Isolation and Characterization of Microsatellite DNA Markers in the Deep-Sea Amphipod Paralicella tenuipes by Illumina MiSeq Sequencing

Abstract

Here, we describe the development of 16 polymorphic microsatellite markers using an Illumina MiSeq sequencing approach in the deep-sea amphipod Paralicella tenuipes. A total of 25 577 844 DNA sequences were filtered for microsatellite motifs of which 197 873 sequences were identified. From these sequences, 64 had sufficient flanking regions for primer design and 16 of these loci were polymorphic. Between 5 and 30 alleles were detected per locus, with an average of 13.63 alleles per locus, across a total of 120 individuals from 5 separate deep sea trenches from the Pacific Ocean. For the 16 loci, observed and expected heterozygosity values ranged from 0.116 to 0.414 and 0.422 to 0.820, respectively, with one locus displaying significant deviation from Hardy–Weinberg equilibrium. The microsatellite loci that have been isolated and described here are the first molecular markers developed for deep sea amphipods and will be invaluable for elucidating the genetic population structure and the extent of connectivity between deep ocean trenches.

The Abyssal and Hadal zones are the 2 deepest marine biozones (3000–6000 m, and 6000–c. 11 000 m respectively). These are perhaps the most poorly explored and least understood ecosystems on Earth, particularly the hadal trenches (Jamieson and Fujii 2011; Jamieson 2015). The abyssal and hadal zones differ in that the former comprises large geographically expansive plains while the latter comprises deep trenches formed by tectonic convergence, the majority of which are located around the Pacific Rim (Jamieson et al. 2010). These trenches disrupt the continental shelf-slope-rise to abyssal plain continuum in that they form an array of spatially disjunct ultradeep habitats which are often isolated by extremely large geographical distances. As a result of this isolation, it was conventionally believed that hadal trenches lacked any level of connectivity through dispersal and gene flow, which would ultimately promote speciation and high levels of local species endemism (Wolff 1960, 1970; Beliaev 1989).

An excellent model system to test such supposition and examine patterns of connectivity, or conversely isolation, between trenches is the amphipods of the superfamily Lysianassoidea. They are a diverse, ubiquitous, and abundant component of the abyssal and hadal fauna that can be readily sampled using baited traps (Blankenship et al. 2006; Jamieson et al. 2010; Fujii et al. 2013). The presence of seemingly cosmopolitan species of lyssianassoid amphipod, such as Eurythenes gryllus, within different trenches and across the adjoining abyssal plains would suggest that dispersal is possible between these habitats, despite the large geographical distance, and argues against any level of isolation among trenches. In contrast, however, apparent morphological and phylogeographic differences between populations of the same putative species suggests that populations are not panmictic and trenches do represent reproductively isolated habitats where gene flow is limited (Havermans et al. 2013).

The use of neutral molecular markers, such as microsatellite DNA length polymorphisms, can provide more direct estimates of the extent of population genetic structure between populations, and as such can be utilized to assess the extent to which hadal trenches represent demographically and evolutionarily independent units. Here, we develop a suite of microsatellite markers for the lysianassoid amphipod Paralicella tenuipes, which is an abundant, cosmopolitan species found on the abyssal plains and hadal trenches around the entire Pacific Rim. We isolate and characterize multiple microsatellite loci by using Illumina MiSeq high-throughput DNA sequencing which is a useful tool for isolating new genetic markers in species with poor genomic resources and large genome sizes (Davey et al. 2011), while also circumventing the technical challenges of more traditional microsatellite enrichment protocols (Ekblom and Galindo 2011).

Materials and Methods

Sample Collection

Specimens were collected from the Kermadec, Japan, Mariana, Peru-Chile, and New Hebrides trenches around the Pacific Rim across multiple cruises (for details see Ritchie et al. 2015). Samples were collected using an autonomous lander vehicle (Jamieson et al. 2009) which was deployed with funnel traps secured to the feet and baited with locally sourced fish. Species were identified from species-specific mitochondrial COI barcodes according to Ritchie et al. (2015).

DNA Extraction and Sequencing

Genomic DNA (gDNA) was extracted from 4 whole individuals (~3–4mm) from the Japan trench using a conventional phenol–chloroform approach. Extractions were treated with 20mg/mL RNAse A (Sigma-Aldrich). DNA was visualized for high molecular weight on a 1% TBE agarose gel then quantified using a Qubit 2.0 Fluorometer (Life Technologies) with subsequent dilution to ~20ng/µL. After purification a total of 5 µg of DNA pooled from 4 individuals was used for library preparation and shotgun sequencing by fragmentation at MWG Eurofins (Germany) following manufacturer’s protocols (Illumina, San Diego, CA).

Microsatellite Discovery

One lane of Illumina MiSeq v3, using a 2×350bp paired-end length configuration, yielded 15 346 Mbp of data that passed Illumina quality control. Paired fragment libraries were aligned before de novo assembly using ABySS v1.5.2 (Simpson et al. 2009). Sequences had an approximate k-mer length of 25bp (k = 25) and a minimum number of 5 pairs were required to join ends (n = 5). A total of 25 577 844 sequences were filtered for microsatellite arrays containing di-, tri-, or tetra-nucleotide motifs with a minimum of 5 repeats using the software mreps v2.6 (Kolpakov et al. 2003). This returned 197 873 microsatellite loci. Of these loci, 64 had over 50 base pairs of flanking sequence both upstream and downstream of the array to allow for primer design using PRIMER 3 (Rozen and Skaletsky 2000).

Primer Testing

A total of 64 primer pairs were synthesized for amplification testing. Initial testing for PCR amplification success was undertaken using a total of 16 individuals from 3 different trench locations in the Pacific Ocean. PCR amplification reaction mixes contained 0.2mM each dNTPs, 2.5mM MgCl₂, 0.5 µM each primer, 0.5U of BioTaq DNA polymerase (Bioline), 5–20ng DNA template in 1× NH₄ buffer (Bioline: 16mM (NH₄)₂SO₄, 67mM Tris–HCl) in a total reaction volume of 10 µL. PCR amplification was performed using a G-storm thermal cycler (G-storm Ltd, Surrey, UK) with the following touch-down conditions: initial denaturation at 94 °C for 2min, followed by; 20 cycles of denaturation at 94 °C for 30s, annealing at 65 °C–55 °C (−0.5°C/cycle) for 30s, extension at 72 °C for 30s; 15 cycles of denaturation at 94 °C for 30s, annealing at 55 °C for 30s, extension at 72 °C for 30s before a final elongation step at 72 °C for 1min.

After initial screening, 40 loci were excluded either due to PCR failure, lack of polymorphism or ambiguous profiles when genotyped. The remaining 16 loci were selected for genotyping across a total of 120 individuals from 5 different trenches to examine the level of polymorphism (Table 1). The selected loci are composed of 1 tetranucleotide repeat, 3 trinucleotide repeats, and 12 dinucleotide repeats. Forward primers were labeled with 1 of 4 fluorescently labeled universal tails A: 6-Fam GCCTCCCTCGCGCCA, B: HEX GCCTTGCCAGCCCGC, C: ATTO 550 CAGGACCAGGCTACCGTG or D: ATTO 565 CGGAGAGCCGAGAGGTG (MWG Eurofins, Germany) following the procedures outlined in Blacket et al. (2012).

Table 1.

Microsatellite loci characteristics including locus name, primer sequences, repeat motif, size range (base pairs), flurochome primer label, and GenBank accession number

Locus	Primer Seq	Primer Seq	Repeat Motif	Size Range (bp)	Flurochrome	GenBank Accession Number
Paraten1	F:TCTTATGAACTCGACTGTACG	R:CACAGTAGATATCCCTGCATC	GCGT	115–151	6-FAM	KT587291
Paraten3	F:CTCAGGAAACCTCAGATATCAC	R:AGATCATCTCCTTGGTGGTT	ACC	121–127	HEX	KT587292
Paraten5	F:AGGTCACCACTCTGTAAATTCT	R:CGCTACCTTACTAACGGATATT	CTA	123–141	6-FAM	KT587293
Paraten7	F:AGCGGTGGACATCTTAATAAC	R:TCAGAAATACAGTCACTTGAGG	GAT	144–168	ATTO550	KT587294
Paraten10	F:AGAATCTCTGTACCATCAGAGG	R:ACAGTATGATAATGGTGAGCTG	GA	151–259	ATTO550	KT587295
Paraten11	F:ACGGCTGAGTCTAAGATATAGG	R:TGAATTAGCCTTTAGTTCAACA	AC	198–218	HEX	KT587296
Paraten13	F:AGAAGTAACGGACTTCACGAT	R:AATTCGACTAGTTGATCCTACG	GA	178–184	ATTO550	KT587297
Paraten14	F:TACTCCGACACAGGATATGTAA	R:ACGAGTTTCCAGAATTACAAGA	TA	140–182	ATTO565	KT587298
Paraten19	F:CTTCCAATGGTCTATCCTAATC	R:GCTACGGTGTATGTGTTCACT	AG	243–249	ATTO565	KT587299
Paraten27	F:GTAACTGCGGCGAGGATCA	R:AGTCAGGCAAGGAAACGGA	AT	134–210	HEX	KT587300
Paraten30	F:GAAGTTTCGATACAGTCGGTAG	R:AAGACATTTGCAAGACTACCAT	TA	117–157	ATTO565	KT587301
Paraten38	F:ATTTACTGCTAATTGCTGCTG	R:TGAATGCAACCAATTATAAACC	AT	127–173	HEX	KT587302
Paraten41	F:GACGATTAACATGCACATCACT	R:ATACTGACACGCCTCAGATGAC	AC	108–128	ATTO565	KT587303
Paraten50	F:ATGATACGTGCAAGAATGTTTA	R:CTTAATTATCGGTTTTCATTGC	AT	115–149	6-FAM	KT587304
Paraten54	F:GATGTTCGTTACTGGAAGAAGT	R:CACGACACAATCTACATATCCT	AG	115–149	6-FAM	KT587305
Paraten57	F:GTCGTTTGACTCATAAAGGACT	R:AGCATTGGAAACCAATAACTT	GT	156–172	ATTO550	KT587306

PCRs were performed in singleplex with a reaction mix containing 0.2mM each dNTPs, 2.5mM MgCl₂, 0.5 µM of reverse primer, 0.3 µM of forward primer with universal tail, 0.2 µM of flurochorome-labeled forward primer, 0.5U of BioTaq DNA polymerase (Bioline), 5–20ng DNA template in 1× NH₄ buffer (Bioline) in a total reaction volume of 10 uL. PCR amplification was performed using a G-storm thermal cycler (G-storm Ltd, Surrey, UK) with the following touch-down conditions: initial denaturation at 94 °C for 2min, followed by; 20 cycles of denaturation at 94 °C for 30s, annealing at 65 °C–55 °C (−0.5°C/cycle) for 30s, extension at 72 °C for 30s; 15 cycles of denaturation at 94 °C for 30s, annealing at 55 °C for 30s, extension at 72 °C for 30s before a final elongation step at 72 °C for 15min.

Genotypes were resolved on an automatic ABI 3730 DNA Capillary DNA sequencer (Dundee DNA Sequencing Services Ltd) using a GS-500 (LIZ) size standard. Individual electropherograms were scored by eye using GeneMarker v 1.4 (SoftGenetics 2010).

Microsatellite Loci Analysis

Micro-Checker 2.2.3 (van Oosterhout et al. 2004) was used to establish whether any observed heterozygote deficiencies were attributable to null alleles, scoring errors, or large allelic dropout. Tests for linkage disequlibrium between all possible loci combinations were conducted using Genepop 4.1 (Rousset 2011). Genetic diversity statistics were calculated using a variety of statistics: observed (H_O) and expected (H_E) heterozygosity, number of alleles per locus (N_A), and the fixation index (F_IS) were calculated alongside a test for deviation from Hardy–Weinberg equilibrium (P_HWE), where p-values were subsequently adjusted after Bonferroni correction, using the diveRsity package in R (Keenan et al. 2013) (Table 2).

Table 2.

Summary statistics of 16 polymorphic microsatellites in Paralicella tenuipes

Locus	N	NA	HO ± SD	HE ± SD	FIS	PHWE
Paraten1	98	10	0.340±0.11	0.438±0.09	0.2356	0.2987
Paraten3	103	5	0.294±0.06	0.466±0.06	0.3712	0.1501
Paraten5	69	13	0.282±0.14	0.738±0.07	0.6155	0.1626
Paraten7	85	10	0.322±0.20	0.446±0.19	0.2729	0.5231
Paraten10	77	18	0.330±0.16	0.676±0.13	0.5139	0.2681
Paraten11	97	15	0.342±0.16	0.576±0.22	0.4044	0.1237
Paraten13	82	8	0.116±0.06	0.422±0.23	0.7289	0.1953
Paraten14	66	16	0.200±0.08	0.644±0.07	0.7348	0.0359*
Paraten19	82	9	0.272±0.18	0.492±0.12	0.3509	0.3848
Paraten27	76	20	0.332±0.18	0.764±0.07	0.6423	0.1106
Paraten30	80	11	0.350±0.13	0.568±0.22	0.3369	0.1349
Paraten38	92	30	0.304±0.19	0.820±0.13	0.6319	0.0014**
Paraten41	86	11	0.246±0.20	0.646±0.09	0.5592	0.1142
Paraten50	89	15	0.366±0.17	0.582±0.08	0.3772	0.3953
Paraten54	88	17	0.352±0.17	0.802±0.08	0.5617	0.0136*
Paraten57	93	10	0.414±0.19	0.734±0.14	0.4620	0.2585

N, number of individuals with successful amplification; N_A, number of alleles per locus; H_O, observed heterozygosity with standard deviation; H_E, expected heterozygosity with standard deviation; F_IS, fixation index; P_HWE, P value for Hardy–Weinberg equilibrium (where * denotes deviation at P < 0.05 and ** denotes deviation with Bonferroni correction P < 0.0025).

Data Availability

In accordance with the Journal of Heredity data archiving policy (Baker 2013), we have deposited the primary data underlying these as analyses as follows: DNA sequences as Genbank accessions KT587291–KT587306.

Results and Discussion

Here, we demonstrate the utility of high throughput sequencing for isolating microsatellite loci in deep sea amphipods. In this case, 197 873 microsatellite containing sequences were identified and from these a total of 64 microsatellite loci primers were designed for testing. Although there is a high attrition rate due to the generation of short reads with insufficient DNA sequence for primer design the number of primer sets that can be developed is still much greater than what would be expected from traditional microsatellite enrichment protocols. Next-generation sequencing produces a much larger, cost-effective database for microsatellite discovery more rapidly than traditional cloning-based approaches (Yu et al. 2011). It also has the additional benefit of limiting sequencing errors due to greater sequence coverage and quality control filters. While other next-generation sequencing approaches, such as RAD-seq, can provide a considerably greater number of markers with 1000s of single nucleotide polymorphisms (SNPs) for inferring population structure the relative amount of high-quality DNA required for each individual can be too high for samples with inherently poor DNA quality. Furthermore, these methods can also result in data with considerable noise where it is difficult to distinguish between true heterozygotes and artificially assembled variants of nonhomologous but similar RAD-tag loci which is a major shortcoming for study systems without a reference genome (Mesak et al. 2014). This Illumina Miseq approach provides a valuable resource for a whole spectrum of taxa regardless of genome size and availability of genomic resources.

Of the 64 microsatellite loci investigated, 16 were polymorphic. The number of alleles (N_A) per locus varied from 5 to 30, with an average of 13.63 which is comparable to other amphipod microsatellite loci (Baird et al. 2012). Observed (H_O) and expected (H_E) heterozygosities ranged from 0.116 and 0.414, and 0.438 and 0.820, respectively. The expected heterozygosities are comparable to microsatellite loci obtained from other amphipod species but the observed heterozygosities are slightly lower (Baird et al. 2012).

Tests for linkage disequilibrium identified no significant interactions between any of the pairwise comparisons indicating that none of the loci are physically linked. Three markers showed significant deviations from HWE at the P < 0.05 level (Paraten14, Paraten38, and Paraten54) due to a deficiency of heterozygotes but this was reduced to a single locus after Bonferroni correction (Paraten38). This lack of statistical significance is probably caused by low power due to small sample sizes and missing data which are both factors associated with deep-sea samples. Notwithstanding, the substantial heterozygote deficiency is probably biologically meaningful (Waples 2015), and might be caused by population subdivision (Wahlund effect) or conflation of multiple cryptic species. These intriguing hypotheses will need to be addressed in a large-scale population study. Micro-Checker provided no support for this excess of homozygosity being due to null alleles, though this is somewhat difficult to reconcile given that only this single locus was not in Hardy Weinberg equilibrium. An alternative explanation is that the locus is linked to a gene under selection (Slatkin 1995).

Amphipod samples retrieved from the deep sea frequently yield poor quality DNA due to the extreme change in hydrostatic pressure experienced by the cells which often results in cell death and DNA damage (Dixon et al. 2004). This is further compounded by a time delay between the death of the animal and its preservation on deck which allows endogenous nucleases to begin degrading the DNA (Hofreiter et al. 2001). As a result of this degradation, the microsatellite primer sets were amplified in singleplex to maximise amplification success but with sufficient DNA quality these could be combined and the PCRs could be performed in multiplex.

Here, we have produced novel polymorphic microsatellite markers from the first time in an abyssal and hadal amphipod species. These microsatellite loci will provide essential data for inferring the true patterns of movement across abyssal plains and between hadal trenches in an important amphipod species. This will allow us to address the long standing question of whether hadal trenches truly represent geographically isolated habitats which in turn have resulted in distinct, isolated populations of species.

Funding

This work was supported by the HADEEP projects, funded by the Nippon Foundation, Japan (2009765188), the Natural Environmental Research Council, UK (NE/E007171/1) and the Total Foundation, France.