FNZ 64 - Pisauridae (Arachnida: Araneae) - Phylogenetic analysis
Vink, CJ; Dupérré, N 2010. Pisauridae (Arachnida: Araneae). Fauna of New Zealand 64, 60 pages.
(
ISSN 0111-5383 (print),
ISSN 1179-7193 (online)
;
no.
64.
ISBN 978-0-478-34722-7 (print),
ISBN 978-0-478-34723-4 (online)
).
Published 13 Jul 2010
ZooBank: http://zoobank.org/References/05FA0262-846F-4216-96B4-D5ED10D3A7D9
Phylogenetic analysis
Methods. Sequences were edited and aligned using Sequencher 4.6 (Gene Codes Corporation). There was no evidence of insertions/deletions or stop codons in the coding sequences or the actin 5C intron and alignment was straight forward. Uncorrected pairwise distances were calculated using PAUP* version 4.0b10 (Swofford 2002). Partitioned Bayesian analyses implemented in MrBayes version 3.1.2 (Ronquist & Huelsenbeck 2003) following the methods of Brandley et al. (2005) were used to estimate the COI phylogenetic tree topology. A partitioned analysis was not used for actin 5C as the only variation occurred in the 3rd codon positions. COI and actin 5C sequences for 27 specimens were combined into one dataset and also analysed with partitioned analyses in MrBayes. MrModeltest version 2.2 (Nylander 2005) implemented in PAUP* version 4.0b10 (Swofford 2002) was used to select the model parameters. Within MrModeltest the Akaike Information Criterion was used for model selection (Posada & Buckley 2004). The COI data were partitioned by codon, using the model HKY (Hasegawa et al. 1985) for the 2nd codon positions and HKY+Γ (Hasegawa et al. 1985, Yang 1994) for the 3rd codon positions. All COI 1st codon positions were identical and were excluded from the analyses. The model F81 (Felsenstein 1981) was applied to the actin 5C data. The same partitions and their models were applied to the combined dataset. Bayesian analyses were conducted by running two simultaneous, completely independent analyses each with four heated chains, sampling every 1000th tree. The analyses were run for at least 2 × 107 generations until the average standard deviation of split frequencies had dropped below 0.005, which indicated that the two tree samples had converged. MrBayes was used to construct majority rule consensus trees, discarding the first 25% of trees generated as burn-in. TreeView 1.6.6 (Page 1996) was used to view and save trees in graphic format.
Results. Thirty-seven COI haplotypes occurred among the 58 specimens of New Zealand Dolomedes; 17 of D. minor, 11 of D. aquaticus, 7 of D. dondalei, and 2 of D. schauinslandi. One COI haplotype was found in both D. aquaticus and D. minor specimens.
Uncorrected pairwise distances between COI sequences of D. minor, D. aquaticus, D. dondalei, and D. schauinslandi are shown in Table 2. Sixty of the 850 COI nucleotide sites were variable. Only one nucleotide change was nonsynonymous, which was a transition in specimen D034 (D. dondalei).
Of the 935 nucleotides of actin 5C data, there were five variable sites, three of which were phylogenetically informative, and the other two sites were heterozygous in some specimens. All variable sites occurred in the exon. There were also three other sites that were heterozygous in some species. All nucleotide changes were synonymous, which was expected as nonsynonymous substitutions have not been observed in actin 5C across all Araneae (Vink et al. 2008a). There were no insertions or deletions in the actin 5C intron.
The phylogenetic analysis of the COI data (Text-fig. 1) showed that D. schauinslandi and D. dondalei are monophyletic with posterior probabilities of 1 and 0.96 respectively. Dolomedes aquaticus was also monophyletic with the exception of two D. minor haplotypes, one of which was identical to D. aquaticus. Within D. minor, D. aquaticus, and D. dondalei, there were some geographically related groups.
The phylogenetic analysis of the actin 5C data (Text-fig. 2) supported the monophyly of D. aquaticus and D. dondalei. All four species have species-specific nucleotide combinations at positions 348, 459, and 1038 of the actin 5C coding sequence.
The combined phylogenetic analysis of the COI and actin 5C data (Text-fig. 3) supported the monophyly of D. aquaticus and D. dondalei. Dolomedes minor formed three geographically related monophyletic clades; northern North Island, southern North Island plus northern South Island, and South Island.
Discussion. Seven D. minor specimens (D005, D024, D018, D019, D022, D025, D040 –Table 1) have mitochondrial DNA (mtDNA) sequences (COI) that are the same as or closely related to those of D. aquaticus. However, the morphology and nuclear DNA sequences (actin 5C) of these seven D. minor specimens are consistent with those of other D. minor specimens. This suggests relatively recent or on-going introgression that resulted from interspecific hybridisation. Introgression of mtDNA has been noted in other spiders (Johannesen & Veith 2001, Croucher et al. 2004, Vink et al. 2008b, Hedin & Lowder 2009) including species in the sister family Lycosidae (Chang et al. 2007). Both D. minor and D. aquaticus males may prefer to mate with newly moulted, virgin females to avoid sexual cannibalism, which has also been documented as a strategy in other Dolomedes species (Arnqvist 1992, Johnson 2001). This could result in interspecific hybridisation as morphological or behavioural mechanisms that might normally prevent mating between D. minor and D. aquaticus may not be as effective in newly moulted, virgin females. It would seem that the introgressions observed have resulted from at least two accidental matings between male D. minor and female D. aquaticus. The likely scenario is that the offspring that resulted from interspecific hybridisation retained more D. minor characteristics than those of D. aquaticus and backcrossed with D. minor. All D. minor specimens that we sampled from five different locations in Central Otago, Southland, and Fiordland (Table 1) had D. aquaticus COI sequences. It is possible that all D. minor in these regions have D. aquaticus mtDNA. It would be interesting to obtain COI sequence data from D. aquaticus in Central Otago, Southland, and Fiordland, which we would predict would be the same as, or close to, D. minor COI haplotypes in these regions.
All four New Zealand Dolomedes species have shared haplotypes among specimens from geographically distant localities, which indicates that they disperse over some geographic boundaries. However, long-range dispersal seems to be limited or uncommon, as there is some genetic structure linked to geographic locations within each species.
Interspecific divergences in COI between New Zealand Dolomedes species are very low (Table 2) and lower than the divergences in 99.6% of 1249 closely related chelicerate species surveyed by Hebert et al. (2003b). However, some Japanese Dolomedes species differed by only 2% (Tanikawa & Miyashita 2008). Some specimens of D. minor differed from D. aquaticus and D. dondalei by only 0.08%, which is as low as the intraspecific variation observed in some spider species (Vink et al. 2009). COI sequence data is not variable enough to study population structure in New Zealand Dolomedes species; however, microsatellites (Ji et al. 2004) could be used to provide the resolution needed.
Dolomedes schauinslandi are 2.2-3.3% divergent in COI from mainland species. Brower (1994) estimated a rate of 2.3% pairwise divergence per million years in mtDNA in arthropods, which also appears to apply to spiders (e.g., Hedin 2001). This would indicate that D. schauinslandi and D. minor diverged approximately one million years ago, which is lower than other divergences (2–6 Ma) between Chatham Island arthropod species and mainland species (Trewick 2000) including the closely related Lycosidae (Vink & Paterson 2003). It is possible that mtDNA in Pisauridae evolves at a slower rate than in other spiders.
Both the mtDNA introgression and the small genetic distances between species, mean that New Zealand Dolomedes species cannot be reliably identified using a DNA barcoding approach (Hebert et al. 2003a) and highlights the importance of using a number of different datasets in the delimitation of species-group taxa.