Distinct Wolbachia localization patterns in oocytes of diverse host species reveal multiple strategies of maternal transmission

Abstract

A broad array of endosymbionts radiate through host populations via vertical transmission, yet much remains unknown concerning the cellular basis, diversity, and routes underlying this transmission strategy. Here, we address these issues, by examining the cellular distributions of Wolbachia strains that diverged up to 50 million years ago in the oocytes of 18 divergent Drosophila species. This analysis revealed 3 Wolbachia distribution patterns: (1) a tight clustering at the posterior pole plasm (the site of germline formation); (2) a concentration at the posterior pole plasm, but with a significant bacteria population distributed throughout the oocyte; and (3) a distribution throughout the oocyte, with none or very few located at the posterior pole plasm. Examination of this latter class indicates Wolbachia accesses the posterior pole plasm during the interval between late oogenesis and the blastoderm formation. We also find that 1 Wolbachia strain in this class concentrates in the posterior somatic follicle cells that encompass the pole plasm of the developing oocyte. In contrast, strains in which Wolbachia concentrate at the posterior pole plasm generally exhibit no or few Wolbachia in the follicle cells associated with the pole plasm. Taken together, these studies suggest that for some Drosophila species, Wolbachia invade the germline from neighboring somatic follicle cells. Phylogenomic analysis indicates that closely related Wolbachia strains tend to exhibit similar patterns of posterior localization, suggesting that specific localization strategies are a function of Wolbachia-associated factors. Previous studies revealed that endosymbionts rely on 1 of 2 distinct routes of vertical transmission: continuous maintenance in the germline (germline-to-germline) or a more circuitous route via the soma (germline-to-soma-to-germline). Here, we provide compelling evidence that Wolbachia strains infecting Drosophila species maintain the diverse arrays of cellular mechanisms necessary for both of these distinct transmission routes. This characteristic may account for its ability to infect and spread globally through a vast range of host insect species.

Introduction

A large class of endosymbionts rely on efficient vertical transmission to ensure their maintenance and spread through host populations. Many insect endosymbionts achieve this by a strict association with the host germline through successive generations. That is, during oogenesis, endosymbionts target the extreme posterior of the oocyte, the site of germline formation. In contrast to this strict germline-to-germline route of maternal transmission, other endosymbionts occupy the germline via invasion from neighboring somatic cells (Russell et al. 2019). Occupation of the germline via a somatic route requires that the endosymbiont has the capacity to undergo cell-to-cell transmission.

Among all known endosymbionts, maternally transmitted Wolbachia pipientis is the most common, infecting 30–60% of all insect species (Hilgenboecker et al. 2008; Zug and Hammerstein 2012; Weinert et al. 2015). Hosts acquire Wolbachia from other species via horizontal (non-sexual) transfer or from sister species via introgressive transfer (O'Neill et al. 1992; Raychoudhury et al. 2009; Gerth and Bleidorn 2016; Conner et al. 2017; Turelli et al. 2018; Cooper et al. 2019). Following establishment, maternal transmission rates in conjunction with Wolbachia effects on host fitness and reproduction determine Wolbachia spread and prevalence in host populations (Hoffmann et al. 1990; Turelli and Hoffmann 1995; Cooper et al. 2017; Cogni et al. 2021).

While sometimes imperfect, the success and prevalence of Wolbachia ultimately depend on vertical transmission through the female (Hoffmann et al. 1990; Turelli and Hoffmann 1995; Carrington et al. 2011; Hague et al. 2022). The ability of Wolbachia to establish and persist in diverse insect species indicates that it has evolved to usurp highly conserved host processes and components. For example, studies of wMel Wolbachia in model organism Drosophila melanogaster suggest that Wolbachia engage conserved host motor proteins dynein and kinesin in order to target the extreme posterior of the oocyte, the site of germline formation (Ferree et al. 2005; Serbus and Sullivan 2007; Russell et al. 2018).

To study the vertical transmission of Wolbachia in Drosophila, we take advantage of the wealth of knowledge regarding the molecular and cell biology of Drosophila oogenesis (for a review, see Bastock and St Johnston 2008). Drosophila ovaries contain 2 clusters of ovarioles, each containing 2–3 germline stem cells and various stages of developing oocytes. The stem cells undergo asymmetric mitotic division in which 1 daughter cell undergoes self-renewal and the other differentiates into a cytoblast. The cytoblast undergoes 4 rounds of mitotic divisions resulting in 16 interconnected cells, 1 of which becomes the oocyte nucleus while the others differentiate into polyploid nurse cells. The developing oocyte is encapsulated by a monolayer of somatically derived follicle cells resulting in the formation of discrete egg chambers as the maturing oocyte progresses through the ovariole. During the late stages of oogenesis, the nurse cell cytoplasm is expelled into the oocyte, resulting in a rapid doubling of the oocyte volume and concomitant shrinking of nurse cell volume (Theurkauf and Hazelrigg 1998; Bastock and St Johnston 2008). This is followed by posterior pole plasm formation and establishment of the germline (Bilinski et al. 2017).

Wolbachia distribution and localization in insect oocytes and early embryos have been well characterized in wMel-infected D. melanogaster (Kose and Karr 1995; Veneti et al. 2004; Ferree et al. 2005; Serbus et al. 2008; Ramalho et al. 2018; Russell et al. 2018; Guo et al. 2019). Early in oogenesis, wMel Wolbachia are evenly distributed throughout all the cells in the chamber. At this stage, microtubules emanate from the oocyte into the nurse cells (Theurkauf 1994; Grieder et al. 2000). By associating with the minus-end–directed microtubule motor protein dynein, wMel Wolbachia are transported from the nurse cells and concentrate at the anterior end of the oocyte (Ferree et al. 2005). As the oocyte matures, microtubules undergo a dramatic reorganization such that they originate from the oocyte cortex with their plus-ends oriented toward the posterior pole (Theurkauf et al. 1992; Steinhauer and Kalderon 2006). This reorganization is accompanied by a release and dispersal of wMel Wolbachia from the anterior pole. Then, wMel Wolbachia associate with the plus-end–directed motor protein kinesin for transport to the posterior pole and the site of germline formation (Serbus and Sullivan 2007). There, Wolbachia (wMel) associate with pole plasm determinants maintaining their position at the site of germline formation (Serbus and Sullivan 2007) in spite of extensive cytoplasmic streaming (Serbus et al. 2005).

Thus, wMel Wolbachia location and migration patterns provide insights into the host factors with which Wolbachia engages throughout oogenesis. The Drosophila simulans-associated wRi strain exhibits a similar anterior distribution in early oogenesis and then becomes distributed across the entire mature oocyte. However, unlike wMel, it does not concentrate at the posterior pole of the oocyte (Veneti et al. 2004; Serbus and Sullivan 2007). This suggests wRi engages host dynein and kinesin, but not pole plasm determinants. wMel Wolbachia strains introgressed into a D. simulans host background exhibit a concentration in the pole plasm, indicating the Wolbachia genome rather than the host genome drives oocyte localization properties (Poinsot et al. 1998; Veneti et al. 2004; Serbus and Sullivan 2007).

With the exception of wMel and wRi, little is known of the extent to which distribution patterns in the oocyte and transmission routes vary among diverse Wolbachia strains and host species. Here, we fill this important knowledge gap, by examining 18 Drosophila host species infected with Wolbachia strains that diverged up to 50 million years ago. These studies also provide a direct read-out of Wolbachia and host motor protein–pole plasm interactions that are highly conserved and those that vary among species. This work complements and extends previous analyses revealing Wolbachia strain-dependent localization patterns in the Drosophila oocyte and embryo (Veneti et al. 2004; Toomey et al. 2013).

We discovered that Wolbachia distribution patterns in the Drosophila oocytes fall into 3 categories: (1) the vast majority of Wolbachia clusters at the posterior pole; (2) a concentration at the posterior pole, but with a large portion of Wolbachia distributed throughout the entire oocyte; and (3) Wolbachia distributed throughout the oocyte, with none or very few located at the posterior pole. This latter class is particularly interesting as it suggests Wolbachia likely enters the oocyte via a somatic route for transmission to the next generation. We also provide evidence that neighboring somatic follicle cells are the likely source of Wolbachia entering the germline. Together, these findings indicate that Wolbachia are capable of achieving efficient vertical transmission either via strict maintenance in the germline from 1 generation to the next or via invasion of the germline through neighboring somatic cells. Thus, Wolbachia must maintain the diverse arrays of cellular mechanisms necessary for both of these distinct transmission routes. We identify candidate surface proteins that may be responsible for these diverse localization patterns in host oocytes.

Materials and methods

Drosophila stocks

All stocks were grown on standard brown food (Sullivan et al. 2000) at 25°C with a 12-h light/dark cycle. Uninfected Drosophila stocks were generated by tetracycline curing of the infected stock (Serbus et al. 2015). Listed in Table 1 are the 19 different Wolbachia strains we examined, which infect 18 different Drosophila host species. This included 6 wMel-like strains (Cooper et al. 2019; Hague et al. 2020a), 5 wRi-like strains (Turelli et al. 2018), 7 other A-group Wolbachia, and the B-group strain wMau that diverged from A-group Wolbachia up to 46 million years ago (Meany et al. 2019). Each genotype was generated as an iso-female line by sampling a single gravid female from the field and placing her individually into a vial.

Table 1.

Qualitative and quantitative quantification of cellular Wolbachia abundance in stage 10 oocytes.

Wolbachia strain	Host species	Localization class	N	Whole oocyte mean CTCF	Posterior mean CTCF	Posterior cortex mean CTCF
wRec	D. recens	Clumped	3	482,174	300,753	58,453
wTris	D. tristis	Localized	4	223,048	129,013	105,641
wSeg	D. seguyi	Dispersed	5	58,552	11,402	557
wCha	D. chauvacae	Localized	5	108,588	42,636	9,786
wBurl	D. burlai	Localized	4	72,806	53,754	36,944
wMel	D. melanogaster	Localized	2	191,451	136,174	92,341
wBic	D. bicornuta	Clumped	3	90,114	83,621	54,900
wRi	D. simulans	Dispersed	5	333,505	90,664	20,494
wPan	D. pandora	Localized	3	45,423	6,486	1,890
wAna	D. ananassae	Localized	4	41,578	16,843	12,927
wTria	D. triauraria	Dispersed	2	487,897	1,809	345
wAura	D. auraria	Dispersed	5	266,478	3,699	645
wHa	D. simulans	Dispersed	5	82,941	3,488	807
wBai	D. baimaii	Dispersed	2	11,359	0	0
wBoc	D. bocki	Clumped	6	137,664	98,339	82,408
wLeo	D. leontia	Clumped	4	997,045	685,260	713,918
wTsa	D. tsacasi	Localized	4	61,314	36,478	20,620
wNik	D. nikananu	Localized	3	44,764	27,099	13,014
wMau	D. mauritiana	Dispersed	3	69,362	799	322

The number of confocal images (N) used to generate means is shown for each Wolbachia-infected host species. Wolbachia abundance was measured in each region of the oocyte as corrected total cell fluorescence (CTCF).

Ovary, embryo, and larva fixation and staining

Newly eclosed flies were transferred to fresh food for 3–5 days for aging. Approximately 10 females were dissected for each slide preparation. Ovaries were fixed using a modification of previously published procedures (Brendza et al. 2000; Russell et al. 2018). Ovaries were removed and separated in phosphate buffer solution and then fixed in a 200-μl devitellinizing solution (2% paraformaldehyde and 0.5% v/v NP40 in 1 × PBS) mixed with 600 μl heptane for 20 min at room temperature, on a shaker. After removal of the organic layer by brief centrifugation for sample isolation, the ovaries were washed 3 times with PBS-T (0.1% Triton X-100 in 1 × PBS) along with three 5-min washes. Samples were treated with RNAse A (10 mg/ml) and left overnight at room temperature. Samples were then washed again a few times with PBS-T and then stained with a dilute solution of phalloidin for actin staining on a rotator for 1 h. Samples were washed again with 3 quick PBS-T rounds and then solution changes with PBS-T over 2 h. Sixty microliters of propidium iodide (PI) in mounting media was added to the sample after removal of the wash solution and left again overnight. Ovaries were then mounted and carefully separated out again for ease of imaging and removal of excess mature eggs. Slides were coated in nail polish and stored at −20°C until imaged.

Embryos were collected for 1–3 h on plates and fixed in equal volume 32% paraformaldehyde and heptane and fixed as previously described (Sullivan et al. 2000). After extraction with methanol, embryos were blocked for 1 h in 5% PBST–BSA (2.5 g bovine serum albumin fraction V in 50 ml PBS-T) at room temperature. Embryos were then incubated with anti-WD_0009 or anti-FtsZ at 1:500 overnight at 4°C. After 3 washes with 1% PBST–BSA over 1 h, embryos were incubated in Alexa 488 Anti-Rabbit at 1:500 for 1 h at room temperature. Samples were then washed 3 more times over 1 h with 1% PBST–BSA. Embryos were mounted with media containing DAPI and stored at −20°C until imaged.

Larval fat bodies were dissected and fixed from the 3rd instar larvae using a modified version of a published protocol (Maimon and Gilboa 2011). Fat bodies were dissected in 1 × PBS then fixed in 5% paraformaldehyde in PBS for 20 min with gentle agitation. Fat bodies were washed for 5 min, 10 min, then 45 min with cold 1% NP40 in PBS. After washes, fat bodies were blocked in 5% PBST–BSA (2.5 g bovine serum albumin fraction V in 50 ml PBS-T) for 1 h at room temperature. Fat bodies were then incubated with anti-FtsZ at 1:500 overnight on a rocker at 4°C, washed 3 times over 1 h with 1% PBST–BSA, and then incubated with Alexa 488 Anti-Rabbit at 1:500 at room temperature for 1 h. After 3 additional washes over 1 h with 1% PBST–BSA, fat bodies were mounted onto a slide with media containing DAPI. Larval ovaries were isolated from fat bodies using 0.1-mm needles before sealing the slide with nail polish. Samples were then stored at −20°C.

Confocal microscopy

Imaging was performed on an inverted Leica DMI6000 SP5 scanning confocal microscope. Optical sections of a Nyquist value of 0.38 were used. A variety of zooms were used to optimize image viewing, with most being set at 1.5×. Propidium iodide was excited with the 514- and 543-nm lasers, and emission from 550 to 680 nm was collected. GFP was imaged with the 488-nm laser, and emission from 488 to 540 nm was collected. Alexa 633 was imaged with the 633 laser, and emission from 606 to 700 nm was collected. All imaging was performed at room temperature. Images were acquired with Leica Application Suite Advanced Fluorescence software.

Wolbachia quantification and analysis

Following Russell et al. (2018), we used the polygon selection tool to select 3 different regions of the oocyte (see Supplementary Fig. 1 in Russell et al. 2018): the whole oocyte, the posterior region, and the posterior cortex. The “area” and “integrated density” were measured for each region. Additionally, the average of a quadruplicate measure of “mean gray value” beside the stained oocytes was measured as background fluorescence. The corrected total cell fluorescence (CTCF) was calculated as integrated density − (area × background mean gray value) for each region.

All statistical analyses were performed in R. We first tested whether our 3 qualitative classes (posteriorly clumped, posteriorly localized, and dispersed) differed significantly in quantitative estimates of Wolbachia fluorescence (CTCF). We used a one-way ANOVA to test if the 3 groups differed based on their mean CTCF values from the whole oocyte, the posterior region, and the posterior cortex. Distribution and leverage analyses indicated that a transformation of x′ = log(x + 1) was needed to meet assumptions of normality. We tested for significance using an F test with type III sum of squares using the “Anova” function in the car package (Fox and Weisberg 2019). We conducted Tukey’s multiple comparisons among the 3 groups using the “TukeyHSD” function.

Genomic data

In order to perform the phylogenomic analysis and generate phylograms, we obtained Wolbachia sequences from publicly available genome assemblies, which included wMel (Wu et al. 2004), wRi (Klasson et al. 2009), wHa (Ellegaard et al. 2013), wAura, wTria, wPan (Turelli et al. 2018), wAna (Salzberg et al. 2005), and wRec (Metcalf et al. 2014). All other Wolbachia sequences included in this study [wBai, wBic, wBoc, wBurl, wCha, wCurt (wTsa), wNik, wSeg, wTris, and wLeo] were obtained using Illumina sequencing and previously described methods (Hague et al. 2020b). Briefly, tissue samples for genomic data were extracted using a DNeasy Blood & Tissue kit (Qiagen). DNA was cleaned using Agencourt AMPure XP beads (Beckman Coulter, Inc.) following manufacturers’ instructions and eluted in 50 μl 1 × TE buffer for shearing. DNA was sheared using an E220 Focused Ultrasonicator (Covaris Inc.) to a target size of 400 bp. We prepared libraries using NEBNext Ultra II DNA Library Prep with Sample Purification Beads (New England BioLabs). We indexed samples using NEBNext Multiplex Oligos for Illumina (Index Primers Set 3 and Index Primers Set 4), and 10 μl of each sample was shipped to Novogene (Sacramento, CA, USA) for sequencing using Illumina HiSeq 4000, generating paired-end 150 bp reads.

We obtained publicly available host sequences for D. melanogaster (Hoskins et al. 2015), D. simulans (Hu et al. 2013), Drosophila ananassae (Clark et al. 2007), Drosophila pandora (Turelli et al. 2018), Drosophila mauritiana (Meany et al. 2019), Drosophila auraria, Drosophila triauraria, Drosophila baimaii, Drosophila bicornuta (aff. bicornuta), Drosophila bocki, Drosophila burlai, Drosophila chauvacae (cf. chauvacae), Drosophila curta (aff. tsacasi), Drosophila nikananu, Drosophila seguyi, Drosophila tristis, and Drosophila leontia (Conner 2021). A Drosophila recens genome assembly was kindly provided by Kelly Dyer and Rob Unckless. We note that 3 of the montium host species listed above have had recent updates to their taxonomic nomenclature: D. bicornuta has been renamed to aff. bicornuta, chauvacae has been renamed to cf. chauvacae, and curta has been renamed to aff. tsacasi. See Conner et al. (2021) for a full discussion of montium species names and relationships.

Wolbachia phylogenetic analysis

Raw Illumina reads from the newly sequenced Wolbachia strains were trimmed using Sickle version 1.33 (Joshi and Fass 2011) and assembled using AbySS version 2.0.2 (Jackman et al. 2017). K-values of 71, 81, and 91 were used, and scaffolds with the best nucleotide BLAST matches to known Wolbachia sequences with E-values less than 10⁻¹⁰ were extracted as the draft Wolbachia assemblies. For each strain, we chose the assembly with the highest N₅₀ and the fewest scaffolds (Supplementary Table 1). The newly assembled genomes and the previously published assemblies were annotated using Prokka version 1.11, which identifies homologs to known bacterial genes (Seemann 2014). To avoid pseudogenes and paralogs, we only used genes present in a single copy with no alignment gaps in all of the genome sequences. Genes were identified as single copy if they uniquely matched a bacterial reference gene identified by Prokka. By requiring all homologs to have identical length in all of the Wolbachia genomes, we removed all loci with indels. A total of 66 genes totaling 43,275 bp met these criteria. We then estimated a Bayesian phylogram using RevBayes 1.0.8 under the GTR + Γ + I model partitioned by codon position (Höhna et al. 2016). Four independent runs were performed, which all converged on the same topology. All nodes were supported with Bayesian posterior probabilities of 1.

Similar methods were used to generate gene trees for candidate Wolbachia loci putatively involved in host interactions during oogenesis. This included the 3 wsp paralogs [WD_1063 (wsp/RS04815), WD_0009 (wspB/RS00060), and WD_0489 (wspC/RS06475)] and 3 other Wolbachia surface proteins [WD_1085 (bamA/RS04910), WD_0501 (p44/MSP2/RS0225), and WD_1041 (peptidase M2/RS04710)]. Here, the WD_XXXX locus tags refer to the wMel genome assembly (Wu et al. 2004). The protein-coding sequences for each locus were extracted from the assemblies using BLAST and the wMel reference sequences. Notably, we found evidence of insertions in the WD_0009 (wspB) sequences of wBoc, wLeo, and wBic (see Results). We removed these large insertions from the alignment to generate the wspB gene tree. In addition, BLAST identified WD_0489 (wspC) in wMel-like Wolbachia and closely related wBai, wLeo, and wBoc, but yielded no hits in the other Wolbachia strains (wNik, wTsa, wPan, wAna, wTria, wAura, wRi, and wMau), indicating that wspC is either too diverged from the wMel reference for BLAST recognition or alternatively wspC is only present in wMel-related Wolbachia strains. Sequences for each gene were aligned with MAFFT 7 (Katoh and Standley 2013). We then used RevBayes and the GTR + Γ + I model partitioned by codon position to generate gene trees for each locus.

Host phylogenetic analysis

Host phylogenies were generated using the same nuclear genes implemented in Turelli et al. (2018): aconitase, aldolase, bicoid, ebony, enolase, esc, g6pdh, glyp, glys, ninaE, pepck, pgi, pgm, pic, ptc, tpi, transaldolase, white, wingless, and yellow. We used BLAST with the D. melanogaster coding sequences to extract orthologs from the genomes of each host species. Sequences were then aligned with MAFFT 7. Finally, we used RevBayes and the GTR + Γ + I model partitioned by codon position and gene to accommodate potential variation in the substitution process among genes, as described in Turelli et al. (2018). Our initial phylogenetic analysis produced a polytomy in the melanogaster species group regarding the relationships among the montium, melanogaster, and ananassae subgroups (Supplementary Fig. 1). These relationships have been resolved in previous genomic studies involving additional taxa (Turelli et al. 2018; Suvorov et al. 2022); therefore, we generated a constrained host tree in RevBayes that enforced previously defined relationships among the 3 subgroups: [(montium, melanogaster) and ananassae].

Tests for phylogenetic signal

The resulting phylograms were used to test whether cellular Wolbachia abundance in oocytes exhibits phylogenetic signal on either the Wolbachia or host phylogenies. We used our quantitative estimates of Wolbachia abundance to test for phylogenetic signal using Pagel's lambda (λ) (Pagel 1999). Here, we used total Wolbachia fluorescence in whole oocytes, the posterior region, and the posterior cortex (log-transformed CTCF) as continuous characters to calculate maximum likelihood values of Pagel's λ. A Pagel's λ of 0 indicates that character evolution occurs independently of phylogenetic relationships, whereas λ = 1 is consistent with a Brownian motion model of character evolution. We used the “fitContinuous” function in GEIGER (Harmon et al. 2008) and a likelihood ratio test to compare our fitted value to a model assuming no phylogenetic signal (λ = 0). We also used a Monte Carlo-based method to generate 95% confidence intervals surrounding our estimates using 1,000 bootstrap replicates in the pmc package (Boettiger et al. 2012). When applicable, we conducted additional analyses to evaluate whether the number of taxa in our phylogeny (N = 19 Wolbachia strains and N = 18 host species) limited our ability to detect significant departures from λ = 0 (Hague et al. 2020b, 2021). Small phylogenies are likely to generate near 0 values simply by chance, not necessarily because the phylogeny is unimportant for trait evolution (Boettiger et al. 2012). To evaluate whether larger phylogenies increase the accuracy of estimation, we simulated trees with an increasing number of Wolbachia strains/host species (N = 25 and 50) and our empirical estimates using the “sim.bdtree” and “sim.char” functions in the geiger R package (Harmon et al. 2008). We then re-estimated confidence intervals using the larger simulated trees.

Our tests for phylogenetic signal using the constrained host phylogram are presented in the main text; however, we note that our analysis using the unconstrained host tree (with a polytomy at the base of the melanogaster species group) produced similar results. Using the unconstrained tree, we found that Wolbachia abundance in whole oocytes [λ = 0.782 (0, 0.987), P = 0.469], the posterior region [λ = 0.951 (0, 1), P = 0.377], and the posterior cortex [λ < 0.001 (0, 0.563), P = 1] did not exhibit significant phylogenetic signal.

Structural analysis of Wolbachia candidate genes

We used HHPred to identify protein domains of WD_1085 and WD_0501 using the COPe70_2.07, Pfam-A_v33.1, COG_KOG_v1.0, and SMART_v6 databases (Zimmermann et al. 2018; Gabler et al. 2020). Here, we used default setting, including an E-value cutoff for MSA generation of 1e−3, a minimum coverage of MSA hits of 20%, a minimum probability in hitlist of 20%, and a maximum of 250 target hits. As expected, both WD_1085 and WD_0501 yielded multiple hits of varying length for domains associated with outer membrane proteins that form beta barrel structures (results summarized in Supplementary Table 2). We used SignalP-6.0 to identify the signal peptide cleavage sites for WD_1085 between aa residues 25 and 26 (Teufel et al. 2022) and removed the signal peptide for homology modeling. SignalP was unable to confidently identify the signal peptide sequence for WD_0501, so we used the full sequence for homology modeling. We then used I-TASSER to model the 3D structure of each protein through sequence homology (Yang and Zhang 2015; Zhang et al. 2017). For WD_1085, we used PDB:4K3B with an amino acid sequence identity of 23% in the threaded aligned region as a template (Noinaj et al. 2103). For WD_0501, we used PDB:2MLH with an amino acid sequence identity of 17% in the threaded aligned region as a template (Fox et al. 2014). PyMOL was used for image generation (Schrodinger 2015). Results are displayed in Supplementary Fig. 2.

Ectopic expression of Wolbachia surface proteins in yeast

SC–leucine yeast media was prepared as described previously (Ros et al. 1990). Where present, benomyl was added to media from a 10 mg/ml stock in DMSO to a final concentration of 2.5 µg/ml. The yeast expression assays are a modified version of the protocol described by Sheehan et al. (2016). Wolbachia genes were codon optimized for expression in Saccharomyces cerevisiae and synthesized by Twist Bioscience (South San Francisco). The wsp, wspB, and wspC genes were gifts of the Free Genes Project (https://stanford.freegenes.org/). These genes were inserted into pGH448 using the New England Biolabs Golden Gate kit according to the manufacturer's instruction. This placed the Wolbachia genes under control of the GAL1 promoter and CYC1 transcriptional terminator. This plasmid, which will be described in more detail elsewhere, carries a CEN/ARS element and also contains the kanamycin resistance and LEU2 genes for selection in bacteria and yeast, respectively.

The indicated plasmids were transformed into yeast strain GHY1934 (MATa  his3Δ200 lys2-128δ leu2Δ1 ura3-52 trp1Δ63 pGAL1-FLO8-HIS3::KANMX) and GHY1551 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 mad1Δ::KANMX). Both of these strains are isogenic to S288C and are GAL2+. Transformed cells were grown to saturation in SC–leucine liquid media containing 2% glucose. These cells were counted, diluted to 1 × 10⁻⁷ cells per ml, 5-fold serial diluted, and spotted to SC–leucine plates with either 2% glucose or 2% galactose as the carbon source and incubated at the times and temperatures indicated in the figure legends.

Results

Wolbachia navigates the developing Drosophila oocyte in distinct stages

As described above, wMel Wolbachia migration and distribution throughout oogenesis in D. melanogaster have been well documented (Kose and Karr 1995; Veneti et al. 2004; Serbus and Sullivan 2007; Serbus et al. 2008; Ramalho et al. 2018; Russell et al. 2018; Guo et al. 2019). As a reference for our analysis of Wolbachia oocyte distributions in a range of Drosophila species, the key distribution patterns of wMel throughout D. melanogaster oocytes are depicted in Fig. 1.

Fig. 1.

Confocal micrographs of D. melanogaster (wMel) oocytes tracked through early oogenesis. Samples are stained with propidium iodide (PI) for DNA (red) and phalloidin for actin (green). Representative stages of development (1–10) are shown for wMel infected (left panels) and uninfected (right panels). Grayscale zoomed-in images around the nucleus of the oocyte are shown in the 3rd column, with red arrows highlighting Wolbachia in the oocyte. Corresponding arrows in the 6th column show the equivalent area in uninfected oocytes. The magenta double arrow in the stage 5 infected oocyte (2nd column) highlights Wolbachia in the nurse cells. Scale bars all set to 25 µM.

Stage 1 includes the germarium, a structure consisting of the germline stem cells and their cystoblast daughter cells that will differentiate into mature oocytes. Progression to stage 2 involves the cystoblast undergoing 4 rounds of mitotic division to produce a syncytium of 16 cells connected by cytoplasmic bridges and surrounded by a layer of somatically derived follicle cells (Bastock and St Johnston 2008). Propidium iodide (PI)-stained Wolbachia (red) are clearly observed as puncta in the cytoplasm of the infected, but not the uninfected 16 cell cyst (Fig. 1, stages 1 and 2, arrows). Actin is labeled in green.

During stage 3, the cyst has become polarized with 1 of the cells differentiating into an oocyte and the others becoming polyploid nurse cells. At this stage, Wolbachia rely on dynein-mediated microtubule transport (Ferree et al. 2005) and are clearly seen in the infected oocyte cytoplasm anteriorly concentrated at 2 cytoplasmic bridges that connect to the nurse cells (Fig. 1, stage 3, arrows).

During stages 4 and 5, the nurse cell–oocyte complex becomes oblong and is fully encapsulated by follicle cells. Wolbachia are clearly observed in both the oocyte and nurse cell cytoplasm (Fig. 1, stages 4 and 5, arrows). The Wolbachia in the oocyte are tightly associated with the anterior cortex.

During stage 6, the number of follicle cells increases and they are more densely packed in a monolayer encompassing the oocyte. Here, we discovered a previously unrecognized aspect of Wolbachia dynamics and localization: the number of anteriorly localized Wolbachia is greatly diminished at this stage (Fig. 1, stage 6, arrow). This pattern was consistent across all 10 of the stage 6 oocytes we observed in our analysis (see Supplementary Fig. 3).

At stage 7, Wolbachia are again observed in the oocyte. However, rather than being tightly associated with the anterior cortex, Wolbachia are evenly distributed throughout the oocyte (Fig. 1, stage 7, arrow). During stage 7, the microtubules reorient such that their plus-ends extend toward the posterior pole (Cha et al. 2001, 2002; Steinhauer and Kalderon 2006). Accordingly, the plus-end motor protein kinesin is required to establish the even distribution of Wolbachia throughout the maturing oocyte (Serbus and Sullivan 2007). Stage 8 heralds the beginning of vitellogenesis, the process of yolk formation and deposition of nutrients into the oocyte (Cummings et al. 1971). Wolbachia are near, but not yet located at, the posterior pole (Fig. 1, stage 8, arrow). At stage 9, anterior follicle cell migration occurs (Rørth 2002). At this stage, Wolbachia displays an arch-like posterior cortical localization (Fig. 1, stage 9, arrow). By stage 10, there is a distinct concentration of Wolbachia along the posterior cortex, the site of germline formation (Fig. 1, stage 10, arrow). Nurse cell dumping will shortly follow, leading into stages 11 and 12 with the oocyte rapidly growing in size (Mahajan-Miklos and Cooley 1994).

The wMel strain in D. melanogaster provides an outline of the molecular and cellular mechanisms driving Wolbachia vertical transmission. It also raises the question of whether the mechanisms and strategies observed in D. melanogaster are conserved across diverse Wolbachia strains and Drosophila species. To explore this issue, the wMel distribution in D. melanogaster serves as the basis for our comparison of 18 other Wolbachia strains. Bayesian phylograms depicting phylogenetic relationships among the Wolbachia strains and the Drosophila species are shown in Fig. 4 and Supplementary Fig. 4. Our analysis comprised 18 A-group Wolbachia strains, including 6 wMel-like strains (Cooper et al. 2019; Hague et al. 2020a) and 5 wRi-like strains (Turelli et al. 2018), as well as the B-group Wolbachia strain wMau, which diverged from A-group Wolbachia up to 50 million years ago (Meany et al. 2019). These Wolbachia strains infect 18 divergent Drosophila host species spanning 3 species groups (Suvorov et al. 2022).

Fig. 4.

Wolbachia abundance in the posterior region of host oocytes exhibits phylogenetic signal on the Wolbachia phylogeny. a) Estimated Bayesian phylogram for the 19 A- and B-group Wolbachia strains included in tests for phylogenetic signal. b) Mean estimates of Wolbachia abundance (log-transformed CTCF) in the whole oocyte, posterior region, and the posterior cortex. Wolbachia abundance in the posterior region has significant phylogenetic signal [λ = 0.982 (0, 0.998), P = 0.021], suggesting Wolbachia-associated factors determine Wolbachia abundance in the oocyte posterior. c) Estimated Bayesian phylogram of Drosophila host species using 20 conserved single-copy genes. Wolbachia and host divergence times in millions of years (MYA) are reproduced from Meany et al. (2019) and Suvorov et al. (2022), respectively. The Wolbachia and host trees are generally discordant (see Supplementary Fig. 4), as expected with frequent horizontal Wolbachia acquisition (Turelli et al. 2018).

An initial concentration of Wolbachia at the oocyte anterior is a conserved feature

As described above, wMel Wolbachia enter the oocyte through a complex of ring canals connecting the nurse cells to the oocyte and concentrate at the anterior cortex at stages 3–5. Initially observed in D. melanogaster, the number of Wolbachia increases dramatically while at the anterior (Ferree et al. 2005). To determine if anterior localization is a conserved aspect of Wolbachia's navigation through the developing oocyte, we first characterized this trait in 8 Wolbachia strains in their respective hosts: wPan in D. pandora, wNik in D. nikananu, wCha in D. chauvacae, wHa in D. simulans, wTsa in D. tsacasi, wBoc in D. bocki, wAna in D. ananassae, and wMel in D. melanogaster. All 8 strains exhibited a distinct anterior localization during mid-oogenesis (Fig. 2). It is interesting that this localization occurs during stage 5 of oogenesis, which, in D. melanogaster, is prior to the localization of all known anterior axis determinants. Thus, Wolbachia likely relies on as yet undiscovered anteriorly concentrated factor(s).

Fig. 2.

Conserved Wolbachia anterior localization during early oogenesis (stages 4 and 5). Wolbachia and host DNA stained with PI (red) and actin with phalloidin (green) shown in the left most column. Grayscale of PI channel shown in the middle. Four times zoom-in around the oocyte shown on the right. All of the 8 Wolbachia strains examined concentrate at the anterior cortex of the oocyte at this stage. Red arrows highlight Wolbachia's anterior localization positioning for a) wPan in D. pandora, b) wNik in D. nikananu, c) wCha in D. chauvacae, d) wHa in D. simulans, e) wTsa in D. tsacasi, f) wBoc in D. bocki, g) wAna in D. ananassae, and h) wMel in D. melanogaster. Oocytes are approximately 50–75 µM in size.

Wolbachia exhibits 3 distinct distributions with respect to posterior localization in the mature Drosophila oocyte

Concomitant with microtubule rearrangement during stages 7 and 8, wMel Wolbachia are released from the oocyte anterior resulting in the dispersal of the bacteria throughout the entire length of the oocyte with a fraction concentrating at the posterior pole (Ferree et al. 2005). The posterior dispersal of Wolbachia requires the plus-end motor protein kinesin (Serbus and Sullivan 2007). Maintenance of those Wolbachia that reach the posterior pole relies on a stable association with key pole plasm components (Serbus et al. 2011).

We next examined the oocytes of 18 Drosophila species infected with 19 diverse Wolbachia strains (N = 72 total oocytes) to test the hypothesis that oocyte posterior distributions vary. Unlike the conserved anterior localization, Wolbachia exhibit 3 distinct patterns of posterior localization in stage 10 oocytes. As shown in Fig. 3 and Supplementary Fig. 5a, 8 Wolbachia strains exhibit a localization pattern similar to that observed for wMel in D. melanogaster with the bulk of Wolbachia evenly distributed throughout the oocyte, and a large fraction concentrated at the posterior pole. We refer to this pattern as posteriorly localized.

Fig. 3.

Three distinct patterns of posterior Wolbachia localization during late-stage oogenesis. Confocal micrographs of Drosophila oocytes DNA-stained with PI (red) and actin-stained with phalloidin (green) show representative examples of a) posteriorly localized wTris in D. tristis, b) posteriorly clumped wBic in D. bicornuta, and c) dispersed localization of wTria in D. triauraria. Second row depicts single-channel images of PI staining. Third row depicts an enlarged PI-stained image of the posterior region of each oocyte. Scale bars set at 25 µM.

Our analysis revealed 4 Wolbachia strains manifested a 2nd pattern in which the vast majority of Wolbachia localized to a distinct cluster at the posterior pole and only few Wolbachia are distributed throughout the remainder of the oocyte (Fig. 3 and Supplementary Fig. 5b). In contrast to the pattern described above, rather than being distributed in a wide arc along the posterior pole, they are tightly centered at the extreme posterior region of the oocyte (Fig. 3b). We refer to this pattern as posteriorly clumped.

Finally, 7 Wolbachia strains exhibit a 3rd pattern in which Wolbachia are distributed throughout the entire oocyte but there are few to no bacteria concentration at the posterior pole (Fig. 3 and Supplementary Fig. 5c). The lack of the Wolbachia at the posterior pole was unexpected as this was thought to be necessary for incorporation into the germline of the next generation. We refer to this pattern as dispersed.

Quantification of Wolbachia oocyte distribution patterns

We used one-way ANOVAs to test whether the 3 classes (posteriorly localized, posteriorly clumped, and dispersed) differ statistically in quantitative estimates of Wolbachia abundance in the whole oocyte, at the posterior region, and at the posterior cortex (Supplementary Fig. 6). We defined the posterior region as the posterior 12.5% of the oocyte and the posterior cortex as the narrow cortical region of Vasa expression (Russell et al. 2018), and we then measured total Wolbachia abundance in each region as the corrected total cell fluorescence (CTCF). We found that Wolbachia abundance differs significantly among the 3 classes (F_(2,69) = 6.2, P = 0.003), with Tukey’s multiple comparisons indicating that the clumped class has significantly more Wolbachia in whole oocytes than the localized group (P = 0.002). The 3 classes also differed significantly in the amount of Wolbachia in the posterior region (F_(2,69) = 23.09, P < 0.001), such that the dispersed class has significantly fewer Wolbachia at the posterior than the clumped (P < 0.001) and localized (P < 0.001) groups. Finally, the 3 classes also differ in the amount of Wolbachia at the posterior cortex (F_(2,64) = 38.12, P < 0.001). Here, multiple comparisons indicate that all 3 classes significantly differ in levels of Wolbachia at the posterior cortex: the clumped class has the most Wolbachia, followed by localized, and then dispersed (Supplementary Fig. 6). Mean estimates of Wolbachia abundance (CTCF) for each Wolbachia strain and host species are shown in Table 1.

Wolbachia abundance in the oocyte posterior has strong phylogenetic signal

We hypothesized that the diversity of Wolbachia localization patterns may be due to Wolbachia-associated factors, implying that closely related Wolbachia strains would exhibit similar localization patterns in host oocytes. We used the phylogram describing the evolutionary relationships among Wolbachia strains to test whether patterns of Wolbachia localization exhibit a phylogenetic signal using Pagel's λ (Pagel 1999; Hague et al. 2020b, 2021) (Fig. 4). A λ value of 1 would be consistent with trait evolution that entirely agrees with the Wolbachia phylogeny (i.e. strong phylogenetic signal supporting a role of Wolbachia factors), whereas a value of 0 would indicate that trait evolution occurs independently of phylogenetic relationships (Pagel 1999; Freckleton et al. 2002). Most notably, we found that Wolbachia abundance (CTCF) in the posterior region exhibits a strong, significant phylogenetic signal [λ = 0.982 (0, 0.998), P = 0.021]. Wolbachia abundance at the posterior cortex also has a high, but non-significant λ value [λ = 0.959 (0, 0.995), P = 0.683]. Here, the large confidence intervals and our simulations suggest that a larger Wolbachia phylogeny (N = 25 and 50 Wolbachia strains) is required to detect a significant signal of λ > 0 (Supplementary Fig. 7). These results highlight differences in patterns of Wolbachia localization between the wMel- and wRi-like Wolbachia clades (Fig. 3). Closely related wMel-like Wolbachia tend to occur at a higher abundance in the posterior region of oocytes, whereas the wRi-like strains are generally more dispersed. We found no evidence that Wolbachia abundance in the whole oocyte exhibits a phylogenetic signal [λ < 0.001 (0, 0.753), P = 1].

The correlation between Wolbachia localization in oocytes and phylogenetic divergence implies that factors in the Wolbachia genome determine Wolbachia abundance at the oocyte posterior. Because these distributions likely involve direct interactions between Wolbachia and the host cytoplasm, we focused on 6 major Wolbachia surface proteins: the 3 wsp paralogs (WD_1063/wsp, WD_0009/wspB, and WD_0489/wspC), as well as WD_0501, WD_1041, and WD_1085 (Supplementary Table 3). We tested whether sequence divergence at any of these candidate loci predicts Wolbachia localization patterns. Specifically, we tested whether gene trees of the Wolbachia surface proteins (Supplementary Fig. 8) exhibit a phylogenetic signal, using the methods described above. One surface protein (WD_1085) stood out with especially strong evidence of a phylogenetic signal, with significant departures from λ = 0 for Wolbachia abundance at the posterior region [λ = 0.974 (0, 0.998), P = 0.001] and the posterior cortex [λ = 0.942 (0, 0.994), P = 0.028]. In addition, Wolbachia abundance at the posterior region also exhibited significant phylogenetic signal on the gene tree of WD_0501 [λ = 0.992 (0.521, 1), P = 0.022]. All tests for phylogenetic signal using Wolbachia gene trees are summarized in Supplementary Table 4.

Motivated by previous successful analysis of Wolbachia protein function through ectopic expression in yeast, we expressed codon-optimized forms of WD_1085 and WD_0501 in yeast using a galactose-inducible promoter (Sheehan et al. 2016; Rice et al. 2017). Under normal growth conditions, ectopic expression of both genes inhibited growth, with WD_1085 exhibiting a more pronounced inhibition (Supplementary Fig. 9). Because Wolbachia exhibits a close association with microtubules, we also assayed whether ectopic expression of these genes enhanced the growth defects in situations in which the microtubule spindle assembly checkpoint was compromised by a mad1 mutation (Luo et al. 2018), or when the microtubules were compromised directly through the addition of benomyl. As indicated in Supplementary Fig. 9, ectopic expression of WD_1085 significantly inhibited growth when both the microtubules and spindle checkpoint were compromised. These results suggest that ectopic WD_1085 expression perturbs microtubule function.

We recently found that the surface protein WD_0009 (wspB) is pseudogenized due to a premature stop codon in a tropical wMel variant sampled from Australia, which seems to influence wMel abundance in stage 10 oocytes when hosts are reared in the cold (Hague et al. 2022). Here, we did not find that the wspB gene tree exhibits phylogenetic signal related to Wolbachia oocyte abundance (Supplementary Table 4); however, we note that wspB appears to be pseudogenized in a number of other Wolbachia strains in addition to tropical wMel. We found derived deletions of varying length in the wspB sequences of wHa and wBai beginning at nucleotide position 256, both of which produce frame shifts that generate multiple downstream stop codons (Supplementary Fig. 10). In addition, wBoc and wLeo share a large insertion starting at bp position 351. Because the insertions contain a contig break, we were unable to resolve the full length of the insertion. Lastly, wBic contains a 104-bp insertion at bp position 647 that creates multiple downstream stop codons. These results suggest that the surface protein wspB has become pseudogenized at least 4 times in different Wolbachia lineages. Notably, the Wolbachia strains with a putatively pseudogenized version of wspB do not differ from the other strains in mean Wolbachia abundance in whole oocytes (Wilcoxon rank sum test, W = 36, P = 0.964), the posterior region (W = 37, P = 0.893), or the posterior cortex (W = 41, P = 0.622). This suggests that pseudogenization of wspB does not influence the diverse patterns of Wolbachia localization observed here. To gain insight into wspB cellular function, we ectopically expressed this gene (and the other wsp paralogs: WD_1063/wsp and WD_0489/wspC) in yeast using the galactose expression system described above. Using this system, we found that ectopic expression of codon-optimized wspB and wsp significantly inhibited yeast growth (Supplementary Fig. 11). Ectopic expression of wspC kills the cells. Based on similar ectopic expression studies (Sheehan et al. 2016), the suppression is likely due to Wolbachia proteins interacting with and disrupting the function of yeast components and cellular processes that are conserved and performing similar functions in insect cells.

Lastly, we ran similar phylogenetic analyses using the host phylogeny (as opposed to Wolbachia) to test whether host-associated factors might also contribute to the diverse Wolbachia distribution patterns in fly oocytes (Supplementary Fig. 1). We generally found that Wolbachia localization patterns do not exhibit a phylogenetic signal on the host tree. Wolbachia abundance in whole oocytes [λ = 0.365 (0, 0.856), P = 0.656], at the posterior region [λ < 0.001 (0, 0.658), P = 1], and at the posterior cortex [λ < 0.001 (0, 0.637), P = 1] did not exhibit phylogenetic signal using the host phylogram, implying that host-associated factors are not as important in determining diverse Wolbachia localization patterns in oocytes.

Vertical transmission in some host species likely relies on Wolbachia entering the germline from neighboring somatic tissues

Because Wolbachia strains with a dispersed localization do not localize to the pole, one would expect maternal transmission to be compromised. However, previous studies demonstrate that some of these Wolbachia, such as the B-group strain wMau in D. mauritiana, have high maternal transmission rates under standard laboratory conditions (i.e. constant 25°C), equivalent to that of wMel in D. melanogaster (Meany et al. 2019; Hague et al. 2022). These findings suggest that, in addition to strict germline-to-germline transmission, Wolbachia may utilize alternative routes of maternal transmission. Because numerous vertically inherited endosymbionts are transmitted via cell-to-cell transmission from the soma to the germline (Frydman et al. 2006; Russell et al. 2019), we explored whether Wolbachia that infect Drosophila exploits this strategy.

To determine the developmental stage during which somatic Wolbachia might invade the germline in dispersed species, we focused on B-group wMau in D. mauritiana. Initially, we examined the female germline (oocytes) of the 3rd instar larva for the presence of Wolbachia. For comparison, we also examined the oocytes of wMel-infected D. melanogaster (a strain with a posteriorly localized concentration of Wolbachia) 3rd instar larvae using anti-FtsZ to mark the Wolbachia and anti-Vasa to mark the germline. As shown in Fig. 5, Wolbachia are abundantly present in ring-like patterns in many of the Vasa-positive cells in the center of the developing oocyte. Imaging the oocytes of wMau-infected D. mauritiana revealed a similar pattern with Wolbachia concentrated in the developing female germline of the 3rd instar larva (Fig. 5). Images of wRi in D. simulans as well as wBic in D. bicornuta—strains with a dispersed localization and posteriorly clumped localization in mature oocytes, respectively—displayed the same ring-like structure in the developing oocyte of the 3rd instar larva, although they appeared to have a much higher titer than wMau or wMel (Fig. 5). This indicates that in these species, in which vertical transmission occurs with a few or no Wolbachia present at the posterior site of germline formation in mature adult oocytes, Wolbachia are able to occupy the germline at some point between the final stages of oocyte maturation and development of the fertilized egg into a 3rd instar larva. Presumably, the source of the Wolbachia is via invasion from neighboring somatic cells.

Fig. 5.

Wolbachia's presence around the 3rd instar larval germline, in multiple Drosophila hosts. wMel Wolbachia stained with anti-FtsZ can be seen for D. melanogaster, in a) shown in pink. Cells are marked with DAPI (blue). a′) Grayscale channel of anti-FtsZ staining. b) Uninfected ovary for D. melanogaster and b′) grayscale channel for uninfected antibody binding. Germline location marked with anti-Vasa shown in green c) and grayscale channel c′). wMau Wolbachia localization in the ovary for D. mauritiana shown in d) and grayscale channel d′). wRi and D. simulans shown in e) and grayscale e′). f) wBic and D. bicornuta and f′) grayscale channel. Scale bar shown at 25 µM.

To determine if Wolbachia occupied the germline prior to the 3rd instar larval stage, we examined the newly formed germline of late blastoderm and cellularized embryos. At this stage, newly formed germline cells, known as the pole cells, form a distinct cluster of cells at the extreme posterior to the embryo. In all the 4 strains described above, using PI or anti-Wolbachia staining, Wolbachia are readily found in these posterior germline cells (Fig. 6). Imaging wMau in D. mauritiana early blastoderm embryos revealed the presence of Wolbachia in the posterior pole cells. This observation narrows the time window in which Wolbachia invades and occupies the germline from between the final stages of oocyte maturation to early blastoderm formation. Images taken of D. mauritiana in 1-h-old embryos also showed wMau positioned in the posterior pole, narrowing the developmental window of Wolbachia germline occupation between the final stages of oocyte maturation to 1 h after fertilization (Fig. 6). Interestingly, Veneti et al. (2004) reported that wMau exhibits an anterior localization pattern during D. mauritiana embryogenesis; however, we found no such pattern for any of the embryos examined.

Fig. 6.

Wolbachia's presence in the newly formed germline cells of late blastoderm and cellularized embryos. a) One hour D. mauritiana (wMau) embryo stained with PI. a′) Grayscale of PI channel. b) Staining of 2–3 h developed D. mauritiana (wMau) embryo with PI. b′) Grayscale of PI channel. c) PI and anti-Vasa (green) staining of wMel Wolbachia in 2–3 h D. melanogaster embryo. c′) Grayscale of PI channel. d) Anti-FtsZ/DAPI staining for wRi Wolbachia in D. simulans. d′) Grayscale of anti-FtsZ channel. Scale bars set at 25 µM.

Our findings motivated us to examine wMau in D. mauritiana oocytes just prior to egg deposition. Previous studies found that in many infected Drosophila species, Wolbachia is concentrated in the most posterior positioned follicle cells (Kamath et al. 2018). These somatically derived cells, known as polar cells, directly contact the oocyte posterior pole plasm. In accord with Kamath et al. (2018), we find that Wolbachia are sporadically distributed in subsets of more anterior positioned follicle cells encompassing the oocyte. wMau presence in the polar follicle cells is a consistent feature in every D. mauritiana oocyte examined (Fig. 7). Given their position, these wMau-infected cells are likely the polar follicle cells. These observations raise the possibility that Wolbachia in the polar follicle cells may be a somatic source of germline wMau in D. mauritiana. In spite of much effort, however, we have not been able to find cytological evidence for this hypothesis.

Fig. 7.

The A-group wMel and B-group wMau Wolbachia strains are frequently found in the polar follicle cells, as well as other follicle cells. a)–c) Schematic of categories used for description of follicle cell Wolbachia. a) Dispersed localization pattern shown, both polar cells infected, with Wolbachia broadly present in follicle cells. b) Posteriorly localized pattern shown, 1 polar cell infected, with Wolbachia broadly present in follicle cells. c) Posteriorly clumped pattern shown, with no Wolbachia present in the polar cells or other follicle cells. d) Wolbachia quantification in the polar follicle cells during oogenesis stages 9 and 10. Wolbachia were quantified in midsections of each oocyte that displays unique posterior localization patterns: dispersed, posteriorly localized, and posteriorly clumped. For each strain, the count is shown for the number of oocytes observed with Wolbachia present in both polar follicle cells (orange), only 1 polar cell (blue), and neither polar cell (gray). e) Counts of oocytes observed with Wolbachia present anywhere in the follicle cells (pink) or with Wolbachia absent (gray). f) Representative anti-FtsZ staining of wMau in D. mauritiana (dispersed class) present in both polar follicle cells (corresponding to orange labels above). g) Representative staining of wMel in D. melanogaster (posteriorly localized class) present in only 1 polar cell (blue labels above). h) Representative staining of wBic in D. bicornuta (posteriorly clumped class) absent in both polar cells (gray labels above). i) Representative staining of wMel in D. melanogaster (posteriorly localized class) broadly present in anterior follicle cells (pink labeling above). j)–m) Grayscale anti-FtsZ channel of respective composite images above. n)–q) Five times zoom of polar cells, with circles highlighting representative staining in the area. Scale bar set at 25 µM.

To determine the extent to which Wolbachia polar follicle cell localization is conserved, we analyzed Wolbachia localization in the polar and more anterior follicle cells for 7 Wolbachia strains with the following patterns: dispersed (N = 3), posteriorly localized (N = 3), and posteriorly clumped (N = 1) (Fig. 7). We scored for the presence of follicle cell Wolbachia in each oocyte (N = 75) using the following criteria: the presence of Wolbachia in either both or 1 of the posterior polar follicle cells (Fig. 7d) and the presence of Wolbachia in the more anteriorly localized follicle cells (Fig. 7e). This analysis revealed that in a high percentage of oocytes from dispersed strains (a lack of the Wolbachia at the oocyte posterior), Wolbachia are frequently present in the posterior polar follicle cells. In contrast, with the exception of wMel, oocytes examined from strains in which Wolbachia are posteriorly concentrated (the posteriorly localized and clumped strains), few possess Wolbachia in the posterior follicle cells. The 7 Wolbachia strains exhibit significant variation in whether Wolbachia cells are found in the polar follicle cells (χ²₍₁₂₎ = 65.15, P < 0.001) (Fig. 7d). In particular, the divergent B-group strain wMau stood out with a large proportion of oocytes containing Wolbachia in the 2 polar follicle cells. The A-group wMel strain also had a large portion of oocytes with Wolbachia in 1 or 2 of the polar cells. We also found that the Wolbachia strains differ in whether Wolbachia cells are found in the other more anterior follicle cells (χ²₍₆₎ = 43.29, P < 0.001) (Fig. 7e). Again, the wMau and wMel strains had a large portion of oocytes with Wolbachia in the other more anterior follicle cells encompassing the oocyte. Of note, for wMau in D. mauritiana, every oocyte had at least 1 of their polar cells filled with Wolbachia, observationally at a higher titer (Fig. 7). This suggests that in the dispersed strains, vertical transmission may be achieved via invasion of Wolbachia from the polar follicle cells, particularly for the divergent B-group wMau strain.

Discussion

Vertical transmission through the female lineage is a common strategy for many endosymbionts. These endosymbionts achieve high efficiencies of maternal transmission either through direct germline-to-germline transmission or invasion of the maternal germline through neighboring somatic cells (Russell et al. 2019). Because the former involves a continuous presence in the germline while the latter requires navigating a soma-to-germline passage, it is expected that each requires distinct interactions with and manipulations of host cellular processes. Thus, it would be expected that a given endosymbiont would have evolved to utilize 1, but not both of these strategies. Here, we explore this issue by examining cellular aspects of Wolbachia vertical transmission. In contrast to expectations, our comprehensive analysis demonstrates that Wolbachia likely rely on both strict germline-to-germline and soma-to-germline vertical transmission strategies. Which strategy employed varies among Wolbachia strains.

The initial anterior Wolbachia localization in the developing oocyte appears to be a conserved aspect of Wolbachia transmission, as all 8 strains examined exhibit this localization. It is striking that this localization occurs prior to any known anterior determinant, indicating Wolbachia may associate with an as yet undiscovered host factor concentrated at the anterior. That this localization is conserved suggests that it has a functionally significant, but currently unknown role in Wolbachia transmission. While anteriorly localized, the amount of Wolbachia dramatically increases. Whether Wolbachia replication, transport from the nurse cells, or both are responsible for this increase is unknown. The anterior location may be a rich source of membrane for Wolbachia replication. The anterior localization also results in a high concentration of Wolbachia that closely associate with, and perhaps influences, the status of the oocyte nucleus.

To our surprise, we discovered that immediately after the release of wMel Wolbachia from its anterior position in the oocyte (stage 6), the amount of Wolbachia in the oocyte dramatically decreases (see Fig. 1, stage 6 and Supplementary Fig. 3). This suggests that the Wolbachia are transported back into the nurse cell complex. The functional significance of this clearing of Wolbachia from the oocyte is unknown. The retreat of Wolbachia occurs during the time when the anteriorly positioned oocyte nucleus and neighboring follicle cells signal one another in order to establish the dorsal ventral axis (Merkle et al. 2020). Perhaps, Wolbachia exits the oocyte in order to not disrupt this process. Given the orientation of microtubules at this stage, it is likely Wolbachia rapidly exits the oocyte through an association with the plus-end–directed motor protein kinesin.

Following microtubule reorientation and the return of Wolbachia to the oocyte, Wolbachia spreads posteriorly throughout the entire oocyte relying on the host motor protein kinesin (Serbus and Sullivan 2007). All of the Wolbachia strains examined are distributed toward the posterior regions suggesting the interaction with kinesin is also conserved. How Wolbachia engages these motor proteins is unknown. Surprisingly, none of the known host kinesin linker proteins are utilized by Wolbachia, suggesting Wolbachia may interact directly with kinesin (Russell et al. 2018).

It is after this stage, in which Wolbachia is distributed throughout the oocyte, that we observe variability among the Wolbachia strains, with each falling into 1 of 3 distinct classes: 2 distinct classes in which Wolbachia concentrate at the posterior (posteriorly localized and posteriorly clumped) and 1 class in which Wolbachia fail to exhibit a posterior concentration (dispersed). The former 2 classes are distinguished by whether the vast majority (posteriorly clumped) or only a small fraction of the Wolbachia localize to the posterior pole (posteriorly localized). Here, our phylogenomic analyses revealed a strong correlation between Wolbachia posterior abundance and Wolbachia phylogenetic relationships. Generally, we find that closely related wMel-like Wolbachia occur at higher abundance at the oocyte posterior, whereas wRi-like Wolbachia exhibit a more dispersed distribution (Fig. 4). These results suggest that factors intrinsic to Wolbachia help determine posterior localization patterns, which is consistent with previous Wolbachia transplantation studies demonstrating that posterior localization is determined by Wolbachia rather than host factors (Poinsot et al. 1998; Veneti et al. 2004; Serbus and Sullivan 2007). Below, we discuss how specific factors might contribute to variation in Wolbachia localization patterns.

The posteriorly localized class, in which only a small fraction localize to the posterior pole, can be explained by previous work demonstrating that wMel is a weak competitor for kinesin; both Wolbachia and germplasm components rely on kinesin for transport to the posterior pole (Russell et al. 2018). It is thought that Wolbachia has evolved to be a weak competitor because interference with the transport of essential germplasm components and thus germplasm formation may be disadvantageous for Wolbachia proliferation. The other posterior class (posteriorly clumped) is strikingly similar to that observed for wMel in D. melanogaster in which the plus-end motor protein kinesin is overexpressed and excess amounts of Wolbachia are transported to the posterior pole (Russell et al. 2018). It may be that the posteriorly clumped strains are much better competitors for host kinesin. Alternatively, these strains could occur in host species with naturally much higher levels of host kinesin, although our phylogenetic analysis did not support a role of host factors. Functional work will be required to formally test these hypotheses. In particular, it would be interesting to investigate whether there are instances in which Wolbachia disrupts the anterior–posterior (A-P) axis because of its accumulation at the posterior pole.

The posterior pole concentration among Wolbachia strains may also differ due to variability in the ability of Wolbachia to stably associate with the pole plasm. Previous studies demonstrated that Wolbachia posterior pole concentration requires intact pole plasm (Serbus and Sullivan 2007). For example, in oskar mutations, a key pole plasm determinant, wMel Wolbachia fails to accumulate at the posterior region (Hadfield and Axton 1999; Serbus and Sullivan 2007). As described, the oocyte experiences tremendous cytoplasmic streaming requiring posterior components to be anchored directly or indirectly to the cortex (Quinlan 2016). This variability could be due to differences in the ability of Wolbachia to compete with other pole plasm components such as mitochondria for binding sites. Alternatively, the host species may vary in the extent to which their pole plasm accommodates Wolbachia. These alternatives can be explored through trans-infection studies.

It is likely that the composition of Wolbachia surface proteins play a major role in the extent of posterior localization. Our phylogenetic analyses revealed that the gene trees of 2 candidate surface proteins, particularly WD_1085, are strongly associated with Wolbachia localization patterns. The WD_1085 protein has sequence and structural homology to the bacterial outer membrane protein BamA, which consists of a large periplasmic domain attached to a 16-stranded β-barrel domain (Supplementary Fig. 2). BamA is the central subunit of the β-barrel assembly machinery (BAM), which is essential for outer membrane protein biogenesis (Noinaj et al. 2013; Doyle and Bernstein 2019). Given this finding, it is intriguing that we find overexpression of BamA is sensitive to disruptions in microtubules, as these cytoskeletal elements are essential for outer membrane biogenesis in mitochondria (Mado et al. 2019). It may be that microtubules are also required for Wolbachia membrane biogenesis. While we highlight these loci as potential candidates, we note that our results should be interpreted with caution since the WD_1085 gene could be in linkage with other causal loci. Nonetheless, Wolbachia surface proteins are generally considered to be strong candidates for interactions with host cells (Werren et al. 2008; Baldridge et al. 2016; Hague et al. 2022).

Interestingly, our phylogenomic analyses did not find a correlation between host phylogenetic relationships and Wolbachia localization patterns, suggesting that host-associated factors are less important in determining Wolbachia posterior localization. This result contrasts with transplantation studies that indicate hosts may also play a significant role in determining Wolbachia abundance (Veneti et al. 2004). In addition, genome-wide screens reveal host factors play a major role in determining Wolbachia intracellular abundance (White et al. 2017b; Grobler et al. 2018). Thus, variation in the host proteins that Wolbachia engages could still plausibly influence Wolbachia localization patterns. This would contrast with other endosymbionts such as Listeria, which relies on a surface protein that binds and polymerizes host actin to propel the bacteria within and between host cells (Kühn and Enninga 2020). Because the interaction between ActA and actin is essential for cell-to-cell transmission, natural variants of this interaction and transmission strategy have not been discovered. In addition to Wolbachia and host factors, nutrients and other environmental factors have also been shown to influence Wolbachia abundance and localization in host oocytes for closely related wMel-like Wolbachia strains (Serbus et al. 2015; Hague et al. 2020a, 2022).

Two classes in which Wolbachia concentrates in the posterior pole (posteriorly localized and posteriorly clumped) provide a ready explanation for the cellular mechanisms by which it is vertically transmitted through generations. In every generation, Wolbachia targets the site of germline formation in the developing oocyte. This also explains why strains exhibiting this pattern, such as wMel, exhibit efficient transmission strategies under typical laboratory conditions (Hague et al. 2022). The 3rd, distributed class, that exhibits no-to-low Wolbachia in the posterior germplasm of the mature oocyte is puzzling. This would be expected to result in embryos lacking Wolbachia in the germline. However, within hours after fertilization, Wolbachia is clearly present in the germlines of both sexes (Fig. 6). This implies that either these Wolbachia strains are capable of invading the germline from neighboring somatic cells or a few undetected Wolbachia are delivered late in oogenesis to the posterior pole. Support for the 1st hypothesis comes from the fact that in filarial nematodes, Wolbachia germline invasion from the soma has been directly observed (Landmann et al. 2012). While germline invasion via cell-to-cell transmission has not been directly observed in insects, a number of lines of evidence indicate that it likely occurs. In Drosophila cell culture, Wolbachia efficiently undergoes cell-to-cell transmission (White et al. 2017a). Wolbachia injected into the abdomen of adult Drosophila females migrate to and occupy the germline and follicle stem cells (Frydman et al. 2006). Collections directly from nature reveal some strains with developing oocytes lacking Wolbachia (Casper-Lindley et al. 2011). In these strains, all of the mature oocytes were infected, suggesting an alternate route of infection via neighboring somatic cells.

As previously hypothesized, given the posterior follicle cells are directly adjacent to the pole plasm, we propose Wolbachia present in these follicle cells may be the source of germline Wolbachia (Kamath et al. 2018). Our results are in accord with and complement previous studies demonstrating Wolbachia concentrated in the posterior polar follicle cells in a number of Wolbachia-infected species (Kamath et al. 2018). For example, in wMau-infected D. mauritiana, a dispersed class, there is a striking concentration of Wolbachia in the posterior follicle pole cells directly adjacent to the pole plasm. Thus, Wolbachia in these somatic cells are well positioned to invade the pole plasm becoming incorporated into the germline of the next generation (Figs. 7 and 8). In contrast, in the posteriorly localized and clumped strains, we found no to few Wolbachia in the posterior polar follicle cells. We note that invasion of Wolbachia from posterior follicle cells into the germline would require Wolbachia to cross the seemingly impenetrable newly formed vitelline membrane, which could suggest alternative sources of germline Wolbachia. Nonetheless, taken together, the data strongly argue for a somatic-to-germline route of vertical transmission in some Drosophila species.

Fig. 8.

Multiple routes of Wolbachia vertical transmission. a) In Wolbachia strains that concentrate at the posterior pole plasm, vertical transmission likely occurs through a continuous maintenance in the germline. b) In strains where no or few Wolbachia localize to the posterior pole plasm, vertical transmission likely occurs through Wolbachia invading the germline from neighboring somatic cells. c) In theory, vertical transmission could occur through a symbiont invading the oocyte from neighboring somatic cells never associating with the germline.

Among the Wolbachia strains examined, wMel stands out as an exception. Not only is it concentrated at the posterior pole of the oocyte, but there is also a large Wolbachia concentration in the posterior polar follicle cells. Thus, wMel may maintain both robust germline-to-germline and soma-to-germline modes of transmission. Accordingly, it has a high efficiency of vertical transmission under standard laboratory conditions (but see Hague et al. 2022). Notably, we found that wMel and wMau are also both quite prevalent in the more anteriorly located follicle cells of stage 10 oocytes (Fig. 7), raising the possibility that vertical transmission could even occur without Wolbachia ever directly associating with the host germline (i.e. soma-to-soma transmission; Fig. 8).

Surprisingly, literature surveys reveal that germline invasion of endosymbionts from the soma every generation is the most common form of vertical inheritance (Russell et al. 2019). Vertical transmission via a continuous presence in the germline is much less common. One explanation is that hosts maintain mechanisms preventing endosymbiont occupation during the formative stages of germline development. In instances in which the endosymbiont is maintained in the germline, this may evolve into an obligate relationship in which development of the germline depends on the presence of the endosymbiont (Sullivan 2017). Examples of this include the leafhopper (Euscelis plebejus), in which the endosymbiont is required for normal embryonic development (Sander 1968). In addition, removal of Wolbachia from some insects results in increased apoptosis and abnormal oocyte development (Dedeine et al. 2001; Pannebakker et al. 2007). Here, we find that Wolbachia exhibit evidence of both the soma-to-germline and germline-to-germline transmission strategies (Fig. 8), suggesting that Wolbachia may stand out as a rarity among endosymbionts.

Data availability

Newly generated genomic data are available on GenBank (BioProject PRJNA907328). All other data are publicly available at figshare: https://doi.org/10.25386/genetics.22233817.

Supplemental material available at GENETICS online.

Funding

Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number R35GM139595 to WS and R35GM124701 to BSC. This work was also supported by National Science Foundation (NSF) grants to WS (1456535) and BSC (2145195).