Identification of a viral gene essential for the genome replication of a domesticated endogenous virus in ichneumonid parasitoid wasps

1 1 Plos Pathogens 2 Identification of a viral gene essential for the genome replication of a domesticated endogenous 3 virus in ichneumonid parasitoid wasps. 4 Short title (70 characters): A viral gene essential for ichneumonid DEV local DNA amplification 5 6 Ange LORENZI 1,2¶, Fabrice LEGEAI 3,4¶, Véronique JOUAN 1, Pierre-Alain GIRARD 1, Michael R. 7 STRAND2, Marc RAVALLEC 1, Magali EYCHENNE 1, Anthony BRETAUDEAU 3,4, Stéphanie ROBIN 3,4, 8 Jeanne ROCHEFORT 1, Mathilde VILLEGAS 1, Denis TAGU 3, Gaelen R. BURKE 2, Rita REBOLLO 5, 9 Nicolas NÈGRE1*, Anne-Nathalie VOLKOFF1*. 10 11 1 DGIMI, Montpellier University, INRAE, Montpellier, France 12 2 Department of Entomology, University of Georgia, Athens, Georgia, 30602, United States 13 3 INRAE, UMR Institut de Génétique, Environnement et Protection des Plantes (IGEPP), BioInformatics 14 Platform for Agroecosystems Arthropods (BIPAA), Campus Beaulieu, 35042 Rennes, France 15 4 INRIA, IRISA, GenOuest Core Facility, Campus de Beaulieu, Rennes 35042, France 16 5 Univ Lyon, INRAE, INSA Lyon, BF2I, UMR 203, 69621 Villeurbanne, France 17 18 * Corresponding authors: 19 Anne-Nathalie VOLKOFF, [email protected] 20 Nicolas NÈGRE, [email protected] 21 22 ¶ These authors contributed equally to this work. 23 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 2 24 Abstract (300 words) 25 Thousands of endoparasitoid wasp species in the families Braconidae and Ichneumonidae harbor 26 "domesticated endogenous viruses" (DEVs) in their genomes. This study focuses on ichneumonid 27 DEVs, named ichnoviruses (IVs), which derive from an unknown virus and produce virions in ovary calyx 28 cells during the pupal and adult stages of female wasps. Females inject IV virions into host insects when 29 laying eggs. Virions infect cells which express IV genes with functions required for wasp progeny 30 development. IVs have a dispersed genome consisting of two genetic components: proviral segment 31 loci that serve as templates for circular dsDNAs that are packaged into capsids, and genes from an 32 ancestral virus controlling virion production. Because of the lack of homology with known viral genes, 33 the molecular control mechanisms of IV genome are largely uncharacterized. We generated a 34 chromosome-scale genome assembly for Hyposoter didymator and identified a total of 67 H. didymator 35 ichnovirus (HdIV) loci distributed across the 12 wasp chromosomes. By analyzing genomic DNA levels, 36 we found that all HdIV loci were locally amplified in calyx cells during the wasp pupal stage, suggesting 37 the implication of viral proteins in DNA replication. We tested a candidate HdIV gene, U16, encoding a 38 protein with a conserved domain found in primases and which is transcribed in calyx cells during the 39 initial stages of replication. Knockdown of U16 by RNA interference inhibited amplification of all HdIV 40 loci, as well as HdIV gene transcription, circular molecule production and virion morphogenesis in calyx 41 cells. Altogether, our results showed that viral DNA amplification is an early step of IV replication 42 essential for virions production, and demonstrated the implication of the viral gene U16 in this process. 43 44 Author Summary (150-200 words) 45 Parasitoid "domesticated endogenous viruses" (DEVs) provide a fascinating example of eukaryotes 46 acquiring new functions through integration of a virus genome. DEVs consist of multiple loci in the 47 genomes of wasps. Upon activation, these elements collectively orchestrate the production of virions or 48 virus-like particles that are crucial for successful parasitism of host insects. Despite the significance of 49 DEVs for parasitoid biology, the mechanisms regulating key steps in virion morphogenesis are largely 50 unknown. In this study, we focused on the ichneumonid parasitoid Hyposoter didymator, which harbors 51 an ichnovirus consisting of 67 proviral loci. Our findings reveal that all proviral loci are simultaneously 52 amplified in ovary calyx cells of female wasps during the early pupal stage suggesting a hijacking of .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 3 53 cellular replication complexes by viral proteins. We tested the implication of such a candidate, U16, 54 encoding a protein with a weakly conserved primase C-terminal domain. Silencing U16 resulted in 55 inhibited viral DNA amplification and virion production, underscoring the key role of this gene for 56 ichnovirus replication. This study provides evidence that genes involved in viral DNA replication have 57 been conserved during the domestication of viruses in the genomes of ichneumonid wasps. 58 59 Introduction 60 Endogenous viral elements (EVEs) refer to viral sequences in eukaryotic genomes that originate from 61 complete or partial integration of a viral genome into the germline [1]. While retroviruses are the best- 62 known sources of EVEs, bioinformatic studies have also identified non-retroviral EVEs across a diverse 63 range of organisms [2]. Although many EVEs become non-functional and decay through neutral 64 evolution [3], some have been preserved and repurposed by their hosts for new functions, often as short 65 regulatory sequences or individual genes [4,5]. A notable exception to this pattern is observed in 66 domesticated endogenous viruses (DEVs) that have been identified in four lineages of endoparasitoid 67 wasps - insects that lay eggs and develop within the bodies of other insects [6]. Parasitoid DEVs consist 68 of numerous genes conserved within the wasp genome that originate from the integration of complete 69 viral genomes. Unlike other EVEs, these genes remain functional and actively interact to produce virus 70 particles in calyx cells, which are located in the apical part of the oviducts of female wasps [7]. Viral 71 particles are produced in the pupal and adult stages, and accumulate in the oviducts of the wasp. Adult 72 female wasps inject these particles along with eggs into insect hosts where they have essential functions 73 in the successful development of wasp offspring [8]. 74 Parasitoid DEVs are prevalent among species in two wasp families named the Braconidae and 75 Ichneumonidae. The DEVs identified in these families have evolved from different virus ancestors but 76 through convergence have been similarly repurposed to produce either virions containing circular 77 double-stranded (ds) DNAs or virus-like particles (VLPs) lacking nucleic acid. The hyperdiverse 78 Microgastroid complex in the family Braconidae harbors DEVs named bracoviruses (BVs). BVs evolved 79 from a virus ancestor in the family Nudiviridae [9]. Wasps harboring BVs produce virions containing 80 circular dsDNAs. Other braconids in the subfamily Opiinae and ichneumonids in the subfamily 81 Campopleginae independently acquired two other distinct nudiviruses that wasps have coopted to 82 produce VLPs [10, 11]. The fourth identified DEV lineage, named ichnoviruses (IVs), is present in two .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 4 83 ichneumonid subfamilies (Campopleginae and Banchinae) which produce virions containing circular 84 dsDNAs. Unlike the other three DEVs, IVs likely originated from a Nucleocytoplasmic Large DNA Virus 85 (NCLDV) but the precise ancestor remains unknown [12, 13]. 86 BVs have been more studied than IVs but the latter are intriguing because of their uncertain origins. 87 Despite differences in ancestry and gene content, BV and IV genomes are similarly organized into two 88 components that have distinct functions [14]. Insights into the genome components of IVs primarily 89 derive from sequencing two campoplegine wasps named Hyposoter didymator and Campoletis 90 sonorensis [15], along with calyx transcriptome studies [12, 13, 16, 17] and proteomic analyses of 91 purified virions [12, 13]. The first genome component of IVs are domains in the wasp genome that show 92 evidence of deriving from the virus ancestor and having essential functions in virion formation. These 93 domains, named "Ichnovirus Structural Protein Encoding Regions" (IVSPERs), contain intronless genes 94 that are specifically transcribed in calyx cells [12, 13, 17]. Most IVSPER genes are transcribed at the 95 onset of pupation in hyaline stage 1 pupae [16], and some genes in IVSPERs encode proteins 96 associated with IV virions [12, 13]. Six genes have been knocked down by RNA interference (RNAi) in 97 H. didymator which demonstrated that they have functions in virion assembly or cell trafficking [16]. Five 98 IVSPERs have been identified in the H. didymator and C. sonorensis genomes [15], while three have 99 been identified in the genome of the more distantly related banchine G. fumiferanae [13]. The content 100 of IVSPER genes is notably similar between ichneumonid wasp species [12, 13, 17], and their gene 101 order is well-conserved among campoplegine species [15]. Additionally, one intronless gene (U37) was 102 identified in the H. didymator and C. sonorensis genomes outside of any IVSPER with features 103 suggesting it also derives from the virus ancestor [15]. Together, these genes, whether found within or 104 outside IVSPERs, represent the fingerprints of the ancestral viral machinery essential for virion 105 production and are designated as IV core replication genes. Notably, none of these genes are packaged 106 in virions, indicating that IV core genes can only be transmitted vertically through the germline of 107 associated parasitoids. 108 The second component of IV genomes are domains referred to as "proviral segments," which are 109 amplified in calyx cells and produce the circular dsDNAs that are packaged into capsids [18, 19]. The 110 number of proviral segments, typically exceeding 50, are widely dispersed in wasp genomes and exhibit 111 considerable variability between wasp species, [15]. Each proviral segment is characterized by flanking 112 direct repeats (DRs) of variable length (1 kb) and homology that identify where homologous .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 5 113 recombination processes occur to produce circularized DNAs [18, 19]. Some IV proviral segments also 114 contain internal repeats that facilitate additional homologous recombination events, and produce 115 multiple overlapping or nested circularized DNAs per proviral segment [15, 18]. Proviral segments 116 encode genes with and without introns that are predominantly expressed in the hosts of wasps after 117 virion infection [20, 21, 22, 23]. While IV core replication genes represent the conserved viral machinery 118 that produces virions in calyx cells, proviral segments constitute the IV genome components that virions 119 transfer to the hosts wasps parasitize. These segments also play a major role in the virulence of IVs, 120 which contributes to the successful development of parasitoid progeny. 121 The replication of IVs, encompassing the processes leading to the production of virions containing IV 122 segments, occurs within the nuclei of calyx cells during pupal and adult developmental stages [7, 24]. 123 Electron microscopy studies of H. didymator ichonovirus (HdIV) shows that fusiform-shaped capsids are 124 individually enveloped in the nuclei of calyx cells during the late pupal stage (pigmented pupae, stage 125 3) [16]. These enveloped "subvirions" exit the nucleus, traverse the cytoplasm, and exit calyx cells by 126 budding, resulting in mature virions with two envelopes that accumulate in the calyx lumen of the ovaries 127 [7, 24]. Earlier findings indicated that IVSPERs and proviral segments undergo amplification in newly 128 emerged adult wasps [12]. However, these data focused on only a subset of IVSPER genes and one 129 proviral segment, leaving our knowledge of whether all IV genome components are amplified in calyx 130 cells incomplete. Similarly, the initiation time of amplification during pupal development and IV virion 131 production remains unknown. The specific role of IV core genes in virion production is also poorly 132 documented when compared to BVs [25, 26]. The limited sequence homology of IVSPER genes with 133 genes in other viruses provides minimal insights into potential functions. To date, only the six genes 134 mentioned above that are involved in subvirion assembly or cell trafficking have been studied [16]. 135 In this work, we explored IV replication using the campoplegine wasp H. didymator. We first generated 136 a chromosome-level assembly for the H. didymator genome. Through this assembly, we determined 137 that all genome components undergo local amplification in calyx cells which initiates between pupal 138 stages 1 and 2. Notably, IVSPERs, isolated IV core genes, and proviral segments were amplified in 139 large regions with non-discrete boundaries. Next, we studied the function of U16 which is located on H. 140 didymator IVSPER-3. U16 is one of the most transcribed IVSPER genes during the initial pupal stage 141 and contains a weakly conserved domain found in the C-terminus of primases. RNAi knockdown of U16 142 inhibited virion formation. Knockdown also significantly reduced DNA amplification of all HdIV genome .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 6 143 components, which decreased transcript abundance of IV core genes and the abundance of circular 144 dsDNA viral molecules. We conclude U16 is an essential gene for amplification of the HdIV genome and 145 virion production, demonstrating that genes from the IV ancestor regulating IV replication have been 146 conserved during virus domestication. Additionally, our results show that viral DNA amplification is 147 essential for IV virion production. 148 Results 149 Genomic localization of Hyposoter didymator IV components in a novel chromosome-level 150 assembly. 151 The genome assembly for H. didymator we previously generated [15] consisted of 2,591 scaffolds with 152 an N50 of 4 Mbp. We concluded this assembly was overly fragmented to evaluate DNA amplification in 153 calyx cells during virion morphogenesis. We therefore used proximity ligation technology to produce a 154 new chromosome level assembly consisting of twelve large scaffolds that corresponds with the haploid 155 karyotype for H. didymator [27]. The sizes of these scaffolds ranged from 6.7 Mbp to 29.3 Mbp (S1 156 Dataset A, B). 157 The five IVSPERs (IVSPER-1 to IVSPER-5), the predicted IV core gene (U37) located outside of an 158 IVSPER, and 53 of the 54 previously identified proviral segment loci (Hd1 to Hd54) [15] were identified 159 in the new assembly. The new assembly did not include the scaffold containing Hd51, possibly due to 160 low-quality sequencing data (S1 Dataset, B). Our chromosome-level assembly revealed that each 161 scaffold contained at least one HdIV locus, but notably, all IVSPERs and 40% of the proviral segment 162 loci resided on two (scaffold 7 and 11) (S1 Dataset, B). 163 While three IVSPERs and the majority of proviral segments were distantly located from each other in 164 the H. didymator genome, there were exceptions to this pattern including certain pairs of proviral 165 segments separated by less than 20 kb (e.g., Hd36 and Hd38; Hd46 and Hd43; Hd44.1 and Hd44.2; 166 Hd12 and Hd16). In all of these cases, the paired segments exhibited significant homology which 167 suggested they derive from recent duplication events (S1 Dataset, C). Additionally, several proviral 168 segments were in proximity to IVSPERs or IV replication genes that resided outside of IVSPERs (e.g., 169 Hd46 near U37; Hd29 and Hd24 on each side of IVSPER-2; Hd15 near IVSPER-1; also see below). 170 Amplification of Hyposoter didymator IV genome components in calyx cells during wasp pupal 171 development. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 7 172 To investigate whether all or only specific components of the HdIV genome undergo amplification in 173 association with virion morphogenesis, we isolated DNA from calyx cells from stage 1 pupae (one day 174 old, hyaline) and stage 3 pupae (five days old, pigmented abdomen). We then generated paired-end 175 libraries, which were sequenced using the Illumina platform, followed by read alignment to the new 176 chromosome-level genome assembly. When analyzing the reads from stage 1 pupae, read coverage 177 per HdIV locus did not differ significantly from the coverage of randomly selected regions of the same 178 size from the rest of the wasp genome (Fig 1A). In contrast, read coverage for stage 3 pupae was higher 179 for all HdIV loci when compared to the rest of the wasp genome or to values obtained for pupal stage 1 180 (Fig 1A, S1 Table). 181 To more precisely investigate the temporal dynamics of amplification, we conducted relative quantitative 182 (q) PCR assays that measured copy number of genes in IVSPER-1, -2, and -3 in calyx DNA samples 183 that were collected from stage 1-4 pupae. We compared these treatments to DNA samples from hind 184 legs of stage 1 pupae where no HdIV replication occurs. We also included a wasp gene (XRCC1) located 185 in close proximity to IVSPER-1. Results showed that copy number of each tested gene was similar in 186 calyx and hind legs in stage 1 pupae, indicating none were amplified during the initial pupal stage. 187 Subsequently, the copy number of each gene increased progressively with each pupal stage (Fig 1B). 188 While exhibiting lower amplification levels than the IVSPER genes we analyzed, a similar trend was 189 observed for the wasp gene XRCC1 (Fig. 1B). These findings indicated IVSPER amplification in calyx 190 cells begins between pupal stage 1 and stage 2, which further increased in pupal stage 3 and 4. 191 Fig 1. DNA amplification of HdIV loci. (A) Coverage of HdIV loci compared to the rest of the wasp 192 genome. Read coverage values per analyzed region (see Materials and Methods) are presented for 193 each locus type (proviral segments and IVSPERs) at pupal stage 1 (hyaline pupa) and pupal stage 3 194 (pigmented pupa). The coverages per HdIV locus are compared to the coverage per random genome 195 regions outside of HdIV loci (wasp). Note that the coverage value for random wasp regions is lower for 196 DNA samples collected from stage 3 versus stage 1 pupae. This difference is attributed to the higher 197 proportion of reads mapping to HdIV regions among the total number of reads in stage 3 compared to 198 stage 1. The significance levels are indicated as follows: ns = non-significant, **p<0.01, and ***p<0.001. 199 (B) qPCR analysis of select IVSPER genes in calyx cells during wasp pupal development. Top panel. 200 A schematic representation of H. didymator IVSPERs-1, -2, and -3 (GenBank GQ923581.1, 201 GQ923582.1, and GQ923583.1); genes selected for qPCR assays are highlighted in white. U1-24 are .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 8 202 unknown protein-encoding genes, while IVSPs are members of a gene family encoding ichnovirus 203 structural proteins. Bottom panel. Genomic (g) DNA amplification levels of IVSPER genes and wasp 204 XRCC1 in calyx cells from pupal stage 1-4. The XRCC1 (X-Ray Repair Cross Complementing 1) 205 encoding gene is located 1,200 bp from U1 (position 3,270,470 to 3,272,519 in Scaffold-11). Data 206 corresponds to gDNA amplification relative to amplification of the housekeeping gene elongation factor 207 1 (ELF1). The Y-axis was transformed using the square root function for better data visualization. 208 Differential levels of amplification across all components of the HdIV genome 209 The qPCR results presented in Fig 1 indicated amplification levels varied, with genes in IVSPER-3 210 exhibiting higher levels of amplification than genes in IVSPER-1 and -2 (Fig 1B). This variability was 211 corroborated genome-wide by analyzing read coverage per position and the ratio between stage 3 and 212 stage 1 (Fig 2, S1 Fig). Amplification levels of IVSPER loci, determined at the summit of the coverage 213 curve, ranged from 10X for IVSPER-5 in Scaffold-7 to over 200X for IVSPER-3 in Scaffold-3 (S1 Table). 214 This observation aligned with the findings from qPCR analyses, indicating that genes in IVSPER-3 were 215 more highly amplified than those in IVSPER-1 and -2 (Fig 1B). Read mapping further indicated that the 216 peak of amplification occurs toward the middle of each IVSPER (Fig 1B, S1 Fig), consistent with qPCR 217 analyses revealing that within each IVSPER, genes closer to the cluster boundary tended to exhibit 218 lower levels of amplification compared to genes situated in the middle of the cluster (Fig 1B). 219 Fig 2. HdIV DNA amplification. DNA amplification in pupal stage 3 was assessed by mapping genomic 220 DNA Illumina reads against the 12 large H. didymator genome scaffolds. In each scaffold, red bars 221 indicate amplified loci, with the intensity of red corresponding to increased values of the CPM ratio 222 between pupal stage 3 and pupal stage 1. The positions of IVSPERs and isolated IV replication genes 223 are indicated by purple squares, while proviral segments are indicated by green circles. For selected 224 HdIV loci, amplification curves (representing the ratio of the CPM values calculated for 10 bp intervals 225 between pupal stage 3 and pupal stage 1) are shown in boxes. Amplification curves for all of the 226 annotated HdIV loci are shown in S1 Fig. Each HdIV locus is indicated in red while 10,000 bp of flanking 227 sequence on each side of the locus is also shown. For proviral segments, loci are defined as the 228 sequence delimited by two direct repeats; IVSPERs are defined as the region between the start and 229 stop codon of the first and last coding sequences in the cluster; isolated IV replication genes are defined 230 by their coding sequence. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 9 231 Proviral segment loci were relatively more amplified than IV replication gene loci, and also variable in 232 intensity (Fig 2, S1 Fig). For example, coverage ratio between stages 3 and 1 ranged from 30X for 233 proviral locus Hd40 in Scaffold-6 to over 1,100X for Hd27 in Scaffold-7 (S1 Table) at the summit of the 234 coverage curves. Variability in the number of reads mapping to a given proviral locus was consistent 235 with earlier studies indicating that the circularized DNAs packaged into IV capsids are non-equimolar in 236 abundance [8, 28]. 237 All proviral segments consistently exhibited a substantial increase in amplification that peaked between 238 the two DRs (as exemplified by Hd14 or Hd12 in S2 Fig). For numerous proviral loci, the reads mapping 239 between the flanking DRs displayed uniform coverage. However, in other cases, peaks with varying 240 read coverage were evident (as exemplified by Hd32 or Hd16 in S2 Fig). This differential coverage 241 usually applied to proviral segments containing more than one pair of DRs, as illustrated by proviral 242 locus Hd11 (Fig 3A) or Hd32 and Hd16 (S2 Fig). Previous studies indicated Hd11 contains two pairs of 243 DRs, enabling the formation of two nested, circularized segments termed Hd11-1 (formed by 244 recombination between DR1Left (DR1L) and DR1Right (DR1R)) and Hd11-2 (formed by recombination 245 between DR2L and DR2R) (Fig 3A). Reads mapping to the Hd11 locus (bounded by DR1L and DR2R) 246 exhibited three relatively uniform plateaus of different values. Two plateaus corresponded to reads 247 mapping to the predicted locations of Hd11-1 (235X) and Hd11-2 (111X), while the central region with 248 higher coverage (311X) corresponded to reads mapping to both nested segments (Fig 3A). This 249 differential coverage would not be expected if reads mapped only to Hd11 chromosomal DNA. 250 Consequently, the pattern of proviral segment coverage suggested part of the coverage values were 251 due to reads mapping to amplification intermediates and/or circularized dsDNAs that were also present 252 in our DNA samples. Some amplified HdIV loci contain both an IVSPER and proviral segments. Two of 253 these loci resided on Scaffold-11 (Hd29, IVSPER-2, Hd24, and Hd33, Hd15, IVSPER-1 (Fig 3B)). For 254 these loci, the amplification curves spanned the length of the amplified region (yellow dotted line in Fig 255 3B) but were interrupted by peaks corresponding to the length of proviral segments. This pattern 256 suggested amplification levels of the chromosomal form of the proviral segments could correspond to 257 the IVSPER amplification curves, but were higher because reads additionally mapped to circular 258 dsDNAs or amplification intermediates. 259 Fig 3. HdIV amplified regions in Scaffold-11. (A) Detail of the amplified region at the Hd11 locus. (B) 260 Detail of two other amplified regions containing IVSPERs and HdIV proviral loci. In (A) and (B), .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 10 261 amplification curves represent the ratio of the CPM values (calculated for 10 bp intervals) obtained in 262 pupal stage 3 compared to pupal stage 1. For each locus, amplification values at the summit of the 263 peaks and at the start and end positions of HdIV segments are indicated. In (B), amplification curves 264 of IVSPERs are highlighted in yellow. Each amplification curve figure was generated by Integrated 265 Genome Viewer (IGV) [29]. 266 Amplification of H. didymator IV genome components in extensive wasp genome domains with 267 undefined boundaries 268 Since our read coverage data indicated amplified regions were larger than the annotated HdIV loci (Fig 269 2, S1 Fig), we used the MACS2 peak calling program, originally developed for chromatin 270 immunoprecipitation sequencing experiments, to identify areas in the H. didymator genome that were 271 enriched for reads when compared to a control [30]. Amplification peaks were called with MACS2 using 272 alignments from stage 3 pupae as the treatment and alignments from stage 1 pupae as the control. 273 MACS2 identified all HdIV genome components that we had annotated in our earlier study [15] plus 274 several previously unrecognized domains (S2 Table). Manual curation (see Materials and Methods 275 section) indicated three of these new domains were proviral segment loci that we named Hd52, Hd53, 276 and Hd54. Five others were intronless genes, suggesting origins from the IV ancestor, that were outside 277 of IVSPERs. We thus named these genes U38, U39, U40, U41, and U42. The remaining domains 278 detected by MACS2 either contained predicted wasp genes or lacked any features that identified them 279 as IV replication genes or proviral segments. Altogether, the MACS2 algorithm predicted a total of 55 280 domains in the H. didymator genome containing HdIV loci. Two proviral segments (Hd45.1 on Scaffold-4 281 and Hd2-like on Scaffold-7) escaped MACS2 detection, possibly because they were located too close 282 to the ends of each scaffold. However, our read mapping data clearly indicated these two segments are 283 amplified in stage 3 (Table 1) with a profile similar to the other segments (S1 Fig). In total, our read 284 mapping and MACS2 data indicated the H. didymator genome contains 67 HdIV loci (56 proviral 285 segments, five IVSPERs, and six predicted IV replication genes that reside outside of IVSPERs) that 286 are amplified in calyx cells at pupal stage 3 (Fig 2, Table 1). 287 Table 1. All HdIV loci amplified in calyx cells from stage 3 pupae identified by read mapping 288 and/or the MACS2 algorithm. For each scaffold, the position and size of the HdIV loci are indicated. 289 Loci newly identified in the present work are marked with asterisks. Corresponding amplified regions 290 (i.e., the peak predicted by the MACS2 algorithm) are provided for each locus or groups of loci. Start .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 11 291 and end positions delimiting the HdIV loci and the amplified regions detected by MACS2 are indicated. 292 The distance between the start or the end of the amplified region and the locus is presented. For each 293 HdIV locus and amplified region detected by MACS2, coverage values are provided for calyx cell 294 samples collected from stage 1 or stage 3 pupae. Coverage is based on the length of the HdIV locus or 295 amplified region. ND indicates amplified regions not detected by MACS2. 296 Our overall results also indicated all amplified regions in the H. didymator genome containing HdIV loci 297 consist of the annotated HdIV locus plus flanking wasp sequence consistent with our detailed analysis 298 of the wasp gene XRCC1 that is located in close proximity to IVSPER-1 (Fig 1B). Across all HdIV loci, 299 we determined that the flanking regions containing wasp sequence that were amplified varied from 7,000 300 to 15,000 bp (Table 1). The total size of the amplified regions ranged from 10,692 bp (Hd28 on Scaffold- 301 12) to 54,005 bp (IVSPER-2 on Scaffold-11). Most amplified regions contained a single HdIV locus, but 302 seven contained a mix of HdIV genome components (Table 1). Three amplified regions contained the 303 neighboring and closely related proviral segments mentioned above (e.g., Hd36 and Hd38 on Scaffold-1, 304 Hd44.1 and Hd44.2 on Scaffold-2, Hd12 and Hd16 on Scaffold-11). In addition to the two examples 305 noted above on Scaffold 11 (see Fig. 3B), two other amplified loci also contained both IVSPERs and 306 proviral segments (U37, Hd46, and Hd43 on Scaffold-2; U40 and Hd39 on Scaffold-9). Lastly, we 307 searched for sequence signatures that potentially identify the amplification boundaries for each HdIV 308 locus. However, our analysis identified only low-complexity A-tract sequences, which were not specific 309 to HdIV components as they were also found in random wasp genomic sequences (S3 Fig). Thus, no 310 motifs were identified that distinguished the amplification boundaries of HdIV loci. 311 RNAi knockdown of U16 inhibits virion morphogenesis. 312 We selected the gene U16 located on H. didymator IVSPER-3 as a factor with potential functions in 313 activating IV replication. U16 is conserved among all IV-producing wasps for which genome or 314 transcriptome data is available (Fig 4A). In H. didymator calyx cells, U16 is also one of the most 315 transcribed IV genes detected in calyx cells from stage 1 pupae [16]. Sequence analysis using the basic 316 local alignment search tool and DeepLoc2.0 predicted all U16 family members contain a C-terminal 317 alpha-helical domain (PriCT-2) of unknown function that is present in several primases [31] (Iyer et al., 318 2005) and a nuclear localization signal (Fig 4A, S2 Dataset). We next assessed the effects of knocking 319 down U16 by RNAi on virion morphogenesis in calyx cells. We injected newly pupated wasps with 320 dsRNAs that specifically targeted U16 using previously established methods [16]. RT-qPCR analysis .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 12 321 indicated transcript abundance in the calyx of newly emerged adult females was reduced more than 322 90% when compared to control wasps that were injected with dsGFP (Fig 4B). Inspection of the ovaries 323 further indicated that the calyx lumen of control wasps contained blue 'calyx fluid' indicative of HdIV 324 virions being present, whereas almost no calyx fluid was seen in dsU16-injected wasps (Fig 4B). 325 Examination of calyx cell nuclei by transmission electron microscopy similarly showed that calyx cells in 326 one day old control females contained an abundance of subvirions, whereas no subvirions were 327 observed in treatment wasps (Fig 4C). We thus concluded that U16 is required for virion morphogenesis. 328 Fig 4. RNAi knockdown of U16. (A) U16 proteins identified in the campoplegine Hyposoter didymator 329 [12], Campoletis sonorensis [15], and Bathyplectes anurus [17], and in two banchine wasps Glypta 330 fumiferanae [13] and Lissonota sp. [32]. For each, protein size, percentage of identity with H. didymator 331 protein and location of the PRiCT_2 domain are indicated. (B) RT-qPCR data showing relative 332 expression of U16 in dsGFP (control) and ds U16 injected females. ** p<0.01. Images of ovaries 333 dissected from newly emerged adult females that were injected with dsGFP (left) or dsU16 (right). Note 334 the blue color in the oviduct of the dsGFP control indicating the presence of HdIV virions. (C) Schematics 335 and electron micrographs showing that (a) calyx cell nuclei (N) from females treated with dsGFP-injected 336 contain subvirions (V) while (b) calyx cell from a dsU16 -injected wasps do not. This results in no 337 accumulation of virions in the calyx lumen as illustrated in the schematic images. CL, calyx lumen; Cyt, 338 cytoplasm. Scale bars = 5 μm, zooms = 1 μm. 339 RNAi knockdown of U16 also disables amplification of HdIV loci 340 Since U16 contained a domain found in primases, we investigated whether RNAi knockdown also 341 disabled amplification of HdIV genome components. We injected newly pupated wasps with dsU16 or 342 dsGFP, followed by isolation and deep sequencing of calyx cell DNA from stage 3 pupae in three 343 independent replicates. Mapping the reads from dsGFP-treated calyx samples to the H. didymator 344 genome indicated all HdIV loci were amplified as evidenced by higher coverage values when compared 345 to random regions of the wasp genome (Fig 5A). Conversely, coverage values did not differ between 346 HdIV loci and other regions of the wasp genome in dsU16-treated calyx samples (Fig 5A). When 347 analyzing coverage per each HdIV genome component (IVSPERs, isolated IV replication genes, or HdIV .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 13 348 proviral segments), we also determined that values were systematically lower for the dsU16 than 349 dsGFP-treatments (Fig 5B and 5C, S3 Table). 350 Fig 5. Impact of U16 RNAi knockdown on DNA proviral amplification. (A) Comparative distribution 351 of read coverages in ds GFP- and dsU16-injected females. For each of the three replicates, coverage 352 values are given per HdIV loci (V) and per random genome regions outside of the HdIV loci (W), both 353 with the same size distribution. IVSPERs and IV replication genes loci are shown in the left panel, while 354 proviral segment loci are shown in the right panel. (B) Coverage values per IVSPERs, and per IV 355 replication genes residing outside an IVSPER, in the three biological replicates of both dsU16- and 356 dsGFP-injected samples. Names of HdIV loci are indicated as well as the scaffold (Scaf-) they are 357 located in. (C) Coverage values for proviral segment loci in the three biological replicates of the dsU16 358 and dsGFP samples. For better visualization, only the scaffold (Scaf-) in which the proviral segments 359 are located is indicated. The list of the proviral segment loci within each scaffold is available in Table 1. 360 The y-axis was transformed by the log function for better data visualization. Statistical analyses are 361 available at https://github.com/flegeai/EVE_amplification. 362 We extended our analysis by injecting dsGFP or dsU16 into newly formed pupae, followed by isolation 363 of DNA from calyx cells and hind legs, where no HdIV replication occurs. We then used specific primers 364 and qPCR assays that measured DNA abundance of three wasp genes, selected HdIV replication genes 365 inside and outside of IVSPERs, and selected HdIV genes in different proviral segments. As anticipated, 366 no genes were amplified in hind legs from either control or treatment wasps (Fig 6). In dsGFP-injected 367 control wasps, all HdIV genes were amplified in calyx cell samples (Fig 6). Among the wasp genes, only 368 XRCC1 exhibited significant amplification, consistent with its location within the IVSPER-1 amplified 369 region (Fig 6). In contrast, when examining calyx cell DNA from wasps injected with dsU16, none of the 370 HdIV genes nor XRCC1 were amplified (Fig 6). Altogether, our results indicated U16 is required for 371 amplification of all HdIV loci. 372 Fig 6. Impact of U16 RNAi knockdown on amplification of select wasp and HdIV genes. Relative 373 genomic amplification of selected HdIV genes in two-day-old females injected with dsGFP or dsU16. 374 The wasp gene XRCC1, located within the amplified region of the IVSPER-1 locus, was incorporated 375 into the analysis. Wasp histone (H1) and ribosomal protein (rpl) genes served as controls. Samples 376 were obtained from calyx cells (where virion are produced) and hind legs (control). Statistical .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 14 377 significance levels are denoted as follows: ns = non-significant, *p<0.05, **p<0.01, and ***p<0.001. The 378 y-axis values were transformed using the square root function for better data visualization. 379 Impact of DNA amplification on IV replication gene transcription levels and abundance of 380 circularized HdIV molecules in calyx cells. 381 We hypothesized that amplification of IV replication genes would increase transcript abundance which 382 in turn would be affected by inhibiting HdIV DNA amplification. We thus compared transcript abundance 383 of various genes in IVSPER-1, -2, and -3, in calyx RNA samples that were collected from wasps treated 384 with dsU16 or dsGFP. U16 knockdown reduced expression of every HdIV replication gene we examined 385 (Fig 7A). Finally, we investigated the impact of U16 knockdown on the abundance of the circularized 386 dsDNAs that are processed from amplified proviral segments. For this assay, we used PCR primers that 387 specifically amplified the proviral form, circularized (episomal) form or both forms of Hd29 (Fig 7B). 388 Results showed a significant reduction in both the proviral and circularized forms of Hd29 in calyx cell 389 DNA from wasps injected with dsU16 when compared to DNA from wasps injected with dsGFP (Fig. 390 7B). Our results thus indicated U16 is required for proviral segment amplification which is also required 391 for production of circularized segments. 392 Fig 7. Impact of U16 RNAi knockdown on HdIV replication gene expression and proviral segment 393 amplification. (A) Relative expression of nine IVSPER genes in 2-day-old adult females injected with 394 dsGFP (control) or dsU16. (B) Relative DNA amplification of the integrated linear (proviral) and 395 circularized (episomal) forms of viral segment Hd29 in 2-day-old adult females injected with dsGFP 396 (control) or dsU16. The left panel illustrates the position of primer pairs designed to selectively amplify 397 the proviral form (Proviral Left and Right, indicated by red and black arrows), the circularized form 398 (Episomal, red arrows), or both (Proviral + Episomal, brown arrows). The right panel presents the relative 399 amplification of each form using DNA from dsGFP- and dsU16-injected females. In both (A) and (B), 400 significance levels are indicated as follows: ns = non-significant, *p<0.005, **p<0.01, and ***p<0.001. 401 The y-axis values were transformed using the square root function for better data visualization. 402 Discussion 403 During parasitism, wasps associated with IVs, BVs and other DEVs simultaneously inject virus-derived 404 particles and eggs into their host. The role of DEV-derived particles in the success of wasp parasitism 405 is well documented in the literature [22, 33, 34]. BVs, which evolved from a nudivirus, share a set of .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 15 406 genes homologous to nudivirus and baculovirus core genes. Functional studies, guided in part by these 407 similarities, have provided insights into several key processes underlying BV virion production. 408 Identification of BV core genes that regulate the expression of other BV core genes encoding structural 409 proteins [25], are involved in BV virion formation [25, 26, 35], or are required for processing proviral 410 segments into circular DNA molecules packaged into capsids [25, 26] have been documented. In 411 contrast, identifying the components of IV genomes and functions of IV genes regulating replication is 412 more challenging because the hypothesized NCLDV ancestor is unknown. In turn, IV genome 413 components with known or hypothesized functions in replication share little or no homology with known 414 viruses. This study significantly advances understanding of IV replication by generating a chromosome 415 level assembly for the H. didymator genome, presenting several lines of evidence showing that all HdIV 416 loci are amplified in calyx cells when virions are being produced, and identifying U16 as an essential 417 gene for amplification of all HdIV loci and virion formation. This study also highlights the critical role of 418 viral DNA amplification for IV virion production. 419 Earlier studies suggested IV proviral segment loci undergo amplification before viral segment processing 420 [18, 19]. Another study indicated amplification of a few IVSPER genes and one proviral segment located 421 in close vicinity of an IVSPER in one-day-old H. didymator adults [12]. However, the question persisted 422 regarding whether all IV genome components were amplified in calyx cells and when amplification 423 initiates during the time-course of virion production. To address these questions, we used our new 424 chromosome-level genome assembly to map domains that undergo amplification in calyx cells during 425 virion morphogenesis. Read mappings to genomic DNA extracted from H. didymator pupal stages 1 and 426 3 revealed that all HdIV genome components are simultaneously and locally amplified in calyx cells in 427 stage 3 pupae. This analysis further identified five proviral segments and five IV replication genes 428 located outside of IVSPERs that were previously unknown, resulting in a total of 67 HdIV proviral loci 429 dispersed among the 12 H. didymator chromosomes. To elucidate the time-course of HdIV loci 430 replication, the amplification of a subset of IV genome components was analyzed by qPCR. Our results 431 show that HdIV loci amplification initiates between stage 1 and stage 2 pupae and reaches its maximum 432 in stage 4 pupae. The temporal pattern observed in H. didymator is similar to BV-associated braconids. 433 In the braconid wasp Chelonus inanitus, where the amplification kinetics of two proviral segments have 434 been studied, local chromosomal amplification does not occur in the initial stages of pupal development 435 [36]. Instead, it is preceded by an increase in DNA content through endoreduplication [37]. The question .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 16 436 of whether calyx cell nuclei undergo polyploidization before local DNA amplification occurs in the case 437 of H. didymator has yet to be investigated. Collectively, our results indicate DNA amplification of IV 438 genome components constitutes one of the initial steps of virion morphogenesis. 439 Our data indicate all HdIV loci and genes located outside of IVSPERs are amplified with non-discrete 440 boundaries that extend variable distances into flanking wasp DNA. In contrast to certain integrated 441 viruses, such as polyomaviruses, which can be amplified in an "onion skin" type of replication with 442 replication forks terminating at discrete boundaries [38], IVSPER amplification more closely resembles 443 the local amplification observed in Drosophila follicle cells. In Drosophila, six loci corresponding to 444 chorion genes or genes related to oogenesis are amplified in large regions of about 100 Kbp beyond 445 the genes themselves, without discrete termination sites [39, 40]. Similar to IVSPERs, levels of DNA 446 amplification in Drosophila follicle cells vary among different amplicons [40, 41]. In Drosophila follicle 447 cells, amplification of these loci is associated with repeated firing of origins of replication (ORs) 448 interspersed within each gene cluster. This results in overlapping bidirectional replication forks 449 progressing outward on either side of the ORs [41]. These similarities between the pattern of DNA 450 amplification of Drosophila genes and H. didymator proviral loci suggest that IVSPERs and IV replication 451 genes may also be amplified through repeated firing of ORs present within the loci. However, additional 452 approaches, such as nascent strand sequencing based on λ-exonuclease enrichment [42], will be 453 necessary to identify ORs within IV genome components and validate this hypothesis. 454 Amplification of proviral segment loci is further characterized by a significant increase in read coverage 455 at the Direct Repeat (DR) positions bordering the proviral segments, which serve as sites for 456 homologous recombination and circularization of the segments. This suggests that a portion of the rapid 457 increase in read coverage is due to reads mapping to amplification intermediates and circularized 458 segments. The presence of circular forms in the sequenced genomic DNA samples is supported by our 459 qPCR results for segment Hd29, which indicate the presence of amplicons specific to the circular form 460 of Hd29 (Fig 7B). Accurately quantifying the proportion of reads mapping to the chromosomal form of 461 HdIV segments, and estimating the actual extent of local DNA amplification presents a challenge. This 462 is because paired-end reads that align within HdIV segment loci cannot discriminate between 463 chromosomal HdIV DNA, potential replication intermediates, or circularized DNA. Nevertheless, 464 considering the observed pattern of amplification in regions containing both IVSPERs and segments 465 (Fig 3B), we propose that proviral segment loci may undergo amplification similar to IVSPERs or HdIV .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 17 466 replication gene loci. The question persists regarding the subsequent processing of chromosomally 467 amplified DNA and the mechanism behind the generation of a large number of circular molecules. The 468 short-read data generated in this study have several limitations in characterizing whether amplification 469 of proviral segment loci generates concatemeric intermediates and, if so, their orientation. Long-read 470 data will be necessary to address these questions. Nonetheless, our results suggest HdIV proviral 471 segment amplification involves both local chromosomal amplification and amplification of intermediates 472 related to producing the circular dsDNAs that are packaged into capsids. 473 Our interest in U16 stemmed from previous results indicating it is transcriptionally upregulated in calyx 474 cells before the appearance of envelope and capsid components [16]. Sequence analysis during this 475 study revealed a PriCT-2 domain in U16, known from primases in herpesviruses, whose function is 476 unknown but may facilitate the association of the large primase domain (AEP) with DNA [31, 43]. 477 Although other known primase domains were not identified in the U16 sequence, the presence of a 478 PriCT-2 domain suggested this protein might play a role in the replication of HdIV genome components. 479 Additionally, our RNAi experiments demonstrate that U16 knockdown resulted in the complete absence 480 of virion production in calyx cell nuclei and calyx fluid. These observations indicated an essential role 481 for U16 in the early stages of viral replication, potentially involved in the amplification of HdIV genome 482 components and/or the transcriptional regulation of IV replication genes. Subsequently, we analyzed 483 the genome-wide impact of RNAi knockdown of U16 on HdIV loci amplification, revealing that this gene 484 is crucial for the amplification of all H. didymator IV genome components. In the case of IV replication 485 genes, reduced amplification was accompanied by a simultaneous significant reduction in transcript 486 abundance, likely resulting in insufficient amounts of HdIV structural proteins. However, amplification 487 and transcription abundance levels did not fully correlate with each other. For instance, U11 and IVSP3- 488 1 (both located on IVSPER-2) exhibit similar amplification patterns (Fig 1), but earlier findings showed 489 that transcript abundances were not the same in calyx cells [15]. Thus, differences in gene expression 490 observed among genes located within the same amplified regions (Fig 1) could also be affected by 491 promoter strength or other factors. On the other hand, inhibition of proviral segment loci amplification 492 had consequences for the abundance of the circularized dsDNA that are packaged into capsids, which 493 were drastically reduced. Thus, our results identify U16 as an essential protein for virion morphogenesis. 494 However, its precise role in viral replication remains to be understood. Questions to be addressed in the 495 future include whether U16 acts at the initiation or elongation step of HdIV DNA replication, whether it .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 18 496 interacts directly with DNA, or with proteins from the replisome complex, which itself could be composed 497 of a mixture of HdIV and wasp proteins. 498 BVs share some features with IVs but also exhibit differences. Notably, in contrast to IVs, where most 499 core genes with functions in virion morphogenesis reside in IVSPERs, many BV core replication genes 500 are widely dispersed in the genomes of wasps [44, 45, 46] and are not amplified in calyx cells during 501 virion morphogenesis [47]. However, the genomes of some BV-producing wasps do contain a ~400 kb 502 DNA domain in which several nudiviral core genes are located, known as the nudivirus-like cluster. This 503 feature potentially identifies a site where the nudivirus ancestor of BVs integrated into the common 504 ancestor of microgastroid braconids [9]. Notably, the nudivirus-like cluster is amplified with non-discrete 505 boundaries [47], similar to what is reported for IV genome components in this study. The observed 506 similarity in the amplification pattern between the BV nudivirus cluster and the proviral components of 507 IVs could suggest they are amplified through a common mechanism, even though the molecules 508 involved differ. 509 BV genomes also contain proviral segment loci with boundaries defined by flanking DRs and amplified 510 in regions that include flanking regions outside of each DR. However, unlike IV proviral segments, the 511 amplified flanking regions in BVs contain very precise nucleotide junctions that identify the boundaries 512 of amplification [47, 48]. It is also known that some BV proviral segments are amplified as head-to-tail 513 concatemers, consistent with a rolling circle amplification mechanism, while others are amplified as 514 head-to-head and tail-to-tail concatemers, suggesting amplification by different mechanisms. However, 515 all of these concatemers are similarly processed into circular DNAs by recombination at a precise site 516 within DRs, which is a tetramer conserved in all BV segments [47, 48]. Nudiviral genes encoding tyrosine 517 recombinases are further known to mediate this homologous recombination event [25, 26]. These types 518 of molecules could also be present in IV genomes and need to be discovered. Currently, a detailed 519 comparison between BV and IV proviral segment amplification is challenging and will require more 520 information about the machinery involved in the processing of IV proviral segments into circular dsDNAs 521 that are packaged into capsids. 522 Collectively, our results identify U16 as a gene deriving from the IV ancestor that is required for HdIV 523 DNA replication. This suggests that viral regulatory factors required for DNA amplification other than 524 U16 have been preserved in parasitoid genomes. U16 may also interact with wasp cellular machinery 525 in regulating DNA amplification, virion morphogenesis or both. Furthermore, this work emphasizes the .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 19 526 value of studying original endogenized viruses, such as those found in parasitoids, to unveil new 527 regulators of DNA processing. 528 Materials and Methods 529 Insects. H. didymator was reared as previously outlined by [49]. Female pupae obtained from cocoons 530 were staged using pigmentation patterns: stage 1, corresponding to hyaline pupae (approximately 3- 531 day-old pupae); stage 2, had a pigmented thorax (4-day-old); stage 3, had a pigmented thorax and 532 abdomen (5-day-old); stage 4, were pharate adults just before emergence. 533 Dovetail Omni-C Library Preparation and Sequencing. DNA from 10 male offspring (i.e., haploid 534 genomes) from a single female H. didymator was sent on dry ice to Dovetail Genomics for Omni-C™ 535 library construction. In the process of constructing the Dovetail Omni-C library, chromatin was fixed in 536 place within the nucleus using formaldehyde and subsequently extracted. The fixed chromatin was 537 digested with DNAse I followed by repair of chromatin ends and ligation to a biotinylated bridge adapter. 538 Proximity ligation of adapter-containing ends ensued. Post-proximity ligation, crosslinks were reversed, 539 and the DNA was purified. The purified DNA underwent treatment to eliminate biotin not internal to 540 ligated fragments. Sequencing libraries were generated utilizing NEBNext Ultra enzymes and Illumina- 541 compatible adapters. Fragments containing biotin were isolated using streptavidin beads before PCR 542 enrichment of each library. The library was sequenced using the Illumina HiSeqX platform, which 543 generated approximately 30x coverage. Subsequently, HiRise utilized reads with a mapping quality 544 greater than 50 (MQ>50) for scaffolding purposes. 545 Scaffolding the Assembly with HiRise. The de novo assembly from [15], and the Dovetail OmniC 546 library reads served as input data for HiRise, a specialized software pipeline designed for leveraging 547 proximity ligation data to scaffold genome assemblies, as outlined by [50]. The sequences from the 548 Dovetail OmniC library were aligned to the initial draft assembly using the bwa tool (available at 549 https://github.com/lh3/bwa). HiRise then analyzed the separations of Dovetail OmniC read pairs mapped 550 within the draft scaffolds. This analysis generated a likelihood model for the genomic distance between 551 read pairs. The model was subsequently employed to identify and rectify putative misjoins, score 552 potential joins, and execute joins above a specified threshold. A contact map was generated from a 553 BAM file by utilizing read pairs where both ends were aligned with a mapping quality of 60. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 20 554 Genomic DNA (gDNA) extraction for high throughput sequencing. Comparative analysis of two 555 pupal stages. Genomic DNA (gDNA) was extracted from pooled calyx samples dissected from H. 556 didymator female pupae at stage 1 (~60 females) and stage 3 (~50 females). Since the aim was to 557 compare the two developmental pupal stages, a single replicate was done for each stage. Impact of 558 U16 knockdown. Genomic DNA from calyces was collected from stage 3 female pupae that were 559 injected with dsGFP and dsU16. This experiment involved three biological replicates, each 560 corresponding to 30 to 50 calyx samples. Genomic DNA was extracted using the phenol-chloroform 561 method. Briefly, calyx samples were incubated in proteinase K (Ambion, 0.5 μg/μl) and Sarkosyl 562 detergent (Sigma, 20%), followed by treatment with RNAse (Promega, 0.3 μg/μl). Total genomic DNA 563 was then extracted through phenol-chloroform extraction and ethanol precipitation. Following extraction, 564 gDNA was quantified using a QBIT fluorometer (ThermoFisher) and subsequently sent for sequencing 565 to Genewiz/Azenta company. Paired-end sequencing was carried out using Illumina technology and 566 NovaSeq 2x150bp platform. 567 NGS data analyses. Illumina reads were aligned to the updated version of the H. didymator genome 568 using bwa mem [51], version 0.7.17, with default parameters. Subsequently, the aligned reads were 569 converted to BAM files utilizing samtools view (version 1.15) [52]. 570 Prediction of the amplified regions. Amplification peaks were identified using MACS2 [30] by comparing 571 the pupal stage 3 alignment file as treatment and the pupal stage 1 alignment file as control. The 572 specified parameters for this analysis were: --broad --nomodel -g 1.8e8 -q 0.01 --min-length 5000. Out 573 of the 165 predicted peaks (i.e., amplified regions), only those with a fold change (FC) higher than 2 574 were retained for further analyses, resulting in a total of 59 peaks. These 59 peaks encompassed all 575 known proviral loci, except for Hd40, which had a slightly lower value than the specified threshold 576 (FC=1.9), and Hd45.1 and Hd2-like, located too close to the scaffold end and potentially missed. For 577 the predicted peaks with FC>2 that did not correspond to known proviral loci, a manual curation was 578 performed to determine whether these regions corresponded to HdIV loci. Proviral segments were 579 identified by their flanking direct repeats (DRs) and gene contents, specifically the presence of genes 580 belonging to IV segment conserved gene families. To identify putative core IV replication genes, genes 581 present in the MACS2 peak were analyzed. Only those with no similarity to wasp proteins and that were 582 transcribed in calyx cells (based on the available transcriptome from [16]) were retained. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 21 583 Read coverage per proviral region (HdIV locus or amplified region). Raw read counts were determined 584 for each proviral region using featureCounts [53] from the Subread package (version 2.0.1) with the 585 parameters (-c -P -s 0 -O). Subsequently, coverage values were computed with a custom script available 586 at https://github.com/flegeai/EVE_amplification. Coverage values for each region were calculated by 587 dividing the number of fragments mapped to the region by the size of the region (expressed in kilobase 588 pairs, kbp), and further normalized by the depth of the library (expressed in million reads). These 589 coverages were computed for various types of genomic regions, including each locus (IVSPERs, IV 590 replication genes outside IVSPERs, proviral segments), each MACS2-detected amplified region, and 591 for each pupal stage (stage 1, St1 and stage 3, St3), as well as for each experiment (dsGFP and dsU16) 592 and each replicate. 593 Genome coverages per position on H. didymator scaffolds (Counts per Million, CPM) and Maximal value 594 of amplification per proviral locus. Genome coverages per position in 10 bp bins were acquired using 595 the BamCoverage tool from the deeptools package [54] with the options: --normalizeUsing CPM and - 596 bs 10. Subsequently, for each 10 bp bin, the pupal stage 3 (St3) versus stage 1 (St1) ratio was computed 597 through an in-house script available at https://github.com/flegeai/EVE_amplification. This script utilized 598 the pyBigWig python library from deeptools [54]. To determine the maximal counts per million (CPM) at 599 each stage for every proviral locus, an in-house script importing the pyBigWig python library was 600 employed. The maximum CPM value for the "stage 3 / stage 1" ratio was then calculated based on the 601 10 bp bin bigwig file, specifically for the position displaying the highest CPM value at stage 3 (summit). 602 Comparison of read coverages between HdIV loci and the rest of the wasp genome. One hundred sets 603 of random regions, each mimicking the size distribution of HdIV loci, were generated using the shuffle 604 tool from bedtools version 2.27 [55]. This was achieved by utilizing the bed file of HdIV loci (56 for 605 proviral segments and 11 for IVSPERs) as parameters for the shuffle tool. Raw read counts for these 606 randomly generated regions were computed in the same manner as for proviral regions, employing 607 featureCounts [53] from the Subread package (version 2.0.1) with the parameters (-c -P -s 0 -O). 608 Subsequently, coverage values per region were calculated using the same methodology as described 609 earlier, with an in-house script available at https://github.com/flegeai/EVE_amplification. 610 Search for motifs at the HdIV amplified regions boundaries. The MEME suite [56] was employed for 611 analyses using default parameters and a search for six motifs. A dataset comprising a total of 110 612 sequences, each spanning 1,000 nucleotides on both sides of the start and end positions of the 55 HdIV .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 22 613 amplified regions predicted by the MACS2 algorithm, was utilized for this analysis. As a control, a parallel 614 analysis was conducted using 110 sequences, each 2,000 nucleotides in length, randomly selected from 615 locations within the H. didymator genome but outside the HdIV loci. This control dataset allowed for the 616 comparison of motif patterns between the HdIV amplified regions and randomly chosen genomic 617 regions. 618 Genomic DNA extraction for gDNA amplification analyses by quantitative real-time PCR. To 619 assess the level of DNA amplification, total genomic DNA (gDNA) was extracted using the DNeasy 620 Blood & Tissue Kit (Qiagen) following the manufacturer's protocol. Ovaries (ovarioles removed) and hind 621 legs, representing the negative control, were dissected from ten pupae at four different stages. Three 622 replicates were generated for each pupal stage. Quantification of target gene amplification was 623 conducted through quantitative PCR, utilizing LightCycler® 480 SYBR Green I Master Mix (Roche) in 624 384-well plates (Roche). The total reaction volume per well was 3 µl, comprising 1.75 µl of the reaction 625 mix (1.49 µl SYBR Green I Master Mix, 0.1 µl nuclease-free water, and 0.16 µl diluted primer), and 1.25 626 µl of each gDNA sample diluted to achieve a concentration of 1.2 ng/µl. Primers used are listed in S4 627 Table. The gDNA levels corresponding to the viral genes and the housekeeping wasp gene (elongation 628 factor (ELF-1)) were determined using the LightCycler 480 System (Roche). The cycling conditions 629 involved heating at 95°C for 10 min, followed by 45 cycles of 95°C for 10 s, 58°C for 10 s, and 72°C for 630 10 s. Each sample was evaluated in triplicate. The obtained DNA levels were normalized with respect 631 to the wasp gene ELF-1. Raw data are provided in S3 Dataset. 632 Total RNA extraction. Total RNA was extracted from ovaries (ovarioles removed) dissected from pupae 633 at different stages using the Qiagen RNeasy extraction kit in accordance with the manufacturer's 634 protocol. To control for gene silencing, total RNAs were also extracted from individual adult wasp 635 abdomens (2 to 4 days old). For this, Trizol reagent (Ambion) was initially used followed by extraction 636 using the NucleoSpin® RNA kit (Macherey-Nagel). Isolated RNA was then subjected to DNase 637 treatment using the TURBO DNA-free Kit (Life Technologies) to assure removal of any residual genomic 638 DNA from the RNA samples. 639 Protein sequence analyses. Conserved domains of U16 were identified using the CD-search tool 640 available through NCBI's conserved domain database resource [57, 58]. Subcellular localization 641 predictions were made using the DeepLoc - 2.0 tool, a deep learning-based approach for predicting the 642 subcellular localization of eukaryotic proteins [59]. For multiple sequence alignment, CLUSTAL Omega .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 23 643 (version 1.2.4) was employed [60]. Structure predictions for U16 were carried out using the MPI 644 Bioinformatics Toolkit [61]. 645 RNA interference (RNAi). Gene-specific double-stranded RNA (dsRNA) used for RNAi experiments 646 was prepared using the T7 RiboMAX™ Express RNAi System (Promega). Initially, a 350-450 bp 647 fragment corresponding to the U16 sequence was cloned into the double T7 vector L4440 (a gift from 648 Andrew Fire, Addgene plasmid # 1654). Subsequently, an in vitro transcription template DNA was PCR 649 amplified with a T7 primer, and this template was used to synthesize sense and antisense RNA strands 650 with T7 RNA polymerase at 37°C for 5 hours. The primers used for dsRNA production are listed in S4 651 Table. After annealing and DNase treatment using the TURBO DNA-free Kit (Life Technologies), the 652 purified dsRNAs were resuspended in nuclease-free water, quantified using a NanoDrop ND-1000 653 Spectrophotometer (Thermo Scientific), and examined by agarose gel electrophoresis to ensure their 654 integrity. Injections were performed in less than one-day-old female pupae using a microinjector 655 (Fentojet® Express, Eppendorf®) and a micromanipulator (Narishige®). Approximately 0.3-0.6 μl of 500 656 ng/μl dsRNA was injected into each individual. Control wasps were injected with a non-specific dsRNA 657 homologous to the green fluorescent protein (GFP) gene. Treated pupae were kept in an incubator until 658 adult emergence, which occurred approximately 5 days after injection. 659 Transmission electron microscopy. Ovaries were dissected from adult wasps between 2 and 3 days 660 after emergence, following the procedures outlined in [17]. To ensure consistency of the observed 661 phenotype, at least three females (taken at different microinjection dates) were observed for each tested 662 dsRNA. For transmission electron microscopy (TEM) observations, calyces were fixed in a solution of 663 2% glutaraldehyde in PBS for 2 hours and then post-fixed in 2% osmium tetroxide in the same buffer 664 for 1 hour. Tissues were subsequently bulk-stained for 2 hours in a 5% aqueous uranyl acetate solution, 665 dehydrated in ethanol, and embedded in EM812 resin (EMS). Ultrathin sections were double-stained 666 with Uranyless (DeltaMicroscopy) and lead citrate before examination under a Jeol 1200 EXII electron 667 microscope at 100 kV (MEA Platform, University of Montpellier). Images were captured with an EMSIS 668 Olympus Quemesa 11 Megapixels camera and analyzed using ImageJ software [62]. 669 Reverse-transcriptase quantitative real-time PCR (RT-qPCR). For RT-qPCR assays, 400 ng of total 670 RNA was reverse-transcribed using the SuperScript III Reverse Transcriptase kit (Life Technologies) 671 and oligo(dT)15 primer (Promega). The mRNA transcript levels of selected IVSPER genes were 672 measured by quantitative reverse transcription-PCR (qRT-PCR) using a LightCycler® 480 System .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 24 673 (Roche) and SYBR Green I Master Mix (Roche). Expression levels were normalized relative to a 674 housekeeping wasp gene (elongation factor 1 ELF-1). Each sample was evaluated in triplicate, and the 675 total reaction volume per well was 3 µl, including 0.5 µM of each primer and cDNA corresponding to 676 0.88 ng of total RNA. The amplification program consisted of an initial step at 95°C for 10 min, followed 677 by 45 cycles of 95°C for 10 s, 58°C for 10 s, and 72°C for 10 s. The primers used for this analysis are 678 listed in S4 Table. 679 qPCR data analysis. Data were acquired using Light-Cycler® 480 software. PCR amplification 680 efficiency (E) for each primer pair was determined by linear regression of a dilution series (5x) of the 681 cDNA pool. Relative expression, using the housekeeping gene ELF-1 as a reference, was calculated 682 through advanced relative quantification (Efficiency method) software provided by Light-Cycler® 480 683 software. For statistical analyses, Levene’s and Shapiro-Wilk tests were employed to verify homogeneity 684 of variance and normal distribution of data among the tested groups. Differences in gene relative 685 expression between developmental stages and between dsGFP and dsU16-injected females were 686 assessed using a two-tailed unpaired t-test for group comparison. In cases where homogeneity of 687 variance was not assumed, the Welch-test was used to compare gene relative expression between 688 groups. A p-value < 0.05 was considered significant. All statistical analyses were conducted using R 689 [63]. Detailed statistical analyses of qPCR results are provided in S3 Dataset. 690 Data availability. The datasets supporting the conclusions in this article are accessible at the NCBI 691 Sequence Read Archive (SRA) under the Bioproject accession number PRJNA589497. Additionally, the 692 new version of the H. didymator genome, annotation, alignments of reads, and coverage information 693 can be found at BIPAA (https://bipaa.genouest.org/sp/hyposoter_didymator/). Raw data and statistical 694 analyses for all the qPCR analyses are provided in S3 Dataset. Furthermore, sequencing raw data, read 695 coverage analyses, statistical analyses, and in-house scripts are available at 696 https://github.com/flegeai/EVE_amplification. 697 698 Acknowledgments 699 The insects used in the experiments were provided by Raphaël BOUSQUET and Gaétan CLABOTS 700 from the DGIMI insect rearing facility. All RNAi experiments were conducted in the insect quarantine 701 platform (PIQ) of DGIMI lab, which is a member of the Montpellier Vectopole Sud network 702 (https://www.vectopole-sud.fr/). Microscopy observations were facilitated by the Montpellier MEA .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 25 703 platform (https://mea.edu.umontpellier.fr/). All qPCR analyses were performed with the assistance of 704 the Montpellier Genomix qPHD platform (http://www.pbs.univ-montp2.fr/). 705 706 References 707 1. Katzourakis A, Gifford RJ. Endogenous viral elements in animal genomes. PLoS Genet. 2010 Nov 708 18;6(11):e1001191. doi: 10.1371/journal.pgen.1001191. 709 2. Kryukov K, Ueda MT, Imanishi T, Nakagawa S. Systematic survey of non-retroviral virus-like 710 elements in eukaryotic genomes. Virus Res. 2019 Mar;262:30-36. doi: 711 10.1016/j.virusres.2018.02.002. 712 3. Frank JA, Feschotte C. Co-option of endogenous viral sequences for host cell function. Curr Opin 713 Virol. 2017 Aug;25:81-89. doi: 10.1016/j.coviro.2017.07.021. 714 4. Feschotte C, Gilbert C. Endogenous viruses: insights into viral evolution and impact on host biology. 715 Nat Rev Genet. 2012;13(4):283-296. doi: 10.1038/nrg3199. 716 5. Gilbert C, Feschotte C. Genomic fossils calibrate the long-term evolution of hepadnaviruses. PLoS 717 Biol. 2010 Sep;8(9):e1000495. doi: 10.1371/journal.pbio.1000495. 718 6. Drezen JM, Bézier A, Burke GR, Strand MR. Bracoviruses, ichnoviruses, and virus-like particles 719 from parasitoid wasps retain many features of their virus ancestors. Curr Opin Insect Sci. 2022 720 Feb;49:93-100. doi: 10.1016/j.cois.2021.12.003. 721 7. Stoltz DB, Vinson SB. Viruses and parasitism in insects. Adv Virus Res. 1979;24:125-71. doi: 722 10.1016/s0065-3527(08)60393-0. 723 8. Webb BA, Strand MR. The biology and genomics of polydnaviruses. In: Comprehensive Molecular 724 Insect Science, Vol. 6, ed. K Iatrou, S Gill, pp. 323–60. Amsterdam: Pergamon. 2005. 725 9. Bézier A, Annaheim M, Herbinière J, Wetterwald C, Gyapay G, Bernard-Samain S, et al. 726 Polydnaviruses of braconid wasps derive from an ancestral nudivirus. Science. 2009 Feb 727 13;323(5916):926-30. doi: 10.1126/science.1166788. 728 10. Pichon A, Bézier A, Urbach S, Aury JM, Jouan V, Ravallec M, et al. Recurrent DNA virus 729 domestication leading to different parasite virulence strategies. Sci Adv. 2015 Nov 730 27;1(10):e1501150. doi: 10.1126/sciadv.1501150. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 26 731 11. Burke GR. Common themes in three independently derived endogenous nudivirus elements in 732 parasitoid wasps. Curr Opin Insect Sci. 2019 Apr;32:28-35. doi: 10.1016/j.cois.2018.10.005. Epub 733 2018 Oct 23. PMID: 31113628. 734 12. Volkoff AN, Jouan V, Urbach S, Samain S, Bergoin M, Wincker P, et al. Analysis of virion structural 735 components reveals vestiges of the ancestral ichnovirus genome. PLoS Pathog. 2010 May 736 27;6(5):e1000923. doi: 10.1371/journal.ppat.1000923. 737 13. Béliveau C, Cohen A, Stewart D, Periquet G, Djoumad A, Kuhn L, et al. Genomic and Proteomic 738 Analyses Indicate that Banchine and Campoplegine Polydnaviruses Have Similar, if Not Identical, 739 Viral Ancestors. J Virol. 2015 Sep;89(17):8909-21. doi: 10.1128/JVI.01001-15. 740 14. Volkoff A-N, Huguet E. Polydnaviruses (Polydnaviridae). In: Bamford DH, Zuckerman M, editors. 741 Encyclopedia of Virology (Fourth Edition). Academic Press, Oxford; 2021. pp. 849-857. DOI: 742 10.1016/B978-0-12-809633-8.21556-2. 743 15. Legeai F, Santos BF, Robin S, Bretaudeau A, Dikow RB, Lemaitre C, et al. Genomic architecture 744 of endogenous ichnoviruses reveals distinct evolutionary pathways leading to virus domestication 745 in parasitic wasps. BMC Biol. 2020 Jul 24;18(1):89. doi: 10.1186/s12915-020-00822-3. 746 16. Lorenzi A, Ravallec M, Eychenne M, Jouan V, Robin S, Darboux I, et al. RNA interference identifies 747 domesticated viral genes involved in assembly and trafficking of virus-derived particles in 748 ichneumonid wasps. PLoS Pathog. 2019 Dec 13;15(12):e1008210. doi: 749 10.1371/journal.ppat.1008210. 750 17. Robin S, Ravallec M, Frayssinet M, Whitfield J, Jouan V, Legeai F, et al. Evidence for an ichnovirus 751 machinery in parasitoids of coleopteran larvae. Virus Res. 2019;263: 189–206. doi: 752 10.1016/j.virusres.2019.02.001. 753 18. Cui L, Webb BA. Homologous sequences in the Campoletis sonorensis polydnavirus genome are 754 implicated in replication and nesting of the W segment family. J Virol. 1997 Nov;71(11):8504-13. 755 doi: 10.1128/JVI.71.11.8504-8513.1997. 756 19. Rattanadechakul W, Webb BA. Characterization of Campoletis sonorensis ichnovirus unique 757 segment B and excision locus structure. J Insect Physiol. 2003 May;49(5):523-32. doi: 758 10.1016/s0022-1910(03)00053-2. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 27 759 20. Blissard GW, Smith OP, Summers MD. Two related viral genes are located on a single superhelical 760 DNA segment of the multipartite Campoletis sonorensis virus genome. Virology. 1987 761 Sep;160(1):120-34. doi: 10.1016/0042-6822(87)90052-3. 762 21. Theilmann DA, Summers MD. Molecular analysis of Campoletis sonorensis virus DNA in the 763 lepidopteran host Heliothis virescens. J Gen Virol. 1986 Sep;67(Pt 9):1961-9. doi: 10.1099/0022- 764 1317-67-9-1961. 765 22. Webb BA, Strand MR, Dickey SE, Beck MH, Hilgarth RS, Barney WE, et al. Polydnavirus genomes 766 reflect their dual roles as mutualists and pathogens. Virology. 2006 Mar 30;347(1):160-74. doi: 767 10.1016/j.virol.2005.11.010. 768 23. Dorémus T, Cousserans F, Gyapay G, Jouan V, Milano P, Wajnberg E, et al. Extensive 769 transcription analysis of the Hyposoter didymator ichnovirus genome in permissive and non- 770 permissive lepidopteran host species. PLoS One. 2014 Aug 12;9(8):e104072. doi: 771 10.1371/journal.pone.0104072. 772 24. Volkoff AN, Ravallec M, Bossy JP, Cerutti P, Rocher J, Cerutti M, Devauchelle G. The replication 773 of Hyposoter didymator polydnavirus: Cytopathology of the calyx cells in the parasitoid. Biology of 774 the Cell. 1995;83(1):1-13. 775 25. Burke GR, Thomas SA, Eum JH, Strand MR. Mutualistic polydnaviruses share essential replication 776 gene functions with pathogenic ancestors. PLoS Pathog. 2013;9(5):e1003348. doi: 777 10.1371/journal.ppat.1003348. 778 26. Lorenzi A, Arvin MJ, Burke GR, Strand MR. Functional characterization of Microplitis demolitor 779 bracovirus genes that encode nucleocapsid components. J Virol. 2023 Oct 25:e0081723. doi: 780 10.1128/jvi.00817-23. 781 27. Rocher J, Ravallec M, Barry P, Volkoff AN, Ray D, Devauchelle G, Duonor-Cérutti M. Establishment 782 of cell lines from the wasp Hyposoter didymator (Hym., Ichneumonidae) containing the symbiotic 783 polydnavirus H. didymator ichnovirus. J Gen Virol. 2004 Apr;85(Pt 4):863-868. doi: 784 10.1099/vir.0.19713-0. 785 28. Krell PJ, Summers MD, Vinson SB. Virus with a multipartite superhelical DNA genome from the 786 ichneumonid parasitoid Campoletis sonorensis. J Virol. 1982 Sep;43(3):859-70. doi: 787 10.1128/JVI.43.3.859-870.1982. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 28 788 29. Robinson JT, Thorvaldsdóttir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP. 789 Integrative Genomics Viewer. Nat Biotechnol. 2011;29:24-26. doi:10.1038/nbt.1754. 790 30. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, et al. Model-based analysis 791 of ChIP-Seq (MACS). Genome Biol. 2008;9(9):R137. doi: 10.1186/gb-2008-9-9-r137. 792 31. Iyer LM, Koonin EV, Leipe DD, Aravind L. Origin and evolution of the archaeo-eukaryotic primase 793 superfamily and related palm-domain proteins: structural insights and new members. Nucleic Acids 794 Res. 2005 Jul 15;33(12):3875-96. doi: 10.1093/nar/gki702. PMID: 16027112. 795 32. Burke GR, Hines HM, Sharanowski BJ. The presence of ancient core genes reveals 796 endogenization from diverse viral ancestors in parasitoid wasps. Genome Biol Evol. 2021 Jul 797 6;13(7):evab105. doi: 10.1093/gbe/evab105. PMID: 33988720. 798 33. Beckage NE. Polydnaviruses as Endocrine Regulators. In: Beckage NE, Drezen J-M, eds. 799 Parasitoid Viruses. Academic Press; 2012. pp. 163-168 (Chapter 13). doi: 10.1016/b978-0-12- 800 384858-1.00013-8. 801 34. Strand MR. Polydnavirus gene products that interact with the host immune system. In Beckage NE, 802 Drezen J-M (eds.), Parasitoid Viruses. Elsevier. Academic Press, San Diego. 2012. pp. 149-161. 803 doi: 10.1016/B978-0-12-384858-1.00012-6. 804 35. Arvin MJ, Lorenzi A, Burke GR, Strand MR. MdBVe46 is an envelope protein that is required for 805 virion formation by Microplitis demolitor bracovirus. J Gen Virol. 2021 Mar;102(3):001565. doi: 806 10.1099/jgv.0.001565. 807 36. Marti D, Grossniklaus-Bürgin C, Wyder S, Wyler T, Lanzrein B. Ovary development and 808 polydnavirus morphogenesis in the parasitic wasp Chelonus inanitus. I. Ovary morphogenesis, 809 amplification of viral DNA and ecdysteroid titres. J Gen Virol. 2003 May;84(Pt 5):1141-1150. doi: 810 10.1099/vir.0.18832-0. 811 37. Wyler T, Lanzrein B. Ovary development and polydnavirus morphogenesis in the parasitic wasp 812 Chelonus inanitus. II. Ultrastructural analysis of calyx cell development, virion formation and 813 release. J Gen Virol. 2003;84:1151-63. doi: 10.1099/vir.0.18830-0. 814 38. Baran N, Neer A, Manor H. "Onion skin" replication of integrated polyoma virus DNA and flanking 815 sequences in polyoma-transformed rat cells: termination within a specific cellular DNA segment. 816 Proc Natl Acad Sci U S A. 1983 Jan;80(1):105-9. doi: 10.1073/pnas.80.1.105. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 29 817 39. Spradling AC. The organization and amplification of two chromosomal domains containing 818 Drosophila chorion genes. Cell. 1981 Nov;27(1 Pt 2):193-201. doi: 10.1016/0092-8674(81)90373- 819 1. 820 40. Kim JC, Nordman J, Xie F, Kashevsky H, Eng T, Li S, et al. Integrative analysis of gene amplification 821 in Drosophila follicle cells: parameters of origin activation and repression. Genes Dev. 2011 Jul 822 1;25(13):1384-98. doi: 10.1101/gad.2043111. 823 41. Tower J. Developmental gene amplification and origin regulation. Annu Rev Genet. 2004;38:273- 824 304. doi: 10.1146/annurev.genet.37.110801.143851. 825 42. Foulk MS, Urban JM, Casella C, Gerbi SA. Characterizing and controlling intrinsic biases of lambda 826 exonuclease in nascent strand sequencing reveals phasing between nucleosomes and G- 827 quadruplex motifs around a subset of human replication origins. Genome Res. 2015 828 May;25(5):725-35. doi: 10.1101/gr.183848.114. 829 43. Weller SK, Kuchta RD. The DNA helicase-primase complex as a target for herpes viral infection. 830 Expert Opin Ther Targets. 2013 Oct;17(10):1119-32. doi: 10.1517/14728222.2013.827663. 831 44. Burke GR, Walden KK, Whitfield JB, Robertson HM, Strand MR. Widespread genome 832 reorganization of an obligate virus mutualist. PLoS Genet. 2014 Sep;10(9):e1004660. doi: 833 10.1371/journal.pgen.1004660. 834 45. Gauthier J, Boulain H, van Vugt JJFA, Baudry L, Persyn E, Aury JM, et al. Chromosomal scale 835 assembly of parasitic wasp genome reveals symbiotic virus colonization. Commun Biol. 2021 Jan 836 22;4(1):104. doi: 10.1038/s42003-020-01623-8. Erratum in: Commun Biol. 2021 Jul 30;4(1):940. 837 46. Mao M, Strand MR, Burke GR. The complete genome of Chelonus insularis reveals dynamic 838 arrangement of genome components in parasitoid wasps that produce bracoviruses. J Virol. 2022 839 Mar 9;96(5):e0157321. doi: 10.1128/JVI.01573-21. 840 47. Burke GR, Simmonds TJ, Thomas SA, Strand MR. Microplitis demolitor Bracovirus proviral loci and 841 clustered replication genes exhibit distinct DNA amplification patterns during replication. J Virol. 842 2015 Sep;89(18):9511-23. doi: 10.1128/JVI.01388-15. 843 48. Louis F, Bézier A, Periquet G, Ferras C, Drezen JM, Dupuy C. The bracovirus genome of the 844 parasitoid wasp Cotesia congregata is amplified within 13 replication units, including sequences 845 not packaged in the particles. J Virol. 2013 Sep;87(17):9649-60. doi: 10.1128/JVI.00886-13. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 30 846 49. Visconti V, Eychenne M, Darboux I. Modulation of antiviral immunity by the ichnovirus HdIV in 847 Spodoptera frugiperda. Mol Immunol. 2019 Apr;108:89-101. doi: 10.1016/j.molimm.2019.02.011. 848 50. Putnam NH, O'Connell BL, Stites JC, Rice BJ, Blanchette M, Calef R, et al. Chromosome-scale 849 shotgun assembly using an in vitro method for long-range linkage. Genome Res. 2016 850 Mar;26(3):342-50. doi: 10.1101/gr.193474.115. 851 51. Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. 852 arXiv:1303.3997v2 [q-bio.GN]. doi: 10.48550/arXiv.1303.3997. 853 52. Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, et al. Twelve years of SAMtools 854 and BCFtools. Gigascience. 2021 Feb 16;10(2):giab008. doi: 10.1093/gigascience/giab008. 855 53. Liao Y, Smyth GK, Shi W. featureCounts: An efficient general-purpose program for assigning 856 sequence reads to genomic features. Bioinformatics. 2014 Apr 1;30(7):923-30. doi: 857 10.1093/bioinformatics/btt656. 858 54. Ramírez F, Ryan DP, Grüning B, Bhardwaj V, Kilpert F, Richter AS, et al. deepTools2: A next- 859 generation web server for deep-sequencing data analysis. Nucleic Acids Res. 2016 Jul 860 8;44(W1):W160-5. doi: 10.1093/nar/gkw257. 861 55. Quinlan AR, Hall IM. BEDTools: A flexible suite of utilities for comparing genomic features. 862 Bioinformatics. 2010 Mar 15;26(6):841-2. doi: 10.1093/bioinformatics/btq033. 863 56. Bailey TL, Johnson J, Grant CE, Noble WS. The MEME Suite. Nucleic Acids Res. 2015 Jul 864 1;43(W1):W39-49. doi: 10.1093/nar/gkv416. 865 57. Marchler-Bauer A, Bo Y, Han L, He J, Lanczycki CJ, Lu S, et al. CDD/SPARCLE: functional 866 classification of proteins via subfamily domain architectures. Nucleic Acids Res. 2017 Jan 867 4;45(D1):D200-D203. doi: 10.1093/nar/gkw1129. 868 58. Lu S, Wang J, Chitsaz F, Derbyshire MK, Geer RC, Gonzales NR, et al. CDD/SPARCLE: the 869 conserved domain database in 2020. Nucleic Acids Res. 2020 Jan 8;48(D1):D265-D268. doi: 870 10.1093/nar/gkz991. 871 59. Thumuluri V, Armenteros JJA, Johansen AR, Nielsen H, Winther O. DeepLoc 2.0: multi-label 872 subcellular localization prediction using protein language models. Nucleic Acids Research. 2022. 873 doi:10.1093/nar/gkac278. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 31 874 60. Madeira F, Pearce M, Tivey ARN, Basutkar P, Lee J, Edbali O, et al. Search and sequence analysis 875 tools services from EMBL-EBI in 2022. Nucleic Acids Res. 2022 Jul 5;50(W1):W276-W279. doi: 876 10.1093/nar/gkac240. 877 61. Gabler F, Nam SZ, Till S, Mirdita M, Steinegger M, Söding J, et al. Protein sequence analysis using 878 the MPI Bioinformatics Toolkit. Curr Protoc Bioinformatics. 2020 Dec;72(1):e108. doi: 879 10.1002/cpbi.108. 880 62. Rasband WS. ImageJ. National Institutes of Health, Bethesda, Maryland, USA. 1997-2018. 881 http://imagej.nih.gov/ij 882 63. R: A language and environment for statistical computing. R Foundation for Statistical Computing, 883 Vienna, Austria. R Core Team. 2023. URL https://www.R-project.org/. 884 885 Supporting information captions 886 S1 Dataset. Hyposoter didymator Hi-C genome assembly. The dataset includes: A. Figure depicting 887 the Hi‐C scaffold contact map; B. Table presenting the Hi-C scaffolds containing HdIV loci; C. Figure 888 displaying the pairwise comparisons of HdIV segments located in close proximity within the H. didymator 889 scaffolds. 890 S2 Dataset. Sequence analysis and alignment of the U16 gene from H. didymator to four other wasp 891 species that harbor IVs. The dataset includes: A. Multiple sequence alignment of U16 proteins from 892 different parasitoid species. B. Detail of the predicted secondary structure of the PricT-2 domain in the 893 H. didymator U16 protein. C. Subcellular localization of U16 predicted by DeepLoc 2.0. 894 S3 Dataset. Raw data and statistical analyses of qPCR analyses. The dataset includes raw data and 895 statistical analyses for: A. Genomic DNA amplification of IVSPER genes at four different H. didymator 896 pupal stages; B. Genomic DNA amplification of IVSPER and HdIV segment genes in dsGFP and dsU16- 897 injected wasps; C. RNA quantification of IVSPER genes in dsGFP and dsU16-injected wasps; D. DNA 898 amplification of Hd29 segment in dsGFP and dsU16-injected wasps. 899 S1 Table. Read coverage of HdIV loci on each scaffold of the H. didymator genome. 900 S2 Table. List of the peaks predicted in H. didymator genome scaffolds using MACS2 algorithm. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 32 901 S3 Table. Read coverage of HdIV amplified regions in calyx cell DNA from dsGFP- and dsU16-injected 902 female pupae. 903 S4 Table. List of primers used in the present work. 904 S1 Fig. DNA amplification patterns of HdIV loci in calyx cells of H. didymator. 905 S2 Fig. HdIV amplified regions in Scaffold-11. 906 S3 Fig. MEME analysis of boundaries of the predicted MACS2 HdIV amplified regions. 907 908 909 910 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint 33 911 Author contribution 912 A. LORENZI: Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, 913 Writing – Original Draft Preparation, Writing – Review & Editing 914 F. LEGEAI: Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, 915 Writing – Original Draft Preparation, Writing – Review & Editing 916 V. JOUAN, P.-A. GIRARD, M. EYCHENNE, M. RAVALLEC, Investigation, Methodology, Validation 917 A. BRETAUDEAU, S. ROBIN, Data Curation 918 J. ROCHEFORT, M. VILLEGAS, Investigation 919 M. R. STRAND, G. R. BURKE, Writing – Review & Editing 920 R. REBOLLO, Funding Acquisition, Validation, Writing – Original Draft Preparation, Writing – Review & 921 Editing 922 N. NÈGRE, Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, 923 Resources, Supervision, Validation, Writing – original draft, Writing – review & editing 924 A.-N. VOLKOFF, Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, 925 Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – 926 original draft, Writing – review & editing 927 928 Keywords: Endogenous viral element, DNA amplification, Hyposoter didymator, Ichnovirus, 929 polydnavirus, viral replication, RNA interference, co-option, co-evolution 930 931 Fundings 932 This work has been financially supported by the INRAE SPE department (EPIHYPO project) and the 933 French National Research Agency (ENDOVIRE project, #ANR-22-CE20-0005-01). The Dovetail 934 sequencing of the H. didymator genome has received funding from the European Union’s Horizon 2020 935 research and innovation program under the Marie Skłodowska-Curie grant agreement no. 764840 for 936 the ITN IGNITE project, with Denis TAGU from IGEPP as a partner. .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted January 18, 2024. ; https://doi.org/10.1101/2024.01.18.576166doi: bioRxiv preprint