The Lyme disease agent co-opts adiponectin receptor-mediated signaling in its arthropod vector

Adiponectin-mediated pathways contribute to mammalian homeostasis; however, little is known about adiponectin and adiponectin receptor signaling in arthropods. In this study, we demonstrate that Ixodes scapularis ticks have an adiponectin receptor-like protein (ISARL) but lack adiponectin, suggesting activation by alternative pathways. ISARL expression is significantly upregulated in the tick gut after Borrelia burgdorferi infection, suggesting that ISARL signaling may be co-opted by the Lyme disease agent. Consistent with this, RNA interference (RNAi)-mediated silencing of ISARL significantly reduced the B. burgdorferi burden in the tick. RNA-seq-based transcriptomics and RNAi assays demonstrate that ISARL-mediated phospholipid metabolism by phosphatidylserine synthase I is associated with B. burgdorferi survival. Furthermore, the tick complement C1q-like protein 3 interacts with ISARL, and B. burgdorferi facilitates this process. This study identifies a new tick metabolic pathway that is connected to the life cycle of the Lyme disease spirochete.


Introduction
. In addition, Zhu et al. (2008)  burgdorferi is acquired when larval or nymphal ticks feed on infected animals, and is 87 transmitted by nymphs or adults to vertebrate hosts (Kurokawa et al., 2020). Lyme 88 disease in humans manifests as a multisystem disorder of the skin and other organs (e.g., 89 joints, heart, and nervous system), resulting in patients experiencing cardiac, neurological, 90 and arthritic complications (Asch et al., 1994;Singh and Girschick, 2004). A human 91 vaccine against Lyme disease was approved by the FDA but is not currently available 92 (Steere et al., 1998). Targeting tick proteins has the potential to disrupt tick feeding and 93 alter B. burgdorferi colonization or transmission (Kurokawa et  As I. scapularis lack an obvious adiponectin homolog, we examined whether 129 expression of ISARL could be stimulated in the feeding vector by allowing ticks to engorge 130 on mice, including uninfected and B. burgdorferi-infected animals. Interestingly a blood 131 meal containing B. burgdorferi resulted in significantly increased expression of ISARL in 132 the nymphal tick guts (P < 0.0001) (Fig. 1A). This suggests that the presence of B. 133 burgdorferi in the blood meal helps to stimulate tick metabolic activity and/or that ISARL 134 may have an important role during B. burgdorferi colonization of the tick gut. 135

136
Since ISARL expression was upregulated upon B. burgdorferi infection, we 137 hypothesized that RNAi-mediated silencing of ISARL would affect B. burgdorferi 138 colonization by nymphal I. scapularis. To this end, ISARL or GFP (control) dsRNA was 139 injected into the guts of pathogen-free nymphs by anal pore injection. Then, the ticks were 140 allowed to feed on B. burgdorferi-infected mice. Quantitative RT-PCR (qPCR) analysis 141 showed a significant decrease of ISARL expression in the guts of ds ISARL-injected ticks 142 (P < 0.01) when compared to that in control ds GFP-injected tick guts (Fig. 1B), indicating 143 that the knockdown was successful. The engorgement weights of ds ISARL-injected 144 nymphs and control ds GFP-injected nymphs were comparable (P > 0.05) (Fig. 1C), 145 suggesting that silencing ISARL had no effect on tick feeding behavior. However, ISARL-146 silenced nymph guts showed a marked reduction of the B. burgdorferi burden (P < 0.001) 147 when compared to that in control ticks (Fig. 1D), demonstrating that ISARL is associated 148 with B. burgdorferi colonization in the nymphal tick gut. injected ticks after engorgement on B. burgdorferi-infected or uninfected mice, using 168

RNA-seq. 169
After feeding on uninfected mice, 18 genes were significantly differentially 171 expressed with six upregulated and 12 downregulated genes in the guts of ds ISARL-172 injected nymphal ticks when compared to that in control ds GFP-injected tick guts (Fig.  173 2A; Table S1). 35 genes were differentially expressed at a significant level, and all these 174 genes were downregulated in the guts of ds ISARL-injected I. scapularis when compared 175 to that in control ds GFP-injected ticks after feeding on B. burgdorferi-infected mice (Fig.  176 2B; Table S2). In addition, the transcriptome analysis further demonstrated that the ISARL 177 gene was successfully silenced by RNAi (Tables S1 and S2) and this was also confirmed 178 by qPCR (Fig. 2C). No common differentially expressed genes except ISARL were 179 observed between control or experimental ticks feeding on uninfected or B. burgdorferi-180 infected mice (Fig. 2D), suggesting that the 34 significantly expressed genes were all 181 induced by B. burgdorferi, or the influence of B. burgdorferi on the host blood components, 182 rather than blood meal itself, in absence of ISARL. 183

184
In response to the blood meal, a significant change of the metabolic pathways in 185 ticks was observed in absence of ISARL. In particular, based on GO functional 186 classification and KEGG pathways analyses, the glutathione metabolic process with nine 187 genes (e.g., gamma glutamyl transpeptidase, G2/mitotic-specific cyclin A, and glutathione 188 S-transferase) was significantly altered. Moreover, the genes involved in metabolic 189 pathways such propanoate metabolism and carbohydrate transport and metabolism (e.g., 190 acyl-CoA synthetase and soluble maltase-glucoamylase) were also significantly changed 191 in absence of ISARL after engorging on uninfected mice. 192 Similarly, many metabolism-associated genes, including multiple amino acids, 194 lipids or sugars synthesis and transport genes (e.g., 3-hydroxyacyl-CoA dehydrogenase, 195 glycogen phosphorylase, and sugar transporter) were significantly downregulated in the 196 absence of ISARL after engorging on B. burgdorferi-infected mice. GO functional 197 classification and KEGG pathways also showed that the most downregulated genes were 198 involved in fatty acid, lipid and phospholipid, purine, amino acid, glycerophospholipid, and 199 carbohydrate metabolism and transport pathways after silencing ISARL (Fig. 2E), vacuolar H+-ATPase V1 sector, subunit G (V-ATPase), and sideroflexin 1,2,3, putative 216 (SFXN) (Fig. S4A). 217 218 Then, we silenced these four genes individually and investigated their potential 219 roles in B. burgdorferi colonization. We also silenced another four genes, whose P-values 220 were very close to significant (Fig. S4B). These four genes included 3-hydroxyacyl-CoA 221 dehydrogenase, putative (3HADH), adenylosuccinate synthetase (ADSS), GMP synthase, 222 putative (GMPS), and alpha-actinin, putative (ACTN). We did not observe a significant 223 decrease of B. burgdorferi burden in nymphal tick guts after silencing NCAM, V-ATPase, 224 SFXN, ADSS, GMPS, and ACTN compared to ds GFP-injected ticks (Fig. S5). Instead, 225 we found that PTDSS1-silenced nymphs showed a marked reduction in the B. burgdorferi 226 burden in the guts when compared to that in control ticks (P < 0.05) (Fig. 2H). Furthermore, 227 a blood meal containing B. burgdorferi resulted in significantly increased expression of 228 PTDSS1 in the nymphal tick guts (P < 0.05) (Fig. 2I), suggesting that PTDSS1 indeed 229 has a critical role during B. burgdorferi colonization of the tick gut. PTDSS1 is involved in 230 phospholipid metabolism, and mainly uses L-serine as the phosphatidyl acceptor to 231 generate the anionic lipid phosphatidylserine (PS), which serves as a precursor for 232 phosphatidylethanolamine (PE) and phosphatidylcholine (PC) synthesis ( pathway in tick is critical for B. burgdorferi, we silenced another enzyme (ISARL-237 unrelated), phosphatidylserine decarboxylase (PSD, ISCI003338), which is an important 238 enzyme in the synthesis of PE in both prokaryotes and eukaryotes (Voelker, 1997). 239 Interestingly, we also found a significantly decreased B. burgdorferi burden in ds PSD-240 injected tick guts (P < 0.05), and PSD and PTDSS1 elicit a similar degree of reduced B. 241 burgdorferi levels (Fig. 2K). Taken together, ISARL-mediated phospholipid metabolic 242 pathways associated with PTDSS1 have a critical role in B. burgdorferi colonization. databases. This suggests that ticks may utilize vertebrate adiponectin to activate the 251 adiponectin receptor during a blood meal, that tick have another ligand that stimulates the 252 receptor, or both. Since ticks are habitually exposed to adiponectin present during a 253 bloodmeal, we examined whether the tick adiponectin receptor could interact with 254 incoming mammalian adiponectin during blood feeding. We injected recombinant mouse 255 adiponectin into unfed ticks and investigated whether mammalian adiponectin could 256 activate downstream signaling of tick adiponectin receptor by RNA-seq (Fig. 3A). The 257 data showed that one classic downstream gene of mammalian adiponectin signaling, tick 258 glucose-6-phosphatase (G6P, ISCW017459), was significantly downregulated in the 259 presence of mammalian adiponectin ( Fig. 3A; Table S3). It has been demonstrated that 260 in mammals, the binding of adiponectin to its receptor suppresses G6P and 261 phosphoenolpyruvate carboxykinase (PEPCK) expression through an AMP-activated 262 protein kinase (AMPK)-dependent mechanism, which further inhibits glycogenolysis and 263 gluconeogenesis ( Fig. 3B) (Tishinsky, 2012). We further searched G6P and PEPCK 264 homologs in I. scapularis genome, and two G6P homologs (ISCW017459 and 265 ISCW018612) and three PEPCK homologs (ISCW001902, and ISCW000524, 266 ISCW001905) were identified. We designated them as G6P1, G6P2, PEPCK1, PEPCK2, 267 and PEPCK3, respectively. We evaluated gene expression of all these five genes after   burgdorferi-infected WT and Adipo -/mice and allowed to feed to repletion (Fig. 3F). No 281 significant difference of the B. burgdorferi burden in ticks feeding on WT and Adipo -/mice 282 was noted (P > 0.05) (Fig. 3G). We also silenced the G6P1 and G6P2 genes to determine 283 whether G6P-mediated glucose metabolic changes affect B. burgdorferi colonization. 284 Consistent with the previous observation, there was no significant difference in the B. 285 burgdorferi burden between control and G6P1-silenced ticks (P > 0.05) (Fig. S6A). G6P2-286 silenced ticks also did not show altered B. burgdorferi levels (P > 0.05) (Fig. S6B). 287 Furthermore, the expression of G6P1 and G6P2 in the nymphs was not influenced by B. 288 burgdorferi infection (P > 0.05) (Fig. S6C), suggesting that G6P1-or G6P2-mediated 289 changes do not affect B. burgdorferi colonization of the tick gut. To assess any changes 290 in the adiponectin concentration in murine serum after B. burgdorferi infection, the mice 291 were injected subcutaneously with 100 uL containing 1x10 4 or 1x10 7 B. burgdorferi, or 292 PBS as a control. We found that B. burgdorferi does not influence the adiponectin 293 concentration in murine blood (Fig. 3H). Taken together, these data suggest that 294 mammalian adiponectin can regulate ISARL-mediated glucose metabolism pathway, 295 however, it has no effect on B. burgdorferi colonization. C1QL3 could be detected on the surface of ISARL-expressed rather than empty plasmid 327 transfected HEK293T cells (Fig. 4G). A pull-down assay also indicated that recombinant 328 C1QL3 interacts with ISARL as demonstrated by the detection of C1QL3 only in ISARL 329 expressed cells pellet (Fig. 4H). In addition, co-immunolocalization demonstrated that the 330 C1QL3 protein specifically binds to the ISARL-expressed cell membrane (Fig. 4I). These 331 results suggest that tick C1QL3 interacts with ISARL. 332 333 Since C1QL3 is a ligand of tick ISARL and also involved in Borrelia colonization, 334 we further investigated whether C1QL3 has a role on the activation of ISARL by Borrelia. 335 We first assessed if silencing of C1QL3 influenced ISARL expression after feeding on B. 336 burgdorferi-infected mice (Fig. 5A). QPCR assessment showed that the ISARL transcript 337 level following RNAi silencing of C1QL3 was significantly lower than that in control ds 338 GFP-injected tick guts after feeding on B. burgdorferi-infected mice (P < 0.05) (Fig. 5B). 339 More importantly, after silencing C1QL3, a blood meal containing B. burgdorferi did not 340 significantly increase expression of ISARL in the nymphal tick guts as compared to 341 feeding on clean mice (P > 0.05) (Fig. 5C), further suggesting that C1QL3 plays a role in 342 the ISARL signaling pathways. rely on its vertebrate and arthropod hosts for nutrients or enzymes that it cannot 392 synthesize (Tilly et al., 2008). Interestingly, we also found that silencing of I. scapularis 393 3HADH, which is involved in fatty acid metabolic processes, decreased the B. burgdorferi 394 burden in tick gut ( Figure S5D). The markedly decreased B. burgdorferi burden in ticks 395 after silencing of PTDSS1, PSD and 3HADH, suggests that the spirochete may require 396 the tick for selected metabolic needs. 397

398
We also found that B. burgdorferi can interact with an adiponectin-related protein, 399 C1QL3, in ticks, which associates with ISARL. The interactions lead to phospholipid 400 metabolism changes in ticks. We propose that C1QL3 in tick is mainly involved in 401 metabolism, rather than complement activation, as demonstrated by the decreased B.  Table S4. Pathogen-free I. scapularis larvae were acquired from the Centers 445 for Disease Control and Prevention. The larval ticks were fed to repletion on pathogen-446 free C3H/HeJ mice and allowed to molt to nymphs. B. burgdorferi-infected nymphs were 447 generated by placing larvae on B. burgdorferi-infected C3H/HeJ mice, and fed larvae 448 were molted to nymphs. Nymphal ticks were maintained at 85% relative humidity with a 449 Fed-nymph gut cDNA was prepared as described above and used as template to 490 amplify segments of targeted genes. The PCR primers with T7 promoter sequences are 491 shown in Table S4. Double-stranded RNA (dsRNA) were synthesized using the 492 MEGAscript RNAi kit (Invitrogen, #AM1626M) using PCR-generated DNA template that 493 contained the T7 promoter sequence at both ends. The dsRNA quality was examined by 494 agarose gel electrophoresis. DsRNA of the Aequorea victoria green fluorescent protein 495 (GFP) was used as a control. Pathogen-free and -infected tick nymphs were injected in 496 the anal pore with dsRNA (6 nL) using glass capillary needles as described by 497

Effects of silenced genes on B. burgdorferi colonization and transmission 500 501
To examine the effects of silencing targeted genes on the colonization of B. 502 burgdorferi in the tick gut, dsRNA microinjected pathogen-free I. scapularis nymphs were 503 placed on B. burgdorferi-infected mice (C3H/HeJ) and allowed to feed to repletion. The 504 ticks were then collected for gut dissection. The B. burgdorferi burden in the tick gut was 505 quantified by amplifying flaB. FlaB was quantified by extrapolation from a standard curve 506 derived from a series of known DNA dilutions of flaB gene, and data was normalized to 507 tick actin. The knockdown efficiency of targeted genes was tested as described above. 508 Specifically, the expression of targeted genes was estimated with the ΔΔCT method 509 (Schmittgen and Livak, 2008) using the reference gene actin. To test the effects of 510 silencing ISARL on the transmission of B. burgdorferi, a group of three to five GFP or 511 ISARL dsRNA injected B. burgdorferi-infected nymphs were placed on each C3H/HeJ 512 mouse (at least five mice each in the GFP or ISARL dsRNA groups) and allowed to feed 513 to repletion. Ticks are placed on the mouse head/back between the ears. At 7 and 14 514 days-post tick detachment, the mice were anesthetized, and skin was aseptically punch 515 biopsied and assessed for spirochete burden by qPCR. Ticks feed in head area and skin 516 punch biopsies are collected from the pinnae /ears. This site is considered distal as it is 517 not at the site of tick bite. Twenty-one days post tick detachment, the mice were sacrificed, 518 and ear skin, heart, and joints were aseptically collected and assessed for spirochete 519 burden by qPCR. 520 521

RNA-seq and bioinformatic analyses 522 523
DsRNA (ds ISARL and ds GFP) microinjected pathogen-free I. scapularis nymphs 524 were placed on clean and B. burgdorferi-infected mice (C3H/HeJ), respectively, and 525 allowed to feed to repletion. Then, the ticks were collected for gut dissection. Total RNA 526 was purified as described above. In addition, to check the transcriptional alterations in the 527 tick gut in the presence of mammalian adiponectin, pathogen-free tick nymphs were 528 injected in the anal pore with recombinant mouse adiponectin (MilliporeSigma, #SRP3297) 529 and GFP proteins (SinoBiological, #13105-S07E). Then, the guts were dissected after 8h 530 injection, and RNA was purified. The RNA samples were then submitted for library 531 preparation using TruSeq (Illumina, San Diego, CA, USA) and sequenced using Illumina 532 HiSeq 2500 by paired-end sequencing at the Yale Centre for Genome Analysis (YCGA). 533 The I. scapularis transcript data was downloaded from the VectorBase 534 and Adipo -/mice (C57BL/6J) and allowed to feed to repletion. The ticks were then 600 collected for gut dissection. The B. burgdorferi burden in the tick gut was quantified as 601 The C1QL3 was PCR amplified from tick nymph cDNA using the primer pair listed 606 in Table S4, then cloned into the BglII and XhoI sites of the pMT/BiP/V5-His vector 607 (Invitrogen, #V413020). The recombinant protein was expressed and purified using the 608 Drosophila Expression System as described previously (Schuijt et al., 2011b). Statistical significance was assessed using a non-parametric Mann-Whitney test (ns, P > 867 0.05; *, P < 0.05; **, P < 0.01). Source data 1. Source data for PTDSS1 protein relative 868 quantification. Source data 2. Source data for PTDSS1 protein relative quantification. The plot profile of co-localization was conducted by Image J software. Source data 1. 915 Source data for ISARL expression. Source data 2. Source data for C1QL3 protein 916 purification. Source data 3. Source data for C1QL3 protein purification. Source data 4. 917 Source data for binding of C1QL3 to ISARL.     (G) ACTN has no effect on B. burgdorferi acquisition. Silencing of (D) 3HADH decreased the B. burgdorferi burden in tick gut. 3HADH is involved in fatty acid metabolic processes, suggesting that tick fatty acid metabolism may also influence acquisition of B. burgdorferi.
3HADH is not significantly regulated by ISARL, it was not considered further in this study.
Each data point represents one nymph gut. Horizontal bars in the above figures represent the median. Statistical significance was assessed using a non-parametric Mann-Whitney test (ns, P > 0.05; *, P < 0.05; **, P < 0.01).   "-" indicates downregulation of genes in the guts of ds ISARL-injected ticks when compared to that in control ds GFP-injected tick guts.