Research Article |
Corresponding author: Peter Decker ( decker@myriapoden-info.de ) Academic editor: Robert Mesibov
© 2016 Peter Decker.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Decker P (2016) Phylogenetic analysis of the Australian trans-Bass Strait millipede genus Pogonosternum (Carl, 1912) (Diplopoda, Polydesmida, Paradoxosomatidae) indicates multiple glacial refugia in southeastern Australia. ZooKeys 578: 15-31. https://doi.org/10.3897/zookeys.578.8052
|
This study documents the first detailed phylogenetic analysis of an Australian paradoxosomatid millipede genus. Two mitochondrial genes (partial COI and 16S) as well as partial nuclear 28S rDNA were amplified and sequenced for 41 individuals of the southeastern Australian genus Pogonosternum Jeekel, 1965. The analysis indicates that five species groups of Pogonosternum occur across New South Wales, Victoria and Tasmania: P. nigrovirgatum (Carl, 1912), P. adrianae Jeekel, 1982, P. laetificum Jeekel, 1982 and two undescribed species. P. coniferum (Jeekel, 1965) specimens cluster within P. nigrovirgatum. Most of these five species groups exhibit a pattern of high intraspecific genetic variability and highly localized haplotypes, suggesting that they were confined to multiple Pleistocene refugia on the southeastern Australian mainland. The phylogenetic data also show that northwestern Tasmania was colonized by P. nigrovirgatum, probably from central Victoria, and northeastern Tasmania by an as yet undescribed species from eastern Victoria.
Invertebrate, COI, 16S, 28S, genetic variability
Pogonosternum Jeekel, 1965 is the most widespread and species-rich genus of the millipede tribe Antichiropodini Brölemann, 1916 in Victoria, with the five described species Pogonosternum nigrovirgatum (Carl, 1902), P. coniferum Jeekel, 1965, P. adrianae Jeekel, 1982, P. laetificum Jeekel, 1982 and the subspecies P. nigrovirgatum infuscum Jeekel, 1982, all hitherto recorded from Victoria only. However,
Thus, Pogonosternum occurs on both sides of Bass Strait, which separates mainland Australia from Tasmania. The paradoxosomatid genus Somethus Chamberlin, 1920 also has a trans-Bass Strait distribution (
Many soil invertebrates, including millipedes, have limited active dispersal capabilities. Phylogenetic studies of southeastern Australian soil invertebrates can give important insights into the impact of glacial periods during the Pleistocene (
The present study documents a molecular phylogenetic analysis of the antichiropodine genus Pogonosternum, using specimens from across the genus range, and with molecular evidence indicating past isolation in multiple Pleistocene refugia. Finally, the identity and origin of Tasmanian Pogonosternum populations are clarified.
Pogonosternum specimens were collected by hand in Victoria and New South Wales in August 2014 by the author, Karin Voigtländer and Robert Mesibov, and by Mesibov in Tasmania in May 2014 and May 2015 (Fig.
Site numbers, localities, GenBank accession numbers and repository accession numbers for all specimens analyzed. (See also Fig.
Species | Site No. | Locality | GenBank Acc. No. COI | GenBank Acc. No. 16S | GenBank Acc. No. 28S | Voucher |
---|---|---|---|---|---|---|
Outgroup | ||||||
Somethus scopiferus Jeekel, 2002 | SA, Martin Washpool Conservation Park | KT948674 | KU833272 |
|
||
Somethus castaneus (Attems, 1944) | SA, Adelaide, Upper Sturt | KT964477 | SAM OM2135 | |||
Archicladosoma magnum Jeekel, 1984 | VIC, N Rawson | KT948681 | KU833273 |
|
||
Ingroup | ||||||
Pogonosternum adrianae | S58 | VIC, S Dargo | KU745235 | KU745194 | KU745185 |
|
Pogonosternum adrianae | S59 | VIC, W Balook | KU745236 | KU745195 |
|
|
Pogonosternum adrianae | S62 | VIC, NE Moe | KU745237 | KU745196 | KU745186 |
|
Pogonosternum coniferum | S67 | VIC, Langwarrin | KU745238 | KU745197 |
|
|
Pogonosternum coniferum | S71 | VIC, NE Cape Schanck | KU745239 | KU745198 |
|
|
Pogonosternum laetificum | S2 | VIC, NE Tyaak | KU745240 | KU745199 |
|
|
Pogonosternum laetificum | S5 | VIC, SE Glenburn | KU745241 | KU745200 |
|
|
Pogonosternum laetificum | S7 | VIC, E Toolangi | KU745242 | KU745201 |
|
|
Pogonosternum laetificum | S9 | VIC, SE Healesville | KU745243 | KU745202 |
|
|
Pogonosternum laetificum | S14 | VIC, SE Narbethong | KU745244 | KU745203 | KU745187 |
|
Pogonosternum laetificum | S15 | VIC, E Narbethong | KU745245 | KU745204 |
|
|
Pogonosternum laetificum | S17 | VIC, N Marysville | KU745246 | KU745205 |
|
|
Pogonosternum laetificum | S18 | VIC, S Eildon | KU745247 | KU745206 |
|
|
Pogonosternum laetificum | S19 | VIC, W Barjarg | KU745248 | KU745207 |
|
|
Pogonosternum laetificum | S88 | VIC, Mt Macedon | KU745249 | KU745208 |
|
|
Pogonosternum nigrovirgatum | S60 | VIC, SE Traralgon South | KU745250 | KU745209 | KU745188 |
|
Pogonosternum nigrovirgatum | S63 | VIC, SW Trafalgar | KU745251 | KU745210 |
|
|
Pogonosternum nigrovirgatum | S64 | VIC, W Nyora | KU745252 | KU745211 |
|
|
Pogonosternum nigrovirgatum | S65 | VIC, SE The Gurdies | KT948680 | KU745212 | KT964478 |
|
Pogonosternum cf. nigrovirgatum | S77 | VIC, NW Lorne | KU745253 | KU745213 |
|
|
Pogonosternum cf. nigrovirgatum | S78 | VIC, W Kennett River | KU745254 | KU745214 |
|
|
Pogonosternum cf. nigrovirgatum | S81 | VIC, N Apollo Bay | KU745255 | KU745215 | KU745189 |
|
Pogonosternum cf. nigrovirgatum | S83 | VIC, SW Staughton Vale | KU745256 | KU745216 |
|
|
Pogonosternum nigrovirgatum | S87 | VIC, W Gisborne | KU745257 | KU745217 |
|
|
Pogonosternum cf. nigrovirgatum | X2 | TAS, S West Montagu | KU745258 | KU745218 | QVMAG:2015:23:1 | |
Pogonosternum sp. A | S21 | VIC, N Glenrowan | KU745259 | KU745219 |
|
|
Pogonosternum sp. A | S22 | VIC, NE Thoona I | KU745260 | KU745220 |
|
|
Pogonosternum sp. A | S23 | VIC, NE Thoona II | KU745261 | KU745221 |
|
|
Pogonosternum sp. A | S24 | VIC, SE Chiltern | KU745262 | KU745222 |
|
|
Pogonosternum sp. A | S25 | VIC, SSW Chiltern | KU745263 | KU745223 | KU745190 |
|
Pogonosternum sp. A | S31 | NSW, E Talbingo I | KU745264 | KU745224 |
|
|
Pogonosternum sp. A | S32 | NSW, E Talbingo II | KU745265 | KU745225 |
|
|
Pogonosternum sp. A | S42 | VIC, NNW Bemm River | KU745266 | KU745226 |
|
|
Pogonosternum sp. A | S47 | VIC, E Orbost | KU745267 | KU745227 |
|
|
Pogonosternum sp. A | S49 | VIC, Buchan | KU745268 | KU745228 |
|
|
Pogonosternum sp. A | S52 | VIC, SW Nowa Nowa | KU745269 | KU745229 |
|
|
Pogonosternum sp. A | X1 | TAS, W Tomahawk | KU745270 | KU745230 | KU745191 |
|
Pogonosternum sp. B | S26 | NSW, SE Holbrook | KU745271 | KU745231 |
|
|
Pogonosternum sp. B | S27 | NSW, W Tumbarumba | KU745272 | KU745232 |
|
|
Pogonosternum sp. B | S28 | NSW, NNE Tumbarumba | KU745273 | KU745233 | KU745192 |
|
Pogonosternum sp. B | S29 | NSW, SE Batlow | KU745274 | KU745234 | KU745193 |
|
Maps were created with ArcGIS 10. The final phylogenetic trees were edited using Adobe Illustrator CS4.
DNA was extracted from 2-4 legs from each of 41 Pogonosternum specimens and from the three paradoxosomatid species Archicladosoma magnum Jeekel, 1984, Somethus scopiferus Jeekel, 2002 and S. castaneus (Attems, 1944), which were chosen as outgroups (Table
Glom primer cocktail pairs (
For PCR protocol and all primer sequences (COI, 28S) see
Primer pairs 16Sar (Fw) (5’-CGCCTGTTTAACAAAAACAT-3’) and 16Sbr (Rv) (5’-CCGGTCTGAACTCAGATCACGT-3’) (
All PCR mixes had a total volume of 10 µl comprising 1 µl template, 0.2 µM of each primer, 4x0.2 mM dNTPs [Peqlab], 1 x PCR Buffer containing 1.5 mM MgCl2 [Peqlab], and 0.05u Polymerase [Peqlab].
All fragments were sequenced in both directions by the BiK-F Laboratory Centre, Frankfurt, Germany. All obtained sequences were checked via BLAST searches of GenBank; no contamination was discovered. The sequences were aligned by hand in ClustalX ver. 1.83 (
Some homologisation problems in the 16S rRNA sequences arose mainly because of the highly variable expansion loops. As a result, selected alignment positions (272-297) were excluded from the 16S rRNA dataset for all further analyses using MEGA6.
The final alignments consisted of 618 bp of COI mtDNA, 540 bp of 16S rRNA and 1206 bp of 28S rRNA. The combined datasets after these exclusions consisted of 1158 bp for COI+16S. Individual partial alignments can be obtained from the author upon request. The alignment of the combined dataset can be found in the Suppl. material
COI and 16S sequences were combined as a single dataset and incongruence assessed between the mtDNA intergenic spacer sequences with the incongruence length difference (ILD) test (
The combined dataset of COI and 16S was analysed under maximum likelihood (ML) using MEGA6 (
Multiple runs of ML and BI converged in trees with the same topology and similar likelihood score so that only the result of the first run is presented. The topology resulting from ML and BI analyses was largely congruent except for the arrangements of several terminal nodes with low support. Thus, results from the ML and BI analyses are shown together based on the ML tree with bootstrap (BP) and posterior probabilities (PP) of the major lineages shown on the corresponding branches with BP values > 70 (Fig.
Maximum likelihood tree for the combined mitochondrial COI+16S dataset, 1000 bootstrap replicates, values below 70 not shown. The bootstrap values of ML and posterior probabilities of BI are given above and below the corresponding branches, respectively, for all major clades. Scale bar = substitutions per site. Coloured blocks indicate species groups. Color of branches refers to the major subregions shown in the map, Tasmanian branches thicker. General differences in male gonopod morphology are shown by sketches of the apical region of the right gonopod not drawn to scale. Coloured lines link those analysed specimens that have similar gonopod morphology. Posterior view = post.; lateral view = lat.; anterior view = ant.
An appropriate DNA substitution model was determined for 28S under the Bayesian Information Criterion (BIC) in Modeltest implemented in MEGA 6 (
A phylogenetic hypothesis was inferred for COI+16S and 28S by using the maximum likelihood method conducted in MEGA6 (
Mean uncorrected pairwise distances between terminals (transformed into percentages) were determined using MEGA6 (
The monophyly of the genus Pogonosternum is strongly supported (MLBP = 97; BIPP = 1.0) in the mitochondrial tree and shows five clades within Pogonosternum, resembling five species groups (Fig.
One main clade includes three species from the mountainous area east and northeast of Melbourne: the undescribed species Pogonosternum sp. B (MLBP = 99; BIPP = 1.0), already mentioned by
Pogonosternum nigrovirgatum sensu lato with a trans-Bass Strait distribution formed a well-supported (MLBP = 89; BIPP = 1.0) sister clade to the new species P. sp. A (MLBP = 98; BIPP = 1.0) that also has a trans-Bass Strait distribution. Pogonosternum sp. A also occurs in New South Wales (
Within the P. nigrovirgatum s. l. species-group, the greatest genetic distances were observed between populations in the Strzelecki Ranges (S60, S63; MLBP = 100; BIPP = 1.0) and more western populations, with values ranging from 5.0 to 6.8%. Specimens from the Otway Ranges (S77, S78, S81) all formed a well-supported cluster (MLBP = 86; BIPP = 1.0). The Tasmanian specimen (X2) was distinct from both the Strzelecki Ranges (5.4–6.0%) and central and western Victorian specimens (3.7–3.8%). In the case of Pogonosternum sp. A the largest distances (4.2–5.8%) were between the Eastern Gippsland populations (S42, S47; MLBP = 100; BIPP = 1.0) and all other specimens. The status of the northeast Tasmanian specimen is not well resolved; it is closest to a population from Kosciuszko National Park (S31, 3.0%), the two forming a poorly supported sister clade with a specimen from Gippsland (S52; MLBP = 55; BIPP = 0.6).
All species show considerable intraspecific genetic distances and high phylogeographic structure, especially P. laetificum, and, except in the case of P. adrianae, no haplotypes are shared between different populations. Additional one to three sequenced specimens from eight sampling sites (S14, S15, S22, S58, S59, S78, S83, S87) always showed the same haplotype in Pogonosternum (data not published).
Interspecific distances within the genus Pogonosternum are moderately large, varying from 5.5% (P. sp. A–P. nigrovirgatum s. l.) to 10.4% (P. nigrovirgatum s. l.–P. laetificum), except P. adrianae to P. laetificum with only 2.9%.
Owing to the general lack of variability within the nuclear 28S rRNA dataset, the phylogenetic relationships among species were largely unresolved. Distances for 28S rRNA within Pogonosternum are very low, with a maximum of three base pair differences noted for P. sp. B (Fig.
In a separate paper (Decker, in preparation), the morphology of the Pogonosternum species groups is described in detail and new species are described, based on the specimens used here and from ca 130 additional localities. Here I note briefly that several common morphological features were observed in the gonopods of P. nigrovirgatum s. l., P. laetificum, and P. sp. A: some specimens also showed intermediate states of those features (Fig.
Surprisingly, gonopod morphology did not appear to agree well with the phylogenetic tree (Fig.
The mitochondrial tree (Fig.
The 28S tree shows little or only little resolution at the species level (Fig.
With the exception of P. adrianae and P. sp. B, Pogonosternum species show significant variability in gonopod form, with local morphs occurring throughout each species’ distribution area.
Interestingly, P. adrianae is morphologically distinct (in size, spiracles, male tibiotarsal brushes and gonopods, female coxal process) from P. laetificum despite their close genetic distance.
Gonopod variability was also documented for some species of Somethus in South Australia (
This study has shown that in the area of southern and southeast Australia, there are at least two genera, Pogonosternum and Somethus (
The results indicate that there is high intraspecific genetic divergence, with high genetic distances and haplotype diversity in the mitochondrial genes between populations of Pogonosternum, even those adjacent to each other. The P. laetificum clade, which has been sampled extensively in the Central Highlands, shows particularly high intraspecific genetic differences (mean genetic distance of 3.9%), apparently without corresponding geographic patterning, or morphological variation (Decker, in preparation).
The phylogenetic patterns with high intraspecific divergence, high genetic distances, and haplotype diversity with unique local haplotypes, resulting in long branches, shown by Pogonosternum, indicate multiple Pleistocene refugia according to
The phylogenetic patterns shown by Pogonosternum suggest that in Victoria and New South Wales there were large areas with multiple local refugia during the Pleistocene. No region in the study area on mainland Australia showed results which indicate rapid postglacial resettlement of Pogonosternum.
Evidence for multiple glacial refugia was also identified in the spirostreptidan millipede Atelomastix bamfordi Edward & Harvey, 2010 in Western Australia (
However, further studies on genetic and morphological variability on a finer geographical scale could lead to a better understanding of the pattern and impact of isolation in multiple glacial refugia during the Pleistocene, also as an evolutionary driving force for morphological variability in some species.
There is a notable high genetic distance gap within P. nigrovirgatum sensu lato between specimens from the Strzelecki Ranges (S60, S63), West Gippsland, and those sampled in the central and western regions in Victoria, but some specimens of adjacent populations from the latter (S64, S65) were morphologically indistinguishable from specimens from the Strzelecki Ranges. A similar genetic gap was observed in P. sp. A for the populations in Eastern Gippsland east of Orbost (S42, S47) and all other populations. These two cases indicate that these areas may have been isolated for long periods from neighboring regions, possibly before the Pleistocene, perhaps during a marine incursion in the Gippsland Basin and other parts of southeast Australia close to the Miocene–Pliocene boundary (
The genus Pogonosternum shows a trans-Bass Strait distribution and most likely originated in mainland southeast Australia, since the highest species diversity is found on the mainland and the two Tasmanian branches occupy only very subordinate positions on the tree (Fig.
Further studies using more sampling localities in Tasmania and its islands could indicate points of origin in Victoria and the timing of millipede settlement of Tasmania.
Sincere thanks to Karin Voigtländer (
Full data of sequenced specimens
Data type: Tab-delimited text file
Explanation note: Full details of sequenced specimens, including locality, date, collector, collection number and coordinates.
Alignment of combined dataset
Data type: FASTA file
Explanation note: Alignment of the combined COI mtDNA and 16S rRNA dataset
P-distances of combined COI and 16S dataset
Data type: CSV File
Explanation note: Mean uncorrected pairwise distances between terminals (transformed into percentages) of the combined COI mtDNA and 16S rRNA dataset.