European Journal of fmmunogenetics (1992), 19,21-31

C O M P A R I S O N O F TCR-CXP S E Q U E N C E S O F CROSS- R E A C T I V E A N T I - D R A L L O R E A C T I V E T - C E L L

C L O N E S : I D E N T I F I C A T I O N OF POSSIBLE C O N T A C T R E S I D U E S B A S E D ON C H A R G E

C O M P L E M E N T A R I T Y B E T W E E N T C R C H A I N S A N D D R D E T E R M I N A N T S

E. cHAMPAGNE,* s. E ~ ~ ~ ~ ~ ~ , * ~ A . H U C H E N Q , * J . F A B R O N , * + H . L . C O P P I N , ' J . S E V I N * & M . THOMSEN+

*Centre de recherches sur le polyrnorphisrne genetique des populations humaines (CRPG), Toulouse, France and illnit& de recherches d'immunologie et d'irnmunogknktique hurnaine,

INSERM 11100, Toulouse, France

(Received 6 September 1991; revised I November 1991; accepted I November 1991)

SUMMARY

We analysed alloreactive T-cell clones selected for their differential recognition of DR variants differing in the third hypervariable region (hvr) of the DRBl gene (amino acid positions 67-70-71). This polymorphism leads to two main hvr3 types: a basic form (Leuh7-G1n7"-ArgfLys7') and an acidic form (Ileh7-Asp7"-G1u7') where residue 70 is probably directly accessible to the TCR on DRP chains. The TCRs have been sequenced. Three DRwl3-reactive clones use similar Va2 and VP13 gene family members but differ mainly by their cross-reactivity towards acidic or basic DR4 variants and by the sequence of CDR3 on their TCRa and/or P chains. One anti-DRwl3 clone cross-reacts with most specificities sharing the DRwl3 type of hvr3 and reciprocally one anti-DRBon (DRBl"0103) clone cross-reacts with DRwl3. These two clones use similar VP genes and share negative charges in CDR2a at position 56. They also share these negative charges in CDR2a with two other clones reacting specifically with DRBon, the acidic variant of DR1. We hypothesized that the selective recognition of positively or negatively charged residues on the DRP chain would necessitate reciprocal charges on the TCR complementarity determining regions (CDRs) responsible for this interaction. This facilitated identification of those residues of the TCR that possibly interact with the hvr3 determinant of HLA-DR. From these observations the mechanisms allowing the recognition of alloantigens by these T-cell clones are discussed.

Correspondence: Dr E. Champagne, CRPG-CNRS, HBpital Purpan, Ave de Grande Bretagne, 31300 Toulouse, France. Tel. (61) 49 60 80. Fax. (61) 49 90 36.

21

22 E. Champagne et al. I N T R O D U C T I O N

T-lymphocyte stimulation is induced by foreign antigens processed by antigen presenting cells and presented on the surface of these cells in the form of peptides bound to major histocompatibility complex (MHC) molecules (Babbitt et al . , 1985; Buus et al . , 1986; Berzofsky, 1986).

The T-cell receptor (TCR) binds to this antigenic complex by means of a heterodimer consisting of clonally variable polypeptides, the TCR-a and f3 chains (Caccia et af., 1988; Clevers et af., 1988; Hedrick, 1989). On a minor subset of T-cells this heterodimer consists of a y and a 6 chain. The molecular bases of alloreactivity are still unclear. Self peptides are continually presented by the MHC molecules of antigen-presenting cells (Buus et af., 1988) and there is growing evidence that peptides are implied in alloreaction as well as in reactions towards conventional antigens. In several instances, T-cell clones specific for self-restricted foreign antigens have been found to react also against alloantigens. Similarly it has been possible to find foreign antigens that stimulate primarily alloreactive clones. This has led to the concept of molecular mimicry: self-MHC+X may be equivalent to an alloantigen (Hunig & Bevan, 1981; Lombardi et a[ . , 1989; Lechler et af., 1990).

A molecular model of the MHC class I1 molecule has been derived by comparison with the class I crystal structure. Although the structure is still hypothetical, many predicted amino acid positions have been rather accurately confirmed by mutagenesis experiments (Brown etal., 1988; Peccoud et af., 1990; Brown et af., 1990). According to this model, the first and second hypervariable regions (hvrl and hvr2) of the DR molecules are located on the floor of the peptide-binding groove and presumably influence peptide binding. The amino acids of the third hvr of the DRP chain are located on one alpha-helix flanking the groove, and can possibly contact the TCR when pointing ‘up’ or the peptide when facing the groove.

According to the peptidic-self model (Kourilsky ef al . , 1987), alloreactive T-cell specificity would focus on the peptides bound in the antigen-binding groove of the allogeneic MHC molecule. Alloreactivity would then result from variations in the array of self peptides presented by different alleles, whereas the allogeneic MHC molecule itself would be viewed as self-equivalent by the T-cell receptor. There is ample evidence that the amino acids inside the groove can influence alloreactivity (Mellins et af., 1990); since these amino acids are unlikely to interact directly with the TCR, peptide recognition must be involved in alloreactive T-cell specificity. However, evidences for the direct recognition of polymorphic residues of the MHC antigens by the TCR also exist (Elliott & Eisen, 1990; Melief et al . , 1980).

The three dimensional structure of the TCR-af3 is not known. A tempting model has been drawn by comparison of TCR and immunoglobulin sequences (Chothia et af., 1988). From these studies three complementarity determining regions (CDRs) have been postulated in each of the TCR-a and f3 chains: CDRl and 2 in the Va and Vp segments and CDR3 encompassing the V( D)J junctions. In murine antigen-specific experimental systems, it has been shown that amino acids at the V(D)J junctions are critical for peptide specificity, possibly because of their direct contact with the peptide (Engel & Hedrick, 1988; Danska et al., 1990). A model of TCR-peptide-MHC interaction has been proposed in which the TCR faces the MHC molecule so that both CDR3s are located over the peptide groove whereas the Va and VP segments (CDR1 and CDR2) contact polymorphic and conserved residues on the alpha helices of the external domains of the MHC molecules (Davis & Bjorkman, 1988; Claverie et af., 1989).

In an attempt to generate cellular reagents against different class I1 antigens, T-cell clones have been produced in different combinations of stimulating and responder individuals. The present study is focused on the analysis of some of the alloreactive clones selected for their differential

23

reactivity on DR variants differing essentially at positions 67,70, and 71 of the DRP chain. The side chain of residue 71 is thought to point towards the peptide binding site whereas the residue 70 points 'up' on the surface of the molecule. The residue 67 would have an intermediate position (Brown etal. , 1988; Lechler etal. , 1990). In the DR variants used here (Fig. l), the polymorphism of the hvr3 sequence reveals two hvr3 types: a basic type (Leuh7-G1n7"-Arg/Lys7') in DR1, DR4/Dw4, and Dw14, and an acidic type (Ileh7-Asp7"-G1u7') in the DRBon (DRB1*0103) variant of DR1 (Cambon-Thomsen ef al . , 1986; Coppin et al . , 1990), DR4/DwlO, DRl l l DwJVM, and DRwl3. The TCRu and @ chains have been sequenced and compared. Taking advantage of the fact that theepitope on the DR molecule ischarged, we have looked for the TCR residues in the CDR regions which bear opposite charges and that might indicate possible sites of interaction between the TCR and the DRP-hvr3.

M A T E R I A L S A N D M E T H O D S

I -Cell-lines

The B lymphoblastoid cell lines (LCL) representing different HLA-DR haplotypes and HLA-Dw specificities were from the Tenth International Histocompatibility Workshop (Yang et a[. , 1989) or were local LCL generated in our laboratory. In this study we have also used a local panel of peripheral blood mononuclear cells (PBMC).

2-Alloreactive T-cell clones

Alloreactive T-cell clones have been generated by a limiting dilution method as previously described (Essaket et al . , 1990). Briefly, responder PBMC were cultured with irradiated stimulator PBMC. After 6d of primary culture, the blast cells were recovered using a 3545% percoll gradient, resuspended in culture medium and plated at a concentration of 0.5 cells/well in round bottomed microtitre plates in the presence of lo4 irradiated (100 Gy) LCLfrom the specific stimulator. Two days later, 10-20% of T-cell growth factor (TCGF) corresponding to 20 IU/ml of IL-2 were added. After 10-14d of culture, growing wells were transferred to 2ml wells, restimulated and expanded by addition of TCGF and culture medium.

Stimulator and responder HLA types were the following. Clones N G : stimulator A1,28; Bw62,w72; DRwll,wl3; DQwl,w3; DPblank; responder A2,ll ; B18,w55; DRwll,wl4; DQwl.w3; DPw4. Clone P602: stimulator A1,3; B44,35; DR-Bon,Bon; DQw1,wl; DPw4;

24 E. Champagne et al. responder A1,23; B5,8; DR3,5; DQw2,w3; DPwl,w4. Clones N749 and N750: same stimulator as P602; responder A11,28; B44,49; DR2,4; DQwl.w3; DPw4.

All clones were CD4-positive. Blocking assays with anti-class I (W6/32), anti-DQ (TU22). anti-DP (B7/21) and anti-DR (L243, GSP41) antibodies have been performed on A/G clones. Only anti-DR reagents inhibited cell proliferation.

3-T-cell proliferation assay

The T-cell clones were screened for specific responses by assaying lo4 rested T-cells for proliferative activity after incubation with 5 x lo4 irradiated (20 Gy) PBMC or 2.5 x 10‘ irradiated (100 Gy) LCL in 0.15ml of culture medium in round bottomed microtitre plates. In all assays autologous stimulating PBMC were included as negative controls and the cultures were done in triplicates. Cultures were pulsed with 1 pCi ‘H-thymidine after 48h and harvested on glass-fibre filters 16h later. The results are expressed as the specific uptake by subtracting the median thymidine uptake without responder cells from the median uptake in cultures with responder cells. All experiments were repeated in order to assure concordant results,

4-TCR-a and f3 cDNA cloning and sequencing

The protocol used was slightly modified from Loh et al. (1989). Briefly, total RNA was extracted from a minimum of 4 x lo5 cells by caesium chloride gradient centrifugation followed by two ethanol precipitations. The RNA was reverse transcribed using AMV-reverse transcriptase (BRL) and a mixture of Cae (5’CAGACAGAC’ITGTCACTG3’) and Cpe (5’-GTGTGGGA- GATCTCTGCTTCTG3’) primers. The first strand cDNA was phenol-chloroform-extracted and reprecipitated twice before tailing with terminal deoxynucleotidyl transferase (BRL) and dCTP. One third of the reaction product was then submitted to 30 cycles of polymerase chain reaction (PCR) amplification using the Cae or Cpe primers, poly-dG14 and Taq DNA polymerase (Cetus corporation) according to the manufacturer’s recommendations. Ten per cent of the amplifi- cation product was run on a 1.5% low melting point agarose gel and a piece of agarose around 450bp was recovered, melted and 5% was submitted to 30 additional PCR cycles with Cai (5’AlTTAGAGTCTCTCAGCTG3’) or Cpi (5’GACCTCGGGTGGGAACAC3’) and poly- dG14. PCR cycles were as follows. Denaturation step: 9 5 T , 30 sec; elongation step: 7 2 T , 2.5 min; annealing step: 1 min at 60°C (Cpe) or 50°C (other primers). The amplification product was extracted from a second agarose gel, treated with Klenow enzyme to ensure completion of both strands and directly cloned into the HincII site of Pucl8 plasmid before transformation of DH5a competent E. coli cells. The plasmids were sequenced in both directions with T7 DNA polymerase (Pharmacia LKB Biotechnology, Uppsala, Sweden) using universal, reverse and/or Cai/CPi primers. The sequences are derived from at least three clones having the same V-D-J junction. From each T-cell RNA, no more than one in-frame sequence was found. About 50% of the cDNA clones obtained with this protocol were not TCR products.

R E S U L T S

I“ Specificity and cross-reactivity pattern of the ailoreactive clones

The first four clones ( N G series) were made from the same stimulation against a DRwlYDwl9 individual. Their specificity for DR molecules has been checked by antibody blocking experiments with anti-DR, DQ, and DP reagents (not shown). All four clones react strongly against the specific stimulating specificity but each clone has a particular pattern of reactivity

Cross reactive anti- DR alloreactive T-cell clones 25

TABLE 1. Proliferative response of anti-DRwl3 clones. One representative experiment on an extended panel is shown. The cross-reactivity pattern has been confirmed by repeated testing

Clones Stimulation cell WN’ HLA-DR/Dw A/G6 AlGlO A/G11 AIG3 POLY’

KT17 DEU WT5 1 PF97387 SAVC YAR JHAF PE117 MTl4B OMW CB6B WDV HHKB SLE005 WT47 EMJ JVM MZ070782 KASl16 BB

Res~.’ Sp.St.6

9024 9025 9029 9027 9034 9026 9030 9028 9098 9058 9060 9062 9065 9059 9063 9097 9039 9002 9003 Local’ Local

DR4/DKT2 DR4/Dw4 DR4/Dw4 DR4/Dw4 D R4/D w4 DR4/Dw 10 DR4/Dw13 DR4/Dw 14 DR4/Dw I4 DRw 13/DwlR DRw 13/Dw 18 D Rw 1 3/D w 1 8 DRwl3/Dw 18 DRw 13/Dw 19 DRwlYDw 19 DRw 13/Dw 19 DRw 1 UDwJVM DRlIDw20 DRllDwl DRBon/DwBon DRII;DRwl3 DRll;DRwl4

3 - - - - -

’ 4720

402 1559 2036

160

1625 33418 19512 44 089

100 355

35 63 1 1212

-

-

-

-

4454 14278 16011 18538 13952

897 905

13692 9995

503

679 630

6438 20048 16426

269

698 1268

13087 600

-

-

- 95 - - -

61 - -

786 37069 22255 29816 35 634 17032 23414 26469

2013

56 I

30 349 658

-

-

665 52712 - 44513

96 31915 86 38058

- 34378 25 784 30410

250 33 324 1662 36 404 - 27 174

22710 21 033 22 I97 17 494 25 570 18 643 20 880 20913 23 500 23 838 20 198 27 152 22 522 29715

3095 15 118 78 38728

1016 42803 10 I04 20727 39861 3001 1

1898 430

’ Tenth international histocompatibility workshop number.

’ Zero or below zero.

’ From local panel. ” Specific stimulator. ’ Responding cells alone.

Polyclonal reagent generated against a pool of stimulators.

Median ‘H-thymidine uptake (cpm). The value obtained with stimulators alone has been subtracted.

(Table 1). N G 6 and N G 1 0 proliferate against DRwlUDwl9 stimulators but not against the Dw18 subtype which differs by a single amino acid substitution at position 86. This discriminating specificity can also be explained by their .recognition of different alleles encoded by the DRB3 locus. However, for reasons indicated in the discussion below we favour a direct influence of residue 86. These two clones also cross-react against DR4 stimulators but with different variants. N G 6 recognizes only the ‘acidic’ subtype DwlO whereas N G 1 0 recognizes the ‘basic’ subtypes Dw4 and Dw14. Interestingly A/G10 does not discriminate between Dw4 and Dw14 although they differ at position 86.

NG3 and A/G11 do not discriminate between the two DRwl3 subtypes. The pattern of reactivity of NG3 is however very peculiar as it cross-reacts with most specificities carrying the ‘acidic’ hvr3 region, DRBon, DRlUDwJVM, and DR4/DwlO.

The remaining clones N749, N750, and P602 have been raised from two different responders against a DRBon stimulator. They all react specifically against the DRBon stimulators but not against DR1. DRBon and DR1 are associated with the same DQwl allele and the stimulator cells do not display a DP antigen absent in the responder. Thus the three clones are very likely to be

26 E. Champagne et al.

TABLE 2. Proliferative response of anti-DR-Bon clones ~~

T-cell clone ~~

P602 N750 N749 Stimulating DR type’ exp. 1 exp. 2 exp. 1 exp. 2 exp. 1 exp. 2

Dr-Bon 1001 12 13759 19 680 34180 7631 13335 DRl 1928 504 2913 598 517 185 1 DRwl3 3447 5065 1774 80 ND’ ND DR3 978 129 1932 439 417 624

I Stimulators are from a local panel. Response of the specific combination (cpm) minus value of stimulators alone. Not done.

DR-specific. The absence of cross-reactivity between DRBon and DR1 is another evidence of the importance of DR-hvr3 for T-cell reactivity and all three clones are specific for the ‘acidic’ form of it (Table 2). Moreover P602 cross-reacts with DRwl3 stimulators. This reactivity is reminiscent of that of N G 3 which was raised against DRwl3. The specificities of these two clones seem to be focused almost exclusively on this epitope of the DRP chain.

2” Correlation between TCR-a and p sequences and reactivity TCR sequences have been determined by PCR-aided TCR cloning and sequencing (Fig. 2 ) . The T-cell preference for an ‘acidic’ or a ‘basic’ hvr3 epitope can help to define which part of the TCR contacts this epitope, assuming that it would involve the putative CDR loops and necessitate reciprocal charges on the TCR chains.

The first striking result is that three of the four anti-DRwl3 clones use Va and Vp gene segments belonging to the same subfamilies (Va2 and Vp13). N G 6 and NGIO which cross-react with DR4 have very similar p chains. At the VDJ junction, an NTEA motive is found in both clones, the NT residues not being encoded by the Jpl.1 segment also used by the two clones. Three CDR regions bear charges that can explain their difference in reactivity: the differences at positions 26 in CDRla (Asn in NG6, Asp in NGIO) and 96 in CDR3P are symmetrical to the difference between DwlO and Dw4 (Asp and Gln at position 70 respectively); in CDR3a, A/G10 has three acidic residues (94, 95, 99) which do not exist in MG6. A/G11 does not react against DR4; the a and p chains of this clone differ from N G 6 and A/G10 mostly in CDR3s on the a and p chains.

The remaining 4 clones all react against DRBon but not DRI. They all use dissimilar Va and J a segments. Among the two clones that also react against DRwl3, P602 uses a Va2 member. However, they share similar p chains with almost identical Vp segments. At the VDJ junction, they share a glutamin at position 97, a position that has been shown to be important in some cases for peptide recognition (Engel & Hedrick, 1988). The remaining clones N749 and N750 both use Vp6.7 segments and dissimilar a chains. However in the absence of cross-reactivity there is no evidence that these two clones recognize directly the ‘acidic’ hvr3 region. Taken together, the four DRBon-reacting clones use dissimilar a chains. On the contrary, the p chains display a striking similarity. Indeed, the Vp6 and VpS families are closely related and this is particularly true in CDRlP and CDR2P regions. If one tries to find a particular sequence shared by these clones that can explain their discriminating specificity for acidic HVR3, several positions appear

Cross reactive anti-DR alloreactive T-cell clones 27

N749 N750 P602 A/G3 A/G6 A/G10 A/G11

N749 N750 P602 A/G3 A/G6 A/G10 A/G11

N749 N750 P602 A/G3 A/G6 A/G10 A/G11

N749 N750 P602 A/G3 A/G6 A/G10 A I G 1 1

1 10 20 30 40 50 60 70

NEVEQSPQNLTAQEGEF I T 1 NCSY SVGI SALH . WLQQHPGGG 1 VSLFML . SSGKKKffiRL I A. T I N IQEKHS e n v e q h p s t l s v q e g d s a v i kctysmasnyfpwykqelgkrpdl i idi rsvvgeKKC9RIAVTLNKTAKHF KEVEQDPGPLSVPEGAIVSLNCTY SNSAFQYFMWYRQYSRKGPELLNYTYSSGT .fU&.RLQHRSINPA.SI ESVGLHLPTLSVQEGDNS I I NCAY SNSASHF I WYKQESGKGPQF I I DI 6NMLXM)GQRVTVLLNKTVKHL

SNSAFQY FMWYRQY SRKGPELLMYTY SSGNKmG . RFTAQVDKSSKY I KEVEQNSGPLSVPEGAIASLNCTYSUGSQSFFWYRQYSGKSPEL I M S I Y S N G l M f f i .RFTAQLNKASQYV KEVEQNSGPLSVPEGAIASLNCTYSmSFFWYRQYSGKSPEL IMSIY SNGaYmG .RFTAQLNKASQYV

k * t k t * -

80 90 100 110 * * * *

SLH I TASHPRDSAVY I CA .VRffiGSQGNL I FGKGTKLSVKPN SLH ITETQPEDSAVYFCAASTGLTGGFKCIFGAGTRLFVKA SLFMRDSQPSDSATYLCA. .MTEGNTPLVFGKGTRLSVIAN SLQIAATQPGDSAVYFCA.. . .GYTGNQFYFGTGTSLT SLF IRDSQPSDSATYLCA.MSYTGTASfiTFGTGTRLQVTLD SLL I RDSQPSDSATY LCAVPP€EGN€RLTFGTGTRLT I I PN SLL IRDSQPSDSATYLCA. .VNR€GNTPLVFGKGTRLSVVAN

HU-DR SPECIFICITY

Val9 h 8oM Vd.1 Jaw Boll Vd.x JIIX BOM+w13 V d . 2 .kU BON+*13+4Lh110 Vd.y .hy r13+4D*10 Vd.1 k w13+4Dw4 Vd.1 .lax w13

-> c-

Va ( H I L

1 10 20 30 4 0 50 60 70 * * * * * * k * -

VSQSPSNKVTEKGKDVELRCDP I SGHTALYWYRQSLGQGLEFL I YFQGNSAPOKSGLPSDRFSAERTGGS I STLT GAGVSQSPSNKVTEKGKDVELRCDPISGHlALYWYRQSLGQGLEFL IYFQGNSAPOKSGLPSORFSAERTGGSVSTLT DAGVIQSPRREVTEMGQEVTLRCKPISGMNLFWYRQTMMRGLELLIYFNNNVP IDDSGMPEDRFSMMPNASFSTLK DAGVIQSPRHEVTEMGQEVTLRCKPISGMNLFWYRQTMMRGLELL IYFNNNVPIDDSGMPEDRFSAKMPNASFSTLK NAGVTQTPKFRVLKTGQSMTLLCAQL”//€YMYWYRQDPGMGLRL IMSVGLXTTAKGEVPDG .YNVSRLKKQNFLLG NAGVTQTPKFQVLKTGQSMTLQCAQU4N//€YMSWYRQDPGMGLRL I MSVGAGITDQGEVPNG .YNVSRSTTEDFPLR

QCA-MYWYRQDPGMGLRL I YYSASffiTTDKGEVPNG .YNVSRLNKREFSLR

80 90 100 110 * * * *

IQRTQQEDSAVYLCASSRNRRLLAIJTUTQYFGPGTRLTVL IQRTQQEDSAVYLCASSW.. . .GWAXNIQYFGAGTRLSVL IQPSEPRDSAVYFCASSSQ.. . . . T P A FFGQGTRLTVV IQPSEPRDSAVYFCASSGQ.. . . .GSQPQHFGDGTRLSIL LESAAPSQTSVYFCASSNG.. . . .NTA.FFGQGTRLTVV LLSAAPSQTSVYFCASTLX.. .GWNTA. FFGQGTRLTVV LESAAPSQTSVYFCASSA.. .MAGSYGYTFGSGTRLTVV

HU-DR SPECIFICITY

Vp6.7~ Jp2.3 Boll Vp6.7a Jfl2.4 Bow Vg8.3b Jpl.1 Bowin13 Vg8.h Jpl.5 Bow+w13+4Dw10 Vpl3.x Jpl.1 w13+41)*10 Vp13.2 Jpl.1 w13+41)*4 Vp13.y Jp1.2 w13

FIG. 2. TCRa and p amino acid sequences of alloreactive T-cell clones. V-gene family assignment has been made according to Wilson et al. (1988). The numberof previously named members of these families is indicated. New variantsof the Vp6.7 and Vp8.3 genes have been found and called V p 6 . 7 ~ and Vp8.3b. The sequences of V013.x and Vp13.y are identical to the published sequences of CEMl (Duby & Seidman, 1986) and clone G36 from Bragado et al. (1990). Va2.x and Va2.y are as yet undescribed Va2 members. The beginning of N750 p-chain is derived from the sequence of only one clone but is identical to that of Va8.1 (Caccia et al., 1988) indicated in lower case letters. Between lines are the putative CDR regions. In CDR regions acidic and basic residues are indicated in bold italics.

28 E. Champagne et al. to be good candidates. On the p chain, His" in CDRl is in fact conserved between many Vp families (Wilson et al . , 1988). Better candidates are Gln/AsnS" and GhS2 in CDR2. On the a chain all the clones have basic residues at position 56 (it should be noted that the Va2 member used by P602 is missing the acidic residue Asps6 found in other Va2 sequences).

D I S C U S S I O N

The molecular bases of T-cell reactivity against alloantigens are still poorly understood. MHC molecules can be antigenic by three different mechanisms at least. (1) They bear surface allodeterminants that are accessible to antibodies. On DR molecules these determinants are localized on one alpha helix walling the peptide binding groove (hvr3). In murine systems. similar determinants can be recognized directly by T cells on the Kh molecule (Ajitkumar e t a f . , 1988) or even on empty MHC molecules (Elliott & Eisen, 1990). (2) Allogeneic MHC molecules can also bear a peptide pool different from the peptide pool presented by autologous MHC due to polymorphic residues pointing to the peptide site: in this case the surface of the alloantigen which is accessible to the TCR could be seen as self molecules presenting antigenic peptides. (3) Finally alloantigens as classical polymorphic proteins can be degraded and presented in the form of MHC peptides. There are several examples of MHC peptides presented by class I or class I1 molecules (DeKoster eral., 1989; Essaket e t a f . , 1990). The presenting MHC molecule itself could again be viewed as if it was a self molecule.

The first and third mechanisms seem to be evidenced by the clones described here. A/G3 recognizes all specificities carrying the DRwl3-type of hvr3. although they differ by their hvrl and hvr2 and are likely to bind different sets of peptides: thus this clone (and maybe also P602 which has been less extensively tested) is likely to directly recognize the allogeneic molecule regardless of the bound peptide. A clone with similar reactivity (C50) has been described by Termijtelen (1990). Such clones could have come through thymic selection because they have a weak affinity for the self MHC molecule itself but recognize a self peptide. Their alloreactivity pattern would result from accidental reactivity of their TCR with the DRwl3-hvr type. If the Davis-Claverie model of TCR-MHC interaction is valid, this allogeneic hvr3 reactivity must involve the CDRl andor the CDR2 and not the CDR3 of the a and/or p chains which would be involved in the self peptide recognition. This hypothesis is strengthened by the fact that the TCRs of A/G3 and P602 bear very different V(D)J junctions on the a and P chains (except perhaps for a shared glutamin at position 97 on the p chain) whereas CDRl and CDR2 P are very similar. Also CDR2a in both cases bear positive charges.

It is difficult to speculate on the mechanism which can explain the reactivity of the clones N749 and N750. For these two clones the recognition of an acidic DRP-hvr3 is evidenced by their non-reactivity to DR1. We cannot exclude that the two closely related antigen DR1 and DRBon bind different sets of peptides due to the influence of aminoacids 67 and 71. However it is striking that the two clones use Vp6 members, a sub-family that is closely homologous to VpS. Again, they are very similar to A/G3 and P602 in CDRl and CDR2P and bear positive charges in CDR2a (Lys/ArgSh). It is tempting to hypothesize that the interaction with the Asp"' on hvr3 is mediated in these four clones by the basic residue at position 56 on CDR2a.

Among the four clones primarily reacting against DRwl3 ( N G clones), two clones, AIG6, and N G 3 react against DR4fDwlO. However several evidences indicate that this cross-reactivity is due to different TCR-MHC interations. Firstly they use very dissimilar TCR chains. As shown previously the structure of A/G3 TCR is more related to that of anti-DRBon clones whereas N G 6 uses TCR chains closely related to the ones of other anti-DRwl3 clones not reacting with

Cross reactive anti-DR alloreactive T-cell clones 29 DR4/DwlO. Secondly N G 3 has a reactivity extended to other antigens sharing the acidic hvr3 with DR4/DwlO. Thirdly they are differently affected by the polymorphism of residue 86 of DRP ( N G 3 reacts against DRwl31Dwl8 and Dw19 whereas N G 6 recognizes only the Dw19 variant).

NG6, NGlO, and N G l l all use Va2 and VP13 members. This similarity is unlikely to be fortuitous and indicates that similar TCR-MHC interactions are involved for the three clones. The cross-reactivity between DR4 and DRwl3 has already been documented with primarily DR4-reacting clones (H.A. Giertsen, personal communication; Weyand et af., 1991) or DRwl3-reacting clones (Termijtelen et al. , 1989), indicating that DRwl3 reacting TCRs frequently have an intrinsic affinity for DR4 molecules. The close analysis of NG10 reactivity shows that DR4 and DRwl3 reactivities are differently influenced by the Gly/ValM polymor- phism (DRwl3 reactivity is abolished whereas DR4 reactivity is not). It has been argued that the discrimination between DRw13/Dw18 and DRwl3IDwl9 might be due to the recognition of molecules encoded at the DRB3 locus instead of the DRwl3 allele itself. However we do not favour this hypothesis since a similar discrimination occurs between DRl/Dwl and DRUDw20 (Eckels et al., 1990) and since this residue also influences peptide binding on DR7 molecules (Krieger et al., 1991). The direct influence of residue 86 on T-cell reactivity is thus very likely. A possible explanation for the DRw13-DR4 cross-reactivity arises from the comparison of the TCR structures of NG6, NG10 and NG11: the three clones react differently with DR4 stimulators: N G 6 recognizes the acidic form DwlO, NG10 the basic forms Dw4 and Dw14, and NG11 recognizes none of these DR4 molecules. Although they use similar V-gene members the TCR sequences of these three clones differ at the V(D)J junctions. N G 6 and NG10 share an NTEA sequence in CDR3P, which may account for their common DR4 reactivity. They differ in CDR3a by the global charge which is positive in N G 6 (Lys""') and negative in NGlO (Gl~"-Asp'~- GIu"-L~s~'"), and this can account for differential reactivity with an acidic or a basic DR4 variant. A/Gl1 does not share any of these CDR3 characteristics with the other two clones. From the relationship between CDR3 sequences and DR4 reactivity it is tempting to explain the DR4-reactivity by the recognition of a peptide from the DR4-hvr3, presented by DR4 itself: DR4/DwlO-positive cells would present the acidic peptide whereas DR4iDw4-14-positive cells would present the basic one. Peptide-binding experiments should allow testing of these hypotheses.

Prochnicka-Chalufour et al. (1991) observed a preferential distribution of charged residues in CDRls and CDR2s: this is a strong evidence for the importance of electrostatic interactions between TCR and MHC. In addition they proposed that multiple configurations are possible for the interaction between the TCR and the peptide-MHC complex. This would explain why it is difficult to demonstrate a bias in Va or Vp gene family usage in T cells of defined specificity. It also implies that TCR sequences can be compared only if they allow similar TCR-MHC interactions. Within each group, the clones that we compared seem to fulfil this condition based on the criterion that they share a specificity pattern and they use either related V, genes (for anti-DRBon clones) or Va and Vf.3 genes (A/G6, 10 and 11). It is predictable that the analysis of larger panels of alloreactive clones will show other preferential V gene segment usage associated with the involvement of different CDR regions in the contact with a defined MHC epitope.

A C K N O W L E D G M E N T S

We are grateful to Drs A. Cambon-Thomsen, C. De Preval, and B. Rubin for fruitful and critical discussions. The secretarial assistance of D. Sicre and L. Hillat and the help of M. Ferrer for artwork have been much appreciated. This work was supported by grants from INSERM (900305) and ARC (6510 and 415-89).

30 E. Champagne et al. R E F E R E N C E S

AJITKUMAR, P. GEIER, S.S., KESARI, K.V., BORRIELLO, F., NAKAGAWA, M., BLUESTONE, J.A., SAPER, M.A., WILEY, D.C. & NATHENSON. S.G. (1988) Evidence that multiple residueson both thea-helicesof the class I MHC molecule are simultaneously recognized by the T cell receptor. Cell, 54,47.

BABBITT, B.P., ALLEN, P.M., MATSUEDA, G., HABER, E. & UNANUE, E.R. (1985) Binding of immunogenic peptides to Ia histocompatibility molecules. Nature, 317, 359.

BERZOFSKY, J.A. (1986) Antigen processing and presentation. Annales de I’lnstitut Pasteur Immunology, 137D, 297.

BRAGADO, R., LAUZURICA, P., LOPEZ, D., & LOPEZ DECASTRO, J.A. (1990) Tcell receptor VP gene usage in a human alloreactive response. Shared structural features among HLA-B27-specific T cell clones. Journal of Experimental Medicine, 171, 1189.

BROWN, J.H., JARDETZKY, T., SAPER, M.A., SAMRAOUI, A., BJORKMAN, P.J. & WILEY, D.C. (1988) A hypothetical model of the foreign antigen binding site of Class I1 histocompatibility molecules. Nature, 332, 845.

BROWN, M.A., GRIFFITH, I.J., & GLIMSHER, L.H. (1990) Functional and molecular characterization ofI-Akp mutants is consistent with the predicted three dimensional structure of class I1 MHC molecules. Molecular Immunology, 27, 645.

BUUS, S., COLON, S., SMITH, C., FREED, J.H., MILES, C. & GREY, H.M. (1986) Interaction between a ‘processed’ ovalbumin peptide and Ia molecues. Proceedings of the National Academy of Sciences, USA, 83, 3968.

Buus, S., SETTE, A,, COLON, S.M. & GREY, H.M. (1988) Autologous peptides constitutively occupy the antigen binding site on Ia. Science, 242, 1045.

CACCIA, N., TOYONAGA, B., KIMURA, N. & MAK, T.W. (1988) The a and P chains of the T-cell receptor. In: The T-cell receptors (ed. by T.W. Mak), p. 9. Plenum Publishing Corporation, New York.

CAMBON-THOMSEN. A, , THOMSEN, M., ABBAL, M., SOMMER, E., CALOT, M. & OHAYON. E. (1986) A new HLA-D specificity associated with DR blank: D-BON. Tissue Antigens, 27,256.

CHOTHIA, C., ROSS BOSWELL, D. & LESK, A.M. (1988) The outline structure of the T-cell a@ receptor. EMBO Journal, 7,3745.

CLAVERIE, J.M., PROCHNICKA-CHALUFOUR, A. & BOUGUELERET, L. (1989) Implications of a Fab-like structure for the T-cell receptor. Immunology Today, 10, 1.

CLEVERS, H., ALARCON, G., WILLEMAN, T. & TERHORST, C. (1988) The T cell receptorKD3 complex: a dynamic protein ensemble. Annual Review of Immunology, 6,629.

COPPIN, H.L., AVOUSTIN, Ph., FABRON, J., HUCHENQ, A., GARNIER, J.M., THOMSEN, M. & DE PREVAL, C. (1990) Evolution of the DRl gene family. Journal of Immunology, 144,984.

DANSKA, J.S., LIVINGSTONE, A.M., PARAGAS, V., ISHIHARA, T. & FATHMAN, C.G. (1990) The presumptive CDR3 regions of both T cell receptor a and P chains determine T cell specificity for myoglobin peptides. Journal of Experimental Medicine, 172, 27.

DAVIS, M.M. & BJORKMAN, P.J. (1988) T-cell antigen receptor genes and T cell recognition. Nature, 334, 395.

DE KOSTER, H.S., ANDERSON, D.C. & TERMIJTELEN, A. (1989) T cells sensitized to synthetic HLA-DR3 peptide give evidence of continuous presentation of denatured HLA-DR3 molecules by HLA-DP. Journal of Experimental Medicine, 169, 1191.

DUBY, A. & SEIDMAN, J . (1986) Abnormal recombination products result from aberrant DNA rearrangements of the human T cell antigen receptor @-chain genes. Proceedings of the National Academy of Sciences, USA, 83,4890.

ECKELS, D.D., GEIGER, M.J., SELL, Th.W. & GORSKI, J.A. (1990) Involvement of class I1 P-chain amino acid residues 85 and 86 in T-cell allorecognition. Human Immunology, 27,240.

ELLIOTT, T.J. & EISEN, H.N. (1990) Cytotoxic T lymphocytes recognize a reconstituted class I histocompati- bility antigen (HLA-A2) as an allogeneic target molecule. Proceedings of the National Academy of Sciences, USA, 87, 5213.

ENGEL, I. & HEDRICK, S.M. (1988) Site directed mutations in the VDJ junctional region of a T-cell receptor p chain cause changes in antigenic peptide recognition. Cell, 54,473.

ESSAKET, S., FABRON, J.. DE PREVAL, C. & THOMSEN, M. (1990) Recognition of HLA-A1 and HLA-DPw3 by a human CD4+ alloreactive T lymphocyte clone. Journal of Experimental Medicine, 172,387.

FLOMENBERG, N., HANSEN, J.A., KLEIN, J.S. & DUPONT, B. (1989) Background, goals, and experimental design. In: Imrnunobiology of HLA (ed. by D. Dupont), vol. 1, p. 443. Springer Verlag, New York.

Cross reactive anti-DR alloreactive T-cell clones 31 HEDRICK, S.M. (1989) T lymphocyte receptors. In: Fundamental Immunology (ed. by W.E. Paul), p. 291.

Raven Press Ltd, New York. HUNIG, T. & BEVAN, M.J. (1981) Specificity of T-cell clones illustrates altered self hypothesis. Nature, 294,

460. KOURILSKY, Ph., CHAOUAT, G. , RABOURDIN-COMBE, Ch. & CLAVERIE, J.M. (1987) Working principles in

the immune system implied by the 'peptidic self' model. Proceedings of the National Academy of Sciences, USA, 84,3400.

KRIEGER, J.I., KARR, R.W., GREY, H.M., Yu, W.-Y., O'SULLIVAN, D., BATOVSKY, L., ZHENG, 2 . - L . , COLON, S.M., GAETA, F.C.A., SIDNEY, J. , ALBERTSON, M., DEL GUERCIO. M.F., CHESTNUT, R. W. & SETTE, A. (1991) Single aminoacid changes in DR and antigen define residues critical for peptide-MHC binding and T-cell recognition. Journal of Immunology, 146,2331.

LECHLER, R.I., LOMBARDI, G., BATCHELOR, J.R., REINSMOEN, N. & BACH, F.H. (1990) The molecular basis of alloreactivity. Immunology Today, 11, 83.

LOH, E.Y., ELLIOTT, J.F., CWIRLA, S., LANIER, L.L. &DAVIS, M.M. (1989) Polymerase chain reaction with single-sided specificity: analysis of T cell receptor 6 chain. Science, 243, 217.

L MBARDI, D., SIDHU, S., BATCHELOR. J.R. & LECHLER, R.I. (1989) AIlorecognitionofDRl by Tcellsfrom ' a DR4/DRw13 responder mimics self-restricted recognition of endogenous peptides. Proceedings ofthe National Academy of Sciences, USA, 86, 4190.

MELIEF, C.J .M., DE WAAL, L.P., VAN DER MEULEN, M.Y., MELVOLD, R.W. & KOHN, H.I. (1980) Fine specificity of alloimmune cytotoxic T lymphocytes directed against H-2K. A study with Kh mutants. Journal of Experimental Medicine, 151, 993.

MELLINS, E., ARP, B., SINGH, D., CARRENO, B., SMITH, L., JOHNSON, A.H. & PIOUS, D. (1990) Point mutations define positions in HLA-DR3 molecules that affect antigen presentation. Proceedings ofthe National Academy of Sciences, USA, 87,4785.

PECCOUD. J . , DELLABONA, P., ALLEN, P., BENOIST, C. & MATHIS, D. (1990) Delineation of antigen contact residues on an MHC class I1 molecule. EMBO Journal, 9, 4215.

PROCHNICKA-CHALUFOUR, A., CASANOVA, J.-L., AVRAMEAS, S., CLAVERIE, J.-M. & KOURILSKY, P. (1991) Biased amino acid distributions in regions of the T-cell receptors and MHC molecules potentially involved in their association. International Immunology, 3, 853.

TERMIJTELEN, A,, ELSEN, P., KONING, F., DEKOSTER, S. , SCHROEIJERS, W. & VANDERKERCKHOVE, B. (1989) A novel T-cell-defined HLA-DR polymorphism not predicted from the linear amino acid sequence. Human Immunology, 26,47.

TERMIJTELEN, A. (1990) T-cell allorecognition of HLA class 11. Human Immunology, 28, 1. YANG, S.Y., MILFORD, E., HAMMERLING, U. & DUPONT, B. (1989) Description of the reference panel of

B-lymphoblastoid cell lines for factors of the HLA system: The B-cell line panel designed for the tenth international histocompatibility workshop. In: lmmunobiology of HLA (ed. by B. Dupont). Springer Verlag. New York.

WEYAND, C.M., HICOK, K.C. & GORONZY, J.J. (1991) Nonrandom selection of T-cell specificities in anti-HLA-DR responses. Sequence motifs of the responder HLA-DR allele influence Tcell recruitment. Journal of Immunology, 141,70.

WILSON, R.K., LAI, E., CONCANNON, P., BARTH, R.K. & HOOD, L.E. (1988) Structure, organization and polymorphism of murine and human T-cell receptor Q and p chain gene families. Immunological Reviews, 101, 149.