Developmental priming increases copper-tolerance in a model fish species via epigenetic-and microbiome-mediated mechanisms

13 Pollution is a significant t hreat to aquatic ecosystems globally and , in order to survive, natural 14 populations depend upon their ability to rapidly develop tolerance to chemical stressors. We 15 examined whether early-life priming enhances life-long copper-tolerance in a model fish species 16 via developmental plasticity. Stickleback (Gasterosteus aculeatus) embryos were pre -exposed 17 to a low concentration of copper (10 µg/L) during early development , reared in clean water for 18 nine months alongside a control group, and then exposed to copper (0,10 and 20 µg/L ) for 96 h 19 as adults . Priming markedly reduced evidence of copper-toxicity in adult gills at the 20 transcriptional level (including reduced cellular stress response (CSR) and disruption of ion -21 homeostasis) and increased inducibility of the metal-binding protein, metallothionein. In 22 parallel, we identified epigenetic and microbiome-mediated mechanisms likely contributing to 23 this tolerance. Pre -exposure induced persistent DNA methylation changes, consistent with 24 priming of CSR and ion-homeostasis pathways. We identified enhanced copper-tolerance in the 25 gill microbiota of primed fish that likely also contributed to host tolerance. These findings provide 26 critical evidence for developmental plasticity induced by chemical stressors in animals, highlight 27 the importance of integrated microbiome and epigenetic responses, and enhance our 28 understanding of how natural populations cope with pollution in their environment. 29 Key words: phenotypic plasticity, conditioning, toxic metal, RNA-Seq, sensitivity, early -life 30 reprogramming, stress priming 31 .CC-BY 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 It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint

32 Aquatic ecosystems are threatened by unprecedented anthropogenic challenges, including 33 chemical pollution. To better understand, predict, and ultimately to mitigate, the impacts of 34 pollution, it is important to consider the relative sensitivity of different species and populations, 35 including their ability to develop tolerance following chemical exposure. Although there is some 36 evidence of local adaptation to pollution in microorganisms, plants and metazoa following long-37 term exposure1–4, for many species survival will ultimately depend on an ability to rapidly acquire 38 tolerance to acute and fluctuating stressors5,6. 39 Phenotypic plasticity allows organisms to rapidly adjust to environmental challenges within their 40 lifetime and encompasses a range of physiological, morphological, and behavioural 41 adjustments5,7,8. Acclimation is usually rapidly induced and reversible, with changes diminishing 42 after stressor -removal5,9,10. In contrast, developmental plasticity occurs when environmental 43 conditions experienced during critical early -development windows induce persistent, often 44 irreversible, changes in phenotyp e10–12. For plants, there is good evidence that pre -exposure to 45 chemicals and other abiotic stressors during early life, or ‘priming’, can induce persistent 46 tolerance via stress memory 13–15. For animals, while temperature -induced developmental 47 plasticity has been documented in fish and other ectotherms 7,10,16,17, it is unclear whether this 48 phenomenon can similarly enhance tolerance to environmental pollutants. Only a handful of 49 studies have examined whether chemical exposure in early development alters subsequent 50 sensitivity18–21. This knowledge is critical to understand the sensitivity of natural populations 51 experiencing fluctuating levels of pollution, a common feature of many natural environments. 52 Epigenetic mechanisms, regulating differences in gene expression, have been shown to underly 53 developmental plasticity across diverse systems 8,22–25. I n plants, multiple epigenetic 54 mechanisms, including chromatin remodelling, DNA methylation and ncRNAs, are known to 55 facilitate primed stress memor ies, conferring tolerance to environmental stressors, including 56 toxic metals26,27. Similarly, for animals, epigenetic mechanisms contribute to increased thermal 57 tolerance following early life exposure17,25,28,29. Regarding chemical stressors in animals, research 58 has so far focused only on epigenetic toxicity. Many classes of chemical pollutant s, including 59 metals, pesticides and endocrine disruptors , are known to induce epigenetic modifications 60 associated with adverse health outcomes, dependent on the timing and nature of exposure30–33. 61 However, the potential for epigenetic mechanisms to contribute to enhanced chemical 62 tolerance in animals remains unexplored. 63 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint Host-associated microbiota also play a critical role in influencing sensitivity to environmental 64 stressors by extending host adaptive phenotypic capacity 34,35. While microbiomes are often 65 sensitive to disruption by environmental stressors, they have an extensive capacity to rapidly 66 develop tolerance34,36,37. Crucially, microbial adaptive plasticity can enhance host tolerance to 67 environmental challenges38,39. Tolerant microbiota can, for example, limit adverse effects in the 68 host associated with microbiome disruption, and/or confer specific benefits, such as enhanced 69 metabolism or sequestration of toxins 36,40,41. As with the epigenome, the microbiome has 70 heightened environmental sensitivity during early development 42. While microbiome priming is 71 emerging as a technique to enhance agricultural productivity and stressor resilience 41, its 72 potential role in environmental chemical-tolerance is largely unexplored. 73 It is unknown whether, in animals, tolerance to environmental chemicals can be acquired via 74 developmental plasticity, the extent to which this occurs, or the specific mechanisms 75 contributing to this effect. We aimed to address these questions by examining whether copper, 76 a widespread aquatic pollutant, can induce developmental plasticity in three-spined stickleback 77 (Gasterosteus aculeatus), a well-established model in evolutionary ecology and ecotoxicology. 78 We tested the hypothesis that priming would induce persistent, elevated tolerance to copper, 79 with both epigenetic and microbiome -mediated mechanisms contributing to this adaptive 80 response. To specifically examine the capacity for developmental plasticity, distinct from 81 acclimation, we exposed s tickleback embryos to an environmentally -relevant concentration of 82 copper during early development, returned them to control conditions for nine months (until 83 maturity), and then compared response to copper exposure in primed and naïve adults. 84 85 86

87 88 Early-life exposure promotes continued copper accumulation in the gill 89 90 Stickleback embryos were pre-exposed to an environmentally relevant concentration of copper 91 (nominal: 10 µg/L, measured 11.4 ±0.3 µg/L) during early development (one-cell stage to hatched 92 larvae; 1-217 hpf), alongside a synthetic freshwater control group (measured 0.2 ±0.004 µg/L Cu). 93 Survival was high in both groups, but copper exposure caused a small increase in embryo/larval 94 mortality rate (Naïve: 0.69%, Pre-exposed: 1.24%; P=0.0384). Pre-exposure also increased larval 95 whole-body copper concentration (t =-9.20, df = 5.58, P<0.001; Figure 1a). 96 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint After nine months depuration in clean water, both naïve and primed adults were (re-)exposed to 97 two concentrations of copper (nominal: 10 µg/L, measured: 13.5 ±0.3 µg/L; and nominal: 20 µg/L, 98 measured: 21.8 ±0.05 µg/L) for 96-h alongside a control (measured: 5.0 ±0.05 µg/L). Pre-exposed 99 adult stickleback had accumulated higher concentrations of copper in their gills compared to 100 the naïve fish, while adult exposure also increased gill copper concentration in both groups (Pre-101 exposure: F1,48=8.55, P=0.005, Adult-exposure: F2,48=21.18, P<0.001, Interaction: F2,48 =0.36, 102 P=0.69; Figure 1b). In contrast, there was no discernible effect of either pre -exposure or adult 103 exposure on the concentration of copper measured in muscle or liver tissue. No mortalities or 104 behavioural changes were observed during the adult copper exposure, and neither pre-exposure 105 nor adult exposure to copper affected fish size. 106 107 108 109 Figure 1. Copper measured in A) whole larvae after copper exposure during embryonic development (n= 110 12 pools of 5 larvae/group) and B) in the gills of stickleback from each group later exposed to copper for 111 96-h as adults after 9 months depuration in clean water (n= 10/group ). 112 113 Pre-exposure substantially reduces and modifies transcriptional stress response to copper 114 We focused the molecular analyses on the gills of adult fish, given their role in metal uptake and 115 the measured differential accumulation of copper in this tissue. We conducted transcriptomic 116 profiling in both the naïve and primed groups following (re-)exposure to 0 and 10 µg/L copper. 117 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint Pre-exposure to copper during embryonic development had minimal lasting effect s on baseline 118 transcription in adult fish, with only two differentially expressed genes (DEGs) identified between 119 primed and naïve fish. Gene Set Enrichment Analysis (GSEA), identified a limited number (seven) 120 of negatively enriched GO terms (Table S2); which were all related to cytoskeleton structure and 121 function (actin, myosin, troponin, calcium binding). 122 We then compared the transcriptomic response to copper exposure of primed (pre-exposed) 123 adults with that of naïve fish (adult fish exposed to copper for the first time ). The magnitude of 124 transcriptional response was far greater in naïve fish than in primed fish (1575 and 45 DEGs, 125 respectively; Figure 2, Table S1). Of these, 18 DEGs were common between groups, including 126 those encoding seven heat shock proteins (HSPs; subtypes 90,70 and 30). These molecular 127 chaperones, critical in cellular stress response, were the most significantly up -regulated genes 128 in response to copper in both groups, but the magnitude of this up -regulation was markedly 129 higher in naïve fish (ranging 7 -805 fold increase) than in pre -exposed fish (ranging 3 -170 fold 130 increase). Metallothionein B, a metal -sequestering protein, was also strongly up -regulated in 131 both groups, although, in this case, by a greater magnitude in the pre -exposed fish (5.9 fold 132 increase) compared to naïve fish (3.3 fold increase). 133 GSEA revealed a greater magnitude of response to copper in naïve fish (63 enriched terms) than 134 in pre-exposed fish (34 terms). In naïve fish there was strong enrichment of ‘DNA replication’, 135 ‘Cell-cycle’ and associated terms, as well as terms related to protein refolding and synthesis. 136 Similar terms were enriched in pre -exposed fish, but to a far lesser extent. Processes regulated 137 exclusively in naïve fish included strong enrichment of those associated with DNA repair and the 138 proteosome, while terms associated with ion homeostasis were supressed, reflecting down -139 regulation of >30 genes encoding potassium, sodium, calcium, magnesium, ammonium and 140 bicarbonate channels and cotransporters. A further marked distinction between the response of 141 each group was that terms associated with cytoskeleton (including actin, myosin, troponin and 142 calcium ion binding) and extracellular matrix (ECM) interactions were supressed in naïve fish but 143 enhanced in pre-exposed fish. 144 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 145 Figure 2. A) Number of significantly differentially expressed genes (DEGs) identified in response to pre -146 exposure alone (baseline) and in naïve or pre-exposed groups exposed to 10 µg/L copper as adults. B) 147 Heatmap visualising the expression of selected DEGs, based on their primary function (N= 135 DEGs out 148 of a total of 1608 identified in response to copper exposure across both groups (see Fig S1). C) Number and 149 the genomic context of differentially methylated regions (DMRs) identified in response to pre -exposure 150 alone (baseline) and in both naïve and pre-exposed groups exposed to 10 µg/L copper. D) Shared enriched 151 GO terms associated with both DMRs and DEGs. 152 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 153 Copper exposure induced marked and long-lasting changes in the gill methylome 154 We conducted genome -wide DNA methylation profiling (RRBS) in the gills of naïve and p rimed 155 fish after (re-)exposure, on the same samples used for transcriptomic analysis. In contrast to that 156 observed for transcription, we measured considerable, lasting changes in the gill methylome of 157 adult fish following developmental pre-exposure to copper and in the absence of any subsequent 158 exposures. A total of 615 differentially methylated regions (DMRs) were identified between the 159 pre-exposed and naïve groups (328 hyper -methylated, 287 hypo-methylated). Of these, 13.5 % 160 overlapped putative promoters (pps), while 20.4% and 42.2% were associated with exons and 161 introns, respectively (Figure 2c, Table S3). Among the genes associated with these DMRs 162 (overlapping pps, exons, introns), notable examples included those involved in ion homeostasis 163 (particularly sodium, potassium and calcium transport), metal transport and binding (including 164 those encoding copper -uptake protein 2, ceruloplasmin and ferritin), those with immune 165 function, and a number of lncRNAs. GSEA, performed separately for different genomic contexts, 166 identified 41, 50 and 45 enriched terms associated with DMRs located within p ps, exons and 167 introns, respectively (Table S 4, Fig ure S3). Among the most enriched terms were those 168 associated with membrane transport and ion homeostasis, including copper -ion transport. 169 Regulation of immune response (interleukin production) and many terms associated with 170 cellular growth and division, cell adhesion and signalling, were also evident. 171 Adult copper exposure also induced considerable changes in the methylome, and these were 172 more extensive in naïve fish. A total of 571 DMRs (287 hyper-methylated, 284 hypo-methylated) 173 were identified in naïve fish exposed to copper, compared to 385 DMRs (202 hyper -methylated, 174 183 hypo-methylated) in pre -exposed fish (Figure 3, Table S 3). Functional enrichment analysis 175 identified a greater number of terms associated with DMRs in the naïve fish (Table S4, Figure S3). 176 These included, most strongly, ‘immune system’, several terms related to chaperone-mediated 177 protein refolding and, more broadly, many terms related to protein, nucleic acid and cellular 178 repair and turnover. In pre -exposed fish, there were some broad similarities in the function of 179 enriched terms to those in the naïve fish , including those related to nucleosome, immune 180 response and cellular turnover but, notably, ‘ ion transport ’ was only enriched in pre -exposed 181 fish. 182 183 184 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint Genes & functional pathways with both epigenetic and transcriptional modifications 185 DMRs were identified within the gene body or putative promoter of 13 DEGs in naïve fish exposed 186 to copper as adults, and one DEG in pre-exposed fish exposed to copper (Table S5). Functions of 187 these genes included protein degradation, synthesis, folding and damage -repair, as well as 188 calcium signalling, cytoskeleton, and the regulation of cell cycle and cell movement. There were 189 also six shared enriched GO terms associated with both altered methylation and transcription in 190 naïve fish (relating to protein synthesis, folding and calcium signalling) and one shared term (DNA 191 replication) in pre-exposed fish (Figure 2D). 192 We hypothesised that persistent methylation differences following pre-exposure influenced the 193 transcriptional responses of adult stickleback (re -)exposed to copper. We identified 13 genes 194 with differential baseline methylation following pre-exposure that showed a different 195 transcriptional response to copper between primed and naïve adult fish; all of these genes were 196 only transcriptionally responsive to copper in the naïve fish (Table S5). The functions of these 197 genes were related to protein degradation and synthesis, DNA repair, regulation of cell cycle and 198 cell movement, as well as cytoskeleton and regulation of ion channels. Eight shared enriched GO 199 terms, related to cytoskeleton and potassium ion transport, were also identified (Figure 2d). 200 201 Early-life priming increases gill microbiota copper-tolerance 202 203 We hypothesised that priming during the early stages of microbiome establishment would 204 promote enrichment of gill-associated microbiota better able to withstand copper exposure. To 205 test this, we characterised the gill microbiomes of primed and naïve adult fish exposed to 0, 10 206 and 20 µg/L. Exposure to the higher concentration of copper disrupted microbiome community 207 structure (Bray-Curtis dissimilarity) in all fish, regardless of priming, but the effects of the lower 208 copper concentration differed between groups ( Adult exposure: F1,56 =5.938, P <0.001, Pre-209 exposure: F1,56 =1.221, P=0.189, Interaction: F1,56 =1.55, P=0.0493; Figure 3). While fish from the 210 naïve group exposed to 10 µg/L showed microbiome disruption similar to those exposed to 20 211 µg/L copper, the microbiomes of primed fish were more resistant to change and remained similar 212 to those of the fish unexposed to copper as adults. 213 We examined the differences in community composition contributing to these structural 214 changes. We identified two amplicon sequence variants (ASVs) with baseline differential 215 abundance between naïve and pre -exposed fish (Vibriomonas and Candidatus Bacilloplasma). 216 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint In adults exposed to copper, p re-exposure reduced the number of differentially abundant ASVs 217 identified (14 and 29 following exposure to 10 and 20 µg/L, respectively ; Table S6) compared to 218 in naïve fish (32 and 40 following exposure to 10 µg/L and 20 µg/L ; Table S6). Notably, the most 219 abundant ASV overall, Comamonas sp., was strongly inhibited by copper in naïve fish (reduced 220 by 13 - and 62 -fold following exposure to 10 and 20 µg/L ), but not in pre-exposed fish (2-fold 221 reduction in response to 20 µg/L only) . At the same time, there was a marked increase in the 222 abundance of ASVs from the genera Deinococcus, Enhydrobacter, Acinetobacter, 223 Flavobacterium and Brevundimonas in both groups, but generally the magnitude of increase was 224 higher in naïve fish. 225 There were no detectable effects of pre -exposure or adult exposure on overall richness or 226 diversity of ASVs present (Chao1 richness - Pre-exposure: F1,56=2.05, P=0.158, Adult-exposure: 227 F1,56=0.49, P=0. 485; Shannon diversity - Pre-exposure: F1,56=0.02, P=0.880, Adult exposure: 228 F1,56=0.09, P=0.768, Interaction: F1,56=1.00, P=0.32). 229 230 231 Figure 3. A) Gill microbial community structure in adult sticklebacks (re -)exposed to 0, 10 or 20 ug/L 232 copper, visualised using Bray -Curtis dissimilarity values, and B) Relative abundance of the top 20 most 233 abundant bacterial genera identified across all samples. 234 235 236 237 238 239 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint

240 241 We examined whether early -life priming induced developmental plasticity in stickleback using 242 copper as a model toxicant , hypothesis ing that epigenetic and microbiome -mediated 243 mechanisms contribute to this phenomenon. We found that priming substantially ameliorated 244 copper toxicity in adult fish, an effect characterised by reduced stress response at the 245 transcriptional level. In parallel, pre-exposure induced considerable, and persistent, changes in 246 the gill methylome, with notable similarities between differentially methylated and expressed 247 gene pathways suggesting a priming effect on cellular stress response pathways . Furthermore, 248 early-life priming markedly increased copper-tolerance in the gill microbiome, likely also 249 reducing the toxic effects of copper exposure on the stickleback host. 250 251 Early-life priming increased tolerance to copper toxicity in adult fish 252 253 Transcriptional response to copper was markedly reduced in adult fish that had been primed in 254 early life. In both naïve and pre-exposed fish, we identified transcriptional changes dominated by 255 genes and pathways associated with the cellular stress response (CSR), but the magnitude of 256 transcriptional changes and enrichment scores were far greater in naïve fish. The CSR is highly 257 and broadly inducible by many stressors, but its nature varies depending on the severity and 258 duration of the stressor 43,44. In both groups, although with a greater magnitude in naïve fish, we 259 characterised an extensive compensatory CSR, reflecting the repair of cellular components. As 260 part of this, we identified a marked increase in the transcription of heat shock proteins and other 261 molecular chaperones responsible for refolding of damaged proteins, as well as genes involved 262 in DNA repair pathways. There was also strong up -regulation of DNA replication, protein 263 synthesis and cell division pathways, consistent with an increase in cellular turnover. In naïve 264 fish only, there was a distinct enrichment of the proteosome, responsible for degradation of 265 irreversibly damaged proteins . This supports the induction of a more severe CSR in naïve fish, 266 characterised by a switch from repair to autophagic pathways , indicating that greater copper -267 induced cellular damage occurred in these fish. 268 269 Further evidence that priming reduc ed copper toxicity included markedly different 270 transcriptional effects associated with ion transport and cytoskeleton dynamics. In naïve fish 271 only, we found down -regulation of a suite of ion transporters, particularly potassium channels, 272 indicating broadscale disruption of ion homeostasis, a well -known mechanism of copper 273 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint toxicity, especially in the gills 45. Metals can induce cytoskeleton toxicity via interference with 274 calcium signalling and its regulation of troponin -tropomyosin-actin dynamics 46–48. Consistent 275 with this, in naïve fish only, we found inhibition of these pathways, indicative of substantive 276 toxicity. However, in pre -exposed fish, there was instead a consistent stimulatory effect, likely 277 associated with a compensatory CSR, encompassing transcripts associated with enhanced 278 cellular signalling, reorganisation and stability43,44. 279 280 In contrast to the reduced magnitude of CSR, we found that metallothionein, a key metal-binding 281 protein that sequesters and reduces the toxicity of free metal ions, was more strongly up -282 regulated in pre-exposed fish in response to copper exposure. A similar priming effect, increasing 283 inducibility of this protein, has been previously associated with enhanced tolerance to toxic 284 metals in populations with different exposure histories 49 and, more widely, the increased 285 inducibility of genes with protective functions contribute to developmental plasticity and 286 increased stressor tolerance 50. Importantly, the minimal differences in baseline transcription 287 identified between the naïve and pre-exposed fish are consistent with persistent developmental 288 plasticity, rather than acclimatory effects (i.e. frontloading transcription)50. 289 290 Importantly, we found no effects on long-term survival or growth in primed fish compared to their 291 naïve counterparts, suggesting that developmental plasticity in copper tolerance was not 292 associated with overt energetic costs. Surprisingly, early-life exposure caused fish to continue to 293 accumulate more copper in their gills during the nine -month depuration phase, indicative of 294 changes in metal-homeostasis physiology. Following pre -exposure, w e identified persistent 295 methylation differences in slc31a2 (copper-uptake protein) , ceruloplasmin (copper -transport 296 protein) as well > 20 sodium and calcium transporters and channels, which are also responsible 297 for a substantial amount of copper uptake in fish gills 51, although there were no differences in 298 their baseline transcription . Increased copper accumulation was only evident in the gills 299 suggesting that primed fish may have an increased tendency to sequester copper in this tissue, 300 likely in a less -toxic and/or bioavailable form, and consistent with the higher inducibility of 301 metallothionein proteins identified. 302 303 Epigenetic mechanisms contribute to developmental plasticity in copper-tolerance 304 305 Pre-exposure to copper caused extensive and persistent changes in the gill methylome, evident 306 even after nine months in control conditions. This has not been reported before for metals but is 307 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint consistent with existing evidence that developmental thermal stress induces long-lasting 308 changes i n DNA methylation in fish 25,52, and, more broadly, with there being heightened 309 epigenetic sensitivity in early life which, when combined with environmental fluctuation, drives 310 phenotypic plasticity 24. There were also extensive changes in the adult gill methylome 311 immediately following copper exposure, including within many of the same gene pathways 312 modulated by pre -exposure. This suggests that DNA methylation plays a similar role in acute 313 stress responses as in developmental priming, an effect previously identified in stickleback 314 challenged with thermal stress25. 315 316 We identified a considerable functional overlap in epigenomic and transcriptomic responses to 317 copper, including in many gene pathways broadly associated with the CSR (especially protein 318 turnover, cytoskeleton regulation and cell cycle) and ion transport . Thirteen genes were both 319 differentially methylated and expressed in acute response to copper exposure in adults , while 320 another 13 genes , that were persistently differentially methylated following pre -exposure, 321 subsequently showed different transcriptional responses to copper in primed and naïve fish. In 322 addition, we identified similarities in modulated gene pathways , which were predominantly 323 associated with the CSR and ion homeostasis . Together, our results suggest that epigenetic 324 regulation has a priming effect on CSR and ion homeostasis pathways and subsequently 325 contributes to the reduction in transcriptional stress response identified in pre -exposed fish. 326 Priming could, for example, facilitate a more efficient and/or readily inducible CSR, similar to that 327 which occurs in plants, where developmental stress priming induces diverse methylation 328 changes within CSR, defence and s ignalling pathways that are associated with improved 329 tolerance to many stressors, including metals53. 330 331 Priming increased copper tolerance of the gill microbiota 332 333 Copper exposure caused a dose -dependent, disruptive effect on gill microbial community 334 structure, that was more severe in naïve fish . Early-life priming reduced the number of 335 differentially abundant ASVs identified in response to both concentrat ions of copper and, in 336 particular, protected against significant community disruption by the lower (10 µg/L) 337 concentration. The disruptive effects of copper were characterised by a substantial decline in 338 the presence of the otherwise most abundant community member, Comamonas sp. In parallel, 339 there was an increased prevalence of genera including Deinococcus, known for its high abiotic 340 stressor tolerance54, as well as Acinetobacter and Flavobacterium, both of which are genera that 341 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint include opportunistic fish pathogens and have previously been associated with microbiome 342 stressor-disruption42,55. Our results support the hypothesis that priming enabled commensal 343 members of the core microbiome, including Comamonas, to develop enhanced copper 344 tolerance, likely through mechanisms such as genetic adaptation, gene transfer or plasticity 41. 345 This may have been further reinforced by elevated copper accumulation occurring in the gills of 346 primed fish, providing a sustained selective environment for tolerant strains. 347 348 We propose that the extensive changes in community structure observed were associated with 349 disrupted gill microbiome function, thereby increasing the toxic effects of copper to the host. 350 Considering the increase in opportunistic pathogens observed, a dverse effects on host 351 physiology could include a disruption of normal microbial contribution to pathogen defence and 352 provision of beneficial metabolites, and exacerbation of host inflammatory stress responses . 353 Having a more tolerant, less -disrupted microbiome, is likely to be beneficial to the host . In the 354 context of the holobiont concept, our results support the hypothesis that microbiome tolerance 355 also contributes to increased stickleback copper-tolerance. Further research should establish 356 whether microbiota additionally provide specific copper-adaptive benefits to the host , such as 357 sequestration of metals, as reported in plants41. 358 359

360 Developmental plasticity, together with acclimation, transgenerational plasticity and genetic 361 adaptation, contribute to variations in the sensitivity of natural populations to environmental 362 stressors, ultimately influencing their ability to survive50. Establishing the capacity for organisms 363 to acquire tolerance, and elucidating underlying molecular mechanisms, is therefore essential 364 to understanding and predicting the risks posed by pollution and other stressors in the natural 365 environment. Here, we provide some of the first evidence for developmental plasticity in 366 chemical tolerance in animals and demonstrate that both epigenetic and microbiome-mediated 367 mechanisms are associated with this effect. 368 369

Summary 370 371 Ethics approval 372 All experiments were approved by the University of Exeter Ethics committee and conducted 373 under licence from the UK Home Office according to ASPA. 374 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 375 Embryo copper experiment 376 Pools of 50 e mbryos were exposed to either a water control (0 µg/L copper) or 10 µg/L copper 377 (added as CuSO4) from 1- 217 hours post fertilisation (covering the period of embryogenesis and 378 hatching, and including the period of epigenetic reprogramming and microbiome colonisation). 379 Exposures were conducted in 500 ml acid -washed glass dishes containing aerated synthetic 380 freshwater56, with four replicates per treatment, repeated three times with different parental fish 381 (see SI). Embryo mortalities and hatching were recorded daily. Water samples were collected at 382 120 and 217h , and 20 larvae from each replicate (n=12) were collected at 217h for copper 383 measurement (see SI). Embryo survival and copper uptake were analysed using a student’s t-test 384 in R (v 4.3.3; 57). 200 larvae from each of the control and copper -exposed groups were then 385 maintained in dechlorinated tap water (in duplicate tanks per group) for nine months. 386 387 Adult copper exposure experiments 388 Adult male stickleback from both the naïve and pre-exposed groups were exposed to either 0, 10 389 or 20 µg/L copper for 96 hours. Each treatment was performed in duplicate 40 L tanks, with nine 390 fish per tank , supplied with flow-through dechlorinated tap water. Fish were not fed for the 391 duration of the exposure. Water samples were collected at 24 and 72h for copper measurements 392 (see SI). After exposure all fish were humanely sacrificed by lethal dose of benzocaine (0.5 g/L; 393 Sigma-Aldrich), followed by destruction of the brain. Gill, liver and muscle tissue w ere snap 394 frozen and stored at -80°C. All left-side gill arches were used for copper-content analysis (see 395 SI), while all right-side gill arches were used for molecular analysis. The effect of both pre -396 exposure and adult exposure on fish weight and tissue copper concentration was examined 397 using ANOVA in R. 398 399 Sequencing & Bioinformatics 400 Transcriptome, methylome and microbiome analys es were conducted on gill tissue from adult 401 fish, with full details in SI. Briefly, for transcriptome and methylome analyses, gill RNA and DNA 402 were co-extracted using Qiagen AllPrep DNA/RNA Mini kits from fish from four treatment groups 403 (naïve and pre-exposed fish exposed to 0 and 10 µg/L Cu; n=6 per group). RNA-seq libraries were 404 prepared using an Illumina TruSeq Stranded RNA Sample Preparation kit and sequenced using 405 an Illumina HiSeq 2500 ( 100 bp paired end). RRBS libraries were prepared using Ovation RRBS 406 Methyl-Seq kit (Tecan Systems) and sequenced using an Illumina NovaSeq (100 bp paired end). 407 For microbiome analysis, DNA was extracted from gill tissue s using the Qiagen PowerSoil DNA 408 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint Isolation Kit (n= 10 from all six treatment groups). Libraries were prepared, amplifying t he 16S 409 rRNA V4 region using primers 515F, 806R 58, based on the Illumina 16S Sequencing Library 410 Preparation protocol59 and sequenced using an Illumina MiSeq (300 bp paired end). 411 412 RNA-seq reads were quality filtered using Fastp (v0.23.1.3;60) and aligned to the Gasterosteus 413 aculeatus reference genome61 using STAR (v2.7.9a;62). Mapped reads were quantified with RSEM 414 (v1.3.1;63) and d ifferential gene expression analysis was performed using DESeq2 (v1.38.3; 64). 415 Gene set enrichment analysis (GSEA) was conducted using Cluster Profiler (v4.7.1; 65), 416 incorporating customised GO term annotations generated using InterProScan (v5.55.88.0;66) and 417 Blast2Go (v1.4.12;67). 418 419 RRBS reads were quality filtered using TrimGalore 68 and aligned to the reference genome with 420 Bismark (v0.23.1; 69. Differentially-methylated regions (DMRs) were identified using DSS 421 (v2.48.0;70). Genomic location of DMRs (classified as putative promoters (within 1000 bp of the 422 transcription start site ), exons, introns, or intergenic regions), and gene annotation s were 423 determined using Genomation (v3.17;71). GSEA was performed using g:Profiler72 using input gene 424 lists ranked by methylation fold change. 425 426 16S rDNA data were processed using DADA2 73 within Qiime2 (v2024.2,74). Reads were quality-427 filtered, merged, de-noised, assigned to amplicon sequence variants (ASVs) and taxonomically 428 classified using the Silva reference database (v132;75). Alpha and beta diversity metrics were 429 calculated using the Vegan package76. The effects of both early-life priming and adult copper 430 exposure on Chao1 richness and Shannon diversity were assessed using ANOVA. Bray-Curtis 431 dissimilarity was analysed using PERMANOVA and differential ASV abundance w as evaluated 432 using DESeq2 64. 433 434 Author contributions: 435 LL, TUW & ES conceived & designed the study . LL conducted the stickleback experiments with 436 help from JF, JP and AL. LL, HL, RMF, AL & TUW conducted molecular work and AF, KM and MH 437 conducted the Illumina sequencing. NB and LL conducted the metal analysis. TUW (RRBS & 16S), 438 JO (RNA-seq) and LL (metal content) led the data analyses with contributions from RvA, JP, NB & 439 ES. ES supervised the study and, together with LL & TUW, led funding acquisition. TUW wrote the 440 original draft. All authors contributed to review and editing of the final manuscript. 441 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint

442 We thank staff from the University of Exeter Aquatic Resources Centre for assistance with fish 443 husbandry. Funding was received from the Fisheries Society of the British Isles (FSBI), the BBSRC 444 (BB/S004300/1), NERC PhD studentships, the Exeter-Cefas strategic alliance, Cefas Seedcorn, 445 and the Swansea University College of Science Research Fund. RNA-seq and RRBS sequencing 446 was performed at the University of Exeter Sequencing Service, which utilised equipment funded 447 by the Wellcome Trust Institutional Strategic Support Fund (WT097835MF), Wellcome Trust 448 Multi-user Equipment Award (WT101650MA) and BBSRC LOLA award (BB/K003240/1). 449 450 451

452 1. Oziolor, E. M., Reid, N. M., Yair, S., Lee, K. M., Guberman VerPloeg, S., Bruns, P. C., 453 Shaw, J. R., Whitehead, A. & Matson, C. W. Adaptive introgression enables evolutionary 454 rescue from extreme environmental pollution. Science 364, 455–457 (2019). 455 2. Whitehead, A., Pilcher, W., Champlin, D. & Nacci, D. Common mechanism underlies 456 repeated evolution of extreme pollution tolerance. Proceedings of the Royal Society B: 457 Biological Sciences 279, 427–433 (2012). 458 3. Mathivanan, K., Chandirika, J. U., Vinothkanna, A., Yin, H., Liu, X. & Meng, D. Bacterial 459 adaptive strategies to cope with metal toxicity in the contaminated environment – A 460 review. Ecotoxicol Environ Saf 226, 112863 (2021). 461 4. Hanikenne, M. & Nouet, C. Metal hyperaccumulation and hypertolerance: a model for 462 plant evolutionary genomics. Curr Opin Plant Biol 14, 252–259 (2011). 463 5. DiGiacopo, D. G. & Hua, J. Evaluating the fitness consequences of plasticity in tolerance 464 to pesticides. Ecol Evol 10, 4448–4456 (2020). 465 6. Rivera, H. E., Aichelman, H. E., Fifer, J. E., Kriefall, N. G., Wuitchik, D. M., Smith, S. J. & 466 Davies, S. W. A framework for understanding gene expression plasticity and its influence 467 on stress tolerance. Mol Ecol 30, 1381–1397 (2021). 468 7. Pottier, P., Burke, S., Zhang, R. Y., Noble, D. W. A., Schwanz, L. E., Drobniak, S. M. & 469 Nakagawa, S. Developmental plasticity in thermal tolerance: Ontogenetic variation, 470 persistence, and future directions. Ecol Lett 25, 2245–2268 (2022). 471 8. Lafuente, E. & Beldade, P. Genomics of Developmental Plasticity in Animals. Front Genet 472 10, 720 (2019). 473 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 9. Saro, L., Lopes, I., Martins, N. & Ribeiro, R. Testing hypotheses on the resistance to 474 metals by Daphnia longispina: differential acclimation, endpoints association, and 475 fitness costs. Environ Toxicol Chem 31, 909–915 (2012). 476 10. Earhart, M. L., Blanchard, T. S., Harman, A. A. & Schulte, P. M. Hypoxia and High 477 Temperature as Interacting Stressors: Will Plasticity Promote Resilience of Fishes in a 478 Changing World? Biological Bulletin 243, 149–170 (2022). 479 11. Burggren, W. W. Phenotypic switching resulting from developmental plasticity: fixed or 480 reversible? Front Physiol 10, 1634 (2020). 481 12. Beldade, P., Mateus, A. R. A. & Keller, R. A. Evolution and molecular mechanisms of 482 adaptive developmental plasticity. Mol Ecol 20, 1347–1363 (2011). 483 13. Avramova, Z. Transcriptional ‘memory’of a stress: transient chromatin and memory 484 (epigenetic) marks at stress‐response genes. The Plant Journal 83, 149–159 (2015). 485 14. Kumar, V., Pandita, S., Singh Sidhu, G. P., Sharma, A., Khanna, K., Kaur, P., Bali, A. S. & 486 Setia, R. Copper bioavailability, uptake, toxicity and tolerance in plants: A 487 comprehensive review. Chemosphere 262, 127810 (2021). 488 15. Liu, X., Quan, W. & Bartels, D. Stress memory responses and seed priming correlate with 489 drought tolerance in plants: an overview. Planta 255, 45 (2022). 490 16. Vagner, M., Zambonino-Infante, J.-L. & Mazurais, D. Fish facing global change: are early 491 stages the lifeline? Mar Environ Res 147, 159–178 (2019). 492 17. Jonsson, B. & Jonsson, N. Phenotypic plasticity and epigenetics of fish: embryo 493 temperature affects later-developing lift-history traits. Aquat Biol 28, 21–32 (2019). 494 18. Hua, J., Jones, D. K. & Relyea, R. A. Induced Tolerance from a Sublethal Insecticide Leads 495 to Cross-Tolerance to Other Insecticides. Environ Sci Technol 48, 4078–4085 (2014). 496 19. Poupardin, R., Reynaud, S., Strode, C., Ranson, H., Vontas, J. & David, J.-P. Cross-497 induction of detoxification genes by environmental xenobiotics and insecticides in the 498 mosquito Aedes aegypti: Impact on larval tolerance to chemical insecticides. Insect 499 Biochem Mol Biol 38, 540–551 (2008). 500 20. Plautz, S. C. & Salice, C. J. Plasticity in offspring contaminant tolerance traits: 501 developmental cadmium exposure trumps parental effects. Ecotoxicology 22, 847–853 502 (2013). 503 21. Ledesma, M., Gorokhova, E., Nybom, I., Sobek, A., Ahlström, D., Garbaras, A. & Karlson, 504 A. M. Does pre-exposure to polluted sediment affect sub-cellular to population-level 505 responses to contaminant exposure in a sentinel species? Environmental Pollution 341, 506 122882 (2024). 507 22. Yagound, B., Sarma, R. R., Edwards, R. J., Richardson, M. F., Rodriguez Lopez, C. M., 508 Crossland, M. R., Brown, G. P., DeVore, J. L., Shine, R. & Rollins, L. A. Is developmental 509 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint plasticity triggered by DNA methylation changes in the invasive cane toad (Rhinella 510 marina)? Ecol Evol 14, e11127 (2024). 511 23. Lo, N., Simpson, S. J. & Sword, G. A. Epigenetics and developmental plasticity in 512 orthopteroid insects. Curr Opin Insect Sci 25, 25–34 (2018). 513 24. Venney, C. J., Anastasiadi, D., Wellenreuther, M. & Bernatchez, L. The Evolutionary 514 Complexities of DNA Methylation in Animals: From Plasticity to Genetic Evolution. 515 Genome Biol Evol 15, evad216 (2023). 516 25. Metzger, D. C. H. & Schulte, P. M. Persistent and plastic effects of temperature on DNA 517 methylation across the genome of threespine stickleback (Gasterosteus aculeatus). 518 Proceedings of the Royal Society B: Biological Sciences 284, 20171667 (2017). 519 26. Liu, H., Able, A. J. & Able, J. A. Priming crops for the future: rewiring stress memory. 520 Trends Plant Sci 27, 699–716 (2022). 521 27. Iqbal, B., Ahmad, N., Li, G., Jalal, A., Khan, A. R., Zheng, X., Naeem, M. & Du, D. 522 Unlocking plant resilience: Advanced epigenetic strategies against heavy metal and 523 metalloid stress. Plant Science 349, 112265 (2024). 524 28. Loughland, I., Little, A. & Seebacher, F. DNA methyltransferase 3a mediates 525 developmental thermal plasticity. BMC Biol 19, 11 (2021). 526 29. Ryu, T., Veilleux, H. D., Munday, P. L., Jung, I., Donelson, J. M. & Ravasi, T. An epigenetic 527 signature for within-generational plasticity of a reef fish to ocean warming. Front Mar Sci 528 7, 284 (2020). 529 30. Abdelnour, S. A., Naiel, M. A. E., Said, M. Ben, Alnajeebi, A. M., Nasr, F. A., Al-Doaiss, A. 530 A., Mahasneh, Z. M. H. & Noreldin, A. E. Environmental epigenetics: Exploring phenotypic 531 plasticity and transgenerational adaptation in fish. Environ Res 252, 118799 (2024). 532 31. Šrut, M. Ecotoxicological epigenetics in invertebrates: Emerging tool for the evaluation of 533 present and past pollution burden. Chemosphere 282, 131026 (2021). 534 32. Adegbola, P. I. & Adetutu, A. Genetic and epigenetic modulations in toxicity: The two-535 sided roles of heavy metals and polycyclic aromatic hydrocarbons from the 536 environment. Toxicol Rep 12, 502 (2024). 537 33. Mäkinen, H., van Oers, K., Eeva, T. & Ruuskanen, S. The effect of experimental lead 538 pollution on DNA methylation in a wild bird population. Epigenetics 17, 625–641 (2022). 539 34. Koskella, B., Hall, L. J. & Metcalf, C. J. E. The microbiome beyond the horizon of 540 ecological and evolutionary theory. Nat Ecol Evol 1, 1606–1615 (2017). 541 35. Lynch, J. B. & Hsiao, E. Y. Microbiomes as sources of emergent host phenotypes. 542 Science (1979) 365, 1405–1409 (2019). 543 36. Claus, S. P., Guillou, H. & Ellero-Simatos, S. The gut microbiota: a major player in the 544 toxicity of environmental pollutants? NPJ Biofilms Microbiomes 2, 16003 (2016). 545 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 37. Evariste, L., Barret, M., Mottier, A., Mouchet, F., Gauthier, L. & Pinelli, E. Gut microbiota 546 of aquatic organisms: A key endpoint for ecotoxicological studies. Environmental 547 Pollution 248, 989–999 (2019). 548 38. Alberdi, A., Aizpurua, O., Bohmann, K., Zepeda-Mendoza, M. L. & Gilbert, M. T. P. Do 549 Vertebrate Gut Metagenomes Confer Rapid Ecological Adaptation? Trends Ecol Evol 31, 550 689–699 (2016). 551 39. Li, J. & King, K. Microbial Primer: Microbiome and thermal tolerance - a new frontier in 552 climate resilience? Microbiology (Reading) 171, 001523 (2025). 553 40. Lee, S., Kim, Y. & Choi, J. Effect of soil microbial feeding on gut microbiome and 554 cadmium toxicity in Caenorhabditis elegans. Ecotoxicol Environ Saf 187, 109777 (2020). 555 41. Fu, S., Iqbal, B., Li, G., Alabbosh, K. F., Khan, K. A., Zhao, X., Raheem, A. & Du, D. The role 556 of microbial partners in heavy metal metabolism in plants: a review. Plant Cell Reports 557 2024 43:4 43, 1–20 (2024). 558 42. Uren Webster, T. M. U., Consuegra, S. & Garcia de Leaniz, C. Early life stress causes 559 persistent impacts on the microbiome of Atlantic salmon. Comp Biochem Physiol Part D 560 Genomics Proteomics 40, 100888 (2021). 561 43. Somero, G. N. The cellular stress response and temperature: Function, regulation, and 562 evolution. J Exp Zool A Ecol Integr Physiol 333, 379–397 (2020). 563 44. Kültz, D. Evolution of cellular stress response mechanisms. J Exp Zool A Ecol Integr 564 Physiol 333, 359–378 (2020). 565 45. Liao, W., Zhu, Z., Feng, C., Yan, Z., Hong, Y., Liu, D. & Jin, X. Toxicity mechanisms and 566 bioavailability of copper to fish based on an adverse outcome pathway analysis. Journal 567 of Environmental Sciences 127, 495–507 (2023). 568 46. Muralidharan, S., Thompson, E., Raftos, D., Birch, G. & Haynes, P. A. Quantitative 569 proteomics of heavy metal stress responses in Sydney rock oysters. Proteomics 12, 906–570 921 (2012). 571 47. Rebuzzini, P., Civello, C., Fassina, L., Zuccotti, M. & Garagna, S. Functional and 572 structural phenotyping of cardiomyocytes in the 3D organization of embryoid bodies 573 exposed to arsenic trioxide. Scientific Reports 2021 11:1 11, 1–12 (2021). 574 48. Go, Y. M., Orr, M. & Jones, D. P. Actin cytoskeleton redox proteome oxidation by 575 cadmium. Am J Physiol Lung Cell Mol Physiol 305, L831 (2013). 576 49. Uren Webster, T. M., Williams, T. D. T. D., Katsiadaki, I., Lange, A., Lewis, C., Shears, J. A. 577 J. A., Tyler, C. R. C. R. & Santos, E. M. E. M. Hepatic transcriptional responses to copper 578 in the three-spined stickleback are affected by their pollution exposure history. Aquatic 579 Toxicology 184, 26–36 (2017). 580 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 50. Rivera, H. E., Aichelman, H. E., Fifer, J. E., Kriefall, N. G., Wuitchik, D. M., Smith, S. J. & 581 Davies, S. W. A framework for understanding gene expression plasticity and its influence 582 on stress tolerance. Mol Ecol 30, 1381–1397 (2021). 583 51. Kwong, R. W. M. Trace metals in the teleost fish gill: biological roles, uptake regulation, 584 and detoxification mechanisms. Journal of Comparative Physiology B 194, 749–763 585 (2024). 586 52. Anastasiadi, D., Shao, C., Chen, S. & Piferrer, F. Footprints of global change in marine 587 life: Inferring past environment based on DNA methylation and gene expression marks. 588 Mol Ecol 30, 747–760 (2021). 589 53. Chmielowska-Bąk, J., Searle, I. R., Wakai, T. N. & Arasimowicz-Jelonek, M. The role of 590 epigenetic and epitranscriptomic modifications in plants exposed to non-essential 591 metals. Front Plant Sci 14, 1278185 (2023). 592 54. Lim, S., Jung, J. H., Blanchard, L. & De Groot, A. Conservation and diversity of radiation 593 and oxidative stress resistance mechanisms in Deinococcus species. FEMS Microbiol 594 Rev 43, 19 (2018). 595 55. Xavier, R., Severino, R. & Silva, S. M. Signatures of dysbiosis in fish microbiomes in the 596 context of aquaculture. Rev Aquac 16, 706–731 (2023). 597 56. ISO. ISO 7346-3:1996 - Water Quality — Determination of the Acute Lethal Toxicity of 598 Substances to a Freshwater Fish [Brachydanio Rerio Hamilton-Buchanan (Teleostei, 599 Cyprinidae). https://www.iso.org/standard/14030.html (2006). 600 57. R Development Core Team, R. & R Core Team. R: A Language and Environment for 601 Statistical Computing. R Foundation for Statistical Computing vol. 1 409 Preprint at 602 https://doi.org/10.1007/978-3-540-74686-7 (2019). 603 58. Walters, W., Hyde, E. R., Berg-Lyons, D., Ackermann, G., Humphrey, G., Parada, A., 604 Gilbert, J. A., Jansson, J. K., Caporaso, J. G., Fuhrman, J. A., Apprill, A. & Knight, R. 605 Improved Bacterial 16S rRNA Gene (V4 and V4-5) and Fungal Internal Transcribed Spacer 606 Marker Gene Primers for Microbial Community Surveys. mSystems 1, e00009-15 (2016). 607 59. Illumina. 16S Metagenomic Sequencing Library Preparation Preparing 16S Ribosomal 608 RNA Gene Amplicons for the Illumina MiSeq System. (2013). 609 60. Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. 610 Bioinformatics 34, i884–i890 (2018). 611 61. Nath, S., Shaw, D. E. & White, M. A. Improved contiguity of the threespine stickleback 612 genome using long-read sequencing. G3 Genes|Genomes|Genetics 11, jkab007 (2021). 613 62. Dobin, A., Davis, C. A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., 614 Chaisson, M. & Gingeras, T. R. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 615 29, 15–21 (2013). 616 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint 63. Li, B. & Dewey, C. N. RSEM: Accurate transcript quantification from RNA-Seq data with 617 or without a reference genome. BMC Bioinformatics 12, 1–16 (2011). 618 64. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion 619 for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014). 620 65. Wu, T., Hu, E., Xu, S., Chen, M., Guo, P., Dai, Z., Feng, T., Zhou, L., Tang, W. & Zhan, L. I. 621 clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The 622 innovation 2, (2021). 623 66. Jones, P., Binns, D., Chang, H. Y., Fraser, M., Li, W., McAnulla, C., McWilliam, H., 624 Maslen, J., Mitchell, A., Nuka, G., Pesseat, S., Quinn, A. F., Sangrador-Vegas, A., 625 Scheremetjew, M., Yong, S. Y., Lopez, R. & Hunter, S. InterProScan 5: genome-scale 626 protein function classification. Bioinformatics 30, 1236 (2014). 627 67. Götz, S., García-Gómez, J. M., Terol, J., Williams, T. D., Nagaraj, S. H., Nueda, M. J., 628 Robles, M., Talón, M., Dopazo, J. & Conesa, A. High-throughput functional annotation 629 and data mining with the Blast2GO suite. Nucleic Acids Res 36, 3420–3435 (2008). 630 68. Krueger, F. Trim galore: A wrapper tool around Cutadapt and FastQC to consistently 631 apply quality and adapter trimming to FastQ files. Preprint at 632 https://doi.org/10.5281/zenodo.5127898 (2015). 633 69. Krueger, F. & Andrews, S. R. Bismark: A flexible aligner and methylation caller for 634 Bisulfite-Seq applications. Bioinformatics 27, 1571–1572 (2011). 635 70. Feng, H., Conneely, K. N. & Wu, H. A Bayesian hierarchical model to detect differentially 636 methylated loci from single nucleotide resolution sequencing data. Nucleic Acids Res 637 42, e69–e69 (2014). 638 71. Akalin, A., Franke, V., Vlahoviček, K., Mason, C. E. & Schübeler, D. Genomation: a toolkit 639 to summarize, annotate and visualize genomic intervals. Bioinformatics 31, 1127–1129 640 (2015). 641 72. Raudvere, U., Kolberg, L., Kuzmin, I., Arak, T., Adler, P., Peterson, H. & Vilo, J. g: Profiler: a 642 web server for functional enrichment analysis and conversions of gene lists (2019 643 update). Nucleic Acids Res 47, W191–W198 (2019). 644 73. Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A. & Holmes, S. P. 645 DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods 13, 646 581–583 (2016). 647 74. Bolyen, E., Rideout, J. R., Dillon, M. R., Bokulich, N. A., Abnet, C. C., Al-Ghalith, G. A., 648 Alexander, H., Alm, E. J., Arumugam, M., Asnicar, F., Bai, Y., Bisanz, J. E., Bittinger, K., 649 Brejnrod, A., Brislawn, C. J., Brown, C. T., Callahan, B. J., Caraballo-Rodríguez, A. M., 650 Chase, J., Cope, E. K., Da Silva, R., Diener, C., Dorrestein, P. C., Douglas, G. M., Durall, 651 D. M., Duvallet, C., Edwardson, C. F., Ernst, M., Estaki, M., Fouquier, J., Gauglitz, J. M., 652 Gibbons, S. M., Gibson, D. L., Gonzalez, A., Gorlick, K., Guo, J., Hillmann, B., Holmes, S., 653 Holste, H., Huttenhower, C., Huttley, G. A., Janssen, S., Jarmusch, A. K., Jiang, L., 654 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint Kaehler, B. D., Kang, K. Bin, Keefe, C. R., Keim, P., Kelley, S. T., Knights, D., Koester, I., 655 Kosciolek, T., Kreps, J., Langille, M. G. I., Lee, J., Ley, R., Liu, Y.-X., Loftfield, E., Lozupone, 656 C., Maher, M., Marotz, C., Martin, B. D., McDonald, D., McIver, L. J., Melnik, A. V, Metcalf, 657 J. L., Morgan, S. C., Morton, J. T., Naimey, A. T., Navas-Molina, J. A., Nothias, L. F., 658 Orchanian, S. B., Pearson, T., Peoples, S. L., Petras, D., Preuss, M. L., Pruesse, E., 659 Rasmussen, L. B., Rivers, A., Robeson, M. S., Rosenthal, P., Segata, N., Shaffer, M., 660 Shiffer, A., Sinha, R., Song, S. J., Spear, J. R., Swafford, A. D., Thompson, L. R., Torres, P. 661 J., Trinh, P., Tripathi, A., Turnbaugh, P. J., Ul-Hasan, S., van der Hooft, J. J. J., Vargas, F., 662 Vázquez-Baeza, Y., Vogtmann, E., von Hippel, M., Walters, W., Wan, Y., Wang, M., 663 Warren, J., Weber, K. C., Williamson, C. H. D., Willis, A. D., Xu, Z. Z., Zaneveld, J. R., 664 Zhang, Y., Zhu, Q., Knight, R. & Caporaso, J. G. Reproducible, interactive, scalable and 665 extensible microbiome data science using QIIME 2. Nat Biotechnol 37, 852–857 (2019). 666 75. Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., Peplies, J. & 667 Glöckner, F. O. The SILVA ribosomal RNA gene database project: improved data 668 processing and web-based tools. Nucleic Acids Res 41, D590–D596 (2012). 669 76. Oksanen, J. Vegan: ecological diversity. R Package Version 2.4-4 1, 11 (2017). 670 671 .CC-BY 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 November 29, 2025. ; https://doi.org/10.1101/2025.11.28.691212doi: bioRxiv preprint