Mitogenomic perspectives into sciaenid fishes' phylogeny and evolution origin in the New World

Gene 539 (2014) 91–98

Contents lists available at ScienceDirect

Gene journal homepage: www.elsevier.com/locate/gene

Mitogenomic perspectives into sciaenid fishes' phylogeny and evolution origin in the New World Tianjun Xu ⁎, Da Tang, Yuanzhi Cheng, Rixin Wang Laboratory of Fish Biogenetics & Immune Evolution, College of Marine Science, Zhejiang Ocean University, Zhoushan 316022, China

a r t i c l e

i n f o

Article history: Accepted 15 January 2014 Available online 29 January 2014 Keywords: Sciaenidae Mitochondrial genome Phylogeny Divergence time Origin

a b s t r a c t Sciaenid fishes are widely distributed throughout the coastal waters and estuaries of the world. A total of 23 genera of this family are endemic to the Old World. However, evolutionary relationships among Old World sciaenid fishes and their origin have remained unresolved despite their diversity and importance. Besides, hypotheses that explain the origin and biogeographical distribution of sciaenid fishes are controversial. In this study, the complete mitochondrial genome sequences of seven representative sciaenid species were determined and a well-resolved tree was recovered. This new timescale demonstrated that the sciaenid originated during the late Jurassic to early Cretaceous Period. The estimated origin time of sciaenid fish is 208 Mya, and the origin of Old World sciaenid is estimated at 126 Mya. Reconstruction of ancestral distributions indicated a plesiomorphic distribution and center of origin in the New World, with at least one lineage subsequently dispersed to the Old World. Moreover, we conclude that the common ancestors of Old World sciaenid fishes were derived from species of New World. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Sciaenid fishes (Croaker or drum fish, family Sciaenidae) are commercially important fish groups in temperate to tropical coastal waters and estuaries throughout the world. They are particularly abundant at the mouths of large continental rivers, but are poorly known as groups that lived in seawater (Sasaki, 1989). This family comprises 67 to 78 genera, of which 23 are endemic to the Old World (Chao, 1986). External morphological features exhibited by the family are diverse, especially in body form and mouth position. Structures related to sound production, gas bladder and otolith structures are also markedly diverse, which is a distinctive feature of this family. Although sciaenid fishes have attracted a great deal of attention from biologists, who have focused on systematics, evolution and zoogeographics, interrelationships of most Old World Sciaenidae genera and the origin of Sciaenidae have remained elusive for a long time. Morphological disparity among the Old World sciaenid fishes has resulted in numerous conflicting phylogenetic hypotheses (Figs. 1A–H). Chu et al. (1963) classified the sciaenid fishes of China and adjacent areas based on five characters including structure of gas bladder, otolith, sensory pores on snout, lower jaw, and dentition. In their taxonomy, subfamily Argyrosominae and Pseudosciaeninae were sister-groups. The former is composed of the genera Argyrosomus and Nibea, and the latter of the genera Atrobucca, Miichthys, Collichthys and Pseudosciaena (Fig. 1A). Whether the two subfamilies are monophyletic groups or just artificial assemblages has been widely debated (Mohan, 1969; ⁎ Corresponding author. Tel./fax: +86 580 2550826. E-mail address: [email protected] (T. Xu). 0378-1119/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gene.2014.01.048

Taniguchi, 1970; Trewavas, 1962). Trewavas (1962) and Mohan (1969) pointed out that Chu et al.'s classification lacks sampling materials from the Indian Ocean. In contrast to Chu et al.'s taxonomy, Argyrosominae was considered to be polyphyletic in subsequent hypotheses based on more samples and more morphological characters. Taniguchi (1970) and Mohan (1969) suggested a close affinity between Atrobucca, Argyrosomus, and Nibea, although the interrelationships among these genera are uncertain (Figs. 1D and F). Moreover, Taniguchi (1969a, 1969b), Trewavas (1977) and Sasaki (1989) placed Argyrosomus and Nibea into separate main clades (Figs. 1B, C, G and H). Besides Chu et al.'s conclusion (1963), monophyly of Pseudosciaeninae was also recovered in the study by Taniguchi (1970) in which sciaenid fishes phylogeny was analyzed based on the feature of dentition (Fig. 1E). These results, however, are in conflict with reports by Taniguchi (1969a, 1969b), Mohan (1969), Trewavas (1977) and Sasaki (1989), in which Miichthys was not grouped with Collichthys and Pseudosciaena, but placed into different main clades (Figs. 1B, C, F, G and H). Because of a lack of clear synapomorphies, and often unclear character homology, the ability of these morphological characters to resolve phylogenetic relationships is limited. Previous molecular investigations of Old World sciaenid fishes phylogeny have relied primarily on partial mitochondrial or nuclear gene sequences, which offered controversial resolutions or support values were in general low (Figs. 1I–Q). Hypotheses based on such short sequences are conflicting and the interrelationship of many genera is still uncertain (Figs. 1I–Q). Deep nodes in the constructed phylogenetic trees were generally weak, suggesting that short sequences might not be sufficient to solve a particularly difficult phylogenetic problem. It has been shown that the use of complete mitochondrial genome

92

T. Xu et al. / Gene 539 (2014) 91–98

sequences minus non-coding sequences is more powerful than single gene sequences. Moreover, complete mitochondrial genome is a small-scale genome suitable for complete sequencing and thus provides

substantial amount of DNA and amino acid data for phylogeny reconstruction. Several phylogenetic studies have significantly advanced our understanding of fish evolution by using sequences derived from

Fig. 1. Alternative phylogenetic hypotheses of the sciaenid fishes. A) Morphology-based hypothesis of Chu et al. (1963). B–D) Morphology-based hypotheses of Taniguchi (1969a, 1969b, 1970). E) Morphology-based hypothesis of Tanguichi in, 1970. F) Morphology-based hypothesis of Mohan (1969). G) Morphology-based hypothesis of Trewavas (1977). H) Morphologybased hypothesis of Sasaki (1989). I-M) Molecular -based hypotheses of Chen (2007), Meng et al (2004) and Tong et al (2007). N, O) Molecular-based hypotheses of Liu et al. (2010). P) Molecular-based hypothesis of Cheng et al. (2010). Q) Molecular-based hypothesis of Ma et al. (2012). Species used in the present study were colored as red (in morphological data-based trees) or blue (in molecular data-based trees).

T. Xu et al. / Gene 539 (2014) 91–98

complete mitochondrial genome data (Inoue et al., 2010; Nakatani et al., 2011; Yamanoue et al., 2011). The family Sciaenidae is a strongly provincialized group of fish, with nearly all extant genera confined to a specific bioprovince (Sasaki, 1989). Distribution of extant sciaenid fishes has revealed the greatest diversity and highest number of species in two primary regions of the world's ocean, the Tropical America (which belong to the New World) and Indio-West Pacific (which belongs to the Old World). Each region may represent a major evolution center for origin and radiation of sciaenid fishes, and such distribution of extant sciaenid fishes is generally thought to aid evolutionary analysis. There are several hypotheses regarding the origin and evolution of sciaenid fishes. Chu et al. (1963) analyzed the forms of gas bladder and otolith of sciaenid fishes and hypothesized that sciaenid fishes of China may be derived from a “primitive” species (Johinus) in the Indian Ocean during the Tertiary or Quaternary Period. Chao (1986) proposed a zoogeographical hypothesis to explain the distribution of extant sciaenid fishes. A common ancestral group of sciaenid fishes, Sciaenid-like, occupied the shallow Tethys Sea between Laurasia and Gondwana. In the Late Jurassic Period, two “secondary ancestral stocks” evolved, one to the west, pre-Atlantic, and the other to the east, pre-Indio-Pacific (Fig. 2A). Sasaki (1989) drew a number of different zoogeographic inferences from the distribution of monophyletic groups. According to Sasaki's hypothesis, the common ancestor of sciaenid fishes had a Tropical American origin, with subsequent eastward dispersal of ancestors of the Indio-West-Pacific groups (Fig. 2B). Time tree analyses are useful to explain the evolutionary history of species (Zhang and Wake, 2009). However, the sciaenid fishes fossils are very rare, and it is difficult to establish convincing evolutionary pattern from ancient sciaenid fishes to modern ones based only on the existing fossil records of highly specialized Sciaenidae species. Thus the question of when and where sciaenid fishes originated and how such two patterns of evolutionary distribution centers were formed in extant sciaenid fishes remains largely unanswered. Using time information obtained from molecular data can advance our knowledge of sciaenid fishes' evolution when fossil records are not sufficient. Compared to other vertebrate groups, estimations of divergence times of sciaenid fishes have rarely been done, and more efforts should be made to generate new data and viewpoints. To infer a time frame of sciaenid fishes evolution and divergence, and to test previously proposed hypotheses on the origin of sciaenid fishes, a robust phylogeny of this group, especially one including geographically disparate taxa, must be generated. Our study was conducted to obtain molecular evidence that may provide novel insights into the origin and evolution of sciaenid fishes.

93

Accordingly, the study had 3 major goals: (1) to reconstruct a robust phylogeny of Old World sciaenid fishes based on complete mitochondrial genome sequences, and compare it to previous hypotheses by using likelihood tests, (2) to reconstruct a convincing phylogeny of sciaenid fishes (including both New World and Old World taxa) based on a combination of three mitochondrial genes (ATP8, ATP6 and Cytb), and (3) to estimate divergence time of sciaenid fishes with the aim of determining to clear the time scale of evolution of this group. Present-day and reconstructed ancestral distributions were used to trace the historical biogeography of sciaenid fishes and to determine a likely center of origin. 2. Methods 2.1. Taxonomic sampling and annotating the mitochondrial genome Seven Sciaenidae fishes represented one major lineage of Old World sciaenids were sampled and sequenced the mitochondrial genomes (Table S1 for supporting information). Total genomic DNA was extracted from fins samples DNA isolation and amplification strategies were conducted as detailed in the previously methods (Miya et al., 2003). The PCR mixture consisted of 0.2 μM of primers, 0.2 mM dNTPS, 1 μl of DNA template, 2 unit of Taq Plus DNA polymerase (TIANGEN, Beijing), and 5.0 μl of 10× Taq Plus polymerase buffer. PCRs were performed on a PTC-200. The effective fish-universal primers (Miya and Nishida, 1999) and species-specific primers were used to amplify the fragments. The sequenced fragments were confirmed on BLAST searches against published mitochondrial genomes. The tRNAscan-SE software and online DOGMA program was used for tRNA annotation (Lowe and Eddy, 1997; Tamura et al., 2007; Wyman et al., 2004). Finally, the obtained mitochondrial genome sequences were deposited in GenBank. Additionally, three mitochondrial gene sequences (ATP6, ATP8 and Cytb) for New World sciaenid fishes were downloaded from GenBank database. The details are listed in supplementary Table S1. 2.2. Phylogenetic analysis and alternative topology test Mitogenome sequences from 15 species (5 species were used as outgroups) were subjected to multiple alignment using MAFFT (Katoh and Toh, 2008). Excluding ambiguously aligned or highly diverged regions of the alignment can make phylogenetic analyses more reliable prior to tree construction. We therefore identified and trimmed such regions for alignment with Gblocks using a “more stringent selection” setting (Castresana, 2000). Additionally, the resulting gaps in aligned sequences were removed manually. The final data (designated as data

Fig. 2. The sketch map of hypotheses of A. Chao (1986) and B. Sasaki (1989). The rough continental maps from the late Jurassic (A) and early Eocene (B) with potential regions of origin for sciaenid fishes indicated by red shadows, and the potential directions of dispersal of ancestral sciaenid fishes are indicated by red arrows.

94

T. Xu et al. / Gene 539 (2014) 91–98

To identify the ancestral biogeographic range of sciaenid fishes, we first reconstructed a phylogenetic tree of sciaenid fishes representing the New World and Old World species based on three mitochondrial genes (ATP8, ATP6 and partial Cytb genes, total length 1508 bp, designated as data set 2). Inference of ancestral distributions was made using Mesquite software packages (Maddison and Maddison, 2010) by likelihood-based analysis on the phylogenetic tree of sciaenid fishes. Each taxon of sciaenid fishes was assigned to one of two distributional regions: New World (Character coded as 0) and Old World (Character coded as 1). The regions were chosen based on the known range of each taxon.

values (BPs) or Bayesian posterior probabilities (BPPs), and no difference was found in those topologies (Fig. 3). Based on this data, Sciaenidae is composed of three major lineages (Fig. 3, Clade 1–3) that are supported by 100% BPs and 1.0 BPPs. Within Sciaenidae, Argyrosomous was placed at the most basal position forming Clade 1, which is sister to the remaining sciaenid fishes. The Argyrosominae subfamily proposed by Chu et al. (1963) is not supported by the present study. Pennahia, Dendrophysa, Nibea and Chrysochir were confidently recovered as monophyletic groups (BPs = 100%, BPPs = 1; Clade 2). Within this group, Pennahia was placed at the most basal position, then, Dendrophysa branched before Chrysochir and Nibea. Chrysochir was recovered as the sister group of Nibea. In Clade 3, Miichthys formed an independent basal sub-clade that was clustered with Larimichthys and Collichthys. Collichthys niveatus was found to be grouped with Larimichthys polyactis instead of its congeneric species Collichthys lucidus. Using the 12nRTn dataset, we performed likelihood-based tests on a total number of 16 competing tree topologies. The results of the likelihood-based tests are shown in Table 1. Within the 16 competing trees, tree5 proposed by Taniguchi (1970) was found to be the second best one (P values for KH, SH and AU tests are greater than 0.05). All other tree topologies, which were proposed based on morphological or molecular data, were significantly rejected by the three tests (P b 0.01). When comparing the accepted alternative topology and rejected alternative topologies with the topology of this study, we found that the reason why Tanguichi's topology (Taniguchi, 1970) was accepted may be the arrangement of five genera (Pseudosciaena, Collichthys, Miichthys, Nibea and Argyrosomus) in this topology, in which Pseudosciaena, Collichthys and Miichthys formed one group, and Nibea and Argyrosomus formed a different group. This phylogenetic relationship is very similar to the basal relationship in the best tree of the present study, although the relationship among genera within the two groups is uncertain. However, the rejected topologies all have very different arrangements of genera from those of the best tree (Fig. 3), which may have led to the rejection of these topologies by the three tests. These results confirm the robustness of the phylogenetic relationships among the Old World sciaenid fishes.

2.4. Divergence time estimation

3.2. Ancestral distribution analyses

Bayesian molecular dating was carried out by using the MCMCTree program in the PAML 4.6 software package (Yang, 2007). The tree inferred from the three mitochondrial genes was used as the reference topology. Bayesian molecular dating was performed with an ingroup root of 192 Mya, which refers to the split of Tetraodontiformes + Perciformes and Gasterosteiformes + Scorpaeniformes (Yamanoue et al., 2006). The independent rates (IR) model, which has been considered more appropriate for divergence time estimation than correlated rates model in recent studies (Zhong et al., 2009), was used to specify the prior of rates among internal nodes. We used 126–235 Mya as constraint points for the Gasterosteiformes + Scorpaeniformes separation, which were estimated based on fossil and molecular evidences (Yamanoue et al., 2006). Loose bound for the root was set at b500 Mya. To diagnose possible failure of the Markov chains to reach stationarity, at least two MCMC analyses were performed with two different seeds. In each analysis, approximation with a burn-in period of 50,000 cycles was obtained, and every 50 cycles was taken to create a total of 10,000 samples. Similar results were observed in two runs.

The consensus trees generated by ML and Bayesian analysis of data set 2 were nearly identical in topology, so only the ML trees are presented here (Fig. 4) with both bootstrap and posterior probability support indices. In this analysis, all sciaenid fishes were recovered as three wellsupported clades (Clade A, B and C in Fig. 4) and the New World species Sciaenops ocellatus (Clade A) was placed at the basal position. Clade B included six species from the New World and Clade C included species from the New World and the Old World. Genera and species (Clade A and Clade B) placed in the basal positions of this tree were restricted to the New World, suggesting that a New World origin of sciaenid fishes may be possible. The ML reconstruction for ancestral distribution (New World = character state 1; Old World = character 0) of sciaenid fishes is shown in Fig. 5. The ancestral distribution of the most common ancestor for these sciaenid fishes is most likely the New World (P1 N 0.7). This ancestral distribution is found in all three clades in this tree and is also dominant in the ancestral nodes (Fig. 5). Furthermore, an obvious ancestral distribution shifting along the ancestral nodes (C, E, F, G and I) was detected by this analysis (Fig. 5). According to these results, Old World sciaenid fishes may be derived from ancestors dispersing from the New World, not vice versa.

set 1) consisted of 7560 positions from the 13 protein-coding genes, 2368 positions from the two rRNA genes, and 1482 positions from the 22 tRNA genes (name of the data matrix was designated as 12nRTn, where 1, 2, R and T represent 1st codon position, 2nd codon position, rRNA gene and tRNA gene, respectively, and the subscript “n” denotes nucleotide). Third codon positions were excluded to reduce substitutional saturations. Unambiguous aligned sequences were divided into four partitions depending on the data set 12nRTn. Best-fit evolutionary models were chosen for each partition with jModeltest under AIC and BIC (Posada, 2008). RAxML ver. 7.2.8 and MrBayes 3.2 were used to conduct data-partitioning-based tree reconstructions (Huelsenbeck and Ronquist, 2001; Stamatakis, 2006). The constrained tree topologies with reference to the alternative hypotheses (Fig. 1) were created manually and then subjected to Tree-Puzzle to perform the Kishino– Haseawa (KH) and Shimodaira–Hasegawa (SH) tests. The site-wise log likelihood scores were computed for each tree (Schmidt et al, 2002). To calculate the statistical significance of the differences in likelihood, the output file containing the site-wise log likelihood score was subject to CONSEL (Shimodaira and Hasegawa, 2001). Probabilities of alternative phylogenetic hypotheses were calculated using the likelihood-based approximately unbiased (AU) test as implemented in CONSEL (Shimodaira, 2002). 2.3. Tracing ancestral biogeographic ranges

3. Results 3.1. Phylogenetic relationships and likelihood tests Partitioned Bayesian and maximum likelihood analyses based on 15 whole mitochondrial genome sequences resulted in a well-resolved tree, with nearly all the internal branches supported by high bootstrap

3.3. Divergence time estimation of New World and Old World sciaenid fishes All estimated dates of divergence for nodes labeled from A to Y in Fig. 5 are indicated in Supplementary Table S1. Bayesian relaxed clock

T. Xu et al. / Gene 539 (2014) 91–98

95

Fig. 3. Phylogenetic trees of sciaenid fishes. Phylogenetic trees of Old sciaenids constructed by Bayesian inference and Maximum likelihood methods based on the mitogenomic data (data set 1). Oryzias latipes, Beryx splendens, Helicolenus hilgendorfi, Gasterosteus aculeatus and Mola mola were used as outgroups. Numbers near internal branches indicate the Bayesian posterior probabilities and maximum likelihood bootstrap support values, respectively. The three main clades of sciaenid fishes were marked as blue (Clade 1), green (Clade 2), and red (Clade 3).

analyses indicated that the common ancestor of the 23 sciaenid fishes occurred 208 million years ago (Mya), in the early Cretaceous Period. This is much older than the oldest known fossil records of this family (Berg, 1958), but is close to that proposed by Chao (1986), who suggested the occurrence of common ancestor of sciaenid fishes to be in the late Jurassic Period. The origin of Old World sciaenid fishes occurred 126 million years ago, in the middle Cretaceous Period, which is much earlier than previous estimates that were based primarily on the form of gas bladder and otolith (Chu et al., 1963). In addition, shifting of ancestral distribution occurred during the early to middle Cretaceous (Fig. 5). 4. Discussion 4.1. Phylogeny of sciaenid fishes Sciaenid fishes have long attracted biologists across diverse fields, such as morphology, toxicology, and molecular biology, and extensive efforts have been made to investigate sciaenid fishes phylogeny. Yet, the highly specialized morphological characters of sciaenid fishes make it difficult to find unambiguous clues to their phylogeny. In

addition, phylogeny reconstruction based on DNA sequences from a single gene or a short gene segment may introduce great systematic bias and/or long-branch attractions (Felsenstein, 1978; Nikaido et al., 1999), and relationships within this group remain unclear. The phylogenetic analyses conducted in this study were not biased because many genes were included in the mitogenomic data matrix, totaling over 10,000 bp in sequence length. The result demonstrated well-resolved relationships with full Bayesian posterior probability or bootstrap support regardless of the phylogenetic method used (Fig. 3). Compared to phylogenetic trees shown in Fig. 3, the tree of New World and Old World sciaenid fishes is less robust because the data matrix is small. However, the topologies produced by different methods are nearly identical, which could be a circumstantial evidence that verifies the phylogenetic relationship of these sciaenid fishes (Fig. 4). Chu et al. (1963) erected Pseudosciaeniae, which includes the genera Miichthys, Collichthys, Larimichthys and Atrobucca. Other morphology-based classification, in contrast, rejects the monophyly of Pseudosciaeniae because the genus Miichthys is grouped with non-Pseudosciaeniae taxa (Fig. 1). However, previous molecular data-based phylogenies, except that proposed by Liu et al. (2010), support the monophyly of Pseudosciaeniae, and topologies constructed in the present study

Table 1 Results of likelihood-based tests of competing topologies. Code Number

Tree topologies (Hypotheses)

lnL

Diff-lnL

KH

SH

AU

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Best tree (BI\ML); Fig. 3 Chu et al (1963) and Taniguchi (1970); Fig. 1—Tree A and Tree D Mohan (1969); Fig. 1—Tree F Sasaki (1989); Fig. 1— Tree H Taniguchi (1970); Fig. 1—Tree E Taniguchi (1969a); Fig. 1—Tree B Taniguchi (1969b); Fig. 1—Tree C Trewavas (1977); Fig. 1—Tree G Chen (2007); Fig. 1—Tree I Chen (2007); Fig. 1—Tree J Chen (2007); Fig. 1—Tree K Chen (2007); Fig. 1—Tree L Chen (2007), Meng et al. (2004) and Tong et al (2007); Fig. 1—Tree M Liu et al (2010); Fig. 1—Tree N Liu et al (2010); Fig. 1—Tree O Cheng et al (2010); Fig. 1—Tree P Ma et al (2012); Fig. 1—Tree Q

−35591.96 −37345.81 −35777.96 −37216.63 −35626.10 −36917.70 −38814.95 −36918.53 −36993.50 −36732.48 −36912.80 −36961.29 −36973.70 −36192.74 −36142.54 −35931.69 −36255.54

Best 89.48 29.88 101.49 21.93 76.01 119.03 76.90 86.44 75.59 88.46 87.23 88.37 51.03 47.46 38.71 56.43

1.00 b0.01 b0.01 b0.01 0.06 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01

1.00 b0.01 0.070 b0.01 0.73 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01

0.94 b0.01 b0.01 b0.01 0.06 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01 b0.01

96

T. Xu et al. / Gene 539 (2014) 91–98

Fig. 4. Phylogenetic trees of New World and Old World sciaenid fishes. Phylogenetic trees constructed by Bayesian inference and maximum likelihood methods based on three mitochondrial genes (ATP6, ATP8 and Cytb) (data set 2). Oryzias latipes, Beryx splendens, Helicolenus hilgendorfi, Gasterosteus aculeatus and Mola mola were used as outgroups. Numbers near internal branches indicate the ML bootstrap support values and Bayesian posterior probabilities, “–” indicates that no such topology was found in Bayesian tree. All the sciaenid fishes were grouped into three main clades (Clade A, Clade B and Clade C). Clades including the New World sciaenid fishes were colored as red, and Clades including the Old World sciaenid fishes were colored as blue.

support this phylogeny (Fig. 1). Both morphological classification and molecular classification reject the monophyly of Argyrosominae (Chu et al., 1963) because Nibea and Pennahia are not merged into one group. Thus, findings of our study support previous classifications and reject the erection of Argyrosominae. Phylogenies of the genera Pennahia, Chrysochir, Nibea and Dendrophysa based on morphological and molecular data were confidently rejected by likelihood tests (KH, SH and AU). These four genera were recovered as a monophyletic group that is a sister group of Pseudosciaeniae. Such topology was subsequently confirmed in phylogenetic analysis based on three mitochondrial genes. Clades A and B only contain New World genera and were placed at the basal position (Fig. 4). Clade C comprises of both New World and Old World sciaenid fishes. This result is comparable to the phylogenetic tree previously constructed based on morphological characters (Sasaki, 1989). Such topology may indicate that New World species placed in clades A and B were more “primitive” than Old World species and a common ancestor of New World species split to at least three lineages, and at least one lineage may have become the common ancestor of Old World sciaenid fishes. 4.2. The origin of sciaenid fishes The time for the origin of sciaenid fishes is still a controversial topic. The earliest geologic occurrence of this family is thought to be from middle to late Eocene in the Gulf region of North America and the family does not appear in European fossil records until the lower Oligocene (Breard and Stringer, 1999; Frizzell and Dante,

1965; Koken, 1888; Nolf, 2003; Nolf and Stringer, 2003). Berg (1958) also reported that the oldest known fossils of sciaenid fishes are from the Eocene epoc (40–50 Mya). Additionally, Sasaki (1989) mentioned that the oldest fossil record of a sciaenid otolith is from the Eocene Gulf. While these results seem to point to an Eocene origin of sciaenid fishes, Chao (1986) reported that the oldest fossil sciaenid is an otolith record of Otholithus bornholmiensis from Lias in Denmark (Malling and Grönwall, 1909) and this record points to a much older origin than the generally accepted late Cretaceous origin of Perciformes (Lagler et al., 1977; McAllister, 1986). Furthermore, Chao (1986) proposed a hypothesis on a late Jurassic origin of Sciaenidae. Regarding the time of origin of extant sciaenid fishes living along coaster waters of China, Chu et al. (1963) hypothesized that sciaenid fishes of China might have come from a “primitive” species (Johinus) in the Indian Ocean during Tertiary or Quaternary Period. Our estimated molecular time for the origin of sciaenid fishes is much older than those of fossil-based predictions. This new timescale indicates that sciaenid fishes originated during the early Cretaceous Period at the latest. This result is compatible with Chao's late Jurassic origin hypothesis. Although otolith-based fossil records of sciaenid fishes are rich in Eocene Gulf (Takeuchi and Huddleston, 2008), we think this only means that there were large populations of many sciaenid species during the Eocene epoch in the Gulf region of North America. Moreover, some otolith-based fossil records found in Eocene Gulf represent highly specialized species (such as Pogonias and Totoaba), not the common ancestral species of this family. Taking these points together, we propose that the origin of sciaenid fishes occurred during the late Jurassic to early Cretaceous.

T. Xu et al. / Gene 539 (2014) 91–98

97

Fig. 5. Timetree derived from the Bayesian relaxed-molecular clock method and Ancestral distributions assigned by maximum likelihood analysis. Horizontal bars indicate 95% credible intervals of the divergence time estimates. Internal nodes were coded with letters A to Y. ML reconstructions of ancestral habitats are indicated on selected nodes with pie charts showing the likelihoods for two character states (red, New World; blue, Old World).

4.3. Origin and biogeographic distribution of sciaenid fishes The region of origin of sciaenid fishes remains a controversial topic. Chao (1986) hypothesized that the shallow Tethy Sea between Laurasia and Gondwanaland is the place where a common ancestral group of sciaenid fishes occurred, and two “secondary ancestral stocks” evolved, one to the west, pre-Atlantic, and the other to the East, pre-Indio-Pacific, from that region. In fact, this hypothesis matches the vicariance model, in which vicariance is proposed to have led to the splitting and subsequent independent evolution of a common ancestral group of sciaenid fishes. If Chao's hypothesis is true, two main clades, of which one contains the New World species and the other contains the Old World species in the phylogenetic tree should be expected. However, a clade containing New World species was found to be first grouped with Old World species in phylogenetic tree, and other clades containing only New World species were found at the basal position of the tree. This

evidence suggests that the New World may be the region where sciaenid fishes originated, and dispersal, rather than vicariance, shaped the two evolutionary center distribution patterns of sciaenid fishes. In contrast Chao's hypothesis, Sasaki (1989) noted that 14 extant sciaenid genera possess an Amphi-American distribution and the Sciaenidae family possibly originated in a restricted area of the New World. Furthermore, the North American otolith-based fossil records of the Sciaenidae strongly support the hypothesis that Tropical America was the major evolutionary center for this family. Our new distribution reconstructing analysis (Fig. 5) supports Sasaki's hypothesis that sciaenid fishes originated in the New World, and also indicated that the common ancestors of Old World sciaenid fishes may be derived from species of the New World. We were unable to collect more Old World sciaenid samples because these species are on the brink of extinction. However, we believe that the molecular phylogenetic analysis and biogeographic reconstruction

98

T. Xu et al. / Gene 539 (2014) 91–98

of sciaenid fishes described in this study will provide important new insights into the evolution of this morphologically diverse group. We appeal for a Worldwide investigation using genetic resource on these endangered sciaenid fishes. Conflict of interest The authors report no conflicts of interest. Acknowledgments This study was supported by the National Natural Science Foundation of China (31272661) and the Zhejiang Provincial Natural Science Foundation (LY13C040001). We thank Prof. Shenglong Zhao for sampling and morphologically identifying. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.gene.2014.01.048. References Berg, L.S., 1958. System de rezenten und fossilen Fischaratigen und Fische. Veb. Deutscher Verlag der Wissenschaften, Berlin. Breard, S.Q., Stringer, G.L., 1999. Integrated paleoecology and marine vertebrate fauna of the Stone City Formation (middle Eocene). Gulf. Coast Assoc. Geol. Soc. 49, 132–143. Castresana, J., 2000. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552. Chao, N.L., 1986. A synopsis on zoogeography of the Sciaenidae. Indo-Pacific fish biology: proceedings of the Second International Conference on Indo-Pacific Fishes. In: Uyeno, T., Arai, R., Taniuchi, T., Matsuura, K. (Eds.), Ichthyol. Soci. Jap. Tokyo, pp. 570–589. Chen, Q.M., 2007. Molecular Phylogeny of the Sciaenidae in China. Jinan University, GuangZhou. Cheng, Y.Z., Xu, T.J., Shi, G., Wang, R.X., 2010. Complete mitochondrial genome of the miiuy croaker Miichthys miiuy (Perciformes, Sciaenidae) with phylogenetic consideration. Mar. Genomics 3, 201–209. Chu, Y.T., Lo, Y.L., Wu, H.L., 1963. A Study on the Classification of the Sciaenid Fishes of China, with Description of New Genera and Species. Science Technology Press. Felsenstein, J., 1978. Cases in which parsimony or compatibility methods will be positively misleading. Syst. Biol. 27, 401–410. Frizzell, D., Dante, J., 1965. Otoliths of some early Cenozoic fishes of the Gulf Coast. J. Paleontol. 39, 687–718. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. Inoue, J.G., et al., 2010. Evolutionary origin and phylogeny of the modern holocephalans (Chondrichthyes: Chimaeriformes): a mitogenomic perspective. Mol. Biol. Evol. 27, 2576–2586. Katoh, K., Toh, H., 2008. Recent developments in the MAFFT multiple sequence alignment program. Brief. Bioinform. 9, 286–298. Koken, E., 1888. Neue untersuchungen an Tertiären fischotolithen. Z. deut. Geol. Ges. 40, 274–305. Lagler, K.F., Bardach, J.E., Miller, R.R., 1977. Ichthyology, 2nd ed. John Wiley and Sons, New York 506. Liu, S.F., Chen, L.L., Dai, F.Q., Zhuang, Z.M., 2010. Application of DNA barcoding gene CO1 for classifying family Sciaenidae. Oceanol. Limnol. Sin. 2, 10. Lowe, T.M., Eddy, S.R., 1997. TRNAscan-SE: a program for improved detection of transfer RNA genes in genome sequence. Nucleic Acids Res. 25, 955–964. Ma, C.Y., Ma, L.B., Ni, Y., Shen, A.L., Zhang, Y., 2012. Molecular phylogenetic relationships of 13 Sciaenidae species in the China Sea based on RAG1 gene sequences. J. Fish. China 36, 9–16.

Maddison, W.P., Maddison, D.R., 2010. Mesquite: a Modular System for Evolutionary Analysis. Version 2.73 Arizona Board of Regents on Behalf of the University of Arizona, Tuscon, Arizona, U.S.A. Malling, C., Grönwall, K.A., 1909. En fauna in Bornholm Lias. Medd. Dan. Geol. Foren. 15, 271–316. McAllister, D.E., 1986. Evolution of branchiostegals and classification of teleostome fishes. Nat. Mus. Can. Bull. 221–239. Meng, Z.N., Zhuang, Z.P., Ding, S.X., Jin, X.S., Su, Y.Q., Tang, Q.S., 2004. Molecular phylogeny of eight Sciaenid species (Perciformes, Sciaenidae) in the China Sea based on mitochondrial 16S rRNA sequence. Prog. Nat. Sci. 14, 514–521. Miya, M., Nishida, M., 1999. Organization of the mitochondrial genome of a deep-Sea fish, Gonostoma gracile (Teleostei: Stomiiformes): first example of transfer RNA gene rearrangements in bony fishes. Mar. Biol. 1, 416–426. Miya, M., et al., 2003. Major patterns of higher teleostean phylogenies: a new perspective based on 100 complete mitochondrial DNA sequences. Mol. Phylogenet. Evol. 26, 121–138. Mohan, R.S.L., 1969. A synopsis of the Indian genera of the fishes of the family Sciaenidae. Indian J. Fish 16, 82–98. Nakatani, M., Miya, M., Mabuchi, K., Saitoh, K., Nishida, M., 2011. Evolutionary history of Otophysi (Teleostei), a major clade of the modern freshwater fishes: Pangaean origin and Mesozoic radiation. BMC Evol. Biol. 11, 177. Nikaido, M., Rooney, A.P., Okada, N., 1999. Phylogenetic relationships among cetartiodactyls based on insertions of short and long interpersed elements: hippopotamuses are the closest extant relatives of whales. Proc. Natl. Acad. Sci. 96, 10261–10266. Nolf, D., 2003. Revision of the American otolith-based fish species described by Koken in 1888. Louvain Geol. Surv. 12, 1–19. Nolf, D., Stringer, G.L., 2003. Late Eocene (Priabonian) fish otoliths from the Yazoo Clay at Copenhagen, Louisiana. Louvain Geol. Surv. 13, 1–23. Posada, D., 2008. JModelTest: phylogeneticmodel averaging. Mol. Biol. Evol. 25, 1253–1256. Sasaki, K., 1989. Phylogeny of the family Sciaenidae, with notes on its zoogeography (Teleostei: Perciformes). Mem. Fac. Fish. Hokkaido Univ. 36, 1–137. Schmidt, H.A., Strimmer, K., Vingron, M., Haeseler, A.V., 2002. Tree-Puzzle: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18, 502–504. Shimodaira, H., 2002. An approximately unbiased test of phylogenetic tree selection. Syst. Biol. 51, 492–508. Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17, 1246–1247. Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–2690. Takeuchi, G.T., Huddleston, R.W., 2008. A new early miocene species of Pogonias (Teleostei: Sciaenidae) based on otoliths from California. Bull. South. Calif. Acad. Sci. 107, 68–80. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596–1599. Taniguchi, N., 1969a. Comparative osteology of the sciaenid fishes from Japan and its adjacent waters—I Neurocranium. Jpn. J. Ichthyol. 16, 55–67. Taniguchi, N., 1969b. Comparative osteology of the sciaenid fishes from Japan and its adjacent waters—II. Vertebrae. Jpn. J. Ichthyol. 16, 153–156. Taniguchi, N., 1970. Comparative osteology of the sciaenid fishes from Japan and its adjacent waters—III. Premaxillary and dentary. Jpn. J. Ichthyol. 17, 135–140. Tong, X., et al., 2007. Sequence analysis of mitochondrial 16S rRNA gene fragment in Chups croaker (Nibea coibor). Mar. Fish. Res. 28, 85–91. Trewavas, E., 1962. A basis for classifying the sciaenid fishes of tropical West Africa. Ann. Mag. Nat. Hist. 5, 167–176. Trewavas, E., 1977. The sciaenid fishes (croakers or drums) of the Indo-West-Pacific. Trans. Zool. Soc. London 33, 253–541. Wyman, S.K., Jansen, R.K., Boore, J.L., 2004. Automatic annotation of organellar genomes with DOGMA. Bioinformatics 20, 3252–3255. Yamanoue, Y., Miya, M., Inoue, J.G., 2006. The mitochondrial genome of spotted green pufferfish Tetraodon nigroviridis (Teleostei: Tetraodontiformes) and divergence time estimation among model organisms in fishes. Gen. Genet. Syst. 81, 29–39. Yamanoue, Y., Miya, M., Doi, H., Mabuchi, K., Sakai, H., Nishida, M., 2011. Multiple invasions into freshwater by Pufferfishes (Teleostei: Tetraodontidae): a mitogenomic perspective. PLoS ONE 6, e17410. Yang, Z., 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol. Biol. Evol. 24, 1586–1591. Zhang, P., Wake, M.H., 2009. A mitogenomic perspective on the phylogeny and biogeography of living caecilians (Amphibia: Gymnophiona). Mol. Phylogenet. Evol. 53, 479–491. Zhong, B., Yonezawa, T., Zhong, Y., Hasagwa, M., 2009. Episodic evolution and adaptation of chloroplast genomes in ancestral grasses. PLoS One 4, e5297.