Proc. Natl. Acad. Sci. USA Vol. 88, pp. 7699-7703, September 1991 Immunology Chicken T-cell receptor a8-chain diversity: An evolutionarily conserved D encoded glycine turn within the hypervariabfe CDR3 domain (N region/junctional diversity/immunologic repertoire/diversity gene segment/complementarity-determining region) WAYNE T. MCCORMACK, LARRY W. TJOELKER, GREGORY STELLA, CHRISTINA E. POSTEMA, AND CRAIG B. THOMPSON* Howard Hughes Medical Institute, Departments of Internal Medicine and Microbiology/Immunology, University of Michigan Medical School, Ann Arbor, MI 48109 Communicated by J. Herbert Taylor, June 10, 1991

ABSTRACT Unlike mammals, chickens generate an im- species (11). The germ-line repertoire of chicken Vat genes munoglobulin (Ig) repertoire by a developmentally regulated consists of only two Va families, Vat and V,2 (12, 13), which process of intrachromosomal gene conversion, which results in bear structural similarities to the mammalian VpI and VP1I nucleotide substitutions throughout the variable regions of the subgroups (14). The chicken Van family consists of six nearly Ig heavy- and light-chain genes. In contrast to chicken Ig genes, identical Vp genes (average amino acid identity, 97%), and the we show in this report that diversity of the rearranged chicken V2 family consists ofthree to five genes (average amino acid T-cell receptor (TCR) a-chain gene is generated by junctional identity, 95%), depending on the chicken strain examined. heterogeneity, as observed in rearranged mammalian TCR Unlike chicken Ig V segments, all chicken Va segments that genes. This junctional diversity increases during chicken de- have been sequenced contain heptamer and nonamer ele- velopment as a result of an increasing base-pair addition at the ments. Multiple distinct rearrangements of the TCR,8 locus Vp-Dp and Dp-Jp joints (where V, D, and J are the variable, are detected when Southern blots of chicken thymic and diversity, and joining gene segments). Despite the junctional T-cell-line DNA are probed with Va1 and V2 probes (11-13), hypervariability, however, almost all functional Vp-Dp-Jp suggesting that most or all chicken Va genes are functional. junctions appear to encode a glycine-containing #-turn. Such Comparison of cDNA sequences revealed three unique Jo a turn may serve to position the amino acid side chains of a sequences (11), and genomic screening with an oligonucleo- hypervariable TCR a-chain loop with respect to the antigen- tide probe identified a fourth Jo gene segment (13). Sequence binding groove of the major histocompatibility complex mol- alignments of TCR3 cDNA and genomic clones revealed ecule. Consistent with this hypothesis, the germ-line Dp nucle- 12-25 nucleotides at the V9-Je junction that were not en- otide sequences of chickens, mice, rabbits, and humans have coded by the germ-line Va and Jp segments, and suggested the been highly conserved and encode a glycine in all three reading presence of Do segments (11). frames. We have examined the molecular mechanisms for the generation of diversity in the chicken TCRE locus by iden- Immunoglobulins (Ig) and T-cell receptors (TCR) are the tifying the germ-line chicken Do gene segment and by ana- antigen-recognition molecules on the surfaces of B and T lyzing rearranged chicken TCRJ3 genes isolated from the cells, respectively (1-3). Whereas B cells recognize soluble developing thymus and mature spleen.t We observe that antigens, T cells recognize antigenic peptides associated with chicken TCR,8 diversity is generated by junctional variabil- cell surface molecules encoded by the major histocompati- ity, rather than by gene conversion, as in chicken Ig genes. bility complex (MHC) (4, 5). Genetic analyses of the mam- Further, the junctional diversity increases during develop- malian Ig and TCR gene families have shown that somatic ment due to increased random base-pair (N-nucleotide) ad- diversity is generated by junctional and combinatorial diver- dition. Within the junctional variability, however, most re- sity during the assembly offunctional Ig and TCR genes from arranged chicken TCR,8 genes have at least one glycine variable (V), diversity (D, for Ig heavy, TCRf3, and TCRS residue, which is encoded in the germ-line by an evolution- chains), and joining (J) gene segments (1-3). In contrast, arily conserved Do segment. The implications of a conserved birds use distinct molecular mechanisms to generate somatic glycine for TCR structure and antigen recognition are dis- diversity in their Ig loci. Instead of using junctional and cussed. combinatorial variation to create somatic Ig gene diversity, chickens rearrange only single V and J segments for both the heavy (H)- and the light (L)-chain Ig lpci (6, 7). After MATERIALS AND METHODS rearrangement, the VH and VL segments undergo sequence Isolation of Rearranged Chicken TCRE Genes. Polymerase diversification by intrachromosomal gene conversion using chain reaction (PCR; ref. 15) amplification, cloning, and families of V pseudogene segments as sequence donors sequencing of rearranged VP1.1-Dp-JP3 genes from 4-week (6-10). Given these differences between mammalian and thymus DNA were performed as described (16). Thirty cycles avian Ig gene diversification, it was of interest to determine of amplification (1.5 min at 940C, 3 min at 720C) were the molecular mechanisms for the diversification of chicken performed in a Coy thermal cycler. PCR primers included a TCRf3 genes. The recent description of chicken TCR8 cDNA and ge- Abbreviations: CDR, complementarity-determining region; MHC, nomic sequences revealed evolutionarily conserved struc- major histocompatibility complex; TCR, T-cell antigen receptor; tural features of the TCR,8 chains in avian and mammalian TdT, terminal deoxynucleotidyltransferase. *To whom reprint requests should be addressed at: Howard Hughes Medical Institute, 1150 West Medical Center Drive, MSRB-I, Room The publication costs of this article were defrayed in part by page charge 3510, Ann Arbor, MI 48109. payment. This article must therefore be hereby marked "advertisement" tThe sequence reported in this paper has been deposited in the in accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession no. M73064).

7699 Downloaded by guest on September 28, 2021 7700 Immunology: McCormack et al. Proc. Natl. Acad Sci. USA 88 (1991) sense primer at the 5' end of V91.1 (5'-CCCCGTCGACGT- probe spanning this core sequence hybridized to a 1.8-kb TCAGACAAAGAGAGTGTAATCCA-3') and an antisense HindIII fragment isolated from a genomic clone encompass- primer in the 3' flanking region of J3 (5'-GGAAGQCGjjC- ing the 20-kb region 5' of the J4 cluster. Nucleotide sequenc- CGAGGTGAGGGAGATGACAAGCACAGAGCAGG- ing of this genomic fragment revealed a single germ-line Do 3'). Sal I and Not I restriction sites (underlined) at the 5' ends segment (Fig. 1B). Rearrangement of only one Dp segment of the primers facilitated cloning into the pBluescript SK(-) with the four J4 segments was detected when Southern blots plasmid vector (Stratagene). ofthymic DNA were hybridized with probes located 5' ofD's Identification of the Chicken Germ-Line Dp Gene Segment. (data not shown). Comparison ofthe chicken, mouse (21, 22), An oligonucleotide probe (5'-GGGACAGGGGGATC-3') rabbit (20), and human (18, 19) germ-line Do segments spanning the putative chicken Do gene segment (11) was revealed 100% nucleotide sequence identity for the first 11 bp end-labeled with [y-32P]ATP by T4 polynucleotide kinase ofchicken Dp and the human and mouse D,91.1 gene segments, (17). Southern blots containing restriction digests of cloned one nucleotide substitution in mouse Dp2.1 and rabbit D>, genomic fragments ofthe germ-line chicken TCRF3 locus (11) and a 3-bp insertion in human D2.1 (Fig. 18). There is some were prehybridized in lOx Denhardt's solution/0.9 M variation in the sequence and length at the 3' ends of the NaCl/90 mM Tris-HCl, pH 8/0.6 mM EDTA/0.1% SDS at germ-line D,9 segments. Except for the conserved heptamer 420C for several hours. Blots were hybridized at 420C over- and nonamer elements ofthe recombination signal sequence, night in the same solution after addition of labeled probe (106 there is no significant sequence homology in the 5' or 3' dpm/ml). Blots were washed twice at room temperature and flanking regions ofthe genomic D,9 gene ofchicken compared twice at 420C in 0.9 M NaCl/0.09 M trisodium citrate, pH with mouse, rabbit, and human (Fig. 1B). 7/0.1% SDS. A 1.8-kilobase (kb) HindIII genomic fragment Non-Germ-Line-Encoded Nuceotide Additons to the Vp positive for hybridization with the Dp oligonucleotide probe and Dp-Jp Juncos. Alignment of the germ-line Dp segment was subcloned into pGEM-7Zf(+) (Promega) and sequenced with our initial rearranged TCRJ3 clones (Fig. UA) revealed a using primers specific for the SP6 and T7 sites of the vector number of non-germ-line base pairs at the Vp-Ds and Ds-J and internal sequences. junctions, suggesting that N-nucleotide addition may account for a large part of the TCR,3 junctional diversity. N nucleo- tides are template-independent G+C-rich polynucleotide ad- RESULTS ditions to V(D)J joints of mammalian Ig and TCR genes, Chicken TCRN Diversity Is Focused at the Vp-j Junction. believed to be mediated by terminal deoxynucleotidyltrans- To assess the diversity ofthe chicken TCRj3 repertoire, gene ferase (TdT) (23-25). This is in contrast to rearranged chicken rearrangements involving the V91.1 and 4J,3 genes were cloned Ig genes, which lack N nucleotides at V(D)Jjoints (6-10). To from the thymus of a 4-week-old chicken after PCR ampli- assess the contribution of N nucleotides to TCR8 diversity, fication (15, 16). Comparison ofthese clones to the germ-line additional rearranged Vpl.1-DP-J3 clones were isolated from Vp and J4 sequences reveals no sequence variation within the thymocyte DNA prepared from 18-day embryos (early) and Vp or J4 regions, but extensive variation is observed at the from a chicken 4 weeks after hatching (late). Considerable Vp-J junction (Fig. LA), which encodes the third comple- junctional diversity is seen at the 3' end of Ven.a, both ends mentarity-determining region (CDR3) of the TCR43 chain. In ofDp, and the 5' end ofJ43 at each developmental stage (Fig. contrast, rearranged chicken Ig genes isolated from the bursa 2A). The amino acid sequence encoded by the rearranged of Fabricius at this age display multiple blocks of nucleotide V91.1-Ds-J3 genes (Fig. 2B) shows that the junctional vari- substitutions throughout the V regions as a result of intra- ation at the 3' end of Vp and at the 5' end ofJ4 does not extend chromosomal gene conversion (6-10). Sequence variation of more than three amino acids into Vp1.1 or J4. Vp gene segments as a result of gene conversion or somatic The average number ofN nucleotides observed at chicken mutation does not appear to contribute to the TCRF3 reper- Vs-Dp and Dp-J junctions in 18-day clones is 1.1 ± 0.27 toire. (mean + SEM). Rearranged TCR genes cloned from the Identification of the Germ-Line Chicken Dp Gene Segment. thymus at 4 weeks of age have significantly longer N se- The VpJ4junctions isolated from thymus DNA of a 4-week- quence insertions, with an average of 5.1 ± 0.41 (mean ± old chicken (Fig. UA) revealed a 12-bp core sequence that SEM, P < 0.001) nucleotides per joint (Fig. 2A). Specific could represent a germ-line Dp segment. An oligonucleotide mononucleotide and possibly dinucleotide additions are ev- A V131.1 N-Dp?-N J13 germl ine GTCCTACTCAATGACTCAGGCACTTATTTCTGTGCTAAGCAAGATA CAAACACACCACTGAACTTTGGACAGGGCACTCGTCTGACAGTGCTTG 4wk #41 ------CGAGACAGGGAGGC ------4wk #42 ------GGAGGGACAGGGGATCGTTCT ------4wk #43 ------CAGGGGGATCGGGTCCCG ------4wk #44 ------CAGGGGGATCCGGGTC ------

B nonamer heptamer Do heptamer nonamer Ch Do ACACAGGTGGAATTGTTTTTGTACGAGACTGTACCGTTGTG GGGACA GGGGGATC CACAATGATTCA//TTTCCCAGGAGGTCTTTACAAAAACCTGTAT/TTGCATAAGCCA Hu Dp1.1 C--TG--A--GGCA------AGA-----A------.... ----C// ------AC-C-ACGG-A-/AC------C-C-GGC-GTCCCAA-T Mu Dp1.1 CT-TG--A--GTCCT ------TA-AG ----A-A ------C// ---GG------AT-C-ATGG-A-/-C------AT-C-G-CTGTCCCAAGG Mu D32.1 TGTGG--AA---ACT------TC-CGA----A-A------T -----GG------AC-GGAAG---T/-CT------G--C--CCCAAA-A--CA-C AGC Hu D032.1 -GCAG--A- ---ACA------TC-TGG ----A-A----|------T -----GG/ ---G--G----- GG-AGAGG --- T/-CT------A-CC-GATGCAGTA-GCATC FIG. 1. (A) Nucleotide sequence alignment of rearranged chicken TCRT genes with germ-line Vpj.l and J.13 segments (11). Dashes indicate identity with the germ-line sequence. A 12-base-pair (bp) core sequence for the putative D13 segment is underlined. (B) Alignment of nucleotide sequences of the chicken (Ch), human (Hu) (18, 19), rabbit (R) (20), and murine (Mu) (21, 22) genomic Dp segments. Dashes indicate identity with the top line of sequence, and slashes denote gaps introduced for alignment. The D£p region is boxed, and conserved heptamer and nonamer elements of the recombination signal sequence are underlined. Downloaded by guest on September 28, 2021 Immunology: McCormack et al. Proc. Natl. Acad. Sci. USA 88 (1991) 7701

A B -+ - 4- DB - - - + N -f J Vp.1- J3 V ,B1- X N -.' Dp -- AAGCAAGATA GGGACAGGGGGATC CAAACACACCA C A K Q D G T § G S N T P L N F 18 day embryo G Q G D D R G I 4 ------G -- ...... 21 ------T - 18 day embryo 9 -- TTT - G------ACG 23 ---- C------TCGGAAT 4- R G S 11 AG ...... 21 - - M Q I 18 CCGG-- GT 9 - - - - I Y R G ID A 25 G QG D R 13 ...... 25 WJ T Q G 26 cc------13 - - R G CGGAA -G------GGA 26 R D R G I 3 ------1 - G R D R Q I G 10 TAGGGC ------ATCGAG ------3 - A GSI 14 -G - 10 V G H R G S R 15 14 G T GG S 16 G ...... - 15 - - - - G Q G D 24 GAACT --- 16 G T GG S - -...GG 24 .. G TD 8 -- - --GG 7 - H G S 7 C G 5 5 - T GGS 12 - TGCGACC ---- GCGG 12 M R P D A D

4 weeks post-hatching 4 weeks post-hatching

31 ------G------GGTCCCG ------31 - - - - - T Q G S G P A - ... 64 CGT TACCCGGGA ...... 64 - - - - - R Q G T RE ...... - - - - - 51 ...... GTATGAGACC --A------GA 67 R G I R V 67 ...... ------CGGGTC 56 - - -- A PG T S 56 CGCCA----- GT ...... 28 -- - - A R Q G G 28 ...... CGA ------AGGC 55 -- -- G T Q G S T 55 ------GT ------GACT 59 - - - - A P G I K N 59 CCCCC - - AAAAAT 53 - - - - A S P G G S HH 53 CTTCCC ------ACACC 52 - - -- A L R R T F 63 GCTTTCCGTCGG ------AGGGTTTG 30 - - - - G R D R G S F S 52 CTTTGAGAC CCTT 57 - - - - E T V 58 GCTT------TAGATTGG 32 - - - - KD R E 30 GGA------ATCGTTCT 65 - - - P T R D R GN L SX 57 GA ------AGT 68 --- K L V R T G T D 32 GA - GGGAA 54 - - - C T G T G GS R D 65 CTACCC------ACCTCTCATA ...... 62 - - - G P F N S T G V P 68 AAGTTAGTAC ------ACGG 60 - - - V S G T SG R 54 TGTACC ------GAGGG 61 - - TSS§ V A 62 GGCCCCTTTAACTCA ----- TGTACCA 60 AGTATCA------TCGGGAAGG 66 CCCCTCCGACTT ------GG 27 TGA --- AC 61 CTTCGTCC --- GTTG

FIG. 2. (A) Nucleotide sequences of43 rearranged Vp11-Dp-J3junctions, isolated from 18-day embryonic and 4-week post-hatching thymus DNA. Clones were picked and numbered at random and are listed in order of decreasing length of germ-line Vp and Jp sequence. Sequences are and possible P regions (16, 26) are are aligned with the germ-line V1 1, Dp, and J,33 sequences (top line). N nucleotides (23-25) indicated, underlined. Dashes indicate identity with the top line of sequence. Similar results were observed in 20 rearranged V,1 1-D,-J4 genes (16 in-frame, 4 out-of-frame). (B) Amino acid sequences of rearranged in-frame chicken V61.1-Dp-J3 junctions isolated from 18-day embryonic and 4-week post-hatching thymus DNA. Dashes indicate identity with the top line of sequence. The germ-line Do-encoded glycine residues are underlined. ident at the ends of some full-length V, Dp, and J segments, mammalian (30-32) TCRB genes. Thirty-three of 35 in-frame which appear to be palindromic to the coding-region nucle- junctions (94%) contain at least one Dp-encoded glycine (Fig. otides adjacent to the signal-sequence heptamer element. 2B). Originally reported for rearranged chicken Ig light-chain In contrast to the conserved glycine in most of the Vp- genes (16), the generality of specific palindromic nucleotide D#-J junctions examined, the amino acid sequences of additions (P regions; ref. 26) has been extended to mamma- CDR3 on each side ofthe central Dwencoded glycine residue lian TCR (26) and Ig heavy-chain (27-29) genes and to are hypervariable. At position -1 from the central glycine, chicken TCRB genes (this report). only three different amino acid residues are encoded by the Out-of-frame and in-frame junctions are observed at both germ-line Dp, yet 10 of the 20 possible amino acids are found rearrangements we examined (Fig. developmental time points with the same average number of in the 51 in-frame TCR/3 2B). At position -2, where two different amino acids are N nucleotides (Fig. 2A). At embryonic day 18, there are 1.1 encoded by 14 different amino acids are observed. N nucleotides per in-frame joint (n = 48) and 1.3 per DO, Similarly, at positions + 1 and +2, rearranged genes encode out-of-frame joint (n = 10); at 4 weeks post-hatching, there 12 and 11 different amino acid residues, respectively. are 4.9 N nucleotides per in-frame joint (n = 54) and 5.9 per Chicken TCR.3 Diversity in the Mature Spleen. The contri- out-of-frame (n = 14). joint bution of early and late TCR,8 gene rearrangements to the A Conserved Glycine Residue Occurs Within the Hypervari- TCR repertoire in the periphery was determined by isolating able CDR3 Domain. Early and late chicken TCRf3 gene rearranged TCR,8 genes from spleen DNA of a 6-week-old rearrangements display extreme junctional diversity at the chicken. The nucleotide and amino acid sequences of rear- Vg-Ds and Dp-J junctions (Fig. 2A). However, one striking ranged splenic V1.-D9-JP3 genes (data not shown) reveal feature of CDR3 is the presence of at least one Dp-encoded that most in-frame junctions (12 of 14) include a Dp-encoded glycine (Fig. 2B). The amino acid sequences of all germ-line glycine residue. Both early and late TCR3 clones, charac- chicken and mammalian Dp segments include at least one terized by little or no N-nucleotide addition and extensive glycine in all three reading frames (Fig. 1B), and all three Dp N-nucleotide addition, respectively, appear to contribute to reading frames are used in rearranged chicken (Fig. 2B) and the adult splenic T-cell repertoire. Downloaded by guest on September 28, 2021 7702 Immunology: McCormack et al. Proc. Natl. Acad. Sci. USA 88 (1991) DISCUSSION studies based on high-resolution x-ray structures ofthe Ig Fv region (37) suggest that these conserved framework regions In this report we demonstrate that in contrast to the chicken may be folded into a P-sheet structure. Ig loci, in which somatic diversity is generated by gene In the case of the TCR molecule, which appears to have a conversion, somatic diversity is generated in the chicken three-dimensional structure similar to Ig, the conserved TCR(3 locus by junctional variation during V(D)J joining. regions of Vp and Jp could serve to position the CDR3 domain Much ofthe TCR.B diversity is also generated by the addition ofthe TCRN8 chain with respect to the antigen-binding groove ofnontemplated N nucleotides, and the length ofN-sequence of the MHC molecule (2, 38). The CDR3 domains of the addition at Vp-Ds and Dp-Js junctions increases during TCRa and -(3 chains may contact antigen/MHC along the development. It appears unlikely that selection at the protein antigen-binding groove of the MHC molecule, whereas the level is responsible for the appearance oflonger N sequences, CDR1 and CDR2 domains may interact predominately with because out-of-frame junctions are observed at both devel- the a-helices of the MHC molecule (Fig. 3B). If the TCR,3 opmental time points with the same average number of N CDR3 encodes a loop that is positioned on the MHC molecule nucleotides as in-framejoints (Fig. 2A). These results suggest in such a way as to interact with peptide antigen, the a-carbon that the neonatal TCR,8 repertoire in the thymus is different backbone of the TCRN chain may be required to extend into, from that generated later in life and is encoded primarily by along, or across the groove and then turn back. We favor a the germ-line TCRP3 gene segments. model in which the loop extends along either side of the The observation that N-segment length increases during antigen-binding groove, because this position can accommo- development suggests that the expression of TdT, the en- date the variable lengths of CDR3 observed in early and late zyme believed to mediate N-nucleotide addition to Ig and TCRJ3 clones. The presence of one or more glycine residues TCR genes, may be developmentally regulated in the chicken may provide enough flexibility to allow the loop to cross over thymus. TdT levels increase gradually in the developing antigen positioned within the groove between the two MHC thymus of mice and rats (33, 34), and the length of N-nucle- a-helices. Such a loop may be formed by a type II or glycine otide addition increases during development in rearranged turn, in which all ofthe a-carbons of adjacent amino acids in mammalian Ig heavy-chain (27-29) and TCR (26, 35) genes. the turn are oriented in one plane (39, 40). Glycine residues Developmentally regulated N-sequence addition therefore are also preferentially located in 2-, 4-, and 5-residue loops appears to be an evolutionarily conserved mechanism for associated with (3-hairpins in globular proteins (41) and in the generating diversity in vertebrate Ig and TCR genes. The lack longer co loop structure, which is 6-16 residues long (42). That of N segments in rearranged chicken Ig genes (6-10) is most chicken TCR,8 rearrangements include at least one probably due to developmentally programmed Ig gene rear- Dp-encoded glycine residue supports the hypothesis that a rangement (36) prior to expression of TdT in chicken lym- glycine-containing turn or loop may be important for the phoid stem cells. structure of the TCRP3 CDR3 domain. If junctional variability and increasing N-sequence addi- Interestingly, most rearranged human and murine TCR8 tion are the mechanisms used to create CDR3 diversity, and genes have a Ds-encoded glycine (35, 43-45). In addition, the TCRs undergo both positive and negative selection in the preferred reading frame ofthe murine Ig heavy-chain DsP2 and thymus, why is a Do element required, and why is it so highly DFL16 gene segments includes a glycine codon (27-29, 46, 47). conserved? The requirement for twojoining events involving These observations suggest that a D-encoded glycine- Dp increases the junctional variability of CDR3, which is containing turn may be important for the structure of the encoded at the Vp-D9-J9 junction. One striking feature of antigen-binding CDR3 loop in TCR(3, TCR8, and Ig heavy CDR3, however, is the presence of at least one D1-encoded chains. glycine residue (Fig. 2B) within the otherwise hypervariable Of the two clones we observed in the thymus that lacked a CDR3 domain. In addition to conservation ofthe Do element, glycine in CDR3 (Fig. 2B), one (clone 12) encodes a proline comparison of the consensus amino acid sequences for residue, which would provide a kink in the a-carbon backbone chicken and mammalian Vat, VP2, and Jp shows that CDR3 is to allow formation of a CDR3 loop. The observed bias for the also flanked by regions encoded by the Vp and Jo segments addition of G nucleotides by TdT during TCR and Ig gene that have been evolutionarily conserved (Fig. 3A). Modeling rearrangement (23-25) may provide an additional mechanism A

VP1 chicken N D S G T Y F C A K Q D + D S + Y + C A - - G T § G S J/3 marmat E D S A VY L C AS S L chicken G Q G D D R G I chicken X N X X L I F G XG T K L T V L Vo2 F G G T + L T V L G T G G X mauna I XX X X X X F G X G T R L T V L chicken N H S A I Y F C AS S F E D manunal G X G X ++.+. Y F C A SS + D X G X mammaal S Q T S VY F C AS S D R G FIG. 3. (A) Comparison of consensus amino acid sequences for chicken and mammalian Vp1, Vp2, Dp, and Jp segments (11, 37). X indicates a position at which no consensus residue has been identified. Alignments were made based on the PAM250 amino acid similarity matrix using the AALIGN program (DNASTAR). A positive relationship (+), a 0-value relationship (blank space), a negative relationship (-), and identity between residues are indicated in the line between the chicken and mammalian consensus sequences. Conserved Vp andJ4 regions and the central glycine of Do are underlined. (B) Schematic model of TCR/antigen/MHC interaction (2, 38). The MHC molecule is depicted as two a-helical regions atop a 13-pleated sheet, and the antigen is denoted by the triangular prism. Potential sites of TCRa- and -(3-chain CDR1 and CDR2 interaction with the MHC molecule, and of the TCRa-chain CDR3 interaction with antigen/MHC, are depicted as dashed ovals. The ends of the conserved Vp, and J4 regions (see A), symbolized by the filled rectangles, may provide anchor points by which the hypervariable loop of the TCR,8 chain (dotted line) is fixed in position. To account for the variable length ofthe CDR3 loop (Fig. 2), the loop is depicted as extending away from the center of the MHC antigen-binding groove. However, the loop may extend into, along, or across the antigen-binding groove and may interact with antigen and/or both a-helices of the MHC molecule. Within the hypervariable TCR13 loop, the G symbolizes the conserved glycine residue encoded by the Dp segment. Downloaded by guest on September 28, 2021 Immunology: McCormack et al. Proc. Nati. Acad. Sci. USA 88 (1991) 7703

to ensure that CDR3 encodes a glycine or proline turn, because Schat, K. A., Thompson, C. B. & Cooper, M. D. (1991) addition of G nucleotides to the sense strand of Vat or Do, or FASEB J. 4, A613. C. H., Bucy, R. P. & Thompson, C. B. or of 13. Cooper, M. D., Chen, to the antisense strand of D,3 J4, results in the creation (1991) Adv. Immunol. 50, 87-117. glycine codons (GGN) or proline codons (CCN), respectively. 14. Schiffer, M., Wu, T. T. & Kabat, E. A. (1986) Proc. Natl. Clone 52 in Fig. 2B lacks both glycine and proline in CDR3. Acad. Sci. USA 83, 4461-4463. However, because TCR gene rearrangement is ongoing in the 15. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, J. J., Higuchi, thymus (2), one might speculate that this clone represents a T R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science cell that has not undergone positive selection for antigen/ 239, 487-491. MHC binding. One prediction ofthis model would be that most 16. McCormack, W. T., Tjoelker, L. W., Carlson, L. M., Petry- C. B. genes isolated from peripheral T cells (e.g., niak, B., Barth, C. F., Humphries, E. H. & Thompson, rearranged TCR,8 (1989) Cell 56, 785-791. spleen) might have a glycine turn in CDR3, having undergone 17. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular selection for interaction with antigen/MHC via the CDR3 Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold loop. Most in-frame rearranged TCR8 genes isolated from the Spring Harbor, NY). splenic DNA from a 6-week-old chicken do include a D- 18. Clark, S. P., Yoshikai, Y., Taylor, S., Siu, G., Hood, L. & encoded glycine (data not shown). Interestingly, each of the Mak, T. W. (1984) Nature (London) 311, 387-389. two splenic TCRf3 clones lacking a glycine turn has a very 19. Tunnacliffe, A. & Rabbitts, T. H. (1985) Nucleic Acids Res. 13, short CDR3 (3 or 4 amino acids encoded by N-D-N), 6651-6661. suggesting that in contrast to clones that encode larger CDR3 20. Harindranath, N., Alexander, C. B. & Mage, R. G. (1991) Mol. domain of these TCRfi chains may not Immunol., in press. domains, the CDR3 M., Strauss, E., Haars, R., Mak, T. W. groove. of 21. Siu, G., Kronenberg, extend into or along the antigen-binding Inspection & Hood, L. (1984) Nature (London) 311, 344-350. human TCR,8 sequences reveals that those sequences with 22. Kavaler, J., Davis, M. M. & Chien, Y. (1984) Nature (London) short CDR3 domains may also lack a glycine turn; however, 310, 421-423. nearly all human TCR"3 sequences with longer CDR3 domains 23. Sakano, H., Kurosawa, Y., Weigert, M. &Tonegawa, S. (1981) encode a glycine or proline in CDR3 (37). Nature (London) 290, 562-565. In summary, although there appears to be limited diversity 24. Alt, F. W. & Baltimore, D. (1982) Proc. Nati. Acad. Sci. USA in the germ-line elements that encode the chicken TCR/3 chain, 79, 4118-4122. extensive junctional diversity is created during TCR,8 gene 25. Desiderio, S. V., Yancopoulos, G. D., Paskind, M., Thomas, rearrangement in the chicken thymus. Junctional diversity E., Boss, M. A., Landau, N., Alt, F. W. & Baltimore, D. Nature (London) 311, 752-755. as a lengths (1984) increases during development result of increasing 26. Lafaille, J. J., DeCloux, A., Bonneville, M., Takagaki, Y. & of N-nucleotide addition at both the Vp-Ds and D3-Jp junc- Tonegawa, S. (1989) Cell 59, 859-870. tions. Increased N-sequence addition during development 27. Gu, H., Forster, I. & Rajewsky, K. (1990) EMBO J. 9, expands the TCR,8 chain repertoire by the addition of amino 2133-2140. acids along each side of the hypervariable CDR3 loop. The 28. Feeney, A. J. (1990) J. Exp. Med. 172, 1377-1390. conservation of the amino acid sequence at the 3' end of Vi, 29. Meek, K. (1990) Science 250, 820-823. at the 5' end of Jo, and in the Do segment suggests that the 30. Concannon, P., Pickering, L. A., Kung, P. & Hood, L. (1986) hypervariable portion ofCDR3 may be positioned with respect Proc. NatI. Acad. Sci. USA 83, 6598-6602. So- to the antigen-binding groove of the MHC molecule by the 31. Barth, R. K., Kim, B. S., Lan, N. C., Hunkapiller, T., Winoto, A., Gershenfeld, H., Okada, C., Hansburg, V,3 and sequences and bieck, N., conserved structural features of the Jo D., Weissman, I. L. & Hood, L. (1985) Nature (London) 316, a glycine-containing turn encoded by Do. This glycine turn is 517-523. conserved in the nucleotide sequence of the germ-line Do 32. Goverman, J., Minard, K., Shastri, N., Hunkapiller, T., Hans- segment of chickens, mice, rabbits, and humans. burg, D., Sercarz, E. & Hood, L. (1985) Cell 40, 859-867. W. & We thank Jeff Leiden for review of the manuscript and Tom 33. Gregoire, K. E., Goldschneider, I., Barton, R. Bollum, Garrett for helpful discussions. We thank Danie Osborn for oligo- F. J. (1979) J. Immunol. 123, 1347-1352. for preparation of figures. 34. Rothenberg, E. & Triglia, D. (1983) J. Immunol. 130, 1627- nucleotide syntheses and Beverly Burck 1633. This work was supported in part by a Public Health Service Award Institutes of Health. 35. Krangel, M. S., Yssel, H., Brocklehurst, C. & Spits, H. (1990) to C.B.T. from the National J. Exp. Med. 172, 847-859. 1. Alt, F. W., Blackwell, T. K., DePinho, R. A., Reth, M. G. & 36. McCormack, W. T., Tjoelker, L. W., Barth, C. F., Carlson, Yancopoulos, G. D. (1986) Immunol. Rev. 89, 5-30. L. M., Petryniak, B., Humphries, E. H. & Thompson, C. B. 2. Davis, M. M. & Bjorkman, P. J. (1988) Nature (London) 334, (1989) Genes Dev. 3, 838-847. 395-402. 37. Kabat, E. A., Wu, T. T., Reid-Miller, M., Perry, H. M. & 3. Hunkapiller, T. & Hood, L. (1989) Adv. Immunol. 44, 1-63. Gottesman, K. S. (1987) Sequences ofProteins ofImmunolog- 4. Blackman, A., Kappler, J. W. & Marrack, P. (1988) Immunol. ical Interest (U.S. Dept. Health Human Serv., Washington). Rev. 101, 5-101. 38. Claverie, J. M., Prochnicka-Chalufour, A. & Bougueleret, L. 5. von Boehmer, H., Karjalainen, K., Pelkonen, J., Borgulya, P. (1989) Immunol. Today 10, 10-14. & Rammensee, H.-G. (1988) Immunol. Rev. 101, 21-37. 39. Chou, P. Y. & Fasman, G. D. (1977) J. Mol. Biol. 115,135-175. 6. Reynaud, C.-A., Anquez, V., Dahan, A. & Weill, J.-C. (1987) 40. Richardson, J. S. & Richardson, D. C. (1989) in Prediction of Cell 40, 283-291. Protein Structure and the Principles ofProtein Conformation, 7. Reynaud, C.-A., Dahan, A., Anquez, V. & Weill, J.-C. (1989) ed. Fasman, G. D. (Plenum, New York), pp. 1-98. Cell 59, 171-183. 41. Sibanda, B. L. & Thornton, J. M. (1985) Nature (London) 316, 8. Reynaud, C.-A., Anquez, V., Grimal, H. & Weill, J.-C. (1987) 170-174. Cell 48, 379-388. 42. Leszczynski, J. F. & Rose, G. D. (1986) Science 234, 849-855. 9. Carlson, L. M., McCormack, W. T., Postema, C. E., 43. Elliot, J. F., Rock, E. P., Patten, P. A., Davis, M. M. & Chien, Humphries, E. H. & Thompson, C. B. (1990) Genes Dev. 4, Y. (1988) Nature (London) 331, 627-631. 536-547. 44. Loh, E. Y., Cwirla, S., Serafini, A. T., Phillips, J. H. & 10. McCormack, W. T. & Thompson, C. B. (1990) Genes Dev. 4, Lanier, L. L. (1988) Proc. Natl. Acad. Sci. USA 85, 9714-9718. 548-558. 45. Asarnow, D. M., Kuziel, W. A., Bonyhadi, M., Tigelaar, 11. Tjoelker, L. W., Carlson, L. M., Lee, K., Lahti, J., McCor- R. E., Tucker, P. W. & Allison, J. P. (1988) Cell 55, 837-847. mack, W. T., Leiden, J. M., Chen, C. H., Cooper, M. D. & 46. Ichihara, Y., Hayashida, H., Miyazawa, S. & Kurosawa, Y. Thompson, C. B. (1990) Proc. Natl. Acad. Sci. USA 87, (1989) Eur. J. Immunol. 19, 1849-1854. 7856-7860. 47. Kaartinen, M. & Makela, 0. (1989) Immunol. Today 6, 324- 12. Lahti, J. M., Chen, C. H., Tjoelker, L. W., Calnek, B. W., 327. Downloaded by guest on September 28, 2021