Transfer of a chromosomal Maverick to endogenous bracovirus in a parasitoid wasp C. Dupuy ��� G. Periquet ��� C. Serbielle ��� A. Bezier �� ��� F. Louis ��� J-M Drezen Received: 4 October 2010 / Accepted: 15 March 2011 / Published online: 31 March 2011 �� Springer Science+Business Media B.V. 2011 Abstract Bracoviruses are used by parasitoid wasps to allow development of their progeny within the body of lepidopteran hosts. In parasitoid wasps, the bracovirus exists as a provirus, integrated in a wasp chromosome. Viral replication occurs in wasp ovaries and leads to for- mation of particles containing dsDNA circles (segments) that are injected into the host body during wasp oviposi- tion. We identified a large DNA transposon Maverick in a parasitoid wasp bracovirus. Closely related elements are present in parasitoid wasp genomes indicating that the element in CcBV corresponds to the insertion of an endogenous wasp Maverick in CcBV provirus. The pres- ence of the Maverick in a bracovirus genome suggests the possibility of transposon transfers from parasitoids to lep- idoptera via bracoviruses. Keywords Bracovirus Wasp genomes Maverick element Horizontal transfer Introduction Polydnaviruses display a unique obligate association with endoparasitoid wasps. They are composed of bracoviruses and ichnoviruses, associated with braconid and ichneumonid wasps, respectively, which originated from different ancestor viruses (Whitfield and Asgari 2003 Volkoff et al. 2010). The viral ancestor of bracoviruses, integrated its genome into the genome of the ancestor of bracovirus- associated wasps approximately 103 �� 10 million years ago (Murphy et al. 2008). Recent data showed that the ancestor virus was related to nudiviruses (Bezier et al. 2009a, b), a group close to baculoviruses. Functionally, bracovirus particles are gene transfer agents used by parasitoid wasps to facilitate development of their progeny within the body of immunocompetent lepidopteran larvae (Kroemer and Webb 2004 Dupuy et al. 2006). The multiple dsDNA circles packaged in the parti- cles exist also as proviral sequences integrated in the wasp genome (Fleming and Summers 1991 Belle et al. 2002). In the wasp ovaries, these proviral sequences are used as a template to produce the dsDNA circles that are then packaged in the bracovirus particles (Herein referred to as the packaged genome). Bracoviral particles are injected during wasp oviposition into the Lepidoptera they enter virtually all tissues and packaged genes are expressed by lepidopteran host cells (Fleming and Summers 1991 Albrecht et al. 1994 Gruber et al. 1996 Savary et al. 1999 Drezen et al. 2003 Belle et al. 2002 Pasquier-Barre et al. 2002). The gene products manipulate host immune defen- ses and development, thereby ensuring the development of parasitoid larvae in an environment that otherwise would be harmful (Webb et al. 2006). Unlike most viruses, PDVs are not transmitted horizontally since no viral progeny are produced in parasitized host tissues. They are exclusively inherited by vertical transmission of proviral sequences integrated in the wasp genome (Belle et al. 2002). During the past few years, the genomes of several polydnaviruses have been sequenced (Espagne et al. 2004 Webb et al. 2006 Lapointe et al. 2007 Desjardins et al. 2008 Choi et al. 2009a) including the packaged form of Cotesia congregata bracovirus (CcBV) that comprises 30 C. Dupuy (&) G. Periquet C. Serbielle A. Bezier �� F. Louis J.-M. Drezen Institut de Recherche sur la Biologie de l���Insecte, UMR CNRS 6035, Universite �� F. Rabelais, UFR Sciences et Techniques, Parc Grandmont, 37200 Tours, France e-mail: catherine.dupuy@univ-tours.fr 123 Genetica (2011) 139:489���496 DOI 10.1007/s10709-011-9569-x

dsDNA circles (genome segments). These segments code mostly for proteins thought to target key pathways of host physiology (Cui and Webb 1997 Espagne et al. 2004 Kroemer and Webb 2004), but they also contain a few genes sharing similarities with viruses or transposable elements (TEs) (Drezen et al. 2006). In a previous study, we identified in a CcBV segment 31 the remnants of tan- demly integrated 2 retroelements, a DIRS and a Dong element (Drezen et al. 2006). In addition, we identified a cluster of 4 genes encoding a DNA polymerase B (POLB), an integrase (INT), a coil-coiled protein and an ATPase 32L protein near those transposable elements. Both POLB and ATPase 32L are used in phylogenetic analyses of viruses (Iyer et al. 2006). However the predicted proteins shared high similarities with a set of proteins encoded by a gene cluster present in two nematode genomes (Drezen et al. 2006). This suggested that this gene cluster might correspond to an exogenous genetic entity not yet identified horizontally transferred between phyla. Here we evidence that these sequences are indeed the remnant of a Maverick element (or Polinton), which belongs to a recently char- acterized class of TEs (Feschotte and Pritham 2005 Kapitonov and Jurka 2006 Pritham et al. 2007). Our analysis shows for the first time the presence of a Maverick in a polydnavirus genome and thus suggests that PDVs could possibly mediate horizontal transfers of Mavericks from hymenopteran to lepidopteran genomes. Mavericks are classified with Helitrons in DNA Class 2, Subclass 2 of transposable elements as they transpose by replication involving the displacement of a single strand (Wicker et al. 2007). Mavericks/Polintons are very large transposons (up to 20 kb) that represent a plausible inter- mediate between eukaryotic viruses and selfish mobile elements (Filee et al. 2008). Indeed, on the one hand they display typical characteristics of TEs such as the integrase gene (INT) and a large (400���700 bp) terminal inverted repeat (TIRs), flanked by 6 bp target site duplications (TSDs). On the other hand, they encode proteins involved in the viral DNA replication cycle, such as DNA-poly- merase B, ATPase 32L, and adenoviral-like protease (PRO) which are encoded by core genes conserved in the genome of Nucleo Cytoplasmic Large DNA Viruses (NCLDVs) (Iyer et al. 2001, 2006 Filee et al. 2008). Some Mavericks also encode capsid-like proteins (Cl) (Kapitonov and Jurka 2006 Pritham et al. 2007). This conservation of virus genes suggests Mavericks might descend from an ancestral virus that adapted to an intragenomic lifestyle by acquisition of an integrase-like protein (Pritham et al. 2007) but it is not possible to exclude that unidentified pathogenic Maverick-like viruses still exist. Recently, Desjardins et al. (2008) described a Maverick in the flanking region of the Glyptapanteles flavicoxis bracovirus proviral locus that is not packaged in the particles. In the present study, we characterize a Maverick present within the CcBV packaged genome and compared it with related elements we found by data mining of insect genomes. Results and discussion A Maverick within CcBV packaged genome In order to characterize the CcBV sequences of the seg- ment 31 (CcBVS31), POLB, INT, coil-coiled, and ATPase proteins were used as queries in CENSOR searches against Repbase (Kohany et al. 2006). Results indicate they belong to a Maverick family, sharing 65.46, 59.24, 36.36 and 70.49% of similarities at amino-acid level respectively with the POLB, INT, coil-coiled and ATPase of the Tribolium castaneum Mavericks 2_Tc or 3_Tc (Table 1). We did not find any TIR(s) in the CcBV Maverick but this CcBV sequence has been subjected to secondary insertions and thus TIRs may have been degraded thereby preventing the identification of the complete element. The organisation of the CcBV Maverick fragment is similar to the corresponding region of the Tribolium cas- taneum Maverick except for the POLB, which is in reverse orientation (Fig. 1). The Maverick of T castaneum (Fig. 1) encodes also two coil-coiled proteins (Cc1 and Cc2), a capsid-like protein (Cl), and a protein sharing similarities with a matrix metalloproteinase from Heliothis zea nudi- virus-1 (HzNV-1) (Pritham et al. 2007). We identified also two additional genes, not previously described in Maver- icks, encoding putative proteins that we named P1 and P2. The genes P1, Cc2 and P2 are conserved in the G. flavic- oxis Maverick (Desjardins et al. 2008) (Fig. 1). By using T. castaneum and G. flavicoxis Maverick proteins as queries against CcBVS31 sequence in a tBLASTN analysis we found a sequence similar to the capsid-like protein (Cl) but we did not find sequences coding for the adenovirus cys- teine protease (PRO) nor P1 and P2 proteins, suggesting that secondary insertions of the DIRS and Dong retro- elements located in close vicinity to the Maverick have resulted in deletion of its downstream region. Accordingly we found a remnant of Cc2 in the DIRS region where a 75 aa peptide is encoded sharing 97% similarity with the 50 end of G. flavicoxis Cc2 (Fig. 1). Interestingly, some bracovirus segments from other braconid species are closely related to CcBVS31. By ana- lysing the closest homologous sequences, we found they do not contain the Maverick, thereby providing insight on the bracovirus CcBVS31 sequence before TE insertions. These sequences have been identified in the bracoviruses of Cotesia sesamiae (Bac 14G12: EF710642.1) and Cotesia vestalis (Segment S41: EF067329) (Choi et al. 2009b). Moreover, another CcBV packaged sequence corresponding to the 490 Genetica (2011) 139:489���496 123