TOR signalling is required for host lipid metabolic remodelling and survival following enteric infection in Drosophila

When infected by enteric pathogenic bacteria, animals need to initiate local and whole-body defence strategies. While most attention has focused on the role of innate immune anti-bacterial responses, less is known about how changes in host metabolism contribute to host defence. Using Drosophila as a model system, we identify induction of intestinal target-of-rapamycin (TOR) kinase signalling as a key adaptive metabolic response to enteric infection. We find TOR is induced independently of the IMD innate immune pathway, and functions together with IMD signalling to promote infection survival. These protective effects of TOR signalling are associated with re-modelling of host lipid metabolism. Thus, we see that TOR switches intestinal metabolism to lipolysis and fatty acid oxidation. In addition, TOR is required to limit excessive infection mediated wasting of adipose lipid stores by promoting an increase in the levels of fat body-expressed de novo lipid synthesis genes. Our data support a model in which induction of TOR represents a host tolerance response to counteract infection-mediated lipid wasting in order to promote survival.


30
Animals are constantly exposed to bacterial pathogens in their environment. As a result, they must 31 be able to sense invading pathogens and then trigger appropriate defence responses. One defence 32 strategy is to decrease pathogen load (Schneider and Ayres, 2008). Central to this mechanism are 33 the innate immune responses. These are responsible for sensing invading bacteria at the sites of    infection response in Drosophila may be due to the different bacterial infections used or because of 89 differences in host metabolic or nutrient status. Nevertheless, they indicate that further work is 90 required to clarify how TOR may play a role in immune and metabolic responses to infection. We 91 address this issue in this paper. We show that enteric infection leads to increased TOR signalling 92 independently of innate signalling, and that this induction is required to remodel host lipid 93 metabolism and promote survival.

99
TOR kinase couples environmental signals to changes in cellular metabolism. Generally, TOR has 100 been shown to be activated by favorable conditions (e.g., abundance of nutrients and growth 101 factors), while being inhibited by stress conditions (e.g., starvation, low oxygen, oxidative stress). 102 We were interested in examining how TOR activity might be affected by enteric bacterial infection. 103 We first infected flies with the gram-negative bacteria Pseudomonas entomophila (P.e) for 4hr and 104 then dissected intestines for western blotting. Ribosomal protein S6 kinase (S6K), is directly 105 phosphorylated and activated by TOR, hence we used western blotting for phosphorylated S6K as a 106 readout for TOR activity. We found that oral P.e. feeding lead to increased phosphorylated S6K 107 levels ( Figure 1A). This increase was blocked by pre-feeding the flies rapamycin, a TOR inhibitor, 108 indicating that the induction of phosphorylated S6K was through an increase in TOR activity 109 ( Figure 1B). We also examined phosphorylation of ribosomal protein S6, a downstream target of 110 ribosomal protein S6 kinase. We saw that 4hrs of oral P.e infection also induced phosphorylated S6 111 levels in the intestine ( Figure 1C). Moreover, when we performed immunostaining with the anti-112 phosphorylated S6 antibody, we saw that the increase in TOR activity was apparent in all cell types samples, we saw that oral P.e. infection led to an increase in tRNA and pre-rRNA levels and an 119 increase in mRNA levels of three ribosome biogenesis genes, Nop5, ppan, fibrillarin upon oral P.e 120 infection ( Figure 1E). We explored these effects of oral bacterial infection further by performing a 121 time course following oral P.e. feeding. We saw that the induction of TOR was rapid (within 4hrs of 122 infection) and persisted for 24hrs of the oral infection period ( Figure S1A). We also found that this 123 induction of TOR was similar in males and females ( Figure S1B). Moreover, the effects of P.e. appear 124 limited to adults since 4hr oral infection in larvae didn't increase phosphorylated S6K levels, and in 125 fact showed a small decrease ( Figure S1C). We also tested two other pathogenic gram-negative 126 bacteria, Vibrio cholera (V.c.) and Erwinia carotovora carotovora (Ecc15). We again used western 127 blotting for phosphorylated S6K to measure TOR and saw that oral infection with V.c. and Ecc15 128 both led to increased intestinal TOR activity ( Figure 2A). Together, these data indicate that 129 induction of intestinal TOR kinase signalling is a rapid response to enteric gram-negative bacterial 130 infection and that it stimulates the protein biosynthetic capacity of intestinal epithelial cells. (PI3K/AKT pathway) and phosphorylated ERK (ERK pathway). We found that oral P.e. had no effect 139 on PI3K/AKT or ERK pathway, suggesting that P.e. effects may be specifically inducing TOR 140 activation ( Figure S1D). We next explored whether other enteric intestine stresses might also 141 regulate intestine TOR signalling. Feeding flies with three known chemical intestine stressors, 142 bleomycin (a DNA damaging agent), dextran sodium sulphate (a detergent) and paraquat (an 143 oxidative stressor), had no effect on intestine TOR activity ( Figure 2B). We also explored two 144 nutrient stresses -high sugar and high fat. However, we saw that feeding the flies either a high 145 sugar (40%) or high fat (30%) supplemented diet, also did not have any effect on TOR signalling in 146 the intestine ( Figure 2C). Together our data suggests that the induction of TOR appears specific to 147 oral bacterial infection. to gram-negative infections. We began by testing the involvement of IMD signalling by using 158 mutants for Imd and Relish, two components of the pathway. We infected either control (w 1118 ) or 159 either rel or imd mutants with P.e. for 4hr and then measured intestinal phosphor S6K levels. We  Inhibiting TOR and IMD pathways simultaneously reduces survival upon P.e. infection.

180
We next examined the consequences of TOR induction upon P.e. infection. We first examined effects 181 on survival following enteric P.e. infection. Under our laboratory fly culture conditions, the strain of 182 P.e. we use is not strongly pathogenic. Thus, when we infected control (w 1118 ) adult flies for 2 days 183 and then monitored their survival over approximately three weeks, we saw little effect on viability 184 compared to uninfected flies ( Figure 3E). When we infected flies and simultaneously inhibited TOR 185 by feeding flies rapamycin, we found that this induced a slight, but significant, decrease in survival 186 compared to flies fed rapamycin alone ( Figure 3E). We next tested the possibility that TOR 187 functions in parallel IMD/Relish signalling to promote infection survival. We found that relish 188 mutants had a generally reduced lifespan on our normal lab food compared to control (w 1118 ) adults 189 ( Figure 3F). Either enteric infection with P.e., or blocking TOR with rapamycin, had no effect on  We next examined what role TOR might play in these lipid effects by measuring TAG levels at 0-, 1-225 and 3-days following infection in control vs rapamycin-treated flies. In control flies, infection led to 226 a transient decrease in TAG levels at the 1-day timepoint, but then TAGs recovered to the same level 227 as uninfected flies at 3 days ( Figure 5D). Rapamycin treatment alone had no significant effect on 228 TAG levels at any timepoint compared to uninfected control flies ( Figure 5D). However, when we 229 infected flies and simultaneously fed them rapamycin to inhibit TOR, we saw a progressive 230 depletion of TAG stores at each timepoint following infection ( Figure 5D). These results suggest 231 TOR is needed to limit excessive loss of lipid stores following infection. To do this, TOR may be 232 blocking excess lipase function (to limit lipolysis) or may be increasing lipid synthesis (to resupply 233 new lipids). We found that infected flies showed a significant upregulation in mRNA expression    Figure 7C). Moreover, we saw that infection-mediated 256 mobilization of glycogen was reduced in rapamycin-fed animals ( Figure 7D). Taken together, these 257 results suggest that infection leads to depletion of stored glycogen in part through TOR signalling. 258 Thus, one possibility is that this TOR-dependent mobilization of glycogen is used to provide glucose 259 for the TOR-induced de novo synthesis of TAGs.

297
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which this version posted July 12, 2021. ; https://doi.org/10.1101/2021.07.12.452110 doi: bioRxiv preprint We found that the TOR induction was not required for induction of the AMPs the main anti-298 bacterial resistance response in flies. The AMPs are primarily induced by IMD/Relish signalling 299 following enteric infection. Interestingly we saw the infection survival was reduced only when we 300 simultaneously blocked both IMD signalling (relish mutants) and TOR signalling (rapamycin 301 feeding). Based on these data, one simple model is that upon infection, the IMD pathway is induced 302 to initiate resistance (antibacterial defences), while TOR induction plays a role in tolerance 303 responses (adaption to pathogen infection).

305
Tolerance responses are defined as alterations in host biology that limit pathology and promote

322
In the context of infection-mediated lipid mobilization, we saw that the main function for TOR 323 appeared to be limit excess lipid loss. Thus when we rapamycin-treated flies we saw that the

395
Immunostaining 396 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The tissues were then transferred to fresh PAT containing the primary antibody, overnight at 4°C.

401
The primary antibody incubation was followed by 3 washes with 1X PBT + 2% fetal bovine serum 402 (FBS) for 30 mins each. The tissues were then incubated with secondary antibody in PBT without     was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.  Adult w 1118 mated females were subjected to 4hr oral P.e. infection (B) or 24hr rapamycin pre-treatment followed by 4hr oral P.e. feeding (B). Dissected intestines were lysed and analyzed by western blotting using antibodies to phosphorylated-S6K and actin or tubulin (shown as loading controls). C, D) Adult w 1118 mated females were subjected to 4hr oral P.e. infection. Intestines were then either lysed and processed for western blotting using antibodies to phosphorylated ribosomal protein S6 and actin, (C) or stained with phosphorylated-S6 antibody for immunofluorescence, (D  Figure 3. TOR and IMD signaling function in parallel to control survival in response to enteric infection. A) Adult w 1118 and Immune deficiency (imd) mutants subjected to 4hrs oral P.e. infection. Dissected intestines were lysed and processed for western blotting using antibodies against phosphorylated-S6K and actin (loading control). b Adult w 1118 and Rel/NF-kB transcription factor relish mutants subjected to 4hrs oral P.e. infection. Dissected intestines were lysed and processed for western blotting using antibodies against phosphorylated-S6K and actin (loading control). C) Adult w 1118 mated females subjected to feeding for 4hrs with P.e. alone (left) and P.e. + antioxidant N-acetyl cysteine, NAC (right). Dissected intestines were lysed and processed for western blotting using antibodies against phosphorylated-S6K levels and actin (loading control). D) Adult w 1118 mated females subjected to 4hr uracil feeding. Dissected intestines were lysed and processed for western blotting using antibodies against phosphorylated-S6K levels and actin (loading control). E) Survival plot of control w 1118 (E) and relish E20 (rel) mutant (F) mated female flies subjected to 48hr oral P.e. infection. Animals were then returned to standard food and the percentage of animals surviving was counted. N = at least 50 animals per experimental condition. *p< 0.05, log rank test.  qRT-PCR analysis on adult w 1118 mated females subjected to a 24hr pre-treatment of rapamycin or DMSO control followed by 24hr oral P.e. feeding along with rapamycin. mRNA transcript levels of anti-microbial peptides (AMPs) are presented as relative changes vs control (corrected for RpS9). The bars represent the mean for each condition, with error bars representing the S.E.M and individual values plotted as symbols. ns = not significant, two-way ANOVA followed by Students t-test,  A) Adult w 1118 mated females were subjected to 24hr oral P.e. infection and then whole-body glycogen levels were measured. The bars represent the mean for each condition, with error bars representing the S.E.M. and individual values plotted as symbols. * p<0.05, Students t-test. B) w 1118 mated females subjected to 24hr of either sucrose (control) or 24hr oral P.e. feeding (orange bars) and then processed for qRT-PCR analysis of genes involved in glycogen breakdown. The bars represent the mean for each condition, with error bars representing the S.E.M. and individual values plotted as symbols. * p<0.05, Students t-test. C) w 1118 mated females were pretreated for 24 hours with either DMSO (control) or rapamycin, followed by 24hr of either sucrose (control) or 24hr oral P.e. feeding (grey bars). Whole animals were then processed for qRT-PCR analysis of GlyP mRNA. The bars represent the mean for each condition, with error bars representing the S.E.M and individual values plotted as symbols. *, p<0.05, two-way ANOVA followed by Students t-test. C) w 1118 mated females were pretreated for 24 hours with either DMSO (control) or rapamycin, followed by 24hr of either sucrose or 24hr oral P.e. feeding. Whole animals were then processed for measurement of total glycogen assays. Left, the bars represent the mean for each condition, with error bars representing the S.E.M. and individual values plotted as symbols. * p<0.05, two-way ANOVA, followed by Students t-test. Right, the data are presented as the percentage decrease in whole-body glycogen levels upon infection in control vs. rapamycin-treated samples. The bars represent the mean for each condition, with error bars representing the S.E.M. and individual values plotted as symbols. *p<0.05, Students t-test.