Unusual biogeographic patterns of closely related groups reflect events in the past, and molecular analyses can help to elucidate these events. While ample research on the origin of disjunct distributions of different organism groups in the Western Paleartic has been conducted, such studies are rare for Eastern Palearctic organisms. In this paper we present a phylogeographic analysis of the disjunct distribution pattern of the extant species of the strongly cool-adapted Epiophlebia dragonflies from Asia. We investigated sequences of the usually more conserved 18 S rDNA and 28 S rDNA genes and the more variable sequences of ITS1, ITS2 and CO2 of all three currently recognised Epiophlebia species and of a sample of other odonatan species. In all genes investigated the degrees of similarity between species of Epiophlebia are very high and resemble those otherwise found between different populations of the same species in Odonata. This indicates that substantial gene transfer between these populations occurred in the comparatively recent past. Our analyses imply a wide distribution of the ancestor of extant Epiophlebia in Southeast Asia during the last ice age, when suitable habitats were more common. During the following warming phase, its range contracted, resulting in the current disjunct distribution. Given the strong sensitivity of these species to climatic parameters, the current trend to increasing global temperatures will further reduce acceptable habitats and seriously threaten the existences of these last representatives of an ancient group of Odonata.
Citation: Büsse S, von Grumbkow P, Hummel S, Shah DN, Tachamo Shah RD, Li J, et al. (2012) Phylogeographic Analysis Elucidates the Influence of the Ice Ages on the Disjunct Distribution of Relict Dragonflies in Asia. PLoS ONE 7(5): e38132. https://doi.org/10.1371/journal.pone.0038132
Editor: Robert DeSalle, American Museum of Natural History, United States of America
Received: March 2, 2012; Accepted: May 4, 2012; Published: May 30, 2012
Copyright: © 2012 Büsse et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This project was partly financed through the DFG (Deutsche Forschungsgemeinschaft) grant HO2306/6-1. TH was directly supported through the DFG Heisenberg grant HO2306/7-1. JL and XZ were supported by a grant from the National Science Foundation of China (No. 41071033). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Disjunct biogeographic patterns with closely related organisms occurring in widely separated areas are puzzling. Such disjunct distributions reflect historical events, and a major goal of research is to deduce which events were responsible for the given distribution pattern. Disjunct ranges of closely related taxa can come into existence either by tectonics, by dispersal or by intervening extinction . For European mountain ranges it has already been shown that today’s disjunct distribution patterns of many cold-adapted species are post ice age relicts . However, the study of disjunct species and their phylogeography in other areas such as e.g. the Himalayas and Southeast Asia is still in its infancy.
In this paper we illuminate the disjunct distribution pattern of Epiophlebia, a unique dragonfly from Asia.
Traditionally, the Odonata have been divided into Anisoptera (dragonflies), Zygoptera (damselflies) and “Anisozygoptera”. Besides the extant species of Epiophlebia the “Anisozygoptera” comprised mainly Jurassic fossils  until it was shown that “Anisozygoptera” are not monophyletic –. Presently, Epiophlebia is considered to be the most basal sistergroup of the Anisoptera, with several extinct lineages nested in between –. A close relationship of Anisoptera and Epiophlebia has also been corroborated by several molecular analyses and the term Epiprocta has been introduced for this grouping –.
The species of Epiophlebia have been considered as “living fossils”  because they display features of both damselflies (Zygoptera) and dragonflies (Anisoptera). In their general body outline they resemble dragonflies, but as in damselflies their fore- and hind wings are similarly shaped and petiolate . Like dragonflies, the larvae use a rectal chamber for respiration, but jet propulsion, which is typical of dragonflies, was never observed .
So far it was assumed that there are three extant species of Epiophlebia but there were doubts on the species differentiation between E. superstes (Sélys, 1889) and E. laidlawi Tillyard, 1921 from early on , –. Recently, a third species E. sinensis Li and Nel, 2011 was described from China .
While the Japanese E. superstes was thoroughly described by Asahina , , E. laidlawi, and E. sinensis are only poorly known. The morphological discrimination between species is based only on the following features , , : adults of E. laidlawi differ from E. superstes in larger size and brownish body coloration, parts of the male genitalia differ slightly in shape, the apical process of the eighth sternite is less developed in the female of E. laidlawi, and the wings are slightly longer . For the larvae of E. laidlawi the following differences to E. superstes have been descibed [16: p. 445]: slightly larger body size; length and form of the third antennal segment; form of the antero-lateral angle of the pronotum, form of the fore femur, development of the lateral spines of abdominal segments 7 to 9, and differences in the shape of the epiproct.
E. sinensis was described on the base of two adult male specimens, which differ from the other species in the hairiness of the epiproct and in the colouration of the abdomen .
Nowadays, the distribution of Epiophlebia is disjunct. E. superstes is restricted to large areas of Japan , , , E. laidlawi is found only in the Himalayas , , ,  and the recently described E. sinensis adds another dot in north-east China to the disjunct pattern. This distribution is due to very specific habitat requirements. Epiophlebia prefers cold mountain streams with temperatures of about 4 to 5°C in winter and about 16–17°C in summer (data published for E. superstes ) and altitudes between 1,300 to approximately 3,000 m (for E. laidlawi ). The recently described E. sinensis also fits into this pattern as it was collected in the vicinity of a mountain stream, however, at an elevation of not more than 500 m . In a recent discussion the biogeography of E. superstes and E. laidlawi  Epiophlebiidae and the extinct closely related Stenophlebiidae are considered as part of an “archeo-palaearctic dragonfly fauna” that formed in the Mesozoic and during the Tertiary was intermingled with oriental faunal elements. The extant Epiophlebia is considered as belonging to this ancient fauna, parts of which survived on the Japanese islands, in the Himalayas and in China , .
The present study aims at clarifying the phylogenetic relationships and the biogeographic history of E. superstes, E. laidlawi and E. sinensis from a genetic point of view. Since DNA sequences so far were only available for E. superstes , ,  we had to acquire additional sequence data for specimens from the other species and from different populations. The specimens available had originally been preserved for morphological and faunistic research, applying preservatives containing, among others, formaldehyde. Thus, we applied techniques used for analysis of degraded DNA  to get sequences of sufficient length. To achieve a good resolution of relationships on all taxonomic levels we investigated conserved genes as well as more variable regions , , .
Here we present the unexpected homogeneity of DNA-sequences of the supposed species E. laidlawi, E. superstes and E. sinensis. In the light of our new data, the extant disjunct distribution of Epiophlebia is explained in a distinctly different scenario than those suggested by  or .
The specimens of Epiophlebia initially were not collected and preserved for subsequent DNA analysis. Thus, amplification and sequencing turned out to be problematic. However, for all targeted genes it was possible to acquire sections of up to 300 bp (Table S1). These were positioned in such a way, that phylogenetic relevant sequences could be expected while allowing straightforward alignment throughout at least the Odonata.
For E. superstes and E. laidlawi sequences of 18 S, and 28 S rDNA, ITS1, ITS2 and CO2 genes were analysed as well as CO1 for E. superstes. For E. laidlawi it was not possible to get a sequence for CO1. Due to contamination with other odonatan DNA we could only acquire the ITS sequences for E. sinensis, because our primers are highly specific for these genes.
All sequences show an extreme degree of similarity between all Epiophlebia species.
Sequences of E. superstes from  were used as reference. The sequences of 18 S rRNA (240 bp) and 28 S rRNA genes (1∶191 bp; 2∶267 bp; 3∶251 bp; 4∶293 bp) did not show any differences between E. laidlawi (specimens NATR3 & NA01) and E. superstes (our sequences and FN356086, EU424328). For CO2 (265 bp) a single difference at position 368 relative to the E. superstes reference sequence (EU055421) was found in all specimens including the Japanese control specimen (Figure S1).
Likewise, the fragment of ITS2 (265 bp) shows one deletion of G at position 2613 (relative to reference sequence FN356086) in all specimens investigated. Additionally, a maximum of three more differences were found in one of the E. laidlawi specimens (NA01) the other specimen (NATR03) showed only one difference and in E. sinensis there are two (Figure S2).
In ITS1 (215 bp) two deletions (one G at position 1943 and one C at position 1944, relative to the E. superstes reference sequence from GenBank) are common to all specimens investigated. Additionally, ITS1 shows a maximum of 11 differences in E. laidlawi and seven in E. sinensis. In specimen NA01 of E. laidlawi also a heterozygous duplication of AAC at position 1935 of ITS1 was detected (Figure 1).
The complete dataset for phylogenetic analysis contained 17 taxa including a representative of Zygentoma as outgroup, two specimens of both Epiophlebia laidlawi and E. superstes and one specimen of E. sinensis. For each gene one sequence for E. superstes was taken from GenBank (Table S1). The 4799 characters were composed of 537 positions from the CO2 gene, 1700 positions from 18 S, 2157 positions from 28 S, 231 nucleotides from ITS2 and 173 nucleotides from ITS1. For some of the taxa in the matrix sequences from different species, or specimens of uncertain species determination, of a certain genus had to be combined into a chimeric sequence. This is true for Zygentoma, Calopterygidae, Lestidae, Gomphidae, Cordulegasteridae, Aeshnidae, Corduliidae and Coenagrionidae (Table S1). Alignment of ITS sequences turned out to be problematic due to their high variability. Eventually, an alignment was accepted, produced with the online version of MAFFT  with standard parameters, except for the scoring matrix set to “20PAM/k = 2” and the offset value set to 0.1. Highly variable sections were removed from the dataset.
Since it was not possible to obtain a sequence of CO1 for E. laidlawi or E. sinensis, this gene was not used in the phylogenetic analysis.
Data were formatted as mixed dataset for analysis with MrBayes 3.1.2. For each data partition a model was selected with MrModeltest 2.3: for CO2, 18 S and 28 S GTR+I+G was used, for ITS2 F81+G and for ITS1 JC+G. A NEXUS file with alignments and all parameters used can be found in Dataset S1.
Bayesian analysis over four million generations produced the phylogram shown in Figure 2A.
(A) Phylogram of bayesian analysis of full dataset (see text). Numbers indicate posterior probability for respective branches. (B) Strict consensus of three trees from maximum likelihood analysis. Numbers indicate bootstrap values. Where these values are missing the respective node collapsed in the bootstrap analysis.
For the objective of this investigation the relationships of the Epiophlebia specimens are most important. These specimens do not appear in a sistergroup relationship with the other specimen of the same species. This result is reproduced when analysing the dataset with the maximum likelihood algorithm (ML) (SYM+I+G used for the complete dataset, Figure 2B) or under maximum parsimony (hsearch, addseq = random, nreps = 100; both in PAUP*). With respect to the relationships of the Epiophlebia specimens also phylogenetic analyses based only on single genes reproduced the same arrangement.
The position of Epiophlebia as sistergroup of Zygoptera and the non-monophyly of Anisoptera are remarkable. These results contradict the widely accepted hypothesis of a monophyletic Anisoptera and a recently published hypothesis based on more comprehensive datasets proposing a sistergroup relationship between Epiophlebia and Anisoptera –. However, in the maximum likelihood analysis (Figure 2B) a sistergroup relationship of Epiophlebia and a clade containing Anisoptera with the exception of Aeshnidae is recovered.
The external morphology of several larval instars of E. superstes and E. laidlawi was compared by . He concluded that they might well be two separate species based on six differing characters. Four of these characters could not be confirmed by our morphological investigation of larvae of both species. Two characters, the lateral posterior corners of the abdominal segments 7 to 9, and the shape of the epiproct consistently show slight differences between these supposed two species. Larvae of E. sinensis were not available for investigation.
Differences in the adults are restricted mainly to coloration. Himalayan specimens seem to be more brownish than the black Japanese specimens and the Chinese ones have a reddish hint in the posterior area of the abdomen. Furthermore, there are small differences in the arrangement of setae and in the shape of the male genitalia , –, .
In total, the visible differences between specimens from the different regions allow identifying where a specimen was found. However, a general problem is the small number of known specimens from the Himalayas and from China. Thus the variability of morphological features cannot be determined.
When assembling and aligning the sequences from Epiophlebia superstes, E. laidlawi and E. sinensis, it quickly became obvious that there are only very few differences between these supposed species.
In order to be able to adequately rate this observation, we compared sequences of several different closely related odonatan species as well as specimens of the same species from different populations (Table 1). Between representatives of different species there usually are significantly more differences over similar sequence length in the same area of a gene than are present between the specimens of Epiophlebia. In most species the number of intraspecific differences (e.g. in CO2 of Ischnura asiatica) is higher than the number of differences found between species of Epiophlebia. Even the usually variable partial 28 S sequences show no polymorphisms between the investigated specimens. We were only able to locate the more or less identical region of ITS from GenBank for other Odonata species. Unfortunately, for 18 S, 28 S and CO2 we were not able to locate the same sections in enough specimens of closely related species or of different specimens of the same species for such a comparison. However, due to the used sequencing method we identify the most variable regions of 18 S and 28 S and even those show no differences.
The occurrence of a heterogeneous insertion of three base pairs as well as the clustering of polymorphisms in ITS1 in the E. laidlawi specimens indicates that this region is comparatively unstable. Slippage events might occur more often than in other regions and it is likely that the observed polymorphisms arose from two slippage events rather than one-by-one through point mutations. Thus, we propose that the 11 differences observed between E. laidlawi and E. superstes can be traced back to only four different events.
Even though there are so few variations between the specimens investigated we even observed a few different positions between the E. superstes sequences from GenBank that was used for comparison and the newly produced sequences. The specimens that were sequenced for the GenBank data are from a population in the vicinity of Tokyo, whereas our specimens are from Hokkaido in the far north of Japan. Therefore the observed differences might just reflect the sequences’ origin form different populations and also indicate that there is not very much gene exchange between these populations.
The comparisons of sequences of different populations and of closely related species (Table 1) reveal that the interspecific variability in DNA sequences within different populations of Odonata can be much higher than between specimens of Epiophlebia. In total this suggests that gene transfer between its species and populations took place in the not so distant past.
Despite the differences in the general topology of trees generated with maximum parsimony, maximum likelihood and bayesian algorithms the specimens of Epiophlebia are always grouped in one well-supported clade but never appear in the arrangement of the currently recognized species. This arrangement, with specimens of Epiophlebia freely mixed on the tree, is independent of the composition of the dataset and of the algorithm applied. Even the fairly variable CO- and ITS-sequences produce this result when analysed separately. These results challenge the assumption of three separate extant species of Epiophlebia. Further investigations especially of the morphology should be done to clarify the variability of discriminating features and the taxonomic status of the three species.
Further results of the present phylogenetic analysis stating a non-monophyletic Anisoptera and a sistergroup relationship of Epiophlebia and Zygoptera are remarkable. Taxa for the analysis were originally selected to represent major clades from the phylogenetic system of Odonata as reconstructed in  and we expected to recover its general topology. However, our taxon sample is not as comprehensive as in any of the investigations that recently confirmed monophyletic Anisoptera with a sistergroup Epiophlebia , , , . Therefore, the differing topology found in the present analysis may well be attributed to effects of the composition of taxa in the dataset. Monophyletic Anisoptera as well as its sistergroup relationship with Epiophlebia are also supported by morphological characters found in extant as well as in fossil Odonata , , . These relationships seem to be more probable than the topology found in our analyses, since our dataset was not compiled to be especially informative in respects of the relationships of high-level taxa of Odonata. However, since our results are reproducible with different algorithms and from different combinations of sequence data, they should be understood as a strong indication that higher level phylogenetic relationships within Odonata might not yet be finally resolved.
Biogeographic History of Epiophlebia
Firm biogeographic connections between the area of the present Himalayas, the Asian mainland and of Japan in former times are well documented by the Sino-Japanese floristic region . However, since when an effective isolation of the Asian mainland and the Japanese populations is established, can only be estimated.
Presently Japan is separated from the mainland by sea-straits with depths of ca. 55 m north of Hokkaido and ca. 130 m between the southern island Kyushu and Korea . The last substantial land bridge to Hokkaido was present during the Würm sea level lowering approximately 20,000 years BP . According to some authors , ,  a land bridge also existed between Kyushu and the Korean peninsula during this time (Figure 3). Assuming that during this period was the most recent possibility for genetic exchange between the Japanese and the mainland populations of Epiophlebia, it also follows that their distribution was significantly different from the present one.
(A) Known present distribution, the star marks the type locality of E. sinensis Li & Nel, 2011. (B) Approximate coastline during the Würm glacial period with land bridges between Japan and the mainland. The black line marks a possible northern boundary of the range of Epiophlebia during this time. (Modified from ).
Epiophlebia larvae apparently are very stenoecious and inhabit only cool headwaters of streams. During the last glacial maximum suitable environmental conditions probably were present throughout the lowlands south, southeast and east of the Himalayas. The area with adequate environmental conditions probably reached at least northwards to the connection between the southern Japanese island Kyushu and the mainland (Figure 3B). When the temperatures rose again at the end of the last ice age Epiophlebia retreated to the cooler higher areas. Consequently, the extant populations of Epiophlebia are small relicts of a formerly much larger and wider distributed population.
Previously the extant distribution of Epiophlebia was dated back to the Jurassic when Pangäa broke apart , . This might be true for the general Asiatic distribution of the ancestors of Epiophlebia. However, our data show that the extant species and most likely the genus are very much younger. Therefore, the interpretation that even the species go back to Jurassic time  can not be maintained.
Current climatic development threatens the existence of these populations . With globally rising temperatures suitable habitats for species like Epiophlebia with such narrow tolerances for environmental parameters are in great danger of extinction.
Status of Epiophlebia Species
In the light of the extreme similarity of sequences in Epiophlebia specimens from Japan and Nepal, the observed morphological differences between their adults as well as their larvae tend to appear as minor local variations. In summary, this indicates that probably the Himalayan and the Japanese populations are in fact representatives of a single biological species. However, with the current climatic and topographic situation the chances of future genetic exchange between these populations are bleak. So, even if at present the different populations just represent one biological species, the probability that they are just on the verge of becoming “real” species is very high.
One might even assume that the populations in the different valleys of the Himalayas are separated from one another by the high mountain ranges as effectively as any of them is separated from the Japanese population by sheer distance . So, perhaps even here there is a certain probability for the future formation of individual species.
For E. sinensis the situation is not very different. Only the amount of available data for this species is much worse than for the Japanese and the Himalayan populations. No larvae are known so far and sequence data are also sparse. Furthermore, the available information on the morphology is based on no more than two male specimens . So, we know nothing about morphological variability. The available sequence data are, however, as similar to those of E. superstes as those from E. laidlawi are. If our biogeographic interpretations are correct, than it even seems to be probable that genetic exchange between E. superstes and E. sinensis was possible until a more recent time than between E. superstes and E. laidlawi. Thus, E. sinensis might also be just another population of this single species of Epiophlebia.
Mutation Rates of Different DNA-sequences of Epiophlebia
The analyses of the different sequences indicate a low variability even in quickly evolving parts of the genome as in the ITS genes. This implies that the range expansion of Epiophlebia during the last glacial maximum was very quick, because these circumstances enhance low genetic diversity in a population .
On the other hand, this might present an opportunity to estimate mutation rates backwards, at least for the highly variable sections of the DNA. Assuming that the genome was very homogeneous after range expansion during the last glacial maximum about 20,000 years ago, the substitution rate can be inferred from the number of differences in the sequences that we were able to acquire for Epiophlebia:
For ITS1 a rate of 4.65×10−7 to 2.56×10−6 substitutions per site per year can be deduced, for ITS2 there are 5.19×10−7, and for CO2 1.89×10−7 substitutions per site per year. Compared to such rates in other organisms ,  the rates for Epiophlebia are clearly among the higher values. Nevertheless, these figures do not contradict the assumption of a comparatively recent separation of the Epiophlebia populations.
Materials and Methods
Eleven larvae of Epiophlebia superstes (Sélys, 1889) of different instars were collected in 2010 in Hokkaido, Japan, fixed and stored in 80% ethanol.
Twelve larvae of different instars of E. laidlawi Tillyard, 1921 were collected in 2008 and 2009 in Nepal, fixed in 4% formaldehyde and stored in 70% ethanol. Specimens are stored in the collection of the Hindu Kush Himalayan Benthological Society, Nepal.
Two adults of E. sinensis Li and Nel, 2011 were collected in 2011 in Heilongjiang province, China as described in . Two femora of these specimens were available for sequencing.
Larvae of Coenagrion spec. were collected in the botanical garden of the Georg-August-University in Göttingen, Germany; fixed in FAE and stored in 70% ethanol (to simulate the conditions of preservation of the other specimens).
For detailed morphological investigation tomography data of four larvae of E. laidlawi and three specimens of E. superstes were acquired at the Swiss Light Source synchrotron (SLS, Viligen, Switzerland, proposal no. 20100088 by TH) and with a v|tome|x s X-ray scanner (GE Sensing & Inspection Technologies GmbH phoenixjx-ray) at the Palaeontological Institut at University Bonn (Germany).
The DNA analysis was carried out under strict safety conditions , such as separation of pre- and post-PCR laboratories and the use of disposable protective clothing, glassware, and disposable gloves. All experiments took place with disposable laboratory ware, such as pipette tips and cups, while workbenches and other laboratory equipment were cleaned with detergents (Alconox™ Detergent, Aldrich, Germany), bi-distilled water, and ethanol before use for each sample to avoid cross-contamination. In accordance with the recommendations of , all disposable ware and solutions, buffers, and MgCl2 were irradiated with ultraviolet light at a short distance employing aluminum foil coating. Negative PCR and extraction controls were employed.
Before DNA extraction the guts of the specimens were removed to avoid possible contamination with foreign DNA. Only thorax and leg muscles were used for the analyses.
For cell lysis, 200 µl ATL buffer (Qiagen, Germany) was added to 10–20 mg of tissue. The mixture was homogenized in a TissueLyser (Qiagen) at 30 Hz for 60 s using a 5 mm steel ball. After removal of the ball, 30 µl of Proteinase K (20 mg/ml) was added to the solution and incubated at 56°C for 18 hours under constant agitating. 200 µl of the supernatant were used for automated DNA extraction with the Biorobot® EZ1 (Qiagen, Germany) following the protocol of the QIAamp DNA FFPE Tissue procedure. The elution volume was 50 µl; the DNA extract was stored at −20°C. We carried out at least two independent DNA extractions and sequencings for each individual to permit authentication of the analysis results by means of comparison. Heterogeneity was only detected in ITS1 of E. laidlawi.
Due to storage conditions and influences of the preservatives, DNA of the E. laidlawi specimens was degraded. Therefore, primers were designed matching the profile of ancient DNA characteristics  and amplifying fragments between 200 and 300 bp each (Table S2). Because of this limitation, we amplified only polymorphic sites instead of the whole genes in the cases of the mainly conserved 18 S and 28 S rRNA genes . To gain information on the polymorphic sites, alignments of sequences from GenBank of numerous different species were carried out using MegAlign (Lasergene, www.DNASTAR.com) and the Clustal V algorithm. The primers were designed to discriminate against human DNA and to amplify as many taxa of Odonata as possible.
PCR Parameters and Sequencing
The reaction volume in each setting was 25 µl, containing 12.5 µl 2x master mix (AmpliTaq® Gold 360, ABI), 0.4 µM of each primer, 5–7.5 µl of DNA extract and filled up with RNAse free water (Qiagen). PCR was carried out under the following conditions: initialization 95°C for 5 min; 40 – 45 cycles at 95°C for 1 min, annealing temperature (Table S2) for 1 min, 72°C for 2 min; final elongation at 72°C for 7 min; and soak at 10°C for 10 min. The PCR success and product quantity were checked by agarose gel electrophoresis. Further purification and sequencing were carried out with commercial kits (MiniElute® PCR Purification Kit, Qiagen, ABI Prism BigDye V 3.1 Terminator Cycle Sequencing Kit and NucleoSeq Kit, Macherey-Nagel) as specified by the manufacturers.
Both the forward and reverse primers used for amplification were also used for the sequencing reaction. The sequencing conditions were: initial at 96°C for 10 min; 25 cycles at 96°C for 10 s, 50°C for 5 s and 60°C for 4 min. For sequencing an ABI 310 genetic analyser with POP6 polymer was used. Sequence reads were checked for quality and assembled using SEQMAN (Lasergene; www.DNASTAR.com).
Sequences were compiled and aligned using MEGA version 5 . Ribosomal DNA sequences were aligned automatically with the clustalW algorithm and standard parameters. Protein coding mitochondrial sequences were aligned manually via the corresponding amino acid sequences. For alignment of ITS sequences the online version of MAFFT ,  was used.
The most appropriate models of DNA substitution for bayesian and maximum likelihood tree searches were selected with MrModeltest 2.3  and PAUP*4.0 b10  separately for each sequence alignment and for a combined dataset containing all sequences except CO1, which was not used in the analysis since it was not possible to obtain sequences of this gene from E. laidlawi or E. sinensis specimens. The complete dataset contained 17 taxa, including a representative of Zygentoma as outgroup, and 4799 positions. Taxa from Odonata were chosen to represent all major clades as present in the phylogeny of . A Nexus file of the alignments are provided as Dataset S1.
To find the most probable phylogenetic relationships of the Epiophlebia species the aligned sequences were combined into a mixed dataset for analysis with MrBayes 3.1.2 , . MrBayes offers the unique possibility to reconstruct a phylogeny based on information form several different genes in a single analysis, while applying the most appropriate model to each gene sequence. Additional tree searches were done for the combined dataset with global parameters with the maximum likelihood (ML) algorithm in PAUP* and for each gene individually with MrBayes and with ML in PAUP*. To check node stability for ML a bootstrap analysis with 1000 replicates was done with the same parameters as for the original analysis.
Alignment of CO2 sequences from different specimens of Epiophlebia species. E. superstes_GB = reference sequence from GenBank.
Alignment of ITS2 sequences from different specimens of Epiophlebia species. E. superstes_GB = reference sequence from GenBank.
GenBank accession numbers for sequences used in phylogenetic analysis. * = this paper.
Primers. # this paper; * ; primers in italics are specific only for Epiophlebia.
Data matrix in NEXUS format including parameters for bayesian and maximum likelihood analysis.
TH and SB want to thank Prof. Dr. R. Willmann for his support of their work. KY wishes to thank Akiko Otsuki for support during fieldwork. Morphological investigation of Epiophlebia larvae was backed up through µCT investigations at the Swiss Light Source (proposal no. 20100088 by TH), Villigen, Switzerland and at the Institut für Paläontologie, University Bonn with the help of Dr. Irina Ruf. The authors wish to thank the anonymous reviewers for the very constructive comments.
Conceived and designed the experiments: SB PvG SH DNS RDTS KY TH. Performed the experiments: SB PvG TH. Analyzed the data: SB PvG TH. Contributed reagents/materials/analysis tools: PvG DNS RDTS JL XZ KY. Wrote the paper: SB PvG TH SW.
- 1. Lomolino MV, Riddle BR, Brown JH (2006) Biogeography. Sunderland: Sinauer Associates. 845 p.MV LomolinoBR RiddleJH Brown2006Biogeography.SunderlandSinauer Associates845
- 2. Habel JC, Assmann T, editors. (2011) Relict species. Phylogeography and conservation biology. Berlin: Springer. 449 p.JC HabelT. Assmann2011Relict species.Phylogeography and conservation biology. BerlinSpringer449
- 3. Nel A, Martinez-Delclos X, Paicheler JC, Henrotay M (1993) Les “Anisozygoptera” fossiles. Phylogenie et classification (Odonata). Martinia 3: 1–311.A. NelX. Martinez-DelclosJC PaichelerM. Henrotay1993Les “Anisozygoptera” fossiles. Phylogenie et classification (Odonata).Martinia31311
- 4. Lohmann H (1996) Das phylogenetische System der Anisoptera (Odonata). Entomologische Zeitschrift 106: 209–252, 253–266, 360–367: H. Lohmann1996Das phylogenetische System der Anisoptera (Odonata).Entomologische Zeitschrift 106: 209–252, 253–266,360–367
- 5. Rehn AC (2003) Phylogenetic analysis of higher-level relationships of Odonata. Syst Ent 28: 181–239.AC Rehn2003Phylogenetic analysis of higher-level relationships of Odonata.Syst Ent28181239
- 6. Fleck G, Bechly G, Martínez-Delclòs X, Jarzembowski E, Coram R, et al. (2003) Phylogeny and classification of the Stenophlebioptera (Odonata: Epiproctophora). Ann Soc Ent France (n.s.) 39: 55–93.G. FleckG. BechlyX. Martínez-DelclòsE. JarzembowskiR. Coram2003Phylogeny and classification of the Stenophlebioptera (Odonata: Epiproctophora).Ann Soc Ent France (n.s.)395593
- 7. Grimaldi D, Engel MS (2005) Evolution of the insects. Cambridge: Cambridge University press. 755 p.D. GrimaldiMS Engel2005Evolution of the insects.CambridgeCambridge University press755
- 8. Dumont HJ, Vierstraete A, Vanfleteren JR (2010) A molecular phylogeny of the Odonata (Insecta). Syst Ent 35: 6–18.HJ DumontA. VierstraeteJR Vanfleteren2010A molecular phylogeny of the Odonata (Insecta).Syst Ent35618
- 9. Fleck G, Ullrich B, Brenk M, Wallnisch C, Orland M, et al. (2008) A phylogeny of anisopterous dragonflies (Insecta, Odonata) using mtRNA genes and mixed nucleotide/doublet models. J Zool Syst Evol Res 46(4): 310–322.G. FleckB. UllrichM. BrenkC. WallnischM. Orland2008A phylogeny of anisopterous dragonflies (Insecta, Odonata) using mtRNA genes and mixed nucleotide/doublet models.J Zool Syst Evol Res464310322
- 10. Bybee SM, Ogden TH, Branham MA, Whiting MF (2008) Molecules, morphology and fossils: a comprehensive approach to odonate phylogeny and the evolution of the odonate wing. Cladistics 23: 1–38.SM BybeeTH OgdenMA BranhamMF Whiting2008Molecules, morphology and fossils: a comprehensive approach to odonate phylogeny and the evolution of the odonate wing.Cladistics23138
- 11. Hovmöller R, Pape T, Källersjö , M (2002) The Palaeoptera Problem: Basal Pterygote Phylogeny Inferred from 18 S and 28 S rDNA Sequences. Cladistics 18: 313–323.R. HovmöllerT. PapeKällersjöM2002The Palaeoptera Problem: Basal Pterygote Phylogeny Inferred from 18 S and 28 S rDNA Sequences.Cladistics18313323
- 12. Asahina A (1954) A morphological study of a relic dragonfly, Epiophlebia superstes Sélys (Odonata, Anisozygoptera). Tokyo: The Japan Society for the Pormotion of Science. 153 p.A. Asahina1954A morphological study of a relic dragonfly, Epiophlebia superstes Sélys (Odonata, Anisozygoptera).TokyoThe Japan Society for the Pormotion of Science153
- 13. Tabaru N (1984) Larval development of Epiophlebia superstes in Kyushu. Tombo 27: 27–31.N. Tabaru1984Larval development of Epiophlebia superstes in Kyushu.Tombo272731
- 14. Tillyard RJ (1921) On an Anisozygopterous larva from the Himalayas (Order Odonata). Records of the Indian Museum 22: 93–107.RJ Tillyard1921On an Anisozygopterous larva from the Himalayas (Order Odonata).Records of the Indian Museum2293107
- 15. Furukawa H (1934) On the nymph of a dragonfly Epiophlebia superstes Sélys. Annot zool Jpn 14: 451–456.H. Furukawa1934On the nymph of a dragonfly Epiophlebia superstes Sélys.Annot zool Jpn14451456
- 16. Asahina A (1961) Is Epiophlebia laidlawi Tillyard (Odonata, Anisozygoptera) a good species? Int Revue ges Hydrobiol 46: 441–446.A. Asahina1961Is Epiophlebia laidlawi Tillyard (Odonata, Anisozygoptera) a good species?Int Revue ges Hydrobiol46441446
- 17. Asahina A (1963) Description of the possible adult dragonfly of Epiophlebia laidlawi from the Himalayas. Tombo 6: 18–20.A. Asahina1963Description of the possible adult dragonfly of Epiophlebia laidlawi from the Himalayas.Tombo61820
- 18. Li J-K, Nel A, Zhang X-P, Fleck G, Gao M-X, et al. (2011) A third species of the relict family Epiophlebiidae discovered in China (Odonata: Epiproctophora). Syst Ent. J-K LiA. NelX-P ZhangG. FleckM-X Gao2011A third species of the relict family Epiophlebiidae discovered in China (Odonata: Epiproctophora).Syst Ent
- 19. Nesemann H, Shah RDT, Shah DN, Sharma S (2011) Morphological characters of Epiophlebia laidlawi Tillyard larvae, with notes on the habitat and distribution of the species in Nepal (“Anisozygoptera”: Epiophlebiidae). Odonatologica 40: 191–202.H. NesemannRDT ShahDN ShahS. Sharma2011Morphological characters of Epiophlebia laidlawi Tillyard larvae, with notes on the habitat and distribution of the species in Nepal (“Anisozygoptera”: Epiophlebiidae).Odonatologica40191202
- 20. de Sélys Longchamps E (1889) Palaeophlebia. Nouvelle légion de Caloptérygines. Suivi de la description d’une nouvelle gomphine du Japon: Tachopteryx Pryeri. Ann Soc Ent Belgique 33: 153–159.E. de Sélys Longchamps1889Palaeophlebia. Nouvelle légion de Caloptérygines. Suivi de la description d’une nouvelle gomphine du Japon: Tachopteryx Pryeri.Ann Soc Ent Belgique33153159
- 21. Brockhaus T, Hartmann A (2009) New records of Epiophlebia laidlawi Tillyard in Bhutan, with notes on its biology, ecology, distribution, biogeography and threat status (Anisozygoptera: Epiophlebiidae). Odonatologica 38: 203–215.T. BrockhausA. Hartmann2009New records of Epiophlebia laidlawi Tillyard in Bhutan, with notes on its biology, ecology, distribution, biogeography and threat status (Anisozygoptera: Epiophlebiidae).Odonatologica38203215
- 22. Hasegawa E, Kasuya E (2006) Phylogenetic analysis of the insect order Odonata using 28S and 16S rDNA sequences: a comparison between data sets with different evolutionary rates. Ent Sci 9: 55–66.E. HasegawaE. Kasuya2006Phylogenetic analysis of the insect order Odonata using 28S and 16S rDNA sequences: a comparison between data sets with different evolutionary rates.Ent Sci95566
- 23. Carle FL, Kjer KM, May ML (2008) Evolution of Odonata, with special reference to Coenagrionoidea (Zygoptera). Arthropod Syst Phyl 66(1), 37–44: FL CarleKM KjerML May2008Evolution of Odonata, with special reference to Coenagrionoidea (Zygoptera).Arthropod Syst Phyl 66(1),37–44
- 24. Hummel S (2003) Ancient DNA Typing: Methods, Strategies and Applications. Berlin: Springer. 310 p.S. Hummel2003Ancient DNA Typing: Methods, Strategies and Applications.BerlinSpringer310
- 25. Kjer KM, Carle FL, Litman J, Ware J (2006) A molecular phylogeny of Hexapoda. Arthropod Syst Phyl 64(1): 35–44.KM KjerFL CarleJ. LitmanJ. Ware2006A molecular phylogeny of Hexapoda.Arthropod Syst Phyl6413544
- 26. Regier JC, Zwick A (2011) Sources of Signal in 62 Protein-Coding Nuclear Genes for Higher-Level Phylogenetics of Arthropods. PLOSone 6(8): e23408.JC RegierA. Zwick2011Sources of Signal in 62 Protein-Coding Nuclear Genes for Higher-Level Phylogenetics of Arthropods.PLOSone68e23408
- 27. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acid Res 30(14): 3059–3066.K. KatohK. MisawaK. KumaT. Miyata2002MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform.Nucleic Acid Res301430593066
- 28. Davies A (1992) Epiophlebia laidlawi – Flying! Kimminsia, 3(2): 10–11.A. Davies1992Epiophlebia laidlawi – Flying! Kimminsia,321011
- 29. Xylander WER, Günther KK (2003) 7. Ordnung Odonata, Libellen. In: Dathe HH, editor. Lehrbuch der Speziellen Zoologie, Band I, 5. Teil: Insecta. Berlin: Spektrum Akademischer Verlag, 167–172: WER XylanderKK Günther20037. Ordnung Odonata, Libellen. In: Dathe HH, editor. Lehrbuch der Speziellen Zoologie, Band I, 5. Teil: Insecta.Berlin: Spektrum Akademischer Verlag,167–172
- 30. Carpenter FM (1992) Superclass Hexapoda. In: Moore RC, Kaesler RL, editors. Treatise on invertebrate paleontology Part R Arthropoda 4 Volume 4. Boulder: The Geological Society of America. pp. 59–89.FM Carpenter1992Superclass Hexapoda.RC MooreRL KaeslerTreatise on invertebrate paleontology Part R Arthropoda 4 Volume 4BoulderThe Geological Society of America5989
- 31. Ikeda H, Ohba H (1998) Himalayan Potentilla and its Relation to the Sino-Japanese Floristic Region. In: Boufford E, Ohba H, editors. H. IkedaH. Ohba1998Himalayan Potentilla and its Relation to the Sino-Japanese Floristic Region.E. BouffordH. OhbaSino-Japanese flora its characteristics and diversification Bulletin No. 37. Tokyo: The University Museum The University of Tokyo. Available http://www.um.u-tokyo.ac.jp/publish_db/Bulletin/no37/no37010.html. Accessed 18 January 2012. Sino-Japanese flora its characteristics and diversification Bulletin No. 37. Tokyo: The University Museum The University of Tokyo. Available http://www.um.u-tokyo.ac.jp/publish_db/Bulletin/no37/no37010.html. Accessed 18 January 2012.
- 32. Millien-Parra V, Jaeger J-J (1999) Island biogeography of the Japanese terrestrial mammal assemblages: an example of a relict fauna. J Biogeogr 26: 959–972.V. Millien-ParraJ-J Jaeger1999Island biogeography of the Japanese terrestrial mammal assemblages: an example of a relict fauna.J Biogeogr26959972
- 33. Clark PU, Dyke AS, Shakun JD, Carlson AE, Clark J, et al. (2009) The last glacial maximum. Science 325: 710–714.PU ClarkAS DykeJD ShakunAE CarlsonJ. Clark2009The last glacial maximum.Science325710714
- 34. Šegota T (1973) Radiocarbon measurements and the Holocene and late Würm sealevel rise. Eiszetalter und Gegenwart 23/ 24: 107–115.T. Šegota1973Radiocarbon measurements and the Holocene and late Würm sealevel rise.Eiszetalter und Gegenwart 23/24107115
- 35. Tushingham AM, Peltier WR (1993) Implications of the radio carbon timescale for ice-sheet chronology and sea-level change. Quat. Res. (N.Y.) 39: 125–129.AM TushinghamWR Peltier1993Implications of the radio carbon timescale for ice-sheet chronology and sea-level change. Quat. Res.(N.Y.)39125129
- 36. Li Z, He Y, An W, Song L, Zhang W, et al. (2011) Climate and glacier change in southeastern China during the past several decades. Environ Res Lett 6: 045404.Z. LiY. HeW. AnL. SongW. Zhang2011Climate and glacier change in southeastern China during the past several decades.Environ Res Lett6045404
- 37. Bernard R, Heiser M, Hochkirch A, Schmitt , T (2011) Genetic homogeneity of the Sedgling Nehalennia speciosa (Odonata: Coenagrionidae) indicates a single Würm glacial refugium and trans-Palaearctic postglacial expansion. J Zool Syst Evol Res. R. BernardM. HeiserA. HochkirchSchmittT2011Genetic homogeneity of the Sedgling Nehalennia speciosa (Odonata: Coenagrionidae) indicates a single Würm glacial refugium and trans-Palaearctic postglacial expansion.J Zool Syst Evol Res
- 38. Ho SWY, Lanfear R, Phillips MJ, Barnes I, Thomas JA, et al. (2011) Bayesian estimation of substitution rates from ancient DNA sequences with low information content. Syst Biol 60: 366–375.SWY HoR. LanfearMJ PhillipsI. BarnesJA Thomas2011Bayesian estimation of substitution rates from ancient DNA sequences with low information content.Syst Biol60366375
- 39. Friedrich M, Tautz D (1997) An episodic change of rDNA nucleotide substitution rates has occurred during the emergence of the insect order Diptera. Mol Biol Evol 14: 644–653.M. FriedrichD. Tautz1997An episodic change of rDNA nucleotide substitution rates has occurred during the emergence of the insect order Diptera.Mol Biol Evol14644653
- 40. Tamariz J, Voynarosvka K, Prinz M, Caragine T (2006) The application of ultraviolet irradiation to exogenous sources of DNA in plasticware and water for the amplification of low copy number DNA. J Forensic Sci 51: 790–794.J. TamarizK. VoynarosvkaM. PrinzT. Caragine2006The application of ultraviolet irradiation to exogenous sources of DNA in plasticware and water for the amplification of low copy number DNA.J Forensic Sci51790794
- 41. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA 5: Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance and maximum parsimony methods. Mol Biol Evol 28: 2731–2739.K. TamuraD. PetersonN. PetersonG. StecherM. Nei2011MEGA 5: Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance and maximum parsimony methods.Mol Biol Evol2827312739
- 42. Katoh K, Asimenos G, Toh H (2009) Multiple Alignment of DNA Sequences with MAFFT. In Posada D (ed.): Bioinformatics for DNA Sequence Analysis. Methods Mol Biol 537: 39–64.K. KatohG. AsimenosH. Toh2009Multiple Alignment of DNA Sequences with MAFFT. In Posada D (ed.): Bioinformatics for DNA Sequence Analysis.Methods Mol Biol5373964
- 43. Nylander JAA (2004) MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. JAA Nylander2004MrModeltest v2.Program distributed by the author. Evolutionary Biology Centre, Uppsala University
- 44. Swofford DL (2001) PAUP*: Phylogenetic analysis using parsimony and other methods, version 4.0b10. Computer program distributed by Sinauer Associates Inc., Sunderland, MA. DL Swofford2001PAUP*: Phylogenetic analysis using parsimony and other methods, version 4.0b10.Computer program distributed by Sinauer Associates Inc., Sunderland, MA
- 45. Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: 754–755.JP HuelsenbeckF. Ronquist2001MRBAYES: Bayesian inference of phylogeny.Bioinformatics17754755
- 46. Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574.F. RonquistJP Huelsenbeck2003MRBAYES 3: Bayesian phylogenetic inference under mixed models.Bioinformatics1915721574
- 47. Hastings DA, Dunbar PK, Elphingstone GM, Bootz M, Murakami H, et al., editors. (1999) DA HastingsPK DunbarGM ElphingstoneM. BootzH. Murakami1999The Global Land One-kilometre Base Elevation (GLOBE) Digital Elevation Model, Version 1.0. Available: http://www.ngdc.noaa.gov/mgg/topo/globe.html. Accessed 25 October 2011. The Global Land One-kilometre Base Elevation (GLOBE) Digital Elevation Model, Version 1.0. Available: http://www.ngdc.noaa.gov/mgg/topo/globe.html. Accessed 25 October 2011.
- 48. Freeland JR, May M, Lodge R, Conrad KF (2003) Genetic diversity and widespread haplotypes in a migratory dragonfly, the common green darner Anax junius. Ecol Entomology 28: 413–421.JR FreelandM. MayR. LodgeKF Conrad2003Genetic diversity and widespread haplotypes in a migratory dragonfly, the common green darner Anax junius.Ecol Entomology28413421