Molecular Phylogeny of Two Slave-Making Ants: Rossomyrmex and Polyergus (Hymenoptera: Formicidae)
Total Page:16
File Type:pdf, Size:1020Kb
Ann. Zool. Fennici 39: 267–271 ISSN 0003-455X Helsinki 10 October 2002 © Finnish Zoological and Botanical Publishing Board 2002 Molecular phylogeny of two slave-making ants: Rossomyrmex and Polyergus (Hymenoptera: Formicidae) Eisuke Hasegawa1, Alberto Tinaut2* & Francisca Ruano2 1) Center for Ecological Research, Kyoto University, Kyoto 606-01 and Labo- ratory of Wildlife Conservation, National Institute for Environmental Studies. Onogawa 16-2, Tsukuba 305, Japan; Present address: Laboratory of Animal Ecology, Department of Ecology and Systematics, Graduate School of Agriculture, Hokkaido University. Kita-ku, 060-8589 Sapporo, Japan (e-mail: [email protected]) 2) Department of Animal Biology and Ecology, University of Granada, ESP-18071 Granada, Spain (e-mail: [email protected]) Received 23 April 2001, accepted 11 May 2002 Hasegawa, E., Tinaut, A. & Ruano, F. 2002: Molecular phylogeny of two slave- making ants: Rossomyrmex and Polyergus (Hymenoptera: Formicidae). — Ann. Zool. Fennici 39: 267–271. Using mitochondrial cytochrome oxidase I gene (COI), we established the phyloge- netic relationship between two slave-making genera of the tribe Formicini: Polyergus and Rossomyrmex. The resulting phylogenetic tree presents two well-defi ned groups: Formica + Polyergus, and Proformica + Cataglyphis + Rossomyrmex. Our result con- tradicts prior classifi cations which considered Polyergus and Rossomyrmex to be sister genera. These results imply that slave-making in the two taxa evolved independently. Introduction ize Formica and Proformica, respectively. All current phylogenies consider Polyergus The genus Polyergus is composed of fi ve spe- a sister genus to Formica, but the phylogenetic cies (Bolton 1995) endemic to the Holarctic position of the genus Rossomyrmex has been region. In Rossomyrmex, only 2 species have repeatedly revised since its initial description been identifi ed: R. proformicarum Arnoldi from by Arnoldi (1928). Arnoldi found similarities the Tian-Shan (Kazakhstan) and Caspian steppes between Rossomyrmex and Proformica, and (Russia), and R. minuchae Tinaut from the Sierra between Polyergus, Formica and Cataglyphis, Nevada mountains (Spain). Polyergus and Ros- but did not establish the phylogenetic relation- somyrmex are slave-making ants, which parasit- ships among these taxa (Arnoldi 1928). In sub- * Corresponding author 268 Hasegawa et al. • ANN. ZOOL. FENNICI Vol. 39 A Formica B Melophorus ? Formica Polyergus Polyergus Rossomyrmex ? Rossomyrmex Bajcaridris ? Gen. nov Proformica Alloformica Proformica Cataglyphis ? Alloformica Fig. 1. Phylogenetic rela- tionships in the tribe Formi- cini hypothesized by Agosti Cataglyphis 1989 (A) and 1994 (B). Table 1. Species, localities and Genbank accession sequent papers, Rossomyrmex was considered number. derived from Formica-like ancestors and phyl- logenetically related near Formica and/or Pol- Species Collecting Genbank accesion yergus (Wilson 1971, Brown 1973, Buschinger location number 1990, Hölldobler & Wilson 1990). Agosti (1989) (Fig. 1A) placed it, with doubt, between Cataglyphis the Formica/Polyergus group and Proformica, C. bicolor Morocco AB010942 C. fl oricola Spain AB010936 Alloformica, Bajcaridris and Cataglyphis. Later C. rosenhaueri Spain AB010941 Agosti (1994) (Fig. 1B) changed Rossomyrmex C. velox Spain AB010933 to a position derived from Polyergus, upholding Formica a classical opinion which considers two groups F. cunicularia Spain AB010926 of genera with close phylogenetic ties: on one F. exsecta Poland AB010927 side, Formica L., Polyergus Latreille and Ros- F. fusca Spain AB010925 somyrmex Arnoldi, and, on the other, Bajcarid- F. truncorum Japan AB010929 ris Agosti, Proformica Ruzsky, Alloformica F. yesensis Japan AB010928 Lasius Dlussky and Cataglyphis Foerster. According L. niger Germany AB007981 to this phylogeny, the two slave-making genera, Plagiolepis Polyergus and Rossomyrmex, are grouped P. pygmaea Spain AB010938 together and are considered to be sister taxa. Proformica Tinaut et al. (1994), however, reported that, P. ferreri Spain AB010935 on the basis of male genitalia, Rossomyrmex P. longiseta Spain AB010934 appears to be more closely related to Proformica Polyergus and Cataglyphis than to Formica. P. rufescens Spain AB010931 In the present paper, we use molecular tech- P. samurai Japan AB010930 Rossomyrmex niques to determine the phylogenetic position of R. minuchae Spain AB010937 Rossomyrmex and Polyergus. ANN. ZOOL. FENNICI Vol. 39 • Molecular phylogeny Rossomyrmex and Polyergus 269 Materials and methods product were sequenced using Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin The species Elmer) on an ABI 373A automated sequencer. The accuracy of the target sequence of a species The species used and their sampling sites are was confi rmed by comparing the sequence from given in Table 1. For outgroup species, we have one of a pair of primers with the complementary considered two species: Plagiolepis pygmaea sequence from the other primer. Latreille (tribe Plagiolepidini) and Lasius niger L. (tribe Lasiini), both from the Formicinae sub- family. Phylogenetic reconstruction The phylogenetic relationship between the spe- DNA extraction and sequencing cies was reconstructed using the software PAUP 4.0b5 (Swofford 1998). We used two different DNA from one individual of each species was algorithms to estimate the phylogeny: neigh- extracted from ants preserved in 99.9% etha- bour joining (NJ), and maximum parsimony nol. A DNA extraction kit (DNeasy Tissue Kit, (MP). For NJ, we selected Kimura’s two-param- QIAGEN), was used, following the manufactur- eter distance as an index of genetic distances er’s instructions. between sequences. In addition, we used a We designed primers for the mitochondrial gamma function with 0.5 shape variable to cor- cytochrome oxidase subunit I (COI) gene by rect between-site-rate variations. For MP, we comparing the sequences of the fruit fl y Dro- equally weighted all substitutions. Robustness sophila melanogaster (de Bruijn 1983) and the of the resulting topologies were examined by a honey bee Apis mellifera L. (Crozier & Crozier bootstrap test with 1000 resamplings. 1993). We searched for conservative regions in which more than 90% of the sequences matched between the two species. Two primer pairs were Results designed (given as 5´–3´): CI13: ATAATTTTTTTTATAGTTATACC The target region was 1104-bp nucleotides CI14: ATTTCTTTTT TTCCTCTTTC; and in length, but the primer region itself and CI21: CTTTATCAACATTTATTTTGATTTTT sequences near the primers could not be deter- CI24: TCCTAA AAAATGTTGAGGAAA. mined accurately. As a result, we obtained a In total, we obtained 1104bp sequence within 974-bp sequence from each species. These COI from these primer pairs. sequences are deposited at the DDBJ database The target fragment was amplifi ed by PCR (see Table 1). in a programmable thermal cycler (PTC-100, Figure 2 shows NJ and MP trees with boot- MJ Research Inc.). Each PCR contained a strap probabilities of each node. We also esti- 50 µl reaction volume of the following compo- mated the maximum-likelihood tree, using the nents: 100 ng DNA, 10 mM Tris-HCl (pH 8.3), HYK 85 model. The ML estimation resulted in 50 mM KCl, 2 mM MgCl2, 0.001% gelatine, the same topology as the unweighted MP, and 20 pmol of primers, and 0.4U Taq DNA polymer- thus the ML tree is not presented. We did not ase (GeneTaq, NipponGene). Amplifi cation was provide a phylogram with branch-length infor- 30 cycles per 1 min at 94 °C, 1 min at 45 °C, mation because our main concern in this study and 3 min at 60 °C. Amplifi cation products can be resolved without branch length. Branch were purifi ed with a DNA-purifi cation kit (PCR lengths can be determined using the above-men- Purifi cation Kit, QIAGEN), and suspended in tioned deposited sequences. 25 µl of TE buffer. Both strands of each PCR Two clades were well supported in all 270 Hasegawa et al. • ANN. ZOOL. FENNICI Vol. 39 A: NJ 60 Formica (Serviformica) fusca 52 Formica (S.) cunicularia 99 Formica (Coptoformica) exsecta Formica (str.) yessensis 91 100 Formica (str.) truncorum Polyergus samurai 100 Polyergus rufescens 59 100 Proformica longiseta 65 Proformica ferreri Rossomyrmex minuchae 41 Cataglyphis velox 24 Cataglyphis bicolor 77 24 Cataglyphis rosenhaueri Cataglyphis floricola Plagiolepis pygmaea Lasius niger Fig. 2. The estimated phylogenetic relation- ships between the exam- B: Unweighted MP 62 Formica (Serviformica) fusca ined species. Topologies 69 Formica (S.) cunicularia were derived from two 100 Formica (str.) yessensis estimation algorithms (A: 94 Formica (str.) truncorum 52 Formica (Coptoformica) exsecta NJ and B: unweighted Polyergus samurai MP). The numbers 100 Polyergus rufescens at branches represent 65 100 Proformica longiseta boot strap probabilities 32 Proformica ferreri 40 Rossomyrmex minuchae from 1000 resamplings. 58 Cataglyphis bicolor Descriptive statistics of 48 Cataglyphis velox the unweighted MP tree 80 Cataglyphis rosenhaueri Cataglyphis floricola were TL = 1059, CI = Plagiolepis pygmaea 0.5467, HI= 0.4533, RI = Lasius niger 0.4888 and RC = 0.2673. cases, one including Formica and Polyergus to be ancestral, although this could change with and the other consisting of Cataglyphis, Pro- the inclusion of other genera and is not currently formica and Rossomyrmex. The bootstrap sup- supported by the bootstrap analysis. Indeed the port for both groups was 57%–94% for the bootstrap support was low for all clades (Fig. 2), Formica/Polyergus clade and 77%–85% for except