Introduction

The high diversity of organisms in Costa Rica has been referred to as a product of diverse ecosystems resulting from the interaction between complex microclimates, soils, topography, and a variety of biological processes, as well as the position of the country in the land-bridge between North and South America. Costa Rica’s biodiversity comprises more than 500, 000 species of organisms, approximately 84% of which are yet to be described. This percentage is even higher (90%) if we take insects, fungi, bacteria and viruses into account (Sánchez-Azofeifa et al. 2001). As for the number of aphid species present in Costa Rica, the list was recently extended (Pérez Hidalgo et al. 2009; Zamora Mejías et al. 2010; Villalobos Muller et al. 2010) and research is ongoing.

During an expedition in 2008 in the area of Cerro de la Muerte (Cordillera de Talamanca), Costa Rica, three apterous viviparous females and several nymphs were collected on Chusquea tomentosa (Fig. 1). At first, they were assigned to the subtribe Rhopalosiphina Mordvilko, 1914 (Aphidini Latreille, 1802). This identification was confirmed in the laboratory when it was verified that the marginal papillae on abdominal segments I and VIII were in dorsal position to the corresponding stigmata. The morphological characters of the specimens resembled those of the genus Rhopalosiphum Koch, 1854, though the length of the setae were reminiscent of species in the subgenus Paraschizaphis Hille Ris Lambers, 1947 (Schizaphis Börner, 1931).

According to Valenzuela et al. (2009), Rhopalosiphum and Schizaphis form a monophyletic group with Melanaphis van der Goot, 1917, the separation between them being unclear. The vein structure of the wings separates Schizaphis from Rhopalosiphum, however, as there were no alates available, molecular analyses were carried out to verify the relationship with the genus Rhopalosiphum through qualitative and quantitative characters. Molecular analyses are normally used to determine species and resolve taxonomic problems in the family Aphididae (Lozier et al. 2008; Foottit et al. 2008, 2009; Lee et al. 2010; Pike et al. 2010).

A microscopic morphotaxonomic study of the specimens enabled the hypothesis that they could not be assigned to any known species, and strengthened the hypothesis that they be assigned to Rhopalosiphum. Molecular phylogenetic analysis based (1) on a fragment of the mitochondrial DNA containing the 5’ region of the cytochrome c oxidase 1 (COI) and (2) on the nuclear gene coding for the Elongation factor-1 alpha (EF1α) were used to verify both hypotheses.

Several species included in the subfamily Aphidinae are known living on plant species of the subfamily Bambusoideae (Poaceae), but only two belong to Rhopalosiphum (Aphidini): Rhopalosiphum arundinariae (Tissot, 1933) and Rhopalosiphum rufiabdominale (Schrank, 1899), and none to the genus Schizaphis.

Material and methods Material studied

Three apterous viviparous female and several nymphs (sample CRI-235) were recorded on Chusquea tomentosa Y. Widmer et L. G. Clark (Poaceae: Bambusoideae: Bambuseae: Chusquinae) in Ojo de Agua (Cerro de la Muerte, Cordillera de Talamanca, Costa Rica) (9°36'N, 83°47'W), 2968 m, 26.ii.2008.

Morphological study

Thirty-three quantitative characteristics and the qualitative features of shape, sclerotization, pigmentation and cuticular ornamentation, were considered. The method used for measurements is that normally employed in our studies (Nieto Nafría and Mier Durante 1998). A camera lucida fitted to the microscope was used for the drawings and the microphotographs were taken with a Leica DC digital camera with IM 1000 version 1.10 software.

DNA extraction and PCR amplification

Total DNA was extracted separately from two samples, one of them containing a single nymph and the second the contents of the abdomen of 3 apterous adults, all kept in 96% ethanol. We followed the HotSHOT (Hot Sodium Hydroxide and Tris) method (Truett et al. 2000).

PCR amplification of the two gene fragments analyzed was carried out on 3 µl of the extracted DNA. A 710 bp fragment of the 5’ region of the mitochondrial cytochrome c oxidase subunit 1 (COI) was amplified using primers LCO1490 and HCO2198, described by Folmer et al. (1994). PCR conditions for COI amplification were as follows: 94°C for 1 min; 35 cycles of 94°C for 30 s, 48°C for 1 min and 68°C for 1 min; a final extension step of 7 min at 68°C was included after cycling. Amplification of the Elongation factor-1 alpha (EF1α) gene fragment was performed using two consecutive PCR reactions with primers Efs175 (Moran et al. 1999) and Efr1 (5’GTGTGGCAATSCAANACNGGAGT3’) in the first reaction and then primers Efs175 and Efr2 (5’TTGGAAATTTGACCNGGGTGRTT3’) in the second hemi-nested reaction. PCR conditions used in the first reaction were: 94°C for 1 min; 40 cycles of 94°C for 30 s, 50°C for 1 min and 68°C for 1.5 min; a final extension step of 7 min at 68°C was included after cycling. The hemi-nested PCR was done similarly but using 52°C for the annealing step and using 1 µl of the first PCR product.

Sequencing and analysis of DNA sequences

PCR products were purified by ammonium precipitation and reconstituted in 10 μL of LTE buffer (10mM Tris, 0, 1mM EDTA). Direct sequencing of amplified fragments was done in both directions using PCR primers (Efr2 was used as reverse primer for sequencing the EF1α fragment). Sequencing was conducted using the Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) following the manufacturer’s instructions, and samples were loaded onto an ABI 3700 automated sequencer.

Chromatograms were revised and sequences corresponding to each sample assembled using the Staden package v1.6.0 (Staden et al. 2000). Multiple alignments were carried out with Clustal X v1.81 (Thompson et al. 2002) with gap opening and gap extension penalties of 10.0 and 0.2, respectively, and subsequently manually revised.

Phylogenetic analysis of COI sequences were done using MEGA version 4 (Tamura et al. 2007). For EF1α sequences ModelTest (Posada and Crandall 1998) was used to find the evolutionary model that best fitted sequence data and phylogenetic reconstruction was done using RAxML (Stamatakis et al. 2008).

Results Morphological data

A study of the qualitative and quantitative (metric and meristic) characters of the specimens enabled us to establish the hypothesis that they belong to the genus Rhopalosiphum as, apart from the above-mentioned character of the marginal papillae on the abdomen, (1) when alive they are ovoid and when preserved the body is not very long and the margins are curved (Figs 2A, 2B), (2) the dorsal cuticle of the thorax and abdomen is membranous, except for the presence of intersegmental sclerites and a pair of large sclerites on abdominal segment VIII (Figs 2A, 2B), (3) the dorsal cuticle has a more or less regular reticulate area formed by coalescent spinules (Figs 2Ba, 2Bb, 2F, 2G), (4) the siphunculi are longer than the cauda and clearly constricted underneath the apical edge (Figs 2A, 2B, 2D), and (5) there are few setae on the cauda.

A comparison of the characters of these specimens with those of apterae in other species of Rhopalosiphum and Schizaphis also strengthened the hypothesis that they could not be assigned to any known species.

Molecular data

A 710 bp DNA fragment containing a portion of the mitochondrial COI gene was amplified through PCR from the two samples analysed. Useful sequences obtained from each sample consisted of 658 nucleotides. Identical sequences were obtained for both samples so that a single sequence was finally assigned and deposited in Genbank (accession number HE604204). The online identification engine available at the Barcode of Life Data Systems (BOLD) (Ratnasingham and Hebert 2007) using the COI species database, failed to find any record corresponding to any identified species that matched our sequence. After a BLASTN search against the non-redundant nucleotide database at the NCBI, sequences from different Rhopalosiphum species were most similar to our sequence (93–94% identical) followed by Schizaphis sequences (92–93% identical). We then aligned our sequence with sequences from all Rhopalosiphum species available at the NCBI database and from species representative of closely related genera (Schizaphis, Melanaphis, etc.) that we had previously retrieved from the database, and built a phylogenetic tree (Fig. 3A). The tree shows that the sequence from our unknown species groups with relatively high support within a monophyletic clade that contains all other Rhopalosiphum and Schizaphis COI sequences occupying a rather basal position within that clade.

For the Elongation factor-1 alpha (EF1α) gene fragment, we obtained an identical sequence from the two analyzed samples of 987 bp which was deposited in the Genbank with accession number HE604205. Using sequences available for EF1α in NCBI for different Rhopalosiphum and closely related species within Aphidini, an ML tree was built that included the sequence obtained for our unknown species (Fig. 3B). As with the COI sequence, our unknown species grouped with strong support within a monophyletic clade that also included sequences from Schizaphis and Euschizaphis. However, unlike the COI tree, both Rhopalosiphum and Schizaphis-related sequences separated into two distinct clades, though with very low bootstrap support.

Discussion and conclusion

Molecular data using both mitochondrial COI and nuclear EF1α gene sequences confirmed that the Rhopalosiphum chusqueae specimens belong to the same monophyletic clade as other Rhopalosiphum species occupying a rather basal position in the group likely closely related to other divergent Rhopalosiphum species such as Rhopalosiphum nymphaeae. Both trees revealed the close relationship between Rhopalosiphum and Schizaphis genera. Although COI sequences are widely used in taxonomy, their utility for phylogeny reconstructions seems rather limited as their phylogenetic signal is somewhat weak in comparison with other markers (Wilson 2010). Contrarily, EF1α is widely used in phylogenetic reconstructions and its use in insect phylogeny has been shown to be informative (Simon et al. 2009; Wilson 2010). In this respect, although the COI analysis did not recover the monophyly of Rhopalosiphum and Schizaphis genera separately, ML analysis of EF1α separated both genera and clearly Rhopalosiphum chusqueae grouped within the Rhopalosiphum clade, which, along with our morphometric data discussed above, supports its assignation to the Rhopalosiphum genus.

Approximately 15 species are classified in the genus Rhopalosiphum (Remaudière and Remaudière 1997; Zhang and Qiao 1997; Eastop and Blackman 2005; Blackman and Eastop 2006) associated with arboreal Rosaceae (Prunus or Pyroidea) as the primary host and with Poaceae, Cyperaceae or, less frequently, other plants as the secondary host if their cycle is dioecious, or only with one of them if their cycle is monoecious. Most of the species probably originate in North America, with a subsidiary centre of dispersal in Central Asia (Blackman and Eastop 1994; Halbert and Voegtlin 1998; Blackman and Eastop 2006). Five of its species have an exclusively Nearctic distribution: Rhopalosiphum arundinariae (Tissot), Rhopalosiphum cerasifoliae (Fitch), Rhopalosiphum enigmae Hottes & Frison, Rhopalosiphum laconae Taber, Rhopalosiphum nigrum Richards, and Rhopalosiphum padiformis Richards; and another four Nearctic species have been introduced in other parts of the world: Rhopalosiphum parvae Hottes & Frison and Rhopalosiphum rufulum Richards in Europe, Rhopalosiphum musae Schouteden has been recorded in areas of Europe, Central Asia, Africa and Australia, and Rhopalosiphum oxyacanthae (Schrank) is known in Central- and South-America, Europe, Asia and Australia. To date, only four species, linked mainly to crops, have been recorded in Central American countries: Rhopalosiphum maidis (throughout Central America), Rhopalosiphum nymphaeae in Panama, Rhopalosiphum padi in Costa Rica and Panama, and Rhopalosiphum rufiabdominale in Honduras, Costa Rica and Panama (Evans and Halbert 2007; Quirós et al. 2009; Villalobos Muller et al. 2010); Rhopalosiphum oxyacanthae is also known in Central America, without country (Blackman and Eastop 2006).

Species of Rhopalosiphum most resembling the new species due to their morphological characters are Rhopalosiphum rufiabdominale and Rhopalosiphum padiformis. The former originated from East Asia (Blackman and Eastop 2006) and is currently widely distributed. Rhopalosiphum padiformis originates from North America. Rhopalosiphum chusqueae sp. nov. coincides with both species in the length of the setae, with Rhopalosiphum padiformis in the number of antennal segments and shape of the cauda, and with specimens of Rhopalosiphum rufiabdominale in the 4 setae on abdominal segment VIII (Rhopalosiphum rufiabdominale has 3 to 8 setae on this segment). It is easily distinguished from them because the antennae in Rhopalosiphum rufiabdominale are five-segmented and the dorsal setae in Rhopalosiphum padiformis are not pointed and abdominal segment VIII only has 2 setae. Bamboo species are the host plants of the mentioned Rhopalosiphum rufiabdominale and Rhopalosiphum arundinariae; this last species can be easily differentiated from Rhopalosiphum chusqueae by the shape of the cauda (short and more or less triangular or rounded) and siphunculus (more or less tapering) and by much shorter setae on body dorsum and appendages.

In view of the above, a new species can be established, the description of which follows.