Identification of potential CTL epitopes of bovine RSV using allele-specific peptide motifs from bovine MHC class I molecules

Veterinary Immunology and Immunopathology 54(1996) 21 I-219

EL-SEWER

Veterinary immunology and immunopathology

Identification of potential CTL epitopes of bovine RSV using allele-specific peptide motifs from bovine MHC class I molecules R.M. Gaddum a**, S.A. Ellis a, A.C. Willis b, R.S. Cook a, K.A. Shines a, L.H. Thomas a, G. Taylor a aInstitute b MRC Immunochemistry

j;7r Animal

Health,

Unit. Biochemistry

Compton.

Department.

Newhury

RG20 7NN. UK

South Parks Road. University

ofOxjbrd.

Oxford,

UK

Abstract Respiratory syncytial virus (RSV) is a major cause of lower respiratory tract infection in young infants and housed calves. Depletion of CD8+ lymphocytes from calves inhibited their ability to clear the virus from the nasopharynx and lungs. To study these cells further, a cytotoxic T lymphocyte (CTL) assay was established. CTL could be demonstrated in the peripheral blood of gnotobiotic calves 7-10 days post infection (p.i.) with RSV and in lungs 10 days p.i. This response was both MHC-restricted and virus-specific. Following separation of the lung lymphocytes by magnetic activated cell sorting, it was shown that the cytolytic activity was mediated by cells of the CD8+ phenotype. To identify epitopes recognised by bovine CTL, the consensus motifs from MHC class I alleles found in the herd at Compton were identified. cDNA libraries were constructed and screened for full length class I sequences. The isolated cDNA clones were then transfected into mouse P815 cells and the expressed product immunoprecipitated and matched with a serological specificity. The bovine MHC class I molecules were isolated from lysed transfected cells by affinity chromatography, using a monoclonal antibody specific for bovine MHC class I, and bound peptides were separated by reverse-phase HPLC. Analysis of the protein sequences of bovine RSV for the defined motifs has identified potential CTL epitopes. Kry%rzord.y: MHC class I; Peptide motifs; Cytotoxic T lymphocytes; Respiratory syncytial virus

* Corresponding author. 0165.2427/96/$15.00 PI/

Copyright 0 1996 Elsevier Science B.V. All rights reserved.

SO 165.2427(96)05686-3

1. introduction Cytotoxic T lymphocytes (CTL) play a decisive role in cell-mediated immunity against intracellular pathogens and tumour cells as they are capable of recognising and destroying cells expressing intracellular foreign proteins such as early viral antigens and viral regulatory proteins (Oldstone. 1991) when presented as short peptides in association with MHC class I molecules. Therefore. there is considerable interest in identifying effective ways of inducing protective CTL responses in vivo, as a means of vaccination. The ability to identify potential CTL epitopes would obviously contribute greatly to this area of research. Human and bovine RSV are the major cause of lower respiratory tract disease in infants and housed calves respectively (Stott and Taylor, 1985). Most severe disease occurs in both babies and calves during the first 6 months of life when maternal antibodies are present, and currently available vaccines are of limited efficacy. Studies of cell-mediated immune responses to human RSV have been carried out in a murine model; passive transfer of T cells and T cell depletion (Cannon et al., 1987; Graham et al.. 1991) have provided information on the role of the T cell subsets in immune protection. The CTL response has been analysed in detail and induction of CTLs which recognise the M2 protein correlate with resistance to RSV, although this is short-lived (Connors et al.. 1991; Kulkarni et al.. 1993a). However, mice are not a natural host for the virus, therefore may not be a very accurate model of the disease. RSV infection in immunodeficient humans can be very severe and this contrasts with studies carried out in mice where depletion of T cells abrogated disease (Graham et al.. 199 I). Depletion studies in calves have shown that CD8’ cells are essential for virus clearance. As part of a study to dissect the immune response to bovine RSV we have obtained evidence that a strong CTL response is mounted which is central to the role of recovery from infection. Molecular characterisation of class I molecules together with recently developed technologies for isolating and sequencing of MHC-bound antigenic peptides has allowed us to identify potential CTL epitopes from bovine RSV. By determining the peptide motif from a number of bovine class I alleles it has been possible to identify a series of potential CTL epitopes from bovine RSV and to increase our understanding of the role of the MHC in determining variability in the specificity of immune responses.

2. Association

of peptides with class 1 molecules

It is well-established that CTLs recognise processed foreign antigens as peptides. usually consisting of eight to ten amino acid residues. when associated with MHC class I molecules (Townsend and Bodmer, 1989). There are three main ways that peptides associating with the MHC class I molecules have been studied: I. X-ray crystallography of peptide-MHC complexes (Fremont et al.. 1992: Gorga et al., 1992; Madden et al.. 1993. erratum 1994): 2. peptide binding studies: 3. biochemical isolation and study of naturally MHC-associated peptides. The latter two methods are reviewed by Rammensee et al. ( 1995).

R.M. Gaddum et al./ Veterinary Immurwlogy und Immunopathology 54 (1996) 211-219

213

Table 1 Examples of MHC class I allele-specific motifs Allele

Ref.

Amino acid position 1

H-2K ’ H-2D b HLA-A2 HLA-B7 BoLA-A20 BoLA-A I 1

Falk et al., 1991 Falk et al., 1991 Falk et al., 1991 Huczko et al., 1993 Bamford et al., 1995 Hegde et al., 1995

2

3

4

5

Y N L P K P

6

7

8

9

IO

I/L M/I V L/F R I/V

The method by which viral peptides could be eluted from MHC I molecules (Van Bleek and Nathenson, 1990; Rotzschke et al., 1990) was adapted to elute the self-peptides from the MHC I molecules on the surface of P815 cells, a mouse mastocytoma cell line, by Falk et al. (1991). Sequencing the pool of peptides by Edman degradation led to the discovery of the principle of allele-specific motifs. The peptides binding to individual class I alleles have a distinct amino acid residue pattern. By studying the frequency of a particular residue at each position it is possible to determine whether the frequency is increased, suggesting preferential binding of that amino acid at that position. On occasions where the frequency greatly increased either exclusively for one amino acid or for two similar residues it was termed an anchor residue(s); in general, a motif consists of two or three anchor residues. In this way, a peptide motif can be derived for an individual MHC allele. Since 1991, the allele-specific motifs for over 40 human and mouse class I alleles have been identified (reviewed by Rammensee et al., 1995) and more recently for two bovine class I alleles (Bamford et al., 1995; Hegde et al., 1995) (Table 1).

3. Identifying

a CTL epitope

Identification of a CTL epitope has until now been based on methods using recombinant viruses expressing whole proteins, truncated proteins or mini-genes, deletion mutants or synthetic peptides (Anderson et al., 1988; Whitton et al., 1988; Kast et motifs has al., 1991; Lawson et al., 1994). The ability to determine allele-specific provided a new approach for the identification of CTL epitopes. Over 120 previously identified CTL epitopes have since been shown to contain a peptide-MHC motif (Rammensee et al., 1995). Therefore, it seems feasible that knowledge of the peptide motif should allow potential CTL epitopes to be identified by screening the primary amino acid sequences for these motifs. Indeed, this has been successfully used by many researchers (Table 2). Furthermore, peptides corresponding to CTL epitopes when used to vaccinate mice have been shown to protect against a number of viral pathogens, e.g. Sendai virus and HPV-16 induced tumours (Kast et al., 1991; Feltkamp et al., 1993). In particular, a peptide which sensitised murine target cells for lysis by M2 specific-CTL was identified following screening of the M2 amino acid sequence of human RSV for

214

R.M. tiddum

et cd./ Vrtrrrnury

Imnrunolo~y

und Immunoi,utholo,~y

54

f 1996)

21 I-219

Table 2 Correct identification of CTL epitopes based on the predicted class

I

MHC-peptide

motifs

Ref.

Pathogen and protein

Restricting element

Influenza A nucleoprotein

HLA-A

Hepatitis B virus surface antigen

HLA-A?

RSV M’2 (22K) protein

H-2 K ”

Kulkami et al., 1993b

Measles virus nucleoprotein

H-2 L. ”

Beauverger et al., 1994

I

DiBrino et al., 1993 Nayersina et al.. 1993

the H-2 Kd peptide motif (Kulkami et al.. 1993b). Immunization with a recombinant vaccinia virus, which encoded the identified M2 epitope, induced CTL and resistance to human RSV infection in BALB/c mice (Kulkami et al.. 1995). By determining allele-specific motifs for bovine MHC class I alleles. we hope to identify the proteins important for induction of a CTL response and identify peptides which could be effective in stimulating a protective CTL response against bovine RSV. This will also give us valuable information on the role of the MHC in determining the specificity of immune responses.

4. Characterisation

of bovine MHC class I molecules

The bovine class I MHC, referred to as BoLA. has provisionally been assigned to chromosome 23 (Fries et al.. 1989). More than 50 MHC class I specificities have been defined serologically (Davies et al., 1994). The number of transcribed class I loci is unclear. but the use of biochemical and molecular methods. including one-dimensionalIEF, peptide mapping and cDNA sequencing has suggested that at least three exist (Bensaid et al., 1991; Ellis et al., 1992: Garber et al.. 1994). Our own data, from studies of the IAH. Holstein-Friesian herd, indicate that the number of genes transcribed and expressed by individual haplotypes is variable. Furthermore. the alloantisera used for serological typing may recognise only one of the gene products of a haplotype. leaving the other(s) undetected: thus, animals which share the same tissue type may differ in one or more of their expressed class I genes. By the construction of bovine cDNA libraries from selected bulls within the IAH herd. it has been possible to identify a number of different clones that encode class I molecules. Following sequencing, these clones were transfected into P815 mouse mastocytoma cells and the successful expression of bovine class I confirmed by ID-IEF gels. Obtaining these transfectants allowed us to use the method of Falk et al. ( 199 1) to determine the peptide motif for each of the identified alleles. The transfectants (10”’ ) were lysed in 0.5% NP40 and passed over a Sepharose CL-4B column to which either glycine or a non-specific mAb was coupled. The MHC-peptide complexes were retained on an immuno-affinity column of CnBractivated-Sepharose coupled to the monoclonal antibody IL-A88 (a gift from ILRI. Kenya) which recognises a monomorphic determinant on bovine MHC class I heavy chains (Toye et al.. 1990). The MHC-peptide complexes were eluted from the IL-A88 column in 0.1% trifluoroacetic acid. The peptide pool was separated from the MHC class I heavy and light chains either by reverse phase HPLC or by using a 3 kDa cut-off

R.M. Gaddum et al./Veterinary

Immunology and Immurwpathology 54 (1996) 211-219

215

membrane filter (Amicon, USA). The peptide pool was then sequenced by Edman degradation and the results interpreted as described by Falk et al. (1991), a baseline level for each amino acid having been determined. Analysis of the transcribed class I genes identified in bull HD, which serologically typed as A18/A31, showed that only one gene from this bull was associated with the Al8 haplotype, whereas three were associated with the A31 haplotype; however, only two of the latter were expressed and only one of these was recognised by A31 specific alloantisera. Preliminary motifs have been obtained for these molecules. In collaboration with S. Srikumaran (University of Nebraska) we have been able to confirm the motif for a bovine allele which serologically types as Al 1 and runs in the same position, on a ID-IEF gel as the product of the recently sequenced Al 1 gene (Sawhney et al., 1995). A proline was an anchor residue at position 2 with either a valine or an isoleucine at position 9 (Hegde et al., 1995). At present, only one other peptide motif has been published for a bovine class I molecule, derived from an animal which typed serologitally as A20 and the sequence of this molecule was not determined (Bamford et al., 1995). The CTL response is known to be important in recovery (Taylor et al., 1995) and knowledge of these motifs has allowed us to identify potential CTL epitopes from bovine RSV.

5. CTL responses

to bovine RSV and identification

of potential

epitopes

In vivo depletion of T cell subsets of calves using monoclonal antibodies has demonstrated an important role for CD8+ T cells in virus clearance (Taylor et al., 1995). Furthermore, flow cytometry studies have shown an increase in the CD8+ cells in the lung and the trachea following infection with bovine RSV. A concurrent increase in IL-2R on T lymphocytes isolated from these two sites has indicated T cell activation (McInnes et al., in preparation). To study these cells further, a CTL assay for bovine RSV was established. A specific CTL response was detected in peripheral blood from 7 days after infection with bovine RSV and in lung lymphocytes from calves killed on day 10. Further characterisation showed the response to be MHC-restricted, virus-specific and mediated by cells of the CD8+ phenotype (Gaddum et al., 1996). The CTLs recognised two strains of bovine RSV which have differences in the G protein (Furze et al., 1994). The amino acid sequences for nine of the ten bovine RSV proteins are known, and these sequences have been screened for the peptide motifs obtained so far. This has identified a number of potential CTL epitopes of bovine RSV. Table 3 shows the results obtained using the Al 1 motif. In the absence of a bovine RSV sequence for the polymerase (L) protein, the human L protein has been screened, with a number of matches found. Based on information from previously identified CTL epitopes and subsequently established peptide motifs, it has been demonstrated that a certain amount of variation may occur, both in length, and in the amino acid which may form an anchor residue. Therefore, if a motif encodes an anchor residue with for example an isoleucine or a leucine at position 9, then 1Omer peptides with a fitting C-terminus should also be considered, as should other aliphatic residues at the C-terminus such as valine or

Table 3 Potential CTL epitopes of bovine RSV proteins presented by the bovine MHC class Bovine RSV protein

1allele

Al

I

Self peptide

* p * * * * * *

1

l

*p******v*

F protein

G protein M protein N protein L protein (Human)

FPQAETCKVQ LPATRKPPIN II P A S LT IPYAGLVLVI YPHYIDVFVH WPTLRNAIVL PPYTGDHIVD KPPIFTGDVD APPYIGDHIV T PL FLTEA LPPYKAEKIV

I W V P

I V

methionine. Allowing for these variations. approximately 40 peptides capable of binding to the Al 1 allele were identified from the nine bovine RSV protein sequences. Having identified all possible CTL epitopes it is necessary to test their ability to sensitise target cells for lysis by effector lymphocytes expressing the same class I molecules. We are currently setting up peptide CTL assays to determine bona fide epitopes. The lower the concentration of peptide required to sensitise a target cell for lysis, the higher the affinity of the peptide for the MHC molecule. CTL epitopes generally have a high binding affinity for MHC molecules. In some instances, preliminary MHC binding assays have been carried out to determine those peptides capable of binding the MHC. Not all the peptides will bind the MHC and not all those which bind MHC will be CTL epitopes; however. all peptides which are CTL epitopes will bind MHC. Biochemical methods using radiolabelled peptide, e.g. affinity columns (Buus et al., 1987) can be used to determine whether or not a peptide will bind MHC, thus reducing the number of peptides which require screening in a CTL assay. Alternative binding assays have employed mutant cell lines, e.g. RMA-S cells (Ljunggren and Karre, 1985) which have defects in their antigen-processing capacity. Under physiological conditions these cells express a low density of MHC class I on their cell surface. When transfected with the MHC gene of interest, and incubated with a peptide with a high affinity for class I, the MHC is stabilised and their surface density is increased to levels detectable with antibodies.

6. Limitations

of this method for predicting

CTL epitopes

Precise characterisation of the epitopes would allow identification of those proteins suitable for use in vaccines. Alternatively, use of the peptide epitopes themselves could be considered. This latter strategy relies on the availability of a sufficient number of allele-specific peptide-motifs being identified and a peptide vaccine is likely to include a cocktail of peptides which are capable of being recognised by the vast majority of

R.M. Gaddum et al./

Veterinary

Immunology

and lmmutwpathology

54 (1996) 21 I-219

217

bovine alleles expressed within a herd. In this way a protective immunity may be induced in most, if not all, animals. Preliminary data from our Holstein/Friesian herd at IAH, suggest that a large proportion of the herd express one or more of ten serological specificities (unpublished data, 1995). Although cattle are outbred, one or more of a relatively small number of commonly expressed class I molecules may be present in the vast majority, making determination of allele-specific motifs a suitable method for studying CTL epitopes, at least within the Holstein/Friesian population. Further studies of bovine class I molecules and associated peptides are required to determine the feasibility of this approach. The method of prediction described in this paper takes account only of the primary peptide structure, although it is widely accepted that secondary structure mediated by non-anchor residues can influence the binding of a peptide to a class I molecule (Gairin and Oldstone, 1993; Ruppert et al., 1993; Kast et al., 1994). The presence of anchor residues alone is not sufficient to achieve high binding of the peptide to class I (Kast et al., 1994). For this reason, it would be advisable to carry out preliminary MHC-peptide binding studies to reduce the number of potential CTL epitopes. Some peptides which possess the appropriate consensus motif and are able to bind class I molecules in binding assays may not be naturally produced owing to a series of selection procedures which operate during antigen processing. For example, the cleavage of polypeptides by proteasomes generally occurs after hydrophobic or charged residues. The exact kind of LMP influence on specificity is controversial (Howard and Seelig, 1993). Selection has also been exhibited by the TAP transporter molecules which transport peptides into the ER (Neefjes and Momburg, 1993). Mouse TAP prefers peptides with hydrophobic C-termini whereas human TAP molecules are permissive for a range of different peptides. Whether similar methods for selection operate within bovine cells is unknown. Not all possible CTL epitopes may be identified using this method as a number of reported CTL epitopes fail to contain the predicted motif (Stauss et al., 1992; Stuber et al., 1992; Calin-Laurens et al., 1993; Frumento et al., 1993; Mandelboim et al., 1994). It is also possible that an identified peptide may be unstable or particularly susceptible to serum proteolytic degradation (Widmann et al., 1991). The peptide may not be immunogenic; for example, a peptide from hen egg lysozyme containing a Kb motif which increased class I expression on the surface of RMA-S cells was not immunogenic. This was possibly because the peptide resembled a self peptide, to which the host was probably tolerant (Calin-Laurens et al., 1993). It is certain that knowledge of the peptides which associate with bovine class I will advance our understanding of class I-restricted immune responses to bovine pathogens and therefore increase the chance of successful vaccine development. Furthermore, it may be necessary to understand and include a means of inducing secondary signals, e.g. cytokine production, accessory molecule stimulation, within the same vaccine.

7. Conclusion A method for determining the consensus peptide motifs of bovine class I alleles has been established and we have identified a number of potential CTL epitopes to bovine

RSV which we soon hope to confirm as bona fide epitopes in CTL assays. Knowledge of these peptide motifs for bovine class I alleles will enable us to identify potential CTL epitopes from the amino acid sequences of any bovine pathogen. This will be of particular interest for pathogens which require involvement of a CTL response for either recovery from. or protection against. infection. In addition, the information obtained may help to identify those residues in sequenced class I molecules that determine the binding of peptide anchor residues.

8. Note added in proof Since submission of this paper we have identified three more peptide motifs from bovine MHC class I molecules (Gaddum et al.. 1996)

Acknowledgements We would like to thank Dr M. Stear. University Veterinary School, Glasgow serologically typing all the calves and transfectants used in this study.

for

References Anderson, R.W.. Tevethla. M.J.. Kalderon. D.. Sm~lh. A.E. and Tevethia, S.S.. 1988. J. Viral., Bamford. A.I.. Douglas, Friede. T., Stevanovic. S.. Rammensee. H.G. and Adair. B.M.. Beauverger. P.. Buchland, R. and Wild, F., IYYJ. Vwolog),

62: 285-296.

1995. Immunol. Lett..

7(H) 2X1

Bensaid. A., Kaushaul, A., Baldwin, C.L.. Clevers. H.. Young, J.R.. Kemp. S.J.. MacHugh. N.D., Toye, P. and Teale. A.J., 1991. Immunogenetics. 3.3. 747. Buus. S.. Sette, A., Colon, S.M.. Miles. C. and (ire). Calin-Laurens,

V., Trescol-Bi&mont,

H.M..

IYX7. Science, 235: 1351.

M-C., Genier. D. and Rabourdin-Combe,

Cannon. M.J., Stall, E.J.. Taylor. G. and Askonas. B.A.,

Connors, M., Collins. P.L.. Firestone, C. Y. and Murphy. B.R.,

1991. _I. Viral.,

Davies, C.J.. Joosten. 1. and Bemoco. D.. 1994. Eur. J. Immunogenet., DiBrino.

M.. Tsuchida. T.. Turner, R.V..

C.. 199.1. Vaccine.

I I:

974.

1987. Immunology, 62: 133. 65. 16.34.

21: 239.

Parker. K.C , Coligan. J.E. and Biddison. W.E.,

1993. J. Immunol..

15 I: 5930. Ellis. S.A.. Braem. K.A. dnd Morrwn.

W.A..

lYY2. Immunopenekx.

37: 4Y.

Falk. K., Rotzschke. 0.. Stevanovic, S.. Jung, G. and Rammensee, H-G.. Feltkamp.

M.C.W..

Smits, H.L.,

Vierboom,

M.P..

Schegget. J.. Melief. C.J.M. and Kast, W.M.. Fremont. D.H..

R.P.,

1991. Nature, 351: 290.

de Jongh, B.M..

Drijthout,

J.W..

ter

1993. Eur. J. Immunol., 23: 2242.

Matsumura. M.. Stura. E.A.. Peterson. P.A. and Wilson, [.A.. 1992. Science. 257, 919.

Fries, R.. Beckmann. J.S.. Geoges,

M.. Soiler.

Frumento, G., Harris. P.E.. Gawinowicz.. Furze. J.. Werv.

Minnaar.

M.A..

M. and Womack. J.. 1989. Anim. Genet., 20: 2. Suciu-Foca. N. and Pemis, B.. 1993. Cell. Immunol.,

G.. Lerch, R. and Taylor, G.. IYYJ. J. (ien. \ lrol., 75: 363.

Gaddum. R.M.. Cook, R.S.. Thomas, L.H. and Taylor, C.. 19%. Immunology. 88, 421-427. Gairin, J.E. and Oldstone, M.B.,

1993. J. Vwl..

67, ?+I?.

Garber, T.L.. Hughes, A.L., Watkins, D.I. and Templeton, J.W.,

lYY4. Immunogenetics. 39: 257.

Gorga. J.C.. Madden. D.R.. Prendergast. J.K.. Wiley. D.C. and Strominger, J.L., 1992. Proteins, 12: 87.

1.52:

R.M. Guddum

et al./

Veterinary

Immunology

and Immunoppathology

54 (1996) 211-219

219

Graham, B.S., Bunton, L.A., Wright, P.F. and Karzon, D.T., 1991. J. Clin. Invest., 88: 1026. Hegde, N.R., Ellis, S.A., Gaddum, R.M., Tregaskes, C.A., Sarath, G. and Srikumaran, S., 1995. Immunogenetics, 42: 302-303. Howard, J.C. and Seelig, A., 1993. Nature, 365: 211. HUCZkO, EL., Bodnar, W.M., Benjamin, D., Sakaguchi, K., Zhu. N.Z., Shabanowitz, J., Henderson, Apella, E., Hunt, D.F. and Engelhard, V.H., 1993. J. Immunol.. 151: 2572. Kast. W.M., ROUX, L., Curren, J., Blom, H.J.J., Voordouw. A.C., Meloen, R.H., Kolakofsky, D. and Melief, C.J.M., 1991. Proc. Natl. Acad. Sci., 88: 2283. Kast, W.M.. Brand& R.M.P., Sidney, J., Drijfhout, J-W., Kubo, R.T., Grey, H.M., Melief, C.J.M., Sette, A., (1994) J. Immunol. 152, 3904. Kulkami, KuIkarm Kulkami, Virol.,

A.B., Connors, M., Firestone, C.Y., Morse III, H.C., Murphy, B.R., (1993a) J. Virol. 67, 1044. A.B., Morse III, H.C., Bennink, J.R., Yewdell, J.W. and Murphy, B.R., 1993b. J. Virol., 67: 4086. A.B., Collins, P.L.. Batik, I.. Yewdell, J.W., Bennink, J.R.. Crowe, J.E. and Murphy, B.R., 1995. J. 69: 1261.

Lawson, C.M., Bennink, J.R., Restifio, N.P., Yewdell, J.W. and Murphy, B.R.. 1994. J. Virol., 68: 3505. Ljunggren, H.G. and Kane, K.J., 1985. J. Exp. Med., 162: 1745. Madden. D.R., Garboczi, D.N. and Wiley, D.C.. 1993. Cell, 75: 693. Mandelboim, O., Berke, G., Fridkin, M., Feldman, M., Eisenstein, M. and Eisenbach, L., 1994. Nature, 369: 67. Nayersina, R., FowIer, P., Guilhot, S., Missale, G., Cemy, A., Schlicht, H-J., Vitiello. A., Chesnut, Person, J.L., Redeker, A.G. and Chisari, F.V., 1993. J. Immunol., 150: 4659. Neefjes, J.J. and Momburg, F., 1993. Curr. Opin. Immunol., 5: 27. Oldstone, M.B.A., 1991. J. Viral., 65: 6381. Rammensee, H-G., Friede, T. and Stevanovic, S., 1995. Immunogenetics, 41: 178. Rotzschke, O., Falk, K., Deres, K., Schild, H., Norda, M., Metzger, J., Jung, G. and Rammensee, H-G., 1990. Nature, 348: 252. Ruppert. J., Sidney, J., Celis, E.. Kubo, R.T., Grey, H.M. and Sette, A., 1993. Cell, 74: 929. Sawhney, S.M.S., Hasima, N.N., Glass, E.J., Al-Murrani, S.W.K., Nichani, A.K., Spooner, R.L., Williams, J.L. and Russell, G.C., 1995. Immunogenetics, 41: 246. Stauss, H.J., Davies, H.. Sadovnikova, E., Chain, B., Horowitz, N. and Sinclair, C., 1992. Proc. Natl. Acad. Sci. USA, 89: 7871. Stott, E.J. and Taylor, G., 1985. Arch. Virol., 84: I. Stuber, G.. Modrow, S.. Hoglund, P., Franksson, L., Elvin, J., Wolf, H., Karre, K. and Klein, G., 1992. Eur. J. Immunol., 22: 2697. Taylor, G., Thomas, L.H., Wyld, S.G.. Furze, J., Sapp, P. and Howard, C.J., 1995. J. Virol., 69: 665886664. Townsend, A. and Bodmer, H., 1989. Ann. Rev. Immunol., 7: 601. Toye, P.G., MacHugh, N.D., Bensaid, A.M., Alberti, S., Teale, A.J. and Morrison, W.I., 1990. Immunology, 70: 20. Van Bleek, G.M. and Nathenson, S.G., 1990. Nature, 348: 213. Whitton, J.L.. Gebhard, J.R., Lewicki, H., Tishon, A. and Oldstone, M.B.A., 1988. J. Virol., 62: 687. Widmann, C.. Maryanski, J.L., Romero, P. and Corradin, G., 1991. J. Immunol., 147: 3745