Cryptic diversity and speciation in an endemic copepod crustacean Harpacticella inopinata within Lake Baikal

Abstract

Ancient lakes are hotspots of species diversity, posing challenges and opportunities for exploration of the dynamics of endemic diversification. Lake Baikal in Siberia, the oldest lake in the world, hosts a particularly rich crustacean fauna, including the largest known species flock of harpacticoid copepods with some 70 species. Here, we focused on exploring the diversity and evolution within a single nominal species, Harpacticella inopinata Sars, 1908, using molecular markers (mitochondrial COI, nuclear ITS1 and 28S rRNA) and a set of qualitative and quantitative morphological traits. Five major mitochondrial lineages were recognized, with model-corrected COI distances of 0.20-0.37. A concordant pattern was seen in the nuclear data set, and qualitative morphological traits also distinguish a part of the lineages. All this suggests the presence of several hitherto unrecognized cryptic taxa within the baikalian H. inopinata, with long independent histories. The abundances, distributions and inferred demographic histories were different among taxa. Two taxa, H. inopinata CE and H. inopinata CW, were widespread on the eastern and western coasts, respectively, and were largely allopatric. Patterns in mitochondrial variation, that is, shallow star-like haplotype networks, suggest these taxa have spread through the lake relatively recently. Three other taxa, H. inopinata RE, RW and RW2, instead were rare and had more localized distributions on either coast, but showed deeper intraspecies genealogies, suggesting older regional presence. The rare taxa were often found in sympatry with the others and occasionally introgressed by mtDNA from the common ones. The mitochondrial divergence between and within the H. inopinata lineages is still unexpectedly deep, suggesting an unusually high molecular rate. The recognition of true systematic diversity in the evaluation and management of ecosystems is important in hotspots, as it is everywhere else, while the translation of the diversity into a formal taxonomy remains a challenge.

Keywords: Copepoda; ancient lakes; endemism; phylogeography; speciation.

FIGURE 1

A female Harpacticella inopinata from the Barguzin sampling site.

FIGURE 2

(a) A tree of relationships of unique mitochondrial COI haplotypes (ML tree with bootstrap values at the nodes), and results of various species delimitation analyses based on complete haplotype data of Harpacticella inopinata (each bar represents an inferred MOTU). Outgroup sequences are Tigriopus fulvus (MK211324), T. brevicornis (EF207720), T. californicus (AF096943), Harpacticella jejuensis (KM272559), H. sp (ZPC043‐13), Harpacticus flexus (MG817135) and H. uniremus (MH242788); (b) COI haplotype networks for each of the five main mitochondrial lineages. Crosslines denote the nucleotide differences between haplotypes. Circle size is proportional to the haplotype frequency in the data; (c) means of model‐corrected distances (below diagonal) and of p‐distances (above diagonal) (in %) between lineages from COI; (d) distribution of sampling sites (Table 1) and the observed distributions of the identified mitochondrial lineages in Lake Baikal, colour coded as in the tree. Arrows indicate surface currents.

FIGURE 3

(a) ML tree from 28S rRNA sequences of Harpacticella inopinata, with bootstrap support values; (b) an MP tree from a pruned and gap‐coded ITS1 alignment, built‐in PAUP using ACCTRAN optimization; tree scale indicates number of differences; (c) mean model‐corrected distances (below diagonal) and mean p‐distances (above diagonal) (in %) between lineages from 28S rRNA; (d) distributions of the ITS1 lineages in Lake Baikal, strong arrows point to samples where individuals with a discrepancy in their COI versus ITS1 identities were observed; arrows within the lake indicate surface currents.

FIGURE 4

Haplotype mismatch distributions (a, c, e, g, h) and Bayesian skyline plots (b, d, f) from the five main COI lineages of Harpacticella inopinata (N, number of sequences). Observed frequencies (coloured line) were compared with the expectations from a model of stable population size (grey dashed line) and the growth‐decline model (black dashed line) obtained by DNAsp. BSPs represent changes in effective population size over time. The solid line depicts the median estimate, and the shaded area represents the 95% posterior density interval. Tentative alternative time scales correspond to COI divergence rates of 2.6% Ma⁻¹, and a fivefold higher rate of 13% Ma⁻¹. Note different scales on horizontal axes in a–d versus e–h.

FIGURE 5

Morphology of Harpacticella inopinata, and definition of morphometric measurements. (a) Gross morphology: lateral view of an adult female. The length of the body is ca. 1 mm; (b) a single P5 leg from each of two CE lineage specimens from the Barguzin site (162, 168); (c) one caudal ramus from each of two CW lineage specimens from Kharantsy (5, 7). Arrows show the length and width measurements for morphometric analyses.

FIGURE 6

Inter‐lineage variation in qualitatively assessed morphological characters (a) Antenna II exopod; (b) P1 endopod, armature of the terminal segment. CW—specimen (7) from Kharantsy, Olkhon island, CE—Barguzin (162), RW2—Sakhyurta (9a), RW—Kotelnikovskii Mys (3kt), RE—Malyi Ushkanii island (6U). The examined segments are encircled.

FIGURE 7

Principal component analysis of morphometric measurements. Scores of (a) PC2 and (b) PC3 are plotted against PC1 on the horizontal axis. The three components account for 49%, 17% and 13% of the variation in the analysis. The polygons enclose observations of each of the mitochondrial lineages, with colours explained in the legend. Specimens from stone habitats are presented by dots and those from sand by open squares. The specimens that had a contradictory CW versus RW lineage identity in COI versus ITS1 gene are shown as brown stars.