| Nucleic Acids Research | Pages |
Drosophila and human RecQ5 exist in different isoforms generated by alternative splicing
Introduction
Materials And Methods
GenBank accession numbers
DNA clones and sequences
Sequence analysis
PCR and 3[prime] RACE
Radiation hybrid mapping
Northern blot hybridization
Immunolocalization of Drosophila RECQ5
Results
RecQ5 is conserved between Drosophila and humans
Drosophila RECQ5 isoforms generated by alternative splicing
Human RECQ5 isoforms generated by alternative splicing
RECQ5 is expressed in many tissues
Drosophila RECQ5 localizes to the nucleus
Discussion
Acknowledgements
References
Drosophila and human RecQ5 exist in different isoforms generated by alternative splicing
Received March 23, 1999; Revised and Accepted July 26, 1999
ABSTRACT Members of the RecQ helicase superfamily have been implicated in DNA repair, recombination and replication. Although the genome of the budding yeast Saccharomyces cerevisiae encodes only a single member of this family, there are at least five human RecQ-related genes: RecQL, BLM, WRN, RecQ4 and RecQ5. Mutations in at least three of these are associated with diseases involving a predisposition to malignancies and a cellular phenotype that includes increased chromosome instability. Metazoan RecQ helicases are defined by a core region with characteristic helicase motifs and sequence similarity to Escherichia coli RecQ protein. This core region is typically flanked by extensive, highly charged regions, of largely unknown function. The recently reported human RecQ5, however, has only the core RecQ-homologous region. We describe here the identification of the Drosophila RecQ5 gene. We recovered cDNAs corresponding to three alternative splice forms of the RecQ5 transcript. Two of these generate nearly identical 54 kDa proteins that, like human RecQ5, consist of the helicase core only. The third splice variant encodes a 121 kDa isoform that, like other family members, has a C-terminal extension rich in charged residues. A combination of RACE and cDNA analysis of human RECQ5 demonstrates extensive alternative splicing for this gene also, including some forms lacking helicase motifs and other conserved regions.
INTRODUCTION
Genome instability is a significant factor in the development of many cancers. Mutations in genes that function in DNA metabolism frequently result in genome instability and, in humans, cancer-associated syndromes (reviewed in 1). One such class of genes is that encoding proteins related to Escherichia coli RecQ helicase. Three of the five known human members of this class have been shown to be associated with hereditary diseases that include predisposition to cancers. Mutations in BLM result in Bloom syndrome (BS), mutations in WRN result in Werner syndrome (WS), and mutations in RECQ4 result in Rothmund-Thomson syndrome (RTS) (2-4). Each of these syndromes is a rare, autosomal recessive disorder with unique clinical features. BS is characterized by dwarfism and immune deficiency (reviewed in 5); WS is characterized by the premature onset of a number of conditions associated with aging (reviewed in 6,7); and RTS is characterized by skin and skeleton abnormalities, but also includes features associated with premature aging (reviewed in 8). Each of these syndromes involves a greatly increased risk for many types of malignancy, and each exhibits a cellular phenotype that includes genome instability. In BS cells, genome instability is manifested as the presence of chromosome aberrations and increased levels of exchange between homologous chromosomes and between sister chromatids (9,10). Chromosome instability in WS cells is revealed as an increased rate of chromosome translocations and deletions (11,12). RTS patients have been observed to acquire somatic mosaicism for chromosome rearrangements and changes in chromosome number (13,14).
Insights into the cellular functions of RecQ helicases have come from studies in model organisms, especially microorganisms. The protein family is named for E.coli RecQ, a 3[prime] to 5[prime] DNA helicase that is required either to promote or to inhibit recombination, depending upon the pathway (15-17). A single RecQ helicase-encoding gene has been found in each of the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe (18-20). Mutations in S.cerevisiae SGS1 result in hyper-recombination and defects in both mitotic and meiotic chromosome segregation (18,21). Sgs1p interacts with topoisomerases II and III in a two-hybrid assay, and sgs1 mutations suppress the slow growth phenotype of top3 mutations (22). These phenotypes suggest that Sgs1p is involved in both replication and recombination.
One issue that studies in microorganisms cannot address is the multiplication and divergence of the RecQ family in animals. In addition to the BLM, WRN and RECQ4 loci described above, two other human genes encoding RecQ-related proteins, RECQL and RECQ5, have been reported (23-25). The genome of the nematode Caenorhabditis elegans encodes at least four RecQ-related proteins, and it is possible to define orthologous relationships between these proteins and human WRN, BLM, RecQL and RecQ5, based on sequence similarities within the region conserved among family members. Thus, it appears that this gene family underwent substantial divergence early in metazoan evolution, with different family members presumably acquiring different functions (26). Genetic characterization of RecQ-related genes in model metazoans is likely to provide important new insights into these functions.
We report here the identification of the Drosophila melanogaster RecQ5 gene. Characterization of cDNAs encoding Drosophila RECQ5 show that alternative splicing produces drastically different isoforms. We also found evidence that human RECQ5 is alternatively spliced to produce different isoforms.
MATERIALS AND METHODS
GenBank accession numbers
AF134239, Drosophila RecQ5 (genomic sequence); AF135183, human RECQ5 (for transcripts A and B).
DNA clones and sequences
We screened 105 clones from a 0-4 h Drosophila embryo cDNA library (27) and recovered 25 RecQ5 cDNAs. We initially characterized four of these further. Two contained the entire coding region and two were truncated at the 5[prime] end. Comparison to genomic sequence showed that three of these had the larger (327 nt) intron removed, and therefore encode the 121 kDa isoform. One of the truncated cDNAs had the smaller intron (75 nt) removed, and therefore encodes the 54 kDa isoform. We analyzed the remaining cDNAs by PCR and found that 19 had the larger intron removed and two had the smaller intron removed. Two RecQ5 EST sequences were subsequently deposited into GenBank (28). One of these, clone LD21474 (GenBank accession no. AA735537), had the larger intron removed. The other, clone GH01404 (GenBank accession no. AI062257), had no intron removed at the alternative splice site, although it carries a polyadenylation tract at the 3[prime] end and other introns were removed.
The original human RECQ5 cDNA in the EST database is I.M.A.G.E. Consortium CloneID 590051 (GenBank accession nos AA155835 and A155882) (29). Using this insert as a probe, we screened 5 × 105 plaques from each of two human brain cDNA libraries and obtained a single positive clone. The sequence of this cDNA and our RACE sequences has been deposited into GenBank under accession number AF135183.
During the course of this work, several I.M.A.G.E. Consortium cDNA clone EST sequences were deposited into GenBank. We obtained and characterized the cDNAs that were available (Table 1 and Fig. 3). CloneID 2191126 carries a 1292 bp insert. The library from which this clone is derived has inserts cloned directionally into the SalI (5[prime]) and NotI (3[prime]) sites of pCMV-SPORT6. The SalI end appears to be intact. However, 162 bp near this end matches the end of a 5[prime] EST read from CloneID 380349, but in the opposite orientation. The last 512 bp of the insert match RECQ5 sequence in a 3[prime] to 5[prime] direction. There is no poly(A) tract at this end, and there is a fragment of rearranged polylinker sequence between the NotI site and the insert. We conclude that this clone carries a cDNA fragment with its 5[prime] end toward the NotI site and its 3[prime] end (which does not include the terminus) at the SalI end.
Table 1. Human RECQ5 cDNAs
| GenBank | I.M.A.G.E. | Library tissue source | Classa | Insertb |
| AF135183 | - | brain | A1 | 1-1356; 1823-2178 |
| AA155835 | 590051 | umbilical vein endothelium | A1 | 985-1356; 1823-2178 |
| AI686448 | 2267876 | serous papillary carcinoma | A1 or B | <1928-2176 |
| AI684041 | 2267435 | serous papillary carcinoma | A2 or B | <1961-2397 |
| AI537883 | 2189520 | serous papillary carcinoma | A3 | 1-1356; 1823-2828 |
| N53539 | 284217 | multiple sclerosis lesions | A3 or B | 2198-2828 |
| AI218469 | 1845920 | mixed fetal lung, testis, B-cell | C1 | (500)-1113; +332 bp |
| AI363275 | 2016056 | glioblastoma | C2 | 311-1113; +396 bp |
| AI310229 | 1914385 | kidney clear cell tumor | F1 | 278-1034; +191 bp |
| AI671940 | 2316345 | kidney | F2 | <866-1225; +47 bp |
| AI567000 | 2191126 | serous papillary carcinoma | F3 | 846-1356; +780 bp |
bCoordinates are relative to GenBank accession number AF135183. Numbers preceded by + denote sequences (of the indicated length) not present in AF135183. Numbers in parentheses are approximations. Some cDNAs were unavailable for analysis; these are marked with < to indicate that the insert continues in the 5[prime] direction. Coordinates for these cDNAs are from the sequence in the corresponding GenBank record.
Sequence analysis
Sequence comparisons were performed with the Wisconsin Package, v.9.1 (Genetics Computing Group). Multiple alignments of the RecQ conserved amino acid sequences (corresponding to human RecQ5 residues 1 through 406) were generated with the PILEUP program. Protein sequence distances were determined with the DISTANCES program (Kimura scoring method), and a phylogenetic tree was constructed with the GROWTREE program (neighbor-joining method). Other methods gave slightly different trees, especially with respect to the placement of human RecQ4, but all methods generated essentially similar orthologous relationships between the metazoan sequences analyzed.
Searches for nuclear localization signals were done on the PSORT server at http://psort.nibb.ac.jp
PCR and 3[prime] RACE
Human RECQ5 introns were confirmed by PCR from genomic DNA (Promega), used at a concentration of 1 ng/µl. Primers used were Q12 (GCGCTTTGGCTTCTCTGA) and Q205a (CAAACCCAAAGCCTTCTT), which amplify a 666 bp fragment, including a 472 bp intron at 113; Q974 (GGGTGTGAACGCCAAGGCTT) and Q1061a (CAGGGACCTTCTC-CTCCAT), which amplify a 626 bp fragment that includes a 538 bp intron at 1011; Q1090 (GGGAGTGGATAAA-GCCAAT) and QC326 (GGGCACAGCACTAGGCAAT), which amplify a 90 bp fragment across the 5[prime] boundary of the intron between motifs V and VI; Q1318 (GGCCTTTGATGCCCTGGTGACC) and Q2081a (CCTCGATCTACCATGAGCTT), which amplify a 765 bp fragment that spans the 467 nt intron position; Q1773 (GGGGAGTCATGTGCTTTGAA) and Q2111a (CTGAAAATCAGGAGACGGG), which amplify a 339 bp fragment; and Q2458 (GGCCAAGTGTTCCTGTTCAT) and Q3157a (GCCCCGCCTCATTAGTTA), which amplify a 700 bp fragment.
To determine whether there are alternative 3[prime] ends on RECQ5 transcripts, we used two gene-specific primers from the 1.7 kb cDNA: Q1173 (GGGCTGGCAGGATGGGAAGCCTT) and Q1318. Amplification was done with the Marathon kit (Clontech), on human placenta second-strand cDNA ligated to adapters, according to the instructions. We obtained bands of 2.7 and 1.7 kb with primer 1173, and 2.5 and 1.5 kb with primer 1318. Several smaller bands were also detected by probing a blot of the RACE products, but these were not characterized. Bands were gel-isolated and reamplified with the same or internal primers for sequencing.
Radiation hybrid mapping
Radiation hybrid mapping was performed using the Stanford G3 Radiation Hybrid Panel (20) obtained from Research Genetics, according to the instructions. Primers used were 1852 (CTGAGGGCTGCTTGGTGTAGTCAGGTT; 3[prime] end of coding) and 2081a (above; 3[prime] UTR). Raw scoring was submitted to the RH server at http://www-shgc.stanford.edu/RH/index.html. The result returned indicated linkage to AFMb054zf9 on chromosome 17, with a LOD score of 21.7, and a distance of 2.17 cR10000.
Northern blot hybridization
A blot of human mRNA was obtained from OriGene Technologies (Rockville, MD). The 229 bp PCR amplimer from the radiation hybrid mapping was 32P-labeled using the Ready-to-Go DNA labeling kit (Pharmacia), and the blot probed and washed according to the manufacturer's instructions. We also probed a similar blot (Clontech) containing mRNA from leukocytes, colon, small intestine, ovary, testis, prostate, thymus and spleen, and obtained similar results to those in Figure 4 (data not shown).
Immunolocalization of Drosophila RECQ5
The protein-coding region of a RecQ5 cDNA encoding the 121 kDa isoform was fused in-frame downstream of a segment encoding the FLAG epitope DYKDDDDK. This was placed under the control of the Ubiquitin promotor through the initiating methionine codon. 5 × 106 Schneider S2 cells were transiently transfected with 5 µg DNA and 10 µl Superfect (Qiagen). Two days after transfection, these cells were stained with a 1:1000 dilution of anti-FLAG antibody M5 (Sigma) and 0.5 µg/ml DAPI.
RESULTS
RecQ5 is conserved between Drosophila and humans
We discovered a D.melanogaster gene encoding a novel member of the RecQ helicase superfamily while cloning an unrelated cell cycle checkpoint gene. Database searches using the deduced protein sequence revealed putative orthologues in C.elegans (E03A3.2) and humans. A sequence of the human ortholog was recently published with the name RECQ5 (25), so we designate the Drosophila gene RecQ5. The predicted human RecQ5 and Drosophila RECQ5 proteins are 49% identical to one another in the RecQ-conserved region. Sequence comparisons between family members using this region place human RecQ5, Drosophila RECQ5 and C.elegans E03A3.2 onto a distinct branch (Fig. 1), suggesting that these genes share a conserved function in DNA metabolism.
Figure 1. The eukaryotic RecQ helicase family. (A) Dendrogram depicting relationships between RecQ helicase family members from E.coli (Ec), S.cerevisiae (Sc), S.pombe (Sp), C.elegans (Ce) and humans (Hs), along with Drosophila RECQ5. The relationship of human RecQ4 to the other sequences is unclear (Materials and Methods), but the other metazoan proteins fall into four branches. Another Drosophila RecQ family member, DmBLM, the ortholog of human BLM, was recently reported (26), but is not indicated here. (B) Schematics of the corresponding protein structures. The shaded region represents the RecQ helicase-related conserved region that was used in sequence comparisons. Black boxes indicate the seven conserved helicase motifs. Arrows in Drosophila and human RecQ5 indicate termination sites of alternative isoforms (see Figs 2 and 3).
We mapped the Drosophila RecQ5 gene to interval 70E1-4 by in situ hybridization to polytene chromosomes. We determined the genomic position of human RECQ5 by radiation hybrid mapping using the Stanford G3 panel (Materials and Methods). Our results place the gene in 17q23-25 near the marker AFMb054zf9.
Drosophila RECQ5 isoforms generated by alternative splicing
Previously described members of the eukaryotic RecQ helicase family range in size from 75 kDa for human RecQL to 162 kDa for human WRN. In most family members, the highly conserved helicase region is flanked by extensive sequences that show similarity to one another primarily in amino acid composition, being rich in charged residues. The reported human RECQ5 sequence encodes a 46 kDa protein that consists of the core region conserved among RecQ family members. Analysis of Drosophila RecQ5 cDNAs reveals predicted isoforms of 54 and 121 kDa (Fig. 2). The smaller isoforms are very similar in structure to human RecQ5, consisting entirely of the region encompassing the seven motifs common to known helicases (30), together with flanking sequences conserved in all RecQ family members. The larger isoform has in addition an extensive C-terminal region abundant in charged residues. Among the 588 residues in this region, 98 (16.7%) are acidic, and 126 (21.4%) are basic. The charged residues are primarily glutamic acid (10.9%) and lysine (14.1%). No sequence similarity is apparent between this region and other RecQ family members, other than in composition.
Figure 2. Alternative splicing at Drosophila RecQ5. (A) Schematic of three alternative RecQ5 transcripts. Each box represents one exon. Protein-coding regions are shaded; black boxes indicate the seven helicase motifs. Transcript A encodes a 121 kDa isoform (RECQ5a); transcripts B and C encode nearly identical 54 kDa isoforms (RECQ5b and RECQ5c). Positions of potential nuclear localization signals (KRPKK, HRRKR and PYYKRKI) are indicated with arrows. (B) DNA sequence corresponding to the region of the alternative splice (indicated by a bracket in 3a), determined by sequencing cDNAs and genomic DNA. The sequence begins near the end of the common second exon. The sequence corresponding to the larger version of the intron is in lower case. Predicted amino acid sequences are shown, with asterisks indicating stops.
Figure 3. Transcripts and partial genomic structure of human RECQ5. (A) The structures of representative cDNAs carrying RECQ5 sequences are shown. Protein-coding regions are shaded; black boxes indicate the seven helicase motifs. Known intron positions are indicated, but other intron positions were not determined. The third intron depicted is ~2.1 kb. The fourth intron depicted is the one not removed in class B transcripts. The uppermost schematic represents our 1710 bp cDNA, and others represent I.M.A.G.E. Consortium cDNAs (Table 1). A vertical line indicates that the cDNA was probably incomplete at that end. A sigmoid line indicates that the cDNA continues for an undetermined distance (these clones were not yet available for analysis). The lower three cDNAs appear to be fusions between RECQ5 at and the 3[prime] end of another gene (hatched boxes).
Figure 4. Expression of human RECQ5. A blot containing human mRNA (OriGene Technologies) was probed for RECQ5 using sequences near the 3[prime] end of transcript A. The probe used should detect transcripts A and B, but not C and D. RNA size markers are indicated to the left. The sources of mRNA are indicated above each lane.
The different Drosophila RECQ5 isoforms are encoded by transcripts that differ by alternative splicing (Fig. 2). The region encoding the helicase motifs is contained entirely on the second exon, which can be joined to either of two alternative third exon start sites. Exon 3a, following removal of a 327 nt intron, maintains an open reading frame into subsequent exons, resulting in a 121 kDa isoform (RECQ5a). Exon 3b, following removal of a 75 nt intron, has an immediate stop codon at the 5[prime] end, resulting in the truncated 54 kDa isoform (RECQ5b). We also found a cDNA in which no intron was removed at this position. Other introns had been removed and there was a polyadenylation tract at the 3[prime] end, indicating that this cDNA is derived from a processed transcript. This cDNA encodes a 54 kDa isoform (RECQ5c) almost identical to RECQ5b.
Human RECQ5 isoforms generated by alternative splicing
We wanted to know whether alternative splicing is a feature conserved in human RECQ5. Human RECQ5 was originally defined by EST sequences from a single cDNA (Table 1, line 2). We obtained and sequenced this cDNA and found it to be incomplete at the 5[prime] end. Using this clone as a probe, we screened human cDNA libraries, and recovered one RECQ5 clone from a brain cDNA library. The 1710 bp sequence of this cDNA is identical to the sequence of the EST clone, except that it is longer at the 5[prime] end and it contains a complete open reading frame. Both cDNAs have a polyadenylation tract 275 bp 3[prime] to the termination codon. The corresponding mRNA encodes a 49 kDa polypeptide, slightly longer than the 46 kDa polypeptide predicted by the published RECQ5 coding sequence (25). Comparison of the published nucleotide sequence to our sequences showed that they are identical except at the extreme 3[prime] end of the open reading frame, at which point the published sequence has a stop codon not found in our sequences.
To investigate the differences between our cDNAs and the published sequence, we performed 3[prime] RACE on human placenta mRNA, using two gene-specific primers to sequences near the end of the protein-coding region (Materials and Methods). With each primer we amplified a product of the size predicted by our cDNA sequences (data not shown). In addition, we obtained two larger products for each primer, which we refer to as RACE products 1 and 2.
Sequencing of these larger RACE products showed that each contains an insertion of 467 bp near the end of the coding region (Fig. 3). The insertion begins with a GT and ends with an AG, so we considered that it might correspond to an intron, and that RACE products 1 and 2 represent amplification of transcripts in which this intron had not been removed. We designed primer pairs that span either the 5[prime] or 3[prime] boundary of the insertion, and that span the entire insertion (Materials and Methods). When we used these primers in PCR reactions using human genomic DNA as a template, products of the predicted sizes were specifically amplified (data not shown), confirming that this insertion is indeed derived from an unspliced intron.
One possible explanation is that RACE products 1 and 2 were from unprocessed primary transcripts. Several observations suggest that this is not the case. First, the template in the RACE reactions was double-stranded cDNA made from polyadenylated mRNA, and both products carry a polyadenylation tract at their 3[prime] ends. Second, the 3[prime] end of product 1 is identical to that of two 3[prime] EST sequences in the database (see below). Third, the 5[prime] end of the unspliced intron has an in-frame stop codon, resulting in a sequence that is identical to the published RECQ5 protein-coding sequence (25). We conclude that this 467 nt intron is removed from some transcripts but not from others. As is the case in Drosophila, failure to remove this intron results in a truncation of the polypeptide, in this case removing 25 residues to yield a 46 kDa isoform.
RACE products 1 and 2 also differ at their 3[prime] ends from our cDNA. Both RACE products contain the entire 3[prime] end of transcript A, followed by additional sequences. Product 1 has an additional 1118 bp, and product 2 has this same sequence followed by an additional 718 bp, ending in an Alu sequence. There is a putative polyadenylation signal near the 3[prime] end of the Product 1 sequence, and database searches revealed two ESTs from cDNAs with identical 3[prime] ends (Table 1 and Fig. 3), confirming that this is a bona fide polyadenylation site.
During the course of this work, additional human RECQ5 EST sequences were deposited into GenBank (Table 1). Analysis of these cDNAs reveals additional polyadenylation sites, splicing patterns and predicted protein isoforms (Fig. 3). We have designated each transcript class according to predicted protein product (distinguished by different letters) and polyadenylation site (distinguished by different numbers). Transcripts in class A encode the 49 kDa isoform. There are at least three different polyadenylation sites for class A transcripts. Transcripts in class B encode the 46 kDa isoform, due to failure to remove the 467 bp intron. This class is presently represented only by our RACE products and the coding sequence reported by Kitao et al. (25).
A third class, designated C, is represented by I.M.A.G.E. Consortium CloneIDs 2016056 and 1845920. The inserts in these cDNAs match our RECQ5 cDNA sequence from their 5[prime] ends to bp 1113 (of our sequence). At this junction, novel sequence begins, extending 332 and 396 bp, respectively, to a polyadenylation tract. The novel sequence begins with GT, suggesting that these cDNAs might correspond to transcripts in which termination and polyadenylation occurred within an intron. PCR from human genomic DNA using primer pairs spanning the junction point yielded the fragment predicted from these two EST sequences, indicating that the sequences across the junction point are contiguous in the genome. Hence, there is an intron of ~2 kb between the regions encoding helicase motifs V and VI; transcripts in class C terminate within this intron. These transcripts encode a 36 kDa protein that lacks motif VI as well as additional sequences conserved among RecQ family members.
Transcript class D is represented by I.M.A.G.E. Consortium CloneIDs 1914385, 2316345 and 2191126 (Table 1 and Fig. 3). These transcripts appear to be hybrids between RECQ5 and another gene. They each have RECQ5 sequences at their 5[prime] ends and another sequence at their 3[prime] ends. CloneID 1914385 has RECQ5 sequence through 1034 (of RECQ5), followed by 191 bp of sequence that is identical to the 3[prime] ends of a group of 19 other EST sequences. These other EST sequence reads, which are up to 601 bp in length, are identical to one another along their entire lengths, and none of this sequence is from RECQ5, even in cases where 5[prime] EST reads are available. Thus, these cDNAs are probably derived from an independent gene, which we will refer to as gene D.
CloneID 2316345 is similar to 1914385, except that the junction point is at 1229 in RECQ5, and there are only 51 bp of the 3[prime] end of gene D. The insert in 2191126 also appears to be a fusion between RECQ5 and gene D. This insert is in a rearranged orientation (Materials and Methods) and is lacking a polyadenylation tract at the 3[prime] end, so it likely represents an internal fragment of a cDNA. The 3[prime]-most sequences from this insert overlap the 5[prime] end of CloneID 380349, whose 3[prime] end identifies it as a cDNA from gene D. Thus, 2191126 appears to identify a fusion point more 5[prime] in gene D.
RECQ5 is expressed in many tissues
To determine the relative abundances and tissue distribution of the different RECQ5 transcripts, we probed blots of human mRNA. We detected a 1.7 kilonucleotide (knt) transcript at a low level in all tissues examined (Fig. 4). If our RACE products correspond to transcripts identical at the 5[prime] end to the 1710 bp cDNA we sequenced, as is suggested by the published RECQ5 sequence, these transcripts would be 2828 and 3646 nt (excluding polyadenylation tracts). However, we were unable to detect these larger transcripts using either probes within the region in common to all transcripts or probes specific to the larger transcripts. This was true even for placenta mRNA, which was the source of our RACE products. In contrast, Kitao et al. (25) detected transcript sizes of 3.6 and 3.8 knt for RECQ5, also present in all tissues examined. However, these authors did not report any signals corresponding to smaller transcripts.
Drosophila RECQ5 localizes to the nucleus
Saccharomyces cerevisiae Sgs1p and human WRN are both found to be concentrated in the nucleolus (31,32). To determine whether Drosophila RECQ5 shows a similar pattern of sub-nuclear localization, we transiently expressed the 121 kDa isoform, tagged with an N-terminal FLAG epitope, in Drosophila Schneider2 cells. When cells were stained with anti-FLAG antibody to detect FLAG-RECQ5, we detected localization to the nucleus, but no sub-nuclear localization was apparent (Fig. 5).
Figure 5. Immunolocalization of Drosophila RECQ5 to the nucleus. (A) DAPI staining of a field of cells transfected with the Ubiq::FLAG-RecQ5 construct. Four nuclei can be seen in their entirety. (B) The same field stained with anti-FLAG antibody. Two cells can be seen to express FLAG-RecQ5. In both cells the protein is strongly localized to the nucleus.
DISCUSSION
We have described a new D.melanogaster RecQ helicase family member that shares a high degree of amino acid identity with human RecQ5. A unique feature of the Drosophila gene is the existence of isoforms that differ by the presence or absence of a highly charged C-terminus. Most other eukaryotic family members have regions of a similar size on one or both sides of the helicase region, but the functions of these regions are largely unknown. Perhaps the best understood case is WRN, which has a recognizable exonuclease motif at its N-terminus (33). This region has been demonstrated to possess a 3[prime] to 5[prime] exonuclease activity in vitro, and this activity is separable from the helicase activity (34). The region N-terminal to the helicase region of Sgs1p has been shown to interact with the C-terminal portion of topoisomerase II in a yeast two-hybrid assay (18). It is possible that the highly charged, C-terminal region specific to the 121 kDa isoform of Drosophila RECQ5 interacts with other proteins, and that the difference between the isoforms is a way of regulating these interactions. The existence of different Drosophila RecQ5 isoforms provides a unique opportunity to study the in vivo functions of these sequences outside the core RecQ-related region. Such dramatically different isoforms have not been reported for any other family members. It is interesting to note that most family members have a dibasic (RR) sequence at the C-terminus of the RecQ-conserved region, suggesting the possibility of proteolytic processing at this location.
Alternative splicing also generates different isoforms of human RecQ5. In one case, the isoform produced depends on whether or not an intron is removed, as was observed for the Drosophila gene. The size difference between the 49 and 46 kDa RecQ5 isoforms, however, is much less than the difference between the Drosophila isoforms. These two human RecQ5 isoforms carry all of the sequences conserved among RecQ family members, and therefore both may have helicase activity. Nonetheless, it is possible that these different isoforms possess different biochemical activities or different capacities to interact with other cellular proteins that may regulate activity.
Another possibility is that the different C-termini regulate protein stability or sub-cellular localization. The gene encoding murine RecQL is alternatively spliced in the testis to remove a short region believed to contain a nuclear localization signal (35). The only recognizable nuclear localization signal in human RecQ5 is the sequence PERRVRS, which begins at residue 12. All predicted isoforms of Drosophila RECQ5 have at least one potential nuclear localization sequence (Fig. 2A, arrows). Consistent with this observation is our finding that at least the larger isoform localizes to the nucleus when expressed in a Drosophila cell line. It seems unlikely that the function of Drosophila RECQ5 alternative splicing is to regulate nuclear localization, but it is possible that this is the case for human RecQ5.
A third human RecQ5 isoform is produced by transcripts in class C, in which polyadenylation occurs 5[prime] to the sequences encoding the last helicase motif. This 36 kDa protein lacks both helicase motif VI and additional sequences conserved among RecQ helicases, and likely has no catalytic function. Transcripts in this class may be important for regulating RECQ5 expression.
Class D transcripts are somewhat perplexing. They appear to be fusions between RECQ5 and another gene. Such fusions could have occurred in the DNA (e.g. by chromosome rearrangement), the RNA (e.g. by aberrant splicing), or the cDNA (e.g. by cloning artifacts). The three fusion cDNAs came from different libraries, and they have different fusion points in both genes, so it seems unlikely that these represent cloning artifacts. It is possible that these transcripts result from aberrant splicing. The fusion point in 2191126 is at a RECQ5 exon boundary. However, the fusion point of 1914385 is not at an exon boundary. Either chromosome rearrangements or aberrant splicing could therefore account for these transcripts. Conceptual translation of available sequences from this second gene does not reveal any similarities to other proteins in the databases.
We used radiation hybrid mapping to map human RECQ5 to the tip of 17q, in agreement with the in situ hybridization mapping of Kitao et al. Loss of heterozygosity in this region has been associated with ovarian cancer and familial breast carcinomas (36-40). Given the phenotypes associated with mutations in other RecQ helicase family members, human RECQ5 may be a good candidate for a tumor suppressor gene.
We mapped Drosophila RecQ5 to polytene region 70E1-4. We were unable to identify probable candidates for mutations in RecQ5 among known mutations in the region. Efforts are now underway to generate mutations in this gene. Preliminary results suggest that Drosophila RecQ5 is not an essential gene. Analysis of these mutations will provide valuable insights into the function of this gene in Drosophila, and perhaps other organisms, and into the functional significance of the different isoforms.
ACKNOWLEDGEMENTS
We thank Ken Burtis and Sarah Wayson for helpful suggestions, Elaine Kwan for assistance with cell culture experiments, and Todd Laverty for the in situ hybridization to polytene chromosomes. J.J.S. was supported in part by a fellowship from the Cancer Research Fund of the Damon Runyon, Walter Winchell Foundation, DRG 1355.
REFERENCES
*To whom correspondence should be addressed at: Department of Biology, CB #3280, Fordham Hall, University of North Carolina, Chapel Hill, NC 27599-3280, USA. Tel: +1 919 843 9400; Fax: +1 919 962 8472; Email: sekelsky{at}email.unc.edu
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Z. Bachrati, R. H. Borts, and I. D. Hickson Mobile D-loops are a preferred substrate for the Bloom's syndrome helicase. Nucleic Acids Res., January 1, 2006; 34(8): 2269 - 2279. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Hu, X. Lu, E. Barnes, M. Yan, H. Lou, and G. Luo Recql5 and Blm RecQ DNA Helicases Have Nonredundant Roles in Suppressing Crossovers Mol. Cell. Biol., May 1, 2005; 25(9): 3431 - 3442. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Z. Ozsoy, H. M. Ragonese, and S. W. Matson Analysis of helicase activity and substrate specificity of Drosophila RECQ5 Nucleic Acids Res., March 1, 2003; 31(5): 1554 - 1564. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Cui, R. Klima, A. Ochem, D. Arosio, A. Falaschi, and A. Vindigni Characterization of the DNA-unwinding Activity of Human RECQ1, a Helicase Specifically Stimulated by Human Replication Protein A J. Biol. Chem., January 10, 2003; 278(3): 1424 - 1432. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Kawasaki, S. Maruyama, M. Nakayama, K. Matsumoto, and T. Shibata Drosophila melanogaster RECQ5/QE DNA helicase: stimulation by GTP binding Nucleic Acids Res., September 1, 2002; 30(17): 3682 - 3691. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Z. Ozsoy, J. J. Sekelsky, and S. W. Matson Biochemical characterization of the small isoform of Drosophila melanogaster RECQ5 helicase Nucleic Acids Res., July 15, 2001; 29(14): 2986 - 2993. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Emmert, T. D. Schneider, S. G. Khan, and K. H. Kraemer The human XPG gene: gene architecture, alternative splicing and single nucleotide polymorphisms Nucleic Acids Res., April 1, 2001; 29(7): 1443 - 1452. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Adams, S. E. Celniker, R. A. Holt, C. A. Evans, J. D. Gocayne, P. G. Amanatides, S. E. Scherer, P. W. Li, R. A. Hoskins, R. F. Galle, et al. The Genome Sequence of Drosophila melanogaster Science, March 24, 2000; 287(5461): 2185 - 2195. [Abstract] [Full Text]
|
|
|
|
 |
|
 G. M. Hansen, D. Skapura, and M. J. Justice Genetic Profile of Insertion Mutations in Mouse Leukemias and Lymphomas Genome Res., February 1, 2000; 10(2): 237 - 243. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||