Molecular and morphological phylogeny of the parasitic wasp genus Yelicones ( Hymenoptera : Braconidae : Rogadinae )

Phylogenetic relationships of the braconid wasp genus Yelicones Cameron are studied using the D2–D3 region of the nuclear 28S rRNA gene, both alone and simultaneously with morphology. The results support a morphology-based phylogeny, presented elsewhere, with Yelicones being divided into two major groups corresponding to the New and Old World faunas. The African and Asian species largely form separate clades except for Yelicones wui Chen & He from China which is associated with the Afrotropical species. Potential molecular synapomorphies are illustrated. 617 * Corresponding author. MATERIAL AND METHODS Laboratory protocols A middle leg was removed from each individual (7 preserved in ethanol and 3 from recently collected dry pinned specimens). Sequence data for Y. wui Chen & He came from Chen et al. (2003). After drying, the tissue was ground in 50 μl of 5% (w/v) Chelex (Bio-Rad, Hercules, CA, USA)/TE containing 12 μg/ml proteinase K, and digested for 100 min at 56°C. Proteinase K was then heat-inactivated by incubation at 96°C for 15 min, and the samples stored at –20°C without separating the supernatant prior to PCR amplification. Samples were thawed on ice, briefly vortexed and the Chelex pelleted by centrifugation at 16,300 g for five min prior to removal of 2 ml of supernatant as template for each PCR reaction. PCR reactions were carried out in a 25 μl final volume using pureTaq ready-to-go PCR beads (Amersham Biosciences, Buckinghamshire, UK) with 400 nM of each of 28S forward (GCG AAC AAG TAC CGT GAG GG; Hancock et al., 1988) and 28S reverse (TAG TTC ACC ATC TTT CGG GTC; Campbell et al., 1993) primers. The PCR program for all the amplifications had an initial 3 min denaturation at 80°C, followed by 40 cycles at 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min. A 10 min extension followed the final cycle. The PCR products were cleaned with the wizard SV gel (Promega, Madison, WI, USA) and PCR clean up system and then sequenced using the dideoxy terminator cycle sequencing (Applied Biosystems, Inc., Foster City, CA, USA) with an ABI 3700 automated DNA sequencer according to manufacturers instructions. All species were sequenced in both directions and all pherogram interpretations checked manually.


INTRODUCTION
Yelicones Cameron is a cosmopolitan and highly distinctive genus of parasitic rogadine braconid wasps.The genus is recognised by its robust legs, highly modified and distinctive tarsi with the telotarsus extremely enlarged and the other tarsal segments greatly reduced (Roman, 1910) and by its tridentate mandibles although these are often difficult to observe when they are closed (Quicke & Kruft, 1995;Areekul & Quicke, in press a).As in all other Rogadinae sensu stricto, Yelicones sp. are koinobiont endoparasitoids of Lepidoptera larvae, specifically Pyralidae and Crambidae (Shenefelt, 1975;Areekul & Quicke, in press a).Yelicones mummify the host before pupating within the host (Quicke & Chishti, 1997;Quicke & Shaw, in press).
Recently, Areekul & Quicke (in press a) revised Yelicones species from North, Central and South America.Sixty-three species were described as new and a species level morphological phylogeny based on 116 characters was presented for the world fauna.The results showed a nearly perfect division into two large groups comprising the Old World (OW) and the New World (NW) species, respectively.In the preferred tree in which colour characters were excluded, nearly all the OW species formed a monophyletic group nested within a basal grade of NW taxa, though within the OW clade species from the Australian, African and Oriental regions were largely intermixed, probably due to a lack of phylogenetic signal.However, inclusion of colour characters had a major impact on tree topology in terms of whether the OW or NW species were derived groups: when colour characters were included in the analysis, the NW species appeared to be derived from within a paraphyletic OW grade.
The above conflict in topologies between analyses including or excluding colour characters could be the result of the latter being more homoplastic at higher taxonomic levels, particularly as none of the outgroup taxa are very close to Yelicones.However, the fit of the colour characters on the trees from the simultaneous analysis of colour and morphological characters was higher than the fit of the morphological characters.This suggested that colour character data should not be discounted immediately, though it is also possible that colour characters might be evolving differently from morphological ones, especially when they are involved in aposematic patterns (Areekul & Quicke, in press b).Here we attempt to resolve the conflict by using molecular data in addition to morphological data.Due to the rarity of this taxon, very few fresh specimens were available for sequencing.A total of 11 species, 3 from the NW and 4 each from Africa and Asia were sequenced.We were not able to obtain fresh material from Australia.

Laboratory protocols
A middle leg was removed from each individual (7 preserved in ethanol and 3 from recently collected dry pinned specimens).Sequence data for Y. wui Chen & He came from Chen et al. (2003).After drying, the tissue was ground in 50 µl of 5% (w/v) Chelex (Bio-Rad, Hercules, CA, USA)/TE containing 12 µg/ml proteinase K, and digested for 100 min at 56°C.Proteinase K was then heat-inactivated by incubation at 96°C for 15 min, and the samples stored at -20°C without separating the supernatant prior to PCR amplification.
Samples were thawed on ice, briefly vortexed and the Chelex pelleted by centrifugation at 16,300 g for five min prior to removal of 2 ml of supernatant as template for each PCR reaction.PCR reactions were carried out in a 25 µl final volume using pureTaq ready-to-go PCR beads (Amersham Biosciences, Buckinghamshire, UK) with 400 nM of each of 28S forward (GCG AAC AAG TAC CGT GAG GG; Hancock et al., 1988) and 28S reverse (TAG TTC ACC ATC TTT CGG GTC; Campbell et al., 1993) primers.The PCR program for all the amplifications had an initial 3 min denaturation at 80°C, followed by 40 cycles at 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min.A 10 min extension followed the final cycle.The PCR products were cleaned with the wizard SV gel (Promega, Madison, WI, USA) and PCR clean up system and then sequenced using the dideoxy terminator cycle sequencing (Applied Biosystems, Inc., Foster City, CA, USA) with an ABI 3700 automated DNA sequencer according to manufacturers instructions.
All species were sequenced in both directions and all pherogram interpretations checked manually.

Taxa investigated
The species included in this study are listed in Table 1, together with their provenances, depositories and EMBL/Gen-Bank sequence accession numbers.The twelve outgroup taxa chosen for this study were as follows: Aleiodes nigricornis Wesmael, 1838, Aleiodes praetor (Reinhard, 1863), Aleiodes dispar (Haliday, 1833), Pseudoyelicones limonensis Areekul & Quicke, Rogas Nees, 1818, Spinaria Brullé, 1846, Conspinaria Schulz, 1906, Bulborogas van Achterberg, 1995, Stiropius Cameron, 1911, Clinocentrus Haliday, 1833, Tebennotoma Enderlein, 1912, Mesocentrus Szépligeti, 1900, and of which Tebennotoma was scored from the literature (van Achterberg, 1995).The use of multiple outgroups reflects the general lack of understanding of rogadine relationships (see Zaldivar-Riverón et al., 2004), and in particular the lack of an obvious sister taxon to Yelicones, though Bulborogas and Pseudoyelicones van Achterberg, Panteado-Dias & Quicke species are similar, perhaps convergently, in body form, being robust with swollen femora and shortened tarsi (see Areekul et al., 2004).Three species of Aleiodes have been included because this is the largest and morphologically most diverse genus of the Rogadinae sensu stricto (Chen & He, 1997).Zaldivar-Riverón et al. (2004)  Achterberg, 1991), A. praetor belongs to the subgenus Neorhogas Szépligeti and A. dispar belongs to the subgenus Heterogamus Wesmael.Three members of the Rogas group of genera (Rogas, Conspinaria and Spinaria) were included as they appear to form the sister group of the Aleiodes group and were expected to help show whether Yelicones, Bulborhogas and Pseudoyelicones form a natural group.The betylobraconine, Mesocentrus Szepligeti, was included because betylobraconines also have shortened fore tarsi and robust femora (van Achterberg, 1995) and it is not certain that they are not derived rogadines and thus could be related to Yelicones.Finally, Stiropius of the Stiropiini and Clinocentrus and Tebennotoma of the Clinocentrini were included as these tribes are putatively most basal within the Rogadinae, the former because of their parasitization of leaf-mining microlepidoptera larvae (Gracillariidae and Lyonetiidae), and the latter because of their more exserted ovipositors.Trees were rooted with Tebennotoma.

Phylogenetic analysis
Sequences were aligned by eye and analyses performed with several ambiguous regions excluded and with remaining gaps treated as uninformative (see Fig. 2).The alignment used with excluded regions indicated is available from http://www.treebase.org/treebase/(study accession number S1356; matrix accession number M2396).Maximum parsimony analyses were implemented in PAUP* version 4.0b10 (Swof-Fig.1.A -strict consensus of 2 MPTS from molecular-based phylogeny of Yelicones; B -successive approximations weighting tree from the combined analysis.Clades congruent with the strict consensus of MPTs from unweighted analysis are shown in bold.Bootstrap values greater than 50% from unweighted analyses are indicated ford, 1998) using 10,000 random additions with tree bisection reconnection (TBR) branch swapping holding a maximum of one tree.Relative branch support was assessed by bootstrapping (Felsenstein, 1985) using 500 replicates, each based on 100 random additions.
Eighty morphological characters were used in a combined analysis of molecular and morphological data.The morphological characters used are a subset of those used by Areekul & Quicke (in press a) that are informative for the Yelicones species investigated here.See appendix 1 for character definitions and appendix 2 for the data matrix.
Successive approximations weighting (SAW: Farris, 1969) using the maximum value of the retention index as the reweighting function (see Quicke et al., 1999;Gauthier et al., 2000) was applied to the combined molecular plus morphological data set to obtain a more resolved tree.

Molecular results
Parsimony analysis of 28s rRNA sequence data based on 564 alignable base pairs resulted in 2 most parsimonious trees (MPTs), the strict consensus of which is shown in Fig. 1A (tree length = 386, CI = 0.642, RI = 0.681).The analysis supports the morphological-based phylogeny of Areekul & Quicke (in press a) in that Yelicones species are divided into 2 large clades, corresponding to OW and NW species, with high bootstrap support (100) (Fig. 1A).Many clear sequence features separate the OW and NW species including several insertions/deletions (indels) that were excluded from or were uninformative in our analyses (Fig. 2, fragments 1, 2, 4, 6).Additionally, a clear resolution of the OW group was obtained, separating Yelicones into largely African and Asian clades (see Fig. 2, fragments 1, 2, 4, 5).The problematic taxon, Y. wui from China, being consistantly recovered with the African clade.Yelicones delicatus from North America has a distinctive sequence relative to other Yelicones, with many unique features such as a two base insertion in fragment 6, a deletion in fragment 1 and a 1 base substitution in fragment 3 (Fig. 2).The outgroup taxa are not well resolved.

Combined analysis
The combined analysis of morphological and molecular data sets resulted in 6 MPTs (tree length = 787, CI = 0.478, RI = 0.544).The strict consensus of these was less resolved than that from the molecular analysis of Yelicones relationships (see thickened branches in Fig. 1B).The NW species were recovered as monophyletic but the OW species were hardly resolved.However, there is more resolution among the basal taxa, with the Rogadini + Yeliconini recovered as monophyletic in addition to the Rogas genus group (Spinaria, Conspinaria and Rogas) being recovered as monophyletic.A more resolved tree was found following successive approximations weighting (SAW): congruence between the SAW tree and the strict consensus tree of 6 MPTs is indicated in bold (Fig. 1B).The OW and the NW species again form two separate clades.However, the Asian species are not recovered as monophyletic except for a clade comprising Y. nipponensis + Y. belokobylskiji.Y. wui was recovered with the otherwise monophyletic African taxa.The SAW tree shows additional resolution among the outgroup taxa compared with the strict consensus tree from molecular data alone including the three Aleiodes species recovered as monophyletic.Pseudoyelicones and Bulborogas formed a clade with 100% bootstrap support as found with the molecular data, and again they formed a sister group to Yelicones with slightly decreased bootstrap support (99%) (Fig. 1).

DISCUSSION
Our molecular-based phylogeny supports the phylogeny based on morphological characters and indicated that Yelicones species are divided into 2 large clades, the OW and NW species groups.DNA sequences show strong signals separating the OW and NW, and also differentiate African from Asian species [except for the problematic Y. wui (Fig. 1A)].In morphology-based analysis including colour characters (Areekul & Quicke, in press a), Y. wui was recovered as sister group to Y. variegatus from Madagascar, and within the NW clade in analysis excluding coloration.Y. wui has more or less the same colour pattern as many species found in Africa (e.g.Y. variegatus and Y. kibaleiensis), their ground-plan body colour being yellow with brown to dark brown distributed throughout the body.This might lead to the recovery of Y. wui closely related to Y. variegatus and Y. kibaleiensis in morphology-based phylogeny with colour characters included.There are a few species from Asia that have more or less the same colour pattern as Y. wui (e.g.Y. elegans and Y. koreanus) but the lack of specimens of these taxa for DNA sequencing prevented us from telling whether these also belong to the same group.
In the morphology-based phylogeny (Areekul & Quicke, in press a), Y. delicatus was often recovered between the OW and NW groups, especially in analysis including coloration.Its morphology is rather intermediate between that of OW and NW species, (e.g., it has an almost complete occipital carinae and the maximum length of eye is more than 1.8 times the width of eye, both of which are characteristic of most OW species).However, the 28S rDNA data clearly show that it belongs to the same monophyletic group as the other two NW species examined (note the apparent synapomorphic substitutions in fragments 3 and 6 and the possibly synapomorphic "AAA" insertion in fragment 6 (Fig. 2).
The molecular and combined analyses presented here do not resolve the initial question (based on morphological analyses including and excluding colour characters) as to whether NW or OW Yelicones species are basal (forming a paraphyletic grade) (Areekul & Quicke, in press a).Instead they both point to a reasonably wellsupported scenario with both NW and OW groups being monophyletic.However, our taxon sampling was poor especially for the NW clade, and additional taxon sampling will be required fully to resolve this matter.
The molecular phylogeny also confirms a close relationship between Pseudoyelicones and Bulborogas as was suggested by Areekul et al. (2004) based on venom appa- ratus morphology.Although Pseudoyelicones and Bulborogas appear as a sister group to Yelicones in both analyses (Fig. 1A, B), the bootstrap support for this in the purely molecular tree (Fig. 1A) is only marginal.The secondary venom duct insertion in Pseudoyelicones and Bulborogas is modified (hard) and thus resembles that of members of the Rogas groups of genera (Zaldivar-Riverón et al., 2004) whereas secondary venom duct of Yelicones is unmodified, suggesting a more basal position, (unless this modification has been secondarily lost).Unfortunately, the hosts and mummification biology of Pseudoyelicones and Bulborogas are unknown, though features of Yelicones mummies (Quicke & Shaw, in press) suggest that Yelicones is probably less derived than the Rogas group.

TABLE 1 .
studied the venom apparatus of 46 species of Aleiodes, six different types of venom apparatus are recognised which partially corresponded with the different subgenera.The three species chosen here represent groups with markedly different types of venom apparatus: A. nigricornis belongs to the subgenus Aleiodes (sensu van 618 * Abbreviations: CAS -California Academy of Sciences, San Francisco; DCBU -Universidade Federal de São Carlos, São Carlos, São Paulo, Brazil; ESUW -University of Wyoming, Laramie, Wyoming; INBC -Instituto Nacional de Bioversidad (INBio), Santo Domingo de Heredia, Costa Rica; NHM -Natural History Museum, London; NMS -National Museums of Scotland, Edinburgh.Species included in this study, collection localities and EMBL/GenBank accession numbers.