3. Origin of Euteleostean Biodiversity
37
assume a priori rates of the molecular clock for bony fishes. Rate calibration was
conducted using three independent data points, two of which are external calibration points based on reliable paleontological and molecular evidence. The other is
an internal calibration point reasonably assuming that cichlids could not disperse
between well-separated African and South American landmasses. Cichlids are secondary freshwater fishes that have undergone remarkable adaptive radiation on
landmasses of Gondwanaland origin (see, e.g., Nishida 1997; Stiassny and Meyer
1999). Recent phylogenetic studies of cichlids based 011 molecular (Stiltmann et al.
1995; Zardoya et al. 1996; Streelman and Karl 1997; Streelman et al. 1998;
Takahashi et al. 1998; Kanie and Nishida, manuscript in preparation) and morphological (Stiassny 1991) evidence consistently indicate that African and neotropical
cichlids are monophyletic relative to each other, and that Indian and Malagasy
species make outgroups of the African + neotropical clade. This phylogenetic framework is in good agreement with a currently well-accepted separation process of
Gondwanaland (Smith et al. 1994; The Plates Project 1998), supporting the close
association between the continental-clade formation of cichlids and the
Gondwanaland breakup. These three calibration points are congruent with each
other and serve to set a rate calibration for fishes, which may be useful for rough
estimation of divergence times among diverse fish groups.
2 Materials and Methods
2.1 Samples and Sequence Determination
Fifteen fish species sequenced in this study are listed in Table 1. Three species were
selected from each of African and neotropical cichlids so as to represent deep lineages among each group, in light of previous work (Kocher et al. 1995; Stiltmann
et al. 1995; Zardoya et al. 1996; Nishida 1997; Streelman et al. 1998; Takahashi et
al. 1998). Representatives of four out of six families in the suborder Labroidei
(Nelson 1994) were employed (i.e. Cichlidae, Embiotocidae, Pomacentridae, and
Labridae). Fish samples for African cichlids and non-cichlid perciforms were captured in native localities of individual species, while the others were obtained from
local shops. Total DNA was extracted from a small amount of muscle tissues or fins
according to Kocher et al. (1989) and used for the amplification by the polymerase
chain reaction (PCR) of genes for NADH dehydrogenase subunit 2 (ND2) and cytochrome b (cytb). They are two of the best represented genes for fishes among mitochondrial protein genes that can provide clearly orthologous sets of sequences.
The sequences and positions of the PCR primers used for the amplification are
shown in Table 2 and Fig. 1, respectively. Amplification of ND2 genes was carried
out either with L4296 and H5635 or with L4157m and H5934m. Amplification of
cytb genes was done with L14724m and H15990 except for T. polylepis cytb gene
that was amplified with L14724m and fcytb-2, and with fcytb-1 and H15915. PCR
reactions consisted of 30 cycles of denaturation at 92°C for 40 s, annealing at 45-
37
assume a priori rates of the molecular clock for bony fishes. Rate calibration was
conducted using three independent data points, two of which are external calibration points based on reliable paleontological and molecular evidence. The other is
an internal calibration point reasonably assuming that cichlids could not disperse
between well-separated African and South American landmasses. Cichlids are secondary freshwater fishes that have undergone remarkable adaptive radiation on
landmasses of Gondwanaland origin (see, e.g., Nishida 1997; Stiassny and Meyer
1999). Recent phylogenetic studies of cichlids based 011 molecular (Stiltmann et al.
1995; Zardoya et al. 1996; Streelman and Karl 1997; Streelman et al. 1998;
Takahashi et al. 1998; Kanie and Nishida, manuscript in preparation) and morphological (Stiassny 1991) evidence consistently indicate that African and neotropical
cichlids are monophyletic relative to each other, and that Indian and Malagasy
species make outgroups of the African + neotropical clade. This phylogenetic framework is in good agreement with a currently well-accepted separation process of
Gondwanaland (Smith et al. 1994; The Plates Project 1998), supporting the close
association between the continental-clade formation of cichlids and the
Gondwanaland breakup. These three calibration points are congruent with each
other and serve to set a rate calibration for fishes, which may be useful for rough
estimation of divergence times among diverse fish groups.
2 Materials and Methods
2.1 Samples and Sequence Determination
Fifteen fish species sequenced in this study are listed in Table 1. Three species were
selected from each of African and neotropical cichlids so as to represent deep lineages among each group, in light of previous work (Kocher et al. 1995; Stiltmann
et al. 1995; Zardoya et al. 1996; Nishida 1997; Streelman et al. 1998; Takahashi et
al. 1998). Representatives of four out of six families in the suborder Labroidei
(Nelson 1994) were employed (i.e. Cichlidae, Embiotocidae, Pomacentridae, and
Labridae). Fish samples for African cichlids and non-cichlid perciforms were captured in native localities of individual species, while the others were obtained from
local shops. Total DNA was extracted from a small amount of muscle tissues or fins
according to Kocher et al. (1989) and used for the amplification by the polymerase
chain reaction (PCR) of genes for NADH dehydrogenase subunit 2 (ND2) and cytochrome b (cytb). They are two of the best represented genes for fishes among mitochondrial protein genes that can provide clearly orthologous sets of sequences.
The sequences and positions of the PCR primers used for the amplification are
shown in Table 2 and Fig. 1, respectively. Amplification of ND2 genes was carried
out either with L4296 and H5635 or with L4157m and H5934m. Amplification of
cytb genes was done with L14724m and H15990 except for T. polylepis cytb gene
that was amplified with L14724m and fcytb-2, and with fcytb-1 and H15915. PCR
reactions consisted of 30 cycles of denaturation at 92°C for 40 s, annealing at 45-