AUSTRALIAN FRONTIERS OF SCIENCE, 2005
Walter and Eliza Hall Institute of Medical Research, Melbourne, 12-13 April
Infection and immunity a bird’s eye view
Dr Jamie Rossjohn, Wellcome Trust Research Fellow, Monash University
(Click on image for a larger version)
What I want to do before delving into the depths of some of the work that we do is to take a step back and ask the question: why do I, and many of my like-minded colleagues, want to spend a lot of our scientific time determining the three-dimensional structures of proteins?
Proteins are the micromachines of our body, and the malfunctioning of proteins is the root cause of many diseases. It is the precise three-dimensional shape of a protein that determines its function. Therefore, if you can visualise its precise three-dimensional shape or structure it will give you great insights into a particular function of that particular protein. So if the protein is an enzyme, it will provide the basis for catalysis. If it is a receptor, it will give you the mode of binding. If it binds nucleic acids, it provides the basis for interaction. If the protein is a channel, it provides the mechanism of channel function. And, of course, added to these potential great advancements of fundamental knowledge there is also the application of structural biology, which is structure-based drug design.
Since this field started in the early 20th century, there have been some wonderful achievements in the field. I want to highlight just a couple of them. Firstly, there is Dorothy Hodgkins, the only female British Nobel Prize winner, who solved the structure of penicillin and insulin. We have Watson and Crick, who solved the structure of DNA. (It is just over 50 years since this fantastic, landmark achievement.) And a home-grown example is the development of the anti-flu drug Relenza, done by Peter Coleman, von Itzstein and colleagues.
When the field started, it could take a number of years, say 30 years, to solve the first protein structure. Structural biology and X-ray crystallography is very much a technology-driven discipline. So with the advance of faster computers, better methodology and better hardware we can solve structures a lot better and a lot faster.
(Click on image for a larger version)
One of the big impacts of structural biology methods of determination is the development and the use of synchrotron radiation. Synchrotron irradiation enables structural biologists to determine structures a lot faster and a lot more accurately. So I think it is a wonderful opportunity that the National Synchrotron is being built at Monash University, Melbourne, currently under construction now and it will be on line in 2007. I think this is a wonderful opportunity for any of the Australian scientific community who wish to embark on structural biology related programs in the future.
(Click on image for a larger version)
The specific aspect of my research that I want to highlight today is in relation to cellular mediated immunity. The video that you have just seen playing in the top left panel of the screen was of a killer T-cell killing the virally infected host cell. It is a critical step of a critical component of the immune system. The panel below that highlights a snapshot of the junction between the killer T-cell and that of the virally infected cell. It is this junction that contains a myriad of receptors that, upon initial engagement, dictate the outcome of whether a cell gets lysed or not is it virally infected, yes or no? And there are a myriad of receptors within this system, termed the synapse or specifically the immunological synapse.
You see now a schematic showing some of the components of the heart of this immunological synapse, which I want to spend a little bit of time discussing.
We have the antigen-presenting cell the cell that may be virally infected and the viral fragments, or the peptide fragments, are co-presented by the major histocompatibility complex molecule, MHC class I, which is a heterodimer comprising a heavy chain and a light chain.
This is specifically co-recognised by an T-cell receptor, which is depicted here in the recognition event. Upon this specific recognition event, there is a host of co-recruitment events, such as the employment of the receptor CD8, and also employment of the CD3 complex, which is a multicomponent system but is monomorphic, and which comprises the εγ heterodimer, the εδ heterodimer, and the CD3 ζζ homodimer.
The CD3 complex is critically important, because these are the components that mediate the signalling, translating the extracellular recognition event to the cell signalling event. The T-cell receptor doesn’t have any signalling function whatsoever.
One of the critical questions in the field and, I think, in any receptor recognition event is how such extracellular recognition events translate to intracellular signalling.
The amazing thing about this system is that it is dominated by weak interactions. The interactions between a T-cell receptor and the MHC are in the micromolar range, which is in stark contrast, say, to that of the antibody-antigen interactions, which are very, very tight interactions. And the interactions between some of these coreceptors are in the 200, 300, 400 micromolar range, so they are very, very weak indeed. So one of the critical questions you have got to ask yourself is then: why is it so extraordinarily weak, and how does it achieve this level of exquisite specificity when you have such weak interactions?
Because it is a sort of a system that is dominated by weak intermolecular interactions, it is also a system that can be manipulated, modulated, in many different ways. For example, you can make subtle changes in the peptide, leading from a peptide that can kill a cell to a peptide that can prevent a cell from being killed from an agonist to an antagonist peptide.
One of the most common forms of modulating immune response is by developing monoclonal antibodies to some of these components, and these are very successful immunosuppressants. For example, OKT3 is one immunosuppressant used since the mid-1970s, and Michelle Dunstan and Lars [inaudible] recently elucidated the structure of this drug OKT3 in complex with this antigen. So this information provided the mode of binding for this drug and also the mode of dimerisation of this essential heterodimer.
But I shall be largely focusing on the MHC interacting with the T-cell receptor just two components but with a large complexity underneath these two components.
(Click on image for a larger version)
The complexity arises from the observation that the MHC comes in two flavours. It comes in the flavour of MHC class I and MHC class II, both of which are heterodimers. The MHC class I comprises a heavy chain and a beta2-microglobulin domain. The antigenic peptide is presented by the heavy chain and it is bound in an extended fashion between the helical jaws of the heavy chain, termed the antigen-binding cleft. MHC class I normally binds residues of about nine amino acids in length, whereas MHC class II, as you can see, is similar it is still a heterodimer but it has a different function: it can bind longer peptides. It can bind peptides about 13 amino acids long.
(Click on image for a larger version)
There is another layer of complexity, though, more flavours of the MHC class I and class II, and there are many, many different alleles within the MHC. These alleles can differ by a single amino acid or by up to 30 amino acids. These polymorphisms, which I have colour coded here, are concentrated in the antigen-binding cleft. The effect of these polymorphisms is to subtly alter the MHC so that you can bind distinct sets of peptides or peptide repertoires. But it also subtly alters the substructure of MHC class I. The combination of these two effects means it can profoundly alter the types of T-cell receptors that can be recognised by these complexes.
(Click on image for a larger version)
I show here an example of the antigenic peptide, where in essence the residues that point down are the residues form the MHC anchor residues, that make the interaction specific with the major histocompatibility complex. It is the residues that point up, however, that are the potential contact points of the T-cell receptor. We have five potential contact points here, but typically there are about three or four interactions per T-cell receptor complex.
(Click on image for a larger version)
As to what the T-cell has to face, or what a given T-cell receptor has to face, on the cell surface, there are a myriad of choices. There is a forest of MHC peptide complexes, and each one is subtly different they can have a subtly different HLA allele, they can have different peptides bound. Yet despite all this diversity the T-cell receptor is very, very much restricted to interact, in the normal healthy state, with only one MHC complex. So the question is: how is this extraordinary specificity, this restriction, achieved against this myriad of opportunities?
(Click on image for a larger version)
Well, one of the ways a T-cell receptor does that is by matching the diversity stakes. So there is a large diversity in the T-cell’s repertoire. Depicted here is a schematic of an T-cell receptor, which comprises two chains, depicted in red and light blue. It comprises variable domains, depicted at the bottom, and constant domains, depicted at the top. The variable domains are the region that interacts with the MHC. The region it interacts with is the hypervariable region, and it is encoded by the CDR loops, the complementarity determining region loops.
Here you see a schematic of how this T-cell receptor diversity is generated.
So for a particular V chain, any one of 46 V gene segments combine with one of 50 J segments to give a certain V chain, and similarly with a V chain, you can have a combination of one of three gene segments to give rise to the V chain.
The CDR loops 1 and 2 are encoded by the V or the V domains only, whereas, as you can see, it is the CDR3 regions that are encoded by a number of the gene segments the V, J and the V, D and J segments. But in addition to that we have non-nucleotide encoded additions and deletions at the boundaries. So all in all, this can give rise to about 107 or 108 unique T-cell receptors.
However, despite this diversity there are some T-cell receptors that don’t make use of this diversity whatsoever. It is identical and unrelated individuals. This is termed immunodominance something I will touch upon lightly.
(Click on image for a larger version)
This is the example. If an individual is infected with Epstein-Barr virus, which is a common, persistent virus with which 90 per cent of the human population is infected, and you have the particular flavour of the MHC termed HLA-B8, then you will have this T-cell receptor in your body at high levels. And it will be identical down to the last amino acid in unrelated individuals. So one of the questions we want to ask from the structural biology is what is driving this remarkable observation of immunodominance.
(Click on image for a larger version)
But in addition to that, this T-cell receptor has another rather unique property, in that it can cross-react. It can react with things that it shouldn’t do, and it can cross-react with another flavour of MHC allele termed HLA-B44. In a normal healthy state, that doesn’t happen. But in the case of a transplant recipient or donor, a transplantation, it can happen. You get cross-reactivity, and the consequences can be quite dire. You can get T-cell mediated transplant rejection.
It is very subtle, though, in its discrimination between what it can react with and what it cannot cross-react with. For example, it can cross-react with HLA B*4402 but not B*4403. Yet there is only one amino acid difference between these two different alleles. In fact, the differences between B*4403 and B*4402 are so marked, it has actually been termed the ‘taboo mismatch’ in the transplantation field, because it can really result in dire consequences. So we wanted to address what is the structural basis of these observations.
(Click on image for a larger version)
What we did was to determine the structure of HLA B*4402, the complex with the peptide, just depicted here. This shows the one amino acid difference, at position 156. The important thing about this mutation, or this difference, polymorphism, is that it is buried. It cannot directly interact with the T-cell receptor, thereby affecting the different TCR recognition properties. The effects are a lot more subtle.
So we solved the structure of B44 and B*4402 and B*4403 with the same ligand. You see here a superposition of these two structures.
The effects of these mutations are very, very subtle. It causes an increase in the breathing movements of B*4403, which also translates to an increase in the breathing movements of the peptide as well. It also relaxes the type of peptides that B*4403 can bind, in comparison with B*4402. Yet it is these subtle structural events that can lead to such drastic observations of transplant-mediated rejection or outcome. Crystallography is a great technique, where you can examine structural differences that are so subtle yet have such profound biological outcomes.
(Click on image for a larger version)
I also want to continue the theme of sensitivity when we also looked at the conformation of the antiviral peptide bound to HLA-B8. Here we have a schematic of how the peptide is bound in the antigen-binding cleft. And there is only one residue, really, that contacts the T-cell receptor it can be seen sticking out like a sore thumb.
But if you make a mutation of tyrosine to phenylalanine, which is just removing its hydroxyl group, you have a ten-fold loss of recognition, Tyr to Phe, which means it goes from a signalling outcome to a non-signalling outcome. So why is the system so exquisitely selected to be so sensitive?
(Click on image for a larger version)
This slide addresses one of the ways we addressed the structural basis of immunodominance and this specificity for determining the crystal structure of the complex. Here you can see the antiviral peptide with the tyrosine sitting centrally at the T-cell receptor interface.
Seen here is a footprint of the T-cell receptor docking onto the MHC. Each loop has been colour coded, which represents the CDR loops depicted here. The essential tyrosine is sticking up, very much exposed, and you can see that it has been taken a firm grasp of by a number of the CDR loops. It interacts with CDR1 and CDR3, and parts of CDR3. So this is explaining the specificity of this interaction. And a lot of these CDR loops chip in to make this large, extensive interface, apart from CDR1. So structural biology can give you great insights into the mode of recognition, in this particular case.
(Click on image for a larger version)
This shows the sort of thing that structural biologists do: to have a very good atomic look at the interactions. So you have the essential tyrosine here forming a myriad of interactions. As I said, this hydroxyl group is very important. It doesn’t directly contact the T-cell receptor; it forms water-mediated contacts. So these water-mediated contacts are very, very critical.
One of the things we did as well, which is depicted in the movie seen at the right of this slide, was to solve the structures of these components in the non-liganded states. So we could actually see the flexibility, the conformational changes that take place upon complexation. It is very much an induced-fit mechanism of binding.
(Click on image for a larger version)
From one of the observations, then, we wondered about the mechanisms of immunodominance. Here we have a T-cell receptor, and in the normal situation you might have about three prominent side chains sticking up as sort of pegs, making specificity contacts with notches in the T-cell receptor. In the case of the system that we are exploring, however, there is only one residue that is sticking up, and only one notch. And we can have similar unusual situations as well. Is this one of the mechanisms that we have for generating immunodominance?
(Click on image for a larger version)
We explored that by looking at the influenza system, where we have again a very flat, featureless epitope which is a restricted response, where we have something that is sticking up which leads to a very diverse response. So we wanted to ask the question: well, what happens if you take this and make it flat? What sort of response do we get in the T-cell repertoire?
Well, we were able to do this by combining the techniques of structural biology with that of reverse genetics.
(Click on image for a larger version)
In a nutshell, when we looked at this system, we mutated the prominent epitope to a flat epitope. And it changed the T-cell repertoire remarkably, such that it now generates an immunodominant response. So it is very much a case of pebbles in a pool having a large ripple effect, in this particular system.
(Click on image for a larger version)
Finally, we also wanted to ask: what is the energetic landscape of this particular interaction? We have a myriad of interactions interfaced; we wanted to ask the question: do each of the highly selected gene segments contribute equally to the energetic load, or are some interactions more important than others? Is there an energetic hotspot, and if so, where and how large would it be?
And so we wanted to establish this, and shown here is the hypothesis we had.
(Click on image for a larger version)
To do this we mutated every residue that was forming a contact to alanine, then we looked at its binding characteristics, its affinity and interactions, by using SPR analysis. You can clearly see that the CDR3 loops and the CDR3 loops, circled here, have a profound effect on recognition if you mutate it, but so do some of these areas here, further to the right and also circled, as well. But these areas do not directly contact the MHC; in fact, all they do is stabilise the conformation of the ligated loops of CDR3 and that of CDR3.
(Click on image for a larger version)
So in fact there is an energetic hotspot, not all interactions at the interface are important, and the energetic hot resides firmly and squarely with the CDR 3 loops. This has profound implications for T-cell immunodominance, T-cell alloreactivity and the central question of MHC restriction.
We were also able to look at this energetic hotspot as well by comparing it with the surface of HLA-B8. So despite the very extensive interface, only a few residues were actually very, very critical for this interaction in terms of an energetic setting.
(Click on image for a larger version)
I hope that I have given you some insights into the antigen recognition process by using structural biology as a tool. It is a very important tool, but if you also combine it with novel technologies and related techniques it can be wholly informative in addressing the biological question that you started out with.
This is just one example of a structural biology program that has been undertaken. There are many structural biology programs that are undertaken throughout the Australian community, and all are geared up to use the synchrotron for 2007 and beyond. And perhaps there is enough scope for us to take on some real ‘big picture science’ to match the structural genomics initiatives that are currently taking place overseas.
(Click on image for a larger version)
It just remains for me to thank my collaborators and the people who have worked so hard within the laboratory. This is a fantastic collaboration with James McCluskey and his team at the University of Melbourne. Also listed here are the people at Monash working in this particular program; there is a collaboration with Scott Borrows at the Queensland Institute of Medical Research as well. I would also like to thank Monash University for their generous support of my research program, the Wellcome Trust, the National Health and Medical Research Council and the Australian Research Council.
