Wolbachia infection at least partially rescues the fertility and ovary defects of several new Drosophila melanogaster bag of marbles protein-coding mutants

The D. melanogaster protein coding gene bag of marbles (bam) plays a key role in early male and female reproduction by forming complexes with partner proteins to promote differentiation in gametogenesis. Like another germline gene, Sex lethal, bam genetically interacts with the endosymbiont Wolbachia, as Wolbachia rescues the reduced fertility of a bam hypomorphic mutant. Here, we explored the specificity of the bam-Wolbachia interaction by generating 22 new bam mutants, with ten mutants displaying fertility defects. Nine of these mutants trend towards rescue by the wMel Wolbachia variant, with eight statistically significant at the fertility and/or cytological level. In some cases, fertility was increased a striking 20-fold. There is no specificity between the rescue and the known binding regions of bam, suggesting wMel does not interact with one singular bam partner to rescue the reproductive phenotype. We further tested if wMel interacts with bam in a non-specific way, by increasing bam transcript levels or acting upstream in germline stem cells. A fertility assessment of a bam RNAi knockdown mutant reveals that wMel rescue is specific to functionally mutant bam alleles and we find no obvious evidence of wMel interaction with germline stem cells in bam mutants.


Introduction
The bag of marbles (bam) gene plays a critical role in female and male gametogenesis in D. melanogaster.In the stem cell niche of the ovaries and testes lies a collection of germline stem cells (GSCs) that asymmetrically divide to self-renew and produce differentiating daughter cells.In females, bam acts as a master regulator and its expression induces daughter cell differentiation [1].In males, bam functions later in gametogenesis where it initiates terminal differentiation of the spermatogonia that eventually become sperm [1][2][3][4].
In female D. melanogaster, bam is largely repressed in GSCs by signaling it receives from the stem cell niche [5][6][7].Bam that escapes silencing in GSCs is antagonized by eukaryotic translation initiation factor 4A (eIF4A) [8].When a GSC divides, a daughter cell is displaced from the stem cell niche, allowing bam to now be expressed.In the daughter cell, Bam functions to bring together various proteins that repress self-renewal and promote differentiation, leading to the establishment of a cystoblast.
In both male and female D. melanogaster, bam null mutants are completely sterile and have a tumorous cytological phenotype [1,9].A partial loss-of-function mutant (bam bw ) has reduced fertility in females and is sterile in males [10,11].In females, the fertility defect of the bam bw mutant is rescued by both wMel and wMelCS variants of the endosymbiont bacteria Wolbachia pipientis [11,12].
W. pipientis, commonly referred to as Wolbachia, is estimated to inhabit 50% or more of insect species and is an obligate bacterium in filarial nematodes [13][14][15].Across both arthropods and nematodes, Wolbachia manipulates reproduction to promote its own maternal transmission (reviewed in [16,17]).For example, Wolbachia famously induces cytoplasmic incompatibility, in which a female's progeny may have reduced viability if they do not harbor the same Wolbachia strain as the male parent.Wolbachia is also known to induce parthenogenesis, feminization, and male-killing of host progeny, all of which promote its maternal transmission.The extent of these parasitic phenotypes varies across Wolbachia variants and their host identity, age, and environment, e.g.[18][19][20].Rescuing the fertility defect of the hypomorphic bam mutant aligns with the evolutionary interests of Wolbachia.
Wolbachia also rescues the sterility of hypomorphs of another GSC gene, Sex lethal (Sxl), in D. melanogaster (Wolbachia variant unspecified) [21].It has thus been proposed that the rescue is dependent on the point mutation induced in the Sxl mutant and not on the severity of the mutant, though the researchers suggest that rescue cannot happen if Sxl protein levels are too low.Of note is that the Sxl mutants they studied were in different genetic backgrounds, which may have affected the results observed and the conclusions made.Such background effects were observed with bam in [22], where females with the same L255F bam mutation had strikingly different fertility across two different genetic backgrounds of D. melanogaster.Nevertheless, further research into one of the Sxl mutants, Sxl f4 , showed that the rescue is partially due to a restoration of GSCs by the Wolbachia protein toxic manipulator of oogenesis (TomO) [23].

PLOS GENETICS
Mutants of additional germline genes meiotic P26 (mei-P26), ovarian tumor (otu), and sans fille (snf) are not rescued by Wolbachia, despite having similar severity of an egg-absence phenotype to some Sxl mutants (Wolbachia variant unspecified) [21].These results suggest the Wolbachia interaction may be specific to a functional pathway in female D. melanogaster, possibly through a direct, physical interaction with a gene product like Sxl that is expressed in the germline.
In this study, we explore the bam-wMel interaction, building off the findings for the bam bw hypomorphic mutant.The bam bw hypomorphic allele is a L255F change in a gene region critical for bam function.Thus far, four functional regions of bam have been documented in D. melanogaster: The CAF40 binding motif (residues 13-36) [23], the Benign gonial cell neoplasm (Bgcn) binding region (residues 100-350) [24], the COP9 signalosome 4 (Csn4) binding region (residues 151-350) [24], and a ubiquitin binding region (residues 200-250) [25].Bam also has a PEST domain (residues 402-434), which marks the protein for rapid degradation [1].Bam's specific binding regions and their interactions with binding partners are illustrated in Fig 1a and 1b respectively.To repress GSC self-renewal, Bam forms a complex with Bgcn, Meiotic-P26 (Mei-P26), Brain tumor (Brat), a deadenylase, and either Argonaute-1 (Ago1) or Sxl to destabilize nanos (nos) mRNA [26][27][28].The destabilization of nos mRNA may occur through Bam's interaction with the CAF40 which is part of the CCR4-NOT deadenylase complex [23,29].Thus, in GSCs, nos mRNA is translated to promote GSC self-renewal; in differentiating daughter cells, nos mRNA is destabilized by the Bam complex, inhibiting nos translation and the self-renewal pathway.
Bam also represses the self-renewal pathway through its interaction with Csn4 [24].In GSCs, the COP9 complex is composed of nine Csn subunits.When Bam is expressed in differentiating daughter cells, it breaks up the complex by sequestering Csn4, thereby inhibiting the GSC self-renewal pathway.
Additionally, Bam promotes differentiation by forming a complex with Otu to deubiquitinate Cyclin A (CycA) [25].In GSCs, CycA is degraded through a process involving the ubiquitin-conjugating enzyme Effete [30].In differentiating GSC daughter cells, it is proposed that Bam binds to both Otu and to proteins such as CycA, which have ubiquitin chains that mark them for degradation.When in a complex with Bam, Otu deubiquitinates CycA, stabilizing CycA levels and thus promoting GSC daughter cell differentiation.
Here, we have generated 22 new bam mutants across the different binding and functional regions of bam.We analyzed the genetic interaction between these mutants and the wMel variant of Wolbachia to better understand the extent of the bam-wMel interaction.We are partly motivated by the observation of rapidly diversifying amino acids at bam across the Drosophila phylogeny [31][32][33].Bam appears novel to the Drosophila genus [30], and we have previously hypothesized that Wolbachia could contribute to the rapid evolution observed at bam due to the observed bam-Wolbachia genetic interaction and the fact that both Wolbachia infection and positive selection at bam occur intermittently throughout the Drosophila phylogeny [11,[32][33][34].While there is no congruence between the patterns of infection and positive selection at bam across species (p = 1.0,S1 Table ), we cannot disregard the potential for historic infections that may have driven positive selection at bam.
In this study, we find that wMel genetically interacts with alanine point mutants throughout the bam gene, suggesting that wMel rescue of bam mutants is not specific to any single binding/functional region of bam.We find no evidence for two hypotheses of how wMel could rescue bam in a non-specific manner: There is no effect of wMel on a hypomorph-like bam knockdown mutant, nor is there an effect of wMel on the frequency of GSC mitosis that occurs upstream of bam.Thus, further consideration on how Wolbachia genetically interacts with bam is valuable for future endeavors that examine potential co-evolution of bam and Wolbachia.

Design of putative bam hypomorph alleles
A Venus-tagged bam allele was designed to facilitate easy analysis of Bam protein expression and to serve as a control to the subsequent 22 Venus-tagged bam mutant alleles (Fig 1a), all of which are described further below.The known Bam binding residues in the CAF40 binding motif were targeted for replacement [23], producing independent alleles with mutations of L17A, M24A, and L28A.The remaining mutant alleles were designed through an alanine scanning approach [35] that targeted the known functional and/or binding regions of bam.Strings of charged residues were changed to alanine by choosing an alanine codon with similar usage bias in D. melanogaster as that of the original residue [36].Additionally, we designed an allele to recreate the bam bw L255F mutation as a transgene with Venus tag.All new alleles are listed in S1 File, with the original and mutated codons.

Generation of transgenic flies
Transgenic fly lines carrying distinct putative hypomorphic bam::Venus alleles and a control bam::Venus transgenic fly line were generated using PhiC-mediated integration [37].Approximately 1.5 kb upstream and 700 bp downstream regions of bam were cloned from genomic DNA of a D. melanogaster bam::Venus line [38].This construct is similar to that of the D. melanogaster bam YFP line from [11] but contains a 36bp glycine-glycine-serine linker sequence between the 3' end of the bam coding sequence and the Venus tag to allow for proper folding of both proteins.The bam::Venus sequence was cloned into a pMiniT 2.0 vector, using the NEB PCR cloning kit.Mutant alleles were subsequently made in the established bam::Venus pMiniT 2.0 plasmid using the NEB Q5 Site-Directed Mutagenesis Kit, with primers designed online using NEBase Changer (S2 Table) [39].Briefly, mutagenic primers were used to amplify the bam::Venus pMiniT 2.0 plasmid using a PCR cycle of 98˚C for 30 seconds; 25 cycles of 98˚C for 10 seconds, 57-72˚C annealing for 30 seconds (specified per primer set in S2 Table ), 72˚C for 3.5 minutes; and 72˚C for 2 minutes.PCR products were treated with kinase, ligase, and DpnI before being transformed into competent cells.The resultant plasmids were subcloned into the pCasper\attB vector, a gift from Dan Barbash that was used by [11] for PhiC transgenesis.Plasmids were verified by sequencing.
Plasmids were injected into embryos of a w, nos-int; P{CaryP}attP40 D. melanogaster line by GenetiVision.G 0 flies were mated with a w 1118 isogenic line and F 1 progeny were screened for successful transgenesis by selecting for those with mini-white eye color.Mutant lines were identified by Sanger sequencing.
All transgenic flies were assessed for function in a bam null background.Thus, the genotype that was assessed in all experiments was w; [w +mc , bam::Venus]/+; bam Δ59 /bam Δ86 , where bam::Venus represents the control bam allele or one of the 22 mutant alleles.For the remainder of this paper, "transgenic females" and "transgenic bam mutants" refer to flies of this genotype.Wolbachia status of the assessed flies (uninfected or wMel-infected) was established by maternal inheritance when crossing in the bam null alleles.

Fertility assays of female transgenic flies
Virgin transgenic females and virgin CantonS males were collected over 48 hours and aged two days.At least 30 individual crosses of one female and two males were set up on yeast-glucose food (minimum of 30, maximum of 88 crosses were set per genotype tested).All crosses were kept in an incubator at 25˚C with a 12-hour light/dark cycle.Parents were removed on the seventh day.Progeny were counted every other day over seven days, starting on the day of first eclosion.If the female parent died by the time of removal, data for that cross was discarded.Five independent fertility assays were conducted to cover all 22 mutants, with a transgenic control (w; [w +mc , bam::Venus]/+; bam Δ59 /bam Δ86 ) present in each assay.
The total progeny after seven days was summed and estimation statistics [40] were performed to determine first, if a fertility defect was present for a mutant, and second, if there was a wMel rescue effect.Estimation statistics and plots were generated using the dabestr package in R (v4.0.1).While this offers an alternative statistical test that allows us to easily visualize effect size, we also report p-values for ease and convention.P-values were calculated through the associated website as results of a two-sided permutation t-test [41], as they were not provided in the R package.

Nurse-cell-positive egg chamber assessment of female transgenic flies
For the ten transgenic mutants that displayed a fertility defect, an assay for nurse-cell-positive egg chambers was also performed, as has been previously done for a supplemental assessment of bam's function [11,12].Briefly, ovaries of 1-2 day old flies were fixed in 4% paraformaldehyde, rinsed and washed with PBST as described above, and then mounted in ProLong Glass with NucBlue (ThermoFisher).Ovaries were imaged on a Zeiss i880 confocal microscope at 40X and 63X (Plan-Apochromat 1.4 NA, oil), using a 405nm laser line (Cornell BRC Imaging Core Facility).Egg chambers with any identifiable large, polyploid nuclei of nurse cells indicated a successful event of Bam's differentiation function and were counted as nurse-cell-positive egg chambers.Estimation statistics were used to assess effect size and a t-test was used for statistical significance as described above for fertility of the female transgenic flies.Correlation between nurse-cell-positive egg chamber data and fertility data was performed with Kendall's rank correlation in R (v4.0.1).

Cytological observations of the ovaries of female transgenic flies
For cytology of the ten transgenic females with fertility defects and the control females, flies were aged 2-3 days and then dissected and prepared for confocal microscopy as in [11,33].Ovaries were dissected in ice-cold 1x PBS and fixed in 4% paraformaldehyde (Electron Microscopy Sciences) for 15 minutes.Tissue was rinsed 3x and washed 4x, respectively, with PBST (1X PBS, 0.1% Triton-X 100), blocked in PBTA (1X PBS, 0.1% Triton-X100, 3% BSA) (Alfa Aesar), and then incubated with primary antibodies overnight at 4˚C.The rinse and wash steps were repeated, followed by an overnight incubation with a goat anti-serum at 4˚C to minimize fluorescent background binding, and then a 1.5-hour incubation with secondary antibodies.The rinse and wash steps were repeated and tissue was mounted in ProLong Diamond AntiFade (Invitrogen).

RNAi knockdown of bam
We performed reciprocal crosses between a wMel-infected nos-Gal4 and an uninfected UASbam HMS00029 line (a TriP bam shRNA line) (Bloomington Drosophila Stock Center lines 4442 and 33631, respectively) to generate the bam knockdown lines.
Bam expression was evaluated by RT-qPCR.RNA of four samples consisting of three 1-2 day old female flies each were prepared using the NEB Monarch Total RNA Miniprep Kit.RT-qPCR was performed with 10μL reactions (5μL Luna Universal One-Step Reaction Mix (2X), 0.5μL Luna WarmStart RT Enzyme Mix (20X), 0.4μL forward primer, 0.4μL reverse primer, 2μL template RNA, 1.7μL water), using three technical replicates and a standard curve.Reverse transcription was performed at 55˚C for 10 minutes, followed by initial denaturation at 95˚C for 1 minute, and then 40 cycles of 95˚C for 10 seconds and 60˚C for 30 seconds.Primers included 5'GCGCAATCGAAACGGAAACT3' and 5'GAGTAGCGGTGCTCCAGATC3' for bam and 5'CCGCTTCAAGGGACAGTATC3' and 5'CAATCTCCTTGCGCTTCTTG3' for

PLOS GENETICS
the reference gene Ribosomal protein L32 (Rpl32) (Rpl32 primers are from [43]).Relative expression to a wildtype bam control was done by the method of [44].
A fertility assay and a nurse-cell-positive egg chamber assay were performed as described above, but at 29˚C to increase the activity of the Gal4-UAS system.Estimation statistics [40] were used to compare the mean fertility of the uninfected and infected bam knockdown mutants.A Wilcoxon signed rank test was used to determine the statistical significance of the mean nurse-cell-positive egg chambers between the uninfected and infected bam knockdown mutants.

Analysis of GSC mitosis frequency
Nurse-cell-positive egg chamber assays were performed as described above to confirm that the L255F bam mutation that we created as a transgene produced a partial bam mutant phenotype and wMel rescue like that of the original bam bw hypomorph (w; bam bw /bam Δ86 ).
Then, we assessed GSC mitosis frequency in the transgenic control (w; [w +mc , bam::Venus]/ +; bam Δ59 /bam Δ86 ) and the transgenic L255F mutant (w; [w +mc , bam L255F ::Venus]/+; bam Δ59 / bam Δ86 ) by comparing the number of actively dividing GSCs indicated by mitosis marker pH3 to the number of non-dividing GSCs.Flies were aged 2-3 days at 25˚C, at which time ovaries were dissected, immunostained and prepared for confocal microscopy, and imaged as described above.
For each genotype, a Fisher's exact test was used to determine if there were more mitotically active GSCs in the wMel-infected germariums than in the uninfected germariums.

Identification of bam sequence amino acid divergences
Bam nucleotide sequences from the D. melanogaster Zambia population of the Drosophila Population Genomics Project [45] were extracted for use in the MK test.Sequences that had any ambiguous nucleotides (Ns) were removed, leaving 187 individual sequences for our analyses.The D. simulans FlyBase reference sequence (r2.02) was used as an outgroup.Sequences were aligned using PRANK v.170427 [46].Nucleotide location of synonymous and nonsynonymous divergences between D. melanogaster and D. simulans were identified using a custom R script (v4.0.1) and mapped against the locations of the alanine mutations generated in this study.

mutants were generated, ten with bam female fertility defects
22 novel mutants were generated as bam transgenes with alanine replacements and a Venus fluorescent tag.They were distributed across the five functional domains/regions of the bam gene (Fig 1a).Each mutant was assayed for fertility defects and results were compared to a control transgenic line of a wildtype bam gene tagged with Venus.Mutants were characterized by comparing the average number of progeny produced by females with and without wMel infection.
Of the 22 bam alanine-scan mutants generated, ten had fertility defects in uninfected females (Fig 2 ).These included all three of the mutants that targeted known binding residues in the CAF40 binding motif of bam, as well as mutants that targeted PLOS GENETICS residues in the Bgcn, Csn4, and ubiquitin binding regions.There were four null-like fertility mutants that produced no progeny when uninfected (transgenic Venus-tagged bam L17A , bam H180A, D181A, H182A , bam K225A, E227A, R228A , and bam E320A, H321A, R322A ).There were six hypomorph-like fertility mutants that produced less progeny on average than the control when uninfected.The six hypomorph-like fertility mutants included three that exhibited severe fertility defects (transgenic Venus-tagged bam L28A , bam E152A, K153A, H155A , and bam R234A, R235A ) and three that exhibited mild fertility defects (transgenic Venus-tagged bam M24A , bam D249A, E250A , and bam E336A, R337A ).The mean effect size between the number of progeny of the mutants with fertility defects compared to the control ranged from a minimum of -28.2 to a maximum of -74., and bam K225A, E227A, R228A ) contain egg chambers with nurse cells, indicating rare occurrences of differentiation that did not translate to adult progeny in the fertility assays.Thus, these mutants are only null-like in their fertility but are not true null alleles in the assessed genetic background.
Two mutants with mild fertility defects (transgenic Venus-tagged bam D249A, E250A and bam M24 ) had mean nurse-cell-positive egg chamber counts that trended lower than the control, but with small effect sizes (-5.54, p = 0.086 and -6.28, p = 0.023; respectively).Overall, fertility and the number of nurse-cell-positive egg chambers correlated across the ten fertilitydefective bam mutants (Kendall's rank correlation τ = 0.828, S3 Fig).
Our transgenic bam::Venus control had no tumorous germaria and had proper Bam expression (Fig 4a and 4b).In parallel with the novel mutants we generated, we recreated the bam bw hypomorphic mutation as a transgenic bam L255F ::Venus mutant.As has been previously characterized for this L255F bam mutation in a bam null background, we see a tumorous germarium with some cells containing a spectrosome (Fig 4c) [8,10,24].We also found that the overproliferating GSC-like cells in this hypomorph expressed Bam, and the lack of Bam expression in the anterior most GSCs suggested that there was no ectopic expression of Bam in the anterior part of the germarium that would cause a loss of GSCs.That these ten new bam phenotypic mutants are distributed throughout the gene demonstrates that multiple regions of the gene and protein are critical for bam function.Additionally, the variation in severity of defects for mutations within one binding or functional region shows that these phenotypes are allele specific.

wMel clearly rescues 8 of 10 fertility defective mutants
To determine if wMel rescues the new bam mutants with fertility defects, we performed both the fertility assessment and nurse-cell-positive egg chamber assessment on the ten transgenic mutants as uninfected and infected with wMel.When infected with wMel, the transgenic bam::Venus control showed a slight increase in fertility across each independent fertility assay (p<0.05,Fig 5 and S3 Table) and in the number of nurse-cell-positive egg chambers (p<0.05,PLOS GENETICS Fig 6).Thus, for a mutant to be classified with wMel rescue, the increase in mean effect size had to be greater than that observed in the control.
Of the four null-like fertility mutants that produced no progeny when uninfected, three produced some progeny when infected with wMel and showed an increase in nurse-cell-positive egg chambers (transgenic Venus-tagged bam L17A , bam H180A, D181A, H182A , and bam E320A, H321A, R322A ) (p<0.05).While the slight production of progeny cannot be classified as wMel fertility rescue because of the small effect size for the three mutants, there is rescue at the cytological nurse-cell level for transgenic bam L17A ::Venus and bam E320A, H321A, R322A ::Venus.Transgenic bam H180A, D181A, H182A ::Venus trends towards rescue at the nurse-cell level, but has a small effect.The fourth null-like fertility mutant (bam E320A, H32A, R322A ::Venus) did not produce progeny or nurse-cell-positive egg chambers regardless of infection status (Fig 4 and S2 Table ).
All six hypomorph-like fertility mutants that had an increase in average fertility (p<0.05,  ), with a mean effect size greater than that of the control).The three mutants with severe fertility defects (transgenic Venus-tagged bam L28A , bam E152A, K153A, H155A , and bam R234A, R235A ) saw an approximate 20-fold increase in the average number of progeny per female when infected with wMel.The three mutants with mild fertility defects (transgenic Venus-tagged bam M24A , bam D249A, E250A , and bam E336A, R337A ) saw an approximate 2.5-fold increase in the average number of progeny per female when infected with wMel, bringing them to a fertility level similar to that of the infected control (p>0.05 for all).The cytological data likewise showed an increase in nurse-cell-positive egg chambers in wMel-infected flies that brought the total numbers to be comparable to the infected control (p>0.05 for all, except bam E152A, K153A, H155A ::Venus where p = 0.042).
Overall, we find that wMel rescues the reduced fertility and/or nurse-cell-positive egg chambers of eight mutants across all the documented binding regions of bam and throughout the predicted 3D protein structure (Fig 7) [47,48].Of the last two mutants, the mutant that contains some nurse-cell-positive egg chambers when uninfected trends towards rescue and the mutant that is null-like at the fertility and nurse-cell level does not.These results show that the wMel rescue is not limited to the L255F bam mutation and that wMel interacts with bam in a manner that is not specific to the known binding regions of the gene.
We plotted the mean progeny for all transgenic females (both uninfected and wMelinfected) to easily compare the above fertility results across all mutants (Fig 8).When comparing the mean fertility for each mutant between the uninfected and wMel-infected lines, we find that the data falls roughly along a logarithmic curve, suggesting there is a relationship between the severity of uninfected mutants and the increase in fertility seen with wMel infection.However, we cannot discount the possibility of the increase in fertility being due to an allelespecific interaction with wMel or variation in Wolbachia abundance.

GSC mitosis in a bam hypomorphic mutant is not affected by wMel
To determine if the increase in fertility and the number nurse-cell-positive egg chambers of bam mutants by wMel is due to an increase in asymmetric division by GSCs, we quantified GSC mitosis in the bam::Venus control and the bam L255F ::Venus mutant.The bam L255F transgenic hypomorph recapitulates the severity of the endogenous bam bw hypomorph, though shows slightly reduced rescue compared to the bam bw hypomorph (S8 Fig) .This difference is likely due to the differences in the genetic backgrounds of these two otherwise identical mutations.
Cells in the apical tip of the germarium that expressed the pH3 mitosis marker but not the Venus-tagged Bam protein were identified as GSCs actively undergoing mitosis (S9 Fig) .The rate of GSC mitosis was not different between wMel-infected and uninfected flies within both the transgenic control and the transgenic hypomorph (p = 1.0) (Table 1).Overall, the bam L255F ::Venus transgenic hypomorph had fewer GSC mitotic events compared to the transgenic control, but it was not statistically significant (p = 0.259 with wMel and p = 0.448 without wMel).

wMel does not rescue a bam knockdown phenotype
To determine if wMel can rescue the phenotype of a reduced expression, but functional Bam protein in females, we tested for the rescue of a bam RNAi knockdown mutant by performing a reciprocal cross using the UAS-Gal4 system.We assessed both the number of egg chambers that have a nurse cell present (nurse-cell-positive egg chambers) per ovary in wildtype and bam knockdown mutants, as in [11,12], as well as the fertility of the knockdown mutants.
We find that the bam knockdown mutants have fewer nurse-cell-positive egg chambers per ovary than the wildtype lines regardless of wMel status (p<0.05)(Fig 9a   time captured in the nurse-cell-positive assay compared to the fertility assay.These results show that wMel does not rescue the tumorous mutant phenotype or the fertility of a bam knockdown mutant.Additionally, there is no increase in bam expression in the knockdown mutant with wMel compared to the uninfected knockdown mutant (Fig 9c).This shows that wMel does not increase the expression levels of bam, consistent with the lack of rescue that we observe.All together, these results suggest that wMel rescue is specific to partial-loss-of-function protein coding bam mutants.

Discussion
The critical reproductive gene bam is one of a few genes that has been shown to genetically interact with Wolbachia in D. melanogaster [11,18].In female and male adults, bam's molecular role is to bring together other proteins into complexes, which then carry out functions vital to the maintenance or progression of gametogenesis.Bam's central and extensive role in early reproduction, as well as Wolbachia's potential for evolutionary impact motivates a question about the specificity of the bam-Wolbachia interaction.Here, we first discuss our results considering the molecular nature of the bam-Wolbachia interaction and then follow up with a discussion on implications for the D. melanogaster host and wMel Wolbachia variant from an evolutionary perspective.
We targeted bam's role of protein binding and minimized affecting the overall conformation of the Bam protein by replacing charged residues with alanine throughout bam's documented binding regions [35].We find no specificity to the presence/absence of rescue or the magnitude of fertility rescue for the different functional regions of the linear bam sequence or with the AlphaFold-predicted Bam protein structure.The strongest evidence for this comes from our finding that wMel rescues fertility defects of the known CAF40 binding residues (bam M24A and bam L28A ), as well as fertility defects of several mutations not involved in CAF40 binding.This suggests that wMel does not rescue the specific interaction between Bam and a single one of its known protein partners in bam fertility mutants.Of note, however, is the possibility that the fertility defects have arisen from a disruption of Bam protein structure despite the practice to mitigate this with an alanine-scanning approach.Additionally, it is possible that the rescued mutants' amino acid sites are involved in a direct interaction between Bam and an unmapped binding partner (e.g., Otu [25]).
The severe and mild mutants were largely rescued to the level of the infected control females.Thus, the severity of the mutation likely plays a large role in wMel's ability to have a substantial impact on host reproductive success.The biochemical nature of the allele interactions and the effect of wMel titer on host reproductive success are areas to be further investigated.Additionally, some egg chambers were noticeably mutant in their cytology, which may explain discrepancies between the nurse-cell-positive egg chamber data and the fertility data.As bam also plays a role in mitotic divisions of the germline [9], nurse-cell-positive egg chambers may be present, but with an incorrect number of nurse cells.
We tested two alternative hypotheses for how wMel could be rescuing bam fertility mutants.The first proposes that wMel induces a fertility rescue by affecting bam's expression.It has previously been shown that there is no difference in bam expression in the bam bw hypomorph as uninfected and infected with wMel [11].Here, we tested the ability of wMel to rescue the mutant phenotype of a wildtype bam gene that is reduced in expression via RNAi.Importantly, we have a knockdown mutant that displays the hypomorph phenotype of tumorous germariums with some wildtype egg chambers.If the knockdown phenotype is too severe, wMel may have no effect as evidenced by the null-like fertility mutant generated in this study (transgenic bam E320A, H321A, R322A ::Venus) and wMel's lack of interaction with the bam null mutant documented previously [11].With the hypomorph-like-phenotype bam knockdown mutant (nos-Gal4/+; UAS-bam HMS00029 /+), we find that wMel does not rescue the number of nurse-cellpositive egg chambers or affect the fertility of female flies; it also does not increase the expression level of bam.
The second hypothesis proposes that wMel increases the rate of GSC mitosis, which increases the output of reproduction.Previous studies have shown wMel to interact with GSCs, including preliminary research in a mei-P26 D. melanogaster mutant that shows increased GSC mitosis in wMel-infected germariums compared to uninfected germariums [49].Other evidence of wMel's interaction with GSCs includes research on the Sxl f4 D. melanogaster mutant that has shown that wMel restores the number of GSCs that were lost due to ectopic bam expression [22] and the necessity for wMel for GSC maintenance in filarial nematodes [50].
We find that wMel does not increase the rate of GSC mitosis in the bam L255F ::Venus hypomorph, and we do not observe any obvious ectopic Bam expression in the GSC niche in any of our transgenic bam mutants.These results are consistent with what we see with the bam knockdown results, as an increase in GSC mitosis rate would likely be reflected in an increase in progeny in the wMel-infected bam knockdown females.Thus, while it is possible for there to be a subtle wMel-GSC interaction that we have not yet assessed, collectively, our results suggest that a large component of the wMel bam interaction is specific to mutant phenotypes that are the result of a disruption of bam function.Biochemical assays are needed to firmly conclude that the wMel bam interaction is specific to bam functions, but it is notable that this proposal is akin to that suggested for Wolbachia's interaction with Sxl [51].The observations that Wolbachia cannot rescue the reduced egg count of a snf mutant, which has no Sxl expression in the germline, and a Sxl f5,fHa mutant, which has a transposon insertion in the second intron [21] parallels the bam knockdown results of our study.
It is possible that the fertility rescue of bam mutants occurs through the same (unknown) mechanism by which the rescue of Sxl mutants occurs.This would be a process beyond the documented restoration of GSC number in Sxl f4 mutants by Wolbachia factor TomO, as it was found that TomO is not sufficient to restore fertility of the Sxl f4 mutants [22].
Bam and Sxl have been proposed to partner together to repress the translation of the stem cell maintenance factor nos to determine GSC daughter cell fate [27,28] and to maintain female sexual identity in the adult germline [52].Alternatively, it is possible that Wolbachia interacts with bam and Sxl independently.Bam has a number of other genetic and physical interacting partners that offer functional pathways for which Wolbachia could be interacting with.For example, Bam promotes differentiation of GSC daughter cells by facilitating deubiquitination of CycA with Otu [25].
Regardless of the manner in which wMel molecularly interacts with bam, that wMel increases fertility of almost all bam protein-coding mutants generated in this study has broader implications for bam and Wolbachia's population and evolutionary dynamics.A recent study modeling potential bam-Wolbachia interactions suggests that wMel's ability to mediate the negative fitness consequences of protein-coding mutants like that which is observed here, is not enough to generate the magnitude of positive selection currently observed at bam [53].The nonsynonymous amino acid replacements that are prevalent in bam across several Drosophila species [31][32][33] may instead be due to arms race dynamics between bam and Wolbachia, an evolutionary process independent of Wolbachia, or some combination of the two.
For Wolbachia, the increased reproductive success for almost all bam protein-coding mutants with wMel infection is an addition to wMel's arsenal of promoting its transmission and spread through populations.It is particularly advantageous for wMel, as, unlike some

PLOS GENETICS
other Wolbachia variants, it exhibits minimal cytoplasmic incompatibility in field-caught D. melanogaster [19].Ultimately, that wMel rescues fertility defects of multiple bam protein-coding mutants, as well as evidence that Wolbachia rescues reproductive defects of other germline genes such as Sxl and mei-P26, may allow for opportunities for an obligate symbiotic relationship between D. melanogaster host and wMel to develop.S1 File.Summary of all generated transgenic mutants.Included for each mutant are the amino acid site(s) and the corresponding functional and/or binding region; the original codon and residue; the number of species that the residue is conserved across from the seven species (D. melanogaster, D. simulans, D. yakuba, D. santomea, D. teissieri, D. erecta, and D. eugracilis) multiple alignment of [38]; the alanine codon that replaced the original codon sequence; and the results of the fertility and nurse-cell-positive egg chamber experiments.(XLSX) S1 Data.Underlying data for fertility and nurse-cell-positive egg chamber figures.(XLSX)

Fig 1 .
Fig 1. Bam is involved in several processes that regulate the differentiation of GSCs in female D. melanogaster.(a) Schematic of the five known binding or functional regions of bam are colored across the coding sequence exons of the gene.Each black vertical bar represents the location of a new mutant generated with alanine replacements in this study.(b) In GSCs (left), Bam is repressed and antagonized by eIF4A.In the differentiating daughter cells (right), Bam is expressed and coordinates other molecules to inhibit the self-renewal pathway or promote the differentiation pathway.Bam has been shown to genetically and physically interact with additional molecules that are not shown.https://doi.org/10.1371/journal.pgen.1011009.g001 and S1 Fig and S3 Table 3 (Fig 2 and S3 Fig).

Fig 2 .
Fig 2. Fertility of transgenic mutants compared to the respective bam::Venus transgenic control.Data is colored according to the documented functional region that the mutation is located in.Below the jitter plots are the mean, two-sided permutation t-test p-value, and the average effect size between the mutant and the respective control.Gray panels indicate mutants with fertility defects (negative average effect size and p<0.05).https://doi.org/10.1371/journal.pgen.1011009.g002

Fig 3 .
Fig 3. Number of nurse-cell-positive egg chambers for transgenic mutants compared to the bam::Venus control.Data is colored according to the documented functional region that the mutation is located in.Below the jitter plots are the mean, two-sided permutation t-test p-value, the average effect size between the mutant and the respective control, and the corresponding fertility classification as described in the text.Gray panels indicate mutants with nurse-cell-positive egg chamber counts that are less than the control (negative average effect size and p<0.05).https://doi.org/10.1371/journal.pgen.1011009.g003 Fig 2 and S1 Fig and S3 Table

Fig 5 .
Fig 5. Fertility of transgenic control and transgenic bam mutants with fertility defects as uninfected and infected with wMel.Two-sided permutation t-test p-values and the average effect size between the uninfected (blue) and infected with wMel (red) data are shown below each data set.Mutant E320A, H321A, R322A was completely sterile with and without wMel and thus is not shown.Gray panels highlight the mutants that are classified as having wMel rescue at the fertility level due having an average effect size greater than that observed in the respective control and p<0.05.https://doi.org/10.1371/journal.pgen.1011009.g005 Fig 5 and S6 Fig and S3 Table) and average number of nurse-cell-positive egg chambers with wMel infection (p<0.05,Fig 6 and S7 Fig and S4 Table

Fig 6 .
Fig 6.Number of nurse-cell-positive egg chambers of the transgenic bam mutants with fertility defects, as uninfected and infected with wMel.Twosided permutation t-test p-values and the average effect size between the uninfected (blue) and infected with wMel (red) data are shown below each jitter plot data set.Mutant E320A, H321A, R322A had no nurse-cell-positive egg chambers with and without wMel and thus is not shown.Gray panels highlight the mutants that are classified as having wMel rescue at the nurse-cell level, due to having an average effect size greater than that observed in the control and p<0.05.https://doi.org/10.1371/journal.pgen.1011009.g006

Fig 7 .
Fig 7. The locations of the mutations generated and their effect on fertility and wMel rescue across the documented binding and functional regions of bam.Shown in (a) a linear schematic of the gene and (b) the predicted protein structure.* indicates fertility defect mutants in which there was no wMel rescue at the fertility and the nurse-cell-positive egg chamber level.https://doi.org/10.1371/journal.pgen.1011009.g007

Fig 8 .
Fig 8. Mean fertility comparison of all transgenic lines generated with and without wMel.The control lines are transgenic bam::Venus (black) and the transgenic mutants are labeled by the first residue that was changed to alanine (orange for mutants with hypomorph-like fertility defects, red for mutants with null-like fertility defects, and blue for mutants without fertility defects).Figure shows data across five independent fertility assays with five different controls.https://doi.org/10.1371/journal.pgen.1011009.g008 and S10 Fig).Bam knockdown mutants with wMel had fewer nurse-cell-positive egg chambers on average than bam knockdown mutants without wMel (p<0.05) but had similar levels of fertility over a period of seven days (p = 0.69, Fig 9b).This difference is likely due to the narrow window of

Table 1 .Fig 9 .
Fig 9. Bam knockdown by RNAi in the UAS-Gal4 system.Bam wildtype 1 and 2 are the UAS and Gal4 stocks, respectively; bam knockdowns are the nos-Gal4/+; UAS-bam HMS00029 /+ reciprocal F1 progeny of the stocks.(a) The number of nurse-cell-positive egg chambers per ovary.*** = significant at p<0.001; NS = not significant.(b) Fertility of the bam knockdown mutants as uninfected and wMel-infected.Jitter plots (left) show the raw data as dots with horizontal lines representing the mean values of each dataset.The mean difference between uninfected and infected lines are plotted as a bootstrap sampling distribution with a 95% confidence interval (right).(c) Bam expression on a log2 scale relative to bam wildtype 1.Each bar represents the dataset of four samples consisting of three female flies, and shows the mean value with standard error.https://doi.org/10.1371/journal.pgen.1011009.g009