The coincidence of ecological opportunity with hybridization explains rapid adaptive radiation in Lake Mweru cichlid fishes

Abstract

The process of adaptive radiation was classically hypothesized to require isolation of a lineage from its source (no gene flow) and from related species (no competition). Alternatively, hybridization between species may generate genetic variation that facilitates adaptive radiation. Here we study haplochromine cichlid assemblages in two African Great Lakes to test these hypotheses. Greater biotic isolation (fewer lineages) predicts fewer constraints by competition and hence more ecological opportunity in Lake Bangweulu, whereas opportunity for hybridization predicts increased genetic potential in Lake Mweru. In Lake Bangweulu, we find no evidence for hybridization but also no adaptive radiation. We show that the Bangweulu lineages also colonized Lake Mweru, where they hybridized with Congolese lineages and then underwent multiple adaptive radiations that are strikingly complementary in ecology and morphology. Our data suggest that the presence of several related lineages does not necessarily prevent adaptive radiation, although it constrains the trajectories of morphological diversification. It might instead facilitate adaptive radiation when hybridization generates genetic variation, without which radiation may start much later, progress more slowly or never occur.

Fig. 1

Geographic setting and cichlid diversity in Lakes Bangweulu and Mweru. Map of the Lake Mweru–Bangweulu region with colonizing lineages of haplochromine cichlids from the genera Serranochromis (Se.), Sargochromis (Sa.), Pseudocrenilabrus (P.) and Orthochromis (O.), and other cichlids. Lake Mweru has always been part of the Congo catchment (dark green in the inset map of Africa). During the Pleistocene, about 1 million years ago, the Luapula River captured the northeastern arm of the Zambezi catchment (light green), the Chambeshi–Bangweulu system. Lake Mweru was likely colonized by 11 species belonging to nine lineages of haplochromine cichlids. The fish photos illustrate the species that evolved from these lineages in radiations restricted to Lake Mweru (red frame) or with species in rivers flowing into Lake Mweru (orange frame). The number of species and overall ecology of the radiations are listed below. Lake Bangweulu was colonized by six haplochromine species belonging to four lineages, all of which are widespread in the Zambezi and none of them speciated in the lake. Mitochondrial haplotype clade numbers following Joyce et al. are given in parentheses. Both lakes were colonized by non-haplochromine cichlids too (four species in Lake Bangweulu, six in Lake Mweru) but none of these diversified in the lakes and they are not part of this study. The photos were taken by Ole Seehausen, except for the “New Kalungwishi” individuals which were photographed by Hans van Heusden (first photo) and by Numel Phiri and Cyprian Katongo (second photo), and the four photos of the O. kalungwishiensis complex which were also taken by Hans van Heusden. The species names are given in Supplementary Fig. 2.

Fig. 2

Morphological variation and morphospace partitioning between Lake Mweru radiations. a Morphological complementarity of the four intralacustrine cichlid radiations, the lacustrine members of the Orthochromis assemblage, the non-endemic Serranochromis altus and Se. angusticeps (Se. alt/ang), the nonendemic Se. robustus and Se. thumbergi (Se. rob/thumb) and Orthochromis stormsi (O. stormsi) in Lake Mweru. The first two principal components of all haplochromine cichlid species we found in Lake Mweru show nearly perfect complementarity in morphospace occupation among the radiations. Of the Orthochromis species, only the taxa occurring in Lake Mweru, O. polyacanthus and O. sp. “red-cheek” are shown, but not the seven riverine species of the Kalungwishi and Luongo Rivers. Sister taxa of each radiation are indicated with filled symbols. Their centroids or that of the radiations themselves (in the absence of sister taxa) are connected by thin lines to each phenotype of the corresponding radiation to visualize approximate trajectories of phenotypic divergence and diversification. The underlying data and the R script for all panels are provided at Zenodo with doi: 10.5281/zenodo.3435419. b Mweru taxa predicted onto morphospace of serranochromines of the radiation of ancient Lake Makgadikgadi. Makgadikgadi Sargochromis include Chetia and Pharyngochromis species which are nested in the genus Sargochromis. In the presence of a diverse Pseudocrenilabrus radiation in Lake Mweru, the serranochromine radiations altogether are confined to a subset of the morphospace this lineage occupies elsewhere (such as in the Okavango region, the modern centre of Makgadikgadi-derived diversity). This is mostly due to much reduced morphological diversity in Sargochromis of Lake Mweru. Compared to Pseudocrenilabrus of Lake Bangweulu, the Pseudocrenilabrus radiation in Lake Mweru expanded into serranochromine morphospace. c In the presence of the Pseudocrenilabrus radiation in the lake, Orthochromis are confined to the epilithic algae and Aufwuchs scraper niche in the rocky wave washed littoral of the lake, an extreme feeding niche that the Pseudocrenilabrus radiation has not invaded. In the rivers, where Pseudocrenilabrus have not radiated and only the ancestral type P. philander is present, two species with partial Orthochromis and partial Pseudocrenilabrus ancestry evolved in lentic, i.e. lake-like, parts of the river. These “New Kalungwishi” species overlap in morphospace with littoral Pseudocrenilabrus from Lake Mweru.

Fig. 3

Cytonuclear discordance and tests of hybridization reveal reticulate ancestry of all Mweru radiations. Nuclear cartoon topology of the “orthochromines” (a) and serranochromines (c) with green vertical lines showing mitochondrial sister relationships that deviate from the nuclear tree (solid lines) and sharing of mitochondrial haplotypes (dashed lines). Lake Mweru taxa are highlighted with red edges and radiations are shown as triangles, whereas Lake Bangweulu taxa are shown with blue edges. Riverine taxa in the drainage system of Lake Mweru are shown with orange edges. Black vertical lines indicate evidence for hybridization from D statistics (numbered as shown in b and d) and other tests of hybridization. Where data allow, arrow heads indicate the putative direction of gene flow. b and d D statistic results supporting hybridization edges shown in a and c for “orthochromines” and serranochromines, respectively. Error bars indicate three standard deviations from the mean and are depicted in grey if overlapping with 0 (non-significant, |z| < 3). D statistics and z-scores are averaged across multiple tests with different outgroups. Tests for edges 4–6 in b show averages of tests using Pseudocrenilabrus from Lake Mweru, Bangweulu or from the Zambezi/Cunene as P2 as they were all very similar. Likewise, tests where two groups are given for P2 or P3 revealed very similar results for both groups and have thus been averaged. All D statistics used to compute the averages are provided as Supplementary Data 2–4 and 7–8. R scripts and the data underlying the figures are provided at Zenodo with 10.5281/zenodo.3435419.

Fig. 4

Lake Bangweulu taxa show no signatures of hybridization. Comparisons of Bangweulu taxa (as P1) with their closest relatives (sister taxon in Lake Mweru as P2) reveals no evidence for excess allele sharing between any Bangweulu taxon with any other taxon (P3) in our dataset, which would lead to positive D statistics. The only exception is Se. robustus, which shows excess allele sharing with Sargochromis both of Lake Mweru and of Lake Bangweulu. Given that the two Sargochromis lake clades do not differ in allele sharing with Se. robustus of Bangweulu (see first test of ‘SaB vs SaM’ in this figure) and that slight excess allele sharing is also observed in the closely related Se. thumbergi of Lake Bangweulu, gene flow likely occurred from Se. robustus Bangweulu into the common ancestor of both Sargochromis lake clades. Therefore, the onxly signature of hybridization involving a Bangweulu taxon is shared by both Bangweulu and Mweru taxa and may reflect hybridization in the distant past. Hence, there is no additional event of hybridization in Lake Bangweulu, whereas there is rampant evidence for hybridization among multiple different Lake Mweru taxa (see Fig. 3). Abbreviations and colour scheme follow those in Fig. 3. Exact values are given in Supplementary Data 10. The R script and the input file used to make this figure are provided at Zenodo with 10.5281/zenodo.3435419.