Variation in lipid synthesis, but genetic homogeneity, among Leptopilina parasitic wasp populations

Abstract Lipid synthesis can have a major effect on survival and reproduction, yet most insect parasitoids fail to synthesize lipids. For parasitic wasps in the genus Leptopilina, however, studies have suggested that there is intraspecific variation in the ability for lipid synthesis. These studies were performed on only few populations, and a large‐scale investigation of both lipogenic ability and population genetic structure is now needed. Here, we first examined lipogenic ability of nine Leptopilina heterotoma populations collected in 2013 and found that five of nine populations synthesized lipids. The 2013 populations could not be used to determine genetic structure; hence, we obtained another 20 populations in 2016 that were tested for lipogenic ability. Thirteen of 20 populations (all Leptopilina heterotoma) were then used to determine the level of genetic differentiation (i.e., haplotype and nucleotide diversity) by sequencing neutral mitochondrial (COI) and nuclear (ITS2) markers. None of the 2016 populations synthesized lipids, and no genetic differentiation was found. Our results did reveal a nearly twofold increase in mean wasp lipid content at emergence in populations obtained in 2016 compared to 2013. We propose that our results can be explained by plasticity in lipid synthesis, where lipogenic ability is determined by environmental factors, such as developmental temperature and/or the amount of lipids carried over from the host.

Here, we first examined lipogenic ability of nine Leptopilina heterotoma populations collected in 2013 and found that five of nine populations synthesized lipids. The 2013 populations could not be used to determine genetic structure; hence, we obtained another 20 populations in 2016 that were tested for lipogenic ability. Thirteen of 20 populations (all Leptopilina heterotoma) were then used to determine the level of genetic differentiation (i.e., haplotype and nucleotide diversity) by sequencing neutral mitochondrial (COI) and nuclear (ITS2) markers. None of the 2016 populations synthesized lipids, and no genetic differentiation was found. Our results did reveal a nearly twofold increase in mean wasp lipid content at emergence in populations obtained in 2016 compared to 2013. We propose that our results can be explained by plasticity in lipid synthesis, where lipogenic ability is determined by environmental factors, such as developmental temperature and/or the amount of lipids carried over from the host.
Unlike many other animals, several insect parasitoids were found to lack the ability for lipid synthesis. These insects fail to synthesize storage lipids following sugar-feeding, which typically stimulates lipid synthesis (Visser & Ellers, 2008). Parasitoids have a parasitic larval lifestyle, where development is spent feeding in or on an arthropod host (Godfray, 1994). The ability for lipid synthesis was lost repeatedly during the evolution of distinct parasitoid taxa, including beetles, flies, and wasps, as a consequence of the parasitic larval lifestyle (Visser et al., 2010). Parasitoid larvae can readily consume the lipid stores of their host, suggesting that lipid synthesis in parasitoids is redundant or even costly to maintain (Visser, Willett, Harvey, & Alborn, 2017). While the majority of parasitoids lack the ability for lipid synthesis, several phylogenetically distinct taxa were found capable of lipid synthesis (Visser et al., 2010). As the lack of lipid synthesis was found to be ancestral in parasitic hymenoptera, lipid synthesis seems to have re-evolved independently in some parasitic wasp species.
Between-species variation in the ability for lipid synthesis became evident by testing a large number of taxonomically distinct parasitoid species (Visser et al., 2010), but only few species were tested repeatedly for the ability to synthesize lipids (Giron & Casas, 2003;Rivero & West, 2002;Visser et al., 2012Visser et al., , 2017). An exception are species in the genus Leptopilina, which have been popular model systems for a multitude of research fields, including, but not limited to, studies on (theoretical) ecology and behavior (e.g., foraging behavior), chemical communication (e.g., host-finding cues), life histories (e.g., time vs egg limitation), and physiology (e.g., host immunity) (Fleury, Gibert, Ris, & Allemand, 2009;Haccou, Vlas, Alphen, & Visser, 1991;Heavner et al., 2017;Janssen, van Alphen, Sabelis, & Bakker, 1995;Visser, van Alphen, & Hemerik, 1992;Wertheim, Vet, & Dicke, 2003). Initially, L. heterotoma (Figure 1) was found to lack lipid synthesis (Eijs, Ellers, & van Duinen, 1998), but data on another population later revealed active lipid synthesis (Le Lann et al., 2014;Visser et al., 2010). In a study using the closely related species Leptopilina boulardi, four populations were tested using the same host species that revealed contrasting lipogenic phenotypes: two populations synthesized lipids, while two populations did not (Moiroux et al., 2010). Later work on these same four populations then revealed a strong genetic structure with populations synthesizing lipids being genetically closer to each other than to populations that lacked lipid synthesis (Seyahooei, van Alphen, & Kraaijeveld, 2011). These results suggest that genetic divergence corresponds to the observed variation in ability for lipid synthesis in L. boulardi

populations.
A large-scale investigation of both the ability for lipid synthesis and population genetic structure (haplotype and nucleotide diversity) in Leptopilina wasps is now needed. Here, we started by collecting nine different L. heterotoma populations from the field in Europe in 2013 and tested these populations for the ability to synthesize lipids. Based on previous results in Leptopilina (Eijs et al., 1998;Le Lann et al., 2014;Moiroux et al., 2010;Visser et al., 2010), we expected to find variation in the ability for lipid synthesis between populations. Intraspecific variation in ability for lipid synthesis was indeed observed between these populations, but all nine cultures perished before genetic structure could be determined. In a renewed effort, a total of 20 populations from Europe and Asia were then obtained from other laboratories or the field in 2016: 19 populations belonging to three Leptopilina species (L. heterotoma n = 13 populations; L. boulardi n = 4 populations; and L. victoriae n = 2 populations), and one population of a closely related species, Ganaspis brasiliensis (Hymenoptera: Figitidae). The latter species is phylogenetically close to Leptopilina, and a potential biocontrol agent against the pest Drosophila suzukii, which has not yet been tested for lipogenic ability. We then established the genetic structure (including measures of haplotype and nucleotide diversity) of all 13 L. heterotoma populations obtained in 2016 by sequencing the mitochondrial COI gene and the nuclear Internal Transcribed Spacer 2 (ITS2) gene region to quantify genetic divergence between populations. While we predicted to observe variation and genetic differentiation between these Leptopilina populations/species, none of the 20 populations tested were found to synthesize lipids and virtually no genetic differentiation was found between the 13 L. heterotoma populations. We discuss how differences between the 2013 and 2016 populations can be explained.

| Insects
In 2013, Drosophila melanogaster (Diptera: Drosophilidae) hosts were obtained from a culture collected in Dwingeloo, the Netherlands (see Supporting information Table S1 for GPS coordinates). Hosts were maintained in flasks with continuous access to food medium (20 g agar, 35 g yeast, 50 g sugar, 5 ml nipagin containing 100 g 4-methyl F I G U R E 1 Model parasitic wasp Leptopilina heterotoma. Photograph courtesy of Hans Smid from BugsinthePicture, www. bugsinthepicture.nl hydroxyl benzoate in 1L 96% alcohol, and 5 ml propionic acid per liter water) that was replaced every 3-4 days at a temperature of 20°C, a relative humidity of 75%, and a photoperiod of L:D 16:8. In 2016, D. melanogaster were obtained from an existing laboratory culture that was originally collected in Sainte-Foy-les-Lyon in France in 1994. Hosts were maintained in cages with continuous access to food medium at a temperature of 24°C, a relative humidity of 30%, and a photoperiod of L:D 16:8.  information Table S1). Wasp cultures were maintained at a temperature of 23°C, a relative humidity of 75%, and a photoperiod of L:D 16:8. We choose to increase the rearing temperature of wasps in 2016 to be able to maintain populations from all geographic areas (i.e., all populations obtained from other laboratory were already maintained at 23°C).

| Testing for lipogenic ability
To test whether wasps synthesize lipids, we conducted feeding experiments similar to those performed in previous studies (Eijs et al., 1998;Le Lann et al., 2014;Moiroux et al., 2010;Visser et al., 2010Visser et al., , 2012. Using this method, a comparison is made between the total amount of storage lipids present right after emergence from the host, that is, teneral lipid levels, and the amount of lipids after feeding on a sugar source (up to 14 days). Lipid extractions were performed using gravimetry as described in Visser et al. (Visser et al., 2010), with the exception that individuals were dried in an oven at 60°C for 3 days before and after extraction of lipids rather than freeze-dried. Lipid levels were then calculated by subtracting the lipid-free dry mass from the lipid-containing dry mass, after which the percentage fat was calculated. In 2013, only females were tested, but in 2016, males were used, because females were used for maintaining cultures of all populations. Although females are typically larger and contain more lipid reserves, there was no a priori assumption that the ability for lipid synthesis would differ between the sexes. To indeed verify that sex did not affect lipogenic ability, similar experiments were performed with females of three of the 2016 L. heterotoma populations (Leiden, the Netherlands; Wilsele, Belgium; Eupen, Belgium; Table 1).

| Statistics
We are primarily interested in testing whether lipid synthesis occurs within populations; hence one-way ANOVAs or Mann-Whitney Utests (in case of non-normal data/heterogeneity of variances) were performed for each population separately. A significant increase in lipid levels after sugar-feeding suggests that lipid synthesis has occurred, whereas lipid synthesis is lacking when lipid levels remain stable or decrease (Eijs et al., 1998;Ellers, 1996;Visser et al., 2010Visser et al., , 2012. We further compared teneral lipid content of female wasps obtained in 2013 and 2016, and between 2013 populations synthesizing and lacking lipid synthesis, to determine whether and when host lipid content may affect lipogenic ability of wasps using oneway ANOVAs. Statistics were performed using R project version 3.4.1 (R Development Core Team, 2016).

| Genetic structure of L. heterotoma populations
DNA extraction-Total DNA was extracted from two to five adult males for each of the thirteen L. heterotoma populations using the Cetyl Trimethyl Ammonium Bromide (CTAB) extraction method [described in (Navajas, Lagnel, Gutierrez, & Boursot, 1998)]. In short, each male was snap-frozen in liquid nitrogen and crushed with a plastic pestle in a 1.5-ml microcentrifuge tube. Two hundred μl of extraction buffer (2% CTAB, 1.4 M NaCl, 0.2% 2-b mercaptoethanol, 20 mM EDTA, 100 mM TRIS-HCL, pH 8.0, 65°C) and 4 μl protein kinase K (10 mg/ml) were then added, after which samples were incubated at 65°C for 1 hr. Proteins were then removed by adding 200 μl of chloroform/isoamyl alcohol (24/1) and DNA precipitated by adding one volume of isopropanol. Samples were then rinsed with ethanol (76% v/v ethanol containing 10 mM ammonium acetate) and resuspended in 20 μl ultra-pure water. Two microliters RNase (100 μg/ml) was then added and samples incubated at 37°C during 30 min.
PCR amplification and sequencing-Two partial DNA fragments of the COI gene and ITS2 DNA region were amplified and sequenced.
Amplifications were performed using a Veriti Thermal Cycler (Applied Biosystems) with an initial denaturation step at 94°C for 2 min, followed by 35 cycles with 30 s at 94°C, 30 s at 48°C, and 1 min at 72°C with a final extension cycle of 10 min at 72°C for COI.
For ITS2, we used an initial denaturation step at 94°C for 2 min, followed by 35 cycles with 30 s at 94°C, 30 s at 59°C, and 1 min at 72°C with a final extension cycle of 7 min at 72°C. Ten microliters of PCR product purified with Illustra ExoProstar (GE Healthcare) was prepared and send out for sequencing in both directions (3730xl DNA Analyzer; Macrogen Inc., Amsterdam). Sequences were aligned, after which consensus sequences were generated using Geneious ® software version 10.0.9 (Kearse et al., 2012). Consensus sequences (Continues)

| Lipogenic ability
Lipid synthesis varied between populations obtained in 2013. Five of nine populations increased lipid levels, whereas lipid levels remained stable or decreased in the other four populations (Table 1). This is in stark contrast with findings for the 2016 populations, where none of the populations were found to synthesize lipids, including one population that was collected at the same location both years (Table 1).  parasitic lifestyle, and that lipid synthesis was a discrete trait, that is, a wasp species either synthesizes lipids or it does not (Visser et al., 2010). Lipid synthesis was then found to vary intraspecifically in the parasitic wasp genus Leptopilina, but only one or few populations were ever tested simultaneously (Eijs et al., 1998;Le Lann et al., 2014;Moiroux et al., 2010;Visser et al., 2010). To gain a better un-   Year obtained % fat TA B L E 2 (Continued) between L. heterotoma from different localities in Japan for COI, but unlike our findings, the COI sequence of a population collected in

| D ISCUSS I ON
France matched with the one collected in Japan (i.e., Sapporo, from which our Japanese population also originated). In another phylogenetic study, little sequence divergence was found between L. heterotoma originating from France and the Netherlands, but here only a single individual was sampled per population and only three populations were compared (Schilthuizen et al., 1998). These authors suggested that studying the biogeography of Drosophila parasitoids, including Leptopilina, is hampered by the potential human-assisted colonization of new geographic areas. This particularly applies to L. heterotoma, a generalist that has been found on most continents (Nordlander, 1980) and is in line with estimates of genetic divergence in D. melanogaster (Schlotterer & Tautz, 1994 (Visser et al., 2010) reported teneral lipid levels of 26% (±0.6, 1 SE) and 23% (±0.9, 1 SE), respectively, where the former was found to lack lipid synthesis, and the latter was found to synthesize lipids. These L. boulardi and L. heterotoma strains were reared on the same D. melanogaster host strain (but a different strain from those used here). Overall, female L. heterotoma wasps thus seem to lack lipid synthesis when teneral lipid content lies between ~14% (±1.5, 1 SE) (population from Vouvray, France) and 31% (±1.3, 1 SE) (population from Wilsele, Belgium), but start synthesizing lipids when teneral lipid levels are between ~13% (±0.8, 1 SE) (population from Sankt Goar, Germany) and 23% (±0.9, 1 SE; see findings of Visser et al., 2010). We now need to explicitly test when and how host lipid content affects lipogenic ability in parasitic wasps.

ACK N OWLED G M ENTS
We are grateful to three anonymous reviewers and associate edi-

DATA ACCE SS I B I LIT Y
Data will be made available as supporting information Data S1. DNA sequences: Genbank accessions MG561215 -MG561267.