Interference of field evidence, morphology, and DNA analyses of three related Lysiphlebus aphid parasitoids (Hymenoptera: Braconidae: Aphidiinae).

This study provides evidence on integrating the morphological, field, and laboratory data, and application of the cytochrome oxidase subunit I (COI) barcoding gene to the three asexual or sexual Lysiphlebus spp., i.e., Lysiphlebus cardui (Marshall), Lysiphlebus confusus Tremblay and Eady and Lysiphlebus fabarum (Marshall) (Hymenoptera: Braconidae: Aphidiinae). New aphid- invasive plant association, Aphis fabae Scopoli (Hemipreta: Aphididae) on Impatiens glandulifera Royle, has been used in the same model area in the Czech Republic under the same sampling and rearing method for several consecutive years and throughout the season. For molecular identification of these three species, we used DNA sequences of the barcoding region of the mitochondrial COI gene. Although our results confirmed ecological and morphological differences among L. cardui, L. confusus, and L. fabarum, genetic analysis on the basis of COI mitochondrial barcoding gene does not support species status of the mentioned Lysiphlebus taxa. The level of morphological differentiation in these Lysiphlebus Förster species is in accordance with the usual species variability within subfamily Aphidiinae. However, it should be examined how appearance of asexual lineages affects the morphological or genetical variability.

Aphidiine parasitoids are one of the few groups of endoparasitoid insects whose fauna can be studied relatively easily by region. The updated approach to faunistical research around the world has been based mainly on the tritrophic (parasitoid–aphid–plant) associations and subsequent derivable interactions of the individual food webs in an ecosystem ( Pike et al. 2000 , Kavallieratos et al. 2004 , Starý 2006 ). The results can be utilized in a set of approaches starting from the fauna, ecology, biodiversity, and ecosystem interactions, among others, up to the ecologically friendly management of pest aphids.

Lysiphlebus spp. belong to the most common aphidiine parasitoids of aphids. Their host range is predominantly associated with the Aphidinae aphids who are phylogenetically the youngest; in addition, most of the species are numerous and abundant in ecosystems ( Starý 1981 , Starý and Rejmánek 1981 , Kavallieratos et al. 2004 , Ortiz-Rivas and Martinez-Torres 2010 ). Species composition of the genus Lysiphlebus is not very rich. Three centers of origin tend to be distinguishable: West Palaearctic, Far East ( Starý et al. 2002 ), and North American ( Pike et al. 2000 ). Some of these species have invaded other areas either as naturally expansive ( Starý 1995 ) or as purposely introduced species (i.e., Australia; Carver and Franzmann 2001 ).

The taxonomy of the genus and the species identification has been examined in this research. Because of the disorder of the molecular, ecological, and morphological data, the status of at least three European species is questionable: Lysiphlebus cardui (Marshall), Lysiphlebus confusus Tremblay and Eady, and Lysiphlebus fabarum (Marshall). Moreover, there are both sexual and asexual populations distinguishable in their distribution range ( Němec and Starý 1985 , Belshaw et al. 1999 , Sandrock et al. 2011 ).

On the basis of their long-term evidence concerning the host range of these three sexual or asexual Lysiphlebus species, their host preferences, and their sympatric occurrence, we have attempted to demonstrate their taxonomical differentiation in a definable model situation: the same model site in Czech Republic, the same sampling and rearing method in several subsequent years, and throughout the season and the same plant–aphid association. This study brings evidence of integrating the morphological, field, and laboratory data, and application of the COI barcoding gene to the three asexual or sexual L . cardui , L. confusus, and L. fabarum .

Materials and Methods

This study provides the background to long-term research on the associations of the exotic invasive Impatiens glandulifera Royle and the local parasitoid guilds from 1996 to 2011. A new, partially exotic aphid– plant association, Aphis fabae Scopoli (Hemiptera: Aphididae) on I.glandulifera , is now quite common all over the country ( Slavík 1996 , 1997 ). It has become widely distributed in wetland habitats, prevailing along the rivers, ponds and brooks, humid lowland forests, and even the edges of wet meadows and bare grounds all over the Czech Republic. It is a progressively expanding species, extending rather quickly even in 1 yr ( Slavík 1996 ) to nearby habitats through the natural dissemination of seeds by birds in flood waters.

The respective field research was realized in the model locality: Czech Republic, southern Bohemia, between Třeboň and Jindřichův Hradec, close to the city Stráž nad Nežárkou, and the village Pístina (faunistic grid 6955), a mixed lowland forest ( Picea abies (L.), Betula pendula Roth, Fraxinus betulus (L.), Alnus glutinosa L., and Salix caprea L.). The length of the model site is about 140 m and contains a fish pondside and a birch avenue along the road, with 5 m wide groves of I.glandulifera on the sides.

Sources of potential target host aphid–plant associations were occasionally found on the site edges or in a close neighboring area : Aphis farinosa J.F. Gmelin (Hemiptera: Aphididae) on S.caprea and Aphis fabae cirsiiacanthoidis Scopoli (Hemiptera: Aphididae) on Cirsium arvense (L.) Scop.

Samples were collected at two weekly intervals. At each sampling date, only one-third of an aphid infested plants were sampled to prevent an adverse effect on the aphid population. Parts of aphid-infested plants were gently cut with scissors and transferred into plastic translucent containers with nylon mesh at their tops. The containers were transported in field cool boxes up to the laboratory, where they were maintained under a 22°C temperature, 18 h of photoperiod (fluorescent light), and 70% relative humidity (RH) and were lightly sprinkled with water twice a day. The containers were inspected daily for the emergence of parasitoids. The aphid species transfer trials were realized in the same environment, and A. fabae on Vicia faba L. association was used. The bean plants were grown from seeds in wet conifer sawdust, in a rearing container of 35- by 35- by 35-cm size, covered with nylon mesh. Interactions of three parasitoid species associated with A. fabae on I. glandulifera and the attending ants were investigated from 2000 to 2006.

The number of A. fabae on I. glandulifera infested plants varied depending on the overall degree of aphid infestation. In most cases, there was a single aphid colony found per plant, but if two or more were found, these were sampled together.

The overwintering of parasitoid mummies sampled in the fall (October) and kept in the open until the next February was examined in air-conditioned cabinets under the conditions mentioned above. All the examined material is deposited in coll. P. Starý (České Budějovice). New records of aphidiines on aphid–host plant associations are indicated with an asterisk in front of each host plant name.

Molecular Analyses

For the molecular identification of these three species, we used DNA sequences of the barcoding region of the mitochondrial cytochrome oxidase subunit I (COI) gene. In total, 15 specimens were sequenced ( Table 1 ). DNA was extracted using the KAPA Express Extract kit (Kapa Biosystems, Inc. Boston, USA) following the manufacturer’s instructions. A barcoding region of COI gene was amplified using primers LCO1490 (5′-GGTCAACAAATCATAAAGATATTGG-3′) and HCO2198 (5′- TAAACTTCAGGCTGACCAAAAAATCA-3′); (Folmer et al. 1994 ). We used polymerase chain reaction protocols and cycling conditions, which were previously published by Petrović et al. (2013) .

Table 1.

Sampling data for specimens used for molecular analyses of three Lysiphlebus species collected in Pístina, Czech Republic

Code	Date	Parasitoid	Aphid	Plant
Lys4	5 Sept. 2000	L . fabarum	A. fabae	I. glandulifera
Lys6	3 July 2003	L. fabarum	A. fabae	I. glandulifera
Lys7	3 July 2003	L. fabarum	A. fabae	I. glandulifera
Lys9	9 July 2003	L. fabarum	A. fabae	C. arvense
Lys11	24 July 2003	L. cardui	A. fabae	I. glandulifera
Lys12	24 July 2003	L. cardui	A. fabae	I. glandulifera
Lys13	24 July 2003	L. cardui	A. fabae	I. glandulifera
Lys14	24 July 2003	L. cardui	A. fabae	C. arvense
Lys15	24 July 2003	L. cardui	A. fabae	C. arvense
Lys16	24 July 2003	L. cardui	A. fabae	C. arvense
Lys19	24 July 2003	L. confusus	A. fabae	I. glandulifera
Lys20	24 July 2004	L. confusus	A. farinosa	Salix sp.
Lys21	24 July 2004	L. confusus	A. farinosa	Salix sp.
Lys22	24 July 2004	L. confusus	A. farinosa	Salix sp.

Sequences were edited using FinchTV ( www.geospiza.com ). After alignments, conducted using CLUSTAL W integrated in the MEGA5 software ( Tamura et al. 2011 ), sequences showed no indels (insertion or deletion) and were trimmed to a length of 627 bp. All sequences are deposited under accession numbers KM408522-KM408535 in GenBank. For the calculation of genetic distances, we used Kimura’s two-parameter method (K2P) of base substitution. Maximum parsimony and neighbor-joining trees were also obtained using the MEGA5 software. The robustness of the trees was assessed using a bootstrap analysis with 1,000 replicates. A COI sequence of Lysiphlebus testaceipes (Cresson), from GenBank (accession no. KC237764.1) Aphidius avenae Haliday from GenBank (accession no. JN164785.1), and Ephedrus blattnyi Starý also from GenBank (accession no. JN164786.1) were used as outgroups for molecular analyses.

Results

Morphology

The three target species can easily be distinguished morphologically using the characters on the marginal side of the forewing in combination with the pubescence of the hind femur as follows, key ( Figs. 1–6 ):

Fig. 1–6.

Forewing and hind femur (females). 1. Forewing of L. cardui (Marshall); 2. Hind femur of L. cardui (Marshall); 3. Forewing of L. confusus Tremblay and Eady; 4. Hind femur of L. confusus Tremblay and Eady; 5. Forewing of L. fabarum (Marshall); 6. Hind femur of L. fabarum (Marshall).

1 Forewing marginal side with short setae, equal to these on the wing surface ( Figs. 1 and 5 ); hind femur with erected ( Fig. 2 ) or adpressed setae ( Fig. 6 ) …………………………………………………………… 2

Forewing marginal side with setae distinctly longer than the surface setae ( Fig. 3 ); hind femur with semi-erected setae ( Fig. 4 ) … … …… ……………………………………………………………………………  L.confusus

2(1) Hind femur with erected setae ( Fig. 2 ) ……..…………………  L.cardui

Hind femur with adpressed setae ( Fig. 6 ) …………………  L.fabarum

Review of Parasitoid–Aphid–Plant Associations in Czech Republic

The list of the hosts is briefly summarized below using original information by Starý (2006) . For all species, only the Aphis spp. hosts are listed. Populations of all three Lysiphlebus species in the A. fabae on I. glandulifera and C. arvense associations dealt with in this article were asexual.

L. cardui (Marshall) .

Börner on Anthriscus sylvestris Hoffmann, Scopoli on Chenopodium sp., C.arvense , Arctium lappa L., Matricaria perforata Merat, * I.glandulifera, V.faba, Rumex sp ., Calendula officinalis L., Philadelphus coronarius L., Helianthus annuus L., Sylibun marianum L., Solanum nigrum L., J.F. Gmelin on S. caprea L., Salix sp., Kaltenbach on Epilobium parviflorum Schreber, Koch on Cichorium intybus L., Kaltenbach on I.glandulifera , L. on Sambucus nigra L., F.P. Müller on Spiraea vanhouttei (Briot), J.F. Gmelin on Urtica dioica L.

A . fabae was the dominant host species for L. cardui , especially in association with C. arvense.

L. confusus Tremblay and Eady .

Scopoli on * I.glandulifera, C.arvense, J.F. Gmelin on S.caprea, Salix viminalis L.

L . confusus attacking A. farinosa feeding on Salix spp. and A. fabae on I. glandulifera have become common associations in the Czech Republic.

L. fabarum (Marshall) .

Börner on An. sylvestris , Koch on Hypericum perforatum L., Walker on Knautia arvensis L., Ferrari on Trifolium repens L., Koch on Medicago sativa L., Onobrychis sativa Lamarck, Lathyrus pratensis L., Vicia sp., Hartig on Laburnum anagyroides Medik, Scopoli on Beta vulgaris L., S.nigrum, Chenopodium album L., Arctium lappa, C.arvense , * I.glandulifera, Euonymus europaeus L., C. officinalis, Carduus acanthoides L., Dahlia variabilis (Willd), P.coronarius, Spiraea sp., Fagopyrum convolvulus (L.), Kaltenbach on E.parviflorum , Kaltenbach on Hedera helix L., Schrank on Hieracium sp., van der Goot on Rubus idaeus L., Koch on C. intybus , (Börner) on Cirsium oleraceum L., (Börner) on Daucus carota L., Kaltenbach on Camelina alyssum (Miller), Capsicum sp., Theobald on Iris germanica L., Passerini on Origanum sp., Goetze on Plantago major L., P. media L., (Nevsky) on Polygonum sp., de Geer on Malus domestica Borkh, Börner on Sanguisorba minor Scopoli , (Hille Ris Lambers) on Potentilla reptans L., L. on Rumex sp., Walker on Salvia pratensis L., L. on Helichrysum sp., (Börner) on Ribes aureum Pursh, F.P. Müller on S. vanhouttei , Mordwilko on Stachys recta L., Schrank on Symphytum officinale L., (Börner) on Taraxacum officinale Weber, (Börner) on Scabiosa columbaria L., J.F. Gmelin on U.dioica , (Börner) on Achillea millefolium L., Schrank on Verbascum austriacum Schott, (Börner) on Carlina vulgaris L.

L. fabarum was a dominant species in association with A. craccivora, A. fabae on B. vulgaris , Chenopodium L. , and thistles in farmland habitats. Its association with A. fabae on I. glandulifera seems less typical for this species because of its prevailing occurrence in open field type of habitats.

I. glandulifera–A. fabae –Parasitoid Associations at the Model Site and Ant Attendance Association

I . glandulifera was infested by the aphid populations somewhat later in the season because of its phenology. The aphid occurrence lasts almost up to the first freeze in the autumn. A. fabae was observed to start populations from a few individuals occurring on upper parts of the stem and terminal leaves, soon increasing to more or less higher populations, terminating in dense colonies covering even the terminal parts and leaves of the plants, and decreasing at the end of the season (increase of alatae still before the freeze). Aphid numbers varied in the course of the season and through the years.

Parasitoids appeared somewhat later than aphids. Parasitoid numbers varied significantly in the course of the season, specifically representing a combination of L. cardui and L. confusus and less frequently even L. fabarum ( Fig. 7 ). It is also necessary to emphasize the highly important role of I. glandulifera in the oversummering of the aphid as its colonies can be found even outbreaking in the course of the whole summer during the period when most of the local alternative herbal hosts become dried-up and free of aphids.

Fig. 7.

Seasonal occurrence and abundance of L. cardui, L. confusus, and L. fabarum parasitizing A. fabae on I. glandulifera association. Model site: Pístina, 2000. Numbers of parasitoid adults counted from field sampled aphid colonies reared in the laboratory.

Altogether (2000–2006), 17.151 Lysiphlebus individuals were identified to species, with respective numbers (in brackets) for these years as follows: 2000 (3,014), 2001 (1,634), 2002 (3,752), 2003 (5,663), 2004 (34), 2005 (1,920), and 2006 (1,134) specimens.

Ant attendance started almost immediately with the appearance of A. fabae colonies and lasted until the end of the season. Several ant species (Hymenoptera: Formicidae) were observed: Formica fusca L., Lasius emarginatus (Olivier), Lasius platythorax Seifert, Myrmica rubra L., and Myrmica ruginodis Nylander (Hymenoptera: Formicidae). All the five identified ant species manifested rigorous defensive responses to mechanical stimuli. However, whereas the ants mostly disregarded the parasitoid adults, if present, in an aphid colony the Lasius species manifested a different behavior: they defended unparasitized colonies, colonies with some mummies, colonies consisting of merely mummies with no live aphids and empty mummies with emergence holes and without any live aphids.

Interactions of Aphids and Parasitoids Within and With Nearby Ecosystems

An environmental analysis of all the available host species of three Lysiphlebus species in the wetland ecosystem and nearby agroecosystems has determined specific differences. L. confusus is almost uniquely associated with A. farinosa on Salix spp. ( Table 2 ). It apparently switches to A. fabae on I. glandulifera subsequent to a decrease in A. farinosa . Its association with A. fabae on I. glandulifera is significant and common in the studied ecosystem, whereas it rearly parasitizes A. fabae on C.arvense .

Table 2.

Occurrence and abundance of Lysiphlebus species on representative samples in wetland and nearby habitats

Date	L. cardui	L. confusus	L. fabarum	Total
Wetland: A. farinose on Salix spp .
10 May 2000		87 (100%)		87
5 June 2000		310 (100%)		310
Pondside: A. fabae on C. arvense
9 July 2003	18 (54%)		15 (46%)	33
Field: A. fabae on C. arvense
17 July 2005	9 (100%)			9
24 July 2005	94 (96%)	4 (4%)		98

Numbers of parasitoid adults obtained from field-sampled aphid colonies reared in the laboratory. Model site: Pístina.

L. cardui manifests a clearly dominant position of A. fabae on C. arvense and switches from there to A. fabae on I. glandulifera . This can be easily derived from species and from the abundance of data on this association both in the wetland ecosystem and in a nearby agroecosystem. Rarely, it also attacks A. farinosa on Salix spp.

L. fabarum is a less common parasitoid of A. fabae on Cirsium spp. and less commonly parasitizes A. fabae on I. glandulifera.

Laboratory Transfer Trials

The trials targeted the verification of the host alternation of parasitoids of A. fabae on I. glandulifera to conspecific or other aphids in the ecosystem.

L. cardui.

A . fabae on I. glandulifera to A. fabae on V. faba (positive: 1998, 2000, 2003, 2004, 2005).

Note: Overwintering A. fabae on I. glandulifera in the open 2003–2004 accepted A. fabae on V.faba when transferred in the laboratory conditions in February 2004.

A . farinosa on Salix sp. to A. fabae on V. faba (positive: 2004; adverse 1995).

L. confusus.

A . fabae on I. glandulifera to A. fabae on V. faba (positive: 1991, 1998, 1999, 2001).

A . farinosa on Salix sp. to A. fabae on V. faba (positive: 1985, 1995, 1996, 1998, adverse 1998).

L. fabarum.

A . fabae on Chenopodium sp., C. arvense, B. vulgaris to A. fabae on V. faba (positive: 1983, 1985, 1987, 1997, 1996, 2005).

DNA Identification

Calculated K2P distances show that all analyzed specimens of L. fabarum belong to a single haplotype. The same is recorded in L. cardui regardless of the aphid-plant association from which it is collected. Specimens of L. confusus occur in two haplotypes with a K2P distance of 0.6%.

The topology of both the maximum parsimony and neighbor-joining trees showed weak separation of specimens into four clades ( Fig. 8 ). Three clades correspond to analyzed morphospecies, while the fourth represented with one specimen of L. confusus . Although these three species did not share any haplotype, the calculated K2P distances between them were very low (0.2–0.8% between L. fabarum and L. confusus ; 0.5% between L. fabarum and L. cardui ; and 0.3–0.8% between L. cardui and L. confusus ).

Fig. 8.

Top: The first among the 96 most parsimonious trees for the cytochrome oxidase I (COI) gene of a priori determined specimens of L. cardui , L. confusus, and L. fabarum . Bottom: neighbor-joining tree based on Kimura-2-Parameter (K2P) distances. L. testaceipes , Aphidius avenae, and Ephedrus blattnyi were used for the outgroup species. Numbers above or below the branches represent the bootstrap values (%).

Discussion

Although commonly distinguished in the respective morphological keys, the validity and separation of L. cardui , L. confusus, and L. fabarum have been targeted in several articles, reporting more or less different approaches and results ( Carver and Franzmann 2001 , Kavallieratos et al. 2004 ). Moreover, a mixture of all three of these species may occur even together per sample, and such phenomena may contribute to opinions on their species conspecificity. However, morphological, biological, and DNA analyses may sometimes contradict the integration of respective results. Several updated studies bring focus to the species, or species-complex, occurrence of individual species up to specifically identical morphospecies in Lysiphlebus . Belshaw et al. (1999) raised doubts about the L. cardui separation from L. fabarum but also found genetical isolation of L. fabarum on Brachycaudus cardui (L.) (Hemiptera: Aphididae) biotype from other L. fabarum biotypes. Barahoei et al. (2011) carried out morphometric analyses of wing venation, petiole, and femur of five L. fabarum biotypes in Iran. Although they used a small number of biotypes, this result demonstrated morphometric analysis as an efficient tool to distinguish host associated lineages. Sandrock et al. (2011) on the basis of COI and ATP6 data sets found that among sexual and asexual L. cardui , L. confusus, and L. fabarum , there is no isolation between reproductive modes, and thus, these three definable morphospecies represent a single species. Derocles et al. (2012) suggested a combination using two or more genes for accurate identification of the Aphidiinae. L.confusus and L. fabarum, were also considered, noting a sole nuclear gene (long wavelength rhodopsin) and mitochondrial COI gene are not variable enough to discriminate both related species.

The invasion and spread of I. glandulifera into local ecosystems has also been connected with the development of new food webs: I. glandulifera , A. fabae, L. cardui, L. confusus, L. fabarum . Correspondingly, its invasion has been secondarily followed by increasing the biodiversity of the target ecosystem represented by two species of aphids, A. fabae and Impatientinum balsamines (Kaltenbach) (Hemiptera: Aphididae) (Starý 1970), including several predators ( Starý and Láska 1999 ) and pollinators ( Starý and Tkalců 1998 ).

The local associations in the wetland ecosystem interacted with I. glandulifera through the L. confusus , attacking A. farinosa on Salix spp. and L. cardui parasitoids and less frequently with L. fabarum . Both L. fabarum and L. cardui are derivable from the A. fabae on C. arvense association. For all three parasitoid species, the new association and respective refugium on A. fabae feeding on I. glandulifera in wetland habitats is important as a refugium for oversummering and overwintering, at rather high population levels. The field-derived interactions between the associations on A. farinosa feeding on Salix spp. and A. fabae feeding on C. arvense though the parasitoids were also verified on grounds of the laboratory transfer trials. Ant attendance of A. fabae on I. glandulifera realized by five ant species did not affect adversely the presence of Lysiphlebus adults in the colonies, irrespective of (morpho) species identity.

The relatively new and common association for all three Lysiphlebus (morpho)species studied in an identical environment through the season for several subsequent years has supported the demonstration of their specific ecological peculiarities manifested in the host range and host alternation illustrated also by their relative abundance during the season. However, using the COI mitochondrial barcoding gene, we could not clearly separate these three species because of small genetic distances. Although these three species are separately clustered, with small overlapping of L. confusus , genetical distances are rather low (0.2–0.8%) and with relatively small bootstrap values.

In conclusion, although our results confirmed ecological and morphological differences between L. cardui, L. confusus, and L. fabarum , genetical analysis on the basis of the COI mitochondrial barcoding gene does not support species status of the studied Lysiphlebus taxa. Although we designated experiments to the same locality and association following these three species, our results are in agreement with Belshaw et al. (1999) and Sandrock et al. (2011) as far as the use of molecular markers is concerned. The level of morphological differentiation in these Lysiphlebus species is in accordance with the usual species variability within subfamily Aphidiinae, but it should be examined how appearance of asexual lineages affects the morphological or genetical variability.