A Local, High-Density, Single-Nucleotide Polymorphism Map Used to Clone Caenorhabditis elegans cdf-1

Corresponding author: Kerry Kornfeld, Department of Molecular Biology and Pharmacology, Washington University School of Medicine, 660 S. Euclid, St. Louis, MO 63110., kornfeld{at}molecool.wustl.edu (E-mail)

Communicating editor: I. GREENWALD

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Ras-mediated signaling is required for induction of vulval cell fates during Caenorhabditis elegans development. By screening for suppressors of the multivulva phenotype caused by constitutively active let-60 ras, we identified the mutation n2527. To clone the gene affected by n2527, we developed a method for high-resolution mapping. We took advantage of the genomic DNA sequence of the N2 strain by using DNA sequencing to scan for single-nucleotide polymorphisms (SNPs) at defined genomic positions of the RC301 strain. An average of one polymorphism per 1.4 kb was detected in predicted intergenic regions. Because of this high frequency, DNA sequencing is an efficient method to scan for SNPs. By alternating between identifying SNPs and mapping n2527 using selected recombinants, we generated an SNP map of progressively higher density. An intensive search for SNPs resulted in a local map with an average marker spacing of ~4 kb. This was used to map n2527 to a 9.6-kb interval. The small size of this interval made it feasible to use DNA sequencing to identify the molecular lesion. In principle, this approach can be used for high-resolution mapping of any C. elegans mutation. Furthermore, this approach can be applied to other species as the genomic sequence becomes available. The n2527 mutation affects a previously uncharacterized gene that we named cdf-1, as it encodes a predicted protein with significant similarity to members of the cation diffusion facilitator family.

SCREENING for mutants is an important method that has been used to identify genes that mediate a wide variety of biological processes. Because cloning these genes is essential for a comprehensive analysis, improved cloning methods are extremely useful. Mutations are usually induced in Caenorhabditis elegans using ethyl methanesulfonate (EMS) or other chemical mutagens that primarily cause single-base substitutions (BRENNER 1974 ; COULONDRE and MILLER 1977 ); by contrast to gross DNA changes or the insertion of a transposable element, these subtle changes do not facilitate cloning.

Because the genetic and physical maps of the C. elegans genome are well characterized, a gene affected by a chemically induced mutation is typically identified using a positional cloning approach that involves the following three phases:

1. The mutation is positioned on the physical map (COULSON et al. 1988 ). This defines an interval that contains the gene.

2. Transgenic animals containing genomic DNA from this interval cloned in cosmid or YAC vectors are generated, and assays for rescue of the mutant phenotype are conducted (MELLO et al. 1991 ). This approach is used to search for a DNA fragment that contains the mutated gene and then to define a minimal rescuing fragment.

3. Candidate open reading frames (ORFs) are sequenced positioned on the minimal rescuing fragment using DNA from mutant animals to identify the nucleotide change that causes the mutant phenotype.

In practice, precise mapping reduces the difficulty of identifying a rescuing fragment, and precise definition of a minimal rescuing fragment reduces the difficulty of identifying the nucleotide alteration.

A mutation is initially positioned on the physical map relative to genome-wide systems of markers. These include mutations that cause a phenotype and affect a cloned gene, deletions with endpoints that can be characterized (BARSTEAD et al. 1991 ), and polymorphic-sequence-tagged sites caused by Tc1 transposons present in C. elegans strains that have diverged from the wild-type N2 strain (WILLIAMS et al. 1992 ; KORSWAGEN et al. 1996 ). After genome-wide systems of markers are utilized, the interval containing the mutation is often relatively large, and additional markers that could be used for further mapping are desirable. However, the genome-wide systems of markers cannot usually be expanded in defined genomic regions because these markers are typically identified in genome-wide searches. Additional markers have been identified in defined genomic regions by using the physical map to search for restriction fragment length polymorphisms (RFLPs; RUVKUN et al. 1989 ; HODGKIN 1993 ; KORNFELD et al. 1995A , KORNFELD et al. 1995B ; KIMURA et al. 1997 ; OGG et al. 1997 ). While this approach has been extremely useful, it has two significant limitations. First, it is likely that many polymorphisms cannot be detected as RFLPs because they do not affect a restriction enzyme site or grossly alter the DNA. Second, the mapping resolution that can be achieved with RFLPs is limited because the precise position of the nucleotide change(s) responsible for an RFLP is generally unknown. Because of these limitations, mutations are not generally mapped to high resolution, and a significant effort is often required for the transformation rescue phase of positional cloning.

Single-nucleotide polymorphisms (SNPs), a term we will use to refer to the substitution or insertion/deletion of one or a small number of nucleotides, appear to be the most common type of polymorphism in vertebrates (KWOK et al. 1996 ). Significant efforts are being directed at the development of high-throughput approaches to scan for and score SNPs (LAI et al. 1998 ; LANDEGREN et al. 1998 ; WONG et al. 1998 ). To develop a method for high-resolution mapping in C. elegans, we took advantage of the complete genomic sequence by using DNA sequencing to scan for SNPs at defined genomic positions. We found that SNPs are abundant in the RC301 strain compared to the N2 strain. By alternating between scanning for SNPs and mapping, we generated a local SNP map with an average interval size of ~4 kb and mapped the cdf-1(n2527) mutation to a 9.6-kb interval. The molecular lesion in the affected gene was identified by sequencing this interval. This approach is likely to facilitate the positional cloning of any gene that is affected by a mutation in C. elegans, and may be applicable to other organisms as genomic sequence becomes available.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

General methods and strains:
C. elegans strains were cultured as described by BRENNER 1974 and grown at 22.5° unless otherwise noted. The parent strain of all the mutant strains was N2, a wild-type isolate from Bristol, England. RC301 is a wild-type isolate from Freiburg, Germany (HODGKIN 1993 ). C. elegans strains maintained in the laboratory are inbred by self-fertilization and homozygous at essentially all loci. Unless otherwise noted, the mutations used in this study are described by RIDDLE et al. 1997 and are as follows. LGIV: let-60(n1046gf); dpy-20(e1282). LGX: lon-2(e678); mup-2(n2346ts); cdf-1(n2527) (this study); unc-6(e78); unc-10(e102); dpy-7(e88); uDf1.

Genetic analyses:
We previously described a screen for suppressors of the let-60(n1046gf) multivulva (Muv) phenotype (LACKNER et al. 1994 ; KORNFELD et al. 1995A , KORNFELD et al. 1995B ; JACOBS et al. 1998 ). In brief, we mutagenized let-60(n1046gf) hermaphrodites with EMS, placed 2794 F₁ self-progeny on separate Petri dishes, and examined the F₂ self-progeny for non-Muv animals at 22.5°. We identified 33 independently derived mutations that reduced the penetrance of the Muv phenotype from 93% to <10%, including the n2527 mutation. Ten additional suppressor mutations that meet these criteria were identified in a related screen (BEITEL et al. 1990 ).

The suppression of the let-60(gf) Muv phenotype caused by n2527 displayed linkage to lon-2 on chromosome X (data not shown). The seven other suppressor mutations that are positioned on chromosome X complemented the n2527 suppression of let-60(gf) Muv phenotype, indicating that the complementation group defined by n2527 contains only one mutation (data not shown). Three-factor crosses were used to more precisely map n2527. Of uncoordinated (Unc) non-dumpy (Dpy) self-progeny from let-60(gf); unc-10 dpy-6/ n2527 hermaphrodites, 0/10 segregated n2527. From let-60(gf); unc-6 dpy-7/n2527 hermaphrodites, 0/17 Unc non-Dpy self-progeny and 17/18 Dpy non-Unc self-progeny segregated n2527. These data suggest n2527 is positioned left of unc-10 and unc-6. Of Unc non-muscle positioning abnormal (non-Mup) self-progeny from let-60(gf); mup-2 unc-6/n2527 hermaphrodites, 18/25 segregated n2527. These data suggest n2527 is positioned between mup-2 and unc-6, an ~0.6-map-unit interval (Fig 1A).

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. High-resolution mapping of n2527 relative to SNPs. (A) A portion of the genetic map of chromosome X. Loci defined by mutations that cause visible phenotypes are indicated above, and approximate distances in map units are shown below. Three-factor mapping experiments were used to position n2527 between mup-2 and unc-6. (B) A total of 201 Lon non-Unc self-progeny were selected from the indicated heterozygote. A total of 164 strains displayed the Muv phenotype (top), indicating they lack n2527, and 37 strains displayed the non-Muv phenotype (bottom), indicating they contain n2527. Each group was subdivided by analyzing the DNA sequence at the position of SNPs. N indicates the sequence of the N2-derived lon-2 n2527 unc-6 strain, and R indicates the sequence of the RC301 strain. Scoring amP9 gave the same results shown for amP10. nd, not determined. (C) Horizontal lines represent the physical map; vertical lines indicate cloned genes and polymorphisms (not drawn to scale). The approximate distance between markers is shown in kilobases. +, wild-type N2 sequence. Polymorphisms are changes in the RC301 strain (shown below), except amP15 and amP16, which are changes in the lon-2 n2527 unc-6 strain (shown above); at these two positions, the RC301 strain has the wild-type (+) sequence. The data shown in B were used to determine the number of crossovers that occurred in each interval. These data position n2527 between amP14 and amP16. *, of these 159 strains, 19 were scored at amP4 but not amP5 (B, line 2); thus, the crossovers in these strains may have occurred between amP5 and amP4.

To generate a lon-2 n2527 unc-6 chromosome, we first identified a lon-2 n2527 chromosome by selecting long (Lon) non-Muv non-Unc self-progeny of let-60(gf); lon-2 unc-6/n2527 hermaphrodites. Second, we identified a n2527 unc-6 chromosome by selecting Unc non-Muv non-Mup self-progeny of let-60(gf); mup-2 unc-6/n2527 hermaphrodites. Third, we identified a lon-2 n2527 unc-6 chromosome by selecting Lon self-progeny of let-60(gf); lon-2 n2527/n2527 unc-6 hermaphrodites and identifying an animal that segregated Unc progeny.

To map n2527 relative to SNPs, we mated RC301 males and let-60(gf); lon-2 n2527 unc-6 hermaphrodites, placed cross-progeny on separate Petri dishes, and then identified self-progeny of genotype let-60(gf); lon-2 n2527 unc-6/RC301. We picked 201 Lon non-Unc self-progeny to separate Petri plates, identified self-progeny homozygous for the recombinant chromosome, and scored the penetrance of the Muv phenotype in these strains.

Polymerase chain reaction and DNA sequencing:
Unless otherwise noted, molecular biology techniques were performed as described by SAMBROOK et al. 1989 . The genomic DNA sequence of the N2 strain (C. ELEGANS SEQUENCING CONSORTIUM 1998) was used to design oligonucleotide primers that were used to scan for polymorphisms and sequence the region between amP4 and amP11. Table 1 shows oligonucleotide primers that can be used to identify polymorphisms. Genomic DNA was prepared and PCR amplified as described by WILLIAMS et al. 1992 . Amplification products were fractionated by electrophoresis in low-melting-temperature agarose, purified using ß-agarase (New England Biolabs, Beverly, MA), and used as templates for cycle sequencing reactions that were analyzed using an automated ABI 373A sequencer (Applied Biosystems, Foster City, CA). DNA sequence data were analyzed using the Sequencher 3.0 computer program (Gene Codes Corp., Ann Arbor, MI).

DNA cloning:
To subclone the region containing the predicted open reading frame C15B12.7, we digested cosmid C15B12 (received from A. Coulson, Sanger Center) with SpeI and SacI to generate a 6239-bp DNA fragment that extends from 1612 bp upstream of the predicted START codon to 384 bp downstream from the predicted STOP codon of C15B12.7; this fragment contains no coding sequences from adjacent predicted genes. This fragment was ligated into pBluescript digested with SpeI and SacI to create pJJ4. To delete the majority of the C15B12.7 open reading frame, we digested pJJ4 with XbaI, purified the plasmid backbone, and ligated to recircularize. This plasmid was named pJJ5. This procedure removed a 3704-bp XbaI fragment that contains predicted exons 2–7 of C15B12.7. The n2527 mutation, a G-to-A transition at cosmid position 39081, was engineered into pJJ4 by replacing a 1427-bp NcoI (cosmid position 38840)/KasI (cosmid position 40267) with the equivalent NcoI/KasI fragment derived from the PCR-amplified DNA of let-60(gf); n2527 animals. This plasmid was named pJJ6. DNA sequencing of the NcoI/KasI fragment of pJJ6 revealed no additional changes compared to wild type.

Germ-line transformation experiments:
Germ-line transformation experiments were performed as described by MELLO et al. 1991 . We coinjected let-60(gf); n2527 mutants with plasmid pRF4 (80–90 µg/ml), which contains the dominant mutation rol-6(su1006), and cosmid C15B12 (50 µg/ml) or plasmids pJJ4, pJJ5, or pJJ6 (10–20 µg/ml). We established an independently derived transgenic strain from each F₁ animal that displayed the Rol phenotype and segregated F₂ progeny that displayed the Rol phenotype. If >60% of the Rol animals from a transgenic strain were Muv, then we concluded that the introduced DNA rescued the n2527 suppression of let-60(gf) Muv phenotype.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of n2527 and mapping using genome-wide systems of markers:
The C. elegans vulva, a specialized epidermal structure used for egg laying and sperm entry, is formed by the descendants of P5.p, P6.p, and P7.p (reviewed by HORVITZ and STERNBERG 1991 ). In wild-type hermaphrodites, the anchor cell of the somatic gonad activates a conserved receptor tyrosine kinase-Ras-mitogen-activated protein (MAP) kinase signaling pathway in P6.p, causing P6.p to adopt the 1° vulval cell fate (eight descendants, reviewed by KORNFELD 1997 ). P6.p then signals to P5.p and P7.p, causing these cells to adopt the 2° vulval cell fate (seven descendants). P3.p, P4.p, and P8.p are capable of adopting vulval fates, but appear to receive neither of these signals; thus, these cells adopt the nonvulval 3° cell fate (two descendants). In hermaphrodites with a gain-of-function mutation that constitutively activates the let-60 ras gene, P3.p, P4.p, and P8.p often inappropriately adopt vulval cell fates; the resulting ectopic tissue forms a series of protrusions along the ventral side of the animal, which is called the multivulva (Muv) phenotype. To identify genes that are involved in this Ras-mediated signaling event, we screened for mutations that suppress the let-60(n1046gf) Muv phenotype (see MATERIALS AND METHODS). We isolated 43 mutations that define 21 complementation groups, including the mutation n2527.

The n2527 mutation is an effective suppressor of the phenotype caused by constitutively active Ras, since it reduced the penetrance of the Muv phenotype from 84 to 2% (Table 2). The n2527 mutation is weakly semidominant. Mutants containing n2527 in trans to a deficiency displayed a phenotype similar to mutants homozygous for n2527 (Table 2), indicating that n2527 is a loss-of-function mutation. In a wild-type genetic background, n2527 did not cause a significant penetrance of vulval defects or other visible phenotypes.

To clone the gene affected by n2527, we first positioned it on the genetic and physical maps using mutations that cause visible phenotypes and affect cloned genes. The n2527 mutation displayed linkage to chromosome X, and three-factor mapping experiments indicated that n2527 is positioned between mup-2 and unc-6 (Fig 1A; see MATERIALS AND METHODS). An analysis of the completed genomic DNA sequence of this region indicates that mup-2 and unc-6 are separated by ~450 kb (C. ELEGANS SEQUENCING CONSORTIUM 1998). We did not utilize odr-10, the one remaining visible marker in this interval that has been cloned, because odr-10 mutations cause a behavioral phenotype that is not readily scored.

Identification of SNPs at defined genomic positions:
We next considered two approaches: generating a collection of transgenic animals containing fragments of genomic DNA spanning this interval cloned in cosmid vectors to identify a fragment that can rescue the n2527 mutant phenotype, or additional mapping to more precisely position n2527 on the physical map. We rejected the first approach because it is relatively laborious and there was a chance that it would not succeed, since some of the DNA in the interval was not present in cosmid vectors. To pursue the second approach, we needed to identify many additional markers in this interval, since only two polymorphisms had been reported. We hypothesized that we could exploit our knowledge of the genomic DNA sequence of this interval to scan for SNPs at defined positions by sequencing small fragments of DNA from an evolutionarily diverged wild-type strain. Compared to scanning for RFLPs, this approach has the advantages that any sequence difference can be detected and the position of the polymorphism is known precisely. However, SNPs must occur relatively frequently for this approach to be practical.

The n2527 mutation was generated in a strain derived from the N2 strain, a wild-type isolate from England (BRENNER 1974 ). We chose to look for polymorphisms in the RC301 strain, a wild-type isolate from Germany, since this strain has been used successfully to identify RFLPs in multiple regions of the genome (HODGKIN 1993 ; KORNFELD et al. 1995A , KORNFELD et al. 1995B ; KIMURA et al. 1997 ; OGG et al. 1997 ). Our standard approach to identify a polymorphism was to use the genomic sequence of the N2 strain to design a pair of oligonucleotide primers that amplify a DNA fragment at a defined genomic position. In the first phase of the mapping experiment, primers were designed to amplify predicted intergenic regions, since we reasoned that polymorphisms would be more frequent in noncoding sequence. In the second phase, predicted exons and introns were also scanned to maximize the density of SNPs. The primers were composed of 20 bases and 50% G/C. Products were amplified using DNA from both RC301 and the N2-derived let-60(n1046gf); lon-2 n2527 unc-6 strain, purified from agarose gels, and used as templates for sequencing using the amplification primers and an ABI automated sequencer. We designed primers that were separated by ~1.2 kb to fully exploit the capability of the automated sequencer—up to 600 bases per primer.

Generation of a progressively higher density SNP map by alternating between identifying SNPs and mapping n2527:
In general, a crossover that occurs between n2527 and a polymorphism can be used to determine marker order. To select crossovers near n2527, we used lon-2 and unc-6, visible markers that flank n2527 and are separated by ~1900 kb (Fig 1A). Although genetic mapping positioned n2527 right of mup-2, the more distal marker lon-2 was used to select recombinants because the mup-2 phenotype is not suitable for this procedure. From let-60(n1046gf); lon-2 n2527 unc-6/RC301 animals, we selected 201 Lon non-Unc self-progeny and then identified hermaphrodites homozygous for the recombinant chromosome. A total of 164 strains displayed the Muv phenotype, indicating that they lost n2527 and the crossover occurred between lon-2 and n2527, while 37 strains displayed the non-Muv phenotype, indicating that they contained n2527 and the crossover occurred between n2527 and unc-6. To define a small interval that contains n2527, we pursued two goals: First, the identification of the recombinants in which the crossover breakpoint occurred closest to the left and right of n2527; second, the identification of polymorphisms positioned close to but outside of these closest crossovers. These polymorphisms define an interval that contains n2527.

To identify a polymorphism close to and left of n2527, we first scanned for an SNP near mup-2 to distinguish crossovers that occurred right of mup-2 and, thus, close to n2527 from crossovers that occurred left of mup-2. Two pairs of primers were used to amplify and sequence ~2 kb of DNA; one A-to-C substitution was detected in RC301 and designated amP5 (Table 1). amP5 is 2 kb left of mup-2. DNA sequencing was used to score amP5 in 145 Lon non-n2527 non-Unc recombinants; 140 recombinants contained amP5 (the RC301 sequence), indicating that these crossovers occurred left of amP5, whereas 5 had the wild-type (N2) sequence, indicating that these crossovers occurred right of amP5 (Fig 1B and Fig C). These results identify five crossovers between amP5 and n2527, and they indicate that n2527 is right of amP5. We next scanned for a polymorphism positioned ~90 kb right of amP5 by sequencing ~1.4 kb of DNA; one C-to-A substitution was detected in RC301 and designated amP4 (Table 1). DNA sequencing was used to score amP4 in the five recombinants with crossovers right of amP5; three contained amP4 and two had the wild-type sequence (Fig 1B and Fig C). These results identify two crossovers that occurred between amP4 and n2527, and they indicate that n2527 is right of amP4. Having established that n2527 is right of amP4, we scored amP4 in the remaining 19 Lon non-n2527 non-Unc recombinants to determine if any of these crossovers occurred right of amP4. However, all 19 contained amP4 (the RC301 sequence; Fig 1B and Fig C).

To identify a polymorphism close to and right of n2527, we scanned for polymorphisms left of unc-6. Four pairs of primers were used to sequence ~4.2 kb of DNA, and four SNPs were detected in RC301: a C-to-A substitution designated amP1, a 2-bp deletion designated amP2, an AG-to-GC substitution designated amP3, and a 1-bp insertion designated amP13 (Table 1). amP3 is positioned ~170 kb left of unc-6. DNA sequencing was used to score amP3 in the 37 Lon n2527 non-Unc recombinants; 16 had amP3 and 21 had the wild-type sequence (Fig 1B and Fig C). These results identify 16 crossovers that occurred between n2527 and amP3, and they indicate that n2527 is left of amP3. Since amP1, amP2, and amP13 are close to or right of amP3, we did not score these polymorphisms. We next scanned for polymorphisms left of amP3. About 3.6 kb of DNA was sequenced, and two SNPs were detected in RC301: a G-to-T substitution designated amP6, and a C-to-A substitution designated amP7 that is positioned 133 kb left of amP3 (Table 1). A total of 10 of the 16 recombinants had amP7, and 6 had the wild-type sequence (Fig 1B and Fig C). Thus, n2527 is left of amP7. We next sequenced ~2 kb of DNA and identified an A-to-T substitution in RC301, designated amP8, that is positioned 28 kb left of amP7 (Table 1). A total of 4 of the 10 recombinants had amP8, and 6 had the wild-type sequence (Fig 1B and Fig C). These data indicate that n2527 is left of amP8, and they identify four crossovers that occurred between n2527 and amP8. We next sequenced ~1.5 kb and identified an insertion of ~300 bp in RC301, designated amP11, that is positioned 8 kb left of amP8. Because this insertion is relatively large, it could be scored by gel electrophoresis of PCR products. All four recombinants had amP11, indicating that n2527 is left of amP11 (Fig 1B and Fig C). To summarize, of 164 crossovers that occurred between lon-2 and n2527, we identified 2 that occurred between amP4 and n2527. Of 37 crossovers that occurred between n2527 and unc-6, we identified 4 that occurred between n2527 and amP11. Thus, amP4 and amP11 define an ~19-kb interval that contains n2527.

Generation of a local, high-density SNP map and identification of the n2527 molecular lesion:
Thus far, we had generated an SNP map of progressively higher density centered on the n2527 mutation. We next used DNA sequencing to pursue two goals: first, the generation of an extremely high density map that could be used for further mapping; second, the identification of the n2527 molecular lesion. The 19-kb interval between amP4 and amP11 is predicted to contain four ORFs by the GeneFinder computer program. We determined the sequence of the coding regions and introns of the two centrally positioned, predicted ORFs, designated C15B12.7 and F22A3.3, as well as most of the intergenic region using DNA from both RC301 and the let-60(n1046gf); lon-2 n2527 unc-6 strain (Fig 2A). After analyzing ~2.8 kb of sequence from coding regions, we identified two polymorphisms in the n2527 strain: a G-to-A substitution designated amP15 that changes the predicted codon 186 of C15B12.7 to STOP, and a C-to-T substitution designated amP16 that affects a codon in exon 7 of F22A3.2 but does not result in an amino acid change (Table 1 and Fig 2A). These nucleotide changes were candidates for the n2527 molecular lesion. After analyzing ~6.1 kb of sequence from introns and intergenic regions, we identified four SNPs in RC301: a 1-bp deletion designated amP14, an A-to-G substitution designated amP9, an A-to-G substitution designated amP12, and an A-to-T substitution designated amP10 (Table 1 and Fig 2A).

View larger version (21K): In this window In a new window Download PPT slide   Figure 2. The gene affected by the n2527 mutation is predicted ORF C15B12.7. (A) The horizontal line represents the physical map; thick boxes are predicted exons, and thin lines are predicted introns or intergenic regions. Except for regions marked by diagonal lines, the map uses the scale shown below. The four predicted open reading frames in this region are labeled above; the 5' and 3' ends of C15B12.7 and F22A3.3 are indicated. We determined the DNA sequence of regions shown in black using DNA from the n2527 and RC301 strains; vertical lines indicate the positions of SNPs. Polymorphisms in regular type were identified in RC301; amP15 and amP16 were identified in the n2527 strain (shown lower in bold type). (B) Plasmids listed on the left contained the indicated fragment of genomic DNA. Vertical lines indicate restriction enzyme cleavage sites used for cloning: Sp, SpeI; X, XbaI; Sa, SacI. pJJ5 lacks the XbaI fragment, and pJJ6 contains a C-to-T substitution that changes predicted codon 186 from Gln to STOP. The number of independently derived transgenic strains that displayed rescue of the n2527 suppression of let-60(gf) Muv phenotype and the total number of strains analyzed are indicated (Table 2; see MATERIALS AND METHODS).

We used the polymorphisms from both strains to further position n2527. We analyzed amP14 in the two recombinants that had crossovers between amP4 and n2527; one contained amP14 and one had the wild-type sequence (Fig 1B and Fig C). These results indicate that n2527 is right of amP14, and they identify one crossover that occurred between amP14 and n2527. We analyzed amP16 in the four recombinants that had crossovers between n2527 and amP11. All four contained the RC301 allele, indicating that n2527 is left of amP16 (Fig 1B and Fig C) and that amP16 is not the n2527 molecular lesion. These results positioned n2527 in a 9.6-kb interval defined by amP14 and amP16 that contains portions of two predicted ORFs, C15B12.7 and F22A3.3 (Fig 2A). The remaining four polymorphisms in this interval, amP15, amP9, amP12, and amP10, were not separated from n2527 by any of the five closest crossovers (Fig 1B and Fig C). These results do not establish the order of n2527 relative to these polymorphisms.

The following evidence suggests that amP15 is the n2527 mutation: (1) amP15 is positioned in the 9.6-kb interval between amP14 and amP16 that contains the n2527 mutation, and amP15 was not separated from n2527 by any crossover; (2) no other changes were detected in the n2527 strain after determining the sequences of all the predicted exons and most of the introns and intergenic regions between amP14 and amP16; (3) amP15 creates a premature STOP codon at predicted amino acid 186 of C15B12.7, resulting in a truncated protein lacking the C-terminal two-thirds of the protein; and (4) amP15 is a G-to-A transition, the most common mutation induced by EMS (COULONDRE and MILLER 1977 ), the mutagen used to generate n2527.

DNA containing C15B12.7 rescues the n2527 mutant phenotype in transgenic animals:
To test the hypothesis that the predicted ORF C15B12.7 is the gene affected by the n2527 mutation, we used a transformation rescue assay. let-60(n1046gf); n2527 animals, which are non-Muv, were transformed with cosmid C15B12 and a plasmid that contains a dominant rol-6 mutation as a transformation marker. Six independently derived transgenic strains that displayed the Rol phenotype were obtained. All six strains displayed the Muv phenotype, indicating that C15B12 rescued the n2527 suppression of the Muv phenotype (data not shown). To determine if predicted ORF C15B12.7 is sufficient to rescue the n2527 mutant phenotype, we constructed a plasmid (pJJ4) that contains the complete predicted ORF C15B12.7 and no other predicted ORFs. All six independently derived transgenic strains containing pJJ4 displayed the Muv phenotype, indicating that C15B12.7 is sufficient for rescuing activity (Fig 2B and Table 2). Plasmids containing a deletion of predicted exons 2–7 of C15B12.7 (pJJ5) or the base change detected in the n2527 strain at predicted codon 186 of C15B12.7 (pJJ6) were used as controls. All three independently derived transgenic strains containing pJJ5 displayed the non-Muv phenotype (Fig 2B and Table 2), indicating that an intact version of C15B12.7 is necessary for rescuing activity. All seven independently derived transgenic strains containing pJJ6 displayed the non-Muv phenotype (Fig 2B and Table 2). This failure to rescue indicates that a nonsense change at predicted codon 186 reduces gene activity, and it supports the hypothesis that this G-to-A nucleotide substitution is the n2527 mutation.

The C. elegans Sequencing Consortium noted that the predicted ORF C15B12.7 encodes a protein that is similar to Saccharomyces cerevisiae COT1. PAULSEN and SAIER 1997 analyzed the sequence of the predicted C15B12.7 protein, COT1, and 11 other similar proteins that comprise the cation diffusion facilitator (CDF) protein family. This analysis showed that the predicted C15B12.7 protein has significant similarity to all the members of this family and contains all six predicted transmembrane domains that are characteristic of CDF proteins. Thus, we named the gene defined by the n2527 mutation cdf-1. Several CDF proteins have been shown to regulate the intracellular concentration of heavy metal ions (KAMIZONO et al. 1989 ; CONKLIN et al. 1992 ; PALMITER and FINDLEY 1995 ; PALMITER et al. 1996 ). Interestingly, CDF proteins have not been reported to be involved in Ras-mediated signaling. Further characterization of cdf-1 is required to determine whether CDF-1 regulates the concentration of heavy metal ions and thereby modulates Ras signaling, or whether CDF-1 has a different mechanism of action.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

SNPs are abundant in RC301:
The data presented here can be used to estimate the abundance and nature of polymorphisms in the strain RC301 compared to N2. We identified 13 polymorphisms in ~18 kb of predicted intergenic DNA—an average of one polymorphism per 1.4 kb. One polymorphism was detected in 3.7 kb of predicted intron DNA, and no polymorphisms were detected in 2.8 kb of predicted exon DNA. The polymorphisms are primarily subtle changes—substitutions or insertions/deletions of one or two nucleotides—although one is a 300-bp insertion. It is likely that polymorphisms occur at a similar frequency throughout the RC301 genome, since RFLPs have been detected in many genomic regions of RC301 and we have begun identifying SNPs at a similar frequency on chromosome II (HODGKIN 1993 ; KORNFELD et al. 1995A , KORNFELD et al. 1995B ; KIMURA et al. 1997 ; OGG et al. 1997 ; J. JAKUBOWSKI and K. KORNFELD, unpublished observations). The frequency of SNPs in other strains of C. elegans has yet to be determined. The C. elegans Sequencing Consortium has initiated a large-scale effort to identify randomly positioned SNPs in the strain CB4856, which was isolated in Hawaii (http://genome.wustl.edu/gsc/). Although many candidate SNPs have been identified, it is difficult to estimate the frequency on the basis of these data, since candidate SNPs have not yet been verified.

These findings have two important implications. First, they demonstrate that DNA sequencing is a practical and efficient method to scan for SNPs at defined genomic positions. Based on an average of one polymorphism per 1.4 kb, only about three oligonucleotide primers and three sequencing reactions are necessary to detect a polymorphism in RC301. Second, these results suggest that a genome-wide search for SNPs could yield a map containing thousands of markers. Such a map would have a significantly higher marker density than the existing genome-wide polymorphism maps, which contain several hundred polymorphisms caused by insertions of Tc1 transposable elements (WILLIAMS et al. 1992 ; KORSWAGEN et al. 1996 ). The C. elegans Sequencing Consortium has initiated such an effort. A genome-wide, moderate-density SNP map could be used to position newly identified mutations to a reasonably small chromosomal interval using unselected recombinants. Such a map could also provide a starting point for the generation of a local, high-density SNP map that can be used for high-resolution mapping with selected recombinants.

A method for high-resolution mapping:
The method presented here involves three main steps: The first is the generation of recombinants with crossovers near the mutation. We used cis-linked visible markers to select recombinants. In our experience, this is the most laborious step. Second, a progressively higher density SNP map is generated by alternating between identifying SNPs at defined genomic locations and mapping. This is both efficient and relatively rapid. By identifying SNPs in the center of the interval, relatively few are required; we used six SNPs to narrow the 450-kb interval between mup-2 and unc-6 to a 19-kb interval between amP4 and amP11. The effort necessary to analyze recombinants decreases with each new SNP as the useful crossovers that occurred closely to the mutation are identified. The third step involves the generation of a local, high-density SNP map and the simultaneous search for the molecular lesion.

Mapping resolution depends on the density of crossovers and markers (in this case, SNPs). We analyzed 201 crossovers between lon-2 and unc-6. This interval is ~1900 kb, and thus we can calculate that crossovers occurred on average every 9.5 kb. We generated a local map with an average interval between SNPs of ~4 kb, considering only RC301 polymorphisms between amP14 and amP11 (Fig 1C). In the region from amP14 to amP16 that was investigated intensely, the average interval between SNPs was ~2.4 kb (Fig 1C). These reagents were used to map the n2527 mutation to a 9.6-kb interval. This resolution is much finer than what has been reported previously in C. elegans. Because the local SNP map was denser than the average interval between crossovers, it was not surprising that the mapping resolution was limited by the density of crossovers rather than by the density of SNPs: three polymorphisms in RC301 could not be positioned relative to n2527. Since C. elegans has an average gene density of about one gene per 5 kb (C. ELEGANS SEQUENCING CONSORTIUM 1998), these results indicate that this method may make it possible to map many mutations to a single gene.

High-resolution mapping is useful and important because it significantly reduces the difficulty of subsequent cloning steps. Because n2527 was mapped to an interval of only 9.6 kb, it was practical to identify the molecular lesion by DNA sequencing and bypass the need for the standard procedure of transformation of mutant worms with genomic DNA to identify a rescuing fragment. This is important because transformation can be laborious and is prone to both false-negative and false-positive results. It is likely that this high-resolution mapping can be used to analyze any C. elegans mutation. High-resolution mapping will be particularly useful for positionally cloning genes identified by mutations that cannot be rescued by injection of wild-type DNA, e.g., gain-of-function mutations or mutations that affect genes that function in the germ line, a tissue in which transformed genes are not expressed efficiently. Furthermore, this approach is likely to be useful in other organisms, such as Drosophila and zebrafish, as genomic sequences become available.

	ACKNOWLEDGMENTS

We thank Jennifer Boots and Drew Syder for assistance with mapping and sequencing and Tim Schedl and Steve Johnson for advice about the manuscript. Some strains were provided by the Caenorhabditis Genetics Center (St. Paul, MN), which is funded by the National Center for Research Resources of the National Institutes of Health. This research was supported by a grant from Monsanto-Searle/Washington University Biomedical Program (K.K.). K.K. is a recipient of the Burroughs Wellcome Fund New Investigator Award in the Basic Pharmacological Sciences.

Manuscript received March 29, 1999; Accepted for publication July 2, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BARSTEAD, R. J., L. KLEIMAN, and R. H. WATERSTON, 1991  Cloning, sequencing, and mapping of an -actinin gene from the nematode Caenorhabditis elegans.. Cell Motil. Cytoskeleton 20:69-78[Medline].

BEITEL, G. J., S. G. CLARK, and H. R. HORVITZ, 1990  Caenorhabditis elegans ras gene let-60 acts as a switch in the pathway of vulval induction. Nature 348:503-509[Medline].

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

Genome sequence of the nematode C. elegans: a platform for investigating biology. (1998) Science 282:2012-2018.

CONKLIN, D. S., J. A. MCMASTER, M. R. CULBERTSON, and C. KUNG, 1992  COT1, a gene involved in cobalt accumulation in Saccharomyces cerevisiae.. Mol. Cell. Biol. 12:3678-3688[Abstract/Free Full Text].

COULONDRE, C. and J. H. MILLER, 1977  Genetic studies of the lac repressor. IV. Mutagenic specificity in the LacI gene of Escherichia coli. J. Mol. Biol. 117:577-606[Medline].

COULSON, A., R. WATERSTON, J. KIFF, J. SULSTON, and Y. KOHARA, 1988  Genome linking with yeast artificial chromosomes. Nature 355:184-186.

HODGKIN, J., 1993  Molecular cloning and duplication of the nematode sex-determining gene tra-1.. Genetics 133:543-560[Abstract].

HORVITZ, H. R. and P. W. STERNBERG, 1991  Multiple intercellular signalling systems control the development of the Caenorhabditis elegans vulva. Nature 351:535-541[Medline].

HORVITZ, H. R., S. BRENNER, J. HODGKIN, and R. K. HERMAN, 1979  A uniform genetic nomenclature for the nematode Caenorhabditis elegans.. Mol. Gen. Genet. 175:129-133[Medline].

JACOBS, D., G. J. BEITEL, S. G. CLARK, H. R. HORVITZ, and K. KORNFELD, 1998  Gain-of-function mutations in the Caenorhabditis elegans lin-1 ETS gene identify a C-terminal regulatory domain phosphorylated by ERK MAP kinase. Genetics 149:1809-1822[Abstract/Free Full Text].

KAMIZONO, A., M. NISHIZAWA, Y. TERANISHI, K. MURATA, and A. KIMURA, 1989  Identification of a gene conferring resistance to zinc and cadmium ions in the yeast Saccharomyces cerevisiae.. Mol. Gen. Genet. 219:161-167[Medline].

KIMURA, K. D., H. A. TISSENBAUM, Y. LIU, and G. RUVKUN, 1997  daf-2, an insulin receptor-like gene that regulates longevity and diapause in Caenorhabditis elegans.. Science 277:942-926[Abstract/Free Full Text].

KORNFELD, K., 1997  Vulval development in Caenorhabditis elegans.. Trends Genet. 13:55-61[Medline].

KORNFELD, K., K.-L. GUAN, and H. R. HORVITZ, 1995a  The Caenorhabditis elegans gene mek-2 is required for vulval induction and encodes a protein similar to the protein kinase MEK. Genes Dev. 9:756-768[Abstract/Free Full Text].

KORNFELD, K., D. B. HOM, and H. R. HORVITZ, 1995b  The ksr-1 gene encodes a novel protein kinase involved in Ras-mediated signaling in C. elegans. Cell 83:903-913[Medline].

KORSWAGEN, H. C., R. M. DURBIN, M. T. SMITS, and R. H. A. PLASTERK, 1996  Transposon Tc1-derived, sequence-tagged sites in Caenorhabditis elegans as markers for gene mapping. Proc. Natl. Acad. Sci. USA 93:14680-14685[Abstract/Free Full Text].

KWOK, P.-Y., Q. DENG, H. ZAKERI, S. L. TAYLOR, and D. A. NICKERSON, 1996  Increasing the information content of STS-based genome maps: identifying polymorphisms in mapped STSs. Genomics 31:123-126[Medline].

LACKNER, M. L., K. KORNFELD, L. M. MILLER, H. R. HORVITZ, and S. K. KIM, 1994  A MAP kinase homolog, mpk-1, is involved in ras-mediated induction of vulval cell fates in Caenorhabditis elegans.. Genes Dev. 8:160-173[Abstract/Free Full Text].

LAI, E., J. RILEY, I. PURVIS, and A. ROSES, 1998  A 4-Mb high-density single nucleotide polymorphism-based map around human APOE. Genomics 54:31-38[Medline].

LANDEGREN, U., M. NILSSON, and P.-Y. KWOK, 1998  Reading bits of genetic information: methods for single-nucleotide polymorphism analysis. Genome Res. 8:769-776[Free Full Text].

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

OGG, S., S. PARADIS, S. GOTTLIEB, G. I. PATTERSON, and L. LEE et al., 1997  The Forkhead transcription factor DAF-16 transduces insulin-like metabolic and longevity signals in C. elegans.. Nature 389:994-999[Medline].

PALMITER, R. D. and S. D. FINDLEY, 1995  Cloning and functional characterization of a mammalian zinc transporter that confers resistance to zinc. EMBO J. 14:639-649[Medline].

PALMITER, R. D., T. B. COLE, and S. D. FINDLEY, 1996  ZnT-2, a mammalian protein that confers resistance to zinc by facilitating vesicular sequestration. EMBO J. 15:1784-1791[Medline].

PAULSEN, I. T. and M. H. SAIER, JR., 1997  A novel family of ubiquitous heavy metal ion transport proteins. J. Membr. Biol. 156:99-105[Medline].

RIDDLE, D. L., T. BLUMENTAL, B. J. MEYER and J. R. PRIESS (Editors), 1997 C. elegans II. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

RUVKUN, G., V. AMBROS, A. COULSON, R. WATERSTON, and J. SULSTON et al., 1989  Molecular genetics of the Caenorhabditis elegans heterochronic gene lin-14.. Genetics 121:501-516[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual, Ed. 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SAVAGE, C., M. HAMELIN, J. G. CULOTTI, A. COULSON, and D. G. ALBERTSON et al., 1989  mec-7 is a ß-tubulin gene required for the production of 15-protofilament microtubules in Caenorhabditis elegans.. Genes Dev. 3:870-881[Abstract/Free Full Text].

WILLIAMS, B. D., B. SCHRANK, C. HUYNH, R. SHOWNKEEN, and R. H. WATERSTON, 1992  A genetic mapping system in Caenorhabditis elegans based on polymorphic sequence-tagged sites. Genetics 131:609-624[Abstract].

WONG, D. G., J.-B. FAN, C.-J. SIAO, A. BERNO, and P. YOUNG et al., 1998  Large-scale identification, mapping, and genotyping of single-nucleotide polymorphisms in the human genome. Science 280:1077-1082[Abstract/Free Full Text].

This article has been cited by other articles:

				L. D. MacDonald, A. Knox, and D. Hansen Proteasomal Regulation of the Proliferation vs. Meiotic Entry Decision in the Caenorhabditis elegans Germ Line Genetics, October 1, 2008; 180(2): 905 - 920. [Abstract] [Full Text] [PDF]

				J. J. Bruinsma, D. L. Schneider, D. E. Davis, and K. Kornfeld Identification of Mutations in Caenorhabditis elegans That Cause Resistance to High Levels of Dietary Zinc and Analysis Using a Genomewide Map of Single Nucleotide Polymorphisms Scored by Pyrosequencing Genetics, June 1, 2008; 179(2): 811 - 828. [Abstract] [Full Text] [PDF]

				M. V. Dinkelmann, H. Zhang, A. R. Skop, and J. G. White SPD-3 Is Required for Spindle Alignment in Caenorhabditis elegans Embryos and Localizes to Mitochondria Genetics, November 1, 2007; 177(3): 1609 - 1620. [Abstract] [Full Text] [PDF]

				K. Yamamoto, J. Narukawa, K. Kadono-Okuda, J. Nohata, M. Sasanuma, Y. Suetsugu, Y. Banno, H. Fujii, M. R. Goldsmith, and K. Mita Construction of a Single Nucleotide Polymorphism Linkage Map for the Silkworm, Bombyx mori, Based on Bacterial Artificial Chromosome End Sequences Genetics, May 1, 2006; 173(1): 151 - 161. [Abstract] [Full Text] [PDF]

				J. L. Goldstein, D. Glossip, S. Nayak, and K. Kornfeld The CRAL/TRIO and GOLD Domain Protein CGR-1 Promotes Induction of Vulval Cell Fates in Caenorhabditis elegans and Interacts Genetically With the Ras Signaling Pathway Genetics, February 1, 2006; 172(2): 929 - 942. [Abstract] [Full Text] [PDF]

				D. J. Eastburn and M. Han A Gain-of-Function Allele of cbp-1, the Caenorhabditis elegans Ortholog of the Mammalian CBP/p300 Gene, Causes an Increase in Histone Acetyltransferase Activity and Antagonism of Activated Ras Mol. Cell. Biol., November 1, 2005; 25(21): 9427 - 9434. [Abstract] [Full Text] [PDF]

				J. Yochem, L. R. Bell, and R. K. Herman The Identities of sym-2, sym-3 and sym-4, Three Genes That Are Synthetically Lethal With mec-8 in Caenorhabditis elegans Genetics, November 1, 2004; 168(3): 1293 - 1306. [Abstract] [Full Text] [PDF]

				P. Syntichaki and N. Tavernarakis Genetic Models of Mechanotransduction: The Nematode Caenorhabditis elegans Physiol Rev, October 1, 2004; 84(4): 1097 - 1153. [Abstract] [Full Text] [PDF]

				A. K. Spartz, R. K. Herman, and J. E. Shaw SMU-2 and SMU-1, Caenorhabditis elegans Homologs of Mammalian Spliceosome-Associated Proteins RED and fSAP57, Work Together To Affect Splice Site Choice Mol. Cell. Biol., August 1, 2004; 24(15): 6811 - 6823. [Abstract] [Full Text] [PDF]

				A. Forche, P. T. Magee, B. B. Magee, and G. May Genome-Wide Single-Nucleotide Polymorphism Map for Candida albicans Eukaryot. Cell, June 1, 2004; 3(3): 705 - 714. [Abstract] [Full Text] [PDF]

				D. L. Church and E. J. Lambie The Promotion of Gonadal Cell Divisions by the Caenorhabditis elegans TRPM Cation Channel GON-2 Is Antagonized by GEM-4 Copine Genetics, October 1, 2003; 165(2): 563 - 574. [Abstract] [Full Text] [PDF]

				B. Lakowski, S. Eimer, C. Gobel, A. Bottcher, B. Wagler, and R. Baumeister Two suppressors of sel-12 encode C2H2 zinc-finger proteins that regulate presenilin transcription in Caenorhabditis elegans Development, May 15, 2003; 130(10): 2117 - 2128. [Abstract] [Full Text] [PDF]

				K. A. Swan, D. E. Curtis, K. B. McKusick, A. V. Voinov, F. A. Mapa, and M. R. Cancilla High-Throughput Gene Mapping in Caenorhabditis elegans Genome Res., July 1, 2002; 12(7): 1100 - 1105. [Abstract] [Full Text] [PDF]

				C. Wen, D. Levitan, X. Li, and I. Greenwald spr-2, a suppressor of the egg-laying defect caused by loss of sel-12 presenilin in Caenorhabditiselegans, is a member of the SET protein subfamily PNAS, December 8, 2000; (2000) 11446498. [Abstract] [Full Text]

				K. L. Hill, B. D. Harfe, C. A. Dobbins, and S. W. L'Hernault dpy-18 Encodes an {alpha}-Subunit of Prolyl-4-Hydroxylase in Caenorhabditis elegans Genetics, July 1, 2000; 155(3): 1139 - 1148. [Abstract] [Full Text]

				C. Wen, D. Levitan, X. Li, and I. Greenwald spr-2, a suppressor of the egg-laying defect caused by loss of sel-12 presenilin in Caenorhabditiselegans, is a member of the SET protein subfamily PNAS, December 19, 2000; 97(26): 14524 - 14529. [Abstract] [Full Text] [PDF]

				R. Koch, H. G.A.M. van Luenen, M. van der Horst, K. L. Thijssen, and R. H.A. Plasterk Single Nucleotide Polymorphisms in Wild Isolates of Caenorhabditis elegans Genome Res., November 1, 2000; 10(11): 1690 - 1696. [Abstract] [Full Text]