|Home | About | Journals | Submit | Contact Us | Français|
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In C. elegans, single nucleotide polymorphisms (SNPs) can function as silent genetic markers, with applications ranging from classical two- and three-factor mapping to measuring recombination across whole chromosomes.
Here, we describe a set of 48 primer pairs that flank SNPs evenly spaced across the C. elegans genome and that work under identical PCR conditions. Each SNP in this set alters a DraI site, enabling rapid and parallel scoring. We describe a procedure using these reagents to quickly and reliably map mutations. We show that these techniques correctly map a known gene, dpy-5. We then use these techniques to map mutations in an uncharacterized strain, and show that its behavioral phenotype can be simultaneously mapped to three loci.
Together, the reagents and methods described represent a significant advance in the accurate, rapid and inexpensive mapping of genes in C. elegans.
Single Nucleotide Polymorphism (SNP) mapping has transformed studies of genetic linkage in C. elegans since its introduction in 2001. In SNP mapping, DNA sequence polymorphisms between the wild-type C. elegans strain (N2 Bristol) and a closely related strain (CB4856 Hawaiian) are used as genetic markers. Compared to other markers that have been used for genetic mapping, SNPs have two distinct advantages. First, unlike conventional marker mutations that cause visible phenotypes, SNPs in general have no associated phenotype. Thus, mutant phenotypes that are masked by conventional marker mutations, such as those with subtle behavioral defects, can be mapped using SNPs. Second, SNPs are far denser than other markers, including both visible markers and DNA polymorphisms such as Tc1 insertions. Because SNPs are approximately as dense as genes, SNP mapping can in theory provide single-gene resolution . Together, these two advantages have made SNP mapping the technique of choice for many C. elegans researchers.
SNP mapping is usually done in two phases. The first phase, chromosome mapping, is similar to traditional two-factor mapping and seeks to identify the relevant chromosome and rough position of the gene of interest. The second phase, interval mapping, seeks to place the gene of interest in an interval between two SNPs, and can be used iteratively to fine map the gene. SNP detection in both phases is typically performed by using only SNPs that alter a restriction site, which are also known as snip-SNPs . Although other SNP detection methods have been used for mapping in C. elegans, such as fluorescence polarimetry and indel detection [3,4], snip-SNPs are attractive because they require low initial investment and do not require specialized equipment. However, snip-SNP detection requires PCR amplification of the SNP region, digestion with the appropriate restriction enzyme, and gel electrophoresis. These multiple steps can be daunting, particularly during chromosome mapping when many SNPs need to be assayed simultaneously.
The method described here streamlines the procedure for detecting snip-SNPs, making faster, more efficient mapping possible. First, we identified 48 SNPs that met our criteria (8 per chromosome). This means that every part of the genome is linked to multiple SNPs, so that adjacent SNPs serve as internal controls, and also that sub-chromosome position can be determined. Second, we simplified the PCR step by identifying primer sets that all work at identical amplification conditions, and that are tolerant of being added to reactions by pin-replication. This enabled us to quickly amplify across 48 snip-SNPs in a single 96-well PCR plate. Third, we streamlined restriction digestion by using only snip-SNPs that can be distinguished by a single restriction enzyme, DraI. This enzyme is relatively inexpensive and tolerant of PCR buffer salts. This allowed us to perform digestion in the original PCR plate, by adding to each well an identical digestion cocktail. Finally, primer locations were designed so that the informative digestion products can be resolved on an agarose gel. The accessibility of this mapping procedure, together with its speed, low cost, robustness, and accuracy, should make it a preferred option for most C. elegans labs.
To facilitate mapping a mutation onto a chromosome, we designed a set of PCR reagents based on modifications of the principle of bulk segregant analysis described by Wicks et al. . Our primary goal was to simplify the procedure by performing all steps in a 96-well format PCR plate, and designing each SNP reaction to be performed under identical conditions. To design a set of primers that can be used for SNP mapping in a 96-well format, two conditions must be met: the primers must all use the same conditions for polymorphism detection, and the primers must all use the same conditions for amplification. First, we simplified SNP detection by using only SNPs that could be detected using a single restriction enzyme. Since SNPs are concentrated in non-coding A/T-rich regions of the genome, we reasoned that good coverage would be obtained from the enzyme DraI, which recognizes the sequence TTT^AAA. We identified all DraI SNPs in a custom database (available as supplementary material) that incorporated all SNPs identified by the Genome Sequencing Center, Washington U, St. Louis, MO (6,333 total SNPs, 248 DraI SNPs in our database)  and by Exelixis, South San Francisco, CA (9,295 total SNPs, 257 DraI SNPs in our database) . From among these we selected eight candidate DraI SNPs on each chromosome that were far enough from nearby DraI sites (typically >200 bp on one side, >50 bp on the other) to enable detection of cleavage at the SNP DraI site. The genetic positions of the resulting 48 DraI SNPs are shown in Figure Figure11.
Next, to enable simultaneous amplification of all selected SNPs in a 96-well format, we chose primer pairs with similar annealing temperatures and product length. The program Primer3  was used to design primers to amplify short sequences (typically 300–500 bp) containing each selected DraI SNP. Optimum Tm was set to 60°C. Primer pairs were tested, and unsatisfactory pairs were redesigned until all 48 primer pairs amplified robustly in simultaneous PCR reactions in a single plate. The resulting primer sequences are shown in Table Table11.
Finally, we devised a set of procedures that maximize speed and minimize the potential for error during reaction set up and gel loading (Figure (Figure2).2). Our chromosome mapping procedure begins with the same genetic manipulations as other SNP mapping protocols. Hawaiian males are crossed into the mutant strain to produce heterozygous F1 animals. Homozygous F2 animals from the heterozygous F1 animals are identified based on their mutant phenotype. At the same time, animals with a non-mutant phenotype, which are enriched for Hawaiian sequences at the locus of interest, are also isolated (Figure (Figure2).2). Thirty to fifty animals of each class are combined into two tubes and lysed using detergent and proteinase K (Figure (Figure2).2). A PCR master mix, not including PCR primers, is then assembled and added to each lysate. These PCR mixes are then dispensed into alternating rows of a 96-well PCR plate using a multi-channel pipettor (Figure (Figure2).2). These three simple steps generate a set of 96 PCR master mixes ready for the final addition of primers.
Because there are 96 separate reactions, each requiring addition of a specific primer pair, we generated a pre-arrayed set of primers, which are then added to the PCR master mixes by pin replication. Primer pairs described above are arrayed in pairs into a microtiter plate at 10 μM each primer ('primers' Figure Figure2A),2A), with each row containing the primer pairs for the eight SNPs along a single chromosome. Adjacent rows contain a duplicate set of primers for a particular chromosome, and the plate of primers is pin-replicated into the master PCR mix. After amplification, PCR products are digested in the plate with DraI in a final volume of 15 μL and loaded onto a 2.5% agarose gel using an 8-channel pipette. Because we use a gel comb with wells spaced half the distance between pipette tips of the multi-channel pipette, we can automatically load the mutant samples from the upper row and wild-type samples from the lower row for each SNP pair in adjacent wells. The resulting gel displays all 48 SNP markers, from left to right and from chromosome I to X (Figure (Figure2B).2B). Each mutant SNP is next to its non-mutant control, so that the whole genome can be quickly scanned for linkage.
To validate the final set of primers for chromosome mapping experiments, we used them to map dpy-5, a mutation with a well known genetic position. We crossed CB4856 Hawaiian males to a triply-marked mapping strain, EG1000 dpy-5(e61) I; rol-6(e187) II; lon-1(e1820) III. We allowed the heterozygous F1s to self, and from the F2 generation we picked 50 Dpy and 50 non-Dpy animals into separate lysis reactions. We performed chromosome mapping PCR on these lysates using primer sets from LGI and LGII (Figure (Figure3A).3A). As expected, we found linkage to the center of LGI and no linkage to LGII.
After determining the rough position of a mutation on a chromosome using chromosome mapping, mutations can be quickly mapped to a genetic interval using the same efficiencies of the 96-well format employed in chromosome mapping. Interval mapping differs from chromosome mapping in that the genotype of individual mutant animals, rather than the genotype of pooled animals, must be determined. Also, it is necessary to assay these mutant DNAs for many SNPs within the interval for which linkage has been established. Therefore, it is most convenient to pin replicate the DNA templates, rather than the primers (Figure (Figure44).
Briefly, we crossed Hawaiian males into our mutant strain (isolated from Bristol N2) to generate a heterozygous strain. From the progeny of these heterozygotes, we singled 96 mutant animals onto worm growth plates and allowed them to lay self-fertilized embryos. After the F3 progeny had grown to the adult stage, we washed about a quarter of the F3 progeny from the plate into individual wells of the 96-well plate (see Methods). Deriving templates from the self progeny of a homozygous mutant has the advantage of allowing each mutant recombinant to be scored as a population, rather than a single animal. Also, having additional template is convenient for additional rounds of PCR if higher resolution is desired. We found that 96-well plates containing the lysed worms can be frozen at -80°C and reused successfully after many rounds of thawing and re-freezing. If more rapid mapping was required, we found it possible to remove the single F2 animal (after it had laid sufficient embryos to ensure propagation of the strain) and to lyse it for analysis of its SNPs. Specifically, a drawn out and sealed Pasteur pipette was used as a pick to place each F2 adult individually into 5 μL of lysis buffer in a 96-well plate. The embryos on the worm plate could later be used as a source of DNA for additional mapping experiments.
Lysed DNA from each well was pin replicated into 96-well PCR plates containing complete PCR cocktail minus template. Each plate included the primer set for a single SNP. Typically we use plates representing four adjacent SNPs from the section of a particular chromosome that had shown linkage in the chromosome mapping experiments. For each plate, PCR, digestion and gel electrophoresis were performed as for chromosome mapping.
To validate the technique for mapping a mutation to an interval, we once again used the previously mapped gene dpy-5. We crossed Hawaiian males into the strain EG1000 as described above, and allowed the heterozygous progeny to self-fertilize. From the F2 progeny, we singled 48 Dpy animals onto individual plates. When these plates had starved, we washed each population of progeny into a well containing lysis buffer in a 96-well plate. The lysed DNA was assayed using three SNP primer pairs from the center of LGI (Figure (Figure3B).3B). Since each well contained progeny of a single F2 animal, it was possible to determine whether that animal was homozygous Bristol, homozygous Hawaiian, or heterozygous Bristol/Hawaiian at each SNP. From these data we could identify dpy-5-containing Bristol chromosomes that have recombined with Hawaiian DNA to the left of dpy-5 (recombinant types 'a' and 'b', Figure Figure3B3B and and3C),3C), and to the right of dpy-5 (recombinant type 'c', Figure Figure3B3B and and3C).3C). Keep in mind that each worm contains two dpy-5 containing chromosomes, but at regions near the mutation, usually only one is recombinant as illustrated in Figure Figure3C.3C. We found, as expected, that our map data placed dpy-5 between -1 and 5 on LGI at approximately 0.3, very close to the known map position of 0.0 for dpy-5 (Figure (Figure3C).3C). Interestingly, one of the 48 Dpy F2 animals showed no linkage to LGI (see * in Figure Figure3B).3B). The plate of worms that had been used to generate that lysate was chunked onto a new plate to verify the Dpy phenotype. Surprisingly, they were Dpy, but less so than dpy-5 homozygotes. We have observed that this Dpy phenotype segregates at a low frequency from several unrelated crosses between Hawaiian CB4856 and Bristol N2. Indeed, the plate segregated Rol-6 animals that were also Dpy. This confirms that the phenotype is not due to a dpy-5 mutation, since dpy-5 is epistatic to rol-6, but apparently the synthetic Dpy phenotype is not.
To illustrate the utility of these methods we mapped a suppressor mutation of a behavioral phenotype. The map data demonstrate that these methods were able to map the original uncoordinated mutation and two other loci that synthetically suppress it in a single experiment. This strain, KY5029, was isolated in a screen for suppressors of unc-31(e928) (gift of Liakot Khan and Kouichi Iwasaki). unc-31 encodes the C. elegans homolog of CAPS, a protein required for dense-core vesicle release . unc-31(e928) mutants are very inactive – almost paralyzed – on food. The suppressor strain KY5029 moves well, in fact it is slightly hyperactive. The strain KY5029 was crossed to Hawaiian males. Heterozygous F1 progeny were singled to plates. 85 Unc-31 animals were singled from among the F2 progeny. Most of these plates of Unc-31 animals segregated active suppressed animals, demonstrating that the suppressors are recessive. Two relatively active animals from each of the 85 plates were singled. From among the 170 plates, 30 plates were found to have the hyperactive phenotype of KY5029. Animals from these 30 plates were combined and used for chromosome mapping. From the chromosome mapping experiment we found that the suppressed unc-31 phenotype of KY5029 animals was linked to three genetic regions, on chromosomes I, II, and IV (data not shown). The region on chromosome IV contains unc-31, and linkage to IV is expected since Unc-31 animals were selected from among the F2 generation. Thus, the regions on chromosomes I and II must contain mutations suppressing the Unc-31 phenotype.
We found that the individual suppressor loci are weak unc-31 suppressors on their own. From the unc-31(e928) plates we singled animals that were less strongly suppressed; specifically, they were slightly sluggish rather than hyperactive. Linkage to the Bristol N2 genotype in these animals was observed on chromosomes I and IV, while linkage to the Hawaiian genotype was observed on chromosome II (data not shown). Thus, chromosome I contains a novel suppressor of unc-31 that partially suppresses the unc-31(e928) phenotype. Chromosome II contains a mutation that, in combination with the mutation on chromosome I, suppresses the unc-31(e928) phenotype to produce hyperactive worms. Further analysis determined that the suppressor on II could also partially suppress unc-31(e928) on its own. To simplify future mapping experiments we mapped the independent suppressing activities of the mutations on chromosome I and II in the unc-31(e928) background. Both suppressors were then fine mapped using primer sets on I and II independently to the genetic intervals depicted in Figure Figure5A5A and and5B.5B. Together, these data suggest that KY5029 contains two mutant loci in addition to unc-31(e928). These complex interactions were deciphered with a minimum of time, effort, and confusion.
In summary, these methods comprise an accurate and fast technique for mapping that has advantages over both traditional mapping experiments and over other SNP mapping approaches that have been previously described. Compared to traditional mapping, this technique offers standardized, efficient mapping to small intervals for large numbers of mutations, such as might result from a genetic screen. In addition, it allows the mapping of subtle phenotypes (such as behavior) and complex genotypes (such as suppressor or synthetic mutations).
Compared to other SNP mapping methods, the technique described here occupies a comfortable midpoint between the simple but less efficient method described by Wicks et al. , and the high-throughput but complex techniques published by Swan et al.  and Zipperlen et al. . Although the SNP detection technique described by Wicks is simple and inexpensive, it requires setting up a number of individual PCR reactions and matching the correct PCR product with the correct restriction enzyme. In the methods described here, once the primers (in chromosome mapping) or templates (in interval mapping) are arrayed into a plate, reaction components are accurately dispensed automatically and repeatably. Errors in matching restriction enzymes and buffers to primer sets have been eliminated by using DraI for all reactions. Further, we have found that assaying 8 SNPs on each chromosome means that every mutation is linked to multiple SNPs, giving a high level of redundancy. In fact, because of its cost and accuracy advantages, our technique has been successfully applied in an undergraduate teaching lab setting (M. Peters, personal communication).
The fluorescent polarimetry technique described by Swan et al., and the indel detection technique described by Zipperlen et al., enable high throughput SNP detection in C. elegans. However, these techniques require specialized equipment (a fluorescence polarimeter or capillary sequencer) that are not accessible to every laboratory and that require significant operator knowledge. Compared to those SNP mapping approaches, the technique described here is cheaper and more accessible, since it relies on methods that most labs already use.
In this paper, we present a happy medium between previous approaches to SNP mapping in worms. We build upon the simple, inexpensive, accessible and robust restriction digestion SNP detection technique of Wicks. However our primer sets, equipment and techniques substantially reduce user effort relative to the Wicks method, and so provide the efficiency and low error rate of the Swan or Zipperlen technologies.
Hawaiian CB4856 males were crossed into EG1000 dpy-5(e61) I; rol-6(e187) II; lon-1(e1820) III. Fifty Dpy animals and fifty non-Dpy animals from among the self-progeny of EG1000/CB4856 heterozygote hermaphrodites were picked into separate tubes, each containing 20 μL single-worm lysis buffer (50 mM KCl, 10 mM Tris pH 8.3, 2.5 mM MgCl2, 0.45% IGEPAL CA-630, 0.45% Tween 20, 0.01% (w/v) gelatin, 60 ug/ml proteinase K). A further 96 Dpy animals were picked to individual plates for use in interval mapping (see below). They were lysed by freezing at -80°C followed by incubation at 65°C 1 hour and proteinase was inactivated by incubation at 95°C 15 minutes. The Dpy lysate DNA templates were then added to a PCR master mix containing 424 μL water, 52 μL 10X PCR buffer (10X: 22.5 mM MgCl2, 500 mM Tris-HCl, 140 mM (NH4)2SO4, pH 9.2 at 25°C), 10.4 μL 10 mM dNTPs, and 3.12 μL Taq (5 units/μl). A similar mix was made with the 50 non-Mutant animals. 9.8 μL of the mutant mix or the non-mutant mix was aliquoted into alternate rows of a 96-well PCR plate (Figure (Figure1A).1A). Primer pairs were arrayed into a microtiter plate at 10 μM each primer, so that neighboring rows contain duplicate pairs, and pin-replicated into the master mix. PCR reactions were done using the cycling conditions: 2' 94°C, 35 cycles of (15" 94°C, 45" 60°C, 1' 72°C), 5' 72°C. After amplification, PCR products were digested in the plate with the restriction enzyme DraI in a final volume of 16 μL (10 μL PCR product, 4.15 μL H2O 1.6 μL 10X DraI buffer (New England Biolabs), 0.25 μL DraI (10 units/μL, New England Biolabs)). This was accomplished by adding 6 μL of the enzyme plus enzyme buffer mix to each well using a multi-channel pipette followed by brief centrifugation in a Sorval RT6000D centrifuge with an H1000B rotor. Digestion reactions were incubated at 37°C at least 4 hours. Samples were then loaded onto a 2.5% agarose gel using an 8-channel pipette. The resulting gel displays all 48 SNP markers, from left to right and from chromosome I to X. Each Mutant SNP is next to its non-Mutant control, so that the whole genome can be quickly scanned for linkage.
PCR templates were generated by cloning mutant animals from among the self-progeny of EG1000/CB4856 F1 hermaphrodites (described above) onto individual seeded plates. After 5 days, self progeny were washed from each plate using water (>100 worms / plate) and placed in a single well of a 96-well plate. Worms were allowed to settle to the bottom of the wells for 15' at 4°C then excess water was pipetted off to leave 45 μl in each well. The plates were frozen and stored at -80°C. The plates were thawed and 15 μl of 4X lysis buffer (200 mM KCl, 40 mM Tris pH 8.3, 10 mM MgCl2, 1.8% IGEPAL CA-630, 1.8% Tween 20, 0.04% (w/v) gelatin, 240 ug/ml proteinase K) was then added to each well to give 1X lysis buffer. The plates were covered with sealing tape and briefly vortexed to break up the worm pellet. The worms were lysed by incubation at 65°C 1 hour and 95°C 15 minutes. These PCR templates were stored frozen at -80°C and thawed prior to each use. For each PCR, each well of the 96-well plate received 9.8 μL of a PCR mix containing 8.5 μL water, 1 μL 10X buffer, 0.2 μL 10 mM dNTP, 0.02 μL each primer (100 μM), and 0.06 μL Taq (5 units/μl). Templates were then pin-replicated from the lysis plate. PCR conditions and DraI digests were the same as in chromosome mapping.
We cloned 85 animals with an unc-31(e928)-like phenotype from the self progeny of KY5029/CB4586 hermaphrodites. unc-31(e928) animals are lethargic and uncoordinated; however, most animals exhibit periodic moments of coordinated movement making it difficult to distinguish between plates with no suppressed progeny and plates with weakly suppressed progeny. Therefore, two animals exhibiting coordinated movement were cloned from each plate. We scored the progeny of these 170 animals and divided them into five classes: uncoordinated, sluggish yet coordinated, coordinated, hyperactive, and mixed. For the sluggish and hyperactive phenotypes, we collected animals for chromosome mapping by combining two animals from each plate. Chromosome mapping was performed as described above, suggesting the presence of two suppressors located on chromosomes I and II (data not shown). To confirm the mapping results, individual recombinants were assayed for SNPs on chromosomes I, II, and IV (data not shown). To simplify further interval mapping experiments these suppressors were crossed away form each other to generate two partially suppressed strains, EG5296 ox300; unc-31(e928) and EG5297 dpy-5(e61); ox305; unc-31(e928). For EG5297, we verified the loss of the ox300 chromosome by homozygosing a dpy-5(e61) marker. Interval mapping of ox300 was carried out by cloning 11 unc-31(e928)-like self progeny from EG5296/CB4586 hermaphrodites. From these self-progeny 35 sluggish yet coordinated animals were cloned and assayed at SNPs flanking the suppressor (Figure (Figure5A).5A). ox300 is located on Chromosome I between W03D8 and B0205. ox305 was mapped by cloning 8 unc-31(e928)-like self progeny from EG5297/CB4586 hermaphrodites. From these self-progeny 21 partially suppressed animals were cloned and assayed at SNPs flanking the suppressor (Figure (Figure5B).5B). The suppressor in ox305 is located on Chromosome II between R52 and T10D4. Four new SNPs that are not part of the chromosome mapping set were used. Wormbase allele snp_F12B6 was amplified with primers 5'-caggttggtttttggcaagt-3' 5'-tgattgaacatatccggcaa -3' and detected with MfeI. Wormbase allele snp_F28H1 was amplified with primers 5'-gcagtaggcaagagtcaggc-3' and 5'-tattgcacttggctcacagc -3' and was detected with HpyCH4V. pkP2135 was amplified with primers 5'-tttgcagatttccgatactgtg -3' and 5'-ttttgtcgtaagacctttggtg -3' and was detected with DraI. A polymorphism on T10D4, referenced at the web page  was amplified with primers 5'-gtacgcctcaaaaagtggag -3' and 5'-accacccaacacaatctctg -3' and was detected with MseI.
SNP: Single nucleotide polymorphism.
MWD MH conceived and designed the SNP methods, carried out and supervised the experiments and drafted the manuscript. MWD wrote the scripts that extracted and formatted the SNP data. TH and SO tested SNP primers PH carried out the unc-31 mapping experiments. EMJ helped coordinate the experiments and draft the manuscript. All authors read and approved the final manuscript.
We would like to thank the Caenorhabditis Genetics Center (CGC) for providing strains, S. Wicks for early advocation of SNP mapping in C. elegans, the St. Louis and Exelixis SNP discovery projects for shotgun sequencing of CB4856, L. Khan and K. Iwasaki for KY5029, other laboratories that provided feedback on early versions of our primer sets, and T. Harris for providing initial analysis of SNP sequence data.