We estimate there are ∼15 IgM H chain loci in the nurse shark genome and have characterized one locus. It consists of one V, two D, and one J germline gene segments, and the constant (C) region can be distinguished from all of the others by a unique combination of restriction endonuclease sites in Cμ2. On the basis of these Cμ2 markers, 22 cDNA clones were selected from an epigonal organ cDNA library from the same individual; their C region sequences proved to be the same up to the polyadenylation site. With the identification of the corresponding germline gene segments, CDR3 from shark H chain rearrangements could be analyzed precisely, for the first time. Considerable diversity was generated by trimming and N addition at the three junctions and by varied recombination patterns of the two D gene segments. The cDNA sequences originated from independent rearrangements events, and most carried both single and contiguous substitutions. The 53 point mutations occurred with a bias for transition changes (53%), whereas the 78 tandem substitutions, mostly 2–4 bp long, do not (36%). The nature of the substitution patterns is the same as for mutants from six loci of two nurse shark L chain isotypes, showing that somatic hypermutation events are very similar at both H and L chain genes in this early vertebrate. The cis-regulatory elements targeting somatic hypermutation must have already existed in the ancestral Ig gene, before H and L chain divergence.

The diversity of the Ab repertoire is initially created by the numerous different rearrangements of the V(D)J gene segments that occur during B lymphocyte differentiation (1). In mammals, the Ig H chain locus can consist of 1–3 Mb of 100–200 tandemly duplicated VH, followed by multiple D and JH gene segments that rearrange into fused VDJ combinations (V regions) which are transcribed with the exons of the constant (C) regions. Diversity is generated by the combinatorial variation in gene segment usage and the imprecise joining resulting from deletion and insertion at the VH/D, D/JH, and VL/JL junctions of the H and L chains, respectively.

In mouse and human beings, somatic hypermutation is a mechanism that hones Ab specificity in the course of an immune response by creating higher affinity ligand-combining sites in the Ag-activated B cells (2, 3, 4, 5, 6, 7). In some species where the B cell repertoire develops in gut-associated tissues, such as sheep, hypermutation operates before Ag stimulation of B cells, and then it serves to expand the primary Ab repertoire (8). In mammals, the mutations consist almost entirely of single-base, transition-biased changes. The mutating activity is centered within a 1- to 2-kb region beginning downstream from the variable (V) region gene promoter (9). Although the rearranged V gene is thus targeted, the actual coding sequence per se is not necessary for the process, the V gene can be replaced by other, unrelated sequences (10). Only experiments involving the removal of the Ig enhancer and the nuclear matrix attachment region (MAR)3 components affect the ability of the region to be mutated (11, 12, 13). The role of the intronic enhancer/MARs in hypermutation is not clear but it seems to be one separable from transcription regulation (14). It has been speculated that these (or other) cis-elements are involved in the recruitment of various factors, including activation-induced cytidine deaminase (AID)3and replication protein A, assembling a “mutasome” that most likely induces DNA lesions which become subject to error-prone repair (12, 15, 16, 17).

Rearranging Ig superfamily genes and adaptive immune systems are only found in the jawed vertebrates (18), and AID has been isolated from chicken, Xenopus, zebrafish, and spotted dogfish (see Ref.19). It is not clear whether hypermutation preceded or followed the acquisition of the RAG recombinase in the primitive vertebrate system, ∼450 million years ago. After the report of a novel shark serum molecule (new Ag receptor (NAR)) that was neither Ig nor TCR but rearranged and accumulated extensive substitutions (20), we began studies characterizing the evolution of the hypermutation process with respect to the conventional Ab genes.

In cartilaginous fishes, such as sharks and skates, studies in hypermutation have been hampered by the alternative gene organization of their Ig and Ig-related molecules. Although TCR is apparently organized as in all other vertebrates (21), secreted immune function molecules such as IgM and NAR are encoded by an estimated 100–200 independently rearranging miniloci (clusters), each locus consisting of few V(D)J gene segments and one set of constant region exons (22). However, in nurse shark, of the 70 L chain loci (23), we were able to distinguish a small family of six germline genes encoding one L chain isotype (NS3) and characterize their mutated products. Unlike Ig mutants from other vertebrates, the nurse shark pattern of mutations not only consisted of point mutations but also of tandem changes of 2–5 bp adjacent substitutions (24). This is a prominent feature, involving half of the mutations in L chains and in the nurse shark NAR (25). NAR does not associate with L chains but forms a dimer, and the NAR V gene segment is equally related to both Ig and TCR, consistent with its ancient origins in elasmobranch fishes. Although NAR and IgL diverged early, they have similar hypermutation patterns in the nurse shark. In contrast, in the related species horn shark, mutations in the H chain (26) apparently targeted guanyl and cytosyl, like the IgH locus of the amphibian Xenopus (27), or mutating cultured cell lines like human Burkitt lymphomas, the chicken DT40 lymphoma, and the mouse 18-81 pre-B cell line (28, 29, 30). In the cell lines the mechanistic basis for the switch from a relatively unbiased nucleotide target distribution in vivo to a restricted GC substitution pattern in vitro is not clear, nor is the relationship of the Xenopus mutator with that of any other tetrapod.

It was of interest to ascertain whether the nurse shark H chain genes mutated like the horn shark H chain genes or like the nurse shark L chain genes. It seemed unlikely the two closely related species would have variant mutating mechanisms but even more improbable that H and L chain genes would mutate differently within a lymphocyte. It was possible to monitor mutation at NS3 L chain loci because there were just a few genes; for the IgM H chain we anticipated many more germline loci. To this end, we began characterization of the nurse shark IgH loci, focusing on identifying the germline members of one particular subset with unique traits in both V and C sequences and proceeded to look for somatic mutants.

As a result of these efforts, we report that (1) despite the many genes, nurse shark IgM loci can be classified into distinguishable subsets based on both the V and C region sequence; (2) with the identification of individual loci, their rearrangements could be analyzed and junctional diversity is shown to be the most important contributor to shark Ab repertoire heterogeneity and (3) shark H chain mutants can be identified for the first time, conventionally, through selection by the C region sequence.

Nurse sharks (Ginglymostoma cirratum) were captured off the coast of the Florida Keys. Whole blood was obtained from the caudal sinus of nurse shark Y, estimated to be 2 years old at the time of bleeding, and the cells were passed through a Ficoll gradient to separate PBL from erythrocytes (RBC). Genomic DNA was extracted from the RBC fraction and used to construct the genomic library. RNA was isolated from the PBL for RT-PCR experiments. Shark 33, a 2.5-foot adult male, was sacrificed and its organs were harvested. Genomic DNA was also extracted from RBC and total RNA was obtained from shark 33 epigonal organs and spleen. Neonatal shark AQ was removed by Caesarian section and sacrificed within 1 wk.

Poly(A) RNA was obtained from shark 33 epigonal organ RNA using a mRNA purification kit (Ambion). A cDNA library was constructed (λZAP ExpressXR; Stratagene and Lofstrand Labs Limited) and 550,000 phages were screened with nurse shark IgM constant (C) region probe (“CH-12”). As a result, 279 phages (1/2000 PFU) were selected for further analysis. Phages chosen for DNA sequencing were purified and the cloned inserts excised into the phagemid vector (pBK-CMV). Seven hundred fifty-five thousand phages (calculated by the area covered by nine filters of 420 cm2) from a primary Sau3A I partial genomic library (shark 33 RBC DNA, λ Fix II; Stratagene and Lofstrand Labs Limited) were similarly screened. This is ∼3.4 nurse shark genomes, calculated from an estimated average phage insert size of 16.9 kb and a haploid genome size of 3.75 × 109 bp.

Oligonucleotide primers were synthesized by Life Technologies. PCR conditions were described previously (24). Primers used were as follows: CH1 is 5′-GGYTGYTTGGCGATGGA-3′; CH2-3′ is 5′-ACCTGGCAKGTATARAC-3′; CH2-5′ is 5′-CTCCTAACTGTGAGTTC-3′; gCH1-5′ is 5′-GAYGGTTCTGTGATYTTTGGT-3′; and gCH1-3′ is 5′-TGGCATTCCARYGCTCTTSTC-3′. All of the primers were derived from a comparison of C region sequences cloned during the preliminary phase of the studies. The CH1/CH2-3′ primer pair was used to obtain a 488-bp fragment in RT-PCR or in the course of screening the cDNA phage library. CH2-5′/CH2-3′ amplified a 239-bp fragment from the Cμ2 exon and gCH1-5′/gCH1-3′ amplified the Cμ1 exon; these were used in screening the genomic library. DNA templates consisting of denatured phage lysate did not exceed 17% of the final volume of the PCR mixture.

The primer combination V18-1 (5′-ACCAGAATGACGACGATG-3′) and JH2 (5′-TCACGGTCACCATGGT-3′) was used to amplify V gene segments from phage lysates that contained group 2 Cμ2.

V18-1/JH2 was used to amplify V gene segments from phage lysates that contained group 2 Cμ2. The 1.6-kb fragment consists of 6 bp of the 5′ UT, the leader, leader intron (124 bp), and the V, D, and J gene segments.

V18-1 and CH2R (at the end of the Cμ2 domain, 5′-CCTGAGGTCCAGTGATGT-3′), both specific for group 2 sequences, were used for RT-PCR from spleen of shark AQ to obtain neonatal Ig sequences.

RT-PCR assays were performed using oligo(dT) for priming first-strand synthesis, and PCR products were cloned into pGEM-T-Easy (Promega) for sequencing.

Sequencing of the plasmids was performed by the DNA Sequencing Laboratory at Rockefeller University (New York, NY).

After an initial comparison of RT-PCR-generated μ sequences from shark Y and shark 33, restriction enzyme site patterns were established for differentiating the C regions by the second C domain. HindIII was discovered as the defining marker for group 2 H chains in both animals. An amplified erythrocyte genomic library (shark Y) was screened to test our classifications of H chains, and two phages (V18, V32) were isolated and characterized as members of group 2 IgM genes. At this point, a genomic erythrocyte library and an epigonal organ cDNA library were constructed from tissue originating from one individual, shark 33. These primary libraries were screened for group 2 genes to establish the germline gene sequences for comparison to their rearranged transcripts.

Based on the literature on horn shark IgM organization, we anticipated that in our model, the nurse shark, there would also exist many miniloci, each one consisting of VH, D, and JH gene segments and a set of μ exons. For our hypermutation studies, we needed to identify a single nurse shark IgM H chain locus that encoded a functional μ-chain with unique, distinguishable sequence characteristics. The Cμ2 exon was targeted because it and Cμ3 encode the least conserved domains in vertebrate μ sequences (31). Initially, degenerate primers to the 5′ end of Cμ1 (primer CH1) and the 3′ end of Cμ2 (primer CH2-3′), based on a nurse shark cDNA sequence (accession number M92851), were used to amplify cDNA from shark Y PBL and shark 33 spleen and epigonal organ (∼488-bp fragment). Twenty PCR clones were sequenced; one (CH-12) carried a unique HindIII site and, in contrast to most of the others, lacked an EcoRI site. The original RT-PCR products were then analyzed with these enzymes. It was confirmed that the majority of Cμ sequences from both sharks carried an EcoRI site, whereas a 5–10% minority carried a HindIII site (RT-PCR sequences and gel not shown).

We then screened a secondary shark Y genomic DNA library to characterize IgM germline genes containing Cμ2 exons with a HindIII site. Twenty-four phages were selected with the CH-12 probe. Cμ2 exon sequences were amplified from the phage lysates using degenerate PCR primers (CH2-5′/CH2-3′) directed to conserved regions of Cμ2, which were based on the RT-PCR sequences described above. PCR products of ∼239 bp were generated from 23 of 24 phage lysates; the one PCR-negative sample carried VH and Cμ1 but not Cμ2. The majority of Cμ2+ sequences (16 of 22) were positive for EcoRI and negative for HindIII; the rest were negative for both (5 of 22) sites or negative for EcoRI and positive for HindIII (2 of 22) (gel not shown).

Of the two phages that carried Cμ2 with HindIII sites, a visibly smaller PCR fragment (by 13 bp) than expected was generated from one phage lysate (V32). On further analysis, the V32 phage was also found to have a defective Cμ1. The other phage, called V18, carried functional sequences and is described below.

The V, D, and J gene segments (Fig. 1) and the first two exons of Cμ (data not shown) of genomic clone V18 were subcloned and sequenced. The V18 Cμ2 can be distinguished by the presence of both a BstEII and HindIII site. The sequence of the genomic Cμ is identical to the RT-PCR clone CH-12.

FIGURE 1.

Germline V, D, and J gene segments from genomic clone V18. Portions of phage V18 containing VH, D1, D2, and JH gene segments and Cμ1 and Cμ2 were subcloned into pUC19, and the sequences were determined. The coding sequences are indicated with the single letter amino acid code. The CDR1 and CDR2 in the V are underlined, and the numbering is from Ref.48 , according to homology with the human VH3 subgroup. Splice signals between leader and VH and 3′ of JH are in boldface, as are the heptamers and nonamers of the RS sequences. D2 carries RS with 12 bp spacers on either side. No conventional octamer sequence was observed in the 200 bp upstream of the sequence available (accession number DO192492).

FIGURE 1.

Germline V, D, and J gene segments from genomic clone V18. Portions of phage V18 containing VH, D1, D2, and JH gene segments and Cμ1 and Cμ2 were subcloned into pUC19, and the sequences were determined. The coding sequences are indicated with the single letter amino acid code. The CDR1 and CDR2 in the V are underlined, and the numbering is from Ref.48 , according to homology with the human VH3 subgroup. Splice signals between leader and VH and 3′ of JH are in boldface, as are the heptamers and nonamers of the RS sequences. D2 carries RS with 12 bp spacers on either side. No conventional octamer sequence was observed in the 200 bp upstream of the sequence available (accession number DO192492).

Close modal

The organization of the V18 Ig gene segments is similar to that described in horn sharks (32). As in horn sharks, no octamer was found in the region of the TATA box-like sequence at 64 nuc (Fig. 1) or in the available region 200 bp upstream of the leader sequence. The VH gene segment is flanked by recombination signal (RS) sequences with a 23-bp spacer (23-RS). Downstream are two D genes, D1 with a 12-RS on the 5′ end and a 23-RS on the 3′ end, and D2 flanked by two 12-RS. The JH gene segment is adjacent to a 23-RS and is separated from the first Cμ exon by ∼6.8 kb of intervening DNA; the Cμ1 exon in turn is 2.5 kb upstream of Cμ2.

The V sequence is identical to the V region from a partial cDNA clone, 1E. 1E and three other related V regions (14E, 35S, 72S) from neonatal shark cDNA had been classified as “group 2” IgM V regions (33). Since there is no somatic hypermutation in pups, these V regions would reflect their germline components. Notably, the unique feature of group 2 VH regions is a one-codon insertion in framework (FR) 1 (labeled as 1a in Fig. 1) distinguishing them from the other four VH groups. Because of this unmistakable feature and the ease of detecting the group 2 C region, we chose to look for group 2 genes in cDNA and genomic libraries derived from one individual, shark 33.

A genomic library made from erythrocyte DNA from shark 33 was screened with the CH-12 probe. Of 755,000 PFU (3.4 genomes, see details in Materials and Methods), 46 CH-12+ phages were subsequently identified as containing the first and/or second exon of a μ sequence by the use of degenerate primers gCH1-5′/gCH1–3′ and CH2–5′/CH2-3′, respectively. We initially estimated 14 IgM loci per nurse shark genome but revised this number to 15, after additional screening, as described below.

The Cμ2 PCR products from 30 phages were treated with EcoRI or HindIII or ClaI. The characteristics of these phages are shown in Table I. Among the five HindIII-containing sequences, the Cμ2 from three (F2, G3, I1) were shorter than expected, like the pseudogene in V32. We looked for associated V gene segments, but only one (E1) of the functional genes included them. This library carried average insert sizes of 16.9 kb, whereas the distance from the leader to Cμ2 in the V18 phage is long, ∼11 kb. As a result, we were obliged to extend our screening to isolate phages carrying both VH and Cμ2.

Table I.

Characterization of Cμ2 and correlation with VH groups

Genomic Group
Cμ2RestrictionSitesCμ2+ PhagesLoci/genomea
HindIII EcoRI Cla  
− − 6/33 3/genome 
− − 5/33 3/genomeb 
− − 6/33 3/genome 
− − − 13/33 6/genomec 
Genomic Group
Cμ2RestrictionSitesCμ2+ PhagesLoci/genomea
HindIII EcoRI Cla  
− − 6/33 3/genome 
− − 5/33 3/genomeb 
− − 6/33 3/genome 
− − − 13/33 6/genomec 
cDNA    
Cμ2 Restriction Sites cDNA Identified by Cμ2 
HindIII EcoRI Cla 
− − 6 cDNA (4 sequenced, all group 1 VH) 
− − 4 cDNA (N13, N27, N40, N42, all group 2 VH) 
− − 25 cDNA (10 sequenced, all group 4 VH) 
− − − 4 cDNA (4 sequenced, all group 5 VH) 
cDNA    
Cμ2 Restriction Sites cDNA Identified by Cμ2 
HindIII EcoRI Cla 
− − 6 cDNA (4 sequenced, all group 1 VH) 
− − 4 cDNA (N13, N27, N40, N42, all group 2 VH) 
− − 25 cDNA (10 sequenced, all group 4 VH) 
− − − 4 cDNA (4 sequenced, all group 5 VH) 
a

Loci per genome is estimated as follows. Of 46 phages where H chain sequence was detected, only 30 (65%) can be characterized for the Cμ2 H chain group restriction sites. The number of loci per genome was estimated as follows. There were 6 of 30 Cμ2 that carried group 1 characteristics, so that there would have been 9 group 1 loci in 3. 4 genomes, if all 46 phages could have been characterized by Cμ2.

b

If using the same extrapolations described in footnote a, one arrives at two group 2 loci/genome; however, it is obvious that the experimentally derived numbers are small and permit error. After additional screenings of the library, we obtained a total of nine group 2 V sequences: one G2-V1, two G2-V2, three G2-V3 and four G2-V4. Seven of the phages also carry the expected group 2 C exons. As explained in the text, G2-V1 and G2-V2 are believed to be alleles; therefore, there are three group 2 loci of a total of 15 IgM genes.

c

This group might include pseudogenes or others that lost the characteristic restriction enzyme site in Cμ2.

More of the genomic library was screened, this time with a VH probe derived from the V18 phage. The positive phages were subjected to PCR amplification for Cμ2. Phages that contained Cμ2 with a HindIII site were examined for V gene segments. Three kinds of functional V region gene segments were found and named G2V1 to V3 (accession numbers DO192491 to DO192494). One is identical to that in phage V18 (G2-V1); a second (G2-V2) is identical to G2-V1 except for a single nucleotide substitution at codon 51 (TCC to TAC) in CDR2. The third V gene (G2-V3) differs from G2-V1 in the coding regions at codons −16 (ATC), −8 (ACT), 5 (ATC), 16 (AGT), 18 (TTG), 53 (GCT), 71 (AAA), and 101 (GGT), but the D sequences were the same. Partial sequences were obtained from the pseudogene-associated V32 types (G2-V4); these were never found as cDNA in our studies. The germline VH sequences of G2-V1, G2-V2, and G2-V3 are identical to the V regions of neonatal cDNA sequences 1E, 72S/35S, and 14E, respectively.

Our initial estimate based on C region analyses was 14 IgM loci per genome. However, we decided that there were actually three group 2 loci, although initial extrapolations with the small numbers resulted in a count of two loci (Table I, footnote b). From two screenings nine phages with group 2 V region genes were characterized, and they consisted of one G2-V1, two G2-V2, three G2-V3, and three G2-V4 (G2-V1 and G2-V2 are alleles, as will be explained in Distinguishing H chain groups in cDNA library). We conclude that there are three loci each of group 1, group 2, and groups 3/4 and six loci of group 5, making 15 IgM loci per nurse shark genome.

To get an idea of the cDNA representation of group 2 sequences, we initially screened one shark 33 cDNA phage library plate (50,000 PFU) with the CH-12 probe. Forty-four phages were selected, and PCR products were obtained using primer pairs CH1/CH2-3′ or T3/CH2-3′ for restriction enzyme analyses. After the clones were classified by their Cμ2 restriction sites, representatives of each group were partly or completely sequenced. We found a correlation of theVH region sequence with Cμ2 characteristics. That is, those cDNA with HindIII sites in Cμ2 carried group 2 VH, those with ClaI sites group 1 VH, those with EcoRI sites groups 3/4 VH, and those with none of these markers, group 5. Table I, bottom, shows the distribution of the 44 cDNA phages. Fig. 2 shows partial Cμ2 sequences with the group-specific sites underlined.

FIGURE 2.

Comparison of partial Cμ2 sequences. Various cDNA sequences classified as groups 1, 3, 4, and 5 are compared with the reference Cμ2 from the V18 phage. They differ from V18 by 13–22% over 212 bp. cDNA N42 is also group 2 but with a variant C region containing the HindIII site but lacking BstEII. Two variants from group 4 are included to show differences at the 5′ end. The restriction enzyme sequences that characterize the groups are shaded and labeled. Group 3 cDNA was not isolated in this study; the sequence was obtained from clone 6E (33 ).

FIGURE 2.

Comparison of partial Cμ2 sequences. Various cDNA sequences classified as groups 1, 3, 4, and 5 are compared with the reference Cμ2 from the V18 phage. They differ from V18 by 13–22% over 212 bp. cDNA N42 is also group 2 but with a variant C region containing the HindIII site but lacking BstEII. Two variants from group 4 are included to show differences at the 5′ end. The restriction enzyme sequences that characterize the groups are shaded and labeled. Group 3 cDNA was not isolated in this study; the sequence was obtained from clone 6E (33 ).

Close modal

The five VH regions defined by Rumfelt et al. (33) share 74–80% identity between groups, making them distinct from each other, although by traditional definition they are all of one VH family (34). We are confident that just about any genomic or cDNA sequence identified by markers in Cμ2 will be predictably associated with a VH gene segment or region assigned to groups 1–5. We now redefine the nurse shark groups as extending to the C region as well.

Of the four phages designated as group 2 cDNAs (Table I, bottom), three (N13, N27, N40) carried both HindIII and BstEII sites in Cμ2; these sequences were isolated and the C regions were identical to each other as well as to those of CH12/V18. The V region of N13 was identical to the germline G2-V1; the V regions of N27 and N40 were mutants of G2-V1 and G2-V2 rearrangements, respectively. The Cμ2 of the fourth phage N42 carried HindIII but not BstEII sites, and its C region differed slightly throughout (see Fig. 2); its V region proved to be a G2-V3 sequence.

We conclude that there are three group 2 loci. One is a pseudogene (G2-V4, as in V32) with defective Cμ1 and Cμ2 exons, another is G2-V3, and the third is G2-V1/G2-V2, which are probably alleles. The C regions of G2-V1 and G2-V2 are identical to each other and differ from G2-V3. Moreover, G2-V2 and G2-V3 can be found in other sharks in the absence of G2-V1 (e.g., neonatal shark AQ discussed below and adult shark JS and shark J, our unpublished results).

In summary, the nurse shark IgM loci can be classified into five groups. Previously defined by the V region sequence, they can also be readily distinguished by restriction sites in their Cμ2 domain. Group 2 genes characterized in the germline could be thus identified among rearranged sequences. We went on to screen for cDNA sequences from shark 33 bearing G2-V1 or G2-V2 rearrangements.

In total, 279 CH+ phages were selected from the shark 33 epigonal cDNA library. PCR products from CH1/CH2-3′ amplification were obtained from 231 lysates, and 36 had HindIII sites. Of these, 23 carried both HindIII and BstEII sites. Twenty-two phages were isolated and the complete Ig sequences were determined; they all contained C regions identical to each other (Fig. 3, bottom). Three carried the G2-V3 type of V region and will not be further discussed.

FIGURE 3.

cDNA sequences from shark 33. Full-length cDNA sequences were selected and isolated from the shark 33 epigonal cDNA library, and those with identical C regions are aligned. The cDNA sequences are compared with a reference sequence that is a composite of the genomic V18 in V, D, J, Cμ1, and Cμ2 and of the composite cDNA sequences in Cμ3, Cμ4, and the 3′UT. Only the codons affected by mutations are shown, and the CDR1 and CDR2 are labeled. Those nucleotides that are part of an RGYW hotspot are bolded. The difference between G2-V1 and G2-V2 at codon 51 is indicated in every cDNA. The tandem mutations are underlined. The C domains are partitioned by slashes; the borders of Cμ1/Cμ2 and Cμ2/Cμ3 are known from the genomic sequence. The HindIII site in Cμ2 is in boldface, as is the polyadenylation signal. The membrane sequence is shown separately. Dots indicate identity with reference sequence, dashes gaps. The junctional sequences consist of N regions as well as P regions (italics), and the portions derived from the D gene are underlined. An asterisk indicates that the underlined sequence is the reverse complement of the D2 sequence (as shown). A tally of the number of mutations and of tandem mutation groups is given. The GenBank accession numbers are: AY594647 (Q10), AY594648 (T16), AY594649 (A5), AY594650 (G10), AY594651 (N40), AY594652 (H3), AY594653 (CH7), AY594654 (S10), AY583356 (E9), AY583357 (N13), AY571276 (J7), and AY571275 (J3).

FIGURE 3.

cDNA sequences from shark 33. Full-length cDNA sequences were selected and isolated from the shark 33 epigonal cDNA library, and those with identical C regions are aligned. The cDNA sequences are compared with a reference sequence that is a composite of the genomic V18 in V, D, J, Cμ1, and Cμ2 and of the composite cDNA sequences in Cμ3, Cμ4, and the 3′UT. Only the codons affected by mutations are shown, and the CDR1 and CDR2 are labeled. Those nucleotides that are part of an RGYW hotspot are bolded. The difference between G2-V1 and G2-V2 at codon 51 is indicated in every cDNA. The tandem mutations are underlined. The C domains are partitioned by slashes; the borders of Cμ1/Cμ2 and Cμ2/Cμ3 are known from the genomic sequence. The HindIII site in Cμ2 is in boldface, as is the polyadenylation signal. The membrane sequence is shown separately. Dots indicate identity with reference sequence, dashes gaps. The junctional sequences consist of N regions as well as P regions (italics), and the portions derived from the D gene are underlined. An asterisk indicates that the underlined sequence is the reverse complement of the D2 sequence (as shown). A tally of the number of mutations and of tandem mutation groups is given. The GenBank accession numbers are: AY594647 (Q10), AY594648 (T16), AY594649 (A5), AY594650 (G10), AY594651 (N40), AY594652 (H3), AY594653 (CH7), AY594654 (S10), AY583356 (E9), AY583357 (N13), AY571276 (J7), and AY571275 (J3).

Close modal

Three of the cDNA sequences were derived from germline transcripts (data not shown). H8 and M3 contain the G2-V1 VH gene segment and Q9 the G2-V2 VH gene segment. All three consist of leader spliced to unrearranged VH gene segment and germline JH gene segment spliced to C sequence. The D sequences were spliced out at the same cryptic sites 3′ of the VH-RS and 5′ of the JH-RS in all three. These clones carry the C region transmembrane sequence, as do two rearranged H chains, N13 and E9 (Fig. 3, bottom). The transmembrane sequence is spliced into the GGT codon just before the tailpiece (labeled, Fig. 3). The remaining cDNAs carry secretory tails.

The 16 rearranged cDNAs consist of 9 G2-V1 and 7 G2-V2 rearrangements, and the nucleotide distinguishing them (C/A) in CDR2 is shown. Thirteen of the cDNAs are mutants. Of the 131 changes observed in the V and J sequences, few are shared between the clones. There appeared to be mutations in CDR3 as well, but these were not counted because extensive N region addition and trimming had taken place, and a conservative tally of the mutations would avoid ambiguous assignments.

None of the rearranged sequences shared a common CDR3 sequence, meaning that they were all independent rearrangements. Likewise, few of the 131 substitutions were shared among the sequences. The changes consisted of 53 point mutations and 78 substitutions occurring mostly as doublets and triplets. Longer strings of mutations occur in heavily mutated sequences and could be the result of multiple events. Table IIshows the breakdown of the changes. As in the nurse shark L chain mutations, the H chain point mutations show a bias for transition changes (53%), whereas the tandem mutations do not (36%).

Table II.

Nature of H chain substitutions

All SubstitutionsAll Tandem SubstitutionsAll Single-Base Substitutions
From From  From  
to G — 17 11 to G — 19 to G — 16 
— 11 10 — 16 — 14 
10 — 16 — 21 — 11 
12 14 — — 22 — 12 
Total 27 35 32 37  18 23 21 16   12 11 21  
Total: 131 changes     Total: 78 changes      Total: 53 changes      
Transitions: 43%     Transitions: 36%      Transitions: 53%      
All SubstitutionsAll Tandem SubstitutionsAll Single-Base Substitutions
From From  From  
to G — 17 11 to G — 19 to G — 16 
— 11 10 — 16 — 14 
10 — 16 — 21 — 11 
12 14 — — 22 — 12 
Total 27 35 32 37  18 23 21 16   12 11 21  
Total: 131 changes     Total: 78 changes      Total: 53 changes      
Transitions: 43%     Transitions: 36%      Transitions: 53%      

The pattern of amino acid replacement over the V region is not punctuated by particularly higher amino acid replacement changes in the CDRs (R:S ratio in FR1 is 3.6, CDR1 is 5, FR2 is 1.5, CDR2 is 4, and FR3 is 2.3), but it is generally more similar to the patterns observed in mammalian mutants than those we had found in NS3 L chains.

Many of the mutations occurred in the RGYW hotspot or its complement, WRCY. There are 22 such motifs in the V and J gene segments, occupying 21.5% of a total of 405 bp, but 39% (51 of 131) of the mutations occurred in these hotspots. Of 33 tandem groups, 14 of them lie within or partly within a RGYW/WRCY motif. In 11 of 13 doublets, a first mutation would have abolished the hotspot for a second, if they had occurred as two sequential events. Of the 19 tandem groups occurring outside of the motifs, in only 4 of 19 would one change have generated a new hotspot motif to promote the second mutation. Thus, hotspots influence but are not responsible for recruiting tandem changes.

We believe that the substitutions are not the result of gene conversion events. For example, codon 31 in CDR1, which includes the AGCT motif, was changed from the amino acid serine S into N, T, R, F, and A in the various mutants. Since all five VH groups have either S or N at that position (33), most of the changes in the mutants cannot originate from a germline template of any known VH gene segment. Likewise, codon 54 in CDR2 is universally occupied by S in every VH group, so that the various changes into G, T, R, and K, all of which occurred through tandem substitutions, did not arise from any germline VH template. No mutant sequence matched with 16 differences of the pseudogene G2-V4 (data not shown).

Although there are four gene segments involved in rearrangement events, the RS configuration also allows the formation of V/D2/J. H3 and possibly T16 (Fig. 3) are such examples. Because there are 12-RS sequences on either side of D2, recombination can take place by inversion as well as by deletion. Thus, V/D1/D2inv/J or V/D2inv/J are also possible. Clone E9 contains the most convincing example of inverted D2, but G10 may carry one as well (Fig 3, asterisk).

As shown in Table III, all three reading frames are used for both D genes as well as two in inverted D2. However, even those combinations in the same reading frame, such as J3, Q10, and CH12, show that trimming and N region addition play a significant role in junctional diversification. There is little P region (italicized in Fig. 3) but extensive N region addition at all three junctions (average, 15.4 + 4.1 bp per sequence). CDR3 sizes range from 8 to 14 codons, with an average of 10.9 ± 2.3 codons. This mean length is close to the one obtained (11.6 codons) for 64 adult H chain CDR3 (33).

Table III.

H chain CDR3

CloneCDR3CDR3 Size Combination
V n D n D n / JCodonsD1   D2
J7          GGAACCTCCAGTGGATACCCATTCTGC/GATTAC 11 RF2 + RF1 
          G  T  S  S  G  Y  P  F  C   D  Y   
J3    ATCCGATATACCACAGTCGTGTGGGGATGGCCG/TTTGATTAC     14 RF3 + RF3 
    I  R  Y  T  T  V  V  W  G  W  P   F  D  Y       
Q10                   GGGGTGGTGGAATGG/TTTGATTAC     RF3 + RF3 
                   G  V  V  E  W   F  D  Y       
T16                       GCCGATTCGAT/GTACTATTTTGATTAC RF1 
                       A  D  S  M   Y  Y  F  D  Y   
CH12           ACATATACTACAGTGGGGGGACT/GTACTATTTTGATTAC 13 RF3 + RF3 
           T  Y  T  T  V  G  G  L   Y  Y  F  D  Y   
08     ACAGGGTATACTATAGTCCACTACGACCC/TTTTGATTAC   13 RF3 + RF1 
     T  G  Y  T  I  V  H  Y  D  P   F  D  Y     
A5         GACCATTACAGTGGCTCTGGGGGGG/ACTATTTTGATTAC   13 RF2 + RF2 
         D  H  Y  S  G  S  G  G  D   Y  F  D  Y     
G10             CATGCGGGTATATATGACGGA/TAC     RF3 + inverted RF2 
             H  A  G  I  Y  D  G   Y       
N40 GGGAAGTACACTGGCGTTACTGGGATCCTCGCC/TAC     12 RF2 + RF2 
 G  K  Y  T  G  V  T  G  I  L  A   Y       
H3                GAAGGCGAACTGGGCCAC/TACTATTTTGATTAC 11 RF3 
                E  G  E  L  G  H   Y  Y  F  D  Y   
E4                    CCGGGGTACGGCCC/TTTTGATTAC     RF2 + RF1? 
                    P  G  Y  G  P   F  D  Y       
CH7 GACGTCAGGTACTACACGGGGAATACTGGGGCC/TTTGATTAC 14 RF2 + RF2 
 D  V  R  Y  Y  T  G  N  T  G  A   F  D  Y   
S10               ATACAGTCCATACTGGGGG/TTGATTAC     RF1 + RF3 
               I  Q  S  I  L  G  V   D  Y       
E9     CAGTATATCTCTATCCCAACAAGAGG/TTTGGGTTAC   12 RF1 + inverted RF3 
     Q  Y  I  S  I  P  T  R  G   L  G  Y     
N13          GGGGGTACAGTGTCGGTCGGGGTA/TAC RF3 
          G  G  T  V  S  V  G  V   Y   
CloneCDR3CDR3 Size Combination
V n D n D n / JCodonsD1   D2
J7          GGAACCTCCAGTGGATACCCATTCTGC/GATTAC 11 RF2 + RF1 
          G  T  S  S  G  Y  P  F  C   D  Y   
J3    ATCCGATATACCACAGTCGTGTGGGGATGGCCG/TTTGATTAC     14 RF3 + RF3 
    I  R  Y  T  T  V  V  W  G  W  P   F  D  Y       
Q10                   GGGGTGGTGGAATGG/TTTGATTAC     RF3 + RF3 
                   G  V  V  E  W   F  D  Y       
T16                       GCCGATTCGAT/GTACTATTTTGATTAC RF1 
                       A  D  S  M   Y  Y  F  D  Y   
CH12           ACATATACTACAGTGGGGGGACT/GTACTATTTTGATTAC 13 RF3 + RF3 
           T  Y  T  T  V  G  G  L   Y  Y  F  D  Y   
08     ACAGGGTATACTATAGTCCACTACGACCC/TTTTGATTAC   13 RF3 + RF1 
     T  G  Y  T  I  V  H  Y  D  P   F  D  Y     
A5         GACCATTACAGTGGCTCTGGGGGGG/ACTATTTTGATTAC   13 RF2 + RF2 
         D  H  Y  S  G  S  G  G  D   Y  F  D  Y     
G10             CATGCGGGTATATATGACGGA/TAC     RF3 + inverted RF2 
             H  A  G  I  Y  D  G   Y       
N40 GGGAAGTACACTGGCGTTACTGGGATCCTCGCC/TAC     12 RF2 + RF2 
 G  K  Y  T  G  V  T  G  I  L  A   Y       
H3                GAAGGCGAACTGGGCCAC/TACTATTTTGATTAC 11 RF3 
                E  G  E  L  G  H   Y  Y  F  D  Y   
E4                    CCGGGGTACGGCCC/TTTTGATTAC     RF2 + RF1? 
                    P  G  Y  G  P   F  D  Y       
CH7 GACGTCAGGTACTACACGGGGAATACTGGGGCC/TTTGATTAC 14 RF2 + RF2 
 D  V  R  Y  Y  T  G  N  T  G  A   F  D  Y   
S10               ATACAGTCCATACTGGGGG/TTGATTAC     RF1 + RF3 
               I  Q  S  I  L  G  V   D  Y       
E9     CAGTATATCTCTATCCCAACAAGAGG/TTTGGGTTAC   12 RF1 + inverted RF3 
     Q  Y  I  S  I  P  T  R  G   L  G  Y     
N13          GGGGGTACAGTGTCGGTCGGGGTA/TAC RF3 
          G  G  T  V  S  V  G  V   Y   

Fig. 4 shows the junctional sequences from an RT-PCR performed on neonatal shark AQ spleen with the primer combination V18-1/CH2-3′. Thirteen G2-V2 sequences, no G2-V1, and four G2-V3 (data not shown) sequences were obtained. The average CDR3 size is 7.4, about three codons shorter than in adults. This appears to be due to correspondingly less N region addition per sequence (4.1 + 3.1 bp).

FIGURE 4.

Neonatal rearrangements of G2-V2. RT-PCR was performed on neonatal nurse shark AQ spleen RNA with primers (V18-1/CH2-3′) targeting group 2 sequences. Thirteen G2-V2 sequences were obtained and the CDR3 are shown. Portions belonging to V and J gene flanks are aligned beneath the reference sequences, and those derived from the D gene are underlined as well. The junctional sequences include P nucleotides (italicized) and N region. Sequence that could have belonged to either V or D are shown in lower case. Dashes indicate gaps. There is 1 substitution in V and J gene segments and 10 (different) ones in 5759 bp of C region sequence.

FIGURE 4.

Neonatal rearrangements of G2-V2. RT-PCR was performed on neonatal nurse shark AQ spleen RNA with primers (V18-1/CH2-3′) targeting group 2 sequences. Thirteen G2-V2 sequences were obtained and the CDR3 are shown. Portions belonging to V and J gene flanks are aligned beneath the reference sequences, and those derived from the D gene are underlined as well. The junctional sequences include P nucleotides (italicized) and N region. Sequence that could have belonged to either V or D are shown in lower case. Dashes indicate gaps. There is 1 substitution in V and J gene segments and 10 (different) ones in 5759 bp of C region sequence.

Close modal

Studies in the horn shark Ig showed that the H and L chains are encoded by 100–200 independent loci, each containing its own rearranging V gene segments and C region exons. For hypermutation studies, the plethora of Ig loci in sharks meant that the main difficulty of our endeavor was to distinguish one or a few IgM genes with unique markers in the C region. Once this was accomplished, the germline V gene segments could be characterized in order to identify somatically mutated VDJ rearrangements. In nurse sharks, we were able to distinguish four kinds of Cμ2 sequences through the presence or absence of three restriction enzyme sites, and we quickly determined that few IgM loci contained a Cμ2 with a HindIII site. We also estimate that there are ∼15 IgM loci in the nurse shark, whose genome is less than half the size of the horn shark.

In the course of sorting out IgM H chain cDNA sequences, we established that the classification of IgM groups 1–5, based on V region sequence differences of 20–26%, could be extended to their C region as well. Thus, a H chain cDNA with a HindIII site in Cμ2 would carry a V region with the group 2 VH sequence, and in addition to sequence homology (96–100% identity) these group 2 VH all possess a distinguishing one-codon insertion in FR1. Altogether, these markers define the three functional group 2 IgM genes in shark 33 (G2-V1, G2-V2, G2-V3) and distinguish them from all of the other H chain genes.

The correlation of distinct VH subsets with distinct CH sequence groups in the nurse shark was somewhat of a surprise since the horn shark cDNA showed no insertion/deletion in Cμ2 and few differences in the overall C region (32). The published horn shark sequences were probably not representative of its repertoire, just as, in our case, any random sample of nurse shark 33 cDNA clones would have been mostly group 4 sequences. The nurse shark Cμ1 regions are ∼85–92% identical between groups (data not shown), although within a group they are 97–100% identical. Cμ2 sequences are more divergent, as predicted, with 75–95% identity among groups (Fig. 2); the greatest stretch of differences lies in the 5′ end of Cμ2, with gaps of one to two codons.

Since there was a correlation of V region group (including J sequence) with C region markers in our cDNA, this suggests that little or no interlocus rearrangement occurs. Thus, somatic diversification of shark Ig is without the combinatorial diversity found in tetrapods. However, this may not be a serious deficiency, as discussed below.

Although there are only four gene segments available for rearrangement per IgH locus, the configuration of the RS sequences allows for varied combinations. The presence of the 12-RS at either flank of the D2 gene allows not only direct recombination to D2 without D1 but also D2 inversions, and both types were found among the cDNA CDR3 (Table III). In the horn shark, RS configurations with D1 flanked by 12-RS have been found; this arrangement probably also exists in the nurse shark and allows further variations in rearrangement. The gene segment organization and RS configuration evolved to generate the most efficient diversity and range of CDR3 with few rearranging elements.

N region addition occurs at all three joints, ranging from 2 to 12 bp. The N region GC nucleotide content is 68% (116 of 170), in agreement with what was obtained from rearrangements of L chains (23). TdT not only generates the N region at 90% of L chain junctions but is also present from birth, as evidenced by pup H chain sequences (Fig. 4). Another publication (33) suggested that the shorter pup CDR3 may be in part due to preferential retention of D1 sequences; we did not find this to be true for pup G2-V2 sequences, where less N addition entirely accounts for the shorter junctions.

The addition of N region at three junctions and the usage of all three reading frames in both D elements, along with the flexibility in obtaining either D alone or in combination or in inverted combinations, essentially allows each individual rearrangement event to form CDR3 with as great diversity as if a series of D elements was available. Moreover, there are very few germline-joined VD or VDD or VDDJ among the nurse shark IgM loci (K. Malecek and E. Hsu, unpublished data), unlike the horn shark (35) where 50% of its 100 IgM loci are partly or fully preassembled. In the nurse shark it is very clear that its Ab-combining site diversity is based on the extensive junctional diversity in both the H and L chains. We have speculated in greater detail elsewhere (18) that the selecting factor for recombining genes during vertebrate evolution was the RAG-mediated generation of sequence length variability (in CDR3), created in a tolerated site in the protein; what we have observed in nurse shark Ig supports the importance of this aspect of V(D)J recombination.

We have previously shown that somatic hypermutation at the loci of the nurse shark L chain-type NS3 consisted of both single and contiguous substitutions. The point mutations resembled those in other vertebrates in a bias for transition changes (51, 62, and 62% in three samples, respectively). The tandem mutations, consisting of doublets or triplets of adjacent changes that arose from a single event (24), did not share this characteristic (27, 42, and 45% transitions, respectively). We speculated that there may exist at least two pathways involving polymerases of different properties. The NS3 V genes are preassembled in the germline, and this pattern was observed for all of the loci monitored: NS3-12 and NS3-46 as well as NS3-8 (our unpublished results).

A second L chain isotype, NS5, has been characterized and there are three functional loci (23). One contains a preassembled VJ (NS5-16) and two contain VL and JL gene segments capable of rearrangement (NS5-2, NS5-48). All mutate in the same way as NS3. Among 249 NS5-2 mutations, 123 were point mutations with 55% transition bias, whereas 126 were tandem mutations without a transition bias, 37% (E. Hsu, unpublished results). Thus, at least six L chain loci of two different isotypes have similar mutation characteristics, and those with germline-joined V regions did not differ from those that rearranged somatically.

As found in L chain mutants, half (78 of 131) of the H chain mutations occurred in stretches of 2–4 bp, and these substitutions involved fewer transition changes (36%) than the point mutations (53%). Thus, these characteristics have been consistent in all Ig loci sampled. Somatic mutants from nurse shark IgH loci therefore do not carry the GC-biased mutations of putative horn shark H chain mutants, and it is not clear what the source of this difference might be. The horn shark mutants were not identified by a nonmutating C region as in our studies but rather by a CDR2 oligonucleotide probe that purportedly detected a single VH gene (26). It is possible that this method may have biased the screening by selecting for some subclass of sequences.

Among the 82 horn shark mutations, there were 7 tandem groups consisting of 16 changes, or 20% of the total mutations; and 8 (50%) of 16 changes are transitions. Among the 66 point mutations, 41 (62%) are transitions. However, 6 (40%) of 16 tandem changes were from G/C nucleotides, compared to 51 (77%) of 66 among the point mutations. If the horn shark sequences actually do represent genuine mutants, the patterns are difficult to reconcile with those observed in the nurse shark.

We conclude that, since the appearance of NAR occurred before divergence of sharks from rays and skates, and also before nurse shark from horn shark, the similarity of mutation patterns at NAR, IgH, and IgL loci in the nurse shark might mean that the balance of enzyme and cofactor activities in its lymphocytes is different from those of horn shark.

The pattern of replacement changes in the V region generally follows that found in mammals—a higher ratio of replacement to synonymous (R:S) amino acid substitutions in the CDR compared with the FR; however, the CDR/FR contrast is not great. As with L chain mutants (24), we cannot discern a strong selection for changes at sites that would affect ligand binding as traditionally expected in mammals. Although replacement mutations do occur more frequently per codon in the CDR, this could be due to higher tolerance for changes in the regions not involved in main-chain folding. The mutations and the frequency of shark Ig mutants increase with age (24, 33), and whether hypermutation acts as a contributor to the primary repertoire (36) or promotes affinity maturation in sharks is not clear; the mutants we observe might be the result of a combination of both.

The reason we have not definitively connected hypermutation with only affinity maturation in the nurse shark is that there appears to us a certain discrepancy between the relatively modest increase of binding affinity in Ag-specific monomeric IgM (37) and the extensive substitutions routinely observed in random H and L chain sequences (e.g.,, see Ref.24 and supplemental Figs. S2 and S3 at http://www.immunity.com/cgi/content/full/16/4/571/DC1)). A tandem change of 2–4 bp substitutions occurs as one single event (24), so that any such event almost necessarily generates an amino acid substitution. It seems odd that, with such an efficacious mutator it was historically so difficult to demonstrate effects on the shark Ab response (reviewed in Ref.38).

We have hypothesized that the tandem mutations in nurse shark Ig arise through an error-prone DNA repair enzyme(s) different from those generating point mutations (24). Probably the common initiator is an AID homolog, like the spotted dogfish sequence (19). The AID/apolipoprotein B-editing catalytic subunit family is conserved among vertebrates but absent from the protochordate Ciona intestinalis genome (39). However, it is part of the superfamily of cytidine deaminases that is present in eukaryotes, bacteria, and plants (19) and may be assumed to have evolved from some such ancestor. In contrast, RAG recombinase is present only in jawed vertebrates, and the demonstration of its latent transposase capabilities is good evidence supporting the hypothesis that RAG entered the ancestral vertebrate genome through horizontal transfer (Refs.40, 41, 42 and reviewed in Ref.18). Although sequences 25–30% identical to the RAG1 N-terminal domain and core have been discovered in the sea urchin (43), presumably RAG became established in vertebrates as a result of its chance introduction into a “context” that led to its selection as VDJ recombinase (44).

Until the primordial non-rearranging Ig superfamily gene is identified, if ever, in cyclostomes, which are the representatives of the more primitive jawless fishes, it will not be clear if hypermutation preceded rearrangement or vice versa. We have argued elsewhere (24, 45) that this ancestral V gene may originally have already been able to diversify by mutation, after examination of the very different natures of changes produced by the two mechanisms (18). But by the time cartilaginous fishes appeared, ∼100 million years later, both rearrangement and hypermutation were established, and in the nurse shark they occur at different kinds of Ig and Ig-like (NAR, IgW) loci.

The means by which the mammalian Ig genes become targeted for hypermutation is not known, other than that cis-acting elements such as the Ig enhancer are required. Some groups have proposed that it is not particular sequences but rather high transcription rates as such that target any gene for AID and hypermutation (46, 47); however, others have argued it was the overexpression of AID in those experimental systems that produced nonspecific effects (17). Those cis-elements that participate in the assembly of a mutasome are all present at eight nurse shark Ig H and L chain genes, and we speculate that they already existed in the ancestral Ig locus, prior to H and L chain divergence. If this speculation is true, the mechanism of somatic hypermutation thus preceded the existence of the Ab as we know it—the 2H/2L chain Ig unit that is today universal in vertebrates.

We thank Louis Du Pasquier for discussion and critical reading of our manuscript and Gillian Wu for help on the RS evaluation.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by grants from the National Science Foundation (MCB 0080098) and the National Institutes of Health (GM068095).

3

Abbreviations used in this paper: MAR, matrix attachment region; AID, activation-induced cytidine deaminase; NAR, nurse shark new Ag receptor; RS, recombination signal; FR, framework.

1
Tonegawa, S..
1983
. Somatic generation of antibody diversity.
Nature
302
:
575
-581.
2
Weigert, M. G., I. M. Cesari, S. J. Yonkovich, M. Cohn.
1970
. Variability in the lambda light chain sequences of mouse antibody.
Nature
228
:
1045
-1044.
3
Bernard, O., N. Hozumi, S. Tonegawa.
1978
. Sequences of mouse immunoglobulin light chain genes before and after somatic changes.
Cell
15
:
1133
-1144.
4
Gearhart, P. J., N. D. Johnson, R. Douglas, L. Hood.
1981
. IgG antibodies to phosphorylcholine exhibit more diversity than their IgM counterparts.
Nature
291
:
29
-34.
5
Selsing, E., U. Storb.
1981
. Somatic mutation of immunoglobulin light-chain variable-region genes.
Cell
25
:
47
-58.
6
McKean, D., K. Huppi, M. Bell, L. Staudt, W. Gerhard, M. Weigert.
1984
. Generation of antibody diversity in the immune response of BALB/c mice to influenza virus hemagglutinin.
Proc. Natl. Acad. Sci. USA
81
:
3180
-3184.
7
Griffiths, G. M., C. Berek, M. Kaartinen, C. Milstein.
1984
. Somatic mutation and the maturation of immune response to 2-phenyl oxazolone.
Nature
312
:
271
-275.
8
Reynaud, C. A., C. Garcia, W. R. Hein, J. C. Weill.
1995
. Hypermutation generating the sheep immunoglobulin repertoire is an antigen-independent process.
Cell
80
:
115
-125.
9
Lebecque, S. G., P. J. Gearhart.
1990
. Boundaries of somatic mutation in rearranged immunoglobulin genes: 5′ boundary is near the promoter, and 3′ boundary is ∼1 kb from V(D)J gene.
J. Exp. Med.
172
:
1717
-1727.
10
Yelamos, J., N. Klix, B. Goyenechea, F. Lozano, Y. L. Chui, A. Gonzalez Fernandez, R. Pannell, M. S. Neuberger, C. Milstein.
1995
. Targeting of non-Ig sequences in place of the V segment by somatic hypermutation.
Nature
376
:
225
-229.
11
Betz, A. G., C. Milstein, A. Gonzalez-Fernandez, R. Pannell, T. Larson, M. S. Neuberger.
1994
. Elements regulating somatic hypermutation of an immunoglobulin κ gene: critical role for the intron enhancer/matrix attachment region.
Cell
77
:
239
-248.
12
Goyenechea, B., N. Klix, J. Yelamos, G. T. Williams, A. Riddell, M. S. Neuberger, C. Milstein.
1997
. Cells strongly expressing Igκ transgenes show clonal recruitment of hypermutation: a role for both MAR and the enhancers.
EMBO J.
16
:
3987
-3994.
13
Kong, Q., L. Zhao, S. Subbaiah, N. Maizels.
1998
. A λ3′ enhancer drives active and untemplated somatic hypermutation of a λ1 transgene.
J. Immunol.
161
:
294
-301.
14
Ronai, D., M.D. Iglesias-Ussel, M. Fan, M.J. Shulman, M.D. Scharff.
2005
. Complex regulation of somatic hypermutation by cis-acting sequences in the endogenous IgH gene in hybridoma cells.
Proc. Natl. Acad. Sci. USA
102
:
11829
-11834.
15
Muramatsu, M., K. Kinoshita, S. Fagarasan, S. Yamada, Y. Shinkai, T. Honjo.
2000
. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme.
Cell
102
:
553
-563.
16
Chaudhuri, J., F. W. Alt.
2004
. Class-switch recombination: interplay of transcription, DNA deamination and DNA repair.
Nat. Rev. Immunol.
4
:
541
-552.
17
Reynaud, C. A., S. Aoufouchi, A. Faili, J. C. Weill.
2003
. What role for AID: mutator, or assembler of the immunoglobulin mutasome?.
Nat. Immunol.
4
:
631
-638.
18
Lewis, S. M., G. E. Wu, E. Hsu.
2004
. The origin of V(D)J diversification. T. Honjo, and F. W. Alt, and M. S. Neuberger, eds.
Molecular Biology of B cells
473
-489. Elsevier Academic Press, Amsterdam.
19
Conticello, S. G., C. J. Thomas, S. K. Petersen-Mahrt, M. S. Neuberger.
2005
. Evolution of the AID/APOBEC family of polynucleotide (deoxy)cytidine deaminases.
Mol. Biol. Evol.
22
:
367
-377.
20
Greenberg, A. S., D. Avila, M. Hughes, A. Hughes, E. C. McKinney, M. F. Flajnik.
1995
. A new antigen receptor gene family that undergoes rearrangement and extensive somatic diversification in sharks.
Nature
374
:
168
-173.
21
Rast, J. P., M. K. Anderson, S. J. Strong, C. Luer, R. T. Litman, G. W. Litman.
1997
. α, β, γ, and δ T cell antigen receptor genes arose early in vertebrate phylogeny.
Immunity
6
:
1
-11.
22
Hinds, K. R., G. W. Litman.
1986
. Major reorganization of immunoglobulin VH segmental elements during vertebrate evolution.
Nature
320
:
546
-549.
23
Fleurant, M., L. Changchien, C. T. Chen, M. F. Flajnik, E. Hsu.
2004
. Shark Ig light chain junctions are as diverse as in heavy chains.
J. Immunol.
173
:
5574
-5582.
24
Lee, S. S., D. Tranchina, Y. Ohta, M. F. Flajnik, E. Hsu.
2002
. Hypermutation in shark immunoglobulin light chain genes results in contiguous substitutions.
Immunity
16
:
571
-582.
25
Diaz, M., J. Velez, M. Singh, J. Cerny, M. F. Flajnik.
1999
. Mutational pattern of the nurse shark antigen receptor gene (NAR) is similar to that of mammalian Ig genes and to spontaneous mutations in evolution: the translesion synthesis model of somatic hypermutation.
Int. Immunol.
11
:
825
-833.
26
Hinds-Frey, K. R., H. Nishikata, R. T. Litman, G. W. Litman.
1993
. Somatic variation precedes extensive diversification of germline sequences and combinatorial joining in the evolution of immunoglobulin heavy chain diversity.
J. Exp. Med.
178
:
815
-824.
27
Wilson, M., E. Hsu, A. Marcuz, M. Courtet, L. Du Pasquier, C. Steinberg.
1992
. What limits affinity maturation of antibodies in Xenopus—the rate of somatic mutation or the ability to select mutants?.
EMBO J.
11
:
4337
-4347.
28
Denepoux, S., D. Razanajaona, D. Blanchard, G. Meffre, J. D. Capra, J. Banchereau, S. Lebecque.
1997
. Induction of somatic mutation in a human B cell line in vitro.
Immunity
6
:
35
-46.
29
Sale, J. E., M. Bemark, G. T. Williams, C. J. Jolly, M. R. Ehrenstein, C. Rada, C. Milstein, M. S. Neuberger.
2001
. In vivo and in vitro studies of immunoglobulin gene somatic hypermutation.
Philos. Trans. R. Soc. London B
356
:
21
-28.
30
Bachl, J., M. Wabl.
1996
. An immunoglobulin mutator that targets G.C base pairs.
Proc. Natl. Acad. Sci. USA
93
:
851
-855.
31
Hsu, E..
1994
. The variation in immunoglobulin constant regions in evolution.
Semin. Immunol.
6
:
383
-391.
32
Kokubu, F., K. Hinds, R. Litman, M. J. Shamblott, G. W. Litman.
1988
. Complete structure and organization of immunoglobulin heavy chain constant region genes in a phylogenetically primitive vertebrate.
EMBO J.
7
:
1979
-1988.
33
Rumfelt, L. L., R. L. Lohr, H. Dooley, M. F. Flajnik.
2004
. Diversity and repertoire of IgW and IgM VH families in the newborn nurse shark.
BMC Immunol.
5
:
8
34
Brodeur, P. H., R. Riblet.
1984
. The immunoglobulin heavy chain variable region (IgH-V) locus in the mouse. I. One hundred IgH-V genes comprise seven families of homologous genes.
Eur. J. Immunol.
14
:
922
-930.
35
Kokubu, F., R. Litman, M. J. Shamblott, K. Hinds, G. W. Litman.
1988
. Diverse organization of immunoglobulin VH gene loci in a primitive vertebrate.
EMBO J.
7
:
3413
-3422.
36
Weill, J. C., S. Weller, C. A. Reynaud.
2004
. A bird’s eye view on human B cells.
Semin. Immunol.
16
:
277
-281.
37
Dooley, H., M.F. Flajnik.
2005
. Shark immunity bites back: affinity maturation and memory response in the nurse shark, Ginglymostoma cirratum.
Eur. J. Immunol.
35
:
935
-945.
38
Hsu, E..
1998
. Mutation, selection and memory in B lymphocytes of exothermic vertebrates.
Immunol. Rev.
162
:
25
-36.
39
Azumi, K., R. De Santis, A. De Tomaso, I. Rigoutsos, F. Yoshizaki, M. R. Pinto, R. Marino, K. Shida, M. Ikeda, M. Arai, et al
2003
. Genomic analysis of immunity in a Urochordate and the emergence of the vertebrate immune system: “waiting for Godot”.
Immunogenetics
55
:
570
-581.
40
Agrawal, A., Q. M. Eastman, D. G. Schatz.
1998
. Transposition mediated by RAG1 and RAG2 and its implications for the evolution of the immune system.
Nature
394
:
744
-751.
41
Hiom, K., M. Melek, M. Gellert.
1998
. DNA transposition by the RAG1 and RAG2 proteins: a possible source of oncogenic translocations.
Cell
94
:
463
-470.
42
Clatworthy, A. E., M. A. Valencia, J. E. Haber, M. A. Oettinger.
2003
. V(D)J recombination and RAG-mediated transposition in yeast.
Mol. Cell
12
:
489
-499.
43
Kapitonov, V. V., J. Jurka.
2005
. RAG1 core and V(D)J recombina-tion signal sequences were derived from Transib transposons.
PLoS Biol.
3
:
e181
44
Du Pasquier, L., I. Zucchetti, R. De Santis.
2004
. Immunoglobulin superfamily receptors in protochordates: before RAG time.
Immunol. Rev.
198
:
233
-248.
45
Du Pasquier, L., M. Wilson, A.S. Greenberg, M. F. Flajnik.
1998
. Somatic mutation in ectothermic vertebrates: musings on selection and origins.
Curr. Top. Microbiol. Immunol.
229
:
199
-216.
46
Yoshikawa, K., I. M. Okazaki, T. Eto, K. Kinoshita, M. Muramatsu, H. Nagaoka, T. Honjo.
2002
. AID enzyme-induced hypermutation in an actively transcribed gene in fibroblasts.
Science
296
:
2033
-2036.
47
Martin, A., M. D. Scharff.
2002
. Somatic hypermutation of the AID transgene in B and non-B cells.
Proc. Natl. Acad. Sci. USA
99
:
12304
-12308.
48
Kabat, E. A., T. T. Wu, H. M. Perry, K. S. Gottesman, C. Foeller.
1991
.
Sequences of Proteins of Immunological Interest
xiii
U.S. Department of Health and Human Services, Washington, DC.