Genetics and Hatchery Management: A Parentage-based Tagging Approach to Blueback Herring Conservation

Blueback Herring Alosa aestivalis in their native range are found along the Atlantic coast of North America and range from Nova Scotia to Florida. This species is anadromous, spending the majority of their adult lives at sea and returning to freshwater systems to spawn (Bigelow and Schroeder 1953; Loesch 1987). Historically, Blueback Herring play important roles in the ecosystems they transition between, acting as a prey source as well as a transporter of nutrients between freshwater and saltwater. Blueback Herring have also served as a reliable resource for humans as food and fertilizer (Messieh 1977; Loesch 1987; West et al. 2010). Once a thriving commercial fishery, numbers of Blueback Herring have been declining since the late 1980s (NCDMF 2000). In 2006, the National Oceanic and Atmospheric Administration (NOAA) listed Blueback Herring as a Species of Concern (NOAA 2009). Probable reasons for population decline include habitat degradation, overfishing, dam impediments, bycatch, and predation. In an effort to restore Blueback Herring populations to historical levels, several states, including Massachusetts, Rhode Island, Connecticut, and North Carolina, enacted stringent regulations and moratoriums on river herring fisheries as early as the year 2000 (NCDMF 2000). Yet a study conducted by the Atlantic States Marine Fisheries Commission in 2012 found populations still depleted throughout their range (Limburg et al. 2012).

Current conservation efforts, including NOAA's River Herring Conservation Plan, emphasize public awareness, collaborative research, and restoration of river herring throughout their native range (NOAA 2016). One method of restoring abundance focuses on fry stockings originating from hatchery-based spawning. A reliable and conclusive method to track the survivability of these fry is vital in evaluating the success of these programs. Tracking methods have traditionally involved physical or chemical tags such as coded-wire tags or oxytetracycline staining of otoliths, respectively. However, these methods face challenges such as tag loss, high cost, and increased mortality (Skalski et al. 2009; Pine et al. 2012). Advances in genetic technologies, along with a subsequent decline in the cost of performing such analyses, make genetic parentage-based tagging (PBT) an appealing alternative (Andreou et al. 2011; Steele et al. 2013).

A critical component of PBT is the correct identification of fish to species when collecting broodfish for spawning and when sampling fish to evaluate hatchery contribution. However, Blueback Herring share many morphological similarities with the closely related Alewife Alosa pseudoharengus, making these two species difficult to distinguish phenotypically (Bigelow and Schroeder 1953; NOAA 2009; Limburg et al. 2012). Blueback Herring and Alewife also share ecological similarities including anadromous life cycles, diet, and spatial distribution (Bigelow and Schroeder 1953; Loesch 1987) and are generally treated as a single unit for management considerations (NOAA 2009) under the umbrella term of “river herring.” In unaltered ecosystems, river herring achieve isolation largely through temporal and spatial spawning differences, with Blueback Herring spawning later in the season in warmer, lotic waters and Alewife preferring cooler, lentic conditions (Loesch 1987). However, the construction of dams and other manmade migration restrictions has increased the likelihood of hybridization between the two species due to enforced spatial overlap, and recent studies have indicated that such hybridization is occurring in the wild (Hasselman et al. 2014; McBride et al. 2014). Moreover, analyses of Blueback Herring and Alewife in North Carolina rivers indicate that spawning runs largely overlap in this area (Walsh et al. 2005; Overton et al. 2012; Potoka 2016), further increasing the possibility of hybridization in these waters.

Given the challenges of phenotypic discrimination, the need for reliable markers capable of genetically distinguishing one species from the other as well as identifying possible hybrids is a priority for future studies and management of river herring (River Herring TEWG 2016). Identification of a species-specific marker within the mitochondrial genome has proven difficult as work by Chapman et al. (1994) and Faria et al. (2006) indicates notable sequence similarities between the mitochondrial genome of these two species, with divergence rates of approximately 2%. We are aware of one nuclear marker, recombination activating gene 2 (rag2), capable of distinguishing between Blueback Herring and Alewife (Berlinsky et al. 2015). Usage of this marker requires polymerase chain reaction (PCR) amplification followed by either genetic sequencing or digestion with the Bam HI restriction enzyme. A genetic marker capable of differentiating between river herring in a single assay would be of benefit to researchers interested in studying these species as well as to hatchery management. Moreover, hybridization between these two species warrants the need for multiple markers to ensure confidence in species discrimination.

To address management needs for Blueback Herring, we developed a robust microsatellite marker suite that can be used for reliable PBT. This maker set will allow fisheries managers to evaluate the contribution of Blueback Herring from hatcheries to rivers of interest, examine population structure between bodies of water, identify migration between rivers, and assess genetic diversity levels and effective population size (Buchholz-Sørensen and Vella 2016; Fazzi-Gomes et al. 2017; Ywasaki Lima et al. 2017). We also report that microsatellite marker AsaC334 (Julian and Bartron 2007) may be used to differentiate Blueback Herring from Alewife or their F1 hybrids, allowing researchers to address both parentage and species identification in a single assay.

We collected broodfish from three tributaries of the Chowan River (Bennett's Creek, Dillard's Creek, and Sarem Creek) using a boat-mounted electrofishing unit (Smith-Root 7.5 GPP; 170–1,000 V pulsed direct current; 3.5–4.5A) with one dip-netter used to capture fish during daylight hours. We began broodfish collections as water temperatures approached 12°C and concluded collections when spawning activity appeared to be complete or when we met broodfish collection goals (typically from late March through early April). We determined sex for each Blueback Herring by applying directional pressure to the abdomen and observing the presence of milt or eggs. To account for similarities between Alewife and Blueback Herring, we checked species identification a second time during careful transfer from boat to hauling trailer; and we then took fish to U.S. Fish and Wildlife Service Edenton National Fish Hatchery or Watha State Fish Hatchery. We collected 451 broodfish in the year 2013, 290 broodfish in 2014, and 49 broodfish in 2015. We collected 50 juvenile fin clips from wild spawned Blueback herring in 2013 during Division of Marine Fisheries routine seine sampling in the Albemarle Sound. We collected a sample of Alewife fin clips from the lower Roanoke River on 16 November 2015, during routine sampling of juvenile American Shad Alosa sapidissima near Plymouth, North Carolina (pulsed direct current; 3.8–4.2A) roughly 30 min after sunset.

We cultured Blueback Herring fry from tank-spawned broodfish at the Edenton National Fish Hatchery in 2013 and 2014 and Watha State Fish Hatchery in 2015. We held Blueback Herring broodfish in a 6-ft-diameter (1.8-m-diam) circular tank supplied with well water (18°C) and collected eggs via external standpipe and filter sock with a mesh size of approximately 200 μm. Upon egg collection, we treated eggs with Fuller's Earth or Tannic Acid and placed them in hatching jars, where fry were allowed to hatch into aquaria (75 L) at 3–4 d after spawn (Evans 2015). We stored fin clips and samples of Blueback Herring fry from each spawning in prelabeled vials containing nondenatured, spectrophotometric-grade ethanol for use as positive controls in PBT.

In order to identify novel microsatellite markers (Gardner et al. 2011; De et al. 2017), NextGeneration sequencing of one Chowan River Blueback Herring was conducted by the North Carolina State Genomic Sciences Laboratory on an Illumina MiSeq using MiSeq Reagent Kit v3 (600 cycle). We assembled approximately nine million paired-end reads with an average read length of 455 base pairs using QIIME (Caporaso et al. 2010) and fed the resulting file into MSATCOMMANDER (Faircloth 2008) to identify possible microsatellite markers. Many of these markers are likely to be duplicates, so we chose a subset of microsatellites with varying motifs to help ensure selection of unique markers for further testing. We additionally opted for loci containing a large number of repeat units because such microsatellites tend to be more unstable, thus leading to increased variability within the marker (Ellegren 2000; Brohede et al. 2002). In total, we initially tested 24 candidate markers for inclusion in our Blueback Herring microsatellite marker suite. We also selected an additional 16 primer pairs from two published papers (Julian and Bartron 2007; A'Hara et al. 2012) for testing on our samples. We chose markers with the greatest numbers of alleles according to their publications: Ap037, Ap070, Aa004, AsaD055, AsaD030, AsaC249, Ap058, AsaD042, AsaD021, Aa082, Aa074, AsaC334, Ap047, Aa093, AsaC051, and Ap071 (Table 1).

We extracted genomic DNA from samples using the Macherey–Nagel NucleoSpin 96 Tissue kit and processed on an Eppendorf Robotic liquid handler (epMotion 5075) or with the Macherey–Nagel 8 Tissue kit and processed with the Macherey–Nagel vacuum manifold. Extracted DNA was quantified and diluted when necessary to a working concentration of approximately 40 ng/μL. Markers identified through NextGeneration sequencing were initially amplified using a three-primer PCR method described by Schuelke (2000) and briefly analyzed for robust amplification and for consistent allele sizes using GeneMapper 4.0 (ThermoFisher Scientific, Waltham, MA). Six of these markers were selected for further analysis with our Blueback Herring samples: AaAC4, AaAG5, AaAG7, AaAC6, AaACAG1, and AaAGAT2, named according to their species and repeating unit. Genbank accession numbers are KY554778, KY554779, KY554780, KY554781, KY554782, KY565243, respectively (Table 1; Figure S1, Supplemental Material).

The 6 NextGeneration identified markers and the 16 published markers were amplified in 4 multiplex PCR reactions (Table 1) with a 1:10 mixture of Takara ExTaq Premix and Promega GoTaq MasterMix with 1 μL of genomic template per reaction and primers labeled with fluorescent dye. The thermal profile employed an initial denaturation of 95°C for 4 min followed by 5 cycles of 95°C for 15 s, 62°C for 15 s, and 72°C for 30 s; then 30 cycles of 94°C for 30 s, 58°C for 30 s, and 72°C for 30 s with a final elongation step at 72°C for 10 min. One μL of amplified product was run on an ABI 3130XL Genetic Analyzer using ABI GeneScan 600 LIZ Size Standard, and resulting fragment sizes were analyzed using GeneMapper 4.0. Genotyping was run for 451 Blueback Herring broodfish collected from the Chowan River in 2013 and spawned in six separate hatchery tanks, 50 wild-spawned juvenile Chowan River Blueback Herring to use as negative controls, and 41 fry from known breeding tanks to use as positive controls in our PBT test.

We analyzed allele frequencies for the 22 described microsatellite markers, including conformity to Hardy–Weinberg equilibrium, and consequent parentage-based tagging using CERVUS 3.0.3 (Kalinowski et al. 2007). This program utilizes a maximum likelihood approach, creating simulations from the data set to assign confidence levels. Simulations assumed 10,000 offspring, 100 candidate parents (with 100% of the parents being sampled), low mistyping error rate (0.001), and low error rate (0.0001). At least three simulations were executed per analysis. Critical delta scores were determined using 95% confidence for the relaxed criterion and 99% for the strict criterion. Parentage analyses were performed without reference to sex determination or spawning tank.

We amplified one μL of genomic DNA in a 10-μL reaction using Promega GoTaq MasterMix and 0.2 μL of AsaC334 primers (10 μM; F: ATG GTT ATG TGG GCT CTT TAT G, R: GTT CAT CCT GCC AGA TCT AAG G). Cycling conditions for AsaC334 amplification followed the same protocol as that for microsatellite amplification. Sequencing was carried out in 10-μL reactions using 2 μL of PCR product, 0.125 μL BigDye (ABI Prism BigDye Terminator v3.1 Cycle Sequencing Kit), 250 nmoles of both AsaC334F and AsaC334R, and 0.875 μL of BDX64 (Molecular Cloning Laboratories). We ethanol-precipitated sequencing reactions, rehydrated with 11 μL of formamide, and ran them on an ABI 3130XL Genetic Analyzer. Geneious software v 7.1 was used to analyze resulting sequence (Kearse et al. 2012).

After performing allele frequency analysis on the 2013 genotyped samples, we discarded from use in PBT any markers not in Hardy–Weinberg equilibrium or exhibiting null allele probabilities >0.05. We also discarded marker AsaC334 because it did not conform to expected allele sizes in several instances. An additional three markers had observed heterozygosity rates of <0.5. We discarded two of these; however, because microsatellite AaAG5 had a large number of possible secondary alleles (18), we felt sufficient discriminatory power remained to retain this marker in the data set. The newly established suite of 14 markers all conformed to HWE, had null allele frequency estimates of <0.024, and mean observed heterozygosity (H_obs) ranging from 0.482 to 0.914 (Table 2). The mean number of alleles per locus was 18.8, with a mean expected heterozygosity of 0.729, a mean observed heterozygosity of 0.722, and a mean polymorphic information content of 0.702. The combined non-exclusion probability for parent pairs was 6.138 × 10⁻¹⁰ and the combined identity non-exclusion probability was 8.030 × 10⁻¹⁶.

We used these 14 markers to perform PBT on samples collected in 2013. All 41 positive controls matched appropriately at 99% confidence levels to male and female parent pairs spawned in the same tank, and none of the 50 negative controls matched to broodfish. As a further test of this microsatellite panel, we genotyped another 290 broodfish from the year 2014 and 49 broodfish from 2015, along with 50 positive control fry collected exclusively in 2015 (Table S1, Supplemental Material). We tested the 50 fry against a collective data set containing both the 2014 and 2015 broodfish for PBT. The program CERVUS matched all 50 positive control fry to male and female parent pairs collected in 2015 at 99% confidence levels. We identified no parent pairs from the 2014 cohort, as expected.

One published marker not used for PBT, AsaC334 (Julian and Bartron 2007), did not show expected allele patterns in three 2013 broodfish suspected of being Alewife. These fish showed alleles not normally observed in Blueback Herring samples and were difficult to amplify at several loci. We sequenced the AsaC334 locus, a tetramer, in four known Blueback Herring samples from the 2014 Chowan broodfish, four Alewives collected in the Roanoke River and confirmed through peritoneal examination (Berlinsky et al. 2015; Hasselman et al. 2015), and one vouchered Chowan Alewife from the North Carolina Museum of Natural Sciences collections. Sequencing revealed a two-base-pair insertion in this marker that tracked exclusively with Blueback Herring when compared with Alewife (Figure 1). The absence of this insertion in Alewife can be viewed easily in GeneMapper as a peak that sits between our two expected bins for Blueback Herring, signifying an allele size that differs by two base pairs (Figure 2). Using GeneMapper, we confirmed this sequence variation in an additional 9 Chowan River Alewife samples and 15 Roanoke River Alewives. We compared these individuals against the Blueback Herring broodfish samples collected in 2013 as well as 10 vouchered Roanoke Blueback Herring samples from the North Carolina Museum of Natural Sciences collection. The species-specific insertion was confirmed in all cases analyzed except the three individuals suspected of being Alewife in the 2013 broodfish. These three fish lacked the two-base-pair insertion, consistent with our known Alewife samples. Additionally, our analysis found one juvenile Roanoke phenotypically identified as an Alewife that contained both possible alleles, with one allele falling in the expected bin and the other sitting in between the tetrameric bins (Figure 2), suggesting the presence of Blueback Herring and Alewife hybrids in the Roanoke River.

The ability to use AsaC334 as a discriminatory marker was further supported by confirmation of our findings using the nuclear marker recombination activating gene 2 (rag2; Berlinsky et al. 2015). The rag2 gene contains a species-specific single-nucleotide polymorphism whereby a cytosine creates a BamHI site specific to Alewives as compared with Blueback Herring. Samples (Blueback Herring n = 15, Alewife n = 16) used to examine AsaC334 were also subjected to rag2 sequencing. In all instances, the rag2 marker confirmed our species designation using AsaC334. Rag2 sequencing also confirmed the juvenile hybrid detected by AsaC334 (Figure 3).

Genetics can be a powerful tool in hatchery management and conservation. To this end, we have characterized microsatellite markers that may be used in these fields for Blueback Herring. A significant challenge in previous years for hatchery management of Blueback Herring has been the ability to readily distinguish between this species and the related alosine, Alewife. Herein we present a microsatellite marker capable of distinguishing these two species. Sequencing of the AsaC334 marker revealed a two-base-pair insertion that occurred exclusively in our Blueback Herring samples when compared with Alewife. The presence of this sequence variant in two separate river systems demonstrates the robustness of AsaC334 as a differentiating marker for these closely related species. This microsatellite marker can distinguish between Blueback Herring and Alewife without the need for additional sequencing or alternate assays, and may easily be inserted into panels used for PBT, allowing researchers to address issues of species and parentage with the same analysis.

Examination of our samples using this marker identified one fish with allelic patterns consistent with those expected from a hybrid. Identification of a hybrid within the Roanoke River could indicate one of several possibilities. Previous studies have indicated the presence of Blueback Herring and Alewife hybrids in the John H. Kerr Reservoir (Hasselman et al. 2014). Therefore, our results may provide the first evidence that such hybrids are escaping the John H. Kerr Reservoir and moving downstream to the lower Roanoke River. Alternatively, hybridization could be occurring naturally in the Roanoke River with Alewife and Blueback Herring in spawning condition at similar times. A third possibility is that hatcheries might unintentionally be stocking hybrids if broodfish were not identified correctly at the species level and subsequently spawned in a hatchery setting. Regardless of the cause, identification of AsaC334 as a microsatellite marker that can be analyzed simultaneously with other microsatellites used for PBT can help track and manage potential hybridization in the future.

We further present the development of a robust Blueback Herring microsatellite marker set suitable for parentage-based tagging. This 14-locus panel was 100% effective at identifying positive controls known to spawn in a hatchery with broodfish cohorts from both 2013 and 2015. Similarly, the panel was successful in eliminating wild-spawned juvenile as possible descendants of hatchery broodfish. Our findings demonstrate that fisheries managers may utilize this panel for parentage-based tagging to assess hatchery contribution of Blueback Herring fry stockings. Data collected from PBT analyses have the ability to tag a fish back to specific parents from known breeding years; therefore, these studies have the ability to yield information beyond hatchery contribution, including optimal stocking locations, migration rates between rivers, age range of fish returning to spawn, and overall genetic diversity of the population of interest. Moreover, the same microsatellite data set may be employed in other genetic analyses of interest to hatchery management such as population structure and effective population estimates. Overall, the wealth of data obtained through the use of these microsatellite markers can allow fishery biologists to not only track hatchery contribution for Blueback Herring, but to address issues such as the appropriateness of mixing stock from multiple rivers and ensuring the maintenance of genetic diversity within stocked rivers.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Fourteen microsatellite markers were selected for use in Blueback Herring Alosa aestivalis parentage-based tagging. Alleles for Blueback Herring samples collected from 2013 through 2015 are recorded based on fragment size after polymerase chain reaction amplification with marker-specific primers and measured using GeneMapper 4.0. The 14 marker names are listed on top, with A and B noting the two distinct alleles per marker. The first four numbers of the sample ID indicate year of collection (sample year − sample number). Red indicates broodfish, blue indicates negative controls, and green indicates positive controls. Positive controls came from broodfish spawned in the hatchery but collected from 3 different creeks; sc = Sarem Creek, bc = Bennett's Creek, and ic = Indian (Dillard's) Creek. Alleles that were unable to be discriminately called by two researchers were scored as 0.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S1 (173 KB XLSX).

Figure S1. Six novel Blueback Herring Alosa aestivalis microsatellite markers were identified in 2014 through NextGeneration sequencing for possible use in parentage-based tagging. Complete sequence for these six markers is shown here. Primer sequences are highlighted in yellow. The previously developed microsatellite marker AsaC334 (Julian and Bartron 2007) may be used for species discrimination between Blueback Herring and Alewife Alosa pseudoharengus. Additional analysis of our samples with the known discriminatory locus recombination activating gene 2 (rag2) confirmed our findings with AsaC334. Genomic sequence for both of these loci is shown. Primer sequence is not included because they have been trimmed for quality, but may be found in Materials and Methods.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S2 (16 KB DOCX).

Reference S1. Evans J. 2015. River herring production report, 2015. Raleigh, North Carolina: North Carolina Wildlife Resources Commission. Federal Aid in Sport Fish Restoration F-108 Report.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S3 (379 KB PDF).

Reference S2. Limburg K, Oliveira K, Wiedenmann J, O'Boyle B. 2012. River Herring benchmark stock assessment, Volume 1. Raleigh, North Carolina: Atlantic States Marine Fisheries Commission. Stock Assessment Report No. 12-02.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S4 (10.626 KB PDF).

Reference S3. [NCDMF] North Carolina Division of Marine Fisheries. 2000. North Carolina fishery management plan: Albemarle Sound Area River Herring. Morehead City, North Carolina: NCDMF.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S5 (548 KB PDF).

Reference S4. [NOAA] National Oceanic and Atmospheric Administration. 2009. Species of concern: River Herring. NOAA.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S6 (1229 KB PDF); also available at http://www.nmfs.noaa.gov/pr/pdfs/species/riverherring_detailed.pdf (1.23 MB PDF).

Reference S5. [NOAA] National Oceanic and Atmospheric Administration. 2016. River herring conservation plan. NOAA Fisheries, Greater Atlantic Region.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S7 (194 KB PDF); also available at https://www.greateratlantic.fisheries.noaa.gov/protected/riverherring/conserv/rh_plan_2015_executive_summary_060716.pdf (191 KB PDF).

Reference S6. Potoka KM. 2016. River Herring monitoring program in the Chowan River Basin, North Carolina—2015. Survey. Raleigh, North Carolina: North Carolina Wildlife Resources Commission. Survey.

Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S8 (318 KB PDF).

Reference S7. River Herring [TEWG] Technical Expert Working Group. 2016. Meeting summary. NOAA Fisheries Greater Atlantic Region. Found at DOI: http://dx.doi.org/10.3996/022017-JFWM-011.S9 (63 KB PDF); also available at https://www.greateratlantic.fisheries.noaa.gov/protected/riverherring/tewg/sept_2016_tewg_meeting_summary_final.pdf (64 KB PDF).

We gratefully acknowledge the work performed by the North Carolina State Genomic Sciences Laboratory (Raleigh, North Carolina) in conducting our NextGeneration sequencing. Thanks go to our collaborators at the Edenton National Fish Hatchery and the Watha State Fish Hatchery who conducted production, stocking, and fin clip collections for samples described in this paper. Funding for analysis of these samples came from the Federal Aid in Sport Fish Restoration program, project F-108. We thank the Associate editor and reviewers for their time and effort on this manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.