MBE Advance Access originally published online on July 25, 2007
Molecular Biology and Evolution 2007 24(9):2069-2080; doi:10.1093/molbev/msm138
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Gene Flow between Species of Lake Victoria Haplochromine Fishes
,1,2
* Biology Department and Center for Natural Sciences and Environmental Research (CENSER), College of Science, De La Salle University-Manila, Manila, Philippines
Department of Biosystems Science, The Graduate University for Advanced Studies (Sokendai), Hayama, Kanagawa, Japan
Department of Anatomy, Tsurumi University, Yokohama, Japan
Tübingen, Germany
|| Department of Biology, Pennsylvania State University
E-mail: satta{at}soken.ac.jp.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
The haplochromine cichlid fishes of Lake Victoria (LV), East Africa, are a textbook example of adaptive radiation—a rapid divergence of multiple morphologically distinguishable forms from a few founding lineages. The forms are generally believed to constitute a "flock" of several hundred reproductively isolated species in a dozen or so genera. This belief has, until now, not been subjected to a test, however. Here, we compare genetic variation at 11 loci in 10 haplochromine populations of 6 different species. Although the genetic diversity in the populations is quite high, using a variety of statistical tests, we find no evidence of genetic differentiation among the populations of LV haplochromines. On genetic distance trees, populations of the same species intermingle with those of different species. At the molecular level, the species are indistinguishable from one another. Genetic comparisons with closely related species in 2 crater lakes indicate that the species within LV continue exchanging genes. These observations have important implications for phylogenetic reconstruction. The approach used in this study is applicable to other instances of adaptive radiation.
Key Words: hybridization • gene flow • cichlid fishes • Lake Victoria • adaptive radiation • ancestral polymorphism
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
The Great East African Rift is a system of cracks and wrinkles in the earth's crust, which runs from the Gulf of Aden (Red Sea) in the north to Mozambique in the south (Schluter 1997
). It is a tectonically highly active region in which a large part of Africa might be breaking away from the continent. The single northern Ethiopian Rift splits into 2 branches in Kenya and Tanzania—the Western and Eastern Rifts—separated by a broad uplifted plateau. Some of the Rift's fissures and volcanic craters have become filled with water to give rise to lakes of different sizes and ages. In the Western Rift, the largest lakes are, from north to south, lakes Albert, Edward, George, Kivu, Tanganyika, and Malawi. On the plateau between the 2 branches is the world's third largest lake, Lake Victoria (LV). In the hills of the Western branch is a series of crater lakes, of which lakes Lutoto and Nshere are of relevance for the present study. The main focus of the study is, however, LV. Although the shallow depression now occupied by LV has been the site of a large water reservoir for more than half a million years, the ancient lake dried up 15,400 years ago and then filled with water again beginning some 800 years later (Johnson et al. 1996). Whether the desiccation of the lake was complete is still controversial (Johnson et al. 1996; Fryer 2001), but if any parts of it survived the desiccation period, they could not have represented more than small, shallow, scattered patches.
Lakes of the Rift System teem with fishes that have entered them from the intricate and changing East African river system. Of the various groups now found in the lakes, the most successful have been the fishes of the family Cichlidae in the order of perch-like fishes, Perciformes. Within the cichlid family, one group, the haplochromines, is renowned for undergoing recent spectacular adaptive radiations in the large lakes of the Western Rift and in LV. In the latter, the radiation has given rise to a large species "flock," an assemblage of morphologically and behaviorally differentiated but closely related forms derived from a single ancestral form (Greenwood 1981). The morphological differences between the forms are primarily in coloration and in tissues and organs associated with the feeding apparatus. Behavioral differences include modes of feeding and reproduction. Several hundreds of haplochromine species are thought to inhabit LV, of which some 200 have been described and named. All of them were originally assigned to a single genus, Haplochromis, but later some authors have reassigned them to several genera (Greenwood 1979, 1981; Seehausen 1996). At least some of the species interbreed in aquaria (Crapon de Caprona and Frittzsch 1984; Sato A, unpublished data), but the occurrence of natural hybrids in the lake has not been documented reliably. The species share genetic polymorphism widely, both at mitochondrial (mt) and nuclear (nuc) DNA loci (Nagl et al. 1998, 2000; Seehausen et al. 2002; Terai et al. 2004). The sharing could be explained either by the retention of ancestral polymorphism or by interspecies hybridization.
The aims of the present study are to, first, assess the extent of genetic variation in selected LV haplocchromine species at the mtDNA and nucDNA loci, and second, attempt to distinguish between the contributions of the 2 potential sources of the shared polymorphism and so determine whether and to what extent the species exchange genes.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
Fishes
H.T. collected all specimens during 1993–1998 expeditions to the lakes of the East African Rift System. Specimens are still in his collection; some have, however, been distributed to other colleagues for further investigation. Pieces of fins from freshly caught fishes were fixed in 70% ethanol, changed after 1 day, and stored. DNA was isolated from them in the laboratory in Tübingen, Germany.
Molecular Methods and DNA Data Analysis
Methods of genomic DNA isolation, polymerase chain reaction (PCR) amplification, sequencing, and cloning were described in our earlier publications (Nagl et al. 2000, 2001; Terai et al. 2004). Primers used are listed in table S1 (see Supplementary Material online). Sequences were aligned by the ClustalW program and revised manually. Nucleotide sites containing gaps were excluded from the analysis. Population parameters such as the number of segregating sites, nucleotide diversity and divergence, and the number of shared haplotypes were determined using our own computer programs (available on request from Y.S.). Measurements such as genetic distances (Nei and Kumar 2000), as well as (net) nucleotide divergences were computed from pairwise nucleotide differences by using the Mathematica program. Genetic distances based on nucleotide differences were corrected by Kimura's 2-parameter method. Phylogenetic trees were drawn by the Neighbor-Joining method (Saitou and Nei 1987).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
Background Information
We collected 218 haplochromine specimens from LV and 2 crater lakes, Lutoto and Nshere, in the Western Branch of the East African Great Rift System. The LV specimens represented 4 species: Haplochromis (Ptyochromis) sauvagei (Hasa), Haplochromis (Ptyochromis) xenognathus (Haxe), Haplochromis (Paralabidochromis) chilotes (Hach), and Haplochromis sp. "rock kribensis" (Haro). Of Hasa and Haxe, we sampled 3 populations: Nyamanga (N), Yetti (Y), and Anyanga (A) located up to 350 km apart. In addition to morphological differences, the 4 species differed also in their ecologies (Greenwood 1981
; Johnson et al. 1996). Haro was typed previously (Nagl et al. 2000) as belonging to a different mt lineage (subgroup VD) than the other 3 species (subgroup VC) and, hence, presumably derived from a separate founder population. The 2 species from the crater lakes have not been formally described and named; we refer to them by the names of the lakes: Haplochromis sp. "lutotu" (Halu) and Haplochromis sp. "nshere" (Hans). They, too, belong to different mtDNA lineages—subgroups VE and VF, respectively, which are closely related to subgroup VB found in lakes Edward and George in the Western Rift (Sato et al. 2003). As their subgroup designations indicate, they are closely related to the VC and VD lineages found in LV (Nagl et al. 2000). In addition to morphological differences, the 4 LV species differ also in their ecologies. Hasa and Haxe are oral shelling and crushing molluscivores, Hach an insectivore, and Haro an Aufwuchs eater feeding on epilithic cyanobacteria, algae, and invertebrates. We tested the genetic (DNA) diversity of the 10 populations in 1 of 2 ways. From 6 of the 11 loci tested, we obtained nucleotide sequences of DNA segments (loci). They were: mtDNA control region (CR) and the nucDNA loci Melanocortin 1 receptor, Mc1r; Hagoromo, Hag; Tyrosinase, Tyr (all 3 involved in the control of fish pigmentation); Opsin 1 long wavelength light-sensitive, Opn1lws; and Short interspersed element 1357, SINE1357 (a transposable element). The remaining 5 loci (SINEs S1918, S1807, S1801, and S1909 as well as an indel in the Dopachrome tautomerase, Dct locus) were tested for the presence or absence of an insert by PCR amplification.
Genetic Diversities within Populations
In the first step of the analysis, we assessed the genetic diversity within each of the 10 populations using various measures (Nei and Kumar 2000) (table 1). Among these, most informative was the nucleotide diversity (
). The values varied from population to population and from locus to locus, but on the whole the observed differences did not show any significant trends. The ranges of the values (expressed in percentages) for the individual nuc loci were as follows: Hag, 0.13–0.20; Opn1lws, 0.11–0.65; Mc1r, 0.02–0.28; Tyr, 0.09–0.46; SINE1357, 0–0.56. The average values obtained from the 5 nuc loci for the individual species were as follows: Hasa, 0.30; Haxe, 0.30; Hach, 0.27; Haro, 0.29; Hans, 0.17; Halu, 0.12. They are comparable to values reported for other fish species: Chinook salmon, 0.19 (6 loci; Ford 1998), and zebrafish, 0.54 (calculated from SNPs in 66.8 kb of noncoding sequence; Guryev et al. 2006). They are also comparable with values reported for mammalian species (except human, whose populations give an order of magnitude lower values than, e.g., the chimpanzee population; Li and Sadler 1991; Kaessmann et al. 1999; Satta 2001). Hence, the nucleotide diversity at nuc loci of the endemic LV haplochromines does not give any indication of pauperization attributable to a recent bottleneck in the demographic histories of the populations. This conclusion is in line with the results of more limited earlier studies on the extent of genetic polymorphism in these species (Nagl et al. 1998; Seehausen et al. 2002). Somewhat unexpectedly, the nucleotide diversity of the mtDNA CR in all the haplochromine species and populations tested was found to be of the same order of magnitude as that of the nuc genes (table 2). The values for the individual species were: Hasa, 0.39; Haxe, 0.34; Hach, 0.42; Haro, 0.77; Hans, 0.40; Halu, 0.17. Because, generally, the mutation rate of mtDNA is higher than that of the nucDNA (Nei and Kumar 2000), one might have expected the values to be higher as well. Indeed, some of the reported values of several other fish species are an order of magnitude higher than the values of the haplochromine fishes: Japanese sea bass (Liu et al. 2006), 3.0; spotted sea bass (Liu et al. 2006), 1.29; Atlantic mackerel (Nesbo et al. 2000), 2.9; swordfish (Alvarado Bremer et al. 2005), 2.5. We return to this result later in the text.
|
|
Genetic Variability between Populations
The second and main goal of the study was to assess the genetic variability between the studied populations. To this end, we used the following measures of interpopulation divergence (tables 2 and 3 and fig. 1): nucleotide divergence (dxY), net nucleotide divergence (d), proportion of shared haplotypes (PH), proportion of shared polymorphic sites (PS), coefficient of gene differentiation (NST), and 3 different estimates of genetic distance (D) (Nei and Kumar 2000
). We applied the tests for these measurements to all 45 pairs into which the 10 populations could be arranged. Although the presence of frequent intragenic recombination prevented to make phylogenetic analysis of nucDNA, the various tests indicated a division of the pairs into 3 categories (I, II, and III), distinguished from one another by different degrees of divergence (fig. 1). We discuss the 3 categories in their turn.
|
|
Category I: Lake Lutoto Fishes
Category I contains those pairs, of which one member is the Halu population. The grouping separates the Halu population from the other 9 populations, and this separation is corroborated by other observations. First, on a phylogenetic tree of mtDNA CR sequences, the Halu sequences form a monophyletic cluster, which is set apart from the LV and the Hans sequences (fig. 2). Second, of the 4 SINE markers, one (SINE1801) is fixed in Halu, but polymorphic in all other populations; another (SINE1807) is fixed in Halu and Hans, but polymorphic in the LV populations; the third (SINE1909) and the Dct deletion are absent in Halu and Hans, but retained in all the LV populations; whereas the fourth (SINE1918) is polymorphic in all 10 populations (table 4). Third, no mtDNA CR haplotypes are shared between Halu (or Hans) and the other populations (table S2, see Supplementary Material online). Fourth, although some haplotypes at the nuc loci are shared between Halu and the other populations, the average proportion of this sharing is significantly lower than among these other populations (table S3, see Supplementary Material online). Fifth, at the mtDNA CR and 2 of the 5 nuc loci (Mc1r and SINE1357), there is virtually no sharing of polymorphism at individual sites between Halu and the other populations. At the Hag, Tyr, and Opn1lws loci, sharing does occur at low frequencies (tables S4 and S5, see Supplementary Material online). Sixth, at both the mtDNA and nucDNA loci, the nucleotide divergence between the Halu and the other populations is significantly larger (P < 0.01) than the nucleotide diversity within the Halu population (table 2). It is this high divergence of Halu from the other populations that is responsible for the clustering of Halu population pairs in category I. Seventh, on all genetic distance trees based on the 5 nuc loci combined, the Halu population assumes an outgroup position to the remaining 9 populations (fig. 3). Taken together, all these observations indicate that the Halu population has not been exchanging genes with the other 9 populations for some time. This time has been long enough to generate Halu-specific polymorphism at the mtDNA CR region. It has, however, not been long enough to erase entirely the contribution of ancestral polymorphism to the extant polymorphism at the nuc loci because 6.4% of haplotypes at these loci are still shared between Halu and the other populations (table S3, see Supplementary Material online). The Halu population is thus physically and genetically close to but isolated from the LV populations and so serves as an outgroup for the interpretation of the genetic variability in the LV flocks. Apparently, in a not too distant past, the Halu, the Hans, and the LV populations originated from a common ancestral population in which some of the nuc gene polymorphisms now found in the descendant populations had existed already.
|
|
|
mtDNA Versus nucDNA.
The foregoing observations indicate that the amount of nucleotide divergence dXY of any of the LV populations from Halu consists of 2 components: the polymorphism inherited from the ancestral population and the substitutions accumulated since the divergence of the 2 populations in a pair (fig. S1, see Supplementary Material online). This compound nature of dXY can be expressed as dXY = 2µ(Ts + tg) where Ts is the species divergence time, t is the average coalescence time in the ancestral population in generations, g is the generation time in years, and µ is the nucleotide substitution rate per site per year. The values of µ and t are different between mtDNA and nucDNA, reflecting the differences in mutation rate and mode of inheritance (fig. S1, see Supplementary Material online). To disentangle the 2 components and so determine how fast the mtDNA CR evolves relative to the nucDNA, it is necessary to subtract from dXY the amount due to the ancestral polymorphism. Under the simplifying assumptions that the µ of the ancestral population was similar in extent to that of the descendant extant populations and that the t value in the ancestral and descendant populations was also similar, the amounts of the ancestral polymorphism that need to be subtracted from dxY of the population pairs are
= 2µtg. The nucleotide divergence then simplifies to the net nucleotide divergence d = 2µTs. On the plot of the computed d values of mtDNA versus nucDNA, the regression line drawn through the origin is given by dmt = 4dnuc approximately, where and subsequently the subscripts mt and nuc stand for mtDNA and nucDNA, respectively (fig. 1). Because dmt = 2µmtTs and dnuc = 2µnucTs, the regression line indicates that the mtDNA of haplochromine fishes evolves about 4 times faster than the nucDNA. This being the case, one might expect the of mtDNA and nucDNA to have similar values because the 4 times higher substitution rate of the mtDNA compensates for the 4 times larger effective population size (Ne) of the nucDNA, so that mt/nuc = Neµmtg/4Neµnucg = 1. (We assume that the effective population size of mtDNA is given by the number of females [Nf] and that the sex ratio is 1:1, so that 2Nf = Ne). Indeed, the mt and nuc values of the Halu population are 0.17 ± 0.14% and 0.12 ± 0.06%, respectively (table 2 for other populations). The faster substitution rate and smaller effective size of mtDNA compared with nucDNA explains why in the mtDNA CR we found no evidence in the Halu and LV populations of polymorphism derived from the common ancestral population. The mtDNA CR polymorphism in the Halu and LV populations is separately monophyletic (in the sense of fig. 2) in contrast to the polyphyletic polymorphism at nucDNA loci. For this situation to arise, the divergence time (Ts in years) between the Halu and LV populations must meet the following rough condition: Neg < Ts < 4Neg. The left- and right-hand sides of the inequality define the mtDNA monophyly and the nucDNA polyphyly, respectively. Rewriting the inequality to Ts/4 < Neg < Ts indicates that the Neg must be in a rather narrow range. By using the 1–2 Myr divergence of the LV from the Lake Malawi haplochromine fishes (Schluter 1997) (node L0 in fig. 2) as a calibration point for the mtDNA substitution rate, the actual divergence time of Halu from the LV populations can be estimated from the dmt value of these populations. The dmt value between the LV and Lake Malawi fishes becomes 6.8 ± 1.0%. For simplicity, if we assume that their divergence time is 1.5 Myr, the nucleotide substitution rate is estimated as 2.3 x 10–8 per site per year, the divergence time of the Halu population from the other populations as 165,000 years with the average dmt value being 0.75%, and the Ne of these populations as 41,000–165,000 with g = 1.
Category II: Lake Nshere Fishes and Haro
Category II contains population pairs in which one member is either the Hans or the Haro population (fig. 1), the former in the Nshere crater lake and the latter in LV. The crater lake has apparently been isolated from LV for some time and this isolation explains why all the Hans mtDNA CR sequences, like all the Halu sequences, form a monophyletic cluster on the phylogenetic tree (fig. 2). It also explains why the SINE1807 marker is fixed and the SINE1901 marker is absent in the Hans population (table 4), why no mtDNA CR haplotypes are shared between Hans and the other 9 populations (table S2, see Supplementary Material online), and why the proportion of shared nucDNA haplotypes between Hans and the LV populations is lower than that between the populations within the LV (table S3, see Supplementary Material online). All these observations indicate that gene flow between the Hans and the LV populations has ceased, but apparently more recently than between the Halu and LV populations. The average dmt value of Hans from the LV species is 0.37%. It is one-half of that of Halu, so that the Hans population became isolated 
83,000 years ago.
The grouping of the Hans with the Haro pairs into the category II suggests that mtDNA CR of Haro diverged from other LV populations as early as that of Hans. This conclusion is also supported by the population tree analysis based on the various distance measures for nucDNA (fig. 3). With one exception, in all of them Haro forms an outgroup to the non-Haro LV clade; in the one exception (the DST tree), the Haxe (A) population takes over the place of Haro. The Haro population is well separated from Halu, somewhat less well from Hans and poorly from the other LV populations. This last separation is indicated by the following observations. First, the Haro fishes carry a single nucleotide substitution in their mtDNA CR, which distinguishes them from fishes of other LV populations examined (Nagl et al. 2000). Second, in pairwise interpopulation comparisons, the Haro-containing pairs differ from all the other LV pairs by either the absence or the lower proportion of shared haplotypes and sites of shared nucleotide polymorphism (tables S2–S5, see Supplementary Material online). These are the differences that are responsible for Haro's placing in an outgroup position relative to the other LV populations on the population trees (fig. 3). Third, the dXY values of the Haro-containing population pairs are similar to those of the Hans-containing pairs (table 2). This similarity is the reason why the Haro- and Hans-containing population pairs group together in category II in figure 1, separately from both category I (Halu-containing pairs) and category III (pairs involving non-Haro LV populations). The separation indicates an isolation of both the Haro and Hans pairs from the non-Haro LV pairs. But while the isolation of the Hans population is complete (as indicated, e.g., by no sharing of mtDNA CR haplotypes), the Haro population shares at least 2 identical mtDNA CR haplotypes with Haxe-A (table S2, see Supplementary Material online). Fourth, for mtDNA CR, the dXY values of the pairs of Haro with other LV populations are significantly higher (in the range from 0.69% to 0.81%) than those of the LV pairs not involving Haro (range from 0.34% to 0.50%). In fact, the former values are as high as those between Hans and the LV populations (range from 0.61% to 0.84%; table 2). These large dXY values of Haro-involving pairs are the consequence of the presence of a distinct Haro-specific haplogroup (present in Haro6157, 6162, 6747, 7042, and 7043 in fig. 2); they suggest that the initial divergence of Haro took place as early as that of Hans, some 83,000 years ago. At the same time, however, the mitochondrial mt value of the Haro population (0.77%) is as high as the mitochondrial dXY values of the pairs of Haro with the other LV populations. If this high value were the consequence of persisting ancient polymorphism, one would expect the nuc to be equal to the mt, but this is not the case. The Haro nuc is significantly (P < 0.05) lower than the Haro mt and this observation favors the interpretation that the high Haro mt value reflects gene flow between Haro and the other LV populations. This conclusion is consistent with the sharing of mtDNA CR haplotypes between Haro and Haxe-A. It does not necessarily imply that gene flow occurred only at the level of mtDNA; nucDNA must have been exchanged as well, but its flow has been overshadowed by the slower substitution rate and the larger extent of unsorted ancestral polymorphism. The genetic variation observed in the Haro population is thus presumably the result of 2 opposing processes, recent gene flow from other LV populations on the one hand and a past isolation of Haro from these populations on the other hand. Gene flow brings into the Haro population distantly related sequence lineages from other LV populations and makes these populations genetically closer to each other than to Hans, whereas the past isolation had worked toward retention of the Haro-specific sequence lineages. As a result, Haro is more closely related to LV populations than it is to Hans (fig. 3).
Category III: Other LV Fishes
Category III contains 7 LV populations: 3 species (Hasa, Haxe, and Hach), of which the first 2 are each represented by 3 populations (A, N, and Y) sampled from different localities. There are few species- or population-specific haplotypes in the data sets (table 5 and fig. S2, see Supplementary Material online). There are no single fixed nucleotide substitutions at either the mtDNA or the nucDNA loci that could serve as identifying markers of the populations (fig. S2, see Supplementary Material online). There are also no significant differences in gene frequencies among these populations, as is apparent from the similarities of the interpopulation genetic distances, regardless whether the comparison is made within or between species (tables 2 and 3), and from the intermingling of populations from different species on the population trees (fig. 3). And on the plot of nucleotide divergences between populations, there is no indication of a greater disparity between species than between populations of the same species. The indistinguishability of the LV non-Haro populations at the DNA level could mean either that the populations diverged from one another very recently and so did not have enough time to sort out their ancient polymorphism and develop species-specific differences or that they are continually homogenized by gene flow. Below we argue that the latter is the more likely explanation of the 2 possibilities.
|
Ancient Polymorphism Versus Gene Flow.
There are genetic differences between the LV populations, but they are small: the average dmt and dnuc values for all the population pairs are 0.06% and 0.015%, respectively. They translate into population divergence of approximately 14,000 years, which is very close to the time of refilling of the LV. This observation supports the notion that the speciation occurred within LV. In an attempt to interpret these findings, we proceed from facts to their interpretation and from it to speculation. The 2 principal facts are: first, the LV haplochromines are phenotypically distinguishable into groups, which have traditionally been referred to as "species." And second, at the genotypic level, the present study fails to provide any evidence for species-specific genetic changes. The 2 implications from these facts are: first, there must be a few genes that are responsible for the phenotypic differences between the species and that evolve under selection leading to differential fixation of existing alleles in the species. Heritability of the phenotypic differences is indicated by their persistence through generation of fishes maintained in aquaria. A low number of genes encoding these differences are suggested by segregation data obtained from interspecies crosses in the laboratory (Sato A, unpublished data). If so, the effect of selective sweep should be limited to small genomic regions and genetic variation at most loci should not be diminished. The second implication is that the time required for the fixation of alleles at the loci under selection should have been long enough to begin to differentiate the species in allelic frequencies at other loci, evolving under neutrality. If Hach, Haxe, and Hasa speciated from each other about 14,000 years ago, the fact that the net nucleotide divergences between these species are similar to those between populations within species (fig. 1) strongly implies that gene flow continues to homogenize genetic variation both within and between the species. The sharing of mtDNA haplotypes between the Haro and Haxe-A populations has provided a direct support for this contention. And if Haro exchanges genes with Haxe, gene flow between Haxe and Hasa or Hach as well as between Hasa and Hach is so much more likely. Indeed, there are nucleotides that are specific to and shared by LV species (table 5). Like the shared mtDNA haplotypes, the shared variable nucleotides and LV-specific shared haplotypes at nuc loci corroborate the presence of gene flow between species.
We speculate that included in the ancestral polymorphism might have been at least some of the alleles, which later became responsible for the development of the phenotypic adaptations characterizing the LV species. The actual adaptations may not have developed in the ancestral population because the environmental conditions that later drove their evolution in LV may not have existed on other places. Alternatively, the adaptations may have developed at these other places but are now interpreted as instances of convergent evolution (Greenwood 1981). In LV, the evolution of the adaptations is driven by interplay of natural selection with sexual selection (Seehausen and van Alphen 1999). The former is largely responsible for the adaptation of groups of fishes to the various ecological niches, the latter for the development of sexual dimorphism. Both types of adaptation lead to a partial isolation of the groups, but the incompleteness of the isolation allows interbreeding between the groups and thus prevents their differentiation in terms of genetic variability at the bulk of loci. This process may had been going on in the ancestral population and simply continued after colonization of LV 14,600 years ago. Should some of these species survive and become reproductively isolated, their phylogenetic relationship (see fig. 4 for a schematic representation) will not be resolvable, no matter how many genes one uses in the analysis. The same conclusion may apply to many other cases of adaptive radiation.
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
Transspecies Polymorphism Versus Gene Flow
The main finding of the present study is that the morphologically and behaviorally differentiated species of the LV haplochromine flock are indistinguishable at the molecular level. Strictly speaking, this conclusion applies only to the 4 species and the 11 markers (genes) tested, but there is reason to believe that it can be extended to include most of the species in the flock and large sections of their genomes. The reason is that Greenwood (1979)
classified 3 of the 4 species as representing 2 different genera, whereas the fourth species (Haro) represents a group that arose from a different founding population than the rest of the species (Nagl et al. 2000 and this study). Similarly, the 11 markers represent different classes of loci, so that both the species and the loci can be considered representative of the whole flock and the genome, respectively. The main inference we draw from the finding is that there is a gene flow between the species as extensive as that between local populations of the individual species. Is this result surprising? Undoubtedly there will be some, who will say: "I always knew it," but knowing something without evidence is not a satisfactory situation. The only hint that there might have been an interspecies gene flow between the LV haplochromines was the observation of transspecies polymorphism in the populations (Nagl et al. 1998, 2000; Seehausen et al. 2002; Terai et al. 2004). However, shared polymorphism can occur also between species that do not exchange genes (Klein et al. 2007), so that this was not by any means a strong hint.
Hybridization and Speciation
Natural hybridization leading to interspecies gene flow is commonly ascertained either by direct observation of tagged individuals or by typing using molecular markers that distinguish the parental species. Both methods have been used to document interspecies hybridization in cichlid fishes of lakes Malawi and Tanganyika (Salzburger et al. 2002), as well as in studies of many other animal groups (Seehausen 2004; Mallet 2007). Neither of the 2 methods is easily applicable to the LV haplochromine flock. Direct observations in LV are difficult because of the murkiness of its water and of other factors; moreover, tagging of individuals is unsuitable for this kind of large-scale study of populations. And the prerequisite of the molecular method is the existence of molecular markers distinguishing the species. Such markers are not available in the LV haplochromines. Because of this limitation, the bulk of the recent studies on natural interspecies hybridizations have been carried out on reproductively isolated and genetically well-differentiated species, in which the isolating mechanisms, for one reason or another, break down occasionally, or in other words, on species in which the speciation has either been completed or has reached an advanced stage (Mallet 2007). The method is not applicable to early stages of speciation such as those found in the LV haplochromines or the ground finches of the Galapagos Islands. Indeed, to the best of our knowledge, natural hybridization in the LV haplochromines has not been documented prior to the present study. In the ground finches, attempts are being made to solve the marker problem by the use of microsatellites (Grant et al. 2005), which, however, are markers of individuality, like tags, and not of species. Their suitability for population studies is controversial and their genealogy obscure. The present study is an attempt to study speciation at the population level by using population genetics methods. The study spotlights the fundamental issue in the investigation of early speciation phases—the problem of distinguishing gene flow from shared ancestral polymorphism. The approach we have chosen does allow us to make this distinction and to come to the conclusion that the 4 tested species have been exchanging genes from their inception at a rate sufficient to homogenize the species genetically. The homogenization affects large parts of the species' genomes, presumably with the exception of the few genes controlling the phenotypic interspecies differences. The challenge is now to develop a model that would account for the maintenance of these differences in face of the continuing interspecies gene flow.
Implications for Phylogenetic Reconstruction
From the viewpoint of our studies, the total of the genes of the LV haplochromine flock can be imagined as a single pool differentiated into subpools only by the putative genes responsible for the phenotypic differences between the species. If the whole flock does not become extinct, at some time in the future, most of the lineages now represented by the different species will terminate and only a few species will continue to persist. Because of their compounded and genetically promiscuous history, their phylogeny will not be resolvable, no matter how many genes future biologists will use in their attempts at phylogenetic reconstruction of this adaptive radiation. Similar situation is, in fact, encountered today in attempts at reconstruction of phylogenies of species that have survived past adaptive radiations. As for the present-day LV haplochromine flock, one must give up hope that its phylogeny can be resolved. Although some might think that once the genes responsible for the phenotypic differences between the species are identified, their sequences will reveal the true phylogeny of the group, we think that this is unlikely to be the case. A phylogenetic tree based on any one of these putative genes will be a tree of that gene and not of the species, and the phylogenies of the different genes will probably not be congruent with one another. A tree based on the combined information extracted from all these genes will be an artifact, an illusion, rather than a depiction of the flock's history.
Implications for the Species Concept
In this communication, we have avoided terms and concepts, such as introgressive hybridization, transgressive segregation, or syngameon, currently fashionable in evolutionary biology. For this infraction, some will undoubtedly accuse us of being "divorced from modern concepts in evolutionary biology," but our intention has been to spotlight the actual data and the message they relay, without dressing them in confounding terminology. Besides, the situation we are dealing with in this study does not correspond exactly to the one for which this terminology has been introduced. One confounding term (Hey 2006), however, we could not have avoided using—the term "species." What Greenwood (1981), who described many of the haplochromine forms as species, meant by the term was probably a group of organisms delineated by morphological (and ecological in some cases) characters alone—what might be referred to as the morphological species concept (Lincoln et al. 1982). It appears, however, that Greenwood assumed a degree of reproductive isolation for these species and so his usage might also be regarded as falling under the biological species concept (Mayr 1942). What contemporary writers mean by "haplochromine species" is difficult to say, but presumably they do not disagree with Greenwood's concept. The one exception on record is the 1975 paper by Sage and Selander, in which the authors found the described morphologically distinguishable Mexican cichlid species of the genus Cichlasoma indistinguishable at the molecular (allozyme) level and suggested that the species should therefore be reclassified as "morphs." The latter term is almost as nebulous as the former, with one of its meanings being "any local population of a polymorphic species exhibiting distinctive morphology or behavior" (Lincoln et al. 1982). Sage and Selander (1975) suggested also that the LV haplochromines might also qualify for a similar reclassification. Note, however, that by this definition, "morphs" are nearly the same as "species" of the morphological species concept. It is in this sense that we use the latter term in the present communication. In his later years Greenwood (1979) went a step farther and revised his single genus (Haplochromis) classification of the LV haplochromine species flock, splitting it into 12 genera. Cichlid experts now commonly use his revised classification. When we published our first cichlid paper (Klein et al. 1993), the reviewers and the editors bullied us into adopting the "splitters" nomenclature. But in view of the data presented in this study, we find no justification for the continuation of the usage and so we adopt the single genus classification.
Implications for Conservation Efforts
The LV haplochromine flock appears to be declining in its diversity, and the downslide is attributed to the eutrophication of the lake as a result of human activities and to the introduction of a voracious predator, the Nile perch (Seehausen et al. 1997). These 2 factors are undoubtedly mainly responsible for the disappearance of the species. A contributing factor might, however, also be interspecies hybridization. Seehausen et al. (1997) have suggested that the eutrophication in the last few decades has disrupted mate selection and led to hybridization. The widespread genetic homogeneity of the flock indicates, however, that the interspecies gene flow must have started long before humans began changing the lake's ecology. Moreover, hybridization may not only lead to disappearance of species in some situations but may also generate new species in others (Seehausen 2004). We emphasize, however, that none of the findings and deductions of the present study undermine the ongoing conservation efforts in LV. The elucidation of the evolutionary processes underway in LV haplochromines is of such importance to science and humanity that indeed no effort should be spared to prevent their disruption.
|
Accession Numbers |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
Hag01–Hag24: AB326187–AB326210, S01–S36: AB325542–AB325577, mt01–mt83: AB325718–AB325800, LWS01–LWS75: AB326112–AB326186, Tyr01–Tyr41: AB326243–AB326283, MC01–MC22: AB326311–AB326332.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
Tables S1–S5 and figures S1 and S2 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
The experimental part of this research was carried at the now defunct Max-Planck Institut für Biologie, Abteilung Immungenetik, Tübingen, Germany. We thank Drs Colm O'hUigin and Masatoshi Nei for their comments on the manuscript, Drs F. Witte and O. Seehausen for introducing one of us (H.T.) to the identification of Haplochromis species, and the Uganda National Council for Science and Technology, the Tanzania Commission for Science and Technology, Ministry of Tourism Natural Resources and Environment, Tanzania, the Kenya Ministry of Tourism and Wildlife, Nairobi, and Kenya Fisheries Research Institute, Kisumu, and its crew for help and support, and Mr Sigi Engelhardt, Kikam Bala for help.
|
Footnotes |
|---|
1 These authors contributed equally to this paper, one (I.E.S.) by generating the data and the other (Y.S.) by analyzing the data statistically.
2 Present address: Department of Biosystems Science, The Graduate University for Advanced Studies (Sokendai), Hayama, Kanagawa, Japan.
William Martin, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAccession NumbersSupplementary MaterialAcknowledgementsReferences |
|---|
Alvarado Bremer JR, Mejuto J, Gómez-Márquez J, Boán F, Carpintero P, Rodriguez JM, Vinas J, Greig TW, Ely B. Hierarchical analyses of genetic variation of samples from breeding and feeding grounds confirm the genetic partitioning of Northwest Atlantic and South Atlantic populations of swordfish (Xiphias gladius L.). J Exp Mar Biol Ecol (2005) 327:167–182.[CrossRef][Web of Science]
Crapon de Caprona M-D, Frittzsch B. Interspecific fertile hybrids of haplochrominae cichlidae (Teleostei) and their possible importance for speciation. Neth J Zool (1984) 34:503–538.
Ford MJ. Testing models of migration and isolation among populations of Chinook salmon (Oncorhyncus tschwaystscha). Evolution (1998) 52:539–557.[CrossRef][Web of Science]
Fryer G. On the age and origin of the species flock of haplochromine cichlid fishes of Lake Victoria. Proc R Soc Lond B (2001) 268:1147–1152.[Medline]
Grant PR, Grant BR, Petren K. Hybridization in the recent past. Am Nat (2005) 166:56–67.[CrossRef][Web of Science][Medline]
Greenwood PH. Towards a phyletic classification of the "genus" Haplochromis (Pisces, Cichlidae) and related taxa. Part II: the species from Lakes Victoria, Nabugabo, Edward, George and Kivu. Bull Br Mus Nat Hist Zool (1979) 39:1–101.
Greenwood PH. The haplochromine fishes of the East African Lakes (1981) Ithaca (NY): Cornell University Press.
Guryev V, Koudijs MJ, Berezikov E, Johnson SL, Plasterk RH, van Eeden FJM, Cuppen E. Genetic variation in the zebrafish. Genome Res (2006) 16:491–497.
Hey J. On the failure of modern species concepts. Trends Ecol Evol (2006) 21:447–450.[CrossRef][Medline]
Johnson TC, Scholz CA, Talbot MR, Kelts K, Ricketts RD. Late Pleistocene desiccation of Lake Victoria and rapid evolution of cichlid fishes. Science (1996) 273:1091–1093.[CrossRef][Web of Science][Medline]
Kaessmann H, Wiebe V, Paabo S. Extensive nuclear DNA sequence diversity among chimpanzees. Science (1999) 286:1159–1162.
Klein D, Ono H, O'hUigin C, Vincek V, Goldschmidt T, Klein J. Extensive MHC variability in cichlid fishes of Lake Malawi. Nature (1993) 364:330–334.[CrossRef][Medline]
Klein J, Sato A, Nikolaidis N. MHC, TSP, and the origin of species. Annu Rev Genet (2007) 41:281–304.[Medline]
Li WH, Sadler LA. Low nucleotide diversity in man. Genetics (1991) 129:513–523.[Abstract]
Lincoln RJ, Boxhall GA, Clark PF. A dictionary of ecology, evolution and systematics (1982) Cambridge: Cambridge University Press.
Liu JX, Gao TX, Yokogawa K, Zhang YP. Differential population structuring and demographic history of two closely related fish species, Japanese sea bass (Lateolabrax japonicus) and spotted sea bass (L. maculates) in Northwestern Pacific. Mol Phylogenet Evol (2006) 39:799–811.[CrossRef][Web of Science][Medline]
Mallet J. Hybrid speciation. Nature (2007) 446:279–283.[CrossRef][Medline]
Mayr E. Systematics and the origin of species (1942) New York: Columbia University Press.
Nagl S, Tichy H, Mayer WE, Takahata N, Klein J. Persistence of neutral polymorphisms in Lake Victoria cichlid fish. Proc Natl Acad Sci USA (1998) 95:14238–14243.
Nagl S, Tichy H, Mayer WE, Takezaki N, Takahata N, Klein J. The origin and age of the haplochromine species flock in Lake Victoria, East Africa. Proc R Soc Lond B (2000) 267:1049–1061.[Medline]
Nagl S, Tichy H, Mayer WE, Samonte IE, McAndrew BJ, Klein J. Classification and phylogenetic relationships of African tilapiine fishes inferred from mitochondrial DNA sequences. Mol Phylogenet Evol (2001) 20:361–374.[CrossRef][Web of Science][Medline]
Nei M, Kumar S. Molecular evolution and phylogenetics (2000) New York: Oxford University Press.
Nesbo CL, Rueness EK, Iversen SA, Skagen DW, Jakobsen KS. Phylogeography and population history of Atlantic mackerel (Scomber scomrbus L.): a genealogical approach reveals genetic structuring among the eastern Atlantic stocks. Proc R Soc Lond B (2000) 267:281–292.[Medline]
Sage RD, Selander RK. Trophic radiation through polymorphism in cichlid fishes. Proc Nat Acad Sci USA (1975) 72:4669–4673.
Saitou N, Nei M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol (1987) 4:406–425.[Abstract]
Salzburger W, Baric S, Sturmbauer C. Speciation via introgressive hybridization in East African cichlids? Mol Ecol (2002) 11:619–625.[CrossRef][Medline]
Sato A, Takezaki N, Tichy H, Figueroa F, Mayer WE, Klein J. Origin and speciation of haplochromine fishes in East African crater lakes investigated by the analysis of their mtDNA, Mhc genes, and SINEs. Mol Biol Evol (2003) 20:1448–1462.
Satta Y. Comparison of DNA and protein polymorphisms between humans and chimpanzees. Genes Genet Syst (2001) 76:159–168.[CrossRef][Web of Science][Medline]
Schluter T. Geology of East Africa (1997) Berlin (Germany): Gebrüder Borntraeger.
Seehausen O. Lake Victoria rock cichlids: taxonomy, ecology, and distribution (1996) Netherlands: Verduijn Cichlids.
Seehausen O. Hybridization and adaptive radiation. Trends Ecol Evol (2004) 19:198–207.[CrossRef][Medline]
Seehausen O, Koetsier E, Schneider MV, Chapman LJ, Chapman CA, Knight ME, Turner GF, van Alphen JJM, Bills R. Nuclear markers reveal unexpected genetic variation and a Congolese-Nilotic origin of the Lake Victoria cichlid species flock. Proc R Soc Lond B (2002) 270:129–137.
Seehausen O, van Alphen JJM. Can sympatric speciation by disruptive sexual selection explain rapid evolution of cichlid diversity in Lake Victoria. Ecol Lett (1999) 2:262–271.[CrossRef][Web of Science]
Seehausen O, van Alphen JJM, Witte F. Cichlid fish diversity threatened by eutrophication that curbs sexual selection. Science (1997) 277:1808–1811.
Terai Y, Takezaki N, Mayer WE, Tichy H, Takahata N, Klein J, Okada N. Phylogenetic relationships among East African haplochromine fishes as revealed by short interspersed elements (SINEs). J Mol Evol (2004) 58:64–78.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  I. van der Sluijs, T. J.M Van Dooren, K. D Hofker, J. J.M van Alphen, R. B Stelkens, and O. Seehausen Female mating preference functions predict sexual selection against hybrids between sibling species of cichlid fish Phil Trans R Soc B, September 12, 2008; 363(1505): 2871 - 2877. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||