Springer 2006
Hydrobiologia (2006) 553:255–266 DOI 10.1007/s10750-005-1301-3
Primay Research Paper
The phylogenetic relationships of the Gobioninae (Teleostei: Cyprinidae) inferred from mitochondrial cytochrome b gene sequences Jinquan Yang1,2, Shunping He1, Jo¨rg Freyhof3, Kai Witte4 & Huanzhang Liu1,* 1
Institute of Hydrobiology, The Chinese Academy of Sciences, 430072, Wuhan, China Graduate School of Chinese Academy of Sciences, 100039, Beijing, China 3 Department of Biology and Ecology of Fishes, Leipniz-Institute of Freshwater Ecology and Inlands Fisheries, Mu¨ggelseedamm 310, 12587, Berlin, Germany 4 Max Planck Institute for Developmental Biology, Spemannstr. 35/3, 72076, Tu¨bingen, Germany (*Author for correspondence: Tel.: +86-27-68780776; Fax: +86-27-68780123; E-mail:
[email protected]) 2
Received 6 April 2005; accepted 4 June 2005
Key words: phylogeny, monophyly, cytochrome b, diversity, Cyprinidae, Gobioninae
Abstract The Gobioninae are a group of morphologically and ecologically diverse Eurasian freshwater cyprinid fishes. The intergeneric relationships of this group are unresolved and the possible monophyly of this subfamily remains to be established. We used complete mitochondrial cytochrome b gene sequences from most genera within the gobionine group, in addition to a selection of cyprinid outgroups, to investigate the possible monophyly of this group and resolve the interrelationships within the group. Our results support the monophyly of the Gobioninae and identify four monophyletic groups within the subfamily; the Hemibarbus group, the Sarcocheilichthys group, the Gobio group, and the Pseudogobio group. The morphologically aberrant genera Gobiobotia, Xenophysogobio and Gobiocypris are included in the Gobioninae, with the latter a sister group of Gnathopogon.
Introduction The Cyprinidae contains more than 2000 species and is one of the largest teleost families (Nelson, 1994). The phylogenetic structure of this family is still under debate, though several subfamilies are widely accepted and appear to form well-supported monophyletic clades (Chen et al., 1984; Howes, 1991). One of these is the Gobioninae, the gudgeons, a speciose subfamily with approximately 130 species in about 30 genera that are widely distributed throughout northern and eastern Eurasia from Spain east to Japan and south to central Vietnam. With the exception of the endemic western Palaearctic genus Romanogobio, and the genus Gobio, which is widespread from Spain to northern China, the gobionine cyprinids
exhibit greatest diversity and species richness in eastern Asia (China, Japan, Korea and Vietnam), particularly on the Chinese mainland (Bana: rescu & Nalbant, 1973; Luo et al., 1977; Bana: rescu, 1992; Yue, 1998). Gobionines are a group of small to mediumsized cyprinids that exhibit great morphological, ecological, and behavioural variation (Bana: rescu & Nalbant, 1973; Bana: rescu, 1992; Yue, 1998). Some are benthic and rheophilic, while others are semi-pelagic. Some species only inhabit fast-running cold headwater streams and others are found in tropical swamps. Some gudgeons possess an inferior mouth with 8 barbels while in some species the mouth is superior and barbels are absent. Some gobionines exhibit parental care of their eggs, while others lay them on the gills of
256 freshwater mussels, and yet others release their pelagic eggs into stream currents. Because of their morphological and ecological diversity, there is debate about which genera should be included in the Gobioninae and whether this subfamily constitutes a monophyletic unit within the Cyprinidae (Liu, 1940; Ramaswami, 1955; Hosoya, 1986). Liu (1940) proposed that the morphologically highly derived genera Gobiobotia and Xenophysogobio should to be excluded from the Gobioninae, forming a distinct subfamily the Gobiobotinae. Hosoya (1986) divided the Gobioninae into two branches, which comprised two monophyletic units, which may constitute two distinct subfamilies. The position of the Gobioninae within the Cyprinidae is poorly understood. Most authors using morphological data associate these fishes with the large cyprinid subfamily, the Cyprininae (also sometimes called the Barbinae) (Mori, 1935; Berg, 1940; Luo et al.; 1977; Arai, 1982; Hosoya, 1986; Howes, 1991). In contrast, more recent authors, such as Cavender & Coburn (1992) and Chen et al. (1984), have grouped the Gobioninae within the large cyprinid subfamily the Leuciscinae, with the cyprinid subfamily the Acheilognathinae as its sister-group. Recent molecular data appear to show that the Gobioninae may be a monophyletic group (Zardoya & Doadrio, 1999; Zardoya et al., 1999; Liu & Chen, 2003; He et al., 2004), though a contrary result was also obtained by Cunha et al. (2002). Molecular data also support the view that the gobionines should be grouped more closely within the cyprinids to the subfamily Leuciscinae rather than to the Cyprininae (Briolay et al., 1998; Zardoya et al., 1999; Zardoya & Doadrio, 1999; Gilles et al., 2001; Liu & Chen, 2003; He et al., 2004). However, with the exception of the study by Liu & Chen (2003), all these studies are based on a limited number of species. Consequently, the systematic position of the Gobioninae within the Cyprinidae, the monophyly of the group, and the phylogenetic relationships of genera within the subfamily are equivocal. The mitochondrial cytochrome b gene is widely used for phylogenetic analyses and is considered to be one of the most reliable mitochondrial markers for phylogenetic studies (Zardoya & Meyer, 1996). Here we use complete mitochondrial DNA cyto-
chrome b gene sequences for 49 gobionine species to obtain a better understanding of their phylogenetic relationships and the position of this group within the cyprinids. The aims of this study were: (1) to test the monophyly of Gobioninae, (2) to estimate the phylogenetic relationships among gobionine genera.
Materials and methods Samples collection Here we adopt the classification of the Gobioninae sensu Yue (1998). Specimens used in this study were collected from major rivers in China and Europe during 2000–2004. Due to the difficulties of obtaining samples, one gobionine genus from China, Ladislavia, one endemic to Vietnam, Parasqualidus, and two endemic to Korea, Pseudopungtungia and Coreoleuciscus, were not included in our analysis. All specimens for sequencing were preserved in 95% ethanol. Specimens of all the species used in this study are deposited in the collection of the Institute of Hydrobiology, Chinese Academy of Sciences and in the Fish Collection of Jo¨rg Freyhof (FSJF, Berlin) and K.-E. Witte (KEW, Tu¨bingen), respectively. DNA sequences for some species were downloaded from GenBank for comparison. To test the systematic position of the Gobioninae within the family Cyprinidae, some species from different subfamilies of the Cyprinidae were also included for analysis. Species names, river drainages and GenBank accession numbers are shown in Table 1. mtDNA extraction, amplification and sequencing Genomic DNA was extracted from muscle tissue following standard phenol/chloroform protocols (Kocher et al., 1989). The mitochondrial Cyt b gene was amplified by the polymerase chain reaction (PCR) in 50 ll reactions containing 0.5 ll dNTPs (1 mM), 5 ll reaction buffer (200 mM Tris-HCl pH 8.4, 500 mM KCl, 50 mM MgCl), 2.5 ll of each primer (10 lM), 0.5 ll (2.5 U) of Taq DNA Polymerase, 3 ll template DNA and 36 ll H2O. The primer pairs used were L14724 (5¢-GAC TTG AAA AAC CAC CGT TG-3¢) and H15915 (5¢-CTC CGA TCT CCG GAT TAC AAG AC-3¢) (Xiao et al.,
257 Table 1. Samples used in the present study, showing GenBank accession number, river drainage and country from which specimens were collected Species
Accession numbers
Drainage
Origin
Abbottina rivularis Acanthogobio guitheri
AY953020 AY953003
Yangtze River Yellow River
China China
Belligobio numifer
AY952987
Yangtze River
China
Biwa zezera*
AF309507
Biwa Lake
Japan
Coreius guichenoti
AY953001
Yangtze River
China
Coreius heterrodon
AY953000
Yangtze River
China
Gnathopogon imberbis
AY952998
Yangtze River
China
Gnathopogon nicholsi
AY952997
Yangtze River
China
Gobio cynocephalus Gobio gobio
AY953005 AY953007
Amur River Rhine River
China Germany
Gobio macrocephalus
AY953006
Amur River
China
Gobio tenuicorpus
AY953004
Yellow River
China
Gobiobotia filifer
AY953002
Yangtze River
China
Gobiobotia meridionalis*
AF375867
Pearl River
China
Gobiocypris rarus*
AF375865
Yangtze River
China
Hemibarbus labeo
AY952988
Yangtze River
China
Hemibarbus maculatus Hemibarbus medius
AY952990 AY952989
Yangtze River Yangtze River
China China
Huigobio chinssuensis
AY953016
Yellow River
China
Mesogobio tumenensis
AY953008
Tumenjiang River
China
Microphysogobio fukiensis
AY953014
Minjiang River
China
Microphysogobio tungtingensis
AY953013
Yangtze River
China
Paracanthobrama guichenoti
AY952994
Yangtze River
China
Paraleucogobio strigatus
AY952999
Amur River
China
Platysmacheilus exiguus Platysmacheilus longibarbatus
AY953015 AY953017
Pearl River Yangtze River
China China
Pseudogobio guilinensis
AY953018
Pearl River
China
Pseudogobio vaillanti
AY953019
Yangtze River
China
Pseudorasbora elongata
AY952996
Yangtze River
China
Pseudorasbora parva
AY952995
Yangtze River
China
Pungtungia herzi*
AF375864
Unkown
Japan
Rhinogobio cylindricus
AY952992
Yangtze River
China
Rhinogobio hunanensis Rhinogobio typus
AY952993 AY952991
Yangtze River Yangtze River
China China
Romanogobio banaticus
AY952329
Danube River
Romania
Romanogobio kesslerii
AY952328
Dniester River
Ukraine
Romanogobio macropterus
AY952332
Arax River
Turkey
Romanogobio uranoscopus
AY952331
Danube River
Romania
Rostrogobio liaohensis
AY953012
Liaohe River
China
Sarcocheilichthys microoculus*
NC004694
Unkown
Japan
Sarcocheilichthys nigripinnis Sarcocheilichthys kiangsiensis
AY952983 AY952984
Yangtze River Yangtze River
China China
Saurogobio dabryi
AY953011
Yangtze River
China
Saurogobio gymnocheilus
AY953009
Yangtze River
China
Saurogobio immaculatus
AY953010
Hainan Island
China Continued on p. 258
258 Table 1. (Continued) Species
Accession numbers
Drainage
Origin
Squalidus gracilis*
AF375866
Unkown
Japan
Squalidus argentatus
AY952985
Yangtze River
China
Squalidus nitens Xenophysogobio boulengeri*
AY952986 AF375868
Yangtze River Yangtze River
China China
Acheilognathus chankaensis*
AF051854
Unkown
China
Aphyocypris chinensis*
AF307452
Liaohe River
China
Barbus barbus*
Y10450
Unkown
Europe
Discogobio tetrabarbatus
AY953022
Pearl River
China
Hemiculter leucisculus*
AF051865
Unkown
China
Hypophthalmichthys molitrix* Leuciscus waleckii
AF051866 AY953021
Unkown Amur River
China China
Megalobrama terminalis*
AF051872
Unkown
China
Opsariichthys uncirostris*
AF308437
Unkown
Japan
Rhodeus ocellatus*
AF051876
Unkown
China
Squaliobarbus curriculus*
AF051877
Unkown
China
Xenocypris davidi*
AF036195
Unkown
China
Zacco platypus*
AF309085
Unkown
China
Other cyprinid species
Asterisks indicate sequences obtained from GenBank.
2001). The PCR reaction was optimized at 35 cycles of 94 C for 40 s (denaturation), 56 C for 45 s (annealing), 72 C for 1 min (extension), and a final 72 C extension for 8 min. Amplified doublestranded DNA was purified using a BioStar glassmilk DNA purification kit following the manufacture’s instructions. Purified DNA was sequenced by a commercial sequencing company with the same primer set L14724 and H15915. Sequence alignment and phylogenetic analysis Nucleotide sequences were aligned using the Clustal X multiple alignments program (Thompson et al., 1997), with default gap penalties, and the output was later improved by eye. No ambiguous alignments were found, and no gaps postulated. DNA sequences were also translated into amino acid sequence in MEGA version 2.1 (Kumar et al., 2001), producing full-length reads with no internal stop codons. The amount of sequence variation and the Kimura’s two parameter distances were also analyzed using MEGA 2.1. Base compositional bias was calculated across all taxa, and a chisquare (v2) test of base heterogeneity was conducted using PAUP* version 4.0b10 (Swofford,
2002) for all positions. Nucleotide saturation was analyzed by plotting pairwise inferred substitutions against corrected GTR-distance values. Phylogenetic trees were estimated with both maximum parsimony (MP) and maximum likelihood (ML) optimality criteria using the computer program PAUP*. MP analysis used tree bisectionreconnection (TBR) branch swapping and a heuristic searches with 20 replicates of random stepwise additions. No topological restraints were enforced and all characters were included with equal weighting. Support for each node was evaluated using bootstrap values (Felsenstein, 1985) and Bremer decay indices (Bremer, 1994). Qualitative support for recovered nodes was assessed using a non-parametric bootstrap analysis with 1000 pseudoreplicates and a heuristic tree search with 20 additional sequence replicates. Bremer decay indices were calculated using TreeRot v2 (Sorenson, 1999). The appropriate model of sequence evolution was chosen on the basis of hierarchical likelihoodratio tests (LRTs) as implemented in ModelTest version 3.06 (Posada & Crandall, 1998). The parameters of the models were computed using PAUP*. Under the Akaike information criterion
259 (AIC), the general time reversible model with some sites assumed to be invariable and with variable sites assumed to follow a discrete gamma distribution (GTR+G+I; Yang, 1994) was consistently selected as the best-fit ML model of nucleotide substitution. The cyt b data sets had the following statistics: estimated nucleotide frequencies A = 0.3276, C = 0.3333, G = 0.0859, and T = 0.2532; nucleotide substitution rate matrix A-C = 0.3554, A-G = 8.0173, A-T = 0.2929, CG = 0.5767, C-T = 6.0624, and G-T = 1.0000; assumed proportion of invariable sites = 0.4745 and the Gamma distribution shape parameter (a)=0.6433. A heuristic tree search strategy with 20 random-addition sequences and TBR branch swapping was used with PAUP* to find the best ML tree. Because bootstrap analysis for node support was too time consuming, Bayesian posterior values were determined for the ML tree using the computer program MrBayes version 3.0b4 (Huelsenbeck & Ronquist, 2001), with the optimal model of sequence evolution determined from the LRTs. Bayesian analyses were initiated from random starting trees and four Markov chain Monte Carlo (MCMC) simulations were run simultaneously for 2 000 000 generations with tree sampling every 100 generations. Chain stationarity was achieved after only 30 000 generations (‘burnin’) and, therefore, 300 trees were subsequently discarded. The frequency that a particular clade occurs within the last 19701 trees after the burn-in was interpreted as a measure of node support. In accordance with Wahlberg et al. (2003), when discussing our results we refer to support values as either giving weak, moderate, good or strong support. We define weak support as Bremer support values of 1–2 (bootstrap values 50–63% and posterior probabilities 50–69%), moderate support as values between 3 and 5 (bootstrap values 64–75% and posterior probabilities 70– 84%), good support as values between 6 and 10 (bootstrap values 76–88% and posterior probabilities 85–94%), and strong support as values >10 (bootstrap values 89–100% and posterior probabilities 95–100%). Because transitions (Ts) often accumulate at rates higher than transversions (Tv), we also conducted weighted parsimony analyses under two Ts/Tv weighting schemes (2.6:1 and 5:1). The Ts/ Tv ratio of 2.6:1 was derived from ModelTest.
Both weighted MP analyses yielded two different consensus trees with most nodes unresolved. Therefore, the two results are not addressed in the discussion.
Results Sequence variations and mutation analysis A total of 44 complete nucleotide sequences of the mitochondrial cytochrome b gene were determined, of which 42 were gobionine taxa. These sequences were aligned with the complete cytochrome b sequences of 7 gobionine and 12 other cyprinid species obtained from GenBank. All protein encoding 1140 positions were analyzed, of which 578 (50.7%) were constant sites and 485 (42.5%) were phylogenetically informative sites using the parsimony criterion. Pairwise sequence divergence among gobionines varied from 0.4 (between Squalidus argentatus (Sauvage and Dabry) and Squalidus nitens (Gu¨nther)) to 23.8% (between Gobiocypris rarus (Ye and Fu) and Squalidus gracilis (Temminck and Schlegel)), with a mean of 18.8% (±0.7%). This figure is higher than that among out-group taxa, which ranged from 4.8 (between Xenocypris davidi (Bleeker) and Acheilognathus chankaensis (Dybowski)) to 24.7% (between Zacco platypus (Temminck and Schlegel) and Barbus barbus L.), with a mean of 17.2% (±0.7%) (table of genetic distance not shown). The mean base composition of cytochrome b gene sequences was similar to that previously reported for cyprinid fishes (Briolay et al., 1998; Durand et al., 2002), with low G content (15.2%) and almost equal A, T and C content (28.1%, 28.9%, 27.7%, respectively). The content of A+T (56.3%) was higher than that of C+G (43.7%), which fell within the range of the GC content typical for vertebrates (Nei & Kumar, 2000). Significant compositional biases exist at the second and especially the third codon position, where there was a marked under representation of G (13.3 and 6.7%, respectively). Base frequencies were homogeneous across all taxa for three positions (v2 = 167.91, df = 177, p = 0.676). However, nucleotide composition at the third position exhibited significant heterogeneity: first position, v2 = 36.29, df=177, p = 1.0; second position,
260 v2 = 2.51, df=177, p = 1.0; and third position, v2 = 474.07, df=177, p < 0.001. Variability among sequences was detected mainly in third codon positions, of which 98.4% were variable. For first and second positions, 37.6 and 13.4% were variable, respectively. Substitutions showed some level of saturation in third codon positions but not in first and second codon positions (not shown). Phylogenetic relationships Maximum likelihood analysis yielded two trees with a likelihood score of )ln L=20779.6. The
consensus Bayesian tree differed slightly for ML trees. Most nodes of ML trees were supported with significant ML Bayesian posterior probabilities (Fig. 1). The phylogenetic position of the Gobioninae closer to the Leuciscine than to the Cyprinine was recovered with a significant posterior probability (PP = 100). The monophyly of the subfamily Gobioninae was also strongly supported (PP = 100), and the gobionine genera were grouped into four discrete assemblages (Fig. 1). The first assemblage, the Hemibarbus group, has strong support as a monophyletic group (PP = 100), comprising the genera Belligobio, Hemibarbus and Squalidus (Fig. 1). However, the
Figure 1. Tree resulting from maximum likelihood analysis of the complete cyt b gene dataset under the GTR+G+I model. Maximum-likelihood score=)20779.58, a=0.64, and I=0.47. Numbers at nodes represent Bayesian posterior probabilities (given for posterior probabilities greater than 50%).
261 monophyly of the second assemblage, the Sarcocheilichthys group, is not supported by ML Bayesian posterior probability values greater than 50%. In addition, interrelationships within the Sarcocheilichthys group are not well resolved. Within this group, Coreius is at the basal position in the Bayesian tree with a posterior probability lower than 50% and forming an independent group in the ML tree. Sarcocheilichthys and the remaining genera form a monophyletic group with moderate ML Bayesian posterior probability values (PP=81) (Fig. 1) and Gobiocypris appears in this group as a sister genus of Gnathopogon with strong support. The genus Pseudorasbora is not monophyletic in the ML analyses, with Pseudorasbora elongata (Temminck & Schlegel) the sister taxon of Pungtungia herzi (Herzenstein). The monophyly of the third assemblage, the Gobio group including Acanthogobio and Mesogobio tumensis (Chang), is strongly supported (PP = 100). A sister-group relationship between the genera Romanogobio and Gobio is strongly supported (PP = 100) (Fig. 1). There is weak support for monophyly in the fourth assemblage, the Pseudogobio group (PP = 56). Within this group, Gobiobotia and Xenophysogobio are situated basally; Abottina and Saurogobio form a sister-group, being the sistergroup to several genera characterized by strongly papillated lips (PP = 59) (Fig. 1). The two most parsimonious trees were recovered in maximum parsimony analysis using uniform weights for all changes. Most of the MP trees are congruent with the ML tree in topology (Fig. 2). The monophyly of the Gobioninae is weakly supported by bootstrap analysis but well supported by Bremer decay indices (BP = 56 and decay score = 7). Only two of the four monophyletic assemblages recovered in ML analyses are supported with moderate bootstrap values and good decay scores (BP = 64 and decay score of 6 in the Hemibarbus group, BP = 69 and decay score of 6 in the Gobio group). The monophyly of the Sarcocheilichthys group is recovered with neither high bootstrap values nor high decay scores. When the genus Coreius is excluded, the remaining genera formed a monophyletic unit with a moderate decay score but with bootstrap values not greater than 50%. Within this group the monophyly of the genus Pseudorasbora is also not supported.
MP trees weakly supported the integrity of the Pseudogobio group. Moderate decay scores supported the formation of sister groups between Gobiobotia and Xenophysogobio as well as Abbottnia and Saurogobio with the other genera of the Pseudogobio group, but were not supported in bootstrap analysis. Like our results with ML trees, when Abbottina, Gobiobotia, Sauragobio and Xenophysogobio were excluded, the monophyly of the remaining genera with strongly papillated lips received moderate support by bootstrap analysis and good support from Bremer decay indices (Fig. 2). Tests of alternative topologies To further test the monophyly of the Pseudogobio and Sarcocheilichthys groups, we constrained these groups as monophyletic and performed further heuristic searches. The resulting trees did not differ significantly from those obtained without constraint enforcement in Kishino–Hasegawa (KH) (Kishino & Hasegawa, 1989) and two-tailed Wilcoxon signed-ranks tests (Templeton, 1983) (p > 0.05 in each case) (Table 2).
Discussion Monophyly and systematic position of Gobioninae Our results concur with those of Bana: rescu & Nalbant (1965, 1973) and Luo et al. (1977), supporting the view that the Gobioninae is a monophyletic subfamiliy within Cyprinidae. Our results contradict the polyphyletic origin of this group proposed by Hosoya (1986) and Howes (1991). Previous molecular work has demonstrated the monophyly of the European and North American subfamily the Leuciscinae, the African, Palearctic and East Asian subfamily the Cyprininae and the East Asian subfamily the Cultrinae, including the former subfamily, the Xenocyprininae (Zardoya & Doadrio, 1999; Zardoya et al., 1999; Gilles et al., 2001; Cunha et al., 2002; Liu & Chen, 2003; He et al., 2004). These studies indirectly demonstrate the gobionines to be monphyletic, since if these groups are all monophyletic within the cyprinid family, the gobionines cannot be members of these groups and vice versa. Within the Cyprinidae, the
262
Figure 2. Strict consensus of two equally parsimonious trees from the equally weighted cyt b gene dataset. The tree length is 5083 steps, CI is 0.192 and RI is 0.808. Numbers given above branches are bootstrap values and numbers below the branch are Bremer support values (given for bootstrap values greater than 50%). Table 2. Summary of the two-tailed Wilcoxon signed-ranks tests and Kishino–Hasegawa (KH) test of alternative topologies Topology
a
Wilcoxon signed-ranks test
Kishino–Hasegawa test
Maximum parsimony
Maximum likelihood
Tree length
n
MP
5099
25
Pseudogobionini monophyletica
5092
Sarcocheilichthyini monophyletica
5096
z
p
)1.4000
0.16
Best 49
)0.6676
0.50
Ln L
Ln L diff.
p 0.144
)25464.76
16.4401
)25448.32
Best
)25456.52
8.2031
0.623
Maximum parsimony tree recovered from constraint search.
Gobioninae seem to be more closely associated with the Leuciscinae than the Cyprinine, as suspected by Arai (1982), Chen et al. (1984), Gilles
et al. (2001), Liu & Chen’s (2003) and He et al. (2004), contradicting the view of Luo et al. (1977) and Hosoya (1986). However, our results do not
263 support the idea that the Gobioninae are the sister group to the Acheilognathinae (the bitterlings) as proposed by Chen et al. (1984). Major groups within Gobioninae Four major evolutionary lineages within the Gobioninae are postulated in the present study, which partly confirms the division of the gobionines proposed in previous studies (Liu, 1940; Ramaswami, 1955; Luo et al., 1977; Kim, 1984; Hosoya, 1986; Rainboth, 1991; B an a: rescu, 1992; Naseka, 1996; Yu & Yue, 1996; Yue, 1998). B an a: rescu (1992) suspected that Hemibarbus and Belligobio might represent the ancestral group of gobionine cyprinids. Hemibarbus has distinct morphological characters considered to be primitive; the last simple dorsal ray is strongly ossified, the vent is positioned immediately in front of the anus, and the pharyngeal teeth are arranged in three rows. The main morphological difference between Hemibarbus and Belligobio is that the former has a strongly ossified last unbranched dorsal ray. Consequently, B an a: rescu & Nalbant (1973) assigned Belligobio as a subgenus of Hemibarbus. Because the type species of Belligobio, B. eristigma (Jordan & Hubbs), is not included in the present analysis, further studies on the genus Belligobio and Hemibarbus may demonstrate that Belligobio should be included within the genus Hemibarbus. The close relationship of Squalidus to Hemibarbus and Belligobio is also supported by sequence data for the mitochondrial control region (Yang et al. unpublished data), whereas, morphological data indicate that Squalidus may be close to Gobio (B an a: rescu & Nalbant, 1965, 1973; Hosoya, 1986; B an a: rescu, 1992) or Gnathopogon (Luo et al., 1977; Yue, 1995). The monophyly of the Sarcocheilichthys group was supported by both ML and MP results, though bootstrap support, posterior probabilities and decay scores for this conclusion are not high. Our analysis showed that our Sarcocheilichthys group comprised most of the genera that were excluded from the Gobioninae by Hosoya (1986), with the exception of Rhinogobio and Paracanthobrama. Paracanthobrama was formerly classified as a subgenus of Hemibarbus by B an a: rescu & Nalbant (1973) and B an a: rescu (1992). Relationships between Paracanthobrama and the remaining
genera of our Sarcocheilichthys group are not certain in the present study. Rhinogobio appears to be a sister group of Gobiocypris or Gnathopogon, although it had neither bootstrap values greater than 50% nor significant posterior probabilities, it did show a high decay score in MP analyses. Assignment to the genus Pseudorasbora is mainly based on characters related to the position of the mouth, with Pseudorasbora and Pungtungia morphologically similar. Further studies may show Pungtungia to be a synonym of Pseudorasbora, alternatively Pseudorasbora elongata (Wu) may be transferred to the genus Pungtungia. Gobiocypris was formerly thought to be allied to Aphyocypris and was treated as a member of the cyprinid subfamily the Danioninae (Ye & Fu, 1983). However, it has six branched anal fin rays, which is a diagnostic morphological character for the Gobioninae. He et al. (2004) and results from the present study support the assignment of Gobiocypris to the Gobioninae, with a relationship to Gnathopogon. Additional molecular data are needed to better resolve the relationships among the genera of the Sarcocheilichthys group. In the present study the position of the genus Coreius appears ambiguous. This genus shows molariform teeth in one row, large body size, small eyes, and long barbels and tongue, which made it inconsistent with members of other groups. It was treated as an isolated aberrant group by Bana: rescu (1992), a conclusion that is supported by our ML analysis. However, it affiliates with the Sarcocheilichthys group in our Bayesian and MP analyses, so we tentatively group it with the Sarcocheilichthys group. Further molecular data are needed to confirm the taxonomic position of this genus. The monophyletic Gobio group comprises the genera Gobio, Romanogobio, Acanthogobio and Mesogobio tumensis. Our data support the view that Acanthogobio and M. tumensis should be included in the genus Gobio. Bana: rescu & Nalbant (1973) highlighted that Acanthogobio seem to be a morphologically derived species of Gobio. We cannot exclude the possibility that Acanthogobio and M. tumensis might by introgressed with mtDNA from Gobio species. Introgressive hybridization is not rare between closely related species (Avise, 2000; Doiron et al., 2002; Rognon & Guyomard, 2003) and is well known from
264 cyprinid fishes (Gilles et al., 1998; Durand et al., 2000; Tsigenopoulos et al., 2002). Resolution of the position of Acanthogobio and Mesogobio will come with further examination of the type species of Mesogobio, M. lachneri (B an a: rescu & Nalbant) and the use of further molecular markers. The monophyly of the Pseudogobio group was not well supported in our study by the uniform weighted MP method. However, the monophyletic constrained tree represented the best tree with our data and was not rejected by the Templeton and KH tests. This group comprises the genera Gobiobotia, Sauragobio, Abbottina, Pseudogobio, Platysmcheilus, Huigobio, Microphysogobio, Rostrogobio, Biwia and Xenophysogobio. Several morphological studies already support the view that these genera form a monophyletic unit (Hosoya, 1986; B an a: rescu, 1992), though some authors excluded Gobiobotia and Xenophysogobio from the group (Luo et al., 1977; Yu & Yue 1996; Yue, 1998). Gobiobotia and Xenophysogobio have had inconsistent systematic positions because of the curious bony cups and four pairs of barbels (contrasting with one pair in other genera). Most authors concur that Gobiobotia and Xenophysogobio belong to the Gobioninae (Taranetz, 1938; Ramaswami, 1955; B an a: rescu & Nalbant, 1973; Hosoya, 1986; B an a: rescu, 1992; Naseka, 1996), while Luo et al. (1977), Kim (1984) and Yue (1998) ascribed it to an independently subfamily. Hosoya (1986) believed Gobiobotia and Xenophysogobio should be placed in the Gobioninae since they share synapomorphies in the lower jaw and in the caudal base with other genera of the Gobioninae (Pseudogobio, Abbottina, Saurogobio, Microphysogobio and Biwia). Our molecular results support the view that Gobiobotia and Xenophysogobio belong to the Pseudogobio group. With the exclusion of Saurogobio, Gobiobotia, Abbottina and Xenophysogobio, all the genera exhibit strongly papillose lips with one or two smooth or papillose pads and a reduced air bladder with an encapsulated anterior chamber. In conclusion, our phylogenetic analysis using complete mitochondrial cytochrome b gene sequences suggests that the subfamily Gobioninae is a monophyletic group. In addition, the subfamily can be divided into a further four monophyletic groups: the Hemibarbus group, the Sarcocheilichthys group, the Gobio group, and the Pseudogobio
group. On the basis of our data, interrelationships among genera within the Sarcocheilichthys and Pseudogobio groups were not well resolved. Further molecular work examining the interrelationships of the Gobioninae should include additional mitochondrial or nuclear genes.
Acknowledgements We thank Dr. C. Smith for English correction and suggestion. Our thanks also to Q. Tang, Y. Zhu, X. Xia, L. Zhang and Y. Zeng for their help in collecting samples and laboratory work. We also would like to extend our thanks to D. He and Z. Peng for discussion and information about the manuscript. Funding supports are from NSFC 40432003, CAS KZCX-SW-126, and Cypriniformes TOL.
References Arai, R. M., 1982. A chromosome study on two cyprinid fishes, Acrossocheilus labiatus and Pseudorasbora pumila pumila, with notes on Eurasian cyprinids and their karyotypes. Bulletin of the National Science Museum, Tokyo (A) 8: 131– 152. Avise, J. C., 2000. Phylogeography. The History and Formation of Species. MA: Harvard University Press, Cambridge. Banarescu, P., 1992. A critical updated checklist of Gobioninae (Pisces, Cyprinidae). Travaux Du Muse´um D’histoire Naturelle ‘‘Grigore Antipa’’ 32: 303–330. Bana: rescu, P. & T. T. Nalbant, 1965. Studies on the systematics of Gobioninae (Pisces, Cyprinidae). Revue Roumaine de Biologie. Serie Biologie Animale 10: 219–229. Banarescu, P. & T. T. Nalbant, 1973. Pisces, Teleostei, Cyprinidae (Gobioninae). Das Tierreich, Lieferung 93. Walter de Guryter, Berlin. Berg, L. S., 1940. Classification of fishes, both recent and fossil. Transactions of Institute of Zoology, Academy of Sciences, USSR 5: 87–345. Bremer, K., 1994. Branch support and tree stability. Cladistics 10: 295–304. Briolay, J., N. Galtier, M.R. Brito & Y. Bouvet, 1998. Molecular phylogeny of Cyprinidae inferred from cytochrome b DNA sequences. Molecular Phylogenetics and Evolution 9: 100–108. Cavendar, T. M. & M. Coburn, 1992. Phylogenetic relationships of North American Cyprinidae. In: Mayden. R. (ed.), Systematics, historical ecology, and North American freshwater fishes. Stanford University Press, Stanford, CA, pp. 293–327. Chen, X., P. Yue & R. Lin, 1984. Major groups within the family Cyprinidae and their phylogenetic relationships. Acta Zootaxonomica Sinica 9: 424–440.
265 Cunha, C., N. Mesquita, T. E. Dowling, A. Gilles & M. M. Coelho, 2002. Phylogenetic relationships of Eurasian and American cyprinids using cytochrome b sequences. Journal of Fish Biology 61: 929–944. Doiron, S., L. Bernatchez & P. Blier, 2002. A comparative mitogenomic analysis of the potential adaptive value of arctic charr mtDNA introgression in brook charr populations. Molecular Biology and Evolution 19: 1902–1909. Durand, D. J., S. C. Tsigenopoulos, E. U¨nlu¨ & P. Berrebi, 2002. Phylogeny and biogeography of the family Cyprinidae in the Middle East inferred from Cytochrome b DNA – evolutionary significance of this region. Molecular Phylogenetics and Evolution 22: 91–100. Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. Gilles, A., G. Lecointre, E. Faure, R. Chappaz & G. Brun, 1998. Mitochondrial phylogeny of European cyprinids: implications for their systematics, reticulate evolution, colonization time. Molecular Phylogenetics and Evolution 10: 132–143. Gilles, A., G. Lecointre, A. Miquelis, M. Loerstcher, R. Chappaz & G. Brun, 2001. Partial combination applied to phylogeny of European cyprinids using the mitochondrial control region. Molecular Phylogenetics and Evolution 19: 22–33. He, S., H. Liu, Y. Chen, M. Kuwahara & T. Nakajima, 2004. Phylogenetic analysis of cyprinid fishes based on mitochondrial cytochrome b gene sequences. Science in China (C) 34: 96–104. Hosoya K., 1986. Interrelationships of the Gobioninae (Cyprinidae) In Uyeno T. et al. (ed.), Indo-Pacific Fish Biology: proceeding of the Second International Conference on Indo-Pacific Fishes Ichthyological Society of Japan, Tokyo: 484–501. Howes, G., 1991. Systematics and biogeography: an overview. In Winfield, I. J. & J. S. Nelson (eds), Cyprinid fishes. Systematics, Biology and Exploitation. Chapman and Hall, New York: 1–33. Huelsenback, J. P. & F. Ronquist, 2001. MrBayes: Bayesian inference of phylogeny. Bioinformatics 17: 754–755. Kim, I. S., 1984. The taxonomic study of gudgeons of the subfamily Gobioninae (Cyprinidae) in Korea. Bulletin of Korean Fishery Society 17: 436–448. Kishino, H. & M. Hasegawa, 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. Journal of Molecular Evolution 29: 170–179. Kocher, T. D., W. K. Thommas, A. Meyer, S. V. Edwards, F. X. Pallablanca & A. C. Wilson, 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences of the United States of America 86: 6196–6200. Kumar, S., K. Tamura, I. B. Jakobsen & M. Nei, 2001. MEGA: Molecular Evolutionary Genetics Analysis. Ver 2.1. Bioinformatics 17: 1244–1245. Liu, C. K., 1940. Preliminary studies on the air-bladder and its adjacent structures in Gobioninae. Sinensia 11 (1/2): 77–104. Liu, H. & Y. Chen, 2003. Phylogeny of the East Asian cyprinids inferred from sequences of the mitochondrial DNA control region. Canadian Journal of Zoology 81: 1938–1946.
Luo, Y., P. Yue & Y. Chen, 1977. Gobioninae. In Wu, X. W. (ed.), Cyprinid Fishes of China, II. People’s Press, Shanghai: 439–549. Mori, T., 1935. Description of two new genera and seven new species of Cyprinidae from chosen. Annotationes of Zoologicae Japonenses 15: 161–181, pls. 11–13. Naseka, A. M., 1996. Comparative study on the vertebral column in the Gobioninae (Cyprinidae, Pisces) with special reference to its systematics. Publicaciones Especiales. Instituto Espanol de Oceanografia 21: 149–167. Nei, M. & S. Kumar, 2000. Molecular Evolution and Phylogenetics. Oxford University Press, New York. Nelson, J. S., 1994. Fishes of the World (3rd edn.). John Wiley and Sons, New York. Posada, D. & K. A. Crandall, 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818. Rainboth, W. J., 1991. Cyprinids of South-East Asia. In Winfield, I. J. & J. S. Nelson (eds), Cyprinid fishes. Systematics, Biology and Exploitation. Chapman and Hall, New York: 156–211. Ramaswami, L. S., 1955. Skeleton of cyprinoid fishes in relation to phylogenetic studies. VI. The skull and Weberain apparatus in the subfamily Gobioninae (Cyprinidae). Acta Zoologica 36: 127–158. Rognon, X. & R. Guyomard, 2003. Large extent of mitochondrial DNA transfer from Oreochromis aureus to O niloticus in West Africa. Molecular Ecology 12: 435–445. Sorenson, M. D., 1999. TreeRot (version 2). MA: Boston University, Boston. Swofford, D. L., 2002. PAUP*: Phylogenetic Analysis Using Parsimony (* and other Methods) (version 4.0). MA: Sinauer Associates, Sunderland. Taranetz, A., 1938. On the relationship and origin of gudgeons of the basin of the Amur River. Moskva: Zoologicheskij Zhurnal 17: 453–471. Templeton, A. R., 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of human and apes. Evolution 37: 221–244. Thompson, J. D., T. J. Gibson, F. Plewwniak, F. Jeanmougin & D. G. Higgins, 1997. The Clustal X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25: 4876–4882. Tsigenopoulos, C. S., P. Ra´b, D. Naran & P. Berrebi, 2002. Multiple origins of polyploidy in the phylogeny of southern African barbs (Cyprinidae) as inferred from mtDNA markers. Heredity 88: 466–473. Wahlberg, N., E. Weingartner & S. Nylin, 2003. Towards a better understanding of the higher systematics of Nymphalidae (Lepidoptera: Papilionoidea). Molecular Phylogenetics and Evolution 28: 473–484. Xiao, W., Y. Zhang & H. Liu, 2001. Molecular systematics of Xenocyprinae (Teleostei: Cyprinidae): taxonomy, biogeography, and coevolution of a special group restricted in East Asia. Molecular Phylogenetics and Evolution 18: 163–173. Yang, Z., 1994. Estimating the pattern of nucleotide substitution. Journal of Molecular Evolution 39: 105–111. Ye, M. & T. Fu, 1983. Description of a new genus and species of Danioninae from China (Cypriniformes: Cyprinidae). Acta Zootaxonomica Sinica 8: 434–437.
266 Yu, L. & P. Yue, 1996. Studies on Phylogeny of the Pseudogobionini fishes (Cypriniformes: Cyprinidae). Acta Zootaxonomica Sinica 21: 244–255. Yue, P., 1995. Revision of the classification on the cyprinid fishes of the genus Squalidus from China. Acta Hydrobiologica Sinica 19: 89–91. Yue, P., 1998. Gobioninae. In Chen, Y. et al. (eds), Fauna Sinica. Osteichthyes. Cypriniformes II Science Press, Beijing: 232–389. Zardoya, R. & I. Doadrio, 1999. Molecular evidence on the evolutionary and biogeographical patterns of Euro-
pean cyprinids. Journal of Molecular Evolution 49: 227– 237. Zardoya, R., P. S. Economidis & I. Doadrio, 1999. Phylogenetic relationships of Greek Cyprinidae: Molecular evidence for at least two origins of the Greek cyprinid fauna. Molecular Phylogenetics and Evolution 13: 122–131. Zardoya, R. & A. Meyer, 1996. Phylogenetic performance of mitochondrial protein-coding genes in resolving relationships among vertebrates. Molecular Biology and Evolution 13: 933–942.