A Werner syndrome protein homolog affects C. elegansdevelopment, growth rate, life span and sensitivity to DNA damage by acting at a DNA damage checkpoint

Werner syndrome (WS) is associated with a rapid acceleration of aging, and is caused by mutations in the RecQ family DNA helicase gene wrn(Yu et al., 1996). wrn encodes a protein with a central domain of seven helicase motifs and two conserved domains (RQC and HRDC) located C-terminal to the helicase domain (Morozov et al., 1997). WRN differs from other members of the RecQ family in that it possesses an unusual exonuclease domain homologous to the 3′→5′exonuclease domain of E. coli DNA polymerase I and of RNaseD(Moser et al., 1997; Huang et al., 1998; Shen et al., 1998; Suzuki et al., 1999). Nevertheless, when mutated, WRN causes genomic instability disorders associated with an elevated risk of cancer, short stature, and/or premature aging, like two other RecQ family DNA helicases in humans: the Bloom syndrome(BS) and the Rothmund-Thomson syndrome (RTS, RecQL4) proteins(Mohaghegh and Hickson,2001).

Several defects at the cellular level have been detected in WS. Cells cultured from WS patients have a reduced life span, an extended S phase, and reduced RNA transcription by RNA polymerases I and II(Martin et al., 1970; Salk et al., 1985; Balajee et al., 1999; Shiratori et al., 2002). Although no direct role of WRN has been established in telomere metabolism, WS fibroblasts expressing a transfected human telomerase (TERT) gene have an increased life span and can be immortalized(Wyllie et al., 2000). These results suggest that a telomerase-dependent pathway is involved in the accelerated cellular senescence caused by the absence of WRN. WS cells are also hypersensitive to certain DNA damaging agents, including the chemical carcinogen 4-NQO (Ogburn et al.,1997), camptothecin (Poot et al., 1999), and DNA cross-linking agents(Poot et al., 2001).

WRN has been shown to form complexes with proteins involved in cellular responses to DNA damage and in DNA replication. The identification of a functional interaction between WRN and the p53 tumor suppressor protein has emphasized the role of the RecQ family in maintaining genomic stability(Spillare et al., 1999). WRN also dramatically stimulates the cleavage reaction catalyzed by the human 5′ flap endonuclease/5′→3′ exonuclease FEN1(Brosh et al., 2001), a DNA structure-specific nuclease implicated in DNA replication and repair(Lieber, 1997). The ability of replication protein A (RP-A; RPA1 – Human Gene Nomenclature Database) to stimulate the unwinding of long stretches of DNA duplex by WRN helicase suggests that WRN may function in replication, a notion supported by interaction of WRN with other replication proteins. Recent evidence points to a direct protein interaction between WRN and the Ku80/70 heterodimer implicated in non-homologous end-joining of double-strand breaks(Cooper et al., 2000). In addition to these functional interactions, WRN has been reported to physically interact with human polymerase delta(Szekely et al., 2000), PCNA and DNA topoisomerase I (Label et al.,1999). These interactions suggest that WRN is a central player in a macromolecular complex essential for DNA replication or repair.

In C. elegans, four RecQ family proteins are predicted from the genomic DNA sequence. Of these proteins, the one encoded by the open reading frame (ORF) F18C5.2 is most homologous to human WRN. To understand the role of this C. elegans WRN homolog (WRN-1), we localized the protein in C. elegans and investigated the phenotypes arising from inhibited expression.

Bristol N2, as a standard wild-type strain, and div-1(or148ts)strains were obtained from the C. elegans Genetics Center (St Paul,MN, USA). An EST clone of the Ce-wrn-1 gene (yk41c3) was provided by Dr Y. Kohara (National Institute of Genetics, Japan). Deoxynucleotide oligomers were synthesized at Genotech (Korea).

The EST clone yk41c3 of the F18C5.2 ORF lacked the first exon (115 nucleotides; nt) of the predicted ORF of 16 exons. Therefore, to obtain a 5′-terminal cDNA clone, we isolated C. elegans total RNA using an RNeasy kit (Qiagen). cDNA synthesis progressed in a reaction mixture (50μl) containing C. elegans total RNA (3 μg), a primer (10 pmoles) of sequence 5′-GTGGACATAAGAACAAATTGGTC-3′ (nt 752-729 in the ORF) from exon 3, and Superscript reverse transcriptase II (200 units,Stratagene), at 42°C for 1 hour. First cDNA strand synthesis was terminated by heating at 70°C, and then template RNA was degraded by RNase H (2 units, Takara). A cDNA fragment was amplified from the first cDNA strand by PCR (polymerase chain reaction) using the SL1 primer(5′-GGTTTAATTACCCAAGTTTGAG-3′) and a primer of sequence 5′-CATTTCTGACAACATCCCACTG-3′ (nt 715-694 in the ORF) from exon 3. The amplified cDNA fragment was cloned into pGEM-T vector (Promega) and sequenced with an ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit(Perkin-Elmer).

To obtain a full-length cDNA construct of wrn-1, PCR was carried out using two primers:5′-CGCGGATCCATGATAAGTGATGATGACGATC-3′, containing nt 1-22 of the ORF, and a BamHI recognition sequence (underlined); and 5′-CATTTCTGACAACATCCCACTG-3′, corresponding to nt 715-694 of the ORF. The amplified DNA product was digested with BamHI and HindIII (nt 641-646 in the ORF) restriction enzymes, electrophoresed on a 1% agarose gel, and then eluted from the gel using a DNA extraction kit(Intron Biotechnology, Korea). The purified cDNA fragment was inserted into plasmid yk41c3, previously digested with BamHI and HindIII,to yield pCeWRN.

A 5′-terminal cDNA fragment was amplified from pCeWRN by PCR using primers: nt 1-22 of the ORF with a BamHI recognition sequence; and nt 627-609 with a SalI recognition sequence (underlined),5′-CAGCTGCGTCATTGATGCCCACTTC-3′. The cDNA fragment was cloned into pGEM-T, excised from the recombinant T-vector, and then inserted between the BamH1 and SalI sites of the pMAL-c2 overexpression vector (New England Biolabs). E. coli XL1-Blue cells harboring pMAL/CeWRN were cultured at 37°C to O.D._600nm 0.5,and isopropyl-thio-β-D-galactoside (Calbiochem) was added to 1 mM. After further incubation for 3 hours, the cells were harvested and sonicated in 10 ml PBS with 10 mM Na₂HPO₄, 2 mM KH₂PO₄ (pH 7.4), 137 mM NaCl and 2.7 mM KCl. Cell lysate was microcentrifuged at 14,000 rpm for 5 minutes, and the supernatant electrophoresed on a 7% SDS polyacrylamide gel. The overexpressed protein band was excised, crushed in PBS, mixed with Freund's adjuvant, and then injected into Balb/c mice four times at weekly intervals (200 μg protein per injection).

C. elegans embryos were immunostained by a slightly modified version of the procedure of Crittenden and Kimble(Crittenden and Kimble, 1999). Embryos were freeze-cracked, fixed, incubated with polyclonal mouse antiserum against the N-terminal 209 amino acids of WRN-1 (1:25 dilution), and then with FITC-conjugated goat anti-mouse immunoglobulin G (1:500 dilution, Santa Cruz Biotechnology) pre-treated with C. elegans acetone powder. After being stained with DAPI (4,6-diamidino-2-phenylindole, 1 μg/ml), specimens were observed with a fluorescence microscope (DMR HC, Leica). Gonads and intestines were extruded by decapitating adult C. elegans, fixed in 3% paraformaldehyde, then immunostained as described(Jones et al., 1996). Whole-worm staining was carried out by the collagenase method of Nonet et al.(Nonet et al., 1993). After fixation in 4% paraformaldehyde, worms were incubated in a reducing solution[5% β-mercaptoethanol, 1% Triton X-100, 0.1 M Tris-Cl (pH 6.9)] at 37°C overnight, and then reacted with collagenase (1000 units/ml, Sigma)in buffer [0.1 M Tris-Cl (pH 7.5), 1 mM CaCl₂] at 37°C for 5 hours. Subsequent reactions with primary and secondary antibodies were as described above for embryos.

The pCeWRN recombinant plasmid was linearized with BamHI and ApaI restriction enzymes at its multicloning site to prepare antisense and sense transcripts of wrn-1, respectively. Antisense RNA was synthesized using BamHI-digested plasmid DNA (2 μg), T7 RNA polymerase (5 units, MBI), ribonucleoside triphosphates (rNTPs, 0.4 mM each)and RNase inhibitor (5 units, Takara) in buffer [40 mM Tris-Cl (pH 8.0), 8 mM MgCl₂, 2 mM spermidine, 50 mM NaCl, 18 mM DTT; total 50 μl], at 37°C for 2 hours. Sense RNA was synthesized under the same reaction conditions as described for antisense RNA, except for the use of ApaI-treated DNA (2 μl) and T3 RNA polymerase (5 units, MBI). After RNA synthesis, RNase-free DNase I (2 units) was added to degrade template DNA, and then phenol (pH 4.5) extraction and ethanol precipitation were carried out. An equivalent amount of sense and antisense RNAs were mixed to a total concentration of 1 μg/μl, and then microinjected into the intestines of young adult N2 worms. The worms were placed on an NGM plate with an E. coli OP50 lawn, and were transferred to new plates after 12 hours.

Twelve hours after microinjection, the worms were allowed to lay embryos for 6 hours. F1 progeny (>100) were grown at 20°C or 25°C, and transferred to fresh plates every one or two days. Death was scored by absence of any movement after several light pokes with a platinum wire.

F1 progeny of the microinjected worms, designated Ce-wrn-1(RNAi)worms, were grown at 20°C or 25°C. Over 500 F1 worms were examined daily, with a stereomicroscope or with Nomarski optics (DMR HC, Leica), from the L1 stage. Worms with abnormal phenotypes were counted to 8 days old. The wild-type N2 strain was used as a control instead of a mock-RNAi strain, as the phenotypes produced by microinjection of dsRNA derived from a 5′-upstream DNA sequence were the same as wild type.

F1 larvae at the L1 stage, derived from the microinjected P0 worms, wereγ-irradiated with a ¹³⁷Cs source (IBL 437C, CIS Biointernational) at 60 Gray (Gy). After being kept at 20°C for 3 days,over 200 worms were examined under a stereomicroscope or with Nomarski optics. Wild-type N2 worms were also γ-irradiated and their phenotypes examined as a γ-irradiation control. In order to measure larval growth rate at 20°C, over 150 L1 stage worms were γ-irradiated at 0, 10 or 20 Gy. Subsequently, worms reaching the L4 stage, as defined by vulva shape, were scored every 12 hours.

The EST clone yk41c3 of the F18C5.2 ORF was digested with NotI and ApaI enzymes, and inserted into the pPD129.36(L4440) vector(Timmons and Fire, 1998),which contains two convergent T7 polymerase promoters in opposite orientations, separated by the multicloning site. Plasmid DNA was transformed into E. coli HT115(DE3) (W3110, rnc14::ΔTn10) cells by electroporation (Invitrogen). Cells harboring plasmid DNA were directly applied onto agar plates, composed of standard NGM/agar medium supplemented with 100 μg/ml ampicillin, 12.5 μg/ml tetracycline and 0.4 mM IPTG, and then cultured overnight at room temperature. L4 stage N2 worms were grown to adults on the plate covered with E. coli cells producing dsRNA of wrn-1, allowed to lay embryos for 1 hour, and then removed from the plate. The embryos were incubated at 20°C until they reached the L4 larval stage, followed by a temperature shift to 25°C. Twenty-four hours later,autofluorescence of adult worms was photographed using a fluorescence microscope (525 nm filter), and their heads were observed with Nomarski optics for 7 days. Control worms were fed with non-transformed HT115(DE3) E. coli cells.

The EST clone yk1302e07 of chk-1 (Y39H10A.7) was digested with XhoI and inserted into the pPD129.36(L4440) vector. On the feeding-plate covered with E. coli cells producing dsRNA of wrn-1, chk-1, or both, wild-type N2 or div-1(or148ts) worms at the L4 stage were placed and grown to adults at 25°C. The adult worms were dissected to isolate 2-cell-stage F1 embryos, which were then observed microscopically with Nomarski optics at appropriate intervals. In order to measure the duration of S and M phase in the early embryos, embryos were photographed every 10 seconds by time-lapse microscopy (Leica IM 1000).

On the feeding-plate covered with E. coli cells producing dsRNA of wrn-1, wild-type N2 worms were grown from L1 to L4 stages at 25°C. L4 stage worms were transferred to a new feeding-plate containing 25 mM hydroxyurea and were dissected to isolate the gonads 12 hour later. After staining with DAPI, the gonads were observed using a fluorescence microscope.

Of the four RecQ family proteins predicted by the C. elegansgenomic DNA sequence, ORF F18C5.2 is most homologous to human WRN; it has therefore been named as wrn-1 in WormBase(http://www.wormbase.org; Fig. 1). As the yk41c3 EST clone was shorter than the predicted ORF at its 5′-end, a 5′-end cDNA clone was amplified by reverse transcription-polymerase chain reaction(RT-PCR) of C. elegans total RNA, using gene-specific primers and an SL1 primer. The trans-splice leader sequence of SL1 (22 nucleotides) is found in two-thirds of C. elegans mRNAs and indicates that precursor mRNAs have been trans-spliced with SL1 RNA(Conrad et al., 1995). The 5′-end cDNA clone was fused to yk41c3 at a HindIII site (nt 641-646 in the ORF) to construct a full-length cDNA clone (pCeWRN) encoding 1056 amino acids. However, two EST clones of wrn-1 (yk1276b08 and yk811e06), which have trans-splice leader sequences of SL2 RNA(5′-GGTTTTAACCCAGTTACTCAAG-3′) and a novel SL RNA(5′-GTTTTTAACCCAGTTAATTGAG-3′) at the 5′-end, respectively,have recently been reported in WormBase. These two EST clones together with our 5′-end cDNA clone indicate that the pre-mRNAs of wrn-1 are trans-spliced at the same splice junction with three different SL RNAs. Trans-splicing with SL2 RNA means that the pre-mRNA is co-transcribed from an upstream gene (Evans et al.,2001), most likely from the F18C5.3 gene, which has a mRNA polyadenylation site about 100 nucleotides away from the trans-splice site of wrn-1. The F18C5.3 ORF is highly homologous to human DRIM(down-regulated in metastasis) protein, the molecular function of which is not clearly known (Schwirzke et al.,1998).

All RecQ family proteins have a helicase domain of 200-300 amino acids that contains five to seven motifs that are conserved among all helicases. The helicase domain with a DEAH box in the WRN-1 ORF shows 43% identity in the amino acid sequence with human WRN (Fig. 1). In addition, the C-terminal sequence succeeding the helicase domain, which consists of an RQC (RecQ helicase conserved) domain and an HRDC(Helicase and RNase D, C-terminal conserved) domain, share 25% sequence identity with human WRN. One major difference between the two WRN proteins is the absence of an exonuclease domain in C. elegans WRN-1. The exonuclease domain in human WRN is most homologous to ZK1098.3 and ZK1098.8(mut-7) of the C. elegans genomic ORFs, suggesting the possibility that one of these proteins is associated with C. elegansWRN-1.

WRN-1 was immunolocalized in C. elegans at various developmental stages (Fig. 2). The protein was non-uniformly distributed in nuclei from the early embryonic stage throughout embryogenesis. In early embryos, the protein level was higher in mitotic cells than in interphase cells, as is clearly shown in 6- and 8-cell stage embryos (mitotic cells marked with arrowheads or an arrow). At metaphase, WRN-1 had an unusual location, overlapping with the periphery of equatorially aligned condensed chromosomes facing the spindle poles (see the cell marked with an arrow in Fig. 2A, and the metaphase cell in Fig. 2D), and also, less frequently, overlapping with the mitotic spindles(Fig. 2A, arrow). This localization to one side of a mitotic sister chromatid was observed for CENP-A and CENP-C homologs in C. elegans embryos, which are centromere-binding proteins of holocentric C. elegans chromosomes(Moore and Roth, 2001), and also for SAN-1, which is involved in the spindle checkpoint(Nystul et al., 2003).

In larval stages WRN-1 was present in the nuclei of numerous cells, and the fraction of WRN-1-positive somatic cells was significantly lower in adults(Fig. 2B). In adult worms, the protein was expressed in hypodermal, intestinal and germ cells, and the protein level decreased with age. When nuclei in the intestine were magnified,an uneven distribution of the protein was observed(Fig. 2C). Both mitotic and meiotic prophase germ cells in the L4 stage gonad contained WRN-1 in the nuclei (Fig. 2C), but the level was significantly reduced in the meiotic prophase germ cells of adults (data not shown). In the oocytes of adult gonads, WRN-1 localized to condensed chromosomes (Fig. 2D). RNAi of wrn-1 eliminated immunologically detectable protein in embryos,supporting effective suppression of the endogenous gene expression and no cross-reactivity of the antibody (Fig. 2A). By contrast, RNAi of a RecQ homolog (RCQ-5) encoded by the E03A35.2 ORF, with 39% amino acid identity in the helicase domain of WRN-1,did not affect WRN-1 expression, thus underlining the specificity of wrn-1 RNAi (S.-J.L., unpublished).

To assess the in vivo function of WRN-1, RNA interference (RNAi) was carried out by microinjecting double-stranded RNA (dsRNA) of the wrn-1 gene into young adult worms (P0). When the progeny (F1) wrn-1(RNAi) worms were grown at 20°C, their life span was not affected: wild type and wrn-1(RNAi) worms lived for 17.1(±0.2) and 16.8 (±0.2) days after birth, respectively (J.-S.Y.,unpublished). By contrast, RNAi significantly reduced their life span at 25°C: the life span was 11.0 (±0.2) days for wrn-1(RNAi)worms and 13.6 (±0.1) days for wild-type worms (P<0.001; Fig. 3). In addition, the brood size of wrn-1(RNAi) worms was reduced to 94% of wild type at 20°C, and to 84% at 25°C (J.-S.Y., unpublished), whereas embryonic hatching was unaffected at either temperature.

To determine whether the reduced life span of wrn-1(RNAi) worms was due to progeria or to sickness, symptoms of normal aging, such as lipofuscin accumulation and tissue deterioration in the head(Garigan et al., 2002), were probed for the worms fed with E. coli cells producing dsRNA of wrn-1. Autofluorescence from intestinal lipofuscin increased with age in adult worms, as shown in Fig. 4A, and wrn-1(RNAi) worms had stronger autofluorescence than wild type at the same adult stage. When the two strains were compared for overall autofluorescence intensity in a worm, wrn-1(RNAi) adults advanced in lipofuscin accumulation by about 2 days relative to wild type. Adult C. elegans heads were photographed as in Fig. 4B, in order to assign scores of 1-5 depending on the extent of tissue damage, such as cavity formation and pharyngeal clogging due to bacterial packing. From the scatter diagrams of Fig. 4C,representing tissue deterioration in the head, wrn-1(RNAi) worms are estimated to age about one day faster than wild type.

Besides reduced life span, wrn-1(RNAi) worms at 25°C showed an increased frequency of abnormal features such as small body size, bag of worms, ruptured body, dumpy shape, growth arrest at larval stages and transparent body (Fig. 5A). These abnormal phenotypes were about three-fold more frequent in wrn-1(RNAi) worms than in wild-type worms(Fig. 5C). Some phenotypes were induced by RNAi even at 20°C, but at a much lower frequency than at 25°C (Fig. 5B).

Small body, dumpy and growth arrest are similar to the short stature characteristics of human Werner syndrome. Small body was designated as the adult body length less than 70% of wild type. Dumpy worms had short and fat bodies. Ruptured body, with internal organs such as the intestine and gonad bursting out of the vulva, followed on from a protruding vulva. This phenotype usually occurs because of defects in vulval morphogenesis(Hurd and Kemphues, 2003). The bag of worms phenotype, i.e. an egg-laying defect resulting in many hatched larvae remaining inside the worm, was probably caused by defects in HSN neurons producing serotonin, or in chemosensory neurons, as the defect was relieved by serotonin treatment (S.-J.L., unpublished). However another possibility is a partial defect in other egg-laying systems, such as vulva and sex muscles. In addition, transparent body lacking gut granules (Fitzgerald and Schwarzbauer, 1998) was induced at a very low frequency by the RNAi. As the small or dumpy aspect of wrn(RNAi) worms was not inherited by their progeny, these phenotypes probably were not caused by germ-line mutations. The combined frequency of bag of worms, ruptured body and growth arrest in wrn-1(RNAi) worms was 6% higher than in wild type and may have contributed to the 19% reduction in life span at 25°C(Fig. 3). Nevertheless, the small or dumpy phenotype was generally not accompanied by premature death,suggesting that the reduced life span mainly resulted from accelerated aging.

L1 stage larvae were γ-irradiated and their growth to adults was examined after 3 days. As shown in Fig. 5D, γ-radiation greatly increased the frequency of the phenotypes listed in Fig. 5Aeven at 20°C, both in wild type and wrn-1(RNAi) strains. Among these phenotypes, small body and bag of worms appeared in 40% and 35% of wrn-1(RNAi) worms, respectively, much higher than the values (close to 10% for each) for wild type. These large increases in the frequency of abnormal phenotypes must be due to enhancement of the RNAi effects byγ-radiation, given the much lower frequency (<2%) of these phenotypes in the absence of ionizing radiation (at 20°C). This finding suggested that WRN-1 participates in cellular responses to DNA damage and that the exacerbated developmental defects in its absence probably resulted from defective cellular signaling or/and DNA repair. The fact that the dumpy phenotype, unlike the other phenotypes listed in Fig. 5, was not significantly increased by ionizing radiation is very intriguing, as it points to a distinctive role of elevated metabolic rate at a higher temperature in inducing dumpiness.

As the enhancement of RNAi phenotypes by ionizing radiation suggested that WRN-1 played a role in cellular responses to DNA damage, larval growth rate was measured after irradiating L1 stage worms with lower doses of γ-ray than those worms shown in Fig. 5D. Even in the absence of γ-irradiation, wrn-1(RNAi) larvae surprisingly grew faster than wild-type N2 larvae,as shown in Fig. 6. Although wrn-1(RNAi) larvae reached the adult stage about 6 hours earlier than wild type at 20°C, this was much less than the difference of 2.6 days(P<0.001) in life span that was observed between the two strains at 25°C. Therefore, there remains a substantial reduction in life span(Fig. 3) even when the difference in larval growth rate is taken into account. Another striking aspect of the growth rate of wrn-1(RNAi) larvae was its independence from ionizing radiation, which contrasted with the growth retardation of wild-type larvae with increasing γ-ray dose(Fig. 6). The fast growth of wrn-1(RNAi) larvae and its insensitivity to DNA damage suggested that the strain probably was defective in a cell cycle checkpoint responding to DNA damage.

Fast development of wrn-1(RNAi) strain was observed during embryogenesis as well as at the larval stage, as shown in Fig. 7. When 2-cell embryos were observed under a microscope with Nomarski optics, the amount of time elapsing between the 2-cell and the 4-cell stage was 16.1 (±0.3)minutes (n=5) for wild type, whereas the corresponding time was 11.9(±0.2) minutes for the wrn-1(RNAi) strain(P<0.001). Ninety minutes after reaching the 2-cell stage,wild-type and wrn-1(RNAi) strains reached the 35-cell and 50-cell stages, respectively, clearly demonstrating a difference between cell division rates in the two strains (Fig. 7A). In order to determine which phase of the cell cycle was accelerated in wrn-1(RNAi) embryos, time-lapse micrographs were taken of the early embryos from 2- to 4-cell stages. S-phase duration was measured from cytokinesis of the P0 cell to nuclear envelope breakdown (NEBD) of the AB or P1 cell (Fig. 7B), as described by Brauchle et al. (Brauchle et al., 2003). As plotted in Fig. 7C, S-phase duration was shortened by 2.3 minutes(P<0.001) in the AB cell and by 2.5 minutes (P<0.001)in the P1 cell, by the RNAi, whereas M-phase duration was not affected (data not shown). chk-1(RNAi) was as effective as wrn-1(RNAi) in reducing the S-phase duration of the AB cell (P=0.44), and was more effective in shortening the same phase of the P1 cell (P=0.005),whereas simultaneous RNAi of both genes had no additive effect [wrn-1 versus wrn-1/chk-1, P=0.60 (AB cell) and 0.93 (P1 cell); chk-1 versus wrn-1/chk-1, P=0.24 (AB) and 0.002(P1)]. Thus, WRN-1 appears to act in the same checkpoint pathway as CHK-1. When wrn-1(RNAi) was performed in div-1(or148ts) C. elegans, which is mutated in the B subunit of the DNA polymeraseα-primase complex (Encalada et al.,2000), S-phase was shortened by 2.6 minutes (P<0.001)in the AB cell and by 4.1 minutes (P<0.001) in the P1 cell. These changes of S-phase duration were more or less similar to the corresponding changes in wild type. By contrast, a greater reduction of S-phase duration was obtained by chk-1(RNAi) in the div-1(or148ts) mutant: 5.8 minutes (P<0.001) in the AB cell and 7.7 minutes(P<0.001) in the P1 cell. Double RNAi of chk-1 and wrn-1 in the div-1(or148ts) strain was no more effective than single RNAi of chk-1 [P=0.76 (AB) and 0.49 (P1) for chk-1 versus wrn-1/chk-1]. Therefore, C. elegans chk-1 is involved in the DNA replication checkpoint pathway induced by inefficient priming for Okazaki fragments, as first demonstrated by Brauchle et al. (Brauchle et al.,2003), and wrn-1 appeared to be less efficient than chk-1 at this checkpoint. The effect of wrn-1(RNAi) on S-phase duration was significantly enhanced in the P1 cell of the div-1(or148ts) mutant relative to that of wild type(P=0.01), whereas the effect was similar in the AB cells of the two strains (P>0.50).

In order to probe the role of wrn-1 in the DNA replication checkpoint, C. elegans gonads were treated with hydroxyurea to interrupt DNA replication by depleting deoxynucleotides, and were then observed by fluorescence microscopy. As shown in Fig. 8, pre-meiotic nuclei in the wild-type gonad were enlarged and reduced in number because of cell cycle arrest combined with continued nuclear growth. By contrast, small nuclei were observed in the wrn-1(RNAi) gonad even after hydroxyurea treatment,and a fraction of the nuclei appeared to be fragmented, suggesting an ineffective checkpoint for replication blockage.

Werner syndrome is accompanied by genomic instability, leading to DNA deletions and translocations (Salk et al.,1981; Fukuchi et al.,1989) resulting from enzymatic malfunction of WRN as a helicase and exonuclease. The sequence similarity between C. elegans and human WRN suggests that these two proteins possess similar functions in maintaining genomic stability, although C. elegans WRN-1 lacks the exonuclease domain of human WRN.

C. elegans WRN-1 is diffusely distributed in the nucleoplasm during interphase as is mouse WRN, whereas human WRN is predominantly localized to the nucleolus (Marciniak et al., 1998; Suzuki et al.,2001). However, human WRN relocates from the nucleolus to nucleoplasmic foci upon induction of DNA damage with UV, ionizing radiation,camptothecin, etoposide or 4-nitroquinoline-1-oxide(Gray et al., 1998; Sakamoto et al., 2001; Blander et al., 2002). The WRN foci overlap with the foci of RP-A almost completely, and overlap with those of RAD51 partially, suggesting that human WRN cooperates with RP-A and RAD51 in response to DNA damage (Sakamoto et al., 2001). Similarly, interruption of DNA synthesis by depleting deoxynucleotides with hydroxyurea results in the movement of WRN from nucleoli to distinct nuclear foci that co-localize with RP-A(Constantinou et al., 2000). However, the localization of WRN-1 to the poleward periphery of metaphase chromosomes in the early C. elegans embryo has not been observed in other organisms. This peculiar localization may be needed to ensure equal partition of the protein into daughter cells, or may mean that WRN-1 functions in a spindle checkpoint involving holocentric chromosomes during mitosis.

RNAi of C. elegans wrn-1 did not cause discernable phenotypes at 20°C, whereas at 25°C the worms had a reduced life span and showed significant increases in such developmental abnormalities as small body, bag of worms, ruptured body, dumpy shape, growth arrest and transparency(Fig. 5). These developmental abnormalities have been induced by deficiency of ATR and RAD51 homologs in C. elegans, and their frequency was increased by ionizing radiation(Aoki et al., 2000; Rinaldo et al., 2002). Therefore, the developmental defects probably resulted from DNA damage during development. Some of these phenotypes resemble the symptoms of human Werner syndrome, such as short stature and premature aging. The enhanced phenotypes of wrn-1(RNAi) at 25°C was very likely due to the fact that the metabolic rate of C. elegans is 1.2 times higher at 25°C than at 20°C (Van Voorhies and Ward,1999), thus generating a greater oxidative stress. The slight reduction in life span was probably due to premature aging, as was demonstrated by faster lipofuscin accumulation in the C. elegansintestine and tissue deterioration in the head, although other phenotypes such as bag of worms, ruptured body and growth arrest certainly contributed to the reduction. Many small C. elegans mutants are known to have defects in the TGF-β signaling pathway (Gumienny and Padgett, 2003), and others are mutated in the spectrin βchain (sma-7), a MAP kinase (sma-5), a fatty acid elongation enzyme (elo-2) or a basement membrane protein (SPARC). Dumpiness in C. elegans is generally due to mutation of genes participating in collagen production or in dosage compensation. The phenotype bag of worms is thought to be caused by defects in HSN or chemosensory neurons, or to partial defects in other egg-laying systems, based on its sensitivity to serotonin. A C. elegans line harboring human WRN cDNA linked to a strong C. elegans promoter was prepared, but the occurrence of wrn-1(RNAi)phenotypes was not reduced in the transgenic line (S.M.H., unpublished). The failure to rescue the phenotypes could be due to non-equivalence of human and C. elegans WRNs, or to inefficient expression of the exogenous gene.

Ionizing radiation increased the frequency of wrn-1(RNAi)phenotypes synergistically, suggesting that C. elegans WRN-1 is involved in the DNA damage response (Fig. 5D). WRN-1 may be involved in non-homologous end-joining by interacting with Ku70/80 (Cooper et al.,2000), or may be involved in recombinational repair. WRN-1 may also play a role in DNA damage signaling, as in the control of p53-mediated transcriptional activation by human WRN(Spillare et al., 1999). Indeed, the fast larval growth of the wrn-1(RNAi) strain, especially its insensitivity to ionizing radiation, strongly suggests that WRN-1 acts in a DNA damage checkpoint pathway. In germ cells treated with hydroxyurea, WRN-1 was required to activate the DNA replication checkpoint, which agreed well with the role of a Saccharomyces cerevisiae homolog, Sgs1, at the same checkpoint (Frei and Gasser,2000; Myung and Kolodner,2002). In addition, S phase was accelerated in wrn-1(RNAi) embryos, indicating a role of WRN-1 as a checkpoint protein. Nevertheless, wrn-1(RNAi) was not as potent as chk-1(RNAi) in reducing the extended S-phase of div-1(or148ts) embryos, in which priming of Okazaki fragments is inefficient. And, double RNAi of wrn-1 and chk-1 was no more effective than single RNAi of chk-1. This suggests that WRN-1 works in a sub-pathway that diverges from CHK-1 in the DNA replication checkpoint pathway, and that is either up- or downstream of CHK-1. In sgs1Δ S. cerevisiae cells, S phase proceeds faster than in wild-type cells, but the termination stage of S phase takes longer, so that the total length of S phase is unchanged(Versini et al., 2003). The fast progression of S phase in S. cerevisiae and C. elegansresulting from the absence of a RecQ homolog contrasts with the extended S phase in human WRN^– cells(Martin et al., 1970; Salk et al., 1985). Recently,human WRN^– cells were found to be defective in the chromosomal decatenation checkpoint in G2 phase(Franchitto et al., 2003),suggesting the possibility that C. elegans WRN-1 may also participate in a DNA damage checkpoint during G2 phase.

There have been conflicting reports concerning the sensitivity of wrn-deficient cells to DNA damaging agents. Human WS patient cells were not hypersensitive to UV or X-ray(Fujiwara et al., 1977), but were sensitive to camptothecin, 4-nitroquinoline-N-oxide and DNA-crosslinking agents (Poot et al., 1999; Poot et al., 2001; Poot et al., 2002). The embryonic stem cells of wrn-knockout mice are hypersensitive to camptothecin and etoposide, which are inhibitors of DNA topoisomerases I and II, respectively, but were not sensitive to γ-radiation, UV or mitomycin(Lebel and Leder, 1998). WRN mutants of the chicken B-cell line DT-40 were sensitive to various DNA damaging agents, such as methylmethanesulfonate, 4-nitroquinoline-N-oxide,etoposide and camptothecin (Imamura et al., 2002). The insensitivity of human and murine WRN^– cells to ionizing radiation(Fujiwara et al., 1977; Lebel and Leder, 1998) is in contrast to the enhancement of the C. elegans wrn-1(RNAi) phenotypes by ionizing radiation demonstrated in this study.

Although a clear explanation for their pleiotropic phenotypes cannot be provided, the similarity in phenotypes between wrn-1(RNAi) worms and Werner syndrome patients suggests that the RNAi worm could be a useful model for the syndrome. The fact that Wrn knockout mice did not recapitulate Werner syndrome phenotypes such as premature aging, small stature, developmental abnormalities and tumor formation(Lebel and Leder, 1998; Lombard et al., 2000) further emphasizes the importance of wrn-deficient C. elegans as a potential model for Werner syndrome.

We thank Dr Yuji Kohara (National Institute of Genetics, Japan) for the yk41c EST clone. N2 and div-1(or148ts) C. elegans strains were obtained from the C. elegans Genetics Center (St Paul, MN, USA),which is supported by the National Center for Research Resources. This work was supported by a Life Phenomena and Function Research Grant[01-J-LF-01-B-83] from the Korean Ministry of Science and Technology, and by the Brain Korea 21 Project in 2003 to H.-S.K.