The Antarctic silverfish (Pleuragramma antarctica) is a critically important forage species with a circumpolar distribution and is unique among other notothenioid species for its wholly pelagic life cycle. Previous studies have provided mixed evidence of population structure over regional and circumpolar scales. The aim of the present study was to test the recent population hypothesis for Antarctic silverfish, which emphasizes the interplay between life history and hydrography in shaping connectivity. A total of 1067 individuals were collected over 25 years from different locations on a circumpolar scale. Samples were genotyped at fifteen microsatellites to assess population differentiation and genetic structuring using clustering methods, F-statistics, and hierarchical analysis of variance. A lack of differentiation was found between locations connected by the Antarctic Slope Front Current (ASF), indicative of high levels of gene flow. However, gene flow was significantly reduced at the South Orkney Islands and the western Antarctic Peninsula where the ASF is absent. This pattern of gene flow emphasized the relevance of large-scale circulation as a mechanism for circumpolar connectivity. Chaotic genetic patchiness characterized population structure over time, with varying patterns of differentiation observed between years, accompanied by heterogeneous standard length distributions. The present study supports a more nuanced version of the genetic panmixia hypothesis that reflects physical-biological interactions over the life history.

A total of 1067 Pleuragramma antarctica individuals from 19 sampling locations collected by multiple institutions between 1989 and 2014 were included in the analysis. Of these, 249 individuals collected between 1989 and 1997 from the Ross Sea (RS96, RS97), Halley Bay (HB89, HB91), and South Shetland Islands (SSI96) were included from a previous study on population structure using mitochondrial DNA (Zane et al. 2006). A further 562 individuals collected between 2001 and 2012 from Larsen Bay (LB07), Charcot Island (CI10), Marguerite Bay (MB01, MB02, MB10, MB11), and Joinville Island (JI07, JI10, JI12) were included from a previous connectivity study focused on the Antarctic Peninsula (Agostini et al. 2015). The remaining 256 individuals collected in 2011 and 2014 from the South Orkney Islands (SOI11), Larsen Bay (LB11), Filchner Trough (FT14), Atka Bay (AB14), and Halley Bay (HB14) have not previously been examined. All samples, including those obtained in 2011 and 2014 that had not been previously examined, were collected during authorized scientific cruises carried out by nations legally committed to the Convention for the Conservation of Antarctic Marine Living Resources (CCAMLR 1980). No manipulations were carried out on live samples.

Muscle tissue or fin clips were preserved in 95-100% ethanol at the time of sampling. Genomic DNA was extracted from specimens using a standard salting-out procedure (Patwary et al. 1994). Genomic DNA was extracted from all previously unexamined samples, as well as from tissue samples remaining from individuals on which the mitochondrial DNA analysis was performed in a previous study (Zane et al. 2006). Concentration and quality of the extracted DNA (260/280 nm and 260/230 nm) was checked using a NanoDrop UV–Vis spectrophotometer (Thermo Scientific) prior to PCR amplification. All extracted DNA was of high enough quality to use in subsequent PCR reactions.

Individuals were genotyped using 16 published EST-linked microsatellites developed in Chionodroco hamatus (Molecular Ecology Resources Primer Development Consortium et al. 2011, Agostini et al. 2013), that had been previously shown to cross-amplify successfully in P. antarctica (Agostini et al. 2015). The 16 loci were amplified in two multiplex PCR reactions. Multiplex PCR reaction volume was 10 µL, containing 1x QIAGEN Multiplex PCR Master mix (HotStartTaq DNA Polymerase, Multiplex PCR Buffer, dNTP Mix; QIAGEN, Hilden, Germany), 0.2 µM primer mix and 100 ng of genomic DNA. The PCR amplification profile for all loci consisted of: (1) an initial activation step of 15 min at 95 ˚C; (2) 30 cycles of denaturation at 94 ˚C for 30 s, annealing at 57 ˚C for 90 s, and extension at 72 ˚C for 60 s; and (3) a final extension of 30 min at 60 ˚C.

PCR products were prepared for microsatellite genotyping and sent to an external service (BMR Genomics, http://www.bmr-genomics.com/), where they were sequenced using an ABI 3730xl automated sequencer (LIZ 500 as internal size ladder, Applied Biosystems, Waltham, MA, USA). Microsatellites were analyzed using GeneMarker ver. 2.6 (SoftGenetics). Genotypes of individuals included in a previous microsatellite-based analysis of P. antarctica (Agostini et al. 2015) were integrated into this analysis. To ensure that the datasets were comparable, two individual samples from this previous study (Agostini et al. 2015) were processed alongside samples from the present study as positive controls during each amplification and sequencing run. This allowed the comparison of the raw microsatellite data between the previous and current study and confirmed that there was no change in microsatellite sizing. Binning was automated with FlexiBin 2 (Amos et al. 2007) and refined by eye to assure accuracy with the corresponding binnings established in the previous study (Agostini et al. 2015).

Pleuragramma2018_Genotypes_15loci_GENEPOP.txt

Text file containing the genotypes of all analyzed individuals at the 15 microsatellite loci considered in the final analysis. Data are divided by the 19 population divisions based on the 19 sampling provenances of samples. File prepared using CREATE ver 1.37 (Coombs et al. 2008) for compatibility with GENEPOP online (Rousset 2008), a common format for downstream bioinformatics analyses.

Pleuragramma2018_Genotypes_metadata.xlsx

Excel file containing all metadata associated with the genotypes produced in this study.

Metadata (columns A – U) headers are:

• Study #: the ID number of the samples used in the present study. Samples without Study # (na) represent samples from a previous study (Agostini et al. 2015) - only the genotypes of these samples were used in the present study;
• Extract-DNA box ID: the DNA extraction ID combined with the ID of the storage box where the genomic DNA was stored (na values refer to samples from Agostini et al. 2015);
• BMR CODE: the ID given to sequences for samples produced by BMR (na values refer to samples from Agostini et al. 2015);
• Age: life stage: adult, juvenile, larva (na values refer to samples from Agostini et al. 2015);
• Tissue type: the type of tissue from which DNA was extracted: muscle, fin clip, fin/m (fin and muscle), whole (for larvae and juveniles) (na values refer to samples from Agostini et al. 2015);
• Year: sampling year;
• Geographic Origin: sampling area;
• Comment: additional information related to the sample, where relevant;
• Dataset: Cecilia (Agostini et al. 2015) or Jilda (present study);
• Population: population labels corresponding to sampling areas, representing the 19 population divisions;
• Pop specific: population labels corresponding to sampling location, further dividing the 19 population divisions into population subdivisions by sampling station within the 19 sampling areas;
• study#_pop ID: individual sample IDs based on Study # and population ID for samples from the present study, and Agostini et al. 2015 sample ID and population ID for samples from Agostini et al. 2015;
• study#_pop specific ID: individual sample IDs based on Study # and pop specific ID for samples from the present study, and Agostini et al. 2015 sample ID and pop specific ID for samples from Agostini et al. 2015;
• Pop 6: population labels corresponding to the 6 population divisions based on sampling region;
• station: sampling station information where available;
• depth: sampling depth in m;
• SL: standard length of sampled fish in cm where available;
• Cruise: name of research cruise during which sample was collected;
• n: total number of individuals in Pop specific group divisions;
• Lat: latitude of sampling location, where available;
• Long: longitude of sampling location, where available.

Genotype data (columns V – BA) correspond to the genotype information for samples at the 16 total microsatellite loci sequenced in this study. Only 15 loci were ultimately used in the analysis, as allele Ch11230, columns AP and AQ, was problematic due to stuttering and interference from other fluorophores used during sequencing. Genotype information is expressed as whole numbers based on the output derived from Flexibin 2 (Amos et al. 2007), which takes microsatellite size information in base pairs and groups these into discrete allele categories based on the repeat length of the microsatellite. In this spreadsheet, each locus has two columns of information, the first providing the genotype of the first allele for that locus for a given individual, the second providing the genotype for the second allele for that locus for a given individual.

Amos W, Hoffman JI, Frodsham A, Zhang L, Best S, Hill AVS (2007) Automated binning of microsatellite alleles: Problems and solutions. Molecular Ecology Notes 7:10-14