Edmond Fischer (2003) - How Proteins Communicate with One Another to Integrate Intracellular Signals

Our paths seem to merge at many, many times and I know Hans for many, many years. My only objection is he called me Doctor or Professor Fischer, why don’t you call me Eddie like anybody else. Well yesterday at our round table, several of us mentioned signalling. And this will be the subject of my talk today. By signalling we understand the very complex series of reactions that induce a cell to grow, to develop different shapes and eventually to die, in answer to many external and internal signals. A cell in the body is exposed to a multitude of signals that it has to integrate into a coherent response. And to do that pathways have to speak with one another, to communicate with one another, to coordinate, to synchronise all the reactions that take place and so that the pathways can carry out their specific functions. Now this subject belongs to the field of proteomics which so to speak represents the other side of the coin of genomics. And you all have heard a lot about genomics in connection with the human genome project. And there is no question that the essential completion of the human genome. And the unravelling of the genetic make-up of many dozen organisms now represents an achievement of enormous proportion. One that will change biology by changing the way we are thinking about various problems. And that will affect the future of biomedicine and medicine in general. Nonetheless it is also clear that the accumulation of these huge amounts of DNA-based information can only take us so far. The genome can predict the entire set of proteins that an organism can potentially produce. But it cannot take into consideration the enormous diversification of structure, of gene structure such as gene insertion or switching or recombination. Other kind of rearrangements as seen for instance in the immune system where less than 1,000 genes, V(D)J genes can potentially give rise to more than 500 million proteins. It cannot account for the co-transcriptional or transcriptional modification such as alternative splicing, the epigenetic changes such as mutations following exposure to radioactivity or chemical carcinogen or processing by limited proteolysis, the 100’s of chemical modifications, glycosylation, methylation, isoprenylation, …, 100’s of these. And therefore even if one disregards totally the immune system that has evolved precisely to produce structural diversity, there are far more proteins in the proteome than there are genes in the genome, contrary to the genome which represents a sort of a fixed entity. A problem with a finite end point more or less. The proteome is constantly in a state of flux, changing constantly with the stage of development, the growth of the cell and even with environmental factors. And finally the genome cannot tell us how proteins interact with one another through adaptor proteins, through scaffolding proteins, anchoring proteins, to produce those huge complexes that actually characterise their function. They can’t tell us the signals that will send a protein to the membrane or the ER or the Golgi. To transport in and out of the nucleus which I'm sure Günter Blobel will discuss later, or the signals that will target a protein for destruction at the appropriate time. And of course it won’t tell us anything about the mechanism by which enzymes are regulated. And this is what I want to focus on, regulation by protein phosphorylation. Because we know today that this is probably the most prevalent mechanism by which cellular events are affected. And indeed protein phosphorylation is involved at just about any aspect of cell behaviour, control of metabolism, transcription, translation, immune response, transformation and so forth. Now most of these phosphorylation reactions at least quantitatively occur on serine and threonine. But one of the great excitements in this area was the discovery just 25 years ago that phosphorylation on tyrosine was intimately implicated in transformation and oncogenicity, bringing into play a variety of tyrosine kinases, of cellular and viral origin, or linked to receptor for mitogenic hormones and growth factors. And since the latter represents some of the important molecules by which external signals are received, let me tell you a little bit about these. We know now I think to date about 59 different growth factor receptors. That can be divided in about, oh I think 20 subcategories. They all have the same architecture, a single trans-membrane segment that separates the catalytic activity inside tyrosine kinase, and then on the outside a great variety of structural motifs. That will allow specific binding of the ligand, cysteine rich domains that you find on the EGF or everywhere, those immunoglobulin-like domains, EGF-like, cadherin-like, factor 8, kringles, leucine protein rich - A great variety of domains that would give the specific character to those receptors. Most of the ligands are circulating hormones even though the FGF requires a protein such as heparin to bind. And then except for the F family, this is the largest family of receptors that recognise membrane bound ligands and they are concentrated on, if not exclusively found in the neuronal system where they play an important role in communication by regulating cell migration and axonal path finding and so forth. We begin to understand now how signal is transduced down from those receptors. In the resting state they are separated from one another and inactive. Upon binding of a ligand, for instance EGF, they dimerise and immediately undergo transphosphorylation. And then a signal is transduced by virtue of the fact that those tyrosine phosphate groups can recognise small modules, in this case called SH2, that stands for Src-homology 2 domain, that have a high affinity for tyrosine phosphate group. Not any but within a specific sequence. So before that those adaptor molecules are let’s say free in the cytoplast and they are recruited upon phosphorylation of the receptor. And then the change in conformation in that adaptor, that stands for growth factor receptor bound, described by Joseph Schlessinger. The change in conformation allow other modules, now the SH3 module, to bind to the next elements because it recognises prolene rich region. And then this binds to the next element and then triggers now a cascade of reaction. So this system now would allow a receptor to communicate with a protein far away, under conditions where it would have no other way to communicate. As you see communication is practically mechanical. Proteins interlock with one another like if one were dealing with a tinker toy or Lego type of assemblage. And this interlocking of protein guarantees that only the right protein will be drawn into a particular pathway and guarantees the fidelity of the transcription system. And today 100’s of those kind of modules have been identified by comparative sequence analysis. And they are found on a plethora of adaptor proteins, regulatory proteins, transcription factors and so forth. And I just want to mention, bring up a few of them. You have the SH2, Src-homology 2, I just mentioned that recognises tyrosine phosphate. SH3 that recognises prolene rich region. WW, it’s a domain sandwiched between 2 tryptophan, approximately 27, 30 residue long. That also recognise prolene blobs but of a different sequence and also recognise serine phosphate. The PH domain for pleckstrin homology domain, its principle role is not protein-protein interaction but it recognises anionic head groups on phospholipid, on phosphoinositide such as PIP2, PIP3. And it also recognises G protein and in this case plays an important role in chemotaxis. The PTB domain for phosphotyrosine bound also recognises phosphotyrosine bound but in a totally different way than the SH2. And it recognises particularly this NPXY module. And it has a very high affinity for upstream sequences. So it will bind even when the tyrosine is not phosphorylated. And then the PDZ domain is probably one of the most important domains for protein-protein interaction. And it recognises hydrophobic C termini, hydrophobic C termini, usually valine or isoleucine and also some very, some beta fingers, very sharp beta fingers that mimic in fact the C terminal segment. And those motifs can either be in different proteins or in the same protein so that PDZ can undergo head to tail oligomerisation. And that is its main characters. I will come back to that. The possibility of forming strings of PDZ domain to cluster, to organise receptors and ion channels. So let me now rapidly give you some schematic examples how some of those modules function. The SH2 domain has been found now in more than 100 proteins and they are usually found in adaptor proteins like RAB 2, this is an analogue. And they usually bind to activated receptor. And in most of these you have several SH3 domains which would allow those adaptors to latch on to different proteins in a signalling pathway. But the SH2 domain can play another and very important role. It can serve as a conformational switch. Many enzymes and this is the case of all the Src kinases, contain a tyrosine phosphate that serves as a negative determinant. So the SH2 domain can bind with it and therefore shield the catalytic site. Or in other cases the SH2 domain can bind to an inhibitory domain that again serve as a negative determinant. So in this form those enzymes are inactive, this is the case of the SIP enzyme and some tyrosine phosphatases. In this case, activation occurs when a specific tyrosine phosphatase knocks out this phosphate group and therefore the molecule opens, frees the catalytic site, or when that enzyme comes into contact of another molecule that has a higher affinity for the SH2 domain. So the SH2 domain switches its allegiance from the enzyme, let’s say to the receptor, in this case. So you see the SH2 is, can be considered as an internal conformational switch, just like the phosphate group can function as an external molecular switch to activate or inhibit the enzymes. The main function of the PH domain and this has been found now in about 500 different protein is to bind the molecules to the membrane and to serve the same function as a myristoyl or as a palmitoyl group. And you find these in substrates, for instance for the fibroblast growth factor or the insulin substrate IRS1234. Those molecules have usually a large string of tyrosine residue. And they serve as annex binding, anchoring, annex docking sites for molecules that have SH2 domain. So the way it works, when you activate for instance the insulin receptor, it recruits its substrate to the membrane. The PTB domain slaps it against the receptor and the receptor immediately phosphorylates many of the tyrosine residue and triggers signalling, down signalling. The importance of the PH domain in this reaction can be seen in a kinase, the bruton kinase, BTK, that plays an important role in B cell development and signalling. A single point mutation in the PH domain, argenine to cysteine mutation, interferes with its binding to the membrane and causes gross functional aberrations resulting in X-linked agammaglobulinemia simply by not being able to recruit that kinase to the membrane. The PDZ domain, as I said, owes its characteristic to the fact that it can form chains, multiple copies on a chain, MAGUK stands for membrane associated protein with one elate kinase-like. They don’t have any quanylate kinase activity. But here is the post-synaptic density protein PSD95 from which PDZ gets its name, InaD protein that I’ll discuss, so, a large number of proteins that all have the sort of a modular structure. And let me show you, taken from the literature, some schematic examples, how these function. PSD95, its two PDZ domains combine to subunits of the NMDA receptor for instance and dimerise the channel. And another PDZ domain, I think it’s the second one as a matter of fact, can also form a heterodimeric liaison with a PDZ domain of the neuronal nitric oxide synthase. It’s an enzyme that requires calcium calmodulin for activity. So localisation now of the synthase with a channel and the synthase is as I said a calcium calmodulin requiring enzyme. Localisation of the two together will bring about the very efficient production of nitric oxide, following calcium entry. InaD has a 5 PDZ domain, no catalytic activity and its main function again is to cluster many enzymes in the phototransducing, the drosophila phototransducing system. The human homologue has eight of those PDZ domains. So upon activation of the system by a photon through the G protein, there’s an activation at PLC beta. And this produces diacylglycerol and calcium that will now activate this tranchant receptor potential calcium channel and bring about calcium entry and cell depolarisation. And the reverse reaction, the deactivation process, is a calcium dependent reaction that involves PKC, an I-specific protein kinase C, calmodulin, arrestin, a calmodulin-dependent protein kinase. But by clustering all those proteins together, you have a huge amplification of the system. To such an extent that activation of rhodopsin by a single photon will bring about the activation of about 100 calcium channels in a millisecond time scale. So this is the importance of putting all those proteins together. Finally PKZ can participate or can serve as a localisation machinery of the cell, in c-elegans for instance three proteins, 2, 7 and 10. Each with its PDZ domain can participate in the transports of receptors, particularly lets 23A tyrosine kinase receptor. So this unit together with some of its targets or some of its additional protein can serve as a cargo to transport proteins inside the cell. The complexity of the regulation of signalling is well demonstrated by P95VAV and this looks like one of those Swiss army knives that can do anything. With all of its domains it can interact with a number of systems. I don’t describe, it is an enzyme, it is a GEF, a guanine nucleotide exchange factor which is always linked with a PH domain. But importantly it has a very hydrophobic region at the N terminus analogous to calponin. Calponin are proteins that regulate smooth muscle contraction. And that calponin like region serves a very negative function because when part of it is deleted, if you chop off the N terminus then VAV becomes oncogenic. And VAV becomes phosphorylated upon receptor activation. It becomes phosphorylated by the fusion protein BCR-Abl, that is implicated in chronic myelogenous leukaemia. Ok so, our chairman tells me that I am exceeding my time. What I wanted to say is that the big problem for the cell is how now does it select among all the things it can do. Because a protein like VAV serves like those roundabouts that you find in traffic that can send cars in every direction. And this is a big difficulty for the cell. And it can only do that when it receives a signal from the outside, when many receptors collaborate together, speak with one another, generating a combinatorial type of response. So this is some of the things that we know about signalling but where do we go from there? I mean what are the important problems that we have to solve? We know several of the important signalling pathways. We have characterised the elements of those pathways. We know their structure, we know their function. But these molecules are only the words that a cell uses to bring about, to modify its behaviour. We know some of those words, we know bits and pieces of sentences they use but we don’t know the language a cell has to use to synchronise, to coordinate all the reactions that take place. Furthermore through 3½ billions of evolution the cells have had all the time in the world to establish some secondary pathways, parallel pathways, a chance, feedback reactions to regulate their growth, to protect them against adversity or to plan their death. And much more importantly, we don’t know the language cells speak with cells. And this has been absolutely crucial in the establishment of those very complex networks of information as you see during embryonic morphogenesis. That you see in the immune system. And the infinitely more complex central nervous system where 1,000million cells speak with one another through something like a million billion synapses to finally establish the generation of memory and thought and consciousness. And as Erwin Neher said yesterday, this will be one of the main challenges that the biologists will have to solve in the years to come. Thank you for your attention. Applause.

Edmond Fischer (2003)

How Proteins Communicate with One Another to Integrate Intracellular Signals

Edmond Fischer (2003)

How Proteins Communicate with One Another to Integrate Intracellular Signals

Comment

„Signaling is very complex“, Edmond Fischer emphasizes at the beginning of this talk, his fourth one at the Lindau Nobel Laureate Meetings. Yet this complexity appears less daunting and more challenging when one listens to Fischer who is one of the pioneers of signaling research and at the age of 83 still an engaging and prolific speaker. Together with Edwin Krebs he had discovered the mechanism of reversible protein phosphorylation, which turned out to be the most prevalent switch of intracellular communication. In this presentation, however, he only briefly mentions his own achievement and turns to the wider field of signal integration within cells. Even if the unraveling of the human genome was an „achievement of enormous proportion“, the knowledge of the genome is not nearly sufficient to explain and understand cellular signaling. „The cell in a body is exposed to a multitude of signals that it has to integrate into a coherent response“, he says. „The pathways have to speak to another – this is a subject that belongs to the field of proteomics.“ Focussing on signal transduction via enzyme-coupled receptors, Fischer summarizes current knowledge on how extracellular signals are relayed into the cell, and then turns to a detailed description of interaction domains, which allow signaling proteins to assemble to signaling complexes that consequently trigger downstream cascades. „Communication is practical mechanical: Proteins interlock with one another, comparable to lego“, Fischer says. Vividly he introduces form and function of interaction domains such as SH2, PH, and PDZ to his audience. It is typical for Fischer that he does not get lost in details but keeps the big questions at the horizon in mind. „How does the cell select among all it can do?“ Doesn’t its situation resemble that of a car driver who finds himself in a roundabout that can send him to many different directions? How can a cell, faced with a plethora of signals, choose the right combinatorial type of response? „We know several important pathways, their structures and elements“, Fischer concludes. „But these molecules are only the words that the cell brings about to modify its behaviour. We know some of these words, even pieces of sentences, but we do not know the language the cell uses to synchronize the signals.“Joachim Pietzsch

Abstract

The types of reaction a cell uses to transduce an external signal into a particular response will be discussed. The subject belongs to the field of proteomics as opposed to genomics. Some characteristics of both will be contrasted, emphasizing that the proteome that remains constantly in a state of flux represents a far more complex and dynamic entity than the genome.

The main focus of the talk will be on signalling by tyrosine phosphorylation which has been directly implicated in the regulation of cell growth, differentiation and transformation. External signals coming in the form of mitogenic hormones and growth factors act on transmembrane receptors that are themselves tyrosine kinases; these, in turn, transduce they signals with the help of a variety of adapter or docking proteins that interact with one another through a diversity of binding domains, thereby initiating different signalling pathways. Some of the properties of these modules will be detailed. The src-homology 2 (SH2) domain, in addition to allowing adapter or linker proteins to bind to activated receptors, can also serve as an internal conformational switch to regulate the activity of various enzymes. The PH domain recruits protein to the membrane while the PDZ domain which is usually found in multiple copies within a protein serves mainly to cluster ion-channels and receptors on the membrane, and bridge them to the cytoskeleton.

Transmembrane protein tyrosine phosphatases which catalyze the reverse reaction also have a modular structure, often containing immunoglobulin-like and/or fibronectin type III repeats. Surprisingly and contrary to the tyrosine kinase growth factor receptors that respond mainly to circulating ligands, the tyrosine phosphatase receptors display all the hallmarks of cell adhesion molecules. This would suggest that they are involved in, or regulated by, cell-cell or cell-matrix interaction, with the exciting possibility that they might be directly implicated in contact inhibition. Some of the problems that confront us in the field of cell signalling and remain to be solved will be summarized.

Cite


Specify width: px

Share

COPYRIGHT

Cite


Specify width: px

Share

COPYRIGHT


Related Content