Three Nuclear Protein-Coding Genes Corroborate a Recent

Home , Anomopoda, Ctenopoda, Entognatha, Onychopoda

Genes Genet. Syst. (2021) 96, p. 1–12 Molecular phylogeny of Branchiopoda 1 Three nuclear protein-coding genes corroborate a recent phylogenomic model of the Branchiopoda (Crustacea) and provide estimates of the divergence times of the major branchiopodan taxa

Taro Uozumi1,2, Keisuke Ishiwata1,2, Mark J. Grygier3,4, La-orsri Sanoamuang5 and Zhi-Hui Su1,2* 1Department of Biological Sciences, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan 2JT Biohistory Research Hall, Takatsuki, Osaka 569-1125, Japan 3Lake Biwa Museum, Kusatsu, Shiga 525-0001, Japan 4Center of Excellence for the Oceans, National Taiwan Ocean University, Keelung 202301, Taiwan 5Applied Taxonomic Research Center and International College, Khon Kaen University, Khon Kaen 40002, Thailand

(Received 24 August 2020, accepted 18 October 2020; J-STAGE Advance published date: 14 March 2021)

The class Branchiopoda (Crustacea) shows great diversity in morphology and lifestyle among its constituent higher-level taxa: Anostraca, Notostraca, Laevicau- data, Spinicaudata, Cyclestherida and Cladocera. The phylogenetic relationships among these taxa have long been controversial. We sequenced three orthologous nuclear genes that encode the catalytic subunit of DNA polymerase delta and the largest and second-largest subunits of RNA polymerase II in the expectation that the amino acid sequences encoded by these genes might be effective in clarify- ing branchiopod phylogeny and estimating the times of divergence of the major branchiopodan taxa. The results of phylogenetic analyses based on these amino acid sequences support the monophyly of Branchiopoda and provide strong molecular evidence in support of the following phylogenetic relationships: (Anostraca, (Notostraca, (Laevicaudata, (Spinicaudata, (Cyclestherida, Cladocera))))). Within Cladocera, comparison of the nucleotide sequences of these same genes shows Ctenopoda to be the sister group of Haplopoda + Anomopoda. Three statistical tests based on the present amino acid sequence data—the approximately unbiased test, Kishino–Hasegawa test and weighted Shimodaira–Hasegawa test—tend to refute most of the previous molecular phylogenetic studies on Branchiopoda, which have placed Notostraca differently than here; however, our results corroborate those of one recent phylogenomic study, thus confirming the effectiveness of these three genes to investigate relationships among branchiopod higher taxa. Diver- gence time estimates calibrated on the basis of fossil evidence suggest that the first divergence of extant branchiopods occurred about 534 Ma during the early Cambrian period and that diversification within the extant branchiopod lineages started in or after the late Permian.

Key words: Branchiopoda, molecular phylogeny and evolution, chronology of cladogenesis, DNA polymerase gene, RNA polymerase gene

Edited by Junko Kusumi INTRODUCTION * Corresponding author. [email protected] DOI: https://doi.org/10.1266/ggs.20-00046 The class Branchiopoda (Crustacea) is a relatively Copyright: ©2021 The Author(s). This is small, primarily freshwater taxon comprising about an open access article distributed under 1,120 described species (Adamowicz and Purvis, 2005; the terms of the Creative Commons BY 4.0 Brendonck et al., 2008; Forró et al., 2008; Ahyong et al., International (Attribution) License (https://creativecommons.org/ licenses/by/4.0/legalcode), which permits the unrestricted distri- 2011). Branchiopods display a wide diversity of morphol- bution, reproduction and use of the article provided the original ogy and lifestyle (Olesen, 2009) and comprise six main source and authors are credited. groups: Anostraca (fairy shrimps); Notostraca (tadpole 2 T. UOZUMI et al. shrimps); Laevicaudata, Spinicaudata and Cyclestherida rently considered by many to be the clade most closely (clam shrimps); and Cladocera (water ﬂeas). Laevicau- related to Hexapoda (Regier et al., 2010; von Reumont data, Spinicaudata and Cyclestherida were previously et al., 2012; Oakley et al., 2013; Schwentner et al., 2018; classiﬁed together as ‘Conchostraca’, but that grouping Lozano-Fernandez et al., 2019; Noah et al., 2020), but is now recognized as paraphyletic (Olesen and Richter, Branchiopoda has also been considered to be the sis- 2013). All non-cladocerans (fairy shrimps, tadpole ter group of Hexapoda (Glenner et al., 2006; Lozano- shrimps and clam shrimps) live in inland freshwaters or Fernandez et al., 2019). For a review of this topic, see salt lakes and are generally known as “large branchio- Giribet and Edgecombe (2012). pods”. Although they share characteristics such as many In recent years, many investigations of branchiopod serially similar phyllopodous trunk limbs (Olesen, 2009), phylogeny and evolution have been conducted, based morphological and molecular analyses suggest that “large on analyses of both morphological and molecular data branchiopods” are paraphyletic (see Olesen and Richter, (Hanner and Fugate, 1997; Olesen, 1998, 2000; Negrea 2013). About half of the known species of Branchiopoda et al., 1999; Taylor et al., 1999; Spears and Abele, 2000; are cladocerans, some of which are marine. The Cladoc- Braband et al., 2002; deWaard et al., 2006; Stenderup et era are further divided into four subgroups, Anomopoda, al., 2006; Richter et al., 2007; Olesen, 2009; Schwentner Ctenopoda, Haplopoda and Onychopoda (Calman, 1909; et al., 2018; Luchetti et al., 2019). Relationships among Fryer, 1987; Walossek, 1993; Olesen et al., 1997; Olesen, the higher-level taxa, especially the phylogenetic posi- 1998, 2004; Negrea et al., 1999). tions of Notostraca and Cyclestherida and the phylogeny The systematics of Branchiopoda has recently received within Cladocera, have long been disputed (Stenderup much attention, not only because of interest in the intrin- et al., 2006; Richter et al., 2007; Olesen, 2009) (see Fig. sic diversity within the group, but also on account of bran- 1). Nonetheless, morphological analyses have succeeded chiopods’ possible importance for understanding the origin in providing a model of branchiopod sister-group relations and evolution of insects (Hexapoda). Molecular system- and phylogeny, namely (Anostraca, (Notostraca, (Laevi- atics has greatly changed our traditional understanding caudata, (Spinicaudata, (Cyclestherida, Cladocera))))) of arthropod phylogeny by revealing that Crustacea and (Richter et al., 2007; Olesen, 2009) (Fig. 1A). Hexapoda form a common clade (Pancrustacea), which As for molecular evidence, mitochondrial 12S rRNA is now widely accepted among zoologists; for a review, data supported Notostraca as the sister group to Laevi- see Giribet and Edgecombe (2019). Among crustaceans, caudata (Braband et al., 2002) (Fig. 1D), whereas maxi- either Remipedia or Remipedia + Cephalocarida is cur- mum likelihood (ML) analysis and Bayesian inference

A Haplopoda B Haplopoda Anomopoda Cladocera Anomopoda Cladocera Ctenopoda Ctenopoda Cyclestherida Cyclestherida Spinicaudata Spinicaudata Laevicaudata Notostraca Notostraca Laevicaudata Anostraca Anostraca

C Haplopoda D Haplopoda Anomopoda Cladocera Anomopoda Cladocera Ctenopoda Ctenopoda Cyclestherida Cyclestherida Notostraca Spinicaudata Spinicaudata Notostraca Laevicaudata Laevicaudata Anostraca Anostraca Fig. 1. Phylogenetic hypotheses of Branchiopoda based on morphological and molecular analyses. A, mainly based on morphological studies (Richter et al., 2007; Olesen, 2009; see also Olesen and Richter, 2013) and supported by a recent phylogenomic study (Schwentner et al., 2018); B, C and D, based on various molecular markers with different analytical methods as described in the Introduction (for B: deWaard et al., 2006; Stenderup et al., 2006; Richter et al., 2007; for C: Stenderup et al., 2006; for D: Braband et al., 2002). Molecular phylogeny of Branchiopoda 3 based on mitochondrial 16S rRNA and nuclear 28S rRNA branchiopod representing all six higher taxa to address data supported Notostraca as the sister group to Cycles- the questions raised above. Phylogenetic analyses were therida + Cladocera (Stenderup et al., 2006) (Fig. 1C) and conducted based on the inferred aa sequences of the maximum parsimony analysis supported Notostraca as encoded proteins, and previously proposed phylogenetic the sister group to Spinicaudata + Cyclestherida + Cla- hypotheses were evaluated statistically. In addition, the docera (Stenderup et al., 2006) (Fig. 1B). This last phy- divergence times of branchiopod lineages were estimated logeny (Fig. 1B) was also supported by an analysis based based on fossil calibration points including a recently dis- on six molecular loci (COI, 16S rRNA, 18S rRNA, EF-1 covered early spinicaudatan fossil (Gueriau et al., 2016). alpha, 12S rRNA and 28S rRNA) (deWaard et al., 2006), but when the same authors restricted their analysis to MATERIALS AND METHODS only three of these loci (COI, 16S rRNA and 18S rRNA), four of the large branchiopod taxa, Anostraca, Notostraca, Taxonomic sampling The nucleotide sequences Laevicaudata and Spinicaudata, were clustered into one reported here have been deposited in the DDBJ/EMBL/ clade. In the analyses of Richter et al. (2007), who used GenBank nucleotide sequence databases with accession sequence data from six loci and 65 morphological char- numbers LC532181–LC532216 as shown in Supplemen- acters, the morphological data supported a sister-group tary Table S1. relationship between Notostraca and Diplostraca (= Lae- Supplementary Table S1 lists all the taxa used in this vicaudata + Spinicaudata + Cyclestherida + Cladocera) study along with their collection localities and the respec- (Fig. 1A), while the molecular data strongly supported tive sequence accession numbers for all three nuclear Notostraca as the sister group to Spinicaudata + Cycl- protein-coding genes (DPD1, RPB1, RPB2). Fifteen estherida + Cladocera (Fig. 1B). Most recently, mito- branchiopod species (Fig. 2) representing all major groups chondrial genomic sequence data provided a topology in were sequenced for these genes, with the exception of which Notostraca, Spinicaudata and Cladocera formed Onychopoda (one of four infraorders within the subor- one group, but with weak bootstrap support and no data der Cladocera), living material of which was not avail- from Laevicaudata and Cyclestherida (Luchetti et al., able. Twenty-two other arthropod species, including 2019). Most analyses have supported a sister-group rela- two chelicerates, one myriapod, 16 hexapods and three tionship between Cyclestherida and Cladocera, but some non-branchiopod crustaceans, were used as outgroups in studies have suggested a position of the former within the the phylogenetic analysis (Supplementary Table S1). We latter (deWaard et al., 2006; Richter et al., 2007). would have liked to include representatives of the pur- Phylogenomic studies have increased in recent years portedly basal crustacean classes Cephalocarida and because it has become easier to obtain genome-wide Remipedia in both sets of outgroups, but no living mate- sequence data. In one such study, Schwentner et al. rial of these rare animals was available and none of their (2018) provided a strongly supported topology that, for relevant nuclear gene sequences has yet been archived. branchiopods, basically coincides with the morphology- based phylogenies of Richter et al. (2007) and Olesen RNA isolation, reverse transcription, PCR and (2009). However, phylogenomic datasets produced by sequencing Live specimens were used for RNA iso- different workers sometimes give contradictory results lation. Total RNA was extracted from a part of the despite showing strong support values. For example, specimen (legs and tails) for anostracans and notostra- one phylogenomic dataset shows the Cephalocarida (a cans and from the whole body for other branchiopod class of Crustacea) as sister to Remipedia (Regier et al., specimens using ISOGEN (Nippon Gene) following the 2010; Noah et al., 2020), while another shows it as sister manufacturer’s instructions. Total extracted RNA was to Branchiopoda + Remipedia + Hexapoda (Schwentner dissolved in 22 μl DNase/RNase-Free Distilled Water et al., 2018). These differences are probably due to the (Invitrogen), and 1 μg was reverse-transcribed to cDNA inclusion of inappropriate genes in the datasets. It is using the SMART RACE cDNA Ampliﬁcation Kit (Clon- important to select molecular markers that are known to tech) and SuperScript III Reverse Transcriptase (Invitro- be appropriate for phylogenetic analyses. Previous stud- gen). The resulting cDNA was then used as a template ies (Ishiwata et al., 2011; Sasaki et al., 2013; Miyazawa et to amplify each of the three genes by PCR with LA-Taq al., 2014) have shown that the amino acid (aa) sequences (Takara Bio) using sense and antisense degenerate prim- encoded by three nuclear genes, RNA polymerase II larg- ers (Supplementary Fig. S1 and Supplementary Table est subunit (RPB1), RNA polymerase II second-largest S2). The experimental procedures for PCR ampliﬁcation subunit (RPB2) and DNA polymerase delta catalytic and sequencing were according to Ishiwata et al. (2011) subunit (DPD1), are effective for resolving the relation- and Sasaki et al. (2013). ships among the higher-level taxa of Hexapoda and For Cyclestheria hislopi, total RNA (822 ng) was used to Myriapoda. Here, we used the sequences of these three construct an RNA library using the TruSeq RNA Library genes (RPB1, RPB2, DPD1) obtained from 15 species of Preparation Kit v2 (Illumina). The library was then 4 T. UOZUMI et al.

Fig. 2. Photographs of branchiopod species included in this study.

used for next-generation sequencing (NGS) on MiSeq MAFFT L-INS-i (Katoh et al., 2005) and then inspected (Illumina). Sequence assembly was performed with by eye. Unambiguously aligned aa sites were selected CLC Genomics Workbench (Filgen). Most of the nucleo- using Gblocks ver. 0.91b (Castresana, 2000; Talavera and tide sequences of the three genes used in this study were Castresana, 2007). obtained from the NGS sequence data, and the missing Selection of an appropriate outgroup often improves parts were amplified by PCR as described above. the results of phylogenetic analysis. Outgroup taxa that represent long branches may cause misplacement of Sequence alignment and phylogenetic analyses long-branched ingroup taxa (Schneider and Cannarozzi, Nucleotide sequences were corrected and edited using 2009), but the use of certain outgroups that are much 4Peaks (Netherlands Cancer Institute) and connected by more closely related to the ingroups may reduce the ATSQ ver. 5.1 (GENETYX Corporation). The consen- effects of such artifacts. With this in mind, we prepared sus sequences of the three genes were translated to the two sequence datasets for phylogenetic analysis. One corresponding encoded aa sequences with GENETYX- consisted of all 37 samples (37-sample dataset) used in MAC ver. 16.0.1 (GENETYX Corporation), aligned by this study and was intended to help confirm or refute Molecular phylogeny of Branchiopoda 5 the monophyly of Branchiopoda. The 22 outgroup taxa of the MrBayes analyses was examined by calculating the included representatives of the Chelicerata, Myriapoda, effective sample sizes of all parameters using Tracer v1.5 Hexapoda and two non-branchiopod crustacean classes; (http://tree.bio.ed.ac.uk/software/tracer/). Bayesian pos- the hexapods included ten species of wingless insect terior probabilities were obtained from the majority rule (two proturans, three diplurans, two bristletails and consensus tree sampled after the initial burn-in period. three silverfishes) and six species of winged insect (Fig. In addition, phylogenetic analyses within Anostraca, 3, Supplementary Table S1). The other dataset com- Notostraca and Cladocera were conducted using the prised 22 samples (22-sample dataset) with only seven nucleotide sequences of the three genes. Two sequence hexapod and non-branchiopod crustacean species used as datasets were constructed: one for Anostraca and Notos- outgroups. The 22-sample dataset was used to analyze traca (11,225 nucleotide sites), and the other for Cladoc- the phylogenetic relationships among branchiopods, sta- era (11,202 nucleotide sites). For both datasets, sequence tistically test the previously proposed hypotheses of their alignment was performed with MAFFT (Katoh et al., relationships and estimate divergence times. 2005) and all gaps were removed. Model testing was Phylogenetic trees were inferred by the ML method and carried out using jModelTest 2.1.5 (Guindon and Gascuel, Bayesian analysis. The ML analysis was carried out on 2003; Darriba et al., 2012), and GTR+G+I was selected RAxML 7.2.8 (Stamatakis, 2006) with the best-fit model, as the best model for the two datasets. The phylogenetic namely the LG+G model that was determined for each trees were inferred with RAxML 7.2.8 (Stamatakis, 2006) dataset using ProtTest 3 under the AIC, BIC and AICc cri- under this model, again with 1,000 bootstrap replicates. teria (Darriba et al., 2011). To assess the node supports, 1,000 nonparametric bootstrap runs were performed for Statistical tests Previous molecular phylogenetic stud- each ML analysis. Bayesian inference was performed ies have suggested three possible placements of Notos- with MCMC analysis in MrBayes v3.1.2 (Ronquist and traca (Braband et al., 2002; Stenderup et al., 2006; Richter Huelsenbeck, 2003) using the LG+G model. Each analy- et al., 2007). In this study, these hypotheses were statis- sis was run for 2,000,000 generations, and the trees were tically tested based on the 22-sample dataset using the sampled every 1,000 generations (burn-in = 500,000 gen- CONSEL program (Shimodaira and Hasegawa, 2001) erations). To test the convergence of chains, the log file with the approximately unbiased test (AU), the Kishino–

88/1.00 Cryptotympana facialis 0.05 98/1.00 Thrips palmi Teleogryllus emma 67/0.97 97/1.00 Karoophasma biedouwensis Mnais pruinosa 100/1.00 100/1.00 Epiophlebia superstes Ectognatha 100/1.00 Isolepisma japonica 100/1.00 Thermobia domestica Hexapoda 68/1.00 Nipponatelura sp. 59/0.94 Petrobiellus takunagae 100/1.00 Pedetontus unimaculatus 100/1.00 Lepidocampa weberi 94/1.00 Metriocampa sp. 100/1.00 Occasjapyx japonicus Entognatha Baculentulus morikawai 100/1.00 Nipponentomon nippon 100/1.00 Moina macrocopa 57/0.84 Daphnia pulicaria 90/1.00 Diaphanosoma sp. 47/1.00 100/1.00 Leptodora richardi 91/1.00 Cyclestheria hislopi 99/1.00 Eulimnadia braueriana 99/1.00 Leptestheria kawachiensis Branchiopoda 100/1.00 Caenestheriella gifuensis 82/0.90 Lynceus biformis Triops cancriformis 100/1.00 54/- Triops granarius 100/1.00 Triops longicaudatus 100/1.00 -/0.84 56/0.80 Artemia franciscana Branchinella kugenumaensis 100/1.00 Eubranchipus uchidai 100/1.00 Oratosquilla oratoria Procambarus clarkii Malacostraca 89/1.00 Cyclops vicinus Copepoda Thereuonema tuberculata Myriapoda Psathyropus tenuipes Chelicerata 100/1.00 Ixodes scapularis Fig. 3. Phylogenetic tree of Branchiopoda based on aa sequences of DPD1, RPB1 and RPB2 genes with 22 outgroup species (37-sample dataset). Phylogenetic analyses were performed with RAxML and MrBayes. Bootstrap values and posterior probabilities are shown at each node. 6 T. UOZUMI et al.

Hasegawa test (KH) and the weighted Shimodaira– (Supplementary Table S3). Hasegawa test (wSH) (Kishino and Hasegawa, 1989; The 37-sample dataset including numerous outgroup Shimodaira and Hasegawa, 2001; Shimodaira, 2002). taxa (22 species) yielded a total of 3,447 aa sites (DPD1, 874 aa; RPB1, 1,434 aa; RPB2, 1,139 aa) selected by Divergence time estimation An ephemeral-pool Gblocks (Castresana, 2000; Talavera and Castresana, branchiopod community from the late Famennian (372.2– 2007) as suitable for phylogenetic analyses. The 358.9 Ma) was recently discovered at the Strud locality in 22-sample dataset, including only seven outgroup spe- Belgium, and two fossil arthropods, Haltinnaias serrata cies, had a total of 3,490 aa sites (DPD1, 886 aa; RPB1, Gueriau et al., 2016 and Gesvesia pernegrei Gueriau et al., 1,467 aa; RPB2, 1,137 aa). Neither of the two datasets 2016, were described as being the earliest known unequivo- contains missing aa sites. Sequence information includ- cal representatives of the groups Anostraca and Spinicau- ing parsimony-informative and variable aa sites in the data, respectively (Gueriau et al., 2016). The minimum two alignment datasets is shown in Supplementary Table age of G. pernegrei (358.5 Ma) provides a calibration point S4. The proportion of parsimony-informative sites was for the time of divergence between Spinicaudata and Cla- highest for the DPD1 gene, and the RPB2 gene was the doceromorpha (= Cyclestherida + Cladocera). With this most highly conserved. These results were consistent and four other fossil-based calibration points (Wolfe et al., with those observed for these genes in insects (Ishiwata 2016), a timetree was constructed on MEGA X (Kumar et al., 2011; Sasaki et al., 2013) and myriapods (Miyazawa et al., 2018; Stecher et al., 2020) based on the 22-sample et al., 2014). dataset using the RelTime method (Tamura et al., 2012, 2018) and the LG model (Le and Gascuel, 2008) with Monophyly of Branchiopoda Using the 37-sample gamma distribution (+G) and evolutionarily invariable dataset including 22 outgroup species, all analytical (+I) settings. The Tao et al. (2019) method was used to methods generated the same tree topology, with the set minimum and maximum time boundaries on nodes exception of the relationships among the three species for which calibration densities were provided. According of Triops (Fig. 3). Monophyly of Branchiopoda was con- to Wolfe et al. (2016), the minimum ages of the other four ﬁrmed with robust support (bootstrap value of 100% and calibration points are 125.71 Ma for crown-group Anos- posterior probability of 1.00). Such monophyly had been traca, 162.5 Ma for crown-group Spinicaudata, 386.9 Ma indicated previously by morphological analyses including for crown-group Diplostraca and 405 Ma for crown-group larval characters (e.g., Olesen, 2007, 2009), and by some Hexapoda. For all calibration points, the maximum age molecular studies based on different molecular data and was set to 521 Ma (Wolfe et al., 2016). sampling sets (Regier et al., 2010; von Reumont et al., 2012; Schwentner et al., 2018). When chelicerates and myriapods were included as out- RESULTS AND DISCUSSION groups, Hexapoda was most closely related to Branchiop- Sequence data and alignment We determined the oda with high posterior probability but low bootstrap sequences of three nuclear protein-coding genes (DPD1, support (Fig. 3). The other two crustaceans, belonging RPB1 and RPB2) from three species of Anostraca, three to Copepoda and Malacostraca, showed a close relation- species of Notostraca, one species of Laevicaudata, three ship to each other with high support (Fig. 3), in agree- species of Spinicaudata, one species of Cyclestherida ment with the results of von Reumont et al. (2012) and and four species of Cladocera. For DPD1, the complete Schwentner et al. (2018). However, to understand the encoded aa sequence was obtained from representatives phylogeny and evolution of Crustacea, it is necessary to of four higher taxa and ranged in length from 1,047 to analyze more crustacean samples including representa- 1,120 aa, although that obtained from the species of tives of the two classes Cephalocarida and Remipedia, Notostraca and Laevicaudata lacked a short part of which were not available in this study. Despite extensive the N-terminal region (Supplementary Table S3). For genomic data-based studies, it remains unclear which taxa RPB1, the C-terminal region (approx. 400 aa) had are most closely related to Branchiopoda; Schwentner et repeated sequences that were not appropriate for phylo- al. (2018) showed that Branchiopoda is the sister group of genetic analysis, and this region was excluded from the Remipedia + Hexapoda, while Noah et al. (2020) indicated analysis, while the N-terminal sequence was determined a sister relationship between Branchiopoda and Multi- completely for ten species representing all six taxa and crustacea (Copepoda + Thecostraca + Malacostraca). In had a length ranging from 1,529 to 1,907 aa (Supple- light of the phylogenetic relationships suggested by mentary Table S3). For RPB2, the complete sequences these studies, the outgroups used in the present study obtained from nine species ranged in length from 1,174 to (Hexapoda, Copepoda and Malacostraca) can be judged 1,179 aa. The remaining six species were each missing as appropriate, if not ideal, for phylogenetic analyses of a short part of the N-terminal region and consequently Branchiopoda. had a sequence length ranging from 1,146 to 1,164 aa Molecular phylogeny of Branchiopoda 7

Phylogenetic relationships among higher-level taxa (Tree No. 2 in Fig. 5); 2) the sister group of Spinicau- of Branchiopoda Both the 37-sample and 22-sample data + Cyclestherida + Cladocera (deWaard et al., 2006; datasets resulted in the same strongly supported infer- Stenderup et al., 2006; Richter et al., 2007) (Tree No. 3 ences of phylogenetic relationships among the sequenced in Fig. 5); or 3) the sister group of Cyclestherida + Cla- branchiopods considered in this study, except for minor docera (Stenderup et al., 2006) (Tree No. 4 in Fig. 5). It differences among the compared species of Anostraca and would simplify matters if we could reject one or more Notostraca (Figs. 3 and 4). The choice of outgroups did of these options with statistical confidence. Statistical not appreciably affect the outcome. From the topology analyses using the AU, KH and wSH tests as outlined obtained, Anostraca was identified as the sister group to above, based on the 22-sample dataset, resulted in P val- all remaining branchiopods, with the latter constituting ues for all tests that were less than 0.01 (Fig. 5B). All a monophyletic Phyllopoda. Within Phyllopoda, Noto- three tree topologies were thus rejected by these tests straca branched off first, followed by Laevicaudata and at the 1% significance level. The present results provide Spinicaudata in that order, and Cyclestherida was the sis- strong molecular evidence to support Notostraca as the ter lineage to Cladocera (Figs. 3 and 4). These relation- sister group of all extant non-anostracan branchiopods, in ships were all supported by very high bootstrap values agreement with the prevailing morphology-based model of (≥ 89%) and posterior probabilities (1.00) (Fig. 4). The branchiopod phylogeny (Richter et al., 2007; Olesen, 2009) present results thus strongly support the monophyly of and also with a recent phylogenomic study (Schwentner several higher-level taxa, namely Phyllopoda, Diplos- et al., 2018). This confirms the effectiveness of using the traca, Onychocaudata, Cladoceromorpha and Cladocera aa sequences encoded by these three genes—as opposed (Fig. 4), which had mainly been proposed or corroborated to more expensive, time-consuming and technically diffi- on the basis of morphological studies (Richter et al., 2007; cult phylogenomic analyses—to investigate relationships Olesen, 2007, 2009; see also Olesen and Richter, 2013), among crustaceans, as well as among the higher taxa of and furthermore agree with the results of a recent phy- Hexapoda (Ishiwata et al., 2011; Sasaki et al., 2013) and logenomic study (Schwentner et al., 2018). Myriapoda (Miyazawa et al., 2014). In most molecular analyses conducted up to now, except for Schwentner et al. (2018), the detailed results have dif- Phylogeny of Cladocera As shown in Figs. 3 and 4, fered from ours on account of the use of different datasets the phylogenetic relationships within Anostraca, Noto- or analytical methods (Braband et al., 2002; deWaard et straca and Cladocera as revealed by the encoded aa al., 2006; Stenderup et al., 2006; Richter et al., 2007). In sequences of the three analyzed nuclear genes were not general, Notostraca has fared the worst in terms of well resolved due to low node-supporting bootstrap val- inconsistent phylogenetic position, appearing as either 1) ues. This is probably due to the lower-level taxa within the sister group of Laevicaudata (Braband et al., 2002) each group being too closely related, and thus having

100/1.00 Moina macrocopa 53/0.67 Daphnia pulicaria Cladoceromorpha 0.02 Anomopoda Cladocera 94/1.00 Diaphanosoma sp. Ctenopoda 100/1.00 Leptodora richardi Haplopoda Onychocaudata Cyclestheria hislopi Cyclestherida 89/1.00 Eulimnadia braueriana 97/1.00 Diplostraca 100/1.00 Leptestheria kawachiensis Spinicaudata 100/1.00 Caenestheriella gifuensis Phyllopoda 89/1.00 Lynceus biformis Laevicaudata -/0.99 Triops longicaudatus Triops granarius 100/1.00 100/1.00 Notostraca 52/- Triops cancriformis 81/0.95 Eubranchipus uchidai 92/1.00 Branchinella kugenumaensis Anostraca Sarsostraca 100/1.00 Artemia franciscana 100/1.00 Petrobiellus takunagae 100/1.00 Pedetontus unimaculatus Hexapoda 94/1.00 Baculentulus morikawai 100/1.00 Nipponentomon nippon Cyclops vicinus Copepoda Oratosquilla oratoria Procambarus clarkii Malacostraca Fig. 4. Phylogenetic tree of Branchiopoda inferred from aa sequences of DPD1, RPB1 and RPB2 genes, with seven outgroup species including four hexapods and three non-branchiopods (22-sample dataset). Bootstrap values and posterior probabilities are shown at each node. 8 T. UOZUMI et al.

A Tree No. 1 (ML) 1 Tree No. 2 1 Tree No. 3 1 Tree No. 4 1 2 Cladocera 2 2 2 Cladoceromorpha 3 3 3 3 4 4 4 4 Onychocaudata 5 Cyclestherida 5 5 5 6 6 6 10 Diplostraca 7 Spinicaudata 7 7 11 Phyllopoda 8 8 8 12 9 Laevicaudata 9 10 6 10 10 11 7 11 Notostraca 11 12 8 12 12 9 9 13 13 13 13 14 Anostraca 14 14 14 15 15 15 15 16 16 16 16 17 17 17 17 18 18 18 18 19 Outgroup 19 19 19 20 20 20 20 21 21 21 21 22 22 22 22

B Tree No. Topology ln L ± δ P(AU) P(KH) P(wSH) 1 (((((((((((1,2),3),4),5),((6,7),8)),9),(10,(11,12))),((13,14),15)),((16,17),(18,19))),20),(21,22)); ML 0.997 0.991 0.998 2 ((((((((((1,2),3),4),5),((6,7),8)),(9,(10,(11,12)))),((13,14),15)),((16,17),(18,19))),20),(21,22)); -38.632 ± 15.972 0.006 0.009 0.019 3 (((((((((((1,2),3),4),5),((6,7),8)),(10,(11,12))),9),((13,14),15)),((16,17),(18,19))),20),(21,22)); -43.632 ± 15.265 <0.001 0.003 0.007 4 ((((((((((1,2),3),4),5),(10,(11,12))),(((6,7),8),9)),((13,14),15)),((16,17),(18,19))),20),(21,22)); -79.639 ± 22.796 <0.001 <0.001 0.001 Taxon name: 1, Moina macrocopa; 2, Daphnia pulicaria; 3, Diaphanosoma sp.; 4, Leptodora richardi; 5, Cyclestheria hislopi; 6, Eulimnadia braueriana; 7, Leptestheria kawachiensis; 8, Caenestheriella gifuensis; 9, Lynceus biformis; 10, Triops cancriformis; 11, Triops granarius; 12, Triops longicaudatus; 13, Eubranchipus uchidai; 14, Branchinella kugenumaensis; 15, Artemia franciscana; 16, Petrobiellus takunagae; 17, Pedetontus unimaculatus; 18, Baculentulus morikawai; 19, Nipponentomon nippon; 20, Cyclops vicinus; 21, Oratosquilla oratoria; 22, Procambarus clarkii. Fig. 5. Statistical testing of hypotheses concerning the phylogenetic position of Notostraca. The AU test, KH test and wSH test were performed based on the same aa sequence dataset as in Fig. 4. A, tested tree topologies: the ML tree obtained in this study (Tree No. 1) and the three previously proposed candidate topologies (Tree Nos. 2–4). The candidate topologies were inferred by ML analysis with the same method and dataset used in Fig. 4, with Notostraca constrained into a single cluster with Laevicaudata (Tree No. 2; see Braband et al., 2002), with Onychocaudata (Tree No. 3; see deWaard et al., 2006; Stenderup et al., 2006; Richter et al., 2007) or with Cladoceromorpha (Tree No. 4; see Stenderup et al., 2006). The nodes within Notostraca, Onychocaudata (in Tree No. 3) and Cladoceromorpha (in Tree No. 4) were not ﬁxed. B, results of statistical tests for each tree topology with tree and specimen numbers corresponding to those shown in A. ΔlnL, the difference of the log-likelihood between ML tree and test tree.

accumulated too few sequence changes in these three A 100 Triops granarius genes to resolve their phylogenetic relationships. Addi- 100 Triops longicaudatus Notostraca tional analyses for these taxa using nucleotide sequences Triops cancriformis that included synonymous substitutions, and not just the 97 Eubranchipus uchidai encoded aa sequences, showed well-resolved phylogenetic Branchinella kugenumaensis Anostraca 0.1 Artemia franciscana relationships with strong bootstrap support (Fig. 6). The branching patterns of the three anostracan species and B Daphnia pulicaria Anomopoda the three species of Triops were (Artemia franciscana, 99 100 Moina macrocopa (Branchinella kugenumaensis, Eubranchipus uchidai)) Cladocera Leptodora richardi Haplopoda and (Triops cancriformis, (T. longicaudatus, T. granarius)) Diaphanosoma sp. Ctenopoda (Fig. 6A), respectively. With Cyclestherida as the out- Cyclestheria hislopi Cyclestherida group, the branching pattern of the four cladoceran spe- 0.1 cies (Fig. 6B) showed Diaphanosoma sp. (Ctenopoda) as Fig. 6. ML trees of Notostraca and Anostraca (A) and Cladocera the sister group of the remaining species (Haplopoda + (B) inferred from nucleotide sequences of DPD1, RPB1 and RPB2 Anomopoda), and the species of Anomopoda clustered genes. Phylogenetic analyses were performed with RAxML together. Despite many studies based on both morpho- (GTR+G+I model). Bootstrap values are shown at each node. logical and molecular data, the relationships among the four higher taxa of Cladocera, namely the Anomopoda, results provide partial resolution, but only four cladoceran Ctenopoda, Haplopoda and Onychopoda, have remained species were included in the analysis, with no represen- unclear (Braband et al., 2002; deWaard et al., 2006; tative of the Onychopoda, and the aa sequence analysis Richter et al., 2007; Olesen, 2009; Olesen and Richter, gave different results than the nucleotide sequence analy- 2013). Similar to our results, Schwentner et al. (2018) sis, albeit with quite weak support for the former. Elu- showed that Ctenopoda is the sister group of the remain- cidation of the phylogeny of Cladocera still requires more ing groups of Cladocera, without, however, including any study. representative of Haplopoda in their study. The present Molecular phylogeny of Branchiopoda 9

Divergence time of branchiopod lineages In the Branchiopoda using various molecular clock methods, timetree produced by using five branchiopod fossil ages namely BEAST (Drummond and Rambaut, 2007), Mul- as calibration constraints (see Materials and Methods), tidivtime (Thorne et al., 1998), MCMCTree (Yang, 2007) the divergence between Anostraca and Phyllopoda is and r8s (Sanderson, 2003). For the divergence times estimated to have occurred 534 Ma, during the early of Anostraca–Phyllopoda and Notostraca–Diplostraca, Cambrian, and the divergence times between Notos- BEAST calculated maximum ages with averages of 606 traca and Diplostraca (= Laevicaudata + Spinicaudata + Ma and 548 Ma, and r8s gave minimum ages with aver- Cyclestherida + Cladocera) and between Laevicaudata ages of 495 Ma and 465 Ma, respectively. Our estimate and Onychocaudata (= Spinicaudata + Cyclestherida + yielded divergence times within these bounds, 534 Ma Cladocera) are estimated to be 496 Ma (late Cambrian) and 496 Ma (Fig. 7), which are similar to Sun et al.’s and 419 Ma (beginning of Devonian), respectively (Fig. (2016) Multidivtime and MCMCTree estimates. These 7). The split between Cyclestherida and Cladocera is divergence ages also roughly support the model of the estimated to have occurred 305 Ma, during the late Car- origin and evolutionary process of Branchiopoda proposed boniferous, and the Cladocera first diverged 261 Ma, dur- by Negrea et al. (1999) on the basis of morphological data ing the late Permian (Fig. 7). and the then-known fossil evidence. Luchetti et al. (2019) estimated the divergence time of Phylogenetic analysis of Anostraca (Weekers et al., four higher branchiopod taxa using a phylogenetic tree 2002) has shown a division of this order into two distinct based on mitochondrial genomic sequences. The diver- monophyletic groups. The first consists of the Artemiidae gence date between Spinicaudata and Cladocera and and Parartemiidae, representing saltwater taxa, whereas the internal divergence ages of various anostracan and the second contains all freshwater taxa. Artemia cladoceran taxa were similar to or older than ours, but franciscana, one of the three anostracan species used in the times of divergence between Anostraca and Phyllop- the present study, belongs to the first group and our other oda, and between Notostraca and Onychocaudata, were two species belong to the second group. Our results sug- younger than our estimates. This may be due to the satu- gest that the two clades of extant anostracans diverged ration of base substitutions accumulated in the mitochon- 142 Ma, during the early Cretaceous (Fig. 7). This, drial DNA sequences of those lineages. If so, including together with the lack of fossil evidence of any additional additional fossil calibration points for higher taxa crown stem- or crown-group anostracan diversity between the groups such as Onychocaudata (Gueriau et al., 2016) and Middle Devonian and the present day (Negrea et al., Diplostraca (Wolfe et al., 2016) might have reduced the 1999; Belk and Schram, 2001), suggests that the anostra- effect of such saturation. Another study (Sun et al., 2016) can lineage was evolutionarily static up until this binary estimated the divergence times of the major lineages of split. A similar case may be seen in Cyclestherida,

199.05 Moina macrocopa Anomopoda 261.52 Daphnia pulicaria Cladocera Cladoceromorpha Diaphanosoma sp. Ctenopoda 258.49 305.86 Leptodora richardi Haplopoda Onychocaudata 376.19 Cyclestheria hislopi Cyclestherida Caenestheriella gifuensis 170.62 Diplostraca 419.05 Leptestheria kawachiensis Spinicaudata Eulimnadia braueriana 496.00 134.03 Lynceus biformis Laevicaudata Phyllopoda 51.73 Triops cancriformis 534.48 Triops granarius Notostraca 36.12 Triops longicaudatus 142.09 Artemia franciscana 647.58 Branchinella kugenumaensis Anostraca 119.59 Eubranchipus uchidai 64.20 Petrobiellus takunagae Pedetontus unimaculatus 520.00 Hexapoda 160.95 Baculentulus morikawai Nipponentomon nippon Outgroup

Proterozoic Paleozoic Mesozoic Cenozoic Cry Ed Cam DeSiOr Car JuTrPe Cre Pa N, Q Divergence Time (Ma) 700 600 500 400 300 200 100 0 Fig. 7. Timetree of branchiopod higher taxa cladogenesis inferred using the RelTime method and the LG+G+I model. Evolution- ary analyses were conducted in MEGA X (Kumar et al., 2018; Stecher et al., 2020). Filled circles indicate the calibration points (see MATERIALS AND METHODS). The estimated log likelihood value is -34133.33. The divergence times between taxa are shown at each node, and the bars represent 95% confidence intervals. The chronology bar is according to the International Chronostratigraphic Chart v2020/01 of the International Commission on Stratigraphy. 10 T. UOZUMI et al. which our data show splitting from the common ances- Ministry of Education’s Higher Education Sprout Project. tor of Cladocera 305 Ma, during the late Carboniferous (Fig. 7). Although the extant populations seem to repre- REFERENCES sent the results of a split into three geographical lineages during the Middle to Late Cretaceous (Schwentner et al., Adamowicz, S. J., and Purvis, A. (2005) How many branchiopod 2013), all are still similar enough to be formally classi- crustacean species are there? Quantifying the components of underestimation. Global Ecol. Biogeogr. 14, 455–468. fied as a single species, Cyclestheria hislopi. The first Ahyong, S. T., Lowry, J. K., Alonso, M., Bamber, R. N., Boxshall, divergence among extant spinicaudatans is estimated to G. A., Castro, P., Gerken, S., Karaman, G. S., Goy, J. W., have occurred 170 Ma, in the middle Jurassic, much later Jones, D. S., et al. (2011) Subphylum Crustacea Brünnich, than that inferred in a previous study (Schwentner et 1772. Zootaxa 3148, 165–191. al., 2020), either 294.6 Ma or 403 Ma depending on the Belk, D., and Schram, F. R. (2001) A new species of anostracan from the Miocene of California. J. Crustac. Biol. 21, 49–55. choice of fossil-based minimum-age calibration point. In Braband, A., Richter, S., Hiesel, R., and Scholtz, G. (2002) Phy- sum, the present results (Fig. 7) suggest that the major logenetic relationships within the Phyllopoda (Crustacea, branchiopodan lineages were established at least before Branchiopoda) based on mitochondrial and nuclear mark- the break-up of Pangaea, and diversification within the ers. Mol. Phylogenet. Evol. 25, 229–244. extant branchiopod lineages started in or after the late Brendonck, L., Rogers, D. C., Olesen, J., Weeks, S., and Hoeh, W. R. (2008) Global diversity of large branchiopods (Crustacea: Permian, perhaps in connection with historical vicariance Branchiopoda) in freshwater. Hydrobiologia 595, 167–176. events such as the break-up of Pangaea into Gondwana Calman, W. T. (1909) Crustacea. In A Treatise on Zoology, Part and Laurasia, until which time they had evolved mini- 7 Appendiculata Fasc. 3. (ed.: Lankester, E. R.). Adam and mally, and silently. This notion should be tested by fur- Charles Black, London, UK. ther analyses using additional species from the lineages Castresana, J. (2000) Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol. of the Notostraca and Laevicaudata. Biol. Evol. 17, 540–552. Darriba, D., Taboada, G. L., Doallo, R., and Posada, D. (2011) Conclusion The aa sequences of three protein-coding ProtTest 3: fast selection of best-fit models of protein evolu- genes (RPB1, RPB2 and DPD1) proved to be highly reliable tion. Bioinformatics 27, 1164–1165. in inferring the phylogenetic relationships of higher-level Darriba, D., Taboada, G. L., Doallo, R., and Posada, D. (2012) jModelTest 2: more models, new heuristics and parallel com- taxa within Branchiopoda. Our results strongly support puting. Nat. Methods 9, 772. the monophyly of Branchiopoda and some of its constitu- deWaard, J. R., Sacherova, V., Cristescu, M. E. A., Remigio, E. A., ent taxa, Phyllopoda, Diplostraca, Onychocaudata, Cla- Crease, T. J., and Hebert, P. D. N. (2006) Probing the rela- doceromorpha and Cladocera, and corroborate Schwenter tionships of the branchiopod crustaceans. Mol. Phylogenet. et al.’s (2018) placement of the Notostraca as the sister Evol. 39, 491–502. Drummond, A. J., and Rambaut, A. (2007) BEAST: Bayesian to all extant non-anostracan branchiopods. Further- evolutionary analysis by sampling trees. BMC Evol. Biol. more, the nucleotide sequence analyses showed that the 7, 214. suborder Ctenopoda is the sister group to Haplopoda + Forró, L., Korovchinsky, N. M., Kotov, A. A., and Petrusek, A. Anomopoda within Cladocera. The divergence time (2008) Global diversity of cladocerans (Cladocera; Crustacea) estimated on the basis of firm data including that pro- in freshwater. Hydrobiologia 595, 177–184. Fryer, G. (1987) A new classification of the branchiopod vided by a recently discovered unequivocal spinicaudatan Crustacea. Zool. J. Linn. Soc. 91, 357–383. fossil suggests that the first divergence of crown-group Giribet, G., and Edgecombe, G. D. (2012) Reevaluating the branchiopods occurred during the early Cambrian period arthropod tree of life. Annu. Rev. Entomol. 57, 167–186. about 534 Ma, and that diversification within the extant Giribet, G., and Edgecombe, G. D. (2019) The phylogeny and evo- branchiopod lineages started in or after the late Permian, lutionary history of arthropods. Curr. Biol. 29, R592–R602. Glenner, H., Thomsen, P. F., Hebsgaard, M. B., Sørensen, M. V., before which they had evolved but little. and Willerslev, E. (2006) The origin of insects. Science 314, 1883–1884. We are grateful to M. Goto, K. Igarashi, Y. Kusuoka, K. Gueriau, P., Rabet, N., Clément, G., Lagebro, L., Vannier, J., Masunaga, T. Ojika, Y. Okamoto, R. J. Smith and T. Yamamoto Briggs, D. E. G., Charbonnier, S., Olive, S., and Béthoux, O. for their invaluable contribution in the supply of samples. We (2016) A 365-million-year-old freshwater community reveals deeply thank P. Dabseepai and her student for sampling morphological and ecological stasis in branchiopod crusta- Cyclestheria hislopi in Khon Kaen, Thailand. Thanks are also ceans. Curr. Biol. 26, 383–390. due to A. Sasaki for sampling and sequencing of Cyclestheria Guindon, S., and Gascuel, O. (2003) A simple, fast and accurate hislopi and producing illustrations of the representative bran- method to estimate large phylogenies by maximum- chiopods. This work was supported in part by grants from likelihood. Syst. Biol. 52, 696–704. JSPS KAKENHI (No. 17570191 and No. 16K07471). Some col- Hanner, R., and Fugate, M. (1997) Branchiopod phylogenetic lections were made as part of Lake Biwa Museum Comprehen- reconstruction from 12S rDNA sequence data. J. Crustac. sive Research Project S06-02, and M. J. G.’s later contributions Biol. 17, 174–183. were facilitated by support for National Taiwan Ocean Univer- Ishiwata, K., Sasaki, G., Ogawa, J., Miyata, T., and Su, Z.-H. sity’s Center of Excellence for the Oceans under the Featured (2011) Phylogenetic relationships among insect orders based Areas Research Center Program within the R.O.C. (Taiwan) on three nuclear protein-coding gene sequences. Mol. Phy- Molecular phylogeny of Branchiopoda 11

logenet. Evol. 58, 169–180. Regier, J. C., Shultz, J. W., Zwick, A., Hussey, A., Ball, B., Wetzer, Katoh, K., Kuma, K., Toh, H., and Miyata, T. (2005) MAFFT ver- R., Martin, J. W., and Cunningham, C. W. (2010) Arthropod sion 5: improvement in accuracy of multiple sequence align- relationships revealed by phylogenomic analysis of nuclear ment. Nucleic Acids Res. 33, 511–518. protein-coding sequences. Nature 463, 1079–1083. Kishino, H., and Hasegawa, M. (1989) Evaluation of the maxi- Richter, S., Olesen, J., and Wheeler, W. C. (2007) Phylogeny of mum likelihood estimate of the evolutionary tree topologies Branchiopoda (Crustacea) based on a combined analysis of from DNA sequence data, and the branching order in Homi- morphological data and six molecular loci. Cladistics 23, noidea. J. Mol. Evol. 29, 170–179. 301–336. Kumar, S., Stecher, G., Li, M., Knyaz, C., and Tamura, K. (2018) Ronquist, F., and Huelsenbeck, J. P. (2003) MrBayes 3: Bayesian MEGA X: Molecular Evolutionary Genetics Analysis across phylogenetic inference under mixed models. Bioinformat- computing platforms. Mol. Biol. Evol. 35, 1547–1549. ics 19, 1572–1574. Le, S. Q., and Gascuel, O. (2008) An improved general amino Sanderson, M. J. (2003) r8s: Inferring absolute rates of molec- acid replacement matrix. Mol. Biol. Evol. 25, 1307–1320. ular evolution and divergence dates in the absence of a Lozano-Fernandez, J., Giacomelli1, M., Fleming, J. F., Chen, A., molecular clock. Bioinformatics 19, 301–302. Vinther, J., Thomsen, P. F., Glenner, H., Palero, F., Legg, Sasaki, G., Ishiwata, K., Machida, R., Miyata, T., and Su, Z.-H. D. A., Iliffe, T. M., et al. (2019) Pancrustacean evolution (2013) Molecular phylogenetic analyses support the mono- illuminated by taxon-rich genomic-scale data sets with phyly of Hexapoda and suggest the paraphyly of Ento- an expanded remipede sampling. Genome Biol. Evol. 11, gnatha. BMC Evol. Biol. 13, 236. 2055–2070. Schneider, A., and Cannarozzi, G. M. (2009) Support patterns Luchetti, A., Forni, G., Skaist, A. M., Wheelan, S. J., and from different outgroups provide a strong phylogenetic sig- Mantovani, B. (2019) Mitochondrial genome diversity and nal. Mol. Biol. Evol. 26, 1259–1272. evolution in Branchiopoda (Crustacea). Zoological Lett. 5, Schwentner, M., Clavier, S., Fritsch, M., Olesen, J., Padhye, S., 15. Timms, B. V., and Richter, S. (2013) Cyclestheria hislopi Miyazawa, H., Ueda, C., Yahata, K., and Su, Z.-H. (2014) (Crustacea: Branchiopoda): a group of morphologically cryp- Molecular phylogeny of Myriapoda provides insights into tic species with origins in the Cretaceous. Mol. Phylogenet. evolutionary patterns of the mode in post-embryonic devel- Evol. 66, 800–810. opment. Sci. Rep. 4, 4127. Schwentner, M., Rabet, N., Richter, S., Giribet, G., Padhye, S., Negrea, S., Botnariuc, N., and Dumont, H. J. (1999) Phylogeny, Cart, J.-F., Bonillo, C., and Rogers, D. C. (2020) Phylogeny evolution and classification of the Branchiopoda (Crustacea). and biogeography of Spinicaudata (Crustacea: Branchiop- Hydrobiologia 412, 191–212. oda). Zool. Stud. 59, 44. Noah, K. E., Hao, J., Li, L., Sun, X., Foley, B., Yang, Q., and Xia, Schwentner, M., Richter, S., Rogers, D. C., and Giribet, G. (2018) X. (2020) Major revisions in arthropod phylogeny through Tetraconatan phylogeny with special focus on Malacostraca improved supermatrix, with support for two possible waves and Branchiopoda: highlighting the strength of taxon- of land invasion by chelicerates. Evol. Bioinform. Online specific matrices in phylogenomics. Proc. R. Soc. 285, 16, 1–12. 20181524. Oakley, T. H., Wolfe, J. M., Lindgren, A. R., and Zaharoff, A. K. Shimodaira, H. (2002) An approximately unbiased test of phylo- (2013) Phylotranscriptomics to bring the understudied into genetic tree selection. Syst. Biol. 51, 492–508. the fold: monophyletic Ostracoda, fossil placement, and pan- Shimodaira, H., and Hasegawa, M. (2001) CONSEL: for assess- crustacean phylogeny. Mol. Biol. Evol. 30, 215–233. ing the confidence of phylogenetic tree selection. Bioinfor- Olesen, J. (1998) A phylogenetic analysis of the Conchostraca and matics 17, 1246–1247. Cladocera (Crustacea, Branchiopoda, Diplostraca). Zool. J. Spears, T., and Abele, L. G. (2000) Branchiopod monophyly Linn. Soc. 122, 491–536. and interordinal phylogeny inferred from 18S ribosomal Olesen, J. (2000) An updated phylogeny of the Conchostraca— DNA. J. Crustac. Biol. 20, 1–24. Cladocera clade (Branchiopoda, Diplostraca). Crustaceana Stamatakis, A. (2006) RAxML-VI-HPC: maximum likelihood- 73, 869–886. based phylogenetic analyses with thousands of taxa and Olesen, J. (2004) On the ontogeny of the Branchiopoda (Crus- mixed models. Bioinformatics 22, 2688–2690. tacea)—contribution of development to phylogeny and Stecher, G., Tamura, K., and Kumar, S. (2020) Molecular Evolu- classification. In Crustacean Issues 15, Evolutionary tionary Genetics Analysis (MEGA) for macOS. Mol. Biol. Developmental Biology of Crustacea. (ed.: Scholtz, G.), pp. Evol. 3, 1237–1239. 217–269. Balkema, Lisse, Netherlands. Stenderup, J. T., Olesen, J., and Glenner, H. (2006) Molecu- Olesen, J. (2007) Monophyly and phylogeny of Branchiopoda, lar phylogeny of the Branchiopoda (Crustacea)—multiple with focus on morphology and homologies of branchiopod approaches suggest a ‘diplostracan’ ancestry of the Notos- phyllopodous limbs. J. Crustac. Biol. 27, 165–183. traca. Mol. Phylogenet. Evol. 41, 182–194. Olesen, J. (2009) Phylogeny of Branchiopoda (Crustacea)—char- Sun, X.-Y., Xia, X., and Yang, Q. (2016) Dating the origin of the acter evolution and contribution of uniquely preserved fos- major lineages of Branchiopoda. Palaeoworld 25, 303–317. sils. Arthropod Syst. Phylogeny 67, 3–39. Talavera, G., and Castresana, J. (2007) Improvement of phylog- Olesen, J., Martin, J. W., and Roessler, E. W. (1997) External enies after removing divergent and ambiguously aligned morphology of the male of Cyclestheria hislopi (Baird, 1859) blocks from protein sequence alignments. Syst. Biol. 56, (Crustacea, Branchiopoda, Spinicaudata), with comparison 564–577. of male claspers among the Conchostraca and Cladocera Tamura, K., Battistuzzi, F. U., Billing-Ross, P., Murillo, O., and its bearing on phylogeny of the ‘bivalved’ Branchiopoda. Filipski, A., and Kumar, S. (2012) Estimating divergence Zool. Scr. 25, 291–316. times in large molecular phylogenies. Proc. Natl. Acad. Sci. Olesen, J., and Richter, S. (2013) Onychocaudata (Branchiopoda: USA 109, 19333–19338. Diplostraca), a new high-level taxon in branchiopod system- Tamura, K., Tao, Q., and Kumar, S. (2018) Theoretical founda- atics. J. Crustac. Biol. 33, 62–65. tion of the RelTime method for estimating divergence times 12 T. UOZUMI et al.

from variable evolutionary rates. Mol. Biol. Evol. 35, as the possible sister group of Hexapoda. Mol. Biol. Evol. 1770–1782. 29, 1031–1045. Tao, Q., Tamura, K., Mello, B., and Kumar, S. (2019) Reliable Walossek, D. (1993) The Upper Cambrian Rehbachiella and the conﬁdence intervals for RelTime estimates of evolutionary phylogeny of Branchiopoda and Crustacea. Fossils Strata divergence times. Mol. Biol. Evol. 37, 280–290. 32, 1–202. Taylor, D. J., Crease, T. J., and Brown, W. M. (1999) Phyloge- Weekers, P. H. H., Murugan, G., Vanﬂeteren, J. R., Belk, D., and netic evidence for a single long-lived clade of crustacean Dumont, H. J. (2002) Phylogenetic analysis of anostracans cyclic parthenogens and its implications for the evolution of (Branchiopoda: Anostraca) inferred from nuclear 18S ribo- sex. Proc. Biol. Sci. 266, 794–797. somal DNA (18S rDNA) sequences. Mol. Phylogenet. Evol. Thorne, J. L., Kishino, H., and Painter, I. S. (1998) Estimat- 25, 535–544. ing the rate of evolution of the rate of molecular evolu- Wolfe, J. M., Daley, A. C., Leff, D. A., and Edgecombe, G. D. (2016) tion. Mol. Biol. Evol. 15, 1647–1657. Fossil calibrations for the arthropod Tree of Life. Earth- von Reumont, B. M., Jenner, R. A., Wills, M. A., Dell’Ampio, E., Sci. Rev. 160, 43–110. Pass, G., Ebersberger, I., Meyer, B., Koenemann, S., Iliffe, T. Yang, Z. (2007) PAML 4: Phylogenetic analysis by maximum M., Stamatakis, A., et al. (2012) Pancrustacean phylogeny in likelihood. Mol. Biol. Evol. 24, 1586–1591. the light of new phylogenomic data: support for Remipedia RNA polymerase II largest subunit (RPB1)

1 3 2 4 5 6 7 8 9 10

1 19 1000 2000 3000 4000 12 5631 15 14 18 17 16 13 11

RNA polymerase II second largest subunit (RPB2)

23 20 26 27 30 22 24 28 31 21 25 29

1 40 1000 37 2000 3000 3528 39 38 35 34 33 32 36

DNA polymerase delta catalytic subunit (DPD1)

41 43 44 / 45 46 47 48 49 42

1 59 1000 57 55 2000 52 3000 3276 58 56 53 51 54 50

500 bp

Supplementary Fig. S1. The coding sequence region of the three genes (RPB1, RPB2 and DPD1) sequenced in this study. The numbers and locations of primers used for PCR amplification and sequencing are indicated. The primer numbers correspond to those listed in Supplementary Table S1. Supplementary Table S1. List of taxa used for phylogenetic analyses in this study Subphylum Order Suborder Family Species Sample ID Locality DDBJ/EMBL/GenBank accession number Reference DPD1 RPB1 RPB2 Chelicerata Opiliones Eupnoi Psathyropus tenuipes 42 Yatsukami, Hamamura Kaigan, Tottori, Japan AB811978 AB811992 AB812006 Sasaki et al., 2013 Acari Ixodida Ixodes scapularis XM_002399417* DS725348* DS979207* Myriapoda Scutigeromorpha Thereuonema tuberculata 9 Yodo River, Torikaishimo, Settsu, Osaka, Japan AB811979 AB811993 AB812007 Sasaki et al., 2013 Crustacea Cyclopoida Cyclops vicinus 12 Yahagi River, Kawashima-cho, Anjo, Aichi, Japan AB811980 AB811994 AB812008 " Stomatopoda Oratosquilla oratoria 88 purchased from Osaka Prefectural Central Wholesale Market, Japan AB811981 AB811995 AB812009 " Decapoda Pleocyemata Procambarus clarkii E-33 Higashino, Matsuyama, Ehime, Japan AB811982 AB811996 AB812010 " Anostraca Artemiidae Artemia franciscana 114 Great Salt Lake, Utah, USA LC532181 LC532193 LC532205 This study Thamnocephalidae Branchinella kugenumaensis 97 Kataoka-cho, Kusatsu, Shiga, Japan AB811985 AB811999 AB812013 Sasaki et al., 2013 Chirocephalidae Eubranchipus (Drepanosurus ) uchidai 119 Ishikari, Hokkaido, Japan LC532183 LC532195 LC532207 This study Notostraca Triopsidae Triops longicaudatus 95 Kusatsu, Shiga, Japan LC532190 LC532202 LC532214 " " Triops granarius E-54 Saijo, Ehime, Japan AB811984 AB811998 AB812012 Sasaki et al., 2013 " Triops cancriformis 126 Shonai, Yamagata, Japan LC532184 LC532196 LC532208 This study Diplostraca Laevicaudata Lynceidae Lynceus biformis 94 Kusatsu, Shiga, Japan LC532189 LC532201 LC532213 " Spinicaudata Cyzicidae Caenestheriella (=Cyzicus ) gifuensis 96 Kusatsu, Shiga, Japan LC532191 LC532203 LC532215 " Leptestheriidae Leptestheria kawachiensis 93 Kusatsu, Shiga, Japan LC532188 LC532200 LC532212 " Limnadiidae Eulimnadia braueriana 116 Otsu, Shiga, Japan LC532182 LC532194 LC532206 " Cyclestherida Cyclestheriidae Cyclestheria hislopi ch2020 Bueng Kaeng Nam Ton, Kham Sa-at, Khon Kaen, Thailand LC532192 LC532204 LC532216 " Cladocera Leptodoridae Leptodora richardi 130 Lake Biwa, Kusatsu, Shiga, Japan LC532185 LC532197 LC532209 " Sididae Diaphanosoma sp. 149 Lake Biwa, Kusatsu, Shiga, Japan LC532187 LC532199 LC532211 " Daphniidae Daphnia pulicaria 36 Takiwaki-cho, Toyota, Aichi, Japan AB811983 AB811997 AB812011 Sasaki et al., 2013 Moinidae Moina macrocopa 132 Kusatsu, Shiga, Japan LC532186 LC532198 LC532210 This study Hexapoda Protura Baculentulus morikawai 4 Takamatsu Hill, Aina, Atsugi, Kanagawa, Japan AB811986 AB812000 AB812014 Sasaki et al., 2013 Nipponentomon nippon 3 Takamatsu Hill, Aina, Atsugi, Kanagawa, Japan AB811987 AB812001 AB812015 " Diplura Japygomorpha Occasjapyx japonicus 63 Saiki, Tsukuba, Ibaraki, Japan AB811991 AB812005 AB812019 " Campodeomorpha Metriocampa sp. 44 Mt. Hokiya, Ueda, Nagano, Japan AB598692 AB596891 AB597582 Ishiwata et al., 2011 Lepidocampa weberi 54 Shimoda, Shizuoka, Japan AB598693 AB596892 AB597583 " Archaeognatha Pedetontus unimaculatus 39 Shiroyama, Shimoda, Shizuoka, Japan AB598694 AB596893 AB597584 " Petrobiellus takunagae 41 Shiroyama, Shimoda, Shizuoka, Japan AB598695 AB596894 AB597585 " Zygentoma Nipponatelura sp. 66-1 Shimoda, Shizuoka, Japan AB598696 AB596895 AB597586 " Thermobia domestica 2 provided by Dr. Teruyuki Niimi AB598697 AB596896 AB597587 " Isolepisma japonica 46 Shirahama, Wakayama, Japan AB598698 AB596897 AB597588 " Odonata Anisozygoptera Epiophlebia superstes 29 Kibune River, Sakyo-ku, Kyoto, Japan AB598699 AB596898 AB597589 " Zygoptera Mnais pruinosa 31 Kibune River, Sakyo-ku, Kyoto, Japan AB598700 AB596899 AB597590 " Mantophasmatodea Karoophasma biedouwensis 157a Namaqualand, South Africa AB598706 AB596905 AB597596 " Orthoptera Ensifera Teleogryllus emma E-57 Kamo River, Furukawaotsu, Saijyo, Ehime, Japan AB598709 AB596908 AB597599 " Thysanoptera Terebrantia Thrips palmi E-63 Shimoidai-machi, Matsuyama, Ehime, Japan AB598717 AB596916 AB597607 " Hemiptera Auchenorrhyncha Cryptotympana facialis 49 Murasaki-cho, Takatsuki, Osaka, Japan AB598719 AB596918 AB597609 " * Sequence data retrieved from GenBank. Supplementary Table S2. List of primers used for PCR and sequencing in this study No.a Primer name Nucleotide sequence Amino acid sequence For RPB1 1 arN0.8 (40) GTCAAGARRGTiCARTTYGGNAT VK(RK)VQFGI 2 aN1.1 (85) CGAATGTCiGTNACiGARGGNGG RMSVTEGG 3 aN1 (223) ACCGAGTGYCCiGGiCAYTTYGG TECPGHFG 4 arN1.2 (226) GAGTGTCCiGGiCAYTTYGGNCA ECPGHFGH 5 aN1.5 (901) ATGCTGCARTTYCAYGTiGCNAC MLQFHVAT 6 N2 (1447) CCGTACAAYGCiGAYTTYGAYGG PYNADFDG 7 N3.2 (2338) GTCGGTCAGCARAAYGTiGARGG VGQQNVEG 8 N4 (3262) GGTGAGCCiGCNACiCARATGAC GEPATQMT 9 aN5 (3805) GACAAGATGGARGAYGAYATGTT DKMEDDMF 10 arN7 (4183) GACGTCATGACiGCNAARGGNCA DVMTAKGH 11 brC1.2a (4527) CCATGGAGTCATYTGiGGRGACAT MSPQMTPW 12 aC1 (4397) AGCAGGTCRAARCAiCCNGTNCC GTGCFDLL 13 arC1.3 (4202) CCTTTCGCiGTCATNACRTCRCA CDVMTAKG 14 aC2 (3446) ACGTTCTTGGCYTTYTCNGCRTC DAEKAKNV 15 C3 (2513) CCTTCACGiCCiCCCATNGCRTG HAMGGREG 16 C5 (1598) GTATCCTGiACDATiCCCATNAC VMGIVQDT 17 arC6 (1091) TTCGGATCiGGiGTDATNACNGT TVITPDPN 18 aC7 (470) TCCATTTCRTCiCCiCCYTCRCA CEGGDEMD 19 arC9 (248) TGACCGAARTGiCCNGGRCAYTC ECPGHFGH

For RPB2 20 aN1.1 (55) TCATCCGARYTiTGGCARGARGC SSELWQEA 21 aN1 (64) CTATGGCARGARGCiTGYTGGAT LWQEACWI 22 aN2 (583) CTGATTGCiCARGARAARATGGC LIAQEKMA 23 N3 (1339) TACTCTCTAGCiACiGGNAAYTGGG YSLATGNW 24 arN3.1 (1348) GCTACiGGiAAYTGGGGNGAYCA ATGNWGDQ 25 aN3.3 (1531) TGTCCAGCiGARACiCCNGARGG CPAETPEG 26 aN6 (1702) ATCTTYGTiAAYGGiTGYTGGGT IFVNGCWV 27 N4 (2167) CGTAACACTTAYCARTCiGCNATGGG RNTYQSAMG 28 N5 (2182) TCGGCTATGGGiAARCARGCNATG SAMGKQAM 29 aN7 (2821) CATGGTCARAARGGiACiTGYGG HGQKGTCG 30 N8 (2986) GGAGAAATHGGiGAYGCiACNCC GEIGDATP 31 N9 (3265) GGTGAGATGGARCGiGAYTGYCA GEMERDCQ 32 C2 (3288) CATGCAATCiCGYTCCATYTCNCC GEMERDCM 33 arC2.4 (2843) CCACACGTiCCYTTYTGNCCRTG HGQKGTCG 34 C3.5 (2207) CCCATAGCYTGYTTiCCCATNGC AMGKQAMG 35 arC3.8 (1736) CTGTGGATNCCiACCCARCANCC GCWVGIHR 36 C4 (1559) GCCTGTCCTTCiGGiGTYTCNGC AETPEGQA 37 arC4.1 (1553) CCTTCAGGiGTYTCNGCNGGRCA CPAETPEG 38 aC5 (905) ACCATCTCCATCATYTCiGGRTC DPEMMEMV 39 aC6 (344) TCGAGCCTCRTTiGGCATCATNG MMPNEAR 40 arC7 (311) CCATCCTTYTCCCARTGNGTNGG PTHWEKDG

For DPD1 41 aN1 (310) GTCATACGiATGTWYGGiGTNAC VIRM(FY)GVT 42 aN1.1 (346) TCTGTCTGYTGYCAYGTiCAYGG SVCCHVHG 43 N (892) AGTTTCGAYATYGARTGYGCNGG SFDIECAG 44 Nc (892) AGCTTCGAYATYGARTGYGCNGG SFDIECAG 45 N6 (1402) GAGCAGAARGARGAYGTiCAYCA EQKEDVHH 46 aN7 (2068) TACGGATTYACiGGiGCNCARGT YGFTGAQV 47 N7.2 (2425) CACGACAARATGGAYTGYAARGG HDKMDCKG 48 N7.5 (2770) AAGGCTGARGAYCCiATNTAYGT KAEDP(IM)YV 49 N8 (3127) TGGACiCAGTGYCARMGiTGYCA WTQCQRCQ 50 C1 (3197) ATAGGACARTCiCGRCTNGTRCA CTSRDCPI 51 aC1.1 (3149) TGACATCGYTGRCAYTSiGTCCA WT(EQ)CQRCQ 52 aC1.5 (2458) CAGTCTCGAKiCCYTTRCARTCCAT MDCKG(IL)ET 53 C2 (2249) ACCATGACAGARTCiGTRTCNCCRTA YGDTDSVMV 54 aC3 (2087) TGTGCGCCiGTRAAiCCRTANAC VYGFTGAQ 55 aC4 (1811) GTGTAGCACARRTTRTGiGCNATCAT M(IM)AHNLCYT 56 C (1574) ACACCAGTNACiCKNGCCATYTC EMARVTGV 57 aC5 (1424) TGATGCACRTCYTCYTTYTGYTC EQKEDVHH 58 C6 (923) CCTTTGCGiCCiGCRCAYTCDAT IECAGRKG 59 C8 (371) AAGCCATGiACRTGRCARCANAC VCCHVHGF a Primer numbers correspond to those shown in Supplementary Fig. S1. Supplementary Table S3. Sequence information of the three genes determined in this study Order Suborder Species DPD1 RPB1 RPB2 Length (aa) NC Length (aa) NC Length (aa) NC Anostraca Artemia franciscana 1109 ○ ○ 1698 ○ 1179 ○ ○ Branchinella kugenumaensis 1120 ○ ○ 1651 ○ 1177 ○ ○ Eubranchipus uchidai 953 ○ 1719 ○ 1177 ○ ○ Notostraca Triops longicaudatus 967 ○ 1756 ○ 1175 ○ ○ Triops granarius 1006 ○ 1627 ○ 1175 ○ ○ Triops cancriformis 967 ○ 1802 ○ 1175 ○ ○ Diplostraca Laevicaudata Lynceus biformis 964 ○ 1732 ○ 1175 ○ ○ Spinicaudata Caenestheriella gifuensis 992 ○ 1577 1149 ○ Leptestheria kawachiensis 960 ○ 1529 1146 ○ Eulimnadia braueriana 1103 ○ ○ 1555 ○ 1174 ○ ○ Cyclestherida Cyclestheria hislopi 1093 ○ ○ 1907 ○ ○ 1174 ○ ○ Cladocera Leptodora richardi 957 ○ 1724 ○ 1146 ○ Diaphanosoma sp. 959 ○ 1616 1164 ○ Daphnia pulicaria 957 ○ 1765 1146 ○ Moina macrocopa 1047 ○ ○ 1746 1146 ○ N , N-terminus of the coding sequence of the gene; C , C-terminus of the coding sequence of the gene; ○, completely determined. Supplementary Table S4. Proportions of parsimony-informative and variable amino acid sites in the alignment data sets Data set Gene Total Parsim-inform Singleton Variable Conserved Variable/total Parsim-info/total 22-sample DPD1 886 392 104 496 390 55.98% 44.24% RPB1 1467 395 157 552 915 37.63% 26.93% RPB2 1137 231 111 342 795 30.08% 20.32% 37-sample DPD1 874 453 88 541 333 61.90% 51.83% RPB1 1434 470 143 613 821 42.75% 32.78% RPB2 1139 275 121 396 743 34.77% 24.14% These data were calculated with MEGA X (Kumar et al., 2018).