Dietary saccharides and sweet tastants have differential effects on colonization of Drosophila oocytes by Wolbachia endosymbionts

Metazoan organisms are increasingly recognized as communities of prokaryotic and eukaryotic cells. Symbiotic interactions within the collective unit of an organism range from mutualistic to parasitic (Dale and Moran, 2006). Endosymbiotic Wolbachia bacteria are unique in that they occupy a wide range of the symbiotic spectrum. Wolbachia are Alphaproteobacteria that reside within the cells of mites, crustaceans, filarial nematodes (Werren et al., 2008) and approximately 52% of all insect species based on a handful of typing loci (Weinert et al., 2015). At least 470 distinct Wolbachia strains have been reported to date (Baldo et al., 2006). Of those, some are reported to provide essential cofactors to the host (Ghedin et al., 2007; Hosokawa et al., 2010; Nikoh et al., 2014), promote host reproduction (Dedeine et al., 2001; Landmann et al., 2011; Starr and Cline, 2002) and protect the host from lethal RNA viruses (Chrostek et al., 2013; Hedges et al., 2008; Martinez et al., 2014; Teixeira et al., 2008). Conversely, the wMel^Pop Wolbachia variant lyses brain cells and shortens insect lifespan (Min and Benzer, 1997). This positions Wolbachia as a uniquely informative system for elucidating the cellular mechanisms of symbiosis.

A consensus requirement for Wolbachia success across diverse hosts is robust vertical transmission. Though Wolbachia occupy the germline stem cells (GSC) of male and female hosts, removal of the bacteria during spermatogenesis creates a ‘dead end’ with respect to transmission (Bressac and Rousset, 1993; Serbus et al., 2008). Thus, persistence of Wolbachia in maternal germline cells is of critical importance for transmission to progeny. In the Drosophila melanogaster model system that naturally carries wMel Wolbachia (O'Neill et al., 1992; Riegler et al., 2005), the GSC are infected with these bacteria. This ensures that differentiating daughter cells (cystoblasts) inherit Wolbachia during mitosis (Ferree et al., 2005; King, 1970; Serbus et al., 2008). While the cystoblast undergoes mitosis to generate an interconnected cyst of 16 germline cells, Wolbachia exiting the nearby somatic cell niche also invade the germline cyst (Toomey et al., 2013). After the cyst is coated with a blanket of somatic follicle cells, creating a unit referred to as an egg chamber (King, 1970), additional horizontal invasion events may also occur (Casper-Lindley et al., 2011). Wolbachia also replicate to populate the germline cells of the egg chamber, including the oocyte cell that ultimately takes over to form a completed egg (King, 1970; Serbus et al., 2011). Similar germline loading mechanisms are expected to apply to other Wolbachia-Drosophila combinations, with differential contributions to germline colonization by GSC loading and horizontal invasion in each case (Toomey et al., 2013).

Maternal transmission relies upon sufficient Wolbachia titer within the germline cells. One strategy of Wolbachia transmission in embryogenesis is the use of mass action to promote inclusion of bacteria in embryonic germline cells (Veneti et al., 2004). A complementary strategy to facilitate bacterial transmission is through strategic subcellular localization (Breeuwer and Werren, 1990; Hadfield and Axton, 1999; Rasgon and Scott, 2003; Stouthamer et al., 1993; Veneti et al., 2004; Zchori-Fein et al., 1998). In D. melanogaster, the host microtubule motor proteins Dynein and Kinesin-1 act sequentially to elevate Wolbachia concentration at the oocyte posterior cortex (Ferree et al., 2005; Serbus and Sullivan, 2007). This is followed by association of Wolbachia with a cortical mixture of components referred to as pole plasm (Ashburner, 1989; Riechmann and Ephrussi, 2001; Serbus and Sullivan, 2007). This positions the bacteria for envelopment by embryonic germline cells specified by the pole plasm (Ashburner, 1989; Hadfield and Axton, 1999; Serbus and Sullivan, 2007). Maternal Wolbachia transmission rates documented in D. melanogaster are near 97% in the field (Hoffmann et al., 1998) and 100% in the lab (Turelli and Hoffmann, 1995), indicating this maternal transmission strategy is effective.

The molecular mechanisms that regulate Wolbachia titer are not well understood. Body-wide Wolbachia titer has been reported to vary up to 180,000-fold in lab-reared offspring of mosquitoes collected from nature (Ahantarig et al., 2008), and 20,000-fold between wild-caught Drosophila innubila individuals (Unckless et al., 2009). This titer variation may be due in part to sensitivity to host temperature (Bordenstein and Bordenstein, 2011; Mouton et al., 2006, 2007; Wiwatanaratanabutr and Kittayapong, 2009, 2006), host crowding (Hoffmann et al., 1998; Wiwatanaratanabutr and Kittayapong, 2009), host genetic background (Boyle et al., 1993; Poinsot et al., 1998; Veneti et al., 2004; Serbus et al., 2011) and host age (Tortosa et al., 2010; Unckless et al., 2009).

A set of studies has particularly highlighted the impact of diet on Wolbachia titers in vivo, implicating roles for dietary cholesterol (Caragata et al., 2013) and other macronutrients (Ponton et al., 2015). It was recently shown that dietary yeast, known to trigger insulin signaling in Drosophila (Géminard, 2009 #1357; Teleman, 2010 #1333), suppresses Wolbachia titer in developing oocytes (Serbus et al., 2015 #1785). By contrast, dietary sucrose, which is expected to induce insulin resistance in Drosophila (Broughton et al., 2010; Morris et al., 2012; Musselman et al., 2011; Norseen et al., 2012; Pasco and Léopold, 2012; Yang et al., 2005), led to elevated oocyte titers (Serbus, 2015 #1785). Understanding how diet affects oocyte Wolbachia titer is expected to inform the mechanisms supporting Wolbachia colonization of host cells and ultimately, Wolbachia transmission. To address the mechanisms underlying Wolbachia titer control in oogenesis, an array of structurally diverse dietary sugars and sweet tastants was selected, and their impact on colonization investigated as described below.

Prior results indicated that yeast-enriched food reduces Wolbachia titers in oogenesis, whereas sucrose-enriched food elevates oocyte Wolbachia titer (Serbus et al., 2015). To confirm this, two-day-old flies were exposed to yeast- and sucrose-enriched food for 3 days (Fig. 1E, Table 1; Table S1). The ovarian tissues were dissected, fixed, stained with propidium iodide and imaged by confocal microscopy. Each punctate nucleoid that is labeled by the DNA stain is interpreted as representing a single bacterium. The resulting images suggested overall more Wolbachia puncta in the sucrose condition, and fewer Wolbachia puncta in the yeast condition (Fig. 2A-C). For finer resolution of oocyte Wolbachia titer, Wolbachia were quantified from representative oocyte focal planes, and analyzed by Kruskal–Wallis ANOVA. According to these criteria, oocyte Wolbachia titer in the yeast-enriched condition was significantly lower than the control [χ²(2)=27.3, P<0.001] (Fig. 2D). Though higher oocyte Wolbachia titers were detected in the sucrose-enriched condition, the values did not differ significantly from the control [χ²(2)=15.6, P=0.056]. Significant oocyte titer differences were detected between the yeast and sucrose conditions, however, with sucrose exhibiting a 577% higher median titer value than yeast [χ²(2)=42.9, P<0.001] (Fig. 2D). Overall, this outcome corroborates opposing effects of dietary sucrose and yeast on oocyte Wolbachia titer.

A surprising finding from prior work was that food enriched in the monosaccharide constituents of sucrose, namely glucose and fructose, did not recapitulate high oocyte titer responses analogous to sucrose (Serbus et al., 2015). This raised questions as to whether any monosaccharide is capable of affecting oocyte Wolbachia titer. To test this, two-day-old flies were collected and exposed to galactose-enriched food for 3 days (Fig. 1A, Table 1; Table S1). Wolbachia quantification indicated that galactose-fed flies carried significantly more Wolbachia than control oocytes [χ²(1)=18.2, P<0.001] (Fig. 2E). This indicates galactose to be the first dietary monosaccharide capable of elevating oocyte Wolbachia titer.

To test the extent to which other dietary disaccharides affect oocyte Wolbachia titer, flies were exposed to lactose-, maltose- and trehalose-enriched foods (Fig. 1B-D, Table 1; Table S1). These treatments elevated oocyte Wolbachia titer, with magnitude increasing from lactose to maltose to trehalose (Fig. 2F). Oocyte Wolbachia titers in disaccharide enriched conditions were also identified as significantly different from the control [lactose χ²(3)=25.0, P=0.002; maltose χ²(3)=28.8, P<0.001; trehalose χ²(3)=39.0, P<0.001] (Fig. 2F). This indicates that exposure to a range of disaccharide-enriched diets increases oocyte Wolbachia titer.

To consider the basis for oocyte titer changes, oocyte size was tested. Specifically, the two-dimensional area of every oocyte image used for titer quantification above was measured. As all sample compression and oocyte focal plane selection were standardized for each experiment, the resulting area values are a proxy estimate for oocyte size. According to this analysis, no significant differences in oocyte area were identified between control and sucrose-enriched conditions [χ²(2)=12.2, P=0.085], nor control and yeast-enriched conditions [χ²(2)=6.6, P=0.811] (Fig. 2G). Significance was detected when comparing oocyte area values between sucrose- and yeast-enriched conditions [χ²(2)=18.7, P=0.004]. The median area of sucrose-treated oocytes was 84% of the yeast condition (Fig. 2G), in contrast to the 577% disparity between median oocyte Wolbachia titers in these conditions (Fig. 2D). Oocyte area did not differ significantly between control and galactose-fed oocytes [χ²(1)=1.27, P=0.26] (Fig. 2H), nor between control, lactose-, maltose-, and trehalose-fed oocytes [χ²(3)=6.72, P=0.083] (Fig. 2I). Thus, oocyte area did not parallel the significantly higher oocyte Wolbachia titer responses to natural saccharides (Fig. 2E,F). This suggests that oocyte size changes are not responsible for sugar-induced increases in oocyte Wolbachia titer.

D. melanogaster ovary size is responsive to nutritional conditions. Through apparent impacts on systemic insulin signaling, sucrose-rich diets have been shown to reduce ovary size, whereas yeast-rich diets increase it (Fig. 3A) (Geminard et al., 2009; LaFever and Drummond-Barbosa, 2005; Morris et al., 2012). Direct measurement of ovary volume in response to these diets confirms that the size changes are substantial (Fig. 3B). Ovary volumes were significantly different between control and sucrose-fed oocytes [χ²(2)=29.1, P<0.001], and control and yeast-fed oocytes [χ²(2)=30.2, P<0.001] as well as sucrose- and yeast-fed oocytes [χ²(2)=59.3, P<0.001] (Fig. 3B). As dietary sucrose and yeast exert opposite impacts on ovary volume and Wolbachia titer, these data open the possibility that oocyte titer reflects ovary size.

To further assess the relationship between ovary size and oocyte Wolbachia titer, ovary volume was assessed across sugar-enriched dietary conditions. This analysis indicated consistently small ovary volumes for galactose-, lactose-, maltose- and trehalose-fed oocytes [χ²(4)≥57.0, P<0.001 for all] (Fig. 3C). As these conditions significantly elevated oocyte Wolbachia titers (Fig. 2E,F), this outcome is consistent with oocyte Wolbachia titer as an inverse correlate of ovary size. Furthermore, median ovary volumes ranged from 25% of control in the maltose condition, to 36% of the control in the galactose condition (Fig. 3C). This is analogous to size reductions seen in sucrose-fed ovaries (39% of the control) (Fig. 3B). Taken together, the data suggest that sugar-enriched diets generally lead to ovary size reduction.

To test whether oocyte titer is selectively responsive to natural dietary sugars, an array of other sweet tastants was tested (Table 1; Table S1). Unlike natural sugars, the synthetic disaccharide lactulose (Fig. 1F) is reportedly indigestible by eukaryotes (Schmidl and Labuza, 2000). Flies exposed to lactulose-enriched diets did not exhibit any significant oocyte titer difference from the control [χ²(1)=1.57, P=0.211] (Fig. 4A). The impact of the sugar alcohols, erythritol and xylitol, as well as the artificial sweeteners aspartame and saccharin, were also tested (Fig. 1G-J). Oocyte Wolbachia titer was significantly different between xylitol and erythritol conditions [χ²(4)=28.9, P=0.034] as well as xylitol and saccharin conditions [χ²(4)=27.6, P=0.040] (Fig. 4B). However, no significant differences were evident when comparing the sweet tastant treatments against the control [χ²(4)≤26.5, P≥0.087 for all] (Fig. 4B). These data suggest that properties outside of taste recognition are responsible for sugar-driven increases in oocyte Wolbachia titer.

To consider the basis for the oocyte titers observed in sweet tastant conditions, oocyte and ovary sizing were also examined. No significant changes in oocyte area were detected in response to lactulose [χ²(1)=3.15, P=0.076], (Fig. 4C) nor artificial sweeteners and sugar alcohols [χ²(4)=1.75, P<0.782] (Fig. 4D). However, consistently small ovary size was detected in response to all sweet tastant treatments, with median ovary volumes ranging from 17% of the control for saccharin [χ²(2)=50.7, P<0.001] to 36% of the control for xylitol [χ²(2)=35.8, P<0.001] (Fig. 4E-G). Thus, ovary volume reduction associated with sweet tastants parallels that induced by sugar-enriched diets. However, as sugar-enriched diets elevate oocyte Wolbachia titer and sweet tastants do not, this indicates that oocyte Wolbachia titer is not specified exclusively by ovary size.

To further investigate the basis for sugar-associated oocyte titer increases, candidate hypotheses were pursued. During food preparation, the sugar solutions were distinctively thick in consistency compared to other treatments. The apparently hygroscopic properties of the sugars opened the possibility that they may act as a desiccant after ingestion. To test the impact of desiccation on oocyte Wolbachia titer, flies were exposed to standard fly food containing dehydrating silica gel in a 2:1 volumetric ratio. After 3 days of exposure, samples were examined. Analysis of oocyte Wolbachia titer indicated that fewer Wolbachia were carried by the silica gel condition, with the median oocyte titer value at 81% of the control [χ²(1)=4.95, P=0.026] (Fig. 5A). Oocyte size assessment indicated that oocyte area was significantly larger in the silica gel condition, with the median area value at 112% of the control [χ²(1)=4.09, P=0.043] (Fig. 5B). No differences in ovary size were observed between control and silica gel conditions [χ²(1)=0.056, P=0.813] (Fig. 5C). These data suggest that desiccation impacts on Wolbachia and oogenesis are entirely distinct from that of the dietary sugars. Thus, desiccation is not responsible for the titer-increasing effects of dietary saccharides.

Another possibility is that titer-elevating sugars generally affect oocyte Wolbachia titer through impact on core nutritional signaling processes. It was previously demonstrated that dietary yeast and sucrose exert opposite effects on oocyte Wolbachia titer in an insulin-dependent manner (Serbus et al., 2015). If dietary sugars are generally antagonistic to insulin signaling, one possibility is that they will rescue the impact of dietary yeast on oocyte Wolbachia titer. To test this, flies were exposed to diets dually enriched in yeast and dietary saccharides. An array of responses was evident (Fig. 6A). No dual feedings of yeast and sugar elevated oocyte titer significantly above control levels. Relative to the control, oocytes exposed to diets enriched in either yeast or yeast+trehalose showed significant depletion of Wolbachia [χ²(6)=75.5, P<0.001 and χ²(6)=47.9, P<0.021, respectively]. Yeast-fed oocytes showed the overall lowest titer levels, differing significantly from all dual yeast-sugar feeding conditions [χ²(6)≥59, P≤0.002 for all] except yeast+trehalose [χ²(6)=27.5, P=1.000]. The yeast+trehalose condition displayed significantly lower oocyte Wolbachia titer as compared to dual feedings of yeast+sucrose, galactose, or lactose [χ²(6)≥53.6, P≤0.008 for all] (Fig. 6A). Thus, the yeast+trehalose oocyte titer profile paralleled many of the outcomes associated with exposure to dietary yeast alone. These data overall indicate that dietary sucrose, galactose, lactose and maltose rescue oocyte titer suppression by dietary yeast, whereas dietary trehalose does not.

To investigate the basis for differential oocyte Wolbachia titer responses to yeast+sugar diets, sizing controls were also performed. No significant differences in oocyte area were observed between any of the feeding conditions used [χ²(6)=9.03, P=0.172] (Fig. 6B). By contrast, ovary size was responsive to nutrient-altered diets. Ovary volumes in all yeast+sugar feeding conditions were significantly lower than in the yeast-fed condition [χ²(6)≥67.5, P<0.001 for all cases] (Fig. 6C). Furthermore, no significant ovary volume differences were observed between the control and any of the yeast+sugar dual feeding conditions [χ²(6)≤36.8, P≥0.397 for all]. This indicates that dietary sugars consistently rescued yeast-driven ovary enlargement. Some variation was detected in the extent of ovary size modification by dual yeast+sugar diets. Ovary volumes in the yeast+galactose condition were distinguished as significantly greater than the yeast+lactose condition [χ²(6)=57.6, P<0.005] and the yeast+trehalose condition [χ²(6)=49.3, P<0.035] (Fig. 6C). As these trends do not parallel oocyte titer outcomes in a consistent manner, the implication is that dietary sugars affect ovary development and Wolbachia colonization dynamics through mechanisms that are at least partially independent.

The impact of diverse dietary sugars on insulin signaling has not been fully defined in D. melanogaster. From the perspective of Wolbachia endosymbiosis, this study suggests that dietary sugars induce different classes of mechanistic responses. This work showed that D. melanogaster diets enriched in galactose, lactose, maltose and trehalose significantly elevated oocyte Wolbachia titer. As oocyte size was also unaffected by most dietary sugar treatments, all observed titer increases are interpreted to represent true elevation of bacterial quantity per oocyte and not a concentration artifact of cell size. No titer-related trends were evident in terms of reported caloric value (Table S1), gustatory preferences (Stafford et al., 2012), nor the magnitude of neural responses to single tastants (Freeman et al., 2014). These outcomes are most readily reconciled with the structural content of the sugars, as maltose and trehalose are both glucose disaccharides, and lactose contains galactose as one of its major constituents. The absence of titer-elevating effects by lactulose is consistent with a possible requirement for glucose as a constituent of titer-influencing disaccharides.

Though these sugars were selected for analysis due to their structural features, D. melanogaster may reasonably encounter some of these sugars in nature. Maltose is reportedly the major starch breakdown product released from chloroplasts at night (Niittayla et al., 2004; Wiese et al., 2004) and is commonly associated with starchy plant products such as grains (Halford et al., 2011). Natural exposure to trehalose is less likely, as it is carried at low levels in higher plants, serving as a signaling cue (Grennan, 2007; Lunn et al., 2014). Galactose is a core component of raffinose-containing oligosaccharides that are widespread in higher plants (Sengupta et al., 2015) and carried within dozens of fruits and vegetables (Gross and Acosta, 1991). Though lactose is not expected to appear in a natural D. melanogaster diet, one highly speculative possibility is that lactose digestion by microbes occupying the food vial or gut microbiome releases galactose, ultimately inducing titer responses. Future analyses of sugar uptake by D. melanogaster are needed to further inform the relevance of ingested doses. At this time, only two (Meyer et al., 2011; Wang and Wang, 1993) of the 26 predicted sugar transporters in D. melanogaster (Adams et al., 2000; Marygold et al., 2016) have been characterized. However, 17 of the predicted transporter proteins are expressed by females and detected in digestive tissues, and are thus potentially relevant for consideration (Table S2).

Another finding of this study was that ovary size responses to dietary cues did not consistently correspond to oocyte titer changes. As ovary size and oocyte Wolbachia titer are oppositely affected by sugar- and yeast-enriched diets, the simplest interpretation is that spatial re-allocation of Wolbachia within the ovary is responsible for increased oocyte titers. This hypothesis invokes horizontal Wolbachia invasion between cells of the ovary as influential in oocyte colonization by Wolbachia. Invasion has been reported to contribute to Wolbachia colonization of the distal tip of the Drosophila ovary (Toomey et al., 2013), early stages of Drosophila oogenesis (Casper-Lindley et al., 2011) and ovarian cells of mosquitoes and nematodes (Hughes et al., 2012; Landmann et al., 2012). Sugar alcohol and artificial sweetener treatments deviated from this invasion paradigm, as their reductions of ovary size were not paralleled by oocyte titer increases. Thus, ovary size may modulate oocyte colonization by Wolbachia titer in some cases, but the data argue against ovary size as a binary predictor of oocyte titer. One possibility is that changes in bacterial loading and replication in the germline offset changes in the extent of horizontal invasion. Reduced oocyte titers seen in xylitol-fed oocytes suggest that antibiotic properties of this sugar alcohol (Katsuyama et al., 2005; Renko et al., 2008) may exaggerate such tendencies. Another possibility is that dietary saccharides and sweet tastants each alter ovary size in a different manner. Ovary size reflects the number of productive ovarioles as well as the rates of egg chamber production, egg development and egg laying by each female (King, 1970). At this time, it cannot be ruled out that the physical basis for small ovary size may differ between saccharide and sweet tastant conditions.

This study further sought to address the mechanistic basis for saccharide impacts on oocyte Wolbachia titer. One possibility was that the concentrated sugar additives may increase oocyte Wolbachia titer as an indirect consequence of ovarian responses to desiccation. However, dietary desiccation tests showed reduction of oocyte Wolbachia titer rather than an increase, suggesting that sugar-based titer responses are unrelated to hydration. Another formal possibility is that Wolbachia responsiveness to dietary sugars is due to uptake of these sugars and/or their derivatives after ingestion. Though Wolbachia are predicted to encode a single hexose phosphate transporter, WD_0619, homologous to GlpT/PgpT/UhpT of Escherichia coli (Fann and Maloney, 1998; Kadner et al., 1992), there is no information to suggest the Wolbachia homolog of this transporter is sufficiently permissive to take up the diverse dietary sugars analyzed in this study.

The impact of dietary sugars on oocyte Wolbachia titer is currently best explained through nutritional impacts on the host. Dietary yeast is expected to activate multiple nutritional signaling branches, including insulin signaling, that converge upon activation of the mTORC1 kinase complex (Geminard et al., 2009; Teleman, 2010). Prior work showed that chemical inhibition of mTORC1 increased oocyte Wolbachia titer analogous to dietary sucrose, while loss of mTORC1 suppression lowered oocyte titers (Serbus et al., 2015). These findings and others implicated insulin as a suppressor of oocyte Wolbachia titer and inherently suggested that yeast-associated phenotypes should be ameliorated by dietary sugars. This study showed that dietary sugars did suppress yeast-associated ovary enlargement across the board, consistent with such a prediction. However, oocyte Wolbachia titers showed a range of responses, with trehalose exerting no impact on yeast-driven titer depletion, whereas sucrose, galactose, lactose and maltose restored oocyte titer to control levels. The disparity is surprising, as trehalose-enriched diets elicited the largest recorded oocyte titer increase to date. Examination of ovary size further indicated that galactose rescue of yeast-driven ovary enlargement was significantly less effective than lactose and trehalose. These findings suggest that saccharide treatments, all singly capable of elevating oocyte Wolbachia titer, may exert distinct functional impacts on oocyte Wolbachia titer. Integrated quantitative analyses will play an important role going forward in elucidating the mechanisms of oocyte colonization by Wolbachia.

The standard food used in this study is based upon a recipe by the Bloomington Drosophila Stock Center (http://flystocks.bio.indiana.edu/Fly_Work/media-recipes/bloomfood.htm). Our fly food was prepared in large batches that consisted of 20 liters water, 337 g yeast, 190 g soy flour, 1325 g yellow corn meal, 96 g agar, 1.5 liters Karo light corn syrup and 94 ml propionic acid. This standard food was used as a base for all nutrient-altered foods that were prepared in this study (Table S3). The sugar-enriched foods were prepared by first making a stock sugar solution of 20 g sugar in 10 ml ddH₂O, solubilized with rounds of 15 s in the microwave and then stirring, repeated until the sugar dissolved. 1.5 ml amounts of these sugar solutions were immediately mixed with 3.5 ml of melted standard food. As aspartame, erythritol, saccharin, and xylitol were not uniformly soluble, the sweetener-enriched foods were generated through direct addition of powder equivalents directly into 5 ml of melted standard food to a final concentration of 1 M (Table S3). Yeast-enriched food was prepared by mixing 1.5 ml of heat-killed yeast paste into 3.5 ml melted standard food. Dually enriched food was prepared through addition of 1.5 ml sugar solution and 1.5 ml heat-killed yeast to 2 ml standard food. Desiccated food was prepared by addition of 2.5 g silica gel (roughly 2.5 ml volume) to vials containing 5 ml standard food (Table S3). To ensure homogeneous suspensions of nutrient-altered diet preparations, all food vials were immediately transferred to an ice bucket to be cooled with additional stirring every 10 min until the food completely solidified. Kimwipe strips were inserted into the food to wick away excess moisture.

All feeding experiments were done using flies of the genotype w; Sp/Cyo; Sb/Tm6B, reared on standard food and in a controlled, 25°C environment. This stock carries the wMel Wolbachia strain as confirmed previously (Christensen et al., 2016). 0−24-hour-old adult flies were selected at random and transferred into new bottles of standard food and aged for 2 days. Then flies were transferred to vials of nutrient-altered food and incubated for 3 days. Controls were run in parallel with all treatment conditions in all experiments.

Ovarian tissues were dissected in PBS and fixed in 2% formaldehyde for 20 min as previously described (Serbus et al., 2015). The ovaries were rinsed with PBS-Triton 0.1% (PBS-T), incubated overnight in 10 mg ml⁻¹ RNAse, and rinsed extensively with PBS-T the next day. Then tissues were infused with 70% glycerol that contained 0.015 mg ml⁻¹ propidium iodide, and mounted on a slide. All replicates were imaged by laser scanning confocal microscopy on either Leica SP2 or an Olympus FV1200 confocal microscope at 63× magnification with 1.5× zoom. The Z-height of oocytes on each slide was standardized against the control slide for each replicate. Z-series images were acquired from randomly selected stage 10 egg chambers at 1.5 μm intervals. Uniform intensity settings were applied to all egg chambers imaged in each replicate.

To quantify oocyte Wolbachia titer, stacks of confocal images were examined to identify the deepest possible focal plane where Wolbachia are clearly visible across all samples of the replicate (Serbus et al., 2015). Images were manually processed in Photoshop to remove extraneous signal outside the oocyte, and remaining oocyte puncta were quantified using the Analyze Particles feature in Image J version 2.0.0-rc-43/1.51d (NIH). Thus, these data quantify the Wolbachia titer carried within one focal plane of each oocyte. This has been verified as a representative measure for comparing oocyte Wolbachia titer between different conditions (Serbus et al., 2011). Three or more experimental replicates were performed for all treatment conditions examined. Significance of differences between conditions was determined by ANOVA analysis of the raw data.

For measurement of oocyte area, the same representative oocyte focal planes used for Wolbachia titer assessment were re-analyzed. Oocytes were manually outlined in Microsoft PowerPoint, and the resulting two-dimensional shapes flood-filled in with color. Screen shots of these ovary fill diagrams were then imported into Fiji (Image J version 2.0.0-rc-43/1.51d, NIH) for conversion into 8-bit, thresholded black and white images. The area of the ovary fill diagrams was determined in terms of pixels² by the Analyze Particles function in Fiji. A scale bar was also used to calculate a pixel² to micron² ratio (9.3025:1) that was applied to all oocyte area data, for presentation and discussion purposes only. Statistical differences were determined through analysis of the primary data in terms of pixel² units.

For measurement of ovary volume, tissues were dissected from adult flies and imaged using an AmScope MD500 5.0 megapixel digital Camera mounted upon a Jenco ST-F803 dissection microscope set at 1× magnification. The pixel length and width of each ovary was assessed with the ‘Measure’ tool in Fiji. These values used to approximate ovary volume using the standard ellipsoid formula for volume; V=4 3πabc⁻¹, where a=½ the length and b and c=½ the width. Three biological replicates were assessed for all treatment conditions. The area of a reference object was measured to determine the pixel to mm ratio (148.62:1) appropriate for describing the volumetric data. This conversion was applied in the context of presentation and discussion purposes only. All statistical analyses of ovary volumes were based upon primary data in terms of the pixel³ units.

All data in this study were analyzed with the IBM SPSS statistics program, v.23. The descriptive statistics function was used to analyze the distribution of the data. According to the residuals as well as metrics such as skewedness and kurtosis, the current data did not fit a normal distribution. However, the data met the assumptions of the two-tailed Kruskal–Wallis ANOVA, and thus this test was systematically applied for data analysis. Variation within each experimental condition is indicated by boxplot format used to graphically display the data. Post hoc data presented here were generated by SPSS as standard outputs of the analysis, including the adjusted P-values reported throughout the manuscript. Though subtle differences in oocyte titer, oocyte area and ovary volume may not be detected by this analysis framework, the data empirically demonstrated that the sample ‘n’ for these experiments was sufficient to identify clear cases of significance.

We thank A. J. M. Zehadee Momtaz, Jessica Pintado Silva, Steen Christensen, Pamela White, Christopher Chin, Yongjun Huang, M. Alejandro Barbieri and Erasmo Perera for their assistance and helpful discussions.