Time-resolved transcriptomic mapping reveals conserved stress programs and metabolic rewiring in Escherichia coli under antimicrobial photodynamic and blue light exposure

23 Antimicrobial photodynamic inactivation (aPDI) and antimicrobial blue light (aBL) are 24 emerging, resistance -agnostic strategies for controlling bacterial pathogens, yet their 25 systems-level impact on prokaryotic physiology remains incompletely understood. Here, we 26 used time -resolved g lobal transcriptomics to define how Escherichia coli reprograms gene 27 expression in response to diverse photodynamic stresses. E. coli K-12 BW25113 was 28 exposed to five phototreatments differing in photosensitizer chemistry and light wavelength, 29 including rose bengal, TMPyP , new methylene blue, aBL alone, and aBL combined with 5 -30 aminolevulinic acid, and transcriptional responses were profiled after short (30 min) and 31 prolonged (7 –8 h) exposure. Short -term phototreatments triggered rapid and extensive 32 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint transcriptional remodeling, affecting up to ~58% of the genes and dominated by conserved 33 stress programs including oxidative defense, sulfur metabolism, and a broad downshift in 34 biosynthesis and energy generation. In contrast, prolonged exposure elicited more rest rained 35 but highly treatment -specific adaptive responses, characterized by suppression of core 36 energy metabolism, including oxidative phosphorylation and the tricarboxylic acid cycle, 37 coupled with activation of alternative catabolic pathways. Together, thes e findings reveal a 38 common acute stress architecture across photodynamic modalities followed by divergent 39 long-term adaptive trajectories, providing a systems -level framework for understanding 40 bacterial responses to light -based antimicrobials and informing the rational optimization of 41 photodynamic therapies. 42 43

44 The growing crisis of antimicrobial resistance (AMR) has emerged as one of the most 45 pressing global health challenges, threatening not only human and animal health but also 46 food production systems and environmental sustainability. In response, the World He alth 47 Organization (WHO) has called for a coordinated, multisectoral strategy to mitigate AMR 48 under the One Health initiative, which emphasizes the cooperation of human, animal, and 49 environmental health sectors [1]. 50 Antimicrobial photodynamic inactivation (aPDI) and antimicrobial blue light (aBL) align with 51 this initiative, offering environmentally friendly, non -antibiotic approaches capable of 52 eradicating pathogens regardless of their antimicrobial susceptibility profile , without inducing 53 resistance [2], [3], [4] . The principal mechanism involves a light, non-toxic photosensitizer 54 (PS) and molecular oxygen to induce an oxidative burst that non -selectively kills bacteria. 55 While both methods share this general mode of action, they differ primarily in the o rigin of the 56 photosensitizers involved. In aPDI, exogenous photosensitizing compounds are introduced 57 and activated by specific light wavelengths matching their absorption spectra [5]. In contrast, 58 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint aBL relies on the excitation of endogenous porphyrins naturally present within bacterial cells, 59 which absorb blue light in the 400 –470 nm range [6]. Following illumination, both exogenous 60 and endogenous PSs reach a long -lived triplet state and can react by three proposed 61 mechanisms. In the Type I mechanism, electron transfer leads to the formation of superoxide 62 radicals ( 02 -• ), hydrogen peroxide (H 202), and hydroxyl radicals (HO•). The Type II 63 mechanism results in energy transfer directly to molecular oxygen , yielding highly reactive 64 singlet molecular oxygen (¹O₂ ) [7]. These ROS (reactive oxygen species) can cause 65 oxidative damage to bacterial membranes, nucleic acids, and essential biomolecules, 66 ultimately leading to bacterial cell death [8], [9]. Additionally, a Type III mechanism has been 67 proposed, in which light -activated photosensitizers exert bacteric idal effects through direct 68 interaction with cellular targets, independently of oxygen availability [10]. 69 Despite their proven antimicrobial efficacy, both aPDI and aBL remain underutilized in clinical 70 practice, largely due to the lack of standardized pr otocols and insufficient understanding of 71 bacterial response mechanisms . The photodynamic treatment strategies can vary widely in 72 photosensitizers, light sources, and treatment durations ; this heterogeneity complicates 73 clinical translation . To date, only a few studies have addressed global transcriptomic or 74 regulatory responses following aPDI or aBL treatment [11], [12], [13], [14] ; these have 75 typically been limited to a single photosensitizer or s pecific time point, without systematic 76 comparison across multiple treatments or exposure durations. 77 To comprehensively elucidate the molecular responses of Escherichia coli to light -based 78 antimicrobial strategies, we designed a phototreatment panel that covered chemical and 79 mechanistic diversity. We therefore s elected five complementary photodynamic approaches 80 applied under both short -term (30 min) and prolonged (7 –8 h) exposure conditions. This 81 combination enabled us to capture both acute stress responses and longer -term adaptive 82 trajectories across fundamentally different modes of photodynamic action. 83 Three of the selected treatments employed exogenous photosensitizers representing distinct 84 and well -characterized chemical classes with different photochemical properties. TMPyP 85 [5,10,15,20-tetrakis(1-methyl-4-pyridinium)porphyrin tetra -(p-toluenesulfonate)] is a tetra -86 cationic porphyrin derivative that acts predominantly via a Type II photodynamic mechanism 87 through high -efficiency singlet oxygen generation [7]. Rose Bengal, a halogenated 88 fluorescein derivative with strong absorption in the green spectral range (480 –550 nm), also 89 primarily drives Type II photochemistry but differs markedly from TMPyP in molecular 90 structure and charge distribution [7]. In contras t, new methylene blue (NMB), a 91 phenothiazinium dye absorbing in the red region, operates mainly via a Type I mechanism 92 involving electron transfer and radical formation [15]. Together, these three photosensitizers 93 span the major classes of clinically and experimentally relevant photodynamic agents and 94 encompass both dominant photochemical reaction pathways. To complement these 95 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint exogenous approaches, we included two strategies based on endogenous bacterial 96 chromophores. Antimicrobial blue light (aBL) exploi ts naturally occurring intracellular 97 porphyrins, providing a photosensitizer -free modality that is increasingly explored for clinical 98 and environmental applications. In parallel, we combined aBL with 5 -aminolevulinic acid 99 (ALA) to enhance intracellular acc umulation of endogenously produced porphyrins, thereby 100 intensifying photodynamic effects while preserving the same light -based activation principle 101 [16]. 102 To our knowledge, this is the first study to systematically characterize the transcriptomic 103 response of E. coli to photodynamic strategies. This study provides a comprehensive, time-104 resolved transcriptomic framework for understanding how E. coli responds to different 105 photodynamic antimicrobial treatments that rely on both exogenous and endogenous 106 photosensitization pathways. By comparing short - and long -term exposures across 107 mechanistically distinct phototreatments, we identify common stress resp onses as well as 108 treatment-specific adaptive changes that influence bacterial survival under photodynamic 109 stress. These findings improve our fundamental understanding of light -based antimicrobial 110 action and support the rational development and standardizat ion of aPDI and aBL for 111 translational and clinical applications 112 113

114 Strain and culture conditions 115 E. coli BW25113 Keio collection parent strain was used [17]. Bacteria were cultured in M9 116 minimal medium supplemented with 0.4% glucose, 0.2 mM MgSO 4, and 0.1 mM CaCl 2 or 117 seeded on M9 agar plates with the same supplements as in liquid medium and 1.5% agar. 118 For overnight culture , a single colony was inoculated into 5 ml of M9 or LB medium and 119 incubated at 37°C for 18 hours under aerobic conditions with shaking (150 rpm). Log -phase 120 culture was prepared by diluting the overnight culture 1:25 into 10 mL of fresh supplemented 121 M9 medium and incubating for approximately 5.5 h under the abovementioned conditions. 122 Light sources 123 Three custom-made light-emitting diode (LED) lamps were used in this study with emission 124 maxima at 415 nm and a radiosity of 25 mW/cm 2 (Cezos, Poland), λmax 522 nm an d a 125 radiosity of 10.6 mW/cm2 (Cezos, Poland), and λmax 633 nm and a radiosity of 138.3 mW/cm2 126 (lab255, Poland). A 415 nm lamp was used for aBL and aBL with ALA treatment, while a 522 127 nm lamp was used for RB, and a 633 nm lamp was used for TMPyP and NMB treatment. 128 Chemical compounds 129 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint All compounds were purchased from Merck (Sigma -Aldrich, Darmstadt, Germany ). Stock 130 solutions were prepared in sterile water and stored in the dark at 4°C - 1 mM for 131 photosensitizers and 5 mg/ml for ALA. Working solutions were fre shly prepared by diluting 132 the stock solutions to the desired concentrations. 133 Determination of the sub -lethal conditions of photodynamic treatments for short - and long -134 term duration 135 For the short-term treatment, overnight cultures were grown in LB medium fo r the aBL group, 136 and in supplemented M9 medium for aPDI and aBL + ALA groups. Then , cultures were 137 diluted 1:25 into 10 mL of fresh, supplemented M9 medium, and 5 mL aliquots were 138 transferred into a 6 -well plate. At this point, for aBL -ALA treatment, cultur es were 139 supplemented with the desired concentration of ALA. The plate was then placed in a 140 ThermoMixer® C (Eppendorf, Germany) at 37 °C for approximately 5.5 h with continuous 141 shaking and heating. After incubation, the OD 600 was measured, and cultures were adjusted 142 to an optical density of 0.3 in M9 medium, and aliquots of 900 μl were transferred to a 24-well 143 plate. For aBL/aBL + ALA samples, cultures were immediately subjected to illumination, while 144 for aPDI samples, the PS was added to the bacterial suspe nsions at the desired 145 concentration, followed by a 15 -minute incubation in the dark at 37 °C prior to illumination. 146 After irradiation, 10 μl aliquots were serially diluted tenfold in PBS to generate dilutions of 147 10−1 to 10 −5 and streaked horizontally onto M9 agar plates. Plates were incubated at 37 °C 148 for 16 –20 h, after which colony -forming units (CFUs) were counted to determine survival 149 rates. Control groups included cells that were not treated with PSs or light. Each experiment 150 was performed in triplicate. 151 For the long -term treatment, overnight cultures were grown in different media depending on 152 the treatment group: in LB medium for the aBL group, in supplemented M9 medium with the 153 addition of ALA for the aBL+ALA group, and in supplemented M9 medium for th e aPDI 154 groups. Then cultures were similarly diluted 1:25 into 10 mL of fresh supplemented M9 155 medium, and 900 μl aliquots were transferred to a 24 -well plate. For the aBL and aBL + ALA 156 groups, the plate was immediately placed in the ThermoMixer® C and subje cted to 157 simultaneous incubation and illumination. In the aPDI groups, photosensitizers were added 158 at the desired concentrations, followed by 15 minutes of dark incubation at 37 °C, after which 159 the samples were illuminated and incubated simultaneously. Opti cal density (OD 600) was 160 measured every hour using a microplate reader (Envision, PerkinElmer) to monitor bacterial 161 growth for 12 h. Control groups included cells that were not treated with PSs or light. Each 162 experiment was performed in triplicate. The detailed experimental setup is presented in 163 Figure 1. 164 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint 165 Figure 1. Experimental setup for photodynamic treatment. 166 RNA extraction and sequencing 167 Total RNA was isolated from E. coli cells after illumination using the RNeasy ® Mini Kit 168 (Qiagen, Germany). For samples exposed to short -term photodynamic treatment, an 169 additional 40 -minute incubation at 37 °C post -irradiation was included to allow for cellular 170 recovery. RNA concentration was measured using a NanoDrop One spectrophot ometer 171 (Thermo Scientific, USA), and samples were stored at −80 °C until further processing. RNA 172 samples were sent to Lexogen GmbH (Austria), where quality control, genomic DNA 173 removal, rRNA depletion (RiboCop), and library preparation were performed using the 174 CORALL Total RNA -Seq Library Prep Kit. Sequencing was conducted on an Illumina 175 Novaseq X platform in paired-end 150 bp read mode, yielding an average of 10 million reads 176 per sample. 177 Transcriptomic data analysis 178 The RNA-seq reads were subjected to qual ity control using FastQC v0.12.1. Subsequently, 179 quality filtering and adapter trimming were performed with BBDUK2 from BBMAP package 180 v36.14, using the following parameters: qtrim = w, trimq = 20, maq = 10, forcetrimleft=20, k = 181 23, mink = 8, hdist = 1, tbo , tpe, minlength = 100, removeifeitherbad = t. High -quality reads 182 were then mapped to the complete sequenced genome of the reference strain K -12 183 (ENSEMBL ASM584v2) using Bowtie 2. The filtered FASTQ files were used as input for read 184 counting using RSEM v1.2.30. Differential expression analysis was performed using DESeq2 185 v1.34.0, and genes with a > 1.5 -fold change in expression and an adjusted P -value < 0.05 186 were considered as differentially expressed (DE). Further, functional analysis of enrichment 187 in Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) 188 pathways was conducted with clusterProfiler, KEGGREST, and org.EcK12.eg.db libraries in 189 Bioconductor/R. 190 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint Specifically, enrichment was assessed across Biological Process (BP), Molecular Function 191 (MF), and Cellular Component (CC) domains using Benjamini-Hochberg (BH) adjusted p -192 values and q -values with a cutoff of 0.05. To ensure accurate database cross -referencing, 193 gene symbols were converted to Entrez IDs using the bitr utility. Significantly enriched terms 194 were visualized as dot plots (clusterProfiler), prioritizing the top 25 categories for each 195 functional domain. 196 Regulon enrichment analysis was performed to identify transcription factors (TFs) 197 significantly associated with diff erentially expressed genes (DEGs) using regulatory 198 interaction data from RegulonDB. DEGs were filtered based on an adjusted p -value 1. Enrichment was calculated using a Fisher ’s exact test, with p -200 values corrected for multipl e testing using the Benjamini -Hochberg procedure. The analysis 201 was restricted to TFs with at least 5 target genes in the background set. 202

203 Optimization of sub -lethal photodynamic treatment parameters for E. coli transcriptomic 204 profiling 205 To obtain biologically meaningful transcriptomic data, it was essential to first establish sub -206 lethal photodynamic conditions that elicit cellular stress responses without causing extensive 207 cell death or lysis. Careful optimization of these parameters ensures that o bserved gene 208 expression changes reflect active bacterial adaptation rather than nonspecific damage or 209 loss of viable cells. This step is therefore critical for accurately interpreting the molecular 210 mechanisms underlying bacterial responses to photodynamic treatments. 211 For short -term treatments, sub -lethal doses were defined as those that resulted in an 212 approximately 1 log 10 reduction in CFU/ml in the mid -log phase of growth during 30 minutes 213 of exposure. The duration of the treatment was selected based, amon g others, on previous 214 study reports that 15 -30 minutes is sufficient for activation of stable gene expression, 215 including initiation of SOS response, and enables monitoring of rapid bacterial response. 216 Moreover, the threshold of 1 log 10 CFU/ml in viability allows for an accurate mechanistic 217 analysis without being biased by irreversible cell damage or lysis. In terms of long -term 218 treatments, sub-lethal doses were defined as continuous exposure until cultures reached the 219 mid-exponential phase of growth, corres ponding to approximately 50% inhibition of growth 220 (based on OD 600 values) compared to untreated controls. Control cultures were grown for 6 221 h, while treated cultures required 7 -8 h to reach the same phase. The duration of treatment 222 was chosen based on obse rved bacterial growth dynamics, ensuring that cells reached the 223 log phase, which is optimal for high RNA yield from metabolically active cells. Additionally, the 224 extended exposure of low-dose aPDI/aBL allows for the assessment of long-term or adaptive 225 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint responses that may develop over several bacterial generations. The estimated sub -lethal 226 conditions for the abovementioned treatments are presented in Table 1. 227 228 Table 1. Experimental conditions for short- and long-term photodynamic treatments. 229 Treatment Short-term Long-term RB + 522 nm light 5 µM + 14.3 J/cm2 5 µM +61.1 J/cm2 TMPyP + 633 nm light 2 µM + 131.9 J/cm2 0.2 µM + 610.3 J/cm2 NMB + 633 nm light 3 µM + 131.9 J/cm2 1 µM + 610.3 J/cm2 aBL 415 nm 42.8 J/cm2 91.8 J/cm2 aBL 415 nm + ALA 22.5 J/cm2 + 1.5 µg/ml 107.1 J/cm2 + 4 µg/ml 230 RB - rose bengal, NMB – new methylene blue, aBL – antimicrobial blue light, ALA – 5-aminolevulenic acid 231 232 Exposure duration determines the strength and specificity of E. coli transcriptomic responses 233 To obtain a comprehensive picture of how E. coli responds at the transcriptomic level to 234 various photodynamic protocols, RNA -seq sequencing was performed. The detailed 235 experimental setup is presented in Figure 1. RNA -seq data were obtained for all conditio ns 236 with high coverage and mapping quality, and detailed sequencing statistics are provided in 237 the Supplementary Materials Table S1 and Figure S1, S2. 238 Principal component analysis (PCA) of log 2-transformed gene expression values from control 239 and treated samples (short exposure conditions, Figure 2A) revealed that the five treatment 240 groups formed distinct clusters, with biological replicates tending to group within those 241 clusters, suggesting a strong transcriptional response to those conditions . Notably, t he 242 treatment clusters also showed a directional tendency away from the control. In the case of 243 long treatment (Figure 2B), the differentiation between control and treated samples is less 244 pronounced, but we observed a strong separation of aBL -treated sample s from the other 245 groups. This suggests that prolonged photodynamic exposure exerts a weaker effect on 246 gene expression, except for blue light. As a complement to the PCA analysis, a hierarchical 247 clustering heatmap of the top 100 most variable genes (selecte d based on variance of mean 248 expression values across three biological replicates per group) revealed a clear separation 249 between control and treated samples following short exposures (Figure 2C), with distinct 250 gene expression profiles induced by treatment. However, samples did not cluster according 251 to the type of treatment (aBL/aBL+ALA vs. aPDI). In long -treated samples (Figure 2D), 252 clustering indicated a less robust transcriptome shift for extended exposure of aBL+ALA and 253 aPDI. Notably, cells subjected to b lue light showed the most distinct transcriptional profile 254 among all samples. E. coli cells exposed to short -duration aBL or aPDI treatments exhibited 255 extensive transcriptomic change (Figure 2E, Table 2), with an average of 2, 105 differentially 256 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint expressed g enes (padj < 0.05; fold change ≥ 1.5), accounting for 49.1% of all actively 257 transcribed genes. The strongest response was observed for TMPyP combined with red light 258 (57.9%), while RB activated by green light elicited the weakest effect (34.9%). In contrast , for 259 all long treatments, we observed markedly reduced response in the number and proportion 260 of DEGs, with an average of only 16.9% of the transcriptome being differentially expressed. 261 The weakest response was marked in the extended TMPyP treatment, affec ting just 6.6% of 262 the transcriptome. It is worth noting that despite the variability in DEG numbers across 263 conditions, the ratio of upregulated to downregulated genes remained relatively balanced in 264 most treatments. 265 Table 2. Differential gene expression profiles under phototreatments. 266 Treatment No. of upregulated genes No. of downregulated genes No. of differentially expressed genes (% of transcriptome) aBL [S] 1210 1223 2433 (56.7%) aBL [L] 415 407 822 (19.2%) aBL + ALA [S] 1130 1075 2205 (51.4%) aBL + ALA [L] 332 269 601 (13.9%) RB [S] 829 667 1496 (34.9%) RB [L] 543 494 1037 (24.1%) NMB [S] 1021 948 1969 (46.1%) NMB [L] 193 92 285 (6.6%) TMPyP [S] 1253 1173 2426 (57.9%) TMPyP [L] 540 349 889 (20.7%) S- short treatment, L-long treatment 267 % of transcriptome was calculated as % number of DEGS relative to number of active genes in the transcriptome 268 (baseMean > 0) 269 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint 270 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint Figure 2. Transcriptional response of E. coli BW25113 to ten photodynamic treatments (short 271 and long exposures – RB + green light, TMPyP and NMB + red light, aBL and aBL + ALA. (A–B) 272 Principal component analysis (PCA) of normalized gene expression data for short (A) and long (B) 273 treatment conditions. Each point represents a biological replicate. (C-D) Heatmaps of the top 100 most 274 variable genes across experimental groups, based on variance of mean expression values (averaged 275 across three biological replicates per group). Gene -wise expression values we re standardized (z -276 scores), and both genes and groups were hierarchically clustered using Euclidean distance, across 277 short- (C) and long -term (D) treatment, respectively. (E) Differentially expressed genes (DEGs) per 278 treatment. Bars show the total number o f DEGs ( padj < 0.05; fold change ≥ 1.5), with red and blue 279 indicating upregulated and downregulated genes, respectively. Results shown are based on at least 280 three independent biological replicates. 281 282 Short-term photodynamic stress triggers rapid defense pat hways, while prolonged exposure 283 drives metabolic adaptation 284 To investigate the functional impact of various photodynamic treatments on E. coli cells, gene 285 ontology (GO) enrichment analysis and KEGG pathway analyses were performed on 286 differentially expressed genes (DEGs) identified for all short - and long -term conditions 287 (Figures 3–4). 288 GO analysis revealed that short -term exposures predominantly affected genes involved in 289 the biosynthesis of core macromolecular precursors, including amino acids, nucleotide 290 bases, and intermediary metabolites (Figure 3A). Notably, processes linked with response to 291 temperature stimulus, heat, and abiotic stress wer e significantly overrepresented for most of 292 the treatments, indicating rapid activation of general stress response pathways. Prolonged 293 exposures instead showed significant overrepresentation of pathways associated with 294 metabolic remodeling across nearly al l experimental conditions (Figure 3B). Observed 295 enrichment in GO terms related to organic acid metabolism, energy derivation by oxidation of 296 organic compounds, and generation of precursor metabolites and energy suggests a 297 pronounced shift in cellular metab olism across all conditions as a prolonged adaptation to 298 stress. The comparative analysis of short - vs long -term responses suggests that bacterial 299 cells after temporary treatment prioritize immediate survival, while prolonged exposure leads 300 to an altered metabolic state. 301 302 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint 303 Figure 3. Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEGs) in 304 E. coli BW25113 following short - and long- term photodynamic treatments. (A) GO terms 305 significantly enriched among DEGs after short -term exposures (30 min) to sub -lethal photodynamic 306 stress. (B) GO terms significantly enriched among DEGs after long -term exposures (8 -9 h). Dot size 307 represents the number of DEGs annotated to the term, and color indicates the adjusted p-value (padj). 308

shown are based on at least three independent biological replicates. 309 While GO enrichment analysis was performed on all differentially expressed genes to capture 310 the overall biological processes affected by phototreatments, KEGG pathway enrichment 311 was conducted separately for up - (Figure 4A and 4C) and down -regulated (Figure 4B and 312 4D) genes to preserve the directionality of transcriptional responses at the pathway level. In 313 short-term treatments (Figure 4A), up-regulated genes showed consistent overrepresentation 314 of pathways related to exopolysaccharide biosynthesis, sulfur metabolism, and two -315 component regulatory systems. These pathways ar e associated with canonical stress 316 response, including biofilm formation and environmental sensing, as well as sulfur 317 metabolism can represent involvement in oxidative stress and detoxification. Notably, RB -318 mediated aPDI was associated with the broadest fu nctional diversity among upregulated 319 pathways, including unique activation of fatty acid degradation, lysine degradation, and 320 methane, nitrogen, and propanoate metabolism, suggesting induction of alternative energy 321 and detoxification routes. In contrast, N MB-mediated aPDI and aBL combined with ALA 322 showed enrichment of fewer distinct pathways, which may reflect a more focused or limited 323 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint transcriptional rearrangement. Downregulated genes in short -term exposures (Figure 4B) 324 exhibited consistent enrichment in A BC transporters, biosynthesis of cofactors, and 325 secondary metabolites biosynthesis across all five treatment conditions. Additionally, most 326 phototreatments showed suppression of amino acid biosynthesis, carbon fixation, and 327 oxidative phosphorylation. These findings suggest that photooxidative stress triggers in E. 328 coli cells global metabolic downshift, likely reflecting a transition from growth to survival state. 329 Interestingly, RB -mediated aPDI, which exhibited the strongest transcriptional activation, 330 showed the lowest number of down -regulated pathways, whereas the remaining treatments 331 exhibited a wider range of suppressed processes. 332 Long-term exposure to the five photodynamic conditions resulted in a distinct transcriptional 333 profile compared to short -term responses. Among up -regulated genes (Figure 4C), lysine 334 degradation was the most consistently enriched pathway, indicating a shift toward amino acid 335 catabolism. Additionally, secondary metabolite biosynthesis and microbial metabolism in 336 diverse environments were enriched in 3 out of 5 conditions, reflecting enhanced metabolic 337 flexibility and environmental adaptation. The blue light irradiation with or without ALA 338 exhibited the smallest set of enriched pathways, while all aPDI treatments triggered more 339 extensive gene activation, reflected by a broader range of enriched pathways. Down -340 regulated genes under long -term treatment (Figure 4D) showed a pattern partially similar to 341 the short -term response . The biosynthesis of amino acids was consistently downregulat ed 342 across all conditions, indicating a suppression of growth -related anabolic processes. 343 Moreover, we observed frequent downregulation of central metabolic pathways, including the 344 TCA cycle, oxidative phosphorylation, nitrogen cycle, and two -component syst ems, 345 suggesting a deep metabolic reprogramming characterized by reduced respiration, nitrogen 346 turnover, and signal transduction activity. All prolonged phototreatments, except NMB -347 mediated aPDI, manifested a similar level of pathway enrichment. 348 Notably, a consistent pattern between GO and KEGG pathway enrichment analyses was 349 observed. Both approaches revealed that short -term treatments were associated with the 350 activation of stress -related processes such as response to heat and response to abiotic 351 stimulus (Figure 3A and Figure S3 A), as well as sulfur metabolism, two -component system, 352 and exopolysaccharide biosynthesis (Figure 4A). These responses were accompanied by the 353 suppression of biosynthe tic and energy -consuming processes , including nucleoside 354 phosphate metabolism, and small molecule biosynthesis (Figure S3B), as well as amino acid 355 biosynthesis, carbon metabolism, and ABC transporters (Figure 4B). In long-term treatments, 356 both analyses supported a shift toward metabolic reprogramming and utilization of alternative 357 energy sources. For instance, GO terms related to ribosome biogenesis, translation, and 358 respiration were significantly downregulated (Figure S3D) , aligning with the suppression of 359 KEGG pathways such as ribosome, oxidative phosphorylation, and the TCA cycle (Figure 360 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint 4D), while upregulated genes showed enrichment in lysine degradation and fatty acid 361 metabolism (Figure 4C). 362 363 Figure 4. KEGG pathway enrichment of differentially expressed genes in E. coli BW25113 364 following photodynamic treatments. Pathway enrichment was analyzed separately for upregulated 365 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint (A, C) and downregulated (B, D) genes. (A–B) Short-term treatments. (C–D) Long-term treatments. 366

shown are based on at least three independent biological replicates. 367 368 Distinct regulatory programs underlie short- and long-term responses to photodynamic stress 369 To further explore the regulatory mechanisms underlying the observed transcriptional 370 changes, we performed transcription factor regulon enrichment analysis across all short- and 371 long-term photodynamic treatments (Figure 5). Under short exposure conditions, regulon 372 enrichment was predominantly focused on regulators associated with acid stress, metal 373 homeostasis, and envelope integrity. Across most treatments, GadX, GadW, and RcsAB 374 emerged as the most consistently enriched regulons. In parallel, multiple regulators 375 associated with envelope and membrane stress responses, most notably RcsAB, CpxR, and 376 BaeR, were detected. Short -term exposure also elicited strong enrichment of global 377 metabolic and redox regulators, including Fur and Cbl. Notably, relatively narrow regulon 378 engagement was observed, with minimal involvement of global transcriptional regulators, 379 suggesting that early responses primarily rely on specialized stress-response modules. 380 In contrast, long -term treatment induced enrichment of global regulators associated with 381 central metabolism, anaerobic respiration, and nitrogen utilization. Across multiple 382 treatments, ArcA, IHF, FNR, and NarL emerged as dominant regulons, indicating a shift 383 toward systemic transcriptional reprogramming. The diversity of regulatory responses under 384 prolonged exposure was exemplified by treatment -specific signals, including the enrichment 385 of OmpR and FhlA under NMB conditions, and of Fur under aBL conditions. 386 Together, these results demonstrate that exposure duration is a primary determinant of 387 transcriptional regulatory architecture. Short -term responses rely on specialized stress 388 regulons, whereas long -term adaptation involves extensive engagement of global metabolic 389 and respiratory regulators. Functional classification of enriched regulons is provided in Table 390 S2. 391 392 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint 393 Figure 5. Enrichment of transcription factor regulons under short- and long-term photodynamic 394 treatments. Dot size represents DEG count, and color indicates –log10 adjusted P-value. 395 396 Short-term convergence and long-term divergence in photodynamic responses 397 To evaluate the overlap and divergence in transcriptional responses to the five treatments, 398 we constructed Venn diagrams of differentially expressed genes for short -term (Figure 6A) 399 and long-term (Figure 6B) exposures. Short -term treatments revealed a substantial common 400 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint response, with 891 shared DEGs across all conditions , indicating a robust common core 401 response. Among treatment -specific profiles, aBL induced the largest set (403 genes), 402 followed by TMPyP (197), aBL+ALA (99), NMB (40), and RB (35). In contrast, long -term 403 exposures showed markedly reduced overlap, with only 46 shared DEGs. Unique responses 404 were more pronounced, particularly for aBL (350), RB (262), and TMPyP (159), whereas 405 aBL+ALA (103) and NMB (15) exhib ited smaller exclusive sets. These patterns suggest that 406 the extensive overlap in shorter treatments reflects conserved stress -responsive 407 mechanisms, while the divergence during prolonged treatment highlights distinct adaptive 408 responses to each phototreatment. 409 410 411 412 Figure 6. Venn diagrams showing the overlap of differentially expressed genes (DEGs) across 413 five photo treatment conditions under (A) short -term and (B) long -term exposure. DEGs were 414 defined as those with an adjusted pvalue (padj) < 0.05 and a fold change ≥ 1.5. Results are based on 415 a minimum three independent biological replicates. 416 417

418 Specifically, t he aim of this study was to characterize the transcriptome response of 419 Escherichia coli to photodynamic inactivation during both short-term and long-term exposure. 420 Analysis of these different treatment conditions is crucial, as microorganisms exhibit different 421 transcriptional responses depending on the duration of the stress factor. The use of a single 422 time point can therefore lead to incomplete or distorted conclusions [18]. 423 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint Sublethal treatment conditions have been carefully defined to enable accurate and correct 424 interpretation of transcriptome data and were consistent with other transcriptome studies on 425 bacteria exposed to sublethal photodynamic or chemical stress [11], [18] . Ultimately, the 426 RNA-seq data obtained confirmed the validity of the experimental conditions, as short -term 427 treatment elicited a broad stress response in the microorganism affecting approximately 35 –428 58% of the t ranscriptome, which is characteristic of acute stress exposure, while long -term 429 exposure triggered a weaker but distinct adaptive response, including metabolic 430 restructuring. Such results would not have been possible if the treatment conditions had 431 represented lethal exposure. The above arguments demonstrate that the sublethal conditions 432 were correctly identified. 433 A global transcriptome analysis showed that E. coli undergoes an extensive and exposure -434 duration-dependent response to photodynamic stress. Short -term exposure triggered 435 transcriptional changes affecting up to 58% of the genes, while long-term exposure triggered 436 a smaller number of transcriptional changes, but with a more pronounced adaptive character. 437 This transition from an acute to an adaptive r esponse reflects observations described for 438 other bacteria exposed to oxidative or photooxidative stress [13], [14], [19]. 439 The comparative transcriptome analysis revealed both common and treatment -specific 440 responses of the bacteria to photodynamic inacti vation (aPDI) and antibacterial blue light 441 (aBL). Short-term exposure triggered a relatively similar stress program under all conditions 442 studied, while long -term exposure revealed different adaptation strategies. Short-term 443 treatment showed a preserved str ess program characterized by the activation of sulfur 444 metabolism, exopolysaccharide biosynthesis, and two -component regulatory systems, 445 accompanied by an inhibition of anabolic processes such as amino acid biosynthesis and 446 oxidative phosphorylation (Fig ure 4AB). These changes indicate a rapid slowing of 447 metabolism and the activation of defense mechanisms, which is consistent with previous 448 studies on the response to blue light -induced stress conditions in Staphylococcus aureus 449 [12], E. coli [20], Campylobacter jejuni [14] and Acinetobacter baumannii [21]. In contrast, 450 longer exposure showed a more diverse response with unique transcriptional signatures for 451 each treatment (Fig ure 6B). In particular , long -term exposure to aBL and RB led to 452 transcriptomic changes in the largest group of treatment -specific genes, suggesting strong 453 adaptive diversification. This likely reflects the utilization of universal defense mechanisms 454 and the emergence of specialized survival strategies. 455 One of the most striking observations is the transcriptional response triggered by aBL. 456 Transcriptome analysis showed that the response of E. coli to aBL varies considerably 457 depending on the duration of exposure. Short -term exposure leads to a strong 458 downregulation of genes for oxidative phosphorylation but has no effect on amino acid 459 biosynthesis (Figure 4B), while long -term exposure has the exact opposite effect, i.e., a 460 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint strong suppression of amino acid biosynthesis without significant changes in oxidativ e 461 phosphorylation (Figure 4D). These opposing effects suggest a time -dependent metabolic 462 reprogramming under the influence of photodynamically induced stress. During short -term 463 exposure to aBL, light -induced endogenous chromophores (e.g., flavins, porphyri ns) 464 generate reactive oxygen species (ROS) and trigger an acute response to oxidative stress. 465 The observed inhibition of oxidative phosphorylation likely reflects a protective mechanism 466 that limits ROS production from the electron transport chain, while bi osynthetic pathways 467 remain active to support protein synthesis and repair. However, long -term exposure to aBL 468 leads to an accumulation of oxidative damage that impairs biosynthesis capacity. A 469 substantial reduction in amino acid biosynthesis suggests a tra nsition to a state of metabolic 470 inactivity or dormancy, in which microorganisms conserve resources and prioritize basic 471 functions. Stable expression of oxidative phosphorylation genes during this phase may 472 represent an adaptation to maintain minimal energy production. Together, these results 473 confirm a two -phase adaptation model: an early phase of redox defense characterized by 474 temporary inhibition of respiration, followed by a late phase of resource conservation 475 characterized by inhibition of anabolic metab olism and stabilization of energy pathways. A 476 comparable light -dependent disruption of electron transport was also observed in C. jejuni 477 exposed to violet -blue light, with inactivation of enzymes containing iron -sulfur clusters and 478 loss of heme from cytoch romes, leading to disruption of energy metabolism [14]. Similarly, 479 under the influence of blue light, S. aureus showed transcriptional changes that led to a 480 reduction in the activity of growth pathways [12], [13]. 481 Another unique response was observed during photodynamic inactivation with TMPyP. In this 482 case, activation of sulfur metabolism was observed (Fig ure 4C). The upregulation of sulfur 483 pathways likely reflects an increased need for thiol -based redox buffering, consistent with the 484 role of glutathione and cysteine metabolism in detoxifying reactive oxygen species [19]. 485 These different transcriptomic response patterns underscore that while all treatments studied 486 exert photooxidative pressure, the specific metabolic adaptations depend on the 487 photosensitizer used and the corresponding wavelength of light. 488 Exposure of microorganisms to various stress factors ha s shown that, despite stress factor -489 specific transcriptome responses, there is a transcription response common to the various 490 factors tha t involves the upregulation of genes involved in stress protection, repair, and 491 detoxification, as well as the downregulation of genes associated with rapid growth, protein 492 synthesis, and energy -intensive biosynthesis pathways [22]. Heat shock is one of the best-493 characterized environmental stressors at the transcriptome level. In bacteria such as E. coli, 494 a sudden increase in temperature rapidly induces the alternative sigma factor σ32, which 495 activates the transcription of classic heat shock genes, includ ing dnaK, groEL, groES, and 496 clpB [23]. These encode chaperone molecules and proteases required for the prevention of 497 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint protein aggregation and the refolding of denatured proteins. Transcriptomic analyses also 498 show repression of ribosomal protein genes and translation-related genes, which, similar to 499 our findings, reflects the transition from growth metabolism to adaptive metabolism [22]. 500 Next, oxidative stress caused by reactive oxygen species (ROS), also triggers a significant 501 transcriptional response. I n E. coli , this leads to the activation of the OxyR and SoxRS 502 regulons, resulting in strong induction of antioxidant and redox -balancing genes, including 503 katG (catalase), ahpC (alkyl peroxide reductase), sodA, and sodB (superoxide dismutase) 504 [24]. Another stress agent, i.e., exposure to ultraviolet (UV) radiation and other DNA -505 damaging factors triggers transcriptome activation of DNA repair and tolerance pathways, 506 which are mainly dependent on the SOS response [25], [26] . At the same time, genes 507 associated with cell cycle progression and replication are suppressed, reflecting a 508 transcriptional strategy that prioritizes genome stability over proliferation [22]. All these are in 509 line with our analyses regarding transcriptomic response over time , revealing that the 510 responses of microorganisms are highly dynamic and characterized by rapid induction of 511 early regulatory and signaling genes, followed by a slower reorganization of metabolism. 512 Understanding which regulatory networks are activated under different photodynamic 513 exposure regimes is essential for deciphering how bacteria sense and ultimately survive 514 light-induced stress. During short -term photodynamic exposure, enrichment analysis mainly 515 identified stress -response regulons linked to acid stre ss (GadW, GadX), envelope integrity 516 (RcsAB, CpxR, BaeR, PspF), and metal or redox homeostasis (Fur, Cbl) (Figure 5) . This 517 regulatory pattern indicates that short -term photodynamic stress primarily activates envelope 518 stress response pathways, which aligns w ith membrane damage being an early and 519 prominent target of photodynamic treatment [8], [27], [28], [29] . Additionally, Fur -controlled 520 regulation likely plays a role in early control of intracellular iron levels, helping to limit Fenton 521 chemistry under ROS-generating conditions [30]. Notably, standard oxidative stress regulons 522 like OxyR and SoxRS were not significantly enriched, probably because induction of classic 523 antioxidant genes occurs early (within 10 minutes) and wanes over time [19], while our RN A 524 samples were taken after about 30 minutes of treatment, possibly after the peak of 525 OxyR/SoxRS-driven transcription. 526 In contrast, long -term exposure preferentially engaged global transcriptional regulators, 527 including ArcA, FNR, IHF, and NarL, indicating a transition from localized stress mitigation 528 toward systemic metabolic and respiratory reprogramming. These regulators coordinate 529 redox balance, anaerobic respiration, nitrogen utilization, and large -scale transcriptional 530 organization under sustained stress conditions [31], [32], [33]. 531 Overall, our data demonstrate that diverse photodynamic treatments elicit a shared acute 532 stress architecture in E. coli, whereas prolonged exposure drives largely treatment -specific 533 adaptive trajectories. The clear divergence between short - and long -term transcriptional 534 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint responses highlights the importance of time -resolved analyses for separating immediate 535 molecular damage from longer -term adaptive remodeling under photooxidative stress. 536 Together, these findings revea l the remarkable transcriptional plasticity of bacteria in 537 response to photodynamic challenge and provide a foundation for the rational refinement of 538 light-based antimicrobial therapies. 539 540 Authors’ contribution 541 NB provided data and performed data analysis a nd interpretation. MWS supported data 542 analysis and interpretation. MWS and MG contributed study set -up, evaluated and 543 interpreted the data. NB and MG wrote the manuscript. All authors read and approved the 544 final manuscript. 545 Competing interests 546 The authors have declared no competing interests. 547 Availability of data 548 The data generated or analyzed during this study are included in this published article and in 549 the supplemental material. RNA -seq data are deposited in the NCBI Gene Expression 550 Omnibus (GEO) database under accession number GSE317397. 551 552

553 [1] WHO Bacterial Priority Pathogens List , 2024. 2024. 554 [2] R. Youf et al., “Antimicrobial photodynamic therapy: Latest developments with a focus 555 on combinatory strategies,” Pharmaceutics, vol. 13, no. 12, pp. 1–56, 2021, doi: 556 10.3390/pharmaceutics13121995. 557 [3] A. Rapacka-Zdonczyk, A. Wozniak, M. Pieranski, A. Woziwodzka, K. P. Bielawski, and 558 M. Grinholc, “Development of Staphylococcus aureus tolerance to antimicrobial 559 photodynamic inactivation and antimicrobial blue light upon sub-lethal treatment,” Sci. 560 Rep., vol. 9, no. 1, pp. 1–18, 2019, doi: 10.1038/s41598-019-45962-x. 561 [4] N. Kashef and M. R. Hamblin, “Can microbial cells develop resistance to oxidative 562 stress in antimicrobial photodynamic inactivation?,” Drug Resistance Updates, vol. 31, 563 pp. 31–42, Mar. 2017, doi: 10.1016/J.DRUP .2017.07.003. 564 [5] M. Lan, S. Zhao, W. Liu, C. S. Lee, W. Zhang, and P. Wang, “Photosensitizers for 565 Photodynamic Therapy,” Adv. Healthc. Mater., vol. 8, no. 13, pp. 1–37, 2019, doi: 566 10.1002/adhm.201900132. 567 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint [6] Y. Wang et al., “Antimicrobial blue light inactivation of pathogenic microbes: State of 568 the art,” Drug Resistance Updates, vol. 33–35, pp. 1–22, Nov. 2017, doi: 569 10.1016/j.drup.2017.10.002. 570 [7] F. Cieplik et al., “Antimicrobial photodynamic therapy–what we know and what we 571 don’t,” Crit. Rev. Microbiol., vol. 44, no. 5, pp. 571–589, 2018, doi: 572 10.1080/1040841X.2018.1467876. 573 [8] E. Alves, M. A. F. Faustino, M. G. P. M. S. Neves, A. Cunha, J. Tome, and A. Almeida, 574 “An insight on bacterial cellular targets of photodynamic inactivation,” Future Med. 575 Chem., vol. 6, no. 2, pp. 141–164, 2014, doi: 10.4155/fmc.13.211. 576 [9] C. dos Anjos et al., “New Insights into the Bacterial Targets of Antimicrobial Blue 577 Light,” Microbiol. Spectr., vol. 11, no. 2, Apr. 2023, doi: 10.1128/spectrum.02833-22. 578 [10] M. R. Hamblin and H. Abrahamse, “Oxygen-independent antimicrobial 579 photoinactivation: Type III photochemical mechanism?,” Antibiotics, vol. 9, no. 2, pp. 580 1–17, 2020, doi: 10.3390/antibiotics9020053. 581 [11] D. Muehler et al., “Stress response in Escherichia coli following sublethal phenalene-582 1-one mediated antimicrobial photodynamic therapy: an RNA-Seq study,” 583 Photochemical and Photobiological Sciences, vol. 23, no. 8, pp. 1573–1586, Aug. 584 2024, doi: 10.1007/S43630-024-00617-3/FIGURES/6. 585 [12] T. L. Adair and B. E. Drum, “RNA-Seq reveals changes in the Staphylococcus aureus 586 transcriptome following blue light illumination,” Genom. Data, vol. 9, pp. 4–6, Sep. 587 2016, doi: 10.1016/j.gdata.2016.05.011. 588 [13] S. B. Snell, A. L. Gill, C. G. Haidaris, T. H. Foster, T. M. Baran, and S. R. Gill, 589 “Staphylococcus aureus Tolerance and Genomic Response to Photodynamic 590 Inactivation,” mSphere, vol. 6, no. 1, Feb. 2021, doi: 10.1128/MSPHERE.00762-591 20/SUPPL_FILE/MSPHERE.00762-20_ST003.XLSX. 592 [14] P. Walker et al., “Exploiting Violet-Blue Light to Kill Campylobacter jejuni : Analysis of 593 Global Responses, Modeling of Transcription Factor Activities, and Identification of 594 Protein Targets,” mSystems, vol. 7, no. 4, Aug. 2022, doi: 10.1128/msystems.00454-595 22. 596 [15] F. Ronzani et al., “Comparison of the photophysical properties of three phenothiazine 597 derivatives: Transient detection and singlet oxygen production,” Photochemical and 598 Photobiological Sciences, vol. 12, no. 12, pp. 2160–2169, 2013, doi: 599 10.1039/c3pp50246e. 600 [16] F. Harris and L. Pierpoint, “Photodynamic therapy based on 5‐ aminolevulinic acid and 601 its use as an antimicrobial Agent,” Med. Res. Rev., vol. 32, no. 6, pp. 1292–1327, Nov. 602 2012, doi: 10.1002/med.20251. 603 [17] T. Baba et al., “Construction of Escherichia coli K-12 in-frame, single-gene knockout 604 mutants: The Keio collection,” Mol. Syst. Biol., vol. 2, 2006, doi: 10.1038/msb4100050. 605 [18] V. Zorraquino, M. Kim, N. Rai, and I. Tagkopoulos, “The genetic and transcriptional 606 basis of short and long term adaptation across multiple stresses in Escherichia coli,” 607 Mol. Biol. Evol., vol. 34, no. 3, pp. 707–717, 2017, doi: 10.1093/molbev/msw269. 608 [19] M. Roth et al., “Transcriptomic Analysis of E. coli after Exposure to a Sublethal 609 Concentration of Hydrogen Peroxide Revealed a Coordinated Up‐ Regulation of the 610 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint Cysteine Biosynthesis Pathway,” Antioxidants, vol. 11, no. 4, 2022, doi: 611 10.3390/antiox11040655. 612 [20] S. A. Khan, M. J. Kim, and H. G. Yuk, “Genome-wide transcriptional response of 613 Escherichia coli O157:H7 to light-emitting diodes with various wavelengths,” Scientific 614 Reports 2023 13:1, vol. 13, no. 1, pp. 1–12, Feb. 2023, doi: 10.1038/s41598-023-615 28458-7. 616 [21] G. L. Müller et al., “Light Modulates Metabolic Pathways and Other Novel 617 Physiological Traits in the Human Pathogen Acinetobacter baumannii.,” J. Bacteriol., 618 vol. 199, no. 10, pp. 1–17, May 2017, doi: 10.1128/JB.00011-17. 619 [22] S. Jozefczuk et al., “Metabolomic and transcriptomic stress response of Escherichia 620 coli,” Mol. Syst. Biol., vol. 6, no. 364, pp. 1–16, 2010, doi: 10.1038/msb.2010.18. 621 [23] G. Nonaka, M. Blankschien, C. Herman, C. A. Gross, and V. A. Rhodius, “Reveals a 622 Multifaceted Cellular Response To Heat Stress,” Genes Dev., vol. 20, pp. 1776–1789, 623 2006, doi: 10.1101/gad.1428206.but. 624 [24] M. Zheng, X. Wang, L. J. Templeton, D. R. Smulski, R. A. LaRossa, and G. Storz, 625 “DNA microarray-mediated transcriptional profiling of the Escherichia coli response to 626 hydrogen peroxide,” J. Bacteriol., vol. 183, no. 15, pp. 4562–4570, 2001, doi: 627 10.1128/JB.183.15.4562-4570.2001. 628 [25] J. Courcelle, A. Khodursky, B. Peter, P. O. Brown, and P. C. Hanawalt, “Comparative 629 gene expression profiles following UV exposure in wild-type and SOS-deficient 630 Escherichia coli,” Genetics, vol. 158, no. 1, p. 41, 2001, doi: 631 10.1093/GENETICS/158.1.41. 632 [26] C. Janion, “Inducible SOS response system of DNA repair and mutagenesis in 633 Escherichia coli,” Int. J. Biol. Sci., vol. 4, no. 6, pp. 338–344, 2008, doi: 634 10.7150/ijbs.4.338. 635 [27] S. Bury-Moné et al., “Global analysis of extracytoplasmic stress signaling in 636 Escherichia coli,” PLoS Genet., vol. 5, no. 9, 2009, doi: 637 10.1371/journal.pgen.1000651. 638 [28] T. L. Raivio, S. K. D. Leblanc, and N. L. Price, “The Escherichia coli Cpx envelope 639 stress response regulates genes of diverse function that impact antibiotic resistance 640 and membrane integrity,” J. Bacteriol., vol. 195, no. 12, pp. 2755–2767, 2013, doi: 641 10.1128/JB.00105-13. 642 [29] G. Jovanovic, L. J. Lloyd, M. P. H. Stumpf, A. J. Mayhew, and M. Buck, “Induction and 643 function of the phage shock protein extracytoplasmic stress response in Escherichia 644 coli,” Journal of Biological Chemistry, vol. 281, no. 30, pp. 21147–21161, 2006, doi: 645 10.1074/jbc.M602323200. 646 [30] S. W. Seo, D. Kim, H. Latif, E. J. O’Brien, R. Szubin, and B. O. Palsson, “Deciphering 647 Fur transcriptional regulatory network highlights its complex role beyond iron 648 metabolism in Escherichia coli,” Nature Communications 2014 5:1, vol. 5, no. 1, pp. 649 4910-, Sep. 2014, doi: 10.1038/ncomms5910. 650 [31] S. Iuchi and E. C. C. Lin, “Adaptation of Escherichia coli to redox environments by 651 gene expression,” Mol. Microbiol., vol. 9, no. 1, pp. 9–15, Jul. 1993, doi: 652 10.1111/J.1365-2958.1993.TB01664.X;WGROUP:STRING:PUBLICATION. 653 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint [32] D. A. Tolla and M. A. Savageau, “Regulation of Aerobic-to-Anaerobic Transitions by the 654 FNR Cycle in Escherichia coli,” J. Mol. Biol., vol. 397, no. 4, p. 893, 2010, doi: 655 10.1016/J.JMB.2010.02.015. 656 [33] C. E. Noriega, H. Y. Lin, L. L. Chen, S. B. Williams, and V. Stewart, “Asymmetric cross-657 regulation between the nitrate-responsive NarX-NarL and NarQ-NarP two-component 658 regulatory systems from Escherichia coli K-12,” Mol. Microbiol., vol. 75, no. 2, pp. 659 394–412, Jan. 2010, doi: 10.1111/J.1365-660 2958.2009.06987.X;PAGE:STRING:ARTICLE/CHAPTER. 661 662 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.02.703343doi: bioRxiv preprint