TCRs recognize peptides displayed by MHC molecules

T cells generally recognize fragments of degraded proteins (peptides), which must be bound to (‘presented by’) specialized antigen-presenting molecules encoded by the major histocompatibility complex (MHC). In humans the MHC was first identified as the human leukocyte antigen (HLA) locus so the molecules may also be called HLA molecules.

MHC molecules can display a wide range of peptides, which have been derived from intracellular proteins, on the cell surface thereby alerting the immune system to the presence of intracellular invaders.

Because MHC molecules are expressed only on the cell surface, engagement of the TCR occurs only in the context of intimate cell–cell interactions.

When a primed T cell bearing an appropriate TCR (capable of recognizing, for example, a particular viral peptide) comes into contact with an infected cell, it can rapidly kill that cell and thereby limit the spread of the viral infection.

The requirement for cell–cell interaction also allows T cells to provide critical regulatory functions. T cells, for example, can couple signals from the innate immune system, such as those arising from ‘professional’ APCs (e.g. dendritic cells) with adaptive (T and B cell) responses, providing a critical layer of integration, coordination, and regulation.

TCRs are similar to immunoglobulin molecules

The TCR was identified much later than immunoglobulin, even though early theoretical considerations strongly suggested that T cells must bear cell surface antigen receptors. It is now clear that there are two varieties of TCR, termed αβ and γδ, and that both molecules:

Fig. 5.1 Similarities and differences between T cell receptors and immunoglobulins

TCRs are very similar to Fab fragments of B cell receptors. Both receptor types are composed of two different peptide chains and have variable regions for binding antigen, constant regions, and hinge regions. The principal differences are that TCRs remain membrane-bound and contain only a single antigen-binding site.

The two types of TCR tend to populate different tissue sites and are thought to perform distinct functions.

The αβ heterodimer is the antigen recognition unit of the αβ TCR

The αβ TCR is the predominant receptor found in the thymus and peripheral lymphoid organs of mice and humans. It is a disulfide-linked heterodimer of α (40–50 kDa) and β (35–47 kDa) subunits and its structural features have been determined by X-ray crystallography (Fig. 5.2).

Fig. 5.2 The T cell antigen receptor

Three-dimensional structure of an αβ TCR – only extracellular domains are shown. The α chain is colored blue (residues 1–213), and the β chain is colored green (residues 3–247). The β strands are represented as arrows and labeled according to the standard convention used for the immunoglobulin fold. The disulfide bonds (yellow balls for sulfur atoms) are shown within each domain and for the C terminal interchain disulfide. The complementarity determining regions (CDRs) lying at the top of the diagram are numerically labeled (1–4) for each chain. These form the binding site for antigen/MHC molecule.

(Adapted from Garcia KC, Degano M, Stanfield RL, et al. Science 1996;274:209–219. Copyright AAAS.)

Each polypeptide chain of the αβ TCR contains two extracellular immunoglobulin-like domains of approximately 110 amino acids, anchored into the plasma membrane by a transmembrane domain that has a very short cytoplasmic tail.

The extracellular portions of the α and β chains fold into a structure that resembles the antigen-binding portion (Fab) of an antibody (see Fig. 3.9). Indeed, as in antibodies, the amino acid sequence variability of the TCR resides in the N terminal domains of the α and β (and also the γ and δ) chains.

The regions of greatest variability correspond to immunoglobulin hypervariable regions and are also known as complementarity determining regions (CDRs). They are clustered together to form an antigen-binding site analogous to the corresponding site on antibodies (see Fig. 5.2). Note, however, that:

The disulfide bond that links the α and β chains is in a peptide sequence located between the constant domain of the extracellular portion of the receptor and the transmembrane domain (shown as the C terminal residue in the α and β chain in Fig. 5.2).

One remarkable feature of the transmembrane portion of the receptor is the presence of positively charged residues in both the α and β chains. Unpaired charges would be unfavorable in a transmembrane segment. Indeed, these positive charges are neutralized by assembly of the complete TCR complex, which contains additional polypeptides bearing complementary negative charges (see below).

The CD3 complex associates with the antigen-binding αβ or γδ heterodimers to form the complete TCR

The αβ or γδ heterodimers must associate with a series of polypeptide chains collectively termed the CD3 complex for the antigen-binding domains of the TCR to form a complete, functional receptor that is stably expressed at the cell surface and is capable of transmitting a signal upon binding to antigen.

The four members of the CD3 complex (γ, δ, ε, and ζ) are sometimes termed the invariant chains of the TCR because they do not show variability in their amino acid sequences. (The γ and δ chains of the CD3 complex should not be confused with the quite distinct antigen-binding variable chains of the TCR that bear the same names.)

The CD3 chains are assembled as heterodimers of γε and δε subunits with a homodimer of ζ chains, giving an overall TCR stoichiometry of (αβ)₂, γ, δ, ε₂, ζ₂. Current data suggest that the TCR complex exists as a dimer (Fig. 5.3).

Fig. 5.3 The T cell receptor complex

The TCR α and β (or γ and δ) chains each comprise an external V and C domain, a transmembrane segment containing positively charged amino acids, and a short cytoplasmic tail. The two chains are disulfide linked on the membrane side of their C domains. The CD3 γ, δ, and ε chains comprise an external immunoglobulin-like C domain, a transmembrane segment containing a negatively charged amino acid, and a longer cytoplasmic tail. A dimer of ζζ, ηη, or ζη is also associated with the complex. Several lines of evidence support the notion that the TCR–CD3 complex exists at the cell surface as a dimer. The transmembrane charges are important for the assembly of the complex. A plausible arrangement that neutralizes opposite charges is shown.

The CD3 γ, δ, and ε chains are the products of three closely linked genes, and similarities in their amino acid sequences suggest that they are evolutionarily related.

Indeed, all three are members of the immunoglobulin superfamily, each containing an external domain followed by a transmembrane region and a substantial, highly conserved, cytopasmic tail of 40 or more amino acids.

As with the transmembrane domains of the variable chains of the TCR, the membrane-spanning regions of these CD3 chains contain charged amino acids.

The negatively charged residues in the transmembrane region of the CD3 chains interact with (and neutralize) the positively charged amino acids in the αβ polypeptides, leading to the formation of a stable TCR complex (see Fig. 5.3).

The CD3ζ gene is on a different chromosome from the CD3γδε gene complex, and the ζ protein is structurally unrelated to the other CD3 components. The ζ chains possess:

An alternatively spliced form of CD3ζ, called CD3η, possesses an even larger cytoplasmic tail (42 amino acids longer at the C terminus).

The γδ TCR structurally resembles the αβ TCR but may function differently

The overall structure of the γδ TCR is similar to that of its αβ counterpart. Each chain is organized into:

One indication that the two types of T cell (i.e. T cells with αβ TCRs and T cells with γδ TCRs) might perform different functions comes from their anatomic distribution:

It is further believed that there are distinct subsets of γδ T cells that can perform different functions.

TCR variable region gene diversity is generated by V(D)J recombination

As with antibody genes, a highly diverse repertoire of TCR variable region genes is generated during T cell differentiation by a process of somatic gene rearrangement termed V(D)J recombination (see Chapter 3). Variable (V), joining (J), and sometimes diversity (D) gene segments are joined together to form a completed variable region gene.

Junctional diversity (imprecise joining of V, D, and J with loss and/or addition of nucleotides) contributes an enormous amount of variability to the TCR repertoire in addition to the variation that results from combinatorial assortment of the various gene segments.

TCRs are encoded by several sets of genes

The general arrangement of the genes encoding the α, β, γ, and δ chains of the TCR is remarkably similar to that of the immunoglobulin heavy chain genes (see Figs 3.21–3.23), suggesting a common origin from a primordial rearranging antigen receptor locus.

Figure 5.w1 illustrates the murine TCR genes, which are similar to those of humans. All four TCR gene families have been strongly conserved across more than 400 million years of evolution of the jawed vertebrates, which suggests a strong selective pressure for the preservation of both αβ and γδ T cell functions.

Fig. 5.w1 Murine T cell recpetor genes

The δ chain loci are embedded within the α loci and tandem duplication has occurred in the β chain loci. The last of each set of Jβ genes and the Vγ3 gene are pseudogenes.

The nomenclature of the TCR genes is simple – the TCRA locus encodes the α gene, TCRB the β gene, and so on. Interestingly, the TCRD locus is nested within the TCRA cluster.

The α and γ loci have sets of V and J gene segments (analogous to immunoglobulin light chain loci), whereas the β and δ loci have V, D, and J gene segments (analogous to immunoglobulin heavy chains).

TCR V genes used in the responses against different antigens

A major area of research in recent years has been to determine which sets of TCR V genes are used in the responses against different antigens. Because T cells recognize antigenic peptides bound to a particular MHC molecule this depends on:

Once the TCRs have been generated, T cells are subjected to thymic selection and may be further selected by interactions with APCs in the periphery. For these reasons, even if TCRs are generated by random recombination of gene segments, the expressed repertoire will be skewed toward the use of particular gene segments. Furthermore, the preference for different V gene segments by distinct T cell subsets may reflect their ontogeny. For example:

It is thought that these populations arise at distinct stages of intrathymic T cell development, and they appear to have distinct functions.

Recognition by the αβ TCR requires antigen to be bound to an MHC molecule

As discussed in Chapter 3, antibodies recognize intact proteins, binding either to linear epitopes derived from contiguous amino acids or discontinuous epitopes produced by amino acids that are not near one another in the primary structure but are brought together in the three dimensional structure of the protein. In contrast, T cell receptors only recognize linear epitopes present in the form of short peptides which are generated by degradation of intact proteins within the cell (a process termed antigen processing). This property of T cell receptors is critical to the way the immune system recognizes intracellular pathogens and tumor antigens. Such antigens present a special challenge to the immune system: intracellular antigens are hidden within the cell, and are not available for recognition by antibodies. How can intracellular antigens be recognized by extracellular receptors? The immune system has solved this problem by evolving an elegant means of displaying internal antigens (including those of intracellular pathogens) on the cell surface, allowing their recognition by T cells:

This antigen processing and presentation pathway, upon which both activation and regulation of the immune response rests, is a complex and fascinating subject (see Chapter 8).

In humans the MHC is known as the HLA

The proteins responsible for presenting antigens to T cells, MHC class I and class II proteins, were originally discovered as histocompatibility (transplantation) antigens. Histocompatibility refers to the ability to accept tissue grafts from an unrelated donor.

The major histocompatibility complex locus (MHC) comprises over 100 separate genes and was discovered when it was recognized that both donor and recipient had to possess the same MHC haplotype to avoid graft rejection.

The principal moieties that determine rejection were identified as MHC class I and class II molecules (see below), but we know now that the main purpose of the MHC is not to prevent graft rejection. The remaining genes in the MHC (sometimes called class III) are very diverse. Some encode:

There are no functional or structural similarities between these other gene products.

All mammalian species possess the MHC, though details of the complex vary from one species to the next. In humans the locus is known as the HLA (an abbreviation for human leukocyte antigen); in mice it is known as the H-2 locus (Fig. 5.4).

Fig. 5.4 Organization of the murine and human MHCs

Diagram showing the locations of subregions of the murine and human MHCs and the positions of the major genes within these subregions. The human organization pattern, in which the class II loci are positioned between the centromere and the class I loci, occurs in every other mammalian species so far examined. The regions span 3–4 Mbp of DNA.

MHC molecules provide a sophisticated surveillance system for intracellular antigens

From a cell’s perspective, there are two types of antigen that must be dealt with:

MHC class I molecules handle endogenous (or intrinsic) antigens, while MHC class II molecules handle exogenous (extrinsic) antigens. In both cases, the antigenic peptides are produced by proteolytic processing of proteins.

In general:

MHC class I molecules consist of an MHC-encoded heavy chain bound to β₂-microglobulin

The overall structure of the extracellular portion of an MHC class I molecule is depicted in Figure 5.5. It comprises a glycosylated heavy chain (45 kDa) non-covalently associated with β₂-microglobulin (12 kDa), which is a polypeptide that is also found free in serum.

Fig. 5.5 A model of an MCH class I molecule

The peptide backbone of the extracellular portion of HLA-A2 is shown. The three globular domains (α₁, α₂, and α₃) of the heavy chain are shown in green or turquoise and are closely associated with the non-MHC-encoded peptide, β₂-microglobulin (β₂m, gray). β₂-Microglobulin is stabilized by an intrachain disulfide bond (red) and has a similar tertiary structure to an immunoglobulin domain. The groove formed by the α₁ and α₂ domains is clearly visible.

The class I heavy chain consists of:

The three extracellular domains each comprise about 90 amino acids:

The extracellular portion of the class I heavy chain is glycosylated, the degree of glycosylation depending on the species and haplotype.

The predominantly hydrophobic transmembrane region comprises 25 amino acid residues and traverses the lipid bilayer, most probably in an α-helical conformation.

The hydrophilic cytoplasmic domain, 30–40 residues long, may be phosphorylated in vivo.

Heavy chain α1, and α2 domains form the antigen-binding groove

X-ray crystallography has shown that the α₁ and α₂ domains constitute a platform of eight antiparallel β strands supporting two α helices (Fig. 5.6). The disulfide bond in the α₂ domain connects the N terminal β strand to the α helix of the α₂ domain. A long groove separates the a helices of the α₁ and α₂ domains.

Fig. 5.6 The antigen-binding site of the MHC class I molecule HLA-A2

The view of the peptide antigen-binding groove in HLA-A2 as ‘seen’ by the TCR. The α₁ and α₂ domains each consist of four antiparallel β strands followed by a long helical region. The domains pair to form a single eight-stranded β sheet topped by α helices. The locations of the most polymorphic residues are highlighted. Residues around the binding site are highly polymorphic. For example, HLA-2 and HLA-Aw68 differ from each other by 13 amino acid residues. Ten of these differences occur around the antigen-binding site (yellow).

(Modified from Bjorkman et al. Nature 1987;329:512–516, with additional data from Parham P. Nature 1989;342:617–618.)

The original crystal structure of the HLA-A2 molecule revealed diffuse extra electron density in the groove, suggesting the presence of bound peptide antigen. This interpretation was supported by the observation that the majority of polymorphic residues and T cell epitopes on class I molecules are located in or near the groove.

Variations in amino acid sequence change the shape of the binding groove

Comparison of the structures of HLA-A2 and HLA-Aw68 have further refined our understanding of the structural basis for the binding of peptide to class I antigens.

The differences between HLA-A2 and HLA-Aw68 result from amino acid side-chain differences at 13 positions:

• six in α₁;

• six in α₂; and

• one (residue 245, which contributes to interactions with CD8) in α₃.

Ten of the twelve differences between HLA-A2 and HLA-Aw68 are at positions lining the floor and side of the peptide-binding groove (see Fig. 5.6). These differences give rise to dramatic differences in the shape of the groove and on the antigen peptides that it will bind.

Seen in detail, the peptide-binding groove forms a number of ridges and pockets with which amino acid side chains can interact. Typically the groove of an MHC class I molecule will accommodate peptides of eight or nine residues.

Amino acid variations within the peptide-binding groove can vary the positions of the pockets, providing the structural basis for differences in peptide-binding affinity that in turn govern exactly what is presented to a T cell (Fig. 5.7).

Fig. 5.7 Peptide-binding grooves of HLA-Aw68 and HLA-A2

The shapes of the antigen-binding groove on each molecule are illustrated. Differences in amino acids around the groove create different antigen-binding sites. For example, residues around position 45 produce a methionine-binding pocket in both molecules, but the aspartate-binding pocket around residue 74 is present only in HLA-Aw68.

MHC class II molecules resemble MHC class I molecules in their overall structure

The products of the MHC class II genes are:

These products are heterodimers of heavy (α) and light (β) glycoprotein chains, and both chains are encoded in the MHC:

A number of lines of evidence indicate that the α and β chains have the same overall structures. An extracellular portion comprising two domains (α₁ and α₂ or β₁ and β₂) is connected by a short sequence to a transmembrane region of about 30 residues and a cytoplasmic domain of about 10 to 15 residues.

The α₂ and β₂ domains are similar to the class I α₃ domain and β₂-microglobulin, possessing the structural characteristics of immunoglobulin constant domains.

The β₁ domain contains a disulfide bond, which generates a 64 amino acid loop.

The difference in molecular weights of the class II α and β chains is due primarily to differential glycosylation:

The β₂ domain does, however, contain a binding site for CD4, and MHC class II molecules on APCs interact with CD4 on T cells in a manner analogous to the interaction of MHC class I molecules with CD8. CD4 and CD8 are important elements in the recruitment of kinases that signal T cell activation.

Despite the differences in length and organization of the polypeptide chains, the overall three-dimensional structure of MHC class II molecules is very similar to that of MHC class I molecules (Fig. 5.8).

Fig. 5.8 Comparison of the extracellular domains of class I and class II molecules

Ribbon diagrams of the extracellular domains of class I HLA-Aw68 (1) and class II HLA-DR1 (2) MHC molecules. The binding cleft is shown with a resident peptide. These diagrams emphasize the similarity in the three-dimensional structures of class I and class II molecules.

(Redrawn from Stern LJ. Structure 1994;2:245–251. Copyright 1994 with permission from Elsevier.)

The MHC class II binding groove accommodates longer peptides than MHC class I

The structures of MHC class I and class II molecules reflect their functional differences.

The binding groove of MHC class II molecules is more open than that of MHC class I molecules, to accommodate longer peptides (Figs 5.9 and 5.10):

Fig. 5.9 Peptide-binding sites of class I (H-2K^b) and class I (HLA-DR1) MHC molecules

The peptide binding sites of class I (H-2K^b) and class II (HLA-DR1) MHC molecules are shown as α carbon atom traces in a top view of the peptide-binding clefts. The similarities between the two sites can clearly be seen, but there are also some differences, some of which account for the difference in peptide length preference between class I (8–10 amino acid residues) and class II (> 12 amino acid residues).

(Redrawn from Stern LJ. Structure 1994;2:245–251. Copyright 1994 with permission from Elsevier.)

Fig. 5.10 Peptides bind non-covalently within the antigen-binding groove

The hydrogen bonds made by the main chain of a bound peptide with class I (HLA-B27) or class II MHC molecules (HLA-DR1) are shown. The major difference between the two hydrogen-bonding patterns is the clustering of conserved class I hydrogen bonds at the ends of the peptide. By contrast, conserved class II hydrogen bonds are distributed throughout the length of the peptide.

(Redrawn from Stern LJ. Structure 1994;2:245–251. Copyright 1994 with permission from Elsevier.)

The structural features of the class II antigen-binding site have been illuminated by determination of the crystal structure of HLA-DR1 complexed with an influenza virus peptide. Pockets clearly visible within the peptide-binding site accommodate five side chains of the bound peptide and explain the peptide specificity of HLA-DR1.

The precise topology of the MHC peptide-binding groove depends partly on the nature of the amino acids within the groove, and thus varies from one haplotype to the next.

Which peptide can bind to a particular MHC molecule depends on the nature of the side chains of the peptide and their complementarity with the MHC molecule’s binding groove. Some amino acid side chains of the peptide stick out of the groove and are available to contact the TCR.

Peptides are held in MHC molecule binding grooves by characteristic anchor residues

It is possible to purify and sequence peptides that have been generated by a cell and then bound by MHC molecules at the cell surface. These peptides include:

Interactions at the N and C termini confine peptides to the binding groove of MHC class I molecules

A number of peptides bound by particular MHC molecules have been sequenced, and characteristic residues identified – one at the C terminus and another close to the N terminus of the peptide. These characteristic motifs distinguish sets of binding peptides for different MHC class I molecules (Fig. 5.11).

Fig. 5.11 Allele-specific motifs in peptides eluted from MHC class I molecules

Class I MHC molecules from either H-2K^b or H-2K^d haplotypes were immunoprecipitated. Peptides bound to these molecules were purified and sequenced. Amino acid residues commonly found at a particular position are classified as ‘dominant’ anchor residues. Residues that are fairly common at a site are shown as ‘strong’. Positions for which no amino acid is shown could be occupied by several different amino acids with equal frequency. The one-letter amino acid code is used. The diagram represents the class I MHC molecule binding groove, viewed from above with anchor positions of each haplotype highlighted.

The significance of the conserved residues has become clear by analysis of the three-dimensional structures of several MHC class I molecules, which have generated a clear picture of the peptide residing in the binding groove:

• the ends of the peptide-binding groove are closed;

• the peptide is an extended (not α-helical) chain of nine amino acids and the N and C terminals are buried at the ends of the groove;

• some of the side chains extend into the pockets formed within the variable region of the class I heavy chain;

• numerous hydrogen bonds are formed between residues in the class I molecule and those of the peptide along its length;

• in particular, tyrosine residues commonly found at the N terminus of the peptide and a conserved lysine in the MHC class I molecule binding groove stabilize peptide binding (see Fig. 5.10);

• the centers of the peptides bulge out of the groove, so presenting different structures to TCRs.

This picture is consistent with the characteristic motifs found at the ends of peptides eluted from class I molecules.

Peptides may extend beyond the ends of the binding groove of MHC class II molecules

The binding groove of the MHC class II molecule:

• also incorporates a number of binding pockets, though the locations are somewhat different from that on class I molecules;

• is not closed at the ends, so bound peptides extend out of the ends of the groove.

Consistent with this observation, peptides eluted from MHC class II molecules tend to be longer (over 15 residues).

Conserved anchor residues in peptides eluted from MHC class II molecules have been identified (Fig. 5.12).

Fig. 5.12 Allele-specific motifs in peptides eluted from MHC class II molecules

HLA-DR molecules of two haplotypes (DRB1*0401 and DRB1*1101) were purified and incubated with a library of peptides (generated in phage M13). After multiple rounds of selection, peptides that bound effectively to the MHC class II molecules were identified and sequenced. Residues having a frequency greater than 20% are shown as ‘dominant’ anchor residues. Other fairly common residues are shown as ‘strong’. Note that the binding site on MHC class II molecules accommodates longer peptides than that on MHC class I molecules.

(Data abstracted from Hammer J, Valsasnini P, Tolba K, et al. Cell 1993;74:197–203.)

Peptides binding MHC class II are less uniform in size than those binding MHC class I molecules

A major difference between MHC class I and class II molecules occurs at the ends of the peptide-binding groove:

The peptides that bind to MHC class I molecules come from endogenous proteins synthesized within the cell, which are broken down and transported to the endoplasmic reticulum. The mechanism of antigen processing is explained fully in Chapter 8, but it should be noted here that the internal antigen processing pathways generally produce peptides of an appropriate size to occupy the MHC class I molecule antigen-binding groove.

Peptides that bind to MHC class II molecules come from exogenous antigens – proteins that have been internalized by the cell and then degraded. These peptides are less uniform in size than those that bind to MHC class I molecules and may be trimmed once they have found their way to the MHC class II molecule.

The MHC class II molecule antigen processing pathway is quite distinct from the MHC class I molecule pathway (see Chapter 7).

Aggregation of TCRs initiates T cell activation

Although basic models of T cell activation by antigen–MHC show one receptor being triggered by one complex (e.g. Fig. 1.12), this is simplistic.

Each T cell may express 10⁵ receptors and each APC has a similar number of MHC molecules. If a T cell engages an APC, only a tiny proportion of the MHC–antigen complexes on its surface will be of the correct type to be recognized by the T cell.

What, then, is the minimum signal for T cell activation? In practice:

The model of the TCR shown in Figure 5.3 suggests that it can form a dimeric structure clustered around the signaling molecules of the CD3 complex.

Interestingly, there is some evidence that MHC molecules can also dimerize, and that TCRs bound to MHC–peptide complexes tend to form dimers or aggregates. These observations have led to the view that T cell activation requires the cooperative aggregation of specific TCRs with MHC–peptide complexes.

The auxiliary molecules CD4 and CD8 are also important in T cell activation, and the presence of CD4 or CD8 can help stabilize the interaction of the TCR and MHC–peptide. In addition, kinases associated with these molecules are brought into proximity with CD3 so they can phosphorylate the ζζ dimer that initiates activation (Fig. 5.14). The ensuing steps are described in Chapter 8.

Fig. 5.14 The role of CD4 and CD8 in T cell activation

After aggregation of MHC–peptide on the APC with the TCR, either CD4 or CD8 can join the complex. CD4 binds to MHC class II molecules and CD8 to MHC class I molecules. The kinase Lck is attached to the intracytoplasmic portion of CD4 or CD8. Binding of the CD4 or CD8 molecules to specific sites of the MHC molecules brings the kinase into proximity with the ITAMs on the CD3 ζζ dimer (see Fig. 5.3). The kinase phosphorylates the motifs, as the first step in T cell activation.

Antigenic peptides can induce or antagonize T cell activation

The affinity of TCRs for antigen peptide–MHC, expressed as an association constant, is typically of the order of 10^− 5 to 10^− 6 M, which is much lower than the typical affinity of an antibody for its epitope.

The affinity of antigen–MHC for the TCR is especially important because it determines the degree to which the T cell becomes activated. Peptides that activate a T cell are called agonist peptides. By changing one or two amino acids in an agonist peptide, it is possible to generate peptides that antagonize the normal activation by the original peptide. The antagonist peptides typically have a lower affinity for the TCR than the original agonist peptide and are thought to act by either:

Occasionally, a modified peptide will be more effective than the agonist in activating the T cell. These strong agonist (or superagonist) peptides produce higher affinity binding in the TCR–MHC–peptide complex.

The distinctions between these three different types of peptide (agonist, antagonist, and strong agonist) have biological ramifications. Researchers are attempting, for example, to use antagonist peptides to block adverse immune responses.

Understanding TCR affinity could provide insight into how T cells become tolerant to self antigens during development (see Chapter 19).

What constitutes T cell specificity?

Finally we should consider the question of what constitutes T cell specificity. When the specificity of lymphocytes was first explained in Chapter 1, it was stated that each lymphocyte binds to just one antigen using its receptor. Although this is a useful starting point for understanding immune responses, it is not strictly true.

In Chapter 3 it was seen that immunoglobulins could bind to different antigens if they had epitopes that were sufficiently similar. This chapter has shown that antigenic peptides can be mutated and still bind and trigger T cell activation.

One question, then, is how far a peptide can be mutated and still bind to its own TCR. In some cases it has been possible to individually change each amino acid in a peptide without destroying its ability to bind to the MHC molecule or the TCR.

Provided the peptide can form part of the TCR–MHC–peptide complex and provide sufficient binding energy, its precise amino acid sequence does not matter. This is a very important observation.

Later, when we consider how antigenic peptides from microorganisms can trigger autoimmune diseases (see Chapter 20), we learn that one possible mechanism is that a self peptide and a foreign peptide are sufficiently similar to bind to the same T cell, so causing a breakdown in self tolerance. One conclusion from the work cited above is that such peptides do not have to be identical to be cross-reactive.

The three principal human MHC class I loci are HLA-A, HLA-B, and HLA-C

The human MHC class I region contains three principal class I loci – called HLA-A, HLA-B, and HLA-C. Each locus encodes the heavy chain of a classical MHC class I molecule and the whole region:

Fig. 5.15 Genes within the human MHC class I region

The human MHC class I region lies telomeric to the MHC class II region (see Fig. 5.5). In addition to the genes encoding the classic transplantation antigens (HLA-A, HLA-B, and HLA-C), several other principal class I-like genes have been identified (HLA-E, HLA-F, and HLA-G). Mutations in the HLA-H gene are associated with hemochromatosis, a disorder that causes the body to absorb excessive amounts of iron from the diet. A number of other non-classical class I genes and pseudogenes are present, mostly with unknown functions.

Closer analysis of this region has revealed multiple additional MHC class I genes.

Mice have two or three MHC class I loci

The mouse MHC (H-2) has two or three MHC class I loci, but the number of MHC class I genes varies between haplotypes (Fig. 5.w2).

Fig. 5.w2 Genes within the murine class I region

The organization of the MHC class I region of two haplotypes, BALB/c (H-2^d) and B10 (H-2^b), is shown. The class II region lies between the H-2K and H-2D regions. The brackets denote gaps added to align alleles between the two haplotypes. H-2^d and H-2^b haplotypes have different numbers of class I genes in the H-2D/H-2L region. Five class I genes map to the D/L region of BALB/c (H-2^d) mice. Two of these genes encode the serologically detectable H-2D^d and H-2L^d antigens. Three additional class I genes are found in the region between the proximal H-2D^d and distal H-2L^d genes. These genes, called D2^d, D3^d, and D4^d, are of unknown function. Only one class I gene has been identified in the H-2D region of B10 mice. The Q, T, and M regions contain a large number of class Ib genes, the functions of which are mostly unknown.

MHC class I genes involved in antigen presentation to T cells are located in H-2K, H-2D, and H-2L loci.

The organization of the H-2K region is similar in all strains that have been studied. It contains two class I genes, termed K and K2:

The numbers of genes in the H-2D and H-2L loci are variable (see Fig. 5.w2).

Qa, Tla, and M genes encode MHC class Ib molecules

The Qa, Tla, and M genes of the mouse MHC class I region (see Figs 5.4 and 5.w2) encode class Ib molecules, of which the functions are mostly unknown:

The class Ib genes exhibit a low degree of polymorphism.

Human MHC class II genes are located in the HLA-D region

The human MHC class II region spans about 1000 kb of DNA, and the order and orientation of the various loci are similar to that of the homologous loci in the mouse MHC class II region.

The HLA-D region encodes at least six α and ten β chain genes for MHC class II molecules (Fig. 5.16). Three loci (DR, DQ, and DP) encode the major expressed products of the human MHC class II region, but additional genes have also been identified:

Fig. 5.16 Genes within the human and mouse class II regions

The arrangement of the genes within the human and murine MHCs is shown. Homologous genes between the two species are indicated. Expressed genes are colored orange and pseudogenes are shown yellow. Mice of the b, s, f, and q haplotypes fail to express class II-E molecules. The b and s haplotypes fail to transcribe the Ea gene, but make normal cytoplasmic levels of Eb chain. Mice of f and q haplotypes fail to make both Ea and Eb chains.

DR, DQ, and DP α chains associate in the cell primarily with β chains of their own loci. For example:

The organization and length of the DRB region varies in different haplotypes (Fig. 5.17), with different numbers of β chains expressed.

Fig. 5.17 The number of DRB loci varies with different haplotypes

The numbers of DRB loci varies between individuals. For example, a person who has a haplotype producing molecules of the type DR1 (see Appendix 1) has three loci for DRB (top line). Not all of these loci produce mRNA for DR β chains.

The MHC class II region also contains genes that encode proteins involved in antigen presentation that are not expressed at the cell surface (see Chapter 8).

Mouse MHC class II genes are located in the H-2I region

The α and β chains of mouse MHC class II molecules are encoded by separate genes located in the I region of the H-2 complex (see Figs 5.4 and 5.16). The gene nomenclature indicates first the locus and then the type of chain it encodes:

Several other class II b and a genes for which no gene product is known, have been cloned:

The Ob and Eb2 genes display a low level of polymorphism and are transcribed, but it is not known whether they are translated.

Different strains of mice vary in their MHC class II genes (see Fig. 5.16).

MHC polymorphism is concentrated in and around the peptide-binding cleft

A hallmark of the MHC is the extreme degree of polymorphism (structural variability) of the molecules encoded within it. The class Ib molecules are much less polymorphic than the classical class I and II molecules. New allelic variants are identified frequently and logged on dedicated databases (see Internet references).

Within a particular MHC class I or class II molecule, the structural polymorphisms are clustered in particular regions. The amino acid sequence variability in class I molecules is clustered in three main regions of the α₁ and α₂ domains. The α₃ domain appears to be much more conserved.

In MHC class II molecules, the extent of variability depends on the subregion and on the polypeptide chain. For example:

In outbred populations in which individuals have two MHC haplotypes, hybrid class II molecules with one chain from each haplotype can be produced. This generates additional structural diversity in the expressed molecules.

Most of the polymorphic amino acids in MHC class I and class II molecules are clustered on top of the molecule around the peptide-binding site (see Fig. 5.6). Variation is therefore centered in the base of the antigen-binding groove or pointing in from the sides of the α helices. This polymorphism affects the ability of the different MHC molecules to bind antigenic peptides.

All nucleated cells of the body express MHC class I molecules

The function of MHC class I molecules is to present antigens that have entered the cell, such as viral peptides. Because any cell of the body may become infected with a virus or intracellular pathogen, all cells need to sample their internal molecules and present them at the cell surface to cytotoxic T cells.

By contrast, MHC class II molecules are used by APCs to present antigens to helper T cells. Consequently the distribution of class II molecules is much more limited (see Fig. 8.4).

MHC molecules are co-dominantly expressed

This means that, in one individual, all of the principal MHC gene loci are expressed from both the maternal and paternal chromosomes. As there are three MHC class I loci in humans (HLA-A, HLA-B, and HLA-C), each of which is highly polymorphic, most individuals will have genes for six different class I molecules, all of which will be present at the cell surface. Each MHC molecule will have:

A similar logic applies to MHC class II molecules. There are three principal class II loci in humans (HLA-DP, HLA-DQ, and HLA-DR), all of which are polymorphic. At first sight, it would appear that an APC could express six different class II molecules as well as its class I molecules. However, this is probably an underestimate. As noted above, hybrid class II molecules (using one polypeptide encoded by the maternal chromosome and one by the paternal chromosome) also occur.

The specificity of the TCR and MHC explains genetic restrictions in antigen presentation

Much of the original work on antigen presentation was carried out using strains of mice that had been inbred to the point where both maternal and paternal chromosomes were identical. Any offspring therefore inherited the same set of autosomes from each parent, and the offspring were genetically identical to their parents. Clearly the level of diversity in the MHC molecules was much less than in an outbred human population.

The artificial simplicity of the inbred mouse system, however, allowed immunologists to dissect how antigens were presented to T cells in a whole animal, when the molecular structures of the MHC molecules and the TCR were completely unknown.

The key experiment that demonstrated the importance of the MHC in antigen presentation revealed a phenomenon called genetic restriction (also known as MHC restriction). In essence, it was noted that cytotoxic T cells from a mouse infected with a virus are primed to kill cells of the same H-2 haplotype infected with that virus; they do not kill cells of a different haplotype infected with the same virus (Fig. 5.18).

Fig. 5.18 MHC restriction of cytotoxic T cells

A mouse of the H-2^b haplotype is primed with virus and the cytotoxic T cells thus generated are isolated and tested for their ability to kill H-2^b and H-2^k cells infected with the same virus. The cytotoxic T cells kill H-2^b, but not H-2^k cells. In this instance, it is the H-2K class I gene product presenting the antigen to the T cells. The T cell is recognizing a specific structure produced by the association of a specific MHC molecule with a specific viral antigen.

These data, and similar experiments using APCs and helper T cells, showed that T cells that have been primed to recognize antigen presented on MHC molecules of one haplotype will normally respond again only when they see the same antigen on the same MHC molecule.

Peter Doherty and Rolf Zinkernagel performed the key experiments delineating the phenomenon of MHC restriction of T cell responses in the mid 1970s and were awarded the Nobel Prize in Physiology or Medicine in 1996 for this work.