MBE Advance Access originally published online on February 6, 2006
Molecular Biology and Evolution 2006 23(4):828-837; doi:10.1093/molbev/msj101
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Nuclear Gene Variation and Molecular Dating of the Cichlid Species Flock of Lake Malawi
* Department of Genetics, Rutgers the State University of New Jersey, Piscataway; Department of Life Sciences, Ewha Womans University, Seoul, Korea; and Hopkins Marine Station, Stanford University
E-mail: hey{at}biology.rutgers.edu.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The cichlid fishes of Lake Malawi are famously diverse. However, evolutionary studies have been difficult because of their recent and uncertain phylogenetic history. Portions of 12 nuclear loci were sequenced in nine rock-dwelling species (mbuna) and three representatives of pelagic nonmbuna species. In contrast to the pattern of variation at mitochondrial genes, which do provide phylogenetic resolution at the level of mbuna versus nonmbuna, and among some genera, the nuclear loci were virtually devoid of phylogenetic signal. Only a small minority of variable positions were phylogenetically informative, and no phylogenetic branches are supported by more than one site. From the nuclear gene perspective the Malawian radiation appears to be a star phylogeny, as if the founding of the lake was accompanied by a partial bottleneck. The pattern is different from that found in Lake Victoria, in which nuclear loci share large amounts of ancestral variation. In the case of nuclear genes of Lake Malawi, the absence of phylogenetically informative variation suggests a relative absence of ancestral variation. Nuclear genes also differed from the mitochondria in having nearly twice the amount of divergence from Oreochromis (tilapia). An approximate splitting time between mbuna and nonmbuna lineages was estimated as 0.7 Myr. Oreochromis is estimated to have diverged from the cichlids in Lake Malawi and Lake Tanganyika about 18 MYA.
Key Words: Lake Malawi • cichlid • species flock • molecular clock
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The cichlid fishes (Teleostei: Cichlidae) of the East African Rift Valley lakes Malawi, Tanganyika, and Victoria and their satellites offer spectacular examples of rapid radiation (Fryer and Iles 1972
; Echelle and Kornfield 1984). For Lake Malawi, over 600 species have been described (including formal and informal descriptions) (Konings 2001). The estimated age of the lake is 4–5 Myr, with periodical desiccations estimated to have occurred perhaps as recently as 570,000 years ago (Delvaux 1996). With the exception of a small number of tilappine species, all of the Malawi cichlids appear to form a monophyletic group (Moran, Kornfield, and Reinthal 1994; Kocher et al. 1995; Shaw et al. 2000; Sturmbauer et al. 2001; Seehausen et al. 2003; Terai et al. 2004) that makes up a large subset of the tribe Haplochromini (Eccles and Trewavas 1989; Salzburger et al. 2005).
All Malawi haplochromine types fall into one of the two widely used, albeit informal, taxa: the mbuna (from the local name for rock-dwelling types) and the nonmbuna. All the mbuna are similar small forms that dwell in rocky shallow regions of the lake. The nonmbuna include a variety of pelagic and benthic forms that live over sandy portions of the lake bottom. "Mbuna" is described by Eccles and Trewavas (1989, p. 10) as a local name first used in the literature by Fryer (Fryer 1959) (Fryer and Iles 1972) and which was adopted by aquarists and researchers. The mbuna include the genera numbered 2 through 10 of Trewavas (1935) in addition to Iodotropheus (Oliver and Loiselle 1972). Recently, one of the genera, Pseudotropheus, has been the subject of revision and discussion (Meyer and Foerster 1984; Konings 1988; Stauffer et al. 1997), so that depending on one's choice of taxonomy there may be 10 or 11 genera of mbuna.
A recurring and challenging quest in the study of Lake Malawi haplochromines has been the search for characters or genetic loci that can be used to help resolve the phylogenetic history of this radiation (Greenwood 1979; Moran and Kornfield 1993; Sueltmann and Mayer 1997; Albertson et al. 1999). Among mbuna, genetic variation has been showed to be shared, both among species and among genera, for several types of genetic markers, including allozymes (Kornfield 1978; McKaye et al. 1982, 1984), mitochondrial haplotype data (Moran and Kornfield 1993; Parker and Kornfield 1997), and microsatellite or short tandem repeat (STR) loci (Kornfield and Parker 1997) and nuclear DNA sequences (Hey et al. 2004). The pattern has largely been interpreted as caused by recent speciation and persistence of ancestral polymorphism (Moran and Kornfield 1993), as well as being caused by occasional gene exchange (Danley et al. 2000; Hey et al. 2004; Won et al. 2005).
To date the clearest phylogenetic portraits that have emerged come from studies on the mitochondrial genome, with studies done on restriction fragment length polymorphisms (Moran and Kornfield 1993; Moran, Kornfield, and Reinthal 1994), control region sequences (Bowers, Stauffer, and Kocher 1994; Meyer, Montero, and Spreinat 1996; Parker and Kornfield 1997; Nagl et al. 2000; Shaw et al. 2000; Sturmbauer et al. 2001; Salzburger et al. 2005), and sequences of the reduced form of nicotinamide adenine dinucleotide dehydrogenase subunit 2 (ND2) gene (Kocher et al. 1995; Shaw et al. 2000; Salzburger et al. 2005). The main points from these studies are that the mbuna are indeed closely related to one another, but that they do not quite form a monophyletic group, and that the genera Rhamphochromis and Diplotaxodon are basal groups within the Lake Malawi radiation. The mitochondrial data do not offer much resolution for resolving relationships within and among mbuna genera. However, success has been had on this front using large numbers of amplified fragment length polymorphisms (AFLPs) (Albertson et al. 1999).
Although a few studies on nuclear genes have included some representatives from Lake Malawi (Terai, Morikawa, and Okada 2002; Sugie et al. 2004; Terai et al. 2004; Sugawara et al. 2005), very little phylogenetic work has been done on Lake Malawi cichlids using nuclear DNA sequences. Here, we inquire of the patterns and levels of DNA sequence variation found among Malawi at nuclear genes. Our general goal was to see what the Malawi radiation looks like from the perspective of sequences from nuclear loci. In particular, we assess whether the pattern that has emerged from studies on the mtDNA also occurs in the nuclear genome.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Mitochondrial Sequences
There have been several studies on variation in the mitochondrial ND2 gene among African great lake cichlids (Kocher et al. 1995
; Shaw et al. 2000; Salzburger et al. 2002, 2005; Duftner, Koblmuller, and Sturmbauer 2005; Schelly et al. 2005). For each represented species of mbuna or nonmbuna, one sequence was downloaded from GenBank (Supplementary Table 3, Supplementary Material online). The aligned partial sequences each had a length of 981 bp. For outgroups, we used sequences from the Lake Tanganyika species Eretmodus cyanostictus and from the tilapia Oreochromis niloticus (see below for discussion of outgroups).
Lake Malawi Species
Species of three genera of rock-dwelling (mbuna) cichlids, Tropheops, Labeotropheus, and Melanochromis were collected from the waters adjacent to the southern peninsula of Lake Malawi. These rock-dwelling cichlids include Tropheops gracilior, Tropheops tropheops, Tropheops sp. "broad mouth," Labeotropheus fuelleborni, Labeotropheus trewavasae, Melanochromis auratus, Melanochromis vermivorus, and Melanochromis melanopterus. Representatives of nonmbuna species, including Lethrinops gossei, Rhamphochromis ferox, and Diplotaxodon greenwoodi, were provided by J. Markert and M. Arnegard.
Outgroups and Dating
Lake Malawi is estimated to have last dried out somewhere between 0.57 and 1 MYA (Delvaux 1996). Because this date is recent and uncertain, we have used the origin of Lake Tanganyika together with a representative of the most basal cichlid lineage from that lake to root and date the radiation of Lake Malawi cichlids. The oldest parts of Lake Tanganyika arose 9–12 MYA (Cohen, Soreghan, and Scholz 1993), and Lake Tanganyika cichlids exhibit an ancient radiation, relative to those found in Lakes Malawi and Victoria. The most basal endemic group in Lake Tanganyika is the tribe Eretmodini (Kocher et al. 1995; Salzburger et al. 2002), and one representative of this small group, E. cyanostictus, was used as an outgroup. Given the uncertainties over the age and depth of the lake, we assumed that the age of this group falls somewhere between 5 and 10 MYA, with a median of 7.5 MYA for dating purposes.
For a more distant outgroup, we used a representative from the Tilapiini, a store-bought "red tilapia" which is of hybrid origin from both the Nile tilapia (O. niloticus) and the Mossambique tilapia (Oreochromis mossambicus). Both these species were included among 13 members of Oreochromis in an mtDNA study of tilapia-like cichlids (Klett and Meyer 2002). In that study, Oreochromis formed a well-supported monophyletic group (Klett and Meyer 2002).
Selected Loci
Twelve different genomic regions were used in the study. Three regions are from protein-coding genes, and nine are noncoding regions that flank STRs and which were accessible by the polymerase chain reaction (PCR) because of previous sequencing of the STR and its flanking regions in mbuna species.
The noncoding loci were obtained using the inverse PCR protocol (Ochman et al. 1989) as previously described (Hey et al. 2004). Locus-specific primer sequences were developed to sequence the flanking region (Supplementary Table 2, Supplementary Material online). The GenBank accession numbers for these sequences are DQ231254–DQ231333.
Portions of three protein-coding loci involved in pigmentation in African cichlid fishes (Sugie et al. 2004) were also included: Aim1 (Fukamachi, Shimada, and Shima 2001), endothelin receptor b1 (ednrb1) (Parichy et al. 2000), and mitfb (Lister et al. 1999; Lister, Close, and Raible 2001). For each locus, reported cDNA sequences from a Lake Malawi cichlid were mapped onto the genome of Takifugu rubripes to identify putative intron/exon boundaries. Primer pairs were designed so as to amplify both exon and intronic sequences. In the case of Aim1, the predicted gene region, based on the Takifugu genome assembly, contained a putative intron that turned out not to be present when the region was amplified from genomic DNA in the Malawi cichlids in this study. Consequently, the Aim1 sequences reported here contain a single open reading frame and do not appear to contain intronic sequence. For both ednrb1 and mitfb, amino acid sequences translated from the exon portion corroborated the reported exon/intron boundaries. The GenBank accession numbers for these loci are DQ239797–DQ239828.
Phylogenetic Analysis
Estimates of phylogenetic trees were generated under maximum parsimony using PAUP* v. 4.0 (Swofford 2002). For these analyses, all positions that were heterozygous within a species' sequence were coded as polymorphisms in the input file. Phylogenetic estimates were also generated using pairwise distances using the Tamura-Nei three-parameter method (Tamura and Nei 1993) as well as other methods. Distance tree estimates were generated under the minimum evolution criterion (Rzhetsky and Nei 1993) as implemented in the program MEGA 3.0 (Kumar, Tamura, and Nei 2004).
To test whether relative branch lengths are similar across loci, we used the relative ratio test of Muse and Gaut (1997) as implemented in the program HyPhy (Pond, Frost, and Muse 2005). These are likelihood ratio tests (LRTs) of the hypothesis that branch lengths for two trees generated for different data from a common set of taxa have the same relative lengths. Branch lengths were estimated for these tests under the Hasegawa-Kishino-Yano (HKY) mutation model (Hasegawa, Kishino, and Yano 1985).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Mitochondrial DNA
In order to establish a useful benchmark for contrasting nuclear gene variation with mitochondrial variation, a single database sequence for ND2 was used for each of as many species of Malawi cichlids as were available. The ND2 gene was selected because it has been sequenced in a large number of species and because the amount of homoplasy evident in the pattern of variation is far less than appears in sequences from the control region of the mitochondria. Figure 1 shows a distance tree, constructed under the minimum evolution criterion (Rzhetsky and Nei 1993
) as implemented in the program MEGA 3.0 (Kumar, Tamura, and Nei 2004), together with a majority rule consensus tree of 1,000 bootstrap replicates. In general, there seems to be good resolution for the deeper nodes of the tree. Very similar trees were found using maximum parsimony, and the parsimony bootstrap consensus tree had all of the same deeper branches with similar levels of support as the minimum evolution bootstrap consensus in figure 1. Additional control region data and ND2 data (Shaw et al. 2000) revealed the main exceptions from mbuna monophyly that are shown in figure 1. Most mbuna (shown in boxed text in figure 1) cluster together as well as with sequences from the genus Aulonocara and Lethrinops, while one mbuna sequence, Labidochromis caeruleus (as well as other Lethrinops sequences), cluster with other nonmbuna sequences.
|
Previous phylogenetic studies using mitochondrial DNA suggested that sequences from the genus Rhamphochromis were basal in the mitochondrial gene tree from Malawi (Meyer 1993
; Moran, Kornfield, and Reinthal 1994; Meyer, Montero, and Spreinat 1996), or that the genera Rhamphochromis and Diplotaxodon were each others' sister groups, and that together they formed a group that is sister to the remainder of Malawi cichlids (Shaw et al. 2000; Salzburger et al. 2005). A nearly basal position for Rhamphochromis was also suggested by AFLP markers (Seehausen et al. 2003), although the most basal taxon (among Malawi cichlids) in that study was Copadichromis virginalis.
The ND2 trees do not show the pattern, reported on the basis of both control region and ND2, in which Rhamphochromis and Diplotaxodon together formed a sister group to the other Malawi cichlids. Rather, the ND2 data clearly suggest a position for Rhamphochromis as a basal group, with Diplotaxodon (including a Pallidochromis sequence) forming the next basal group. This pattern did not change when trees where generated using other outgroup sequences from Lake Tanganyika (not shown). The most parsimonious trees had a length of 505 steps, whereas those that were constrained for monophyly of the Rhamphochromis, Diplotaxodon, and Palidochromis lineages had a minimum length of 528 steps. The difference was highly significant, with a Wilcoxon signed ranks test statistic of 4.27 (P < 0.0001) (Templeton 1983) and the "winning-sites" test (Prager and Wilson 1988) (P < 0.0001) as implemented in PAUP* 4.0.
If we assume that the divergence between Eretmodus and the Malawi cichlids began 7.5 MYA, then the estimated rate of divergence at ND2 is 0.008 changes per site per million years. Assuming a linearized tree, this rate places the base of the Malawi clade at just over 2 MYA, a date that is somewhat older than what might be expected if the radiation began within the past million years. A similar calculation for the base of the tree places the divergence of Oreochromis at 9.4 MYA, which is in accord with a previous estimate of this date based on ND2 data (Kocher et al. 1995). These rates and estimated time points change very little when alternative distance measures that correct for multiple hits in different ways are used.
It is possible that the 7.5 MYA figure is too large by a factor of two or more. If so, then the actual base of the Malawi clade for mitochondria would be closer to 1 MYA. Probably, a more likely explanation is that the current mtDNA tree includes considerable depth that has persisted and that was already present in the founding population of the Malawi clade.
Nuclear Genes
We examined DNA sequence variation across nearly 6 kbp from 12 loci (table 1), with the large majority of the sequence being noncoding (either intron or intergenic). A total of 48 variable base positions were found among eight mbuna and three nonmbuna cichlids from Lake Malawi. The variable base positions are shown in figure 2, together with two variable indel positions that were not associated with repeats (and which could be unambiguously aligned). Most individuals were heterozygous at one or more positions (as indicated by ambiguity codes in figure 2), and in many cases the heterozygous bases were also represented as differences between other sequences in the panel.
|
|
Visual inspection of figure 2 reveals a fairly striking observation in that there are few variable positions that seem to offer phylogenetic signal. For example, on the basis of the mtDNA, it might be expected that some sites would reveal an apparently derived base (i.e., different from the outgroup) that is limited to the mbuna sequences. Not one site reveals a base that is fixed or even nearly fixed among the mbuna. Similarly, on the basis of the mtDNA, which suggests that Rhamphochromis is basal within Malawi cichlids, it might be expected that some sites would support a branch between the pair of Rhamphochromis and Eretmodus and the remaining mbuna and nonmbuna sequences. One such position is found (gene 10 in figure 2) offering weak nuclear gene support for this mtDNA pattern. In the majority rule consensus of all most parsimonious trees, there are only two branches: one consisting of the mbuna species M. auratus, M. melanopterus, T. gracilior, and Tropheops broad mouth and the other consisting of the mbuna species L. fuelleborni and T. tropheops. Not included in any groups are the other mbuna species, L. trewavasae and M. vermivorus, and all three of the nonmbuna, L. gossei, R. ferox, and D. greenwoodi. In effect, the nuclear sequence data offer only very weak support for relatedness among some mbuna species.
Among the 48 variable positions found among Malawi cichlids, only 10 were parsimony informative (i.e., the less frequent base occurs at least twice), and all of these correspond to different partitions among the operational taxonomic units (OTUs). Six of these sites were found to be shared among mbuna genera, and two sites in the locus UNH130 were found to be shared between mbuna and nonmbuna. Surprisingly, none of the variable positions were found to be fixed in any type of comparisons (fig. 2). The absence of consistent patterns and fixed nucleotide differences among groups suggests that the nuclear loci are essentially devoid of phylogenetic signal among the species examined.
A DNA distance tree constructed for the concatenated nuclear gene data using the minimum evolution criterion as for ND2 (Rzhetsky and Nei 1993) is shown in figure 3. The branch that separates the outgroups from the Malawi cichlids was supported with 100% bootstraps. The branch that separates Rhamphochromis and the outgroups from the remaining Malawi cichlids had weak bootstrap support (54%). No other branches had more than 20% bootstrap support. If it is assumed that the divergence between Eretmodus and the Malawi cichlids began 7.5 MYA, the substitution rate is 1.1 x 10–9 changes per base per year. A linearized tree places the base of the Malawi clade at 0.70 MYA, which is about one-third of that estimated for ND2.
|
The same calculation for the base of the tree finds an estimated divergence time for Oreochromis of 17.7 MYA, which is nearly double that found for ND2. The ND2 sequences came from O. niloticus, whereas the nuclear gene sequences are from the commercial red tilapia which is a hybrid of O. niloticus and O. mossambicus. Given that the genus Oreochromis appears to be monophyletic, on the basis of mtDNA sequences from 13 species (Klett and Meyer 2002
), the fact that a species hybrid was used for the nuclear gene sequence probably does not explain the different degrees of divergence observed between mtDNA and nuclear genes.
Relative Ratio Tests
The relative branch lengths in figures 1 and 3 give the appearance that the Malawi radiation is older from the perspective of the mitochondrial ND2 gene than it is from the perspective of the nuclear loci (i.e., assuming that the time of separation of the Malawi cichlids from the outgroup Eretmodus is the same for both groups of genes). To test for a disparity in relative branch lengths, we used the relative ratio test of Muse and Gaut (1997) with simple three-taxon trees that included a representative mbuna (M. auratus) and a representative of the possibly basal genus Rhamphochromis, as well as the outgroup Eretmodus. This comparison is highlighted in the top panel of figure 4, which also shows the estimated dates for the nuclear loci and the mtDNA assuming a 7.5 MYA date for ancestry with Eretmodus. For this test, the nuclear loci were concatenated and treated as a single locus to compare with the ND2 data. The HKY substitution model (Hasegawa, Kishino, and Yano 1985) was applied globally for each data set (i.e., a single transition/transversion ratio was determined for each data set).
|
The LRT of the hypothesis that the relative branch lengths from the outgroup Eretmodus to the ingroup (the two Malawi taxa) are the same in the two data sets yielded a test statistic (two times the likelihood ratio or 2LR) of 4.061 which is statistically significant for the number of degrees of freedom (P = 0.04389, 1 df). In effect the apparent disparity in relative branch lengths receives modest support from this test. If the separation of Eretmodus was really at the same time for both nuclear and mitochondrial genes, then the base of the mitochondrial tree for the Malawi radiation (represented by M. auratus and Rhamphochromis) is deeper than it is in the case of the nuclear tree.
Another apparent disparity in relative branch lengths concerns the position of Oreochromis. The ND2 analysis suggests a divergence date a little less than 10 MYA, which is consistent with other reports and summaries based on mitochondrial data (Kocher et al. 1995; Salzburger and Meyer 2004). However, the nuclear data show a greater relative divergence of Oreochromis (table 1 and figs. 3 and 4) consistent with a divergence date of about 18 MYA. Evidence from fossils is not particularly helpful. A morphological analysis using cichlid fossils suggested that the divergence of tilapiine (which includes Oreochromis) and haplochromine cichlids was older than the 45 MYA date of the fossils described in the study (Murray 2001), a history that is not suggested by any molecular data (Albertson et al. 1999; Klett and Meyer 2002; Sugie et al. 2004; Salzburger et al. 2005). It also does not appear to be the case that the relative divergence between Oreochromis and Lake Malawi cichlids varies greatly among the nuclear loci, relative to the divergence between Eretmodus and the Malawi cichlids. From table 1, which shows the mean divergence per base pair for these comparisons, the Oreochromis divergence varies from 1.8 times (in the case of mitfb) to 4.6 times (in the case of UNH001) the value for Eretmodus (values obtained by dividing OD2 by OD1 in table 1). The ratio of the weighted average across loci is 2.3. In contrast, the relative divergence for Oreochromis in the case of ND2 is only 1.3 times that found for Eretmodus (fig. 1).
A relative ratio test between the ND2 data and concatenated nuclear loci was conducted using a Malawi cichlid, D. greenwoodii, in addition to Eretmodus and Oreochromis. This comparison is highlighted in the lower panel of figure 4, which also highlights the estimated dates for Oreochromis coancestry, assuming a 7.5 MYA date for Eretmodus. Assuming the HKY model for both loci and a global transition/transversion ratio for each of the data sets, the test was highly significant (2LR = 16.425, P = 5.0613 x 10–5, 1 df). To test for variation among the nuclear loci in the relative branch length from Oreochromis to the pairing of Eretmodus and Diplotaxodon, a LRT was done in which a model that constrained the Oreochromis branch to be a constant multiple of the branch lengths of Eretmodus and Diplotaxodon was compared to one without that constraint. The test for the seven loci for which sequence data could be obtained for all three taxa was not significant (2LR = 8.08843, P = 0.2317, 6 df). This is consistent with a relatively constant position of Oreochromis to Eretmodus and the Malawi cichlids, across nuclear loci.
It is possible that the mitochondrial ND2 tree, which is based on protein-coding sequence, is different from the nuclear loci because of natural selection. This was tested by estimating the branch lengths for a simple tree (again using M. auratus, Rhamphochromis, and Eretmodus) and estimating the synonymous and nonsynonymous branch lengths under the codon substitution model of Muse and Gaut (1994). The likelihood of this unconstrained tree was then compared to that for a model in which the nonsynonymous branch lengths were all a constant multiple of the corresponding synonymous branch lengths using the HyPhy program. The test statistic did not approach statistical significance (2LR = 0.819123, P = 0.66394, 2 df).
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
In contrast with studies based on the mitochondrial genome, we found essentially zero phylogenetic resolution among 12 nuclear regions. This is not simply because the data set is small. Not only do we not find phylogenetic signal among the mbuna, which are a particularly closely related group on the basis of mtDNA and AFLP studies and which have long been known to share genetic variation (Moran and Kornfield 1993
), but also we find no consistent pattern between mbuna and nonmbuna species or among nonmbuna species. If a phylogenetic tree interpretation is demanded of the data, then it appears as if the 11 species and six genera of Malawi cichlids form a star phylogeny. When a distance approach is used, weak support is found for Rhamphochromis, forming a basal connection to the rest of the Malawi cichlids (fig. 3).
Previous studies of Lake Malawi cichlids using the mitochondrial genome had suggested that members of the informal mbuna taxon formed a closely related group that, with some exceptions, was close to being monophyletic. This pattern was also found in the analysis of published ND2 gene sequences (fig. 1). These sequences also yield strong support for a basal position of the nonmbuna genus Rhamphochromis, among the Malawi cichlids, which is consistent with some previous work (Meyer 1993; Moran, Kornfield, and Reinthal 1994; Meyer, Montero, and Spreinat 1996) but is partly in conflict with some reports that had placed Rhamphochromis together with the genus Diplotaxodon as the basal group in the Malawi radiation (Shaw et al. 2000; Salzburger et al. 2005). These two genera are the only ones among those that have multiple species represented in the study that appeared to be monophyletic or nearly so.
The pattern of variation found among Malawi cichlids at nuclear loci is similar in some respects to that observed at nuclear loci among haplochromine cichlids from Lake Victoria. Nagl et al. (1998) studied four randomly selected nuclear loci and discovered that variation was widely shared among 12 endemic Lake Victoria species as well as among some closely related riverine species. Something like this is observed among the mbuna of Lake Malawi, which share mitochondrial (Moran and Kornfield 1993) and microsatellite variation (Van Oppen et al. 2000; Won et al. 2005) as well as nuclear locus variation (table 1). Nagl et al. (1998) explained their observation of shared variation in Lake Victoria by recent colonization of the lake by a large number of riverine founders. In the case of the Malawi mbuna, the relative roles of recent shared ancestry and gene flow, as causes of shared variation, still remain to be resolved, though speciation was clearly quite recent in some cases (Won et al. 2005).
Like Lake Malawi, Lake Victoria has had a history of desiccation, possibly as recently as 15,000 years ago (Johnson et al. 1996; Sturmbauer et al. 2001). However, a study of the mitochondrial control region in a large number of species suggests that the major radiation dates to 100,000 years ago (Verheyen et al. 2003). This same analysis also suggested that present day mitochondrial variation descended from two mitochondrial lineages that were present among the founders of the Lake Victoria flock. This is consistent with the report from nuclear genes that showed widespread shared variation between species from Lake Victoria and related riverine species (Nagl et al. 1998).
The estimated time depth of the tree for the mitochondrial ND2 gene among Malawi cichlids (near 2 MYA—based on a linearized tree, see Results) also suggests that multiple mtDNA lineages persist since the founding of the Malawi flock, which may have occurred within the past 1 Myr since the last desiccation of the lake (Delvaux 1996). However, the nuclear gene variation among Malawi cichlids does not suggest the presence of multiple lineages that predate the founding of the lake. If this had occurred, then we would expect that some variation would persist and appear as phylogenetically informative sites among the nonmbuna and mbuna genera. Rather, the paucity of informative sites is suggestive of a small founding population for the Malawi species flock (i.e., a bottleneck) that radiated rapidly into different groups.
It is important to emphasize that the presence of just a small number of phylogenetically informative sites cannot simply be explained by gene exchange or by persistence of ancestral variation. If ancestral variation had persisted in multiple lineages of the Malawi radiation, then this would be expected to appear as phylogenetically informative sites. This is true even if there was widespread gene exchange among the lineages early in the radiations, as has been suggested. Gene exchange could well have occurred and contributed to the radiation (Seehausen 2004). Hybridization would also be expected to lead to a lack of phylogenetic resolution, but it would not necessarily remove parsimony-informative variation at individual base positions. In order for hybridization and gene flow across species to contribute to the loss of old informative variation, those processes would have had to occur at a rate sufficient that genetic drift within and among diverging species lineages could lead to the loss of polymorphisms from the entire species flock.
Dating Discrepancies
We have focused on the time point of coancestry for E. cyanostictus because this species is a member of the oldest extant lineage of cichlids that are endemic to Lake Tanganyika and is therefore likely to have an age that approaches the age of this lake. If this date is assumed to be shared by both nuclear and mitochondrial loci, then relative ratio tests indicate, with a test statistic that is on the margin of statistical significance, that the Malawi radiation as viewed from the mitochondrial ND2 gene is deeper than the same radiation as viewed from the nuclear loci. The greater depth of the mitochondrial tree is not in the direction that would be expected on the basis of population genetic considerations. Genes of sex-limited inheritance are expected to have a coalescent time roughly one-quarter of diploid autosomal genes (assuming an even sex ratio and equal variance in reproductive success by the two sexes), so the expected time depth of the mitochondrial tree, at each node that joins separate taxa, is less than for autosomal loci by a factor that is a function of the effective population size of the ancestral species at that node. Consideration of the coalescent time depth of different genes also raises the point that the normal stochastic variance among genes can be quite large (Hudson and Coyne 2002; Hudson and Turelli 2003). The only variance in the relative ratio test is that which arises under the mutational model that is assumed. If coalescent variance were also included in the test, then it is probable that the difference between nuclear and mitochondrial tree depths would not be statistically significant.
The nuclear and mitochondrial trees differ even more for the apparent depth of the node that joins Oreochromis to the other taxa in the study. Nuclear genes suggest a time depth of approximately 18 MYA, which is nearly double the estimate for the mitochondria. Also the nuclear loci were consistent with each other in showing a greater divergence for Oreochromis than does the mitochondria. The fact that the nuclear sequences came from a hybrid of two Oreochromis species probably does not account for this difference as mitochondrial-based phylogenetic studies have shown Oreochromis to be a monophyletic genus (Klett and Meyer 2002). The time point discrepancy important for the general understanding of the radiation of tilapiine cichlids in Africa.
Given the apparent relative time discrepancies, we must consider the possibility that the time point of divergence of Eretmodus and the lineage that led to the Malawi radiation may not be shared by the nuclear and mitochondrial genes. For example, it is possible that this node is actually much older for the mitochondrial tree than for the nuclear tree. For instance, the ND2 tree includes considerable depth that has persisted beyond the node and that was already present in the founding population of the Malawi clade. This is a valid interpretation of the apparent discrepancies around the dating of the Oreochromis divergence. However, if this were the case, it would also have strong implications for the timing of the base of the Malawi radiation. If the time point of Eretmodus on the ND2 tree is actually much older, then so would the apparent age of the Malawian radiation on the ND2 tree be. If the time point of the Eretmodus on the nuclear gene tree is actually much younger, then so would be the Malawian radiation for nuclear genes. Given that this interpretation generates even larger discrepancies and that the nuclear gene tree time depth for the Malawi radiation (0.7 MYA) roughly corresponds to the estimated time since the last desiccation of the lake (Delvaux 1996), it seems simpler to consider the divergence time point between Eretmodus and the Malawian lineage as being shared by the two types of loci. It is also important to keep in mind the current uncertainty over dates when the lake might not have been habitable for fish. For example, it is possible that Lake Malawi did not in fact dry out within the past 1 Myr. Preliminary results of a scientific drilling suggest that the lake has not dried out in the past 1.5 Myr (Scholz et al. 2005).
The finding that the relative divergence between Eretmodus, Oreochromis, and the Malawi cichlids is roughly consistent across the nuclear loci suggests that the cause of the discrepancy may lie with the mtDNA and that for some reason the mtDNA history has not accurately tracked the phylogenetic history of these species. However, this suggestion is tentative at this stage, and a better understanding of the discrepancy must await additional data from more nuclear loci, and particularly for more genera of Malawi cichlids.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary Tables 2 and 3 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Thanks to Tom Kocher and Axel Meyer for helpful comments on the manuscript. This research was supported by the National Science Foundation.
|
Footnotes |
|---|
Sudhir Kumar, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Albertson, R. C., J. A. Markert, P. D. Danley, and T. D. Kocher. 1999. Phylogeny of a rapidly evolving clade: the cichlid fishes of Lake Malawi, East Africa. Proc. Natl. Acad. Sci. USA 96:5107–5110.
Bowers, N., J. R. Stauffer, and T. D. Kocher. 1994. Intra- and interspecific mitcochondrial DNA sequence variation within two species of rock-dwelling cichlids (Teleostei: Cichlidae) from Lake Malawi, Africa. Mol. Phylogenet. Evol. 3:75–82.[CrossRef][Medline]
Cohen, A. S., M. J. Soreghan, and C. A. Scholz. 1993. Estimating the age of formation of lakes: an example from Lake Tanganyika. Geology 21:511–514.
Danley, P., J. Markert, M. Arnegard, and T. Kocher. 2000. Divergence with gene flow in the rock-dwelling cichlids of Lake Malawi. Evolution 54:1725–1737.[CrossRef][Web of Science][Medline]
Delvaux, D. 1996. Age of Lake Malawi (Nyasa) and water level fluctuations. Mus. R. Afr. Centr. Tervuren (Belg.) Dept. Geol. Min. Rapp. Ann. 1995–1996:99–108.
Duftner, N., S. Koblmuller, and C. Sturmbauer. 2005. Evolutionary relationships of the limnochromini, a tribe of benthic deepwater cichlid fish endemic to Lake Tanganyika, East Africa. J. Mol. Evol. 60:277–289.[CrossRef][Web of Science][Medline]
Eccles, D. H., and E. Trewavas. 1989. Malawian cichlid fishes: the classification of some haplochromine genera. Lake Fish Movies, Herten, Germany.
Echelle, A. A., and I. Kornfield. 1984. Evolution of fish species flocks. University of Maine at Orono Press, Orono, Maine.
Fryer, G. 1959. The trophic interrelationships and ecology of some littoral communities in Lake Nyasa with especial reference to the fishes, and a discussion of the evolution of a group of rock-frequenting Cichlidae. Proc. Zool. Soc. Lond. 132:153–281.
Fryer, G., and T. D. Iles. 1972. The cichlid fishes of the great lakes of Africa. T.F.H. Publications, Neptune City, N.J.
Fukamachi, S., A. Shimada, and A. Shima. 2001. Mutations in the gene encoding B, a novel transporter protein, reduce melanin content in medaka. Nat. Genet. 28:381–385.[CrossRef][Web of Science][Medline]
Greenwood, P. H. 1979. Towards a phyletic classification of the ‘genus’ Haplochromis (Pisces, Cichlidae) and related taxa. Part I. Bull. Br. Mus. (Nat. Hist.) Zool. 45:265–322.
Hasegawa, M., H. Kishino, and T. Yano. 1985. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22:160–174.[CrossRef][Web of Science][Medline]
Hey, J., Y.-J. Won, A. Sivasundar, R. Nielsen, and J. A. Markert. 2004. Using nuclear haplotypes with microsatellites to study gene flow between recently separated cichlid species. Mol. Ecol. 13:909–919.[CrossRef][Medline]
Hudson, R. R., and J. A. Coyne. 2002. Mathematical consequences of the genealogical species concept. Evolution 56:1557–1565.[CrossRef][Web of Science][Medline]
Hudson, R. R., and M. Turelli. 2003. Stochasticity overrules the "three-times rule": genetic drift, genetic draft, and coalescence times for nuclear loci versus mitochondrial DNA. Evolution 57:182–190.[CrossRef][Web of Science][Medline]
Johnson, T. C., C. A. Scholz, M. R. Talbot, K. Kelts, R. D. Ricketts, G. Ngobi, K. Beuning, I. Ssemmanda, and J. W. McGill. 1996. Late Pleistocene dessication of Lake Victoria and rapid evolution of cichlid fishes. Science 273:1091–1093.[Abstract]
Klett, V., and A. Meyer. 2002. What, if anything, is a Tilapia?—mitochondrial ND2 phylogeny of tilapiines and the evolution of parental care systems in the African cichlid fishes. Mol. Biol. Evol. 19:865–883.
Kocher, T. D., J. A. Conroy, K. R. McKaye, J. R. Stauffer, and S. F. Lockwood. 1995. Evolution of NADH dehydrogenase subunit 2 in East African cichlid fish. Mol. Phylogenet. Evol. 4:420–432.[CrossRef][Web of Science][Medline]
Konings, A. 1988. Comments on the classification of species of the genus Melanochromis. Cichlid News 7:6–16.
———. 2001. Malawi cichlids in their natural habitat. Cichlid Press, El Paso, Tex.
Kornfield, I. 1978. Evidence for rapid speciation in African cichlid fishes. Experientia 34:335–336.[CrossRef][Web of Science]
Kornfield, I., and A. Parker. 1997. Molecular systematics of a rapidly evolving species flock: the mbuna of Lake Malawi and the search for phylogenetic signal. Pp. 25–37 in T. D. Kocher and C. A. Stepien, eds. Molecular systematics of fishes. Academic Press, New York.
Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief. Bioinform. 5:150–163.
Lister, J. A., J. Close, and D. W. Raible. 2001. Duplicate mitf genes in zebrafish: complementary expression and conservation of melanogenic potential. Dev. Biol. 237:333–344.[CrossRef][Web of Science][Medline]
Lister, J. A., C. P. Robertson, T. Lepage, S. L. Johnson, and D. W. Raible. 1999. Nacre encodes a zebrafish microphthalmia-related protein that regulates neural-crest-derived pigment cell fate. Development 126:3757–3767.[Abstract]
McKaye, E. R., T. Kocher, P. Reinthal, and I. Kornfield. 1982. A sympatric species complex of Petrotilapia Trewavas from Lake Malawi analysed by enzyme electrophoresis (Pisces, Cichlidae). Zool. J. Linn. Soc. 76:91–96.[CrossRef]
McKaye, K. R., T. Kocher, P. Reinthal, R. Harrison, and I. Kornfield. 1984. Genetic evidence of allopatric and sympatric differentiation among color morphs of a Lake Malawi cichlid fish. Evolution 38:215–219.[CrossRef][Web of Science]
Meyer, A. 1993. Phylogenetic relationships and evolutionary processes in East African cichlid fishes. Trends Ecol. Evol. 8:279.[CrossRef]
Meyer, B. J., C. M. Montero, and A. Spreinat. 1996. Molecular phylogenetic inferences about the evolutionary history of the East African cichlid fish radiations. Pp. 303–323 in T. C. Johnson and E. O. Odada, eds. Limnology, climatology and paleoclimatology of the East African lakes. Gordon and Breach Scientific, London.
Meyer, M. K., and W. Foerster. 1984. Un nouveau Pseudotropheus du lac Malawi avec des remarques sur le complexe Pseudotropheus-Melanochromis (Pisces, Perciformes, Cichlidae). Rev. Fr. Aquariol. Herpetol 10:107–112.
Moran, P., and I. Kornfield. 1993. Retention of an ancestral polymorphism in the Mbuna species flock (Teleostei: Cichlidae) of Lake Malawi. Mol. Biol. Evol. 10:1015–1029.[Web of Science]
Moran, P., I. Kornfield, and P. N. Reinthal. 1994. Molecular systematics and radiation of the haplochromine cichlids (Teleostei: Perciformes) of Lake Malawi. Copeia 2:274–288.[CrossRef]
Murray, A. M. 2001. The oldest fossil cichlids (Teleostei: Perciformes): indication of a 45 million-year-old species flock. Proc. R. Soc. Lond. B Biol. Sci. 268:679–684.[Medline]
Muse, S. V., and B. S. Gaut. 1994. A likelihood approach for comparing synonymous and non-synonmous nucleotide substitution rates with applications to the chloroplast genome. Mol. Biol. Evol. 11:715–724.[Abstract]
———. 1997. Comparing patterns of nucleotide substitution rates among chloroplast loci using the relative ratio test. Genetics 146:393–399.[Abstract]
Nagl, S., H. Tichy, W. Mayer, N. Takezaki, N. Takahata, and J. Klein. 2000. The origin and age of haplochromine fishes in Lake Victoria, East Africa. Proc. R. Soc. Lond. B Biol. Sci. 267:1049–1061.[Medline]
Nagl, S., H. Tichy, W. E. Mayer, N. Takahata, and J. Klein. 1998. Persistence of neutral polymorphisms in Lake Victoria cichlid fish. Proc. Natl. Acad. Sci. USA 95:14238–14243.
Ochman, H., J. W. Ajioka, D. Garza, and D. L. Hartl. 1989. Inverse polymerase chain reaction. Pp. 105–111 in H. A. Erlich, ed. PCR technology, principles and application for DNA amplification. Stockton Press, New York.
Oliver, M. K., and P. V. Loiselle. 1972. A new genus and species of cichlid of the mbuna group (Pisces: Cichlidae) from Lake Malawi. Rev. Zool. Bot. Afr. 85:309–320.
Parichy, D. M., E. M. Mellgren, J. F. Rawls, S. S. Lopes, R. N. Kelsh, and S. L. Johnson. 2000. Mutational analysis of endothelin receptor b1 (rose) during neural crest and pigment pattern development in the zerafish Danio rerio. Dev. Biol. 227:294–306.[CrossRef][Web of Science][Medline]
Parker, A., and I. Kornfield. 1997. Evolution of the mitochondrial DNA control region in the mbuna (Cichlidae) species flock of Lake Malawi. J. Mol. Evol. 45:70–83.[CrossRef][Web of Science][Medline]
Pond, S. L., S. D. Frost, and S. V. Muse. 2005. HyPhy: hypothesis testing using phylogenies. Bioinformatics 21:676–679.
Prager, E. M., and A. C. Wilson. 1988. Ancient origin of lactalbumin from lysozyme: analysis of DNA and amino acid sequences. J. Mol. Evol. 27:326–335.[CrossRef][Web of Science][Medline]
Rzhetsky, A., and M. Nei. 1993. Theoretical foundation of the minimum-evolution method of phylogenetic inference. Mol. Biol. Evol. 10:1073–1095.[Abstract]
Salzburger, W., T. Mack, E. Verheyen, and A. Meyer. 2005. Out of Tanganyika: genesis, explosive speciation, key-innovations and phylogeography of the haplochromine cichlid fishes. BMC Evol. Biol. 5:17.[CrossRef][Medline]
Salzburger, W., and A. Meyer. 2004. The species flocks of East African cichlid fishes: recent advances in molecular phylogenetics and population genetics. Naturwissenschaften. 91:227–290.
Salzburger, W., A. Meyer, S. Baric, E. Verheyen, and C. Sturmbauer. 2002. Phylogeny of the Lake Tanganyika cichlid species flock and its relationship to the Central and East African haplochromine cichlid fish faunas. Syst. Biol. 51:113–135.[CrossRef][Medline]
Schelly, R., W. Salzburger, S. Koblmuller, N. Duftner, and C. Sturmbauer. 2005. Phylogenetic relationships of the lamprologine cichlid genus Lepidiolamprologus (Teleostei: Perciformes) based on mitochondrial and nuclear sequences, suggesting introgressive hybridization. Mol. Phylogenet. Evol. 38:426–438.
Scholz, C. A., T. C. Johnson, J. W. King, A. S. Cohen, R. P. Lyons, L. Kalindekafe, S. L. Forman, L. R. McHargue, and B. S. Singer. 2005. Initial results of scientific drilling on Lake Malawi, East African rift. EOS Trans. Am. Geophys. Union 86:PP13C-03.
Seehausen, O. 2004. Hybridization and adaptive radiation. Trends Ecol. Evol. 19:198–207.
Seehausen, O., E. Koetsier, M. V. Schneider, L. J. Chapman, C. A. Chapman, M. E. Knight, G. F. Turner, J. J. van Alphen, and R. Bills. 2003. Nuclear markers reveal unexpected genetic variation and a Congolese-Nilotic origin of the Lake Victoria cichlid species flock. Proc. R. Soc. Lond. B Biol. Sci. 270:129–137.[Medline]
Shaw, P. W., G. F. Turner, M. R. Idid, R. L. Robinson, and G. R. Carvalho. 2000. Genetic population structure indicates sympatric speciation of Lake Malawi pelagic cichlids. Proc. R. Soc. Lond. Ser. B 267:2273–2280.[Medline]
Stauffer, J. R., N. J. Bowers, K. A. Kellogg, and K. R. McKaye. 1997. A revision of the blue-black Pseudotropheus zebra (Teleostei: Cichlidae) complex from Lake Malawi, Africa, with a description of a new genus and ten new species. Proc. Acad. Nat. Sci. Phila. 148:189–230.
Sturmbauer, C., S. Baric, W. Salzburger, L. Ruber, and E. Verheyen. 2001. Lake level fluctuations synchronize genetic divergences of cichlid fishes in African lakes. Mol. Biol. Evol. 18:144–154.
Sueltmann, H., and W. E. Mayer. 1997. Reconstruction of cichlid fish phylogeny using nuclear DNA markers. Pp. 39–51 in T. D. Kocher and C. A. Stepien, eds. Molecular systematics of fishes. Academic Press, New York.
Sugawara, T., Y. Terai, H. Imai, G. F. Turner, S. Koblmuller, C. Sturmbauer, Y. Shichida, and N. Okada. 2005. Parallelism of amino acid changes at the RH1 affecting spectral sensitivity among deep-water cichlids from Lakes Tanganyika and Malawi. Proc. Natl. Acad. Sci. USA 102:5448–5453.
Sugie, A., Y. Terai, R. Ota, and N. Okada. 2004. The evolution of genes for pigmentation in African cichlid fishes. Gene 343:337–346.[CrossRef][Web of Science][Medline]
Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4. Sinauer Associates, Sunderland, Mass.
Tamura, K., and M. Nei. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10:512–526.[Abstract]
Templeton, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37:221–244.[CrossRef][Web of Science]
Terai, Y., N. Morikawa, and N. Okada. 2002. The evolution of the pro-domain of bone morphogenetic protein 4 (Bmp4) in an explosively speciated lineage of East African cichlid fishes. Mol. Biol. Evol. 19:1628–1632.
Terai, Y., N. Takezaki, W. E. Mayer, H. Tichy, N. Takahata, J. Klein, and N. Okada. 2004. Phylogenetic relationships among East African haplochromine fish as revealed by short interspersed elements (SINEs). J. Mol. Evol. 58:64–78.[CrossRef][Web of Science][Medline]
Trewavas, E. 1935. A synopsis of the cichlid fishes of Lake Nyasa. Ann. Mag. Nat. Hist. 16:65–118.
Van Oppen, M., C. Rico, G. Turner, and G. Hewitt. 2000. Extensive homoplasy, nonstepwise mutations, and shared ancestral polymorphism at a complex microsatellite locus in Lake Malawi cichlids. Mol. Biol. Evol. 17:489–498.
Verheyen, E., W. Salzburger, J. Snoeks, and A. Meyer. 2003. Origin of the superflock of cichlid fishes from Lake Victoria, East Africa. Science 300:325–329.
Won, Y. J., A. Sivasundar, Y. Wang, and J. Hey. 2005. On the origin of Lake Malawi cichlid species: a population genetic analysis of divergence. Proc. Natl. Acad. Sci. USA 102:6581–6586.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. S. Cohen, J. R. Stone, K. R. M. Beuning, L. E. Park, P. N. Reinthal, D. Dettman, C. A. Scholz, T. C. Johnson, J. W. King, M. R. Talbot, et al. Ecological consequences of early Late Pleistocene megadroughts in tropical Africa PNAS, October 16, 2007; 104(42): 16422 - 16427. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.D Hulsey, M.C Mims, and J.T Streelman Do constructional constraints influence cichlid craniofacial diversification? Proc R Soc B, August 7, 2007; 274(1620): 1867 - 1875. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Genner, O. Seehausen, D. H. Lunt, D. A. Joyce, P. W. Shaw, G. R. Carvalho, and G. F. Turner Age of Cichlids: New Dates for Ancient Lake Fish Radiations Mol. Biol. Evol., May 1, 2007; 24(5): 1269 - 1282. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||