## Abstract

Cartilaginous fish express canonical B and T cell recognition genes, but their lymphoid organs and lymphocyte development have been poorly defined. Here, the expression of Ig, TCR, recombination-activating gene (Rag)-1 and terminal deoxynucleosidase (TdT) genes has been used to identify roles of various lymphoid tissues throughout development in the cartilaginous fish, Raja eglanteria (clearnose skate). In embryogenesis, Ig and TCR genes are sharply up-regulated at 8 weeks of development. At this stage TCR and TdT expression is limited to the thymus; later, TCR gene expression appears in peripheral sites in hatchlings and adults, suggesting that the thymus is a source of T cells as in mammals. B cell gene expression indicates more complex roles for the spleen and two special organs of cartilaginous fish—the Leydig and epigonal (gonad-associated) organs. In the adult, the Leydig organ is the site of the highest IgM and IgX expression. However, the spleen is the first site of IgM expression, while IgX is expressed first in gonad, liver, Leydig and even thymus. Distinctive spatiotemporal patterns of Ig light chain gene expression also are seen. A subset of Ig genes is pre-rearranged in the germline of the cartilaginous fish, making expression possible without rearrangement. To assess whether this allows differential developmental regulation, IgM and IgX heavy chain cDNA sequences from specific tissues and developmental stages have been compared with known germline-joined genomic sequences. Both non-productively rearranged genes and germline-joined genes are transcribed in the embryo and hatchling, but not in the adult.

## Introduction

The evolutionary origins of adaptive immunity are of fundamental significance to our understanding of the basis for self versus non-self recognition and host defense. One approach that has been employed in investigating this broad issue involves phylogenetic comparisons of genes involved in lymphocyte function among various vertebrate groups. Studies to date in both mammalian and non-mammalian model systems have revealed major differences in the manner in which the immune repertoire is derived. Cartilaginous fishes, which include sharks, skates and ratfishes, provide a particularly striking dichotomy with regard to the function, genetic structure and regulation of immune receptors (1,2). Whereas the organization of TCR, and MHC class I and II genes found in various species of cartilaginous fish bear a striking resemblance to higher vertebrate counterparts (35), their Ig genes are arrayed in hundreds of independent loci that do not undergo combinatorial joining (1,6). Furthermore, depending on the particular species of cartilaginous fish, varying percentages of the individual Ig gene loci are either partially or fully joined in the germline (712). Sharks are capable of mounting a hapten-specific immune response but it is associated with a lack of affinity maturation and minimal fine specificity, even over extended courses of immunization (1315).

Despite extensive characterizations of many of these genes, relatively little is known with regard to the cell lineage-specific expression of Ig and TCR, their respective patterns of development, and the development of the immune repertoire. These issues are of particular significance in cartilaginous fish as these species lack an obvious equivalent of bone marrow and possess additional, unique lymphoid tissues, including the epigonal and Leydig organs, which have been described histologically as potential equivalents of bone marrow (16).

Among cartilaginous fish the oviparous clearnose skate, Raja eglanteria, is the best-characterized model that is presently available for developmental studies (17). Embryos can be obtained without sacrifice to the breeding stock and, despite a season-limited reproductive cycle, it is possible to obtain developmental-staged specimens. The studies described herein define the developmental stage- and tissue-specific expression patterns of a number of different genes that function in primary antigen recognition as well as the somatic rearrangement process, and relate these to the ontogenetic development of the immune (B cell) repertoire.

## Methods

### Tissue and blood collection

Embryonic, hatchling and adult skates were euthanized using tricaine methanesulfonate (Argent, Redmond, WA and Crescent Research Chemicals, Phoenix, AZ), and liver, spleen, thymus, spiral intestine, epigonal (gonad in hatchling and embryo), nidamental (shell) gland (adult females only), kidney (adult only), muscle (hatchling and adult only), rectal gland and Leydig organ were snap-frozen in liquid nitrogen for future use, as were whole embryos after removal of the yolk sac. Peripheral blood leukocytes (PBL) were recovered by low-speed centrifugation from whole blood, obtained from the caudal vein.

### DNA isolation, RNA isolation and Northern blot analysis

DNA was prepared by embedding purified erythrocytes in agarose blocks and extraction (in situ) with lithium dodecylsulfate (LDS) (18). DNA prepared in this manner consists only of the high-mol.-wt form; low-mol.-wt DNA and RNA are eliminated by passive diffusion during the extraction and block processing steps. RNA extraction using RNAzol B (Tel-Test, Friendswood, TX), mRNA selection and Northern blotting were carried out using standard methods. S26, a ribosomal mRNA, was used to normalize Northern blots as it demonstrates the most consistent relative expression of mRNA among multiple tissues (19). UV-cross-linked blot transfer membranes were prehybridized in Ultrahyb hybridization buffer (Ambion, Austin, TX) at 42°C for 30 min and hybridized with various probes that were labeled to uniform sp. act. employing random hexamers. Blots were washed under conditions of moderate stringency and exposed either to autoradiographic film or to a phosphor screen (20).

### Ribonuclease protection assay

The constant regions from TCR α, β, γ and δ were subcloned into pBluescript (Stratagene, La Jolla, CA) (21). Probes and standards were constructed from templates using the MaxiScript kit (Ambion) following the manufacturer's protocol. RNA (10 μg) from thymus, spleen, rectal gland, epigonal, Leydig organ, spiral intestine and liver was used for each ribonuclease protection assay (RPA II kit; Ambion). Autoradiographic signals were converted into densitometric values. The highest signal for a specific mRNA type was taken as 100% relative abundance and related by fractional percentage to the determinations made for all other tissues.

### Real-time PCR

Real-time PCR analysis was carried out using a GeneAmp 5700 sequence detection system (PE Biosystems, Foster City, CA). One microgram of total RNA from each tissue was reverse transcribed into cDNA using 5.5 mM MgCl2, 500 μM of each dNTP, 2.5 μM random hexamers, 0.4 U/μl RNase inhibitor and 1.25 U/μl MultiScribe reverse transcriptase (PE Biosystems). Then 1 μl of each 100 μl reaction was subsequently used in each PCR along with SYBR Green PCR Master Mix (PE Biosystems) and 300 nM of each primer, constructed to optimize amplification for detection by SYBR Green. Controls with no template or cDNA controls with no reverse transcriptase were used to test each primer pair. Duplicate determinations made in the presence of DNase I gave identical results. Relative expression for each different transcript was determined by comparison to plasmid standards and cDNA values were normalized to expression of 28S rRNA (22), which was found to be abundant in all tissues at all developmental stages examined. Although the sensitivity of real-time PCR can extend to detect single-copy genes using SYBR Green, the application in these studies is the determination of relative abundance (as indicated), without assignment of specific quantifiable parameters. In multiple instances where repetitive determinations have been made (including temporal variation), the variance in results is statistically insignificant.

### Genomic and cDNA libraries

A genomic library was constructed from R. eglanteria purified red blood cell DNA, which was embedded in agarose and extracted with LDS (see above). Embedded DNA was digested partially with MboI (18) and ligated into a λDASH–BamHI vector (Stratagene). The library is equivalent to a single genome and was amplified. Spleen, epigonal and Leydig organ cDNA libraries were constructed as described (23), and amplified from the RNA isolated from the same animal that was used as the DNA source for the genomic library.

### IgM and IgX cDNA clones

Several different strategies were employed in order to selectively examine specific regions of Ig. In order to amplify cDNAs that would include the entire third complementarity-determining region (CDR3), primers were designed on the basis of known IgM and IgX sequences that spanned the variable (V) region of framework 2 (FR2) (5′-TTGGTCCGTCAGGTCCCCGGGCAG-3′) to the first constant region (Cμ1) (5′-TTGATCCTCGCAGGTGAAGAGAAT-3′) for IgM. The corresponding region of IgX was amplified from the V region of FR2 (5′-GGGTGAAACAGGTCCCCGGGAAAG-3′) to Cx1 (5′-GAAGAGGTGATGTGGACTGAAGGC-3′). Approximately 8×105 p.f.u. from each cDNA library was plated onto nitrocellulose and replica lifts were screened with probes specific for the first constant regions of IgM and IgX. Positive clones were used as templates and amplified using high-fidelity PfuI polymerase, and the respective FR2- and Cμ1- or Cx1-specific primer sets. Spleen, gonad (which presumably contains undifferentiated epigonal tissue), liver (embryos only) and Leydig organ tissues from hatchling and 8-week embryos were isolated and cDNAs generated from DNase I-treated total RNA used in individual amplification reactions. IgM or IgX V region FR2 and the corresponding Cμ1- or Cx1-specific primers were used to amplify cDNA sequences containing CDR3. All PCR reactions were T/A subcloned. The combined PCR amplification/DNA sequencing error rate is estimated to be ~1/5000 (24).

### Joined genomic IgM and IgX genes

Owing to the limited number of available heavy chain germline sequences in R. eglanteria, the FR2 primers that were used for the cDNA priming were paired with primers complementing the joining (J) region of IgM and IgX that were designed on the basis of available published and unpublished sequences. Genomic DNA that was embedded in agarose and extracted in LDS from the same animal that was used for cDNA library construction also was used for deriving genomic DNA for amplification of germline-joined genes. Cycling parameters (94°C for 1 min, 58°C for 45 s and 72°C for 30 s for 30 cycles, with PfuI polymerase) were optimized for short (250 bp) products, corresponding to fully joined genes. PCR products were size-selected (essentially only short products form), subcloned and sequenced as described for cDNA clones.

### DNA sequencing and sequence analysis

Automated sequencing using the LI-COR system and ThermoSequenase (Amersham Pharmacia, Arlington Heights, IL) chemistry and analysis of sequence data were performed as described (20).

## Results

### Expression of antigen binding receptor genes is up-regulated at 8 weeks of embryonic development

Owing to limited amounts of available tissue, several different approaches were utilized for examining gene transcription. Initially, expression of both TCR and Ig was examined over the course of the 11-week period of embryonic development. The greatest relative abundance of TCR α, δ and γ mRNA expression is at 8 weeks; in contrast, TCR β expression peaks at 7 weeks, and then plateaus over weeks 8 and 9 (Fig. 2A). The maximum relative expression of IgM and IgX as well as both type I (9,26) and II (26) light chain (LCI and LCII) gene transcripts also occurs at the 8-week point in embryonic development (Fig. 2B). LCI clusters, which are joined in the germline in Raja (9), exhibit a low level of expression throughout the course of sampling; however, again, their relative abundance peaks at 8 weeks. The significance of the basal level (background) of LCI expression is unclear and could relate to a lower threshold used to interpret amplification. Although unlikely, it is possible that the higher relative level of expression at 8 weeks could relate to a general mRNA quantitative effect at that stage in development. Expression of the myelin gene, a non-lymphocyte-specific gene, peaks at 11 weeks and several transcription factors, including Raja Ikaros (20), are expressed at relatively higher levels several weeks earlier in development (A. Miracle, unpublished observations; data not shown).

### The patterns of tissue-specific expression of TCR change during development

Taken together, these results indicate that the expression of TCR genes is restricted to the thymus in the embryo and is regulated in a highly tissue-dependent manner in the hatchlings. In the hatchling, only minimal levels of TCR expression occur in the thymus, whereas TCR expression in the adult is predominantly in the thymus and spleen. Despite the changes in TCR expression in various tissues, TCR are not expressed at significant levels in the Leydig and are expressed at barely detectable levels in the epigonal throughout development.

### Differential expression of Ig heavy chain genes in multiple tissues

Two forms of IgX have been described, which share a V region and the first two conserved constant region domains (Cx1 and Cx2) (27,28). Quantitation of the short (VxCx1Cx2) and long (VxCx1Cx2Cx3Cx4) forms of IgX was approximated by relating combined (long and short forms) and long-form-specific (Cx3Cx4) real-time PCR analyses (Fig. 5B). These results indicate that the short form predominates in the 8-week embryo.

Although significant levels of IgM expression are evident in adult spleen, epigonal and spiral intestine, the normalized expression of IgM in the adult is 3-fold higher in Leydig organ than in the spleen (Fig. 7A). The apparent slight increase seen in spleen possibly is significant based on the level of amplification above threshold. Notably, the ratio of secretory to transmembrane forms resembles that described previously in Northern blot analysis of adult spleen RNA in another cartilaginous fish (29). Peripheral blood, which exhibits considerable levels of Ig expression, could account for the significant expression of IgM in the highly vascularized kidney (Fig. 7A) (30). The two bands that appear in the PBL sample for IgM possibly reflect differential RNA processing. As in the hatchling, the relative expression of the two forms of IgX is approximately equivalent, confirming previous descriptions in skate (28). Other cartilaginous fish possess long orthologous forms of IgX, NARC (31); a short form of NARC equivalent to the short form of IgX has been identified (M. Flajnik, pers. commun.). The expression of IgX in spleen is considerably lower relative to IgM expression, which could be functionally significant. Only minimal levels of expression of the transmembrane form of IgM are observed in the adult (Fig. 7B).

### B cell gene expression in the embryo and hatchling thymus

Although the thymus is the first site of TCR gene expression, it is not exclusively a T cell organ in the embryo and hatchling. In the embryo, the thymus is a minor site of IgM expression and a major site of expression of IgX (short form), accompanied by low-level expression of LCI and LCII. In the hatchling, IgX expression decreases in the thymus but IgM expression and LCI expression continue. B cell gene expression is lost from this organ only in the adult.

### Non-productive rearrangements and expression of germline-joined gene clusters occurs in embryos and hatchlings

Relating the sequences of expressed genes to specific genomic clusters is a formidable undertaking, as described previously (9,34). However, relating germline-joined heavy or light chain genes to specific transcripts is facilitated by the uniqueness of CDR3 sequences. Although it is not possible at this time to define the entire heavy chain gene family in skate, which would require the analysis of several hundred additional clusters, it is possible to compare the FR2 → J sequences of germline-joined genes to the corresponding regions of cDNAs derived from different embryonic, hatchling and adult tissues.

Of the 28 distinctive IgM cDNA sequences derived from embryos, six are out-of-frame. Notably, the sequence across the FR2 → J region of one of these out-of-frame transcripts is identical to a corresponding sequence of an out-of-frame genomic clone, indicating that this gene can be expressed. Similarly, two of 22 in-frame cDNA sequences, both of which derive from embryonic liver, can be matched to the nucleotide sequences of two different in-frame germline-joined genomic clones. Twenty-five IgM cDNA sequences from hatchlings also were compared with the genomic sequence database. Two of the in-frame sequences from hatchling gonadal tissue are identical at the nucleotide level to two different in-frame germline-joined genomic clones. However, none of the 99 adult IgM cDNA sequences, all of which are in-frame, match the sequences in the germline-joined gene database. None of the adult IgM cDNAs that were examined match the sequences that were derived in parallel analyses of the embryonic or hatchling IgM cDNAs.

Of the IgX embryonic cDNA sequences, nine out of 10 (on the basis of the FR2 → J sequence) represent productive (in-frame) transcripts; of 25 hatchling sequences, 21 were in-frame. All of the 158 total IgM and IgX cDNA sequences in the adult potentially are productive. Finally, no matches across this region were found between embryo, hatchling and adult cDNAs or between these partial cDNAs and the database of germline-joined genes.

### CDR3 diversity is equivalent at different developmental stages

cDNA sequences were examined for CDR3 diversity. No significant differences in overall CDR3 diversity and length were observed between the cDNAs recovered from different tissues at the three different developmental stages that were examined (Fig. 10). Specifically, the lengths of CDR3 for IgM range from 3 to 12 residues in the embryos, 3 to 13 residues in the hatchlings and 3 to 15 residues in the adults. Similarly, the lengths for CDR3 for IgX range from 3 to 10 residues in the embryos, 3 to 11 residues in the hatchlings and 2 to 14 residues in the adults. Although several cDNAs were identified in hatchling and adult that were longer than those recovered in embryos, this difference is minimal and largely reflects individual outliers as well as the considerably larger sample size in adults.

## Discussion

Despite considerable effort, it has not been possible to identify homologous forms of Ig or TCR in vertebrate species that diverged prior to the emergence of the jawed vertebrates (1). From a phylogenetic perspective, the cartilaginous fish are the earliest extant forms in which an adaptive immune system, as defined by rearranging immune receptor genes, has been identified. Investigations into the regulation of the uniquely organized Ig loci of cartilaginous fish are of fundamental interest and significance in terms of gaining an overall view of the evolution of lymphoid function (1,2). The work described here lays a foundation for the cell biology of immune system development in a cartilaginous fish.

### TCR and Ig expression in ontogeny

In order to approach the analysis of gene expression in these species at periods in development in which only limited amounts of certain tissues are available, it was necessary to utilize several different complementary approaches, including real-time PCR, ribonuclease protection assays and Northern blot analysis. Taken together, these approaches have revealed both conserved and divergent patterns of expression of genes that are critical to lymphoid development and function.

During embryonic development, all four classes of TCR are expressed primarily in the thymus starting at a point two-thirds of the way through embryonic development, which is similar to the timing of the onset of TCR expression during the embryonic development in mouse (35). High levels of expression of both Rag-1 and TdT, which function in segmental rearrangement, are evident in the embryonic thymus, consistent with rearrangement and junctional diversification of TCR. In hatchlings, TCR expression is detected in the spleen and intestine, implicating these tissues as potential secondary lymphoid sites. The lack of abundant expression of TCR α, β, γ and δ in the thymus at the hatchling stage may relate to a lull' in the waves of thymic precursor gene expression, such as has been documented in an avian model (36). In the adult, the coordinate expression of TCR α and β in the thymus and spleen correspond to the patterns evident in mammals where these two tissues are the sites of primary αβ T cell development and αβ T cell function respectively.

TCR δ expression cannot be detected in the adult unless a higher sensitivity ribonuclease protection analysis is employed; studies using this method indicate that a reciprocal relationship exists between TCR α and δ in thymus and spleen (Fig. 4B). This observation also is interesting from the standpoint of cis regulation of gene expression (37); however, the genomic relationship of TCR α and δ has not yet been defined in skate, which limits further interpretation of this finding. It remains to be seen whether the unique expression pattern of TCR δ corresponds to a lymphoid or non-lymphoid function. Although it is tempting to relate the expression of TCR genes in skate to that of the orthologous forms of TCR found in mammals, it is important to recognize that the dimerization patterns of the former are not known. Furthermore, it remains possible that the specific function of these molecules may differ from those of the orthologous forms seen in higher vertebrates.

A far more complex pattern emerges for the tissue-specific expression of Ig genes during ontogeny. The highest relative abundance of both IgM and IgX also occurs at 8 weeks of embryonic development and then falls off dramatically. Significant lymphoid gene expression is seen in the embryonic but not adult liver of skate, paralleling the tissue-specific expression pattern that occurs in mammals (38,39). At 8 weeks of embryonic development, the highest relative abundance of IgM and IgX is seen in the spleen; IgX is expressed in greater abundance in more tissues relative to IgM at this stage. It presently is difficult to speculate as to the significance of this observation in that the function of IgX is not understood and the expression of both classes of light chain genes, which are germline joined, is uncoupled from the segmental rearrangement process. However, the distinctly regulated patterns of IgM and IgX expression as well as expression of light chain genes suggests that the variation observed is functionally significant, and that separate lineages of cells may express IgM and IgX at these stages.

The abrupt coincidental expression of Ig and Rag-1 genes, in the spleen, liver, Leydig organ and gonad in the 8-week embryo suggests that B cell development occurs at multiple sites in the developing skate embryo in contrast to the apparent restriction of T cell development to the thymus. Moreover, in the embryonic and hatchling skate there is substantial Ig gene expression in the thymus, raising the possibility that the thymus also could be a site of B cell development in these early stages. In other vertebrates, variation in the sites of B cell development, such as the avian bursa (40), are in marked contrast to T cell development.

Although both Rag-1 and TdT are involved in the generation of antigen recognition site diversity, the tissue-specific expression patterns of these genes differ. The overall pattern for Rag-1 expression is relatively uniform in a number of different lymphoid tissues in the embryo, and is followed by a disproportionate increase in abundance in the thymus of the hatchling and then exhibits a reduced level of expression in a number of different tissues in the adult. However, the apparent uniformity of Rag-1 expression may or may not reflect protein levels. The less restricted nature of Rag expression is potentially significant against the background of germline-joined gene clusters. Furthermore, differences in Rag expression seen in cartilaginous fish may relate to the 2–3 orders of magnitude reduction in intronic length separating recombining elements in cartilaginous fish and the lack of combinatorial diversity during genetic rearrangement in these species. TdT expression is elevated in thymus in the embryo, but is hardly detectable in only three lymphoid tissues at the hatchling stage and in the intestine of the adult. In these studies, there is no apparent correlation between TdT expression and the diversification of the repertoire, although it plays a highly significant role in repertoire development in mammals (41). However, distinct differences are apparent in terms of Ig repertoire diversity in skate versus mammalian ontogeny (see below).

### Roles of the Leydig and epigonal organs

The highest levels of expression of IgM and IgX heavy chains as well as LCI and LCII are found in the Leydig organ, which is unique to certain species of cartilaginous fish. Even employing a highly sensitive ribonuclease protection assay, it was not possible to detect significant levels of TCR mRNAs in either the Leydig or epigonal organs, which both express high relative levels of Ig heavy and light chain mRNAs. The absence of TCR expression distinguishes the Leydig organ from other lymphoid tissues, including avian bursal tissue in which a similar inability to detect TCR mRNAs using Northern blotting has been noted; however, TCR expression in the bursa can be detected using RT-PCR. T cell seeding to bursa or contamination with PBL represents possible sources for the signals (C.-L. Chen, pers. commun.).

In terms of understanding the role of the Leydig organ in B cell development, it is notable that Rag-1 is expressed most abundantly in the adult Leydig organ but TdT expression is reduced significantly relative to the levels that are expressed in the adult intestine, which are roughly equivalent to the levels of TdT that are expressed in the thymus. The relative diversity of the CDR3 region of IgM and IgX cDNAs from Leydig organ is not significantly different from that observed in cDNAs recovered from other tissues. Inspection of the Leydig sequences reveals no unusual patterns of predicted residues (data not shown). Although the most distinctive feature of the Leydig organ is largely of a quantitative nature, it is likely that the organ plays a significant role in the adaptive immune response.

Recently we identified a new member of the PU-1 family of transcription regulatory factors and have designated it as SpiD (42). Analyses of expression patterns using real-time PCR have shown that SpiD, as well as PU.1, are expressed abundantly in the Leydig organ and epigonal but not in spleen; whereas SpiC is not expressed in Leydig organ and epigonal but is expressed abundantly in spleen. In that both T and B cells are abundant in the spleen and B cells are highly abundant in Leydig organ and epigonal, it is probable that B cells in Leydig organ and epigonal are either at a different stage of development or represent different lineages than those found in spleen, analogous to B1 versus B2 lineage (43) and other functional distinctions in repertoires of mature B cells (44).

Interestingly the epigonal, another lymphoid organ found in those species, is notably larger in cartilaginous fish that lack a Leydig organ (45) and to some degree the tissue-specific expression patterns defined here for the Leydig organ resemble those seen in the epigonal organ. Furthermore, both tissues express equivalent amounts of both light chain types in the adult, but heavy chain gene expression is much higher in Leydig organ. Although it is beyond the present scope of these studies, the results are consistent with the possibility that the epigonal and Leydig organs serve either redundant or complementary roles in lymphoid development.

### Differential regulation of Ig gene cluster expression in development

The relative levels of IgM and IgX do not vary in relation to each other throughout development; however, significant variation in the expression of the two different families of germline-joined light chain genes is apparent. Specifically, the relative abundance of LCI expression is consistent during ontogeny, in marked contrast to the relatively low level of expression of LCII in the embryo and hatchling. However, LCII in the adult skate is the predominantly expressed form of light chain that is expressed in the Leydig organ. Efforts presently are underway to determine whether or not there is any specificity in utilization of the various subfamilies of joined light chain genes. Although interpretation of these data is confounded by the paucity of information regarding the association patterns of Ig heavy and light chains in cartilaginous fish, the observations described here provide critical information that will facilitate a better understanding of such interactions and the role of allelic exclusion in the regulation of Ig gene expression in cartilaginous fish.

In order to provide an initial estimate of the diversity of the immune repertoire and in particular to examine the expression of germline-joined genes, a large number of Ig cDNAs from spleen, Leydig organ and epigonal derived from a single adult animal were sequenced and their CDR3s were compared. The basis for these types of studies lies in the studies of CDR3 diversity that have been conducted in a number of different higher vertebrate species, and have established an age-dependent increase in CDR3 length and diversity (4649). Within the sample size represented, no structurally significant differences are apparent in terms of overall CDR3 length and diversity. Several in-frame cDNA sequences (FR2 → Cμ1), which were identified in both the embryonic liver and hatchling epigonal organ, are indistinguishable from the corresponding regions of in-frame germline-joined genomic sequences. None of the significantly greater number of FR2 → Cμ1 and FR2 → Cx1 (n = 158) cDNA sequences recovered from an adult appear to have derived from germline-joined clusters based on comparison of CDR3 sequences using the database of genomic germline-joined CDR3 sequences that was generated specifically for this study. The findings indicate that transcription of germline-joined IgM and IgX clusters may constitute ~10–15% of the Ig expression seen in embryo and hatchling; these genes are not expressed in the adult skate. These latter observations are consistent with past failures to detect transcripts of germline-joined heavy chain genes that have been carried out using tissues derived from adults (10). However, it also is possible that later in development transcripts of somatically rearranged genes are too abundant to permit detection of the rare potential transcripts that derive from germline-joined genes.

### Developmental regulation selection for in-frame rearrangements

The increase in Ig and TCR gene expression at the 8-week developmental stage correlates with the apparent surge in the expression of Rag-1 and TdT. Based on the coincidental increases in transcriptional activity and high proportion of non-productive transcripts at 8 weeks, mass transcription of antigen receptor gene clusters may be taking place. The nature of Ig organization in the skate (and other cartilaginous fish) is possibly prone to non-productive transcription as distances between promoters and the coding segments of Ig loci are orders of magnitude closer than are found in mammals (1). Widespread run-off of Ig clusters in early development may establish the B cell antigen binding receptor repertoire, which is regulated further as cells with productively rearranged Ig mature. Establishment of a repertoire early in development by such a shotgun' approach to transcription would negate a requirement for precise regulation of selective transcription of clusters in the mature adult, which are present on different chromosomes (27). Examples of the use of strategies that differ from those used in man and mouse to establish an Ig repertoire have been defined in other classes of vertebrates, including the bursa of Fabricius in avians (40,50) and ovine Peyer's patches (51,52), as well as in the lymphoid tissues of rabbit (53) and other mammals (54,55).

General similarities in the ontogeny of T cell development are evident between observations made in this study and those established in other model systems. On the other hand, the unique clustered genomic organization and expression of Ig differ markedly in terms of both the expression patterns and involvement of unique lymphoid tissues. An explanation, at least in part, for these observations is that T cells and TCR (α and β) may be constrained in an evolutionary sense owing to their obligatory interactions with MHC for antigen recognition, necessitating a parallel, cooperative evolution of the two (TCR and MHC) systems. In contrast, B cells may lack direct dependence on a separate multigene family, thus explaining the marked differences in organization, isotype and diversification mechanisms. Such differences may accompany the variation seen in the sites of primary lymphopoiesis that have been observed throughout the different classes jawed vertebrates (1).

Table 1.

Sequences of 3′ portions of variable regions obtained from (top) genomic DNA, and (bottom) cDNAs from various tissues from 8 week embryos, hatchlings and an adult clearnose skate

Source N In-frame Out-of-frame
Germline-joined genomic IgM 53 29 24
Germline-joined genomic IgX 21 16
Source N In-frame Out-of-frame
Germline-joined genomic IgM 53 29 24
Germline-joined genomic IgX 21 16
Source N In-frame Non-productive
The total number of different sequences obtained from each source is given as N, with the proportions of those sequences designated as either in-frame or out-of-frame or non-productive (out-of-frame or stop codons).
Embryonic IgM 28 22
spleen
Leydig
liver
Embryonic IgX 10
spleen
Leydig
liver
Hatchling IgM 25 21
spleen
Leydig 10
Hatchling IgX 25 21
spleen 10
Leydig
spleen 36 36
Leydig 33 33
epigonal 30 30
spleen 17 17
Leydig 22 22
epigonal 20 20
Source N In-frame Non-productive
The total number of different sequences obtained from each source is given as N, with the proportions of those sequences designated as either in-frame or out-of-frame or non-productive (out-of-frame or stop codons).
Embryonic IgM 28 22
spleen
Leydig
liver
Embryonic IgX 10
spleen
Leydig
liver
Hatchling IgM 25 21
spleen
Leydig 10
Hatchling IgX 25 21
spleen 10
Leydig
spleen 36 36
Leydig 33 33
epigonal 30 30
spleen 17 17
Leydig 22 22
epigonal 20 20
Fig. 1.

Ventral dissection of adult skate (A) indicating locations of: Leydig organ (L), epigonal organs (E), spiral intestine (I), rectal gland (R) and spleen (S). Hematoxylin & eosin A staining of: (B) spleen (×100), (C) thymus (×100), (D) Leydig organ (×200), (E) epigonal (×200).

Fig. 1.

Ventral dissection of adult skate (A) indicating locations of: Leydig organ (L), epigonal organs (E), spiral intestine (I), rectal gland (R) and spleen (S). Hematoxylin & eosin A staining of: (B) spleen (×100), (C) thymus (×100), (D) Leydig organ (×200), (E) epigonal (×200).

Fig. 2.

Real-time PCR analyses of the expression of (A) TCR α, β, γ and δ; and (B) IgM, IgX, LCI and LCII at various time points in development. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Data points at weeks 1–6, 10 and 11 represent single embryos; weeks 7–9 represent the average of two non-sibling embryos. Interindividual variation at the same time points is indistinguishable within the limits of test error.

Fig. 2.

Real-time PCR analyses of the expression of (A) TCR α, β, γ and δ; and (B) IgM, IgX, LCI and LCII at various time points in development. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Data points at weeks 1–6, 10 and 11 represent single embryos; weeks 7–9 represent the average of two non-sibling embryos. Interindividual variation at the same time points is indistinguishable within the limits of test error.

Fig. 3.

Real-time PCR analyses of the expression of TCR α, β, γ and δ using primer sets that are specific for each constant region isotype. (A) Eight-week embryos and (B) hatchlings. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Each data point represents tissue from a single embryo or hatchling.

Fig. 3.

Real-time PCR analyses of the expression of TCR α, β, γ and δ using primer sets that are specific for each constant region isotype. (A) Eight-week embryos and (B) hatchlings. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Each data point represents tissue from a single embryo or hatchling.

Fig. 4.

Analysis of TCR expression in adult tissues. (A) RNA blots contained 10 μg of total RNA per lane. Probes specific for the constant regions of TCR α, β and γ were used for each hybridization. Exposure to X-ray film was 11 days for TCR α, 9 days for TCR β, and 13 days for TCR γ and S26. Size references are in kbp; the blot was hybridized in the order TCRγ → TCRβ → TCRα → S26. TCR δ hybridization failed to detect a signal (not shown). (B) Ribonuclease protection assay of relative TCR expression. Relative mRNA levels are compared for each TCR between the tissues listed and cannot be compared among mRNA specifying different TCR constant regions.

Fig. 4.

Analysis of TCR expression in adult tissues. (A) RNA blots contained 10 μg of total RNA per lane. Probes specific for the constant regions of TCR α, β and γ were used for each hybridization. Exposure to X-ray film was 11 days for TCR α, 9 days for TCR β, and 13 days for TCR γ and S26. Size references are in kbp; the blot was hybridized in the order TCRγ → TCRβ → TCRα → S26. TCR δ hybridization failed to detect a signal (not shown). (B) Ribonuclease protection assay of relative TCR expression. Relative mRNA levels are compared for each TCR between the tissues listed and cannot be compared among mRNA specifying different TCR constant regions.

Fig. 5.

Real-time PCR analyses of the expression of IgM and IgX in tissues from 8-week embryos. (A) Total IgM and IgX expression, (B) IgM transmembrane (TM) and IgM secretory (SEC) expression; IgX short- and IgX long-form expression. Expression of the IgX short form is inferred by subtraction of relative abundance of long form from total IgX expression. Amplification was quantified by comparison to a standard curve for each primer set and normalized to 28S. Each data point represents tissue from a single embryo.

Fig. 5.

Real-time PCR analyses of the expression of IgM and IgX in tissues from 8-week embryos. (A) Total IgM and IgX expression, (B) IgM transmembrane (TM) and IgM secretory (SEC) expression; IgX short- and IgX long-form expression. Expression of the IgX short form is inferred by subtraction of relative abundance of long form from total IgX expression. Amplification was quantified by comparison to a standard curve for each primer set and normalized to 28S. Each data point represents tissue from a single embryo.

Fig. 6.

Analyses of Ig expression in a hatchling. (A) RNA blot analysis of IgM and IgX expression in different tissues. Each lane contained 10 μg of total RNA. Probes specific for the first constant region of IgM and the first two constant regions of IgX were used for each hybridization. Blots were exposed to a PhosphorImager screen for 14 days; the S26 blot was exposed for 7 days. Size references are in kbp; the blot was hybridized in the order IgX → IgM → S26. (B) Real-time PCR analysis of the expression of transmembrane (TM) and secretory (SEC) forms of IgM. Amplification was quantified by comparison to a standard curve for each primer set and normalized to 28S. Each data point represents tissue from a single hatchling.

Fig. 6.

Analyses of Ig expression in a hatchling. (A) RNA blot analysis of IgM and IgX expression in different tissues. Each lane contained 10 μg of total RNA. Probes specific for the first constant region of IgM and the first two constant regions of IgX were used for each hybridization. Blots were exposed to a PhosphorImager screen for 14 days; the S26 blot was exposed for 7 days. Size references are in kbp; the blot was hybridized in the order IgX → IgM → S26. (B) Real-time PCR analysis of the expression of transmembrane (TM) and secretory (SEC) forms of IgM. Amplification was quantified by comparison to a standard curve for each primer set and normalized to 28S. Each data point represents tissue from a single hatchling.

Fig. 7.

Analyses of Ig expression in an adult. (A) RNA blot analyses of IgM and IgX expression. Each lane contained 10 μg of total RNA. Probes specific for the first constant region of IgM and the first two constant regions of IgX were used for each hybridization. The blot was exposed to X-ray film for 14 days for IgX, 7 days for IgM and 12 days for S26. Size references are in kbp; the blot was hybridized in the order IgX → IgM → S26. (B) Real-time PCR analysis of the expression of transmembrane (TM) and secretory (SEC) forms of IgM. Amplification was quantified by comparison to a standard curve for each primer set and normalized to 28S. Each data point represents tissue from a single adult.

Fig. 7.

Analyses of Ig expression in an adult. (A) RNA blot analyses of IgM and IgX expression. Each lane contained 10 μg of total RNA. Probes specific for the first constant region of IgM and the first two constant regions of IgX were used for each hybridization. The blot was exposed to X-ray film for 14 days for IgX, 7 days for IgM and 12 days for S26. Size references are in kbp; the blot was hybridized in the order IgX → IgM → S26. (B) Real-time PCR analysis of the expression of transmembrane (TM) and secretory (SEC) forms of IgM. Amplification was quantified by comparison to a standard curve for each primer set and normalized to 28S. Each data point represents tissue from a single adult.

Fig. 8.

Analysis of light chain expression. Real-time PCR analyses of the expression of LCI and LCII from: (A) 8-week embryo, (B) hatchling and (C) adult tissues. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Each data point represents tissue recovered from a single animal. (D) RNA blot analyses of LCI and LCII in an adult. Duplicate blots contained 10 μg of total RNA. Probes specific for the constant regions of LCI and LCII were used for each hybridization. One representative panel is shown for S26 hybridization, which was performed after light chain hybridization. Exposure to PhosphorImager was 7 days for LC blots and 14 days for S26. Size references are in kbp.

Fig. 8.

Analysis of light chain expression. Real-time PCR analyses of the expression of LCI and LCII from: (A) 8-week embryo, (B) hatchling and (C) adult tissues. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Each data point represents tissue recovered from a single animal. (D) RNA blot analyses of LCI and LCII in an adult. Duplicate blots contained 10 μg of total RNA. Probes specific for the constant regions of LCI and LCII were used for each hybridization. One representative panel is shown for S26 hybridization, which was performed after light chain hybridization. Exposure to PhosphorImager was 7 days for LC blots and 14 days for S26. Size references are in kbp.

Fig. 9.

Real-time PCR analyses of the expression of Rag-1 and TdT from: (A) 8-week embryo, (B) hatchling and (C) adult. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Each data point represents tissues from single animals.

Fig. 9.

Real-time PCR analyses of the expression of Rag-1 and TdT from: (A) 8-week embryo, (B) hatchling and (C) adult. Amplification was quantified by comparison to a standard curve for each primer set and normalized to coincident amplification of 28S. Each data point represents tissues from single animals.

Fig. 10.

Relative sequence variation (56) plots of in-frame segments of IgM and IgX variable regions, isolated from: (A) 8-week embryonic spleen, Leydig organ, gonad and liver; (B) hatchling spleen, Leydig organ and gonad; and (C) adult spleen, Leydig organ and epigonal. Open bars indicate CDR2 regions and solid bars indicate CDR3 regions (232733).

Fig. 10.

Relative sequence variation (56) plots of in-frame segments of IgM and IgX variable regions, isolated from: (A) 8-week embryonic spleen, Leydig organ, gonad and liver; (B) hatchling spleen, Leydig organ and gonad; and (C) adult spleen, Leydig organ and epigonal. Open bars indicate CDR2 regions and solid bars indicate CDR3 regions (232733).

Transmitting editor: J. F. Kearney

We would like to thank Barbara Pryor for editorial assistance. This work was supported by grants R37 AI23338 to G. W. L. from the National Institutes of Health, and NAG2-1370 to M. K. A. and E. V. R. from NASA. A. L. M has been supported in part by the Institute for Biomolecular Science, University of South Florida and M. K. A. is a fellow of the Stowers Institute. C. A. L. received partial support from the Henry L. and Grace Doherty Charitable Foundation, the Vernal W. and Florence Bates Foundation, and the Disney Wildlife Conservation Fund.

## References

1
Litman, G. W., Anderson, M. K. and Rast, J. P.
1999
. Evolution of antigen binding receptors.
Annu. Rev. Immunol.

17
:
109
.
2
Flajnik, M. F. and Rumpfelt, L. L. 2000. The immune system of cartilaginous fish. In Du Pasquier, L. and Litman, G. W., eds, Current Topics in Microbiology and Immunology: Origin and Evolution of the Vertebrate Immune System, p. 249. Springer, Berlin.
3
Okamura, K., Ototake, M., Nakanishi, T., Kurosawa, Y. and Hashimoto, K.
1997
. The most primitive vertebrates with jaws possess highly polymorphic MHC class I genes comparable to those of humans.
Immunity

7
:
777
.
4
Kasahara, M., Vazquez, M., Sato, K., McKinney, E. C. and Flajnik, M. F.
1992
. Evolution of the major histocompatibility complex: isolation of class II a cDNA clones from the cartilaginous fish.

89
:
6688
.
5
Bartl, S. and Weissman, I. L.
1994
. Isolation and characterization of major histocompatibility complex class IIB genes from the nurse shark.

91
:
262
.
6
Hinds, K. R. and Litman, G. W.
1986
. Major reorganization of immunoglobulin VH segmental elements during vertebrate evolution.
Nature

320
:
546
.
7
Kokubu, F., Litman, R., Shamblott, M. J., Hinds, K. and Litman, G. W.
1988
. Diverse organization of immunoglobulin VH gene loci in a primitive vertebrate.
EMBO J.

7
:
3413
.
8
Hohman, V. S., Schuchman, D. B., Schluter, S. F. and Marchalonis, J. J.
1993
. Genomic clone for the sandbar shark lambda light chain: generation of diversity in the absence of gene rearrangement.

90
:
9882
.
9
Anderson, M. K., Shamblott, M. J., Litman, R. T. and Litman, G. W.
1995
. The generation of immunoglobulin light chain gene diversity in Raja erinacea is not associated with somatic rearrangement, an exception to a central paradigm of B cell immunity.
J. Exp. Med.

181
:
109
.
10
Yoder, J. A. and Litman, G. W. 2000. Immune-type diversity in the absence of somatic rearrangement. In Du Pasquier, L. and Litman, G. W., eds, Current Topics in Microbiology and Immunology: Origin and Evolution of the Vertebrate Immune System, p. 271. Springer, Berlin.
11
Lee, S. S., Fitch, D., Flajnik, M. F. and Hsu, E.
1999
. Rearrangement of immunoglobulin genes in shark germ cells.
J. Exp. Med.

191
:
1637
.
12
Diaz, M., Greenberg, A. S. and Flajnik, M. F.
1998
. Somatic hypermutation of the new antigen receptor gene (NAR) in the nurse shark does not generate the repertoire: possible role in antigen-driven reactions in the absence of germinal centers.

95
:
14343
.
13
Mäkelä, O. and Litman, G. W.
1980
. Lack of heterogeneity in anti-hapten antibodies of a phylogenetically primitive shark.
Nature

287
:
639
.
14
Litman, G. W., Erickson, B. W., Lederman, L. and Mäkelä, O.
1982
. Antibody response in
Heterodontus. Mol. Cell Biochem.

45
:
49
.
15
Shankey, T. V. and Clem, L. W.
1980
. Phylogeny of immunoglobulin structure and function. IX. Intramolecular heterogeneity of shark 19S IgM antibodies to the dinitrophenyl hapten.
J. Immunol.

125
:
2690
.
16
Zapata, A. G. and Cooper, E. L. 1990. The Immune System: Comparative Histophysiology. Wiley, Chichester.
17
Luer, C. A. 1989. Elasmobranchs (sharks, skates, and rays) as animal models for biomedical research. In Woodhead, A. D., ed., Nonmammalian Animal Models for Biomedical Research, p. 121. CRC Press, Boca Raton, FL.
18
Amemiya, C. T., Ota, T. and Litman, G. W. 1996. Construction of P1 artificial chromosome (PAC) libraries from lower vertebrates. In Lai, E. and Birren, B., eds, Analysis of Nonmammalian Genomes, p. 223. Academic Press, San Diego, CA.
19
Vincent, S., Marty, L. and Fort, P.
1993
. S26 ribosomal protein RNA: an invariant control for gene regulation experiments in eucaryotic cells and tissues.
Nucleic Acids Res.

21
:
1498
.
20
Haire, R. N., Miracle, A. L., Rast, J. P. and Litman, G. W.
2000
. Members of the Ikaros gene family are present in early representative vertebrates.
J. Immunol.

165
:
306
.
21
Rast, J. P., Anderson, M. K., Strong, S. J., Luer, C., Litman, R. T. and Litman, G. W.
1997
. α, β, γ, and δ T cell antigen receptor genes arose early in vertebrate phylogeny.
Immunity

6
:
1
.
22
Le, H. L., Lecointre, G. and Perasso, R.
1993
. A 28S rRNA-based phylogeny of the gnathostomes: first steps in the analysis of conflict and congurence with morphologically based cladogrms.
Mol. Phylogenet. Evol.

2
:
31
.
23
Harding, F. A., Amemiya, C. T., Litman, R. T., Cohen, N. and Litman, G. W.
1990
. Two distinct immunoglobulin heavy chain isotypes in a primitive, cartilaginous fish,
Raja erinacea. Nucleic Acids Res.

18
:
6369
.
24
Haire, R. N., Buell, R. D., Litman, R. T., Ohta, Y., Fu, S. M., Honjo, T., Matsuda, F., de la Morena, M., Carro, J., Good, R. A. and Litman, G. W.
1993
. Diversification, not utilization, of the immunoglobulin VH gene repertoire is restricted in DiGeorge syndrome.
J. Exp. Med.

178
:
825
.
25
Honma, Y., Okabe, K. and Chiba, A.
1984
. Comparative histology of the Leydig and epigonal organs in some elasmobranchs.
Jap. J. Ichthyol.

31
:
47
.
26
Rast, J. P., Anderson, M. K., Ota, T., Litman, R. T., Margittai, M., Shamblott, M. J. and Litman, G. W.
1994
. Immunoglobulin light chain class multiplicity and alternative organizational forms in early vertebrate phylogeny.
Immunogenetics

40
:
83
.
27
Anderson, M., Amemiya, C., Luer, C., Litman, R., Rast, J., Niimura, Y. and Litman, G.
1994
. Complete genomic sequence and patterns of transcription of a member of an unusual family of closely related, chromosomally dispersed immunoglobulin gene clusters in
Raja. Int. Immunol.

6
:
1661
.
28
Anderson, M. K., Strong, S. J., Litman, R. T., Luer, C. A., Amemiya, C. T., Rast, J. P. and Litman, G. W.
1999
. A long form of the skate IgX gene exhibits a striking resemblance to the new shark IgW and IgNARC genes.
Immunogenetics

49
:
56
.
29
Kokubu, F., Hinds, K., Litman, R., Shamblott, M. J. and Litman, G. W.
1988
. Complete structure and organization of immunoglobulin heavy chain constant region genes in a phylogenetically primitive vertebrate.
EMBO J.

7
:
1979
.
30
Lacy, E. R. and Reale, E.
1985
. The elasmobranch kidney. I. Gross anatomy and general distribution of nephros.
Anat. Embryol.

173
:
23
.
31
Greenberg, A. S., Hughes, A. L., Guo, J., Avila, D., McKinney, E. C. and Flajnik, M. F.
1996
. A novel chimeric' antibody class in cartilaginous fish: IgM may not be the primordial immunoglobulin.
Eur. J. Immunol.
,
26
:
1123
.
32
Blackwell, T. K. and Alt, F. W.
1989
. Mechanism and developmental program of immunoglobulin gene rearrangement in mammals.
Annu. Rev. Genet.

23
:
605
.
33
Harding, F. A., Cohen, N. and Litman, G. W.
1990
. Immunoglobulin heavy chain gene organization and complexity in the skate,
Raja erinacea. Nucleic Acids Res.

18
:
1015
.
34
Hinds-Frey, K. R., Nishikata, H., Litman, R. T. and Litman, G. W.
1993
. Somatic variation precedes extensive diversification of germline sequences and combinatorial joining in the evolution of immunoglobulin heavy chain diversity.
J. Exp. Med.

178
:
825
.
35
Fowlkes, B. J. and Pardoll, D. M.
1989
. Molecular and cellular events of T cell development.

44
:
207
.
36
Coltey, M., Bucy, R. P., Chen, C. H., Cihak, J., Loseh, U., Char, D., Le Douarin, N. M., Cooper, M. D.
1989
. Analysis of the first two waves of thymus homing stem cells and their T cell progeny in chick–quail chimeras.
J. Exp. Med.

170
:
543
.
37
Chien, Y.-H., Iwashima, M., Kaplan, K. B., Elliott, J. F. and Davis, M. M.
1987
. A new T-cell receptor gene located within the alpha locus and expressed early in T-cell differentiation.
Nature

327
:
677
.
38
Owen, J. J., Cooper, M. D. and Raff, M. C.
1974
. In vitro generation of B lymphocytes in mouse foetal liver, a mammalian bursa equivalent'.
Nature

249
:
361
.
39
Velardi, A. and Cooper, M. D.
1984
. An immunofluorescence analysis of the ontogeny of myeloid, T, and B lineage cells in mouse hemopoietic tissues.
J. Immunol.

133
:
672
.
40
Reynaud, C.-A., Anquez, V., Grimal, H. and Weill, J.-C.
1987
. A hyperconversion mechanism generates the chicken light chain preimmune repertoire.
Cell

48
:
379
.
41
Benedict, C. L., Gilfillan, S., Thai, T. H. and Kearney, J. F.
2000
. Terminal deoxynucleotidyl transferase and repertoire development.
Immunol Rev

175
:
150
.
42
Anderson, M. K., Sun, X., Miracle, A. L., Litman, G. W. and Rothenberg, E. V.
2001
. Evolution of hematopoiesis: three members of the PU.1 transcription factor family in a cartilaginous fish,
Raja eglanteria. Proc. Natl Acad. Sci. USA
,
98
:
553
.
43
Borrello, M. A. and Phipps, R. P.
1996
. The B/macrophage cell: an elusive link between CD4+ B lymphocytes and macrophages.
Immunol. Today

17
:
471
.
44
Martin, F. and Kearney, J. F.
2000
. B-cell subsets and the mature preimmune repertoire. Marginal zone and B1 B cells as part of a `natural immune memory'.
Immunol. Rev.

175
:
70
.
45
Mattisson, A. and Fänge, R.
1982
. The cellular structure of the Leydig organ in the shark, Etmopterus spinax (L.).
Biol. Bull.

162
:
182
.
46
Feeney, A. J.
1990
. Lack of N regions in fetal and neonatal mouse immunoglobulin V–D–J junctional sequences.
J. Exp. Med.

172
:
1377
.
47
Schwager, J., Burckert, N., Courtet, M. and Du Pasquier, L.
1991
. The ontogeny of diversification at the immunoglobulin heavy chain locus in
Xenopus. EMBO J.

10
:
2461
.
48
Feeney, A. J.
1992
. Predominance of VH–D–JH junctions occurring at sites of short sequence homology results in limited junctional diversity in neonatal antibodies.
J. Immunol.

149
:
222
.
49
Bangs, L. A., Sanz, I. E. and Teale, J. M.
1991
. Comparison of D, JH, and junctional diversity in the fetal, adult, and aged B cell repertoires.
J. Immunol.

146
:
1996
.
50
Reynaud, C.-A., Dahan, A., Anquez, V. and Weill, J.-C.
1989
. Somatic hyperconversion diversifies the single VH gene of the chicken with a high incidence in the D region.
Cell

59
:
171
.
51
Reynaud, C.-A., Mackay, C. R., Muller, R. G. and Weill, J.-C.
1991
. Somatic generation of diversity in a mammalian primary lymphoid organ: the sheep ileal Peyer's patches.
Cell

64
:
995
.
52
Reynaud, C.-A., Garcia, C., Hein, W. R. and Weill, J.-C.
1995
. Hypermutation generating the sheep immunoglobulin repertoire is an antigen-independent process.
Cell

80
:
115
.
53
Knight, K. L.
1992
. Restricted VH gene usage and generation of antibody diversity in rabbit.
Annu. Rev. Immunol.

10
:
593
.
54
Parng, C. L., Hansal, S., Goldsby, R. A. and Osborne, B. A.
1996
. Gene conversion contributes to Ig light chain diversity in cattle.
J Immunol

157
:
5478
.
55
Sun, J. and Butler, J. E.
1996
. Molecular characterization of VDJ transcripts from a newborn piglet.
Immunology

88
:
331
.
56
Kabat, E. A., Wu, T. T., Foeller, C., Perry, H. M. and Gottesman, K. 1991. Sequences of Proteins of Immunological Interest, 5 edn. US Department of Health and Human Services, Washington, DC.