Cotesia congregata Bracovirus Circles Encoding PTP and Ankyrin Genes Integrate into the DNA of Parasitized Manduca sexta Hemocytes

Abstract

Polydnaviruses (PDVs) are essential for the parasitism success of tens of thousands of species of parasitoid wasps. PDVs are present in wasp genomes as proviruses, which serve as the template for the production of double-stranded circular viral DNA carrying virulence genes that are injected into lepidopteran hosts. PDV circles do not contain genes coding for particle production, thereby impeding viral replication in caterpillar hosts during parasitism. Here, we investigated the fate of PDV circles of Cotesia congregata bracovirus during parasitism of the tobacco hornworm, Manduca sexta, by the wasp Cotesia congregata Sequences sharing similarities with host integration motifs (HIMs) of Microplitis demolitor bracovirus (MdBV) circles involved in integration into DNA could be identified in 12 CcBV circles, which encode PTP and VANK gene families involved in host immune disruption. A PCR approach performed on a subset of these circles indicated that they persisted in parasitized M. sexta hemocytes as linear forms, possibly integrated in host DNA. Furthermore, by using a primer extension capture method based on these HIMs and high-throughput sequencing, we could show that 8 out of 9 circles tested were integrated in M. sexta hemocyte genomic DNA and that integration had occurred specifically using the HIM, indicating that an HIM-mediated specific mechanism was involved in their integration. Investigation of BV circle insertion sites at the genome scale revealed that certain genomic regions appeared to be enriched in BV insertions, but no specific M. sexta target site could be identified.IMPORTANCE The identification of a specific and efficient integration mechanism shared by several bracovirus species opens the question of its role in braconid parasitoid wasp parasitism success. Indeed, results obtained here show massive integration of bracovirus DNA in somatic immune cells at each parasitism event of a caterpillar host. Given that bracoviruses do not replicate in infected cells, integration of viral sequences in host DNA might allow the production of PTP and VANK virulence proteins within newly dividing cells of caterpillar hosts that continue to develop during parasitism. Furthermore, this integration process could serve as a basis to understand how PDVs mediate the recently identified gene flux between parasitoid wasps and Lepidoptera and the frequency of these horizontal transfer events in nature.

Keywords: parasitoid wasps; polydnavirus; viral symbionts; virulence; virus integration.

FIG 1

Alignment of host integration motifs (HIMs) of 12 MdBV circles and of CcBV homologous sequences. Alignment of HIM sequences with similarity for each site colored in shades of blue for MdBV (A) and CcBV (B). All these sequences show boundary sequences forming indirect repeats of 9 nucleotides in MdBV and 8 nucleotides in CcBV. In MdBV HIMs, the sites ACtAGG, forming the J1 border of the insertion with the host genome, and CTaGT, forming the J2 border, are shown. For CcBV HIMs, similar conserved sites AccaGT and CTgGT were identified. (C and D) Model of integrase binding on HIM sites. HIM sites are composed of 29 (CcBV) or 30 (MdBV) nucleotide palindromic extremities separated by a stretch of 40 to 50 nucleotides (depending on the circle) that is lost during integration. This organization is compatible with a model in which a monomer of an integrase protein (here named Int) binds to each extremity. The two monomers then dimerize to form a synaptic complex, allowing J1 and J2 to come into contact. During circle integration into host DNA, one of the strand exchanges results in the loss of the loop. The bases conserved in MdBV or CcBV HIMs are shown in dark or light blue (bases conserved among 80% and 60% of the different HIMs, respectively). Nucleotides interacting directly with the integrase (IBS, integration binding site) are likely to be the ones that are the most conserved between circles. (C) Note that in MdBV HIMs most of the bases conserved in the J1 plus strand correspond to those conserved in the J2 minus strand, which is compatible with the binding of monomers from the same protein on both extremities. (D) Several positions are conserved between CcBV and MdBV HIMs, which suggests a common origin of the integration mechanism; however, CcBV extremities are more divergent from one another, which questions whether they still bind a single protein or two different monomers.

FIG 2

PCR-based detection method of CcBV circular or linear forms. Different forms of CcBV DNA are expected. CcBV circles containing one direct repeat junction (DRJ) and an HIM are produced in female wasps and encapsidated in viral particles that are injected into the host. (A) These circles originate from a proviral segment integrated in the wasp genome that is flanked by DRJ sequences. (B) Reintegrated forms of CcBV circles may also be present in wasp genomic DNA and in lepidopteran host DNA that are flanked by J1 and J2 sequences contained within HIMs. To detect the different CcBV forms, primers were designed to allow amplification of a region in a specific gene of the circle (set a), a region that included the DRJ motif (set b), and a region that spanned both the DRJ and the J1/J2 junctions (set c). Lack of amplification of a PCR product was expected with the third set of primers in the case of integration of the circles in M. sexta genomic DNA at the J1/J2 borders and also for male wasps, because there is no detectable production of viral circles in male wasps. Lack of amplification of a PCR product was expected with the second set of primers in male wasps because there is no detectable production of viral circles in male wasps unless a reintegrated form of the circle is present in the genome, as has been described for C. sesamiae strains. (C) Representative results obtained for the different primer couples allowing amplification of circles C10 and C26 are shown. Lack of amplification of primer couples (set c) in M. sexta organisms parasitized for 12 days suggest CcBV circles C10 and C26 are integrated in the M. sexta genome and are no longer detectable as circular forms. Note that amplification of primer couples (set c) could be obtained using a more processive polymerase, indicating that not all circles are integrated at this time point (data not shown).

FIG 3

Principle of the primer extension capture (PEC) method used to detect insertions of CcBV circles in the M. sexta genome. (A) DNA was extracted from hemocytes of one caterpillar 12 days after oviposition. This DNA potentially consists of circular CcBV DNA and M. sexta DNA with integrated CcBV DNA. (B) The DNA was used to construct an MiSeq library, which was amplified. Biotinylated primers corresponding to sequences near viral J1 and/or J2 junctions were hybridized to the DNA (C) and extended by amplification (D). (E) Biotinylated primer/target duplexes were captured with streptavidin-coated magnetic beads. (F) Captured targets were eluted, amplified with adaptor primer sites, and subjected to a second round of extension capture before MiSeq NGS sequencing. (G) Sequenced reads were mapped against the CcBV genomic circle, and aligned reads were then mapped to the M. sexta genome for IE identification.

FIG 4

Chimeric read coverage of CcBV HIM regions for each of the nine CcBV circles. J1 and J2 junction sites are indicated by red and blue boxes. The x axis corresponds to a 100-bp sequence spanning the HIM region with J1 and J2 sites. The y axis corresponds to read coverage for each base. To experimentally define J1 and J2 sequences, colored areas differentiate nucleotides above (pink) or below (gray) a chosen threshold that corresponds to one-fourth of the maximum read coverage for each circle at each junction J1 or J2. Indeed, in a chimeric read the first nucleotide after the J1 or J2 junction can be by chance (with a probability of 0.25) the same in the M. sexta DNA insertion site. These reads were therefore excluded to define J1 and J2 sites.

FIG 5

IE distribution over the M. sexta genome. The x axis corresponds to 4,195 windows of 100 kb of M. sexta genome. The y axis corresponds to the cumulative IE number counted for each of the nine circles inside each 100-kb window. Because of the fragmentation of the M. sexta genome assembly (20,891 scaffolds from 3.2 Mb to 500 bp), the 100-kb windows spanning several M. sexta genome scaffolds were highlighted in red, whereas windows fully present in one scaffold are in black. The number of 5 representative scaffolds is indicated below the x axis. Note that no specific integration sites could be detected in the M. sexta DNA sequence.

FIG 6

Proposed evolutionary scenario for the acquisition of HIMs in bracoviruses. Key dates (in millions of years) for the origin of braconid microgastroid subfamilies are from Murphy et al. (9). So far HIMs have only been identified in bracoviruses belonging to the Microgastrinae. No HIMs have been identified so far in Cheloninus inanitus (Cheloninae), and whether they are present in Toxoneuron nigriceps (Cardiochilinae) awaits investigation. A proposed evolutionary scenario is that acquisition of HIMs by BV occurred before the divergence of the Microgastrinae 54 Mya.