Molecular Biology

Rna Surveillance Growth Disorders Evolution Chondrocyte Psach Chondrodysplasia



Tweet
Jeffrey Graf's image for:
"Rna Surveillance Growth Disorders Evolution Chondrocyte Psach Chondrodysplasia"
Caption: 
Location: 
Image by: 
©  

Introduction

 In molecular biology, there has long been what has been referred to as the "central dogma", which holds  that DNA stores genomic information, DNA is transcribed, or copied to messenger RNA which is in turn used by molecular machines composed of RNA called ribosomes which use messenger RNA as templates to manufacture sequences of amino acids called proteins. It is the tradition of the central dogma that it is proteins that are the work-horses of the cell. As long as decades ago, biologists found that they could introduce nucleotide sequences that were complementary to messenger RNA ( mRNA ) of  a single gene, and thereby stop the expression of that gene.

Perhaps surprisingly, from the standpoint of "central-dogma" principles, complementary sequences of mRNA not only disrupt expression of a gene, or translation into protein, but they disrupt transcription of the gene or production of the mRNA. This implies that a message has been sent in the other direction, that is, back to the DNA. This process is referred to as RNA interference, or RNAi and may be mediated by small interfering RNA transcripts, or siRNA. There has been an explosion of research both on the mechanism of RNAi and it's use as a research tool. One particularly noteworthy aspect was the discovery of argonaut proteins [1], and in particular, that when these proteins were deleted or blocked, the shutdown of transcription by RNAi could also be prevented in mammalian cells. [2]

 Along with the discovery of Argonaut proteins and their associated complexes, that have the ability to regulate transcription based upon the results of downstream processes in the cell, a new phrase entered the molecular biology lexicon, which is "defense of the genome". It has long been known that DNA damage or infection by virus could trigger apoptosis, but new studies are showing that even genomic sequences, if they are "errant" can be recognized. At present, these studies involve Argonaut proteins, associated micro RNA's and have predominantly been observed in germline cells[3].

 In this paper, I will discuss further the hypothesis of "defending the genome", but from a different perspective than current studies more directly related to the details of small RNA. In specific, I will advance the hypothesis that something called "RNA surveillance" assures that an organisms genome will be stable not only from DNA damage and invading pathogens, but from small random mutations that might take place in the course of DNA replication. This hypothesis might propose to answer the flip side of the evolution question. We know that new species arise, but how do our "living fossils" remain essentially constant  since they first entered the fossil record.

The hypothesis is not of purely theoretical interest, but medical. Pediatric institutions around the world are in the process of turning their attention to something called "RNA surveillance". I believe (hypothesize)  that RNA surveillance and RNA interference as mediated Argonaut type proteins will converge to become the same DNA silencing process. The benefit of hastening this convergence is that small RNA researchers already know many of the things that pediatric genetic laboratories wish to find out.

Some Personal Background

In the summer of 1999 I had the opportunity to visit Harbor  Branch Oceanographic Institution and take a course on marine  finfish aquaculture with Dr Daniel Benetti.  The course  that convened a would  be historic in  the budding field of aquaculture. 

The purpose of the two week seminar was to introduce the latest techniques of fin fish aquaculture  that had been developed at Harbor Branch Oceanographic Institution in Fort Pierce, Florida. Due to declining fisheries located around the state of Florida, Harbor Branch had taken the initiative do develop farming techniques to replace seafood that could no longer be produced at customary levels by the surrounding sea. ACTED,  the institutions aquaculture division was in charge of developing new protocols for the farming of snapper, flounder and grouper among others, and Dr. Daniel Benetti was offering interested community members the opportunity to benefit from that  research.

The attendees came from around the country. including Brian O'Hanlon and Joseph Ayvazian, who would latter from Snapperfarm. From that collaboration between Benetti and O'Hanlon would emerge the flagship of sustainable aquaculture, a venture that would pioneer open sea farming,not just as a concept demonstration, but as an operating business that would be featured in  National Geographic as well as CNN Money. As exciting as it was to witness this pivotal partnership form and develop I became more interested in one of the key biological observations that Benetti and his colleagues had made.

An important part of completing the cycle in sustainable aquaculture is maintenance of brood stock, spawning, and rearing hatchlings. For each new species for which an aquaculture protocol is to be developed, the nature and behavior of the hatchlings must be observed and detailed, to see if it is feasible to recreate the necessary conditions in the laboratory. The reproduction cycle involves generation of eggs, which hatch to become yolk sack fry, and as the yolk sack is used up, they begin their first food. One peculiarity that Benetti and Grube noticed was that with each brood there seemed to be a number of fry which failed to thrive, and in fact seemed programmed for defeat. Since their measure of success was the yield, or percent of eggs that they could raise to maturity, they were troubled to see the "athletes" actually consume their siblings that seemed to be on a different growth program. Benetti and partners labeled the peculiar observations from their hatchery " athletes and runts ", the runts being the fry and larva that failed to thrive, even when conditions provided seemed to be satisfactory for their siblings. There must be some evolutionary advantage underlying this consistent observation.  Could this be a way in which nature provides food for the "athletes" at the expense of overall yield? The issue would be set aside as an as the pair of researchers proceeded to blaze a trail of breakthroughs in aquaculture and marine science. My track, somewhat unfortunately was to go back to a cold remote laboratory and work on forming new foundations for a hypothesis about human growth disorders. i would not get a chance to pursue the hatchery problem, though in many senses, that's where my heart was.

RNA makes medical news

  David Clayton has focused his career around detailing the structure, function and particularly the genetics of the mitochondria, a cellular organelle which serves as the cell's power plant. The mitochondria is probably most well known as the cells source of ATP, the energetic molecule which powers all cellular processes which require energy. He not only discovered that mitochondria have their own DNA, which was a primary factor in the development of endosymbiotic theory,  but he played a primary role in sequencing and classifying each of those 17,000 mitochondrial DNA genes.

In the spring of 2001, Dr. Clayton had big news[4], at least in the field of small RNA, an emerging field of study driven by the increasing realization that many genes, maybe even most genes that are expressed have no associated protein. One of those genes that  has no associated protein is the RNA component of  Mitochondrial RNA Processing (RNase MRP). What was peculiar was that mutations in this small RNA were found to be associated with a rare growth disorder known as cartilage hair hypoplasia. [5]  It was not known what role RNase MRP plays in the disorder, but the discovery that it was  a small RNA that was causing the problem was seen as an important lead by those interested in discovering the role of enzymatic RNA in cell function. Rarely, it seems does a "non explanation", or merely an observation such as this qualify as "big news".  Such is the nature of our emerging understanding of the diverse functions of RNA. We know that many genes are producing RNA transcripts that never seem to be translated into proteins, maybe even most active genes, but science is making little progress determining what these unknown functions may be. The molecular tools for studying RNA are not yet as sophisticated as those used for studying proteins and DNA.

RNase MRP is so named because it has been found to play a role in the processing if pre-messenger RNA into messenger RNA. This is a splicing process common to all higher animals which serves as a quality control step for the production of new mRNA. In particular, if a virus inserts a gene of it's own into the genomic DNA of a host, a complicated splicing step will presumably prevent a "simplistic" viral gene from being transcribed into a viable mRNA and thus translated into protein. Thus, the eukaryotic system of introns and exons is thought to be a defense system against viral attack for simple organisms as well as multicellular organisms. Because of the splicing step required to produce a viable mRNA, single cell, and simple multicellular organisms can defend themselves without the benefit of a well developed immune system. But the so called "big news" here a splicing protein could be the source of a growth disorder in humans.

Growth disorders are usually named according to their severity, and hopefully, the genes they are associated with if they are known.  In scientific terms,  "plasia" refers to growth,  achondroplasia refers to no cartilage growth, chondrodysplasia refers to bad or malformed cartilage growth and there is even a "pseudoachondroplasia", or PSACH which translates from medical terminology  to "yet another source of no cartilage growth".

The hunt for nothing

 And of course, pseudoachondorplasia ( PSACH ) seemed to be packing a few surprises of it's own. Researcher Jacqueline Hecht of Texas Medical had discovered that PSACH was closely associated with a gene she named cartilage oligomeric matrix protein ( COMP ) referring to the characteristic of forming little 5 pointed stars ( oligomers ) that locate themselves outside the cell ( in the matrix ) of cartilage. Presumably then, COMP plays some critical role in developing cartilage, and when they contain one of a few mutations, that critical role is impacted and a severe form of dwarfism occurs. Well not exactly. One of the first experiments they executed after discovering comp was to develop a transgenic mouse which mimics the behavior of COMP mutations, that is, develops with a growth disorder. But before even that, they developed a mouse which had no COMP at all, or a NULL mouse. To the surprise of researchers, the COMP NULL mouse developed normally [8]. This was surprising because researchers were now challenged with "fixing" something which had no necessary purpose in development. Certainly COMP plays some role in the consistency of cartilage over a normal life, but it is not a component which contains a necessary signal for proper bone development.

As an additional observation in affected chondrocytes, or cartilage cells, the culprit COMP never leaves the cell, instead it triggers a stress response in an organelle known as the endoplasmic reticulum, and the "choking chondrocyte" dies. A polymer of exracellular proteins, which should be secreted to the cartilage matrix, instead solidify inside the cell, eventually killing the chondrocyte.

 Well then, the road to decoding the mystery of the COMP gene and the associated PSACH disorder began with partnering with the University of Delaware in order to develop tools to selectively control the expression of COMP in model cells. The tool chosen was the ribozyme, or a gene that can be inserted into a cell by a viral vector, and produce an enzymatically active RNA transcript that will cleave, and thus disable the mRNA of the targeted transcript. The design of the targeting ribozyme is somewhat of a complex process.  The mRNA of the gene that we desire to target has a characteristic shape which is dependent upon its sequence. The prediction of the 2 dimensional shape, or secondary structure can be done with bioinformatics tools which are either downloadable, or if they are particularly computationally intensive, can be done online by a supercomputer. One such is the SFOLD website hosted by the Department of Health of New York State.

The Computer Modelling

   SFOLD stands for statistical RNA folding prediction. The secondary structure of an RNA is not a constant shape, but a probabilistic distribution, somewhat like an actively writhing snake. For the sequence of the COMP mRNA, the SFOLD returns between a dozen and two dozen possible structures listed by boltzmann distribution probability.  Since the most active sites, in terms of potential ribozyme targeting are on the loops, and not on the paired, or Watson-Crick bound stems, those are the ones that are the best targets for ribozymes, or any "antisense" targeting technology.

But I was faced with a question here. Should I use the sequence of the normal gene, or the sequence of the mutant, or PSACH causing gene. Simple enough. First I submitted one and then the other. Hopefully, they would produce the same results and there would be no decision to make. But, of course, Murphy ruled and the mutant and normal were different, and in fact impossible to compare side by side. Both produced a dozen or two possible strucures depending on how low you wished to do a cut off, or threshold based upon it's calculated probability of the structure being representative at a particular time.

Great, after a couple weeks, maybe months of drinking coffee, twirling my hair around my finger, and going to graduate classes, I decided I would write a computer program that totals up the binding probabilities of between each pair of nucleotides, and if they were bound in a single pair more than 50 percent of the time, I would label them paired. Otherwise, I would label them loop nucleotides, regardless of what they might be paired with from "time to time". My hope was that I could get a single picture that I would call my "maximum probability structure". Curiously or not, the maximum probability structures of the normal and PSACH COMP  mRNA were different. The deletion of a single GAC trinucleotide caused a "switch" in the structure of the mRNA.

  The question now was  whether the calculated switch real or merely a computational aberation. The structure of mRNA is difficult to determine in the laboratory, but if there was in fact a switch in mRNA structure due to the PSACH trinucleotide deletion, one could design ribozymes that would target one structure and not the other. In fact, this was done and laboratory results confirmed that there had been a switch [7] [8] .

The papers from the Hecht group represent something called antisense technology. What Hecht has done is create a family of ribozymes that selectively interact with different parts of the COMP transcript. The activity of a ribozyme ceated in this nature is dependent upon the shape of the target, that is is the cleave site bound or is it on a loop? If ribozymes have differential activity between he normal and mutant COMP transcript, then it confirms that a destabilization, or "flip" has taken place in the target.

 Is it possible that the severe PSACH condition is not caused by a bad matrix protein, but by a bad mRNA structure? Is there such a thing as a surveillance system that can recognize a genetic mutation by the shape of the mRNA, a characteristic that can now be predicted by computer modelling without laboratory procedures? Let me just say that the thought of doing molecular biology without a laboratory is a total non-starter in a molecular biology laboratory. For me then, this was the end of the line. None the less, surveillance of mRNA was becoming the focus of growth disorder research on the other side of the globe.

RNA Surveillance gains fans

 Across the globe from Texas Medical center, in a pediatrics research laboratory in Australia, researchers were were confronting laboratory results as puzzling as Dr. Hecht's. In a laboratory run by John Bateman, mutations on collagen 10, also a product of long bone growth plates, would not only cause a condition similar to Dr. Hecht's Pseudoachondroplasia, but in fact the offending genes could be recognized by the cell, and transcription of the offending allele was shut down by the cell. [9]  The full name of Dr. Bateman's subject pathology is   Schmid metaphyseal chondrodysplasia. (MCDS)  Metaphyseal is an anatomical term which refers to the ends of long bones and chondrodysplasia, as I have mentioned, refers to poor cartilage growth.

One of the first things we might ask, if we were to consider COMP and collagen 10 together is "how do the mouse models compare". Is a Collagen 10 deficient mouse normal or does it display characteristics of a chondrodysplasia?  Unfortunately, the answer is slightly more complicated than the COMP mouse model. [10] .  Collagen 10, l when completely deleted from a mouse model, causes  visible changes in the phenotype.  But, I think it is worth while to take another close look at the Gressa paper. The article notes severe developmental repression in the immune system, specifically the thymus and the spleen. In the histology of the long bone growth plates, the histology, or microscopic anatomy, seemed to pair with the normal, not the COL 10 deficient mice. Given our particular interest, and hindsight of over a decade, we could have almost titled the article "Collagen 10 deficient mouse long bone growth plates appear normal". This would have been a surprise given the most likely motive for the paper was to provide background for research on MCDS.

Two Types of surveillance

Lets take a look at two proposed types of surveillance. In the case of COMP, we are talking about a trinucleotide deletion, so the transcription process stays in frame. In the case of the Collagen 10 mutation, it is a single nucleotide deletion that causes a nonsense transcript and an early stop codon. The surveillance method for nonsense mutations has been fairly thoroughly studied and the references are discussed in Tan [11]. In highly simplified terms, the process of splicing together the exons of a messenger RNA leaves protein assemblies at the exon-exon junctions, the exon-exon junction complex (EJC).  There is then a "pioneer run" of the transcription process executed in the nucleus and during this process the EJCs are removed from the transcript.  Presumably then, a transcript cannot leave the nucleus unless all  of the EJC's are removed, and that serves as a surveillance system against nonsense mutations. That is, unless the nonsense mutation is in the last exon, in which case there we will be no remaining EJC to be left on the transcript. So then, the question is, if a MCDS causing mutation is in the last codon of of a collagen 10 transcript, will it be recognized by the chondrocytes RNA surveillance system? The presumed answer is no, but of course the surprise answer is yes.  [11].  The title of the Tan paper ( from the Bateman Group ), introduces the term 3' UTR, which refers to the portion of the RNA transcript located after the stop codon. Once again this fits into our title concept, the diverse functions of RNA. In other words, there are at least two, and in fact probably multiple methods of RNA surveillance that exist in chondrocytes.

 Once again RNA stability is an issue.

 When we discussed my computer modeling of the COMP transcript and its potential relationship to the actual medical condition related to chondrodyplasia, we were on left field as far as providing a realistic scientific hypothesis.  When we model an RNA, we always get a "snake bites tail" type of structure where the 5' cap, or beginning of the transcript Watson-Crick binds to the tail of the transcript, or 3' UTR. So then, mutations in the 3' UTR could serve to destabilize the mRNA in a thermodynamic sense. Now Tan and Bateman [11] are in fact verifying that stability is a factor in mRNA surveillance associated with this particular phenotype of growth disorder.

The cartilage connection

 The Bateman paper [9] introduces a lot of new concepts and terminology to the world of pediatric growth disorders.  Among these is that chondrocytes, or cells that produce cartilage extracellular matrix have special RNA surveillance properties that don't exist in other cells. For example, kidney cells also used in [9] as a control were not affected by transcripts carrying the mutation. How could we use this information to better understand David Claytons "Big News"?  Claytons gene, RNase MRP is expressed and necessary to all cells, but it only causes a disorder in cartilage and hair. Now we know that the RNA surveillance that is increasingly viewed as the cause of the disorder is located in chondrocytes only.

So then, what role could RNase MRP play in RNA surveillance? Actually that question is one that is better put to David Clayton. But, at least I could say there are two general possibilities. One is that the mutated RNase MRP is itself triggering a RNA surveillance system, or RNase MRP is part of an RNA surveillance system that is malfunctioning in a chondrocyte specific way, as is collagen 10 in MCDS.

The Big Picture

So what do these pediatric disorders have to do with the culture of marine finfish? In biology, we tend to think there must be an underlying reason for everything, even if it is not immediately apparent. Why do  people have an appendix if it serves no immediate purpose. The answer is that the appendix serves a purpose to our relatives in the animal kingdom. To us, an appendix is a hazard with no immediate purpose.  Perhaps the same is true for the growth disorders we call the displasias, or cartilage disorders. The small insignificant anomalies in DNA that cause them are legacies from the cartilage properties we share with early ancestors of verebrates. Cartilage disorders serve to freeze the genome in vertebrates.

 What do I mean by "freeze the genome". In the Darwinian view of evolution there is a concept called incrementalism, the notion that many small changes will accumulate to a large change, and in fact a new species. According to this theory we would expect all species to be gradually changing over time. In fact, that is not what we observe. Paleontologists, those who study evolution from fossil evidence inform us that the process of evolution is more of a discontinuous process called "punctuated equilibrium"' There are exceptions to the observed process of punctuated equilibrium, most notably the shark. Fossil evidence of the shark, most commonly the teeth, show us that one of the first vertibrates has changed very little from it's first appearance, and in fact there is no clear differentiation between early sharks and today's sharks. That is, they have only changed incrementally, with no observed "puntcutated equilibrium" which is seen in the fossil record of later inhabitants of the land and oceans. 

The mechanism of freezing, or defending the genome is as follows: When random mutations occur, as they would during a large reproductive event such as a fish spawn, there are naturally many zygotes that contain replication errors. In order to prevent these errors from propagating, "tissue specific surveillance" recognizes them in chondrocytes, the critical cell in the organisms morphological development. The recognition process is as previously described. If a particular transcript develops errors during either the splicing stage, or possibly even malformed mRNA that does not seem native according the rules of surveillance, a stress response is triggered in the endoplasmic reticulum and the chondrocyte dies. The result is a malformed phenotype, or "runt"  in the terminology of the hatchery. The runt is transformed from a prospective member of it's species community into a meal with fins. It most likely does not matter what the particulars of the actual mutation are. As scientists have discovered in pediatric genetics, there is no rhyme or reason that can be attached  the translated function of affected proteins, and their intended use by the organism.

Is there a good , practical, reason for me to have brought all of this up? As you might have guessed, I am going to answer my own question in the affirmative. Science laboratories are usually devoted to doing very specific work. A matrix biology, or bone and cartilage laboratory, will be configured much differently than a laboratory looking at RNA secondary structure, enzymatic action, and even splicing and RNA surveillance. A pediatrics laboratory will be located at a children's hospital, while a fish hatchery may be located at a marine science institute, or even a station on an island archipelago. The practical justification for writing this then is to encourage scientists from different origins to cooperate more efficiently when it becomes apparent that the solution to a particular problem will come from well outside the confines of a particular scientific community.

 If the molecular dynamics of chondrodysplasias are elucidated, it may be relatively simple to design a drug to relieve many of the affects. When the molecules are known, and there are good model systems to recreate the phenomenon in the laboratory, libraries of drug candidates can be screened to find a treatment. When the molecular dynamics are not known, and the condition cannot be modeled, it is difficult to focus in on drug candidates. 

I if there is an uplifting aspect of these curious findings, it is that the genes affected are most often themselves not necessary to normal development, and a such, simpler treatments, such as drugs, remain a viable option than future untested therapeutic concepts such as gene therapy.

References

1. Lasse Peters1 and Gunter Meister1, ,
Argonaute Proteins: Mediators of RNA Silencing
Mol Cell. 2007 Jun 8;26(5):611-23.  Pub Med 

2. Janowski BA, Huffman KE, Schwartz JC, Ram R, Nordsell R, Shames DS, Minna JD, Corey DR.
Involvement of AGO1 and AGO2 in mammalian transcriptional silencing.
Nat. Struct. Mol. Biol. 13 (2006), pp. 787–792 Pub Med
 

3. Tasleem Samji
PIWI, piRNAs, and Germline Stem Cells: What’s the link?
Yale J Biol Med. 2009 September; 82(3): 121–124.  Pub Med

 4. David A. Clayton
Molecular biology: A big development for a small RNA
Nature 410, 29-31 (1 March 2001)  Pub Med

 5. Maaret Ridanpää,et. al.  Mutations in the RNA Component of RNase MRP Cause a Pleiotropic Human Disease, Cartilage-Hair Hypoplasia  abstract

 6. Merritt, Thomas M., Ph. D. Dissertation - Studies of pseudoachondroplasia chondrocytes and knockdown of the cartilage oligomeric matrix protein
The University of Texas Graduate School of Biomedical Sciences at Houston, 2008, 159 pages;  Abstract

7. Alcorn JL, Merritt TM, Farach-Carson MC, Wang HH, Hecht JT.
Ribozyme-mediated reduction of wild-type and mutant cartilage oligomeric matrix protein (COMP) mRNA and protein.
RNA. 2009 Apr;15(4):686-95. Epub 2009 Feb 23. Full text

 8. Svensson L, Aszódi A, Heinegård D, Hunziker EB, Reinholt FP, Fässler R, Oldberg A.
Cartilage oligomeric matrix protein-deficient mice have normal skeletal development.
Mol Cell Biol. 2002 Jun;22(12):4366-71. Full text

  9. Bateman JF, Freddi S, Nattrass G, Savarirayan R.
Tissue-specific RNA surveillance? Nonsense-mediated mRNA decay causes
collagen X haploinsufficiency in Schmid metaphyseal chondrodysplasia cartilage. Hum Mol Genet. 2003 Feb 1;12(3):217-25. Full Text

10. Catherine J. Gressa and Olena Jacenkoa
Growth Plate Compressions and Altered Hematopoiesis in Collagen X Null Mice
The Journal of Cell Biology, Volume 149, Number 4, , 2000 983-993 Full Text

 11. Jacqueline T. Tan, Friederike Kremer, Susanna Freddi,
 Katrina M. Bell, Naomi L. Baker, Shireen R. Lamandé,
 and John F. Bateman
Competency for Nonsense-Mediated Reduction in Collagen X mRNA Is Specified by
the 3' UTR and Corresponds to the Position of
 Mutations in Schmid Metaphyseal Chondrodysplasia
Am J Hum Genet. 2008 March 3; 82(3): 786–793. PDF



Tweet
More about this author: Jeffrey Graf

From Around the Web




ARTICLE SOURCES AND CITATIONS
  • InfoBoxCallToAction ActionArrowhttp://www.helium.com/items/1208533-gene-expression-promoter-transcription-factor-drosophila-wingless-wnt-hedgehog-transgenic-biotech
  • InfoBoxCallToAction ActionArrowhttp://www.fau.edu/hboi/
  • InfoBoxCallToAction ActionArrowhttp://www.fau.edu/hboi/Aquaculture/AQacted.php
  • InfoBoxCallToAction ActionArrowhttp://www.snapperfarm.com/2006/aboutsnapperfarm.htm
  • InfoBoxCallToAction ActionArrowhttp://money.cnn.com/2009/04/27/smallbusiness/farming_the_open_oceans.fsb/
  • InfoBoxCallToAction ActionArrowhttp://www.hhmi.org/research/groupleaders/clayton_bio.html
  • InfoBoxCallToAction ActionArrowhttp://www.helium.com/items/1209408-mitochondria-endosymbiotic-theory-training-atp-adp-krebs-cycle
  • InfoBoxCallToAction ActionArrowhttp://sfold.wadsworth.org/
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pubmed/17560368?itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=4
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pubmed/16936728?itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=2
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pubmed/19774126?ordinalpos=1&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_SingleItemSupl.Pubmed_Discovery_PMC&linkpos=3&&log$=citedinpmcreviews&logdbfrom=pubmed
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pubmed/11242026
  • InfoBoxCallToAction ActionArrowhttp://www.cell.com/abstract/S0092-8674(01)00205-7
  • InfoBoxCallToAction ActionArrowhttp://proquest.umi.com/pqdlink?Ver=1&Exp=01-20-2015&FMT=7&DID=1529775631&RQT=309&attempt=1&cfc=1
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2661830/?tool=pubmed
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC133870/?tool=pubmed
  • InfoBoxCallToAction ActionArrowhttp://hmg.oxfordjournals.org/cgi/content/full/12/3/217
  • InfoBoxCallToAction ActionArrowhttp://jcb.rupress.org/cgi/content/abstract/149/4/983?ijkey=2294d380156ff6c5742c4a726026f1a52a9e8b58&keytype2=tf_ipsecsha
  • InfoBoxCallToAction ActionArrowhttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2427218/pdf/main.pdf