| Edited by Ali Masoudi-Nejad. Baozhong Liu: Corresponding author. E-mail: bzliu@qdio.ac.cn |
Microsatellites or simple sequence repeats (SSRs) are highly polymorphic sequences distributed throughout the genome. SSR markers are co-dominant, multi-allelic and easy for scoring, with a wide range of applications including genetic mapping, quantitative trait loci (QTL) association, kinship analysis, population genetics, and evolutionary studies (Pérez et al., 2005). Although their applications in genetic analysis have already been widely demonstrated, the traditional approaches to develop microsatellite loci from non-model species require considerable investment. SSR isolation relies on the creation and screening of enriched microsatellite libraries including cloning and hybridization to detect positive clones, plasmid isolation and Sanger sequencing.
The introduction of a massively parallel pyrosequencing technology developed by 454 Life Sciences Technology has made sequencing cheaper and more efficient. This new technology has been widely applied to the sequencing of microbial genomes, genotyping, genome resequencing, transcriptome profiling and methylation studies (Sithichoke et al., 2009). Fast identification of microsatellites from genome shotgun sequences has been recently acknowledged and investigated in several species, e.g. copperhead snake (Agkistrodon contortrix) (Castoe et al., 2010), blue duck (Hymenolaimus malacorbyncbos) (Abdelkrim et al., 2009) and mungbean (Vigna radiata) (Sithichoke et al., 2009).
Most of the markers developed by genome shotgun sequences correspond to type II markers lacking known functions (Weber, 1990). Similar to genome DNA, EST sequences also contain SSR sequences, which can be used to developed SSR markers (Liu et al., 1999; Whan et al., 2000; Eujayl et al., 2002; Karsi et al., 2002). Therefore, mining microsatellite sequences from the EST sequences of the transcriptome is readily accessible and can be immediately used in the development of specific markers such as EST-SSRs, consuming less time and cost. Furthermore, EST-SSRs, as rich resources of type I markers associated with genes of known functions, can be more useful for comparative gene mapping (Liu et al., 1999) and physical mapping.
The clam M. meretrix is an important commercial bivalve and can be widely found in the coastal and estuarine areas of South and Southeast Asia (Tang et al., 2006). Like other bivalves, the production of M. meretrix mainly relies on the collection of natural spat, despite their economical importance. Recently, selective breeding program have been initiated in clam M. meretrix. The quantitative genetic studies of economical characters such as growth suggested that significant improvement could be achieved by selective breeding (Wang et al., 2011). Therefore, the development of genetic and genomic data are likely to contribute to the development of selective breeding programs, and, more generally, to further the understanding about the genome of this species. In light of all these, we generated and characterized a total of 751,970 original reads covering 310.82 Mb of the M. meretrix larvae transcriptome profiling using 454 sequencing technology. The objective of this study was to identify EST-SSR markers and investigate the type and distribution of repeat motifs in different larval stages of M. meretrix. The results will facilitate the use of molecular markers in M. meretrix genetic analysis and breeding. Moreover, the EST-SSRs specific to different larval stages were compared, which would be of great value to the understanding of the repeat function of gene expression and to the studies of larval development.
The M. meretrix larvae in four developmental stages, i.e., trochophore, D-veliger, pediveliger and postlarva, were used as the source of cDNA libraries in this study. The larvae were collected from a hatchery in July 2009. Then they were frozen in liquid nitrogen then stored at –80°C till use. Samples of M. meretrix larvae in each larval stage were used for RNA purification and sequencing. Total RNA was extracted using phenol-chloroform extraction, and mRNA was purified through affinity chromatography. Four microgram cDNA were synthesized by SMART cDNA synthesis method. A total of four cDNA libraries were ultimately sequenced using 454 Life Sciences technology on the Genome Sequencer (GS) FLX System.
A total of 751,970 reads were generated with the average read length of 413.35 bp covering 310.82 M. They were obtained from the samples of M. meretrix larvae in four larval stages (trochophore, D-veliger, pediveliger and postlarva). All the original reads have been deposited in the NCBI Sequence Read Archive (SRA, www.ncbi.nlm.nih.gov/Traces/sra/sra.cgi?) (Accession No.: SRA021052). A total of 35,205 contigs were yielded with 89,532 reads remaining at singletons after the assembly. Further analysis revealed that 94.66% of the unique sequences showed a comparatively lower expression level (depth < 10), indicating a deep-coverage sequencing (Huan et al., 2011).
In order to identify microsatellite markers, non-redundant sequences were screened for SSRs using MISA software (http://pgrc.ipk-gatersleben.de/misa). In the search for SSR standard, we defined SSRs as dinucleotide repeat (DNP) ≥ 12 bases; trinucleotide repeat (TNP) ≥ 15 bases; tetranucleotide repeat (TTNP) ≥ 20 bases; pentanucleotide repeat (PNP) ≥ 20 base; hexanucleotide repeat (HNP) (and more) ≥ 24 bases (Cardle et al., 2000). Reverse-complement repeat motifs and translated or shifted motifs were grouped together (e,g. AC representing AC, CA, TG and GT) due to the double-stranded nature of DNA and the fact that the start site of a SSR could be considered arbitrary (Jurka and Pethiyagoda, 1995). Primer pairs were designed to amplify microsatellite regions using BatchPrimer3 (Frank et al., 2008). The compound repeat type microsatellite sequences were not included in the designing of amplification primers. We termed such loci with PCR primer as potentially amplifiable loci or PAL.
The expression level of EST-SSRs was considered in the evaluation of the SSR distribution in larval stages. The total number of EST-SSRs in each developmental stage (trochophore, D-veliger, pediveliger and postlarva) was calculated using the following formula: NSSR = Σni=11×D(i), where D(i) is the depth of each contig and singleton containing the SSRs. The depth of contigs containing SSRs ranged between 2 and 375, and the depth of singletons was one. The same formula NT =Σni=11×D(i) was also used in the evaluation of the total number of unique sequences, where D(i) is the depth of all contigs and singletons. The depth of contigs ranged between 2 and 1613.
Chi-square (χ2) goodness-of-fit tests with three degrees of freedom were applied to test whether SSR density was significantly different in four larval stages. The expected number of SSRs was calculated using the following formula: Ei = NSSR/L*Li (Lawson and Zhang, 2008), based on the assumption that the SSR density was the same in the four larval stages. In this formula, N is the total number of SSRs in the four larval stages; Ei is the expected number of SSRs in the four larval stages, trochophore, D-veliger, pediveliger and postlarva, respectively; L is the total length of base pairs in the four larval stages; and Li is the length of base pairs in the four larval stages, trochophore, D-veliger, pediveliger and postlarva, respectively.
2,970 EST-SSRs (2.38%) were isolated from a total of 124,737 contigs and singletons from the above transcriptome, using the Troll software. Among the 2,970 identified EST-SSRs, 1,126 EST-SSRs were identified from contigs and 1,844 EST-SSRs from singletons. These EST-SSRs could be classified into 1194 dinucleotide repeats, 1123 trinucleotide repeats, 357 tetranucleotide repeats, 60 pentanucleotide repeats and 236 other motif types of SSRs including hexanucleotide repeats or repeats with more than 6 bases and compound repeat types (Fig. 1).
 View Details | Fig. 1 Frequency distribution of different repeat types of SSR, and subset of these containing PAL (gray), which were identified in Uni-sequences (214,737 contigs and singletons) of clam M. meretrix. The numbers on the bars indicate the percentage of each repeat type of SSR in total number. |
Except for 233 compound repeat type loci, 2343 PAL loci had flanking regions and could be used for primer design, proving themselves to be promising candidates for PCR amplification. EST-SSRs containing PAL accounted for 78.9% of all microsatellites. The trinucleotide repeats had the most PAL (979), accounting for 33% of the total EST-SSRs. Within each repeat class, besides the compound repeat type, four repeat types (DNP, TNP, TTNP and PNP) had the percentage of PAL higher than 78% (Fig. 1).
Through the comparison across the five classes of repeats, dinucleotide SSR was the dominant repeat type (40.2%). TNPs were found in 1123 SSR loci (37.8%), which was close to the number of DNPs, followed by tetranuleotide (12.0%) and pentanucleotide (2.0%) SSR. There were significant differences in the relative abundances of specific repeat motifs (Fig. 2). The most common motif type of DNPs was TA/AT (71.3% of DNPs), followed by AC/TG (18.3% of DNPs) and TC/AG (10.4% of DNPs). The GC/CG motif was not found in the data set. TA/AT motif also had the largest number of PAL loci in the dinucleotide repeat. Trinucleotide repeat motifs were dominated by TTG repeat, which was found in 512 loci (45.6% of TNPs). TTG repeats contained the largest number of PAL loci in TNPs. The least frequent TNP motif were GC-rich (CAG/TGG/TGC), and the accumulative total value was no more than 9% of TNPs. The number of loci identified for TTNPs was 357, which was less than one third of that of DNPs or TNPs. For tetranucleotides, AAAC, ACAG and TGAT were the three most common motifs, half of all the TTNP motifs were observed fewer than 30 times (data not shown).
 View Details | Fig. 2 Frequency distribution of 4 di-nucleotide repeat mortifs (A) and 7 tri-nucleotide repeat mortifs (B) in Uni-sequences (214,737 contigs and singletons) of clam M. meretrix. The subset of these containing PAL is denoted in gray. The numbers on the bars indicate the percentage of each di-nucleotide repeat mortif in all di-nucleotide repeat types and each tri-nucleotide repeat mortif in all tri-nucleotide repeat types, respectively. |
The repeat number of motif mainly ranged between 5 and 50. There were 909 SSRs with 6 tandem repeats in DNPs, TNPs and TTNPs, followed by 5 tandem repeats in TNPs (623). The two most common repeat number accounted for 51.6% of total EST-SSRs, followed by 7 tandem repeats (15.1%, 450) and 8 tandem repeats (6.9%, 204). More details about different repeat motifs of di-, tri- and tetranucleotide distribution in EST-SSRs were listed in Fig. 3. It showed that with the increase of repetition number, the number of microsatellites declined rapidly. Variations in EST-SSRs in most cases were due to the variations in repeat number. When slippage mutations happen, expansion occurs more frequently to short microsatellites and contraction occurs more frequently to long microsatellites. And the length-dependent mutation pattern explained the scarcity of long microsatellites (Weber, 1990; Temnykh et al., 2000).
 View Details | Fig. 3 The distribution of the number of perfect tandem repeat units for di-nucleotide repeats (A), tri-nucleotide repeats(B) and tetra-nucleotide repeats (C). The numbers on the bars indicate the percentage of repeat numbers in each repeat type. |
The transcriptome in each larval stage was also generated, respectively. The proportion of EST-SSRs in D-veliger, pediveliger, postlarva stage was similar, being 2.7%, and that in the trochophore stage was 3.5% (Table 1). However, the Chi-square goodness-of-fit test showed that there were considerable variations in the density of EST-SSRs in the four developmental stages (Table 1). The densities of di-nucleotide repeats in D-veliger, pediveliger and postlarva stages were slight lower by 4%–6% than expected density, whereas in trochophore stage the density of di-nucleotide repeats was higher by 13.7% than expected density. The density of tri-nucleotide repeats showed the same tendency as di-nucleotide repeats in the four developmental stage, but with more significant variation, e.g., tri-nucleotide repeats density in the trochophore stage was higher by 47.1% than expected density. For the tetra-nucleotide, the repeats densities in the trochophore and D-veliger stage were higher by 26.4% and 39.9% than expected density, respectively, while in pediveliger and postlarva stage, the repeat density was significantly lower than expected density, especially in postlarva stage, it was lower by 55.5% than expected density. The density of EST-SSRs showed no significant variation in four developmental stages for pentanucleotide repeat. The total number of hexanucleotide repeats was small, and therefore it was not analyzed in the goodness-of-fit test.
 View Details | Table 1 EST-SSR densities in different larval stages of M. meretrix |
From each SSR type, we randomly selected a subset of 40 primer pairs including 6 dinucleotides, 11 trinucleotides, 19 tetranucleotides and 4 pentanucleotides to assess the PAL availability rate. Approximately 65% of PAL gave PCR products of expected size and half of PAL were found to be polymorphic (Table 2). The trinucleotides SSR showed the highest PAL availability rate (72.72%), followed by dinucleotides (66.67%) and tetranucleotides (42.11%) types (Table 2). No polymorphic locus was detected in the 4 pentanucleotides (Table 2). Beside three loci (MM9504, MM573 and MM1031) significantly deviated from HWE (P < 0.001) were excluded from the genetic analysis, the characterization of 17 polymorphic EST-SSRs was evaluated in 30 individuals randomly selected from a natural population in Northern China (Lu et al., 2011).
 View Details | Table 2 Primer sequences, characteristics (type and number of repeats), amplification conditions (optimized annealing temperature) and Polymorphic (Yes or No) in the 40 microsatellite loci developed from ESTs in M. meretrix |
Using the next-generation sequencing technology, we obtained more than 700,000 high-quality reads, which could be assembled into 124,737 unique sequences. Although some genomic SSRs of M. meretrix have been developed based on standard enrichment-based approaches, its small quantity still limits its application in genetic mapping and MAS (marker assisted selection). Here an extensive set of 2,970 microsatellites over 12 bp was identified from M. meretrix EST dataset without previously knowing anything about the target genome sequence. The rate of successful amplification of EST-SSRs was high in this batch of PAL, and their primers were available for evolutionary and population genetic research as well as high-resolution chromosome linkage mapping studies of M. meretrix.
The frequency data of SSRs in protein-encoding or noncoding DNA have been widely reported in vertebrates (Moran, 1993; Jurka and Pethiyagoda, 1995; Edwards et al., 1998; Castoe et al., 2010) and plants (Wang et al., 1994; Varshney et al., 2002; Zeng et al., 2010). For example, EST-SSR density of dicotyledonous species ranged between 2.65% and 16.82% (Kumpatla and Mukhopadhyay, 2005), very few studies were conducted on the bivalve. Cruz et al. (2005) once concatenated data from 3165 DNA sequences of 326 bivalve species, accounting for 101 Mb, to determine taxonomic patterns of microsatellite distribution in genomic regions of markedly different functionality. They found that the microsatellite density was 2.9 times higher in introns than in exons. Tanguy et al. (2008) found that less than 0.5% of EST-SSRs were identified in four bivalve species, Bathymodiolus azoricus, Crassostrea gigas, Mytilus edulis and Ruditapes decussatus. The frequency of EST-SSRs detected in M. meretrix was much higher than that of the bivalve species studied. However, it was in agreement with the findings by Li et al. (2010), who identified 3.1% of microsatellite-containing sequences in 1796 unigenes of M. meretrix. The species difference might be the main reason for the variations across species. The EST-SSR frequency is also dependent on other factors, such as different software used in detecting SSRs and customized parameters including the repeat length threshold and the number of repeat unit in microsatellites. For example, some SSR identifying tools can detect imperfect and compound SSRs, while others can only identify perfect SSRs or perfect and compound SSRs (Sharma et al., 2007). In the present study, mononucleotide repeat motifs were excluded in our strategy for identifying microsatellites. The most dominant motifs were di- and tri-nucleotides, which totally accounted for 78.0% of EST-SSRs. It was not consistent with the results reported for other clam species (Tanguy et al., 2008). This result also suggested that there was significant heterogeneity in the most dominant type of repeats, i.e. di- vs tri- nucleotide across bivalve species (Tanguy et al., 2008). In addition, it showed that the most abundant motif in M. meretrix was TA/AT, and the ratio of the (TA/AT)n motif to the second most abundant motif (in our case (AC/TG)n) was nearly 4:1. These data were completely contrary to those of M. galloprovincialis in the distribution type of dinucleotide motif (Cruz et al., 2005). The great variance in motif abundance among species has been reported in bivalves (Cruz et al., 2005; Tanguy et al., 2008).
In all vertebrates, GC-rich motif of trinucleotide is common in exons, but they are less common in intronic sequences (Toth et al., 2000). It was interesting that there was no CG/GC repeat motif and very few CCG/CGG repeats in our results. The rarity of the CCG/CGG repeat units had been reported in a large number of plants (Kumpatla and Mukhopadhyay, 2005). Selection might also be against the CCG/CGG repeat unit due to the requirements of splicing mechanism (Li et al., 2004). Long CCG/CGG sequences was competitive against splicing machinery components, thus giving rise to inadequate splicing. Moreover, CCG/CGG repeats might form potential higher structures such as hairpin and quadruplex, and thus affects the efficiency and accuracy of splicing and the formation of mature mRNA (Coleman and Roesser, 1998; Toth et al., 2000).
The larvae of M. meretrix undergoes a planktonic stage including trochophore, D-veliger, pediveliger, and then metamorphose into postlarval stage. Furthermore, trochophores live on yolks and the larvae start to feed since D-veliger stage. A preliminary differential analysis revealed that several unique sequences showed very different expression patterns in different stages (Huan et al., 2011). It is unclear why di-, tri- and tetra- nucleotides EST-SSR densities showed significant variation in four developmental stages. Microsatellites were non-randomly distributed throughout eukayotic genomes, showing different properties in genomic regions of different functionality (Katti et al., 2001). Moreover, more and more studies suggested that the microsatellites, which were regarded as ‘junk DNA’ in the past, may play a more active role in affecting gene expression (Toutenhoofd et al., 1998; Kenneson et al., 2001; Persico et al., 2006; Lawson and Zhang, 2008). The knowledge of the patterns of microsatellite distribution may help us understand its role in gene regulation and its evolutionary properties. More researches on microsatellite distribution combined with the gene expression profiles in larval stages of M. meretrix need to be done to validate its biological functions.
In conclusion, the molecular markers, such as EST-SSRs, can be applied in the research fields of biodiversity, taxonomy, and population genetics etc. This EST dataset of M. meretrix larvae specific to different developmental stages is a powerful resource for developmental biology and genetics studies.
This project was financially supported by National Basic Research Program of China (2010CB126403) and the Key Program of National Natural Science Foundation of China (30730071).
|
Index