A new method for SNP discovery

Introduction

The identification of genes affecting complex traits is a very difficult and challenging task. For many complex traits, the observable variation is quantitative, and loci affecting such traits are generally termed quantitative trait loci (QTLs). In contrast with monogenic traits, it is impossible to identify all the genomic regions responsible for complex trait variation without a genetic map constructed with molecular markers. Single nucleotide polymorphisms (SNPs) can be used as genetic markers for constructing high-density genetic maps and to carry out association studies related to diseases (1). As a result of their abundance, heredity stability, and the availability of high-throughput analysis technologies, SNP markers have begun to replace traditional markers such as restriction fragment length polymorphisms (RFLPs), amplified fragment length polymorphisms (AFLPs), and simple sequence repeat markers (SSRs or microsatellites) for fine mapping and association studies. SNPs have become an important application in the development and research of genetic markers.

Studies utilizing SNPs have become feasible with the availability of a variety of methods of SNP detection and genotyping. As the demand for genetic analysis increases, SNP detection technologies are being developed at an accelerated pace. SNP detection encompasses two broad areas: (i) scanning DNA sequences for previously unknown polymorphisms, and (ii) screening (genotyping) individuals for known polymorphisms. Scanning for new SNPs can be further divided into the global (or random) approach, and the regional (or targeted) approach (2).

There are several strategies that can be applied to new SNP discovery. The most straightforward method is direct sequence comparison using public or other sequence databases (3,4) and locus-specific amplification of target genome regions followed by sequence comparison (5,6). For SNP discovery in candidate genomic regions, a prescreening of SNPs prior to sequence determination is needed. There are a number of methods to prescreen SNPs (7–9), such as single-strand conformational polymorphism (10), denaturation kinetics (11), chemical cleavage (12–14), enzyme cleavage (15–18), array hybridization (19), mismatch repair detection (20), and bacteriophage Mu DNA transposition (21,22).

The main drawback of these methods is the requirement for prior sequence information (i.e., at the very least, the region for primer design should be clear). Such sequences are usually the product of whole-genome shotgun sequencing applications and have been mostly limited to model organisms or humans. In non-model organisms, there have been certain research efforts focusing on SNP discovery in candidate genes or genomic regions (23). However, these efforts are limited to highly homologous regions in which the sequence can be obtained by homologous cloning.

There are certain methods which have been developed for exploiting SNPs randomly in the genome, such as representation shotgun sequencing (24), primer-ligation–mediated PCR (25) and degenerate oligonucleotide–primed PCR (26,27). Despite the advantage of reducing the number of clones required in the analysis, these methods still require the sequencing of tens of thousands of clones to obtain data suitable for SNP discovery, and the efficiency of SNP discovery has been much lower than expected. Therefore, there is still room for considerable improvement in the specificity, sensitivity, and cost-effectiveness of SNP detection methods. Here we describe a novel and effective method to develop SNPs randomly throughout the entire genome in essentially any organism. We were able to specifically isolate the sequences containing SNPs using the half-smooth tongue sole (Cynoglossus semilaevis) as the experiment model. The half-smooth tongue sole is a newly exploited and commercially important cultured marine flatfish in China. Moreover, the females grow 2–4 times faster than males, which is useful for studying both sex determination and development (28,29). A high-density genetic map is necessary for this purpose, and SNPs represent an excellent choice for fulfilling this requirement.

Materials and methods

Purification and preparation of CEL I nuclease

CEL I nuclease was purified as described (18). Briefly, celery stalks (500 g) were juiced at 4°C, adjusted to 0.1 M Tris-HCl, pH 7.7, 100 µM phenylmethanesulphonylfluoride (PMSF), and spun for 20 min at 2600× g to pellet debris. The supernatant was brought to 25% saturation in (NH₄)₂SO₄, mixed for 30 min at 4°C and spun at 16,000× g at 4°C for 40 min. The resulting supernatant was adjusted to 80% (NH₄)₂SO₄, mixed for 30 min at 4°C and spun at 16,000× g for 1.5 h. The pellet was suspended in 0.1 M Tris-HCl, 0.5 M KCl pH 7.7, 100 µM PMSF (1/10 starting volume). The suspension was transferred to a dialysis tube and dialyzed against a total of 2 L of the same buffer, with four changes over 4 h. Aliquots of extract were stored at -20°C. Different dilutions of CEL I were tested and an optimal concentration of 1 U/µL was determined so that the CEL I cleaved SNPs efficiently but did not nonspecifically cleave dsDNA.

Sample and DNA extraction

Liver tissues were collected from Cynoglossus semilaevis individuals and frozen in liquid nitrogen. For DNA extraction, a piece of liver tissue of ∼20 mg was homogenized in 500 µL lysis buffer [10mM Tris-HCl pH 8.0, 100 mM EDTA pH 8.0, 100 mM NaCl, 0.5% sodium dodecyl sulfate (SDS), and 100 µg/mL freshly added proteinase K]. Then the homogenate was lysed at 55°C for 60–90 min. DNA was extracted with the phenol:chloroform method (phenol/chloroform/isoamyl alcohol 25:24:1 solution; Shanghai Shenggong, Shanghai, China). After one phenol, one phenol-chloroform, and one chloroform extraction, DNA was precipitated with two volumes of ethanol. The DNA was pelleted, washed once in 70% ethanol, dried, and dissolved in TE buffer. DNA quality and concentration were assessed by agarose electrophoresis and measured with a GENEQUANT Pro RNA/DNA spectrophotometer (Amersham Pharmacia Biotech Ltd., Cambridge, England).

MseI digestion, adaptor ligation and pre-amplification

DNA extracted from ten individuals was mixed equally as a DNA pool. 1000 ng of pool DNA was digested with 50 U MseI (New England Biolabs, Ipswitch, MA, USA) for 2 h at 65°C. The digested DNA was purified (QIAquick PCR Purification Kit; Qiagen, Valencia, CA, USA) and ligated to MseI adaptors (5′-TACTCAGGACTCAT-3′/5′-GACGATGAGTCCTGAG-3′) with T4 ligase (Takara Bio Inc., Shiga, Japan) at 16°C overnight. The ligation mixture was diluted 10 times and then used for PCR pre-amplification with adaptor-specific primers (5′-GATGAGTCCTGAGTAAN-3′). The conditions of the PCR were as follows: pre-denaturaion at 94°C for 5 min; 20 cycles of denaturation at 94°C for 30 s, annealing at 53°C for 1 min, and extension at 72°C for 1 min; final extension at 72°C for 7 min; and maintenance at 4°C.

Heteroduplex formation, CEL I digestion and Bst polymerase elongation

To form heteroduplexes of the PCR product, the denaturation and annealing were performed by using PCR conditions as follows: 1 cycle of 95°C for 10 min; 95°C to 85°C (−2°C/sec); 85°C to 25°C (-0.1°C/sec); maintenance at 4°C.

For CEL I digestion of 10 µL PCR product, 20 µL of CEL I buffer [10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES; pH 7.5), 10 mM MgSO₄, 0.002% (w/v) Triton X-100, 20 ng/mL of bovine serum albumin (BSA)] and 1 µL optimized CEL I enzyme were added on ice. The mixture was incubated at 45°C for 30 min. Reaction was stopped by the addition of 5 µL 0.15 M EDTA (pH 8.0).

The CEL I–digested DNA was purified with the QIAquick PCR Purification Kit and used for Bst elongation. The elongation mixture contained 20 mM Tris-HCl (pH 8.8), 10 mM KCl, 10 mM (NH4)₂SO₄, 2 mM MgSO₄, 1 mM MgCl₂, 0.1% Tween X-100, 20 nM dATP, 20 nM dCTP, 20 nM dGTP, 13 nM dTTP, 7 nM biotin-dUTP (Roche, Basel, Switzerland) and 4 units of Bst DNA polymerase (New England Biolabs). The mixture was incubated at 65°C for 60 min. The elongation product was purified (QIAquick PCR Purification Kit) and used for specific capture by magnetic streptavidin-coated beads.

Isolation of DNA containing SNPs

The DNA molecules containing biotin-dUTP were captured using streptavidin-coated beads (Catalog no. Z5481; Invitrogen, Carlsbad, CA, USA). First, the beads were prepared by washing 1 mg of beads three times in 500 µL TEN₁₀₀ (10 mM Tris-HCl, 1 mM EDTA, 100 mM NaCl, pH 7.5) after which they were re-suspended in 40 µL of TEN₁₀₀. The purified Bst elongation product was added to the prepared beads and incubated for 30 min at room temperature with constant gentle agitation. The nonspecific DNA was removed by 3 non-stringent washes followed by 3 stringent washes. Non-stringent washes were performed by adding 400 µL of TEN₁₀₀₀ (10 mM Tris-HCl, 1 mM EDTA, 1 M NaCl, pH 7.5). The stringent washes were performed by adding 400 µL of 0.2× sodium chloride sodium citrate (SSC)/0.1% SDS to the DNA. Finally, the beads were washed with 1× SSC. All of the washes were carried out at room temperature for 5 min, with recovery of the DNA by magnetic-field separation. The DNA was then separated from the beads-DNA complex by adding 50 µL TE (10mM Tris-HCl, 1mM EDTA, pH 8.0) and incubated at 95°C for 10 min, after which the supernatant was removed quickly and stored at −20°C.

PCR amplification and cloning

The isolated DNA was amplified in a 50-µL reaction containing 5 pmol MseI-N primer, 1× Taq DNA polymerase buffer, 1 mM dNTP, in the presence of 1 unit Taq DNA polymerase (Takara Bio, Inc.). The thermal cycling parameter comprised an initial denaturation for 5 min at 94°C; followed by 20 cycles of 94°C for 30 s, 53°C for 60 s, and 72°C for 60 s; with a final elongation step at 72°C for 7 min. The PCR product was purified (QIAquick PCR Purification Kit), ligated to pBS-T vectors (Tiangen, Beijing, China), and then propagated in Escherichia coli DH5α. Positive clones were confirmed by PCR with the vector-specific primers M13 (5′-CGCCAGGGTTTTCCCAGTCACGAC-3′) and RV-M (5′-AGCGGATAACAATTTCACACAGG-3′). The clones with DNA inserts ranging from 600 bp to 1500 bp were sequenced by an automated DNA sequencer (ABI 3730; Applied Biosystems, Foster City, CA, USA).

SNP validation

Primers were designed according to the obtained sequences. DNA fragments were amplified in four Cynoglossus semilaevis individuals with PCR, purified on 1% agarose gel, extracted via a QIAEX II Gel Extraction Kit (Qiagen), ligated to pBS-T vectors (Tiangen), propagated in E. coli DH5α cells and sequenced. For every DNA fragment, 4 positive clones per individual were sequenced. Sequences were aligned by DNAMAN software (version 5.2.2.0; Lynnon Corp., Pointe-Claire, Quebec, Canada) to validate SNP sites. The polymorphic sites, which appeared in two individuals and twice in one individual, were considered to be SNP sites.

Results

DNA extracted from five female and five male Cynoglossus semilaevis were mixed equally and used for SNP discovery. After isolation of the DNA molecules containing SNPs, they were cloned and sequenced. Finally, 55 positive clones with DNA inserts estimated from 600 bp to 1500 bp were sequenced. Among these 55 sequences, three contained microsatellites [(CA)₁₆, (TC)₇ and (CT)₉, respectively], nine contained polyT, and two contained polyC. Ultimately, 41 sequences were deemed appropriate to design primers and 23 pairs of primers were designed for validating SNPs. There were 18 sequences for which we did not design primers, because these sequence reads were longer than 600 bp. Such sequences cannot be sequenced correctly by a single sequencing reaction. To achieve the goal of cost-effectiveness, we first chose fragments less than 600 bp, from which it is possible to obtain SNPs at an efficient cost. Fifteen of 23 primer pairs produced specific amplification products, and the other eight pairs had either nonspecific amplification or no amplification (which could result from defective primers or non-optimal amplification conditions). Ten of these 15 primerpairs were cloned and sequenced to confirm the existence of SNP sites in four individuals, and nine fragments contained SNPs (Table 1 and Figure 1). The method used here to confirm SNPs might miss SNPs of very low frequency; however, it may enhance accurate validation that the DNA fragment isolated did contain SNPs, since certain artifacts could appear due to PCR or sequencing errors.

Table 1.  Primers Used for SNP Validation

At the beginning, we wanted to validate SNPs in pooled DNA, which could reduce the amount of sequencing and cost. We found a total of 35 SNPs in these 9 fragments, but only 18 SNPs were validated among them, and 4 new SNPs were discovered in another 4 individuals. One possibility is that the other 17 SNPs did exist but not in these particular 4 individuals we used. However, another possibility is that these 17 SNPs were artifacts resulting from PCR (a method in which artifacts can be enriched).

Discussion

SNPs have become the markers of choice for biological studies, but their isolation for non-model organisms with unsequenced genomes is often difficult. Here, we describe a rapid and cost-effective strategy to isolate SNPs which is applicable to any organism.

The DNA from ten Cynoglossus semilaevis individuals were mixed and amplified by PCR. After denaturing and annealing, the fragments containing SNPs formed mismatches and were recognized by CEL I nuclease. CEL I is specific for DNA distortions and mismatches in a range of pH 6.0–9.0. Incision occurs on the 3′-side of the mismatch site in one of the two DNA strands, with the production of a 3′-OH group in a hetero-duplex (15). CEL I nicks one stand of DNA in a mismatch heteroduplex at the site of the mismatch. Using limited CEL I digestion, one can avoid the second cut in the opposite strand of the same DNA molecule after the first nick. Therefore, the non-cut strand can be used as the template, and the cut strand with the 3′-OH can be used as the primer. The DNA polymerase Bst elongates the primer with strand displacement activity and adds biotin-dUTP to the synthesized DNA strands. The DNA molecules containing biotin-dUTP were captured using streptavidin-coated beads and separated by magnetic field from background DNA (30). The complementing single DNA strand of the biotin-dUTP–containing DNA strand was then isolated after DNA denaturation. The isolated DNA contains SNPs or other sites that can be cleaved by CEL I, which results were confirmed by sequencing and alignment (Figure 2).

To increase the efficiency and decrease the cost during SNP discovery, pooled DNA was used as a template, eliminating the necessity of amplifying and sequencing many individual samples. All SNP-containing DNA fragments isolated with this method were validated in four individuals. Finally, 9 out of 10 fragments were shown to contain at least one SNP. One SNP per fragment is in fact informative for certain applications, such as genetic map construction or association analysis. SNPs located closely in the genome usually exhibit similar behavior, especially in the high-linkage disequilibrium region.

Some sequences isolated contained microsatellite, polyT, and polyC regions, because these regions form mismatches easily and thus can be recognized by CEL I. In any event, a method which can exploit SNPs throughout the whole genome in non-model organisms has been developed. It will prove to be an ideal and promising technique for rapid and efficient identification of SNP sites in a wide variety of organisms. After the SNP isolation in the method we have developed, the confirmation step by sequencing can be replaced by Ecotilling and thus reduce the cost of sequencing. Targeting induced local lesions in genomes (TILLING) and Ecotilling are closely related, powerful high-throughput methods, and are useful in the rapid detection of small mutations or natural polymorphisms (31–34). This application appears highly likely to reduce time and labor, thus improving cost-effectiveness.