Population genetic structure of the abyssal grenadier (Coryphaenoides armatus) around the mid-Atlantic ridge

Abstract

Understanding the factors that affect the levels and distribution of genetic diversity in oceanic and deep sea environments is a central focus in marine population genetics. Whilst it has been considered that the oceans represent a homogenous environment that would facilitate dispersal and minimise population structure, it is now clear that topographical features and current patterns can influence the extent of spatial gene flow and promote significant population genetic divergence even at local scales. Here we examine patterns of population genetic structure among N. Atlantic populations of the cosmopolitan abyssal grenadier Coryphaenoides armatus in relation to two hypothesised barriers to gene flow—the mid-Atlantic Ridge and the Charlie-Gibbs Fracture Zone/Sub-Polar Front. A suite of microsatellite markers were developed to examine the spatial pattern of allelic variation among 210 individuals from ten sampling locations encompassing sites east and west of the MAR and north and south of the CGFZ, plus a geographically distinct sample of individuals from the Crozet Islands in the Indian Ocean. Considerable genetic diversity was detected among individuals (n_a=5−13 and H_O=0.46−0.69 across populations) but with an overall lack of genetic divergence between populations. Pairwise estimates of divergence among NE Atlantic samples were small and non-significant (max F_ST=0.04) and Structure-based Bayesian analysis of genetic clusters returned no distinct population structure. The only indication of genetic structure was between the Atlantic and Indian Oceans, with F_ST estimates of ca. 0.05. Patterns of genetic diversity and divergence are discussed in relation to what has been resolved in Coryphaenoides congeners, and what is known about the life history and ecology of C. armatus.

Introduction

A long-held perception in marine population genetics was that intra-specific genetic differentiation should be minimal given the homogeneous nature of the marine environment (Pampoulie et al., 2004) and a concomitant high potential for gene flow (McMillen-Jackson et al., 1994). However, the burgeoning number of population genetics studies focused on marine species have overturned this view, resolving a variety of patterns of genetic structure underpinned by the interplay of life history, oceanography, seabed topology and variation in selection pressure over macro- and micro-geographic scales (Bowen and Avise, 1990, Banks et al., 2007, Papetti et al., 2012, Andrade and Solferini, 2007).

Studies on species from bathyal, abyssal and hadal zones are scarce relative to intertidal, coastal and near-coastal species, but the highly speciose nature of the deep sea environment implies that there are barriers to gene flow which partition genetic variation. What data are available provide a mixed picture. For example, blue hake (Antimora rostrata), orange roughy (Hoplostethus atlanticus) and the hydrothermal vent shrimp species (Rimicaris exoculata) show no genetic structure over large spatial scales (White et al., 2011b, White et al., 2009, Teixeria et al., 2012, respectively). Conversely, the deep-water demersal fish tusk (Brosme brosme) shows genetic heterogeneity indicative of spatial differentiation where bathymetric barriers may be limiting adult migration (Knutsen et al., 2009). Genetic structure has also been shown among populations of the bluemouth (Helicolenus dactlopterus) (Abiom et al., 2005) and the hydrothermal vent tubeworm, Ridgeia piscescie (Young et al., 2008). In many cases where genetic structure has been detected, the often limited nature of most deep-sea sampling compared with shallow water studies precludes an ability to tease apart the different stochastic and deterministic processes that may drive the observed patterns.

There is, however, a growing understanding of how major oceanographic processes and topographic factors could influence the patterns of dispersal and gene flow in abyssal species. The distribution and dispersal of faunal communities are significantly influenced by deep-water circulation (Bullough et al., 1998) and current systems which can become modified by seabed topography (Read et al., 2010, Søiland et al., 2008). Within the North Atlantic Ocean the most prominent topographical feature is the Mid-Atlantic Ridge (MAR) located between 40°N (Azores) and 63°N (Iceland), and extending from abyssal depths on the flanks to ca. 800 m depth on the crest. The MAR is interrupted by three main fracture zones between 48° and 54°N—the Faraday, Maxwell and Charlie-Gibbs Fracture Zones and this complex bathymetry has considerable influence on North Atlantic circulation (Søiland et al., 2008). The North Atlantic Current (NAC) extends eastward from the Gulf Stream until it encounters the MAR at 52°N (Read et al., 2010) where the bathymetry of the CGFZ splits the NAC into several branches. The most northerly branch is the Sub-Polar Front (SPF) (Read et al., 2010). Whilst the CGFZ itself is not considered a barrier to gene flow it causes noticeable changes in the water mass properties of the NAC (Read et al., 2010) which may in turn affect the dispersal potential of larvae (White et al., 2010).

One species that has been examined to test such hypotheses is the roundnose grenadier, Coryphaenoides rupestris. A population structure study across the MAR and Northeast Atlantic resolved both an isolation-by-distance pattern alongside structure around the CGFZ (White et al., 2010). It was suggested that the SPF is causing a boundary where larvae are distributed into two current systems (north and south of the front) which promote differentiation by genetic drift (White et al., 2010). It was also noted that the majority of the resolved divergence could be attributed to one locus that might be under selection due to genetic hitch-hiking. A second study encompassing samples from across the North Atlantic showed significant population structuring overall, with less prominent, but still significant, structure occurring across the MAR (Knutsen et al., 2012). It was proposed that dispersal may be mediated by a relatively small number of adult grenadiers or that larval dispersal is restricted to short-range movements due to the larvae not acting as passive particles.

Here we compare and contrast the patterns of population genetic structure in the roundnose grenadier with a sister species, the abyssal grenadier (Coryphaenoides armatus). Whilst C. armatus is superficially similar to C. rupestris, they have very different life histories and ecology. C. armatus inhabits the lower slopes and abyssal plains typically between 2000 m and 4700 m (Merrett and Haedrich, 1997), with observations and captures recorded between 282 m and 5180 m (King and Priede, 2008), whereas C. rupestris occur at depths between 700 m and 1800 m. Coryphaenoides rupestris have distinct spawning times in late autumn to mid-winter (Knutsen et al., 2012) whereas C. armatus are believed to be semelparous given their late maturation, a lack of observed egg-bearing females (Priede et al., 1994), relatively high fecundity (approximately 2,500,000 eggs per female; Stein and Pearcy, 1982) and a lack of seasonal energy expenditure associated with a reproductive event (Drazen, 2002).

Little information is available on early life stage development in C. armatus. It is thought that eggs are buoyant so they float upwards, with the larvae developing at shallower depths and juveniles then returning to the abyss (Marshall, 1973). This hypothesis is consistent with individuals switching from a visual to an olfactory-based scavenging strategy at approximately 450 mm total length (Wagner, 2003) and associated differences in diet. Smaller fish typically feed on copepods and amphipods whereas larger fish consume decapods, benthic prey, and carrion, such as squid and fish (McLellan, 1977, Mauchline and Gordon, 1984). In contrast however, a second school of thought suggests the outer egg ornamentation that is characteristic of Coryphaenoides can slow the ascent of the eggs, affecting their trajectory, and this has been used to argue against a pelagic stage (Robertson, 1981).

A lack of definitive information on key aspects of the life history of C. armatus and an incomplete picture of how the bathymetry and oceanography of the MAR affects dispersal offers little scope for predicting patterns of gene flow and spatial genetic structure in the region. Here we examine patterns of population genetic structure by developing a suite of microsatellite DNA markers to examine the spatial distribution of polymorphism among populations collected on three research cruises at a network of sampling sites along the MAR, both north and south of the CGFZ and east and west of the MAR from which barriers to gene flow can be inferred. Patterns are compared with an outgroup population from the Crozet Islands, Southern Indian Ocean to better understand how genetic divergence increases with geographic distance in C. armatus.