1. Introduction
Lipolexis, a small genus in the subfamily Aphidiinae, is widely distributed across Europe and Asia. Like all members of this subfamily, species of
Lipolexis are solitary koinobiont endoparasitoids of aphids (Aphididae). Förster (1862) [
1] described it as monotypic with
Lipolexis gracilis as the type species. However, its taxonomic distinctiveness was subsequently questioned by some authors who considered it as a synonym of the genus
Diaeretus [
2,
3]. Starý [
3] re-established
Lipolexis on the basis of type material and specimens collected in former Czechoslovakia. A second member of this genus,
Lipolexis oregmae Gahan, was first described as
Diaeretus oregmae by Gahan [
4]. However, in his revision of generic boundaries in the family Aphidiidae (now subfamily Aphidiinae), Starý [
5] stated that, aside from its type species, the genus
Diaeretus was monotypic and other taxa should be transferred to other genera. Consequently, he reassigned
D. oregmae to
Lipolexis [
6]. Two years later, Mackauer [
7] described the third species,
Lipolexis scutellaris Mackauer, from the Oriental region. However, after re-examining
L. oregmae and
L. scutellaris, Starý concluded that
L. scutellaris was a junior synonym of
L. oregmae [
8]. Since this time, additional species of
Lipolexis have been reported.
Lipolexis chinensis Chen was described by Chen [
9], but three years later Starý and Ghosh [
10] placed it as a junior synonym of
L. gracilis. However, three more species were subsequently described from the Oriental region:
Lipolexis wuyiensis Chen [
11] from China, and both
Lipolexis myzakkaiae Pramanik and Raychaudhuri and
Lipolexis pseudoscutellaris Pramanik and Raychaudhuri [
12] from India. Although these three species are still regarded as valid, their descriptions leave doubt. Moreover, it seems that neither species has been mentioned in the scientific literature since their description.
Lipolexis gracilis has a Palaearctic distribution and it is the only
Lipolexis currently thought to be widely distributed in Europe [
13,
14,
15]. By contrast,
L. oregmae, an Oriental species, has been introduced to USA (Florida) [
16], Jamaica [
17], Dominica [
18] and Costa Rica [
19]. While the potential value of
L. gracilis in biocontrol of black bean aphid,
Aphis fabae Scopoli [
20], and soybean aphid,
Aphis glycines Matsumura [
21] has been considered,
L. oregmae has proven effective against the brown citrus aphid,
Toxoptera citricida (Kirkaldy). The latter aphid causes serious damage to citrus plantations both by transmitting the tristeza closterovirus [
22] and by direct damage to the plant.
Lipolexis oregmae has been successfully introduced to Florida from Guam in a biological control program directed against this aphid pest [
16]. Hoy et al. [
17] reported the unintentional establishment of
L. oregmae in Jamaica, where the invasion of
T. citricida has led to serious losses in citrus production. Furthermore, Cocco et al. [
18] and Zamora et al. [
19] reported the same situation for Dominica and Costa Rica, respectively.
As Mackauer [
23] considered
Lipolexis as phylogenetically close to the genera
Trioxys and
Binodoxys, he placed it within the subtribe Trioxina. This subtribe is classified by different authors within tribe Trioxini [
24] or tribe Aphidiini [
23]. However, morphological examination of its immature stages led Finlayson [
25] to recognize that larvae of
Lipolexis possess an epistoma, a unique character for Aphidiinae genera. Based on this observation, she suggested that
Lipolexis should be placed in a separate subtribe despite the morphological similarity of the adult to those of Trioxina species.
Although the genus
Lipolexis is widely distributed and one species is an important biological control agent, the last taxonomic study on it was conducted more than 50 years ago by Starý [
3]. Several molecular studies on Aphidiinae included representatives of this genus, but conclusions concerning the position of
Lipolexis were ambiguous. Based on phylogenetic analysis of the mitochondrial NADH 1 dehydrogenase gene using neighbor joining analysis, Smith et al. [
26] positioned
L. gracilis as a sister group to the rest of Aphidiina and Trioxina species. In the same study, both weighted and unweighted parsimony analysis positioned
L. gracilis within the tribe Aphidiini. Furthermore, in the weighted parsimony analysis of the 16S RNA sequence conducted by Kambhampati et al. [
27],
L. gracilis was placed as a sister group to the genera
Binodoxys and
Trioxys with high bootstrap value supporting placement of
Lipolexis within the subtribe Trioxina. However, the same study placed
L. gracilis with the subtribe Aphidiina with low bootstrap support in the unweighted parsimony analysis. The results of phylogenetic analysis of 18S RNA region [
28] positioned
L. gracilis as a sister group to the genus
Trioxys, supporting Mackauers’ [
23] classification. The study of Shi and Chen [
29] based on three genes (16S RNA, 18S RNA, ATPase 6) also supports the traditional classification of
Lipolexis within the subtribe Trioxina. Furthermore, analysis of the barcode region of cytochrome
c oxidase subunit I (COI) yielded a phylogenetic tree with
L. gracilis grouped as a sister taxon to
Falciconus pseudoplatani Marshall within species in the subtribes Trioxina and Monoctonina [
30]. Based on these results, it is evident that the position of
Lipolexis within Aphidiinae remains uncertain and merits further research.
Our exhaustive studies on populations of L. gracilis across Europe and L. oregmae in the far East and USA (Florida) revealed considerable variation in taxonomically important morphological characters, such as the number of maxillary and labial palpomeres, as well as in the shape of flagellomere 1 (F1) and petiole. Motivated by these observations, both morphology and molecular analysis (COI barcode region) were used to re-examine specimens of L. gracilis and L. oregmae from various sites across their distributions. Our research sought to ascertain if L. gracilis and L. oregmae include cryptic species and to reconstruct phylogenetic relationships within the genus. It led to recognition of six new species and a key for the identification to all known Lipolexis species.
4. Discussion
Until recently, taxonomists relied on morphological and ecological studies as a basis for describing new species or identifying newly collected specimens. However, easy access to DNA sequence information has now enabled integrative taxonomy which combines morphological/ecological and molecular data to clarify species boundaries. The barcode region of COI is by far the most widely utilized molecular marker used for molecular studies on the subfamily Aphidiinae [
30,
44,
45]. Aside from aiding the identification of the newly collected specimens, this gene region aids the delineation of species boundaries, an approach of particular value for aphidiine genera that are suspected to include cryptic species complexes [
46,
47,
48,
49]. Our study reinforced the value of COI as it separated
Lipolexis species with high resolution and bootstrap support. The molecular results were concordant with those from morphological study leading to the description of six new species of
Lipolexis. Each of these species possessed diagnostic morphological characters which separates it from its congeners which makes the fact that they went unnoticed surprising.
Based on morphological and molecular differences, Lipolexis can be separated into two main groups, i.e., gracilis and oregmae. The average genetic distance between species in these two groups is high, ranging from 21.7% to 23.2%. Furthermore, gracilis and oregmae groups are morphologically distinguished by the differing shape of the petiole: members of the L. gracilis group have a petiole with prominent central carinae, while those in the L. oregmae group have a petiole that is smooth dorsally, but with noticeably crenulated lateral longitudinal carinae. Additionally, members of the gracilis group have a slightly shorter metacarpal vein (R1) than members of the oregmae group (proportion between length of pterostigma and metacarpal vein is 0.90–1.1 in gracilis group versus 0.75–0.90 in oregmae group).
Five new species belonging to the
L. gracilis group were discovered in this study:
Lipolexis labialis sp. n.,
L. pelopsi sp. n.,
L. takadai sp. n.,
L. pakistanicus sp. n. and
L. peregrinus sp. n. Furthermore, within the
gracilis group, two additional clusters are formed, one constituted of
L. gracilis s. str. and
L. labialis sp. n., and the second of the remaining species (
L. pelopsi sp. n.,
L. takadai sp. n.,
L. pakistanicus sp. n. and
L. peregrinus sp. n.) with the average between group distance of 11.4%. Although
L. labialis sp. n. is most closely related to
L. gracilis (average genetic distance of 7.3%), the two species are readily distinguished by their differing number of labial palpomeres.
L. gracilis possesses one labial palpomere while
L. labialis sp. n. has two. These two species have overlapping distributions in Europe, but may use different hosts. All reared specimens of
L. labialis employed Macrosiphini aphid hosts (
Myzus Passerini,
Roepkea Hille Ris Lambers) excepting one reared from
Anoecia corni (F.). Although
L. gracilis has also been reported from
Myzus [
43,
50], in the light of the new species discoveries, it is not clear whether the specimens reared from
Myzus are indeed
L. gracilis or might be some other
Lipolexis species, such as
L. labialis sp. n. or
L. peregrinus sp. n.
In this phylogenetic analysis, within the second cluster of the gracilis group, L. takadai sp. n. wasrepresented by just a single specimen from China. However, morphological examination of other Lipolexis revealed two specimens from Japan which fully correspond to it. Although molecular data indicate that L. takadai sp. n. is most closely related to L. peregrinus sp. n., L. pakistanicus sp. n. and L. pelopsi sp. n. (3.9%, 4.5%, 6.8% average COI divergence respectively), it possesses several morphological differences from other species of Lipolexis. Although it is only known from China and Japan, its actual distribution may be much broader. Lipolexis takadai sp. n. was reared from Aphis gossypii Glover, a common host for other Lipolexis.
A second species from this cluster, Lipolexis pelopsi sp. n., is most closely related to L. peregrinus sp. n., L. pakistanicus sp. n. and L. takadai sp. n., with 6.9%, 8.6%, 6.8% average COI distance, respectively. It has a Mediterranean distribution as it was reared from aphids collected in Greece, Bosnia and Herzegovina, Croatia, and Montenegro. It is a parasitoid of aphids in the genus Aphis attacking various plant species. Lipolexis pelopsi sp. n. coccurs in this region with L. gracilis and L. labialis sp. n., and also shares an aphid host (Aphis sp.) with L. gracilis. It is characterised by a pubescent body (specimens from Greece possess heavily setose mesoscutum), a feature that distinguishes it from all other species.
Lipolexis peregrinus sp. n., another species in the same cluster, is most closely related to
L. pakistanicus sp. n.,
L. takadai sp. n. and
L. pelopsi sp. n. (2.2%, 3.9%, 6.9% average COI distance, respectively). Its specimens were collected in both Asia (Japan, China) and Europe (Spain, Slovenia). Further investigations are needed to ascertain if
L. peregrinus sp. n., is present across the Palaearctic or if it was introduced from Asia to Europe or vice versa. Given its phylogenetic position within the
gracilis group and its number of palpomeres, it probably originates from the Oriental region.
Lipolexis peregrinus sp. n. is phylogenetically closest to Asian species and its maxillary palps have 3 palpomeres, a trait that characterizes Asian species, while the European species (i.e.,
L. gracilis,
L. pelopsi sp. n. and
L. labialis sp. n.) possess maxillary palps with 4 palpomeres. In Asia,
L. peregrinus sp. n. was found to parasitize aphid genera
Aphis L.,
Melanaphis and
Toxoptera Koch [
51,
52] (in both studies identified as
L. gracilis), while in Europe it was reared from
M. persicae and
Aphis sp.
Lipolexis pakistanicus sp. n. is phylogenetically most closely related to L. peregrinus sp. n., L. pelopsi sp. n. and L. takadai sp. n. (2.2%, 8.6%, 4.5% average COI distances, respectively). With the exception of two specimens from Bangladesh and two from Moldova, all material was collected in Pakistan. As most specimens were collected in Malaise traps, the aphid host spectrum is unknown, but one specimen from Pakistan was reared from A. gossypii. This species is likely distributed across the Oriental region. In Bangladesh, it co-occurs with L. oregmae and L. bengalensis sp. n., but more information is needed on host use to ascertain the extent of overlap.
Although L. bengalensis sp. n. is a member of the oregmae group, based on both sequence results and morphology, it is genetically distant from L. oregmae (19.9% average), implying the species have long been separate. At the moment, L. bengalensis sp. n. is only known from Bangladesh where it was collected in a vegetable crop field where L. oregmae was also collected, indicating their co-occurence. Its aphid host is unknown because it was collected by Malaise trap.
Starý [
53] reported that
L. gracilis is common in steppe habitat, orchards, and wood edges. Our material of
L. gracilis originates from high montane habitat, mixed forests, steppes, orchards, and urban area. Although
L. gracilis mostly parasitizes hosts from the tribes Aphidini (genus
Aphis Linnaeus) and Macrosiphini (genera
Brachycaudus and somewhat less
Myzus Passerini), it has been reared from other unrelated aphid genera, such as
Anoecia Koch,
Myzocallis and
Therioaphis [
43]. By comparison, the Oriental
L. oregmae is mostly associated with aphids from the tribes Aphidini (genera
Aphis,
Rhopalosiphum Koch and
Toxoptera Koch) and Macrosiphini (genera
Cavariella,
Liosomaphis,
Myzus,
Pentalonia,
Semiaphis and
Sitobion), but it also attacks representatives of some tribes ignored by
L. gracilis such as Cerataphidini (
Ceratovacuna lanigera Zehntner), Greenideini (
Greenidea psidii van der Goot (=
formosana (Maki)) and Tuberolachnini (
Tuberolachnus salignus (J. F. Gmelin)) [
43]. However, in light of our discovery of six new species, host use might not be that broad as previously thought so trophic associations for these two species require careful validation.
With the exception of L. labialis sp. n., all other species of Lipolexis parasitize Aphis despite their sympatry in the same habitat. For instance, in Europe, both L. gracilis and L. labialis sp. n. parasitize Aphis spp. Furthermore, three species were found in Montenegro, i.e., L. labialis sp. n., L. pelopsi sp. n. and L. gracilis, while, L. labialis sp. n., L. gracilis and L. peregrinus sp. n. were recorded from Slovenia. As a result, it seems clear that speciation in the genus Lipolexis has not been driven by host specialization.
Several studies have reported unusually high COI distances between species within the subfamily Aphidiinae [
47,
49]. For example Kocić et al. [
47] found COI average distances of up to 20.7% among species in the phylogenetically old aphidiine genus
Ephedrus. Similarly, some species of
Lipolexis showed even higher average distances (i.e., 23.2% between
L. bengalensis sp. n. and
L. peregrinus sp. n.). While adults of
Ephedrus possess several plesiomorphic taxonomic characters, this is not the case for
Lipolexis. In fact,
Lipolexis is characterized by several apomorphic taxonomic characters, such as reduced wing venation, elongated petiole, and shape of the ovipositor sheath. Its apomorphies and phylogenetic studies at the subfamily level (where it is clustered with evolutionary younger genera
Binodoxys and
Trioxys) suggest this genus evolved quite recently. However, the average genetic distances of COI, which are even higher than those in
Ephedrus, might indicate that
Lipolexis diversification was not a recent event. Furthermore, Finlayson [
25] stated that the larval morphology of
Lipolexis exhibits plesiomorphic morphological characters. One possibility might be that
Lipolexis separated early in the evolution of Aphidiinae and acquired adult apomorphic traits independently. Further research is needed in order to try to resolve the complex and poorly investigated position of
Lipolexis and its relationships with the other aphidiine members.