Genomic and karyotypic variation in Drosophila parasitoids (Hymenoptera, Cynipoidea, Figitidae)

Abstract Drosophila melanogaster Meigen, 1830 has served as a model insect for over a century. Sequencing of the 11 additional Drosophila Fallen, 1823 species marks substantial progress in comparative genomics of this genus. By comparison, practically nothing is known about the genome size or genome sequences of parasitic wasps of Drosophila. Here, we present the first comparative analysis of genome size and karyotype structures of Drosophila parasitoids of the Leptopilina Förster, 1869 and Ganaspis Förster, 1869 species. The gametic genome size of Ganaspis xanthopoda (Ashmead, 1896) is larger than those of the three Leptopilina species studied. The genome sizes of all parasitic wasps studied here are also larger than those known for all Drosophila species. Surprisingly, genome sizes of these Drosophila parasitoids exceed the average value known for all previously studied Hymenoptera. The haploid chromosome number of both Leptopilina heterotoma (Thomson, 1862) and Leptopilina victoriae Nordlander, 1980 is ten. A chromosomal fusion appears to have produced a distinct karyotype for Leptopilina boulardi (Barbotin, Carton et Keiner-Pillault, 1979)(n = 9), whose genome size is smaller than that of wasps of the Leptopilina heterotoma clade. Like Leptopilina boulardi, the haploid chromosome number for Ganaspis xanthopoda is also nine. Our studies reveal a positive, but non linear, correlation between the genome size and total chromosome length in Drosophila parasitoids. These Drosophila parasitoids differ widely in their host range, and utilize different infection strategies to overcome host defense. Their comparative genomics, in relation to their exceptionally well-characterized hosts, will prove to be valuable for understanding the molecular basis of the host-parasite arms race and how such mechanisms shape the genetic structures of insectcommunities.


Keywords
Drosophila, Figitidae, parasitoid, genome size, karyotype introduction Each species has a characteristic genome size and chromosome number. This information often serves as a starting point for obtaining whole genome sequence. It is also useful for cytological or PCR-based genotyping and comparative genomics. Drosophila melanogaster Meigen, 1830 is by far the best-studied insect. Availability of its annotated sequence data (Flybase 2011) is facilitating rapid progress as details of novel gene functions are uncovered and analysis of gene interaction networks and pathways is becoming possible. Sequencing of the D. melanogaster genome also provided the baseline for the analysis of eleven additional Drosophila Fallen, 1823 species, spurring detailed investigation of the evolution of biological processes (Crosby et al. 2007).
Many species of Drosophila serve as hosts to parasitic wasps (Schlenke et al. 2007). In spite of spectacular progress on the model organism itself, practically nothing is known about the genomics or genetics of the parasitic wasps. Leptopilina Förster, 1869 and Ganaspis Förster, 1869 species (Figitidae) attack larval stages, emerge as free-living adults from the pupal cases of their hosts (Schilthuizen et al. 1998, Melk and Govind 1999, Allemand et al. 2002. L. boulardi (Barbotin, Carton et Keiner-Pillault, 1979) is a specialist parasitoid, while L. heterotoma (Thomson, 1862) is a generalist; these species exhibit distinct strategies to evade or overcome host defense (Schlenke et al. 2007, Kraaijeveld and Godfray 2009, Lee et al. 2009). Drosophila-Leptopilina interactions have increasingly become important in understanding innate immunity against natural metazoan parasites and the molecular basis of the arms race between insect host/parasites (Chiu et al. 2006;Kraaijeveld and Godfray 2009, Lee et al. 2009, Paddibhatla et al. 2010. Karyotypes of only two parasitic wasps attacking Drosophila spp., namely, Leptopilina heterotoma with n = 10 (Crozier 1975) and L. clavipes (Hartig, 1841) with n = 5 (Pannebakker et al. 2004) have been previously reported. These initial results indicate considerable karyotypic diversity within the Leptopilina genus, and related taxa. Here we describe the genome sizes and karyotypes of Leptopilina species from the L. heterotoma and L. boulardi clades, as well as that of Ganaspis xanthopoda (Ashmead, 1896), and discuss the relationship and significance of these observations.

Material and methods
Wasps were cultured on the yw strain of D. melanogaster as described in Sorrentino et al. (2004). Origins of the four larval parasitoids of D. melanogaster, namely: Leptopilina boulardi, L. heterotoma, L. victoriae Nordlander, 1980 and Ganaspsis xanthopoda are given in Table 1.
Flow cytometric analysis of genome size, based on nuclei isolated from heads of females of three species of Leptopilina, and Ganaspsis xanthopoda was carried out as described before (Johnston et al. 2004), except that propidium iodide (PI) was added to each sample to a final concentration of 50 µg/ml (not 5µg/ml). Samples were prepared as follows: (A) Each wasp species alone, (B) Drosophila alone, and (C) 4-6 replicates of a wasp head and a Drosophila head prepared together, with mean genome size estimates and standard errors of those estimates based on the 4-6 co-preparations. DNA amount was determined as the ratio of the mean fluorescence of the sample 2C divided by the mean fluorescence of the Drosophila standard, multiplied by the genome size of the standard (1C D. melanogaster = 175 Mb, 1C D. virilis Sturtevant, 1916 = 333 Mb).
Chromosomal preparations for karyology were obtained from cerebral ganglia of male and female prepupae of parasitic wasps according to the technique used by Imai et al. (1988) with modifications. For an initial assessment of the main karyotypic features of Leptopilina spp., metaphase plates from a few preparations of L. boulardi and L. heterotoma were stained with Hoechst 33258 (0.2 µg/ml, Molecular Probes) for five minutes. Images were acquired with a Zeiss Laser 510 Scanning Confocal Microscope and formatted with Zeiss LSM5 software. For detailed karyotype analysis, haploid and diploid mitotic divisions were stained with Giemsa and photographed using Zeiss Axioskop 40 FL optic microscope fitted with an AxioCam MRc camera. Metaphase plates with the best chromosomal morphology were used to obtain karyograms. Chromosomes were classified into four groups (metacentrics, submetacentrics, subtelocentrics and acrocentrics) according to Levan et al. (1964). To obtain karyograms, digital table 1. Origins, genome sizes, and gross karyotypic data of Drosophila parasitoids. Genome size of wasp species correlates with total chromosomal length deduced from karyotypic analysis. The total length of the haploid G. xanthopoda chromosome set differs from both L. boulardi and L. heterotoma at p<0.001, and from L. victoriae at p <0.05; L. boulardi differs from both L. heterotoma and L. victoriae at p<0.001 (T-tests for independent samples). images of metaphase plates were processed with Adobe Photoshop. Measurements of chromosomes were taken using Zeiss AxioVision and then processed with STATIS-TICA (StatSoft Inc. 1995). Relative lengths of chromosomes (RL) were calculated as percentages of the ratio of a particular chromosome relative to total length of the haploid set. Centromere index (CI) is the percentage of the ratio of length of the short arm relative to total length of the particular chromosome.

Genome sizes
The results of the study of genome sizes of the Drosophila parasitoids show almost no intraspecific variation, yet greater than 2.5-fold interspecific variation ( Fig. 1; Table 1). The gametic genome size of Ganaspis xanthopoda (1 C = 971 Mb) is larger than that of any of the three Leptopilina species (370 Mb < 1C < 520 Mb) studied ( Fig. 1). In turn, the genome sizes of all parasitic wasps studied in this paper are also larger than those known for all Drosophila species, which range from 1C = 136.5 to 331.5 Mb (Gregory and Johnston 2008). Our results provide the first information on genome sizes not only of the family Figitidae, but of the superfamiy Cynipoidea as a whole. It is intriguing that the genome sizes of all these parasitoids exceed the average value known for previously studied Hymenoptera, i.e., 360.75 Mb (Tsutsui et al. 2008, Ardila-Garcia et al. 2010), but are fairly close to those of many Chalcidoidea (Tsutsui et al. 2008, Ardila-Garcia et al. 2010, the closest group to cynipoids (see Sharkey 2007).

Karyotypes
Total lengths of haploid chromosome sets of above mentioned species are given in Table 1. The relative lengths and centromere indices of all chromosomes are given in Table 2.
Ganaspis xanthopoda. Nine chromosomes were found in the haploid karyotype of this species (n = 9; Fig. 2a). Chromosomes are long relative to Leptopilina spp.  (see Table 1 and below); most of them are of similar size. However, the first meta-or submetacentric chromosome is about twice as long as the remaining ones. Most other chromosomes are subtelocentric, except for the sixth submetacentric, eighth metacentric, and last acrocentric ones. Leptopilina boulardi. As in G. xanthopoda, n = 9 (and 2n = 18; Figs 2b and c; Fig. 3a). Moreover, the karyotype of G. xanthopoda is superficially similar to that of L. boulardi in that the very large first metacentric chromosome is more than twice as long as the second. However, the length of all remaining L. boulardi chromosomes is roughly half that of the G. xanthopoda chromosomes. Furthermore, chromosomes of the second and third pairs are submetacentric, those of the fourth, fifth, sixth and eighth pairs are subtelocentric, and chromosomes of the seventh and ninth pairs are acrocentric.
L. heterotoma. Consistent with previous observations (Crozier 1975), we found n = 10 and 2n = 20 in this species (Figs 2d and e; Fig. 3b). All chromosomes of the karyotype are submetacentrics that gradually decrease in size.
L. victoriae. This species belongs to the L. heterotoma clade (Allemand et al. 2002), and unsurprisingly, its karyotype is similar to that of L. heterotoma. The haploid karyotype of L. victoriae contains ten submetacentric or metacentric chromosomes (n = 10) of similar size (Fig. 2f ). The first chromosome of L. victoriae is significantly longer and the fifth and tenth chromosomes are significantly shorter than the corresponding chro-mosomes of L. heterotoma. In addition, the centromere position in the first and eighth submetacentric chromosomes is significantly different than observed for the apparently metacentric chromosomes of L. victoriae.

Discussion
Parasitic wasps make up a significant number of species of all insects (LaSalle and Gauld 1993). However, because of the complete absence of genomic information, the molecular biology and genetics of parasitic wasps of Drosophila have lagged behind, even though it is now possible to rapidly sequence genomes of organisms without prior genetic or genomic information.
In a study of genome size of 89 species of bees, wasps, and ants, Ardila-Garcia et al. (2010) hypothesized that genome sizes are constrained by traits associated with parasitism or eusociality. They however found that not all parasitoids have small genomes (Ardila-Garcia et al. 2010). So while it is not altogether surprising that the genomes of the koinobiont parasitoids of Drosophila studied here are as large as that of nonparasitic Hymenoptera, it is intriguing that they have such large genomes. Koinobionts keep their host alive; and must develop and emerge before their host is exhausted and dies. Small genomes replicate faster and require fewer resources, which imposes a selection cost on a bloated genome. An antagonist selective force must act on the parasitoid genome. Because of their obligate and intimate relationship with their fly hosts, it is possible that parasitic wasps take up, or share genetic information via transposons. Widespread transfer of genes laterally has recently been documented from Wolbachia Hertig, 1936 to insect or nematode genomes (Hotopp et al. 2007). Wolbachia has been associated with many parasitic wasps of Drosophila (Vavre et al. 2009). Genomic sequence information will reveal if horizontal transmission of transposons, facilitated by the parasitic life style, may have contributed to the large genome size. In this scenario, different transposon types, with rapid turnover in the genome are expected.
Our karyotypic study provides new insights into the genome structure of Drosophila parasitoids. First, the study demonstrates an obvious positive correlation between the genome size and total chromosome length in those parasitic wasps (Table 1; Fig. 4). However, chromosome length in G. xanthopoda increases relatively slower than might be expected from its larger genome size (Fig. 4). This observation suggests that a significant portion of the bloated Ganaspis genome is repeat sequence that is highly condensed at metaphase. High copy number of satellite DNA is associated with genome size variation in Drosophila species (Bosco et al. 2007) and it is possible that a similar discrepancy in transposon or satellite DNA in the G. xanthopoda genome accounts for smaller than expected increase in chromosome length (Fig. 4).
Second, our study reveals that genome sizes vary independently of the chromosome number in Drosophila parasitoids. This may not be surprising if the large metacentric chromosomes of L. boulardi ( Fig. 2b; Fig. 3a) and G. xanthopoda (Fig. 2a) have an independent origin via chromosomal fusions. Parallel chromosomal fusions are relatively frequent within various lineages of parasitic Hymenoptera (Gokhman 2004(Gokhman , 2009). In addition to Ganaspis and Leptopilina, chromosome numbers of Phaenoglyphis villosa (Hartig, 1841) (n = 10) and Callaspidia defonscolombei Dahlbom, 1842 (n = 11) from the same family have been studied (see Gokhman 2009). This information indicates that n = 10 (or a value close to 10) is likely to be an initial chromosome number for species of the Leptopilina/Ganaspis clade. If this is true, karyotypes with n = 9 found in G. xanthopoda and L. boulardi as well as that with n = 5 found in L. clavipes are likely to have resulted from chromosomal fusions and are therefore derived from a chromosome set that was probably similar to the karyotypes of L. heterotoma or L. victoriae (see also e.g. Gokhman 2010).
Third, the karyotype provides the scaffold number for future sequencing effort in these insects. When the karyotypic features of the species studied here are superimposed onto their phylogeny (Schilthuizen et al. 1998), clear correspondence is revealed: L. heterotoma and L. victoriae share very similar karyotypes, and are the most closely related species. In contrast, L. boulardi belongs to a distinct clade of the Leptopilina genus. Cytogenetic mapping of Expressed Sequence Tags, combined with restrictionsite associated DNA (RAD) sequencing (Baird et al. 2008) based on the karyotype would ensure the highest quality genomic assembly, and pave the way for comparative genomics of parasitoid wasps of Drosophila. Such comparative genomics will provide insights into the organization of the host and parasitoid genomes and the co-evolution of these insects in nature.