Confirmation through Genetic Analysis of the Existence of Many Local Phyloclades of the Genus Simocephalus (Crustacea, Cladocera) in China

Previously, a series of Simocephalus taxa (Cladocera: Daphniidae) from China were described. Most were proposed to be junior synonyms in the last revision of the genus. Using original material from China and data from GenBank, we investigate the biodiversity and phylogeny of Simocephalus using sequences of the cytochrome c oxidase subunit I (COI) and the nuclear 18S genes. In both cases, neighbor-joining, maximum likelihood and Bayesian inference analyses led to highly congruent tree topologies. The grouping of the deeper clades agrees with the inter-generic classification of Orlova-Bienkowskaja (2001). Only the populations of S. serrulatus from Eurasia and North America seem to be closely related, and there are no other shared species between the two continents. Our study unambiguously confirms the existence of many lineages from the subgenera of Simocephalus (Echinocaudus) and Simocephalus s.str. in China, but their morphology needs to be reexamined by taking a wider range of characters (e.g., of female thoracic limbs and adult males) into consideration.


Introduction
Cladocera (Crustacea: Branchiopoda) is an important group of micro-crustaceans predominantly inhabiting continental water bodies of different, if not all, types [1]. Among the most famous peculiarities of these animals are their sexually produced diapausing eggs, which are resistant to desiccation and other unfavourable conditions and are important propagules for passive dispersal by different modes, i.e. by birds [1], [2]. Their strong ability to survive passive dispersal was one reason why cladoceran species' distributions were for a long time accepted as cosmopolitan, but since the 1970's this concept has changed radically to the so-called non-cosmopolitanism, or ''continental endemism'' [3], [4], [5], [6]. The correctness of this idea is now confirmed for some genera and species groups [6], [7], [8], [9], although the real diversity and distribution of taxa in other groups needs to be accurately studied.
Some cladocerans, such as species of the genus Simocephalus Schödler, 1858 (family Daphniidae Straus, 1820), are used as environmental indicators and ''standard'' test objects in toxicological studies [10], [11]. Representatives of this genus are very common in vegetation, the open littoral zones of ponds and lakes, the semi-static affluents of rivers and pools and puddles of various types. Based on morphological characters, Orlova-Bienkowskaja [9] recognized 20 valid species in this genus belonging to five subgenera: Simocephalus s. str., Simocephalus (Coroncephalus), Simocephalus (Acutirostratus), Simocephalus (Echinocaudus), and Simocephalus (Aquipiculus). Many of the taxa were regarded by Orlova-Bienkowskaja [9] as junior synonyms of species described earlier.
Near the end of the 20 th century, a powerful new tool for testing taxonomic hypotheses, molecular phylogenetics, became available. In cladocerans, it was mainly applied to species of different Daphnia groups [5], [19]. However, molecular phylogenetic studies were subsequently conducted for some other genera and families of the cladocerans [6], [20], [21], [22]. COI barcoding studies for the Simocephalus genus were started by Elías-Gutiérrez et al. [23]. These authors recognized eight taxa in tropical Mexico and Guatemala, including two species that are habitually similar to S. mixtus, two species habitually similar to S. exspinosus, and two species similar to S. punctatus. Then, Jeffrey et al. [24] detected six species in Arctic Canada including two different clades of ''S. cf. serrulatus'' and four clades of ''S. cf. punctatus''. Young et al. [25], in contrast, found that all of the populations from Taiwan classified as S. vetulus, S. vetuloides and S. mixtus actually belonged to a single species, which compromises the taxonomy according to Orlova-Bienkowskaja [9].
The aim of this paper was to investigate the biodiversity and phylogeny of Simocephalus in China using the sequences of cytochrome c oxidase subunit I (COI) and nuclear 18S genes. Specimens were preserved in absolute ethanol (100%) or were brought to the laboratory alive. They were initially examined using a Leica DM 6000 B Digital-Microscope (Germany) with a CTR6000 electric cabinet, Leica LAS software, and Leica DFC 495 CCD. The determination was first made by following Orlova-Bienkowskaja [9]. However, populations were then differentiated according to their morphological characters as proposed in the Chinese literature [13], [16], [17], [18]. The specimens from the studied populations were deposited in the collection of the Hydrobiology Laboratory in Hangzhou Normal University (HZNU), Zhejiang province, People's Republic of China (Table 1).
The 25-ml PCR reaction consisted of 2 ml of genomic DNA, 8.5 ml of double-distilled H 2 O, 1 ml of each primer (10 mM) and 26 Taq PCR Master Mix (12.5 ml). The thermal conditions used to amplify the COI gene included an initial denaturing step of 5 min at 94uC, 35 cycles of 30 seconds at 94uC, 45 seconds at 51uC, 50 seconds at 72uC, and a final extension of 72uC for 7 min. The thermal conditions used to amplify the 18S gene consisted of two cycles of 30 seconds at 94uC, 45 seconds at 60uC, and 45 seconds at 72uC; five cycles of 30 seconds at 93uC, 45 seconds at 55uC, and 45 seconds at 72uC; followed by 35 cycles of 30 seconds at 93uC, 30 seconds at 50uC, and 3 min at 72uC.
The PCR products were gel-purified and sequenced on an ABI 37306l sequencer using both the forward and reverse primers. The HZNU collection sequences comprised S. vetulus, S. vetuloides, S. beianensis, S. serrulatus, S. heilongjiangensis, S. himalayensis microdus, S. sibiricus, and S. himalayensis from China, S. cf. congener from Norway, S. sp. (''S. serrulatus'' in Young et al. [25]) from Hangzhou province in China, and Daphnia cf. similoides from China ( Table 1). The nucleotide sequences of the newly analysed specimens were deposited in GenBank database (Table 1).
The alignment was created using ClustalW [32] and manually edited. The nucleotide composition, conserved sites, variable sites, parsimony-informative sites, transition/transversion ratio, and average genetic distances between each pair of species were determined using MEGA 5.1 [33]. A 658-bp COI fragment and  Table 3. The genetic distances (Dxy) between groups of the genus Simocephalus based on the COI gene.

Intra-group
Inter-group   [37] was used to generate Bayesian inferences (BI). The program was run for two million generations and sampled every 100 generations, and the first 25% of all of the trees sampled before convergence was discarded as burn-in. The 50% majority rule consensus tree was generated from the remaining trees, and the posterior probability of each node was calculated as the percentage of the trees that recovered the particular node.
The NJ, ML, and BI phylogenetic analyses led to highly congruent tree topologies (Fig. 1). In all of the trees, the terminal branches represent 100% support for presumed biological species of Simocephalus. Sometimes such biological species are undistinguishable if morphological identification is used (see for example Simocephalus congener and S. cf. congener in Europe, Fig. 1A) Their number is not fully clear, as indicated by clades A2 and E, for example. Although the statistical support for the deep branches is low, the grouping of the deeper clades generally agrees with the intra-generic classification of Orlova-Bienkowskaja [9], namely for Simocephalus (Echinocaudus), S. (Coronocephalus), Simocephalus s. str., and S. (Aquipiculus).
Clade B is the Eurasian portion of Simocephalus (Echinocaudus). It contains sub-clade B1 (S. cf. congener and S. exspinosus from Europe, most likely a single taxon, and S. himalayensis) and has high posterior probability and high bootstrap support (BI/ ML/NJ, 100/95/87). The other clade, B2, contains S. congener from Europe and S. himalayensis microdus (as a subspecies whose separate status is questionable). Clade B3 consists of ''S. serrulatus'' from Taiwan. The fourth clade (B4) contains only S. sibiricus. It is important that all of the Eurasian taxa of S. (Echinocaudus) form a monophyletic group that is well-supported (BI/ML/NJ, 100/88/91, respectively) by different statistical analyses.
Clade C, the American portion of Simocephalus (Echinocaudus), contains two taxa, S. cf. exspinosus 1 and 2 from North America, with moderate support.
Clade E, S. (Coronocephalus), contains various clades of S. serrulatus from North America and Eurasia, and the number of taxa in this complex is unclear. S. (Coronocephalus) is resolved as basal to other species of Simocephalus that are distant one from another, but support for this position is relatively weak.
The genetic distances between groups which were formed by the sequences were calculated ( Table 3). The intra-group genetic distance of Simocephalus varies from 0.070 to 0.224, and the intergroup genetic distance of Simocephalus is not exceeding 0.008. The greatest genetic distance is between S. himalayensis and S. serrulatus, while the smallest is between Simocephalus cf. congener and S. himalayensis (7.0%).
The tree (Fig. 2) contains four well-supported clades (A, B, E, D) that correspond to the subgenera identified by Orlova-Bienkowskaja [9]. As in the case of COI, ''S. serrulatus'' from Young et al. [25] appears within the S. (Echinocaudus) subgenus, which confirms the misclassification of this specimen. Simocephalus s.str. (clade A, abbreviations as in the COI tree) is represented only by the subclade A1. It is a sister group of S. (Coronocephalus) (clade E), and the clade containing these two subgenera is a sister group to S. (Echinocaudus) (clade B) which contains clades B1, B2 and B3+B4, corresponding to the clades from the COI tree. S. (Aquipiculus) (clade D) is the basal-most taxon of the genus Simocephalus in this analysis with strong statistical support for this position. No representative of clade D from the COI tree is present in this tree. See the genetic distances in Table 4.
The position of clade E differs between the 18S and COI trees. In both cases statistical support of its grouping with other branches is moderate or definitively insufficient for a final verdict. Therefore we need additional studies (using other genes?) for understanding the exact position of S. (Coronocephalus) in the genus.

Discussion
Our study confirms the opinion [1] that a real diversity of the cladocerans is several times higher than is accepted now, owing to the existence of many cryptic species complexes instead of ''traditional'' taxa. Our study also supported the concept of ''continental endemism'' [6], [38]. In the case of Simocephalus, only the populations of S. serrulatus from Europe and North America seem to be closely related; there are no other species shared between the two continents. We propose that the differentiation of some clades, such as the Eurasian and North American sections of S. (Echinocaudus), most likely each took place within the continent to which they are now largely confined.
In the case of Simocephalus, the COI barcoding approach was very effective for the discrimination of cryptic species. This might be explained by the age of the genus which is known since the Mesozoic [39], [40]. Each subgenus of Simocephalus has recent taxa on different continents (except Antarctica), which could be regarded as confirmation of an ancient, possibly Mesozoic, differentiation between subgenera that occurred before the continental break up, similarly to the subgenera of Daphnia [40]. We believe that the continental endemism of Simocephalus taxa is also mainly explained by their old age. At the same time, we also found some cases of later, inter-continental, differentiation, see above.
According to the rule-of-thumb of the barcoding approach [41], two clades are considered as distinct species if the divergence between them in COI sequences is greater than 3% while lower (0.7-2.2%) values suggest recent divergence of a clade. Of course, these values seem to vary in different groups of the Daphniidae; the mutation rate is much faster in halophilic cladocerans, for example [41], [42]. However, all of the terminal branches revealed by the 3% criterion are potentially separate species, which are thereby quite numerous in China.
In many cases, appropriately naming such taxa is impossible. Due to greater, recent activity by molecular phylogeneticists in North America [23], [24], [31], the continent is simply better studied. In contrast, type localities of the majority of the ''non-Chinese'' species are located in Europe (S. vetulus, S. congener, S. exspinosus, and S. serrulatus) or Eastern Siberia (S. mixtus, S. vetuloides, and S. sibiricus). These regions have not been adequately studied genetically except in the preliminary work of Kohout et al. [29] in Central Europe.
Orlova-Bienkowskaja [9] proposed to differentiate the five subgenera within the genus Simocephalus based on the shape of the frontal part of the head, rostrum shape, ocellus shape, length of the postero-dorsal valve prominence, expression of the pre-anal angle, anal teeth on postabdomen and presence of basal or distal pecten of spines on the postabdominal claw. In our study, the COI and 18S trees support this classification. As usual, the statistical support for the deeper branches of the COI tree is insufficient to draw any conclusions [23].
The characters used by Orlova-Bienkowskaja [9] for species discrimination are less successful as noted earlier by Hann [43]. Some characters seem to be too variable and originate many times in different clades, such as: 1) Ocellus shape. This character was found to be very variable even in a single population of S. vetulus [44]. A minute ocellus appears several times in the evolution of the genus, see Fig. 1, clade A2. Table 4. The genetic distances (Dxy) between each pair of species of genus Simocehalus involved 12 nucleotide sequences include outgroups are shown based on the 18S gene.   2) Shape of the postero-dorsal valve prominence. Earlier, Young et al. [25] showed that S. vetulus, S. mixtus and S. vetuloideslike morphotypes from Taiwan belong to a single species, and the size of the postero-dorsal prominence is too variable to be used in the taxonomy of, at least, this clade. In our tree, there are two clades conforming to the diagnosis of S. cf. vetulus in clade A1. Therefore, the shape of the postero-dorsal prominence does not work well for species determination. 3) Size and number of spines in the basal pecten on the postabdominal claw. According to the species determination scheme of Orlova-Bienkowskaja [9] the main differences between S. exspinosus and S. congener concern the anal teeth and the basal pecten of the spines on the postabdominal claw. Simocephalus exspinosus has 12 to 22 teeth while Simocephalus congener bears 9 to 18 teeth, according to Orlova-Bienkowskaja [9] and 7 to 9 in our material. The former has 8 to 12 moderately-sized postabdominal spines while the latter has 20 to 25 fine spines or 18 according to Orlova-Bienkowskaja [9] and 17 to 20 in our material from Norway. Earlier, Hann [43], based both on morphology and the electrophoretic analysis of allozymes, proposed that there are ''S. exspinosus'' and ''S. congener'' hybrids in Canada. In addition, the spectra of variability seem to overlap. Therefore, the significance of the size and number of the spines in the basal pecten must be regarded as unknown to date. In our tree, S. exspinosus and S. cf. congener from Europe look to belong to a single taxon (clade B1) in contrast to other, morphologically similar, forms, such as S. himalayensis microdus from China, S. congener from Europe (clade B2), and others. There are even two congener-like taxa in Europe, clades B1 and B2.
Orlova-Bienkowskaja [9] proposed that S. sibiricus and S. himalayensis are junior synonyms of S. exspinosus. In contrast, Chen et al. [14] and Shi et al. [18] found some differences among S. himalayensis, S. himalayensis microdus, and S. exspinosus. Table 5 summarizes the differences between the taxa of the Simocephalus (Echinocaudus) subgenus in China based on information from Chinese sources (see also Fig. 3). Unfortunately, most of these ''differences'' are very dubious and appear to have been proposed despite insufficient information on the variability in such characters throughout the whole Eurasian range. Characters such as presence-absence of small teeth on the anal embayment and the expression of the preanal angle of the postabdomen seem to be more promising (Fig. 3), but variability in the former and the latter must be studied. We believe that male characters could be more important for taxonomy, but they have not yet been adequately described.

Conclusion
Our study unambiguously confirmed the existence of both local and widely distributed lineages from the subgenera of S. (Echinocaudus) and Simocephalus s.str. in China. To date, their determination based on morphological characters is difficult. But it is a consequence of their inadequate study instead of morphology ''lacking resolution'' [41]. Morphology of different cladoceran taxa needs to be reexamined by taking a wider range of characters into consideration (e.g., of female thoracic limbs and of adult males). However, keeping in mind that many species were previously described using European populations as the type specimen, a new revision of the European taxa that combines molecular and morphological methods is also urgently needed.