Frequent Germline Mutations and Somatic Repeat Instability in DNA Mismatch-Repair-Deficient Caenorhabditis elegans

Corresponding author: Ronald H. A. Plasterk, Uppsalalaan 8, 3584 CT, Utrecht, The Netherlands., plasterk{at}niob.knaw.nl (E-mail)

Communicating editor: B. J. MEYER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mismatch-repair-deficient mutants were initially recognized as mutation-prone derivatives of bacteria, and later mismatch repair deficiency was found to predispose humans to colon cancers (HNPCC). We generated mismatch-repair-deficient Caenorhabditis elegans by deleting the msh-6 gene and analyzed the fidelity of transmission of genetic information to subsequent generations. msh-6-defective animals show an elevated level of spontaneous mutants in both the male and female germline; also repeated DNA tracts are unstable. To monitor DNA repeat instability in somatic tissue, we developed a sensitive system, making use of heat-shock promoter-driven lacZ transgenes, but with a repeat that puts this reporter gene out of frame. In genetic msh-6-deficient animals lacZ⁺ patches are observed as a result of somatic repeat instability. RNA interference by feeding wild-type animals dsRNA homologous to msh-2 or msh-6 also resulted in somatic DNA instability, as well as in germline mutagenesis, indicating that one can use C. elegans as a model system to discover genes involved in maintaining DNA stability by large-scale RNAi screens.

DNA mismatch repair (MMR) mutants were originally found in screens directed at the identification of bacterial mutants that had a mutator phenotype and thus had elevated levels of spontaneous mutants in their progeny. Subsequent genetic as well as biochemical studies identified the MMR machinery as an enzymatic complex that could identify DNA mismatches resulting from single-nucleotide substitutions or small insertions/deletions, recognize the parental from the newly synthesized strand, excise the new strand around the lesion, and initiate repair to close the gap (reviewed in KOLODNER 1996 ; MODRICH and LAHUE 1996 ; JIRICNY 1998 ).

One of the greatest success stories of model organism genetics came when a human syndrome of cancer predisposition, hereditary nonpolyposis colon cancer (HNPCC), was found to result from a defect in human homologs of genes encoding components in the bacterial MMR machinery (FISHEL et al. 1993 ; LEACH et al. 1993 ; BRONNER et al. 1994 ; LIU et al. 1994 ; NICOLAIDES et al. 1994 ; PAPADOPOULOS et al. 1994 ). The fact that these cancers are characterized by an increased instability of simple DNA repeats provided the first clue that a replication-associated repair mechanism was involved (PEINADO et al. 1992 ; AALTONEN et al. 1993 ; IONOV et al. 1993 ; PELTOMAKI et al. 1993 ). The notion that MMR defects are associated with human cancer provides strong support for the hypothesis that a so-called mutator phenotype, here as a result of elevated levels of unrepaired somatic DNA mismatches, can promote tumorigenesis (LOEB 1991 ). This model has been further supported by mouse knockouts of the MMR genes Msh2, Msh6, Pms2, or Mlh1 that show enhanced cancer frequencies and repeat instability (DE WIND et al. 1995 ; REITMAIR et al. 1995 ; BAKER et al. 1996 ; EDELMANN et al. 1996 , EDELMANN et al. 1997 ; NARAYANAN et al. 1997 ; PROLLA et al. 1998 ).

Also in humans that do not contain germline mutations in DNA MMR genes, tumors are often found that display repeat instabilities. Upon analysis these are often defective in known components of the MMR machinery; either they carry mutations within the genes themselves or the expression of these MMR genes is epigenetically downregulated as a result of hypermethylation (BUERMEYER et al. 1999 and references therein). However, not all sporadic human tumors with repeat instability show a defect in the known DNA MMR genes. In addition, in numerous HNPCC cases no germline mutations are found in the known MMR genes (PELTOMAKI and DE LA CHAPELLE 1997 ). This suggests that there are additional genes in humans whose loss results in this specific type of genetic instability. Indeed, recently, two other genes, i.e., MLH3 and EXO1, that are involved in the correction of mismatched bases (SZANKASI and SMITH 1995 ; FLORES-ROZAS and KOLODNER 1998 ) have been found to be mutated in the germline of patients suffering from HNPCC (LIPKIN et al. 2001 ; WU et al. 2001A , WU et al. 2001B ).

Relatively little is known about effects of DNA mismatch repair mutations upon germline mutagenesis in animals. Hypermutability of simple sequence repeats in human sperm can be measured (BACON et al. 2001 ), and sperm of patients with HNPCC have been analyzed for aneuploidy frequencies (MARTIN et al. 2000 ). However, the spontaneous levels of mutagenesis in common laboratory animals such as the mouse are too low to measure with a degree of precision. Here, we analyzed the effect of MMR on transmission of the genetic material to subsequent generations by studying MMR-defective Caenorhabditis elegans strains and found that the male as well as the female germline is protected against spontaneous mutagenesis. This was monitored by (i) the segregation of different visible mutant phenotypes, (ii) an enhanced level of mutations in essential genes in a balanced segment of the genome, and (iii) an enhanced level of mutations in a monitor gene (unc-93). In addition, we followed the fate of mono- and dinucleotide repeats in the C. elegans genome over many generations in wild-type as well as in MMR-deficient animals.

Tumorigenesis as a result of somatic DNA changes is not easily scored in mutator mutants, due to the fact that C. elegans lives only 2 weeks and has only 959 somatic cells. Therefore, we constructed transgenic animals that contained hundreds of tandemly repeated copies of an out-of-frame lacZ gene, reasoning that replication errors during development that would bring only one copy back in frame could be visualized by staining of such tissues using 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside (X-gal). To enhance the chances of a frameshift we cloned mono- or dinucleotide repeats between the initiation ATG and the ß-galactosidase-encoding lacZ gene. Such constructs were found to be stable in wild-type animals (i.e., no or almost no ß-galactosidase expression observed), while clear positive patches were seen in virtually every individual animal in a msh-6 genetic background. Identical phenotypes were observed after RNAi of msh-6 and of msh-2. RNAi is the experimental silencing of the expression of a given gene by administration, in this case feeding, of double-stranded RNA corresponding to that gene (FIRE et al. 1998 ). Thus feeding on dsRNA homologous to a MMR gene is sufficient to trigger DNA instability in somatic cells. This suggests that we can use C. elegans as a model system to discover genes involved in maintaining DNA stability by large-scale RNAi screens. These genes may be homologs of the yet unknown non-msh genes protecting DNA against mutations.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and maintenance:
General methods for culturing C. elegans strains were as described in BRENNER 1974 . Strains used in this study were CB1500 [unc-93(e1500)], MT765 [unc-93(e1500 n224)], and BC1958 [dpy-18(e364)/eT1 III; unc-46(e177)/eT1 V]. A deletion mutant of msh-6: pk2504 was isolated from a chemical deletion library as described (JANSEN et al. 1997 ).

Spontaneous mutation frequency:
Growing cultures of msh-6 strains segregate a plethora of visible mutants indicative of a mutator phenotype. From the brood of four msh-6 hermaphrodites, 300 progeny animals were picked that had a wild-type appearance. These worms were grown individually and the progeny were inspected for Mendelian segregation of visible phenotypes. Plates were screened a second time 2 days after food deprivation; this allows the scoring of an embryonic lethal phenotype, here interpreted as the abundant presence of dead eggs on the culture dish.

To determine whether msh-6 animals have a high incidence of male (him) phenotype, the broods of three to five animals of genotype msh-6 or wild type were inspected for the presence of a male: msh-6, 1/1209 (0.08%); wild type, 1/1059 (0.09%). The genetic recombination frequency was analyzed by determining the genetic distance between the visible marker unc-32 and dpy-18 on LGIII in an msh-6 and wild-type genetic background. For animals of genotype msh-6; unc-32 dpy-18/+ +, the brood consisted of 412 wild type, 20 Unc, 21 Dpy, and 112 Unc Dpy, resulting in a recombination frequency of 0.075 (map distance 7.5 cM). In a mismatch-proficient genetic background the frequency was 0.072 (map distance 7.2 cM): A total of 527 wild type, 26 Unc, 24 Dpy, and 140 Unc Dpy segregated from animals of genotype unc-32 dpy-18/+ +.

The mutator phenotype of msh-6 C. elegans was quantified using the reciprocal translocation eT1(III;V) as a balancer, as described by ROSENBLUTH et al. 1983 . First, msh-6 males were crossed with hermaphrodites that were homozygous for the translocated eT1 chromosomes (this genotype results in a visible phenotype because the translocation disrupts the unc-36 locus). F₁ males were subsequently crossed with hermaphrodites of genotype dpy-18; unc-46 (to mark the nontranslocated chromosomes) and cross-progeny of genotype msh-6/+ I; dpy-18/eT1 III; unc-46/eT1 V were selected. Next generation animals homozygous for msh-6 and segregating both Dpy-18 Unc-46 and Unc-36 animals were used as starting strains in the following experimental setup: Phenotypically wild-type progeny of hermaphrodites of the above-described genotype were picked onto individual plates and scored for segregation of the Dpy-18 Unc-46 phenotype. The frequency of recessive lethal mutations induced in the balanced area of the genome is reflected by the percentage of animals that fail to segregate this phenotype: A lethal mutation in the crossover-suppressed region of the canonical chromosomes prevents embryos homozygous for these chromosomes from developing into adult Dpy Unc worms. Clonal lines that were positive in this screen were grown further and confirmed as carrying a lethal mutation inside one of the crossover-suppressed regions if showing no such Dpy Unc markers in at least 250 offspring.

For determining the germline frequency in male sperm of msh-6 animals, males of genotype msh-6 I; dpy-18/eT1 III; unc-46/eT1 V were crossed to hermaphrodites of genotype eT1(III/V). Phenotypically wild-type progeny were analyzed for segregation of the marked chromosomes as described above. The germline mutation frequency of hermaphrodite oocytes was determined by analyzing the phenotypically wild-type cross-progeny of dpy-18/eT1; unc-46/eT1 males crossed to msh-6; dpy-18; unc-46 hermaphrodites. In both crossing schemes, the msh-6 deficient animals that were used to start the analysis were homozygous for more than one generation. Therefore, to prevent scoring mutations that occurred in earlier generations (that result in so-called "Jackpots") >30 cross-progeny animals were tested from a single hermaphrodite.

RNAi of msh-6 and msh-2 was done by injecting hermaphrodites of strain BC1958 with cognate dsRNA and subsequent analysis of the mutator phenotype in the phenotypically wild-type F₁. Thus the F₂ was inspected for segregation of the Dpy Unc phenotype. In addition, RNAi was measured by culturing BC1958 animals on msh-2 or msh-6 dsRNA-producing bacteria (described below).

Mutation spectrum of msh-6 worms:
Phenotypic reversion of the uncoordinated "rubber-band," egg-laying-defective phenotype conferred by unc-93(e1500) was used to determine the nature of mutations that occurred in a msh-6 genetic background. Cultures started with single hermaphrodites of genotype msh-6 unc-93(e1500) were inspected regularly for revertants that were recognized by their wild-type movement and normal egg-laying behavior. Intragenic reversion events [mutations in at least four other loci can suppress the unc-93(e1500)-associated phenotype] were identified by the failure of these alleles to complement unc-93(e1500n224). Subsequently, the coding region of the unc-93 locus was sequenced.

Microsatellite repeat instability in msh-6 worms:
From a single hermaphrodite (msh-6 and Bristol N2), 55 progeny were picked to start lines that were maintained by transferring several L4 animals every 3–4 days to new plates. After 10 generations DNA was isolated from cultures started with a single animal (due to the mutator phenotype of msh-6, mutations will accumulate and often a sterile phenotype is observed when individual animals are cloned out). From these cultures different genomic loci were analyzed by sequencing PCR products. Primers used are (5'-3') as follows: R03C1_A, cggcaaacaatttttccg; R03C_C, acggaggtgttcacggag; F59A3_A, cgtttgaaggatgatgtc; F59A3_C, gatgctcgatgacttcgg; C41D7_A, gattctcaagtccacccg; C41D7_C, gacccgttctcctactcc; M03F4_A, cgaaatggatctgagtggg; M03F4_C, atatcccatgatgacccc; C24A3_A, gagtgcgcttgaagagactg; C24A3_C, cggaactcggagagagatag; Y54G11A_A, ggatcttggctcctggaacg; and Y54G11A_C, cattgagtgatactcggccg.

Detection of somatic repeat instability:
To allow detection of somatic repeat instability we created several constructs that contained stretches of either mono- or dinucleotide repeats between the start of translation and the lacZ open reading frame (ORF), under the control of a heat-shock promoter.

In brief, vector L2681 (kindly provided by A. Fire), which has a green fluorescent protein (GFP)/LacZ fusion under the control of a heat-shock promoter, was digested with BamHI and allowed to close to create pRP1820; this cloning step removes two upstream ATG sequences without affecting essential promoter sequences. Then, the original starting codon was removed by site-directed mutagenesis to create pRP1821. This construct was subsequently used as a recipient for insertion of DNA fragments containing different types of repeats: Partially complementing oligonucleotides were annealed and inserted into a KpnI site near the beginning of the fusion protein-encoded sequences. All constructs had a similar molecular architecture: heat-shock promoter-(KpnI)-ATG-(A or CA)n-GFP/lacZ ORF (sequences and cloning details available upon request). The different types of repeat used in this study were pRP1822, (A)₁₆; pRP1823, (A)₁₇; pRP1840, (A)₁₅; pRP1841, (CA)₁₅; pRP1842, (CA)₁₄; and pRP1843, (CA)₁₃. pRP1822 and pRP1842 contain an in-frame lacZ construct encoding functional ß-galactosidase.

All constructs were injected separately (together with pRF4 containing the dominant marker rol-6) into the canonical C. elegans strain Bristol N2 to establish transgenic lines (MELLO et al. 1991 ). The transgenic array containing pRP1822 was integrated by -irradiation and used for further detailed analysis of somatic reversion events.

To identify expression of ß-galactosidase, nematodes were fixed and stained with X-gal.

RNAi: By injection:
PCR fragments of msh-6 [Y47G6A, nucleotides (nt) 22458–23143] and msh-2 (H26D21, nt 467–946, GenBank accession no. AF106587) coding sequences were cloned into vector pCCM114 (kind gift of Craig Mello) that contains oppositely oriented T7 promoters. Plasmid DNA was isolated, linearized, and used as template to synthesize dsRNA in vitro with T7 RNA polymerase (Boehringer Mannheim, Indianapolis) according to the manufacturer's conditions. Hermaphrodites were injected with 500 ng/µl dsRNA.

By feeding:

msh-6 and msh-2 DNA segments (identical sequences as described above) were cloned into the "feeding vector" L4440 and subsequently transformed to HT115 bacterial cells that were used for RNAi by feeding using the protocol described by Ahringer and co-workers (FRASER et al. 2000 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutator phenotype in mismatch-repair-defective C. elegans:
We screened the genome sequence of C. elegans for homologs of bacterial and human DNA MMR genes and found msh-2 and msh-6 (H26D21.2 and Y47G6A.11, respectively; homologous to prokaryotic mutS) and mlh-1 and pms-2 (T28A8.7 and H12C20.2; homologous to prokaryotic mutL). Surprisingly, an ortholog of msh-3 was not detected. We then knocked out the msh-6 gene, using a mutant library approach previously developed in our laboratory (JANSEN et al. 1997 ). Fig 1 shows the genomic organization of the C. elegans msh-6 gene, the human and Saccharomyces cerevisiae homologs aligned to msh-6 of C. elegans, and the deletion mutant that was used in this study.

View larger version (82K): In this window In a new window Download PPT slide   Figure 1. The C. elegans msh-6 gene. (A) Structure of the C. elegans msh-6 gene deduced from genomic sequences and cDNA generated by reverse transcription-PCR from Bristol N2 RNA. The genomic region that is deleted in pk2504 (nt 24180–25956 of Y47G6A, GenBank accession no. AC024791) and takes out exon 5 and part of exon 6 is indicated. (B) Alignment of the amino acid sequence of C. elegans, human, and S. cerevisae MSH-6 using the CLUSTALW algorithm. Black background indicates amino acid identity and gray background indicates conserved amino acid substitutions. The amino acids deleted in pk2504 are underlined. Possible alternative splicing of exon 4 onto exon 7 predicts an out-of-frame product.

Homozygous msh-6 mutants are viable, and the first indication of a mutator phenotype was the frequent occurrence of readily recognizable mutants (such as dpy and unc) among the progeny. Since C. elegans lines are maintained as self-fertilizing hermaphrodites, spontaneous new mutations can homozygose in self-progeny, so that recessive mutations are easily observed. At least 20 phenotypic mutations could be found in the brood (n = 300) of phenotypically wild-type hermaphrodites. In the parental strain such a level of spontaneous mutations is not seen (0 mutant phenotypes in the brood of 300 cloned individuals). Apart from the increased mutation frequency in the msh-6 mutant, no other phenotypes that are indicative of specific defects in genome stability were noted: X-chromosomal nondisjunction is not affected by the msh-6 deletion, indicated by the absence of a high incidence of male (him) phenotype. Also, no effect was observed on genetic recombination: The genetic distance between visible markers is similar in wild-type and msh-6 animals (see MATERIALS AND METHODS for details).

Increased germline mutagenesis in msh-6 defective C. elegans in both the male and female germline:
To validate and quantify this mutator effect, we measured the frequency of recessive lethal mutations induced within a sizable region of the C. elegans genome (ROSENBLUTH et al. 1983 ; see MATERIALS AND METHODS for details). In a wild-type strain we detect spontaneous mutations in this region below a frequency of 10^-3, which is in line with the numbers reported in the literature for wild-type C. elegans, i.e., 6 x 10^-4 (ROSENBLUTH et al. 1983 ). In msh-6 mutants this level is 1.6 x 10^-2 (Fig 2), at least 28 times elevated. These mutations could theoretically arise from mutations that occur uniquely in the sperm or in the oocytes of the hermaphrodite parent. In humans, mutation frequencies determined for microsatellite loci revealed a clear paternal excess; males mutate about five times as often as females (reviewed in ELLEGREN 2000 ). To test whether the MMR machinery protects the male and female germlines equally, we performed experiments that scored for spontaneous mutants in progeny from crosses between males and hermaphrodites, in which either one of the parents was mutant and the other wild type for msh-6 (see MATERIALS AND METHODS for details). As shown in Fig 2, both the oocytes of the hermaphroditic mother and the sperm from male fathers show a similar increase in the level of spontaneous mutagenesis in the msh-6 mutant. We conclude two things: The frequency of original DNA replication errors is probably comparable in sperm and oocytes, and the level of protection by the MMR machinery is also similar. (Note that we did not test the sperm produced by hermaphrodites, since mutations there cannot be distinguished from those that arose in the oocytes of the same animal.)

View larger version (38K): In this window In a new window Download PPT slide   Figure 2. Mismatch repair proteins MSH-6 and MSH-2 protect the C. elegans germline from spontaneous mutagenesis. The experimental set-up that is used to measure the level of spontaneous mutagenesis is described in MATERIALS AND METHODS. This assay determines the absolute number of loss-of-function mutations in essential genes in a region that covers 7% of the C. elegans genome (estimated number of target genes, 300). The y-axis reflects the percentage of animals that acquire such a lethal mutation within one generation.

As a second measure of mutation frequencies we scored for loss-of-function mutations in the unc-93 monitor gene. Animals homozygous for the unc-93(e1500) mutation are uncoordinated and absolutely egg-laying defective, while complete loss of the unc-93 gene has no visible phenotype, and thus loss-of-function mutants of the unc-93(e1500) gene can be scored by recognizing normally moving animals among contracted ones (GREENWALD and HORVITZ 1980 ). Therefore, this gene has been previously used to assay mutagenesis levels (DE STASIO et al. 1997 ). We found that the levels of mutations in unc-93(e1500) increase at least 30-fold in msh-6 mutants compared to wild type (data not shown). The advantage of using the unc-93 monitor gene is that once obtained, these mutants can also be identified at the molecular level by direct sequencing of the relatively small genomic unc-93 gene. It is known that loss of four other genes (sup-9, sup-10, sup-11, and sup-18) also reverts the unc-93(e1500) phenotype, so we first sorted out the mutations that mapped to unc-93 and sequenced only those (from 64 revertants that were independently isolated, 18 mapped to the unc-93 locus). The nature of the mutations is shown in Table 1: Mostly we find frameshifts in short monomeric runs and single-base substitutions, which is similar to the spectrum seen in S. cerevisiae MSH6 strains (MARSISCHKY et al. 1996 ; GREENE and JINKS-ROBERTSON 1997 ) and in human cell lines defective for hMSH6 (BHATTACHARYYA et al. 1994 ).

Microsatellite repeat instability in msh-6-defective C. elegans:
As discussed in the Introduction, microsatellite instability is a hallmark of tumors derived from HNPCC patients [replication errors in repeats (RER⁺)]. To see if C. elegans defective for msh-6 display microsatellite instability, we started 55 parallel lines by cloning the progeny of one msh-6 hermaphrodite. These lines were maintained for 10 generations, and then we picked one animal per line and sequenced various genomic loci containing microsatellites. As shown in Table 2, mono-, di-, and trinucleotide repeats become highly unstable in the absence of functional MSH-6, while no instability was observed in wild-type animals.

Visualization of somatic repeat instability in C. elegans:
Having observed these frequent repeat length changes in the germline of msh-6 mutants we wondered if these could also be observed in somatic cells. With a life span of only 2 weeks, and most somatic cells being only a few cell divisions removed from the zygote, one may not expect too many mutations. Therefore we devised a sensitive system for scoring repeat length instability. Inspired by the work of Petes and co-workers (HENDERSON and PETES 1992 ; STRAND et al. 1993 ), we cloned a repeat into a reporter gene in such a way that the repeat was between the ATG initiation triplet and the domain of the gene encoding the enzymatic activity and would keep the latter out of frame. Unrepaired replication errors in the repeat will bring the gene in frame, which can be visualized using the reporter. To enhance the chances of finding such events, we took advantage of the fact that transgenes in C. elegans are usually tandem repeats of hundreds of copies of the injected DNA; one would hope that a frameshift in only one of those copies could be scored. Initial attempts to use GFP for this purpose failed. We then switched to using lacZ, which was present as a fusion protein in the same construct. A disadvantage of this reporter is that the animal needs to be impregnated with the reagent X-gal, which kills the animal. An advantage is that lacZ staining is more sensitive, especially because one can prolong the staining to increase the detection level.

Transgenic animals with the lacZ gene cloned "in frame" under the control of a heat-shock promoter show, as expected, a broad ß-galactosidase expression pattern (Fig 3A). In contrast, in transgenic animals that have the lacZ gene out of frame, downstream of a repeat sequence, virtually no staining is observed after heat shock (Fig 3B). However, in msh-6 mutant background, as shown in Fig 3C, almost every worm shows one or more blue patches. We conclude that these arise from repeat instability and restoration of the lacZ reading frame in different cell lineages. Unfortunately the fixed and stained worms have not allowed us to recognize specific sublineages, but we could see blue patches of multiple tissues.

View larger version (72K): In this window In a new window Download PPT slide   Figure 3. Genetic instability in MMR-defective somatic cells. A schematic representation of the constructs that are used to measure somatic repeat instability is depicted above the images of the nematodes. (a) Transgenic C. elegans that carry multiple in-frame copies of heat-shock-driven lacZ. (b) MMR-proficient transgenic C. elegans (N2) that carry multiple copies of a lacZ-containing construct in which a repeat sequence is cloned immediately downstream of the ATG that puts the downstream-positioned ß-galactosidase ORF out of frame. (c) The identical transgenic array crossed into an msh-6 genetic background.

To check the role of the repeat in this msh-6-dependent frameshift, we generated transgenic animals that contained identical constructs without the repeat and saw that the level of somatic cells that stain blue was not different in wild-type and msh-6-deficient animals (data not shown). This suggests that the increased staining that we observed for msh-6 cells results from replication errors that occur at the repeat position, leading to one in-frame lacZ copy in that cell and its descendants.

Destabilizing the germline by feeding msh-2 and msh-6 dsRNA:
RNA interference is the silencing of gene expression by administration of dsRNA that corresponds to exonic sequences of that gene (FIRE et al. 1998 ). One of the striking features of RNAi in C. elegans is that the dsRNA can be administered by feeding Escherichia coli that contain a plasmid that transcribes both strands, which combines to form dsRNA (TIMMONS and FIRE 1998 ). We fed worms on E. coli that produced dsRNA for msh-6 and measured spontaneous mutation frequencies by scoring for mutants in the progeny. The results are shown in Fig 2: As for the genetic knockout, RNAi of msh-6 resulted in an increased spontaneous mutation frequency. The administration of the msh-6 dsRNA via another route, i.e., direct injection of purified dsRNA into the gonad of the animal, had the identical outcome. We then also performed RNAi for another known MMR component, the msh-2 gene, and found a comparable increase in mutation frequency (Fig 2).

Destabilizing the genetic contents of somatic cells by msh-2 and msh-6 RNAi:
Combining the lacZ reporter with msh-2 and msh-6 RNAi, we fed dsRNA to worms and scored for restoration of the lacZ reading frame in somatic cells, and, as shown in Fig 4A, the effect is the same as that of the genetic null: Almost every animal has patches of ß-galactosidase expression. The fact that we can phenocopy the genetic null mutation using RNAi allows us to analyze whether the staining was dependent on cell division as a marker for DNA replication. This prediction should be met if the patched phenotype is indeed the result of frameshift errors in somatic cells during development. We placed animals at different developmental stages on RNAi food and measured the level of staining after the animals developed to adulthood. We reasoned that because young larvae undergo more cell division cycles than mature animals, they are more prone to frameshift errors. Indeed, we found a reduction in the number of patches when the animals were exposed at later developmental stages (Fig 4B), consistent with a dependence on cell division.

View larger version (31K): In this window In a new window Download PPT slide   Figure 4. Somatic repeat instability induced by msh-2 and msh-6 RNAi. (a) Animals are fed E. coli that produce dsRNA homologous to the C. elegans MMR genes msh-2 and msh-6 or a non-MMR control gene, unc-22. (b) Nematodes synchronized at different developmental stages (L1 larvae or L4 young adults) were exposed to RNAi by feeding, cultured to adulthood, and analyzed for somatic repeat instability. P0 indicates that progeny animals were analyzed from cultures in which the mother was transferred to the RNAi plates. The y-axis indicates the percentage of animals that display blue patches. (c) The stability of different types of repeats was assayed for their dependence on the MMR genes msh-2 and msh-6. The absence of blue patches is indicated by -, while a frequent occurrence of blue-patched animals is indicated by +.

We also tested a number of constructs that differed in the type or length of the repeat present between the ATG and the ß-galactosidase coding regions (Fig 4C). Whereas MMR-proficient C. elegans transgenics failed to display significant somatic frameshift errors, RNAi of msh-2 or msh-6 resulted in the patched phenotype for all four transgenes tested, further demonstrating the requirement for MMR to stabilize repeated sequences in the C. elegans genome.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Most research on MMR function in vivo has focused either on unicellular organisms (because in those one can easily monitor mutator effects in large numbers of progeny) or on somatic cells or tissue culture cells of higher animals. The number of progeny animals that need to be inspected to recognize spontaneous mutants (that are not induced by chemicals or radiation) is prohibitively large. C. elegans, however, is a multicellular organism that can be studied in large numbers, allowing a systematic assessment of MMR in both somatic and germline tissues. Another advantage of this genetic system is that animals can be maintained as self-fertilizing hermaphrodites (spontaneous new mutations can homozygose in self-progeny) so that recessive mutations that lead to a phenotype will become visible in the next generation.

We show that loss of MMR in C. elegans results in a drastic increase in the level of spontaneous mutagenesis in both the male and female germlines. This severely affects the integrity of the hereditary material, thus threatening the existence of subsequent generations. We can extrapolate the total number of mutations that occur in MMR-deficient C. elegans, considering that the target region we assayed to quantify the mutation frequency covers 7% of the total diploid genome. Because a high number of target genes are analyzed simultaneously, a possible sequence bias is ruled out. Considering that loss of essential genes is only part of the analysis (roughly estimated to account for 10% of all genes), we calculate the total number of loss-of-function mutations that occur in MMR-deficient animals to be approximately one to two per animal per generation. This high number predicts that growth of the animals (by self-fertilization, thus preventing mating and the exchange of genetic material) leads to rapid loss in fecundity. Indeed, when 50 individual msh-6 lines were grown separately (by transferring single worms each generation), 57% became sterile before the tenth generation. Recently and independently, such reduced survival after serial passage was also shown—with elevated levels of microsatellite instability and increased mutation frequency—in a strain in which the C. elegans msh-2 gene is disrupted by a transposable element (DEGTYAREVA et al. 2002 ).

One could in principle culture such strains (in population) for multiple generations and achieve quite significant accumulated levels of mutations, while maintaining selection pressure for viability. Possibly a strain like this may be of use in experiments aiming at experimental quantitative genetics, where genetic adaptation to specific environmental challenges can be studied more efficiently than in a wild-type isolate, because the rate of evolution is enhanced.

We found that inactivation of msh-2 or msh-6 results in a similar increase in the frequency of spontaneous mutagenesis in both the germline and somatic tissue. This situation differs from yeast and mammals where, in addition to msh-6, loss of msh-3 is required to phenocopy the msh-2 knockout to result in a complete MMR defect. In these systems two distinct protein complexes (both containing MSH2) with different substrate specificity function in damage recognition: MSH2/MSH6 acts primarily on single-base mispairs whereas the MSH2/ MSH3 complex has a higher affinity for insertion/deletion loops (STRAND et al. 1993 ; JOHNSON et al. 1996 ; MARSISCHKY et al. 1996 ; SIA et al. 1997 ; UMAR et al. 1998 ). The phenotypic similarity in msh-2- and msh-6-defective C. elegans, together with the notion that the sequenced C. elegans genome does not contain an ortholog of the msh-3 gene, strongly suggest that this species has only one damage recognition complex: Inactivating one partner of this MSH2/MSH6 heterodimer knocks out MMR. At present we do not know whether MSH-6 compensates for the absence of MSH-3. In other words, is the substrate specificity of C. elegans MSH2/MSH-6 greater than that in other animal systems? Alternatively, C. elegans just fails to repair larger mispairs that require the action of a specific MSH2/MSH3 complex. We found that MSH-6 is required to suppress tract expansions/contractions of endogenous mono-, di-, and trinucleotide microsatellite repeats, suggesting a role for this protein in the repair of insertion/deletion loops of at least 3 nt.

As stated, a significant fraction of human tumors is apparently caused by somatic mutations in genes that affect genome stability, but not in all cases are these mutations in genes of the known MMR system. There seems to be no direct way to identify these genes, while they may be highly relevant as causative agents of human cancers. We here describe a C. elegans system that mimics the somatic repeat stability in human cancers, in that there is also an msh-2- and msh-6-dependent repeat length change, but here it can be recognized by an intense blue staining. Since feeding dsRNA homologous to MMR genes can also induce this effect, we now have a system to test any C. elegans gene for its role in repressing repeat length changes. Recently genome-wide libraries of dsRNAs of C. elegans have been described (FRASER et al. 2000 ; GONCZY et al. 2000 ), and we are currently testing all genes in this animal's genome for their mutator effect. If additional classes of mutator genes exist, possibly not at all related to MMR, but perhaps to replication factors, chromatin proteins that protect the genome, or totally novel protection systems, they can now be discovered. Eventually, possible human homologs can be tested for their role in human cancer etiology.

	ACKNOWLEDGMENTS

We thank Rik Korswagen, Edwin Cuppen, and Rene Ketting for critically reading the manuscript; members of the Plasterk Lab for discussions; Andy Fire and Julie Ahringer for reagents and protocols; and the Caenorhabditis Genetic Stock Centre for providing some of the strains used in this study.

Manuscript received January 14, 2002; Accepted for publication March 9, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AALTONEN, L. A., P. PELTOMAKI, F. S. LEACH, P. SISTONEN, and L. PYLKKANEN et al., 1993  Clues to the pathogenesis of familial colorectal cancer. Science 260:812-816[Abstract/Free Full Text].

BACON, A. L., M. G. DUNLOP, and S. M. FARRINGTON, 2001  Hypermutability at a poly(A/T) tract in the human germline. Nucleic Acids Res. 29:4405-4413[Abstract/Free Full Text].

BAKER, S. M., A. W. PLUG, T. A. PROLLA, C. E. BRONNER, and A. C. HARRIS et al., 1996  Involvement of mouse Mlh1 in DNA mismatch repair and meiotic crossing over. Nat. Genet. 13:336-342[Medline].

BHATTACHARYYA, N. P., A. SKANDALIS, A. GANESH, J. GRODEN, and M. MEUTH, 1994  Mutator phenotypes in human colorectal carcinoma cell lines. Proc. Natl. Acad. Sci. USA 91:6319-6323[Abstract/Free Full Text].

BRENNER, S., 1974  The genetics of Caenorhabditis elegans.. Genetics 77:71-94[Abstract/Free Full Text].

BRONNER, C. E., S. M. BAKER, P. T. MORRISON, G. WARREN, and L. G. SMITH et al., 1994  Mutation in the DNA mismatch repair gene homologue hMLH1 is associated with hereditary non-polyposis colon cancer. Nature 368:258-261[Medline].

BUERMEYER, A. B., S. M. DESCHENES, S. M. BAKER, and R. M. LISKAY, 1999  Mammalian DNA mismatch repair. Annu. Rev. Genet. 33:533-564[Medline].

DEGTYAREVA, N. P., P. GREENWELL, E. R. HOFMANN, M. O. HENGARTNER, and L. ZHANG et al., 2002  Caenorhabditis elegans DNA mismatch repair gene msh-2 is required for microsatellite stability and maintenance of genome integrity. Proc. Natl. Acad. Sci. USA 99:2158-2163[Abstract/Free Full Text].

DE STASIO, E., C. LEPHOTO, L. AZUMA, C. HOLST, and D. STANISLAUS et al., 1997  Characterization of revertants of unc-93(e1500) in Caenorhabditis elegans induced by N-ethyl-N-nitrosourea. Genetics 147:597-608[Abstract].

DE WIND, N., M. DEKKER, A. BERNS, M. RADMAN, and H. TE RIELE, 1995  Inactivation of the mouse Msh2 gene results in mismatch repair deficiency, methylation tolerance, hyperrecombination, and predisposition to cancer. Cell 82:321-330[Medline].

EDELMANN, W., P. E. COHEN, M. KANE, K. LAU, and B. MORROW et al., 1996  Meiotic pachytene arrest in MLH1-deficient mice. Cell 85:1125-1134[Medline].

EDELMANN, W., K. YANG, A. UMAR, J. HEYER, and K. LAU et al., 1997  Mutation in the mismatch repair gene Msh6 causes cancer susceptibility. Cell 91:467-477[Medline].

ELLEGREN, H., 2000  Microsatellite mutations in the germline: implications for evolutionary inference. Trends Genet. 16:551-558[Medline].

FIRE, A., S. XU, M. K. MONTGOMERY, S. A. KOSTAS, and S. E. DRIVER et al., 1998  Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans.. Nature 391:806-811[Medline].

FISHEL, R., M. K. LESCOE, R. M. RAO, N. G. COPEL, and N. A. JENKINS et al., 1993  The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75:1027-1038[Medline].

FLORES-ROZAS, H. and R. D. KOLODNER, 1998  The Saccharomyces cerevisiae MLH3 gene functions in MSH3-dependent suppression of frameshift mutations. Proc. Natl. Acad. Sci. USA 95:12404-12409[Abstract/Free Full Text].

FRASER, A. G., R. S. KAMATH, P. ZIPPERLEN, M. MARTINEZ-CAMPOS, and M. SOHRMANN et al., 2000  Functional genomic analysis of C. elegans chromosome I by systematic RNA interference. Nature 408:325-330[Medline].

GONCZY, P., G. ECHEVERRI, K. OEGEMA, A. COULSON, and S. J. JONES et al., 2000  Functional genomic analysis of cell division in C. elegans using RNAi of genes on chromosome III. Nature 408:331-336[Medline].

GREENE, C. N. and S. JINKS-ROBERTSON, 1997  Frameshift intermediates in homopolymer runs are removed efficiently by yeast mismatch repair proteins. Mol. Cell. Biol. 17:2844-2850[Abstract].

GREENWALD, I. S. and H. R. HORVITZ, 1980  unc-93(e1500): a behavioral mutant of Caenorhabditis elegans that defines a gene with a wild-type null phenotype. Genetics 96:147-164[Abstract/Free Full Text].

HENDERSON, S. T. and T. D. PETES, 1992  Instability of simple sequence DNA in Saccharomyces cerevisiae.. Mol. Cell. Biol. 12:2749-2757[Abstract/Free Full Text].

IONOV, Y., M. A. PEINADO, S. MALKHOSYAN, D. SHIBATA, and M. PERUCHO, 1993  Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363:558-561[Medline].

JANSEN, G., E. HAZENDONK, K. L. THIJSSEN, and R. H. PLASTERK, 1997  Reverse genetics by chemical mutagenesis in Caenorhabditis elegans.. Nat. Genet. 17:119-121[Medline].

JIRICNY, J., 1998  Replication errors: challenging the genome. EMBO J. 17:6427-6436[Medline].

JOHNSON, R. E., G. K. KOVVALI, L. PRAKASH, and S. PRAKASH, 1996  Requirement of the yeast MSH3 and MSH6 genes for MSH2-dependent genomic stability. J. Biol. Chem. 271:7285-7288[Abstract/Free Full Text].

KOLODNER, R., 1996  Biochemistry and genetics of eukaryotic mismatch repair. Genes Dev. 10:1433-1442[Free Full Text].

LEACH, F. S., N. C. NICOLAIDES, N. PAPADOPOULOS, B. LIU, and B. J. JEN et al., 1993  Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 75:1215-1225[Medline].

LIPKIN, S. M., V. WANG, D. L. STOLER, G. R. ANDERSON, and I. KIRSCH et al., 2001  Germline and somatic mutation analyses in the DNA mismatch repair gene MLH3: evidence for somatic mutation in colorectal cancers. Hum. Mutat. 17:389-396[Medline].

LIU, B., R. E. PARSONS, S. R. HAMILTON, G. M. PETERSEN, and H. T. LYNCH et al., 1994  hMSH2 mutations in hereditary nonpolyposis colorectal cancer kindreds. Cancer Res. 54:4590-4594[Abstract/Free Full Text].

LOEB, L. A., 1991  Mutator phenotype may be required for multistage carcinogenesis. Cancer Res. 51:3075-3079[Free Full Text].

MARSISCHKY, G. T., N. FILOSI, M. F. KANE, and R. KOLODNER, 1996  Redundancy of Saccharomyces cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair. Genes Dev. 10:407-420[Abstract/Free Full Text].

MARTIN, R. H., J. GREEN, E. KO, L. BARCLAY, and A. W. RADEMAKER, 2000  Analysis of aneuploidy frequencies in sperm from patients with hereditary nonpolyposis colon cancer and an hMSH2 mutation. Am. J. Hum. Genet. 66:1149-1152[Medline].

MELLO, C. C., J. M. KRAMER, D. STINCHCOMB, and V. AMBROS, 1991  Efficient gene transfer in C. elegans: extrachromosomal maintenance and integration of transforming sequences. EMBO J. 10:3959-3970[Medline].

MODRICH, P. and R. LAHUE, 1996  Mismatch repair in replication fidelity, genetic recombination, and cancer biology. Annu. Rev. Biochem. 65:101-133[Medline].

NARAYANAN, L., J. A. FRITZELL, S. M. BAKER, R. M. LISKAY, and P. M. GLAZER, 1997  Elevated levels of mutation in multiple tissues of mice deficient in the DNA mismatch repair gene Pms2.. Proc. Natl. Acad. Sci. USA 94:3122-3127[Abstract/Free Full Text].

NICOLAIDES, N. C., N. PAPADOPOULOS, B. LIU, Y. F. WEI, and K. C. CARTER et al., 1994  Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 371:75-80[Medline].

PAPADOPOULOS, N., N. C. NICOLAIDES, Y. F. WEI, S. M. RUBEN, and K. C. CARTER et al., 1994  Mutation of a mutL homolog in hereditary colon cancer. Science 263:1625-1629[Abstract/Free Full Text].

PEINADO, M. A., S. MALKHOSYAN, A. VELAZQUEZ, and M. PERUCHO, 1992  Isolation and characterization of allelic losses and gains in colorectal tumors by arbitrarily primed polymerase chain reaction. Proc. Natl. Acad. Sci. USA 89:10065-10069[Abstract/Free Full Text].

PELTOMAKI, P. and A. DE LA CHAPELLE, 1997  Mutations predisposing to hereditary nonpolyposis colorectal cancer. Adv. Cancer Res. 71:93-119[Medline].

PELTOMAKI, P., R. A. LOTHE, L. A. AALTONEN, L. PYLKKANEN, and M. NYSTROM-LAHTI et al., 1993  Microsatellite instability is associated with tumors that characterize the hereditary non-polyposis colorectal carcinoma syndrome. Cancer Res. 53:5853-5855[Abstract/Free Full Text].

PROLLA, T. A., S. M. BAKER, A. C. HARRIS, J. L. TSAO, and X. YAO et al., 1998  Tumour susceptibility and spontaneous mutation in mice deficient in Mlh1, Pms1 and Pms2 DNA mismatch repair. Nat. Genet. 18:276-279[Medline].

REITMAIR, A. H., R. SCHMITS, A. EWEL, B. BAPAT, and M. REDSTON et al., 1995  MSH2 deficient mice are viable and susceptible to lymphoid tumours. Nat. Genet. 11:64-70[Medline].

ROSENBLUTH, R. E., C. CUDDEFORD, and D. L. BAILLIE, 1983  A rapid eukaryotic mutagen test system using the reciprocal translocation, eT1(III;V). Mutat. Res. 110:39-48.

SIA, E. A., R. J. KOKOSKA, M. DOMINSKA, P. GREENWELL, and T. D. PETES, 1997  Microsatellite instability in yeast: dependence on repeat unit size and DNA mismatch repair genes. Mol. Cell. Biol. 17:2851-2858[Abstract].

STRAND, M., T. A. PROLLA, R. M. LISKAY, and T. D. PETES, 1993  Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair. Nature 365:274-276[Medline].

SZANKASI, P. and G. R. SMITH, 1995  A role for exonuclease I from S. pombe in mutation avoidance and mismatch correction. Science 267:1166-1169[Abstract/Free Full Text].

TIMMONS, L. and A. FIRE, 1998  Specific interference by ingested dsRNA. Nature 395:854[Medline].

UMAR, A., J. I. RISINGER, W. E. GLAAB, K. R. TINDALL, and J. C. BARRETT et al., 1998  Functional overlap in mismatch repair by human MSH3 and MSH6. Genetics 148:1637-1646[Abstract/Free Full Text].

WU, Y., M. J. BERENDS, R. H. SIJMONS, R. G. MENSINK, and E. VERLIND et al., 2001a  A role for MLH3 in hereditary nonpolyposis colorectal cancer. Nat. Genet. 29:137-138[Medline].

WU, Y., M. J. BERENDS, J. G. POST, R. G. MENSINK, and E. VERLIND et al., 2001b  Germline mutations of EXO1 gene in patients with hereditary nonpolyposis colorectal cancer (HNPCC) and atypical HNPCC forms. Gastroenterology 120:1580-1587[Medline].

This article has been cited by other articles:

				H. Feitsma, A. Akay, and E. Cuppen Alkylation damage causes MMR-dependent chromosomal instability in vertebrate embryos Nucleic Acids Res., July 1, 2008; 36(12): 4047 - 4056. [Abstract] [Full Text] [PDF]

				J. L. Yanowitz Genome Integrity Is Regulated by the Caenorhabditis elegans Rad51D Homolog rfs-1 Genetics, May 1, 2008; 179(1): 249 - 262. [Abstract] [Full Text] [PDF]

				P. D. Hoffman, J. M. Leonard, G. E. Lindberg, S. R. Bollmann, and J. B. Hays Rapid accumulation of mutations during seed-to-seed propagation of mismatch-repair-defective Arabidopsis Genes & Dev., November 1, 2004; 18(21): 2676 - 2685. [Abstract] [Full Text] [PDF]

				J. Pothof, G. van Haaften, K. Thijssen, R. S. Kamath, A. G. Fraser, J. Ahringer, R. H.A. Plasterk, and M. Tijsterman Identification of genes that protect the C. elegans genome against mutations by genome-wide RNAi Genes & Dev., February 15, 2003; 17(4): 443 - 448. [Abstract] [Full Text] [PDF]

				A. Wei, A. Yuan, G. Fawcett, A. Butler, T. Davis, S.-y. Xu, and L. Salkoff Efficient isolation of targeted Caenorhabditis elegans deletion strains using highly thermostable restriction endonucleases and PCR Nucleic Acids Res., October 15, 2002; 30(20): e110 - e110. [Abstract] [Full Text] [PDF]