Posts Tagged ‘Proteins’

Ferritin Cage Enzyme Encapsulation as a New Platform for Nanotechnology

 Reporter: Irina Robu, PhD

In bionanotechnology, biological systems such as viruses, protein complexes, lipid vesicles and artificial cells, are being developed for applications in medicine and materials science.  However, the paper published by Stephan Tetter and Donald Hilvert in Angewandte Chemie International Edition show that it is possible to encapsulate proteins such as ferritin by manipulating electrostatic interactions with the negatively charged interior of the cage.The primary role of ferritin is to protect cells from the damage caused by the Fenton reaction; where, in oxidizing conditions, free Fe(II) produces harmful reactive oxygen species that can damage the cellular machinery.

The ferritin family proteins are protein nanocages that evolved to safely store iron in an oxidizing world. Since ferritin family proteins are able to mineralize and store metal ions, they have been the focus of much research for the production of metal nanoparticles and as prototypes for semiconductor production. The ferritin cage itself is highly symmetrical, and is made up of 24 subunits arranged in an octahedral symmetry. Ferritins are smaller than other protein used for protein   encapsulation.   Their  outer  diameter is only 12 nm, whereas engineered lumazine synthase variants form cages with diameters ranging from about 20 to 60 nm.The ferritin cage displays remarkable thermal and chemical stability it is likely to modify the surface of the ferritin cage through the addition of peptide and protein tags. These characteristics have made ferritins attractive vectors for the delivery of drug molecules and as scaffolds for vaccine design.

In summary, the paper published in Angewandte Chemie International Edition is the first example of protein incorporation by a ferritin.  Dr. Donald Hilvert and colleagues have shown that AfFtn not only complexes positively charged guest proteins within its naturally negatively charged luminal cavity, but that the in vitro mixing technique can be extended to the encapsulation and protection of other functional  fusion proteins.

Hence, the recent discovery of encapsulated ferritins has identified an exciting new platform for use in bio nanotechnology. The use of synthetic biology tools will allow their rapid implementation in materials science, bio-nanotechnology and medical applications.



Read Full Post »

Robotically Driven System Could Reduce Cost of Discovering Drugs

Reporter: Irina Robu, PhD

However, their approach had only been tested using synthetic or previously acquired data, the team’s current model builds on this by letting the computer choose which experiments to do. The experiments were then carried out using liquid-handling robots and an automated microscope.

A total of 9,216 experiments were done, each consisting of acquiring images for a given cell clone in the presence of a given drug. The challenge for the algorithm was to learn how proteins were affected in each of these experiments, without performing all of them.

The originality of this work was to identify new phenotypes on its own as part of the learning process. To do this, it clustered the images to form phenotypes. The phenotypes were used to form a predictive model, so the learner could estimate the outcomes of unmeasured experiments. The basis of the model was to identify different sets of proteins that responded similarly to sets of drugs, so that it could predict the trend in the unmeasured experiments. The learner repeated the process for a total of 30 rounds, completing 2,697 out of the 9,216 possible experiments. As it progressively performed the experiments, it identified more phenotypes and more patterns in how sets of proteins were affected by sets of drugs.

Using an assortment of calculations, the team determined that the algorithm was able to learn a 92% accurate model for how the 96 drugs affected the 96 proteins, from only 29% of the experiments conducted.

Read Full Post »

This content is password protected. To view it please enter your password below:

Read Full Post »


Larry H. Bernstein, MD, FCAP, Curator



Human Exomes Galore

A new database includes complete sequences of protein-coding DNA from 60,706 individuals.

By Karen Zusi | November 16, 2015

The ability to sequence a person’s entire genome has led many researchers to hunt for the genetic causes of certain diseases. But without a larger set of genomes to compare mutations against, putting these variations into context is difficult. An international group of researchers has banked the full exomes of 60,706 individuals in a database called the Exome Aggregation Consortium (ExAC). The team’s analaysis, posted last month (October 30) on the preprint server bioRxiv, was presented at the Genome Science 2015 conference in Birmingham, U.K. (September 7).

Led by Daniel MacArthur from the Broad Institute of MIT and Harvard, the research team collected exomes from labs around the world for its dataset. “The resulting catalogue of human genetic diversity has unprecedented resolution,” the authors wrote in their preprint. Many of the variants observed in the dataset occurred only once.

“This is one of the most useful resources ever created for medical testing for genetic disorders,” Heidi Rehm, a clinical lab director at Harvard Medical School, told Science News.

Among other things, the team found 3,230 genes that are highly conserved across exomes, indicating likely involvement in critical cellular functions. Of these, 2,557 are not associated with diseases. The authors hypothesized that these genes, if mutated, either lead to embryonic death—before a problem can be diagnosed—or cause rare diseases that have not yet been genetically characterized.

“We should soon be able to say, with high precision: If you have a mutation at this site, it will kill you. And we’ll be able to say that without ever seeing a person with that mutation,” MacArthur said during his Genome Science talk, according to The Atlantic.

This is not the complete set of essential genes in the human body, David Goldstein, a geneticist at Columbia University in New York City, pointed out to Nature. Only by studying more exomes will researchers be able to refine that number, he noted.


Analysis of protein-coding genetic variation in 60,706 humans

, , , , ,  et al.      doi:

Large-scale reference data sets of human genetic variation are critical for the medical and functional interpretation of DNA sequence changes. Here we describe the aggregation and analysis of high-quality exome (protein-coding region) sequence data for 60,706 individuals of diverse ethnicities. The resulting catalogue of human genetic diversity has unprecedented resolution, with an average of one variant every eight bases of coding sequence and the presence of widespread mutational recurrence. The deep catalogue of variation provided by the Exome Aggregation Consortium (ExAC) can be used to calculate objective metrics of pathogenicity for sequence variants, and to identify genes subject to strong selection against various classes of mutation; we identify 3,230 genes with near-complete depletion of truncating variants, 79% of which have no currently established human disease phenotype. Finally, we show that these data can be used for the efficient filtering of candidate disease-causing variants, and for the discovery of human knockout variants in protein-coding genes.

Analysis of protein-coding genetic variation in 60,706 humans



Over the last five years, the widespread availability of high-throughput DNA sequencing technologies has permitted the sequencing of the whole genomes or exomes (the 18 protein-coding regions of genomes) of over half a million humans. In theory, these data represent a powerful source of information about the global patterns of human genetic variation, but in practice, are difficult to access for practical, logistical, and ethical reasons; in addition, the inconsistent processing complicates variant-calling pipelines used by different groups. Current publicly available datasets of human DNA sequence variation contain only a small fraction of all sequenced samples: the Exome Variant Server, created as part of the NHLBI Exome Sequencing Project (ESP)1, contains frequency information spanning 6,503 exomes; and the 1000 Genomes (1000G) Project, which includes individual-level genotype data from whole-genome and exome sequence data for 2,504 individuals2.

Databases of genetic variation are important for our understanding of human population history and biology1–5, but also provide critical resources for the clinical interpretation of variants observed in patients suffering from rare Mendelian diseases6,7. The filtering of candidate variants by frequency in unselected individuals is a key step in any pipeline for the discovery of causal variants in Mendelian disease patients, and the efficacy of such filtering depends on both the size and the ancestral diversity of the available reference data.

Here, we describe the joint variant calling and analysis of high-quality variant calls across 60,706 human exomes, assembled by the Exome Aggregation Consortium (ExAC; This call set exceeds previously available exome-wide variant databases by nearly an order of magnitude, providing unprecedented resolution for the analysis of very low-frequency genetic variants. We demonstrate the application of this data set to the analysis of patterns of genetic variation including the discovery of widespread mutational recurrence, the inference of gene-level constraint against 10 truncating variation, the clinical interpretation of variation in Mendelian disease genes, and the discovery of human “knockout” variants in protein-coding genes.


Deleterious variants are expected to have lower allele frequencies than neutral ones, due to negative selection. This theoretical property has been demonstrated previously in human population sequencing data18,19 and here (Figure 1d, Figure 1e). This allows inference of the degree of natural selection against specific functional classes of variation: however, mutational recurrence as described above indicates that allele frequencies observed in ExAC-scale samples are also skewed by mutation rate, with 10 more mutable sites less likely to be singletons (Figure 2c and Extended Data Figure 4d). Mutation rate is in turn non-uniformly distributed across functional classes – for instance, stop lost mutations can never occur at CpG dinucleotides (Extended Data Figure 4e). We corrected for mutation rates (Supplementary Information) by creating a mutability-adjusted proportion singleton (MAPS) metric. This metric reflects (as expected) strong selection against predicted PTVs, as well as missense variants predicted by conservation-based methods to be deleterious (Figure 2e).

The deep ascertainment of rare variation in ExAC also allows us to infer the extent of 19 selection against variant categories on a per-gene basis by examining the proportion of 20 variation that is missing compared to expectations under random mutation. Conceptually similar approaches have been applied to smaller exome datasets13,20 but have been underpowered, particularly for the analysis of depletion of PTVs. We compared the observed number of rare (MAF <0.1%) variants per gene to an expected number derived from a selection neutral, sequence-context based mutational model13. The model performs extremely well in predicting the number of synonymous variants, which should be under minimal purifying selection, per gene (r = 0.98; Extended Data Figure 5).


Critically, we note that LoF-intolerant genes include virtually all known severe haploinsufficient human disease genes (Figure 3b), but that 79% of LoF-intolerant genes have not yet been assigned a human disease phenotype despite the clear evidence for extreme selective constraint (Supplementary Information 4.11). These likely represent either undiscovered severe dominant disease genes, or genes in which loss of a single copy results in embryonic lethality.

The most highly constrained missense (top 25% missense Z scores) and PTV (pLI ≥0.9) genes show higher expression levels and broader tissue expression than the least constrained genes24 (Figure 3c). These most highly constrained genes are also depleted for eQTLs (p < 10-9 for missense and PTV; Figure 3d), yet are enriched within genome-wide significant trait-associated loci (χ2 p < 10-14, Figure 3e). Intuitively, genes intolerant of PTV variation are dosage sensitive: natural selection does not tolerate a 50% deficit in expression due to the loss of single allele. It is therefore unsurprising that these genes are also depleted of common genetic variants that have a large enough effect on expression to be detected as eQTLs with current limited sample sizes. However, smaller changes in the expression of these genes, through weaker eQTLs or functional variants, are more likely to contribute to medically relevant phenotypes. Therefore, highly constrained genes are dosage-sensitive, expressed more broadly across tissues (as expected for core cellular processes), and are enriched for medically relevant variation.

Finally, we investigated how these constraint metrics would stratify mutational classes according to their frequency spectrum, corrected for mutability as in the previous section (Figure 3f). The effect was most dramatic when considering stop-gained variants in the LoF-intolerant set of genes. For missense variants, the missense Z score offers information additional to Polyphen2 and CADD classifications, indicating that gene-level measures of constraint offer additional information to variant-level metrics in assessing potential pathogenicity.

We assessed the value of ExAC as a reference dataset for clinical sequencing approaches, which typically prioritize or filter potentially deleterious variants based on functional consequence and allele frequency6. To simulate a Mendelian variant analysis, we filtered variants in 100 ExAC exomes per continental population against ESP (the previous default reference data set for clinical analysis) or the remainder of ExAC, removing variants present at ≥0.1% allele frequency, a filter recommended for dominant 16 disease variant discovery6. Filtering on ExAC reduced the number of candidate protein-altering variants by 7-fold compared to ESP, and was most powerful when the highest 18 allele frequency in any one population (“popmax”) was used rather than average (“global”) allele frequency (Figure 4a). ESP is not well-powered to filter at 0.1% AF without removing many genuinely rare variants, as AF estimates based on low allele counts are both upward-biased and imprecise (Figure 4b). We thus expect that ExAC will provide a very substantial boost in the power and accuracy of variant filtering in Mendelian disease projects.


The above curation efforts confirm the importance of allele frequency filtering in analysis of candidate disease variants. However, literature and database errors are prevalent even at lower allele frequencies: the average ExAC exome contains 0.89 reportedly Mendelian variants in well-characterized dominant disease genes at <1% popmax AF and 0.20 at <0.1% popmax AF. This inflation likely results from a combination of false reports of pathogenicity and incomplete penetrance, as we show for PRNP in the accompanying work [Minikel et al, submitted]. The abundance of rare functional variation in many disease genes in ExAC is a reminder that such variants should not be assumed to be causal or highly penetrant without careful segregation or case-control analysis28,7.

We investigated the distribution of PTVs, variants predicted to disrupt protein-coding genes through the introduction of a stop codon or frameshift or the disruption of an essential splice site; such variants are expected to be enriched for complete loss-of-function of the impacted genes. Naturally-occurring PTVs in humans provide a model for the functional impact of gene inactivation, and have been used to identify many genes in 6 which LoF causes severe disease31, as well as rare cases where LoF is protective against disease32.

Among the 7,404,909 HQ variants in ExAC, we found 179,774 high-confidence PTVs (as 10 defined in Supplementary Information Section 6), 121,309 of which are singletons. This 11 corresponds to an average of 85 heterozygous and 35 homozygous PTVs per individual (Figure 5a). The diverse nature of the cohort enables the discovery of substantial numbers of novel PTVs: out of 58,435 PTVs with an allele count greater than one, 33,625 occur in only one population. However, while PTVs as a category are extremely rare, the majority of the PTVs found in any one person are common, and each individual 16 has only ~2 singleton PTVs, of which 0.14 are found in PTV-constrained genes (pLI 17 >0.9). The site frequency spectrum of these variants across the populations represented in ExAC recapitulates known aspects of demographic models, including an increase in intermediate-frequency (1%-5%) PTVs in Finland33 and relatively common (>0.1%) PTVs in Africans (Figure 5b).



Discussion  Here we describe the generation and analysis of the most comprehensive catalogue of 29 human protein-coding genetic variation to date, incorporating high-quality exome sequencing data from 60,706 individuals of diverse geographic ancestry. The resulting call set provides unprecedented resolution for the analysis of very low-frequency protein-coding variants in human populations, as well as a powerful resource for the clinical interpretation of genetic variants observed in disease patients. The complete frequency CC-BY-ND 4.0 International license for this preprint is the author/funder. It is made available under a bioRxiv preprint first posted online October 30, 2015; ; The copyright holder and annotation data from this call-set has been made freely available through a public website []

The ExAC resource provides the largest database to date for the estimation of allele frequency for protein-coding genetic variants, providing a powerful filter for analysis of candidate pathogenic variants in severe Mendelian diseases. Frequency data from ESP1 have been widely used for this purpose, but those data are limited by population diversity and by resolution at allele frequencies ≤0.1%. ExAC therefore provides 21 substantially improved power for Mendelian analyses, although it is still limited in power at lower allele frequencies, emphasizing the need for more sophisticated pathogenic variant filtering strategies alongside on-going data aggregation efforts. ExAC also highlights an unexpected tolerance of many disease genes to functional variation, and reveals that the literature and public databases contain an inflated number of reportedly pathogenic variants across the frequency spectrum, indicating a need for stringent criteria for assertions of pathogenicity.

Finally, we show that different populations confer different advantages in the discovery of gene-disrupting PTVs, providing guidance for projects seeking to identify human “knockouts” to understand gene function. Individuals of African ancestry have more PTVs (140 on average), with this enrichment most pronounced at allele frequencies above 1% (Figure 5b). Finnish individuals, as a result of a population bottleneck, are depleted at the lowest (<0.1%) allele frequencies but have a peak in frequency at 1-5% (Figure 5b). However, these differences are diminished when considering only LoF-constrained (pLI > 0.9) genes (Extended Data Figure 10). Sampling multiple populations would likely be a fruitful strategy for a researcher investigating common PTV variation. However, discovery of homozygous PTVs is markedly enhanced in the South Asia samples, which come primarily from a Pakistani cohort with 38.3% of individuals self- reporting as having closely related parents, emphasizing the extreme value of consanguineous cohorts for “human knockout” discovery (Figure 5d) [Saleheen et al., to 8 be co-submitted].


While the ExAC dataset dramatically exceeds the scale of previously available frequency reference datasets, much remains to be gained by further increases in sample size. Indeed, the fact that even the rarest transversions have mutational rates13 on the order of 1 x 10-9 implies that almost all possible non-lethal SNVs likely exist in some person on Earth. ExAC already includes >70% of all possible protein-coding CpG transitions at well-covered sites; order of magnitude increases in sample size will eventually lead to saturation of other classes of variation.

Read Full Post »

The biochemistry of S amino acids

Larry H. Bernstein, MD, FCAP, Curator


Amino Acid and Sulfur Metabolism

Dr. Rainer Höfgen

 Sulfur is together with nitrogen, phosphorous and potassium a plant macronutrient and a crucial element affecting plant growth, plant performance and yield. The group of Dr. Rainer Hoefgen focuses on characterising the regulation of cysteine and methionine as a result of sulfate uptake and assimilation in the model plant Arabidopsis thaliana.

Cysteine and methionine are two essential amino acids which contain sulfur. We are also looking at interconnections between sulfur metabolism and other plant nutrients. Further, we are investigating means of improving the nutritional quality of crops, with a current focus on rice (Oryza sativa) with respect to a balanced amino acid composition.

In our studies of plant sulfur metabolism, we use two mutually supporting approaches as the basis of our research portfolio. The first is a targeted, pathway-oriented approach aimed at understanding pathway architecture and coordination, and the regulation of the sulfur-containing metabolites as such. The second is a non-biased approach in which functional genomics is used to work out how sulfur metabolism is embedded and controlled within the whole plant system.

sulfur uptake and assimilation

sulfur uptake and assimilation

Zoom Image

Sulfur is a required macronutrient, sulfur uptake and assimilation are crucial determinants in how quickly plants grow and cope with various stresses, and therefore, in how well crops yield.

Inorganic sulfate is taken up through plant roots and, via cysteine biosynthesis, incorporated as organic sulfur. Our investigations focus on fundamental questions about cysteine (cys) and methionine (met) biosynthesis and on the possibility of engineering crop plants enriched in these sulfur-containing amino acids. Methionine is essential for non-ruminant mammals (including man) and uptake of cysteine reduces the methionine requirement. We have used transgenic strategies to generate many plant lines affected in cysteine and methionine biosynthesis, and subjected them to detailed molecular and biochemical analyses. Recently, we embarked on a course to study sulfur metabolism in a holistic way, rather than focusing on single pathways as such. By applying functional genomic tools like transcript, metabolite, and protein profiling in our analysis of transgenic potato (Solanum tuberosum) and of the model plant Arabidopsis thaliana, we are heading for a better understanding of the sulfur metabolism network in plants.

To learn about the control mechanisms involved in sulfur-containing amino acid biosynthesis, we are isolating and studying the involved genes and their promoters. The model plant systems of our investigations are potato and Arabidopsis, although a limited amount of work is also dedicated to rice (Oryza sativa), cucumber (Cucumis sativus), and tomato (Lycopersicon esculentum). Various transgenic plants exhibiting reduced or increased expression of relevant genes in the pathway have been produced and analysed. Fundamental knowledge of pathway regulation has been obtained as well as an improvement of the nutritional quality of a crop plant: Nutritional quality is largely determined by methionine, which is often the most limited of the essential amino acids.

The main thrust of our research recently shifted to analysing sulfur metabolism networks. In a systems biology approach, we investigate the response of Arabidopsis to different periods or degrees of sulfur starvation by applying non-biased, multiparallel tools including transcript, protein, and metabolite profiling. Our results are integrated to form working models for further detailed investigations with a focus on regulatory aspects of metabolism. This work entails the detailed analysis of Arabidopsis mutants and pulls many of our earlier results together into biological context (eg. the increased thiol levels seen during SAT over-expression, glutathione involvement in stress response mechanisms towards active oxygen species, etc.). Our long-term goal is to imbed sulfur metabolism in a broader context such as carbohydrate and nitrogen metabolic networks, which will occur through close collaborations with external and in house research groups.


metabolite profiling

metabolite profiling


Plants are sessile organisms; if they are to survive and reproduce, they must adapt to the growth conditions in which they find themselves. We use variations in sulfur levels as a stimulus and analyse the complex response using diverse multiparallel techniques, particularly transcript and metabolite profiling, trying to piece together the total system response. The plant of choice here is, obviously, Arabidopsis thaliana, although results obtained in this model system are likely to be transferable to other plant species and crop plants. The goal is to provide a consistent and holistic description of plant sulfur metabolism and its regulation.

H Hesse and R Höfgen (2001) Application of Genomics in Agriculture. In: Molecular analysis of plant adapatation to the environment. Eds: Malcolm J. Hawkesford, Peter Buchner. Kluwer AP, Dordrecht, The Netherlands, 61-79

V Nikiforova, J Freitag, S Kempa, M Adamik, H Hesse, R Hoefgen (2003) Transcriptome analysis of sulfur depletion in Arabidopsis thaliana: Interlacing of biosynthetic pathways provides response specificity. The Plant Journal, 33, 633-650.



Plants adapt to available sulfur soil levels by regulating gene expression and protein activity to maintain homeostasis. Sulfur availability in the environment is not static, nor is the plant’s dependence on sulfur at various developmental stages. Thus, one can assume not only that the activities of regulatory proteins are dynamic, but also that changes in the expression of transcription factors involved in triggering downstream gene expression change temporally. Sulfur deprivation triggers a slow adaptive process that resets the level of sulfur homeostasis. Using transcript profiling, we have been able to identify a number of transcription factors involved in this process, which are now the target of further investigations.


Metabolome analysis and bioinformatics

system response

system response

Zoom Image

Gene expression, metabolite spectrum and enzyme activities change under sulfur-limiting conditions.

The response of steady state transcription levels to the sulfur stimulus is but the first chapter of the story. To understand the system response, we have to turn the page and look at protein profiles – levels and activities – and before closing the book, at metabolite profiles, which adjust rapidly in response to changes in protein expression. We are now focusing on metabolome analysis: The same samples used for transcriptome analysis are examined using element analysis (ICP-AES) and metabolite analyses (HPLC, CE, GC/MS, GC/TOF, LC/MS), either in house or in collaboration with outside research groups.

Malcolm J. Hawkesford, Rothamsted Research, UK

As these analyses are refined and data accumulates, it will become more and more important to overlay and compare transcript and metabolite profiles in order to try to generate an in silico representation of the plant sulfur regulatory complement. Various approaches are and will be followed here: bioinformatic tools have to be developed and/or adapted to fully mine the data. Otherwise, it will not be possible to fully describe the system: by looking only at the most highly expressed genes in isolation, we would simply be scratching at the surface.


Transcriptome Analysis

gene expression

gene expression

Zoom Image

Scatterplots of gene expression of the ratio -/+ S

Plants and some photoautotrophic bacteria assimilate inorganic sulfur from sulfates into cysteine, the first sulfur-containing organic compound and, effectively, the sole molecular doorway for reduced sulfur in all living beings. This essential process has been as finely tuned through millennia of evolution as photosynthesis. Cysteine is subsequently converted to methionine, and then into a variety of other sulfur-containing organic compounds. Sulfur assimilation is even more spendy in terms of reduction equivalents than nitrogen assimilation. Obviously, such a costly enterprise is highly controlled in juxtaposition with the rest of metabolism.

To elucidate this network of interactions, we stimulate Arabidopsis with sulfur (i.e. sulfate) at its rhizosphere with various concentrations and at different developmental stages to institute periods of starvation and replenishment. The plant tissue samples (roots, shoots) are then subjected to array hybridisation/transcript profiling after RNA extraction using either macro-arrays of 7,200 non-redundant genes on nylon filters and now full genome chips. The expression profiles are processed to select differentially expressed genes. Depending on the duration of treatment, anything between a handful and thousands of genes exhibit altered expression mirroring the gradual response of the system to conditions of altered sulfur availability. Among these responsive genes we expect to find sulfur-regulated genes; genes involved in perception, signalling, and immediate responses; and genes further down the line involved in more pleiotropic mechanisms like general stress responses. Since they arise in response to sulfur stimulation, the latter are still regarded as sulfur-responsive genes.

Sulfur-responsive genes are grouped by functional category or biosynthetic pathway. As expected, genes of the sulfur assimilation pathway are altered in expression. Furthermore, genes involved in the flavonoid, auxin, and jasmonate biosynthesis pathways are up regulated when sulfur is limiting. We focus most of our attention, however, on the regulatory elements, transcription factors.

V Nikiforova, J Freitag, S Kempa, M Adamik, H Hesse, R Hoefgen (2003) Transcriptome analysis of sulfur depletion in Arabidopsis thaliana: Interlacing of biosynthetic pathways provides response specificity. The Plant Journal, 33, 633-650

Further reading

MY Hirai, T Fujiwara, M Awazuhara, T Kimura, M Noji, K Saito (2003) Global expression profiling of sulfur-starved Arabidopsis by DNA macroarray reveals the role of O-acetyl-L-serine as a general regulator of gene expression in response to sulfur nutrition. Plant Journal. 33(4)651-663

A Maruyama-Nakashita, E Inoue, A Watanabe-Takahashi, T Yarnaya, and H Takahashi (2003) Transcriptome profiling of sulfur-responsive genes in Arabidopsis reveals global effects of sulfur nutrition on multiple metabolic pathways. Plant Physiology. 132(2)597-605

Sulfur and Other Plant Nutrients

The plant sulfur assimilation pathway is intricately interconnected with various other pathways and regulatory circuits.

Systems Analysis of Plant Sulfur Metabolism

Every organism is a complex, multi-elemental, multi-functional system living in an ever-changing environment. The viability of the system is provided by, and likewise dependent upon, flexible, effective control circuits of multiple informational fluxes, which interconnect in a dense network of metabolic physiological responses.



L-cysteine L-Met

L-cysteine L-Met

Methionine is synthesised from cysteine and phosphohomoserine

Methionine is synthesised from cysteine and phosphohomoserine


Pathway Analysis of Sulfur Containing Amino Acids

To learn about the control mechanisms involved in the biosynthesis of sulfur-containing amino acids, we are isolating and studying genes involved and their promoters. Methionine is synthesised from cysteine and phosphohomoserine via the enzymes cystathionine gamma-synthase (CgS), cystathionine beta-lyase (CbL), and methionine synthase (MS); we have cloned and characterised these three genes in potato.

Biosynthesis of Sulfur-Containing Amino Acids

Biosynthesis of Sulfur-Containing Amino Acids

Genes from Arabidopsis and potato and, when appropriate, E. coli involved in cysteine and methionine biosynthesis have also been cloned, including various isoforms of O-acetylserine (thiol)-lyase, the enzyme that converts O-acetylserine to cys; ATP-sulfurylase, the enzyme activating the inert sulfate through binding to ATP; and serine acetyltransferase (SAT), the enzyme catalysing the activation of serine to O-acetylserine. We manipulated the expression of these genes in an attempt to create conditions in which flux to either cysteine or methionine is increased.

For example, the over-expression of SAT using an E. coli gene targeted to plastids resulted in cysteine and glutathione (a tripeptide containing glutamic acid, cysteine, and glycine) levels almost twice as high as usual. By blocking the competing pathway to threonine using the partial antisense inhibition of threonine synthase in Arabidopsis and potato, we were able to increase leaf and tuber methionine levels significantly. Moreover, analysis of these transformants made it clear that there are species-specific differences in the regulation of methionine biosynthesis.

Our results in Nicotiana plumbaginifolia and potato have established the essential, but not rate-limiting, role of CbL in plant methionine biosynthesis. Furthermore, we found that regulation at the level of CgS differs between the plant species Arabidopsis and potato. Our objective now is to deepen our understanding of the regulation of methionine biosynthesis and to exploit what we learn in order to improve the nutritional quality of crop plants, which is largely determined by methionine content.

Cysteine Biosynthesis

Cysteine biosynthesis represents the essential step in the incorporation of inorganic sulfide to organic sulfur in plants. In order to gain insight into the control mechanisms involved in cysteine biosynthesis, we are isolating and studying the involved genes and their promoters, including genes coding for O-acetylserine(thiol)-lyase (OAS-TL), the enzyme which converts O-acetylserine to cysteine, and serine acetyltransferase (SAT), the enzyme catalysing the activation of serine to O-acetylserine.

Serine acetyltransferase

Serine acetyltransferase

Zoom Image

Serine acetyltransferase

In addition, spatial and developmental aspects of regulation are investigated with respect to gene expression and enzyme activity. We are manipulating the expression of various genes in transgenic potato plants in an attempt to create conditions in which flux to either cysteine or methionine is increased. For example, the heterologous over-expression of an E. coli SAT gene targeted to plastids resulted in a doubling of both cysteine and glutathione (a tripeptide containing glutamic acid, cysteine, and glycine that is involved in stress tolerance) levels. However, these alterations had no effect on outward plant appearance or on the expression and enzymatic activity of OAS-TL. This example demonstrates the importance of SAT in plant cysteine biosynthesis and shows that the accumulation of cysteine and related sulfur-containing compounds is limited by the supply of activated carbon backbones derived from serine. We are currently investigating this and other transgenic plants affected in cysteine and methionine biosynthesis in respect to sulfur assimilation and glutathione-mediated stress tolerance.

Despite the increase of reduced organic sulfur in our potato SAT over-producers, we did not observe an increase in methionine, although other groups reported methionine increases when using a similar approach in maize (Tsakraklides et al., 2002). Again, species specific differences, probably as a result of adaptation to specific environmental or physiological conditions, have to be taken into account, especially when generalising and transferring these data to plant breeding.

V Nikiforova, S Kempa, M Zeh, S Maimann, O Kreft, A P Casazza, K Riedel, E Tauberger, R Hoefgen, H Hesse. (2002) Engineering of cysteine and methionine biosynthesis in potato. Amino Acids 22(259-278).

K Harms, P von Ballmoos, C Brunold, R Höfgen, and H Hesse (2000) Expression of a bacterial serine acetyltransferase in transgenic potato plants leads to increased levels of cysteine and glutathione. Plant J. 22, 335-343

Further reading

MJ Hawkesford (2003) Transporter gene families in plants: the sulphate transporter gene family – redundancy or specialization? Physiologia Plantarum, 117,155-163

G Tsakraklides, M Martin, R Chalam,, MC Tarczynski, A Schmidt, and T Leustek (2002) Sulfate reduction is increased in transgenic Arabidopsis thaliana expressing 5′-adenylylsulfate reductase from Pseudomonas aeruginosa. Plant J. 32, 879

Annu Rev Nutr. 1986;6:179-209.
Metabolism of sulfur-containing amino acids.

Met metabolism occurs primarily by activation of Met to AdoMet and further metabolism of AdoMet by either the transmethylation-transsulfuration pathway or the polyamine biosynthetic pathway. The catabolism of the methyl group and sulfur atom of Met ultimately appears to be dependent upon the transmethylation-transsulfuration pathway because the MTA formed as the co-product of polyamine synthesis is efficiently recycled to Met. On the other hand, the fate of the four-carbon chain of Met appears to depend upon the initial fate of the Met molecule. During transsulfuration, the carbon chain is released as alpha-ketobutyrate, which is further metabolized to CO2. In the polyamine pathway, the carboxyl carbon of Met is lost in the formation of dAdoMet, whereas the other three carbons are ultimately excreted as polyamine derivatives and degradation products. The role of the transamination pathway of Met metabolism is not firmly established. Cys (which may be formed from the sulfur of Met and the carbons of serine via the transsulfuration pathway) appears to be converted to taurine and CO2 primarily by the cysteinesulfinate pathway, and to sulfate and pyruvate primarily by desulfuration pathways in which a reduced form of sulfur with a relatively long biological half-life appears to be an intermediate. With the exception of the nitrogen of Met that is incorporated into polyamines, the nitrogen of Met or Cys is incorporated into urea after it is released as ammonium [in the reactions catalyzed by cystathionase with either cystathionine (from Met) or cystine (from Cys) as substrate] or it is transferred to a keto acid (in Cys or Met transamination). Many areas of sulfur-containing amino acid metabolism need further study. The magnitude of AdoMet flux through the polyamine pathway in the intact animal as well as details about the reactions involved in this pathway remain to be determined. Both the pathways and the possible physiological role of alternate (AdoMet-independent) Met metabolism, including the transamination pathway, must be elucidated. Despite the growing interest in taurine, investigation of Cys metabolism has been a relatively inactive area during the past two decades. Apparent discrepancies in the reported data on Cys metabolism need to be resolved. Future work should consider the role of extrahepatic tissues in amino acid metabolism as well as species differences in the relative roles of various pathways in the metabolism of Met and Cys.

The Sulfur-Containing Amino Acids: An Overview1,2

John T. Brosnan3 and Margaret E. Brosnan

J. Nutr. June 2006; 136(6): 1636S-1640S

Methionine and cysteine may be considered to be the principal sulfur-containing amino acids because they are 2 of the canonical 20 amino acids that are incorporated into proteins. However, homocysteine and taurine also play important physiological roles (Fig. 1). Why does nature employ sulfur in her repertoire of amino acids? The other canonical amino acids are comprised only of carbon, hydrogen, oxygen, and nitrogen atoms. Because both sulfur and oxygen belong to the same group (Group 6) of the Periodic Table and, therefore, are capable of making similar covalent linkages, the question may be restated: why would methionine and cysteine analogs, in which the sulfur atom is replaced by oxygen, not serve the same functions? One of the critical differences between oxygen and sulfur is sulfur’s lower electronegativity. Indeed, oxygen is the second most electronegative element in the periodic table. This accounts for the use of sulfur in methionine; replacement of the sulfur with oxygen would result in a much less hydrophobic amino acid. Cysteine readily forms disulfide linkages because of the ease with which it dissociates to form a thiolate anion. Serine, on the other hand, which differs from cysteine only in the substitution of an oxygen for the sulfur, does not readily make dioxide linkages. The difference results from the fact that thiols are much stronger acids than are alcohols, so that the alcohol group in serine does not dissociate at physiological pH. Substitution of oxygen for sulfur inS-adenosylmethionine would produce so powerful a methylating agent that it would promiscuously methylate cellular nucleophiles without the need for an enzyme.


Structures of the sulfur-containing amino acids.

Methionine and cysteine in proteins.

Although both methionine and cysteine play critical roles in cell metabolism, we suggest that, in general, the 20 canonical amino acids were selected for the roles they play in proteins, not their roles in metabolism. It is important, therefore, to review the role played by these amino acids in proteins. Methionine is among the most hydrophobic of the amino acids. This means that most of the methionine residues in globular proteins are found in the interior hydrophobic core; in membrane-spanning protein domains, methionine is often found to interact with the lipid bilayer. In some proteins a fraction of the methionine residues are somewhat surface exposed. These are susceptible to oxidation to methionine sulfoxide residues. Levine et al. (1) regard these methionine residues as endogenous antioxidants in proteins. In E. coli glutamine synthetase, they tend to be arrayed around the active site and may guard access to this site by reactive oxygen species. Oxidation of these methionine residues has little effect on the catalytic activity of the enzyme. These residues may be reduced to methionine by means of the enzyme methionine sulfoxide reductase (2). Thus, an oxidation–reduction cycle occurs in which exposed methionine residues are oxidized (e.g., by H2O2) to methionine sulfoxide residues, which are subsequently reduced:FormulaFormula

It is considered that the impaired activity of methionine sulfoxide reductase and the subsequent accumulation of methionine sulfoxide residues are associated with age-related diseases, neurodegeneration, and shorter lifespan (2).

Methionine is the initiating amino acid in the synthesis of eukaryotic proteins; N-formyl methionine serves the same function in prokaryotes. Because most of these methionine residues are subsequently removed, it is apparent that their role lies in the initiation of translation, not in protein structure. In eukaryotes, translation initiation involves the association of the initiator tRNA (met-tRNAimet) with eIF-2 and the 40S ribosomal subunit together with a molecule of mRNA. Drabkin and Rajbhandary (3) suggest that the hydrophobic nature of methionine is key to the binding of the initiator tRNA to eIF-2. Using appropriate double mutations (in codon and anticodon), they were able to show that the hydrophobic valine could be used for initiation in mammalian cells but that the polar glutamine was very poor.

Cysteine plays a critical role in protein structure by virtue of its ability to form inter- and intrachain disulfide bonds with other cysteine residues. Most disulfide linkages are found in proteins destined for export or residence on the plasma membrane. These disulfide bonds can be formed nonenzymatically; protein disulfide isomerase, an endoplasmic reticulum protein, can reshuffle any mismatched disulfides to ensure the correct protein folding (4).


S-Adenosylmethionine (SAM)4 is a key intermediate in methionine metabolism. Discovered in 1953 by Cantoni (5) as the “active methionine” required for the methylation of guanidioacetate to creatine, it is now evident that SAM is a coenzyme of remarkable versatility (Fig. 2). In addition to its role as a methyl donor, SAM serves as a source of methylene groups (for the synthesis of cyclopropyl fatty acids), amino groups (in biotin synthesis), aminoisopropyl groups (in the synthesis of polyamines and, also, in the synthesis of ethylene, used by plants to promote plant ripening), and 5′-deoxyadenosyl radicals. SAM also serves as a source of sulfur atoms in the synthesis of biotin and lipoic acid (6). In mammals, however, the great bulk of SAM is used in methyltransferase reactions. The key to SAM’s utility as a methyl donor lies in the sulfonium ion and in the electrophilic nature of the carbon atoms that are adjacent to the sulfur atom. The essence of these methyltransferase reactions is that the positively charged sulfonium renders the adjoining methyl group electron-poor, which facilitates its attack on electron-rich acceptors (nucleophiles).

Metabolic versatility of S-adenosylmethionine.

Metabolic versatility of S-adenosylmethionine.


Metabolic versatility of S-adenosylmethionine.

SAM can donate its methyl group to a wide variety of acceptors, including amino acid residues in proteins, DNA, RNA, small molecules, and even to a metal, the methylation of arsenite (7,8). At present, about 60 methyltransferases have been identified in mammals. However, the number is almost certainly much larger. A bioinformatic analysis of a number of genomes, including the human genome, by Katz et al. (9) has suggested that Class-1 SAM-dependent methyltransferases account for 0.6–1.6% of open reading frames in these genomes. This would indicate about 300 Class 1 methyltransferases in humans, in addition to a smaller number of Class 2 and 3 enzymes. In humans, it appears that guanidinoacetate N-methyltransferase (responsible for creatine synthesis) and phosphatidylethanolamine N-methyltransferase (synthesis of phosphatidylcholine) are the major users of SAM (10). In addition, there is substantial flux through the glycine N-methyltransferase (GNMT) when methionine intakes are high (11). An important property of all known SAM-dependent methyltransferases is that they are inhibited by their product, S-adenosylhomocysteine (SAH).

Methionine metabolism.

Methionine metabolism begins with its activation to SAM (Fig. 3) by methionine adenosyltransferase (MAT). The reaction is unusual in that all 3 phosphates are removed from ATP, an indication of the “high-energy” nature of this sulfonium ion. SAM then donates its methyl group to an acceptor to produce SAH. SAH is hydrolyzed to homocysteine and adenosine by SAH hydrolase. This sequence of reactions is referred to as transmethylation and is ubiquitously present in cells. Homocysteine may be methylated back to methionine by the ubiquitously distributed methionine synthase (MS) and, also, in the liver as well as the kidney of some species, by betaine:homocysteine methyltransferase (BHMT). MS employs 5-methyl-THF as its methyl donor, whereas BHMT employs betaine, which is produced during choline oxidation as well as being provided by the diet (10). Both MS and BHMT effect remethylation, and the combination of transmethylation andremethylation comprise the methionine cycle, which occurs in most, if not all, cells.

Major pathways of sulfur-containing amino acid metabolism.

Major pathways of sulfur-containing amino acid metabolism.

Major pathways of sulfur-containing amino acid metabolism.

The methionine cycle does not result in the catabolism of methionine. This is brought about by the transsulfuration pathway, which converts homocysteine to cysteine by the combined actions of cystathionine β-synthase (CBS) and cystathionine γ-lyase (CGL). The transsulfuration pathway has a very limited tissue distribution; it is restricted to the liver, kidney, intestine, and pancreas. The conversion of methionine to cysteine is an irreversible process, which accounts for the well-known nutritional principle that cysteine is not a dietary essential amino acid provided that adequate methionine is available, but methionine is a dietary essential amino acid, regardless of cysteine availability. This pathway for methionine catabolism suggests a paradox: is methionine catabolism constrained by the need for methylation reactions? If this were so, the methionine in a methionine-rich diet might exceed the methylation demand so that full catabolism could not occur via this pathway. GNMT methylates glycine to sarcosine, which may, in turn, be metabolized by sarcosine dehydrogenase to regenerate the glycine and oxidize the methyl group to 5,10-methylene-THF.

Application of sophisticated stable isotope tracer methodology to methionine metabolism in humans has yielded estimates of transmethylation, remethylation, and transsulfuration. Such studies reveal important points of regulation. For example, the sparing effect of cysteine on methionine requirements is evident as an increase in the fraction of the homocysteine pool that is remethylated and a decrease in the fraction that undergoes transsulfuration (12). In young adults ingesting a diet containing 1–1.5 g protein·kg−1·d−1, about 43% of the homocysteine pool was remethylated, and 57% was metabolized through the transsulfuration pathway (transmethylation = 9.7, transulfuration = 5.4, remethylation = 4.4 μmol·kg−1·h−1) (13).

Methionine metabolism affords a remarkable example of the role of vitamins in cell chemistry. MS utilizes methylcobalamin as a prosthetic group, 1 of only 2 mammalian enzymes that are known to require Vitamin B-12. The methyl group utilized by MS is provided from the folic acid 1-carbon pool. Methylenetetrahydrofolate reductase (MTHFR), which reduces 5,10-methylene-THF to 5-methyl-THF, contains FAD as a prosthetic group. Both of the enzymes in the transsulfuration pathway (CBS and CGL) contain pyridoxal phosphate. It is hardly surprising, therefore, that deficiencies of each of these vitamins (Vitamin B-12, folic acid, riboflavin, and pyridoxine) are associated with elevated plasma homocysteine levels. The oxidative decarboxylation of the α-ketobutyrate produced by CGL is brought about by pyruvate dehydrogenase so that niacin (NAD), thiamine (thiamine pyrophosphate), and pantothenic acid (coenzyme A) may also be regarded as being required for methionine metabolism.

Not only are vitamins required for methionine metabolism, but methionine metabolism plays a crucial role in the cellular assimilation of folate. MS has 2 principal functions. In addition to its role in methionine conservation, MS converts 5-methyl-THF to THF, thereby making it available to support DNA synthesis and other functions. Because 5-methyl-THF is the dominant circulating form that is taken into cells, MS is essential for cellular folate assimilation. Impaired MS activity (e.g., brought about by cobalamin deficiency) results in the accumulation of the folate coenzymes as 5-methyl-THF, the so-called methyl trap (14). This hypothesis explains the fact that Vitamin B-12 deficiency causes a functional cellular folate deficiency.

The combined transmethylation and transsulfuration pathways are responsible for the catabolism of the great bulk of methionine. However, there is also evidence for the occurrence of a SAM-independent catabolic pathway that begins with a transamination reaction (15). This is a very minor pathway under normal circumstances, but it becomes more significant at very high methionine concentrations. Because it produces powerful toxins such as methane thiol, it has been considered to be responsible for methionine toxicity. The identity of the initiating transaminase is uncertain; the glutamine transaminase can act on methionine, but it is thought to be unlikely to do so under physiological conditions (15). In view of the likelihood that this pathway plays a role in methionine toxicity, more work is warranted on its components, tissue distribution, and physiological function.

Regulation of methionine metabolism.

The major means by which methionine metabolism is regulated are 1) allosteric regulation by SAM and 2) regulation of the expression of key enzymes. In the liver, SAM exerts powerful effects at a variety of loci. The liver-specific MAT has a highKm for methionine and, therefore, is well fitted to remove excess dietary methionine. It exhibits the unusual property of feedback activation; it is activated by its product, SAM (16). This property has been incorporated into a computer model of hepatic methionine metabolism, and it is clear that it renders methionine disposal exquisitely sensitive to the methionine concentration (17). SAM is also an allosteric activator of CBS and an allosteric inhibitor of MTHFR (18). Therefore, elevated SAM promotes transsulfuration (methionine oxidation) and inhibits remethylation (methionine conservation). Many of the enzymes involved in methionine catabolism (MAT 1, GNMT, CBS) are increased in activity on ingestion of a high-protein diet (18).

In addition to its function in methionine catabolism, the transsulfuration pathway also provides cysteine for glutathione synthesis. Cysteine availability is often limiting for glutathione synthesis, and it appears that in a number of cells (e.g., hepatocytes), at least half of the cysteine required is provided by transsulfuration, even in the presence of physiological concentrations of cysteine (19). Transsulfuration is sensitive to the balance of prooxidants and antioxidants; peroxides increase the transsulfuration flux, whereas antioxidants decrease it (20). It is thought that redox regulation of the transsulfuration pathway occurs at the level of CBS, which contains a heme that may serve as a sensor of the oxidative environment (21).


Taurine is remarkable, both for its high concentrations in animal tissues and because of the variety of functions that have been ascribed to it. Taurine is the most abundant free amino acid in animal tissues. Table 1 shows that, although taurine accounts for only 3% of the free amino acid pool in plasma, it accounts for 25%, 50%, 53%, and 19%, respectively, of this pool in liver, kidney, muscle, and brain. The magnitude of the intracellular taurine pool deserves comment. For example, skeletal muscle contains 15.6 μmol of taurine per gram of tissue, which amounts to an intracellular concentration of about 25 mM. In addition to its role in the synthesis of the bile salt taurocholate, taurine has been proposed, inter alia, to act as an antioxidant, an intracellular osmolyte, a membrane stabilizer, and a neurotransmitter. It is an essential nutrient for cats; kittens born to mothers fed taurine-deficient diets exhibit retinal degeneration (24). Taurine is found in mother’s milk, may be conditionally essential for human infants, and is routinely added to most infant formulas. Recent work has begun to reveal taurine’s action in the retina. It appears that taurine, via an effect on a glycine receptor, promotes the generation of rod photoreceptor cells from retinal progenitor cells (25).

View this table:


Taurine concentrations in rat tissues (22,23)


The sulfur-containing amino acids present a fascinating subject to the protein chemist, the nutritionist, and the metabolic scientist, alike. They play critical roles in protein synthesis, structure, and function. Their metabolism is vital for many critical functions. SAM, a remarkably versatile molecule, is said to be second, only to ATP, in the number of enzymes that require it. Vitamins play a crucial role in the metabolism of these amino acids, which, in turn, play a role in folic acid assimilation. Despite the great advances in our knowledge of the sulfur-containing amino acids, there are important areas where further work is required. These include methionine transamination and the molecular basis for the many functions of taurine.

Disorders of Sulfur Amino Acid Metabolism

  • Generoso Andria,  Brian Fowler,  Gianfranco Sebastio

Chapter  Inborn Metabolic Diseases  pp 224-231


Several defects can exist in the conversion of the sulfur-containing amino acid methionine to cysteine and the ultimate oxidation of cysteine to inorganic sulfate (Fig. 18.1). Cystathionine-β-synthase (CBS) deficiency is the most important. It is associated with severe abnormalities of four organs or organ systems: the eye (dislocation of the lens), the skeleton (dolichostenomelia and arachnodactyly), the vascular system (thromboembolism), and the central nervous system (mental retardation, cerebrovascular accidents). A low-methionine, highcystine diet, pyridoxine, folate, and betaine in various combinations, and antithrombotic treatment may halt the otherwise unfavorable course of the disease. Methionine adenosyltransferase deficiency and γ-cystathionase deficiency usually do not require treatment. Isolated sulfite oxidase deficiency leads (in its severe form) to refractory convulsions, lens dislocation, and early death. No effective treatment exists.

  1. 1.

    Rubba P, Faccenda F, Pauciullo P, Carbone L, Mancini M, Strisciuglio P, Carrozzo R, Sartorio R, Del Giudice E, Andria G (1990) Early signs of vascular disease in homocystinuria: a noninvasive study by ultrasound methods in eight families with cystathionine ß-synthase deficiency. Metabolism 39: 1191–1195 PubMedCrossRef

  2. 2.

    Kang S-S, Wong PWK, Malinow MR (1992) Hyperhomocyst(e)inemia as a risk factor for occlusive vascular disease. Annu Rev Nutr 12: 279–288 PubMedCrossRef

  3. 3.

    Boushey CJ, Beresford SA, Omenn GS, Motulsky AG (1995) A quantitative assessment of plasma homocysteine as a risk factor for vascular disease. Probable benefits of increasing folic acid intakes. JAMA 274: 1049–1057

  4. 4.

    Mudd SH, Skovby F, Levy HL, Pettigrew KD, Wilcken B, Pyeritz RE, Andria G, Boers GHJ, Bromberg IL, Cerone R, Fowler B, Grobe H, Schmidt H, Schweitzer L (1985) The natural history of homocystinuria due to cystathionine (3-synthase deficiency. Am J Hum Genet 37: 1–31 PubMed

  5. 5.

    de Franchis R, Sperandeo MP, Sebastio G, Andria G. The Italian Collaborative Study Group on Homocystinuria (1998) Clinical aspects of cystathionine ß-synthase deficiency: how wide is the spectrum? Eur J Pediatr 157: S67–7o

  6. 6.

    Kraus JP (1994) Molecular basis of phenotype expression in homocystinuria. J Inherited Metab Dis 17: 383–390 PubMedCrossRef

  7. more…

Read Full Post »

Protein-binding, Protein-Protein interactions & Therapeutic Implications

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

7.3  Protein-binding, Protein-Protein interactions & Therapeutic Implications

7.3.1 Action at a Distance. Allostery_Delabarre_allostery review

7.3.2 Chemical proteomics approaches to examine novel histone modifications

7.3.3 Misfolded Proteins – from Little Villains to Little Helpers… Against Cancer

7.3.4 Endoplasmic reticulum protein 29 (ERp29) in epithelial cancer

7.3.5 Putting together structures of epidermal growth factor receptors

7.3.6 Complex Relationship between Ligand Binding and Dimerization in the Epidermal Growth Factor Receptor

7.3.7 IGFBP-2.PTEN- A critical interaction for tumors and for general physiology

7.3.8 Emerging-roles-for-the-Ph-sensing-G-protein-coupled-receptor

7.3.9 Protein amino-terminal modifications and proteomic approaches for N-terminal profiling

7.3.10 Protein homeostasis networks in physiology and disease

7.3.11 Proteome sequencing goes deep

7.3.1 Action at a Distance. Allostery_Delabarre_allostery review

DeLaBarre B1Hurov J1Cianchetta G1Murray S1Dang L2.
Chem Biol. 2014 Sep 18; 21(9):1143-61

Cancer cells must carefully regulate their metabolism to maintain growth and division under varying nutrient and oxygen levels. Compelling data support the investigation of numerous enzymes as therapeutic targets to exploit metabolic vulnerabilities common to several cancer types. We discuss the rationale for developing such drugs and review three targets with central roles in metabolic pathways crucial for cancer cell growth: pyruvate kinase muscle isozyme splice variant 2 (PKM2) in glycolysis, glutaminase in glutaminolysis, and mutations in isocitrate dehydrogenase 1 and 2 isozymes (IDH1/2) in the tricarboxylic acid cycle. These targets exemplify the drugging approach to cancer metabolism, with allosteric modulation being the common theme. The first glutaminase and mutant IDH1/2 inhibitors have entered clinical testing, and early data are promising. Cancer metabolism provides a wealth of novel targets, and targeting allosteric sites promises to yield selective drugs with the potential to transform clinical outcomes across many cancer types.

Based on knowledge acquired to date, there is no doubt that cancer metabolism provides a wealth of novel therapeutic targets and multiple innovative ways in which to exploit metabolic vulnerabilities for therapeutic benefit. More comprehensive reviews cover the breadth of metabolic targets that are currently under investigation (Stine and Dang, 2013; Vander Heiden, 2011). The following sections of this review focus on PKM2, glutaminase, and mutated IDH1/2 as exemplary metabolism targets under investigation for development of cancer therapies.
Drugging Glycolysis: Targeting Pyruvate Kinase Muscle Isozyme Alternative Splice Variant 2 PK catalyzes the last step of glycolysis, converting phosphoenolpyruvate (PEP) to pyruvate, while producing one molecule of ATP. The reaction encompasses two chemical steps: the first involves a phosphoryl transfer from PEP to ADP, forming an enolate intermediate and ATP, and the second involves protonation of the enolate intermediate, forming pyruvate (Robinson and Rose, 1972). PKM2 is one of four PK isoforms in humans. PKM1 and PKM2 result from the alternative splicing of exons 9 and 10 of the PKM gene, which encode a stretch of amino acids that differ at 23 positions between PKM1 and PKM2. PKM1 is constitutively active in skeletal muscle and brain tissue, but is not allosterically regulated. PKM2 is expressed in fetal and proliferating tissues, has low basal activity compared with PKM1, and is allosterically regulated. R-type pyruvate kinase (PKR) and L-type pyruvate kinase (PKL) are transcribed via different promoters from the PKLR gene. PKR is expressed in erythrocytes and PKL in the liver. PKR, PKL, and PKM1 exist as stable tetramers,whereas PKM2 forms tetramers (high activity form), dimers (low activity form), and monomers (Mazurek, 2011).

Figure 1. Central Metabolic Pathways Utilized by Cancer Cells *denotes mutated isoenzyme.

Pyruvate Kinase Muscle Isozyme Alternative Splice Variant 2 in Cancer Cell Metabolism Cancer cells predominantly express PKM2, which can be downregulated by tyrosine kinase growth factor signaling pathways, allowing metabolic flexibility. Phosphotyrosine peptides have been shown to suppress PKM2 activity by binding tightly to PKM2, thereby catalyzing the release of fructose 1,6-bisphosphate (FBP), resulting in a switch to the low activity dimer state (Christofk et al., 2008b; Hitosugi et al., 2009). This downregulation is thought to support tumor growth and proliferation by allowing for the shunting of glycolytic intermediates toward other biosynthetic pathways (i.e., pentose phosphate and serine pathways). In keeping with this model, the activation of PKM2 in cancer cells using small molecule agonists resulted in serine auxotrophy (Kung et al., 2012). Consistent with the hypothesis that PKM2 is a critical metabolic switch, there is growing evidence that, depending on the cellular stress environment, PKM2activity canberegulated byposttranslational modification such as acetylation (Lv et al., 2011), phosphorylation (Hitosugi et al., 2009), cysteine oxidation (Anastasiou et al., 2011), and proline hydroxylation (Luo et al., 2011). The utility of PKM2 activators in the clinic has yet to be determined, but recent work with tumor xenografts with a PKM2 activator suggests that this may be a viable approach (Parnell et al., 2013). As PKM2 tetramers show greater than 50-fold higher activity than PKM2 monomers (Anastasiou et al., 2012), one consideration when designing drugs to activate PKM2 for therapeutic means would be the need for small-molecule ligands capable of driving the enzyme toward its optimally active tetrameric form, thus forcing cancer cells into a less flexible metabolic state.

Structure of Pyruvate Kinase Muscle Isozyme Alternative Splice Variant 2 The structure of the PKM2 tetramer is summarized in Figure 2A. PKM2 is allosterically activated in a ‘‘feedforward’’ manner by the upstream glycolytic metabolite, FBP, which induces a shift to the active tetrameric conformation (Christofk et al., 2008b; Dombrauckas et al., 2005). PKM2 can be independently allosterically activated by serine (Chaneton et al., 2012), which binds in a distinct pocket that can also accommodate the inhibitor phenylalanine (Protein Data Bank [PDB] ID: 4FXJ). The binding of phenylalanine results in a tetrameric form distinct from the active conformer (Morgan et al., 2013). It is not clear how the change from serine to phenylalanine elicits such a dramatic change in protein behavior, or whether there is any biological interaction between serine activation and phenylalanine inhibition of PKM2 in cancer cells. Of note, PKM1 and PKL/R are not activated by serine, despite the conservation of the serine binding site in all PK isoforms.
Figure 2. Three Different Metabolic Enzymes and Their Allosteric Inhibitors Protomers are depicted as cartoon ribbons in blue, green, yellow, and cyan. Synthetic allostery is depicted in stick format with red highlight. (A) Structure of tetrameric PKM2:AGI-980 (4:2 complex) from PDB 4G1N. AGI-980 is shown in stick rendering near the center of tetramer. Each PK monomer consists of four domains, usually designated A, B, C, and N (Dombrauckas et al., 2005). The tetramer is a dimer-of-dimers with approximate D2 symmetry. The dimer is formed between the A domains of each monomer, while the tetramer is formed via dimerization along the C subunit interfaces of each dimer. The active site of PKM2 lies within a cleft between the A and B domain, illustrated by a PEP analog (red spheres). The FBP binding pocket is located entirely within the C domain (pink spheres). The natural allosteric site of serine is also shown (black spheres). (B)Tetrameric GAC:BPTES (4:2 complex) from PDB 3UO9. Glutamate molecules are shown as spheres. (C) Dimeric IDH2R140Q:AGI-6780 (2:1 complex) from PDB 4JA8 (Wang et al., 2013). NADP molecules are shown as spheres.
Discovery of Allosteric Activators of Pyruvate Kinase Muscle Isozyme Alternative Splice Variant 2 A number of small molecules that potently activate PKM2 have been discovered by various groups (Table 1). Interestingly, all seven X-rayco-complexescurrentlyavailableshowcompoundsbound at a novel binding pocket distinct from the FBP and serine binding sites, which would otherwise allow cells to overcome negative regulation by phosphotyrosines (Kung et al., 2012). The compounds found in structures 3GQY, 3GR4 (Boxer et al., 2010), 3H6O (Jiang et al., 2010), 3ME3, and 3U2Z (Anastasiou et al., 2012) were identified by screening the NIH Small Molecule Repository, and can be classified into two distinct chemical series, both of which establish very similar interactions with PKM2 (Table 1). Analogues in these two classes selectively activated PKM2 allosterically with good selectivity against PKM1, PKL, and PKR (Anastasiou et al., 2012; Boxer et al., 2010; Jiang et al., 2010). The molecule found in the structure 4JPG (Guo et al., 2013) is similar to the two series described above, where the pyrimidone ring is found between the two Phe26 residues (Table 1). Interestingly, the activator found in the 4G1N structure (Kung et al., 2012) sits in the same pocket as the NIH compounds. However, the interactions are quite different, with the side chains of the two Phe26 that line the pocket assuming distinct conformations. This activator wraps around the two aromatic residues, which pushes it closer to the walls of the pocket, allowing for a richer series of interactions with PKM2 (Table 1). There are two additional series of PKM2 activators that have been reported for which no structural information is available (Table 1)(Parnell et al., 2013; Xu et al., 2014; Yacovan et al., 2012). Members of this series were shown to have an activation level comparable to that of FBP, with selectivity for PKM2 over PKL, PKR, and PKM1. PKM2 offers a very interesting example of an allosterically regulated enzyme. Different allosteric sites have so far been identified for three different types of activator (FBP, serine, and small-molecule ligands) and all activate PKM2 by stabilizing the tetrameric form. It is remarkable that molecules as small as serine can dramatically alter this protein’s conformational landscape and aggregation state and lead to an active enzyme. This unusual allosteric site revealed by the small-molecule ligands is of particular curiosity, largely because neither its function nor its native ligands are known. All of the drug-like activators described above bind at the dimer–dimer interface and seem to act by displacing water from the mainly apolar pocket, thus contributing to the stabilization of the tetramer. While these PKM2 activators show promising preclinical data, none have yet entered clinical development.

Table 1. Biochemical Properties of Small Molecule PKM2 Inhibitors Series PDB ID Ligand Reference Binding Characteristics

Substituted N,N’diarylsulfonamide 3GQY (Boxer et al., 2010)

  •  All completely buried within A-A’ interface, 35A ˚ from FBP pocket
  •  Binding pocket lined with residues equivalent to those of PKM2 molecules forming A-A’ interface
  •  All sandwiched between phenyl rings of the two Phe26 from different monomers
  •  All additionally interact with side chain of Phe26 through slightly distorted T-shaped p-p interactions (two such interactions for substituted N,N0diarylsulfonamides and one for thieno[3,2-b]pyrrole[3,2-] pyridazinones)
  1. 3GR4 (Boxer et al., 2010) 3ME3 (Anastasiou et al., 2012)
  2. Thieno[3,2-b]pyrrole [3,2-d]pyridazinone 3H6O (Jiang et al., 2010)
  3. 3U2Z (Anastasiou et al., 2012)
  4. 2-((1H-benzo[d]imidazol1-yl)methyl)-4H-pyrido [1,2-a]pyrimidin-4-ones
  5. 4JPG (Guo et al., 2013)
  • Pyrimidone ring found between the two Phe26 residues forming p-p interactions with the aromatic rings
  • Carbonyl interacts with a bridging water molecule
  • Benzimidazole reaches a region of the activator pocket that is not occupied in any of the published crystal structures
  • One of the imidazole nitrogens forms an H-bond with Lys311, which is normally part of a salt bridge to Asp354

Quinolone sulfonamides 4G1N (Kung et al., 2012)

  •  Quinoline moiety sits on a flat, mainly apolar surface defined by Phe26, Leu27 and Met30 from chain A and Phe26, Tyr390 and Leu394 from chain A’
  •  One of the two oxygen atoms of the sulfonamide accepts an H bond from the backbone oxygen of Tyr390, the other interacts with a water molecule
  •  The oxygen of the amide moiety forms an H-bond with side-chain nitrogen of Lys311
  •  Terminal aromatic ring sits in the other copy of the quinoline pocket d Aromatic rings of the side chains of the two Phe26 lining the pocket almost perpendicular (not parallel); activator wrapped around the two aromatic residues

3-(trifluoromethyl)-1Hpyrazole-5-carboxamide (Parnell et al., 2013; Xu et al., 2014)

  • Cocrystal structure of one compound bound to tetrameric PKM2 obtained but file not available for download from PDB: described as bound to the allosteric site at the dimer–dimer interface

5-((2,3-dihydrobenzo[b] [1,4]dioxin-6-yl)sulfonyl)-2methyl-1-(methylsulfonyl) indoline scaffold (Yacovan et al., 2012)

  • Cocrystal structure of one compound bound to PKM2 obtained but not available for download from the PDB: described as bound to dimer interface
  • Interactions very similar to those established by thieno [3,2-b]pyrrole[3,2-d]pyridazinone series above

Drugging Glutaminolysis: Targeting the Glutaminase C Variant Glutaminase catalyzes the conversion of glutamine to glutamate and ammonia. Glutamate can be oxidized to a-ketoglutarate (aKG), which then anaplerotically feeds into the TCA cycle as a means of providing proliferating cells with biosynthetic intermediates and ATP (Figure 1); glutamate is also used as a substrate for the generation of glutathione, which provides protection from redox stress (Hensley et al., 2013; Shanware et al., 2011). The ammonia produced during the reaction can be used in certain tissues like the kidney to provide pH homeostasis, and nitrogen derived from glutamine is utilized in nucleotide biosynthetic and glycosylation pathways.

Table 2. Characteristics of Small Molecule Glutaminase Inhibitors

BPTES N-(5–[1,3,4]thiadiazol-2yl)-2-phenylacetamide 6 (Shukla et al., 2012)

  • Similar potency but better water solubility vs. BPTES d Attenuated growth of P493 human lymphoma B cells in vitro d Diminished tumor growth in P493 tumor xenograft SCID mice with no apparent toxicity

CB-839 (Calithera) (Gross et al., 2014)

  • Orally bioavailable d Binds at allosteric sites of GLS1 KGA and GAC d Potent, selective, time-dependent reversible inhibition with slow recovery time
  • Anti-proliferative activity (double-digit nM potency) in cellular proliferation assays in wide range of tumors
  • Currently in Phase I trials of locally-advanced/metastatic refractory solid tumors (triple negative breast cancer, NSCLC, RCC, mesothelioma) and hematological cancers [ NCT02071927, NCT02071862, NCT02071888]

Dibenzophenanthridines Compound 968 (Katt et al., 2012; Wang et al., 2010)

  • Modest potency in the low mM concentrations d Loses all inhibitory activity against glutaminase activated by preincubation with inorganic phosphate (phosphate does not affect BPTES potency)
  • Anti-proliferative activity in breast cancer cell line at 10 mmol/L concentration

There are three isoforms of IDH. IDH1 is located in both the peroxisome and the cytosol, whereas IDH2 and IDH3 are located in mitochondria. It is unclear what the relative contributions of the IDH2 and IDH3 isoforms are to overall mitochondrial TCA function. IDH1 and IDH2 are both obligatory homodimeric proteins and use NADP+ as a cofactor, whereas IDH3 uses NAD+ as a cofactor and is a heterotrimeric protein comprising alpha, beta, and gamma subunits. All three isozymes require either Mg2+ or Mn2+ asdivalent metal cofactors for catalysis.The dimeric structure of IDH2 is shown in Figure 2C.

Mutant Isocitrate Dehydrogenase in Cancer Cell Metabolism The role of IDH mutations in cancer metabolism was recognized following the observation of frequent and recurrent mutations of IDH1 and IDH2 in patients with glioma and AML, initially identified by genomic deep sequencing and subsequent comparative genetic analyses (Parsons et al., 2008; Yan et al., 2009). These mutations were originally characterized as loss of function (Mardis etal.,2009; Parsonsetal.,2008; Yanet al.,2009), suggesting that mutated IDH acts as a tumor suppressor due to the loss of catalytic conversion of isocitrate to aKG (Zhaoetal., 2009). However, with the exception of cases of haploinsufficiency, the heterozygous mutation pattern of IDH is more consistent with an oncogene role. Subsequently, IDH mutations were shown to possess the neomorphic activity to generate the oncometabolite, 2-hydroxyglutarate (2HG) (Dang et al., 2009; Gross et al., 2010; Ward et al., 2010). With a single codon substitution, the kinetic properties of the mutant IDH isozyme are significantly altered, resulting in an obligatory sequential ordered reaction in the reverse direction (Rendina et al., 2013). Indeed, the critical kinetic observation of mutant IDH was not only the loss of affinity for isocitrate, but also a dramatic increase in NADPH affinity by three orders of magnitude (Dang et al.,2009), suggesting a substantial change in protein dynamics imparted by the mutation. The only known homeostatic 2HG clearance mechanism is the relatively inefficient reconversion of 2HG back to aKG by D-2hydroxyglutarate dehydrogenase. Therefore, 2HG accumulates when over-produced by mutant IDH. 2HG itself has been shown to be sufficient to drive the malignant phenotype (Rakheja et al., 2013). Abnormally high 2HG levels impair aKG-dependent dioxygenases through competitive inhibition, including those that modify DNA and histones (i.e., Jumonji domain-containing histone demethylases and the ten-eleven translocation (TET) family of 50-methylcytosine hydroxylases) (Chowdhury et al., 2011; Figueroa et al., 2010), as well as EglN prolyl hydroxylase in regulating hypoxia-inducible factor (Losman et al., 2013). This results in altered epigenetic status that blocks cell differentiation. These findings, combined with the inhibitory effects of fumarate and succinate on the same families of aKG-dependent enzymes, highlight a critical and fascinatingnetwork that ties together central metabolic pathways and epigenetic control. Remarkably, mutations in TET2 are mutually exclusive with IDH mutations in AML, strongly suggesting that, in this context, the tumorigenic effects of 2HG are at least in part driven by inhibition of TET2. The precise targets of IDH mutations with associated 2HG production (and TET2 mutations) that promote tumorigenesis are currentlyunknown;however,itisclearthatIDH1/2andTET2mutations lead to a block in hematopoietic cell differentiation (Figueroa et al., 2010; Lu et al., 2012; Moran-Crusio et al., 2011; Wang et al., 2013). To date, no IDH3 mutation associated with cancer has been reported (Krell et al., 2011; Reitman and Yan, 2010), suggesting that the role of IDH1/2 has a greater impact on tumorigenesis. Targeting mutated isoforms of IDH1/2 therefore presents a logical approach to cancer therapy. A consideration in designing suchdrugsistheheterozygoussomaticnatureoftheIDH1/2mutation, which likely yields a mixture of homo- and heterodimers; statistically, heterodimers should be the major species in vivo. Mutant homodimers and wild-type-mutant heterodimers both efficiently catalyze the production of 2HG from aKG (Dang et al., 2009; Rendina et al., 2013). However, the heterodimer is potentially more oncogenic, as it is more efficient at producing 2HG than homodimeric mutants (Pietrak et al., 2011) due to an increased local concentration of substrate while conserving NADPH. The heterodimer as a molecular target therefore becomes an important consideration in this scenario.

Structure of Isocitrate Dehydrogenase Structurally, both IDH1 and IDH2 comprise three main domains: the large domain, the small domain, and the clasp region (Yang et al., 2010). A simplified description of protein motion is provided in Figure 3 (Rendina et al., 2013; Xu et al., 2004). The dynamic of motion may differ slightly between IDH1 and IDH2 mutants. IDH1 mutants appear to open wider than IDH2 mutants to the point of unwinding a helix termed ‘‘seg2’’ (Yang et al., 2010). In contrast, the open form of IDH2 does not involve the melting of any secondary structure, and as a consequence has a much narrower range of motion (Taylor et al., 2008; Wang et al., 2013). This differential in protein dynamics could possibly explain the differential responses of IDH1 and IDH2 to inhibitors. X-ray structures of IDH3 have not yet been reported, but it appears to be distinct from IDH1 and IDH2 in terms of primary sequence and predicted quaternary organization (Kim et al., 1995; Ramachandran and Colman, 1980). There are three arginine residues in the enzyme active site that are predicted to play a central role in electrostatic stabilization and proper geometric orientation of isocitrate via its acidic moieties as the substrate binds in the active site. With the exception of the novel G97D or G97N codon mutation (Ward et al., 2012), virtually all confirmed IDH mutations that generate high levels of 2HG occur in one of these arginines (i.e., IDH1-R132 and IDH2-R172/R140) (Losman and Kaelin, 2013) and have in common a substitution of one of the diffuse positive charges of the respective arginine’s guanidinium moiety.
Discovery of Inhibitors against Mutated Isocitrate Dehydrogenase Several inhibitors of mutant IDH isoforms that block 2HG production in vitro and in vivo have been recently described. The first potent and specific IDH1 inhibitors reported were the phenylglycine series, specifically AGI-5198 (Popovici-Muller et al., 2012; Rohle et al., 2013) and subsequently ML309 (Davis et al., 2014)(Table 3), which were shown to be rapid-equilibrium inhibitors specific for IDH1-R132-codon mutations. These compounds inhibited IDH1-R132H competitively with respect to aKG and uncompetitively with respect to NADPH, suggesting that they preferably bind to the enzyme-NADPH ternary complex. Notably, they do not appreciably cross-react against the IDH2-R140Q mutant isozyme, suggesting a unique binding mode in IDH1-R132 that does not favorably exist in IDH2R140. Because no X-ray co-complex has been reported for this series, the exact mode of binding cannot be ascertained at this time. Preclinical data indicated 2HG inhibition and antitumor effects in vitro and in vivo (Table 3). These phenylglycine compounds appear to be excellent chemical tools for tumor biology investigation, but optimization of their properties is likely required for further therapeutic development. Co-complexes of IDH1-R132H with two different 1-hydroxypyridin-2-one inhibitors have been reported (Zheng et al., 2013), but the quality of the crystal structure data supporting the mechanism of inhibition is poor. AG-120, a selective, potent inhibitor of mutated IDH1, is currently in clinical development for the treatment of cancers with IDH1 mutations (Table 3), but there is currently no published information on this inhibitor. Another inhibitor of mutated IDH1 has been reported recently (Table 3) (Deng et al., 2014). Co-complex X-ray studies revealed that Compound1 binds mutated IDH1 allosterically at the dimer interface resulting in an asymmetric open conformation. Distinctively, Compound 1 displaces the conserved catalytic Tyr139 and further disrupts the Mg2+ binding network, consistent with kinetic results of competitive inhibition with respect to Mg2+, but not with aKG substrate. Others have reported modeling of inhibitors into the active site of IDH1, but experimental evidence is lacking (Chaturvedi et al., 2013; Davis et al., 2014). The first reported potent and selective IDH2 inhibitor was the urea-sulfonamide series, AGI-6780 (Wang et al., 2013), a timedependent slow-tight binder to IDH2-R140Q exhibiting noncompetitive inhibition with respect to substrate and uncompetitive inhibition with respect to NADPH, and nanomolar potency for 2HG inhibition (Table 3). This compound showed good inhibitory selectivity for IDH2-R140Q, with no effect on the closely related IDH1 and IDH1-R132H isozymes. At doses that effectively blocked 2HG to basal levels, AGI-6780 induced differentiation of TF-1 erythroleukemia and primary human AML cells in vitro, suggesting potential to reverse leukemic phenotype in AML tumors harboring the IDH2 mutation. Unlike the case of IDH1 above, the published structure of AGI-6780 co-complexed with IDH2-R140Q allows for detailed analysis of its inhibitory mechanism (Wang et al., 2013). In the X-ray structure, a single molecule
of AGI-6780 binds at the interface of two protomers (Figure 2C). The allosteric inhibition appears to arise from the ability of AGI6780 to keep the IDH2-R140Q mutant enzyme in an open orientation, thereby preventing the NADPH cofactor and substrate aKG from coming close to the catalytic Mg2+ binding site (see Figure 3). The highly symmetric AGI-6780 binding pocket extends deep into the protein interface and is closed over by loops composed of residues 152–167, which also fold over the binding pocket, providing anexplanation for the time-dependent inhibition kinetics. AGI-6780 makes several direct H-bond interactions from its urea group and amide nitrogen to Gln316, but a significant amount of binding energy arises from van der Waals contacts between the protein and hydrophobic surfaces of AGI-6780. The in vivo potential for this compound is not known, since its pharmacokinetic properties were not reported. Nevertheless, this effective mode of inhibition serves as an important molecular model for the design of bioisosteric compounds. OtherIDH2inhibitorsareunderdevelopment,notablyAG-221, a first-in-class, orally available inhibitor (Table 3) which demonstrated a survival advantage in a preclinical study of a primary human IDH2 mutant AML xenograft mouse model (Yen et al., 2013). Early phase I clinical trial data for AG-221 show promise, with meaningful clinical responses in evaluable AML patients harboring IDH2 mutations (Stein et al., 2014). To date, there is no published example of a molecule that inhibits both IDH1 and IDH2 mutant isoforms with equipotency.

Table 3.Characteristics of Small Molecule Inhibitors of Mutant IDH

PhenylglycineAGI-5198 (Popovici-Mulleretal., 2012; Rohleetal.,2013)
N-cyclohexyl-2-(N-(3-fluorophenyl)-2(2-methyl-1H-imidazol-1-yl)acetamido)2-(o-tolyl)acetamide IDH1-R132H

  • Good potency against enzyme and in U87cell line overexpressing R132H mutation (IC50= 70nM)
  • Good oral exposure in rodents at high doses (>300mg/kg), which were likely at levels saturating hepatic clearance mechanisms
  • Plasma 2HG inhibition > 90% (BID dosing) in xenograft model of U87-R132H tumors
  • Promoted differentiation of glioma cells via induced demethylation of histone H3K9me3 and expression of genes associated with gliogenic differentiation at near-complete 2HG inhibition
  • inhibited plasma 2HG and delayed growth of IDH1-mutant but not wild-type glioma xenografts in mice

ML309 (Davis et al.,2014)
2-(2-(1H-benzo[d]imidazol-1-yl)-N-(3fluorophenyl)acetamido)-N-cyclopentyl2-o-tolylacetamide IDH1-R132H IDH1-R132C dIC50=68nM(R132H)

  • Inhibited 2HG production in glioblastoma cell line (IC50 = 250 nM) with minimal cytotoxicity
  • 1-hydroxypyridin2-one Compounds2and3 (Zhengetal.,2013)
    6-substituted1-hydroxypyridin-2-oneIDH1-R132H IDH1-R132C
  • K i= 190 and 280 nM (forR132H)
  • Inhibited production of 2HG in IDH1 mutated cells

AG-120 (Agios)

  • Orally available, selective, potent inhibitor
  • PhaseI studies ongoing in advanced solid tumors (NCT02073994; NCT02074839)

Allostery as an Approach to Drugging Metabolic Enzymes Is Important in Cancer All enzymes discussed in this article are allosterically targeted by small molecule modulators. With the exception of the enzymes of lipid metabolism, it is striking that there are very few examples of the regulation of metabolic enzymes by drug-like molecules at the catalytic site. We believe that this observation will hold true for the wider set of metabolic enzymes. Metabolic pathways are typically regulated by upstream and downstream metabolites through feedforward and feedback mechanisms. This regulation occurs typically through binding at allosteric sites, which have distinctly different properties relative to active sites. Therefore regulation can come from effectors that may have very different properties to the substrate. This review describes the potential therapeutic impact of specific allosteric regulators of PKM2, glutaminase, and IDH. Additionally, preclinical studies of tool compounds demonstrated that allosteric regulators of other enzymes involved in cancer cell metabolism could provide more therapeutic opportunities (Table 4). Substrates and products of metabolic enzymes tend to be small and very polar, and often include crucial metal ions and their ligands, so it is likely that targeting their catalytic pockets will yield molecules with similar properties. From a drug-discovery point of view, targeting allosteric sites is appealing as hydrophilic substrate-binding sites are generally not hospitable to strong interactions with small molecule drugs, which gain potency to a large extent through hydrophobic interactions. In addition, as activity of most metabolic enzymes is regulated by multimerization, the formation of multimers provides opportunity for binding sites to form at protein–protein interfaces.

Table 4. Examples of Allostery in Cancer Cell Metabolism

TH           Tyrosine hydroxylase         Haloperidol                                           Activator             Catecholamine metabolism               (Casu and Gale, 1981)
PDK1      Pyruvate dehydrogenase
kinase isozyme1                  3,5-diphenylpent-2-enoicacids                         Activator             TCAcycle                                                (Stroba et al., 2009)
BCKDK  Branched chain keto acid
dehydrogenase kinase   (S)-a-chloro-phenylpropionicacid[(S)-CPP]     Inhibitor              Branch-chain amino acid                   (Tso et al., 2013)
ACACA   Acetyl-CoA carboxylase
alpha                                 5-tetradecyloxy-2-furoicacid (TOFA)                  Inhibitor              Fatty acid  synthesis                            (Wang et al.,2009)

FBP1     Fructose-1,6
bisphosphatase1               Benzoxazole benzene sulfonamide1                    Inhibitor              Glycolysis                                        (von Geldern et al., 2006)
ALADA minolevulinate
dehydratase                     wALAD in1 benzimidazoles                                     Inhibitor              Haem synthesis                                    (Lentz et al., 2014)
TYR       Tyrosinase         2,3-dithiopropanol                                                   Inhibitor              Melanin metabolism                    (Wood and Schallreuter, 1991)
DBHD  opamine beta
hydroxylase-2H-phthalazinehydrazone (hydralazine;HYD)
2-1H-pyridinonehydrazone (2-hydrazinopyridine;HP)
2-quinoline-carboxylicacid (QCA)
1H-imidazole-4-aceticacid (imidazole-4-aceticacid;IAA)                             Inhibitor         Neurotransmitter synthesis                    (Townes et al.,1990)
deaminase        5-iodo-2’-deoxyuridine5’-triphosphate                                 Inhibitor          Nucleotide metabolism                      (Prusoff and Chang, 1968)
TYMP  Thymidine
phosphorylase     5’-O-tritylinosine (KIN59)                                                    Inhibitor          Nucleotide metabolism                         (Casanova et al.,2006)
TYMS Thymidylate
synthase         1,3-propanediphosphonicacid (PDPA)                                     Inhibitor          Nucleotide   metabolism                        (Lovelace et al.,2007)

Figure 3. Simplified Description of IDH Protein Motion The large domain (residues 1–103 and 286–414) forms nearly all of the NADPH cofactor binding residues and roughly half of the substrate binding residues.The small domain(residues 104–136 and 186–285) contains the remaining substrate binding residues and the metal binding residues. The interface between the two protomers is formed by both the small domain and the clasp region (residues 137–185). The large domain moves away from the small domain to facilitate NADPH cofactor exchange and substrate binding. The large domain then closes up against the small domain, thereby completing the substrate binding pocket and bringing the cofactor, substrate, and metal into close contact with each other and with the key catalytic residues to facilitate hydride transfer between substrate and cofactor and enzyme-assisted carboxylation/decarboxylation. Subsequent opening of the large domain from the small domain would enable product release and cofactor exchange to complete the catalytic cycle (Rendina et al., 2013; Xu et al., 2004).

7.3.2 Chemical proteomics approaches to examine novel histone modifications

Xin LiXiang David Li
Current Opinion in Chemical Biology Feb 2015; 24:80–90


  • A variety of novel histone PTMs have been identified by MS-based methods.
  • Regulatory mechanisms and cellular functions of most novel histone PTMs remain unknown, due to lack of knowledge about their readers, erasers and writers.
  • Chemical proteomics approaches provide valuable tools to characterize novel histone PTMs.
  • The application of photoaffinity probes helps the profiling of histone PTMs’ readers, erasers and writers.

Histone posttranslational modifications (PTMs) play key roles in the regulation of many fundamental cellular processes, such as gene transcription, DNA damage repair and chromosome segregation. Significant progress has been made on the detection of a large variety of PTMs on histones. However, the identification of these PTMs’ regulating enzymes (i.e. ‘writers’ and ‘erasers’) and functional binding partners (i.e. ‘readers’) have been a relatively slow-paced process. As a result, cellular functions and regulatory mechanisms of many histone PTMs, particularly the newly identified ones, remain poorly understood. This review focuses on the recent progress in developing chemical proteomics approaches to profile readers, erasers and writers of histone PTMs. One of such efforts involves the development of the Cross-Linking-Assisted and SILAC-based Protein Identification (CLASPI) approach to examine PTM-mediated protein–protein interactions.

Table 1    Novel histone PTMs                      functions
1             Lysine formylation             Arising from oxidative damage of DNA modification sites overlap with lysine acetylation and methylation, potentially interfere with normal regulation of these PTMs

2      Lysine propionylation  p300,c CREB-binding protein,c Sirt1,c Sirt2,c Sirt3c
Structurally similar with lysine acetylation, regulated by same set of enzymes, H3K23pr may be regulatory for cell metabolism
3    Lysine butyrylation       p300,c CREB-binding protein,c Sirt1,c Sirt2,c Sirt3c
Structurally similar with lysine acetylation, regulated by same set of enzymes
4    Lysine malonylation    Sirt5c
Changing the positively charged lysine to negatively charged residue, likely to affect the chromatin structure
5   Lysine succinylation    Sirt5c
A  mutation to mimic crotonyl lysine that changes lysine to glutamic acid of histone H4K31, reduces cell viability
6  Lysine crotonylation   Sirt1,c Sirt2,c Sirt3
Enriched at active gene promoters potential enhancers in mammalian genomes, male germ cell differentiation
7 Lysine 2-hydroxyiso
butyrylation                     HDAC1-3c
Associated with gene transcription
8  Lysine 4-oxononoylation    Modified by 4-oxo-2-nonenal, generated under oxidative stress, prevents nucleosome assembly in vitro
9 Lysine 5-hydroxylation   JMJD6
suppress lysine acetylation and methylation
10 Glutamine methylation   Nop1  (yeast), fibrillarin (huma)
human histone H2AQ105
11 Serine and
threonine GlcNAcylation  O-GlcNAc transferase
H2BS112 GlcNAcylation promotes K120 monoubiquitination, H3S10 GlcNAcylation suppresses phosphorylation of site
12 Serine and threonine acetylation
13 Serine palmitoylation   Lpcat1
catalyzed H4S47 palmitoylation, Ca2+-dependent, regulates global RNA synthesis
14  Cysteine glutathionylation
H3.2 and H3.3
conserved cysteine, but not H3.1, destabilize the nucleosomal structure
15 Cysteine fatty-acylation
H3.2 C110
16 Tyrosine hydroxylation

Fig. 1. Schematic description of a MS-based method for the identification of novel histone PTMs.

Fig. 2. Chemical proteomics approaches to profile readers and erasers of histone PTMs.
(a) Photo-cross-linking strategy to capture proteins recognizing histone PTMs.
(b) Chemical structure of photoaffinity peptide probes.
Modifications of interest were labeled in green; photo-cross-linkers were labeled in red; chemical handles (alkyne) were labeled in blue; the sequence of probe C and probes 1–5 were derived from the
histone H3 1–15 amino acids residues, the sequence of probe 6 was derived from the histone H4 1–19 amino acids residues.
(c) Schematic for the CLASPI strategy to profile proteins that bind certain histone mark in whole-cell proteomes

Consistent with our findings, Tate and coworkers [57] recently reported the development of a photoaffinity probe based on a succinylated glutamate dehydrogenase (GDH) peptide for capturing Sirt5
as the corresponding desuccinylase. In addition to the application of photo-cross-linking strategy for examining the histone PTMs with known erasers, we recently used CLASPI with a photoaffinity
probe (probe 5, Figure 2b) to profile proteins that recognize a novel histone mark, crotonylation at histone H3K4 (H3K4cr, Table 1, Entry 6) [25], whose erasers were unknown. This study revealed,
for the first time, that Sirt3 can recognize the H3K4cr mark and efficiently catalyze the removal of histone crotonylation marks. More importantly, Sirt3 was found to regulate histone Kcr level in
cells and may potentially modulate gene transcription through its decrotonylase activity [58]. By converting bisubstrate inhibitors of HATs (histone peptides with certain lysine residues covalently
attached to Ac-CoA) to clickable photoaffinity probes (for example, probe 6, Figure 2b), they carried out the first systematic profiling of HATs in whole-cell proteomes [59].  We  anticipate  that  similar methods can be used to search for writers of novel histone PTMs such as Kmal, Ksucc, Kcr and Khib (Table 1) since the corresponding acyl-CoAs are presumed to be the acyl donors.

We have shown, in this review, the applications and recent advances of chemical tools, in combination with MS-based proteomics approaches, for the detection and characterization of histone
PTMs and their readers, erasers and writers.

This article belongs to a special issue

Omics Edited By Benjamin F Cravatt and Thomas Kodadek

Editorial overview: Omics: Methods to monitor and manipulate biological systems: recent advances in ‘omics’

Benjamin F Cravatt, Thomas Kodadek
Current Opinion in Chemical Biology Feb 2015; 24:v–vii

7.3.3 Misfolded Proteins – from Little Villains to Little Helpers… Against Cancer

Ansgar Brüning1,* and Julia Jückstock
Front Oncol. 2015; 5: 47

The application of cytostatic drugs targeting the high proliferation rates of cancer cells is currently the most commonly used treatment option in cancer chemotherapy. However, severe side effects and resistance mechanisms may occur as a result of such treatment, possibly limiting the therapeutic efficacy of these agents. In recent years, several therapeutic strategies have been developed that aim at targeting not the genomic integrity and replication machinery of cancer cells but instead their protein homeostasis. During malignant transformation, the cancer cell proteome develops vast aberrations in the expression of mutated proteins, oncoproteins, drug- and apoptosis-resistance proteins, etc. A complex network of protein quality-control mechanisms, including chaperoning by heat shock proteins (HSPs), not only is essential for maintaining the extravagant proteomic lifestyle of cancer cells but also represents an ideal cancer-specific target to be tackled. Furthermore, the high rate of protein synthesis and turnover in certain types of cancer cells can be specifically directed by interfering with the proteasomal and autophagosomal protein recycling and degradation machinery, as evidenced by the clinical application of proteasome inhibitors. Since proteins with loss of their native conformation are prone to unspecific aggregations and have proved to be detrimental to normal cellular function, specific induction of misfolded proteins by HSP inhibitors, proteasome inhibitors, hyperthermia, or inducers of endoplasmic reticulum stress represents a new method of cancer cell killing exploitable for therapeutic purposes. This review describes drugs – approved, repurposed, or under investigation – that can be used to accumulate misfolded proteins in cancer cells, and particularly focuses on the molecular aspects that lead to the cytotoxicity of misfolded proteins in cancer cells.


How Do Proteins Fold and What Makes Misfolded Proteins Dangerous?

For an understanding of misfolded proteins, it is necessary to understand how cellular proteins attain and then further maintain their native conformation and how mature proteins and unfolded proteins are generated and converted into each other.

The principles and mechanisms of protein folding were one of the major research topics and achievements of biochemical research in the last century. For decades, Anfinsen’s model, which explained protein structure by thermodynamic principles applying to the polypeptide’s inherent amino acid sequence (1), was to be found in the introductory sections of all textbooks in protein biochemistry. According to Anfinsen’s thermodynamic hypothesis, the structure with the lowest conformational Gibbs free energy was finally taken by each single polypeptide due to a thermodynamic and stereochemical selection for side chain relations that form most stable and effective enzymes or structural proteins (1). Beyond this individual selection for the energetically most optimized conformation, evolution also selected for amino acid sequences that energetically allowed the smoothest and most “frustration-free” folding processes via a thermodynamic “folding funnel” (1–3).

Whereas Anfinsen’s model preferred the side chain elements as preferential organizing structures, recent hypotheses have inversely proposed the backbone hydrogen bonds as the driving force behind protein folding (4). According to the former theory, the finally folded protein was assumed to attain a single defined structure and shape (1, 4), and the unfolded conditions were described as being represented by a structureless statistical coil with nearly indefinite conformations – a so-called “featureless energy landscape” (4). The latter model assumes that a protein selects during its folding process from a limited repertoire of stable scaffolds of backbone hydrogen bond-satisfied α-helices and β-strands (4). This also implies that unfolded proteins are not structureless, shoelace-like linear amino acid alignments as often depicted in cartoons for graphical reasons, but actually, at least in part, retain discrete and stable scaffolds.

Once the protein has attained its final conformation, the problem of stabilizing this structure arises. Hydrophobic interactions that press non-polar side chains into the center of the protein are assumed to be a major force in protein stabilization (5, 6). At the protein surface, polar interactions, mainly by hydrogen bonds of polar side chains and backbone structure, are assumed to be of similar importance (6). Salt bridges and covalent disulfide bonds were identified as further forces supporting the stability of proteins (6). Accordingly, all conditions that interfere with these stabilizing forces, including extreme temperature, salt concentrations, and redox conditions, may lead to protein misfolding.

Another aspect that must be taken into account when studying protein folding relates to the very different conditions found in viable cells when compared to test tube conditions. Considering the life-cycle of a protein, each protein begins as a growing polypeptide chain protruding from the ribosomal exit tunnel and with several of its future interacting amino acid binding partners not even yet attached to the growing chain of the nascent polymer. In these ribosomal exit tunnels, first molecular interactions and helical structures are formed, and evidence exists to support the notion that the speed of translation is regulated by slow translating codon sequences just to optimize these first folding processes (7). After leaving the ribosomal tunnel, nascent polypeptides are also directly welcomed by chaperoning protein complexes, which facilitate and further guide the folding process of newly synthesized proteins (8). It is believed that a high percentage of nascent proteins are subject to immediate degradation due to early folding errors (9). Since many nascent proteins are synthesized in parallel at polysomes, the temporal and spatial proximity of unfolded peptides brings the additional risk of protein aggregation (10). Moreover, as mentioned above, even incomplete folding intermediates and partially folded states may form energetically but not physiologically active metastable structures (11, 12). An immediate, perinatal guidance and chaperoning of newborn proteins is therefore essential to creating functional, integrative proteins and to avoiding misfolded, function-less polypeptides with potentially cytotoxic features.

Since protein structure and function are coupled, misfolded proteins are, at first, loss-of-function proteins that might reduce cell viability, in particular when generated in larger quantities. A more dangerous feature of misfolded proteins, however, lies in their strong tendency toward abnormal protein–protein interactions or aggregations, which is reflected by the involvement of misfolded proteins and their aggregates in several amyloidotic diseases, including neurodegenerative syndromes such as Alzheimer’s disease and Parkinson’s disease (13, 14). The fact that several of these intracellular and extracellular protein aggregates contain β-sheet-like structures and form filamentous structures also supports the notion that misfolded proteins are not necessarily structureless protein coils or unspecific aggregates, at least when they are formed by homogenous proteins as in the case of several neurodegenerative diseases (13). Paradoxically, these larger aggregates appear to reflect a cell protective mechanism so as to sequester or segregate smaller, but highly reactive, nucleation cores of condensing protein aggregates (13).

Unspecific hydrophobic interactions, in particular, have been held responsible for protein aggregations that form when terminally folded proteins lose their native conformation and expose buried hydrophobic side chains on their surface (15, 16). These hydrophobic interactions are also believed to be the most problematic issues with newly synthesized polypeptides on single ribosomes or polysomes (12). Once exposed to the surface, the hydrophobic structures will quickly find possible interaction partners. The intracellular milieu can be regarded as a “crowded environment” (17), fully packed with proteins in close contact and near to their solubility limit (8, 12). Thus, misfolded proteins not only aggregate among each other but may also attach to normal native proteins and inhibit their function and activity. Since such misfolding effects and interactions can also include nuclear DNA replication and repair enzymes (18), misfolded proteins may not only exert proteotoxic but also genotoxic effects, thereby endangering the entire cellular “interactome” (19) by interfering both with the integrity of the proteome (proteostasis) and the genome. Therefore, a misfolded protein is not simply a loss-of-function protein but also a promiscuous little villain that might act like a free radical, exerting uncontrolled danger to the cell.

The way in which cells deal with misfolded proteins strongly depends on the nature, strength, length, and location of the damage induced by the various insults. Management of misfolded proteins can be achieved by heat shock protein (HSP)-mediated protein renaturation (repair); proteasomal, lysosomal, or autophagosomal degradation (recycling); intracellular disposal (aggregation); or – in its last consequence if overwhelmed – by programed cell death (despair). In the following paragraphs, the cellular management of misfolded proteins is described and therapeutic options to induce misfolded proteins in cancer cells are presented.

Hsp90 and Hsp90 Inhibitors

The best-known and evolutionarily most-conserved mechanism to protect against protein misfolding is the binding and refolding process mediated by so-called heat shock proteins (HSPs). HSPs recognize unfolded or misfolded proteins and facilitate their restructuring in either an ATP-dependent (large HSPs) or energy-independent manner (low weight HSPs). HSP of 90 kDa (hsp90) is a constitutively expressed HSP and is regarded as the most common and abundantly expressed HSP in eukaryotic cells (20, 21). Although commonly referred to as hsp90, it consists of a variety of isoforms that are encoding for cytosolic (hsp90α1, α2, β), mitochondrial (TRAP1), or endoplasmic reticulum (ER)-resident (GRP94) forms. Its primary function is less that of a stress response protein and more to bind to a certain group of client proteins unable to maintain a stable configuration without being assisted by hsp90 (20, 22, 23). Steroid hormone receptors (estrogen receptor, glucocorticoid receptor), cell cycle regulatory proteins (CDK4, cyclin D, polo-like kinase), and growth factor receptors and their downstream targets (epidermal growth factor receptor 1, HER2, AKT) are among the best-studied client proteins of hsp90 (20–22). Also, several cancer-specific mutations generating otherwise instable oncoproteins, such as mutant p53 or bcr-abl, rely on hsp90 chaperoning to keep them in a soluble form, thereby facilitating the extravagant but vulnerable “malignant lifestyle” of hsp90-addicted cancer cells (21, 24). Accordingly, hsp90 has been assumed to be a prominent target, in particular for hormone-responsive and growth factor receptor amplification-dependent cancer types.

The microbial antibiotics geldanamycin and radicicol are the prototypes of hsp90 inhibitors. Based on intolerable toxicity, these molecules had to be chemically modified for application in humans, and most of the ongoing clinical studies with hsp90 inhibitors are aimed at identifying semi-synthetic derivatives of these lead compounds with an acceptable risk profile. Unfortunately, most recent studies using geldanamycin derivatives have provided disappointing results because of toxicities and insufficient efficacy (22, 25–27). Studies with radicicol (resorcinol) derivatives, in particular with ganetespib, appear to be more promising because of fewer adverse effects (22, 25–27). Liver and ocular (retinal) toxicities have been described as main adverse effects of hsp90 inhibition, and appeared to be experienced less with ganetespib than with most of the first generation hsp90 inhibitors (28).

Since both geldanamycin and radicicol target the highly conserved and unique ATP-binding domain of hsp90, new synthetic inhibitors have also been generated by rational drug design (22, 25–27). However, none of the various natural or synthetic hsp90 inhibitors under investigation have yet provided convincing clinical data, and future studies will show whether hsp90 can eventually be added to the list of effective cancer targets.

Hsp70, Hsp40, Hsp27, and HSF1

Hsp90 is assisted by several other HSPs and non-chaperoning co-factors, finally forming a large protein complex that recruits and releases client proteins in an energy-dependent manner (21, 22, 29). Client proteins for hsp90 are first bound to hsp70, which transfers the prospective client to hsp90 through the mediating help of an hsp70–hsp90 organizing protein (HOP). Binding of potential hsp90 client proteins to hsp70 is facilitated by its co-chaperone hsp40 (23, 30). Exposed hydrophobic amino acids, the typical feature of misfolded proteins, have been described as the main recognition signal for hsp70 proteins (15, 16, 31). Hsp70 proteins are not only supporter proteins for hsp90 but also represent a large chaperone family capable of acting independently of hsp90 and that can be found in all cellular compartments, including cytosol and nucleus (hsp70, hsp72, hsc70), mitochondria (GRP75 = mortalin), and the ER (GRP78 = BiP). Hsp70 chaperones may act on misfolded or nascent proteins either as “holders” or “folders” (31), which means that they prevent protein aggregation either by sheltering these aggregation-prone protein intermediates or by allowing these proteins to fold/refold into their native form in an assisted mechanism within a protected environment (31). Hsc70 (HSPA8) is a constitutively expressed major hsp70 isoform that is an essential factor for normal protein homeostasis even in unstressed cells (16). Misfolded proteins can also be destined by hsp70 proteins for their ultimate degradation. Proteins that expose KFERQ amino acid motifs on their surface during their unfolding process are preferentially bound by hsc70 and can be directed to lysosomes in a process called chaperone-mediated autophagy (CMA) (32, 33). In another mechanism of targeted protein degradation, interaction of hsc70 with the E3 ubiquitin ligase CHIP (carboxyl terminus of Hsc70-interacting protein) leads to ubiquitination of misfolded proteins and thus their destination of the ubiquitin-proteasome protein degradation pathway (34, 35). Since hsc70 is essential for normal protein homeostasis and its knock-out is lethal in mice (16, 36), hsc70 inhibition might not be an optimal target for cancer-specific induction of misfolded proteins. This contrasts with the inducible forms of hsp70 such as hsp72 (HSPA1), which are upregulated in a cell stress-specific manner and are often found to be constitutively overexpressed in cancer tissues (16, 36). Transcriptional activation of these inducible HSPs is mediated by the heat shock factor 1 (HSF1), which also regulates expression of hsp40 and the small HSP hsp27 by sharing a common promoter consensus sequence (heat shock response element) for HSF1 binding (37). HSF1 was also found to be constitutively activated in cancer tissues, modulating several cell cycle- and apoptosis-related pathways via its target genes (38–40). HSF1 itself is kept inactive in the cytosol by binding to hsp90, and the recruitment of hsp90 to misfolded proteins is considered a main activation mechanism to release monomeric HSF1 for its subsequent trimerization, post-translational activation, and nuclear translocation (24, 41). Also, since hsp90 inhibition causes hsp70 induction by HSF1 activation as a compensatory feed-back mechanism (24), combined inhibition of hsp90 and hsp70, or of hsp90 and HSF1 might be a more effective therapeutic approach for cancer treatment than single HSP targeting alone.

Indeed, several small-molecule inhibitors and aptamers for hsp70, hsp40, and hsp27 have been designed (16, 42–44), but most of them remain in pre-clinical development, or are either not applicable in humans or associated with intolerable side effects (16, 42–44). Notably, the natural bioflavonoid quercetin was shown to inhibit phosphorylation and transcriptional activity of the heat shock transcription factor HSF1, thus reducing HSP expression at its most basal level (45–48). This HSP and HSF1 inhibition may also contribute to the observed cancer-preventing effects of a flavonoid-rich diet, which includes fruits and vegetables. However, due to their low bioavailability, the concentrations of flavonoids needed to induce direct cytotoxic effects in cancer cells for (chemo-)therapeutic reasons are obviously not achievable in humans, even when applied as nutritional supplements (49). More effective and clinically more easily applicable inhibitors of HSF1 are therefore urgently sought. Promising HSF1 targeting strategies are currently under development, although are apparently not yet suited for clinical applications (24, 50, 51).

SP Williams Comment:

There is a new hsp90- inhibitor, ganetespib, which is active against ovarian cancer in vitro and in vivo. Clinical trials are looking at this in cisplatin refractory cases. This was identified by a network analysis from a previous siRNA screen on ovarian cancer cells for pathways related to growth inhibition in an effort to find possible targets against CP resistance. The reference is

Protein Ubiquitination and Proteasomal Degradation

Ubiquitin is a 76 amino acid polypeptide that can covalently be attached via its carboxy-terminus to free (lysyl) amino groups of proteins. Ubiquitination of proteins generates a cellular recognition motif that is involved in various functions ranging from transcription factor and protein kinase activation to DNA repair and protein degradation – depending on the extent and exact location of this post-translational modification (52, 53). Monoubiquitination of peptides of more than 20 amino acids was found to be a minimal requirement for protein degradation, but the canonical fourfold (poly-)ubiquitination with three further lysine (K48) side chain-linked ubiquitins appears to be most apt for an effective and rapid substrate recognition by the proteasome (54). This canonical polyubiquitin structure, as well as several other mixed polyubiquitin structures, can be recognized by the external 19S subunits of the 26S proteasome complex (54, 55). Prior to degradation of ubiquitinated proteins by the proteasomal 20S core subunit, the attached ubiquitin chains are released by the external 19S subunits for recycling, although they can also be co-degraded by the proteasome (56). After first passing the 19S subunit, the proteasomal target proteins are then unfolded in an energy-dependent manner and introduced into the narrow enzymatic cavity of proteasome for degradation. The barrel-shaped 20S proteasomal core complex contains three different proteolytic activities in duplicate (β1: caspase-like-, β2: tryptic-, and β5: chymotryptic activity), which initiate an efficient cleavage of the proteasomal target proteins into smaller peptides (57).

It is important to note that specific ubiquitination and ensuing proteasomal degradation is not an exclusive degradation mechanism of misfolded proteins but is also used to regulate the expression level of several native cell cycle regulatory proteins [cyclins, proliferating cell nuclear antigen (PCNA), p53], signaling pathway molecules (β-catenin, IκB), and survival factors (mcl-1) during the course of normal protein homeostasis and cell cycle progression (53, 55, 57, 58). Moreover, proteasomes are involved in protein maturation, including the processing and maturation of the NF-κB transcription factor subunit p50 and the drug-resistant protein MDR1 (57). Therefore, targeting proteasomal activity has not only been of interest for the generation of misfolded, cytotoxic proteins but also for interfering with the expression of proteins involved in several hallmarks of cancer, including cell cycle progression, signal transduction, and apoptosis.

Proteasome Inhibitors

Bortezomib (PS-341, Velcade ™) has long been known as a paragon of a clinically applicable proteasome inhibitor. Bortezomib has been approved for the treatment of multiple myeloma and mantle cell lymphoma (55, 59, 60). The great expectations of transferring the success of bortezomib to non-hematological solid cancer types have unfortunately not yet been fulfilled. It has been suggested that the high antibody-producing capacity of myeloma cells and thus the need for an efficient proteasomal degradation system to cope with the recycling process of misfolded ER-generated antibodies [ER-associated degradation process (ERAD); see below] might contribute to the high sensitivity of myeloma cells to bortezomib (9, 60, 61). Originally, bortezomib was developed to inhibit the proteasomal degradation of the NF-κB inhibitor IκB, thus targeting the pro-inflammatory, but also cancer-promoting, effect of the NF-κB transcription factor (55, 60, 62). Recent insights indicate that the anti-tumoral effect of bortezomib is not only mediated by its NF-κB inhibitory activity but also by its ability to induce accumulation of misfolded proteins in the cytosol and the ER (60, 62–65). However, the use of bortezomib, even for highly sensitive multiple myeloma, is limited by its strong tendency to induce a proteasome inhibition-independent peripheral neuropathy by acting on neuronal mitochondria (61). Since neurodegenerative diseases are associated with protein misfolding and aggregation, the neuropathological effects of bortezomib might also be assumed to be mediated by the possible proteotoxic effects of bortezomib in neuronal cells. However, although proteasome inhibitor-induced neurodegeneration and inclusion body formation have been described in animal models, similarities between proteasome inhibitor-induced neurodegeneration and Parkinson’s disease-like histopathological features could not be established (66).

Table 1 Drugs described in this review and their mechanism of action (MOA), status of approval, and main adverse effects.

Aggresome Formation and Re-Solubilization: Role of HDAC6

As depicted above, proteasome and HSP inhibition will eventually lead to the accumulation of misfolded and polyubiquitinated proteins. Based on their inherent cohesive properties mediated by their exposed hydrophobic surfaces, both ubiquitinated and non-ubiquitinated misfolded proteins tend to adhere as small aggregates (Figure ​(Figure1).1). Individual ubiquitinated proteins and small ubiquitinated aggregates can be recognized by specific ubiquitin-binding proteins such as HDAC6 via its zinc finger ubiquitin-binding domain. HDAC6 is an unusual histone deacetylase located in the cytosol that regulates microtubule acetylation and is also able to bind ubiquitinated proteins. Based on HDAC6’s additional ability to bind to microtubule motor protein dynein, these aggregates are actively transported along the microtubular system into perinuclear aggregates around the microtubule organizing center (MTOC) (108384). Recognition of small, scattered ubiquitinated aggregates by HDAC6 has been described as being mediated by unanchored ubiquitin chains, which are generated by aggregate-attached ubiquitin ligase ataxin-3 (85). Whereas proteasomal target proteins are primarily tagged by K-48 (lysine-48) linked ubiquitins; K-63 linked ubiquitin chains appear to be a preferential modification for aggresomal targeting by HDAC6 and were assumed to mediate a redirection from proteasomal degradation to aggresome formation in the case of proteasomal inhibition or overload (86). Accordingly, aggresome formation is not an unspecific protein aggregation but a specific, ubiquitin-controlled sorting process. Furthermore, these aggresomes consist not only of misfolded and deposited proteins but have also been shown to contain a large amount of associated HSPs and ubiquitin-binding proteins, including HDAC6 [Figure ​[Figure1;1; (108384)]. Aggresomes contain, and are also surrounded by, large numbers of proteasomes (108384), which help to resolubilize these aggregates not only through their intrinsic proteasomal digestion but also by generating unanchored K63-branched polyubiquitin chains, which then stimulate HDAC6-mediated autophagy, another cellular disposal mechanism in involving HDAC6 (87). Notably, HDAC6 has also been shown to control further maturation of autophagic vesicles by stimulating autophagosome–lysosome fusion (Figure ​(Figure1)1) in a manner different from the normal autophagosome–lysosome fusion process (88).

Figure 1

Drugs that inhibit folding or disposal of misfolded proteins. Native mature proteins, nascent proteins, or misfolded proteins can be prevented from folding or refolding by small and large heat shock protein inhibitors, of which the hsp90 inhibitors based 

The HDAC6 multitalent also exerts its deacetylase activity on hsp90 and modifies hsp90 client binding by facilitating its chaperoning of steroid hormone receptors and HSF1 (8991). Recruitment of HDAC6 to ubiquitinated proteins leads to the dissociation of the repressive HDAC6/hsp90/HSF1 complex (91) and allows the release of transcriptionally active HSF1 to the nucleus. The engagement of HDAC6 at the aggresome–autophagy pathway hence also indirectly facilitates HSF1 activity. p97/VCP (valosin-containing protein), another binding partner of HDAC6 and itself a multi-interactive, ATP-dependent chaperone (9294), is assumed to be involved not only in the specific separation of hsp90 and HSF1 by its “segregase” activity but also in the binding and remodeling of polyubiquitinated proteins before their delivery to the proteasome (9395). Additionally, p97/VCP dissociates polyubiquitinated proteins bound to HDAC6 (91). Accumulation of polyubiquitinated proteins thus leads to HDAC6-dependent HSF1 activation and HSP induction, p97/VCP-dependent recruitment and “preparation” of polyubiquitinated proteins to proteasomes, and, in the case of pharmacological proteasome inhibition or physiological overload, to an HDAC6-dependent detoxification of polyubiquitinated proteins by the aggresome/autophagy pathway.

Pharmacological Inhibition of Aggresome Formation: HDAC6 Inhibitors

The central involvement of HDAC6 in aggresome formation and clearance makes HDAC6 one of the most interesting druggable targets for the induction of proteotoxicity in cancer cells. Also, HDAC6 has been found to be overexpressed in various cancer tissues, associated with advanced cancer stages and increased neoplastic transformation (96). Several pan-histone deacetylase inhibitors have been developed and tested in clinical studies for a variety of diseases, including different types of cancer (9798). Although hematological malignancies responded best to most of the already clinically tested pan-histone deacetylase inhibitors, the efficacy on solid cancer types was disappointingly poor and also associated with intolerable side effects (98). The unforeseeable pleiotropic epigenetic mechanism caused by non-specific (nuclear) histone deacetylase inhibitors may also limit their application for use in cancer treatment or HDAC6 inhibition, and has led to the search for selective HDAC6 inhibitors with no inhibitory effects on transcription modifying histone deacetylases. Through screening of small molecules under the rationale of selecting for tubulin deacetylase inhibitors with no cross-reactive histone deacetylase activity, the HDAC6 inhibitor tubacin was identified, and suggested for use in the treatment of neurodegenerative diseases or to reduce cancer cell migration and angiogenesis (99). Hideshima et al. then proved the hypothesis that the combined use of bortezomib with tubacin leads to an accumulation of non-disposed cytotoxic proteins and aggregates in cancer cells (100). Indeed, a synergistic effect of these two drugs against multiple myeloma cells could be observed with no detectable toxic effect on peripheral blood mononuclear cells (100). This and follow-up studies also revealed the efficacy of tubacin as a single agent against leukemia cells (100101) and a chemo-sensitizing effect on cytotoxic drugs in breast- and prostate-cancer cells (102).

Endoplasmic Reticulum Stress

Besides the cytosol, the ER is a major site for protein synthesis, in particular for those proteins destined for extracellular secretion, the cell membrane, or their retention within the endomembrane system. At the rough ER, nascent proteins are co-translationally transported across the ER membrane into the ER lumen (107), where they immediately encounter ER-resident chaperones, most prominently represented by hsp70 family member BiP/GRP78 and hsp90 family member GRP94 to help proper protein folding (15108). Most of these proteins also undergo post-translational modifications, including N- or O-linked glycosylation or protein disulfide bridge-building (109110), thereby adding further mechanisms of protein stabilization but also challenges for proper protein folding.

Similar to the situation in cytosolic protein biosynthesis, a large proportion of nascent proteins in the ER are assumed to misfold and to go “off-pathway” even under normal physiological conditions. Furthermore, the ER lumen, narrowly sandwiched between two phospholipid membranes, has been described as an even more densely crowded environment than the cytosol, additionally facilitating unspecific protein attachments and aggregations (15). Since, with the exception of bulk reticulophagy, the lumen of the ER contains no endogenous protein degradation system, and the anterograde transport of ER proteins to the Golgi, lysosomes, endosomes, or the extracellular environment requires properly folded proteins, a retrograde transport of ER proteins into the cytosol remains the only possible mechanism of preventing misfolded protein accumulation within the ER. In this ERAD, misfolded proteins are re-exported across the ER membrane by a specific multi protein complex, ubiquitinated by ER membrane-integrated ubiquitin ligases, and finally become degraded by cytosolic proteasomes (111112). Notably, association of the cytosolic p97/VCP protein, an important interacting partner with HDAC6, has also been described as being an essential factor for driving the luminal proteins through the ER membrane pore complex into the cytosol (92,112).

Accordingly, all agents and conditions that interfere with these folding, maturation, and retranslocation processes can lead to protein misfolding and aggregation within this sensitive organelle. Chemicals that act as glycosylation inhibitors (tunicamycin), calcium ionophore inhibitors (A23187, thapsigargin), heavy metal ions (cadmium, lead), reducing agents (dithiothreitol), as well as conditions like hypoxia or oxidative stress, all lead to a phenomenon called ER stress (113116). In the ER-stress response, a triad of ER membrane-resident signaling receptors and transducers, IRE1, ATF6, and PERK1, become activated and lead to the transcriptional activation of cytosolic and ER-resident chaperones to cope with the increasing number of misfolded proteins. Induction of autophagy (reticulophagy; ER-phagy) may also occur and supports the removal of damaged regions of the ER (117). Under very intensive or even unmanageable ER-stress conditions, a variety of pro-apoptotic pathways ensue, including CHOP induction, c-JUN-kinase activation, and caspase cleavage (118120), which eventually prevails over the cytoprotective arm of the ER-stress response and may lead to apoptosis. Targeting of protein folding within the ER is therefore a very promising strategy to induce apoptosis in cancer cells, in particular in those cancer cells characterized by an unphysiologically high protein secretion rate, such as, for example, multiple myeloma cells. Whereas the above-mentioned drugs such as tunicamycin or thapsigargin are valuable tools for cell biology studies, they display unacceptable toxicities in humans and are not suited for therapeutic applications. Interestingly, several already established drugs used for non-cancerous diseases have been described as inducing ER stress at pharmacologically relevant concentrations in humans as an off-target effect (113116). The non-steroidal anti-inflammatory COX-2 inhibitor celecoxib is an approved drug to treat various forms of arthritis and pain, but has also been described as exerting ER stress by functioning as a SERCA (sarco/ER Ca2+ ATPase) inhibitor (113116). However, although well tolerated in humans, the ER-stress-inducing ability of celecoxib seems to be weaker than that of direct SERCA inhibitors such as thapsigargin, and the usefulness of celecoxib against advanced cancer has been questioned (116). Various HIV protease inhibitors have been described as inducing ER stress in human tissue cells as a side effect (121123). In particular the HIV drugs lopinavir, saquinavir, and nelfinavir appear to be potent inducers of the ER-stress reaction, leading to a focused interest in these drugs for the induction of ER stress and apoptosis in cancer cells (116124128). In fact, with currently over 27 clinical studies in cancer patients2, nelfinavir, either used as a single agent or in combination therapy, is on the list of the most promising prospective candidates to induce selective proteotoxicity in cancer cells at pharmacologically relevant concentrations. Although the exact mechanism by which nelfinavir induces ER stress is not yet clear, it was shown that nelfinavir causes the upregulation of cytosolic and ER-resident HSPs, and induces apoptosis in cancer cells associated with caspase activation and induction of the pro-apoptotic transcription factor CHOP (125126). Nelfinavir was also shown to be combinable with bortezomib to enhance its activity on cancer cells (129). Since the retrograde transport of misfolded ER proteins is inhibited by the p97/VCP inhibitor eeyarestatin (130131), we recently tested the combination of eeyarestatin with nelfinavir but found no synergistic effect between these two agents in cervical cancer cells (132). In contrast, eeyarestatin markedly sensitized cervical cancer cells to bortezomib treatment (132), which was also observed in preceding studies in which eeyarestatin was used to augment the ER-stress-inducing ability of bortezomib in leukemia cells (131).

Induction of proteotoxicity through the accumulation of misfolded proteins has evolved as a new treatment modality in the fight against cancer. Clinically approved drugs such as bortezomib and carfilzomib provide evidence of the functionality of this approach. Newly developed agents like the HDAC6 inhibitor ACY-1215 or repurposed drugs like nelfinavir or disulfiram are currently being tested in clinical trials with cancer patients and will hopefully further broaden our arsenal of anti-cancer drugs. Notably, most proteotoxic agents that have been approved or are in clinical trials target the ubiquitin-proteasome-system (UPS) and are mainly effective in multiple myeloma cells, which rely on a functional ER/ERAD/UPS for excessive and proper antibody production. Similarly, it can be assumed that other cancer cell types with a marked secretory phenotype may also be affected by ER/ERAD/UPS inhibitors. In accordance with this notion, a recent dose-escalating Phase Ia study with nelfinavir as a single agent, that covered a large variety of solid cancer entities, revealed response rates primarily in patients with neuroendocrine tumors (140). In most other solid cancer types, however, the chemo-sensitizing or combination effects of proteotoxic drugs may prevail, and have become the focus of an increasing number of very promising clinical and pre-clinical studies.

7.3.4 Endoplasmic reticulum protein 29 (ERp29) in epithelial cancer

Friend or Foe: Endoplasmic reticulum protein 29 (ERp29) in epithelial cancer

Chen S1Zhang D2

FEBS Open Bio. 2015 Jan 30; 5:91-8

The endoplasmic reticulum (ER) protein 29 (ERp29) is a molecular chaperone that plays a critical role in protein secretion from the ER in eukaryotic cells. Recent studies have also shown that ERp29 plays a role in cancer. It has been demonstrated that ERp29 is inversely associated with primary tumor development and functions as a tumor suppressor by inducing cell growth arrest in breast cancer. However, ERp29 has also been reported to promote epithelial cell morphogenesis, cell survival against genotoxic stress and distant metastasis. In this review, we summarize the current understanding on the biological and pathological functions of ERp29 in cancer and discuss the pivotal aspects of ERp29 as “friend or foe” in epithelial cancer.

The endoplasmic reticulum (ER) is found in all eukaryotic cells and is complex membrane system constituting of an extensively interlinked network of membranous tubules, sacs and cisternae. It is the main subcellular organelle that transports different molecules to their subcellular destinations or to the cell surface [10,85].

The ER contains a number of molecular chaperones involved in protein synthesis and maturation. Of the ER chaperones, protein disulfide isomerase (PDI)-like proteins are characterized by the presence of a thioredoxin domain and function as oxido-reductases, isomerases and chaperones [33]. ERp29 lacks the active-site double-cysteine (CxxC) motif and does not belong to the redox-active PDIs [5,47]. ERp29 is recognized as a characterized resident of the cellular ER, and it is expressed ubiquitously and abundantly in mammalian tissues [50]. Protein structural analysis showed that ERp29 consists of N-terminal and C-terminal domains [5]: N-terminal domain involves dimerization whereas the C-terminal domain is essential for substrate binding and secretion [78]. The biological function of ERp29 in protein secretion has been well established in cells [8,63,67].

ERp29 is proposed to be involved in the unfolded protein response (UPR) as a factor facilitating transport of synthesized secretory proteins from the ER to Golgi [83]. The expression of ERp29 was demonstrated to be increased in cells exposed to radiation [108], sperm cells undergoing maturation [42,107], and in certain cell types both under the pharmacologically induced UPR and under the physiological conditions (e.g., lactation, differentiation of thyroid cells) [66,82]. Under ER stress, ERp29 translocates the precursor protein p90ATF6 from the ER to Golgi where it is cleaved to be a mature and active form p50ATF by protease (S1P and S2P) [48]. In most cases, ERp29 interacts with BiP/GRP78 to exert its function under ER stress [65].

ERp29 is considered to be a key player in both viral unfolding and secretion [63,67,77,78] Recent studies have also demonstrated that ERp29 is involved in intercellular communication by stabilizing the monomeric gap junction protein connexin43 [27] and trafficking of cystic fibrosis transmembrane conductance regulator to the plasma membrane in cystic fibrosis and non-cystic fibrosis epithelial cells [90]. It was recently reported that ERp29 directs epithelial Na(+) channel (ENaC) toward the Golgi, where it undergoes cleavage during its biogenesis and trafficking to the apical membrane [40]. ERp29 expression protects axotomized neurons from apoptosis and promotes neuronal regeneration [111]. These studies indicate a broad biological function of ERp29 in cells.

Recent studies demonstrated a tumor suppressive function of ERp29 in cancer. It was found that ERp29 expression inhibited tumor formation in mice [4,87] and the level of ERp29 in primary tumors is inversely associated with tumor development in breast, lung and gallbladder cancer [4,29].

However, its expression is also responsible for cancer cell survival against genotoxic stress induced by doxorubicin and radiation [34,76,109]. The most recent studies demonstrate other important roles of ERp29 in cancer cells such as the induction of mesenchymal–epithelial transition (MET) and epithelial morphogenesis [3,4]. MET is considered as an important process of transdifferentiation and restoration of epithelial phenotype during distant metastasis [23,52]. These findings implicate ERp29 in promoting the survival of cancer cells and also metastasis. Hence, the current review focuses on the novel functions of ERp29 and discusses its pathological importance as a “friend or foe” in epithelial cancer.

2. ERp29 regulates mesenchymal–epithelial transition

2.1. Epithelial–mesenchymal transition (EMT) and MET

The EMT is an essential process during embryogenesis [6] and tumor development [43,96]. The pathological conditions such as inflammation, organ fibrosis and cancer progression facilitate EMT [16]. The epithelial cells after undergoing EMT show typical features characterized as: (1) loss of adherens junctions (AJs) and tight junctions (TJs) and apical–basal polarity; (2) cytoskeletal reorganization and distribution; and (3) gain of aggressive phenotype of migration and invasion [98]. Therefore, EMT has been considered to be an important process in cancer progression and its pathological activation during tumor development induces primary tumor cells to metastasize [95]. However, recent studies showed that the EMT status was not unanimously correlated with poorer survival in cancer patients examined [92].

In addition to EMT in epithelial cells, mesenchymal-like cells have capability to regain a fully differentiated epithelial phenotype via the MET [6,35]. The key feature of MET is defined as a process of transdifferentiation of mesenchymal-like cells to polarized epithelial-like cells [23,52] and mediates the establishment of distant metastatic tumors at secondary sites [22]. Recent studies demonstrated that distant metastases in breast cancer expressed an equal or stronger E-cadherin signal than the respective primary tumors and the re-expression of E-cadherin was independent of the E-cadherin status of the primary tumors [58]. Similarly, it was found that E-cadherin is re-expressed in bone metastasis or distant metastatic tumors arising from E-cadherin-negative poorly differentiated primary breast carcinoma [81], or from E-cadherin-low primary tumors [25]. In prostate and bladder cancer cells, the nonmetastatic mesenchymal-like cells were interacted with metastatic epithelial-like cells to accelerate their metastatic colonization [20]. It is, therefore, suggested that the EMT/MET work co-operatively in driving metastasis.

2.2. Molecular regulation of EMT/MET

E-cadherin is considered to be a key molecule that provides the physical structure for both cell–cell attachment and recruitment of signaling complexes [75]. Loss of E-cadherin is a hallmark of EMT [53]. Therefore, characterizing transcriptional regulators of E-cadherin expression during EMT/MET has provided important insights into the molecular mechanisms underlying the loss of cell–cell adhesion and the acquisition of migratory properties during carcinoma progression [73].

Several known signaling pathways, such as those involving transforming growth factor-β (TGF-β), Notch, fibroblast growth factor and Wnt signaling pathways, have been shown to trigger epithelial dedifferentiation and EMT [28,97,110]. These signals repress transcription of epithelial genes, such as those encoding E-cadherin and cytokeratins, or activate transcription programs that facilitate fibroblast-like motility and invasion [73,97].

The involvement of microRNAs (miRNAs) in controlling EMT has been emphasized [11,12,18]. MiRNAs are small non-coding RNAs (∼23 nt) that silence gene expression by pairing to the 3′UTR of target mRNAs to cause their posttranscriptional repression [7]. MiRNAs can be characterized as “mesenchymal miRNA” and “epithelial miRNA” [68]. The “mesenchymal miRNA” plays an oncogenic role by promoting EMT in cancer cells. For instance, the well-known miR-21, miR-103/107 are EMT inducer by repressing Dicer and PTEN [44].

The miR-200 family has been shown to be major “epithelial miRNA” that regulate MET through silencing the EMT-transcriptional inducers ZEB1 and ZEB2 [13,17]. MiRNAs from this family are considered to be predisposing factors for cancer cell metastasis. For instance, the elevated levels of the epithelial miR-200 family in primary breast tumors associate with poorer outcomes and metastasis [57]. These findings support a potential role of “epithelial miRNAs” in MET to promote metastatic colonization [15].

2.3. ERp29 promotes MET in breast cancer

The role of ERp29 in regulating MET has been established in basal-like MDA-MB-231 breast cancer cells. It is known that myosin light chain (MLC) phosphorylation initiates to myosin-driven contraction, leading to reorganization of the actin cytoskeleton and formation of stress fibers [55,56]. ERp29 expression in this type of cells markedly reduced the level of phosphorylated MLC [3]. These results indicate that ERp29 regulates cortical actin formation through a mechanism involved in MLC phosphorylation (Fig. 1). In addition to the phenotypic change, ERp29 expression leads to: expression and membranous localization of epithelial cell marker E-cadherin; expression of epithelial differentiation marker cytokeratin 19; and loss of the mesenchymal cell marker vimentin and fibronectin [3] (Fig. 1). In contrast, knockdown of ERp29 in epithelial MCF-7 cells promotes acquisition of EMT traits including fibroblast-like phenotype, enhanced cell spreading, decreased expression of E-cadherin and increased expression of vimentin [3,4]. These findings further substantiate a role of ERp29 in modulating MET in breast cancer cells.

Fig. 1  ERp29 triggers mesenchymal–epithelial transition. Exogenous expression of ERp29 in mesenchymal MDA-MB-231 breast cancer cells inhibits stress fiber formation by suppressing MLC phosphorylation. In addition, the overexpressed ERp29 decreases the

2.4. ERp29 targets E-cadherin transcription repressors

The transcription repressors such as Snai1, Slug, ZEB1/2 and Twist have been considered to be the main regulators for E-cadherin expression [19,26,32]. Mechanistic studies revealed that ERp29 expression significantly down-regulated transcription of these repressors, leading to their reduced nuclear expression in MDA-MB-231 cells [3,4] (Fig. 2). Consistent with this, the extracellular signal-regulated kinase (ERK) pathway which is an important up-stream regulator of Slug and Ets1 was highly inhibited [4]. Apparently, ERp29 up-regulates the expressions of E-cadherin transcription repressors through repressing ERK pathway. Interestingly, ERp29 over-expression in basal-like BT549 cells resulted in incomplete MET and did not significantly affect the mRNA or protein expression of Snai1, ZEB2 and Twist, but increased the protein expression of Slug [3]. The differential regulation of these transcriptional repressors of E-cadherin by ERp29 in these two cell-types may occur in a cell-context-dependent manner.

Fig. 2  ERp29 decreases the expression of EMT inducers to promote MET. Exogenous expression of ERp29 in mesenchymal MDA-MB-231 breast cancer cells suppresses transcription and protein expression of E-cadherin transcription repressors (e.g., ZEB2, SNAI1 and Twist), ..

2.5. ERp29 antagonizes Wnt/ β-catenin signaling

Wnt proteins are a family of highly conserved secreted cysteine-rich glycoproteins. The Wnt pathway is activated via a binding of a family member to a frizzled receptor (Fzd) and the LDL-Receptor-related protein co-receptor (LRP5/6). There are three different cascades that are activated by Wnt proteins: namely canonical/β-catenin-dependent pathway and two non-canonical/β-catenin-independent pathways that include Wnt/Ca2+ and planar cell polarity [84]. Of note, the Wnt/β-catenin pathway has been extensively studied, due to its important role in cancer initiation and progression [79]. The presence of Wnt promotes formation of a Wnt–Fzd–LRP complex, recruitment of the cytoplasmic protein Disheveled (Dvl) to Fzd and the LRP phosphorylation-dependent recruitment of Axin to the membrane, thereby leading to release of β-catenin from membrane and accumulation in cytoplasm and nuclei. Nuclear β-catenin replaces TLE/Groucho co-repressors and recruits co-activators to activate expression of Wnt target genes. The most important genes regulated are those related to proliferation, such as Cyclin D1 and c-Myc [46,94], which are over-expressed in most β-catenin-dependent tumors. When β-catenin is absent in nucleus, the transcription factors T-cell factor/lymphoid enhancer factors (TCF/LEF) recruits co-repressors of the TLE/Groucho family and function as transcriptional repressors.

β-catenin is highly expressed in the nucleus of mesenchymal MDA-MB-231 cells. ERp29 over-expression in this type of cells led to translocation of nuclear β-catenin to membrane where it forms complex with E-cadherin [3] (Fig. 3). This causes a disruption of β-catenin/TCF/LEF complex and abolishes its transcription activity. Indeed, ERp29 significantly decreased the expression of cyclin D1/D2 [36], one of the downstream targets of activated Wnt/β-catenin signaling [94], indicating an inhibitory effect of ERp29 on this pathway. Meanwhile, expression of ERp29 in this cell type increased the nuclear expression of TCF3, a transcription factor regulating cancer cell differentiation while inhibiting self-renewal of cancer stem cells [102,106]. Hence, ERp29 may play dual functions in mesenchymal MDA-MB-231 breast cancer cells by: (1) suppressing activated Wnt/β-catenin signaling via β-catenin translocation; and (2) promoting cell differentiation via activating TCF3 (Fig. 3). Because β-catenin serves as a signaling hub for the Wnt pathway, it is particularly important to focus on β-catenin as the target of choice in Wnt-driven cancers. Though the mechanism by which ERp29 expression promotes the disassociation of β-catenin/TCF/LEF complex in MDA-MB-231 cells remains elusive, activating ERp29 expression may exert an inhibitory effect on the poorly differentiated, Wnt-driven tumors.

Fig. 3  ERp29 over-expression “turns-off” activated Wnt/β-catenin signaling. In mesenchymal MDA-MB-231 cells, high expression of nuclear β-catenin activates its downstream signaling involved in cell cycles and cancer stem cell

3. ERp29 regulates epithelial cell integrity

3.1. Cell adherens and tight junctions

Adherens junctions (AJs) and tight junctions (TJs) are composed of transmembrane proteins that adhere to similar proteins in the adjacent cell [69]. The transmembrane region of the TJs is composed mainly of claudins, tetraspan proteins with two extracellular loops [1]. AJs are mediated by Ca2+-dependent homophilic interactions of cadherins [71] which interact with cytoplasmic catenins that link the cadherin/catenin complex to the actin cytoskeleton [74].

The cytoplasmic domain of claudins in TJs interacts with occludin and several zona occludens proteins (ZO1-3) to form the plaque that associates with the cytoskeleton [99]. The AJs form and maintain intercellular adhesion, whereas the TJs serve as a diffusion barrier for solutes and define the boundary between apical and basolateral membrane domains [21]. The AJs and TJs are required for integrity of the epithelial phenotype, as well as for epithelial cells to function as a tissue [75].

The TJs are closely linked to the proper polarization of cells for the establishment of epithelial architecture[86]. During cancer development, epithelial cells lose the capability to form TJs and correct apico–basal polarity [59]. This subsequently causes the loss of contact inhibition of cell growth [91]. In addition, reduction of ZO-1 and occludin were found to be correlated with poorly defined differentiation, higher metastatic frequency and lower survival rates [49,64]. Hence, TJs proteins have a tumor suppressive function in cancer formation and progression.

3.2. Apical–basal cell polarity

The apical–basal polarity of epithelial cells in an epithelium is characterized by the presence of two specialized plasma membrane domains: namely, the apical surface and basolateral surface [30]. In general, the epithelial cell polarity is determined by three core complexes. These protein complexes include: (1) the partitioning-defective (PAR) complex; (2) the Crumbs (CRB) complex; and (3) the Scribble complex[2,30,45,51]. PAR complex is composed of two scaffold proteins (PAR6 and PAR3) and an atypical protein kinase C (aPKC) and is localized to the apical junction domain for the assembly of TJs [31,39]. The Crumbs complex is formed by the transmembrane protein Crumbs and the cytoplasmic scaffolding proteins such as the homologue of Drosophila Stardust (Pals1) and Pals-associated tight junction protein (Patj) and localizes to the apical [38]. The Scribble complex is comprised of three proteins, Scribble, Disc large (Dlg) and Lethal giant larvae (Lgl) and is localized in the basolateral domain of epithelial cells [100].

Fig. 4  ERp29 regulates epithelial cell morphogenesis. Over-expression of ERp29 in breast cancer cells induces the transition from a mesenchymal-like to epithelial-like phenotype and the restoration of tight junctions and cell polarity. Up-regulation and membrane

The current data from breast cancer cells supports the idea that ERp29 can function as a tumor suppressive protein, in terms of suppression of cell growth and primary tumor formation and inhibition of signaling pathways that facilitate EMT. Nevertheless, the significant role of ERp29 in cell survival against drugs, induction of cell differentiation and potential promotion of MET-related metastasis may lead us to re-assess its function in cancer progression, particularly in distant metastasis. Hence, it is important to explore in detail the ERp29’s role in cancer as a “friend or foe” and to elucidate its clinical significance in breast cancer and other epithelial cancers. Targeting ERp29 and/or its downstream molecules might be an alternative molecular therapeutic approach for chemo/radio-resistant metastatic cancer treatment

7.3.5 Putting together structures of epidermal growth factor receptors

Bessman NJ, Freed DM, Lemmon MA
Curr Opin Struct Biol. 2014 Dec; 29:95-101


  • Several studies suggest flexible linkage between extracellular and intracellular regions. • Others imply more rigid connections, required for allosteric regulation of dimers. • Interactions with membrane lipids play important roles in EGFR regulation. • Cellular studies suggest half-of-the-sites negative cooperativity for human EGFR.

Numerous crystal structures have been reported for the isolated extracellular region and tyrosine kinase domain of the epidermal growth factor receptor (EGFR) and its relatives, in different states of activation and bound to a variety of inhibitors used in cancer therapy. The next challenge is to put these structures together accurately in functional models of the intact receptor in its membrane environment. The intact EGFR has been studied using electron microscopy, chemical biology methods, biochemically, and computationally. The distinct approaches yield different impressions about the structural modes of communication between extracellular and intracellular regions. They highlight possible differences between ligands, and also underline the need to understand how the receptor interacts with the membrane itself.

Growth factor receptor tyrosine kinases (RTKs) such as the epidermal growth factor receptor (EGFR) have been the subjects of intense study for many years [1,2]. There are 58 RTKs in the deduced human
proteome, and all play key roles in regulating cellular processes such as proliferation, differentiation, cell survival and metabolism, cell migration, and cell cycle control [3].  Importantly, aberrant activation
of RTK signaling by mutation, gene amplification, gene translocation or other mechanisms has been causally linked to cancers, diabetes, inflammation, and other diseases. These observations have prompted
the development of many targeted therapies that inhibit RTKs such as EGFR [4], Kit, VEGFR, or their ligands — typically employing therapeutic antibodies [5] or small molecule tyrosine kinase inhibitors [6].
Following the initial discoveries for EGFR [7] and the platelet-derived growth factor receptor (PDGFR) [8] that ligand-stabilized dimers are essential for RTK signaling, structural studies over the past decade
or so have guided development of quite sophisticated mechanistic views[1]. Each RTK has a ligand-binding extracellular region (ECR) that is linked by a single transmembrane a-helix to an intracellular
tyrosine kinase domain (TKD). Structures of the isolated ECRs and TKDs from several RTKs point to surprising mechanistic diversity across the larger family [1]. Unliganded RTKs exist as an equilibrium
mixture of inactive monomers, inactive dimers and active dimers (Figure 1), except for the extreme case of the insulin receptor (IR), which is covalently dimerized [9]. Extracellular ligand can bind to monomers,
to inactive dimers, or to active dimers — in each case pushing the equilibria shown in Figure 1 towards the central ligand-bound active dimer. Thus, ligand binding can drive receptor dimerization (Figure 1,
upper), or can promote inactive-to-active conformational transitions in dimers (Figure 1, lower). Regardless of pathway, the intracellular TKD of the ligand-stabilized dimer becomes activated either through
trans-autophosphorylation or through induced allosteric changes [1,10]. Roles for other parts of the receptor in RTK activation, including the juxtamembrane (JM) and transmembrane (TM) segments, have
also become clearer. The key current challenge for the field is to assemble data from many studies of isolated RTK parts into coherent views of how the intact receptors are regulated in their native membranes.
We will focus here on recent efforts to do this for the EGFR (or ErbB receptor) family. The missing links in intact RTKs: flexible or rigid? A central goal in extrapolating to the intact RTKs from studies of
isolated soluble domains is to understand how the individual parts of the receptor communicate with one another. The methods that have been used to produce and study the isolated domains inevitably
yield the impression that inter-domain linkers are flexible and disordered. For example, extracellular juxtamembrane regions have typically only been observed as C-terminal extensions of  the soluble ECR.
Similarly, intracellular juxtamembrane regions have been encountered predominantly as N-terminal extensions of TKD constructs, or as short peptides. In each of these contexts, the JM regions are incomplete,
and may appear disordered and flexible simply because key structural restraints have been removed. Nonetheless, this possible artifact has strongly influenced thinking about linkages between the extracellular
and intracellular regions [11], and in turn about mechanisms of RTK signaling. Highly flexible linkages between extracellular and intracellular regions of RTKs are fully consistent with simpler ligand-induced
dimerization models for transmembrane signaling by RTKs. It is more difficult, however, to understand how subtle allosteric communication across the membrane could be achieved if the linkages are truly
flexible. For example, since flexible linkage implies structural independence of the extracellular and intracellular regions, it is difficult to envision how a transition from inactive to active dimer in Figure 1
could be controlled precisely by ligand without more rigid (or restricted) connections.

Recent experimental studies with intact — or nearly intact — EGFR differ in the impressions they provide about how flexibly or rigidly the extracellular and intracellular regions are linked. Springer’s laboratory used cysteine crosslinking and mutagenesis approaches to investigate this issue for EGFR expressed in Ba/F3 cells [12]. They were unable to identify any specific JM or TM region interfaces
that were required for EGFR signaling, leading them to argue that the linkage across the membrane is too flexible to transmit a specific orientation between the extracellular and intracellular regions.
Consistent with this, negative-stain electron microscopy studies of (nearly) full-length EGFR in dodecylmaltoside micelles showed that a given extracellular dimer can be linked to several different
arrangements of the intracellular kinase domain [13,14]. Similarly, dimers driven by inhibitor binding to the intracellular TKD could couple to multiple different ECR conformations [13]. Biochemical
studies are also consistent with such structural independence of the extracellular and intracellular  regions [15,16]. Contrasting with these observations, however, Schepartz and colleagues have
reported that different precise conformations within the EGFR intracellular region can be induced by distinct activating ligands [17]. They used a method called bipartite tetracysteine display that
reports on formation of a chemically detectable tetracysteine motif when two cysteine pairs come together at  the dimer  interface. EGF activation of the receptor led to formation of a  tetracysteine
motif that requires the intracellular JM helix  [18] shown in Figure 2a to form antiparallel coiled-coil dimers  (Figure 2b/c) as proposed by Kuriyan and colleagues [19,20]. Surprisingly, transforming
growth factor-a (TGFa),which also activates EGFR, did not bring these two cysteine pairs together in the same way — arguing that TGFa does not induce formation of the same intracellular antiparallel
coiled-coil. Instead, activation of EGFR with TGFa (but not EGF) stabilized an alternative tetracysteine motif, consistent with a different intracellular JM structure. Evidence for ‘inside-out’ signaling
in EGFR has also been reported, where alterations in the intracellular JM region directly influence allosteric EGF binding to the ECR of the intact receptor analyzed in CHO cells [21–23]. The contradictory
views of flexibility versus rigidity  in linkages between the domains leave the path to understanding the intact receptor unclear, although it seems  reasonable doubt that  the inactive dimers known to
form in the absence of ligand [24–26] could be regulated by extracellular ligand if all linkages are always highly flexible.
Does the membrane hold the key?
All of the studies that support direct conformational communication between the extracellular and intracellular regions of EGFR were performed in cells [17,21,22]. By contrast, most of those that
explicitly suggest otherwise were performed in detergent micelles [13,14,15] — where the potentially important influences of specific membrane lipids (or membrane geometry) are absent. Studies of intact  EGFR in liposomes with defined lipid compositions [27] have shown that the ganglioside GM3 inhibits ligand-independent activation (and dimerization) of the receptor, apparently through interactions with a  site in its extracellular JM region. McLaughlin and colleagues [28,29] also proposed a model in which interaction of the intracellular JM region (and TKD) with anionic phospholipids in the inner leaflet of  the plasma membrane (notably PtdIns(4,5)P2) exerts an inhibitory effect that must be overcome in order for EGFR to signal. Association of the JM and TM regions with specific membrane lipids is likely to  define specific structures in the linkages between the EGFR extracellular and intracellular regions that are more well-defined (and potentially rigid) than is typically appreciated. Recent studies have begun to  shed some structural light on how membrane interactions with the intracellular JM region of EGFR might influence the signaling mechanism. Endres et al. [20] found that simply tethering the complete  intracellular region of EGFR to the inner leaflet of the plasma membrane maintains the TKD in a largely monomeric state and inhibits its kinase activity. Parallel computational studies [30] suggest that this  results from the previously proposed [29] inhibitory interaction of the JM and TKD regions of EGFR with the negatively charged membrane surface. The data of Endres et al. [20] further indicated that TM-mediated dimerization reverses this inhibitory effect. Moreover, NMR studies of a 60-residue peptide containing the TM and part of  the JM region solubilized in lipid bicelles led them to conclude that specific  TM dimerization through an N terminal GxxxG motif stabilizes formation of an antiparallel coiled-coil between the two JM fragments in the dimer — the same JM coiled-coil shown in Figure 2b/c that was  investigated in the bipartite tetracysteine display studies of  intact EGF-bound EGFR described above [17,19]. Independent solid-state NMR studies of a similar TM-JM peptide from the EGFR relative
ErbB2 in vesicles containing acidic phospholipids [31] further suggested that an activating mutation in the TM domain leads to release of  the JM region from the anionic membrane surface. Collectively,
these data suggest that ligand-induced dimerization of the receptor (or reorientation of receptors within a dimer) may engage the TM domain in a specific dimer that promotes both the formation of activating
interactions in the JM region and the disruption of inhibitory interactions between the JM region (and possibly TKD) and the membrane surface.

Negative cooperativity 
A key characteristic of ligand binding at the cell surface to EGFR [36], IR [37], and other receptors [38] is negative cooperativity — which is lost when soluble forms of the ECR from human EGFR [39]
or IR [40] are studied in isolation. Several studies have shown that intracellular and/or transmembrane regions are required for this negative cooperativity to be manifest [21,22,40,41], implying that
these parts of the receptor contribute to breaking the symmetry of the dimer — as required for the two sites to have distinct binding properties [42]. Such propagation of dimer asymmetry across the
membrane would surely require defined structures in the regions that connect extracellular and intracellular regions, and is difficult to reconcile with highly flexible JM linkers.
In brief, binding of one ligand stabilizes a singly-liganded asymmetric dimer in which the unoccupied ligand-binding site is compromised [43]. The binding affinity of the second ligand is thus reduced,
constituting a half-of-the-sites mode of negative cooperativity [44]. Leahy’s group has provided important evidence consistent with a similar mechanism in the cases of human EGFR and ErbB4 [16].
By comparing human ErbB receptor ECR dimer crystal structures with different bound ligands, Leahy and colleagues went on to identify two types of dimer interface [16], a ‘flush’ interface that resembles
the asymmetric (singly-liganded) dimer seen for the Drosophila EGFR [43] and a ‘staggered’ interface seen in the ECRs from EGFR (with bound EGF [12]) and ErbB4 (with bound neuregulin1b[16]).
These observations suggest that the ‘flush’ interface drives the most  stable dimers, which are singly liganded (Figure 2b). Binding of the second ligand is weaker, and also forces the dimer interface
into the less stable ‘staggered’ conformation (Figure 2c). Taken together, these findings suggest both a structural basis for negative cooperativity and a possible structural distinction between singly-liganded
and doubly-liganded ErbB receptor dimers.

A model for EGFR activation
The model shown in Figure 2 summarizes key proposed steps in the activation of human EGFR. In the absence of ligand, the ECR exists in a tethered conformation with the domain II ‘dimerization
arm’ engaged in an intramolecular interaction with domain IV that occludes the dimer interface [49]. The TKDs and the N-terminal portions of each intracellular JM region are thought to be engaged
in autoinhibitory interactions with the membrane surface [20,28,29,30].

Figure 2. More detailed view of EGF-induced activation of EGFR, as described in the text.
In the absence of ligand (a), the ECR adopts a tethered conformation, with an autoinhibitory tether interaction between domains II and IV. The TKD and JM regions lie against the membrane, making what
are believed to be additional autoinhibitory interactions. Domains I and III of the ECR are colored red, and domains II and IV are green. The JM helix is shown as a short cylinder and labeled in magenta.
The N-lobes and C-lobes of the kinase are also labeled, and both helix aC (blue) and the short helix in the activation loop (green) that interacts with aC to inhibit the TKD [50] are shown. The C-tail is
also depicted as a curve bearing 5 tyrosines. As described in the text, binding of a single ligand (b) induces formation of a singly-liganded dimer with a ‘flush’ (presumed asymmetric) ECR dimer interface.
The JM region forms an anti-parallel helix, as labeled in magenta, and the TKDs form an asymmetric dimer in which the activator (grey) allosterically activates the receiver (shown with an amber N-lobe).
It is not clear how the extracellular and intracellular asymmetry is structurally related, if at all. Finally, a second ligand binds to yield a more symmetric dimer with the ‘staggered’ ECR interface (c) described
in the text.

Conclusions Our mechanistic understanding of EGFR and its relatives has advanced dramatically in recent years, and the past year or two has seen substantial progress in putting the results of studies
with isolated domains together into initial views of how the intact receptor works. New insights into the origin of allosteric regulation of EGFR have been gained through a combination of innovative
structural, biochemical, cellular, and computational studies. A self-consistent picture is beginning to emerge. Two key issues remain unclear, however, and represent the current frontiers in studies of EGFR.
The first — for which we describe progress in this review — centers on the influence of specific interactions of the receptor with membrane lipids, which seem likely to define the structural ‘connections’
between extracellular and intracellular regions of the receptor. The second centers on the role of the carboxy-terminal 230 amino acids, which is believed to play a regulatory role for which little detail has
so far been defined [55].

7.3.6 Complex Relationship between Ligand Binding and Dimerization in the Epidermal Growth Factor Receptor

Bessman NJ1Bagchi A2Ferguson KM2Lemmon MA3.
Cell Rep. 2014 Nov 20; 9(4):1306-17.


  • Preformed extracellular dimers of human EGFR are structurally heterogeneous • EGFR dimerization does not stabilize ligand binding
    • Extracellular mutations found in glioblastoma do not stabilize EGFR dimerization • Glioblastoma mutations in EGFR increase ligand-binding affinity

The epidermal growth factor receptor (EGFR) plays pivotal roles in development and is mutated or overexpressed in several cancers. Despite recent advances, the complex allosteric regulation of EGFR remains incompletely understood. Through efforts to understand why the negative cooperativity observed for intact EGFR is lost in studies of its isolated extracellular region (ECR), we uncovered unexpected relationships between ligand binding and receptor dimerization. The two processes appear to compete. Surprisingly, dimerization does not enhance ligand binding (although ligand binding promotes dimerization). We further show that simply forcing EGFR ECRs into preformed dimers without ligand yields ill-defined, heterogeneous structures. Finally, we demonstrate that extracellular EGFR-activating mutations in glioblastoma enhance ligand-binding affinity without directly promoting EGFR dimerization, suggesting that these oncogenic mutations alter the allosteric linkage between dimerization and ligand binding. Our findings have important implications for understanding how EGFR and its relatives are activated by specific ligands and pathological mutations.

X-ray crystal structures from 2002 and 2003 (Burgess et al., 2003) yielded the scheme for ligand-induced epidermal growth factor receptor (EGFR) dimerization shown in Figure 1. Binding of a single ligand to domains I and III within the same extracellular region (ECR) stabilizes an “extended” conformation and exposes a dimerization interface in domain II, promoting self-association with a KD in the micromolar range (Burgess et al., 2003, Dawson et al., 2005, Dawson et al., 2007). Although this model satisfyingly explains ligand-induced EGFR dimerization, it fails to capture the complex ligand-binding characteristics seen for cell-surface EGFR, with concave-up Scatchard plots indicating either negative cooperativity (De Meyts, 2008, Macdonald and Pike, 2008) or distinct affinity classes of EGF-binding site with high-affinity sites responsible for EGFR signaling (Defize et al., 1989). This cooperativity or heterogeneity is lost when the ECR from EGFR is studied in isolation, as also described for the insulin receptor (De Meyts, 2008).

Figure 1

Structural View of Ligand-Induced Dimerization of the hEGFR ECR

(A) Surface representation of tethered, unliganded, sEGFR from Protein Data Bank entry 1NQL (Ferguson et al., 2003). Ligand-binding domains I and III are green and cysteine-rich domains II and IV are cyan. The intramolecular domain II/IV tether is circled in red.

(B) Hypothetical model for an extended EGF-bound sEGFR monomer based on SAXS studies of an EGF-bound dimerization-defective sEGFR variant (Dawson et al., 2007) from PDB entry 3NJP (Lu et al., 2012). EGF is blue, and the red boundary represents the primary dimerization interface.

(C) 2:2 (EGF/sEGFR) dimer, from PDB entry 3NJP (Lu et al., 2012), colored as in (B). Dimerization arm contacts are circled in red.

Here, we describe studies of an artificially dimerized ECR from hEGFR that yield useful insight into the heterogeneous nature of preformed ECR dimers and into the origins of negative cooperativity. Our data also argue that extracellular structures induced by ligand binding are not “optimized” for dimerization and conversely that dimerization does not optimize the ligand-binding sites. We also analyzed the effects of oncogenic mutations found in glioblastoma patients (Lee et al., 2006), revealing that they affect allosteric linkage between ligand binding and dimerization rather than simply promoting EGFR dimerization. These studies have important implications for understanding extracellular activating mutations found in EGFR/ErbB family receptors in glioblastoma and other cancers and also for understanding specificity of ligand-induced ErbB receptor heterodimerization

Predimerizing the EGFR ECR Has Modest Effects on EGF Binding

To access preformed dimers of the hEGFR ECR (sEGFR) experimentally, we C-terminally fused (to residue 621 of the mature protein) either a dimerizing Fc domain (creating sEGFR-Fc) or the dimeric leucine zipper from S. cerevisiae GCN4 (creating sEGFR-Zip). Size exclusion chromatography (SEC) and/or sedimentation equilibrium analytical ultracentrifugation (AUC) confirmed that the resulting purified sEGFR fusion proteins are dimeric (Figure S1). To measure KD values for ligand binding to sEGFR-Fc and sEGFR-Zip, we labeled EGF with Alexa-488 and monitored binding in fluorescence anisotropy (FA) assays. As shown in Figure 2A, EGF binds approximately 10-fold more tightly to the dimeric sEGFR-Fc or sEGFR-Zip proteins than to monomeric sEGFR (Table 1). The curves obtained for EGF binding to sEGFR-Fc and sEGFR-Zip showed no signs of negative cooperativity, with sEGFR-Zip actually requiring a Hill coefficient (nH) greater than 1 for a good fit (nH = 1 for both sEGFRWT and sEGFR-Fc). Thus, our initial studies argued that simply dimerizing human sEGFR fails to restore the negatively cooperative ligand binding seen for the intact receptor in cells.

One surprise from these data was that forced sEGFR dimerization has only a modest (≤10-fold) effect on EGF-binding affinity. Under the conditions of the FA experiments, isolated sEGFR (without zipper or Fc fusion) remains monomeric; the FA assay contains just 60 nM EGF, so the maximum concentration of EGF-bound sEGFR is also limited to 60 nM, which is over 20-fold lower than the KD for dimerization of the EGF/sEGFR complex (Dawson et al., 2005, Lemmon et al., 1997). This ≤10-fold difference in affinity for dimeric and monomeric sEGFR seems small in light of the strict dependence of sEGFR dimerization on ligand binding (Dawson et al., 2005,Lax et al., 1991, Lemmon et al., 1997). Unliganded sEGFR does not dimerize detectably even at millimolar concentrations, whereas liganded sEGFR dimerizes with KD ∼1 μM, suggesting that ligand enhances dimerization by at least 104– to 106-fold. Straightforward linkage of dimerization and binding equilibria should stabilize EGF binding to dimeric sEGFR similarly (by 5.5–8.0 kcal/mol). The modest difference in EGF-binding affinity for dimeric and monomeric sEGFR is also significantly smaller than the 40- to 100-fold difference typically reported between high-affinity and low-affinity EGF binding on the cell surface when data are fit to two affinity classes of binding site (Burgess et al., 2003, Magun et al., 1980).

Mutations that Prevent sEGFR Dimerization Do Not Significantly Reduce Ligand-Binding Affinity

The fact that predimerizing sEGFR only modestly increased ligand-binding affinity led us to question the extent to which domain II-mediated sEGFR dimerization is linked to ligand binding. It is typically assumed that the domain II conformation stabilized upon forming the sEGFR dimer in Figure 1C optimizes the domain I and III positions for EGF binding. To test this hypothesis, we introduced a well-characterized pair of domain II mutations into sEGFRs that block dimerization: one at the tip of the dimerization arm (Y251A) and one at its “docking site” on the adjacent molecule in a dimer (R285S). The resulting (Y251A/R285S) mutation abolishes sEGFR dimerization and EGFR signaling (Dawson et al., 2005, Ogiso et al., 2002). Importantly, we chose isothermal titration calorimetry (ITC) for these studies, where all interacting components are free in solution. Previous surface plasmon resonance (SPR) studies have indicated that dimerization-defective sEGFR variants bind immobilized EGF with reduced affinity (Dawson et al., 2005), and we were concerned that this reflects avidity artifacts, where dimeric sEGFR binds more avidly than monomeric sEGFR to sensor chip-immobilized EGF.

Surprisingly, our ITC studies showed that the Y251A/R285S mutation has no significant effect on ligand-binding affinity for sEGFR in solution (Table 1). These experiments employed sEGFR (with no Fc fusion) at 10 μM—ten times higher than KD for dimerization of ligand-saturated WT sEGFR (sEGFRWT) (KD ∼1 μM). Dimerization of sEGFRWT should therefore be complete under these conditions, whereas the Y251A/R285S-mutated variant (sEGFRY251A/R285S) does not dimerize at all (Dawson et al., 2005). The KD value for EGF binding to dimeric sEGFRWT was essentially the same (within 2-fold) as that for sEGFRY251A/R285S (Figures 2B and 2C; Table 1), arguing that the favorable Gibbs free energy (ΔG) of liganded sEGFR dimerization (−5.5 to −8 kcal/mol) does not contribute significantly (<0.4 kcal/mol) to enhanced ligand binding. …

Thermodynamics of EGF Binding to sEGFR-Fc

If there is no discernible positive linkage between sEGFR dimerization and EGF binding, why do sEGFR-Fc and sEGFR-Zip bind EGF ∼10-fold more strongly than wild-type sEGFR? To investigate this, we used ITC to compare EGF binding to sEGFR-Fc and sEGFR-Zip (Figures 3A and 3B ) with binding to isolated (nonfusion) sEGFRWT. As shown in Table 1, the positive (unfavorable) ΔH for EGF binding is further elevated in predimerized sEGFR compared with sEGFRWT, suggesting that enforced dimerization may actually impair ligand/receptor interactions such as hydrogen bonds and salt bridges. The increased ΔH is more than compensated for, however, by a favorable increase in TΔS. This favorable entropic effect may reflect an “ordering” imposed on unliganded sEGFR when it is predimerized, such that it exhibits fewer degrees of freedom compared with monomeric sEGFR. In particular, since EGF binding does induce sEGFR dimerization, it is clear that predimerization will reduce the entropic cost of bringing two sEGFR molecules into a dimer upon ligand binding, possibly underlying this effect.

Possible Heterogeneity of Binding Sites in sEGFR-Fc

Close inspection of EGF/sEGFR-Fc titrations such as that in Figure 3A suggested some heterogeneity of sites, as evidenced by the slope in the early part of the experiment. To investigate this possibility further, we repeated titrations over a range of temperatures. We reasoned that if there are two different types of EGF-binding sites in an sEGFR-Fc dimer, they might have different values for heat capacity change (ΔCp), with differences that might become more evident at higher (or lower) temperatures. Indeed, ΔCp values correlate with the nonpolar surface area buried upon binding (Livingstone et al., 1991), and we know that this differs for the two Spitz-binding sites in the asymmetric Drosophila EGFR dimer (Alvarado et al., 2010). As shown in Figure 3C, the heterogeneity was indeed clearer at higher temperatures for sEGFR-Fc—especially at 25°C and 30°C—suggesting the possible presence of distinct classes of binding sites in the sEGFR-Fc dimer. We were not able to fit the two KD values (or ΔH values) uniquely with any precision because the experiment has insufficient information for unique fitting to a model with four variables. Whereas binding to sEGFRWT could be fit confidently with a single-site binding model throughout the temperature range, enforced sEGFR dimerization (by Fc fusion) creates apparent heterogeneity in binding sites, which may reflect negative cooperativity of the sort seen with dEGFR. …

Ligand Binding Is Required for Well-Defined Dimerization of the EGFR ECR

To investigate the structural nature of the preformed sEGFR-Fc dimer, we used negative stain electron microscopy (EM). We hypothesized that enforced dimerization might cause the unliganded ECR to form the same type of loose domain II-mediated dimer seen in crystals of unliganded Drosophila sEGFR (Alvarado et al., 2009). When bound to ligand (Figure 4A), the Fc-fused ECR clearly formed the characteristic heart-shape dimer seen by crystallography and EM (Lu et al., 2010, Mi et al., 2011). Figure 4B presents a structural model of an Fc-fused liganded sEGFR dimer, and Figure 4C shows a calculated 12 Å resolution projection of this model. The class averages for sEGFR-Fc plus EGF (Figure 4A) closely resemble this model, yielding clear densities for all four receptor domains, arranged as expected for the EGF-induced domain II-mediated back-to-back extracellular dimer shown in Figure 1 (Garrett et al., 2002, Lu et al., 2010). In a subset of classes, the Fc domain also appeared well resolved, indicating that these particular arrangements of the Fc domain relative to the ECR represent highly populated states, with the Fc domains occupying similar positions to those of the kinase domain in detergent-solubilized intact receptors (Mi et al., 2011). …

Our results and those of Lu et al. (2012)) argue that preformed extracellular dimers of hEGFR do not contain a well-defined domain II-mediated interface. Rather, the ECRs in these dimers likely sample a broad range of positions (and possibly conformations). This conclusion argues against recent suggestions that stable unliganded extracellular dimers “disfavor activation in preformed dimers by assuming conformations inconsistent with” productive dimerization of the rest of the receptor (Arkhipov et al., 2013). The ligand-free inactive dimeric ECR species modeled by Arkhipov et al. (2013) in their computational studies of the intact receptor do not appear to be stable. The isolated ECR from EGFR has a very low propensity for self-association without ligand, with KD in the millimolar range (or higher). Moreover, sEGFR does not form a defined structure even when forced to dimerize by Fc fusion. It is therefore difficult to envision how it might assume any particular autoinhibitory dimeric conformation in preformed dimers. …

Extracellular Oncogenic Mutations Observed in Glioblastoma May Alter Linkage between Ligand Binding and sEGFR Dimerization

Missense mutations in the hEGFR ECR were discovered in several human glioblastoma multiforme samples or cell lines and occur in 10%–15% of glioblastoma cases (Brennan et al., 2013, Lee et al., 2006). Several elevate basal receptor phosphorylation and cause EGFR to transform NIH 3T3 cells in the absence of EGF (Lee et al., 2006). Thus, these are constitutively activating oncogenic mutations, although the mutated receptors can be activated further by ligand (Lee et al., 2006, Vivanco et al., 2012). Two of the most commonly mutated sites in glioblastoma, R84 and A265 (R108 and A289 in pro-EGFR), are in domains I and II of the ECR, respectively, and contribute directly in inactive sEGFR to intramolecular interactions between these domains that are thought to be autoinhibitory (Figure 5). Domains I and II become separated from one another in this region upon ligand binding to EGFR (Alvarado et al., 2009), as illustrated in the lower part of Figure 5. Interestingly, analogous mutations in the EGFR relative ErbB3 were also found in colon and gastric cancers (Jaiswal et al., 2013).

We hypothesized that domain I/II interface mutations might activate EGFR by disrupting autoinhibitory interactions between these two domains, possibly promoting a domain II conformation that drives dimerization even in the absence of ligand. In contrast, however, sedimentation equilibrium AUC showed that sEGFR variants harboring R84K, A265D, or A265V mutations all remained completely monomeric in the absence of ligand (Figure 6A) at a concentration of 10 μM, which is similar to that experienced at the cell surface (Lemmon et al., 1997). As with WT sEGFR, however, addition of ligand promoted dimerization of each mutated sEGFR variant, with KD values that were indistinguishable from those of WT. Thus, extracellular EGFR mutations seen in glioblastoma do not simply promote ligand-independent ECR dimerization, consistent with our finding that even dimerized sEGFR-Fc requires ligand binding in order to form the characteristic heart-shaped dimer. …

We suggest that domain I is normally restrained by domain I/II interactions so that its orientation with respect to the ligand is compromised. When the domain I/II interface is weakened with mutations, this effect is mitigated. If this results simply in increased ligand-binding affinity of the monomeric receptor, the biological consequence might be to sensitize cells to lower concentrations of EGF or TGF-α (or other agonists). However, cellular studies of EGFR with glioblastoma-derived mutations (Lee et al., 2006, Vivanco et al., 2012) clearly show ligand-independent activation, arguing that this is not the key mechanism. The domain I/II interface mutations may also reduce restraints on domain II so as to permit dimerization of a small proportion of intact receptor, driven by the documented interactions that promote self-association of the transmembrane, juxtamembrane, and intracellular regions of EGFR (Endres et al., 2013, Lemmon et al., 2014, Red Brewer et al., 2009).

Setting out to test the hypothesis that simply dimerizing the EGFR ECR is sufficient to recover the negative cooperativity lost when it is removed from the intact receptor, we were led to revisit several central assumptions about this receptor. Our findings suggest three main conclusions. First, we find that enforcing dimerization of the hEGFR ECR does not drive formation of a well-defined domain II-mediated dimer that resembles ligand-bound ECRs or the unliganded ECR from Drosophila EGFR. Our EM and SAXS data show that ligand binding is necessary for formation of well-defined heart-shaped domain II-mediated dimers. This result argues that the unliganded extracellular dimers modeled by Arkhipov et al. (2013)) are not stable and that it is improbable that stable conformations of preformed extracellular dimers disfavor receptor activation by assuming conformations that counter activating dimerization of the rest of the receptor. Recent work from the Springer laboratory employing kinase inhibitors to drive dimerization of hEGFR (Lu et al., 2012) also showed that EGF binding is required to form heart-shaped ECR dimers. These findings leave open the question of the nature of the ECR in preformed EGFR dimers but certainly argue that it is unlikely to resemble the crystallographic dimer seen for unligandedDrosophila EGFR (Alvarado et al., 2009) or that suggested by computational studies (Arkhipov et al., 2013).

This result argues that ligand binding is required to permit dimerization but that domain II-mediated dimerization may compromise, rather than enhance, ligand binding. Assuming flexibility in domain II, we suggest that this domain serves to link dimerization and ligand binding allosterically. Optimal ligand binding may stabilize one conformation of domain II in the scheme shown in Figure 1 that is then distorted upon dimerization of the ECR, in turn reducing the strength of interactions with the ligand. Such a mechanism would give the appearance of a lack of positive linkage between ligand binding and ECR dimerization, and a good test of this model would be to determine the high-resolution structure of a liganded sEGFR monomer (which we expect to differ from a half dimer). This model also suggests a mechanism for selective heterodimerization over homodimerization of certain ErbB receptors. If a ligand-bound EGFR monomer has a domain II conformation that heterodimerizes with ErbB2 in preference to forming EGFR homodimers, this could explain several important observations. It could explain reports that ErbB2 is a preferred heterodimerization partner of EGFR (Graus-Porta et al., 1997) and might also explain why EGF binds more tightly to EGFR in cells where it can form heterodimers with ErbB2 than in cells lacking ErbB2, where only EGFR homodimers can form (Li et al., 2012).

7.3.7 IGFBP-2/PTEN: A critical interaction for tumours and for general physiology?

The insulin-like growth factor family of proteins, together with insulin, form an evolutionarily conserved system that helps to coordinate the metabolic status and activity of organisms with their nutritional environment. When food is abundant, the IGF/insulin signalling pathway is switched on and cell proliferation and other activities are enhanced; while when food is limited, such activities are suppressed to conserve energy and resources [1,2]. The IGF axis consists of two ligands IGF-I and -II, a series of heterotetrameric tyrosine kinase receptors and six high affinity binding proteins IGFBP-1 to-6. These IGFBPs not only regulate the reservoir, availability and functions of IGFs but also have direct actions upon cell behaviour that are independent of IGF-binding [3]. The six IGFBPs are conserved in all placental mammals having evolved from serial duplication of genes that were present throughout vertebrate evolution [4]. Each of the six IGFBPs has evolved unique functions that presumably have conferred some evolutionary advantage and hence have been conserved across mammalian evolution. After IGFBP-3, IGFBP-2 is the second most abundant binding protein in the circulation throughout adult life in humans. While circulating IGFBP-3 levels peak during puberty and decrease thereafter, IGFBP-2 levels are highest in infancy and old age. Together with the other five IGFBPs, IGFBP-2 regulates IGF availability and actions and has pleiotropic effects on normal and neoplastic tissues [3]. One of the clear distinctive structural features of IGFBP-2 is that it contains an Arg-Gly-Asp (RGD) sequence that enables functional interactions with integrin receptors [4]. This structural element is only present in one of the other IGFBPs, IGFBP-1. Although the RGD sequence was only acquired in IGFBP-1 during mammalian evolution it was present within IGFBP-2 from early vertebrate evolution indicating that it has been a long retained functional characteristic of IGFBP-2 [4]. The integrin receptors are critical for the anchorage of cells to the extracellular matrix (ECM) within tissues and hence for maintaining tissue architecture [5,6]. In solid tissue an important safeguard is imposed by linking normal cell functions and proliferation to appropriate cues from the ECM that are mediated by signals from attachment receptors such as the integrin receptors. Anchoragedependent growth is a common feature of normal cells and loss of attachment results in a form of apoptosis called anoikis. The integrin receptors interact with growth factor receptors in an ancillary and permissive manner to ensure that the signals for growth and survival occur in the appropriate setting and not inappropriately in detached cells. It has also become clear that integrin receptors serve many other roles in regulating cell functions and integrating cues from the surrounding ECM [5,6]. Over the last few decades, as the role of IGFBPs as extracellular modulators of IGF-availability and actions has emerged, there has also been a gradual characterization of the intracellular counter-regulatory components that modulate the signals initiated by IGF-receptor activation. There has been considerable progress in charting the signalling cascades initiated from these receptors but it is evident that the reason needs to be mechanisms for inactivating the pathways in intervening periods in preparation for subsequent activation. Throughout the canonical kinase cascades, activated by receptor ligation, at each node there is a corresponding phosphatase that returns the pathway to the inactive state and modulates the signal. The extracellular regulators of these phosphatases have however received much less attention than the activating kinases. That the extracellular counter-regulators may act in synchrony and be linked to the intracellular counter-regulators has only recently started to be revealed. Transgenic over-expression of IGFBP-2 at supra-physiological levels in mice results in reduced somatic growth [7] and this growth deficit is more pronounced when these mice were crossed with mice with raised growth hormone/IGF-I [8] implying that the growth inhibitory effect was due to sequestration of IGF-I. As with most of the IGFBP-family [3], there are also however multiple lines of evidence that IGFBP-2 has cellular actions that are independent of its ability to bind IGFs. There is evidence that IGFBP-2 initiates intrinsic cellular signalling through specific binding of its RGD-motif to integrin receptors, particularly the α5β1 integrin.In addition IGFBP-2 appears to modulate IGF and epidermal growth factor signalling through interactions with α5β3 integrins [9]. A heparin binding domain also exists in IGFBP-2 and it has been shown to bind to glycosaminoglycans [10], heparin [11], and other proteoglycans such as the receptor protein tyrosine phosphatase-β (RPTPβ) [12,13]. In addition,IGFBP-2has been reported to be localized on the cell surface, in the cytoplasm and on the nuclear membrane[14]. Several groups have now reported nuclear localization of IGFBP-2 [15–17]. A functional nuclear localization sequence in the central domain of IGFBP-2 has been reported that appears to interact with importin-α [18]. In the nucleus IGFBP-2 has been reported to regulate the expression of vascular endothelial growth factor [19].
IGFBP-2 and metabolic regulation
Epidemiological studies of human populations have indicated that IGFBP-2 levels are reduced in obesity, metabolic syndrome and type 2 diabetes and are inversely correlated with insulin sensitivity [20]. That these associations were due to a metabolic role for IGFBP-2, rather thanitjustbeingamarkerofdisturbance,hasbeenconfirmedinanumber of animal models. Using a transgenic IGFBP-2 over-expressing mouse model, Wheatcroft and coworkers found that IGFBP-2 was able to protect mice from high-fat/high-energy induced obesity and insulin resistance, and also protected the mice from the age-related development of glucose intolerance and hypertension [21]. Over-expression of IGFBP-2 induced by Leptin in wild type or obese mice similarly resulted in reduced plasma glucose and insulin levels [22]. All these data indicate a metabolic role for IGFBP-2 in glucose homeostasis.
IGFBP-2 and cancer
As indicated above, the early reports had implied that IGFBP-2 was generally a negative regulator of IGF-activity; the systemic growth restriction observed in transgenic mice over-expressing IGFBP-2 was followed by observations that chemically induced colorectal cancers were inhibited in this model [23]. Despite this there has been an accumulation of evidence indicating that IGFBP-2 is positively associated with the malignant progression of a wide range of cancers, as has been reviewed previously [24]. Raised serum levels of IGFBP-2 have been reported and positive associations between tumor IGFBP-2 expression and malignancy or metastasis have been observed for a variety of cancers, including glioma [25], breast [26], prostate [27], lung [28], colon [29] and lymphoid tumor [30]. Subsequent work has generally been consistent with this association between IGFBP-2 and cancer progression. In view of the majority of evidence, indicating that IGFBP-2 interacting with IGFs generally inhibited cell growth, it was suggested thatIGF-independentactionswereprobablyresponsibleforpositiveassociations between IGFBP-2 and tumourgrowth and progression [24]. The explanation for the increased expression of IGFBP-2 that has beenreportedformanydifferentcancersappearstocomefromthefactorsthat have been shown to regulate IGFBP-2 expression.The tumor suppressor gene p53, which is the most mutated gene in many human cancers, has been reported to transcriptionally regulate IGFBP-2 [31].

There also appears to again be reciprocal feedback as p53 mRNA in the breast cancer cell line Hs578T increased significantly after treatment with human recombinant IGFBP-2, suggesting a close interaction between IGFBP-2 and p53 [14]. A number of hormonal regulators of IGFBP-2 expression have been described including hCG, FSH, TGF-β, IL1, estradiol, glucocorticoids, EGF, IGF-I, IGF-II and insulin [24]. The stimulation of IGFBP-2 expression by EGF, IGF-I, IGF-II and insulin has been shown to be via the PI3K/AKT/mTOR pathway in breast cancer cells [32] and in adipocytes [33]. This is one of the most well characterisedsignallingpathwaysactivatedbyinsulinandIGFs.Inaddition the critical counter-regulatory phosphatase that deactivates this pathway the phosphatase and tensin homologue PTEN has been shown to downregulate the expression of IGFBP-2 [34]. This suggests another autoregulatory loop in which activation of the PI3K/AKT/mTOR pathway by IGFs induces the expression of IGFBP-2 that then sequesters the IGFs and modulates the signal. As activating mutations in the PI3K pathway or loss of PTEN are very common across a variety of human cancers, this plus the effect of p53, probably accounts for the common dysregulation of IGFBP-2 observed across many cancers. Using prostate cancer cell lines it has also been shown that local IGFBP-2 expression is metabolically regulated; IGFBP-2 expression was increased in hyperglycemic conditions through acetylation of histones H3 and H4 associated with the IGFBP-2 promoter, furthermore this up-regulation of IGFBP-2 mediated hyperglycemia-induced chemo-resistance [35].

The signaling kinase PI3K plays a fundamental role that has been maintained throughout most of evolution. The ability to control growth and development according to the availability of nutrients provides a survival advantage and therefore has been selectively retained throughout evolution. Evidence has accumulated to indicate that the PI3K pathway provides this control in all species from yeast to mammals.Various forms of the PI3K enzyme exist that are classified into three groups (classes I, II, and III). Only one of these forms is present in yeast and is equivalent to mammalian class III PI3K: this acts as a nutrient sensor and is directly activated by the availability of amino acids and then itself activates mTOR/S6K1 to regulate cell growth and development [36]. In mammals class 1API3K has evolved: this form is not directly activated by nutrients but consists of heterodimers comprising a catalytic p110 subunit and a regulatory p85 subunit that enables the enzyme to be controlled by receptor tyrosine kinases, classically the insulin and insulin-like growth factor receptors (IR and IGF-IR) [37]. This enables the regulation of PI3K by social nutritionally dependent signals rather than by nutrients directly. It is not by chance that the insulin/IGF/PI3K pathway plays a fundamental role in regulating both metabolism and growth as it clearly is an advantage to synchronize the set processes and this synchronized control has been maintained throughout evolution.

Phosphatase and tensin homolog (PTEN)
Of all the intracellular counter-regulators of the IGF-pathway the one that has received the most attention in relation to cancer is PTEN. PTEN is a lipid tyrosine phosphatase that negatively regulates the Akt/ PKB signaling pathway by specifically dephosphorylating phosphatidylinositol (3,4,5)-trisphosphate and thereby reduces AKT activation to reduce signals for cell metabolism, proliferation and survival [37]. PTEN is the second most often mutated tumor suppressor in human cancers, after p53[38]. Aberrant PTEN activity occurs due to mutation, homozygous deletion, loss of heterozygosity, or epigenetic silencing. Lost or reduced activity of PTEN has been observed in a great variety of cancers, including breast [39], prostate [40,41], colorectal [42], lung[43], glioblastoma [44], endometrial [45], etc. It has been demonstrated that deregulation of PTEN is involved in tumorigenesis, tumor progression and also the predisposition of many cancers [46]. AsPI3K/Akt signaling is critical for the metabolic effects of insulin. It is clear that PTEN will also play a role in modulating the metabolic actions of insulin. Consistent with this mice genetically modified to have haploinsufficiency of PTEN were observed to be hypersensitive to insulin [47]. Similarly humans with haplo-insufficiency due to mutations in PTEN were found to have enhanced insulin sensitivity [48]. Recently an increase in insulin sensitivity due to suppression of PTEN has been described in grizzly bears in preparation for hibernation, indicating that this is a mechanism for physiological adaptation [49]. Although the genetic defects resulting in PTEN loss in cancers and the intrinsic mechanisms for regulation of PTEN have been well characterised; there have been relatively few reports of external cell regulators. Of interest one of the few extrinsic regulators that has been described is IGF-II [50]. IGF-II is the most abundant growth factor present in most human tissues and activates the PI3K/AKT/mTOR pathway. Just as the induction of IGFBP-2 by activation of the PI3K pathway suggests an autoregulatory feedback loop extrinsic to the cell;the induction of PTEN by IGF-II via PI3K suggests an additional feedback loop that is intrinsic within the cell (Fig. 1). Activation of the pathway by IGF-II induces expression of PTEN that then attenuates the signal; conversely when the pathway is not activated then PTEN expression is reduced making the cell more responsive for when an activation signal is next received.One of the mechanisms that has emerged,to explain this feedback loop, indicates that the signaling output of the PI3K/PTEN pathway is balanced by asynchronous regulation of the activity of both PI3K and PTEN. The p85α regulatory subunit of PI3K that binds to and represses the activity of the p110 catalytic subunit also binds directly to PTEN at a regulatory site within PTEN where serine/threonine phosphorylation occurs to inactivatePTEN.The p85α subunit binds to unphosphorylated PTEN at this site and enhances its lipid phosphatase activity 3-fold [51]. The nature of this feedback loop has been further extended by reports that PTEN can suppress the expression of IGF-II [52,53]. As IGF-II induces PTEN, the ability of PTEN to subsequently reduce IGF-II expression may enable the cell to protect itself from over-stimulation. In contrast loss of PTEN may increase the expression of IGF-II resulting inactivation of the PI3K/AKT/mTOR pathway that is then unopposed.

PTEN/IGFBP-2 interactions
In view of the recognized importance of loss of PTEN for a variety of cancers there has been considerable interest in identifying biomarkers that could be used clinically to indicate loss of PTEN within tumors. An unbiased screen of human prostate cancer xenografts and human glioblastoma samples using microarray-based expression profiling found that the most significant gene was IGFBP-2 and it was confirmed in experimental models that IGFBP-2 was inversely regulated by PTEN [54]. This was consistent with all of the subsequent studies indicating that the expression of IGFBP-2 was regulated by the PI3K/AKT/mTOR pathway. An increase in tumor IGFBP-2 has also been associated with loss of PTEN in human breast cancer samples[55]. In the same year that a screen revealed IGFBP-2 as the best marker for loss of PTEN; the nature of the interaction between these two proteins was extended by the demonstration that at the cellular level IGFBP-2 can inversely regulate PTEN. With human breast cancer cells it was confirmed that IGF-II stimulated PTEN expression and that this was modulated by the binding of IGF-II to IGFBP-2, but when IGFBP-2 was not bound to IGF-II it was able to suppress PTEN via an interaction with cell surface integrin receptors (Fig. 1) [56]. Subsequent work with human prostate cancer cells indicated that the interaction of IGFBP-2 with integrin receptors could also result in PTEN inactivation via increasing its phosphorylation [57].

Fig.1. A proposed autoregulatory feedback loop of IGFBP-2/PTEN interaction. Binding of IGF-II to the IGF-IR activates the PI3K pathway. Induction of PI3K activates Akt and mTOR signaling and leads to cell proliferation and cell survival. The regulatory subunit of PI3K,p85, also binds and activates PTEN through dephosphorylation. This increased PTEN subsequently blocks IGFII production to avoid overstimulation. On the other hand, activated PI3K pathway induces IGFBP-2 expression, which when unbound to IGF-II, suppresses PTEN via an interaction with integrin receptors and/or the receptor protein tyrosine phosphatase β(RPTPβ). Thus the negative control of PTEN on PI3K signaling is counteracted. These feedback loops enable the extrinsic balance between IGF-II and IGFBP-2 to be tightly integrated to the intrinsic balance between PI3K and PTEN.

The ability of IGFBP-2 to regulate PTEN, originally observed in human cancer cell lines has subsequently been confirmed in a variety of normal cell types from different tissues. In IGFBP-2 knock-out mice a decrease in hematopoietic stem cell survival and cycling has been associated with an increase in PTEN and this appeared to be mediated by the heparin binding domain (HBD) within IGFBP-2 as the administration of a peptide analogue could restore the deficit [58]. Similarly a decrease in bone mass in the IGFBP-2 knock-out mice has been attributed to an increase in PTEN and again the use of a peptide analogue appeared to implicate the IGFBP-2HBD [59]. It was subsequently reported that the IGFBP-2HBD mediated an interaction with the RPTPβ resulting in dimerization and consequent inactivation of RPTPβ and that this reduction in phosphatase activity cooperated with IGF-I activation of the IGF-IR to enhance the phosphorylation and inactivation of PTEN [12]. Recently IGFBP-2 has been reported to also suppress PTEN in human skeletal muscle cells [60] and human visceral adipocytes [61] by interacting with integrin receptors. A similar association between IGFBP-2 and PTEN has been implicated as playing a role in murine skeletal muscle cell differentiation, although the functional regulation was not directly investigated in that study [62].

Evidence from a variety of different sources have indicated a close regulatory feedback loop between IGFBP-2 and PTEN. Work using a variety of different cell types from different tissues and different species has indicated that IGFBP-2 inversely regulates PTEN. There are reports that this is mediated via the IGFBP-2 RGD domain interacting with integrin receptors and by the IGFBP-2 HBD interacting with proteoglycans; the relative involvement of each of these domains and their functional interactions will require further work to elucidate. These studies however suggest a general mechanism that plays a role in a variety of normal physiological processes in addition to having important implications for the progression of many different cancers. The phosphatase PTEN has an important role in determining insulin sensitivity and the extent that IGFBP-2 exerts a metabolic role in regulating PTEN to determine insulin-sensitivity is yet to be examined. The extracellular balance between IGF-II and IGFBP-2 seems tightly linked with the intracellular balance between PI3K and PTEN (Fig. 1). When driving, in order to move forward there is a synchronous application of the accelerator and a removal of the brake. It appears that the cell also synchronizes activation of an essential regulatory pathway with the removal of the tightly linked inactivation pathway.

[1] B.C. Melnik, S.M. John, G. Schmitz, Over-stimulation of insulin/IGF-1 signaling by western diet may promote diseases of civilization: lessons learnt from Laron syndrome, Nutr. Metab. (Lond.) 8 (2011) 41. [2] J.M. Holly, C.M. Perks, Insulin-like growth factor physiology: what we have learned from human studies, Endocrinol. Metab. Clin. North. Am. 41 (2012) 249–263.
[3] J.Holly,C.Perks, The role ofinsulin-like growth factor binding proteins, Neuroendocrinology 83 (3–4) (2006) 154–160.
[4] D.O.Daza, etal.,Evolution of the insulin-like growth factor binding protein (IGFBP) family, Endocrinology 152 (6) (2011) 2278–2289.
[5] A.R. Ferreira, J.Felgueiras, M. Fardilha, Signaling pathways inanchoringjunctionsof epithelial cells: cell-to-cell and cell-to-extracellular matrix interactions, J. Recept. Signal Transduct. Res. (2014) 1–9.
[6] S.H. Kim, J. Turnbull, S. Guimond, Extracellular matrix and cell signalling: the dynamic cooperation of integrin, proteoglycan and growth factor receptor, J. Endocrinol. 209 (2) (2011) 139–151.
[7] A.Hoeflich,etal.,Overexpression ofinsulin-like growth factor-bindingprotein-2 in transgenic mice reduces postnatal body weight gain, Endocrinology 140 (12) (1999) 5488–5496.
[8] A. Hoeflich, et al., Growth inhibition in giant growth hormone transgenic mice by overexpression of insulin-like growth factor-binding protein-2, Endocrinology 142 (5) (2001) 1889–1898.
[9] G.K.Wang,etal., Aninteraction betweeninsulin-likegrowthfactor-bindingprotein 2 (IGFBP2) and integrin alpha5 is essential for IGFBP2-induced cell mobility, J. Biol. Chem. 281 (20) (2006) 14085–14091. [10] T.Arai,W.BusbyJr.,D.R.Clemmons,Bindingofinsulin-likegrowthfactor(IGF)IorII to IGF-binding protein-2 enables it to bind to heparin and extracellular matrix, Endocrinology 137 (11) (1996) 4571–4575. [11] J. Lund, et al., Heparin-binding mechanism of the IGF2/IGF-binding protein 2 complex, J. Mol. Endocrinol. 52 (3) (2014) 345–355.
[12] X. Shen, et al., Insulin-like growth factor (IGF) binding protein 2 functions coordinately with receptor protein tyrosinephosphatase βandtheIGF-Ireceptorto regulate IGF-I-stimulated signaling, Mol. Cell. Biol. 32 (20) (2012) 4116–4130.
[13] V.C.Russo, etal.,Insulin-like growth factor binding protein-2 bindingto extracellularmatrixplaysacriticalroleinneuroblastomacellproliferation,migration,andinvasion, Endocrinology 146 (10) (2005) 4445–4455.
[14] K.W. Frommer, etal., IGF-independent effects of IGFBP-2 on the human breast cancer cell line Hs578T, J. Mol. Endocrinol. 37 (1) (2006) 13–23.
[15] K. Miyako, et al., PAPA-1 Is a nuclear binding partner of IGFBP-2 and modulates its growth-promoting actions, Mol. Endocrinol. 23 (2) (2009) 169–175.
[16] X.Terrien,etal.,IntracellularcolocalizationandinteractionofIGF-bindingprotein-2 with the cyclin-dependent kinase inhibitor p21CIP1/WAF1 during growth inhibition, Biochem. J. 392 (Pt 3) (2005) 457–465.
[17] R.M. Villani, et al., Patched1 inhibits epidermal progenitor cell expansion and basal cell carcinoma formation by limiting Igfbp2 activity, Cancer Prev. Res. (Phila.) 3 (10) (2010) 1222–1234.
[18] W.J. Azar, et al., IGFBP-2 nuclear translocation is mediated by a functional NLS sequence and is essential for its pro-tumorigenic actions in cancer cells, Oncogene 33 (5) (2014) 578–588.
[19] W.J.Azar,etal.,IGFBP-2enhancesVEGFgenepromoteractivityandconsequentpromotion of angiogenesis by neuroblastoma cells, Endocrinology 152 (9) (2011) 3332–3342.
[20] S.B. Wheatcroft, M.T. Kearney, IGF-dependent and IGF-independent actions of IGFbinding protein-1 and -2: implications for metabolic homeostasis, Trends Endocrinol. Metab. 20 (4) (2009) 153–162. [21] S.B. Wheatcroft, et al., IGF-binding protein-2 protects against the development of obesity and insulin resistance, Diabetes 56 (2) (2007) 285–294.

7.3.8 Emerging roles for the pH-sensing G protein-coupled receptors in response to acidotic stress

Edward J Sanderlin, Calvin R Justus, Elizabeth A Krewson, Li V Yang
Cell Health & Cytoskel Mar 2015; 2015(7): 99—109

Protons (hydrogen ions) are the simplest form of ions universally produced by cellular metabolism including aerobic respiration and glycolysis. Export of protons out of cells by a number of acid transporters is essential to maintain a stable intracellular pH that is critical for normal cell function. Acid products in the tissue interstitium are removed by blood perfusion and excreted from the body through the respiratory and renal systems. However, the pH homeostasis in tissues is frequently disrupted in many pathophysiologic conditions such as in ischemic tissues and tumors where protons are overproduced and blood perfusion is compromised. Consequently, accumulation of protons causes acidosis in the affected tissue. Although acidosis has profound effects on cell function and disease progression, little is known about the molecular mechanisms by which cells sense and respond to acidotic stress. Recently a family of pH-sensing G protein-coupled receptors (GPCRs), including GPR4, GPR65 (TDAG8), and GPR68 (OGR1), has been identified and characterized. These GPCRs can be activated by extracellular acidic pH through the protonation of histidine residues of the receptors. Upon activation by acidosis the pH-sensing GPCRs can transduce several downstream G protein pathways such as the Gs, Gq/11, and G12/13 pathways to regulate cell behavior. Studies have revealed the biological roles of the pH-sensing GPCRs in the immune, cardiovascular, respiratory, renal, skeletal, endocrine, and nervous systems, as well as the involvement of these receptors in a variety of pathological conditions such as cancer, inflammation, pain, and cardiovascular disease. As GPCRs are important drug targets, small molecule modulators of the pH-sensing GPCRs are being developed and evaluated for potential therapeutic applications in disease treatment.

Cellular metabolism produces acid as a byproduct. Metabolism of each glucose molecule by glycolysis generates two pyruvate molecules. Under anaerobic conditions the metabolism of pyruvate results in the production of the glycolytic end product lactic acid, which has a pKa of 3.9. Lactic acid is deprotonated at the carboxyl group and results in one lactate ion and one proton at the physiological pH. Under aerobic conditions pyruvate is converted into acetyl-CoA and CO2 in the mitochondria. CO2in water forms a chemical equilibrium of carbonic acid and bicarbonate, an important physiological pH buffering system. The body must maintain suitable pH for proper physiological functions. Some regulatory mechanisms to control systemic pH are respiration, renal excretion, bone buffering, and metabolism.14 The respiratory system can buffer the blood by excreting carbonic acid as CO2 while the kidney responds to decreased circulatory pH by excreting protons and electrolytes to stabilize the physiological pH. Bone buffering helps maintain systemic pH by Ca2+ reabsorption and mineral dissolution. Collectively, it is clear that several biological systems require tight regulation to maintain pH for normal physiological functions. Cells utilize vast varieties of acid-base transporters for proper pH homeostasis within each biological context.58 Some such transporters are H+-ATPase, Na+/H+exchanger, Na+-dependent HCO3/C1 exchanger, Na+-independent anion exchanger, and monocarboxylate transporters. Cells can also maintain short-term pH homeostasis of the intracellular pH by rapid H+ consuming mechanisms. Some such mechanisms utilize metabolic conversions that move acids from the cytosol into organelles. Despite these cellular mechanisms that tightly maintain proper pH homeostasis, there are many diseases whereby pH homeostasis is disrupted. These pathological conditions are characterized by either local or systemic acidosis. Systemic acidosis can occur from respiratory, renal, and metabolic diseases and septic shock.14,9 Additionally, local acidosis is characterized in ischemic tissues, tumors, and chronically inflamed conditions such as in asthma and arthritis caused by deregulated metabolism and hypoxia.1015

Acidosis is a stress for the cell. The ability of the cell to sense and modulate activity for adaptation to the stressful environment is critical. There are several mechanisms whereby cells sense acidosis and modulate cellular functions to facilitate adaptation. Cells can detect extracellular pH changes by acid sensing ion channels (ASICs) and transient receptor potential (TRP) channels.16 Apart from ASIC and TRP channels, extracellular acidic pH was shown to stimulate inositol polyphosphate formation and calcium efflux.17,18 This suggested the presence of an unknown cell surface receptor that may be activated by a certain functional group, namely the imidazole of a histidine residue. The identity of the acid-activated receptor was later unmasked by Ludwig et al as a family of proton-sensing G protein-coupled receptors (GPCRs). This group identified human ovarian cancer GPCR 1 (OGR1) which upon activation will produce inositol phosphate and calcium efflux through the Gq pathway.19 These pH-sensing GPCR family members, including GPR4, GPR65 (TDAG8), and GPR68 (OGR1), will be discussed in this review (Figure 1). The proton-sensing GPCRs sense extracellular pH by protonation of several histidine residues on their extracellular domain. The activation of these proton-sensing GPCRs facilitates the downstream signaling through the Gq/11, Gs, and G12/13 pathways. Their expression varies in different cell types and play critical roles in sensing extracellular acidity and modulating cellular functions in several biological systems.

Figure 1 Biological roles and G protein coupling of the pH-sensing GPCRs.
Abbreviation: GPCRs, G protein-coupled receptors.

Role for the pH-sensing GPCRs in the immune system and inflammation

Acidic pH is a main characteristic of the inflammatory loci.14,20,21 The acidic microenvironment in inflamed tissue is predominately due to the increased metabolic demand from infiltrating immune cells, such as the neutrophil. These immune cells increase oxygen consumption and glucose uptake for glycolysis and oxidative phosphorylation. When oxygen availability is limited, cells often undergo anaerobic glycolysis. This process generates increasing amounts of lactic acid, thereby creating a local acidic microenvironment within the inflammatory loci.22 This presents a role for the pH-sensing GPCR GPR65 (TDAG8) in inflammation and immune cell function.23 TDAG8 was originally identified by cloning as an orphan GPCR which was observed to be upregulated during thymocyte apoptosis.24,25GPR65 (TDAG8) is predominately expressed in lymphoid tissues such as the spleen, lymph nodes, thymus, and leukocytes.2426 It was demonstrated that GPR65 inhibited pro-inflammatory cytokine secretion, which includes IL-6 and TNF-α, in mouse peritoneal macrophages upon activation by extracellular acidification. This cytokine inhibition was shown to occur through the Gs-cAMP-protein kinase A (PKA) signaling pathway.23,27 Treatment with dexamethasone, a potent glucocorticoid, increased GPR65 expression in peritoneal macrophages. Following dexamethasone treatment, there was an inhibition of TNF-α secretion in a manner dependent on increased expression of GPR65.28Another report provides an anti-inflammatory role for GPR65 in arthritis.29 Type II collagen-induced arthritis was increased in GPR65-null mice in comparison to wild-type mice. These studies taken together suggest GPR65 serves as a negative regulator in inflammation.30 However, one study provided a function for GPR65 as a positive modulator in inflammation.31 GPR65 was reported to increase eosinophil viability in the acidic microenvironment by reducing apoptosis through the cAMP pathway. As eosinophils are central in asthmatic inflammation and allergic airway disease, GPR65 may play a role in increasing asthmatic inflammation.31 On the other hand, GPR65 has shown little involvement in immune cell development. One report indicates that GPR65 knockout mice had normal immune development and function.26 Modulation of inflammation by GPR65 is complex and must be examined within each specific pathology.23

In addition to GPR65, GPR4 is also involved in the inflammatory response. Endothelial cells compose blood vessels that often penetrate acidic tissue microenvironments such as the inflammatory loci. Among the pH-sensing GPCR family, GPR4 has the highest expression in endothelial cells. Response to inflammation by vascular endothelial cells facilitates the induction of inflammatory cytokines that are involved in the recruitment of leukocytes for adherence and transmigration into inflamed tissues. Activation of GPR4 by acidosis in human umbilical vein endothelial cells, among other endothelial cell types, increased the expression of a broad range of pro-inflammatory genes including chemokines, cytokines, PTGS2, NF-κB pathway genes, and adhesion molecules.32 Moreover, human umbilical vein endothelial cells, when treated with acidic pH, increased GPR4-mediated endothelial adhesion to leukocytes.32,33 Altogether, GPR65 and GPR4 provide differential regulation of the inflammatory response through their acid sensing capabilities. GPR65 predominately demonstrates function in the inhibition of the inflammatory response whereas GPR4 activation exacerbates inflammation.

Role for the pH-sensing GPCRs in the cardiovascular system

Taken together, both GPR4 and GPR68 play roles in regulating the function of the cardiovascular system. GPR4 regulates blood vessel stability and endothelial cell function and GPR68 increases cardiomyogenic and pro-survival gene expression while also mediating aortic smooth muscle cell gene expression.

Role for the pH-sensing GPCRs in the renal system

GPR4 is expressed in the kidney cortex, isolated kidney collecting ducts, inner and outer medulla, and in cultured inner and outer medullary collecting duct cells.59 In mice deficient for GPR4, renal acid excretion and the ability to respond to metabolic acidosis was reduced.59 In response to acidosis, inner and outer medullary collecting duct cells produced cAMP, a second messenger for the Gs G-protein pathway, through the GPR4 receptor.59 In renal HEK293 epithelial cells GPR4 overexpression was found to increase the activity of PKA.60 In addition, the protein expression of H+-K+-ATPase α-subunit (HKα2) was increased following GPR4 overexpression dependent on increased PKA activity.60

GPR68 has also been reported to alter proton export of HEK293 cells by stimulating the Na+/H+exchanger and H+-ATPase.58 The activation of GPR68 by acidosis was found to stimulate this effect through a cluster of extracellular histidine residues and the Gq/PKC signaling pathway.58 In GPR68-null mice the expression of the pH-sensitive kinase Pyk2 in the kidney proximal tubules was upregulated which might compensate for GPR68 deficiency.58 Taken together, GPR4 and GPR68 may both be necessary for successful systemic pH buffering by controlling renal acid excretion.

Role for the pH-sensing GPCRs in the respiratory system

Aoki et al demonstrated that GPR68-deficient mice were resistant to asthma along with inhibiting Th2 cytokine and immunoglobulin E production.68 This study concludes that GPR68 in dendritic cells is crucial for the onset of asthmatic responses.68 Moreover, GPR65 has been implicated as having a role in respiratory disorders as it is highly expressed in eosinophils, hallmark cells for asthmatic inflammation.69 Kottyan et al showed that GPR65 increased the viability of eosinophils within an acidic environment through the cAMP pathway in murine asthma models.31 In summary, GPR68 and GPR65 play important roles in the respiratory system and asthma. GPR68 regulates gene expression in airway epithelial, smooth muscle and immune cells while GPR65 enhances the survival of airway eosinophils in response to acidosis.

Role for the pH-sensing GPCRs in the skeletal system

GPR65 has also been reported as a pH sensor in bone. GPR65 is expressed in osteoclasts and its activity may inhibit Ca2+ resorption.81 Disruption of GPR65 gene exacerbated osteoclastic bone resorption in ovariectomized mice.81 The relative bone density of GPR65-null mice was less than control mice.81 In cultured osteoclast cells from mice deficient for GPR65, the normal inhibition of osteoclast formation in response to acidosis was abrogated.81 Taken together, this data suggest that the activation of GPR65 may enhance bone density, thus the GPR65 signaling may be important for disease processes such as osteoporosis and other bone density disorders.

Role for the pH-sensing GPCRs in the endocrine system

GPR68 has also been found to modify insulin production and secretion. In GPR68 knockout mice insulin secretion in response to glucose administration was reduced when compared to wild-type mice although blood glucose was not significantly altered.84 GPR68 deficiency in this respect may reduce insulin secretion but at the same time increase insulin sensitivity. In addition, stimulation of GPR68 in islet cells by acidosis increased the secretion of insulin through the Gq/11 G-protein signaling.84

Role for the pH-sensing GPCRs in the nervous system and nociception

Acidosis causes pain by exciting nociceptors located in sensory neurons. Several types of ion channels and receptors, such as ASICs, TRPV1, and proton-sensing GPCRs, have been identified as nociceptors in response to acidosis. ASICs and TRPV act as proton-gated membrane-bound channels, which are activated by acidic pH and mediate multimodal sensory perception including nociception.8688  GPR65 activation sensitized the response of TRPV1 to capsaicin. The results suggest high accumulation of protons post inflammation may not only stimulate nociceptive ion channels such as TRPV1 to trigger pain, but also activate proton-sensing GPCRs to regulate heightened sensitivity to pain.89 Furthermore, Hang et al demonstrated GPR65 activation elicited cancer-related bone pain through the PKA and phosphorylated CREB (pCREB) signaling pathway in the rat model.90 Collectively, GPR4, GPR65, and GPR68 are all expressed in the dorsal root ganglia; GPR65 is a functional receptor involved in nociception and the nervous system by sensitizing inflammatory pain and the evocation of cancer-related bone pain.

Role for the pH-sensing GPCRs in tumor biology

The tumor microenvironment is highly heterogeneous. Hypoxia, acidosis, inflammation, defective vasculature, poor blood perfusion, and deregulated cancer cell metabolism are hallmarks of the tumor microenvironment.9193 The acidity in the tumor microenvironment is owing to the altered cancer cell metabolism termed the “Warburg Effect”. This metabolic phenotype allows the cancer cells to preferentially utilize glycolysis over oxidative phosphorylation as a primary means of energy production.94 This process occurs even in normoxic tissue environments where sufficient oxygen is available. Due to this phenomenon, the Warburg Effect is often termed “aerobic glycolysis”. This unique metabolic phenotype produces vast quantities of lactic acid, which serve as a proton source for acidification. Upon disassociation of lactic acid to one lactate molecule and one proton, the monocarboxylate transporter and proton transporters export lactate and protons into the extracellular tumor microenvironment.95 The proton-sensing GPCRs are activated by acidic pH and facilitate tumor cell modulation in response to extracellular acidification. GPR4, GPR65, and GPR68 play roles in tumor cell apoptosis, proliferation, metastasis, angiogenesis, and immune cell function.19,27,32,33,44,45,96,97

GPR4 has had conflicting reports in terms of tumor suppressing or promoting activities. One study demonstrated that GPR4 could act as a tumor metastasis suppressor, when overexpressed and activated by acidic pH in B16F10 melanoma cells, by impeding migration and invasion of tumor cells.45 GPR4 overexpression also significantly inhibited the lung metastasis of B16F10 melanoma cells in mice.45 Another study utilizing the B16F10 melanoma cell line which overexpressed GPR4 showed an increase in mitochondrial surface area and a significant reduction in membrane protrusions by quantification of 3D morphology.98 These data point to a decrease in cancer cell migration when GPR4 is overexpressed and provides another example of GPR4 as exhibiting tumor metastasis suppressor function.98 However, in another report GPR4 malignantly transformed immortalized NIH3T3 fibroblasts.99 This presents GPR4 with tumor-promoting capabilities. The conflicting reports seem to indicate the functional ability of GPR4 to act as a tumor promoter and a tumor suppressor depending on the context of certain cell types and biological systems.

Reports with GPR65 involvement in cancer cells provide evidence in favor for cancer cell survival; however, opposing evidences suggest GPR65 functions as a tumor suppressor. In the same report suggesting GPR4 is oncogenic due to GPR4 transforming immortalized NIH3T3 fibroblasts, GPR65 overexpression was able to transform the mouse NMuMG mammary epithelial cell line.99 Another group demonstrated in NCI-H460 human non-small cell lung cancer cells that GPR65 promotes cancer cell survival in an acidic microenvironment.100 Conversely, a recent study showed that GPR65 inhibited c-Myc oncogene expression in human lymphoma cells.101 Furthermore, GPR65 messenger ribonucleic acid expression was reduced by more than 50% in a variety of human lymphoma samples when compared to normal lymphoid tissues, therefore implying GPR65 has a tumor suppressor function in lymphoma.101 GPR65 has also been shown to increase glucocorticoid-induced apoptosis in murine lymphoma cells.102 These reports highlight cell type dependency and biological context for GPR65 activity as a tumor suppressor or promoter.

GPR68 also has roles in tumor biology as a potential tumor suppressor or a tumor promoter. Reports have shown that GPR68 can inhibit cancer metastasis, reduce cancer cell proliferation, and inhibit migration. One study showed that when GPR68 was overexpressed in prostate cancer cells, metastasis to the lungs, diaphragm, and spleen was inhibited.97 When GPR68 was overexpressed in ovarian cancer (HEY) cells, cellular proliferation and migration were significantly reduced, and cell adhesion to the extracellular matrix was increased.96 Another study reported GPR68 expression was critical for the tumor cell induced immunosuppression in myeloid-derived cells. This study proposed that GPR68 promotes M2 macrophage development and inhibits T-cell infiltration, and thereby facilitates tumor development.103 In summary, the biological roles of GPR4, GPR65, and GPR68 in tumor biology are complex and both tumor-suppressing and tumor-promoting functions have been reported, primarily dependent on cell type and biological milieu.

Development of small molecule modulators of the pH-sensing GPCRs

GPCRs are critical receptors for the regulation of many physiological operations. It is of little surprise that GPCRs have become a central focus of pharmaceutical development. In fact, 30%–50% of therapeutics focuses on modulating GPCR activity.104,105 In view of the diverse roles of the pH-sensing GPCRs in the context of multiple biological systems, targeting these receptors with small molecules and other modulators could serve as potential therapeutics for diseases associated with deregulated pH homeostasis. There have been recent developments in the characterization of GPR4 antagonists along with agonists for GPR65 and GPR68.29,32,50,106 The GPR4 antagonist demonstrated effectiveness in vitro to reduce the GPR4-mediated inflammatory response to acidosis in endothelial cells.32 The GPR65 agonist, BTB09089, showed in vitro effects in GPR65 activation of immune cells to inhibit inflammatory response; however, the activity of BTB09089 was not strong enough for the use in animal models in vivo.29 The GPR68 agonist, lsx, exhibited pro-neurogenic activity and induced hippocampal neurogenesis in young mice.107 It was also demonstrated that lsx suppressed the proliferation of malignant astrocytes.108 To date, however, much advancement needs to be done in development of efficacious agonists and antagonists of the pH-sensing GPCRs coupled with a capacity to target specific tissue dysfunction in the midst of systemic drug administration to optimize therapeutic effects and minimize potential adverse effects.

Concluding remarks

Cells encounter acidotic stress in many pathophysiologic conditions such as inflammation, cancer, and ischemia. Intricate molecular mechanisms, including a large array of acid/base transporters and acid sensors, have evolved for cells to sense and respond to acidotic stress. Emerging evidence has demonstrated that a family of the pH-sensing GPCRs can be activated by extracellular acidotic stress and regulate the function of multiple physiological systems (Table 1). The pH-sensing GPCRs also play important roles in various pathological disorders. Agonists, antagonists and other modulators of the pH-sensing GPCRs are being actively developed and evaluated as potential novel treatment for acidosis-related diseases.

Table 1 The main biological functions of the pH-sensing GPCRs

7.3.9 Protein amino-terminal modifications and proteomic approaches for N-terminal profiling

Lai ZW1, Petrera A2, Schilling O3.
Curr Opin Chem Biol. 2015 Feb; 24:71-9

Amino-/N-terminal processing is a crucial post-translational modification affecting almost all proteins. In addition to altering the chemical properties of the N-terminus, these modifications affect protein activation, conversion, and degradation, which subsequently lead to diversified biological functions. The study of N-terminal modifications is of increasing interest; especially since modifications such as proteolytic truncation or pyroglutamate formation have been linked to disease processes. During the past decade, mass spectrometry has played an important role in facilitating the investigation of N-terminal modifications. Continuous progress is being made in the development and application of robust methods for the dedicated analysis of native and modified protein N-termini in a proteome-wide manner. Here we highlight recent progress in our understanding of protein N-terminal biology as well as outlining present enrichment strategies for mass spectrometry-based studies of protein N-termini.


    • N-terminal acetylation, pyroglutamate formation, N-degrons and proteolysis are reviewed.• N-terminomics provide comprehensive profiling of modification at protein N-termini in a proteome-wide manner.• We outline a number of established methodologies for the enrichment of protein N-termini through positive and negative selection strategies.• Peptidomics-based approach is beneficial for the study of post-translational processing of protein N-termini.

 Introduction The life of every protein begins at the amino-terminus, also known as the N-terminus. During the initiation of mRNA translation into proteins or polypeptides, newly synthesized amino
acid chains form the N-termini and are the first to exit the ribosomes into the cytosol or the endoplasmic reticulum. The N-termini of these proteins or protein precursors often contain a signaling peptide
sequence proximal to the N-terminus, which may function as a ‘zip-code’ to direct the delivery of a protein to a cellular compartment as well as orchestrating protein maturation via different post-translational
modifications (PTMs) such as acetylation or proteolysis. These modifications often determine protein activity or stability; thus being crucial for the tight regulation of cellular homeostasis (Figure 1).
Mass spectrometry (MS) based analyses of protein N-termini, termed N-terminomics, is a promising tool to tackle these problems. In the past decade, we have witnessed significant progress in the
area of mass spectrometric investigation of post-translational modifications such as phosphorylation or glycosylation [1].  Similarly, MS-based studies of protein N-termini are gaining momentum.
Recent progress in positional proteomics using advanced MS platforms combined with a number of effective enrichment strategies has reinforced significant interest in N-terminomics.
Here we outline some of the most current highlights on proteomics-based studies on N-terminal modifications, including N-acetylation, pyroglutamate formation, proteolysis, and N-terminal degrons
(Figure 2). We also present a number of recent N-terminomic methodologies for the study of protein N-termini.

Acetylation of protein N-termini represents an abundant post-translational modification in eukaryotes, affecting nearly all cytoplasmic proteins. This  modification is catalyzed by the N-terminal
acetyltransferase (Nat) enzyme complex, which transfers an acetyl group to the N-termini of newly synthesized proteins during translation (Figure 2). Initial findings highlighted that N-terminal
acetylation protects proteins from degradation [2–4]. Recent studies however yield a more diverse picture. N-terminal acetylation may also play a role in protein delivery and localization [5–7],
protein complex formation and generation of specific degradation signals in cellular proteins via the N-degron pathway [9,10]. Loss of N-terminal acetylation through inactive acetyltransferases leads to
smaller aggregates of prion proteins [11]. In addition, N-terminal acetyltransferases have been described to also function as N-terminal proprionyltransferases [12].  Genetic mutation in the Naa10 gene,
encoding the NatA catalytic subunit, is known to cause N-terminal acetyltransferase deficient phenotypes. This genetic mutation has also been linked to X-linked disorder of infancy, causing lethality in
male infants[13]. The multifunctional roles of N-acetyltransferases as well as the importance of  N-terminal acetylation have been previously reviewed in [14]. Few MS-based studies have emerged that
specifically investigate acetylated N-termini in a proteome wide manner. The structural and functional integrity of actomyosin fibers depends on active NatB. A novel methodology determines the
extent of N-terminal acetylation in vivo through chemical, stable-isotope coded acetylation of proteins before their mass spectrometric analysis [16].

Pyroglutamate conversion of N-terminal glutamate and glutamine Many proteins and biologically active peptides exhibit an N-terminal pyroglutamic acid (pGlu) residue. This post
translational modification originates from the conversion of N-terminal glutamate and glutamine into pyroglutamic acid by glutaminyl cyclase or isoglutaminyl cyclase (Figure 2). N-terminal
pGlu influences structural stability as well as biological activity of peptides and proteins [17]. pGlu protects proteins from degradation by aminopeptidases [18] as well as regulating the
biological activity of peptide hormones, neuropeptides or chemokines [19]. Examples include thyrotropin releasing hormone (TRH), gonadotropin-releasing hormone, and the human
chemokines MCP-1 and 2. The presence of N-terminal pGlu in some amyloidogenic peptides, such as amyloid-b peptides, increases their hydrophobicity, resulting in an accelerated
aggregation [20]. Modulating the extent of N-terminal pGlu formation through pharmaceutical inhibition of glutaminyl cyclase is considered a promising strategy, for example, to
increase the degradation of inflammatory and neurotoxic peptides. Inhibition of glutaminyl cyclase has alleviated liver inflammation by destabilizing the chemokine MCP1 (CCL2) [21].
Proteolytic degradation of this promigratory chemokine by inhibiting glutaminyl cyclase was also proposed as an attractive novel strategy in preventing thyroid cancer metastasis [22].
Given the functional relevance of N-terminal pGlu in pathological conditions, an MS-based approach to profile this modification may be particularly useful.

N-terminal degrons N-terminal residues have a strong impact on protein stability and half-life. Firstly described in 1986 by Varshavsky and colleagues [25], the N-end rule pathway
has been identified in a broad range of species, from mammals to bacteria, and from yeast to plants [26]. This control of protein degradation in eukaryotes and bacteria is governed
by the formation and recognition of specific sequences at protein N-termini, called N-degrons. The main determinant of an N-degron is an N-terminal destabilizing residue. In eukaryotes,
two N-end rule pathways are being distinguished: the Ac/N-end rule pathway targets proteins through their N-terminally acetylated residues while the Arg/N-rule pathway targets
unacetylated N-terminal residues and involves N-terminal arginylation [26]. Proteolytic processing leading to new protein N-termini is increasingly recognized to play an important
role in the formation of N-degrons. In eukaryotes, N-degron mediated protein degradation occurs through the  ubiquitin–proteasome system. N degrons are recognized by E3
ubiquitin ligases called N-recognins, which induce protein ubiquitylation. Recent studies showed that the N-end rule pathway can be regulated by various mechanisms [26].
Hemin, the ferric (Fe3+) counterpart of heme, and short peptides can bind to components of the N-end rule pathway and impede their functionality [26]. Although the N-end rule
pathway has been molecularly dissected in great detail, numbers of identified physiological substrates undergoing N-end rule degradation have remained limited. A recent study
has expanded the range of substrates targeted by the Arg/N-end rule. Kim and colleagues have shown that N terminal Met followed by a hydrophobic residue functions as an N-degron
[27]. N-terminal Met followed by a small residue is typically removed by aminopeptidases in a cotranslational manner (Figure 2). However, approximately 15% of the genes in mammals
or yeast encode for an N-terminal Met followed by a larger hydrophobic residue. This specific N-degron is targeted by the Ac/N-end rule pathway when the N-terminal Met is acetylated.
The Arg/N-end rule acts instead on the non-acetylated N-terminal Met. As previously mentioned, novel N-degrons can be generated by preceding proteolysis. Piatkov and colleagues
investigated this concept for proteolytic cleavage products that occur during apoptosis [28]. They find that numerous proapoptotic fragments are short lived substrates of Arg/N-end
rule pathway, attributing to this pathway an anti-apoptotic role. Notably, the corresponding N-degron sequences are evolutionary conserved.

Figure 1 Protein N-termini are susceptible to various post-translational modification.
For a more comprehensive overview of all possible N terminal modification, see [60].

Figure 2 Examples of N-terminal mofications: acetylation, pyroglutamate conversion, proteolysis and N-degron processing via deamidation and amino acid conjugation.

Proteolytic processing of N-termini Proteolysis has long been regarded a degradation process. It is now increasingly recognized as an important posttranslational modification
with an array of proteases mediating cellular signaling via the precise processing of bioactive proteins and peptides. The study of cleavage events using N-terminomics is particularly
useful for the identification of proteolytic substrates. Proteolytic cleavage of proteins and polypeptides results in the generation of cleavage fragments with new N-termini and
C-termini. Numerous recent proteomic studies highlighted differential regulation of proteases in different disease settings. MALDI-TOF in combination with enzymatic assays
established reduced levels of dipeptidyl-peptidase (DPP)4 in the serum of patients suffering from metastatic prostate cancer [31]. Another proteomic based study,  using isotope
coded affinity tag (ICAT) labeling showed bacterial leucine aminopeptidase from Plasmodium chabaudi to be significantly upregulated in periodontal disease [32]. Mass spectrometry
was also used for the functional characterization of proteases.

7.3.10 Protein homeostasis networks in physiology and disease

Although most text books of biochemistry describe the process of protein folding to a three dimensional native state as an intrinsic property of the primary sequence, it is becoming increasingly clear that this process can go wrong in an almost infinite number of ways. In fact, many different diseases are caused by the misfolding and aggregation of certain proteins without genetic mutations in the primary sequence. An integrative view of the mechanisms that maintain protein folding homeostasis is emerging, which could be thought as a balanced and dynamic network of interconnected processes tightly regulated by a series of quality control mechanisms. This protein homeostasis network involves families of folding catalysts, co-factors under specific environmental and metabolic conditions. Maintaining protein homeostasis is particularly challenging in specialized secretory cells where the high demand for protein synthesis generates a constant source of stress that could lead to proteotoxicity.

Protein folding is assisted and monitored by diverse interconnected processes that follow a sequential pattern over time. The calnexin/calreticulin cycle ensures the proper folding of glycosylated proteins through the secretory pathway, which establishes the final pattern of disulfide bond formation through interactions with the disulfide isomerase ERp57. Coupled to this cycle is the ER-associated degradation (ERAD) pathway, which translocates terminally misfolded proteins to the cytosol for degradation by proteasomes. In addition, macroautophagy is becoming a relevant mechanism for the clearance of damaged proteins and abnormal protein aggregates through lysosomal hydrolysis, a process also referred to as ERAD-II. The folding status at the ER is constantly monitored by the Unfolded Protein Response (UPR), a specialized signaling pathway initiated by the activation of three types of stress sensors. The process underlying the surveillance of protein folding stress by the UPR is not fully understood, but it may require coupling to key folding mediators such as BiP or the direct recognition of the misfolded peptides by stress sensors. The UPR regulates genes and processs related to almost every folding step in the secretory pathway to reduce the load of misfolded proteins, including protein translation into the ER, translocation, folding, quality control, ERAD, the redox status, and many other related functions. Protein folding stress is observed in many disease conditions such as cancer, diabetes, and neurodegeneration. For example, abnormal protein aggregation and the accumulation of protein inclusions is associated with Parkinson’s and Alzheimer’s Disease, and amyotrophic lateral sclerosis. In those diseases and many others, neuronal dysfunction and disease progression correlates with the presence of a strong ER stress response; however, the direct in vivo role of the UPR in the disease process has been experimentally defined in only a few cases. Therapeutic strategies are currently being developed to increase protein folding and clearance of misfolded proteins, with the goal of alleviating ER stress.

In this issue of Current Opinion in Cell Biology we present a series of focused reviews from recognized experts in the field, that provide an overview of mechanisms underlying protein folding and quality control, and how balance of protein homeostasis is maintained in physiology and deregulated in diseases. Daniela Roth and William Balch integrate the concept of protein homeostasis networks into an interesting model termed FoldFx, showing how the interconnection between different pathways in the context of the cellular proteome determines the energetic barrier required to generate a functional folded peptide. The authors have previously proposed the term Proteostasis to refer to the set of interacting activities that maintain the health of the proteome and the organism (protein homeostasis). The ER is a central subcellular compartment for protein synthesis and quality control in the secretory pathway. Yukio Kimata and Kenji Kohno give an overview of the signaling pathways that control adaptation to ER stress and maintenance of protein folding homeostasis. The authors summarize the models proposed so far for the activation of UPR stress sensors, and discuss how this directly or indirectly relates to the accumulation of unfolded proteins in the ER lumen. Chronic or irreversible ER stress triggers cell death by apoptosis. Gordon Shore, Feroz Papa, and Scott Oakes summarize the complex signaling pathways initiating apoptosis by ER stress, where cross talk between the ER and the mitochondria play a central role. The authors focus on addressing the role of the BCL-2 protein family on the activation of intrinsic mitochondrial apoptosis pathways, highlighting different cytosolic and transcriptional events that determine the transition between adaptive responses to apoptosis programmed by the UPR to eliminate irreversibly injured cells.

Although diverse families of chaperones, foldases and co-factors are expressed at the ER, only a few protein folding networks have been well defined. However, molecular explanations for specific substrate recognition and quality control mechanisms are poorly defined. Here we present a series of reviews covering different aspects of protein maturation. Amy Lee summarizes what is known about the biology of the key ER folding chaperone BiP/Grp78, and its emerging role in diverse pathological conditions including cancer. In two reviews, David B. Williams and Linda M. Hendershot describe the best characterized mechanism of protein quality control at the ER, the calnexin cycle. In addition, they give an overview of the function of a family of ER foldases, the protein disulfide isomerases (PDIs), in folding, quality control and degradation of abnormally folded proteins. PDIs are also becoming key factors in establishing the redox tone of the ER. Riccardo Bernasconi and Maurizio Molinari overview the ERAD process and how this pathway affects the efficiency of the protein folding process at the ER and its relation to pathological conditions.

Lysosomal-mediated degradation is becoming a fundamental process for the control of the haft-life of proteins and the degradation of misfolded, aggregate prone proteins. Ana Maria Cuervo reviews the relevance of Chaperone-mediated autophagy in the selective degradation of soluble cytosolic proteins in lysosomes, and also points out a key role for Chaperone-mediated autophagy in the cellular defense against proteotoxicity. David Rubinsztein and Guido Kroemer present two reviews highlighting the emerging relevance of macroautophagy in maintaining the homeostasis of the nervous system. They also discuss the actual impact of macroautophagy in the clearance of protein aggregates related to neurodegenerative diseases, including Parkinson’s disease, amyotrophic lateral sclerosis, Huntington’s disease among others. In addition, recent evidence suggesting an actual impairment of macroautophagy as a causative factor in aging-related disorders is also discussed.

Alterations in protein homeostasis underlie the etiology of many diseases affecting the nervous system, in addition to cancer and diabetes. Fumiko Urano summarizes the impact of ER stress in β cell dysfunction and death during the progression of type 1 and type 2 diabetes, as well as in genetic forms of diabetes such as Wolfram syndrome. The occurrence of basal ER stress is observed in specialized secretory cells and organs, including plasma B cells. Roberto Sitia covers several aspects of how proteotoxic stresses physiologically contribute to regulate the biogenesis, function and lifespan of B cells, and speculates about the possible impact of ER stress in the treatment of multiple myeloma. Claudio Soto describes the specific role of calcineurin, a key phosphatase in the brain, in the occurrence of synaptic dysfunction and neuronal death in prion-related disorders. We also present provide a review summarizing the emerging role of ER stress and the UPR in most neurodegenerative diseases related to protein misfolding. We also discuss the particular mechanisms currently proposed to be involved in the generation of protein folding stress at the ER in these pathologies, and speculate about possible therapeutic interventions to treat neurodegenerative diseases.

Strategies to increase the efficiency of quality control mechanisms, to reduce protein aggregation and to enhance folding are suggested to be beneficial in the setting of diseases associated with the disruption of protein homeostasis. Finally, Jeffery Kelly overviews recent chemical and biological therapeutic strategies to restore protein homeostasis, which could be achieved by enhancing the biological capacity of the proteostasis network or through small molecule to stabilize misfolding-prone proteins. In summary, this volume ofCurrent Opinion in Cell Biology compiles the most recent advances in understanding the impact of protein folding stress in physiology and disease, and integrates a variety of complex mechanisms that evolved to maintain protein homeostasis in a dynamic way in the context of a changing environment. The biomedical applications of developing strategies to cope with protein folding stress have profound implications for the treatment of the most prevalent diseases in the human population.

7.3.11 Proteome sequencing goes deep
Advances in mass spectrometry (MS) have transformed the scope and impact of protein characterization efforts. Identifying hundreds of proteins from rather simple biological matrices, such as yeast, was a daunting task just a few decades ago. Now, expression of more than half of the estimated ∼20,000 human protein coding genes can be confirmed in record time and from minute sample quantities. Access to proteomic information at such unprecedented depths has been fueled by strides in every stage of the shotgun proteomics workflow-from sample processing to data analysis-and promises to revolutionize our understanding of the causes and consequences of proteome variation.
    • Recent MS advances have transformed the depth of coverage of the human proteome.• Expression of half the estimated human protein coding genes can be verified by MS.• MS sample preparation, instrumentation, and data analysis techniques are highlighted.

Mammalian proteomes  are complex [3]. The human proteome contains ~20,300 protein-coding genes; however, non-synonymous single nucleotide polymorphisms (nsSNPs), alternative
splicing events, and post-translational modifications (PTMs) all occur and exponentially increase the number of distinct proteoforms [4–6]. Detection of 5000 proteins in a proteomic
experiment was a considerable achievement just a few years ago [7–9]. More recently, two groups identified over 10,000 protein groups in a single experiment. Through extensive protein
and peptide fractionation (72 fractions) and digestion with multiple enzymes, Nagaraj et al. identified 10,255 protein groups from HeLa cells over 288 hours of instrument analysis [10].
A comparison with paired RNA-Seq data revealed nearly complete overlap between the detected proteins and the expressed transcripts. In that same year, a similar strategy enabled
the identification of 10,006 proteins from the U2OS cell line [11]. Kim and co-workers analyzed 30 human tissues and primary cells over 2000 LC–MS/MS experiments, resulting
in the detection of 293,000 peptides with unique amino acid sequences and evidence for 17,294 gene products [16]. Wilhelm et al. amassed a total of 16,857 LC–MS/MS experiments
from human cell lines, tissues, and body fluids. These experiments produced 946,000 unique peptides, which map to 18,097 protein coding genes [17]. Together, these two studies
provide direct evidence for protein translation of over 90% of  human genes (Figure 2). New developments in mass spectrometer technology have increased the rate at which proteomes
can be analyzed. We describe developments in sample preparation, MS instrumentation, and bioinformatics that have been key to obtaining comprehensive proteomic coverage.
Further, we consider how access to such proteomic detail will impact genomic  research.

Aurelian Udristioiu


Aurelian Udristioiu

Lab Director at Emergency County Hospital Targu Jiu

Mg²+ is critical for maintaining the positional integrity of closely clustered phosphate groups. These clusters appear in numerous and distinct parts of the cell nucleus and cytoplasm. The Mg²+ ion maintains the integrity of nucleic acids, ribosomes and proteins. In addition, this ion acts as an oligo-element with role in energy catalysis. Biological cell membranes and cell walls exhibit poly-anionic charges on the surface. This finding has important implications for the transport of ions, particularly because different membranes preferentially bind different ions. Both Mg²+ and Ca²+ regularly stabilize membranes by cross-linking the carboxylated and phosphorylated head groups of lipids.

Notable document –

Theor Biol Med Model. 2010 Jun 9;7:19.
Native aggregation as a cause of origin of temporary cellular structures needed for all forms of cellular activity, signaling and transformations.
Matveev VV1.
Cell physiologist at Institute of Cytology, Russian Academy of Sciences

According to the hypothesis explored in this paper, native aggregation is genetically controlled (programmed) reversible aggregation that occurs when interacting proteins form new temporary structures through highly specific interactions. It is assumed that Anfinsen’s dogma may be extended to protein aggregation: composition and amino acid sequence determine not only the secondary and tertiary structure of single protein, but also the structure of protein aggregates (associates). Cell function is considered as a transition between two states (two states model), the resting state and state of activity (this applies to the cell as a whole and to its individual structures). In the resting state, the key proteins are found in the following inactive forms: natively unfolded and globular. When the cell is activated, secondary structures appear in natively unfolded proteins (including unfolded regions in other proteins), and globular proteins begin to melt and their secondary structures become available for interaction with the secondary structures of other proteins. These temporary secondary structures provide a means for highly specific interactions between proteins. As a result, native aggregation creates temporary structures necessary for cell activity.”One of the principal objects of theoretical research in any department of knowledge is to find the point of view from which the subject appears in its greatest simplicity.”Josiah Willard Gibbs (1839-1903).

To date, numerous mechanisms, signal pathways, and different factors have been found in the cell. Researchers are naturally eager to find commonalities in the mechanisms of cellular regulation. I would like to propose a substantial approach to problems of cell physiology – the structural ground that produces signals and underlies the diversity of cellular mechanisms.

The methodological basis for the proposed hypothesis results from studies by the scientific schools of Dmitrii Nasonov [1] and Gilbert Ling [26], which have gained new appreciation over the last 20-30 years owing to advances in protein physics [7] in the study of properties of globular proteins, their unfolding and folding, as well as the discovery of novel states of the protein molecule: the natively unfolded and the molten globule. The key statement for the rationale of the present paper is that the specificity of interactions of polypeptide chains with each other (at the intra- and inter-molecular levels) can be provided only by their secondary structures, primarily α-helices and β-sheets.

Nasonov’s school discovered and studied a fundamental phenomenon — the nonspecific reaction of the cell to external actions [1], while works by Ling [5] and his followers allow the mechanisms of this phenomenon to be understood.

The above-mentioned cell reaction has been called nonspecific because diverse physical and chemical factors produce the same complex of structural changes in the cell: an increase in the turbidity and macroscopic viscosity of the cytoplasm and in the adsorption of hydrophobic substances by cytoplasmic proteins. It is of primary importance that the same changes also occur in the cell during its transition into the active state: muscle contraction, action potential, enhancement of secretory activity (for details, see [8]). Hence, from the point of view of structural changes, there is no fundamental difference between the result of action on the cell of hydrostatic pressure and, for instance, muscle contraction. In both cases, proteins are aggregated.

Nasonov called the cause of these changes the stages of cell protein denaturation, as the changes of properties of isolated proteins during denaturation are very similar to the changes in the cytoplasm during the nonspecific reaction. As a result, the denaturational theory of cell excitation and damage was created [1]. The structural changes of protein denaturation were unclear in Nasonov’s time. Nowadays, it is assumed that the denaturation is the destruction of the tertiary and secondary structure of a protein. Below I give two definitions, for the denaturation of natively folded (globular) proteins and for natively unfolded proteins.

A key notion in physiology is the resting state of the cell. This is implicit in the concept of the threshold character of the action of stimuli on the cell, which has played a historical role in the development of physiological science. It is the threshold that is the boundary between two states — rest and activity. But in effect, all our knowledge about cells concerns active cells, not cells in the resting state. It is in the active cell that variable changes occur that can be recorded. Nothing happens in the resting cell, so there is nothing to be recorded in it. Nevertheless, it is obvious that the resting state is the initial cell state, the starting point for all changes occurring in the cell.

What characterizes the structural aspect of the cell in the state of rest? It is only in Ling’s work [5] that I have found a clear answer to this question. The answer can be interpreted as follows: if all resting cell proteins were arranged in one line, it would turn out that most of the peptide bonds in this superpolypeptide would be accessible to solvent (water), while only a few would be included in secondary structures. When the cell is activated, the ratio between the unfolded and folded areas is changed sharply to the opposite: the proportion of peptide bonds accessible to solvent decreases markedly, whereas the proportion included in secondary structures rises significantly. These two extreme states of cell proteins, suggested by Ling, provide a basis for further consideration.

If Ling’s approach is combined with Nasonov’s theory, we obtain several interesting consequences. First of all, it is clear that proteins with maximally unfolded structures form the structural basis of resting cells because they are inactive, i.e., do not interact with other proteins or other macromolecules. The situation changes when an action on the cell exceeds the threshold: completely or partially unfolded key proteins begin to fold when new secondary protein structures are formed. Owing to these new secondary structures, the proteins become capable of reacting, i.e., intramolecular aggregation (folding of individual polypeptides into globules) and intermolecular aggregation (interaction of some proteins with others) begin. A distinguishing feature of these aggregational processes is their absolutely specific character, which is ensured by the amino acid composition, shape, and size of the secondary structures. The structures appearing have physiological meaning, so such aggregation is native and the secondary structures causing it are centers of native aggregation. Another source of secondary structures necessary for native aggregation is the molten globule.

The ability of cells to return to the initial state, the state of rest, means that native aggregation is completely reversible, and the structures appearing in the course of native aggregation are temporary and are disassembled as soon as they cease to be necessary. Native aggregation can involve both the whole cell and individual organelles, compartments, and structures, and activation of proteins is of a threshold rather than a spontaneous character.

The meaning of the proposed hypothesis of native aggregation is that the primary cause of any functional changes in cell is the appearance, as a result of native aggregation, of temporary structures, continually appearing and disintegrating during the life of the cell. Since native aggregation is initiated by external stimuli or regulatory processes and the structures appearing have a temporary character, these structures can be called signal structures.

Signal structures can have different properties: (i) they can be centers of binding of ions, molecules (solutes), and proteins; (ii) they can have enzymatic activity; (iii) they can form channels and intercellular contacts; (iv) they can serve as matrices organizing the interactions of molecules in synthetic and transport processes; (iv) they can serve as receptors for signal molecules; (v) they can serve as the basis for constructing even more complex supramolecular structures. These structures “flash” in the cell space like signal lights, perform their role, and disappear, to appear in another place and at another time. The meaning of the existence of the structural “flashes” is that during transition into the active state the cell needs new resources, functions, mechanisms, regulators, and signals. As soon as the cell changes to the resting state, the need for these structures disappears, and they are disassembled. Extreme examples of native aggregation are muscle contraction, condensation of chromosomes, the appearance of the division spindle, and interactions of ligands with receptors.

Thus, the present paper will consider the meaning and significance of native aggregation as the universal structural basis of the active cell. The basis of pathological states is the inability of the cell to return to the resting state and errors in the formation of signal structures. The presentation of native aggregation is based on three pillars: (i) reversible protein aggregation is a structural basis of cell activity (Nasonov’s School); (ii) the operation of the living cell or its individual structures can be regarded as a repetitive sequence of transitions between two states (active and resting), a key role in which belongs to natively unfolded proteins (Ling’s approach); (iii) the specificity of interactions of separate parts of a single polypeptide chain with each other (folding) or the interaction of separate polypeptide chains among themselves (self-assembly, aggregation) can be provided only by protein secondary structures.

The goal of this paper is the enunciation of principles, rather than a review of facts corresponding to these principles.

Read Full Post »

The Colors of Life Function

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

2.5.1 Type 1 Copper Proteins

The Cu(II) state of this category has an intense blue color due to a thiolate ligand
to Cu(II) charge transfer, and unusual EPR properties arising from the asymmetrical
Cu site (distorted trigonal-pyramidal). The proteins all have a low molecular
mass and have, so far, rather arbitrarily been divided into sub-groups, such as
azurins, plastocyanins, pseudoazurins, amicyanins and various other blue
proteins. Of these the azurins, amicyanins, pseudo-azurins and plastocyanins
apparently have similar copper coordination by two histidine, one cysteine and
one methionine residue. Where the function of Type I copper proteins is known,
it is invariably electron transfer. As yet the names for these proteins are all trivial
and are often derived from source, function or color. The different classes are
usually discerned on the basis of their primary and tertiary structure.

The first bacterial blue proteins to be described were called azurins. Rusticyanin is
another example of a bacterial protein. It has unusual properties with a reduction
potential of 680 mV, and is functional at pH 2. The azurins have well-defined electron
-transfer functions.

The so-called pseudo-azurins differ from the azurins in the N-terminal amino acid
sequence and the optical spectra, which resemble those of plastocyanins.

The blue proteins known as plastocyanins occur in plants, blue-green and green
algae. Their electron transfer role is well defined, i.e. from the bc1 complex
(EC to the photooxidized P-700.

Amicyanins are electron carriers between methylamine dehydrogenase and
cytochrome c, with a characteristic amino acid sequence.

Of the remaining blue proteins stellacyanin is a well- known example. Umecyanin,
plantacyanin and mavicyanin are also considered to belong to this group.
Although these proteins undergo redox reactions in vitro, their true biological
function remains unknown. Most of these proteins exhibit an unusual EPR signal
in which the copper hyperfine splitting pattern is poorly resolved. There is good
evidence that at least for stellacyanin, methionine does not function as a ligand
for copper.

2.5.2 Type 2 Copper Proteins

The copper centres in these proteins are spectroscopically consistent with square
planar or pyramidal coordination, containing oxygen and/or nitrogen ligation.
The Cu(II) is EPR active, with a ‘normal’ signal. There is no intense blue color.
This group includes the copper/zinc superoxide dismutase (EC,
dopamine b-monooxygenase (EC, galactose oxidase (EC
and the various copper-containing amine oxidases. Some members of this last
group may also contain an organic prosthetic group, such as PQQ
(see section 10), or a modified amino-acid residue.

2.5.3 Type 3 Copper Proteins

In this group a pair of copper atoms comprise a dinuclear centre, with no EPR
activity as for single Cu’s. The best known example of an enzyme containing a
single Type 3 centre is tyrosinase (catechol oxidase, EC This protein
contains a metal center which is a structural analogue of the dinuclear copper
center in hemocyanin (ref 31).

2.5.4 Multi-Copper Oxidases

In addition to the above, there are several proteins with catalytic activity that
contain Types 1, 2 and 3 centres in various stoichiometric ratios. These
include L-ascorbate oxidase (EC, laccase (EC and
ceruloplasmin (ferro-oxidase, EC, the latter two having aromatic diamine
and diphenol oxidase activity. There is growing evidence that in these proteins
the Type 2 and Type 3 copper centres are juxtaposed. Recently it has been
shown that in L-ascorbate oxidase, a trinuclear copper site is present, consisting
of a type 3 copper site, very close (3.9 Å) and possibly bridged to a type 2 copper
site (ref 32). There is a view that ceruloplasmin functions as a ferro-oxidase
and the Fe(III) produced in this reaction can then oxidize the same substrates
as laccase.

2.5.5 Copper Centres in Cytochrome Oxidase

There are two copper centres that appear to be unique. Both are present in
cytochrome-c oxidase (EC The first appears to be an isolated metal ion
and has been referred to as Cud and CuA. The second appears to be part
of a dinuclear centre with cytochrome a3. It has been referred to as Cuu,
Cua3 and CuB. At the moment the ascriptions CuA and CuB are most frequently
used; however, the recent discovery (ref 33) of a cytochrome oxidase in which
cytochrome a has been replaced by cytochrome b, leads to the recommendation
that CuB shall be referred to as Cua3.

There is a striking similarity between two of the Cu centres of N2O reductase
and CuA (ref 34, 35).

2.5.6 Molybdenum enzymes (general)

Molybdenum enzymes contain molybdenum at the catalytic center responsible
for reaction with substrate. They may be divided into those that contain
the iron-molybdenum cofactor and those that contain the pterin-molybdenum

2.5.7 Additional centers

If a molybdenum enzyme contains flavin, it may be called either a molybdenum
flavoprotein or a flavomolybdenum protein, as indicated above. Other centers
should be treated similarly, e.g. an iron-sulfur molybdenum protein.

2.5.8 Molybdenum enzymes containing the iron-molybdenum cofactor

The only enzymes at present known to belong to this group are the nitrogenases
(EC; and EC see pp 89-116 in (ref 36) and pp 91-100 in (ref 37).

2.5.9 Molybdenum enzymes containing the pterin-molybdenum cofactor

These enzymes [see pp 411-415 in (ref 36) and (ref 38)] may be divided
into those in which the molybdenum bears a cyanide-labile sulfido (or thio
– see Note 1) ligand i.e. containing the S2- ligand as Mo=S) and those
lacking this ligand. The former group includes xanthine oxidase (EC,
xanthine dehydrogenase (EC, aldehyde oxidase (EC and
purine hydroxylase (EC: see Note 2 and 3). These may be called ‘molybdenum-
containing hydroxylase’ as is widely done. Molybdenum enzymes lacking the
sulfide (thio) ligand include sulfite oxidase (EC, NAD(P)+-independent
aldehyde dehydrogenase and nitrate reductases (assimilatory and dissimilatory)

2.5.10 Molybdenum enzymes containing the pterin-molybdenum cofactor

These enzymes [see pp 411-415 in (ref 36) and (ref 38)] may be divided into those
in which the molybdenum bears a cyanide-labile sulfido (or thio – see Note 1)
ligand i.e. containing the S2- ligand as Mo=S) and those lacking this ligand. The
former group includes xanthine oxidase (EC, xanthine dehydrogenase
(EC, aldehyde oxidase (EC and purine hydroxylase. These
may be called ‘molybdenum-containing hydroxylase’ as is widely done.
Molybdenum enzymes lacking the sulfide (thio) ligand include sulfite oxidase
(EC, NAD(P)+-independent aldehyde dehydrogenase and nitrate
reductases (assimilatory and dissimilatory) (EC

2.5.11 Metal-Substituted Metalloproteins

Scientists from several areas, dealing with spectroscopy and electron-transfer
mechanisms, often use metalloproteins in which a metal at the active site has
been substituted by another metal ion, like Co, Zn, Hg, Cd. Examples are zinc-
substituted cytochromes and cobalt-substituted ferredoxins.

The names for such modified proteins are easily given by using indications
like: ‘zinc-substituted ….’. In case of multi-metal proteins, where ambiguity might
arise about which metal has been substituted, one could easily add in parentheses
the name of the metal that has been replaced, such as: cobalt- substituted [Fe]

In formulae fragments or short names one could use the following notation:
[3Fe1Co-4S]2+, cytochrome c'[Fe[arrow right]CoFe], plastocyanin[Cu
[arrow right]Hg].

Ambler, R.P. (1980) in From Cyclotrons to Cytochromes (Kaplan, N.O. &
Robinson, A., eds) Academic Press, New York

Moore, G. & Pettigrew, F.(1987) Cytochromes c, Springer-Verlag, Berlin

Bartsch, R.G. (1963) in Bacterial Photosynthesis (Gest, H., San Pietro, A. &
Vernon, L.P., ed.) p. 315, Antioch Press, Yellow Springs, Ohio.

Stiefel, E.I. & Cramer, S.P. (1985) in Molybdenum Enzymes (Spiro, T.G., ed.),
Wiley-Interscience, New York, 89-116.

Smith B.E. et al. (1988), in Nitrogen Fixation Hundred Years After (Bothe,
H., de Bruijn, F.J. & Newton, W.E., ed.), Gustav Fischer, Stuttgart, New York,

Type-2 copper-containing enzymes.
MacPherson IS1, Murphy ME.
Cell Mol Life Sci. 2007 Nov;64(22):2887-99.

Type-2  Cu sites are found in all the major branches of life and are often
involved in the catalysis of oxygen species. Four type-2 Cu protein
families are selected as model systems for review: amine oxidases,
Cu monooxygenases, nitrite reductase/multicopper oxidase, and
CuZn superoxide dismutase. For each model protein, the availability
of multiple crystal structures and detailed enzymological studies provides
a detailed molecular view of the type-2 Cu site and delineation of the
mechanistic role of the Cu in biological function. Comparison of these
model proteins leads to the identification of common properties of the
Cu sites and insight into the evolution of the trinuclear active site found
in multicopper oxidases.

Copper proteins and copper enzymes.
Cass AE, Hill HA.
Ciba Found Symp. 1980;79:71-91.

The copper proteins that function in homeostasis, electron transport, dioxygen
transport and oxidation are discussed. Particular emphasis is placed on the
role of the ligands, their type and disposition which, in conjunction with other
residues in the active site, determine the role of the copper ion. It is proposed that
copper proteins can be considered in four groups. Those in Group I contain a
single copper ion in an approximately tetrahedral environment with nitrogen and
sulphur-containing ligands. Group II proteins have a single copper ion in a square-
planar-like arrangement. Group III proteins have two copper ions in close
proximity. Group IV consists of multi-opper proteins, composed of sites
representative of the other three groups.

Such centers owe their name to the intense blue coloration of the corresponding
Cu(II) proteins. The color is particularly distinctive since the metal centers are
so optically diluted in these metalloenzymes that only intense absorption in the
visible region, resulting from symmetry allowed electronic transitions, can give
rise to conspicuous colors. In contrast, the comparatively pale blue color of normal
Cu(II)) is the result of forbidden electronic transitions between d-orbitals of
different symmetry; in Cu2+(aq) this gives a molar extinction coefficient of
10 M-1cm-1 from a broad absorption between 10,000 cm-1 and 15,000 cm-1
compared to about 3000 M-1cm-1 observed for blue Cu(II) centers.  For the
T1 centers the intense absorption is attributed to a ligand-to-metal charge
transfer between the Cu2+ and a bonded cysteinate ligand. Typically, as in
azurin or plastocyanin this occurs around 16,000 cm-1. Ceruloplasmin has
three T1 centers, and the blue absorption is at 16,400 cm-1 (610nm).

Plastocyanine geometry

around the copper Crystal structures show a very irregular ‘tetrahedral’ coordination
with two sulphurs from methionine and cysteinate, and two histidine nitrogens.
However a comparison of azurin with plastocyanin shows that the geometry
is in some ways closer to a trigonal bipyramid, with and without one extra apical
ligand, so that azurin has a weakly bound glutamine oxygen, and plastocyanine
does not. The T1 coppers in caruloplasmin are in plastocyanine-type domains.
Each of these are coordinated to two histidines and a cysteine, in two of the T1
domains there is also a methionine residue, the third T1 domain has a leucine
residue which may only have a van der Waals type contact with the copper.

T1 copper centers are functional in the reversible electron transfer:

Cu2+ + e-   =   Cu+

The strongly distorted geometry represents a compromise (entactic-state
situation) between d10 Cu(I), with its preferred tetrahedral or trigonal
coordination through soft sulfur ligands, and d9 Cu(II) with preferential
square planar or square pyramidal geometry and nitrogen ligand
coordination.   This irregular, high energy arrangement at the metal
center resembles the transition-state geometry between the tetrahedral
and square planar equilibrium configurations of the two oxidation states
involved and permits enhanced rates of electron transfer. The potential
range for proteins with T1 copper centers runs from 180 mV in
stellacyanin to 680 mV in rusticyanin.

Zinc proteins: enzymes, storage proteins, transcription factors, and replication
Coleman JE.
Annu Rev Biochem. 1992;61:897-946.

In the past five years there has been a great expansion in our knowledge of
the role of zinc in the structure and function of proteins. Not only is zinc
required for essential catalytic functions in enzymes (more than 300 are known
at present), but also it stabilizes and even induces the folding of protein
subdomains. The latter functions have been most dramatically illustrated
by the discovery of the essential role of zinc in the folding of the DNA-binding
domains of eukaryotic transcription factors, including the zinc
finger transcription factors, the large family of hormone receptor proteins,
and the zinc cluster transcription factors from yeasts. Similar functions are
highly probable for the zinc found in the RNA polymerases and the zinc-
containing accessory proteins involved in nucleic acid replication. The rapid
increase in the number and nature of the proteins in which zinc functions
is not unexpected since zinc is the second most abundant trace metal found in
eukaryotic organisms, second only to iron. If one subtracts the amount of iron
found in hemoglobin, zinc becomes the most abundant trace metal found
in the human body.

Zinc Coordination Spheres in Protein Structures
ACS ChemWorx
Mikko Laitaoja , Jarkko Valjakka , and Janne Jänis
Inorg. Chem., 2013, 52 (19), pp 10983–10991
Sept 23, 2013

A statistical analysis in terms of zinc coordinating amino acids, metal-to-ligand
bond lengths, coordination number, and structural classification was performed,
revealing coordination spheres from classical tetrahedral cysteine/histidine binding
sites to more complex binuclear sites with carboxylated lysine residues. According
to the results, coordination spheres of hundreds of crystal structures in the PDB
could be misinterpreted due to symmetry-related molecules or missing electron
densities for ligands.

Protein-folding location can regulate manganese-binding versus copper- or
Tottey S, Waldron KJ, Firbank SJ, Reale B, Bessant C, Sato K, Cheek TR, et al.
Nature. 2008 Oct 23;455(7216):1138-42.

Metals are needed by at least one-quarter of all proteins. Although metallo-
chaperones insert the correct metal into some proteins, they have not been
found for the vast majority, and the view is that most metalloproteins acquire
their metals directly from cellular pools. However, some metals form more
stable complexes with proteins than do others. For instance, as described
in the Irving-Williams series, Cu(2+) and Zn(2+) typically form more stable
complexes than Mn(2+). Thus it is unclear what cellular mechanisms manage
metal acquisition by most nascent proteins. To investigate this question, we
identified the most abundant Cu(2+)-protein, CucA (Cu(2+)-cupin A), and the
most abundant Mn(2+)-protein, MncA (Mn(2+)-cupin A), in the periplasm of
the cyanobacterium Synechocystis PCC 6803. Each of these newly identified
proteins binds its respective metal via identical  ligands within a cupin fold.
Consistent with the Irving-Williams series, MncA only binds Mn(2+) after
folding in solutions containing at least a 10(4) times molar excess of Mn(2+)
over Cu(2+) or Zn(2+). However once MncA has bound Mn(2+), the metal
does not exchange with Cu(2+). MncA and CucA have signal peptides for
different export pathways into the periplasm, Tat and Sec respectively. Export
by the Tat pathway allows MncA to fold in the cytoplasm, which contains only
tightly bound copper or Zn(2+) (refs 10-12) but micromolar Mn(2+) (ref. 13). In
contrast, CucA folds in the periplasm to acquire Cu(2+). These results reveal
a mechanism whereby the compartment in which a protein folds overrides its
binding preference to control its metal content. They explain why the cytoplasm
must contain only tightly bound and buffered copper and Zn(2+).

Predicting copper-, iron-, and zinc-binding proteins in pathogenic species of the
Paracoccidioides genus
GB Tristão, L do Prado Assunção, LPA dos Santos, CL Borges, MG Silva-Bailão,
CM de Almeida Soares, G Cavallaro and AM Bailão*
Front. Microbiol., 9 Jan 2015

Approximately one-third of all proteins have been estimated to contain at least
one metal cofactor, and these proteins are referred to as metalloproteins. These
represent one of the most diverse classes of proteins, containing metal ions that
bind to specific sites to perform catalytic, regulatory and structural functions.
Bioinformatic tools have been developed to predict metalloproteins encoded by
an organism based only on its genome sequence. Its function and the type of
metal binder can also be predicted via a bioinformatics approach.  Paracoccidioides
complex includes termodimorphic pathogenic fungi that are found as saprobic
mycelia in the environment and as yeast, the parasitic form, in host tissues. They
are the etiologic agents of Paracoccidioidomycosis, a prevalent systemic mycosis
in Latin America. Many metalloproteins are important for the virulence of several
pathogenic microorganisms. Accordingly, the present work aimed to predict the
copper, iron and zinc proteins encoded by the genomes of three phylogenetic species
of Paracoccidioides (Pb01, Pb03, andPb18). The metalloproteins were identified
using bioinformatics approaches based on structure, annotation and domains. Cu-,
Fe-, and Zn-binding proteins represent 7% of the total proteins encoded by
Paracoccidioides spp. genomes. Zinc proteins were the most abundant metallo-
proteins, representing 5.7% of the fungus proteome, whereas copper and iron
proteins represent 0.3 and 1.2%, respectively. Functional classification revealed that
metalloproteins are related to many cellular processes. Furthermore, it was observed
that many of these metalloproteins serve as virulence factors in the biology of the
fungus. Thus, it is concluded that the Cu, Fe, and Zn metalloproteomes of the
Paracoccidioides spp. are of the utmost importance for the biology and virulence
of these particular human pathogens.

Zinc finger proteins: new insights into structural and functional diversity
John H Laity, Brian M Lee, Peter E Wright
Current Opinion in Structural Biology Feb 2001; 11(1): 39–46

Zinc finger proteins are among the most abundant proteins in eukaryotic genomes.
Their functions are extraordinarily diverse and include DNA recognition, RNA
packaging, transcriptional activation, regulation of apoptosis, protein folding
and assembly, and lipid binding. Zinc finger structures are as diverse as their
functions. Structures have recently been reported for many new zinc finger
domains with novel topologies, providing important insights into structure/function
relationships. In addition, new structural studies of proteins containing the
classical Cys2His2 zinc finger motif have led to novel insights into mechanisms
of DNA binding and to a better understanding of their broader functions in
transcriptional regulation.

Zinc Finger Proteins

Zinc finger (ZnF) proteins are a massive, diverse family of proteins that serve a
wide variety of biological functions. Due to their diversity, it is difficult to come up
with a simple definition of what unites all ZnF proteins; however, the most common
approach is to define them as all small, functional domains that require coordination
by at least one zinc ion (Laity et al., 2001). The zinc ion serves to stabilize the
integration of the protein itself, and is generally not involved in binding targets.
The “finger” refers to the secondary structures (α-helix and β-sheet) that are
held together by the Zn ion. Zinc finger containing domains typically serve
as interactors, binding DNA, RNA, proteins or small molecules (Laity et al., 2001).

ZnF Protein Families

Cys2His2 was the first domain discovered (also known as Krüppel-type). It was
initially discovered as a repeating domain in the IIIA transcription factor in
Xenopus laevis (Brown et al., 1985; Miller et al., 1985). IIIA has nine repeats
of the 30 amino acids that make up the Cys2His2 domain. Each domain forms
a left-handed ββα secondary structure, and coordinates a Zn ion between
two cysteines on the β-sheet hairpin and two histidines in the α-helix, hence
the name Cys2His2 (Lee et al., 1989). These resides are highly conserved,
as well as a general hydrophobic core that allows the helix to form. The other
residues can show great sequence diversity (Michael et al., 1992). Cys2His2
zinc fingers that bind DNA tend to have 2-4 tandem domains as part of a
larger protein. The residues of the alpha helices form specific contacts with a
specific DNA sequence motif by “reading” the nucleotides in major groove
of DNA (Elrod-Erickson et al., 1996; Pavletich and Pabo, 1991). Cys2His2
proteins are the biggest group of transcription factors in most species. Non-
DNA binding proteins can have much more flexible tertiary structure.
Examples of Cys2His2 proteins include the Inhibitor of Apoptosis (IAP) family
of proteins and the CTFC transcription factor.

Treble clef fingers are a very diverse group of ZnF protiens both in terms of
structure and function. What makes them a family is a shared fold at their core
that looks a little like a musical treble clef, especially if you squint (Grishin,
2001). Most treble clef finger motifs have a β hairpin, a variable loop region,
a β hairpin, and an α helix. The “knuckle” of the β hairpin and the α helix contain
the Cys-x-x-Cys sequence necessary to coordinate the Zn ion. Treble clef
fingers often form the core of protein structures, for example the L24E and
S14 ribosomal proteins and the RING finger family.

Zinc ribbons are a little less structurally complex than the other two major groups.
Zinc ribbons contain two zinc knuckles, often β hairpins, coordinating a zinc ion via
a two Cys residures separated by 2-4 other residues on one knuckle, and a Cys-x-x-
Cys on the other (Hahn and Roberts, 2000). Examples of zinc ribbon-containing
proteins include the basal transcription factors TFIIS and TFIIB that for a complex
with RNAPII to bind DNA, and the Npl4 nuclear core protein that uses a zinc ribbon
to bind ubiquitin (Alam et al., 2004). Cys2His2, treble clef fingers, and zinc ribbons
form the majority of zinc fingers, but there are several other smaller groups that
don’t fit neatly into these three. Green fluorescent protein as a marker for gene

Metallothionein proteins expression, copper and zinc concentrations, and lipid
peroxidation level in a rodent model for amyotrophic lateral sclerosis
E Tokuda, Shin-Ichi Ono,  K Ishige, A Naganuma, Y Ito, T Suzuki
Toxicology Jan 2007; 229(1–2): 33–41

It has been hypothesized that copper-mediated oxidative stress contributes to the
pathogenesis of familial amyotrophic lateral sclerosis (ALS), a fatal motor neuron
disease in humans. To verify this hypothesis, we examined the copper and zinc
concentrations and the amounts of lipid peroxides, together with that of the
expression of metallothionein (MT) isoforms in a mouse model [superoxide
dismutase1 transgenic (SOD1 Tg) mouse] of ALS. The expression of MT-I and
MT-II (MT-I/II) isoforms were measured together with Western blotting, copper
level, and lipid peroxides amounts increased in an age-dependent manner in the
spinal cord, the region responsible for motor paralysis. A significant increase was
already seen as early as 8-week-old SOD1 Tg mice, at which time the mice had not
yet exhibited motor paralysis, and showed a further increase at 16 weeks of age,
when paralysis was evident. Inversely, the spinal zinc level had significantly
decreased at both 8 and 16 weeks of age. The third isoform, the MT-III level,
remained at the same level as an 8-week-old wild-type mouse, finally increasing
to a significant level at 16 weeks of age. It has been believed that a mutant SOD1
protein, encoded by a mutant SOD1, gains a novel cytotoxic function while
maintaining its original enzymatic activity, and causes motor neuron death
(gain-of-toxic function). Copper-mediated oxidative stress seems to be a probable
underlying pathogenesis of gain-of-toxic function. Taking the above current
concepts and the classic functions of MT into account, MTs could have a disease
modifying property: the MT-I/II isoform for attenuating the gain-of-toxic function
at the early stage of the disease, and the MT-III isoform at an advanced stage.

Prion protein expression level alters regional copper, iron and zinc content in
the mouse brain
MJ Pushie,  IJ Pickering, GR Martin, S Tsutsui, FR Jirik and GN George
Metallomics, 2011,3, 206-214

The central role of the prion protein (PrP) in a family of fatal neurodegenerate
diseases has garnered considerable research interest over the past two decades.
Moreover, the role of PrP in neuronal development, as well as its apparent role
in metal homeostasis, is increasingly of interest. The host-encoded form of the
prion protein (PrPC) binds multiple copper atoms via its N-terminal domain
and can influence brain copper and iron levels. The importance of PrPC to the
regulation of brain metal homeostasis and metal distribution, however, is not
fully understood. We therefore employed synchrotron-based X-ray fluorescence
imaging to map the level and distributions of several key metals in the brains of
mice that express different levels of PrPC. Brain sections from wild-type, prion
gene knockout (Prnp−/−) and PrPC over-expressing mice revealed striking
variation in the levels of iron, copper, and even zinc in specific brain regions as
a function of PrPC expression. Our results indicate that one important function
of PrPC may be to regulate the amount and distribution of specific metals within
the central nervous system. This raises the possibility that PrPC levels, or its
activity, might regulate the progression of diseases in which altered metal
homeostasis is thought to play a pathogenic role such as Alzheimer’s,
Parkinson’s and Wilson’s diseases and disorders such as hemochromatosis.

Zinc & Copper Imbalances: Immense Biochemical Implications
Mar 27, 2013 by Michael McEvoy

The status of zinc and copper levels may have profound implications for
many people. Much has been written about the significance of these two
trace elements for many, many years. Many health conditions may be
directly caused by abnormal zinc and copper levels.

With all of the recent attention given to methylation status, gene mutations,
MTHFR, and the associated neurological and mental/behavioral disorders
that may ensue, zinc and copper status remains a pivotal ratio in these regards.

While zinc toxicity and copper deficiency are possible, the subject of this
article is on the more common imbalance: copper toxicity and zinc deficiency.

The Physiological Roles Of Zinc & Copper

Zinc and copper are antagonists. The balance between these two trace
elements is an example of the effects of biological dualism. While zinc
toxicity is possible, far more common is zinc deficiency and copper toxicity.
Both zinc and copper play essential roles in the body, and there can be a
number of causes for why imbalances ensue.

It may be easier to identify the roles that zinc doesn’t play in the body,
than the roles it does play. Zinc is an essential trace element that activates
several hundred enzymatic reactions. These reactions are fundamental
to life and biological activity. Some of the activities that zinc are involved in:

  • DNA & RNA synthesis
  • Gene expression
  • Nervous system function
  • Immune function & immune signaling such as cell
  • Neuronal transmission
  • Brain function
  • Zinc possesses powerful anabolic activities in the cells
  • Formation of zinc proteins known as “zinc fingers”
  • Zinc is essential for blood clotting and platelet formation
  • Zinc is involved in Vitamin A synthesis
  • Folate is made available through zinc enzyme reactions
  • Along with copper, Zinc makes up the antioxidant
    system, ZnCu superoxide dismutase
  • Steroidal hormone synthesis
  • Growth & development of children
  • Testosterone and semen formation
  • The highest concentration of zinc is found in the
    male prostate gland

Copper is an essential trace element serving many important functions
as well. However, copper is well documented to induce several toxic effects
in the body, when elevated. Because copper is a pro-oxidant when free and
unbound, it can quickly generate free radicals.

The major sources for copper toxicity are: exposure to industrial forms
of copper such as copper pipes, copper cookware, birth control, exposure
to copper-based fungicides. Diets high in copper and low in zinc may play
a role in copper toxicity. Pyrrole disorder, which causes depletion of zinc,
may result in elevated levels of copper.

Some of the essential roles copper plays in the body:

  • Connective tissue formation
  • ATP synthesis
  • Iron metabolism
  • Brain health via neurotransmitter synthesis
  • Gene transcription
  • Synthesis of the antioxidant superoxide dismutase
  • Skin pigmentation
  • Nerve tissue: myelin sheath formation
  • Copper tends to rise when estrogen is dominant

Perhaps one of the first reports that zinc and copper imbalances play
a role in human health and disease was their detection in mental
disorders made by Carl Pfeiffer, MD, PhD. Dr. Pfeiffer identified a
condition known as pyrrole disorder, sometimes referred to as
pyrroluria or “mauve factor”.

As it turns out, pyrrole disorder is a major biochemical imbalance
in many people with chronic illnesses such as chronic Lyme disease,
autism, schizophrenia, depression, bi-polar, and chronic fatigue
syndrome. Pyrroles are a byproduct of hemoglobin synthesis.
Apparently, some individuals are more predisposed towards producing
higher amounts of pyrroles. When pyrroles are excessive, they irreversibly
bind to zinc and vitamin B6, causing their excretion. Consequently,
it is common that once zinc levels become depleted, copper levels tend to rise.

Copper Toxicity

Problems associated with copper toxicity include: pyrrole disorder,
estrogen dominance, schizophrenia, depression, anxiety disorder,
chronic fatigue, migraines, liver toxicity, thyroid conditions, chronic
candida yeast infections, PMS, to name a few. Some research has
even implicated copper toxicity with Alzheimer’s Disease and with
cardiovascular disease. Perhaps one of the primary mechanisms
through which copper toxicity can damage tissues is through its
initiation of oxidative stress and free radical formation. Free copper
ions that are not bound to copper proteins such as ceruloplasmin,
are pro-oxidants, and are highly reactive.

Empirical research from clinicians, indicates that there are different
types of copper imbalances. For example, if there is a lot of free,
unbound copper present, this may cause a situation of nutritive
copper deficiency. Another copper imbalance is when high pyrroles
depress zinc levels, and copper levels concomintantly rise. If high
pyrroles are present, B6 will also be lost in high amounts. In a general
but very real sense, all forms of copper excess will affect zinc status,
due to the dualistic nature of zinc and copper.

Copper & Estrogen

It has been known for many years that copper can cause a rise in
estrogen, and conversely estrogen may raise copper. Estrogen
dominance has been extensively studied in its role in breast
cancer development. One possible, critical role that can cause
estrogen to become carcinogenic, is through its oxidation induced by
Once oxidized, estrogen forms volatile hydroxyl radicals and
the associated DNA damage and “mutagenesis”.

Zinc Deficiency

As mentioned previously, pyrrole disorder will directly depress
zinc status, causing high levels of its excretion. When zinc is
lost, copper rises. Because of their essential roles in neuro-
transmitter synthesis, low zinc and high copper levels can
directly effect cognition, behavior and thought processes.
Zinc has been studied in biochemical reactions involving
calcium-driven, synaptic neurotransmission, as well as in
glutamate/GABA balance and with limbic brain function.

Zinc & Reproduction

Zinc is essential for steroidal hormone synthesis, and is a
well known catalyst for testosterone synthesis, as well as
leutinizing hormone. Zinc has demonstrated its ability to
prevent miscarriage and toxicity during pregnancy. The male
prostate gland reportedly contains the highest concentration
of zinc in the body.

Zinc & Brain Function

Much attention has been given to excitotoxicity, such as the
effects induced by MSG (monosodium glutamtate). Excess
stimulation of the excitatory neurotransmitter glutamate,
may cause severe physical and psychological reactions in
certain individuals. Zinc has been studied for its ability to
enhance GABA 
(glutamate’s antagonistic neurotransmitter)
activity and to suppress excess glutamate.

Studies on mice demonstrated that when depleted of zinc
for two weeks, the mice developed seizures, most likely due
to GABA deficiencies and glutamate excess.

There is an emerging body of evidence that demonstrates
that Alzheimer’s disease may involve copper toxicity and
zinc deficiency. Not only can excess copper cause zinc
depletion, but so can excess lead.

The hippocampus, a major part of the limbic brain, records
memories and is responsible for processing meaningful
experiences. Numerous studies site that if hippocampal
cells are deprived of zinc, the hippocampal cells die. In
addition to hippocampus cell death induced by zinc
deprivation, the amygdala, the other major limbic gland
experiences cell death as well, when deprived of zinc.

Green Fluorescent Protein

Chalfie M, Tu Y, Euskirchen G, Ward WW, Prasher DC.
Science. 1994 Feb 11;263(5148):802-5.

A complementary DNA for the Aequorea victoria green fluorescent protein (GFP)
produces a fluorescent product when expressed in prokaryotic (Escherichia coli)
or eukaryotic (Caenorhabditis elegans) cells. Because exogenous substrates and
cofactors are not required for this fluorescence, GFP expression can be used
to monitor gene expression and protein localization in living organisms.

The green fluorescent protein (GFP) is a protein composed of 238 amino acid
residues (26.9 kDa) that exhibits bright green fluorescence when exposed
to light in the blue to ultraviolet range. Although many other marine organisms
have similar green fluorescent proteins, GFP traditionally refers to the protein
first isolated from the jellyfish Aequorea victoria. The GFP from A. victoria
has a major excitation peak at a wavelength of 395 nm and a minor one at
475 nm. Its emission peak is at 509 nm, which is in the lower green portion
of the visible spectrum. The fluorescence quantum yield (QY) of GFP is 0.79.
The GFP from the sea pansy (Renilla reniformis) has a single major excitation
peak at 498 nm.

In cell and molecular biology, the GFP gene is frequently used as a reporter of
expression. In modified forms it has been used to make biosensors, and many
animals have been created that express GFP as a proof-of-concept that a gene
can be expressed throughout a given organism. The GFP gene can be introduced
into organisms and maintained in their genome through breeding, injection with a
viral vector, or cell transformation. To date, the GFP gene has been introduced
and expressed in many Bacteria, Yeast and other Fungi, fish (such as zebrafish),
plant, fly, and mammalian cells, including human. Martin Chalfie, Osamu Shimomura,
and Roger Y. Tsien were awarded the 2008 Nobel Prize in Chemistry on 10 October
2008 for their discovery and development of the green fluorescent protein.

In Aequorea victoria a protein called aequorin releases blue light upon binding
with calcium. This blue light is then totally absorbed by the GFP, which in turn
gives off the green light as in the animation below.

In 1994 GFP was cloned. Now GFP is found in laboratories all over the world where
it is used in every conceivable plant and animal. Flatworms, algae, E. coli and pigs
have all been made to fluoresce with GFP.

The importance of GFP was recognized in 2008 when the Nobel Committee awarded
Osamu Shimomura, Marty Chalfie and Roger Tsien the Chemistry Nobel Prize ”
for the discovery and development of the green fluorescent protein, GFP.”

Why is it so popular? Well, I like to think of GFP as the microscope of the twenty-
first century. Using GFP we can see when proteins are made, and where they can go.
This is done by joining the GFP gene to the gene of the protein of interest so that
when the protein is made it will have GFP hanging off it. Since GFP fluoresces, one
can shine light at the cell and wait for the distinctive green fluorescence associated
with GFP to appear.

A variant of yellow fluorescent protein with fast and efficient maturation for
cell-biological applications
T Nagai, K Ibata, E Sun Park, M Kubota, K Mikoshiba & A Miyawaki
Nature Biotechnology 20, 87 – 90 (2002)

The green fluorescent protein (GFP) from the jellyfish Aequorea victoria
has provided a myriad of applications for biological systems. Over the last
several years, mutagenesis studies have improved folding properties of GFP.
However, slow maturation is still a big obstacle to the use of GFP variants for
visualization. These problems are exacerbated when GFP variants are expressed
at 37°C and/or targeted to certain organelles. Thus, obtaining GFP variants that
mature more efficiently is crucial for the development of expanded research
applications. Among Aequorea GFP variants, yellow fluorescent proteins (YFPs)
are relatively acid-sensitive,and uniquely quenched by chloride ion (Cl−)3. For
YFP to be fully and stably fluorescent, mutations that decrease the sensitivity
to both pH and Cl− are desired. Here we describe the development of an
improved version of YFP named “Venus”. Venus contains a novel mutation,
F46L, which at 37°C greatly accelerates oxidation of the chromophore, the rate-
limiting step of maturation. As a result of other mutations, F64L/M153T/
V163A/S175G, Venus folds well and is relatively tolerant of exposure
to acidosis and Cl−. We succeeded in efficiently targeting a neuropeptide
Y-Venus fusion protein to the dense-core granules of PC12 cells. Its secretion
was readily monitored by measuring release of fluorescence into the medium.
The use of Venus as an acceptor allowed early detection of reliable signals of
fluorescence resonance energy transfer (FRET) for Ca2+ measurements in brain
slices. With the improved speed and efficiency of maturation and the increased
resistance to environment, Venus will enable fluorescent labelings that were not
possible before.

Rhodopsin-like Protein from the Purple Membrane of Halobacterium halobium
Nature New Biology 29 Sep 1971; 233, 149-152  |

HALOPHILIC bacteria require high concentrations of sodium chloride and lower
concentrations of KCl and MgCl2 for growth. The cell membrane dissociates into
fragments of varying size when the salt is removed1. One characteristic fragment—
termed the “purple membrane” because of its characteristic deep purple colour—
has been isolated in relatively pure form from Halobacterium halobium. We can
now show that the purple colour is due to retinal bound to an opsin-like protein,
the only protein present in this membrane fragment.


Stoeckenius, W. , and Rowen, R. , J. Cell Biol., 34, 365 (1967).

Stoeckenius, W. , and Kunau, W. H. , J. Cell Biol., 38, 337 (1968).

Blaurock, A. E. , and Stoeckenius, W. , Nature New Biology, 233, 152 (1971).

Sehgal, S. N. , and Gibbons, N. E. , Canad. J. Microbiol., 6, 165 (1960).

Kelly, M. , Norgård, S. , and Liaach-Jensen, S. , Acta Chem. Scand., 2A, 2169 (1970).

Shapiro, A. L. , Vinnela, E. , and Maizel, jun., J. V. , Biochem. Biophys. Res.
Commun., 28, 815 (1967).

The monomerization of the Purple protein, a member of the GFP-family
Corning, Brooke

Green fluorescent protein (GFP) has been used extensively since its discovery
in the 1960s to report and visualize gene expression. For years it has been the only
known naturally occurring fluorescent pigment that is encoded by a single gene,
making it extremely useful in various fields of biology, because the expression of
this gene directly leads to the appearance of the fluorescent green color. Recently,
however, many more proteins with similar properties to GFP, and available in a
variety of colors, have been isolated from the class of marine organisms called
Anthozoa, which includes the corals. This increase in the availability of colored
proteins in GFP family in turn has expanded the number of available biotech-
nology applications. However, some of these newly discovered GFP-like
proteins do not have wild-type forms that readily allow for the creation of
fusion proteins, particularly because of oligomerization. It is widely accepted
that almost all members of the GFP-family form dimers or tetramers in their
functional forms.

This study investigates a GFP-ike protein, Purple, isolated from two species,
Galaxea fascicularis and Montipora efflorescens. Purple protein forms oligomers
when expressed, which would then interfere with the normal expression of a  protein
to be tagged in gene fusion experiments. We selectively mutated 3 amino acids,
which we believed were responsible for oligomerization in Purple. These 3
residues were chosen based on sequence similarities to a very similar protein,
a mutant form of the Rtms5 chromoprotein from Montipora efflorescens. While
we had hoped that the resulting triple-mutant Purple protein would form
monomers in vivo while retaining its purple coloration, this turned out to
be incorrect. The resulting mutants had lost their ability to turn purple. However,
we also determined that we had successfully changed the oligomerization
state of Purple by examining the relative molecular mass of one our
mutant proteins, which turned out to be half the size of the original
purple protein. It is possible that by adding additional mutations in
the future, the original spectral properties could be recovered. If
successful, this would further expand the utility of the GFP family.

Rhodopsin, also known as visual purple, from Ancient Greek ῥόδον
(rhódon, “rose”), due to its pinkish color, and ὄψις (ópsis, “sight”), is
a light-sensitive receptor protein. It is a biological pigment in photo-
receptor cells of the retina. Rhodopsin is the primary pigment found
in rod photoreceptors. Rhodopsins belong to the G-protein-coupled
receptor (GPCR) family. They are extremely sensitive to light, enabling
vision in low-light conditions. Exposed to light, the pigment
immediately photobleaches, and it takes about 45 minutes to regenerate
fully in humans. Its discovery was reported by German physiologist
Franz Christian Boll in 1876.

Read Full Post »

Older Posts »