D. simulans (Riverside) (DSR) stocks were maintained at room temperature on a standard diet of cornmeal, dextrose and yeast. Stocks were stably infected with a single Wolbachia strain (wRi) and have been maintained at the University of Alberta laboratory for approximately 6 years. The presence of Wolbachia was confirmed at regular intervals using 81F and 691R wsp primer pairs 21 .

DNA from whole flies and gonads was extracted from animals that were less than 5 days post eclosion. Newly eclosed flies were separated by sex and allowed to develop to the appropriate age. Gonad DNA samples were obtained from 4 groups of 100-150 testes each and 4 groups of 75-150 ovaries each. Whole fly DNA was taken from 4 groups of 15 flies each. Three groups of 15-minute AEL (after-egg-laying) embryos were obtained by allowing females to oviposit on egg laying dishes made from fruit juice. Embryos were collected and chilled every 15 minutes until approximately 200 μl of packed embryos were obtained per replicate. The eggs were stored at -80C until DNA extractions could be performed.

Synchronization of larvae was accomplished by allowing several hundred females to oviposit on egg-laying dishes for one hour. The eggs were collected and seeded onto standard media. From these, third instar (3') larvae were collected and stored at -80C until DNA extraction.

DNA was extracted from all tissues and flies with the DNeasy Blood and Tissue Kit (Qiagen) using the manufacturer's protocol with an extended, overnight proteinase K digestion. DNA purity and concentration was determined using a Nanodrop ND1000.

Relative copy numbers of Wolbachia and WO phage in .D. simulans were obtained using the MiniOpticon System (Bio-Rad). The relative Wolbachia infection level was measured by comparing the copy number of the gene for Wolbachia surface protein, wsp, to a single copy gene in the Drosophila genome, CuZn superoxide dismutase (sod). Phage copy numbers were measured by comparing the adenine methyltransferase (wMTase) (WORiB), lyzozyme (WORiA), and tail tube protein (WORiC) genes to wsp in wRi (see table 1 for locus tags and primer sequences).

Reactions were performed in low profile 48-well white plates with flat cap strips (Bio-Rad). Ten microliter reactions included 400nM of each forward and reverse primer, 5 μl of 2× Dynamite qPCR mastermix (Molecular Biology Service Unit - University of Alberta) which included SYBR green (Molecular Probes) and Platinum Taq (Invitrogen), and 125ng of DNA. The thermal cycling conditions were 95°C for 2 minutes, 40 cycles of 95°C, 55°C, and 72°C for 30 seconds each, and a final 2 minute 72°C extension. Fluorescent data were acquired after every 72°C extension. A 60-95°C melting curve was performed to confirm the specificity of the products. No template controls were included to account for DNA contamination. All samples were analyzed in technical and biological triplicates. Standard curves were constructed through a dilution series to validate the primer pairs; all primers were found to have efficiencies that were roughly equal, using the equation E = 10^(-1/slope) -1, and suitable for relative comparisons, and r² values >0.99. The primer sequences were designed using PerlPrimer v1.1.14 [http://perlprimer.sourceforge.net] and are described in table 1. All primers were synthesized by Integrated DNA Technologies and were purified by standard desalting. PCR products were sequenced to confirm specificity of the primers and all amplified a single, specific target. Data were analyzed by the Opticon Monitor 3 software (Bio-Rad) which uses the ΔCT method. The average copy number of integrated phage was compared to the expected number based on published sequence data and the difference was statistically analyzed with a two-tailed t-test. The correlation tests between the three WO phages and wRi were performed using the Pearson Product Moment Correlation test. When determining the relative copy number for each of the phage types, it was assumed that integrated prophage sequences would amplify with the same efficiency as sequences from mature virus particles.

Annotated genomes of Wolbachia strains wMel [GenBank:NC_002978] 10 and wRi [GenBank:NC_012416] 4 , and phage strains WOCauB2 [GenBank:AB478515] 9 , and WOVitA [GenBank:HQ906662] 12 were retrieved 22 . The phage regions [WRi_005250-005970] (WORiB) and [WRi_006570-WRi_007250] (WORiC) from the wRi genome were used for whole phage genome alignments. The region [WD0562-WD0646] from the wMel genome was used for WOMelB genome alignments. Whole genome comparisons were performed using the Mauve plug-in v.2.2.0 20 for Geneious v5.4.4 23 . The predicted amino acid sequences for the large terminase subunit and baseplate assembly gene W were used for phylogenetic analysis.

Proteins were aligned using the ClustalW multiple alignment algorithm implemented in Geneious v5.4.4. 23 . Model selection was performed using Prottest 2.4 24 with Akaike's information criterion (AIC) used to select for an appropriate evolutionary model for each data set [terminase (JTT+I+Γ+F) and baseplate assembly protein W (JTT+Γ)] prior to analysis. The evolutionary history was inferred for both genes using the maximum likelihood method. Phylogenetic trees generated by PHYML used 1000 bootstrap replicated datasets and estimated gamma distribution and proportion of invariable sites 25 .

When lytic viruses replicate and lyse host cells, they do so through an enzymatic process involving a two component cell lysis system of a holin and lysozyme 26 . To date, there is no direct evidence that the WO phages of wRi are capable of enzymatic lysis of bacterial hosts. Therefore, the term "lytic" is not used here to describe phage or phage DNA detected in excess of the integrated prophage genomes. Instead, replicating WO is referred to as a mature, extrachromosomal, or active phage. WO phages in wMel and wRi have been classically referred to as WO-A, WO-B, and WO-C 4 10 . However, the less ambiguous nomenclature WORiA, WORiB, and WORiC respectively is used here to describe the prophage types in wRi and WOMelA and WOMelB for prophage types in wMel.

Quantitative PCR was used to test whether Wolbachia prophages were replicating extrachromosomally. Specific primers that differentiate between the prophage types in wRi were designed (table 1) and Wolbachia titer was determined by comparing the wsp gene copy number to the Drosophila nuclear sod gene. Integrated and extrachromosomal viral copy numbers were determined using primers specific to Wolbachia genes lysozyme (WORiA), MTase (WORiB), and tail tube protein (WORiC). The amplification of the WO-specific primers was compared to Wolbachia copy number using wsp (wRi-specific primers).Values reported are the combination of integrated plus extrachromosomal phages.

WORiA is found once in the wRi genome. The relative copy number of the ORF which encodes a putative lyzozyme [WRi _012650] was measured in young males and females (three replicates of 15 flies each), testes and ovaries, and 15 minute AEL embryos. The relative lyzozyme (WORiA) copy number in these tissues ranged from 0.94 - 1.16 per Wolbachia cell (figure 1A). This is consistent with the single integrated copy in the genome and indicates no extrachromosomal WORiA (all p values > 0.05; two-tailed t-test).

In wRi, there are two integrated copies of the WORiB prophage and each contains one copy of the MTase gene [WRi_005640; WRi_010300] 4 . In DSR males, females, testes, ovaries, and two-hour embryos, the relative MTase copy number ranged from 1.83-2.10 and was not significantly different than two per Wolbachia genome (all p values > 0.05, two-tailed t-test) (figure 1B). There is no evidence of extrachromosomal WORiB phage genomes.

The gene encoding the phage tail tube protein is present once in the wRi genome on the WORiC insert. In males, females, testes, ovaries, and 15 minute AEL embryos, the relative tail tube protein copy number was significantly greater than the expected one copy per Wolbachia genome (p < 0.05 in all cases, two- tailed t-test) (figure 1C). Therefore, WORiC is the extrachromosomal phage in wRi. The average density of all samples tested ranged from 1.29 - 1.61 copies of WORiC per wsp copy.

Occasionally, a DNA sample showed no evidence of extra-chromosomal WORiC DNA (data not shown). This indicates that DNA extracted from groups of flies may mask variation with respect to the amount of replicating phage per individual. Thus, third instar larvae were synchronized to a 1 hour age difference and wRi, WORiA, WORiB, and WORiC numbers were measured for each individual to determine whether the WO copy number varied between individuals (figure 2). Relative phage densities were also compared to Wolbachia densities to determine whether variations in phage copy numbers were related to the bacterial density as observed by Bordenstein et al 15 in N. vitripennis. Among 16 third instar larvae tested, the Wolbachia densities ranged from 6.67 to 19.21 copies per host sod gene, with the exception of one outlier at 34.88. WORiA relative numbers averaged 0.97 and varied from 0.86 to 1.13 copies per Wolbachia. WORiB densities for the larvae averaged 2.02 copies per wRi and ranged between 1.56 and 2.78. Finally, WORiC copy numbers averaged 1.17 and ranged between 0.91 and 1.50 per wsp. None of the densities of the three phage types correlated significantly with the Wolbachia density (Pearson correlation; p = 0.256, 0.12, and 0.16 for WORiA, WORiB, and WORiC, respectively) among the 16 samples tested. Removing the outlier individual (34.88 Wolbachia per host cell) from the analyses did not change the statistical outcome of the correlation test in WORiC (Pearson correlation; p > 0.7).

The genome of the WORiC prophage is predicted to be 77,261 bp containing 56 ORFs [WRi _006570 to WRi_007250]. The core genome containing a DNA packaging and head assembly module and a tail morphogenesis module is shown in table 2 and is 24.2 kbp [WRi_006910 to WRi_007210]. The 35% GC content is identical to the GC content of the wRi genome indicating a long period of co-evolution between prophage and bacteria.

When temperate bacteriophages infect sensitive bacteria their viral genomes direct DNA replication of the phage, cell lysis and the release of progeny, or, if the lytic state is suppressed, they integrate into the bacterial chromosome in the form of a prophage, in what is known as the lysogenic state. In Escherichia coli, lambdoid prophages are stably integrated into the host chromosome and do not undergo lytic induction until the bacterial SOS response is activated 27 . Gavotte et al 17 used a filtration-based purification method accompanied by TEM and ORF7-specific PCR to show that mature phage particles form in Wolbachia-infected tissues in both D. simulans and D. melanogaster, but the specific identity of these virus particles and the regulation of their induction was not addressed.

In this study, the activity of the three distinct prophages found in wRi infecting D. simulans was measured using quantitative PCR. Phage type-specific primers were used to determine how many copies of the phage genomes were present in addition to the integrated forms. The only phage chromosome to appear in excess of the integrated copy number was WORiC. The average number of copies of WORiC in all tissues tested ranged from 1.29 - 1.61 copies per Wolbachia, consistently above the one copy integrated into the wRi genome. Thus, WORiC appears to be the only actively replicating phage in D. simulans.

wRi is considered to be a high CI strain of Wolbachia in D. simulans; embryonic lethality resulting from crosses between infected males and uninfected females is typically between 90 - 100% 28 29 . In N. vitripennis infected with wVitB, which is also a high CI-inducing strain of Wolbachia, Bordenstein et al 15 reported an average WOVitB copy number of 1.6 ± 0.12 per Wolbachia. In the present study, a similar relative density of WORiC suggests that this phage is the active virus observed in past TEM micrographs of Drosophila tissues 5 17 . WORiC genes have been reported as actively transcribed in previous literature. Specifically, the ankyrin related genes in WORiC are expressed in males, females, ovaries, testes, early (2 hour AEL) and late (overnight) embryos 4 .

WORiA and WORiB did not show any evidence of extrachromosomal DNA beyond the one and two copies, respectively, found within the wRi genome. Alignments to WOCauB and WOVitA1 show that both WORiA and WORiB lack the core structural components necessary for virion assembly. The persistence of WORiA and WORiB within the wRi genome suggests that there may be selective pressures maintaining these two prophages. There is evidence that WORiB is actively transcribing at least one ORF located within the prophage genome 30 and so this region may be necessary for another, unrelated, aspect of Wolbachia biology.

The average density of WORiC was derived from pooled samples of multiple individuals and tissues. When 16 third instar larvae were individually measured for phage density, WORiA and WORiB did not significantly deviate from the expected means of one and two copies, respectively. Individual larva, however, had a much wider distribution of WORiC copy numbers, ranging from individuals that appeared to have no extrachromosomal viruses to individuals having more than 1.5 WORiC per Wolbachia. This indicates that not every individual within the larval population is experiencing viral replication, although most are. Currently, the signals which induce viral replication within the confines of an endosymbiotic bacterium are unknown.

Along with the WO density in individual third instar larvae, the relative Wolbachia wRi density per D. simulans host cell was also measured. The wRi density did not significantly correlate with WORiA, WORiB, or WORiC relative densities. However, the WORiC density trends toward a slight inverse association with wRi density. It is possible that with a larger sample population, more statistical significance would emerge. This lack of correlation does not refute the phage density model postulated by Bordenstein et al 15 , whereby the Wolbachia copy number and CI in N. vitripennis was found to be inversely related to phage activity. Rather, it raises the notion that phage density is a population and strain-specific factor. Low levels of replicating phage, as seen here for WORiC, may not significantly impact Wolbachia wRi density and the strength of CI in Drosophila. The effect of phage copy number on CI level in D. simulans has yet to be examined.

Since WORiC in this study was the only wRi prophage capable of extrachromosomal replication, a comparative genomic approach was taken to identify the core genome conserved between WORiC and two known temperate bacteriophages WOVitA1 and WOCauB2. This approach identified essential regions required for phage generation. The genomes of WORiC, WOVitA1, and WOCauB2 show considerable sequence homology which supports the view that WORiC is the active form of phage in wRi. In contrast, the WORiB genome and the WOMelB genome lacking the upstream pyocin region share few homologous sequences with WORiC. Genes with sequence homology in WORiB, WOMelB, and WORiC belong to the DNA packaging and head assembly region. However, the core structural/tail region of WORiC aligns with WOMelB once the pyocin region is included in the analysis. WORiB lacks the pyocin-like region and is therefore deficient in most tail morphogenesis genes.

The chimeric nature of WO phages was initially described by Masui et al 6 , who identified the large terminase subunit, portal protein and minor capsid protein of the packaging region in WOKue as lambda-like, and the baseplate assembly proteins of the structural region as P2-like. This hybridization of lambda and P2 sequences is not exclusive to WO phages, since chimeric phages have been described in other systems; for example Xylella fastidiosa phages XfP1 and XfP2 are also lambda/P2 chimeras 16 . Due to recombination and genetic mosaicism, different parts of a bacteriophage genome can have different evolutionary histories 31 . In the chimeric WO phages (figure 4), the large terminase subunit sequence from the DNA packaging and head assembly regions shows a different phylogenetic relationship than the baseplate assembly protein W sequence from the tail morphogenesis regions. This modular nature of WO phages has been described previously 19 .

The two conserved modules shared by WORiC and the temperate phages WOCauB2 and WOVitA1 include the DNA packaging and head assembly region and the tail morphogenesis region. The genome encoding the DNA packaging and head assembly module includes ORFs that putatively code for a portal protein, a minor capsid protein and the large subunit of the terminase protein. This large terminase subunit contains a DNA-dependent ATPase domain and site-specific nuclease domain which are both involved in DNA translocation during packaging. In double stranded DNA phages, terminases are generally accompanied by a small subunit involved in DNA binding 32 33 . However, no homolog of this small subunit has been identified in any WO genome. The portal protein of tailed bacteriophages forms a complex with the terminase proteins which translocates phage DNA into the prohead during phage replication 33 . The conservation of these packaging genes suggests that DNA packaging in WO phages is driven by an ATP-dependent DNA translocation motor similar to other tailed bacteriophages.

Similarly, the organization of the tail morphogenesis module is conserved among WOVitA, WOCauB, and WORiC. Genes involved in tail assembly include the tail proteins, tail tape measure protein, the tail sheath protein, the contractile tail tube protein and baseplate assembly proteins J,W, and V. Tail morphogenesis in the subfamily Myoviridae, which have long contractile tails, is the most complex of all tailed bacteriophages. In the Myoviridae T4, P2 or Mu, baseplate assembly occurs first and is required for sheath and tail polymerization. It is from the baseplate that the tube polymerizes to a length determined by the tail-tape measure protein and this is followed by the tail sheath which extends the length of the tail 34 .

The presence of the tail sheath gene in active WO genomes suggests that, with respect to tail structure and assembly, these phages are more similar to Myoviridae than to the subfamily Siphoviridae, which includes lambda and lacks contractile tails. The phage tail mediates genome delivery into host cells, and is required for the generation of infectious phages. The absence of this region in the WORiB genome may contribute to the inability of WORiB to form infectious particles.

Unlike WORiC, where the packaging region is located adjacent to the structural proteins, in WOMelB the structural proteins are divided in the genome and separated from the packaging region by approximately 18kbp 35 . One region of structural genes found in WOMelB was initially characterized as a pyocin-like region. Therefore, active phage generation in D. melanogaster wMel could result from the coordinate replication of both packaging and structural regions. Despite much previous interest in Wolbachia's ankyrin containing genes 35 36 , and the suggestion that they may influence phage function, the ORFs encoding ankyrin-containing motifs are outside the core conserved regions of WORiC, WOVitA1 and WOCauB3. The role of ankyrin coding genes in the WO-Wolbachia-host relationship remains elusive 37 38 .

Our results suggest that Wolbachia phages WORiC and known active phages WOCauB and WOVitA1 represent a conserved class of Wolbachia phages. Interest in the conserved genetic modules of the lambda-like DNA packaging and head assembly genes and P2-like tail morphogenesis genes led to the investigation of the relatedness of the Wolbachia phages. Phylogenetic analysis shows similarity between WORiC and WO-B's found in wMel and wRi (based on large terminase subunit phylogeny) and similarity between WORiC and WOCauB2 and WOCauB3 (based on the baseplate assembly protein W phylogeny). These divergent topologies are indicative of the horizontal transfer events occurring between phage genomes. Similarity of genomes of active WO phages may be due to the fact that they have a common, recent origin, or because active WO phages are operating within a limited framework of endosymbiotic bacteria, where opportunities for incorporating novel gene sequences by recombination are limited. Given the present level of knowledge of active WO bacteriophages, we cannot distinguish between these and other possible evolutionary scenarios.