Genetic, Ecological and Morphological Distinctness of the Blue Mussels Mytilus trossulus Gould and M. edulis L. in the White Sea

Abstract

Two blue mussel lineages of Pliocene origin, Mytilus edulis (ME) and M. trossulus (MT), co-occur and hybridize in several regions on the shores of the North Atlantic. The two species were distinguished from each other by molecular methods in the 1980s, and a large amount of comparative data on them has been accumulated since that time. However, while ME and MT are now routinely distinguished by various genetic markers, they tend to be overlooked in ecological studies since morphological characters for taxonomic identification have been lacking, and no consistent habitat differences between lineages have been reported. Surveying a recently discovered area of ME and MT co-occurrence in the White Sea and employing a set of allozyme markers for identification, we address the issue whether ME and MT are true biological species with distinct ecological characteristics or just virtual genetic entities with no matching morphological and ecological identities. We find that: (1) in the White Sea, the occurrence of MT is largely concentrated in harbors, in line with observations from other subarctic regions of Europe; (2) mixed populations of ME and MT are always dominated by purebred individuals, animals classified as hybrids constituting only ca. 18%; (3) in terms of shell morphology, 80% of MT bear a distinct uninterrupted dark prismatic strip under the ligament while 97% of ME lack this character; (4) at sites of sympatry MT is more common on algal substrates while ME mostly lives directly on the bottom. This segregation by the substrate may contribute to maintaining reproductive isolation and decreasing competition between taxa. We conclude that while ME and MT are not fully reproductively isolated, they do represent clearly distinguishable biological, ecological and morphological entities in the White Sea. It remains to be documented whether the observed morphological and ecological differences are of a local character, or whether they have simply been overlooked in other contact zones.

Fig 1. Map of study area and sampling sites.

(A). Location of the Kandalaksha Bay of the White Sea, and general distribution of three mussel taxa in Europe: ME (Mytilus edulis, blue), MT (M. trossulus, red) and M. galloprovincialis (MG, yellow). Bi- or tri-colored circles indicate zones of sympatry (see references in the text). (B-D) Sampling locations in the White Sea: (B) Kandalaksha Bay. (C) Umba town area. (D) Top of Kandalaksha Bay. Pie diagrams depict estimates of the proportions of ME (blue sector) and MT (red sector) genomes in samples from the genetic dataset (GDS), obtained by STRUCTURE analysis of four-locus genotype data (PSS, see text for details). Data on paired local subsamples collected from the algal and the bottom substrates are shown above and below the algae pictogram, respectively. Black pins indicate sampling sites of the MDS (morphology only). Detailed sampling locality data are in S1 Table. The green lines in part D are isohalines of surface water (after [34]).

Fig 2. Mussel morphotypes and their distribution between different genotypes, and between algal and bottom substrates.

(A). Mussel shells of different morphotypes: T-morphotype with an unbroken prismatic dark strip under ligament (lower shell) and E-morphotypes with the dark strip broken (middle) or absent (upper shell). Measurements indicated: L–total shell length, l–distance from shell umbo to the posterior end of the ligament, a–distance from umbo to the anterior end of the dark strip. The index Z = a/l; values of Z corresponding to the different morphotypes are shown. (B). The kernel density function of Z-values within the three genotypic classes (all samples pooled, the genotypic classes defined on the basis of STRUCTURE analysis, see text for details). Yellow dots indicate the medians. (C). Mussels growing on different substrates: on bottom ground vs. fucoid thalli. (D). The mean frequencies of T-morphotype (Z = 0) ± standard error on the algae (horizontal axis) plotted against that on the bottom in samples from 17 sites of MDS. If frequencies were identical on both substrates, the dots would fall on the diagonal (black line).

Fig 3. Genetic composition of mussel samples from Chupa (Ch), Umba (Um) and Ryazhkov (Ry).

Subsamples collected from different substrates (A–algae, B–bottom) are treated separately. Left panel: Frequency distributions of individual T-scores (the sum of T-alleles across the 4 loci). Numbers of individuals are plotted on the abscissas, with T-scores as ordinates. Dark and light bars indicate T- and E-morphotypes, correspondingly. Green lines display the expected distributions under local random mating (dotted lines) and a mixture of parental genotypes without interbreeding (continuous lines). Central panel: Pie charts in the middle illustrate T-frequencies (red sector) vs. E-frequencies (blue sector). The estimated PSS values and the disequilibrium estimates R’ and F_IS are shown. Right panel: ISS distributions. Each symbol represents an individual, ranked along the horizontal axis by ISS. Dark symbols correspond to T-morphotypes, open symbols–to E-morphotypes. The shapes of the symbols reflect an individual T-score: circles– 0–1 (“ME”), triangular– 2–6 (“hybrid”), diamonds– 7–8 (“MT”). Horizontal lines reflect the thresholds chosen to delimit the genotype classes of ME (lower threshold) and MT (upper threshold) on the basis of the analysis of simulated samples (see text, S2 Fig and S3 Fig for details).

Fig 4. Distribution of genetic disequilibrium measures in samples of different genetic composition.

PSS, characterizing samples’ genetic composition, are on abscissas, genotypic disequilibrium measures are on ordinate axes: (A). Intra-locus heterozygote deficit F_IS. (B). Average inter-locus correlation R’. Green crosses correspond to empirical samples, orange crosses mark simulated mixed samples. Circles, squares and triangles mark Um, Ry and Ch samples, correspondingly (small symbols correspond to subsamples from different substrates, large symbols to pooled samples). The curves show the maximum disequilibrium in case of physical mixing without interbreeding. Expectations for equilibrium panmictic populations would be close to 0.

Fig 5. Taxonomic composition and morphological features of putative purebred and hybrid mussels in samples of different genetic composition.

PSS on abscissas are plotted against: (A). frequencies of ME (blue symbols), MT (red symbols) and hybrids (green symbols) in samples; (B). ratios between sample means of shell length L of MT and hybrids (red symbols) and ME and hybrids (blue symbols) in samples; (C). frequencies of T-morphotypes among ME (blue symbols), MT (red symbols) and hybrid (green symbols) genetic classes in samples. Polynomial (A) or linear (B, C) functions were fitted to the data. Groups of less than 4 genotypes are not included. Classification of genotypes to ancestry classes in all cases was ISS-based.