The Wolbachia WalE1 effector alters Drosophila endocytosis

The most common intracellular bacterial infection is Wolbachia pipientis, a microbe that manipulates host reproduction and is used in control of insect vectors. Phenotypes induced by Wolbachia have been studied for decades and range from sperm-egg incompatibility to male killing. How Wolbachia alters host biology is less well understood. Previously, we characterized the first Wolbachia effector – WalE1, which encodes a synuclein domain at the N terminus. Purified WalE1 sediments with and bundles actin and when heterologously expressed in flies, increases Wolbachia titer in the developing oocyte. In this work, we first identify the native expression WalE1 by Wolbachia infecting both fly cells and whole animals. WalE1 appears as aggregates, separate from Wolbachia cells. We next show that WalE1 co-immunoprecipitates with the host protein Past1 and that WalE1 manipulates host endocytosis. Yeast expressing WalE1 show deficiency in uptake of FM4-64 dye, and flies harboring mutations in Past1 or overexpressing WalE1 are sensitive to AgNO3, a hallmark of endocytosis defects. Finally, we also show that Past1 null flies harbor more Wolbachia overall and in late egg chambers. Our results identify interactions between a Wolbachia secreted effector and a host protein and point to yet another important host cell process impinged upon by Wolbachia.

These proteins, referred to as effectors, often act to manipulate or usurp host cell processes to 45 promote bacterial infection (10, 11). These modes include (but are not limited to) attacking the 46 host cell surface to form pores, inactivating host cytosol machinery to collapse the 47 cytoskeleton, or entering the nucleus to manipulate host gene regulation (12). At each stage of 48 attack, the bacterial effectors often interact directly and specifically with host proteins to 49 perturb a biological process that enables pathogen entry into or defense from the host cell (13-50 19). Understanding how bacterial effectors function, therefore, has taught scientists not only 51 how pathogens cause disease, but also fundamental cell biological mechanisms work in healthy 52 tissue (20). While effectors are bacterial in origin, they act within eukaryotic cells and hence 53 often encode domains that share structural, functional, and sequence similarity with eukaryotic 54 proteins (10,11,21,22). 55 56 The first characterized Wolbachia effector, WalE1, is an actin bundling protein that increases 57 Wolbachia titer in the next generation when over-expressed in transgenic flies (8). The WalE1 58 protein contains an N-terminal synuclein domain (8) which may mediate some interactions with 59 host proteins and pathways. WalE1 is upregulated by Wolbachia during host pupation and 60 purified WalE1 protein co-sediments with filamentous actin and increases actin bundling in 61 vitro and in vivo. As the actin cytoskeleton is important for Wolbachia's maternal transmission 62 (23), and for its internalization by host cells (24), the WalE1 effector likely plays an important 63 role in Wolbachia's biology. 64 65 In this study, we sought to characterize the patterns of expression of native WalE1, its host 66 targets (beyond actin) and identify specific host pathways influenced by the effector. We used 67 antibody made to the full-length protein to visualize the effector during Wolbachia infection of 68 Drosophila cells and ovaries. The native protein appears as large aggregates in host cells and 69 early egg chambers. We use co-immunoprecipitation and mass spectrometry to determine host 70 proteins with which WalE1 interacts and identify Past1 as a target of WalE1. Past1 was 71 previously shown to influence endocytosis (25); Garland cells from homozygous Past1 mutant 72 larvae were defective in their ability to endocytose fluorescently labeled avidin. We therefore 73 characterized endocytosis defects upon both WalE1 expression in yeast and flies and Past1 74 abrogation in flies. Finally, we show an interaction between Wolbachia titers and Past1 titers in 75 whole animals, where Wolbachia titer increases in Past1 mutant flies, suggesting that 76 Wolbachia targets the protein to alter its function. Our results shed light on the molecular 77 mechanisms used by a ubiquitous symbiont to alter host biology. 78 79 Results: 80 81 Native WalE1 localizes to aggregates in the host cell cytosol 82 83 We had previously shown phenotypes for overexpression of RFP-tagged WalE1 in whole flies 84 (8). However, this experiment could have been affected by the use of tags and non-native 85 expression; tags can alter protein localization and we were necessarily studying an artificial 86 system and not native expression and secretion of the effector by Wolbachia. We therefore 87 generated an antibody to full length WalE1 and used it to probe cells and flies infected with 88 Wolbachia. The a-WalE1 antibody does not stain Drosophila JW18 cells cleared of their 89 Wolbachia infection with tetracycline. However, in JW18 cells infected with strain wMel, a-90 WalE1 localizes to large aggregates that do not colocalize with Wolbachia (stained with DAPI, 91 Figure 1). In ovaries, we observe similar aggregates of WalE1 in Wolbachia-infected flies only. 92 These aggregates appear most numerous in early stages of oogenesis (in the germarium and 93 stages 2-4), when Wolbachia is most numerous (23), and seem to disappear in later egg 94 chambers ( Figure 2). 95 96 WalE1 interacts with host Past1 and the endocytosis pathway 97 98 Because native WalE1 localized to aggregates in the cell, we sought to identify what host 99 proteins may be localized to those aggregates, interacting with WalE1. Towards that end, we 100 performed a co-immunoprecipitation experiment using a-WalE1 antibody. Infected Drosophila 101 JW18 cells were lysed and incubated with a-WalE1 antibody-coated magnetic beads. As 102 controls we included beads alone and uninfected JW18 cells were also subjected to the same 103 immunoprecipitation. After washes on a magnetic column, bound proteins were eluted and 104 subjected to mass spectrometry. We identified one protein, significantly enriched in our pull 105 down, which came down in the JW18 lysate but was absent from the JW18-tet lysate: Putative 106 Achaete Scute Target 1 (Past1) ( Past1 mutants were previously characterized to suffer from fertility defects, exhibit sensitivity 123 to temperature, and die early (25). We began by acquiring the most commonly used alleles 124 (Past1 60-4 and Past1 110-1 ); both of these alleles were made by imprecise excision of a p-element 125 insertion, which result in Past1 nulls (25). Our first western blot experiments confirmed that 126 these mutants are indeed nulls for Past1 ( Figure S1). We next focused on using deficiency 127 stocks (Df), containing chromosomal ablations, covering the Past1-containing region, to confirm 128 phenotypes previously published for these alleles (25). The Df used in the previous study was 129 not available to us (Df(3R)Kar-Sz37 87C5-87D14), so we used two deficiencies from the 130 Bloomington Drosophila Stock Center, both of which covered Past1 entirely, and the larger of 131 which (Df(3R)BSC486) led to greater levels of lethality. The smaller deficiency (Df(3R)Kar-Sz29) 132 allowed for high levels of viability in animals lacking Past1, and in our extensive study of ovaries 133 from these flies to study Wolbachia titer effects, oogenesis defects were never observed 134 ( Figure 4). These results suggest that something else in the background of the Past1 60-4 and 135 Past1 110-1 flies contributes to the viability and fertility defects previously observed. Importantly, 136 prior work did not specify if the allele used (Past1 110-1 ) was in a whitebackground. The P-137 element used (EY01852) carries mini-white + , and white mutants impair several biological 138 functions -from mobility to lifespan to stress tolerance (31). Therefore, it is entirely possible 139 that the viability and fertility defects previously published were confounded by the background 140 of these mutations. 141 142 We sought to genetically isolate any effects that may be related to Past1 in our assays and 143 therefore generated new CRISPR-Cas9 based mutants in Past1 (Past1 M10 and Past1 F15 ). These 144 mutants generate an identical two basepair deletion that generates a frameshift and eventual 145 stop codon after 83 amino acids, resulting in production of no stable protein ( Figure S1). We 146 used transheterozygotes of the previously generated allele Past1 60-4 and our two CRISPR 147 generated null alleles (Past1 M10 and Past1 F15 ) in these experiments. Our genetic controls were 148 progenitor flies from the Cas9 process, that are wild-type for Past1. 149 150 Past1 mutant flies exhibit endocytosis defects 151 152 Previously, endocytosis defects were observed for past1 mutant flies (25). In that prior study, 153 the authors had used Drosophila Garland cells to evaluate endocytosis in the fly as they are the 154 equivalent of a kidney, filtering the fly haemolymph. One straightforward way to test for fly 155 Garland cell function is to expose flies to silver nitrate (AgNO3); if the fly harbors mutations that 156 alter function of the Garland cells, they will be more sensitive to AgNO3 toxicity (32). We 157 therefore subjected larvae expressing walE1 and larvae carrying mutations in Past1 to AgNO3 158 throughout all of development and monitored viability over time, counting the number of 159 pupae and adults derived from these conditions. Control flies did not show any phenotype 160 when reared in the presence of AgNO3 and indeed, no statistical differences were observed 161 across all fly backgrounds reared in normal food without AgNO3. However, flies expressing the 162 Wolbachia effector were exquisitely sensitive to AgNO3, to the same extent as those carrying 163 Past1 null mutations ( Figure 5). While 90% of control flies survive AgNO3, WalE1 overexpression 164 resulted in 5% survival on average ( Figure 5). This result confirms prior observations regarding 165 the effect of Past1 on endocytosis (25) and supports a role for WalE1 in modifying endocytosis 166 as well. Importantly, all flies used in this experiment did not have the white mutation nor 167 balancer chromosomes, to remove confounding background effects (Supplementary Table 2). 168 169 Wolbachia attain higher titer in Past1 null flies 170 171 Because we identified Past1 as a potential target of WalE1, and because both proteins are 172 involved in modifying endocytosis in Drosophila, we wondered if there would be an interaction 173 between Past1 null mutants and Wolbachia titer. We reasoned that if Wolbachia was using 174 WalE1 to modify Past1, reduction of the dosage of Past1 might influence Wolbachia biology. 175 Therefore, we began by examining Wolbachia titer in flies with different copy numbers of 176 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ; Past1. We used western blot targeting the Wolbachia surface protein and observed a clear and 177 statistically significant effect of Past1 dosage on Wolbachia titer (ANOVA; df = 2, F = 40.611, p < 178 0.001). Flies that are null for Past1 have the highest Wolbachia titer (t = -7.675, df = 13, p < 179 0.001) followed by flies with a half dose of Past1, although the titer in these flies is not 180 significantly different from wild-type (t = -2.632, df = 10, p = 0.457) ( Figure 6). 181 182 We next wondered if this whole-animal increase in Wolbachia load would be observed in the 183 ovary tissue as well, the tissue best studied for Wolbachia titer and localization. We dissected 184 ovaries from Past1 60-4 /Df(3R)Kar-Sz29 flies and sibling controls with a half dose of Past1. 185 Wolbachia titer was quantified by Wolbachia-specific antibody staining (a-FtsZ) in ovarioles as 186 indicated in the methods. We noticed that Wolbachia staining was most intense in the 187 germarium for both backgrounds (as expected) but as oogenesis progressed, flies null in Past1 188 harbored more Wolbachia in older egg chambers (Figure 7, GLM c2 = 10.477; df = 3; p = 0.001). 189 The most differentiation in Wolbachia titer between these backgrounds was observed for 190 stages 7-8 ( Figure 7; GLM c2 = 39.449; df = 3; p < 0.001). This result is reminiscent of the 191 increase in Wolbachia titer observed upon over-expression of WalE1 in our prior work (8) Proteins were separated on 4-20% Tris-Glycine NB precast minigels (NuSep) and transferred to 263 PVDF membrane in Tris-Glycine transfer buffer with 15% methanol at 40v on ice for 3 hours. 264 The membranes were blocked for 5 minutes in SuperBlock™ (TBS) Blocking Buffer 265 (ThermoFisher Scientific), followed by incubation with antibodies diluted in SuperBlock™ (TBS) 266 Blocking Buffer (for 1 hour at room temperature or overnight at 4 o C) according to standard 267 protocols. PageRuler Prestained Protein Ladder (ThermoFisher Scientific) was used as a 268 molecular mass marker. Antibodies utilized include mouse anti-actin at 1:10,000 (Seven Hills 269 Bioreagents, LMAB-C4); rabbit anti-Past1 at 1:1000 (from Mia Horowitz at Tel Aviv University 270 (25)) and mouse anti-Wsp at 1:10,000 (BEI Resources, Inc. NR-31029). F(ab')2-Goat anti-Rabbit 271 IgG (H+L) (A24531) and goat anti-mouse IgG (G-21040) secondary antibodies conjugated to 272 horseradish peroxidase (HRP) were used at 1:5,000 (ThermoFisher Scientific Invitrogen). 273 SuperSignal West Pico Chemiluminescent Substrate was used to detect HRP on immunoblots 274 per manufacturer instructions (ThermoFisher Scientific). Blots were re-probed after stripping 275 (100mM Glycine, 0.15 ND-40, 1% SDS, pH 2) for 1 hour at room temperature, then overnight or 276 up to three days at 4 o C. The rabbit Past1 anti-sera also detects a protein band of ~ 50 KD in 277 wMel-containing lysates that is not observed when the sera are preabsorbed with strips of 278 PVDF membrane containing total E. coli lysate. 279 280 Drosophila cell and ovary immunochemistry and microscopy 281 282 The Drosophila melanogaster JW18 cell line, naturally infected with Wolbachia strain wMel, 283 was used to visualize WalE1 in an infection (40). JW18-TET cells were passaged with 10 ug/ul 284 tetracycline added to the culture medium and were used as an uninfected control cell line. 285 Confluent monolayers were harvested from 25 cm 2 non-vented tissue culture flasks, counted 286 using a disposable hemocytometer (Fisher Scientific), and overlaid as 100 ul of 2 x 10 6 or 200 ul 287 of 1 x 10 6 cells onto Concanavalin-A coated No. 1.5 coverslips (ConA 0.5 mg/ml applied and 288 dried on to sterile, acid-washed coverslips and leaving a 2 mm ConA-free border) (41, 42). Cells 289 were allowed to settle, attach, and spread out overnight before cells were fixed in 4% 290 paraformaldehyde in PBS for 20 minutes, followed by four 1 mL washes of 1X PBST (0.2% 291 Tween-20 added to 1X PBS). Coverslips were then exposed to blocking solution (PBST 0.2% 292 Tween-20 and 0.5% BSA) for 15 minutes at room temperature before transfer to primary 293 antibody (at concentrations indicated below) overnight in blocking solution at 4 o C. In the 294 morning, coverslips were washed in PBST and then allowed to incubate for 1-2 hours at room 295 temperature in the dark in 100 microliters of secondary antibodies diluted in the blocking 296 solution. Coverslips were washed for 5 minutes 3-4 times with one mL PBST, followed by a final 297 dip in a 500 mL beaker of distilled water. Excess moisture was wicked off the edge of the 298 coverslip with a tissue followed by mounting in 10 microliters of Prolong Gold Antifade Reagent 299 with DAPI (Invitrogen) per coverslip on glass slides. 300 301 302 For immunolocalization of proteins in developing egg chambers, ovaries from mated four-day 303 old females were dissected in cold PBS before transfer to 1.5 mL microcentrifuge tubes with 304 100ul Devitt's solution (5.3% paraformaldehyde, 0.5% NP-40, 1X PBS) and 600 ul heptane (43). 305 The ovary tubes were vigorously shaken by hand for 30 seconds to create an emulsion, before 306 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 rocking for 10 minutes at room temperature. After three rinses of ovaries, gently pelleted by 5-307 second spins in a mini microcentrifuge (6000 rpm, 2,000g), in PBST (0.2% Tween-20 added to 1X 308 PBS), the ovaries were soaked in blocking solution (PBST 0.2% Tween-20 and 0.5% BSA) for 15 309 minutes at room temperature before addition of primary antibody and overnight incubation 310 rocking at 4 o C. The next day, four 500-750 microliter rinses with blocking solution, after 311 pelleting with the mini microcentrifuge, preceded the addition of secondary antibodies in the 312 dark, for two hours, at room temperature. Three short rinses with blocking solution after 313 pelleting then followed before mounting 5 to 10 ovary pairs. The ovaries were teased apart on 314 glass slides with tungsten needles or insect pins, and excess buffer was carefully removed by 315 wicking with a tissue before addition of room temperature Prolong Gold Antifade Reagent with 316 DAPI (Invitrogen) and No 1.5 glass coverslips. 317 318 Full-length wMel WalE1 and wAna FtsZ protein-encoding constructs were synthesized by 319 GenScript using codons optimized for E. coli and cloned into pCR-TOPO before subcloning into 320 the pUC57 expression vector (using BamHI and XhoI restriction enzymes). Sequence-verified 321 constructs were expressed in E. coli BL21* and purified using Ni-NTA columns (Sigma-Aldrich). 322 Because pUC57 contains a TEV cleavage site, N-terminal His-tags was removed from these 323 proteins before sending them to Cocalico Biologicals for antibody generation. Rabbit polyclonal 324 antibody sera against both full-length purified proteins were generated (Cocalico Biologicals, 325 Inc) and used separately for immunohistochemistry (1:500 and 1:150, respectively). Secondary 326 antibodies to rabbit and mouse IgGs that were highly purified to reduce cross reactivity were 327 used with 488, 594 and 647 AlexaFluor conjugates at 1:1000 dilution in PBST with 0.5% BSA 328 (Invitrogen ThermoFisher A11070 Images were taken as Z-series stacks at 0.3 to 1.0 um intervals using a Nikon Ti2 fluorescent 334 microscope with 60x oil objective and processed using NIS Elements software (Nikon). Care 335 was taken such that exposure times were normalized across all experimental conditions. For 336 quantification of Wolbachia within the developing oocyte, maximum projections stacks 337 generated were used, excluding the peritoneal sheath. The irregular blob tool was used to 338 outline entire egg chambers, using cortical actin staining as a guide, and DAPI DNA staining 339 along with Actin staining was used to determine egg chamber stages. 340

Co-Immunoprecipitation and Mass Spectrometry 341
Confluent wMel-infected JW18 cells and uninfected TET-JW18 cells were harvested from 25 342 cm 2 , 50 mL non vented tissue-culture culture flasks. Cells were pelleted and washed three times 343 with 1 ml volumes of PBS to remove culture medium and each cell type pellet was resuspended 344 in 200 ul Lysis Buffer II (10mM Tris pH 7.4, 150mM NaCl, 10mM NaH2PO4, 1% Triton X-100) 345 with added 1X Halt™ Protease Inhibitor Cocktail (ThermoFisher Scientific) and 5 mM EDTA on 346 ice for 10 minutes, vortexing every 2 minutes. Debris and nuclei were pelleted at 10K rpm for 347 10 minutes in a microfuge (9391 g), 4 o C rotor. Five microliters of rabbit pAb-WalE1 antisera 348 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made repopulated with other microbiome members as previously reported (45). 388 Drosophila silver nitrate toxicity assays 389 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 Sets of 10-20 virgin females (uninfected, w[+] eyes) and 15-30 male flies were mated in small 390 cages on standard fly food for two days and then allowed to lay eggs for 24 hours on grape 391 agar (Genesee Scientific) plates (60 x 15 mm) with a little pile of yeast paste (1 g live baker's 392 yeast per 2 ml water) for 4-5 days. The 24 hour old plates were incubated at room temperature 393 in a moist chamber another day until 48 hours, at which point 1st instar larvae were collected in 394 a 100 micron disposable sieve, rinsed with room temperature water to remove food and debris, 395 and were placed in groups of 10 larvae per plastic vial of 5 milliliters of minimal medium (5 g 396 agar, 5 g dextrose, 360 milliliter distilled water, solubilized using heat and aliquoted) with ~60 397 microliters of yeast paste made with water or a 0.003 % silver nitrate solution (Ag2NO3) applied 398 to the surface of the minimal medium (based on method described in (32) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ; 419 420 Figure 1. WalE1 native expression in Wolbachia-infected JW18 cells. Antibody made against full-length WalE1 was used to probe Wolbachia-infected (JW18) and uninfected (JW18-tet) cells. Native WalE1 expression is only seen in JW18 cells, where the protein localizes in large aggregates, outside of Wolbachia cells (stained with DAPI). Arrowheads point to WalE1 aggregates in overlay.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ; https://doi.org/10.1101/2023.02.26.530160 doi: bioRxiv preprint Figure 2. WalE1 native expression in Wolbachia-infected fly ovarioles. Antibody made against fulllength WalE1 was used to probe Wolbachia-infected and uninfected ovaries. Native WalE1 expression is seen as green puncta in Wolbachia-infected germaria through stage 9 egg chambers.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  Figure 3. Yeast expressing WalE1 internalize less FM464 dye than control yeast carrying vector alone. Yeast carrying control vector or expressing WalE1 were exposed to FM464 membrane dye for a pulse and internalization of dye was quantified after 60 minutes based on fluorescence intensity (N > 40 for each; c2= 24.706, df = 1, p = 6.676e-07). . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ; https://doi.org/10.1101/2023.02.26.530160 doi: bioRxiv preprint Figure 4. Past1 null mutants are viable. Male genotype for each cross is shown at top right, and females on the lefthand side. Total number of flies counted is noted in parentheses and viability is calculated as a percentage. Severely reduced viability is observed with the large deficiency Df(3R)BSC486, which covers adjacent genes, but other combinations of hemizygous past1 null alleles lead to full or higher percentages of viable offspring. Past1 nulls shown in light orange.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023.  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 27, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023