| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,
,
* Department of Medicine,
Department of Biochemistry and Biophysics,
Department of Microbiology and Immunology, and
James P. Wilmot Cancer Center, University of Rochester School of Medicine and Dentistry, Rochester, NY 14642;
¶ GSF-National Research Centre for Environment and Health, Institute of Molecular Radiobiology, Munich, Germany; and
|| Department of Haematology, East Anglia Blood Centre, University of Cambridge, Cambridge, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Unlike SHM and CSR, AID-dependent DNA deaminase activity in experimental systems shows limited sequence and locus specificity, although some site preference is detectable (8, 11, 12, 13). Specific sequence requirements for functional S regions are also not stringent (14, 15, 16). Based on analysis of AID mutants (17, 18, 19), and by analogy with other members of the APOBEC-1-related family of enzymes (4), it has been hypothesized that AID is directed to its natural substrates through its interaction with specific targeting factors. In particular, CSR activity has been linked to a C-terminal nuclear export signal motif (NES) in the noncatalytic domain of AID (17, 18), which mediates chromosome region maintenance 1 (CRM1)-dependent nuclear export (20, 21, 22). AID SHM activity appears to be dependent instead on separate motifs mapping to the catalytic domain of the molecule (19).
Phylogenetic evidence indicates that SHM and CSR appeared separately during vertebrate evolution (23, 24). SHM is known to occur already in cartilaginous fish, both in Ig and in nurse shark Ag receptor genes, albeit with variable patterns and frequencies, and is likely present in all jawed vertebrates (25, 26, 27, 28). Class switching has been found so far only in land vertebrates (tetrapods), starting with amphibians (23, 24, 29). Although mechanistic differences between vertebrate classes are observed in SHM and CSR, such as in the pattern of mutation and the sequence of S region substrates, the underlying mechanisms are thought to be largely conserved (23, 24). AID homologs have been recently identified by sequence prediction or direct cDNA cloning in several teleosts and in shark (30, 31, 32), suggesting that AID activity was already involved in SHM before CSR appeared. The evolutionary origin of AID from related precursors, as well as the events by which it acquired different activities in the course of vertebrate evolution, are still unknown.
To shed light on the phylogenetic and functional properties of AID during vertebrate evolution, we cloned AID cDNAs from Takifugu and Xenopus, and compared their structural and functional features with those of mammalian AID.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Two complete Xenopus laevis AID (XlAID) cDNAs were cloned from two independent spleen cDNA libraries, using a PCR approach based on homologous sequences present in Assembly V2.0 of the Xenopus tropicalis genome (
http://genome.jgi-psf.org
) (positions 6975–7397 of scaffold 44401) and the Xenopus EST project (www.genome.wustl.edu) database.
Takifugu rubripes AID (TrAID) cDNA was obtained by PCR from a T. rubripes spleen cDNA library (a gift from G. Elgar of the Fugu Genomics Group, London, U.K.) using DNA sequence data from Build 2c of the Takifugu genome (www.ensembl.org/Fugu_rubripes) for primer design.
The Takifugu-human AID fusion proteins were created by overlapping PCR from partial TrAID (aa 1–145) and Homo sapiens AID (HsAID) (aa 139–198) PCR fragments such that the flap region of TrAID was retained. All cDNAs were subcloned via pGEM7zf (Promega) and sequenced before subcloning into the appropriate vectors for functional analysis.
Analysis of AID expression in Xenopus
Outbred Xenopus were obtained from our breeding colony at the University of Rochester. For immunization, Escherichia coli (XL1-blue; Stratagene) were cultured overnight, boiled 30 min, spun, and resuspended in 0.1 vol of Xenopus cell culture medium (33). Two- to 3-year-old Xenopus were injected i.p. with 200 µl of this preparation. Total or cytoplasmic RNA was isolated from Xenopus tissue and 10–20 µg of RNA were analyzed by Northern blotting with a complete XlAID cDNA as probe, or used for reverse transcription. For RT-PCR, samples were amplified using AID primers 5'-TCACGACCCCCATAGGAACTAC-3', 5'-GGATTTAGGAGACTTTGCCTCAAG-3', or 
2-microglobulin primers 5'-CCCTTGTGGTGTAACTGTGCTC-3', reverse 5'-GCACACACCAATCAGAAAAAGGAC-3' for 40 cycles at 58°C annealing temperature.
AID-deficient mouse B cell complementation
All animal work was performed humanely according to federal and local guidelines and approved by the University Committee on Animal Research. cDNAs were cloned into the enhanced GFP (EGFP)-expressing bicistronic pMIG retrovirus vector (34) and retroviral particles generated by packaging into the Phoenix-Eco cell line. Spleen B cells from AID-deficient mice (1) were activated in vitro with 20 µg/ml LPS, 10 µg/ml dextran sulfate, 25 ng/ml IL-4 in complete RPMI 1640 medium, and incubated with viral supernatant at days 1 and 2 of culture using the flow-through technique of Chuck and Palsson (35). Cells were harvested at day 5 and stained with Abs to mouse B220 and IgG1 (BD Pharmingen), and analyzed by flow cytometry using a BD Biosciences FACSCalibur machine. Cells were live-gated based on forward- and side-scatter profile, and the number of IgG1+ cells was measured in the B220+, GFP+ population. In some experiments, GFP+ cells were further subdivided according to high-, medium- or low-GFP expression, and the frequency of IgG1+ cells was measured in each subset.
Bacterial mutation assays
DNA mutator activity by AID orthologs was assayed in BW310 strain E. coli using cDNAs cloned into pTRc99A vectors. Transformant clones were picked, grown in liquid culture, and induced with isopropyl -D-thiogalactoside overnight before plating onto rifampicin selection or nonselective plates, essentially as described (8, 12).
SHM assays
The plasmid pCMV-E(X) was stably cotransfected in the transformed kidney cell line 293 with one of the AID-expressing vectors pCMV-TrAID, pCMV-XlAID, or pCMV-HsAID, respectively. Plasmid pCMV-E(X) is identical with pEGFP-C1 (Clontech Laboratories) but contains a premature TAG stop codon in the GFP coding region (36). For each AID expression construct, 12 independent transfected clones were analyzed by flow cytometry. Before flow cytometry, cells were washed and resuspended in PBS containing propidium iodine to exclude dead cells from the analysis.
Cellular localization of AID
TrAID-EGFP and HsAID-EGFP fusion proteins were generated by cloning TrAID and HsAID cDNAs in-frame into a pEGFP-N1 plasmid (Clontech). Transiently transfected NIH 3T3 cells were cultured for 30 h and transferred onto Nunc Lab-Tek II culture chamber slides overnight, and duplicates were then either left untreated, or incubated with 100 ng/ml leptomycin B (Sigma-Aldrich) for an additional 2 h. Cells were fixed with 4% formaldehyde and nuclei were stained for 10 min with 10 µg/ml propidium iodide in PBS after RNase treatment (10 min at 37°C). Slides were mounted using Gel/Mount medium and analyzed on an Olympus BX-40 fluorescence microscope using ImagePro-Plus 5.0 software (Media Cybernetics).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Takifugu and Xenopus AID cDNAs, were cloned from cDNA libraries based on genomic database information from the fugu and X. tropicalis genome projects. Compared with HsAID (37), XlAID contains two potential additional amino acid residues at the N terminus and an additional glycine at position 70 (human equivalent), for a total of 201 residues, a predicted molecular mass of 24 kDa, and an identity of 67.5% between the two species (Fig. 1A). XlAID mRNA expression in adult frogs is detectable mainly in spleen and was up-regulated following immunization with heat-killed E. coli (not shown), supporting a role for AID in humoral immune responses in amphibians analogous to that in mammals.
|
2) to the third sheet (3), distant from the active site, and therefore predicted not to affect catalytic activity. The first methionine in the cloned TrAID cDNA corresponds to residue 4 in mammalian AID, in agreement with the other Takifugu AID cDNA (32), but unlike the sequences predicted based on genome databases (30, 31). Regardless of whether the discrepancy is due to the presence of partial or alternative cDNA isoforms in Takifugu or to errors in sequence predictions, the observation that the TrAID catalytic domain is functional in mammalian cells shows that this specific difference does not significantly affect catalytic activity (see below). To map potential salient structural-functional motifs onto the AID molecules, we also generated reciprocal fusion cDNAs in which the human and Takifugu CDs and flap regions were fused to the NCDs of the other species (TrCD/HsNCD and HsCD/TrNCD proteins) as well as a set in which 16-aa C-terminal amino acid segments, containing NES elements, were swapped (TrAID/HsNES and HsAID/TrNES) (Fig. 1B).
Mutator activity of TrAID and XlAID in bacteria
We evaluated whether the cloned cDNAs encode functional cytidine deaminase enzymes using a bacterial mutator assay, in which AID-mediated DNA deamination at cytidine residues results in increased frequency of rifampicin (Rif)-resistant bacterial colonies bearing mutations at the gene coding for the Rif target, RNA polymerase (8, 12). XlAID expression in this system resulted in a small but significant increase in the generation of Rif-resistant mutants (
3-fold) compared with bacteria transformed by vector alone, although significantly lower than HsAID-expressing transformants (Fig. 2, A and B), despite comparable protein expression levels (not shown). TrAID was also capable of inducing a mutator phenotype in bacterial cells, 4-fold higher than background (Fig. 2, A and B), which was further increased, as reported by Conticello et al. (32) by incubation of bacterial cultures at room temperature (not shown).
|
CSR activity of AID orthologs and fusion proteins in mammalian cells
To test for CSR activity of the various AID orthologs and fusions, cDNAs were expressed in LPS and IL-4-stimulated B cells from AID-deficient mice (1) by transfection with pMIG retroviral vectors (34), which express bicistronic transcripts also encoding EGFP. Activated B cells were stained at day 5 of culture (day 4 posttransduction) for expression of IgG1, and analyzed by flow cytometry. As expected, HsAID successfully complemented AID deficiency, inducing class switching in 9–37% of B cells (Fig. 3A, Table I). Surprisingly, XlAID was almost as active as the human ortholog, with IgG1+ cells ranging from 6 to 20% (Fig. 3A, Table I). In contrast, transduction of AID-deficient mouse B cells with retroviral pMIG-TrAID vectors did not complement the CSR deficiency above background levels (Fig. 3A, Table I). To test whether our switch assays reflect true quantitative differences in AID activity, we measured the percentage of IgG1+ cells among day 5 pMIG-AID-transduced lymphoblasts expressing low, medium, or high level of retrovirus (estimated based on GFP mean fluorescence intensity) (Table II). This analysis shows that the rate of IgG1 switching in pMIG-HsAID-expressing cells is proportional to the per-cell level of retroviral expression, indicating that the amount of AID in our system is limiting for CSR, and therefore that the assays are approximately quantitative.
|
|
|
60% of wild-type) (Fig. 3C, Table I). In contrast, the reciprocal TrAID/HsNES fusion, with the human NES, could not induce significant CSR activity, because modest CSR activity was detectable in only one of three experiments (Fig. 3C, Table I). Taken together, these data suggest that elements critical for AID activity in this system map to the HsAID NCD upstream of the NES, but are absent in the equivalent region of TrAID. Because an intact NES region is necessary for CSR induction by mammalian AID (17, 18) and, as shown above, the Takifugu AID NES can functionally replace its human counterpart in CSR, we considered whether it also shares with the mammalian NES the ability to mediate AID export from the nucleus via a CRM1-dependent mechanism (20, 21, 22). The NES region of AID from mammals, chicken, Xenopus, Takifugu and other teleosts (catfish and zebrafish) shows significant evolutionary conservation, particularly of the hydrophobic residues critical for nuclear export function (Fig. 4A). To test whether the Takifugu NES is indeed functional, we analyzed the ability of TrAID to be actively exported from the nucleus, as already shown for mammalian AID (20, 21). Like HsAID-EGFP fusion proteins, TrAID-EGFP displayed a predominantly cytoplasmic localization in transfectant NIH 3T3 cells, but accumulated in the nucleus upon treatment with the CRM1 inhibitor leptomycin B (Fig. 4B). We conclude that TrAID is actively exported from the nucleus via a CRM1-dependent mechanism.
|
The inability of TrAID to initiate CSR in mammalian cells could in principle be due to a lack of catalytic activity in mammalian cells, or to more specific differences in target recognition or regulation. We compared the ability of AID orthologs and fusions to induce SHM using a GFP gene stop-codon reversion assay in 293 cells (36). As shown in Fig. 5, all AID orthologs and fusions can induce detectable increases in GFP-stop reversion rates in this system, indicating that all proteins are catalytically active in SHM.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
AID’s phylogenetic history must have involved several critical functional shifts. One is at its emergence as an enzyme for Ab variable region diversification, which must have occurred basally in jawed vertebrate evolution. Ability to induce CSR arose later, after divergence of the bony fish and tetrapod lineages. Third, a GC-related activity may have arisen one or more times, depending on whether the known GC mechanisms in birds and mammals are homologous. These new functions could have been acquired by changes involving the AID protein itself, its ability to interact with regulatory cofactors, and/or by modifications in pathways downstream of AID activity (e.g., leading to differential repair of AID-induced DNA lesions). The latter may be the case for GC in birds, because recent evidence indicates that the mechanism is dependent on the differential resolution of AID-induced mutations in the primary VH segment in the presence or absence of GC donor 
V sequences (7). The same has been proposed for CSR, based on the observation that S region mutations accompany class switching in vitro (41). In contrast, the finding that separate regions within AID are required for CSR and SHM, possibly via association with distinct cofactors, supports the hypothesis that specific changes in the AID protein itself may have been involved in CSR emergence (17, 18, 19).
As this paper was in preparation, two groups have reported that zebrafish (Danio rerio) AID is capable of mediating normal CSR and SHM in mammalian cells (42, 43), whereas AID from both Takifugu (42) and catfish (Ictalurus punctatus) (43) have limited CSR and SHM activity. These data strongly argue that AID evolved the ability to mediate CSR before emergence of the class switch mechanism, although the molecular basis for the observed differences between fish species was not addressed. Together with data indicating that Xenopus S regions are recognized as switch targets by mammalian AID despite significant sequence divergence (16), the observation of zebrafish AID CSR activity and our finding that XlAID is capable of replacing mammalian AID to initiate CSR in activated B lymphocytes indicate a strict conservation of the entire CSR mechanism over many hundreds of millions of years and across divergent vertebrate lineages.
The inability of Takifugu AID to complement AID-deficient mouse B cells in our experiments is also in substantial agreement with the recent reports, although in our hands the deficiency is more pronounced than what was observed by Barreto et al. (42) with Takifugu, and by Wakae et al. (43) with catfish proteins. Although we cannot presently pinpoint the cause of this difference, we note that our retroviral AID expression system is not at functional saturation, as shown by the linear correlation between GFP expression and switch efficiency in transduced B cells (Table II), and may therefore be more sensitive to variations of activity. Together, however, the data presented here and from other labs (41, 42) paint a consistent picture of significant variation in the ability of AID from various fish species to complement mammalian AID activity in SHM and CSR. Simple variation in expression levels and/or stability are unlikely to account for the functional differences, because comparable expression levels are found in mammalian cells for Takifugu and zebrafish AID (42), mouse and catfish AID (43), and human and Takifugu AID (our data; not shown).
From an evolutionary standpoint, zebrafish and catfish are more closely related to each other than to Takifugu (44), with tetrapods, which include Xenopus and mammals, as an outgroup. This suggests two alternative, equally parsimonious evolutionary scenarios, i.e., that zebrafish and tetrapods acquired their activity independently, or that Takifugu and catfish independently lost theirs. Regardless of whether the activity differences between AID orthologs are due to independent gain or loss of function, however, convergent evolution of this kind would seem more plausible based on quantitative/regulatory differences, e.g., ability to associate with cofactors, as opposed to structural changes affecting AID catalytic properties such as substrate specificity, but more detailed studies will be required to address this issue.
Our finding that replacement of the TrAID noncatalytic domain with the human counterpart restores activity in mammalian cells rules out that the difference between Takifugu and mammalian AID involves critical differences in the AID catalytic site, or other elements in the catalytic domain, such as the peculiar teleost-specific 8-residue insertion, or the absence of some N-terminal residues in TrAID compared with mammalian AID. The primary candidate for any novel element in the AID NCD that could have been evolutionarily acquired during CSR emergence is the C-terminal NES motif (17, 18). Although the only confirmed NES function is CRM1-mediated nuclear export (20, 21), this element has been proposed to also mediate interaction with CSR-specific cofactors, acting either at the targeting stage, or after the AID-induced C-to-U DNA deamination (17, 18). Importantly, our results clearly show that both CRM1-mediated nuclear export and any additional putative CSR-specific function of the AID NES must have actually existed before the onset of CSR.
This finding implies that AID nuclear export must also play a selectable CSR-unrelated, ancestral function in fish. Because it has been shown that nuclear export may limit AID’s activity as a generalized mutator (21), it is possible for instance that the AID NES originally evolved as a control mechanism for potential genotoxicity. The role of the NES in switching may represent therefore a form of molecular preadaptation, later exploited for CSR evolution. How could the NES evolve for one function, and later be co-opted and become necessary for CSR? The first possibility is that CRM1 itself, or a factor with the same binding specificity for the NES motif as CRM1, acts as a CSR cofactor. Alternatively, the NES role in CSR may be indirect, that is, CSR evolution may have required the act of nuclear export per se. Perhaps nuclear export allowed AID to target a new substrate by homing to a specific compartment, to undergo posttranslational modifications (e.g., phosphorylation by a cytoplasmic kinase), or to associate with cofactors in the cytoplasm. The latter would imply the existence of at least one yet-uncharacterized CSR-specific element within AID.
The ability of the HsAID NCD to confer high activity to the TrAID CD in mammalian and bacterial cells is also of interest. Although it is possible that the effect is entirely mediated by structural/folding properties of the protein, it would seem rather peculiar for a trans-specific fusion protein to be structurally more favorable than either wild-type forms. A more interesting possibility is that the effect relates to specific regulatory function(s) of the NCD over AID activity, either by restricting substrate access, or by regulating catalytic specificity. Thus, any more "relaxed" protein configuration may increase overall catalytic activity, in agreement with observations on the catalytic effect of both C-terminal truncations (17, 18) and NES/NCD swaps (this paper) and with the proposed AID model (38). This would raise the intriguing possibility that AID was evolutionarily under strong selection for catalytic suboptimality, again, as discussed above, because of its potential mutator role.
Further studies, combining more extensive functional phylogenetic analysis and structural modeling will be required to highlight these and other issue surrounding AID’s complex evolutionary history and functional specialization with respect to the origin of class switching and somatic hypermutation.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported in part by National Institutes of Health (NIH) Grants AI45012 and CA107355 (to A.B.), DK43739 and AI54369Z (to H.C.S.), RR15934 and GM63162 (to J.E.W.), and AI059830 (to J.R.); Air Force Office of Scientific Research Grant F49620 (to H.C.S. and M.P.S.), and Howard Hughes Medical Institute Pilot Project Grant 76200-537802 (to J.E.W.). H.T.I. and S.H.M. were supported by NIH Grants T32 HL07152 and R25 GM64133, respectively. 
2 XlAID and TrAID cDNA sequences have been deposited to the GenBank database under accession numbers AY621657 (Takifugu rubripes) and AY621658 (Xenopus laevis).
3 H.T.I. and M.P.S. contributed equally to this work.
4 Current address: Genzyme Europe Research, 310 Cambridge Science Park, Milton Road, Cambridge CB4 0WG, U.K.
5 Address correspondence and reprint requests to Dr. Andrea Bottaro or Dr. Harold C. Smith, University of Rochester School of Medicine and Dentistry, Rochester, NY 14642. E-mail addresses: andrea_bottaro{at}urmc.rochester.edu and harold_smith{at}urmc.rochester.edu
6 Abbreviations used in this paper: SHM, somatic hypermutation; AID, activation-induced deaminase; CSR, class switch recombination; APOBEC-1, apolipoprotein B mRNA-editing catalytic polypeptide 1; NES, nuclear export signal; CRM1, chromosome region maintenance 1; XlAID, Xenopus laevis AID; TrAID, Takifugu rubripes AID; HsAID, Homo sapiens AID; EGFP, enhanced GFP; CD, catalytic domain; NCD, noncatalytic domain; GC, gene conversion.
Received for publication February 16, 2006. Accepted for publication April 21, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Roy, K. Yu, and M. R. Lieber Mechanism of R-Loop Formation at Immunoglobulin Class Switch Sequences Mol. Cell. Biol., January 1, 2008; 28(1): 50 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Marr, H. Morales, A. Bottaro, M. Cooper, M. Flajnik, and J. Robert Localization and Differential Expression of Activation-Induced Cytidine Deaminase in the Amphibian Xenopus upon Antigen Stimulation and during Early Development J. Immunol., November 15, 2007; 179(10): 6783 - 6789. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F.-T. Huang, K. Yu, B. B. Balter, E. Selsing, Z. Oruc, A. A. Khamlichi, C.-L. Hsieh, and M. R. Lieber Sequence Dependence of Chromosomal R-Loops at the Immunoglobulin Heavy-Chain S{micro} Class Switch Region Mol. Cell. Biol., August 15, 2007; 27(16): 5921 - 5932. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
