Enterococcus faecalis redox metabolism activates the unfolded protein response to impair wound healing

58 Enterococcus faecalis is a gut commensal and opportunistic pathogen that causes difficult -to-treat 59 biofilm-associated infections, including catheter -associated urinary tract infection, infective 60 endocarditis, and chronic wound infections1,2. In wound settings, E. faecalis infection is associated with 61 delayed epithelial migration and immune dysregulation 3. The success of E. faecalis in these 62 environments is often attributed to its metabolic adaptability, including survival under nutrient limitation 63 and oxidative stress4. However, the extent to which E. faecalis metabolism actively interferes with host 64 repair mechanisms is poorly understood. 65 66 One way that host cells respond to environmental and infection-induced insults is through the unfolded 67 protein response (UPR), an evolutionarily conserved signalling pathway triggered by endoplasmic 68 reticulum (ER) stress. The UPR integrates signals related to protein misfolding, membrane 69 perturbations, and redox imbalance to restore homeostasis or induce apoptosis if stress persists5–8. 70 Pathogens have evolved diverse strategies to manipulate the host UPR, often targeting its three main 71 pathways (IRE1, PERK, and ATF6) to subvert host cell function, modulate immune responses, or even 72 exploit UPR-regulated products as a nutrient source9–14. While some bacterial toxins and effectors can 73 induce the UPR, in most cases, the microbial mechanisms by which the UPR is activated or 74 dysregulated are undefined11,15–17. 75 76 E. faecalis generates substantial extracellular reactive oxygen species (ROS), including superoxide and 77 hydrogen peroxide, in the absence of aerobic respiration or fumarate reduction 18,19. These ROS can 78 inflict DNA damage and tissue injury in infection models, as well as modulate host signalling pathways 79 including redox-sensitive pathways like the UPR20–22. These ROS can inflict DNA damage and tissue 80 injury in infection models, as well as modulate host signalling pathways including redox -sensitive 81 pathways like the UPR20,21,23. However, the bacterial source of ROS in E. faecalis and its mechanistic 82 consequences for host cell function have not been fully elucidated. 83 84 In this study, we show that E. faecalis activates the host UPR during wound infection. Through a forward 85 genetic screen and functional validation, we identify extracellular electron transport (EET) as a 86 previously unrecognised mechanism by which E. faecalis generates ROS, which in turn activates the 87 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 3 UPR in epithelial cells and impedes their migration following wounding . We show that EET and 88 associated demethylmenaquinone (DMK) biosynthesis pathways are required for superoxide and 89 hydrogen peroxide generation, mutants of which produce less ROS, fail to activate the UPR, and do 90 not impair epithelial cell migration. These findings not only establish a novel function for EET in ROS 91 generation but also through its interaction with host UPR, as a novel metabolic virulence mechanism 92 by which E. faecalis disrupts epithelial repair, thereby presenting new opportunities for targeting chronic 93 E. faecalis-driven pathologies. 94 95

96 E. faecalis infection activates the UPR in a mouse model 97 We previously show ed that E. faecalis infection impairs wound healin g3. T o gain insight into the 98 mechanisms that influence delayed wound repair , we re-analysed our published single-cell RNA-seq 99 dataset of E. faecalis infected mice wounds at 4 days post-infection22 (dpi) (Figure 1A, GSE229257). 100 For each class we calculated enrichment scores for a panel of stress-response signatures24 (Table S1). 101 Gene-set enrichment for canonical UPR targets revealed that the ER stress-response is not global but 102 concentrated in immune cells (macrophages and neutrophils) and, most strikingly, in an infection -103 specific cluster of keratinocytes ( Figure 1B and 1C). By contrast, oxidative -stress response (OSR) 104 genes were upregulated not only in UPR-elevated keratinocytes and immune cells but also in fibroblasts 105 (Figure 1 D, S1A, and S1 B), consistent with high fibroblast redox activity in infected tissue. We 106 previously found that E. faecalis infection interferes with wound closure signatures, drives a partial 107 epithelial-to-mesenchymal transition (EMT) in keratinocytes, and skews macrophages toward an anti-108 inflammatory phenotype22. The robust UPR response in keratinocytes offers a mechanistic clue where 109 excessive ER stress could amplify the infection -induced EMT shift in these cells, undermining their 110 migratory role and thereby hindering wound repair during E. faecalis infection. 111 112 To corroborate the scRNA-seq findings, we infected full thickness excisional wounds in mice with E. 113 faecalis strain OG1RF and enumerated bacterial colony forming units (CFU) from the wounds at 6 days 114 post-infection (6 dpi), chosen to target the proliferation and remodelling phase of healing, where we 115 observed a 2-log reduction in E. faecalis bacteria burden in wounds compared to the starting inoculum, 116 with no significant animal weight loss or attrition, consistent with our previous studies3 (Figure S1C and 117 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 4 S1D). We quantified UPR-associated mRNA levels of spliced Xbp1 (Xbp1s), Chop, and Herpud1 as 118 markers of Ire1, Perk1, and Atf6 activity, respectively, in whole wound tissue (Figure 1 E). Xbp1s 119 transcripts were significantly higher in uninfected wound s at 6 d pi compared to unwounded skin, 120 indicating that the UPR is activated in wounds regardless of infection state (Figure 1F). A similar 121 observation was reported in wounded transgenic mice expressing a XBP1-Luc fluorescence marker 8-122 10 days post-wounding25. By contrast, both Xbp1s and Chop transcript levels were significantly higher 123 in E. faecalis-infected wounds at 6 dpi compared to unwounded skin samples, while Herpud1 levels 124 remained unchanged, indicating a lack of ATF6 activation (Figure 1F). Taken together, t hese data 125 suggest that E. faecalis infection results in UPR dysregulation, which could impact normal wound 126 healing26. 127 128 IRE1 activation by E. faecalis impedes keratinocyte migration in vitro 129 To corroborate our in vivo and in silico findings, we examined UPR activation in keratinocytes (HaCaT) 130 and fibroblasts (NIH-3T3) which are the dominant cell types in healthy skin that contribute to wound 131 healing27. E. faecalis infection significantly increased mRNA expression of all three UPR pathway 132 markers in NIH-3T3 cells (Figure 2A), whereas in HaCaT cells only XBP1s and CHOP were significantly 133 upregulated (Figure 2B). As a positive control, we treated both cell lines with tunicamycin (Tm), which 134 induces the UPR by inhibiting protein glycosylation in the ER leading to an accumulation of unfolded 135 proteins28. Tm treatment significantly upregulated all three UPR pathway markers in both cell lines to a 136 greater extent compared to E. faecalis infection. Since IRE1 is the most evolutionarily conserved branch 137 of the UPR, we further examined its activation by E. faecalis by assessing the expression of XBP1s 138 target genes29,30. These included the ER chaperone BiP (encoded by HSPA5) and EDEM1, which 139 promote ER homeostasis and are hallmarks of IRE1 activation31,32. Infected cells exhibited a significant 140 increase in EDEM1 transcripts along with elevated levels of XBP1s and BiP proteins (Figure 2C and 141 2D), confirming that E. faecalis activates the conserved IRE1 pathway in vivo and in vitro. 142 143 We next investigated the impact of infection-induced UPR on wound closure using an in vitro HaCaT 144 scratch wound assay. By 15 hpi, both E. faecalis-infected and Tm-treated cells exhibited significantly 145 slower migration compared to uninfected cells, with the difference becoming more pronounced by 27 146 hpi (Figure 2E and S2A; Video S1 ). Notably, neither infected nor Tm-treated cells displayed 147 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 5 appreciable migration throughout the assay period, and infected cells even show ed signs of wound 148 edge retraction as early as 6 hpi, consistent with cell shrinkage or detachment which can be early 149 indicators of apoptosis. To determine whether UPR induction via IRE1 was responsible for impaired 150 migration, we treated both uninfected and infected cells with IRE1 inhibitor (4µ8c), which blocks the 151 RNase activity of IRE133 (IRE1i). Treatment with 50 µM IRE1i was sufficient to block E. faecalis-induced 152 IRE1 in HaCaT cells (Figure 2F). IRE1i also slowed migration in uninfected wells, with the delay evident 153 by 18 hpi and more pronounced at 27 hpi. (Figure 2G and S2B; Video S2). In infected cells, migration 154 was minimal regardless of IRE1i treatment, showing no significant migration at 27 hpi relative to the 3 155 hpi baseline. 156 157 We did not use proliferation inhibitors, such as mitomycin C, which are typically used to differentiate 158 between proliferation and migration in these assays, because the compound caused widespread cell 159 detachment in infected cells during preliminary experiments. Nonetheless, we attribute the observed 160 wound closure primarily to cell migration rather than proliferation for several reasons: (i) the short 24-h 161 duration of the assay, (ii) the long doubling time of confluent HaCaT cells (approx. 32 -36 h), and (iii) 162 previous studies have established that migration is the dominant factor in similar short -term scratch 163 assays26. Thus, our findings suggest that neither UPR hyperactivation (during infection) nor 164 hypoactivation (after IRE1 inhibition) alone fully explains keratinocyte migration; rather, both extremes 165 of UPR activity are implicated. 166 167 Extracellular electron transport drives E. faecalis UPR activation and migration arrest 168 To dissect how E. faecalis induces the UPR, we designed a high-throughput assay for UPR induction: 169 a NIH-3T3 reporter line (3T3R) that fluoresces when IRE1 splices a 26-nt intron from a truncated human 170 XBP1 fused to mApple (Figure 3A and 3B). While IRE1 is quiescent, a premature stop codon between 171 XBP1 and mApple blocks translation, causing cells to remain non-fluorescent. When IRE1 is activated, 172 intron excision shifts the reading frame, removes the stop codon, and allows production of the full 173 XBP1s-mApple fusion, generating a red signal quantifiable by wide -field microscopy. Using this 174 reporter, we screened a defined transposon (Tn) library of 14,976 E. faecalis OG1RF mutants (Figure 175 3A) to identify mutants that failed to trigger UPR activation. 176 177 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 6 Applying an upper threshold of 30% XBP1s-mApple positive (XBP1s+) cells in a given population, we 178 identified 457 UPR defective mutants, corresponding to 369 unique genetic loci (Figure S3A). We 179 filtered out mutants with multiple transposon insertions, insertions outside of coding regions, and those 180 with significant growth defects after overnight culture in BHI media. Pathway analysis did not show 181 consistent enrichment patterns; however, w e notice d that a substantial number of mutants had 182 insertions in genes associated with carbohydrate catabolism and respiration, key components of redox 183 metabolism. This observation prompted us to focus on pathways involved in the synthesis of 184 components and co -factors that facilitate electron flow from carbohydrate catabolism to terminal 185 electron acceptors. Transposon insertions mapped to three respiratory pathways of E. faecalis: aerobic 186 respiration (ndh2, cydD), fumarate reduction (frd), and extracellular electron transport (EET) (ndh3, 187 eetB) (Figure 3C). Additional insertions were identified in genes involved in the biosynthesis of the 188 quinone electron carrier demethylmenaquinone (DMK) (menF, menE), as well as upstream precursors 189 such as chorismate and geranyl pyrophosphate, generated either by the shikimate pathway (aroD) or 190 the mevalonate pathway (ispA). DMK is an integral part of all three respiratory pathways by mediating 191 electron transfer between their membrane-associated components. 192 193 We validated mutants of these respiratory pathways identified in the primary screen and found that only 194 mutants disrupted in quinone electron carrier synthesis (menE, menF, aroD), and EET ( ndh3, eetB) 195 displayed significantly reduced UPR induction compared to wild-type E. faecalis (WT) (Figure 3D). 196 Since these genes are involved in central metabolic processes, we quantified the growth of each mutant 197 to ensure that reduced UPR induction was not simply due to impaired bacterial replication during 198 infection. However, all Tn mutants grew similarly to WT in cell culture media (Figure S3B and S3C), 199 eliminating growth defects as a confounding factor . Given the strong association between EET and 200 UPR induction, we also tested a deletion mutant lacking entire EET operon (ΔEET) (Figure 3C and 201 3D). As expected, the ΔEET mutant was defective in UPR induction, yet retained growth and antibiotic 202 susceptibility profiles similar to the parental OG1RF WT strain (Figure S3B-S3D). To rule out the 203 possibility that lower UPR induction by ΔEET could be due to higher cytotoxicity causing cell loss 204 resulting in weaker XBP1s-mApple fluorescent signals, we assessed cytotoxicity in HaCaT cells at 3 205 and 24 hpi following infection with WT and ΔEET. There was no difference in cytotoxicity between 206 uninfected, WT- or ΔEET-infected cells at 3 hpi (Figure S3E). However, WT but not ΔEET-infected cells 207 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 7 had significantly higher cytotoxicity compared to uninfected cells at 24 hpi. The significantly higher 208 cytotoxicity in WT-infected cells at a delayed timepoint of 24 hpi but not immediately after infection at 3 209 hpi suggests that WT infection was not directly causing cell death but rather indirectly via UPR 210 hyperactivation. T he lack of difference at 24 hpi between uninfected and ΔEET -infected cells also 211 confirms the UPR defective nature of ΔEET, as delayed cell death is a hallmark characteristic of chronic 212 UPR hyperactivation34,35. To further confirm that the loss of UPR induction was specific to the EET 213 pathway and not to secreted toxins, as is the case for Group A Streptococcus14,36, we tested a deletion 214 mutant of the Type 7 secretion system (ΔT7SS), encoding predicted secreted toxins 37. The ΔT7SS 215 mutant induced the UPR to the same extent as WT (Figure S3F), supporting the conclusion that the 216 phenotype of the ΔEET mutant is a direct consequence of its function in redox metabolism. 217 218 Based on these characteristics, ΔEET was selected as the model UPR defective mutant for downstream 219 studies. We used this mutant to assess whether E. faecalis UPR induction via the EET pathway 220 contributes to the inhibition of keratinocyte cell migration. Unlike WT infection, ΔEET did not significantly 221 impair migration, instead showing similar migration, comparable to uninfected controls (Figure 3E and 222 S3G; Video S3 ). However, treating ΔEET-infected cells with IRE1i resulted in significantly slower 223 migration at the 27 hpi endpoint (Figure 3 F and S3H). The se findings suggest that lack of UPR 224 induction by ΔEET allows for physiological levels of UPR induction that support normal cell migration. 225 Furthermore, IRE1i treatment induces UPR hypoactivation in ΔEET-infected cells (Figure 3F and S3H; 226 Video S4), dysregulating UPR homeostasis and impairing cell migration, similar to that observed in 227 uninfected cells treated with IRE1i (Figure 2G, Figure 3G). The recovery of cell migration with ΔEET 228 infection and its reversal by IRE1i treatment demonstrate that E. faecalis EET is associated with UPR 229 induction and impaired cell migration. 230 231 EET-derived ROS is sufficient to activate the UPR, disrupting epithelial migration 232 Disruption of DMK synthesis in E. faecalis impairs both EET function and extracellular superoxide 233 generation18,38. OG1RF mutants disrupted in genes involved in quinone electron carrier synthesis (aroE, 234 aroC, aroA, menB, menD, menE ) produce less superoxide (O2•–) than WT19. Superoxide radicals 235 undergo pH-dependent spontaneous dismutation to generate hydrogen peroxide (H2O2) which can then 236 participate in Fenton chemistry to generate hydroxyl radicals and other ROS in the presence of 237 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 8 transition metals (Figure 4A). H2O2 has been shown to induce the UPR in myotubule and epithelial 238 cells at 50 µM and 200 µM , respectively20,23. However, this may not be a universal response among 239 host cells as another group reported no significant increase in XBP1s expression for fibroblast when 240 treated with 1 mM H2O239. We hypothesised that superoxide generation via EET drives UPR induction 241 in epithelial cells via H2O2. To test whether H2O2 alone is sufficient to induce the UPR in our models, we 242 treated 3T3R cells with increasing concentrations of H2O2 for 3 hours followed by a 21 -hour recovery 243 period. XBP1s fluorescence increased significantly at 250 and 500 µM (Figure 4B). This finding 244 confirms that NIH-3T3 cells can mount a UPR in response to H2O2, supporting the idea that ROS 245 generated via EET contributes to UPR induction during infection. 246 247 Next, to test whether EET indeed generates ROS , we quantified superoxide generation by validated 248 UPR-defective mutants (ΔEET, and transposon insertion mutants in menE, menF, arodD, ndh3, eetB). 249 At the same time, we generated in-frame deletion mutants for each of these genes. All exhibited 250 significantly lower superoxide generation compared to WT and non-UPR defective mutants (ispA, frd, 251 ndh2, cydD), and the in-frame deletion mutants were similar to their respective transposon mutants 252 (Figure 4C and S4A), supporting a link between EET -dependent superoxide production and UPR 253 activation. 254 255 Superoxide dismutase (SOD) and catalase are antioxidants that catalyse the dismutation of superoxide 256 radicals into H2O2 or H2O2 into water and oxygen, respectively (Figure 4A). To determine whether E. 257 faecalis-derived ROS drives UPR induction, we treated infected 3T3R cells with exogenous SOD and 258 catalase. In control experiments, neither enzyme altered XBP1s expression in uninfected cells, whereas 259 treatment with 250 µM H 2O2 to simulate ROS -induced stress robustly increased XBP1s expression 260 (Figure 4D). In cells stimulated with H 2O2, the addition of catalase significantly reduced XBP1s 261 expression, whereas SOD had no effect. This result confirms that H2O2 is the specific reactive oxygen 262 species driving UPR induction in this assay. 263 264 In WT-infected cells, catalase again significantly reduced XBP1s expression, while SOD had no effect 265 (Figure 4E). Adding H2O2 to WT-infected cells did not further increase UPR activation, suggesting that 266 E. faecalis-generated ROS levels are already saturating whereas co-treatment with catalase reversed 267 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 9 this effect. By contrast, catalase had no impact on ΔEET-infected cells, consistent with their low ROS 268 generation (Figure 4F). However, functional complementation of the ΔEET with H2O2 partially restored 269 UPR induction, which was again reversed by catalase, confirming that H2O2 is sufficient to rescue the 270 UPR defective phenotype when EET is disrupted . These findings were replicated in HaCaT cells 271 (Figure 4G-4I). Since lipid peroxidation is a known cause of UPR activation40. We next measured it in 272 infected cells using a BODIPY 581/591 C11 fluorescent probe. WT infected cells exhibited significantly 273 more lipid peroxidation than uninfected controls, while cells infected with the ΔEET mutant showed no 274 significant change (Figure 4J). These data support a model where EET-derived ROS from E. faecalis 275 induces the UPR via lipid peroxidation. 276 277 To determine whether E. faecalis generated ROS impairs wound healing, we assessed in vitro wound 278 closure after treatment with H 2O2 or/and catalase. H2O2 treatment of uninfected cells significantly 279 slowed cell migration at 27 hpi, although no wound edge retraction was observed, unlike in WT-infected 280 cells (Figure 4K and S4B; Video S5 ). Co-treatment with catalase restored cell migration to 281 physiological levels . Similarly, catalase treatment of WT-infected cells significantly improved cell 282 migration, which approached that of uninfected controls and without retraction at 27 hpi (Figure 4L and 283 S4C; Video S6). Adding H2O2 to untreated WT-infected cells had no further effect, consistent with UPR 284 saturation seen earlier (Figure 4E and 4H ). Finally, co-treatment with H2O2 and catalase mirrored 285 catalase-treatment alone, reinforcing the role of E. faecalis-derived H2O2 in cell migration arrest (Figure 286 4L and S4C; Video S6). Altogether, these findings demonstrate that the EET pathway of E. faecalis 287 generates ROS which oxidises lipids in epithelial cells, hyperactivating the UPR , and inhibiting cell 288 migration. 289 290

291 Our findings identify E. faecalis extracellular electron transport (EET) as a novel virulence mechanism 292 that links bacterial redox metabolism to host stress responses and impaired tissue repair. We show that 293 EET-dependent production of reactive oxygen species (ROS), particularly H2O2, activates the unfolded 294 protein response (UPR) in epithelial cells. UPR dysregulation by E. faecalis disrupts normal epithelial 295 function and cell migration , revealing a direct mechanistic connection between bacterial energy 296 metabolism and host healing processes. 297 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 10 298 E. faecalis is known to produce ROS, which have been previously linked to host cellular injury and even 299 carcinogenesis21. In this study, we refine the genetic basis of this activity in terms of a dedicated electron 300 transport system. Previous studies showed that E. faecalis generates extracellular superoxide through 301 a process that requires demethylmenaquinone (DMK) , which is diminished when terminal 302 quinol/cytochrome oxidases are functional, suggesting that respiratory disruption changes the dynamics 303 of electron flow to one that favours the reduction of oxygen to superoxide, resulting in ROS production19. 304 In this model, NADH incompletely reduces DMK, producing semiquinone intermediates which, in the 305 absence of classical aerobic or anaerobic respiration, donate electrons univalently to molecular oxygen, 306 generating superoxide which spontaneously dismutate into hydrogen peroxide which can contribute to 307 host oxidative stress 19. This movement of electrons may be mediated by soluble shuttles, such as 308 flavins41,42. In this study, we refine the mechanistic basis of E. faecalis ROS generation by identifying a 309 critical role for extracellular electron transport (EET) in this process. 310 311 While disrupting aerobic respiration is required for ROS generation and EET function19,38, our data show 312 that the disruption of aerobic respiration alone is not sufficient and that ROS generation also depends 313 on an intact EET system. However, the mechanistic contribution of specific EET components involved 314 in ROS generation remains to be determined. Whether these components overlap with those required 315 for extracellular metal reduction or electrode respiration or represent a distinct branch of EET machinery 316 activated under redox stress, remains an important question for future work. 317 318 The requirement for EET in E. faecalis ROS production aligns with growing evidence that extracellular 319 electron transfer in Gram -positive bacteria extends beyond classical anaerobic respiration. While 320 canonical diderm systems in general like Shewanella and Geobacter use cytochromes and conductive 321 pili to transfer electrons to external acceptors 43,44, re cent studies show that monoderm bacteria, 322 including Listeria monocytogenes and Lactobacillus plantarum, use flavin-based EET to support energy 323 conservation, redox homeostasis, and virulence under host-relevant conditions41,42,45. In particular, E. 324 faecalis and L. monocytogenes rely on EET for fitness in the mouse gastrointestinal tract 41,46, and L. 325 plantarum employs EET to enhance ATP yield via substrate-level phosphorylation in the absence of a 326 classical respiratory chain47. Our findings position E. faecalis within this emerging framework and reveal 327 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 11 a distinct facet: in the presence of oxygen within a heme-free environment, EET contributes to univalent 328 electron transfer from reduced DMK to oxygen, generating superoxide and hydrogen peroxide. Unlike 329 environmental microbial EET systems that avoid oxygen to prevent radical formation, E. faecalis 330 appears to exploit this interaction, linking EET to oxidative stress at the host-pathogen interface. 331 332 The UPR is increasingly recognized as a central node in host stress responses during infection, 333 particularly in epithelial and immune cells. Multiple pathogens including Salmonella enterica , 334 Helicobacter pylori, Pseudomonas aeruginosa, Streptococcus pyogenes, and Brucella melitensis have 335 been shown to manipulate the UPR through secreted toxins or effector proteins, often to promote 336 intracellular survival or dampen immune response s9–11,14,48. Here w e demonstrate that E. faecalis 337 selectively activates the IRE1 and PERK arms of the UPR both in vivo and in vitro, manifesting in 338 impaired epithelial cell migration. Unlike previously described examples, E. faecalis induces the UPR 339 independent of dedicated virulence factors, instead leveraging metabolic ROS production via EET. 340 Similar redox-based virulence strategies have been described in other pathogens. S. pneumoniae 341 produces hydrogen peroxide via SpxB, contributing to epithelial damage49,50. P. aeruginosa phenazines 342 generate intracellular ROS in airway cells 51,52. Yet E. faecalis is the first example, to our knowledge, 343 where a defined EET system is shown to drive ROS production that directly alters host stress signalling 344 and function. 345 346 A compelling hypothesis arising from our data is that the observed UPR hyperactivation and impaired 347 cell migration are consequences of ferroptosis. Ferroptosis is a regulated, iron-dependent form of cell 348 death distinct from apoptosis, driven by the catastrophic accumulation of lipid peroxides. This process 349 is typically restrained by the antioxidant enzyme GPX4, and its failure leads to membrane damage53,54. 350 Ferroptosis is increasingly recognized as a critical factor in diverse pathologies and host -pathogen 351 interactions55–57. Importantly, t he accumulation of lipid peroxides, the biochemical hallmark of 352 ferroptosis is a known trigger of severe endoplasmic reticulum stress, providing a direct mechanistic 353 link to UPR activation40,58,59. Our findings align remarkably well with this framework. The lipid 354 peroxidation we observed in infected keratinocytes is the defining feature of ferroptosis, and the 355 subsequent cell retraction is a classic morphological correlate of this death pathway53,54,60. This model 356 mechanistically connects E. faecalis EET-driven ROS production to the downstream cellular 357 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 12 pathologies of lipid stress, UPR activation, and inhibition of cell migration. While our experiments with 358 catalase confirm ROS as the primary trigger, they do not exclude ferroptosis as the ultimate executioner 359 pathway. We therefore propose that the E. faecalis-host interaction may represent a novel model of 360 infection-induced ferroptosis. Testing this hypothesis, for which our study provides the foundational 361 rationale, will be a critical next step and could be directly addressed by employing specific inhibitors like 362 ferrostatin-1. 363 364 Our data further show that UPR activation must be tightly regulated for effective wound healing. These 365 findings are consistent with previous studies in aged keratinocytes and fibroblasts displaying higher 366 baseline levels of UPR markers and slower in vitro wound cell migration, which could be reversed upon 367 treatment with 4-phenylbutyrate, a broad-acting UPR inhibitor26. Similarly, we show that pharmacologic 368 inhibition of IRE1, the key UPR sensor, led to hypoactivation and impaired cell migration even in 369 uninfected cells, underscoring the importance of physiological UPR signalling during repair. Regulated 370 UPR induction is also necessary for other wound healing processes like the differentiation of dermal 371 fibroblasts to myofibroblasts which promote wound contracture and collagen deposition61. Furthermore, 372 UPR inhibitors especially PERK inhibitors have demonstrated significant cytotoxicity to pancreatic islet 373 cells which depend on mild UPR induction to perform their secretory function62. Even if UPR induction 374 is not lowered to below baseline levels, UPR inhibition will still be counterproductive in restoring 375 homeostasis in pathologies where host cells are already dependent on some level of UPR Induction to 376 perform physiological processes. Therefore, targeting the source of UPR dysregulation i.e. bacterial 377 ROS production, rather than host UPR, may constitute a more effective therapeutic strategy. 378 379 In summary, this work reveals that E. faecalis leverages its respiratory machinery not only for metabolic 380 flexibility but also to perturb host cell physiology. While EET has been linked to efficient infection of the 381 gastrointestinal tract, this work presents molecular details that may contribute to its role in pathogenesis. 382 Future studies should examine the role of EET in vivo , its regulation and contribution within 383 polymicrobial settings, and the potential for targeting redox metabolism to mitigate E. faecalis infections 384 that are increasingly recalcitrant to antibiotic therapy. 385 386 387 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 13 RESOURCES AVAILABILITY 388 389 Lead contact 390 Further information and requests for resources and reagents should be directed to and will be fulfilled 391 by the lead contact, Kimberly Kline ([email protected]). 392 393

availability 394 This study did not generate new unique reagents. 395 396 Data and code availability 397 This paper did not generate any new original datasets nor codes. 398 Any additional information required to reanalyse the data reported in this paper is available from the 399 lead contact upon request. 400 401

402 We are grateful to Thibault and Kline lab members for helpful discussions and critical reading of the 403 manuscript. We thank Gary Dunny and Jennifer Dale for providing the E. faecalis transposon mutants. 404 We would also like to thank the Centre for Biomedical Informatics (Drs. James A. Miller and Bernett 405 Teck Kwong Lee) and the NTU Optical Bio-imaging Centre (NOBIC) for their support. This work was 406 supported by funds from the National Medical Research Council Open Fund (MOH-000566 to GT and 407 KAK), the Singapore Ministry of Education Academic Research Fund Tier 1 (RG31/24 to GT), NTU 408 Research Scholarship to SYTL and RMKA (predoctoral fellowship), and the Swiss National Science 409 Foundation (SNSF grant 310030_212262 to KAK) . Parts of this work were also supported by the 410 National Research Foundation and Ministry of Education Singapore under its Research Centre of 411 Excellence Program (SCELSE). 412 413 AUTHOR CONTRIBUTIONS 414 Conceptualization, A.M.Z.T., G.T. and K.A.K.; Resources A.M.Z.T., M.V., C.S.M., and R.M.K.A.; Data 415 curation, C.C. and R.M.K.A.; Formal analysis, A.M.Z.T., C.C., S.Y.T.L., and R.M.K.A.; Validation, 416 A.M.Z.T.; Investigation, A.M.Z.T., C.C., S.Y.T.L.; Visualization, A.M.Z.T., C.C., S.Y.T.L. and G.T. ; 417 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 14 Methodology, A.M.Z.T., C.C., S.Y.T.L., G.T. and K.A.K.; Writing – original draft, A.M.Z.T., G.T. and 418 K.A.K.; Writing – review and editing, A.M.Z.T., C.C., G.T. and K.A.K.; Project administration, G.T. and 419 K.A.K. 420 421 DECLARATION OF INTERESTS 422 The authors declare no competing interests. 423 424

425 426 Bacterial strains and growth conditions 427 All bacterial strains used in this study are listed in Table S2. E. faecalis strains were routinely cultured 428 on brain-heart infusion (BHI; Neogen #NCM0016A) agar plates and grown in BHI broth. E. coli strains, 429 used for DNA isolation and plasmid manipulation, were cultured in Luria -Bertani (LB) broth or on LB 430 agar plates at 37°C. When required, antibiotics were added at the following final concentrations: 431 erythromycin (Em), 500 µg/ml for E. coli and 25 µg/ml for E. faecalis; rifampin (Rif), 25 µg/ml; and 432 chloramphenicol (Cm), 10 µg/ml. To prepare overnight cultures, a single E. faecalis colony was 433 inoculated into a 14 ml tube containing 4 ml of BHI broth, with the lid tightly sealed. Cultures were grown 434 statically for 18-24 hours at 37°C. The following day, the overnight culture was centrifuged (4,000 x g, 435 10 min), and the bacterial pellet was washed once with PBS before being resuspended in 1 ml of 436 complete DMEM (see Cell Culture section). The bacterial suspension was then normalized by optical 437 density (OD 600) to a concentration equivalent to 8x10⁸ CFU/ml (OD 600 = 1) and further adjusted 438 depending on the specific experimental application. 439 440 Mouse wound excisional model 441 All in vivo procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at 442 Nanyang Technological University, Singapore (Protocol #ARF SBS/NIEA -0314), in accordance with 443 national guidelines. Male C57BL/6J mice (6–7 weeks old, 22–25 g; InVivos, Singapore) were housed 444 under specific-pathogen-free conditions. The wound infection model was adapted from a previous 445 study3. Briefly, mice were anesthetized with 3% isoflurane, and dorsal hair was removed using clippers 446 and a depilatory cream (Nair). The skin was disinfected with 70% ethanol, and a 6 mm full -thickness 447 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 15 excisional wound was created using a sterile biopsy punch (Integra Miltex #33 -36). The wound was 448 immediately inoculated with 10 µl of an E. faecalis OG1RF suspension containing 2x10⁶ CFU. The 449 wound was then sealed with a transparent dressing (Tegaderm, 3M #7100252702) to prevent 450 contamination. Post-procedure, mice were housed individually to prevent wound disruption. At the 451 experimental endpoint, mice were euthanized, and a 1 cm2 piece of skin tissue centered on the wound 452 was excised. For bacterial enumeration, tissues were collected in sterile PBS, homogenized, and plated 453 on BHI agar supplemented with rifampin to confirm infection by the inoculated strain. For molecular 454 analysis, tissues were homogenized in either TRIzol reagent (Thermo Fisher #15596026) for RNA 455 extraction or ice-cold RIPA buffer (Thermo Scientific #89900) supplemented with a protease inhibitor 456 cocktail (Roche #11697498001) for protein extraction. 457 458 scRNA-seq integration and downstream analysis 459 Single-cell datasets (GSE229257) were re-processed in R 4.3.263 following the workflow established in 460 our original study22. Raw matrices were imported with Seurat 5.1.064–67. Cells expressing 6,000 genes or > 12 % mitochondrial reads were removed; genes detected in < 5 cells were discarded. 462 Library size variation was normalised with SCTransform (method = “glmGamPoi”, vst.flavour = “v2”). 463 Batch effects between biological replicates (uninfected vs. E. faecalis-infected wounds) were corrected 464 with Seurat’s reciprocal PCA integration (30 PCs). Principal components ( n = 30) were used for 465 FindNeighbors/FindClusters (resolution = 0.4) and RunUMAP (dims = 1–30). Broad cell classes were 466 assigned on canonical markers as previously reported22, sub-clustering of keratinocytes and fibroblasts 467 employed a second round of PCA/UMAP at resolution = 0.6. All UMAPs use colour-blind-safe palettes 468 generated with RColorBrewer 1.1-368 (brewer.pal, palette = “Set2”). 469 470 Differential expression between infected and uninfected cells within each cluster was ranked by the 471 Wilcoxon area-under-curve statistic using wilcoxauc69 (presto 1.0.0). Ranked lists served as input for 472 gene-set enrichment analysis with fgseaMultilevel (fgsea 1.28.0, minSize = 15, maxSize = 5,000, eps 473 = 0)70,71 Gene sets for unfolded protein response (UPR), oxidative -stress response (OSR) and heat 474 shock response (HSR) were curated from MSigDB (v2023.1). Enrichment was considered significant at 475 Benjamini–Hochberg FDR < 0. 0572. For each cluster, the positive, significant normalised enrichment 476 scores (NES) for infected cells were projected onto the UMAPs. 477 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 16 RNA extraction, reverse transcription, and quantitative real-time PCR 478 Total RNA was extracted from cell lines and homogenised mice wound samples using the EZ-10 479 DNAaway RNA Miniprep Kit (Biobasic #BS88136) following the manufacturer’s protocol. RNA 480 concentrations were quantified using Qubit Broad Range RNA Quantification Assay (Thermo Fisher 481 #Q10210) together with a Qubit 3 Fluorometer following manufacturer’s protocol. Complementary DNA 482 (cDNA) was synthesised from extracted RNA normalised to 1 µg per sample using RevertAid Reverse 483 Transcriptase (Thermo Fisher #EP0441) following manufacturer’s protocol. qPCR was performed using 484 Luna Universal qPCR Master Mix (New England Biolabs #M3003E) together with a CFX-96 (Bio-Rad) 485 or a QuantStudio3 (Thermo Fisher) Real-Time PCR system following manufacturer’s protocol. Each 20 486 µl reaction has a final concentration of 2.5 ng/µl of cDNA and 0.25 µM of primer pairs (Table S3) for 487 target genes. Relative mRNA was normalized to the housekeeping gene Gapdh/GAPDH using the 2-488 ∆∆Ct method73. 489 490 Cell culture 491 Murine embryonic fibroblasts (NIH -3T3) and human keratinocytes ( HaCaT) were cultured in DMEM 492 (Gibco # 11995065) supplemented with 10% heat -inactivated foetal bovine serum (FBS, Cytiva 493 #SV30160.03) and 4 mM GlutaMAX (Gibco #35050061) . This medium is referred to as "complete 494 DMEM. The lentivirus packaging line, 293FT, was cultured in complete DMEM further supplemented 495 with 0.1 mM MEM Non -Essential Amino Acids (Gibco #11140050), 6 mM L -glutamine (Gibco 496 #25030081), 1 mM Sodium Pyruvate (Cytiva #SH30239.01), and 500 μg/ml Geneticin (Gibco Gibco 497 #10131027). All cells were maintained at 37°C in a humidified 5% CO2 incubator. Cells were washed 498 once with PBS (Gibco #14190144) and detached using 0.25% Trypsin-EDTA. Trypsinization times were 499 15 min for HaCaT, 5 min for 293FT, 4 min for 3T3R, and 3 min for NIH-3T3 cells. The reaction was 500 neutralized with an equal volume of complete DMEM. Cells were pelleted by centrifugation , 501 resuspended in fresh medium, and counted using a Countess 3 Automated Cell Counter. For 502 experiments, cells were seeded in 12-well plates at densities of 2.85x104 cells/cm2 (NIH-3T3), 1.15x104 503 cells/cm2 (3T3R) or 1.15x105 cells/cm2 (HaCaT) and in cubated for 24 hours prior to use. For the 504 transposon scree, 3T3R cells were seeded in 96 -well plates at 2.85x104 cells/cm2. Unless otherwise 505 specified, the following final concentrations of reagents were used: tunicamycin at 0.2 µg/ml (for qPCR), 506 2.5 µg/ml (for scratch wound assays), or 5 µg/ml (for immunoblot/microscopy); the IRE1 inhibitor 4µ8c 507 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 17 at 50 µM (MedChemExpress #HY-19707), added 1 hour prior to infection; H2O2 at 250 µM; catalase at 508 100 units/ml; and superoxide dismutase (SOD, Sigma-Aldrich #S5395) at 100 units/ml. 509 510 Optimisation of in vitro infection 511 Assay conditions were optimised across (i) multiple timepoints and (ii) multiplicities of infection (MOI), 512 to identify those that maximise in vitro XBP1s expression (data not shown). 513 514 In vitro infection 515 Confluent NIH-3T3 and HaCaT cells were infected with E. faecalis at a MOI of 800 (800 CFU per host 516 cell) and 600 (600 CFU per host cell), respectively. Following a 3-hour infection period, the medium was 517 removed, and cells were washed three times with PBS. To eliminate extracellular bacteria, fresh 518 complete medium supplemented with a gentamicin-penicillin antibiotic cocktail (50 µg/ml) was added, 519 and the cells were incubated for an additional 21 hours74. 520 521 Immunoblotting 522 Cells were lysed with RIPA buffer (Thermo Fisher #89901) supplemented with reconstituted cOmplete 523 protease inhibitor cocktail ( Roche # 11697498001) by gentle agitation on ice for 5 min before 524 centrifugation for 15 min at 12,000 x g at 4°C. A mixture of 15 µg of total proteins was separated on 10% 525 SDS-PAGE and transferred on nitrocellulose membranes. Immunoblotting was performed with 526 appropriate primary antibodies and IRDye-conjugated secondary antibodies (Table S4). Proteins were 527 visualized using the NIR fluorescence system (Odyssey CLx Imaging System). 528 529 In vitro scratch wound assay model 530 Scratch assays were performed in 12-well plates, adapting a previously published protocol to facilitate 531 automated microscopy75. Confluent HaCaT cell monolayers were scratched with a sterile P200 pipette 532 tip and subsequently infected as described in the In vitro infection section. Following the post-infection 533 wash, fresh complete medium supplemented with antibiotics and 25 mM HEPES (Gibco #15630080) 534 was added to each well. Wound closure was monitored on a Zeiss Axio Observer 7 microscope (10x 535 magnification), acquiring brightfield images every 30 min for 45 h in a controlled environment (37°C, 5% 536 CO2). The resulting time -lapse images were analyzed using a customi sed CellProfiler pipeline to 537 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 18 quantify the scratch area76. To ensure accuracy, images where the wound area was misidentified by 538 the automated pipeline were manually measured in ImageJ using the Wound Healing Size Tool plugin77. 539 540 XBP1s reporter cell line 541 pLVX-XBP1-mNeonGreen-NLS plasmid was a gift from David Andrews ( Table S5). Codon-optimized 542 mApple cDNA was synthesized as a gBLOCK fragment (IDT) and inserted in pLVX -XBP1-543 mNeonGreen-NLS to replace mNeonGreen using Gibson Assembly (New England Biolabs #E2611L) 544 according to manufacturer protocol. The resulting plasmid, pLVX-XBP1-mApple-NLS, was sequence-545 verified and then co -transfected into 293FT cells with the pLP1, pLP2, and pLP/VSVG packaging 546 plasmids (Table S5). Supernatants containing lentiviral particles were harvested at 36 - and 60-hour 547 post-transfection, pooled, and filtered (0.45 μm). NIH-3T3 cells were then transduced with the filtered 548 virus for 24 h ours in the presence of 8 µg/ml of polybrene (Sigma -Aldrich #H9268). After 24 hours 549 recovery, infected cells were selected with 2 µM puromycin. Clonal cell lines were established by 550 seeding single cells into 96-well plates and expanding them for two weeks. Finally, positive clones were 551 validated by assessing homogenous fluorescent signal upon tunicamycin treatment. One validated 552 clone, designated the 3T3 reporter (3T3R) line, was selected for this study. 553 554 Transposon screen 555 A high-throughput screen was performed using an established E. faecalis OG1RF mariner transposon 556 library containing 14,976 mutants arrayed in 96-well plates78. Following overnight growth, the optical 557 density (OD 600) of each mutant culture was measured with a Tecan M200 microplate reader . For 558 infection, 5 µl of each culture was added to 3T3R cells s eeded in 96 -well plates, and plates were 559 centrifuged at 300 x g for 5 minutes to synchronize contact. Following infection, cells were stained with 560 2.5 µg/ml Hoechest 33342 (Thermo fisher #H21492) for 15 minutes, and the medium was then replaced 561 with phenol-red free DMEM (Gibco #31053028) supplemented with 25 mM HEPES , 1 mM sodium 562 pyruvate, and 4 mM GlutaMAX . Plates were imaged on a Zeiss CellDiscoverer 7 microscope (10x 563 magnification) using two fluorescence channels to detect the XBP1s-mApple reporter (Ex/Em: 570/594 564 nm) and Hoechst-stained nuclei (Ex/Em: 348/455 nm). The images were subsequently analysed with 565 CellProfiler 4.2.1 to quantify the percentage of UPR-positive (mApple-expressing) cells in each well. 566 567 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 19 Construction of in-frame deletion mutants in E. faecalis 568 General molecular biology reagents were sourced as follows: genomic DNA from E. faecalis was 569 isolated using the Wizard Genomic DNA Purification Kit (Promega #A1120), while plasmid DNA was 570 isolated from E. coli using the PureLink Plasmid Miniprep Kit (Invitrogen #K210011). All primers used 571 in this study (Table S6) were designed based on the E. faecalis OG1RF genome (NC_017316). Gene 572 fragments were amplified with Phusion High-Fidelity DNA Polymerase (Thermo Scientific #F530), and 573 routine screening was performed with Taq DNA Polymerase (NEB #M0273). T4 DNA ligase and all 574 restriction enzymes used according to the manufacturer's protocols (NEB). In-frame deletion mutants 575 were generated using the temperature -sensitive shuttle vector pGCP213 (Table S5 ), following a 576 previously described protocol79. Deletion constructs were created using two main strategies. For most 577 single genes and smaller operons, regions of approximately 450 bp flanking the target were amplified 578 from OG1RF gDNA; the upstream region was amplified with primer pair P1/P2 and the downstream 579 region with P3/P4. These fragments were then fused by overlap extension PCR using the outer primers 580 P1 and P4 and subsequently cloned into the PstI site of pGCP213. For the larger ΔEET and ΔT7SS 581 operons, a gBlock Gene Fragment (IDT) containing the fused upstream and downstream flanking 582 regions was synthesized and cloned into the vector. The resulting deletion constructs were transformed 583 into the appropriate E. faecalis parent strain by electroporation. Transformants were first selected on 584 BHI-erythromycin agar at the permissive temperature of 30°C. To promote chromosomal integration, 585 colonies were then passaged at the non-permissive temperature of 42°C with erythromycin selection. 586 Curing of the integrated plasmid was achieved by passaging the bacteria at 30°C in antibiotic-free BHI. 587 Finally, the successful deletion of the target gene or operon was verified by colony PCR using external 588 (Screen F/R) and internal (Intern F/R or Intern R) primer pairs (Table S6). 589 590 XBP1s fluorescent reporter assay 591 Following infection or treatment, 3T3R cells were stained with Hoechst 33342 (2.5 µg/ml) and 592 transferred to a phenol-red-free imaging medium. For each well, a 3x3 grid of images was acquired on 593 a Zeiss CellDiscoverer 7 microscope. The intensity of the XBP1s -mApple signal within each nucleus 594 (identified by Hoechst staining) was quantified using a CellProfiler pipeline, and the average intensity 595 per well was used to gauge the level of UPR induction. 596 597 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 20 Growth curve assay 598 To assess bacterial growth, overnight bacterial cultures were processed as described in bacterial strains 599 and growth conditions. After pelleting, they were normalised to a starting OD600 of 0.05 in phenol red 600 free complete media on a 96 -well plate which was sealed with a Breathe -Easy sealing membrane 601 (Sigma-Aldrich #Z380059) following manufacturer’s protocols . OD600 readings were taken at 10-min 602 intervals over 20 h. The rate of change of the OD600 readings was calculated at 50 -min intervals and 603 the highest rate of change was used to calculate the doubling time. 604 605 Antibiotic time-kill assay 606 To assess antibiotic killing, the supernatant of infected HaCaT cells were collected at 4, 5, 6, and 24 607 hpi. These were serially diluted on 96-well plates and 5 µl of the dilutions were spotted onto BHI-Agar 608 plates which were incubated at 37°C for 24 hours. Plates were imaged using a ProtoCOL3 Plus system 609 (Don Whitely Scientific) and bacterial colonies were manually enumerated on ImageJ using the multi -610 point tool. 611 612 Cytotoxicity assay 613 To quantify total cytotoxicity, both detached (floating in the medium) and attached cells were collected 614 and analyzed from infected HaCaT cell cultures. First, to collect the detached cell fraction, the culture 615 medium was harvested, and each well was washed once with 1 ml of PBS. This wash was pooled with 616 the collected medium, and the mixture was centrifuged (300 x g, 5 min). The resulting cell pellet was 617 carefully resuspended in 20 µl of complete medium. Next, to collect the attached cell fraction, the 618 remaining cells in the well were trypsinised for 15 min, neutralized with complete medium, pelleted by 619 centrifugation, and resuspended in 500 µl of fresh complete medium. The viability of both the detached 620 and attached cell suspensions was determined separately using a Countess 3 Automated Cell Counter 621 with trypan blue staining. Total cytotoxicity was then calculated by summing the number of non-viable 622 cells from both fractions and dividing by the total number of cells (viable and non -viable) from both 623 fractions (Equation 1). 624 Cytotoxicity = No. of dead cells (detached + adhered) No. of cells (detached + adhered) × 100% Eq (1) 625 626 627 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 21 Superoxide assay 628 Extracellular superoxide generation was measured by adapting a previously published cytochrome C 629 reduction assay for a 96-well plate format18. Briefly, E. faecalis cultures were normalized to an an OD600 630 of 0.005 in 200 µl of phenol red free complete media containing 20 µM of cytochrome C (Sigma-Aldrich 631 #C3131). The reduction of cytochrome C was measured as the change in absorbance at 550 nm (with 632 650 nm as the reference wavelength) every 2 minutes for 90 minutes at 37°C using a Tecan M200 633 microplate reader. The rate of superoxide generation was calculated from the maximal rate of change 634 in absorbance, after correcting for pathlength. This rate was determined using Beer's Law with an 635 extinction coefficient of 21.5 mM ⁻¹cm⁻¹ for reduced cytochrome C (Equation 2). To determine the 636 amount of superoxide specifically, the rate measured in a parallel reaction containing 100 units/ml of 637 superoxide dismutase (SOD; Sigma -Aldrich #S5395) was subtracted from the rate measured in its 638 absence. 639 Rate of superoxide generation (nmol per minute per 109CFU) = ∆MAXOD550 21.5 × 1 × 1×109 CFU 2×107 CFU × 60 600 × 109 106 Eq (2) 640 641 Lipid peroxidation assay 642 Lipid peroxidation was assessed in HaCaT cells using the Image-iT Lipid Peroxidation Kit (Thermo 643 Fisher #C10445) at 24 hpi following manufactu rer’s protocols. Imaging was performed on a Zeiss 644 CellDiscoverer 7 microscope using two fluorescence channels to detect fluorescence from reduced 645 (Ex/Em: 592/614 nm) and oxidised (Ex/Em: 495/519 nm) BODIPY 581/591 C11. The mean intensity of 646 the reduced and oxidised fluorescent reporters was quantified using a CellProfiler pipeline , with lipid 647 peroxidation calculated based on the ratio of reduced:oxidised signals for each condition. 648 649 Statistics 650 Statistical analyses were performed using GraphPad Prism 9 and 10. In bar-dot plots, dots represent 651 individual replicates, and the bar height indicates the median. Statistical significance was determined 652 using either a one-way ANOVA with Dunnett’s multiple comparisons test or, for scratch wound assays, 653 a two-way ANOVA with Tukey’s multiple comparisons test. An adjusted p-value < 0.05 was considered 654 significant. Unless otherwise stated in the figure legends, all experiments were performed with a 655 minimum of three independent replicates. 656 657 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 22 SUPPLEMENTAL INFORMATION 658 Document S1. Figures S1-S4, Tables S2-S6 659 Table S1. Stress Response Gene Set. Related to Figure 1. Excel Spreadsheet. 660 Video S1. E. faecalis infection and tunicamycin treatment impair HaCaT cell migration. Related 661 to Figure 2E. 662 Video S2. IRE1 inhibition alters HaCaT cell migration in uninfected and E. faecalis -infected 663 conditions. Related to Figure 2G. 664 Video S3. The ΔEET mutant does not impair HaCaT cell migration. Related to Figure 3E. 665 Video S4. The effect of IRE1 inhibition on HaCaT cell migration during infection with WT or ΔEET 666 E. faecalis. Related to Figure 3F. 667 Video S5. Catalase rescues H2O2-induced migration defects in uninfected HaCaT cells. Related 668 to Figure 4K. 669 Video S6. Catalase restores migration in E. faecalis-infected HaCaT cells. Related to Figure 4L. 670 671 FIGURES 672 673 Figure 1. E. faecalis infection activates the UPR in a mouse model 674 (A) Uniform manifold approximation and projection (UMAP) of ~23,000 single-cell transcriptomes from 675 uninfected and E. faecalis-infected wounds22, recoloured here into the six broad cell classes used for 676 downstream stress-response analyses. (B) Per-cell enrichment score for a curated unfolded protein 677 response (UPR) gene set projected onto the UMAP in ( A). (C) Same UPR enrichment as in ( B) but 678 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 23 displayed only for keratinocyte clusters. (D) Enrichment of an oxidative stress response (OSR) gene 679 set plotted for fibroblast clusters. (E) Schematic representation of the UPR in mice, where spliced Xbp1 680 (Xbp1s), Chop, and Herpud1 serve as downstream markers of the IRE1, PERK, and ATF6 pathways, 681 respectively. (F) Gene expression of UPR markers (Xbp1s, Chop, and Herpud1) at 6 dpi for uninfected 682 and E. faecalis -infected 6 –7 week-old C57BL/6J mouse skin wounds, normalized to intact skin . 683 Significance was determined using one-way ANOVA, Dunnett’s test (unwounded skin, n = 16; wounded, 684 uninfected, n = 20; wounded, infected n = 33; ***p < 0 .001, ****p < 0.0001). 685 686 Figure 2. IRE1 activation by E. faecalis impedes keratinocyte migration in vitro 687 (A-B) Gene expression of UPR markers (Xbp1s, Chop, and Herpud1) in infected with E. faecalis at MOI 688 of 800 (A) or 600 (B) or tunicamycin (Tm) treated NIH-3T3 mouse fibroblasts (A) or HaCaT human 689 keratinocytes (B) (n = 3). See Methods for MOI optimization description. (C) Gene expression of IRE1 690 downstream gene ( EDEM1) in cells treated as in (A) ( n = 3). (D) Quantitative analysis and 691 representative immunoblots showing levels of XBP1 s and BiP in HaCaT cells treated as in (A). (E) 692 Scratch wound assay quantification for uninfected, infected and Tm-treated cells (positive control) (n = 693 4 biological replicates). (F) Gene expression of XBP1s from HaCaT cells treated with the DMSO control 694 (open circles) or the IRE1 inhibitor (IRE1i) 4µ8c (closed circles) (n = 3, one-way ANOVA, Dunnett’s test) 695 (G) Scratch wound assay quantification for uninfected and infected cells treated with 0.5% DMSO 696 control (open circles) or IRE1i (close circles) (n = 4 biological replicates). Significance was determined 697 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 24 using one-way ANOVA Dunnett’s test (A, B, and F) or two-way ANOVA Tukey’s test (E and G) (*p < 698 0.05, **p < 0.01, ***p < 0 .001, ****p < 0.0001). 699 700 Figure 3. Extracellular electron transport drives E. faecalis UPR activation and migration arrest 701 (A) E. faecalis OG1RF transposon screen in which a library of 14976 mutants was screened against a 702 NIH-3T3 cell line expressing the Xbp1 -mApple reporter system (3T3R ). (B) Representative 703 epifluorescnece microscopy images of 3T3R under different conditions (uninfected, WT, and Tm-704 treated) at 21 hpi. Scale bar, 100 μ m. (C) Diagram showing the pathways in which a subset of UPR 705 defective mutants (genes/proteins with red font) were identified. (D) Validation of UPR defective mutants 706 with 3T3R (n = 3). (E-F) Scratch wound assay quantification for HaCaT infected with (E) WT and ΔEET, 707 which were also (F) treated with either 0.5% DMSO control (open circles) or IRE1i (close circles). 708 Uninfected and WT data are identical to Figure 2E and 2F and replicated here for ease of comparison 709 (n = 4 biological replicates). (G) UPR induction at 24 hpi in HaCaT after treatment with 0.5% DMSO 710 control (open circles) or IRE1i (closed circles) under uninfected, WT -infected, and ΔEET -infected 711 conditions. Uninfected and WT-infected findings are identical to Figure 2F and replicated here for ease 712 of comparison (n = 3). Significance was determined using one-way ANOVA Dunnett’s test (D and G) or 713 two-way ANOVA Tukey’s test (E and F) (***p < 0 .001, ****p < 0.0001). 714 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 25 715 Figure 4. EET-derived ROS is sufficient to activate the UPR, disrupting epithelial migration 716 (A) Diagram showing the dismutation of superoxide radical (O2•–) into hydrogen peroxide (H2O2) via the 717 catalytic activity of superoxide dismutase (SOD), or spontaneously via a pH-dependent process. H2O2 718 can be converted into a highly reactive hydroxyl radical in the presence of transition metals like ferrous 719 ions (Fe2+) via the Fenton reaction or neutralised into water and oxygen if catalase is present. (B) Dose-720 dependent H2O2 mediated UPR induction in 3T3R (n = 3). (C) In-frame deletion mutants corresponding 721 to the Tn mutants of the EET and DMK synthesis pathways which were identified as UPR defective hits 722 during the transposon screen. The ΔMEN mutant had its entire operon deleted, which contained the 723 menF/D/C/E/B genes (n = 3). (D-F) UPR induction in 3T3(R) for (D) uninfected (E) WT infection, and 724 (F) ΔEET infection, in the presence (closed circles) or absence (open circles) of H2O2, SOD, and 725 catalase (n = 4). (G-I) UPR induction in uninfected (G), WT infected (H), and ΔEET infected (I) HaCaT 726 cells in the presence (closed circles) or absence (open circles) of H 2O2 and catalase (n ≥ 3). (J) Lipid 727 oxidation in uninfected, WT infected, and ΔEET infected HaCaT cells using BODIPY 581/591 C11 (n = 728 3). (K-L) Scratch wound assay quantification in uninfected (K) or WT infected (L) HaCaT cells treated 729 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 26 with catalase and/or H2O2 (n = 3 biological replicates). Significance was determined using one -way 730 ANOVA Dunnett’s test (B-J) or two-way ordinary ANOVA, Tukey’s multiple comparisons test (K-L) (*p 731 < 0.05, **p < 0.01, ***p < 0 .001, ****p < 0.0001). 732 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 27

733 1. Chng, J.-H., L Chong, K.K., Ning Lam, L., Jie Wong, J., and Kline, K.A. Biofilm-associated infection 734 by enterococci. https://doi.org/10.1038/s41579-018-0107-z. 735 2. Fiore, E., Van Tyne, D., and Gilmore, M.S. Pathogenicity of Enterococci. Microbiol Spectr 7, 736 10.1128/microbiolspec.gpp3-0053–2018. https://doi.org/10.1128/microbiolspec.gpp3-0053-2018. 737 3. Chong, K.K.L., Tay, W.H., Janela, B., Yong, A.M.H., Liew, T.H., Madden, L., Keogh, D., Barkham, 738 T.M.S., Ginhoux, F., Becker, D.L., et al. (2017). Enterococcus faecalis Modulates Immune 739 Activation and Slows Healing During Wound Infection. The Journal of Infectious Diseases 216, 740 1644–1654. https://doi.org/10.1093/infdis/jix541. 741 4. Ramsey, M., Hartke, A., and Huycke, M. (2014). The Physiology and Metabolism of Enterococci. 742 In Enterococci: From Commensals to Leading Causes of Drug Resistant Infection, M. S. Gilmore, 743 D. B. Clewell, Y. Ike, and N. Shankar, eds. (Massachusetts Eye and Ear Infirmary). 744 5. Fun, X.H., and Thibault, G. (2020). Lipid bilayer stress and proteotoxic stress -induced unfolded 745 protein response deploy divergent transcriptional and non-transcriptional programmes. Biochimica 746 et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids 1865, 158449. 747 https://doi.org/10.1016/j.bbalip.2019.04.009. 748 6. Halbleib, K., Pesek, K., Covino, R., Hofbauer, H.F., Wunnicke, D., Hänelt, I., Hummer, G., and 749 Ernst, R. (2017). Activation of the unfolded protein response by lipid bilayer stress. Molecular cell 750 67, 673-684. e8. 751 7. Hetz, C., Zhang, K., and Kaufman, R.J. (2020). Mechanisms, regulation and functions of the 752 unfolded protein response. Nat Rev Mol Cell Biol 21, 421–438. https://doi.org/10.1038/s41580-020-753 0250-z. 754 8. Ho, N., Yap, W.S., Xu, J., Wu, H., Koh, J.H., Goh, W.W.B., George, B., Chong, S.C., Taubert, S., 755 and Thibault, G. (2020). Stress sensor Ire1 deploys a divergent transcriptional program in response 756 to lipid bilayer stress. Journal of Cell Biology 219, e201909165. 757 https://doi.org/10.1083/jcb.201909165. 758 9. Baruch, M., Belotserkovsky, I., Hertzog, B.B., Ravins, M., Dov, E., McIver, K.S., Le Breton, Y.S., 759 Zhou, Y., Cheng, C.Y., and Hanski, E. (2014). An Extracellular Bacterial Pathogen Modulates Host 760 Metabolism to Regulate Its Own Sensing and Proliferation. Cell 156, 97 –108. 761 https://doi.org/10.1016/j.cell.2013.12.007. 762 10. Smith, J.A., Khan, M., Magnani, D.D., Harms, J.S., Durward, M., Radhakrishnan, G.K., Liu, Y.-P., 763 and Splitter, G.A. (2013). Brucella Induces an Unfolded Protein Response via TcpB That Supports 764 Intracellular Replication in Macrophages. PLOS Pathogens 9, e1003785. 765 https://doi.org/10.1371/journal.ppat.1003785. 766 11. Celli, J., and Tsolis, R.M. (2015). Bacteria, the ER and the Unfolded Protein Response: Friends or 767 Foes? Nat Rev Microbiol 13, 71–82. https://doi.org/10.1038/nrmicro3393. 768 12. Keestra-Gounder, A.M., and Tsolis, R.M. (2017). NOD1 and NOD2: Beyond Peptidoglycan 769 Sensing. Trends in Immunology 38, 758–767. https://doi.org/10.1016/j.it.2017.07.004. 770 13. Bronner, D.N., Abuaita, B.H., Chen, X., Fitzgerald, K.A., Nuñez, G., He, Y., Yin, X. -M., and 771 O’Riordan, M.X.D. (2015). Endoplasmic Reticulum Stress Activates the Inflammasome via NLRP3- 772 and Caspase -2-Driven Mitochondrial Damage. Immunity 43, 451 –462. 773 https://doi.org/10.1016/j.immuni.2015.08.008. 774 14. Vajjala, A., Biswas, D., Tay, W.H., Hanski, E., and Kline, K.A. (2019). Streptolysin -induced 775 endoplasmic reticulum stress promotes group A Streptococcal host -associated biofilm formation 776 and necrotising fasciitis. Cellular Microbiology 21, e12956. https://doi.org/10.1111/cmi.12956. 777 15. Choi, J.-A., and Song, C. -H. (2019). Insights Into the Role of Endoplasmic Reticulum Stress in 778 Infectious Diseases. Front Immunol 10, 3147. https://doi.org/10.3389/fimmu.2019.03147. 779 16. Tsutsuki, H., Yahiro, K., Ogura, K., Ichimura, K., Iyoda, S., Ohnishi, M., Nagasawa, S., Seto, K., 780 Moss, J., and Noda, M. (2016). Subtilase cytotoxin produced by locus of enterocyte effacement -781 negative Shiga -toxigenic Escherichia coli induces stress granule formation. Cell Microbiol 18, 782 1024–1040. https://doi.org/10.1111/cmi.12565. 783 17. van ‘t Wout, E.F.A., van Schadewijk, A., van Boxtel, R., Dalton, L.E., Clarke, H.J., Tommassen, J., 784 Marciniak, S.J., and Hiemstra, P.S. (2015). Virulence Factors of Pseudomonas aeruginosa Induce 785 Both the Unfolded Protein and Integrated Stress Responses in Airway Epithelial Cells. PLoS 786 Pathog 11, e1004946. https://doi.org/10.1371/journal.ppat.1004946. 787 18. Huycke, M.M., Joyce, W., and Wack, M.F. (1996). Augmented Production of Extracellular 788 Superoxide by Blood Isolates of Enterococcus faecalis. Journal of Infectious Diseases 173, 743–789 745. https://doi.org/10.1093/infdis/173.3.743. 790 19. Huycke, M.M., Moore, D., Joyce, W., Wise, P., Shepard, L., Kotake, Y., and Gilmore, M.S. (2001). 791 Extracellular superoxide production by Enterococcus faecalis requires demethylmenaquinone and 792 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 28 is attenuated by functional terminal quinol oxidases. Molecular Microbiology 42, 729 –740. 793 https://doi.org/10.1046/j.1365-2958.2001.02638.x. 794 20. Pallepati, P., and Averill -Bates, D.A. (2011). Activation of ER stress and apoptosis by hydrogen 795 peroxide in HeLa cells: Protective role of mild heat preconditioning at 40 °C. Biochimica et 796 Biophysica Acta (BBA) - Molecular Cell Research 1813, 1987 –1999. 797 https://doi.org/10.1016/j.bbamcr.2011.07.021. 798 21. Huycke, M.M., Abrams, V., and Moore, D.R. (2002). Enterococcus faecalis produces extracellular 799 superoxide and hydrogen peroxide that damages colonic epithelial cell DNA. Carcinogenesis 23, 800 529–536. https://doi.org/10.1093/carcin/23.3.529. 801 22. Celik, C., Lee, S.T.T., Tanoto, F.R., Veleba, M., Kline, K., and Thibault, G. (2024). Decoding the 802 complexity of delayed wound healing following Enterococcus faecalis infection. eLife 13, RP95113. 803 https://doi.org/10.7554/eLife.95113. 804 23. Pierre, N., Barbé, C., Gilson, H., Deldicque, L., Raymackers, J. -M., and Francaux, M. (2014). 805 Activation of ER stress by hydrogen peroxide in C2C12 myotubes. Biochemical and Biophysical 806 Research Communications 450, 459–463. https://doi.org/10.1016/j.bbrc.2014.05.143. 807 24. Madrazo, N., Khattar, Z., Powers, E.T., Rosarda, J.D., and Wiseman, R.L. (2024). Mapping stress-808 responsive signaling pathways induced by mitochondrial proteostasis perturbations. MBoC 35, 809 ar74. https://doi.org/10.1091/mbc.E24-01-0041. 810 25. Wang, Z., Butler, P., Ly, D., Spiotto, M., Koong, A., and Yang, G. (2010). Activation of the Unfolded 811 Protein Response in Wound Healing. Journal of Surgical Research 158, 209. 812 https://doi.org/10.1016/j.jss.2009.11.111. 813 26. Bachar-Wikstrom, E., Manchanda, M., Bansal, R., Karlsson, M., Kelly-Pettersson, P., Sköldenberg, 814 O., and Wikstrom, J.D. (2021). Endoplasmic reticulum stress in human chronic wound healing: 815 Rescue by 4 -phenylbutyrate. International Wound Journal 18, 49 –61. 816 https://doi.org/10.1111/iwj.13525. 817 27. Qiang, L., Yang, S., Cui, Y. -H., and He, Y. -Y. Keratinocyte autophagy enables the activation of 818 keratinocytes and fibroblastsand facilitates wound healing. Autophagy 17, 2128 –2143. 819 https://doi.org/10.1080/15548627.2020.1816342. 820 28. Yoo, J., Mashalidis, E.H., Kuk, A.C.Y., Yamamoto, K., Kaeser, B., Ichikawa, S., and Lee, S. -Y. 821 (2018). GlcNAc-1-P-transferase–tunicamycin complex structure reveals basis for inhibition of N -822 glycosylation. Nat Struct Mol Biol 25, 217–224. https://doi.org/10.1038/s41594-018-0031-y. 823 29. Mori, K. (2009). Signalling pathways in the unfolded protein response: development from yeast to 824 mammals. J Biochem 146, 743–750. https://doi.org/10.1093/jb/mvp166. 825 30. Hetz, C. (2012). The unfolded protein response: controlling cell fate decisions under ER stress and 826 beyond. Nat Rev Mol Cell Biol 13, 89–102. https://doi.org/10.1038/nrm3270. 827 31. Calfon, M., Zeng, H., Urano, F., Till, J.H., Hubbard, S.R., Harding, H.P., Clark, S.G., and Ron, D. 828 (2002). IRE1 couples endoplasmic reticulum load to secretory capacity by processing the XBP -1 829 mRNA. Nature 415, 92–96. https://doi.org/10.1038/415092a. 830 32. Hosokawa, N., Wada, I., Hasegawa, K., Yorihuzi, T., Tremblay, L.O., Herscovics, A., and Nagata, 831 K. (2001). A novel ER α-mannosidase-like protein accelerates ER-associated degradation. EMBO 832 reports 2, 415–422. https://doi.org/10.1093/embo-reports/kve084. 833 33. Cross, B.C., Bond, P.J., Sadowski, P.G., Jha, B.K., Zak, J., Goodman, J.M., Silverman, R.H., 834 Neubert, T.A., Baxendale, I.R., and Ron, D. (2012). The molecular basis for selective inhibition of 835 unconventional mRNA splicing by an IRE1 -binding small molecule. Proceedings of the National 836 Academy of Sciences 109, E869–E878. 837 34. Demay, Y., Perochon, J., Szuplewski, S., Mignotte, B., and Gaumer, S. (2014). The PERK pathway 838 independently triggers apoptosis and a Rac1/Slpr/JNK/Dilp8 signaling favoring tissue homeostasis 839 in a chronic ER stress Drosophila model. Cell Death Dis 5, e1452. 840 https://doi.org/10.1038/cddis.2014.403. 841 35. Ong, G., Ragetli, R., Mnich, K., Doble, B.W., Kammouni, W., and Logue, S.E. (2024). IRE1 842 signaling increases PERK expression during chronic ER stress. Cell Death Dis 15, 276. 843 https://doi.org/10.1038/s41419-024-06663-0. 844 36. Baruch, M., Hertzog, B.B., Ravins, M., Anand, A., Catherine Youting, C., Biswas, D., Tirosh, B., 845 and Hanski, E. (2014). Induction of endoplasmic reticulum stress and unfolded protein response 846 constitutes a pathogenic strategy of group A streptococcus. Frontiers in Cellular and Infection 847 Microbiology 4. 848 37. Chatterjee, A., Willett, J.L.E., Dunny, G.M., and Duerkop, B.A. (2021). Phage infection and sub -849 lethal antibiotic exposure mediate Enterococcus faecalis type VII secretion system dependent 850 inhibition of bystander bacteria. PLOS Genetics 17, e1009204. 851 https://doi.org/10.1371/journal.pgen.1009204. 852 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 29 38. Hederstedt, L., Gorton, L., and Pankratova, G. (2020). Two Routes for Extracellular Electron 853 Transfer in Enterococcus faecalis. J Bacteriol 202, e00725-19. https://doi.org/10.1128/JB.00725-854 19. 855 39. Xue, X., Piao, J.-H., Nakajima, A., Sakon-Komazawa, S., Kojima, Y., Mori, K., Yagita, H., Okumura, 856 K., Harding, H., and Nakano, H. (2005). Tumor Necrosis Factor α (TNFα) Induces the Unfolded 857 Protein Response (UPR) in a Reactive Oxygen Species (ROS)-dependent Fashion, and the UPR 858 Counteracts ROS Accumulation by TNFα. Journal of Biological Chemistry 280, 33917 –33925. 859 https://doi.org/10.1074/jbc.M505818200. 860 40. Vladykovskaya, E., Sithu, S.D., Haberzettl, P., Wickramasinghe, N.S., Merchant, M.L., Hill, B.G., 861 McCracken, J., Agarwal, A., Dougherty, S., Gordon, S.A., et al. (2012). Lipid Peroxidation Product 862 4-Hydroxy-trans-2-nonenal Causes Endothelial Activation by Inducing Endoplasmic Reticulum 863 Stress *. Journal of Biological Chemistry 287, 11398 –11409. 864 https://doi.org/10.1074/jbc.M111.320416. 865 41. Light, S.H., Su, L., Rivera-Lugo, R., Cornejo, J.A., Louie, A., Iavarone, A.T., Ajo-Franklin, C.M., and 866 Portnoy, D.A. (2018). A flavin -based extracellular electron transfer mechanism in diverse Gram -867 positive bacteria. Nature 562, 140–144. https://doi.org/10.1038/s41586-018-0498-z. 868 42. Light, S.H., Méheust, R., Ferrell, J.L., Cho, J., Deng, D., Agostoni, M., Iavarone, A.T., Banfield, J.F., 869 D’Orazio, S.E.F., and Portnoy, D.A. (2019). Extracellular electron transfer powers flavinylated 870 extracellular reductases in Gram -positive bacteria. Proceedings of the National Academy of 871 Sciences 116, 26892–26899. https://doi.org/10.1073/pnas.1915678116. 872 43. Ueki, T. (2021). Cytochromes in Extracellular Electron Transfer in Geobacter. Applied and 873 Environmental Microbiology 87, e03109-20. https://doi.org/10.1128/AEM.03109-20. 874 44. Reguera, G., McCarthy, K.D., Mehta, T., Nicoll, J.S., Tuominen, M.T., and Lovley, D.R. (2005). 875 Extracellular electron transfer via microbial nanowires. Nature 435, 1098 –1101. 876 https://doi.org/10.1038/nature03661. 877 45. Keogh, D., Lam, L.N., Doyle, L.E., Matysik, A., Pavagadhi, S., Umashankar, S., Low, P.M., Dale, 878 J.L., Song, Y., Ng, S.P., et al. (2018). Extracellular Electron Transfer Powers Enterococcus faecalis 879 Biofilm Metabolism. mBio 9, e00626-17. https://doi.org/10.1128/mBio.00626-17. 880 46. Ho, F.K., Lam, L.N., Matysik, A., Watts, T.D., Wong, J.J., Chong, K.K.L., Choo, P.Y., Tolar, J., Low, 881 P.M., Chua, Z.S., et al. (2025). Role of sortase -assembled Ebp pili in Enterococcus faecalis 882 adhesion to iron oxides and its impact on extracellular electron transfer. Microbiol Spectr 13, 883 e02337-24. https://doi.org/10.1128/spectrum.02337-24. 884 47. Stevens, E.T., Van Beeck, W., Blackburn, B., Tejedor-Sanz, S., Rasmussen, A.R.M., Carter, M.E., 885 Mevers, E., Ajo -Franklin, C.M., and Marco, M.L. (2023). Lactiplantibacillus plantarum uses 886 ecologically relevant, exogenous quinones for extracellular electron transfer. mBio 14, e02234-23. 887 https://doi.org/10.1128/mbio.02234-23. 888 48. Cabello, A.-L., Wells, K., Peng, W., Feng, H. -Q., Wang, J., Meyer, D.F., Noroy, C., Zhao, E. -S., 889 Zhang, H., Li, X., et al. (2024). Brucella-driven host N-glycome remodeling controls infection. Cell 890 Host & Microbe 32, 588-605.e9. https://doi.org/10.1016/j.chom.2024.03.003. 891 49. Hernandez-Morfa, M., Olivero, N.B., Zappia, V.E., Piñas, G.E., Reinoso-Vizcaino, N.M., Cian, M.B., 892 Nuñez-Fernandez, M., Cortes, P.R., and Echenique, J. (2023). The oxidative stress response of 893 Streptococcus pneumoniae: its contribution to both extracellular and intracellular survival. Front 894 Microbiol 14, 1269843. https://doi.org/10.3389/fmicb.2023.1269843. 895 50. Barichello, T., Generoso, J.S., Simões, L.R., Elias, S.G., and Quevedo, J. (2013). Role of oxidative 896 stress in the pathophysiology of pneumococcal meningitis. Oxid Med Cell Longev 2013, 371465. 897 https://doi.org/10.1155/2013/371465. 898 51. Lau, G.W., Ran, H., Kong, F., Hassett, D.J., and Mavrodi, D. (2004). Pseudomonas aeruginosa 899 Pyocyanin Is Critical for Lung Infection in Mice. Infect Immun 72, 4275 –4278. 900 https://doi.org/10.1128/IAI.72.7.4275-4278.2004. 901 52. Sinha, S., Shen, X., Gallazzi, F., Li, Q., Zmijewski, J.W., Lancaster, J.R., and Gates, K.S. (2015). 902 Generation of Reactive Oxygen Species Mediated by 1 -Hydroxyphenazine, a Virulence Factor of 903 Pseudomonas aeruginosa. Chem Res Toxicol 28, 175–181. https://doi.org/10.1021/tx500259s. 904 53. Dixon, S.J., Lemberg, K.M., Lamprecht, M.R., Skouta, R., Zaitsev, E.M., Gleason, C.E., Patel, D.N., 905 Bauer, A.J., Cantley, A.M., Yang, W.S., et al. (2012). Ferroptosis: An Iron -Dependent Form of 906 Nonapoptotic Cell Death. Cell 149, 1060–1072. https://doi.org/10.1016/j.cell.2012.03.042. 907 54. Stockwell, B.R., Angeli, J.P.F., Bayir, H., Bush, A.I., Conrad, M., Dixon, S.J., Fulda, S., Gascón, S., 908 Hatzios, S.K., Kagan, V.E., et al. (2017). Ferroptosis: A Regulated Cell Death Nexus Linking 909 Metabolism, Redox Biology, and Disease. Cell 171, 273 –285. 910 https://doi.org/10.1016/j.cell.2017.09.021. 911 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 30 55. Amaral, E.P., Costa, D.L., Namasivayam, S., Riteau, N., Kamenyeva, O., Mittereder, L., Mayer -912 Barber, K.D., Andrade, B.B., and Sher, A. (2019). A major role for ferroptosis in Mycobacterium 913 tuberculosis–induced cell death and tissue necrosis. Journal of Experimental Medicine 216, 556–914 570. https://doi.org/10.1084/jem.20181776. 915 56. Dar, H.H., Tyurina, Y.Y., Mikulska-Ruminska, K., Shrivastava, I., Ting, H.-C., Tyurin, V.A., Krieger, 916 J., Croix, C.M.S., Watkins, S., Bayir, E., et al. (2019). Pseudomonas aeruginosa utilizes host 917 polyunsaturated phosphatidylethanolamines to trigger theft -ferroptosis in bronchial epithelium. J 918 Clin Invest 128, 4639–4653. https://doi.org/10.1172/JCI99490. 919 57. Qiang, L., Zhang, Y., Lei, Z., Lu, Z., Tan, S., Ge, P., Chai, Q., Zhao, M., Zhang, X., Li, B., et al. 920 (2023). A mycobacterial effector promotes ferroptosis-dependent pathogenicity and dissemination. 921 Nat Commun 14, 1430. https://doi.org/10.1038/s41467-023-37148-x. 922 58. Lee, Y.-S., Lee, D. -H., Choudry, H.A., Bartlett, D.L., and Lee, Y.J. (2018). Ferroptosis -Induced 923 Endoplasmic Reticulum Stress: Cross-talk between Ferroptosis and Apoptosis. Molecular Cancer 924 Research 16, 1073–1076. https://doi.org/10.1158/1541-7786.MCR-18-0055. 925 59. Zheng, Y., Sun, L., Guo, J., and Ma, J. (2023). The crosstalk between ferroptosis and anti -tumor 926 immunity in the tumor microenvironment: molecular mechanisms and therapeutic controversy. 927 Cancer Communications 43, 1071–1096. https://doi.org/10.1002/cac2.12487. 928 60. Xie, Y., Hou, W., Song, X., Yu, Y., Huang, J., Sun, X., Kang, R., and Tang, D. (2016). Ferroptosis: 929 process and function. Cell Death Differ 23, 369–379. https://doi.org/10.1038/cdd.2015.158. 930 61. Matsuzaki, S., Hiratsuka, T., Taniguchi, M., Shingaki, K., Kubo, T., Kiya, K., Fujiwara, T., 931 Kanazawa, S., Kanematsu, R., Maeda, T., et al. (2015). Physiological ER Stress Mediates the 932 Differentiation of Fibroblasts. PLoS One 10, e0123578. 933 https://doi.org/10.1371/journal.pone.0123578. 934 62. Yu, Q., Zhao, B., Gui, J., Katlinski, K.V., Brice, A., Gao, Y., Li, C., Kushner, J.A., Koumenis, C., 935 Diehl, J.A., et al. (2015). Type I interferons mediate pancreatic toxicities of PERK inhibition. Proc 936 Natl Acad Sci U S A 112, 15420–15425. https://doi.org/10.1073/pnas.1516362112. 937 63. R: The R Project for Statistical Computing https://www.r-project.org/. 938 64. Hao, Y., Hao, S., Andersen-Nissen, E., Mauck, W.M., Zheng, S., Butler, A., Lee, M.J., Wilk, A.J., 939 Darby, C., Zager, M., et al. (2021). Integrated analysis of multimodal single -cell data. Cell 184, 940 3573-3587.e29. https://doi.org/10.1016/j.cell.2021.04.048. 941 65. Butler, A., Hoffman, P., Smibert, P., Papalexi, E., and Satija, R. (2018). Integrating single -cell 942 transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 36, 411–943 420. https://doi.org/10.1038/nbt.4096. 944 66. Satija, R., Farrell, J.A., Gennert, D., Schier, A.F., and Regev, A. (2015). Spatial reconstruction of 945 single-cell gene expression data. Nat Biotechnol 33, 495–502. https://doi.org/10.1038/nbt.3192. 946 67. Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W.M., Hao, Y., Stoeckius, 947 M., Smibert, P., and Satija, R. (2019). Comprehensive Integration of Single -Cell Data. Cell 177, 948 1888-1902.e21. https://doi.org/10.1016/j.cell.2019.05.031. 949 68. Wong, B. (2011). Points of view: Color blindness. Nature Methods 8, 441 –441. 950 https://doi.org/10.1038/nmeth.1618. 951 69. Korsunsky, I., Nathan, A., Millard, N., and Raychaudhuri, S. (2019). Presto scales Wilcoxon and 952 auROC analyses to millions of observations. Preprint at bioRxiv, https://doi.org/10.1101/653253 953 https://doi.org/10.1101/653253. 954 70. Subramanian, A., Tamayo, P., Mootha, V.K., Mukherjee, S., Ebert, B.L., Gillette, M.A., Paulovich, 955 A., Pomeroy, S.L., Golub, T.R., Lander, E.S., et al. (2005). Gene set enrichment analysis: A 956 knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the 957 National Academy of Sciences 102, 15545–15550. https://doi.org/10.1073/pnas.0506580102. 958 71. Korotkevich, G., Sukhov, V., Budin, N., Shpak, B., Artyomov, M.N., and Sergushichev, A. (2021). 959 Fast gene set enrichment analysis. Preprint at bioRxiv, https://doi.org/10.1101/060012 960 https://doi.org/10.1101/060012. 961 72. Benjamini, Y., and Hochberg, Y. (1995). Controlling the False Discovery Rate: A Practical and 962 Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B 963 (Methodological) 57, 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x. 964 73. Livak, K.J., and Schmittgen, T.D. (2001). Analysis of Relative Gene Expression Data Using Real-965 Time Quantitative PCR and the 2−ΔΔCT Method. Methods 25, 402 –408. 966 https://doi.org/10.1006/meth.2001.1262. 967 74. da Silva, R.A.G., Tay, W.H., Ho, F.K., Tanoto, F.R., Chong, K.K.L., Choo, P.Y., Ludwig, A., and 968 Kline, K.A. (2022). Enterococcus faecalis alters endo-lysosomal trafficking to replicate and persist 969 within mammalian cells. PLoS Pathog 18, e1010434. https://doi.org/10.1371/journal.ppat.1010434. 970 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint 31 75. Liang, C. -C., Park, A.Y., and Guan, J. -L. (2007). In vitro scratch assay: a convenient and 971 inexpensive method for analysis of cell migration in vitro. Nat Protoc 2, 329 –333. 972 https://doi.org/10.1038/nprot.2007.30. 973 76. Ariffin, N.S. (2023). The CellProfiler pipeline analysis of cell migration. Acta Histochemica 125, 974 152074. https://doi.org/10.1016/j.acthis.2023.152074. 975 77. Suarez-Arnedo, A., Torres Figueroa, F., Clavijo, C., Arbeláez, P., Cruz, J.C., and Muñoz-Camargo, 976 C. (2020). An image J plugin for the high throughput image analysis of in vitro scratch wound 977 healing assays. PLoS One 15, e0232565. https://doi.org/10.1371/journal.pone.0232565. 978 78. Kristich, C., vy, N., Le, T., Barnes, A., Grindle, S., and Dunny, G. (2008). Development and Use of 979 an Efficient System for Random mariner Transposon Mutagenesis To Identify Novel Genetic 980 Determinants of Biofilm Formation in the Core Enterococcus faecalis Genome. Applied and 981 environmental microbiology 74, 3377–3386. https://doi.org/10.1128/AEM.02665-07. 982 79. Nielsen, H.V., Guiton, P.S., Kline, K.A., Port, G.C., Pinkner, J.S., Neiers, F., Normark, S., 983 Henriques-Normark, B., Caparon, M.G., and Hultgren, S.J. (2012). The Metal Ion -Dependent 984 Adhesion Site Motif of the Enterococcus faecalis EbpA Pilin Mediates Pilus Function in Catheter -985 Associated Urinary Tract Infection. mBio 3, 10.1128/mbio.00177 -12. 986 https://doi.org/10.1128/mbio.00177-12. 987 988 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted August 8, 2025. ; https://doi.org/10.1101/2025.08.07.669218doi: bioRxiv preprint