Diversity of Ooencyrtus spp. (Hymenoptera: Encyrtidae) parasitizing the eggs of Stenozygum coloratum (Klug) (Hemiptera: Pentatomidae) with description of two new species

Ooencyrtus spp. (Hymenoptera, Chalcidoidea, Encyrtidae) are important natural enemies of agricultural and forest insect pests, and are distributed worldwide. Their reduced dimensions, highly variable morphological characters and possible effect of wide host range and abiotic factors, make correct identification at the species level particularly difficult. This paper combined molecular, morphological, and biological data to characterize a group of Ooencyrtus spp. emerging from the eggs of the variegated caper bug, Stenozygum coloratum in the east Mediterranean area. COI and ITS2 sequencing revealed the presence of six and five divergent clades, respectively. Three clades were identified as Ooencyrtus telenomicida, Ooencyrtus pityocampae and O. pistaciae. Two clades represent new species which are here described and named Ooencyrtus zoeae and Ooencyrtus mevalbelus. These features were combined with reliable morphological characters to facilitate the separation of these species. A dichotomous key and a new synonymy are proposed. Ooencyrtus pistaciae had two distinct COI clades but only one ITS2 clade. Crossbreeding trials that included Ooencyrtus telenomicida, Ooencyrtus melvabelus sp. nov. and Ooencyrtus zoeae sp. nov. confirmed their reproductive isolation. COI sequences showed 0–0.8% and 4–9% within and between-species genetic differences, respectively. ITS2 showed 0.4–5.9% genetic differences between species, with no genetic differences within species. Haplotype diversity of Israeli and Turkish populations of the various species was 0–0.98 and was particularly low in Ooencyrtus pityocampae, whose Israeli population showed no diversity. The discovery of the Ooencyrtus spp. on the eggs of the caper bug, and their abundance support the idea that the bug can be used as an alternative host for augmentation of populations of these parasitoids in agricultural and forestry systems.


Introduction
In recent years molecular tools have been widely applied to facilitate the identification of important biocontrol agents [1] and to separate closely related parasitoid species [2][3][4][5][6]. The variation of morphological characters is often associated with host range, but abiotic factors may also play a significant role [7,8]. An example of the influence of temperature on the colour of adult parasitoids is given by Meteorus pulchricornis (Hymenoptera: Braconidae) [9].
The genus Ooencyrtus includes more than 300 known species, worldwide [10], and many of them are morphologically similar and often are difficult to distinguish [11,12]. Several Ooencyrtus species are considered important natural enemies of pest species [13][14][15][16][17]. However, the identification of species in this genus according to morphological features can be extremely difficult due to their small size and to the intraspecific variation of morphological characters generally used for identification of other congeners, e.g. leg colour, wing venation, antennal proportions and genitalia [10] [Noyes & Guerrieri pers. comm.].
Furthermore, there is a surprising lack of molecular information about species of this genus that could help in a correct identification at the species level.
Ooencyrtus pityocampae Mercet is a well-known egg parasitoid of the pine processionary moth (PPM) Thaumetopoea pityocampae Den and Schiff/T. wilkinsoni Tams species complex (Lepidoptera: Notodontidae), a major defoliator of pine throughout the Mediterranean basin [18][19][20]. This parasitoid is considered a generalist because it parasitizes eggs of a variety of species belonging to the Lepidoptera and the Heteroptera [21][22][23][24]. In the laboratory, O. pityocampae is easily reared on eggs of many other species of both insect orders [16,[25][26][27]. In the east Mediterranean area it was recently identified in the eggs of the variegated caper bug (CB) Stenozygum coloratum (Klug) (Hemiptera: Pentatomidae) [28]. The CB is found in the Middle East and East Africa [29,30], and in Israel it is especially common on caper plants (Capparis spp.) growing within and on the edges of pine forests [31]. The CB oviposition period lasts throughout the spring and summer, mainly from May through September, and parasitism by O. pityocampae occurs throughout this period [31]. In Israel, the eggs of the PPM are found mainly in September through November [32], therefore it was assumed that O. pityocampae population alternated seasonally between these two hosts [30]. These data suggested that the CB plays a role in conservation of the O. pityocampae population, and potentially could be used for augmentation of OP populations in order to improve biological control of the PPM [30]. However, it soon became evident that the Ooencyrtus population on the CB eggs comprises of several additional closely related species, yet their identity was unclear. In an attempt to identify these species, preliminary molecular work was made. The analysis revealed the presence of several highly divergent genetic clades.
The present study aimed to obtain a clear picture of the identity of the Ooencyrtus species obtained from CB eggs by using an integrative approach, combining morphological, biological, and genetic data. We also compared the levels of genetic diversity of the various species. Finally, we examined the potential importance and use of this parasitoid guild for biological control, with emphasis on the PPM.

Insect collection and rearing
CB egg clusters were sampled in 2010-2013 at various sites in Israel and southern Turkey (Table 1). No specific permissions were required for these locations/activities because they were not private properties or parks. Each egg cluster was placed in a separate glass tube (15mm wide and 100mm long) closed with cotton wool, and kept in the laboratory at 25˚C, 40-60% RH and 14:10 L:D. Most emerging parasitoid individuals were transferred directly to 96% ethanol in 1.5-mL Eppendorf tubes and stored at -20˚C, pending subsequent sampling for the molecular study. Some individual parasitoids (male/female pairs) from various sites were kept alive and reared in the laboratory, in order to obtain additional specimens for the crossbreeding experiments and for identification purposes (see below Section 2.3). Each female was placed inside a glass tube (sizes as above), together with a single male that emerged from the same egg cluster. To obtain maximum diversity, no more than one female/male pair was sampled from each egg cluster. For rearing, female parasitoids were given either CB or silk moth eggs (Bombyx mori L., Lepidoptera: Bombycidae) taken from our laboratory culture. A few offspring of each pair were transferred to 96% ethanol, as described above, and used for the molecular identification; a few others were preserved in 80% ethanol and used for morphological identification. In this way we ensured that the identity of each culture could be confirmed both morphologically

Molecular characterization
A total of 274 Ooencyrtus individuals were sampled for the genetic analysis; all had emerged from CB eggs collected at various sites in Israel and southern Turkey (Table 1) However, the forward primer often attached to the middle of the COI segment, so that only the second half of the segment was obtained. To solve this problem, another forward primer was designed, based on the previous one: We designed the primers for ITS2 sequences, based on an alignment of ITS2 sequences of Chalcidoidea superfamily members, which were available in NCBI. The forward and reverse primers were, respectively, 5' GAACTGCAGGACACATGAACA 3' and 5' CTTGTTCGCTAT CGGTCTCGTGGT 3'.
Amplification of the COI and ITS2 used the following conditions: Initial denaturation at 95˚C for 3 min, followed by 36 cycles of 95˚C for 30 s, 54˚C for 30 s, and 72˚C for 75 s (COI) or 1 min (ITS2), followed by final extension at 72˚C for 5 min.
The PCR for COI was performed in a 30-μL reaction volume containing a 3-μL sample, 0.3 μL (30 pmole) of each primer, 0.6 μL of dNTPs (10 mM of each), 0.15 μL (0.75 units) of Taq polymerase (Fermentas, Vilnius, Lithuania), 0.6 μL of MgCl 2 (to a final concentration of 2.5 mM) and 22.05 μL of dd H 2 O. The PCR for ITS2 was performed with the same reagent concentrations except for MgCl 2 , which was at a final concentration of 2 mM.
To ensure good-quality sequencing results, PCR products were electrophoresed through 1% Agarose gel, and the DNA segment was cut from the gel and later extracted and purified with the RBC-YDF HiYield Gel/PCR DNA Fragments Extraction Kit (RBC, Banqiao City, Taiwan). Sequencing was done by Macrogen Corp. (Seoul, S. Korea). The PCR products of each individual, obtained with the reverse and forward primers, were aligned and compared by means of the ChromasPro software, Version 1.6 (Technelysium Pty Ltd, South Brisbane, Queensland, Australia), and were checked manually for errors.
The COI sequences were aligned manually and were translated into amino acids in order to verify that none of the sequences contained any stop codon, which could indicate the presence of numts. There were no gaps in the COI alignment. The ITS2 sequences were aligned by means of the Mafft (ver. 7) Q-ins-I method [33]. Additionally, the COI sequence of Nasonia longicornis Darling (Hymenoptera: Pteromalidae) (Accession No. EU746612) was downloaded from GeneBank and used as a second outgroup (together with O. kuvanae Howard) in the COI phylogenetic analysis.
The Mega program (ver. 6) was used to calculate genetic distances within and between species, according to the Kimura 2-parameter model [34]; the same program was used to construct Maximum Likelihood phylogenetic trees of COI and ITS2 sequences, with bootstrap support (1000 replicates). The DNASP v5.10.1 program [35] and Arlequin v.3.5.1.3 software [36] were used to compute haplotype and nucleotide diversity indices.

Morphological characterization
Live parasitoids were killed in ethanol 80% and kept at -20˚C until preparation. All parasitoids were critically point dried [37] and mounted on card for further examination. Selected card mounted specimens were slide mounted following Noyes [37]. Unless otherwise specified all data on distribution and hosts of Ooencyrtus are from Universal Chalcidoidea Database [10].

Crossbreeding trials
This experiment was conducted to determine whether interspecific breeding between the studied species could occur. We tested the three most similar morphospecies obtained in the present study, which later were identified as O. telenomicida, O. mevalbelus and O. zoeae, of which the second and third were also the most closely related. Glass-tube arenas, each containing three virgin females of one species and two males of another species, were set up in the laboratory, at 25˚C,40-60% RH and 14:10 L:D. These individuals were provided with honey-water solution on a piece of cloth, and left for 3 days before adding 100 silk moth eggs. After 3-5 days adult parasitoids were removed and eggs were kept at the same laboratory conditions till the emergence of new adults. Because the three studied species reproduce sexually (Shahar unpubl. observations), females are expected among the offspring only in case of successful crossbreeding. Five repetitions were made for each crossbreeding treatment, e.g., ♀♀ species1 with ♂♂ species 2, etc. Females and males of the same species were placed together under similar conditions served as controls.

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank.org/". The LSID for this publication is: urn:lsid:zoobank.org:pub:1B2BAB83-8146-4AE3-A28D-EFF873083A12. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

Phylogenetic analysis
In general, molecular analysis revealed the presence of six (COI, Fig 1) or five (ITS2, Fig 2) genetically distinct clades. The partition of individuals into groups was generally similar for both DNA fragments, although there were some exceptions (described below in this Section). Three of these groups belonged to known species, identified as O. pistaciae, O. telenomicida and O. pityocampae. Two groups of morphologically similar individuals had two distinct COI clades, but identical ITS2 sequences, and both are here identified as O. pistaciae. The other two species presently identified, are the most closely related, both morphologically and genetically, and were named O. zoeae. and O. mevalbelus). All species and groups were found in Israel, but only O. zoeae, O. telenomicida and O. pityocampae were also found in Turkey.
The COI mean genetic distances within species ranged between 0.000 and 0.008, i.e., differences of 0-0.8% (Table 2); and genetic distances between species ranged between 0.044 and 0.090 (Table 3). The most closely related were O. pistaciae groups a and b, separated by a mean genetic distance of 0.044, and, as mentioned above, they are suspected of belonging to the same species. The next most closely related pair were O. mevalbelus and O. zoeae, separated by 0.050. The greatest COI mean genetic distance-of 0.090-was obtained between O. pityocampae and the pair O. mevalbelus and O. zoeae. Genetic distances of ITS2 sequences were generally lower: 0.004-0.059 between species, with no observed genetic variability within species (Table 4). Furthermore, the two O. pistaciae COI clades had identical ITS2 sequences. Slightly different relationships were obtained with the ITS2 sequences from those obtained with the COI ones. According to the ITS2 analysis, the most prominent difference was assigned to O. pityocampae. The latter species was the most distinct from all other examined congeners; with genetic distances of 0.052-0.059, compared with 0.004-0.044 between all the others (Table 3). Whereas according to the COI sequences, O. pityocampae was more closely related to O. telenomicida and O. pistaciae, with separations of 0.052 and 0.055, respectively, than O. mevalbelus and O. zoeae, with separations of 0.073 and 0.083, respectively. Therefore, the last two were placed separately from the other three species in the COI phylogenetic analysis (Fig 1)

Genetic diversity analysis
A total of 70 COI haplotypes were obtained: 1-15 for each population (species and country; Table 5). Haplotype diversity ranged between 0 and 0.955, nucleotide diversity was 0-0.037, and Өs values were 0-6.289 for all populations of species from Turkey and Israel. O. pityocampae was the least diverse, with lower haplotype diversity (0.610) than all other species-in which it ranged from 0.800 to 0.955-apart from O. pistaciae group b, of which only three individuals, with identical haplotype, were obtained. This was even more conspicuous among the Israeli O. pityocampae population, in which, in spite of the relatively large sample, no genetic diversity was found in the COI sequences, with only a single haplotype obtained, compared with five haplotypes in the Turkish population, and 3-15 haplotypes in the populations of all the other species from either Israel or Turkey.

Crossbreeding trials
All interspecies crossings between O. mevalbelus, O. zoeae, and O. telenomicida resulted in male offspring only, whereas all control (within-species crossings) trials resulted in both male and female offspring (29-62% females in all replicates). In conclusion, successful cross-mating and hybridization between these three species did not occur.
-Mesoscutum with a slightly brassy, moderate metallic dark green sheen, a little coppery purple anteriorly and along posterior margin; F1 always shorter than F2 (higher half of interantennal prominence purple shiny; sculpture on middle scutellum striate/elongate) Female. Holotype, length 0.9mm. Head black with very faint metallic lustre on frontovertex, interantennal prominence with green-blue reflections; mesothorax slightly shiny, scutellum dull with a faint green lustre at apex; antennal scape ( Fig 3A) yellow with a narrow brown stripe along dorsal margin (except the basal third), remaining parts of antenna brown except the apical half of pedicel appearing paler brown; thorax black, setae on dorsum of thorax slightly silvery; tegula black; legs yellow except apices of all tarsi, brown; wings hyaline, venation brown; gaster black, the first 2 segments yellowish.
Head on frontovertex with fine, shallow, fairly regular, polygonally reticulate sculpture of mesh size clearly less smaller than an eye facet; ocellar angle about 75˚; occipital margin sharp but not carinate; head about 3.7× as broad as frontovertex; antenna (Fig 3A) with scape about 6× as long as broad; all funicular segments longer than broad and with linear sensilla; clava basal suture perpendicular, apical one oblique and with a sensory area along ventral surface of the last segment so that giving it has a slightly obliquely truncate appearance; mandible with 2 teeth and a broad truncation; mesoscutum with sculpture shallow, made of polygonal irregular cells; scutellum with sculpture coarse, made of elongate cells at sides and rounded in the middle; fore wing ( Fig 3B) about twice as long as broad with marginal vein (Fig 3C) 1.5× as long as broad, postmarginal vein punctiform; mid tibial spur shorter than mid basitarsus; gaster with hypopygium extending not more than 0.6× the gaster length; ovipositor hardly exserted. Relative measurements: HW 27, FVW 7.5, POL 3, OOL <1, OCL 4, OD 1, SL 14, SW 2.5, FWL 64, FWW 30, SMV 24, MV 2, PMV 1, SV 3.
Paratype: Hypopygium ( Fig 3D) very transverse about 3× as broad as long, its posterior margin with a median invagination; ovipositor (Fig 3E) about as long as mid tibia or about 5× as long as gonostylus. Gonostylus (Fig 3E) about 0.7× as long as mid tibial spur; second valvifer with 4 or 5 subapical setae. Relative measurements: OL 76, MT 77, GL 14 Variation. Virtually none in the material at hand except for the body length, which varies from 0.8 to 1.1mm. Male (length 0.8mm): generally similar in appearance to female except for antenna ( Fig 3F) and genitalia ( Fig 3G); antenna ( Fig 3F) with all funicular segments much longer than broad and clothed in setae that are generally much longer than diameter of segments; genitalia ( Fig  3G) without parameres and with a single pair of setae at apex; digitus about 3.5× as long as broad and with 1 apical hook; aedeagus slender, apically rounded and about 0.6× as long as mid tibia.
Hosts  Comments: The new species is virtually undistinguishable from O. mevalbelus with which shares a number of morphological features including relative widths of frontovertex, scape, wing and ovipositor proportions. The small differences noted in the key could generally be considered to fall within intraspecific variation but molecular and biological data confirm the different identity of this species. The species is named in honour of Zoe Elizabeth Good.
Ooencyrtus telenomicida (Vassiljev) (Fig 4) Encyrtus telenomicida Vassiliev, 1904: 104. Type material Urkaine. Probably lost. Not examined.   (Fig 4A) brown except scape yellow with a brown stripe along dorsal margin and apical half of pedicel yellow; thorax black with a very faint dark green sheen on mesoscutum; tegula black; mesopleuron black with a faint purple sheen; legs yellow, tarsi a little darker; wings hyaline, venation brown, marginal vein appearing darker; basal half of gaster yellow (up to third tergite), remaining part brown; gonostylus black.
Head about 4× as wide as frontovertex; ocellar angle about 60˚; antenna ( Fig 4A) with scape slender, 7.4× as long as broad; all funicular segments distinctly longer than broad, F2-F6 with linear sensilla, clava slender, pointed at apex, sensory area at apex only; mandible with 1 tooth and a broad truncation. Thorax with fine, polygonally reticulate sculpture on mesoscutum that is shallower than that on frontovertex; scutellum with minute and longitudinally elongate sculpture that is deeper than that on mesoscutum and giving a striate appearance; setae on mesoscutum thin and longer than those on scutellum; scutellum in profile virtually flat; mesopleuron overreaching the base of gaster; fore wing about (Fig 4B) 2.4× as long as broad, with marginal vein (Fig 4C) about 1× as long as broad, postmarginal vein rudimental.
Comments: The type of telenomicida is missing and our concept of the species is based on material in the BMNH identified authoritatively as such. In general, the species has been separated from closely related ones by the extension of the lineolate sculpture on the scutellum. However, material obtained by rearing the species on a laboratory host (silk moth eggs) shows that this character is extremely variable and is probably unreliable for species discrimination. The species may be highly polyphagous and widely distributed throughout the Palaearctic region and may have been misidentified on several occasions. Diagnosis: Female-Head with frontovertex dark brown, with coppery shine; inter-torular area violet; greenish on face and malar space; mouth margin coppery; mesoscutum with faint bluish reflection, a purplish lustre on scutellum and golden green reflection at sides of gaster; antenna ( Fig 5A) brown with apex of pedicel paler; legs extensively brown the following parts yellowish or appearing paler: joints and base of all tarsi; distal 2/3 rd of mid and hind tibiae; wings hyaline (our specimens have a distinct brown spot below the marginal vein), venation pale brown to brown. Head 3.5× as broad as frontovertex, with shallow reticulate sculpture made of cell clearly smaller than eye facet; ocelli forming an angle of about 90˚; mandible with 2 teeth and a truncation; antenna (Fig 5A) with scape about 7× as long as broad, all funicular segments longer than broad and with linear sensilla except F1; mesoscutum with shallow reticulate sculpture made of irregular cells, clearly different from that of scutellum, coarse, and made of cells elongate and narrow, apical third of scutellum smooth and shiny; fore wing ( Fig 5B) 2.2× as long as broad, marginal vein (Fig 5C) about 1.4× as long as broad and as long as postmarginal vein; hypopygium (Fig 5D) transverse, and extending not more than 0.6× the gaster length, its posterior margin with a median invagination; ovipositor ( Fig 5E) hidden or hardly exserted, about 0.9× as long as mid tibia or 6× as gonostylus; gonostylus about 0.5× as long as mid basitarsus. Relative measurements: HW 28, FVW 8, POL 4, OOL 1, OCL 2, OD 1, SL 14, SW 2, FWL 66, FWW 29, SMV 24, MV 1.5, PMV 1.5, SV 3; OL 74, MT 85, GL 13 Male: generally similar in appearance to female except for antenna ( Fig 5F) and genitalia (Fig 5G).
Comments: The species has been described and illustrated by the authors [40] in detail. Key characters of the specimens examined here fit with those of description particularly for the colour of the body (including the different metallic shine on different parts of the head), colour and features of antenna (in both sexes), hypopygium and ovipositor.  Female. Holotype, length 1.1 mm. Head black with very faint metallic lustre on frontovertex, lower part of interantennal prominence with green-purplish reflections; antenna ( Fig 6A) light brown, scape yellow with a brown dorsal stripe along the apical 2/3 rd ; mesothorax faintly shiny, scutellum dull with a green lustre at apex and sides; gaster with base yellow, remaining brown; setae on dorsum of thorax whitish; legs yellow except apices of all tarsi, brown; wings hyaline, venation brown; gaster black, the first 2 tergites yellowish;.
Head with shallow reticulate sculpture made of cell clearly smaller than eye facet, 3.7× as broad as frontovertex; ocelli forming an angle of about 90˚; mandible with 2 teeth and a truncation; antenna (Fig 6A) with scape about 6.5× as long as broad, all funicular segments longer than broad and with linear sensilla; mesoscutum with shallow reticulate sculpture made of irregular cells, clearly different from that of scutellum, coarse, and made of cells elongate and narrow, apical third of scutellum smooth and shiny; fore wing ( Fig 6B) about 2× as long as broad, marginal vein (Fig 6C) GL 14 Variation: virtually none in the material at hand except for the body length, which in females varies from 0.9 to 1.2mm.
Male: generally similar in appearance to female except for antenna ( Fig 6F) and genitalia (Fig 6G).
Hosts Ooencyrtus pityocampae abdominalis (Mercet); Ö ncüer, 1991: 207 Diagnosis: Female-Head with frontovertex dark brown, with green lustre; inter-torular area with green shine; mesoscutum and scutellum with green reflection; antenna (Fig 7A) with scape yellowish with a dorsal brow stripe at apical 2/3 rd , pedicel brown apex paler, remaining segments brown; legs yellow with apex of tarsi and base of hind coxa brown; wings hyaline sometimes with a small brown spot around the marginal vein, venation brown; gaster brown with basal segments paler. "Maroc, iii.59 Mekhés Voegelé Ex oeufs de Aelia B [in pencil]" "Encyrtidae: Ooencyrtus fecundus n.sp. Ch. Ferriére det."; 5♂ (on slide) "Maroc, iii.59 Mekhés Voegelé Ex oeufs de Aelia" "Encyrtidae: Ooencyrtus fecundus n.sp. ♂ Ch. Ferriére det."; 5♀ (dismembered on 1 slide) "Maroc, iii.59 Mekhés Voegelé Ex oeufs de Aelia" "Encyrtidae: Ooencyrtus fecundus n.sp. B[in pencil] Ch. Ferriére det."; 9♀, 6♂ (on minuten pins, only 1♀ 2♂ intact, remainder mostly lost head or other body parts; the female specimen treated similarly to lectotype and is in good condition but the head has slightly collapsed), "Ooencyrtus Aeliae ♀" "COTYPE"; 5♂ (on card points), " Comments: The species is one of the easiest to recognize amongst palaearctic ones because of the distinctive colour/sculpture of mesoscutum and scutellum coupled with yellow legs. According to Ferriére & Voegelé the holotype, allotype and 20 paratype males and females of O. fecundus were deposited in MHNG with a further 20 paratypes (♂♂and ♀♀) deposited in Direction de la Recherche Agronomique in Rabat, Morocco. In the type series there is no indication of the holotype but there are four males labelled as "TYPUS". None of them could be the holotype because the authors explicitly stated that the holotype is female. We then assume that the authors just forgot to properly label the type material and thus, according to ICZN rules, designate here a Lectotype of O. fecundus. We have examined the type series of O.fecundus and have no doubt in synonymize it in pityocampae.

Discussion
The present study focuses on the Ooencyrtus spp. parasitoid complex occurring in the CB eggs. A major challenge in identification of these species arose from the great lack of distinguishable and reliable morphological differences, coupled with high morphological polymorphism. Therefore, a molecular approach was essential. It is likely that without the use of molecular screening, the presence of at least one of the species, O. zoeae, would have passed unnoticed. Generally, the molecular approach enabled a clear separation between the congeners, which was also supported by the crossbreeding trials. Although O. pityocampae and O. pistaciae were not included in these trials, clear genetic separation suggests that they, also, are reproductively isolated from the rest. Furthermore, two species that were tested in the crossbreeding trials and could not crossbreed-O. mevalbelus and O. zoeae-were found to be the most closely related, which suggests that the other congeners, which are more genetically and morphologically distinct, are even less likely to crossbreed.
Nonetheless, further steps are needed in order to complete the identification process and determine phylogenic relationships. Several problems remain to be resolved. First, two distinct COI haplotypes of O. pistaciae were obtained, which raises the question of whether these two clades represent a single species or two separate ones. Furthermore, although the studied species were well separated according to both DNA fragments, it was difficult to determine their phylogenetic relationships, because of the similar distances separating some of the species, and the inconsistency between the relative genetic distances obtained from the respective ITS2 and COI analyses. Consequently, the relationships between species cannot be fully determined. Interesting findings were obtained by phylogenetic analysis with a slightly larger COI fragment of 1237 bases-which included the presently used 946-bp fragment-that was obtained with LCO1490 and the reverse primer used in the present study. These results show a more similar picture to that obtained with the ITS2 (Samra, unpubl. data), in which O. pityocampae was placed outside the grouping of the other four species, thus supporting the hypothesis of an earlier divergence of O. pityocampae from the others. However, the presence of pseudogenes prevented sequencing of O. zoeae with this primer set, and a poly-t region near the 5' end of this DNA segment caused considerable difficulty in obtaining clear sequences also for the other species. Thus, only a few sequences were obtained from four species and, therefore, we were obliged to use the shorter 946-bp fragment.
In Israel of the Ooencyrtus spp. obtained from CB eggs, O. pityocampae is the only uniparental species. Its thelytoky is almost certainly induced by Wolbachia [41]. This endosymbiont bacterium is known to interfere with the reproductive mode of many insect species, and to induce thelytoky in other hymenopteran parasitoids [42,43]. There is evidence that it also may cause a reduction in mitochondrial haplotype diversity through a process of selective sweep [44]. Therefore, Wolbachia also may account for the low genetic diversity observed in COI haplotypes of O. pityocampae. However, both Israeli and Turkish O. pityocampae populations are infected with the same Wolbachia strain (Samra, unpubl. data), and it is not clear why the Israeli population is much less diverse than the Turkish one. Most probably, uniparental populations of O. pityocampae from Morocco are similarly infected [16,45]. Therefore, the exceptionally low genetic diversity in the Israeli O. pityocampae population is probably associated with a different process; possibly it results from a relatively recent expansion of this species, following the arrival and spread of the PPM in Israel [46]. However, biparental poplations of O. pityocampae have been collected and it would be worthy to characterize them to unravel their relation with uniparental ones. In contrast, the level of genetic diversity of O. telenomicida and O. zoeae is similarly high in both Turkish and Israeli populations, which indicates that the situation in O. pityocampae can be considered unique.
Some of the Ooencyrtus spp. identified in the present study are known natural enemies of various Lepidopteran and Heteropteran pests, e.g., Nezara viridula and Aelia spp. (Heteroptera: Pentatomidae) and PPM [14,16,17]. Their frequent occurrences on CB eggs suggest that this bug might be important for conservation of these egg-parasitoid populations in the east Mediterranean area. Although the CB is known to attack various agricultural plant species [47], it is not considered a significant pest, because damage to agricultural crops is usually quite rare and localized [30,47]. In fact, CB individuals probably cannot reproduce on plants other than capers [30], therefore, switching to other plants is most likely to occur only in late summer when capers' vital foliage becomes scarce. Hence, we suggest that enhancing CB populations may lead to an increase in the populations of its egg parasitoids, which, in turn, might help to control the populations of more significant pest species that occur in the same area. This seems especially relevant to the case of O. pityocampae and the PPM. The eggs of the latter species are laid mainly in September-November [14], therefore, the O. pityocampae population most likely relies on alternative hosts to survive through the spring and summer months (April-August) which coincide closely with the reproductive period of CB [31]. Host switching may also be essential for survival of the other discussed Ooencyrtus spp., though, they do not parasitize PPM eggs (Samra unpubl. data).
Finally, we suggest that the molecular data on the various Ooencyrtus species gathered in the present work could help to elucidate the identities of other Ooencyrtus populations, found on various other hosts in the area. Accurate identification is vital to enable validation of their true distributions and host ranges, and evaluation of their role as mortality agents of their hosts.
Supporting information S1 Appendix. List of Ooencyrtus species from Stenozygum coloratum eggs which were misidentified in previous publications. (DOCX)