfur (Fe-S) group or in a heme group. (Recall the heme group from the Chem 151 tutorial “Hemoglobin and the Heme Group: Metal Complexes in the Blood“.) The iron in these groups is alternately oxidized and reduced between Fe(II) (reduced) or Fe(III) (oxidized) states.

Table 2 shows that the electrons are transferred through the electron-transport chain because of the difference in the reduction potential of the electron carriers. As explained in the green box below, the higher the electrical potential (e) of a reduction half reaction is, the greater the tendency is for the species to accept an electron. Hence, in the electron-transport chain, electrons are transferred spontaneously from carriers whose reduction results in a small electrical potential change to carriers whose reduction results in an increasingly larger electrical potential change.

Reduction Potentials and Relationship to Free Energy

An oxidation-reduction reaction consists of an oxidation half reaction and a reduction half reaction. Every half reaction has an electrical potential (e). By convention, all half reactions are written as reductions, and the electrical potential for an oxidation half-reaction is equal in magnitude, but opposite in sign, to the electrical potential for the corresponding reduction (i.e., the opposite reaction). The electrical potential for an oxidation-reduction reaction is calculated by

erxn = eoxidation + ereduction (12)

For example, for the overall reaction of the oxidation of NADH paired with the reduction of O2, the potential can be calculated as shown below.

Reduction Potentials ereduction
NAD+ + 2H+ + 2e –> NADH + H+ -0.32 V
(1/2) O2 + 2H+ + 2e –> H2O +0.82 V

The overall reaction is

NADH + H–> NAD+ + 2H+ + 2e eoxidation = 0.32 V
(1/2) O2 + 2H+ + 2e –> H2O ereduction = 0.82 V
net: NADH + (1/2)O2 + H+ –>
H2O + NAD+
erxn = 1.14 V

The electrical potential (erxn) is related to the free energy (DG) by the following equation:

DG= -nFerxn (13)

where n is the number of electrons transferred (in moles, from the balanced equation), and F is the Faraday constant (96,485 Coulombs/mole). (Using this equation, DG is given in Joules; one Joule = 1 Volt x 1 Coulomb.)

Hence the overall reaction for the oxidation of NADH paired with the reduction of O2 has a negative change in free energy (DG =-220 kJ); i.e., it is spontaneous. Thus, the higher the electrical potential of a reduction half reaction, the greater the tendency for the species to accept an electron.

Just as in the box above, the electrical potential for the overall reaction (electron transfer) between two electron carriers is the sum of the potentials for the two half reactions. As long as the potential for the overall reaction is positive the reaction is spontaneous. Hence, from Table 2 below, we see that cytochrome c1 (part of the cytochrome reductase complex, #3 in Figure 9) can spontaneously transfer an electron to cytochrome c (#4 in Figure 9). The net reaction is given by Equation 16, below.

reduced cytochrome c–> oxidized cytochrome c+ e eoxidation = – .220 V (14)
oxidized cytochrome c + e –> reduced cytochrome c ereduction = .250 V (15)
NET: reduced cyt c1 + oxidized cyt c –>
oxidized cyt c+ reduced cyt c
erxn = 0.030 V (16) Spontaneous

We can also see from Table 2 that cytochrome c1 cannot spontaneously transfer an electron to cytochrome b (Equation 19):

reduced cyt c–> oxidized cyt c+ e eoxidation = – .220 V (17)
oxidized cyt b + e –> reduced cyt b ereduction = – 0.34 V (18)
NET: reduced cyt c1 + oxidized cyt c –>
oxidized cyt c+ reduced cyt c
erxn = – 0.56 V (19) NOT Spontaneous

Table 2 lists the reduction potentials for each of the cytochrome proteins (i.e., the last three steps in the electron-transport chain before the electrons are accepted by O2) involved in the electron-transport chain. Note that each electron transfer is to a cytochrome with a higher reduction potential than the previous cytochrome. As described in the box above and seen in Equations 14-19, an increase in potential leads to a decrease in DG (Equation 13), and thus the transfer of electrons through the chain is spontaneous.

Complex Name Half Reaction Reduction Potential
Cytochrome reductase

(also known as cytochrome b-c1 complex)

(3 in Figure 9)

Cytochrome b (Fe(III) center)
+ e –>
Cytochrome b (Fe(II) center)
-0.34 V
(at pH 7, T=30oC)
Cytochrome c1 (Fe(III) center)
+ e– –>
Cytochrome c1 (Fe(II) center)
+0.220 V
(at pH 7, T=30oC)
Cytochrome c

(4 in Figure 9)

Cytochrome c (Fe(III) center)
+ e– –>
Cytochrome c (Fe(II) center)
+0.250 V
(at pH 7, T=30oC)
Cytochrome oxidase

(5 in Figure 9)

Cytochrome oxidase
( Fe(III) center) + e– –>
Cytochrome oxidase
(Fe(II) center)
+0.285 V
(at pH 7.4, T=25oC)
Table 2

To view the cytochrome molecules interactively using RASMOL, please click on the name of the complex to download the pdb file.

Hence, the electron-transport chain (which works because of the difference in reduction potentials) leads to a large concentration gradient for H+. As we shall see below, this huge concentration gradient leads to the production of ATP.

Questions on Electron Carriers: Steps in the Electron-Transport Chain; Reduction Potentials and Relationship to Free Energy

  • Briefly, explain why electrons travel from NADH-Q reductase, to ubiquinone (Q), to cytochrome reductase, rather than in the opposite direction.
  • One result of the transfer of electrons from NADH-Q reductase down the electron transport chain is that the concentration of protons (H+ ions) in the intermembrane space is increased.  Could cells move protons (H+ ions) from the matrix to the intermembrane space without transporting electrons?  Why or why not?

 ATP Synthetase: Production of ATP

We have seen that the electron-transport chain generates a large proton gradient across the inner mitochondrial membrane. But recall that the ultimate goal of oxidative phosphorylation is to generate ATP to supply readily-available free energy for the body. How does this occur? In addition to the electron-carrier proteins embedded in the inner mitochondrial membrane, a special protein called ATP synthetase (Figure 9, the red-colored protein) is also embedded in this membrane. ATP synthetase uses the proton gradient created by the electron-transport chain to drive the phosphorylation reaction that generates ATP (Figure 7c).

ATP synthetase is a protein consisting of two important segments: a transmembrane proton channel, and a catalytic component located inside the matrix. The proton-channel segment allows H+ ions to diffuse from the intermembrane space, where the concentration is high, to the matrix, where the concentration is low. Recall from the Kidney Dialysis tutorial that particles spontaneously diffuse from areas of high concentration to areas of low concentration. Thus, since the diffusion of protons through the channel component of ATP synthetase is spontaneous, this process is accompanied by a negative change in free energy (i.e., free energy is released). The catalytic component of ATP synthetase has a site where ADP can enter. Then, using the free energy released by the spontaneous diffusion of protons through the channel segment, a bond is formed between the ADP and a free phosphate group, creating an ATP molecule. The ATP is then released from the reaction site, and a new ADP molecule can enter in order to be phosphorylated.

Questions on ATP Synthetase: Production of ATP

  • A scientist has created a phospholipid-bilayer membrane containing ATP-synthetase proteins. Instead of a proton gradient, this scientist has created a large Cs+ gradient (many Cs+ ions on the side of the membrane without the catalytic unit, and few Cs+ ions on the side of the membrane with the catalytic unit). Would you expect the ATP-synthetase proteins in this membrane to be able to generate ATP, given an abundant supply of ADP and phosphate? Briefly, explain your answer. (HINT: Draw on your knowledge of the structure of protein channels to predict what effect replacing H+ ions with Cs+ ions would have.)
  • Certain toxins allow H+ ions to move freely across the inner mitochondrial membrane (i.e., without needing to pass through the channel in ATP synthetase). What effect do you expect these toxins to have on the production of ATP? Briefly, explain your answer.


In this tutorial, we have learned that the ability of the body to perform daily activities is dependent on thermodynamic, equilibrium, and electrochemical concepts.   These activities, which are typically based on nonspontaneous chemical reactions, are performed by using free-energy currency. The common free-energy currency is ATP, which is a molecule that easily dephosphorylates (loses a phosphate group) and releases a large amount of free energy. In the body, the nonspontaneous reactions are coupled to this very spontaneous dephosphorylation reaction, thereby making the overall reaction spontaneous (DG < 0). As the coupled reactions occur (i.e., as the body performs daily activities), ATP is consumed and the body regenerates ATP by using energy from the food we eat (Figure 3). As seen in Figure 4, the breakdown of glucose (glycolysis) obtained from the food we eat cannot by itself generate the large amount of ATP that is needed for metabolic energy by the body. However, glycolysis and the subsequent step, the citric-acid cycle, produce two easily oxidized molecules: NADH and FADH2. These redox molecules are used in an oxidative-phosphorylation process to produce the majority of the ATP that the body uses. This oxidative-phosphorylation process consists of two steps: the oxidation of NADH (or FADH2) and the phosphorylation reaction which regenerates ATP. Oxidative phosphorylation occurs in the mitochondria, and the two reactions (oxidation of NADH or FADHand phosphorylation to generate ATP) are coupled by a proton gradient across the inner membrane of the mitochondria (Figure 9). As seen in Figures 7 and 9, the oxidation of NADH occurs by electron transport through a series of protein complexes located in the inner membrane of the mitochondria. This electron transport is very spontaneous and creates the proton gradient that is necessary to then drive the phosphorylation reaction that generates the ATP. Hence, oxidative-phosphorylation demonstrates that free energy can be easily transferred by proton gradients. Oxidative-phosphorylation is the primary means of generating free-energy currency for aerobic organisms, and as such is one of the most important subjects in the study of bioenergetics (the study of energy and its chemical changes in the biological world).

Additional Link:

  • This fun description of oxidative phosphorylation by Dr. E.J.Oakeley contains step-by-step animated illustrations of the redox reactions involved, as well as a quiz to test your understanding of the material.


Alberts, B. et al. In Molecular Biology of the Cell, 3rd ed., Garland Publishing, Inc.: New York, 1994, pp. 653-684.

Becker, W.M. and Deamer, D.W. In The World of the Cell, 2nd ed., The Benjamin/Cummings Publishing Co., Inc.: Redwood City, CA, 1991, pp. 291-307.

Fasman, G.D. In Handbook of Biochemistry and Molecular Biology, 3rd ed., CRC Press, Inc.: Cleveland, OH, 1976, Vol. I (Physical and Chemical Data), pp. 132-137.

Guex, N. and Peitsch, M.C. Electrophoresis, 1997, 18, 2714-2723. (SwissPDB Viewer) URL:

Moa, C., Ozer, Z., Zhou, M. and Uckun, F. X-Ray Structure of Glycerol Kinase Complexed with an ATP Analog Implies a Novel Mechanism for the ATP-Dependent Gylcerol Phosphorylation by Glycerol Kinase.Biochemical and Biophysical Reaearch Communications. 1999, 259, 640-644.

Persistence of Vision Ray Tracer (POV-Ray). URL:

Stryer, L. In Biochemistry, 4th. ed., W.H. Freeman and Co.: New York, 1995, pp. 490, 509, 513, 529-557.

Zubay, G. Biochemistry, 3rd. ed., Wm. C. Brown Publishers: Dubuque, IA, 1983, p. 42.


The authors thank Dewey Holten (Washington University in St. Louis) for many helpful suggestions in the writing of this tutorial.

The development of this tutorial was supported by a grant from the Howard Hughes Medical Institute, through the Undergraduate Biological Sciences Education program, Grant HHMI# 71199-502008 to Washington University.

Copyright 1999, Washington University, All Rights Reserved.




Read Full Post »

« Newer Posts - Older Posts »