Posts Tagged ‘conformational change’

Metformin, Thyroid-Pituitary Axis, Diabetes Mellitus, and Metabolism

Metformin, Thyroid-Pituitary Axis, Diabetes Mellitus, and Metabolism

Larry H, Bernstein, MD, FCAP, Author and Curator
and Aviva Lev-Ari, PhD, RN

http://pharmaceuticalintelligence.com/9/27/2014/Metformin,_thyroid-pituitary_ axis,_diabetes_mellitus,_and_metabolism

The following article is a review of the central relationship between the action of
metformin as a diabetic medication and its relationship to AMPK, the important and
essential regulator of glucose and lipid metabolism under normal activity, stress, with
its effects on skeletal muscle, the liver, the action of T3 and more.

We start with a case study and a publication in the J Can Med Assoc.  Then we shall look
into key literature on these metabolic relationships.

Part I.  Metformin , Diabetes Mellitus, and Thyroid Function

Hypothyroidism, Insulin resistance and Metformin
May 30, 2012   By Janie Bowthorpe
The following was written by a UK hypothyroid patient’s mother –
Sarah Wilson.

My daughter’s epilepsy is triggered by unstable blood sugars. And since taking
Metformin to control her blood sugar, she has significantly reduced the number of
seizures. I have been doing research and read numerous academic medical journals,
which got me thinking about natural thyroid hormone and Hypothyroidism. My hunch
was that when patients develop hypothyroid symptoms, they are actually becoming
insulin resistant (IR). There are many symptoms in common between women with
polycystic ovaries and hypothyroidism–the hair loss, the weight gain, etc.

A hypothyroid person’s body behaves as if it’s going into starvation mode and so, to
preserve resources and prolong life, the metabolism changes. If hypothyroid is prolonged
or pronounced, then perhaps, chemical preservation mode becomes permanent even
with the reintroduction of thyroid hormones. To get back to normal, they need
a “jump-start” reinitiate a higher rate of metabolism. The kick start is initiated through
AMPK, which is known as the “master metabolic regulating enzyme.”
(http://en.wikipedia.org/wiki/AMP-activated protein kinase).

Guess what? This is exactly what happens to Diabetes patients when Metformin is
introduced. http://en.wikipedia.org/wiki/Metformin
Suggested articles: http://www.springerlink.com/content/r81606gl3r603167/  and

Note the following comments/partial statements:
“Hypothyroidism is characterized by decreased insulin responsiveness”;
“the pivotal regulatory role of T3 in major metabolic pathways”.

The community knows that T3/NTH (natural thyroid hormone [Armour]) makes
hypothyroid patients feel better – but the medical establishment is averse to T3/NTH
(treating subclinical hypoT (T3/T4 euthyroid) with natural dessicated thyroid (NDT).
The medical establishment might find an alternative view about impaired metabolism
more if shown real proof that the old NDT **was/is** having the right result –i.e., the
T3 is jump-starting the metabolism by re-activating

If NDT also can be used for hypothyroidism without the surmised “dangers” of NTH,
then they should consider it. [The reality in the choice is actually recombinant TH
(Synthroid)]. Metformin is cheap, stable and has very few serious side effects. I use the
car engine metaphor, and refer to glucose as our petrol, AMPK as the spark plug and
both T3 and Metformin as the ignition switches. Sometimes if you have flat batteries in
the car, it doesn’t matter how much you turn the ignition switch or pump the petrol
pedal, all it does is flatten the battery and flood the engine.

Dr. Skinner in the UK has been treating “pre-hypothyroidism” the way that some
doctors treat “pre-diabetes”. Those hypothyroid patients who get treated early
might not have had their AMPK pathways altered and the T4-T3 conversion still works.
There seems to be no reason why thyroid hormone replacement therapy shouldn’t
logically be given to ward off a greater problem down the line.

It’s my belief that there is clear and abundant academic evidence that the AMPK/
Metformin research should branch out to also look at thyroid disease.

Point – direct T3 is kicking the closed -down metabolic process back into life,
just like Metformin does for insulin resistance.
There is serotonin resistance! http://www.ncbi.nlm.nih.gov/pubmed/17250776

Metformin Linked to Risk of Low Levels of Thyroid Hormone

CMAJ (Canadian Medical Association Journal) 09/22/2014

Metformin, the drug commonly for treating type 2 diabetes,

  • is linked to an increased risk of low thyroid-stimulating hormone
    (TSH) levels
  • in patients with underactive thyroids (hypothyroidism),

according to a study in CMAJ (Canadian Medical Association Journal).

Metformin is used to lower blood glucose levels

  • by reducing glucose production in the liver.

previous studies have raised concerns that

  • metformin may lower thyroid-stimulating hormone levels.

Study characteristics:

  1. Retrospective  long-term
  2. 74 300 patient who received metformin and sulfonylurea
  3. 25-year study period.
  4. 5689 had treated hypothyroidism
  5. 59 937 had normal thyroid function.

Metformin and low levels of thyroid-stimulating hormone in
patients with type 2 diabetes mellitus

Jean-Pascal Fournier,  Hui Yin, Oriana Hoi Yun Yu, Laurent Azoulay  +
Centre for Clinical Epidemiology (Fournier, Yin, Yu, Azoulay), Lady Davis Institute,
Jewish General Hospital; Department of Epidemiology, Biostatistics and Occupational
Health (Fournier), McGill University; Division of Endocrinology (Yu), Jewish General
Hospital; Department of Oncology (Azoulay), McGill University, Montréal, Que., Cananda

CMAJ Sep 22, 2014,   http://dx.doi.org:/10.1503/cmaj.140688


  • metformin may lower thyroid-stimulating hormone (TSH) levels.


  • determine whether the use of metformin monotherapy, when compared with
    sulfonylurea monotherapy,
  • is associated with an increased risk of low TSH levels(< 0.4 mIU/L)
  • in patients with type 2 diabetes mellitus.


  • Used the Clinical Practice Research Datalink,
  • identified patients who began receiving metformin or sulfonylurea monotherapy
    between Jan. 1, 1988, and Dec. 31, 2012.
  • 2 subcohorts of patients with treated hypothyroidism or euthyroidism,

followed them until Mar. 31, 2013.

  • Used Cox proportional hazards models to evaluate the association of low TSH
    levels with metformin monotherapy, compared with sulfonylurea monotherapy,
    in each subcohort.


  • 5689 patients with treated hypothyroidism and 59 937 euthyroid patients were
    included in the subcohorts.

For patients with treated hypothyroidism:

  1. 495 events of low TSH levels were observed (incidence rate 0.1197/person-years).
  2. 322 events of low TSH levels were observed (incidence rate 0.0045/person-years)
    in the euthyroid group.
  • metformin monotherapy was associated with a 55% increased risk of low TSH
    in patients with treated hypothyroidism (incidence rate 0.0795/person-years
    vs.0.1252/ person-years, adjusted hazard ratio [HR] 1.55, 95% confidence
    interval [CI] 1.09– 1.20), compared with sulfonylurea monotherapy,
  • the highest risk in the 90–180 days after initiation (adjusted HR 2.30, 95% CI
  • No association was observed in euthyroid patients (adjusted HR 0.97, 95% CI 0.69–1.36).

Interpretation: The clinical consequences of this needs further investigation.


Crude and adjusted hazard ratios for suppressed thyroid-stimulating hormone
levels (< 0.1 mIU/L) associated with the use metformin monotherapy, compared
with sulfonylurea monotherapy, in patients with treated hypothyroidism or
euthyroidism and type 2 diabetes
Variable No. events
TSH levels
Person-years of
Incidence rate,
per 1000 person-years (95% CI)
Adjusted HR*(95% CI)
Patients with treated hypothyroidism, = 5689
= 762
18 503 35.8
1.00 1.00
= 4927
130 3 633 35.8
1.05 0.99
Euthyroid patients, = 59 937
= 7980
12 8 576 1.4
1.00 1.00
= 51 957
75 63 047 1.2
0.85 1.03


Part II. Metabolic Underpinning 
(Source: Wikipedia, AMPK and thyroid)

5′ AMP-activated protein kinase or AMPK or 5′ adenosine monophosphate-activated protein kinase
is an enzyme that plays a role in cellular energy homeostasis.
It consists of three proteins (subunits) that

  1. together make a functional enzyme, conserved from yeast to humans.
  2. It is expressed in a number of tissues, including the liver, brain, and skeletal
  3. The net effect of AMPK activation is stimulation of
    1. hepatic fatty acid oxidation and ketogenesis,
    2. inhibition of cholesterol synthesis,
    3. lipogenesis, and triglyceride synthesis,
    4. inhibition of adipocyte lipolysis and lipogenesis,
    5. stimulation of skeletal muscle fatty acid oxidation and muscle
      glucose uptake, and
    6. modulation of insulin secretion by pancreatic beta-cells.

The heterotrimeric protein AMPK is formed by α, β, and γ subunits. Each of these three
subunits takes on a specific role in both the stability and activity of AMPK.

  • the γ subunit includes four particular Cystathionine beta synthase (CBS) domains
    giving AMPK its ability to sensitively detect shifts in the AMP:ATP ratio.
  • The four CBS domains create two binding sites for AMP commonly referred to as
    Bateman domains. Binding of one AMP to a Bateman domain cooperatively
    increases the binding affinity of the second AMP to the other Bateman domain.
  • As AMP binds both Bateman domains the γ subunit undergoes a conformational
    change which exposes the catalytic domain found on the α subunit.
  • It is in this catalytic domain where AMPK becomes activated when
    phosphorylation takes place at threonine-172by an upstream AMPK kinase
    (AMPKK). The α, β, and γ subunits can also be found in different isoforms.

AMPK acts as a metabolic master switch regulating several intracellular systems

  1. the cellular uptake of glucose,
  2. the β-oxidation of fatty acids and
  3. the biogenesis of glucose transporter 4 (GLUT4) and
  4. mitochondria

The energy-sensing capability of AMPK can be attributed to

  • its ability to detect and react to fluctuations in the AMP:ATP ratio that take
    place during rest and exercise (muscle stimulation).

During muscle stimulation,

  • AMP increases while ATP decreases, which changes AMPK into a good substrate
    for activation.
  • AMPK activity increases while the muscle cell experiences metabolic stress
    brought about by an extreme cellular demand for ATP.
  • Upon activation, AMPK increases cellular energy levels by
    • inhibiting anabolic energy consuming pathways (fatty acid synthesis,
      protein synthesis, etc.) and
    • stimulating energy producing, catabolic pathways (fatty acid oxidation,
      glucose transport, etc.).

A recent JBC paper on mice at Johns Hopkins has shown that when the activity of brain
AMPK was pharmacologically inhibited,

  • the mice ate less and lost weight.

When AMPK activity was pharmacologically raised (AICAR see below)

  • the mice ate more and gained weight.

Research in Britain has shown that the appetite-stimulating hormone ghrelin also
affects AMPK levels.

The antidiabetic drug metformin (Glucophage) acts by stimulating AMPK, leading to

  1. reduced glucose production in the liver and
  2. reduced insulin resistance in the muscle.

(Metformin usually causes weight loss and reduced appetite, not weight gain and
increased appetite, ..opposite of expected from the Johns Hopkins mouse study results.)

Triggering the activation of AMPK can be carried out provided two conditions are met.

First, the γ subunit of AMPK

  • must undergo a conformational change so as to
  • expose the active site(Thr-172) on the α subunit.

The conformational change of the γ subunit of AMPK can be accomplished

  • under increased concentrations of AMP.

Increased concentrations of AMP will

  • give rise to the conformational change on the γ subunit of AMPK
  • as two AMP bind the two Bateman domains located on that subunit.
  • It is this conformational change brought about by increased concentrations
    of  AMP that exposes the active site (Thr-172) on the α subunit.

This critical role of AMP is further substantiated in experiments that demonstrate

  • AMPK activation via an AMP analogue 5-amino-4-imidazolecarboxamide
    ribotide (ZMP) which is derived fromthe familiar
  • 5-amino-4-imidazolecarboxamide riboside (AICAR)

AMPK is a good substrate for activation via an upstream kinase complex, AMPKK
AMPKK is a complex of three proteins,

  1. STE-related adaptor (STRAD),
  2. mouse protein 25 (MO25), and
  3. LKB1 (a serine/threonine kinase).

The second condition that must be met is

  • the phosphorylation/activation of AMPK on its activating loop at
    Thr-172of the α subunit
  • brought about by an upstream kinase (AMPKK).

The complex formed between LKB1 (STK 11), mouse protein 25 (MO25), and the
pseudokinase STE-related adaptor protein (STRAD) has been identified as

  • the major upstream kinase responsible for phosphorylation of AMPK
    on its activating loop at Thr-172

Although AMPK must be phosphorylated by the LKB1/MO25/STRAD complex,

  • it can also be regulated by allosteric modulators which
  • directly increase general AMPK activity and
  • modify AMPK to make it a better substrate for AMPKK
  • and a worse substrate for phosphatases.

It has recently been found that 3-phosphoglycerate (glycolysis intermediate)

  • acts to further pronounce AMPK activation via AMPKK

Muscle contraction is the main method carried out by the body that can provide
the conditions mentioned above needed for AMPK activation

  • As muscles contract, ATP is hydrolyzed, forming ADP.
  • ADP then helps to replenish cellular ATP by donating a phosphate group to
    another ADP,

    • forming an ATP and an AMP.
  • As more AMP is produced during muscle contraction,
    • the AMP:ATP ratio dramatically increases,
  • leading to the allosteric activation of AMPK

For over a decade it has been known that calmodulin-dependent protein kinase
kinase-beta (CaMKKbeta) can phosphorylate and thereby activate AMPK,

  • but it was not the main AMPKK in liver.

CaMKK inhibitors had no effect on 5-aminoimidazole-4-carboxamide-1-beta-4-
ribofuranoside (AICAR) phosphorylation and activation of AMPK.

  • AICAR is taken into the celland converted to ZMP,
  • an AMP analogthat has been shown to activate AMPK.

Recent LKB1 knockout studies have shown that without LKB1,

  • electrical and AICAR stimulation of muscleresults in very little
    phosphorylation of AMPK and of ACC, providing evidence that
  • LKB1-STRAD-MO25 is the major AMPKK in muscle.

Two particular adipokines, adiponectin and leptin, have even been demonstrated
to regulate AMPK. A main functions of leptin in skeletal muscle is

  • the upregulation of fatty acid oxidation.

Leptin works by way of the AMPK signaling pathway, and adiponectin also
stimulates the oxidation of fatty acids via the AMPK pathway, and

  • Adiponectin also stimulates the uptake of glucose in skeletal muscle.

An increase in enzymes which specialize in glucose uptake in cells such as GLUT4
and hexokinase II are thought to be mediated in part by AMPK when it is activated.
Increases in AMPK activity are brought about by increases in the AMP:ATP ratio
during single bouts of exercise and long-term training.

One of the key pathways in AMPK’s regulation of fatty acid oxidation is the

  • phosphorylation and inactivation of acetyl-CoA carboxylase.
  1. Acetyl-CoA carboxylase (ACC) converts acetyl-CoA (ACA) to malonyl-CoA
    (MCA), an inhibitor of carnitine palmitoyltransferase 1 (CPT-1).
  2. CPT-1 transports fatty acids into the mitochondria for oxidation.
  3. Inactivation of ACC results in increased fatty acid transport and oxidation.
  4. the AMPK induced ACC inactivation  and reduced conversion to MCA
    may occur as a result of malonyl-CoA decarboxylase (MCD)
  5. MCD as an antagonist to ACC, decarboxylatesmalonyl-CoA to acetyl-CoA
    (reversal of ACC conversion of ACA to MCA)
  6. This resultsin decreased malonyl-CoA and increased CPT-1 and fatty acid oxidation.

AMPK also plays an important role in lipid metabolism in the liver. It has long been
known that hepatic ACC has been regulated in the liver.

  1. It phosphorylates and inactivates 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR)
  2. acetyl-CoA(ACA) is converted to mevalonic acid (MVA) by ACC
    with inhibition of CPT-1
  3. HMGR converts 3-hydroxy-3-methylglutaryl-CoA, which is made from MVA
  4. which then travels down several more metabolic steps to become cholesterol.

Insulin facilitates the uptake of glucose into cells via increased expression and
translocation of glucose transporter GLUT-4. In addition, glucose is phosphorylated
by hexokinase wheni iot enters the cell. The phosphorylated form keeps glucose from
leaving the cell,

  • The decreasedthe concentration of glucose molecules creates a gradient for more
    glucose to be transported into the cell.
AMPK and thyroid hormone regulate some similar processes. Knowing these similarities,
Winder and Hardie et al. designed an experiment to see if AMPK was influenced by thyroid
hormone. They found that all of the subunits of AMPK were increased in skeletal muscle,
especially in the soleus and red quadriceps, with thyroid hormone treatment. There was
also an increase in phospho-ACC, a marker of AMPK activity.
  •  Winder WW, Hardie DG (July 1999). “AMP-activated protein kinase,
    a metabolic master switch: possible roles in type 2 diabetes”. J. Physiol. 277
    (1 Pt 1): E1–10. PMID 10409121.
  • Winder WW, Hardie DG (February 1996). “Inactivation of acetyl-CoA
    carboxylase and activation of AMP-activated protein kinase in muscle
    during exercise”. J. Physiol. 270 (2 Pt 1): E299–304. PMID 8779952.
  • Hutber CA, Hardie DG, Winder WW (February 1997). “Electrical stimulation
    inactivates muscle acetyl-CoA carboxylase and increases AMP-activated
    protein kinase”. Am. J. Physiol. 272 (2 Pt 1): E262–6. PMID 9124333
  • Durante PE, Mustard KJ, Park SH, Winder WW, Hardie DG (July 2002).
    “Effects of endurance training on activity and expression of AMP-activated
    protein kinase isoforms in rat muscles”. Am. J. Physiol. Endocrinol.
    Metab. 283 (1): E178–86. doi:10.1152/ajpendo.00404.2001. PMID 12067859
  • Corton JM, Gillespie JG, Hardie DG (April 1994). “Role of the AMP-activated
    protein kinase in the cellular stress response”. Curr. Biol. 4 (4):
    315–24. doi:10.1016/S0960-9822(00)00070-1. PMID 7922340
  • Winder WW (September 2001). “Energy-sensing and signaling by
    AMP-activated protein kinase in skeletal muscle”. J. Appl. Physiol. 91 (3):
    1017–28. PMID 11509493
  • Suter M, Riek U, Tuerk R, Schlattner U, Wallimann T, Neumann D (October
    2006). “Dissecting the role of 5′-AMP for allosteric stimulation, activation,
    and deactivation of AMP-activated protein kinase”.  J. Biol. Chem.
    281 (43): 32207–6. doi:10.1074/jbc.M606357200. PMID 16943194


Part III. Pituitary-thyroid axis and diabetes mellitus
The Interface Between Thyroid and Diabetes Mellitus

Leonidas H. Duntas, Jacques Orgiazzi, Georg Brabant   Clin Endocrinol. 2011;75(1):1-9.
Interaction of Metformin and Thyroid Function

Metformin acts primarily by

  • suppressing hepatic gluconeogenesis via activation of AMPK
  • It has the opposite effects on hypothalamic AMPK,
    • inhibiting activity of the enzyme.
  • the metformin effects on hypothalamic AMPK activity will
    • counteractT3 effects at the hypothalamic level.
  • AMPK therefore represents a direct target for dual regulation
    • in the hypothalamic partitioning of energy homeostasis.
  • metformin crossesthe blood–brain barrier and
    • levels in the pituitary gland are substantially increased.
  • It convincinglysuppresses TSH

A recent study recruiting 66 patients with benign thyroid nodules furthermore
demonstrated that metformin significantly decreases nodule size in patients with
insulin resistance.[76] The effect of metformin, which was produced over a
6-month treatment period, parallelled a fall in TSH concentrations and achieved a
shrinkage amounting to 30% of the initial nodule size when metformin was
administered alone and up to 55% when it was added to ongoing LT4 treatment.

These studies reveal a

  • suppressive effect of metformin on TSH secretion patterns in
    hypothyroid patients, an effect that is apparently
  • independent of T4 treatment and does not alter the TH profile.
  • A rebound of TSH secretion occurs at about 3 months following metformin

It appears that recommendations for more frequent testing, on an annual to
biannual basis, seems justified in higher risk groups like patients over 50 or 55,
particularly with suggestive symptoms, raised antibody titres or dylipidaemia.
We thus would support the suggestion of an initial TSH and TPO antibody testing
which, as discussed, will help to predict the development of hypothyroidism in
patients with diabetes.

Hypothalamic AMPK and fatty acid metabolism mediate thyroid
regulation of energy 
M López,  L Varela,  MJ Vázquez,  S Rodríguez-Cuenca, CR González, …, & Vidal-Puig
Nature Medicine  29 Aug 2010; 16: 1001–1008 http://dx.doi.org:/10.1038/nm.2207

Thyroid hormones have widespread cellular effects; however it is unclear whether
their effects on the central nervous system (CNS) contribute to global energy balance.
Here we demonstrate that either

  • whole-body hyperthyroidism or central administration of triiodothyronine
    (T3) decreases

    • the activity of hypothalamic AMP-activated protein kinase (AMPK),
    • increases sympathetic nervous system (SNS) activity and
    • upregulates thermogenic markers in brown adipose tissue (BAT).

Inhibition of the lipogenic pathway in the ventromedial nucleus of the hypothalamus
(VMH) prevents CNS-mediated activation of BAT by thyroid hormone and reverses
the weight loss associated with hyperthyroidism. Similarly, inhibition of thyroid
hormone receptors in the VMH reverses the weight loss associated with hyperthyroidism.

This regulatory mechanism depends on AMPK inactivation, as genetic inhibition of this
enzyme in the VMH of euthyroid rats induces feeding-independent weight loss and
increases expression of thermogenic markers in BAT. These effects are reversed by
pharmacological blockade of the SNS. Thus, thyroid hormone–induced modulation
of AMPK activity and lipid metabolism in the hypothalamus is a major regulator of
whole-body energy homeostasis.

Metabolic Basis for Thyroid Hormone Liver Preconditioning:
Upregulation of AMP-Activated Protein Kinase Signaling
LA Videla,1 V Fernández, P Cornejo, and R Vargas
1Molecular and Clinical Pharmacology Program, Institute of Biomedical Sciences,
Faculty of Medicine, University of Chile, 2Faculty of Medicine, Diego Portales University,
Santiago, Chile
Academic Editors: H. M. Abu-Soud and D. Benke
The Scientific World Journal 2012; 2012, ID 475675, 10 pp

The liver is a major organ responsible for most functions of cellular metabolism and

  • a mediator between dietary and endogenous sources of energy for extrahepatic tissues.
  • In this context, adenosine-monophosphate- (AMP-) activated protein kinase (AMPK)
    constitutes an intrahepatic energy sensor
  • regulating physiological energy dynamics by limiting anabolism and stimulating
    catabolism, thus increasing ATP availability.
  • This is achieved by mechanisms involving direct allosteric activation and
    reversible phosphorylation of AMPK, in response to signals such as

    • energy status,
    • serum insulin/glucagon ratio,
    • nutritional stresses,
    • pharmacological and natural compounds, and
    • oxidative stress status.

Reactive oxygen species (ROS) lead to cellular AMPK activation and

  • downstream signaling under several experimental conditions.

Thyroid hormone (L-3,3′,5-triiodothyronine, T3) administration, a condition
that enhances liver ROS generation,

  • triggers the redox upregulation of cytoprotective proteins
    • affording preconditioning against ischemia-reperfusion (IR) liver injury.

Data discussed in this work suggest that T3-induced liver activation of AMPK

  • may be of importance in the promotion of metabolic processes
  • favouring energy supply for the induction and operation of preconditioning

These include

  1. antioxidant,
  2. antiapoptotic, and
  3. anti-inflammatory mechanisms,
  4. repair or resynthesis of altered biomolecules,
  5. induction of the homeostatic acute-phase response, and
  6. stimulation of liver cell proliferation,

which are required to cope with the damaging processes set in by IR.

The liver functions as a mediator between dietary and endogenous sources
of energy and extrahepatic organs that continuously require energy, mainly
the brain and erythrocytes, under cycling conditions between fed and fasted states.

In the fed state, where insulin action predominates, digestion-derived glucose is
converted to pyruvate via glycolysis, which is oxidized to produce energy, whereas
fatty acid oxidation is suppressed. Excess glucose can be either stored as hepatic
glycogen or channelled into de novo lipogenesis.

In the fasted state, considerable liver fuel metabolism changes occur due to decreased
serum insulin/glucagon ratio, with higher glucose production as a consequence of
stimulated glycogenolysis and gluconeogenesis (from alanine, lactate, and glycerol).

Major enhancement in fatty acid oxidation also occurs to provide energy for liver
processes and ketogenesis to supply metabolic fuels for extrahepatic tissues. For these
reasons, the liver is considered as the metabolic processing organ of the body, and
alterations in liver functioning affect whole-body metabolism and energy homeostasis.

In this context, adenosine-monophosphate- (AMP-) activated protein kinase (AMPK)
is the downstream component of a protein kinase cascade acting as an

  • intracellular energy sensor regulating physiological energy dynamics by
  • limiting anabolic pathways, to prevent excessive adenosine triphosphate (ATP)
    utilization, and
  • by stimulating catabolic processes, to increase ATP production.

Thus, the understanding of the mechanisms by which liver AMPK coordinates hepatic
energy metabolism represents a crucial point of convergence of regulatory signals
monitoring systemic and cellular energy status

Liver AMPK: Structure and Regulation

AMPK, a serine/threonine kinase, is a heterotrimeric complex comprising

  1. a catalytic subunit α and
  2. two regulatory subunits β and γ .

The α subunit has a threonine residue (Thr172) within the activation loop of the kinase
domain, with the C-terminal region being required for association with β and γ subunits.
The β subunit associates with α and γ by means of its C-terminal region , whereas

  • the γ subunit has four cystathionine β-synthase (CBS) motifs, which
  • bind AMP or ATP in a competitive manner.

75675.fig.001 (not shown)

Figure 1: Regulation of AMP-activated protein kinase (AMPK) by
(A) direct allosteric activation and
(B) reversible phosphorylation and downstream responses maintaining
intracellular energy balance.

Regulation of liver AMPK activity involves both direct allosteric activation and
reversible phosphorylation. AMPK is allosterically activated by AMP through

  • binding to the regulatory subunit-γ, which induces a conformational change in
    the kinase domain of subunit α that protects AMPK from dephosphorylation
    of Thr172, probably by protein phosphatase-2C.

Activation of AMPK requires phosphorylation of Thr172 in its α subunit, which can be
attained by either

(i) tumor suppressor LKB1 kinase following enhancement in the AMP/ATP ratio, a
kinase that plays a crucial role in AMPK-dependent control of liver glucose and
lipid metabolism;

(ii) Ca2+-calmodulin-dependent protein kinase kinase-β (CaMKKβ) that
phosphorylates AMPK in an AMP-independent, Ca2+-dependent manner;

(iii) transforming growth-factor-β-activated kinase-1 (TAK1), an important
kinase in hepatic Toll-like receptor 4 signaling in response to lipopolysaccharide.

Among these kinases, the relevance of CaMKKβ and TAK1 in liver AMPK activation
remains to be established in metabolic stress conditions. Both allosteric and
phosphorylation mechanisms are able to elicit

  • over 1000-fold increase in AMPK activity, thus allowing
  • the liver to respond to small changes in energy status in a highly sensitive fashion.

In addition to rapid AMPK regulation through allosterism and reversible phosphorylation

  • long-term effects of AMPK activation induce changes in hepatic gene expression.

This was demonstrated for

(i) the transcription factor carbohydrate-response element-binding protein (ChREBP),

  • whose Ser568 phosphorylation by activated AMPK
  • blocks its DNA binding capacity and glucose-induced gene transcription
  • under hyperlipidemic conditions;(ii) liver sterol regulatory element-binding
    protein-1c (SREBP-1c), whose mRNA and protein expression and those of
    its target gene for fatty acid synthase (FAS)
  • are reduced by metformin-induced AMPK activation,
  • decreasing lipogenesis and increasing fatty acid oxidation due to
    malonyl-CoA depletion;

(iii) transcriptional coactivator transducer of regulated CREB activity-2 (TORC2),
a crucial component of the hepatic gluconeogenic program, was reported
to be phosphorylated by activated AMPK.

This modification leads to subsequent cytoplasmatic sequestration of TORC2 and
inhibition of gluconeogenic gene expression, a mechanism underlying

  • the plasma glucose-lowering effects of adiponectin and metformin
  • through AMPK activation by upstream LKB1.

Activation of AMPK in the liver is a key regulatory mechanism controlling glucose
and lipid metabolism,

  1. inhibiting anabolic processes, and
  2. enhancing catabolic pathways in response to different signals, including
    1. energy status,
    2. serum insulin/glucagon ratio,
    3. nutritional stresses,
    4. pharmacological and natural compounds, and
    5. oxidative stress status

Reactive Oxygen Species (ROS) and AMPK Activation

The high energy demands required to cope with all the metabolic functions
of the liver are met by

  • fatty acid oxidation under conditions of both normal blood glucose levels and
    hypoglycemia, whereas
  • glucose oxidation is favoured in hyperglycemic states, with consequent
    generation of ROS.

Under normal conditions, ROS occur at relatively low levels due to their fast processing
by antioxidant mechanisms, whereas at acute or prolonged high ROS levels, severe
oxidation of biomolecules and dysregulation of signal transduction and gene expression
is achieved, with consequent cell death through necrotic and/or apoptotic-signaling

Thyroid Hormone (L-3,3′,5-Triiodothyronine, T3), Metabolic Regulation,
and ROS Production

T3 is important for the normal function of most mammalian tissues, with major actions
on O2 consumption and metabolic rate, thus

  • determining enhancement in fuel consumption for oxidation processes
  • and ATP repletion.

T3 acts predominantly through nuclear receptors (TR) α and β, forming

  • functional complexes with retinoic X receptor that
  • bind to thyroid hormone response elements (TRE) to activate gene expression.

T3 calorigenesis is primarily due to the

  • induction of enzymes related to mitochondrial electron transport and ATP
    synthesis, catabolism, and
  • some anabolic processes via upregulation of genomic mechanisms.

The net result of T3 action is the enhancement in the rate of O2 consumption of target
tissues such as liver, which may be effected by secondary processes induced by T3

(i) energy expenditure due to higher active cation transport,

(ii) energy loss due to futile cycles coupled to increase in catabolic and anabolic pathways, and

(iii) O2 equivalents used in hepatic ROS generation both in hepatocytes and Kupffer cells

In addition, T3-induced higher rates of mitochondrial oxidative phosphorylation are
likely to induce higher levels of ATP, which are partially balanced by intrinsic uncoupling
afforded by induction of uncoupling proteins by T3. In agreement with this view, the
cytosolic ATP/ADP ratio is decreased in hyperthyroid tissues, due to simultaneous
stimulation of ATP synthesis and consumption.

Regulation of fatty acid oxidation is mainly attained by carnitine palmitoyltransferase Iα (CPT-Iα),

  • catalyzing the transport of fatty acids from cytosol to mitochondria for β-oxidation,
    and acyl-CoA oxidase (ACO),
  • catalyzing the first rate-limiting reaction of peroxisomal β-oxidation, enzymes that are
    induced by both T3 and peroxisome proliferator-activated receptor α (PPAR-α).

Furthermore, PPAR-α-mediated upregulation of CPT-Iα mRNA is enhanced by PPAR-γ
coactivator 1α (PGC-1α), which in turn

  • augments T3 induction of CPT-Iα expression.

Interestingly, PGC-1α is induced by

  1. T3,
  2. AMPK activation, and
  3. ROS,

thus establishing potential links between

  • T3 action, ROS generation, and AMPK activation

with the onset of mitochondrial biogenesis and fatty acid β-oxidation.

Liver ROS generation leads to activation of the transcription factors

  1. nuclear factor-κB (NF-κB),
  2. activating protein 1 (AP-1), and
  3. signal transducer and activator of transcription 3 (STAT3)

at the Kupffer cell level, with upregulation of cytokine expression (TNF-α, IL-1, IL-6),
which upon interaction with specific receptors in hepatocytes trigger the expression of

  1. cytoprotective proteins (Figure 3(A)).

These responses and the promotion of hepatocyte and Kupffer-cell proliferation
represent hormetic effects reestablishing

  1. redox homeostasis,
  2. promoting cell survival, and
  3. protecting the liver against ischemia-reperfusion injury.

T3 liver preconditioning also involves the activation of the

  1. Nrf2-Keap1 defense pathway
  • upregulating antioxidant proteins,
  • phase-2 detoxifying enzymes, and
  • multidrug resistance proteins, members of the ATP binding cassette (ABC)
    superfamily of transporters (Figure 3(B))

In agreement with T3-induced liver preconditioning, T3 or L-thyroxin afford
preconditioning against IR injury in the heart, in association with

  • activation of protein kinase C and
  • attenuation of p38 and
  • c-Jun-N-terminal kinase activation ,

and in the kidney, in association with

  • heme oxygenase-1 upregulation.



Figure 2: Calorigenic response of thyroid hormone (T3) and its relationship with O2
consumption, reactive oxygen species (ROS) generation, and antioxidant depletion in the liver.
Abbreviations: CYP2E1, cytochrome P450 isoform 2E1; GSH, reduced glutathione; QO2, rate
of O2 consumption; SOD, superoxide dismutase.


genomic signaling in T3 calorigenesis and ROS production 475675.fig.003

genomic signaling in T3 calorigenesis and ROS production 475675.fig.003


Figure 3: Genomic signaling mechanisms in T3 calorigenesis and liver reactive oxygen
species (ROS) production leading to
(A) upregulation of cytokine expression in Kupffer cells and hepatocyte activation of genes
conferring cytoprotection,
(B) Nrf2 activation controling expression of antioxidant and detoxication proteins, and
(C) activation of the AMPK cascade regulating metabolic functions.Abbreviations: AP-1, activating protein 1; ARE, antioxidant responsive element; CaMKKβ,
Ca2+-calmodulin-dependent kinase kinase-β; CBP, CREB binding protein; CRC, chromatin
remodelling complex; EH, epoxide hydrolase; HO-1, hemoxygenase-1; GC-Ligase,
glutamate cysteine ligase; GPx, glutathione peroxidase; G-S-T, glutathione-S-transferase;
HAT, histone acetyltransferase; HMT, histone arginine methyltransferase; IL1,
interleukin 1; iNOS, inducible nitric oxide synthase; LKB1, tumor suppressor LKB1 kinase;
MnSOD, manganese superoxide dismutase; MRPs, multidrug resistance proteins; NF-κB,
nuclear factor-κB; NQO1, NADPH-quinone oxidoreductase-1; NRF-1, nuclear respiratory
factor-1; Nrf2, nuclear receptor-E2-related factor 2; PCAF, p300/CBP-associated
factor; RXR, retinoic acid receptor; PGC-1, peroxisome proliferator-activated receptor-γ
coactivator-1; QO2, rate of O2 consumption; STAT3, signal transducer and activator
of transcription 3; TAK1, transforming-growth-factor-β-activated kinase-1; TNF-α, tumor
necrosis factor-α; TR, T 3 receptor; TRAP, T3-receptor-associated protein; TRE,  T3 responsive element; UCP, uncoupling proteins; (—), reported mechanisms;
(- - - -), proposed mechanisms.


T3 is a key metabolic regulator coordinating short-term and long-term energy needs,
with major actions on liver metabolism. These include promotion of

(i) gluconeogenesis and hepatic glucose production, and

(ii) fatty acid oxidation coupled to enhanced adipose tissue lipolysis, with

  • higher fatty acid flux to the liver and
  • consequent ROS production (Figure 2) and
  • redox upregulation of cytoprotective proteins

affording liver preconditioning (Figure 3).

Thyroid Hormone and AMPK Activation: Skeletal Muscle and Heart

In skeletal muscle, T3 increases the levels of numerous proteins involved in

  1. glucose uptake (GLUT4),
  2. glycolysis (enolase, pyruvate kinase, triose phosphate isomerase),
  3. fatty acid oxidation (carnitine palmitoyl transferase-1, mitochondrial thioesterase I),
    and uncoupling protein-3,

effects that are achieved through enhanced transcription of TRE-containing genes

Skeletal muscle AMPK activation is characterized by

(i) being a rapid and transient response,

(ii) upstream activation by Ca2+-induced mobilization and CaMKKβ activation,

(iii) upstream upregulation of LKB1 expression, which requires association with STRAD
and MO25 for optimal phosphorylation/activation of AMPK, and

(iv) stimulation of mitochondrial fatty acid β-oxidation.

T3-induced muscle AMPK activation was found to trigger two major downstream

signaling pathways, namely,

(i) peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) mRNA
expression and phosphorylation, a transcriptional regulator for genes related to

  • mitochondrial biogenesis,
  • fatty acid oxidation, and
  • gluconeogenesis and

(ii) cyclic AMP response element binding protein (CREB) phosphorylation, which

  • in turn induces PGC-1α expression in liver tissue, thus
  • reinforcing mechanism (i).

These data indicate that AMPK phosphorylation of PGC-1α initiates many of the
important gene regulatory functions of AMPK in skeletal muscle.

In heart, hyperthyroidism increased glycolysis and sarcolemmal GLUT4 levels by the
combined effects of AMPK activation and insulin stimulation, with concomitant increase
in fatty acid oxidation proportional to enhanced cardiac mass and contractile function.

Thyroid Hormone, AMPK Activation, and Liver Preconditioning

Recent studies by our group revealed that administration of a single dose of 0.1 mg T3/kg
to rats activates liver AMPK (Figure 4; unpublished work).

  1. enhancement in phosphorylated AMPK/nonphosphorylated AMPK ratios in T3-
    treated rats over control values thatis significant in the time period of 1 to 48
    hours after hormone treatment
  2. Administration of a substantially higher dose (0.4 mg T3/kg) resulted in
    decreased liver AMPK activation at 4 h to return to control values at 6 h
    after treatment

Activation of liver AMPK by T3 may be of relevance in terms of

  • promotion of fatty acid oxidation for ATP supply,
  • supporting hepatoprotection against IR injury (Figure 3(C)).

This proposal is based on the high energy demands underlying effective liver
preconditioning for full operation of hepatic

  • antioxidant, antiapoptotic, and anti-inflammatory mechanisms,
  • oxidized biomolecules repair or resynthesis,
  • induction of the homeostatic acute-phase response, and
  • promotion of hepatocyte and Kupffer cell proliferation,

mechanisms that are needed to cope with the damaging processes set in by IR.
T3 liver preconditioning , in addition to that afforded by

  • n-3 long-chain polyunsaturated fatty acids given alone or
  • combined with T3 at lower dosages, or
  • by iron supplementation,

constitutes protective strategies against hepatic IR injury.

Studies on the molecular mechanisms underlying T3-induced liver AMPK
activation (Figure 4) are currently under assessment in our laboratory.


Fernández and L. A. Videla, “Kupffer cell-dependent signaling in thyroid hormone
calorigenesis: possible applications for liver preconditioning,” Current Signal
Transduction Therapy 2009; 4(2): 144–151.

Viollet, B. Guigas, J. Leclerc et al., “AMP-activated protein kinase in the regulation
of  hepatic energy metabolism: from physiology to therapeutic perspectives,” Acta
Physiologica 2009; 196(1): 81–98.

Carling, “The AMP-activated protein kinase cascade – A unifying system
for energy control,” Trends in Biochemical Sciences, 2004;. 29(1): 18–24.

E. Kemp, D. Stapleton, D. J. Campbell et al., “AMP-activated protein kinase,
metabolic regulator,” Biochemical Society Transactions 2003; 31(1):

G. Hardie, “AMP-activated protein kinase-an energy sensor that
regulates all ;aspects of cell function,” Genes and Development,
2011; 25(18): 1895–1908.

Woods, P. C. F. Cheung, F. C. Smith et al., “Characterization of AMP-activated
protein kinase βandγ subunits Assembly of the heterotrimeric complex in vitro,”
Journal of Biological Chemistry 1996;271(17): 10282–10290.

Xiao, R. Heath, P. Saiu et al., “Structural basis for AMP binding to mammalian AMP-
activated protein kinase,” Nature 2007; 449(7161): 496–500.


Impact of Metformin and compound C on NIS expression and iodine uptake in vitro and in vivo: a role for CRE in AMPK modulation of thyroid function.
Abdulrahman RM1, Boon MRSips HCGuigas BRensen PCSmit JWHovens GC.
Author information 
Thyroid. 2014 Jan;24(1):78-87.  Epub 2013 Sep 25.  PMID: 23819433

Although adenosine monophosphate activated protein kinase (AMPK) plays a crucial role
in energy metabolism, a direct effect of AMPK modulation on thyroid function has only
recently been reported, and much of its function in the thyroid is currently unknown.

The aim of this study was

  1. to investigate the mechanism of AMPK modulation in iodide uptake.
  2. to investigate the potential of the AMPK inhibitor compound C as an enhancer of
    iodide uptake by thyrocytes.

Metformin reduced NIS promoter activity (0.6-fold of control), whereas compound C
stimulated its activity (3.4-fold) after 4 days. This largely coincides with

  • CRE activation (0.6- and 3.0-fold).

These experiments show that AMPK exerts its effects on iodide uptake, at least partly,
through the CRE element in the NIS promoter. Furthermore, we have used AMPK-alpha1
knockout mice to determine the long-term effects of AMPK inhibition without chemical compounds.
These mice have a less active thyroid, as shown by reduced colloid volume and reduced
responsiveness to thyrotropin.

NIS expression and iodine uptake in thyrocytes

  • can be modulated by metformin and compound C.

These compounds exert their effect by

  • modulation of AMPK, which, in turn, regulates
  • the activation of the CRE element in the NIS promoter.

Overall, this suggests that AMPK modulating compounds may be useful for the
enhancement of iodide uptake by thyrocytes, which could be useful for the
treatment of thyroid cancer patients with radioactive iodine.

AMPK: Master Metabolic Regulator

© 1996–2013 themedicalbiochemistrypage.org, LLC | info
@ themedicalbiochemistrypage.org

AMPK-activating drugs metformin or phenformin might provide protection against cancer 1741-7007-11-36-5

AMPK-activating drugs metformin or phenformin might provide protection against cancer 1741-7007-11-36-5


AMPK and AMPK-related kinase (ARK) family 1741-7007-11-36-4

AMPK and AMPK-related kinase (ARK) family 1741-7007-11-36-4


central role of AMPK in the regulation of metabolism



AMP-activated protein kinase (AMPK) was first discovered as an activity that

AMPK induces a cascade of events within cells in response to the ever changing energy
charge of the cell. The role of AMPK in regulating cellular energy charge places this
enzyme at a central control point in maintaining energy homeostasis.

More recent evidence has shown that AMPK activity can also be regulated by physiological stimuli, independent of the energy charge of the cell, including hormones and nutrients.


Once activated, AMPK-mediated phosphorylation events

These events are rapidly initiated and are referred to as

  • short-term regulatory processes.

The activation of AMPK also exerts

  • long-term effects at the level of both gene expression and protein synthesis.

Other important activities attributable to AMPK are

  1. regulation of insulin synthesis and
  2. secretion in pancreatic islet β-cells and
  3. modulation of hypothalamic functions involved in the regulation of satiety.

How these latter two functions impact obesity and diabetes will be discussed below.

Regulation of AMPK

In the presence of AMP the activity of AMPK is increased approximately 5-fold.
However, more importantly is the role of AMP in regulating the level of phosphorylation
of AMPK. An increased AMP to ATP ratio leads to a conformational change in the γ-subunit
leading to increased phosphorylation and decreased dephosphorylation of AMPK.

The phosphorylation of AMPK results in activation by at least 100-fold. AMPK is
phosphorylated by at least three different upstream AMPK kinases (AMPKKs).
Phosphorylation of AMPK occurs in the α subunit at threonine 172 (T172) which

  • lies in the activation loop.

One kinase activator of AMPK is

  • Ca2+-calmodulin-dependent kinase kinase β (CaMKKβ)
  • which phosphorylates and activates AMPK in response to increased calcium.

The distribution of CaMKKβ expression is primarily in the brain with detectable levels
also found in the testes, thymus, and T cells. As described for the Ca2+-mediated
regulation of glycogen metabolism,

  • increased release of intracellular stores of Ca2+ create a subsequent demand for

Activation of AMPK in response to Ca fluxes

  • provides a mechanism for cells to anticipate the increased demand for ATP.

Evidence has also demonstrated that the serine-threonine kinase, LKB1 (also called
serine-threonine kinase 11, STK11) which is encoded by the Peutz-Jeghers syndrome
tumor suppressor gene, is required for activation of AMPK in response to stress.

The active LKB1 kinase is actually a complex of three proteins:

  1. LKB1,
  2. Ste20-related adaptor (STRAD) and
  3. mouse protein 25 (MO25).

Thus, the enzyme complex is often referred to as LKB1-STRAD-MO25. Phosphorylation
of AMPK by LKB1 also occurs on T172. Unlike the limited distribution of CaMKKβ,

  • LKB1 is widely expressed, thus making it the primary AMPK-regulating kinase.

Loss of LKB1 activity in adult mouse liver leads to

  • near complete loss of AMPK activity and
  • is associated with hyperglycemia.

The hyperglycemia is, in part, due to an increase in the transcription of gluconeogenic
genes. Of particular significance is the increased expression of

  • the peroxisome proliferator-activated receptor-γ (PPAR-γ) coactivator 1α
    (PGC-1α), which drives gluconeogenesis.
  • Reduction in PGC-1α activity results in normalized blood glucose levels in
    LKB1-deficient mice.

The third AMPK phosphorylating kinase is transforming growth factor-β-activated
kinase 1 (TAK1). However, the normal physiological conditions under which TAK1
phosphorylates AMPK are currently unclear.

The effects of AMP are two-fold:

  1. a direct allosteric activation and making AMPK a poorer substrate for

Because AMP affects both
the rate of AMPK phoshorylation in the positive direction and
dephosphorylation in the negative direction,

the cascade is ultrasensitive. This means that

  1. a very small rise in AMP levels can induce a dramatic increase in the activity of

The activity of adenylate kinase, catalyzing the reaction shown below, ensures that

  • AMPK is highly sensitive to small changes in the intracellular [ATP]/[ADP] ratio.

2 ADP ——> ATP + AMP

Negative allosteric regulation of AMPK also occurs and this effect is exerted by
phosphocreatine. As indicated above, the β subunits of AMPK have a glycogen-binding domain, GBD. In muscle, a high glycogen content

  • represses AMPK activity and
  • this is likely the result of interaction between the GBD and glycogen,
  • the GBD of AMPK allows association of the enzyme with the regulation of glycogen metabolism
  • by placing AMPK in close proximity to one of its substrates glycogen synthase.

AMPK has also been shown to be activated by receptors that are coupled to

  • phospholipase C-β (PLC-β) and by
  • hormones secreted by adipose tissue (termed adipokines) such as leptinand adiponectin (discussed below).

Targets of AMPK

The signaling cascades initiated by the activation of AMPK exert effects on

  • glucose and lipid metabolism,
  • gene expression and
  • protein synthesis.

These effects are most important for regulating metabolic events in the liver, skeletal
muscle, heart, adipose tissue, and pancreas.

Demonstration of the central role of AMPK in the regulation of metabolism in response
to events such as nutrient- or exercise-induced stress. Several of the known physiologic
targets for AMPK are included as well as several pathways whose flux is affected by
AMPK activation. Arrows indicate positive effects of AMPK, whereas, T-lines indicate
the resultant inhibitory effects of AMPK action.

The uptake, by skeletal muscle, accounts for >70% of the glucose removal from the
serum in humans. Therefore, it should be obvious that this event is extremely important
for overall glucose homeostasis, keeping in mind, of course, that glucose uptake by
cardiac muscle and adipocytes cannot be excluded from consideration. An important fact
related to skeletal muscle glucose uptake is that this process is markedly impaired in
individuals with type 2 diabetes.

The uptake of glucose increases dramatically in response to stress (such as ischemia) and
exercise and is stimulated by insulin-induced recruitment of glucose transporters
to the plasma membrane, primarily GLUT4. Insulin-independent recruitment of glucose
transporters also occurs in skeletal muscle in response to contraction (exercise).

The activation of AMPK plays an important, albeit not an exclusive, role in the induction of
GLUT4 recruitment to the plasma membrane. The ability of AMPK to stimulate
GLUT4 translocation to the plasma membrane in skeletal muscle is by a different mechanism
than that stimulated by insulin and insulin and AMPK effects are additive.

Under ischemic/hypoxic conditions in the heart the activation of AMPK leads to the
phosphorylation and activation of the kinase activity of phosphofructokinase-2, PFK-2
(6-phosphofructo-2-kinase). The product of the action of PFK-2 (fructose-2,6-bisphosphate,
F2,6BP) is one of the most potent regulators of the rate of flux through
glycolysis and gluconeogenesis.

In liver the PKA-mediated phosphorylation of PFK-2 results in conversion of the
enzyme from a kinase that generates F2,6BP to a phosphatase that removes the
2-phosphate thus reducing the levels of the potent allosteric activator of the glycolytic
enzyme 6-phosphfructo-1-kinase, PFK-1 and the potent allosteric inhibitor
of the gluconeogenic enzyme fructose-1,6-bisphosphatase (F1,-6BPase).

It is important to note that like many enzymes, there are multiple isoforms of PFK-2
(at least 4) and neither the liver or the skeletal muscle isoforms contain the AMPK
phosphorylation sites found in the cardiac and inducible (iPFK2) isoforms of PFK-2.

Inducible PFK-2 is expressed in the monocyte/macrophage lineage in response to pro-
inflammatory stimuli. The ability to activate the kinase activity by phosphorylation of
PFK-2 in cardiac tissue and macrophages in response to ischemic conditions allows these
cells to continue to have a source of ATP via anaerobic glycolysis. This phenomenon is
recognized as the Pasteur effect: an increased rate of glycolysis in response to hypoxia.

Of pathological significance is the fact that the inducible form of PFK-2 is commonly
expressed in many tumor cells and this may allow AMPK to play an important role in
protecting tumor cells from hypoxic stress. Indeed, techniques for depleting AMPK in
tumor cells have shown that these cells become sensitized to nutritional stress upon loss
of AMPK activity.

Whereas, stress and exercise are powerful inducers of AMPK activity in skeletal muscle,
additional regulators of its activity have been identified.

Insulin-sensitizing drugs of the thiazolidinedione family (activators of PPAR-γ, see
below) as well as the hypoglycemia drug metformin exert a portion of their effects
through regulation of the activity of AMPK.

As indicated above, the activity of the AMPK activating kinase, LKB1, is critical for
regulating gluconeogenic flux and consequent glucose homeostasis. The action of
metformin in reducing blood glucose levels

  • requires the activity of LKB1 in the liver for this function.

Also, several adipokines (hormones secreted by adipocytes) either stimulate or inhibit
AMPK activation:

  1. leptin and adiponectin have been shown to stimulate AMPK activation, whereas,
  2. resistininhibits AMPK activation.

Cardiac effects exerted by activation of AMPK also include

AMPK-mediated phosphorylation of eNOS leads to increased activity and consequent
NO production and provides a link between metabolic stresses and cardiac function.

In platelets, insulin action leads to an increase in eNOS activity that is

  • due to its phosphorylation by AMPK.

Activation of NO production in platelets leads to

  • a decrease in thrombin-induced aggregation, thereby,
  • limiting the pro-coagulant effects of platelet activation.

The response of platelets to insulin function clearly indicates why disruption in insulin
action is a major contributing factor in the development of the metabolic syndrome

Activation of AMPK leads to a reduction in the level of SREBP

  • a transcription factor &regulator of the expression of numerous
    lipogenic enzymes

Another transcription factor reduced in response to AMPK activation is

  • hepatocyte nuclear factor 4α, HNF4α
    • a member of the steroid/thyroid hormone superfamily.
    • HNF4α is known to regulate the expression of several liver and
      pancreatic β-cell genes such as GLUT2, L-PK and preproinsulin.
  • Of clinical significance is that mutations in HNF4α are responsible for
    • maturity-onset diabetes of the young, MODY-1.

Recent evidence indicates that the gene for the carbohydrate-response-element-
binding protein (ChREBP) is a target for AMPK-mediated transcriptional regulation
in the liver. ChREBP is rapidly being recognized as a master regulator of lipid
metabolism in liver, in particular in response to glucose uptake.

The target of the thiazolidinedione (TZD) class of drugs used to treat type 2 diabetes is
the peroxisome proliferator-activated receptor γPPARγ which

  • itself may be a target for the action of AMPK.

The transcription co-activator, p300, is phosphorylated by AMPK

  • which inhibits interaction of p300 with not only PPARγ but also
  • the retinoic acid receptor, retinoid X receptor, and
  • thyroid hormone receptor.

PPARγ is primarily expressed in adipose tissue and thus it was difficult to reconcile how
a drug that was apparently acting only in adipose tissue could lead to improved insulin
sensitivity of other tissues. The answer to this question came when it was discovered that the TZDs stimulated the expression and release of the adipocyte hormone (adipokine),
adiponectin. Adiponectin stimulates glucose uptake and fatty acid oxidation in skeletal
muscle. In addition, adiponectin stimulates fatty acid oxidation in liver while inhibiting
expression of gluconeogenic enzymes in this tissue.

These responses to adiponectin are exerted via activation of AMPK. Another
transcription factor target of AMPK is the forkhead protein, FKHR (now referred to as
FoxO1). FoxO1 is involved in the activation of glucose-6-phosphatase expression and,
therefore, loss of FoxO1 activity in response to AMPK activation will lead to reduced
hepatic output of glucose.

This concludes a very complicated perspective that ties together the thyroid hormone
activity, the hypophysis, diabetes mellitus, and AMPK tegulation of metabolism in the
liver, skeletal muscle, adipose tissue, and heart.  I also note at this time that there
nongenetic points to be made here:

  1. The tissue specificity of isoenzymes
  2. The modulatory role of AMP:ATP ratio in phosphorylation/dephosphorylation
    effects on metabolism tied to AMPK
  3. The tie in of stress or ROS with fast reactions to protect harm to tissues
  4. The relationship of cytokine activation and release to the above metabolic events
  5. The relationship of effective and commonly used diabetes medications to AMPK
    mediated processes
  6. The preceding presentation is notable for the importance of proteomic and
    metabolomic invetigations in elucidation common chronic and nongenetic diseases


Read Full Post »

Selective Ion Conduction

Reviewer and Curator: Larry H Bernstein, MD, FCAP


This is Part III of a series of articles on Translational Medicine in water transport or paracellelar flow, and ion conductance.  The first article was solely on the aquaporins (Part I), and the second goes from paracellular flow (dealing with paracellin-1 and the familay of claudins.  These proteins factor into a number of diseases of kidney function and Mg(2+) homeostasis, as well as a relationship between congestive heart failure related to infarct remodeling and the sodium-calcium transporter, with a model for treatment (Part II).  The last explores the basis of selective ion conduction, based on the 2003 Nobel Prize presentation by Roderick MacKinnon (Part III).


Nobel Lecture, December 8, 2003
Roderick MacKinnon
Howard Hughes Medical Institute, Laboratory of Molecular Neurobiology and Biophysics, Rockefeller University ,NewYork, NY 

Water is an electrically polarizable substance, which means that its molecules rearrange in an ion’s electric field, pointing negative oxygen atoms in the direction of cations and positive hydrogen atoms toward anions. These electrically stabilizing interac­tions are much weaker in a less polarizable substance such as oil. Thus, an ion will tend to stay in the water on either side of a cell membrane rather than en­ter and cross the membrane. And yet numerous cellular processes, ranging from electrolyte transport across epithelia to electrical signal production in neurons, depend on the flow of ions across the membrane. To mediate the flow, specific protein catalysts known as ion channels exist in the cell mem­brane. Ion channels exhibit the following three essential properties: (1) they conduct ions rapidly, (2) many ion channels are highly selective, meaning only certain ion species flow while others are excluded, (3) their function is regu­lated by processes known as gating, that is, ion conduction is turned on and off in response to specific environmental stimuli. Figure 1 summarizes these properties (figure 1).

MacKinnon. Fig 1  Ion channels exhibit three basic properties

Figure 1. Ion channels exhibit three basic properties depicted in the cartoon. They conduct specific ions (for example K ) at high rates, they are selective (a K  channel essentially excludes Na ), and conduction is turned on and off by opening and closing a gate, which can be regulated by an external stimulus such as ligand-binding or membrane voltage. The relative size of K and Naions is shown.

The modern history of ion channels began in 1952 when Hodgkin and Huxley published their seminal papers on the theory of the action potential in the squid giant axon (Hodgkin and Huxley, 1952a; Hodgkin and Huxley, 1952b; Hodgkin and Huxley, 1952c; Hodgkin and Huxley, 1952d). A funda­mental element of their theory was that the axon membrane undergoes changes in its permeability to Na+ and K+ ions. The Hodgkin-Huxley theory did not address the mechanism by which the membrane permeability changes occur: ions could potentially cross the membrane through channels or by a carrier-mediated mechanism. In their words ‘Details of the mecha­nism will probably not be settled for some time’ (Hodgkin and Huxley, 1952a). It is fair to say that the pursuit of this statement has accounted for much ion channel research over the past fifty years.

As early as 1955 experimental evidence for channel mediated ion flow was obtained when Hodgkin and Keynes measured the directional flow of K+ ions across axon membranes using the isotope 42K+ (Hodgkin and Keynes, 1955). They observed that K+ flow in one direction across the membrane depends on flow in the opposite direction, and suggested that ‘the ions should be con­strained to move in single file and that there should, on average, be several ions in a channel at any moment’. Over the following two decades Armstrong and Hille used electrophysiological methods to demonstrate that Na+ and K+ ions cross cell membranes through unique protein pores – Na+ channels and K+ channels – and developed the concepts of selectivity filter for ion discrim­ination and gate for regulating ion flow (Hille, 1970; Hille, 1971; Hille, 1973; Armstrong, 1971; Armstrong et al., 1973; Armstrong and Bezanilla, 1977; Armstrong, 1981). The patch recording technique invented by Neher and Sakmann then revealed the electrical signals from individual ion channels, as well as the extraordinary diversity of ion channels in living cells throughout nature (Neher and Sakmann, 1976).

The past twenty years have been the era of molecular biology for ion chan­nels. The ability to manipulate amino acid sequences and express ion chan­nels at high levels opened up entirely new possibilities for analysis. The ad­vancement of techniques for protein structure determination and the devel­opment of synchrotron facilities also created new possibilities. For me, a sci­entist who became fascinated with understanding the atomic basis of life’s electrical system, there could not have been a more opportune time to enter the field.

The past twenty years have been the era of molecular biology for ion channels. The ability to manipulate amino acid sequences and express ion channels at high levels opened up entirely new possibilities for analysis. The advancement of techniques for protein structure determination and the development of synchrotron facilities also created new possibilities. For me, a scientist who became fascinated with understanding the atomic basis of life’s electrical system, there could not have been a more opportune time to enter the field. 


The cloning of the Shaker K+ channel gene from Drosophila melanogaster by Jan, Tanouye, and Pongs revealed for the first time a K+ channel amino acid se­quence and stimulated efforts by many laboratories to discover which of these amino acids form the pore, selectivity filter, and gate (Tempel et al., 1987; Kamb et al., 1987; Pongs et al., 1988). At Brandeis University in Chris Miller’s laboratory I had an approach to find the pore amino acids. Chris and I had just completed a study showing that charybdotoxin, a small protein from scor­pion venom, inhibits a K+ channel isolated from skeletal muscle cells by plug­ging the pore and obstructing the flow of ions (MacKinnon and Miller, 1988). In one of those late night ‘let’s see what happens if’ experiments while taking a molecular biology course at Cold Spring Harbor I found that the toxin – or what turned out to be a variant of it present in the charybdotoxin prepara­tion – inhibited the Shaker K+ channel (MacKinnon et al., 1988; Garcia et al., 1994). This observation meant I could use the toxin to find the pore, and it did not take very long to identify the first site-directed mutants of the Shaker K+ channel with altered binding of toxin (MacKinnon and Miller, 1989). I continued these experiments at Harvard Medical School where I began as as­sistant professor in 1989. Working with my small group at Harvard, including Tatiana Abramson, Lise Heginbotham, and Zhe Lu, and sometimes with Gary Yellen at Johns Hopkins University, we reached several interesting conclusions concerning the architecture of K+ channels. They had to be tetramers in which four subunits encircle a central ion pathway (MacKinnon, 1991). This conclusion was not terribly surprising but the experiments and analysis to reach it gave me great pleasure since they required only simple measure­ments and clear reasoning with binomial statistics. We also deduced that each subunit presents a ‘pore loop’ to the central ion pathway (figure 2) (MacKinnon, 1995).

MacKinnon. Fig 2.  tetramer K channel

Figure 2. Early picture of a tetramer K+ channel with a selectivity filter made of pore loops. A linear representation of a Shaker K+ channel subunit on top shows shaded hydrophobic segments S1 to S6 and a region designated the pore loop. A partial amino acid sequence from the Shaker K+ channel pore loop highlights amino acids shown to interact with ex-tracellular scorpion toxins (*), intracellular tetraethylammonium (↑) and K+ ions (+). The pore loop was proposed to reach into the membrane (middle) and form a selectivity filter at the center of four subunits (bottom).

This ‘loop’ formed the binding sites for scorpion toxins (MacKinnon and Miller, 1989; Hidalgo and MacKinnon, 1995; Ranganathan et al., 1996) as well as the small-molecule inhibitor tetraethylammonium ion (MacKinnon and Yellen, 1990; Yellen et al., 1991), which had been used by Armstrong and Hille decades earlier in their pioneering analysis of K+ channels (Armstrong, 1971; Armstrong and Hille, 1972). Most important to my thinking, mutations of certain amino acids within the ‘loop’ affected the channel’s ability to discriminate between K+ and Na+, the selectivity hallmark of K+ channels (Heginbotham et al., 1992; Heginbotham et al., 1994). Meanwhile, new K+ channel genes were discovered and they all had one ob­vious feature in common: the very amino acids that we had found to be im­portant for K+ selectivity were conserved (figure 3). We called these amino acids the K+ channel signature sequence, and imagined four pore loops somehow forming a selectivity filter with the signature sequence amino acids inside the pore (Heginbotham et al., 1994; MacKinnon, 1995).

Figure 3. The K+ channel signature sequence shown as single letter amino acid code (blue) is highly conserved in organisms throughout the tree of life. Some K+ channels contain six membrane-spanning segments per subunit (6TM) while others contain only two (2TM). 2TM K+ channels correspond to 6TM K+ channels without the first four membrane-span­ning segments (S1-S4 in figure 2).

When you consider the single channel conductance of many K+ channels found in cells you realize just how incredible these molecular devices are. With typical cellular electrochemical gradients, K+ ions conduct at a rate of 107 to 108 ions per second. That rate approaches the expected collision fre­quency of K+ ions from solution with the entryway to the pore. This means that K+ ions flow through the pore almost as fast as they diffuse up to it. For this to occur the energetic barriers in the channel have to be very low, some­thing like those encountered by K+ ions diffusing through water. All the more remarkable, the high rates are achieved in the setting of exquisite selectivity: the K+ channel conducts K+, a monovalent cation of Pauling radius 1.33 Å, while essentially excluding Na+, a monovalent cation of Pauling radius 0.95 Å. And this ion selectivity is critical to the survival of a cell. How does nature ac­complish high conduction rates and high selectivity at the same time? The an­swer to this question would require knowing the atomic structure formed by the signature sequence amino acids, that much was clear. The conservation of the signature sequence amino acids in K+ channels throughout the tree of life, from bacteria (Milkman, 1994) to higher eukaryotic cells, implied that nature had settled upon a very special solution to achieve rapid, selective K+ conduction across the cell membrane. For me, this realization provided in­spiration to want to directly visualize a K+ channel and its selectivity filter.


I did not know how we would ever reach the point of ob­taining enough K+ channel protein to attempt crystallization, but the K+ channel signature sequence continued to appear in a growing number of prokaryotic genes, making expression in Escherichia coli possible. We focused our effort on a bacterial K+ channel called KcsA from Streptomyces lividans, dis­covered by Schrempf (Schrempf et al., 1995). The KcsA channel has a simple topology with only two membrane spanning segments per subunit corre­sponding to the Shaker K+ channel without S1 through S4 (figure 2). Despite its prokaryotic origin KcsA closely resembled the Shaker K+ channel’s pore amino acid sequence, and even exhibited many of its pharmacological prop­erties, including inhibition by scorpion toxins (MacKinnon et al., 1998). This surprised us from an evolutionary standpoint, because why should a scorpion want to inhibit a bacterial K+ channel! But from the utilitarian point of view of protein biophysicists we knew exactly what the scorpion toxin sensitivity meant, that KcsA had to be very similar in structure to the Shaker K+ channel.

The KcsA channel produced crystals but they were poorly ordered and not very useful in the X-ray beam. After we struggled for quite a while I began to wonder whether some part of the channel was intrinsically disordered and in­terfering with crystallization. Fortunately my neighbor Brian Chait and his postdoctoral colleague Steve Cohen were experts in the analysis of soluble proteins by limited proteolysis and mass spectrometry, and their techniques applied beautifully to a membrane protein. We found that KcsA was as solid as a rock, except for its C-terminus. After removing disordered amino acids from the c-terminus with chymotrypsin the crystals improved dramatically, and we were able to solve an initial structure at a resolution of 3.2 Å (Doyle et al., 1998). We could not clearly see K+ in the pore at this resolution, but my years of work on K+ channel function told me that Rb+ and Cs+ should be valuable electron dense substitutes for K+, and they were. Rubidium and Cs+ difference Fourier maps showed these ions lined up in the pore – as Hodgkin and Keynes might have imagined in 1955 (Hodgkin and Keynes, 1955).

The KcsA structure was altogether illuminating, but before I describe it, I will depart from chronology to explain the next important technical step. A very accurate description of the ion coordination chemistry inside the selec­tivity filter would require a higher resolution structure. With 3.2 Å data we could infer the positions of the main-chain carbonyl oxygen atoms by apply­ing our knowledge of small molecule structures, that is our intu­ition, but we needed to see the selectivity filter atoms in detail. A high-resolu­tion structure was actually quite difficult to obtain. After more than three ad­ditional years of work by João and then Yufeng (Fenny) Zhou we finally man­aged to produce high-quality crystals by attaching monoclonal Fab fragments to KcsA. These crystals provided the information we needed, a structure at a resolution of 2.0 Å in which K+ ions could be visualized in the grasp of selectivity filter protein atoms (figure 4) (Zhou et al., 2001b). What did the K+ channel structure tell us and why did nature conserve the K+ channel signa­ture sequence amino acids?

MacKinnon Fig 4. Electron density KcsA K channel

Figure 4. Electron density (2Fo-Fc contoured at 2 ir) from a high-resolution structure of the KcsA K+ channel is shown as blue mesh. This region of the channel features the selectivity filter with K+ ions and water molecules along the ion pathway. The refined atomic model is shown in the electron density. Adapted from (Zhou et al., 2001b).

Not all protein structures speak to you in an understandable language, but the KcsA K+ channel does. Four subunits surround a central ion pathway that crosses the membrane (figure 5A). Two of the four subunits are shown in fig­ure 5B with electron density from K+ ions and water along the pore. Near the center of the membrane the ion pathway is very wide, forming a cavity about 10 Å in diameter with a hydrated K+ ion at its center. Each subunit directs the C-terminal end of a ‘pore helix’, shown in red, toward the ion. The C-termi­nal end of an á-helix is associated with a negative ‘end charge’ due to car­bonyl oxygen atoms that do not participate in secondary structure hydrogen bonding, so the pore helices are directed as if to stabilize the K+ ion in the cavity. At the beginning of this lecture I raised the fundamental issue of the cell membrane being an energetic barrier to ion flow because of its oily inte­rior. KcsA allows us to intuit a simple logic encoded in its structure, and elec­trostatic calculations support the intuition (Roux and MacKinnon, 1999): the K+ channel lowers the membrane dielectric barrier by hydrating a K+ ion deep inside the membrane, and by stabilizing it with á-helix end charges.

MacKinnon Fig 5.  KcsA K+ channel   pore-helices (red) and selectivity filter (yellow)

Figure 5. (A) A ribbon representation of the KcsA K+ channel with its four subunits colored uniquely. The channel is oriented with the extracellular solution on top. (B) The KcsA K+ channel with front and back subunits removed, colored to highlight the selectivity filter (yellow). Electron density in blue mesh is shown along the ion pathway. Labels identify the pore, outer, and inner helices and the inner helix bundle. The outer and inner helices correspond to S5 and S6 in figure 2

How does the K+ channel distinguish K+ from Na+? Our earlier mutagene-sis studies had indicated that the signature sequence amino acids would be re­sponsible for this most basic function of a K+ channel. Figure 6 shows the structure formed by the signature sequence – the selectivity filter – located in the extracellular third of the ion pathway. The glycine amino acids in the se­quence TVGYG have dihedral angles in or near the left-handed helical region of the Ramachandran plot, as does the threonine, allowing the main-chain carbonyl oxygen atoms to point in one direction, toward the ions along the pore. It is easy to understand why this sequence is so conserved among K+ channels: the alternating glycine amino acids permit the required dihedral angles, the threonine hydroxyl oxygen atom coordinates a K+ ion, and the side-chains of valine and tyrosine are directed into the protein core sur­rounding the filter to impose geometric constraint.

MacKinnon Figure 6. Detailed structure of the K+ selectivity filter

Figure 6. Detailed structure of the K+ selectivity filter (two subunits). Oxygen atoms coordi­nate K+ ions (green spheres) at positions 1 to 4 from the extracellular side. Single letter amino acid code identifies select signature sequence amino acids. Yellow, blue and red cor­respond to carbon, nitrogen and oxygen atoms, respectively. Green and gray dashed lines show oxygen-K+ and hydrogen bonding interactions.

The end result when the subunits come together is a narrow tube consisting of four equal spaced K+ binding sites, labeled 1 to 4 from the extracellular side. Each binding site is a cage formed by eight oxygen atoms on the vertices of a cube, or a twisted cube called a square antiprism (figure 7). The binding sites are very similar to the single alkali metal site in nonactin, a K+ selective antibiotic with nearly identical K+-oxygen distances (Dobler et al., 1969; Dunitz and Dobler, 1977). The principle of K+ selectivity is implied in a subtle feature of the KcsA crystal structure. The oxygen atoms surrounding K+ ions in the selectivity filter are arranged quite like the water molecules surrounding the hydrated K+ ion in the cavity. This comparison conveys a visual impression of binding sites in the filter paying for the energetic cost of K+ dehydration. The Na+ ion is appar­ently too small for these K+-sized binding sites, so its dehydration energy is not compensated.

MacKinnon Fig 7 K+ channel mimics the hydration shell surrounding a K+ ion

Figure 7. A K+ channel mimics the hydration shell surrounding a K+ ion. Electron density (blue mesh) for K+ ions in the filter and for a K+ ion and water molecules in the central cav­ity are shown. White lines highlight the coordination geometry of K+ in the filter and in wa­ter. Adapted from (Zhou et al., 2001b).

The question that compelled us most after seeing the structure was exactly how many ions are in the selectivity filter at a given time? To begin to under­stand how ions move through the filter we needed to know the stoichiometry of the ion conduction reaction, and that meant knowing how many ions can occupy the filter. Four binding sites were apparent, but are they all occupied at once? Four K+ ions in a row separated by an average center-to-center dis­tance of 3.3 Å seemed unlikely for electrostatic reasons. From an early stage we suspected that the correct number would be closer to two, because two ions more easily explained the electron density we observed for the larger al­kali metal cations Rb+ and Cs+ (Doyle et al., 1998; Morais-Cabral et al., 2001). Quantitative evidence for the precise number of ions came with the high-res­olution structure and with the analysis of Tl+ (Zhou and MacKinnon, 2003). Thallium is the most ideally suited ‘K+ analog’ because it flows through K+ channels, has a radius and dehydration energy very close to K+, and has the favorable crystallographic attributes of high electron density and an anom­alous signal. The one serious difficulty in working with Tl+ is its insolubility with Cl. Fenny meticulously worked out the experimental conditions and de­termined that on average there are between two and two and a half conduct­ing ions in the filter at once, with an occupancy at each position around one half.

We also observed that if the concentration of K+ (or Tl+) bathing the crys­tals is lowered sufficiently (below normal intracellular levels) then a reduc­tion in the number of ions from two to one occurs and is associated with a structural change to a ‘collapsed’ filter conformation, which is pinched closed in the middle (Zhou et al., 2001b; Zhou and MacKinnon, 2003). At concentrations above 20 mM the entry of a second K+ ion drives the filter to a ‘conductive’ conformation, as shown in figure 8. Sodium on the other hand does not drive the filter to a ‘conductive’ conformation even at concentra­tions up to 500 mM.

MacKinnon Figure 8. The selectivity filter can adopt two conformations

Figure 8. The selectivity filter can adopt two conformations. At low concentrations of K+ on average one K+ ion resides at either of two sites near the ends of the filter, which is col­lapsed in the middle. At high concentrations of K+ a second ion enters the filter as it changes to a conductive conformation. On average, two K+ ions in the conductive filter re­side at four sites, each with about half occupancy.

The K+-induced conformational change has thermodynamic consequences for the affinity of two K+ ions in the ‘conductive’ filter. It implies that a frac­tion of the second ion’s binding energy must be expended as work to bring about the filter’s conformational change, and as a result the two ions will bind with reduced affinity. To understand this statement at an intuitive level, rec­ognize that for two ions to reside in the filter they must oppose its tendency to collapse and force one of them out, i.e. the two-ion ‘conductive’ conforma­tion is under some tension, which will tend to lower K+ affinity. This is a de­sirable property for an ion channel because weak binding favors high con­duction rates. The same principle, referred to as the ‘induced fit’ hypothesis, had been proposed decades earlier by enzymologists to explain high speci­ficity with low substrate affinity in enzyme catalysis (Jencks, 1987).

In the ‘conductive’ filter if two K+ ions were randomly distributed then they would occupy four sites in six possible ways. But several lines of evidence hint­ed to us that the ion positions are not random. For example Rb+ and Cs+ ex-hibit preferred positions with obviously low occupancy at position 2 (Morais-Cabral et al., 2001; Zhou and MacKinnon, 2003). In K+ we observed an un­usual doublet peak of electron density at the extracellular entryway to the se­lectivity filter, shown in figure 9 (Zhou et al., 2001b). We could explain this density if K+ is attracted from solution by the negative protein surface charge near the entryway and at the same time repelled by K+ ions inside the filter. Two discrete peaks implied two distributions of ions in the filter.

MacKinnon Fig 9  Figure 9. Two K+ ions in the selectivity filter are hypothesized to exist predominantly in two specific configurations 1,3 and 2,4 as shown.

Figure 9. Two K+ ions in the selectivity filter are hypothesized to exist predominantly in two specific configurations 1,3 and 2,4 as shown. K+ ions and water molecules are shown as green and red spheres, respectively. Adapted from (Zhou et al., 2001b).

Discrete configurations of an ion pair suggested a mechanism for ion con­duction (figure 10A) (Morais-Cabral et al., 2001). The K+ ion pair could dif­fuse back and forth between 1,3 and 2,4 configurations (bottom pathway), or alternatively an ion could enter the filter from one side of the membrane as the ion-water queue moves and a K+ exits at the opposite side (the top path­way). Movements would have to be concerted because the filter is no wider than a K+ ion or water molecule. The two paths complete a cycle: in one com­plete cycle each ion moves only a fraction of the total distance through the fil­ter, but the overall electrical effect is to move one charge all the way.

MacKinnon Fig 10 Figure 10. The selectivity filter is represented as five square planes of oxygen atoms.

Figure 10. (A) Through-put cycle for K+ conduction invoking 1,3 and 2,4 configurations. The selectivity filter is represented as five square planes of oxygen atoms. K+ and water are shown as green and red spheres, respectively. (B) Simulated K+ flux around the cycle is graphed as a function of the energy difference between the 1,3 and 2,4 configurations. Adapted from (Morais-Cabral et al., 2001).

A simulation of ions diffusing around the cycle offers a possible explanation: maximum flux is achieved when the energy difference between the 1,3 and 2,4 configurations is zero because that is the condition under which the ‘energy landscape’ for the con­duction cycle is smoothest (figure 10B). The energetic balance between the configurations therefore might reflect the optimization of conduction rate by natural selection (Morais-Cabral et al., 2001). It is not so easy to demonstrate this point experimentally but it is certainly fascinating to ponder.


The focus of this lecture is K+ channels, but for a brief interlude I would like to show you a Cl selective transport protein. By comparing a K+ channel and a Cl ‘channel’ we can begin to appreciate familiar themes in nature’s solu­tions to different problems: getting cations and anions across the cell mem­brane. ClC Cl channels are found in many different cell types and are associ­ated with a number of physiological processes that require Cl ion flow across lipid membranes (Jentsch et al., 1999; Maduke et al., 2000). As is the case for K+ channels, ClC family genes are abundant in prokaryotes, a fortunate cir­cumstance for protein expression and structural analysis. When Raimund Dutzler joined my laboratory he, Ernest Campbell and I set out to address the structural basis of Cl ion selectivity. We determined crystal structures of two bacterial members of the ClC Cl channel family, one from Escherichia coli (EcClC) and another from Salmonella typhimurium (StClC) (Dutzler et al., 2002). Recent studies by Miller on the function of EcClC have shown that it is actually a Cl – proton exchanger (Accardi and Miller, 2004). We do not yet know why certain members of this family of Cl transport proteins function as channels and others as exchangers, but the crystal structures are fascinating and give us a view of Cl selectivity. Architecturally the ClC proteins are unre­lated to K+ channels, but if we focus on the ion pathway certain features are similar (figure 11).

MacKinnon Fig 11 CIC Cl transport protein

Figure 11. The overall architecture of K+ channels and ClC Cl transport proteins is very dif­ferent but certain general features are similar. One similarity shown here is the use of á-he-lix end charges directed toward the ion pathway. The negative C-terminal end charge (red) points to K+. The positive N-terminal end charge (blue) points to Cl.

As we saw in K+ channels, the ClC proteins have a-helices pointed at the ion pathway, but the direction is reversed with the positive charge of the N-terminus close to Cl. This makes perfect sense for lowering the dielectric barrier for a Cl ion. In ClC we see that ions in its selectivity fil­ter tend to be coordinated by main chain protein atoms, with amide nitrogen atoms surrounding Cl instead of carbonyl oxygen atoms surrounding K+ (figure 12). We also see that both the K+ and Cl selectivity filters contain multiple close-spaced binding sites and appear to contain more than one ion, perhaps to exploit electrostatic repulsion between ions in the pore. I find these simi­larities fascinating. They tell us that certain basic physical principles are im­portant, such as the use of á-helix end charges to lower the dielectric barrier when ions cross the lipid membrane.


channels conduct when called upon by a specific stimulus such as the binding of a ligand or a change in membrane voltage (Hille, 2001). The processes by which ion conduction is turned on are called gating. The con­duction of ions occurs on a time scale that is far too rapid to involve very large protein conformational changes.

Figure 12. K+ and Cl- selectivity filters make use of main chain atoms to coordinate ions

Figure 12. K+ and Cl selectivity filters make use of main chain atoms to coordinate ions: car­bonyl oxygen atoms for K+ ions (green spheres) and amide nitrogen atoms for Cl ions (red spheres). Both filters contain multiple close-spaced ion binding sites. The Cl selectivity fil­ter is that of a mutant ClC in which a glutamate amino acid was changed to glutamine (Dutzler et al., 2003).

In the KcsA K+ channel gating is controlled by intracellular pH and lipid membrane composition, but unfortunately the KcsA channel’s open proba­bility reaches a maximum value of only a few percent in functional assays (Cuello et al., 1998; Heginbotham et al., 1998). At first we had no definitive way to know whether a gate was open or closed in the crystal structures. In the 1970s Armstrong had proposed the existence of a gate near the intracel­lular side of the membrane in voltage dependent K+ channels because he could ‘trap’ large organic cations inside the pore between a selectivity filter near the extracellular side and a gate near the intracellular side (Armstrong, 1971; Armstrong, 1974). Following these ideas we crystallized KcsA with a heavy atom version of one of his organic cations, tetrabutyl antimony (TBA), and found that it binds inside the central cavity of KcsA (Zhou et al., 2001a). This was very interesting because the ~10 Å diameter of TBA far exceeds the pore diameter leading up to the cavity: in KcsA the intracellular pore entry­way is constricted to about 3.5 Å by the inner helix bundle (figure 5B). Seeing TBA ‘trapped’ in the cavity behind the inner helix bundle evoked Arm-strong’s classical view of K+ channel gating, and implied that the inner helix bundle serves as a gate and is closed in KcsA. Mutational and spectroscopic studies in other laboratories also pointed to the inner helix bundle as a pos­sible gate-forming structural element (Perozo et al., 1999; del Camino et al., 2000).

We subsequently determined the crystal structure of MthK, complete K+ channel containing RCK domains, from Methanobacterium thermoautotrophicus (figure 13) (Jiang et al., 2002a). This structure was extremely informative. The RCK domains form a ‘gating ring’ on the intracellular side of the pore. In clefts between domains we could see what appeared to be divalent cation binding sites, and the crys­tals had been grown in the presence of Ca2+. In functional assays we discov­ered that the open probability of the MthK channel increased as Ca2+ or Mg2+ concentration was raised, giving us good reason to believe that the crystal structure should represent the open conformation of a K+ channel.

In our MthK structure the inner helix bundle is opened like the aperture of a camera (figure 14) (Jiang et al., 2002b). As a result, the pathway leading up to the selectivity filter from the intracellular side is about 10 Å wide, explaining how Armstrong’s large organic cations can enter the cavity to block a K+ chan-nel, and how K+ ions gain free access to the selectivity filter through aqueous diffusion. By comparing the KcsA and MthK channel structures it seemed that we were looking at examples of closed and opened K+ channels, and could easily imagine the pore undergoing a conformational change from closed to open.

In the crystal of KvAP the voltage sen­sors, held by monoclonal Fab fragments, adopted a non-native conformation. This observation in itself is meaningful as it underscores the intrinsic flexibil­ity of voltage sensors: in contrast Fab fragments had little effect on the more rigid KcsA K+ channel and ClC Cl channel homolog, both of which we deter­mined in the presence and absence of Fab fragments (Doyle et al., 1998; Zhou et al., 2001b; Dutzler et al., 2002; Dutzler et al., 2003). KvAP’s voltage sensors contain a hydrophobic helix-turn-helix element with arginine residues beside the pore (Jiang et al., 2003a).

The KvAP structure and associated functional studies have provided a conceptual model for voltage-dependent gating – one in which the voltage sensors move at the protein-lipid interface in response to a balance between hydrophobic and electrostatic forces. Rees and colleagues at the California Institute of Technology determined the structure of a voltage regulated mechanosensitive channel called MscS, and although it is unrelated to traditional voltage-dependent channels, it too con­tains hydrophobic helix-turn-helix elements with arginine residues apparent­ly against the lipid membrane (Bass et al., 2002). MscS and KvAP are fascinat­ing membrane protein structures. They do not fit into the standard category of membrane proteins with rigid hydrophobic walls against the lipid mem­brane core. I find such proteins intriguing.

Selected Refernces

Accardi, A. and Miller, C. (2004) Proton-coupled chloride transport mediated by ClC-ec1, a bacterial homologue of the ClC chloride channels. Biophys.J. 86[1], 286a.
Armstrong, C. M. (1971). Interaction of tetraethylammonium ion derivatives with the potassium channels of giant axons. J. Gen. Physiol. 58, 413–437.
Armstrong, C. M. (1981). Sodium channels and gating currents. Physiol. Rev. 61, 645–683.
Armstrong, C. M. and Hille, B. (1972). The inner quaternary ammonium ion receptor in potassium channels of the node of Ranvier. J. Gen. Physiol. 59, 388–400.
Armstrong, C. M. (1974). Ionic pores, gates, and gating currents. Q. Rev. Biophys. 7, 179–210
Armstrong, C. M. and Bezanilla, F. (1977). Inactivation of the sodium channel. II. Gating current experiments. J. Gen. Physiol 70, 567–590.
Atkinson, N. S., Robertson, G. A., and Ganetzky, B. (1991). A component of calcium-actvated potassium channels encoded by the Drosophila slo locus. Science 253, 551–555
Bass, R. B., Strop, P., Barclay, M., and Rees, D. C. (2002). Crystal structure of Escherichia coli MscS, a voltage-modulated and mechanosensitive channel. Science 298, 1582–1587.
Bezanilla, F. (2000). The voltage sensor in voltage-dependent ion channels. Physiol Rev. 80, 555–592
del Camino, D., Holmgren, M., Liu, Y., and Yellen, G. (2000). Blocker protection in the pore of a voltage-gated K+ channel and its structural implications. Nature 403, 321–325.
Doyle, D. A., Morais Cabral, J. H., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., Cohen, S. L., Chait, B. T., and MacKinnon,R. (1998). The structure of the potassium channel: molecular basis of K+ conduction and selectivity. Science 280, 69–77.
Dutzler, R., Campbell, E. B., Cadene, M., Chait, B. T., and MacKinnon,R. (2002). X-ray structure of a ClC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity. Nature 415, 287–294.
Dutzler, R., Campbell, E. B., and MacKinnon, R. (2003). Gating the selectivity filter in ClC chloride channels. Science 300, 108–112.
Garcia, M. L., Garcia-Calvo, M., Hidalgo, P., Lee, A., and MacKinnon, R. (1994). Purification and charaterization of three inhibitors of voltage-dependent K+ channels from Leiurus quinquestriatus var. hebraeus venom. Biochemistry 33, 6834–6839.
Gulbis, J. M., Mann, S., and MacKinnon, R. (1999). Structure of a voltage-dependent K+ channel beta subunit. Cell 97, 943–952.
Gulbis, J. M., Zhou, M., Mann, S., and MacKinnon, R. (2000). Structure of the cytoplasmic f3 subunit-T1 assembly of voltage-dependent K+ channels. Science 289, 123–127.
Heginbotham, L., Abramson, T., and MacKinnon, R. (1992). A functional connection between the pores of distantly related ion channels as revealed by mutant K+ channels. Science 258, 1152–1155.
Heginbotham, L., Lu, Z., Abramson, T., and MacKinnon, R. (1994). Mutations in the K+ channel signature sequence. Biophys. J. 66, 1061–1067.
Hidalgo, P. and MacKinnon, R. (1995). Revealing the architecture of a K+ channel pore through mutant cycles with a peptide inhibitor. Science 268, 307–310.
Hille, B. (1970). Ionic channels in nerve membranes. Prog. Biophys. Mol. Biol. 21, 1–32.
Hille, B. (1973). Potassium channels in myelinated nerve. Selective permeability to small cations. J. Gen. Physiol. 61, 669–686.
Hille, B. (2001). Ion Channels of Excitable Membranes. (Sunderland, MA: Sinauer Associates, Inc.).
Hille, B. (1971). The permeability of the sodium channel to organic cations in myelinated nerve. J. Gen. Physiol. 58, 599–619.
Hodgkin, A. L. and Huxley, A. F. (1952a). A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. 117, 500–544.
Hodgkin, A. L. and Huxley, A. F. (1952b). Currents carried by sodium and potassium ions through the membrane of the giant axon of Loligo. J. Physiol. 116, 449–472.
Hodgkin, A. L. and Huxley, A. F. (1952c). The components of membrane conductance in the giant axon of Loligo. J. Physiol. 116, 473–496.
Hodgkin, A. L. and Huxley, A. F. (1952d). The dual effect of membrane potential on sodium conductance in the giant axon of Loligo. J. Physiol. 116, 497–506.
Hodgkin, A. L. and Keynes, R. D. (1955). The potassium permeability of a giant nerve fibre. J. Physiol. (Lond) 128, 61–88.
Jencks, W. P. (1987). Catalysis in Chemistry and Enzymology. (Dover Publications, Inc.).
Jentsch, T. J., Friedrich, T., Schriever, A., and Yamada, H. (1999). The CLC chloride channel family. Pflugers Arch. 437, 783–795.
Jiang, Y., Lee, A., Chen, J., Ruta, V., Cadene, M., Chait, B., and MacKinnon, R. (2003a). X-ray structure of a voltage-dependent K+ channel. Nature 423, 33–41.
Jiang, Y., Ruta, V., Chen, J., Lee, A., and MacKinnon, R. (2003b). The principle of gating charge movement in a voltage-dependent K+ channel. Nature 423, 42–48.
Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B. T., and MacKinnon, R. (2002a). Crystal structure and mechanism of a calcium-gated potassium channel. Nature 417, 515–522.
Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B. T., and MacKinnon, R. (2002b). The open pore conformation of potassium channels. Nature 417, 523–526.
Jiang, Y., Pico, A., Cadene, M., Chait, B. T., and MacKinnon, R. (2001). Structure of the RCK domain from the E. coli K+ channel and demonstration of its presence in the human BK channel. Neuron 29, 593–601.
MacKinnon, R. (1999). The cavity and pore helices in the KcsA K+ channel: electrostatic stabilization of monovalent cations. Science 285, 100–102.
Ruta, V., Jiang, Y., Lee, A., Chen, J., and MacKinnon, R. (2003). Functional analysis of an archeabacterial voltage-dependent K+ channel. Nature 422, 180–185.
Zhou, M. and MacKinnon, R. (2004). A mutant KcsA K+ channel with altered conduction properties and selectivity filter ion distribution. J.Mol.Biol. 338, 839–846.
Zhou, M., Morais-Cabral, J. H., Mann, S., and MacKinnon, R. (2001a). Potassium channel receptor site for the inactivation gate and quaternary amine inhibitors. Nature 411, 657–661.
Zhou, Y. and MacKinnon, R. (2003). The occupancy of ions in the K+ selectivity filter: Charge balance and coupling of ion binding to a protein conformational change underlie high conduction rates. J. Mol. Biol. 333, 965–975.
Zhou, Y., Morais-Cabral, J. H., Kaufman, A., and MacKinnon, R. (2001b). Chemistry of ion coordination and hydration revealed by a K+ channel- Fab complex at 2.0 Å resolution. Nature 414, 43–48.

Related articles in Pharmaceutical Intelligence:

Part I: Identification of Biomarkers that are Related to the Actin Cytoskeleton

Larry H Bernstein, MD, FCAP


Part II: Role of Calcium, the Actin Skeleton, and Lipid Structures in Signaling and Cell Motility

Larry H. Bernstein, MD, FCAP, Stephen Williams, PhD and Aviva Lev-Ari, PhD, RN


Part III: Renal Distal Tubular Ca2+ Exchange Mechanism in Health and Disease

Larry H. Bernstein, MD, FCAP, Stephen J. Williams, PhD
 and Aviva Lev-Ari, PhD, RN


Part IV: The Centrality of Ca(2+) Signaling and Cytoskeleton Involving Calmodulin Kinases and Ryanodine Receptors in Cardiac Failure, Arterial Smooth Muscle, Post-ischemic Arrhythmia, Similarities and Differences, and Pharmaceutical Targets

Larry H Bernstein, MD, FCAP, Justin Pearlman, MD, PhD, FACC and Aviva Lev-Ari, PhD, RN


Part V: Heart, Vascular Smooth Muscle, Excitation-Contraction Coupling (E-CC), Cytoskeleton, Cellular Dynamics and Ca2 Signaling

Larry H Bernstein, MD, FCAP, Justin Pearlman, MD, PhD, FACC and Aviva Lev-Ari, PhD, RN


Part VI: Calcium Cycling (ATPase Pump) in Cardiac Gene Therapy: Inhalable Gene Therapy for Pulmonary Arterial Hypertension and Percutaneous Intra-coronary Artery Infusion for Heart Failure: Contributions by Roger J. Hajjar, MD

Aviva Lev-Ari, PhD, RN


Part VII: Cardiac Contractility & Myocardium Performance: Ventricular Arrhythmiasand Non-ischemic Heart Failure – Therapeutic Implications for Cardiomyocyte Ryanopathy (Calcium Release-related Contractile Dysfunction) and Catecholamine Responses

Justin Pearlman, MD, PhD, FACC, Larry H Bernstein, MD, FCAP and Aviva Lev-Ari, PhD, RN


Part VIII: Disruption of Calcium Homeostasis: Cardiomyocytes and Vascular Smooth Muscle Cells: The Cardiac and Cardiovascular Calcium Signaling Mechanism

Justin Pearlman, MD, PhD, FACC, Larry H Bernstein, MD, FCAP and Aviva Lev-Ari, PhD, RN


Part IX: Calcium-Channel Blockers, Calcium Release-related Contractile Dysfunction (Ryanopathy) and Calcium as Neurotransmitter Sensor

Justin Pearlman, MD, PhD, FACC, Larry H Bernstein, MD, FCAP and Aviva Lev-Ari, PhD, RN

Part X: Synaptotagmin functions as a Calcium Sensor: How Calcium Ions Regulate the fusion of vesicles with cell membranes during Neurotransmission

Larry H Bernstein, MD, FCAP and Aviva Lev-Ari, PhD, RN

Read Full Post »

%d bloggers like this: