The Journal of Immunology, 2004, 172: 6764-6767.
Copyright © 2004 by The American Association of Immunologists
Evidence That the Long Murine Terminal Deoxynucleotidyltransferase Isoform Plays No Role in the Control of V(D)J Junctional Diversity
Noëlle Doyen,
Jean-Baptiste Boulé1,
François Rougeon and
Catherine Papanicolaou2
Unité de Génétique et Biochimie du Développement, Unité de Recherche Associée Centre National de la Recherche Scientifique 2581, Institut Pasteur, Paris, France
 |
Abstract
|
|---|
Two TdT isoforms have been found in the mouse. The short isoform is known to add N regions to gene segment junctions during V(D)J recombination, but the role of the long (TdTL) isoform is controversial. We have shown that TdTL, although endowed with terminal transferase activity, is thermally unstable and unable to add N regions in vivo. In this study, we demonstrate that TdTL is devoid of 3'-5' exonuclease activity, and provide an analysis of nucleotide deletion and addition patterns in large series of V(D)J coding joins, arguing against a role of TdTL in the control of junctional diversity in Igs and TCRs.
|
Introduction
|
|---|
Terminal deoxynucleotidyltransferase (TdT) participates in the diversification of the immune repertoire by adding nucleotides (N regions) to the junctions of DNA segments assembled during V(D)J recombination (1, 2, 3). Our laboratory first reported the expression through alternative splicing of two isoforms of TdT in the mouse (4), the only species with two such isoforms. Unlike murine TdT short (TdTS)3 isoform, murine TdT long (TdTL) isoform, which has a 20-aa insertion, cannot add N regions to V(D)J junctions and tends to remain in the cytoplasm of transfected cells, where it is rapidly degraded (5). In transgenic mice expressing only TdTL, the protein is localized in the nucleus of splenocytes, but is still unable to add N regions to V(D)J junctions (6). Interestingly, the 20-aa insertion renders TdTL thermosensitive. Nevertheless, TdTL is a true terminal transferase, as it can polymerize nucleotides efficiently in vitro (7).
In a recent report, murine TdTS and TdTL were described as having "distinct and opposite diversifying activities" (8). In this study, we conducted additional experiments on the catalytic properties of the two TdT isoforms and compared the junctional diversity in sets of sequences recombined in transfected cells, in the absence of TdT and in the presence of TdTS or TdTL. Our results do not support the hypothesis that the two murine TdT isoforms play competing roles in the processing of coding junctions in B cell receptor and TCR genes (8).
|
Materials and Methods
|
|---|
Assay for site-specific recombination
NIH 3T3 mouse embryo fibroblasts (ATCC CRL 6442; American Type Culture Collection, Manassas, VA) (3 x 106 cells) were transfected with 2.5 µg of pBlueRec, 6 µg of pRAG1, 4.8 µg of pRAG2, and 4.5 µg of pTdTS or pTdTL. COS 1 African Green Monkey SV40-transformed kidney cells (ATCC CRL 1650) (3 x 106 cells) were transfected with pH2Rec instead of pBlueRec and with the same other vectors. Controls were done in the absence of the vector expressing TdT. Plasmid DNAs were recovered after incubation of the cells at 37°C/5% CO2 for 48 h and transformed into Escherichia coli competent cells. Rearrangement of pBlueRec and pH2Rec recombination substrates gives rise to blue colonies. Plasmid DNAs were prepared, and the recombination junctions were sequenced. Details of the different procedures and a map of the recombination plasmid have been given (1, 5, 9).
Protein purification
Strains, plasmids, and methods used for overexpression of TdTS and TdTL in bacteria have been described previously (7, 10). The presence of the additional 20 aa had proven to make murine TdTL more difficult to both produce and purify than TdTS, and the procedure had been modified to obtain sufficient amounts of this isoform (7, 10). Two additional purification steps were applied in the present work to eliminate traces of a bacterial 3'
5' contaminant only detectable in our TdTL preparation and at high enzyme/primer ratio (data not shown). The protocol used for protein purification was as follows: cell pellets from 1 L (TdTS) or 12 L (TdTL) bacterial culture were resuspended in 1/30 vol buffer A (50 mM phosphate buffer, pH 7.0, 300 mM NaCl) and lysed in a homogenizer (EmulsiFlex-C5; Avestin, Ottawa, Canada) at 10,000 
. The supernatant was loaded at a rate of 1 ml/min on a 5-ml cobalt-affinity resin (Talon resin; Clontech Laboratories, Palo Alto, CA). The column was washed with 3–5 vol buffer A + 10 mM imidazole, and proteins were eluted with buffer A + 200 mM imidazole. Fractions containing TdT were pooled, dialyzed against buffer B (NaOAc, pH 4.6, 200 mM NaCl, 50 mM MgOAc, 50 mM (NH4)2SO4), and loaded on a 2-ml cation exchange resin (resource 15S; Amersham Pharmacia Biotech, Piscataway, NJ) equilibrated in buffer B. Proteins were eluted with a 0–1 M NaCl gradient. Fractions containing TdT were pooled, desalted, and loaded on a 1-ml heparin column (Hitrap heparin; Amersham Pharmacia Biotech) equilibrated in buffer B. Proteins were eluted with a 0–1 M NaCl gradient, and the elution fractions were analyzed by SDS-PAGE. Pure fractions (>98%) were dialyzed against buffer C (HEPES-NaOH, pH 6.9, 200 mM NaCl, 50 mM (NH4)2SO4, 50 mM MgOAc), concentrated, and kept frozen at –70°C after addition of 50% glycerol. Protein concentration was estimated by absorbance at 280 nM using a theoretical extinction coefficient of 54,870 M–1 cm–1.
Primer elongation/degradation assay
Terminal transferase activity was detected in protein preparations by incorporation of dATP into ssDNA using the following standard assay: 130 nM TdTS or 130 nM TdTL was incubated at 35°C in 200 nM potassium cacodylate, 25 mM Tris-HCl, pH 6.6, 0.25 mg/ml BSA, 4 mM MgCl2, 4 µM ZnSO4, 5% glycerol, 1 mM dATP, and 20 nM 5'-32P-labeled (dA)10 primer. The 3'5' exonuclease activity in protein preparations was searched using the same assay in the absence of dATP. Aliquots were withdrawn at 0, 5, 15, 30, and 60 min, supplemented with a formamide dye mix, and electrophoresed on a 16% acrylamide denaturating gel. Products were visualized after exposure of the wet gel under a Kodak film (Biomax MR) at –70°C.
|
Results
|
|---|
TdTL does not modify V(D)J recombination coding ends
We had previously compared the functional properties of the two murine TdT isoforms by analyzing the nucleotide additions at the V(D)J recombination junctions of an episomic substrate transfected into NIH 3T3 fibroblasts. This work established the role of TdTS in N region addition (1) and the lack of N region addition by TdTL (5). Following the report by Thai et al. (8) that murine TdTS and TdTL have "distinct and opposite diversifying activities," we conducted a retrospective analysis of various sets of junction sequences obtained in transfection experiments from NIH 3T3 and COS cell transfectants (1, 5). The junction sequences from NIH 3T3 fibroblasts and COS cells are displayed in Fig. 1, a and b, respectively, with their characteristics summarized in Table I. In the presence of TdTS, N regions are present in 
75% of the rearranged plasmids (79 and 67%, respectively, in 3T3 and COS cells), whereas in the absence of TdT or in the presence of TdTL, only 4–14% of the junctions contain nucleotide additions (see Table I and Fig. 1, a and b). As previously reported (5), the nucleotide insertion frequencies observed with TdTL are in the same range as those observed in the absence of TdT or in the junctions retrieved from TdT knockout mice (2, 3). Presumed palindromic (P) nucleotides, which are believed to result from asymmetric opening of coding end recombination intermediate hairpins, are present in no-TdT, TdTS, and TdTL sequences, and in both 3T3 (4, 44, and 19%, respectively) and COS cells (4, 21, and 18%, respectively). They are more frequent at recombined junctions in TdTS- or TdTL-transfected 3T3 or COS cells than at junctions retrieved from cells not expressing TdT, as if the presence of TdT influenced hairpin or P nucleotide processing, but there is no significant difference in the percentage of P nucleotides added when sequences from TdTS-transfected cells and TdTL-transfected cells are compared. The analyses of deletion patterns at recombination junctions in 3T3 and COS cells reveal no significant differences among no-TdT, TdTS, and TdTL sequences (Fig. 1 and Table I). The nucleotide loss, calculated on average per sequence, is the same (4 nt) whether recombination occurred in the absence of TdT or in the presence of TdTS or TdTL. About one-half of the TdTS and TdTL sequences have deletions >4 nt (40–52%). Deletions >4 nt are also present in about one-half of the no-TdT sequences in COS cells (46%) and in 22% of the sequences in 3T3 cells.
View larger version (75K):
[in this window]
[in a new window]
|
FIGURE 1. Sequences of the junctions formed in pBlueRec or pH2Rec recombination substrates after cotransfection with plasmids coding for RAG1, RAG2, and either no TdT, TdTS, or TdTL in NIH 3T3 fibroblasts (a) and in COS cells (b). Recombined sequences are aligned with the underlined unmodified coding ends of pBlueRec (a) and pH2Rec (b). Numbers in parentheses indicate repeated sequences recovered from independent transfections: 9 and 5 transfections in the absence of TdT, 13 and 5 transfections in the presence of TdTS, and 6 and 5 transfections in the presence of TdTL, in 3T3 and COS cells, respectively. Sequences already published (1 ) are marked with an asterisk. Putative P nucleotides are underlined. nt deleted, Represent the number of nucleotides deleted from the coding ends. ¶, Runs of nucleotides are indicated as follows: nN = Nn.
|
|
Murine TdTL exhibits terminal transferase activity in vitro and is devoid of 3'-5' exonuclease activity
Comparison of the two murine TdT isoform activities was performed on a 32P-labeled (dA)10 primer substrate in enzyme excess (enzyme/primer > 5), with proteins highly purified from bacterial extracts (see Materials and Methods for details). The chain length distributions of the products obtained in kinetics studies over a 60-min time course, in the absence or presence of nucleotides, are shown in Fig. 2. In the presence of nucleotides, both TdTS and TdTL exhibit terminal transferase activity, with TdTS synthesizing more of the longer polymers. Synthesis by TdTL is less efficient, but within 15 min almost all of the (dA)10 primer has been elongated. In the absence of nucleotides, there is no degradation of the primer by TdTS or TdTL. Preincubation of TdTS or TdTL at 35°C for 1 h leads to the reduction of terminal transferase activity, as previously reported (7), but does not reveal any cryptic exonuclease activity (data not shown).
View larger version (65K):
[in this window]
[in a new window]
|
FIGURE 2. Product analyses of TdTS and TdTL reactions. A total of 130 nM TdTS or TdTL (2.5 µl) was incubated with 20 nM 5'-32P-labeled (dA)10 oligonucleotide, in the presence of 1 mM dATP or in the absence of nucleotide, in a 25 µl final volume. Aliquots were withdrawn at 0, 5, 15, 30, and 60 min, and products were analyzed on a 16% polyacrylamide denaturing gel.
|
|
|
Discussion
|
|---|
It has been shown that in transfected cells, only TdTS adds N nucleotides to V(D)J recombination junctions (5, 8). The absence of N region addition by TdTL is not due to a lack of terminal transferase activity (7). As we showed in this study, TdTL is, under appropriate in vitro conditions, also able, like TdTS, to synthesize very long DNA chains. However, TdTL is much more unstable than TdTS in transfected cells (5) and is thermosensitive in vitro (7). This fragility could explain why TdTL, which has the same nuclear localization sequence as TdTS, does not accumulate in the nucleus of transfected cells and does not add N regions (5). The lack of terminal transferase activity recently reported for TdTL in vitro cannot be explained by sequence differences in the clones used,4 and is most likely due to the experimental conditions that are, considering the thermosensitivity of the enzyme, highly unfavorable (8).
Our exonuclease assay was free of nucleotides because the competing polymerase activity of TdTS and TdTL would in the presence of nucleotides mask any associate 3'-5' exonuclease activity, whether intrinsic or contaminant. In these conditions, no 3'-5' exonuclease activity is detected in vitro, even at high enzyme/primer ratio, with either TdTS or TdTL. In the work of Thai et al. (8), the presence of nucleotides in the assay precludes to assert that only TdTL has exonuclease activity because the temperature-sensitive polymerase activity of TdTL is, as explained above, suppressed under the experimental conditions used.
Using a recombination assay, we found no significant difference among no-TdT, TdTS, and TdTL deletion patterns of V(D)J recombination junctions in two different transfected cell lines (3T3 or COS cells), indicating that TdTL does not have any effect on coding end trimming in these cells. Interestingly, a lack of effect of TdTL on coding end processing was also found in sequences retrieved from transgenic mice only expressing TdTL (6). Thai et al. (8) had deduced from results obtained in similar transfection experiments, but using Chinese hamster ovary cells, that murine TdTS and TdTL have "distinct and opposite diversifying activities." It is conceivable that TdTL recruits an exonuclease specifically expressed in Chinese hamster ovary cells. However, the observation that the effect of TdTL on V(D)J junctions is limited to specific experimental conditions makes it unlikely that such properties contribute to shaping the immune repertoire in the mouse.
TdT is conserved across the vertebrate phylum, from cartilaginous fish to humans (12), yet the mouse is the only species in which expression of TdTS and TdTL isoforms has been demonstrated. Three alternatively spliced mRNAs have been isolated by RT-PCR from calf thymus (13), and potential splicing sites corresponding to the bovine splicing sites have been identified in human genomic sequences (8). However, the two bovine TdTL isoforms, with none identical with murine TdTL, have not been shown to be expressed during lymphocyte differentiation. Although it is now believed that a high percentage of human genes undergoes alternative splicing (see Sorek et al. (14) and references therein), identification of potential alternative splice sites is not sufficient to ascertain a functional and regulated expression of spliced variants. As discussed previously (7), TdTL may result from an evolutionary happenstance or represent an ancestral, maybe vestigial, form of the enzyme. It is interesting to note that in a few members of the family X of DNA polymerases, a subclass of an ancient nucleotidyltransferase superfamily to which TdT belongs, the position of the TdTL 20-aa insertion coincides with an insertion zone of sequences of variable length (15). Structural analyses localize the position of the insertion within the thumb subdomain, in a region, between two conserved domains, which forms a short loop pointing outward the structured catalytic core (16, 17, 18). In the case of TdTL, the extension of the loop does not modify the catalytic activity, but has functional consequences. The underlying mechanisms of the functional inhibition caused by the presence of the loop in TdTL are unknown, but may involve specific protein-protein interactions (14).
|
Footnotes
|
|---|
1 Current address: Department of Molecular Biology, Lewis Thomas Labs, Princeton University, Princeton, NJ 08544. 
2 Address correspondence and reprint requests to Dr. Catherine Papanicolaou, Unité de Génétique et Biochimie du Développement, Unité de Recherche Associée Centre National de la Recherche Scientifique 2581, Institut Pasteur, 25 rue du Dr Roux, 75015 Paris, France. E-mail address: papanico{at}pasteur.fr
3 Abbreviations used in this paper: TdTS, TdT short; P, palindromic; RAG, recombination-activating gene; TdTL, TdT long.
4 Errors in the original published sequence of TdTS isolated from BALB/c mice thymus (4 ) were corrected. The correct sequence has been published (7 ) and is accessible online (Genbank accession number CAA48634). Crystal structure of the catalytic core of BALB/c TdTS, which comprises the sequence encoded by exon IX identical with the sequence of the C57BL/6 TdTS used by Thai et al (8 ), has been resolved at 2.35 Å (11 ). Thai et al., misled by the sequence error in exon IX (4 ), relegate the discrepancies between their results and ours to differences in this exon in the TdT clones used (8 ). However, the only differences that exist are confined to 2 aa at positions 26 and 99, both localized in a region that is not essential to terminal transferase activity in vitro and neither of which is conserved among all known TdT sequences (7 12 ).
Received for publication August 29, 2003.
Accepted for publication March 18, 2004.
|
References
|
|---|
- Kallenbach, S., N. Doyen, M. Fanton d’Andon, F. Rougeon. 1992. Three lymphoid-specific factors account for all junctional diversity characteristic of somatic assembly of T-cell receptor and immunoglobulin genes. Proc. Natl. Acad. Sci. USA 89:2799.[Abstract/Free Full Text]
- Gilfillan, S., A. Dierich, M. Lemeur, C. Benoist, D. Mathis. 1993. Mice lacking TdT: mature animals with an immature lymphocyte repertoire. Science 261:1175.[Abstract/Free Full Text]
- Komori, T., A. Okada, V. Stewart, F. W. Alt. 1993. Lack of N regions in antigen receptor variable region genes of TdT-deficient mice. Science 261:1171.[Abstract/Free Full Text]
- Doyen, N., M. Fanton d’Andon, L. A. Bentolila, Q. T. Nguyen, F. Rougeon. 1993. Differential splicing in mouse thymus generates two forms of terminal deoxynucleotidyl transferase. Nucleic Acids Res. 21:1187.[Abstract/Free Full Text]
- Bentolila, L. A., M. Fanton d’Andon, Q. T. Nguyen, O. Martinez, F. Rougeon, N. Doyen. 1995. The two isoforms of mouse terminal deoxynucleotidyl transferase differ in both the ability to add N regions and subcellular localization. EMBO J. 14:4221.[Medline]
- Benedict, C. L., S. Gilfillan, J. F. Kearney. 2001. The long isoform of terminal deoxynucleotidyl transferase enters the nucleus and, rather than catalyzing nontemplated nucleotide addition, modulates the catalytic activity of the short isoform. J. Exp. Med. 193:89.[Abstract/Free Full Text]
- Boulé, J. B., F. Rougeon, C. Papanicolaou. 2000. Comparison of the two murine terminal deoxynucleotidyltransferase terminal isoforms: a 20-amino acid insertion in the highly conserved carboxyl-terminal region modifies the thermosensitivity but not the catalytic activity. J. Biol. Chem. 275:28984.[Abstract/Free Full Text]
- Thai, T. H., M. M. Purugganan, D. B. Roth, J. F. Kearney. 2002. Distinct and opposite diversifying activities of terminal transferase splice variants. Nat. Immunol. 3:457.[Medline]
- Kallenbach, S., M. Goodhardt, F. Rougeon. 1990. A rapid test for V(D)J recombinase activity. Nucleic Acids Res. 18:6730.[Free Full Text]
- Boulé, J. B., E. Johnson, F. Rougeon, C. Papanicolaou. 1998. High-level expression of murine terminal deoxynucleotidyl transferase in Escherichia coli grown at low temperature and overexpressing argU tRNA. Mol. Biotechnol. 10:199.[Medline]
- Delarue, M., J. B. Boulé, J. Lescar, N. Expert-Bezançon, N. Jourdan, N. Sukumar, F. Rougeon, C. Papanicolaou. 2002. Crystal structures of a template-independent DNA polymerase: murine terminal deoxynucleotidyltransferase. EMBO J. 21:427.[Medline]
- Bartl, S., A. L. Miracle, L. L. Rumfelt, T. B. Kepler, E. Mochon, G. W. Litman, M. F. Flajnik. 2003. Terminal deoxynucleotidyl transferases from elasmobranches reveal structural conservation within vertebrates. Immunogenetics 55:594.[Medline]
- Takahara, K., N. Hayashi, K. Fujita-Sagawa, T. Morishita, Y. Hashimoto, A. Todo. 1994. Alternative splicing of bovine terminal deoxynucleotide transferase cDNA. Biosci. Biotechnol. Biochem. 58:786.[Medline]
- Sorek, R., R. Shamir, G. Ast. 2004. How prevalent is functional alternative splicing in the human genome?. Trends Genet. 20:68.[Medline]
- Taladriz, S., T. Hanke, M. J. Ramiro, M. Garcia-Diaz, M. Garcia De Lacoba, L. Blanco, V. Laraga. 2001. Nuclear DNA polymerase
from Leishmania infantum: cloning, molecular analysis and developmental regulation. Nucleic Acids Res. 29:3822.[Abstract/Free Full Text]
- Sawaya, M. R., H. Pelletier, A. Kumar, S. H. Wilson, J. Kraut. 1994. Crystal structure of rat DNA polymerase : evidence for a common polymerase mechanism. Science 264:1930.[Abstract/Free Full Text]
- Maciejewski, M. W., R. Shin, B. Pan, A. Marintchev, A. Denninger, M. A. Mullen, K. Chen, M. R. Kryk, G. P. Mullen. 2001. Solution structure of a viral DNA repair polymerase. Nat. Struct. Biol. 8:936.[Medline]
- Showalter, A. K., I. J. Byeon, M. I. Su, M. D. Tsai. 2001. Solution structure of a viral DNA polymerase X and evidence for a mutagenic function. Nat. Struct. Biol. 8:942.[Medline]
Top
Abstract
Introduction
