bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 available under aCC-BY 4.0 International license.

1 Dual proteomics of hemolymph infected

2 with the heritable poulsonii

† † 3 Florent Masson , Samuel Rommelaere , Alice Marra, Fanny Schüpfer, Bruno Lemaitre*

4 Global Health Institute, School of Life Sciences, École Polytechnique Fédérale de Lausanne (EPFL),

5 Lausanne, Switzerland

6

† 7 These authors contributed equally

8

9 *Corresponding author:

10 Phone number: +41 21 693 18 31

11 Email address: [email protected]

12

13 Running title: Drosophila/Spiroplasma proteomics

14

15 Key words: Spiroplasma, endosymbiosis, proteomics

16

17 ORCID numbers:

18 Florent Masson: 0000-0002-5828-2616

19 Samuel Rommelaere: 0000-0003-3201-0847

20 Bruno Lemaitre: 0000-0001-7970-1667 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 available under aCC-BY 4.0 International license.

21 Abstract

22

23 Insects are frequently infected with heritable bacterial .

24 Endosymbionts have a dramatic impact on their host physiology and evolution. Their

25 tissue distribution is variable with some species being housed intracellularly, some

26 extracellularly and some having a mixed lifestyle. The impact of extracellular

27 endosymbionts on the biofluids they colonize (e.g. insect hemolymph) is however

28 difficult to appreciate because biofluids composition can depend on the contribution

29 of numerous tissues.

30 Here we address the question of Drosophila hemolymph proteome response to the

31 infection with the endosymbiont Spiroplasma poulsonii. S. poulsonii inhabits the fly

32 hemolymph and gets vertically transmitted over generations by hijacking the

33 oogenesis in females. Using dual proteomics on infected hemolymph, we uncovered

34 a low level and chronic activation of the Toll immune pathway by S. poulsonii that

35 was previously undetected by transcriptomics-based approaches. Using Drosophila

36 genetics, we also identified candidate proteins putatively involved in controlling S.

37 poulsonii growth. Last, we also provide a deep proteome of S. poulsonii, which, in

38 combination with previously published transcriptomics data, improves our

39 understanding of the post-transcriptional regulations operating in this bacterium.

40

41 Summary statement

42

43 We report the changes in Drosophila melanogaster hemolymph proteome upon

44 infection with the heritable bacterial endosymbiont Spiroplasma poulsonii.

1 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

45 Introduction

46

47 Insects frequently harbor vertically transmitted bacterial symbionts living within their

48 tissues, called endosymbionts (Douglas, 2015). Endosymbionts have a major impact

49 on the host physiology and evolution as they provide ecological benefits such as the

50 ability for the host to thrive on unbalanced diets (Douglas, 2016), protection against

51 viruses or parasites (Hedges et al., 2008; Oliver et al., 2003; Scarborough et al.,

52 2005; Teixeira et al., 2008) or tolerance to heat (Montllor et al., 2002). A peculiar

53 group of insect endosymbionts also directly affects their host reproduction. This

54 group is taxonomically diverse and includes from the Wolbachia,

55 Spiroplasma, Arsenophonus, Cardinium and Rickettsia genera (Duron et al., 2008;

56 Medina et al., 2019). Four reproduction-manipulative mechanisms have been

57 unraveled so far, namely cytoplasmic incompatibility, male-killing, parthenogenesis

58 and male feminization (Werren et al., 2008), all of them leading to an evolutionary

59 drive that favors infected individuals over non-infected ones and promotes the

60 endosymbiont spread into populations. Their ease of spread coupled with the virus-

61 protecting ability of some species (Hedges et al., 2008; Teixeira et al., 2008) make

62 them promising tools to control insect pest populations or to render them refractory to

63 human arboviruses (Hoffmann et al., 2011). Reproductive manipulators are however

64 fastidious bacteria that are difficult to manipulate, hence slowing down research on

65 the functional aspects of their interaction with their hosts (Masson and Lemaitre,

66 2020). Their real impact on host physiology thus remains largely elusive.

67

68 Spiroplasma are long helical bacteria belonging to the class, which

69 are devoid of cell wall (Gasparich, 2002). They infect a wide range of arthropods and

2 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

70 plants where they act as pathogens, commensals or vertically transmitted

71 endosymbionts depending on the species (Gasparich, 2002). S. poulsonii (hereafter

72 Spiroplasma) is a vertically transmitted endosymbiont infecting the fruit fly Drosophila

73 (Haselkorn, 2010; Mateos et al., 2006). It lives free in the host hemolymph where it

74 feeds on circulating lipids (Herren et al., 2014) and achieves vertical transmission by

75 infecting oocytes (Herren et al., 2013). Most strains cause male-killing, that is the

76 death of the male progeny during early embryogenesis through the action of a toxin

77 named Spaid (Harumoto and Lemaitre, 2018). Spiroplasma infection also confers

78 protection to the fly or its larvae against major natural enemies such as parasitoid

79 wasps (Ballinger and Perlman, 2017; Xie et al., 2014) and nematodes (Hamilton et

80 al., 2016; Jaenike et al., 2010).

81

82 Spiroplasma has long been considered as untractable, but recent technical

83 advances such as the development of in vitro culture (Masson et al., 2018) and

84 transformation (Masson et al., 2020b) make it an emergent model for studying the

85 functional aspects of insect-endosymbiont interactions. Some major steps have been

86 made in the understanding of the male killing (Harumoto et al., 2014; Harumoto et al.,

87 2016; Harumoto et al., 2018; Veneti et al., 2005) and protection phenotypes

88 (Ballinger and Perlman, 2017; Hamilton et al., 2016; Paredes et al., 2016) or on the

89 way the bacterial titer was kept under control in the adult hemolymph (Herren et al.,

90 2014; Paredes et al., 2015). Such studies, however, relied on single-gene studies

91 and did not capture the full picture of the impact of Spiroplasma on its host

92 physiology. We tackled this question using dual-proteomics on fly hemolymph

93 infected by Spiroplasma. This non-targeted approach allowed us to get an extensive

94 overview of the end-effect of Spiroplasma infection on the fly hemolymph protein

3 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

95 composition and to identify previously unsuspected groups of proteins that where

96 over- or under-represented in infected hemolymph. Using Drosophila genetics to

97 knock-down the corresponding genes, we identified new putative regulators of the

98 bacterial titer. This work also provides the most comprehensive Spiroplasma

99 proteome to date. By comparing this proteome to the existing transcriptomics data,

100 we draw conclusions about Spiroplasma post-transcriptional regulations.

4 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

101 Results

102 1. Drosophila hemolymph protein profiling

103 We investigated the effect of Spiroplasma infection on Drosophila hemolymph protein

104 content using Liquid Chromatography-tandem Mass Spectrometry (LC-MS/MS). To

105 this end, we extracted total hemolymph from uninfected and infected 10 days old

106 females. At this age, Spiroplasma is already present at high titers in the hemolymph

107 but does not cause major deleterious phenotypes to the fly (Herren and Lemaitre,

108 2011; Masson et al., 2020a). Extraction was achieved by puncturing the thorax and

109 drawing out with a microinjector. This procedure induces some tissue damage but

110 recovers very few hemocytes (circulating immunes cells).

111 Peptides were mapped to both the Drosophila and Spiroplasma predicted proteomes,

112 hence allowing having an overview of i) differentially represented Drosophila proteins

113 in infected versus uninfected hemolymph and ii) an in-depth Spiroplasma proteome

114 in the infected hemolymph samples. These two datasets will be analyzed in separate

115 sections.

116 The mapping on Drosophila proteome allowed the identification of 909 quantified

117 protein groups (a protein group contains proteins that cannot be distinguished based

118 on the identified peptides). The complete list of quantified Drosophila proteins and

119 their fold-change upon Spiroplasma infection is available in Supplementary Table S1.

120 The hemolymph extraction process involves tissue damage (e.g. the subcuticular

121 epithelium, muscles and fat body) that leads to the release of intracellular proteins in

122 the sample. Hence we first sorted protein groups depending on whether the main

123 protein is predicted to be extracellular or not. Proteins were defined as extracellular if

124 they bore a predicted signal peptide or were annotated with a Gene Ontology (GO)

125 term “extracellular”, or intracellular if no signal peptide nor any “extracellular” GO

5 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

126 term was predicted. Based on this localization criteria the Drosophila dataset could

127 be split in 555 intracellular protein groups representing 61% of the total proteome and

128 354 extracellular proteins representing the remaining 39% of the total proteome

129 (Figure 1A).

130

131 We then compared the relative amounts of protein groups between infected and

132 uninfected hemolymph. Of the intracellular protein groups, 35 were differentially

133 represented in the Spiroplasma infected samples, including 8 overrepresented and

134 27 depleted groups compared to the uninfected control (Figure 1B). A functional GO

135 analysis on the differentially represented protein groups revealed that a significant

136 part of these protein groups were related to the proteasome (subunits α2, 3, 4 and 7

137 and subunits β3 and 6) and the Toll immune pathway (Figure 1C). The Toll pathway

138 GO enrichment comprises only proteasome subunits, probably because of their

139 involvement in the degradation of Toll pathway intermediates (e.g. Cactus

140 (Palombella et al., 1994)). While differentially abundant intracellular protein groups

141 were mostly underrepresented in Spiroplasma infected hemolymph, analysis of the

142 extracellular protein subset revealed an opposite trend. Among the 71 extracellular

143 protein groups having a significantly different abundance, 4 were depleted and 67

144 were overrepresented in the infected samples compared to the uninfected ones

145 (Figure 1D). Surprisingly, the functional GO annotation highlighted an

146 overrepresentation of serine-endopeptidase molecular function. Serine proteases are

147 involved in the regulation of the melanization and the Toll pathways and indeed the

148 GO terms are enriched for these biological processes as well (Figure 1E).

6 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

149 2. Immune-related protein enrichment is a consequence of a mild

150 transcriptional activation

151

152 Since Spiroplasma is devoid of cell wall, it lacks microbe-associated molecular

153 patterns such as peptidoglycan and was not expected to interact with pattern-

154 recognition receptors regulating the fly immune system. This idea was supported by

155 previous work showing that Spiroplasma do not trigger a strong level of Toll and Imd

156 pathway, the two main immune pathways in Drosophila (Herren and Lemaitre, 2011;

157 Hurst et al., 2003). Previous work also showed that flies defective for the inducible

158 humoral immune response have normal Spiroplasma titer, suggesting that immune

159 pathways are not required to control Spiroplasma growth (Herren and Lemaitre,

160 2011). Our observation that several immune-related proteins are more abundant in

161 the hemolymph of infected flies (Figure 2A and Supplementary Table S1) led us to

162 investigate further on this point. We first tested if the changes in immune-related

163 protein abundance were a consequence of altered gene expression. We measured

164 the corresponding mRNA levels of several differentially represented proteins (Figure

165 2B). Immune-related genes were slightly upregulated in infected flies, although their

166 induction levels did not compare with those observed after systemic infection by a

167 pathogenic bacteria (De Gregorio et al., 2001; Troha et al., 2018). These results

168 confirm that Spiroplasma does not trigger a strong immune response (Herren and

169 Lemaitre, 2011; Hurst et al., 2003). But our data reveals that the presence of

170 Spiroplama still provokes a mild and chronic production of immunity proteins.

171

172 3. A majority of enriched hemolymphatic proteins are not involved in

173 Spiroplasma growth control

7 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

174

175 A systemic infection or a genetic induction of the humoral immune response

176 promotes Spiroplasma growth in flies (Herren and Lemaitre, 2011). We therefore

177 wondered if the presence of small amounts of antimicrobial peptides (AMPs) could

178 be beneficial for Spiroplasma growth. To test this hypothesis, we measured

179 Spiroplasma titer in flies over-expressing different AMP genes (Tzou et al., 2002).

180 We found that AMP gene overexpression did not increase Spiroplasma titer (Figure

181 3A). Similarly, flies lacking ten AMP coding genes (Hanson et al., 2019) had a

182 Spiroplasma titer comparable to that of control flies (Figure 3B). These results

183 suggest that AMPs released during the humoral immune response do not promote

184 Spiroplasma growth.

185 We then tested if other immune-related proteins found more abundant in infected

186 hemolymph could alter Spiroplasma growth. We first compared the Spiroplasma

187 titers of control OregonR flies with that of hayan, attD and tep2 mutants. We observed

188 no difference in bacterial titer, indicating that these genes are not required neither

189 detrimental to Spiroplasma growth (Figure 3B). We then used an in vivo RNAi

190 approach to silence immune genes coding for the most enriched proteins in

191 Spiroplasma-infected flies using the ubiquitously expressed Act5C-GAL4 driver.

192 Silencing these genes did not provoke an increase in Spiroplasma titer, further

193 reinforcing the idea that immune proteins do not prevent Spiroplasma growth (Figure

194 3C).

195 Among the most differentially represented proteins in Spiroplasma infected

196 hemolymph, we also detected several proteins of unknown function. To test their role

197 in Spiroplasma growth control, we used the same RNAi-mediated knockdown

198 approach and quantified Spiroplasma titer in these flies. Most of the RNAi lines

8 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

199 tested showed no change in Spiroplasma titer as compared to control lines

200 (Supplementary Figure S1). Two RNAi lines targeting CG18067 and CG14762,

201 respectively, showed a slight yet not significant reduction in Spiroplasma titer. These

202 results suggest that all the tested genes do not facilitate Spiroplasma growth nor

203 control its titer. Instead, these proteins are likely to be induced as a consequence of

204 Spiroplasma infection and may participate in the host adaptation to the bacteria. A

205 mutant line was available for only one of the most enriched uncharacterized proteins

206 (CG15293). We found that this mutation leads to a significant increase in bacteria

207 titer (Supplementary Figure S1). This gene codes for a 37 kDa secreted protein with

208 no known function. Our results raised the hypothesis that CG15293 participates in

209 the control of Spiroplasma titer in vivo.

210

211 4. Transcriptome-proteome correlation in Spiroplasma poulsonii

212

213 The proteomics analysis of infected hemolymph samples also allowed the detection

214 and quantification of Spiroplasma proteins. With a total of 503 proteins, this is the

215 most comprehensive Spiroplasma proteome to date. The complete list of

216 Spiroplasma quantified proteins is available in Supplementary Table S2.

217 We took advantage of a previously published transcriptomics dataset of Spiroplasma

218 (Masson et al., 2018) to compare the expression level of Spiroplasma genes to their

219 corresponding protein abundance by building a linear model of the proteomics signal

220 as a function of the transcript level (Figure 4). The two measures are poorly

221 correlated (Kendall’s τ = 0.40). Analyzing the normalized residuals of the model also

222 allowed us to identify proteins that are particularly discrepant with the model, that is

223 with absolute standardized residuals greater that 2. This approach uncovered 27

9 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

224 proteins, of which 11 have a significantly higher abundance and 16 a lower

225 abundance in the proteome that what was predicted from the transcriptomics data

226 (Supplementary Table S3). Over-represented proteins with a reliable annotation

227 include the membrane lectin Spiralin B, the cytoskeletal protein Fibrilin, the glucose

228 transporter Crr, the potassium channel KtrA, a ferritin-like protein Ftn, the

229 chromosome partitioning protein ParA and the DNA polymerase subunit DnaN.

230 Under-represented proteins include the transporters SteT and YdcV, the

231 serine/threonine phosphatase Stp, the helicase PcrA, the ATP synthase subunit

232 AtpH, the GMP reductase GuaC and the tRNA-(guanine-N(7)-)-methyltransferase

233 TrmB. Other proteins have no predicted function.

10 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

234 Discussion

235 This study provides an extensive characterization of the proteome of

236 Spiroplasma-infected Drosophila hemolymph. This dual-proteomics approach

237 provides a deep proteome of Spiroplasma in its natural environment and pinpoints

238 host proteins which abundance is altered by the presence of the bacteria. The power

239 of Drosophila genetics allowed us to test the involvement of these candidate genes in

240 regulating endosymbiosis. A targeted genetic screen revealed that most of the

241 differentially abundant proteins upon Spiroplasma infection do not participate in the

242 control of symbiont growth but may rather be involved in host adaptation to

243 Spiroplasma.

244 Spiroplasma is devoid of cell wall, hence devoid of peptidoglycan, which is the

245 main immune elicitor for the insect immune system (Lemaitre and Hoffmann, 2007).

246 This led to the assumption that Spiroplasma was undetectable by the canonical

247 innate immune pathways. Furthermore, the elicitation of the fly immune system (by

248 an infection or using genetics) has no impact on Spiroplasma titer, suggesting that

249 the bacteria are not only invisible by also resistant to the fly immune effectors (Herren

250 and Lemaitre, 2011). Flies that over-express AMPs and the ΔAMP10 mutant line

251 have normal bacterial titer that confirms the host immune system does not affect

252 Spiroplasma growth. However, numerous immune-related proteins (mostly

253 associated to the Toll pathway) were enriched in Spiroplasma infected hemolymph,

254 including soluble receptors (GNBP1, GNBP-like3), signal transduction intermediates

255 (Spätzle-Processing Enzyme and Hayan) and effectors or putative effectors (Attacin,

256 Bomanin Bc3 and CG33470). The Spiroplasma titer was not altered in several flies

257 carrying loss-of function alleles of these genes, indicating that the immune elicitation

258 is a consequence of the infection but not a control mechanism. It is worth noting that

11 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

259 the expression levels of the genes were extremely low as compared to a fully-fledged

260 immune response to pathogenic bacteria (De Gregorio et al., 2001). Such low

261 induction of the immune response has been reported in flies experiencing stress,

262 such as cold or heat stress (MacMillan et al., 2016; Telonis-Scott et al., 2013).

263 Therefore, the mild induction of immune proteins in infected hemolymph may be an

264 unspecific stress response associated to the presence of bacteria. Another attractive

265 hypothesis would be that proteases released by Spiroplasma could trigger the

266 soluble sensor Persephone and activate the Toll pathway in a peptidoglycan-

267 independent fashion (Gottar et al., 2006; Issa et al., 2018).

268 Another uncovering of this study is the depletion of proteasome

269 components in the hemolymph upon Spiroplasma infection. As the ubiquitin-

270 proteasome is a major degradation pathway for intracellular proteins (Kleiger and

271 Mayor, 2014), the components that we detected in the hemolymph are presumably

272 released from cells that broke upon hemolymph collection (epithelial, fat or muscular

273 cells, but also possibly hemocytes). However, functionally active 20S proteasome

274 units have been discovered circulating in extracellular fluids in mammals, including in

275 serum (Sixt and Dahlmann, 2008) hence we cannot exclude the existence of

276 circulating proteasome units in Drosophila hemolymph. Host ubiquitin-proteasome

277 systems have long been suspected to be a key element in symbiotic homeostasis. It

278 is upregulated in cells harboring intracellular symbionts in weevils, presumably to

279 increase protein turnover and provide free amino-acids to the bacteria (Heddi et al.,

280 2005). In vitro work on cell cultures infected by the facultative endosymbiont

281 Wolbachia also revealed that it induces the host proteasome presumably also to

282 support its own growth (Fallon and Witthuhn, 2009; He et al., 2019; White et al.,

283 2017). Remarkably, one proteasome subunit was detected as enriched upon

12 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

284 Spiroplasma infection in the aphid Aphis citricidus when the insect was feeding on a

285 suboptimum but not on an optimum host plant, suggesting an interaction between

286 symbiosis, nutrition and the proteasome-ubiquitin degradation system (Guidolin et al.,

287 2018). In the case of Drosophila-Spiroplasma interaction, the depletion of

288 proteasome subunits upon infection could thus be a titer regulation mechanism: by

289 decreasing its proteolysis, the host would limit the release of amino acids in the

290 extracellular space that would be usable by Spiroplasma.

291 Our approach also produced an in-depth Spiroplasma proteome on total

292 proteins regardless of their cell localization. This gives a quantitatively unbiased

293 overview of each protein abundance which was not achieved by previous data based

294 on detergent extractions (Paredes et al., 2015). Such quantitative approach allowed

295 us to make a comparison between the transcript and protein levels on about one

296 fourth of the total number of predicted coding sequences (Masson et al., 2018).

297 Although the correlation between transcriptomics and proteomics data greatly varies

298 among models, tissues and experimental set-ups (Haider and Pal, 2013), our data

299 indicate a rather low correlation in the case of Spiroplasma. The correlations between

300 transcript and protein levels depends on the interaction between transcript stability

301 and protein stability (Schwanhäusser et al., 2011). It also depends on the intrinsic

302 chemical properties of each transcript that affect the ribosome binding or the

303 translation speed, for example the Shine-Dalgarno sequence in prokaryotes

304 (Bingham et al., 1986) or more importantly the codon adaptation index of the coding

305 sequence (Lithwick and Margalit, 2003). An explanation for the overall lack of

306 correlation between transcripts and proteins regardless of the model could be that

307 selection operates at the protein level, hence changes in mRNA levels would be

308 offset by post-transcriptional mechanisms to maintain stable protein levels over

13 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

309 evolutionary times (Schrimpf et al., 2009). This hypothesis entails that genes having

310 a low mRNA level compared to their protein levels are likely to be under strong

311 selective pressure to maintain high protein abundancy through post-transcriptional

312 regulations. Intriguingly, this is the case for S. poulsonii Spiralin B, a protein

313 homologous to the S. citri Spiralin A, a membrane lectin suspected to be essential for

314 insect to plant transmission (Duret et al., 2003; Killiny et al., 2005). Spiralin B has

315 been identified as a putative virulence factor in S. poulsonii as it is upregulated when

316 the bacteria are in the fly hemolymph compared to in vitro culture (Masson et al.,

317 2018) and could be involved in oocyte infection for vertical transmission. Similarly,

318 the Crr glucose transporter has an unexpectedly high protein/mRNA ratio, possibly in

319 connection with its role in bacteria survival in the hemolymph. Spiroplasma has a

320 pseudogenized transporter for trehalose, the most abundant sugar in the hemolymph

321 (Paredes et al., 2015). This could have been selected over host-symbiont coevolution

322 to prevent Spiroplasma overgrowth, hence increasing the stability of the interaction

323 by limiting the metabolic cost of harboring the bacteria. Maintaining high Crr levels

324 could thus be an offset response to maintain bacteria proliferation despite trehalose

325 inaccessibility. Last, the ferritin-like protein Ftn has also a bias towards high protein

326 abundancy, suggesting that iron could be a key metabolite in Spiroplasma-

327 Drosophila symbiotic homeostasis.

328 All tissues bathed by hemolymph contribute to its composition. As a

329 consequence, studying the impact of symbionts on this biofluid is unapproachable by

330 transcriptomic methods only. Dual proteomics proved to be an efficient approach to

331 overcome this hurdle and gain novel insights into the Drosophila-Spiroplasma

332 symbiosis. This method is also promising for the study of other symbioses,

333 particularly those where symbionts inhabit biofluids rather than cells.

14 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

334 Material and Methods

335 Spiroplasma and Drosophila stocks

336 Flies were kept at 25°C on cornmeal medium (35.28 g of cornmeal, 35.28 g of

337 inactivated yeast, 3.72 g of agar, 36 ml of fruits juice, 2.9 ml of propionic acid and

338 15.9 ml of Moldex for 600 ml of medium). The complete list of fly stocks is available

339 in Supplementary Table S4.

340 Spiroplasma poulsonii strain Uganda-1 (Pool et al., 2006) was used for all infections.

341 Fly stocks were infected as previously described (Herren and Lemaitre, 2011).

342 Briefly, 9 nL of Spiroplasma-infected hemolymph was injected in their thorax. The

343 progeny of these flies was collected after 5-7 days using male killing as a proxy to

344 assess the infection (100% female progeny). Flies from at least the 3rd generation

345 post-injection were used experimentally.

346

347 Hemolymph extraction procedure

348 Embryos were collected from conventionally reared flies and dechorionated using a

349 bleach-based previously published method (Broderick et al., 2014) to ensure that

350 they were devoid of horizontally transmitted pathogens. One generation was then left

351 to develop in conventional rearing conditions to allow for gut microbiota

352 recontamination. Hemolymph was extracted from 10 days-old flies from the second

353 generation after bleach treatment using a Nanoject II (Drummond) as previously

354 described (Herren et al., 2014). 1 µL of hemolymph was collected for each sample

355 and frozen at -80°C until further use. Hemolymph was then diluted 10 times with PBS

356 containing Protease Inhibitor Cocktail 1X (Roche). 1 µl was used for protein

357 quantification with the Pierce BCA Protein Assay Kit (Thermofisher). The remaining 9

15 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

358 µl were mixed with SDS (0.2% final), DTT (2.5mM final) and PMSF (10µM final).

359 Aliquots of 15µg were used for proteomics analysis.

360

361 Proteomics sample preparation and LC-MS/MS

362 Each sample was digested by Filter Aided Sample Preparation (FASP) (Wiśniewski

363 et al., 2009) with minor modifications. Dithiothreitol (DTT) was replaced by Tris (2-

364 carboxyethyl)phosphine (TCEP) as reducing agent and iodoacetamide by

365 chloracetamide as alkylating agent. A proteolytic digestion was performed using

366 Endoproteinase Lys-C and Trypsin. Peptides were desalted on C18 StageTips

367 (Rappsilber et al., 2007) and dried down by vacuum centrifugation. For LC-MS/MS

368 analysis, peptides were resuspended and separated by reversed-phase

369 chromatography on a Dionex Ultimate 3000 RSLC nanoUPLC system in-line

370 connected with an Orbitrap Q Exactive HF Mass Spectrometer (Thermo Fischer

371 Scientific). Database search was performed using MaxQuant 1.5.1.2 (Cox and Mann,

372 2008) against a concatenated database consisting of the Ensemble Drosophila

373 melanogaster protein database (BDGP5.25) and a custom Spiroplama poulsonii

374 proteome predicted from the reference genome (Genbank accession number

375 JTLV00000000.2). Carbamidomethylation was set as fixed modification, whereas

376 oxidation (M) and acetylation (Protein N-term) were considered as variable

377 modifications. Label-free quantification was performed by MaxQuant using the

378 standard settings. Hits were considered significant when the fold-change between

379 infected and uninfected hemolymph was >2 or <-2 and the FDR<0.05.

380 Complete proteomics data are available on the ProteomExchange repository with the

381 accession number #[data under submission].

382

16 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

383 Spiroplasma quantification

384 Spiroplasma quantification was performed by qPCR as previously described (Masson

385 et al., 2020a). Briefly, the DNA of pools of 5 whole flies was extracted and the copy

386 number of a single-copy bacterial gene (dnaA or dnaK) was quantified and

387 normalized by that of the host gene rsp17. Primers sequences are available in

388 Supplementary Table S4. Each experiment has been repeated 2 or 3 independent

389 times with at least 3 technical replicates each. Data were analyzed by one-way

390 ANOVA.

391

392 RT-qPCR

393 Gene expression measurement were performed by RT-qPCR as previously

394 described (Iatsenko et al., 2020; Romeo and Lemaitre, 2008). Briefly, 10 whole flies

395 were crushed and their RNA extracted with the Trizol method. Reverse transcription

396 was carried out using a PrimeScript RT kit (Takara) and a mix of random hexamers

397 and oligo-dTs. qPCR was performed on a QuantStudio 3 (Applied Biosystems) with

398 PowerUp SYBR Green Master Mix using primers sequences available in

399 Supplementary Table S4. The expression of the target gene was normalized by that

400 of the housekeeping gene rp49 (rpL32) using the 2-∆∆CT method (Pfaffl, 2001).

401 Each experiment has been repeated 2 or 3 independent times with at least 3

402 technical replicates each. Data were analyzed by Mann-Whitney tests.

403

404 Acknowledgments

405 We are grateful to Florence Armand, Romain Hamelin and the EPFL Proteomics

406 Core Facility for sharing their expertise on proteomics.

407

17 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

408 Competing interests

409 The authors declare no competing interests.

410

411 Funding

412 This work was funded by the Swiss National Science Foundation grant

413 N°310030_185295.

18 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

414 References

415 Ballinger, M. J. and Perlman, S. J. (2017). Generality of toxins in defensive symbiosis:

416 Ribosome-inactivating proteins and defense against parasitic wasps in Drosophila.

417 PLoS Pathog. 13, e1006431.

418 Bingham, A. H., Ponnambalam, S., Chan, B. and Busby, S. (1986). Mutations that reduce

419 expression from the P2 promoter of the Escherichia coli galactose operon. Gene 41,

420 67–74.

421 Broderick, N. a., Buchon, N. and Lemaitre, B. (2014). Microbiota-induced changes in

422 Drosophila melanogaster host gene expression and gut morphology. MBio 5, 1–13.

423 Cox, J. and Mann, M. (2008). MaxQuant enables high peptide identification rates,

424 individualized p.p.b.-range mass accuracies and proteome-wide protein quantification.

425 Nat. Biotechnol. 26, 1367–1372.

426 De Gregorio, E., Spellman, P. T., Rubin, G. M. and Lemaitre, B. (2001). Genome-wide

427 analysis of the Drosophila immune response by using oligonucleotide microarrays.

428 Proc. Nat. Acad. Sci. 98, 12590–12595.

429 Douglas, A. E. (2015). Multiorganismal Insects: Diversity and Function of Resident

430 Microorganisms. Annu. Rev. Entomol. 60, 17–34.

431 Douglas, A. E. (2016). How multi-partner endosymbioses function. Nat. Rev. Microbiol. 14,

432 731–743.

433 Duret, S., Berho, N., Danet, J. L., Garnier, M. and Renaudin, J. (2003). Spiralin is not

434 essential for helicity, motility, or pathogenicity but is required for efficient transmission of

435 Spiroplasma citri by its leafhopper vector Circulifer haematoceps. Appl. Environ.

436 Microbiol. 69, 6225–6234.

437 Duron, O., Bouchon, D., Boutin, S., Bellamy, L., Zhou, L., Engelstadter, J. and Hurst, G.

438 D. (2008). The diversity of reproductive parasites among arthropods: Wolbachia do not

439 walk alone. BMC Biol. 6, 27.

440 Fallon, A. M. and Witthuhn, B. A. (2009). Proteasome activity in a naïve mosquito cell line

19 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

441 infected with Wolbachia pipientis wAlbB. In Vitro Cell. Dev. Biol. Anim. 45, 460–466.

442 Gasparich, G. E. (2002). : evolution, adaptation and diversity. Front. Biosci. 7,

443 619–640.

444 Gottar, M., Gobert, V., Matskevich, A. A., Reichhart, J.-M., Wang, C., Butt, T. M., Belvin,

445 M., Hoffmann, J. A. and Ferrandon, D. (2006). Dual detection of fungal infections in

446 Drosophila via recognition of glucans and sensing of virulence factors. Cell 127, 1425–

447 1437.

448 Guidolin, A. S., Cataldi, T. R., Labate, C. A., Francis, F. and Cônsoli, F. L. (2018).

449 Spiroplasma affects host aphid proteomics feeding on two nutritional resources. Sci.

450 Rep. 8, 2466.

451 Haider, S. and Pal, R. (2013). Integrated analysis of transcriptomic and proteomic data.

452 Curr. Genomics 14, 91–110.

453 Hamilton, P. T., Peng, F., Boulanger, M. J. and Perlman, S. J. (2016). A ribosome-

454 inactivating protein in a Drosophila defensive symbiont. Proc. Natl. Acad. Sci. U. S. A.

455 113, 350–355.

456 Hanson, M. A., Dostálová, A., Ceroni, C., Poidevin, M., Kondo, S. and Lemaitre, B.

457 (2019). Synergy and remarkable specificity of antimicrobial peptides in vivo using a

458 systematic knockout approach. Elife 8, e44341.

459 Harumoto, T. and Lemaitre, B. (2018). Male-killing toxin in a bacterial symbiont of

460 Drosophila. Nature 557, 252–255.

461 Harumoto, T., Anbutsu, H. and Fukatsu, T. (2014). Male-killing Spiroplasma induces sex-

462 specific cell death via host apoptotic pathway. PLoS Pathog. 10, e1003956.

463 Harumoto, T., Anbutsu, H., Lemaitre, B. and Fukatsu, T. (2016). Male-killing symbiont

464 damages host’s dosage-compensated sex chromosome to induce embryonic apoptosis.

465 Nat. Commun. 7, 12781.

466 Harumoto, T., Fukatsu, T. and Lemaitre, B. (2018). Common and unique strategies of

467 male killing evolved in two distinct Drosophila symbionts. Proc. R. Soc. B Biol. Sci. 285,.

468 Haselkorn, T. S. (2010). The Spiroplasma heritable bacterial endosymbiont of Drosophila.

20 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

469 Fly (Austin). 4, 80–87.

470 He, Z., Zheng, Y., Yu, W.-J., Fang, Y., Mao, B. and Wang, Y.-F. (2019). How do Wolbachia

471 modify the Drosophila ovary? New evidences support the “titration-restitution” model for

472 the mechanisms of Wolbachia-induced CI. BMC Genomics 20, 608.

473 Heddi, A., Vallier, A., Anselme, C., Xin, H., Rahbe, Y. and Wackers, F. (2005). Molecular

474 and cellular profiles of insect bacteriocytes: mutualism and harm at the initial

475 evolutionary step of symbiogenesis. Cell Microbiol 7, 293–305.

476 Hedges, L. M., Brownlie, J. C., O’Neill, S. L. and Johnson, K. N. (2008). Wolbachia and

477 virus protection in insects. Science (80-. ). 322, 702.

478 Herren, J. K. and Lemaitre, B. (2011). Spiroplasma and host immunity: Activation of

479 humoral immune responses increases endosymbiont load and susceptibility to certain

480 Gram-negative bacterial pathogens in Drosophila melanogaster. Cell. Microbiol. 13,

481 1385–1396.

482 Herren, J. K., Paredes, J. C., Schüpfer, F. and Lemaitre, B. (2013). Vertical Transmission

483 of a Drosophila Endosymbiont Via Cooption of the Yolk Transport and Internalization

484 Machinery. MBio 4 (2), e00532-12.

485 Herren, J. K., Paredes, J. C., Schüpfer, F., Arafah, K., Bulet, P. and Lemaitre, B. (2014).

486 Insect endosymbiont proliferation is limited by lipid availability. Elife 3, e02964.

487 Hoffmann, A. A., Montgomery, B. L., Popovici, J., Iturbe-Ormaetxe, I., Johnson, P. H.,

488 Muzzi, F., Greenfield, M., Durkan, M., Leong, Y. S., Dong, Y., et al. (2011).

489 Successful establishment of Wolbachia in Aedes populations to suppress dengue

490 transmission. Nature 476, 454.

491 Hurst, G. D. D., Anbutsu, H., Kutsukake, M. and Fukatsu, T. (2003). Hidden from the host:

492 Spiroplasma bacteria infecting Drosophila do not cause an immune response, but are

493 suppressed by ectopic immune activation. Insect Mol. Biol. 12, 93–97.

494 Iatsenko, I., Marra, A., Boquete, J.-P., Peña, J. and Lemaitre, B. (2020). Iron

495 sequestration by transferrin 1 mediates nutritional immunity in Drosophila melanogaster.

496 Proc. Natl. Acad. Sci. 117, 7317 LP-7325.

21 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

497 Issa, N., Guillaumot, N., Lauret, E., Matt, N., Schaeffer-Reiss, C., Van Dorsselaer, A.,

498 Reichhart, J.-M. and Veillard, F. (2018). The Circulating Protease Persephone Is an

499 Immune Sensor for Microbial Proteolytic Activities Upstream of the Drosophila Toll

500 Pathway. Mol. Cell 69, 539–550.e6.

501 Jaenike, J., Unckless, R., Cockburn, S. N., Boelio, L. M. and Perlman, S. J. (2010).

502 Adaptation via symbiosis: recent spread of a Drosophila defensive symbiont. Science

503 (80-. ). 329, 212–215.

504 Killiny, N., Castroviejo, M. and Saillard, C. (2005). Spiroplasma citri Spiralin acts in vitro as

505 a lectin binding to glycoproteins from its insect vector Circulifer haematoceps.

506 Phytopathology 95, 541–548.

507 Kleiger, G. and Mayor, T. (2014). Perilous journey: a tour of the ubiquitin-proteasome

508 system. Trends Cell Biol. 24, 352–359.

509 Lemaitre, B. and Hoffmann, J. (2007). The host defense of Drosophila melanogaster. Annu

510 Rev Immunol 25, 697–743.

511 Lithwick, G. and Margalit, H. (2003). Hierarchy of sequence-dependent features associated

512 with prokaryotic translation. Genome Res. 13, 2665–2673.

513 MacMillan, H. A., Knee, J. M., Dennis, A. B., Udaka, H., Marshall, K. E., Merritt, T. J. S.

514 and Sinclair, B. J. (2016). Cold acclimation wholly reorganizes the Drosophila

515 melanogaster transcriptome and metabolome. Sci. Rep. 6, 28999.

516 Masson, F. and Lemaitre, B. (2020). Growing Ungrowable Bacteria: Overview and

517 Perspectives on Insect Symbiont Culturability. Microbiol. Mol. Biol. Rev. 84, e00089-20.

518 Masson, F., Calderon Copete, S., Schüpfer, F., Garcia-Arraez, G. and Lemaitre, B.

519 (2018). In Vitro Culture of the Insect Endosymbiont Spiroplasma poulsonii Highlights

520 Bacterial Genes Involved in Host-Symbiont Interaction. MBio 9(2), e00024-18.

521 Masson, F., Calderon-Copete, S., Schüpfer, F., Vigneron, A., Rommelaere, S., Garcia-

522 Arraez, M. G., Paredes, J. C. and Lemaitre, B. (2020a). Blind killing of both male and

523 female Drosophila embryos by a natural variant of the endosymbiotic bacterium

524 Spiroplasma poulsonii. Cell. Microbiol. 22(5), e13156.

22 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

525 Masson, F., Schüpfer, F., Jollivet, C. and Lemaitre, B. (2020b). Transformation of the

526 Drosophila sex-manipulative endosymbiont Spiroplasma poulsonii and persisting

527 hurdles for functional genetics studies. Appl. Environ. Microbiol. AEM.00835-20.

528 Mateos, M., Castrezana, S. J., Nankivell, B. J., Estes, A. M., Markow, T. A. and Moran,

529 N. A. (2006). Heritable Endosymbionts of Drosophila. Genetics 174, 363–376.

530 Medina, P., Russell, S. L. and Corbett-Detig, R. (2019). Deep data mining reveals variable

531 abundance and distribution of microbial reproductive manipulators within and among

532 diverse host species. bioRxiv 679837.

533 Montllor, C. B., Maxmen, A. and Purcell, A. H. (2002). Facultative bacterial endosymbionts

534 benefit pea aphids Acyrthosiphon pisum under heat stress. Ecol. Entomol. 27, 189–195.

535 Oliver, K. M., Russell, J. A., Moran, N. A. and Hunter, M. S. (2003). Facultative bacterial

536 symbionts in aphids confer resistance to parasitic wasps. Proc Natl Acad Sci U S A 100,

537 1803–1807.

538 Palombella, V. J., Rando, O. J., Goldberg, A. L. and Maniatis, T. (1994). The ubiquitin-

539 proteasome pathway is required for processing the NF-kappa B1 precursor protein and

540 the activation of NF-kappa B. Cell 78, 773–785.

541 Paredes, J. C., Herren, J. K., Schüpfer, F., Marin, R., Claverol, S., Kuo, C.-H., Lemaitre,

542 B. and Béven, L. (2015). Genome sequence of the Drosophila melanogaster male-

543 killing Spiroplasma strain MSRO endosymbiont. MBio 6(2), e02437-14.

544 Paredes, J. C., Herren, J. K., Schüpfer, F. and Lemaitre, B. (2016). The Role of Lipid

545 Competition for Endosymbiont-Mediated Protection against Parasitoid Wasps in

546 Drosophila. MBio 7(4), e01006-16.

547 Pfaffl, M. W. (2001). A new mathematical model for relative quantification in real-time RT-

548 PCR. Nucleic Acids Res. 29, e45–e45.

549 Pool, J. E., Wong, a and Aquadro, C. F. (2006). Finding of male-killing Spiroplasma

550 infecting Drosophila melanogaster in Africa implies transatlantic migration of this

551 endosymbiont. Heredity (Edinb). 97, 27–32.

552 Rappsilber, J., Mann, M. and Ishihama, Y. (2007). Protocol for micro-purification,

23 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

553 enrichment, pre-fractionation and storage of peptides for proteomics using StageTips.

554 Nat. Protoc. 2, 1896–1906.

555 Romeo, Y. and Lemaitre, B. (2008). Drosophila immunity: methods for monitoring the

556 activity of Toll and Imd signaling pathways. Methods Mol. Biol. 415, 379–394.

557 Scarborough, C. L., Ferrari, J. and Godfray, H. C. (2005). Aphid protected from pathogen

558 by endosymbiont. Science (80-. ). 310, 1781.

559 Schrimpf, S. P., Weiss, M., Reiter, L., Ahrens, C. H., Jovanovic, M., Malmström, J.,

560 Brunner, E., Mohanty, S., Lercher, M. J., Hunziker, P. E., et al. (2009). Comparative

561 Functional Analysis of the Caenorhabditis elegans and Drosophila melanogaster

562 Proteomes. PLOS Biol. 7, e1000048.

563 Schwanhäusser, B., Busse, D., Li, N., Dittmar, G., Schuchhardt, J., Wolf, J., Chen, W.

564 and Selbach, M. (2011). Global quantification of mammalian gene expression control.

565 Nature 473, 337–342.

566 Sixt, S. U. and Dahlmann, B. (2008). Extracellular, circulating proteasomes and ubiquitin —

567 Incidence and relevance. Biochim. Biophys. Acta - Mol. Basis Dis. 1782, 817–823.

568 Teixeira, L., Ferreira, A. and Ashburner, M. (2008). The bacterial symbiont Wolbachia

569 induces resistance to RNA viral infections in Drosophila melanogaster. PLoS Biol 6, e2.

570 Telonis-Scott, M., van Heerwaarden, B., Johnson, T. K., Hoffmann, A. A. and Sgrò, C.

571 M. (2013). New Levels of Transcriptome Complexity at Upper Thermal Limits in Wild

572 Drosophila Revealed by Exon Expression Analysis. Genetics 195, 809–830.

573 Troha, K., Im, J. H., Revah, J., Lazzaro, B. P. and Buchon, N. (2018). Comparative

574 transcriptomics reveals CrebA as a novel regulator of infection tolerance in D.

575 melanogaster. PLoS Pathog. 14, e1006847.

576 Tzou, P., Reichhart, J.-M. and Lemaitre, B. (2002). Constitutive expression of a single

577 antimicrobial peptide can restore wild-type resistance to infection in immunodeficient

578 Drosophila mutants. Proc. Natl. Acad. Sci. U. S. A. 99, 2152–2157.

579 Veneti, Z., Bentley, J. K., Koana, T., Braig, H. R. and Hurst, G. D. D. (2005). A Functional

580 Dosage Compensation Complex Required for Male Killing in Drosophila. Science (80-.

24 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

581 ). 307, 1461–1463.

582 Werren, J. H., Baldo, L. and Clark, M. E. (2008). Wolbachia: master manipulators of

583 invertebrate biology. Nat. Rev. Microbiol. 6, 741–751.

584 White, P. M., Serbus, L. R., Debec, A., Codina, A., Bray, W., Guichet, A., Lokey, R. S.

585 and Sullivan, W. (2017). Reliance of Wolbachia on High Rates of Host Proteolysis

586 Revealed by a Genome-Wide RNAi Screen of Drosophila Cells. Genetics 205, 1473–

587 1488.

588 Wiśniewski, J. R., Zougman, A., Nagaraj, N. and Mann, M. (2009). Universal sample

589 preparation method for proteome analysis. Nat. Methods 6, 359–362.

590 Xie, J., Butler, S., Sanchez, G. and Mateos, M. (2014). Male killing Spiroplasma protects

591 Drosophila melanogaster against two parasitoid wasps. Heredity (Edinb). 112, 399–408.

592

593

25 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

594 Figure Legends

595 Figure 1. Drosophila hemolymph protein profiling upon Spiroplasma infection 596 (A) Representation of the dataset repartition between intracellular and 597 extracellular protein groups. Extracellular groups are detected either by the 598 presence of a predicted signal peptide using the SignalP algorithm or the 599 annotation with a GO term referring to the extracellular space. The overlap 600 category corresponds to protein groups with both a predicted signal peptide 601 and a GO term “extracellular”. Numbers in brackets indicate the number of 602 protein groups in each category. 603 (B) Volcano plot of the intracellular protein group subset. The log fold-change 604 indicates the differential representation of the protein group in the infected 605 samples versus the uninfected samples. Bold purple dots indicate significance 606 as defined by an absolute fold change value over 2 or under -2 and a p-value 607 below 0.05. 608 (C) Functional GO annotation of the intracellular protein groups. Only significant 609 (p-value < 0.05) GO terms of hierarchy 4 or less are displayed. CC: Cell 610 Component; MF: Molecular Function; BP: Biological Process. 611 (D) Volcano plot of the extracellular protein group subset. The log fold-change 612 indicates the differential representation of the protein group in the infected 613 samples versus the uninfected samples. Bold purple dots indicate significance 614 as defined by an absolute fold change value over 2 or under -2 and a p-value 615 below 0.05. 616 (E) Functional GO annotation of the extracellular protein groups. Only significant 617 (p-value < 0.05) GO terms of hierarchy 4 or less are displayed. CC: Cell 618 Component; MF: Molecular Function; BP: Biological Process. 619 620 621 Figure 2. Spiroplasma infection induces a mild transcriptional immune 622 response 623 624 (A) Changes in a selection of immune-related protein group abundances 625 quantified by LC-MS/MS. Data are expressed as fold change of the 626 Label-Free Quantification (LFQ) values of Spiroplasma infected

26 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

627 hemolymph over the uninfected control (log2 scale). *, p<0.05; ** 628 p<0.005, ***; p<0.005, ****, p<0.0005 upon Student t-test. 629 (B) mRNA quantification of a selection of candidate genes in uninfected and 630 Spiroplasma-infected flies. Results are expressed as the fold change of 631 target mRNA normalized by rpL32 mRNA in Spiroplasma-infected flies as 632 compared to uninfected flies (log2 scale). *, p<0.05; ** p<0.005, ***; 633 p<0.0005 upon Mann-Whitney test on ΔCt values. 634 635 Figure 3. Targeted genetic screening of candidate proteins 636 (A-C) Quantification of Spiroplasma titer in two weeks old flies in various 637 genetic backgounds. Titer is expressed as the fold-change over the 638 appropriate control line. All graphs represent the mean +/- standard deviation 639 of a pool of at least 2 independent experiments. (A) Control flies (Act5C- 640 GAL4 driver crossed w1118, white bars) are compared to flies overexpressing 641 a single antimicrobial peptide gene driven by the ubiquitous Act5C-GAL4 642 driver (black bars). Titer is expressed as the fold-change over control. (B) 643 Quantification of Spiroplasma titer in wild-type flies (OregonR, white bar) and 644 hayan, attD and tep2 null mutants (black). Spiroplasma titer in AMP10 iso 645 flies were compared to that of their isogenic wild-type counterpart (w1118 iso, 646 white bar). (C) Quantification of the impact of RNAi-mediated knockdown on 647 Spiroplasma titer with the Act5C-GAL4 driver. Titer is expressed as the fold- 648 change over the control (Act5C-GAL4 driver crossed to w1118). 649 650 651 Figure 4. Spiroplasma transcription-translation correlation

652 Each dot represents a protein positioned according to its log2(LFQ; Label-free 653 Quantification) value in the proteome versus its log2(CPM ; Count Per Million) 654 in the transcriptomics dataset from (Masson et al., 2018). The black line 655 represents the linear model log2(LFQ) = 0.57892 x log2(CPM) + 21.38325, 656 with an adjusted R2 = 0.2909. Dot size and color are adjusted to the residuals 657 of the model, with bigger bluer dots indicating a higher residual hence a 658 stronger deviation from the linear model for the considered protein. 659

27 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.

660 Supplementary Figure 1.

661 (A) Quantification of Spiroplasma titer in two weeks old flies. Titer is expressed 662 as the fold-change of Act5C-GAL4>UAS-RNAi over Act5C-GAL4> w1118. Bars 663 represent the mean +/- standard deviation of a pool of at least 2 independent 664 experiments. (B) Quantification of Spiroplasma titer in CG15293 mutant flies 665 compared to control yw flies. Bars represent the mean +/- standard deviation 666 of a pool of at least 3 independent experiments. ***; p<0.0005 upon Mann- 667 Whitney test on ΔΔCt values. 668 669

670 Supplementary Table 1.

671 Drosophila genes quantified by LC-MS/MS.

672

673 Supplementary Table 2.

674 Spiroplasma genes quantified by LC-MS/MS.

675

676 Supplementary Table 3.

677 Spiroplasma genes outlying the linear model between mRNA and protein

678 levels.

28 bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license. bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license. bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license. bioRxiv preprint doi: https://doi.org/10.1101/2021.02.22.432267; this version posted February 22, 2021. The copyright holder for this preprint (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 available under aCC-BY 4.0 International license.