Study site and species
The field collections were carried out in the Lopé Reserve region (Middle Ogooué, Gabon). The Lopé Reserve is constituted by a mosaic of forest and savannah, primarily formed during the last glaciations (-18000, -12000) and then maintained by human savannah burnings, resulting in a fragmented landscape. The sample area was composed of three forest sites in the Reserve, characterized by vegetation age and 2–5 km apart (Okoumé, Chameau and Rocher). The Okoumé site consisted of 75 years-old Milletia, Aucoumea and Marantacea stands. The Rocher and Chameau sites were older (800 years-old) and consisted in mature forest scattered by rock outcrops and classical pluvial forest stands, respectively. Finally, a fourth sampled site, Doda, was an isolated gallery-forest in a savannah landscape, dating up to 2500 years and situated outside the Reserve. In total, 19 colonies of Cubitermes sp. affinis subarquatus and two colonies of Apilitermes longiceps and Crenetermes albotarsalis were sampled (Table 1). Immediately following collection, individuals were placed in absolute ethanol and stored at 4°C in the laboratory until DNA extraction.
DNA extraction
Total genomic DNA from Cubitermes, Apilitermes and Crenetermes individuals was isolated using an extraction method with Wilson buffer (Tris hydrogen chloride 1 M, ethylenediaminetetraacetic acid 0.5 M, sodium chloride 4.5 M, sodium dodecylsulfate 20%, dithiothreitol, proteinase K) followed by a salting-out procedure.
Mitochondrial and ITS2 sequence analyses
DNA analyses were based on sequences from partial mitochondrial cytochrome oxidase subunit II gene (COII) and nuclear internal transcribed spacer 2 (ITS2). PCR was performed for one individual per colony (N = 19 Cubitermes + 1 Apilitermes + 1 Crenetermes) in a total volume of 40 μL (50 μL for ITS2), composed of 20 μL (25 μL) of Taq PCR Master Mix (Qiagen), 1.6 μL (4 μL) of each primer (10 pM), 15.2 μL (18.5 μL) of distilled water and 1.6 μL (2.5 μL) of template DNA. Primers for COII amplification were forward modified A-tLeu 5'-CAGATAAGTGCATTGGATTT-3' and reverse B-tLys 5'-GTTTAAGAGACCAGTACTTG-3' [48,49], modified by Miura et al. [16]. ITS2 sequences were amplified using forward ITS2F 5'-TGTGAACTGCAGGACACAT-3' and reverse ITS2Rcub 5'-ATTCGGCGGGTAGTCTCG-3' primers modified in this study from Jenkins et al. [23]. The amplification conditions were adapted from Miura et al. [16]. The amplification products were purified using a DNA and Gel band Purification kit (GFX TM PCR kit, Amersham Biosciences USA). Sequence reactions were performed using BigDye Terminator Cycle Sequencing kit version 1.1 (Applied Biosystems), then purified with an ethanol-Na acetate method. Sequence data were obtained using an automatic DNA sequencer (Applied Biosystems, ABI PRISM 310) and analysed with Sequencing Data software (Applied Biosystems). All sequences were registered in GenBank database with accession numbers listed in Table 1.
COII and ITS2 sequences were aligned using CLUSTALW with the default settings [50] and sequences from Apilitermes longiceps and Crenetermes albotarsalis (Termitidae, Termitinae) were added to both datasets in order to root the trees. Phylogenetic analyses were performed using Maximum Parsimony (MP), Bayesian Inference (BI) and Maximum Likelihood (ML) methods.
MP trees were constructed using Phylip (Phylogeny Inference Package) Version 3.572 [51] with the SEQBOOT, DNAPARS and CONSENSE programs with 1000 repetitions of bootstrap.
For probabilistic methods, a nucleotide substitution model was selected for each sequence dataset using the Akaike Information Criterion (AIC) implemented in Modeltest v3.7 [52]. The best-fit substitution model selected was TrN+G (Nst = 6, Rates = gamma, Pinvar = 0) for COII sequences and TVM (Nst = 6, Rates = equal, Pinvar = 0) for ITS2 sequences. BI trees were constructed using MrBayes 3.1 [53]. We ran four Markov chains (one cold, three heated) for 1,000,000 generations, sampled every 100 generations (burnin = 2500, according to the convergence diagnostic). Consensus trees were generated including posterior probability of clades and branch lengths. ML trees were constructed using thePAUP 4.0b10 program [54] and the reliability of the inferred trees was tested by 100 bootstrap resamplings for COII and 40 for ITS2. Tree topologies were congruent across all methods and therefore, only BI trees are reported with posterior probabilities and bootstrap support values for ML and MP.
Microsatellite analyses
The genotypes for 18 to 48 sterile individuals from each colony except two (T14A, n = 6 and T14B, n = 12) i.e. a total of 447 individuals genotyped, were assayed at five microsatellite loci (P14, P19, P32, P34 and P41) by means of Polymerase Chain Reaction (PCR) amplification. Primer sequences and amplification conditions are given in Harry et al. [38]. PCR products were electrophoresed on an ABI Prism 310 DNA sequencer (Applied Biosystems) and microsatellite allele sizes were scored using the GENSCAN and GENOTYPER programs (Applied Biosystems).
We investigated whether genetic differentiation occurs among groups detected in the phylogenetic analyses. In order to do this, we estimated FST coefficients with individuals nested in colonies and colonies nested in major lineages using a three-level hierarchy in the GDA program [55], for each pair of putative cryptic species. 95% CIs were constructed by bootstrapping over loci with 1000 replications. Values for which 95% CIs did not overlap zero were considered as significantly different at the 0.05 level. We compared the allele differentiation among colonies of the same genetic group and among colonies of the same geographical site using the FCT coefficient [56] calculated with a two-level hierarchy in the GDA program.
Phylogenetic distances between colonies were estimated with the Populations 1.2.00 program [57] by using the genetic distance of shared alleles, the minimum genetic distance of Nei and the chord distance of Cavalli-Sforza. The resulting genetic distances and bootstrapping procedures (1000 replicates) were used to construct an unrooted consensus tree. Tree topologies were congruent across the three methods and therefore, only one reconstruction is presented here (reconstruction based on the chord distance of Cavalli-Sforza).
Authors' contributions
VR and CD carried out the molecular genetic studies, participated at the data analysis and the manuscript draft. AL intensively helped VR and CD in DNA sequencing. MH initiated the study, realized the sampling, contributed to the analysis of the results and to the writing of the paper. All authors read and approved the final manuscript.
Acknowledgements
We are grateful to the authorities of Gabon represented by Mr E. Mamfoumbi of the 'Ministère des Eaux et Forêts' for permission to work in the Lopé Reserve. We thank K. Abernethy and L. White for field facilities and scientific collaboration. We also would like to thank J. Harris and M. E. Cruz de Carvalho for editing the English text and anonymous reviewers for helpful comments on the manuscript. Partial financial support of the ABI Prism automated sequencer came from the University of Paris XII – Val de Marne via BQR grant.
Answers to all your Biology Questions
