# Source paper archive - Source title: Combined heat and drought stress alters chromatin accessibility and primes potato abscisic acid biosynthesis - Author(s): Godec, T., Kriznik, M., Zagorscak, M., Stare, K., Ramsak, Z., Sonnewald, S., Abdelhakim, L., Panzarova, K., Rodriguez Granados, N., Hancock, R. D., Taylor, M. A., Campbell, R., Sasidharan, R., Prat, S., Shaikh, M. A., Teige, M., Petek, M., Gruden, K. - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.11.765576v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.11.765576v1.full.pdf - PDF SHA-256: 8a5a1c833cbf175fbc12d9f1b44b9eb0c8e5bace021854ce46a16858e858969e - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1 **Combined heat and drought stress alters chromatin accessibility and primes potato** 2 **abscisic acid biosynthesis** 3 4 Tim Godec [1,2,*], Maja Križnik [1,3], Maja Zagorščak [1,3], Katja Stare [1], Živa Ramšak [1], Sophia 5 Sonnewald [4], Lamis Abdelhakim [5], Klára Panzarová [5], Natalia Rodriguez Granados [6], Robert D 6 Hancock [7], Mark A Taylor [7], Raymond Campbell [7], Rashmi Sasidharan [6], Salomé Prat [8], Maroof 7 Ahmed Shaikh [8], Markus Teige [9], Marko Petek [1,3,*], and Kristina Gruden [1,3,*] 8 9 1 Department of Biotechnology and Systems Biology, National Institute of Biology, Večna pot 121, 1000 Ljubljana, 10 Slovenia 11 2 Jožef Stefan International Postgraduate School, Jamova cesta 39, 1000 Ljubljana, Slovenia 12 3 Biotechnical Faculty, University of Ljubljana, Jamnikarjeva ulica 101, 1000 Ljubljana, Slovenia 13 4 Chair of Biochemistry, Department of Biology, Friedrich-Alexander-Universität Erlangen-Nürnberg, Staudtstraße 14 5, 91058 Erlangen, Germany 15 5 PSI (Photon Systems Instruments), spol. s r.o., Průmyslová 470, 664 24 Drásov, Czech Republic 16 6 Plant Stress Resilience, Institute of Environmental Biology, Utrecht University, Padualaan 8, 3584 CH Utrecht, 17 The Netherlands 18 7 Cell and Molecular Sciences, The James Hutton Institute, Invergowrie, Dundee DD2 5DA, UK 19 8 Centre for Research in Agricultural Genomics (CRAG), Barcelona, 08193 Spain 20 9 Molecular Systems Biology (MOSYS), Department of Functional and Evolutionary Ecology, University of Vienna, 21 Djerassiplatz 1, 1030 Vienna, Austria 22 - Corresponding authors: Tim Godec (tim.godec@nib.si), Marko Petek (marko.petek@nib.si), Kristina Gruden 23 (kristina.gruden@nib.si) 24 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 25 **ABSTRACT** 26 27 Due to climate change, abiotic stresses get increasingly more frequent, in a combined and 28 recurrent manner in the same growing season. Chromatin accessibility enables efficient 29 responses to combined stress and priming for reoccurring stresses. To study the regulatory 30 basis of priming, we profiled chromatin accessibility and gene expression in leaves of two 31 genetically distinct potato cultivars after prolonged heat, drought, waterlogging, and combined 32 heat and drought. We show that most accessibility changes are stress-specific, however, 33 drought stress elicits a distinct transcriptional response as compared to combined heat and 34 drought stress, with strong chaperone induction and dampened abscisic acid signalling, 35 despite water deficit. A second cycle of combined stress reveals transcriptional memory of 36 abscisic acid biosynthesis gene _StNCED3_ and its transcriptional activator StWRKY33b, 37 providing a new mechanistic insight of abiotic stress memory in a staple crop important for 38 global food security. 39 40 Keywords: _Solanum tuberosum_, epigenomics, ATAC-seq, RNA-seq, ABA, transcription factor 41 motif, acclimation, chromatin accessibility 42 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 43 **INTRODUCTION** 44 45 Potato ( _Solanum tuberosum_ ) is one of the world's most important food crops and a key 46 contributor to global food security (Devaux et al., 2020). Yet, potato is particularly vulnerable 47 to abiotic stress, including heat, drought and waterlogging, which can disrupt plant growth and 48 tuber development and reduce yield (Zagorščak et al., 2025). Climate change is expected to 49 increase the frequency and severity of such stresses, making improved crop resilience an 50 important breeding target (Zandalinas & Mittler, 2022). These stresses rarely occur in isolation 51 under field conditions; periods of high temperature often coincide with declining soil water 52 availability, and drought or heat may follow periods of excessive rainfall (Zhang & Sonnewald, 53 2017). This creates a need to understand plant responses not only to individual stresses, but 54 also to their occurrence in combination and sequence (Zagorščak et al., 2025). Responses to 55 combined or successive stress cannot necessarily be inferred from responses to each 56 individual component (Jiang et al., 2025). In _Arabidopsis_, combined heat and drought produce 57 distinct physiological and transcriptional responses specific to the stress combination (Jiang 58 et al., 2026; Morales et al., 2022; Rizhsky et al., 2004). Similar combination-specific responses 59 have been observed across multifactorial and sequential stress treatments in other plants 60 (Jiang et al., 2026; Zandalinas et al., 2021; Zandalinas & Mittler, 2022; Zhang & Sonnewald, 61 2017), suggesting that interactions among stress signals can generate regulatory states that 62 are not captured by single-stress conditions. However, the regulatory basis of these 63 combination-specific transcriptional responses remains poorly understood. 64 65 Chromatin accessibility represents an important regulatory layer underlying transcriptional 66 responses to environmental stress. Regulatory elements such as promoters and enhancers 67 must be accessible to transcription factors and the transcriptional machinery, and changes in 68 chromatin accessibility can therefore alter the potential for gene regulation (Lu et al., 2019). 69 Assay for transposase-accessible chromatin using sequencing (ATAC-seq) enables genome70 wide profiling of these accessible regions (Buenrostro et al., 2013). ATAC-seq offers a broader 71 regulatory read-out than other epigenetic approaches because it directly identifies regions of 72 accessible chromatin, whereas DNA methylation and histone-based assays measure 73 additional regulatory features. A study in Arabidopsis mapping accessible chromatin identified 74 thousands of environmentally responsive regulatory elements and enabled reconstruction of 75 the transcription factor circuits underlying the heat-shock response (Sullivan et al., 2014). 76 Similar coupling between accessibility and expression has been reported under heat and 77 waterlogging in rice (Liang et al., 2021; Reynoso et al., 2019), and across heat, salt and 78 drought in the xerophyte _Zygophyllum xanthoxylum_ (T.-Q. Wang et al., 2025). In tomato, 79 integrated ATAC-seq and RNA-seq analysis identified differential accessibility associated with 80 natural variation in the promoter of _SlWRKY34_ affecting cold tolerance (Guo et al., 2024). 81 These studies illustrate the link between regulatory variation and chromatin accessibility in 82 stress, while regulation of plant acclimatisation to combined stress and priming for reoccurring 83 stress conditions is less well understood (Gallusci et al., 2023). Priming is defined as the 84 enhanced capacity to respond to a reoccurring stress following an initial stress exposure, 85 mediated by stress memory (H. Liu et al., 2022). One mechanism explaining this long-term 86 (memory) phenomenon is that stress is imprinted in changes of chromatin organization. For 87 example, it was shown that DNA methylation patterns play an important role in immune 88 priming in Arabidopsis (Kurtz et al., 2025; Parker et al., 2025). Transcriptional signatures of 89 abiotic stress priming in potato remain poorly characterised, although mild heat priming has 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 90 been shown to alter transcript dynamics following subsequent heat shock, with chromatin 91 remodelling among the implied mechanisms (Trapero-Mozos et al., 2018). 92 93 Responses at the chromatin level after prolonged abiotic stresses and their combinations in 94 potato remain largely unexplored. Existing potato stress-treatment chromatin accessibility 95 studies have focused on cold acclimation in a potato wild relative, or drought and salt 96 treatments (X. Li et al., 2024; Wen et al., 2025) in the potato DM genotype (Wen et al., 2025). 97 Recent analysis of two contrasting cultivars found substantial genotype-dependent DNA 98 methylation differences, but limited heat-induced methylation changes, and no consistent 99 correspondence between differentially methylated regions and heat-responsive gene 100 expression (Yeo et al., 2026). Moreover, to our knowledge, chromatin accessibility under 101 waterlogging stress, or under combined or sequential heat and drought, has not yet been 102 investigated at the genome-wide level in potato. 103 104 Here, we address this gap by profiling chromatin accessibility with ATAC-seq in leaves of two 105 genetically distinct potato cultivars, Désirée and Lady Rosetta, subjected to prolonged heat, 106 drought and waterlogging, as well as to a combined heat and drought stress. RNA-seq was 107 used to complement the accessibility data and identify transcriptional responses associated 108 with chromatin remodelling. Because heat and drought frequently recur and coincide within a 109 single growing season (Zhang & Sonnewald, 2017), this combination is the most relevant for 110 studying plant responses to repeated stress. To test whether the observed responses reflect 111 a priming effect, we exposed Désirée, the cultivar with the stronger chromatin response, to 112 two cycles of combined drought and heat and followed its transcriptional response. Using this 113 approach, we identified novel regulators of abscisic acid (ABA) biosynthesis in plants carrying 114 a memory of the first stress. 115 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 116 **RESULTS** 117 118 **Prolonged abiotic stresses induce distinctive changes in the accessible chromatin** 119 **landscape of potato leaves** 120 121 **Figure 1. Accessible chromatin landscape of potato leaves in response to diverse abiotic** 122 **stresses.** **A)** Experimental design of initial abiotic stress experiment showing stress treatments, 123 treatment durations, and sampling time points. **B)** Sample diversity based on detected accessible 124 chromatin regions: Multidimensional scaling (MDS) of the filtered peaks across samples. MDS 125 dimensions 1 and 2 (left) and dimensions 2 and 3 (right) are shown. Percentages of variation are given 126 in parentheses. **C)** Distribution of absolute log ₂ fold changes (|log ₂ FC|) across 56,724 filtered peaks for 127 each stress and cultivar. Boxes indicate the median and interquartile range (IQR), and whiskers extend 128 to 1.5 × IQR. **D)** Overlap of differentially accessible regions (|log ₂ FC| > 0.5, FDR < 0.05; response 129 averaged over both cultivars) among treatments. **E)** Pearson correlations of log ₂ fold changes for 130 accessible regions (lower left) and expressed genes (upper right) between different stress treatments. 131 **F)** Relationship between promoter accessibility and transcript log ₂ fold change for genes showing 132 significant responses in both assays (FDR < 0.05), shown separately by cultivar and stress. For each 133 gene, the promoter region with the lowest FDR in the differential accessibility analysis was selected as 134 the representative region. Shaded quadrants indicate concordant responses, and the percentage of 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 135 genes showing concordant responses (i.e., changes with the same sign in both assays) is indicated in 136 each plot. n denotes the number of genes. 137 ATAC-seq, assay for transposase-accessible chromatin with sequencing; FDR, false discovery rate; H, 138 heat; D, drought; W, waterlogging; HD, combined heat and drought. 139 140 Plants that have experienced stress are better prepared to withstand the reoccurring adverse 141 conditions. Here we asked whether diverse abiotic stresses leave an imprint in chromatin 142 organisation, which can determine potato acclimation to upcoming abiotic stress. We exposed 143 mature plants to prolonged heat, drought or waterlogging, and after heat followed by combined 144 heat and drought (heat–drought), designed to reflect the stress progression under climate 145 change conditions (Fig. 1A). To capture universal chromatin structure rearrangement across 146 different potato genotypes, we analysed two cultivars, Désirée and Lady Rosetta, which have 147 distinct breeding histories and therefore distinct genomic backgrounds (van Berloo et al., 148 2007). Across the experiment we identified ~147,000 accessible chromatin regions (ACR) 149 covering ~6% of the potato genome. ACRs were detected in promoters of over 80% of 150 transcribed genes, indicating high sensitivity of the applied methodology (Suppl. Fig. S1). 151 ACRs were enriched around the transcription start sites (TSS) (Suppl. Fig. S2B). Nearly one 152 third of ACRs were distal intergenic (located more than 3 kb from TSS), confirming that 153 regulatory regions extend well beyond proximal promoters (Suppl. Fig. S2C). 154 155 We compared the four regimes to test whether chromatin reorganization is shared or specific 156 for the stress responses in our study. Waterlogging caused the largest accessibility changes 157 in both cultivars, followed by heat–drought and drought. (Fig. 1C). Heat had the weakest effect 158 in Désirée, whereas Lady Rosetta heat-treated plants were barely distinguishable from 159 controls. Beyond differing in magnitude, the accessibility changes were also highly stress160 specific: about two thirds of responsive regions responded to only one regime, and fewer than 161 a hundred responded to all four (Fig. 1D). The largest shared sets involved drought, together 162 with heat–drought, waterlogging, or both. The overlap with heat–drought is expected from their 163 shared water deficit, whereas the overlap with waterlogging suggests it shares part of the 164 drought response. Regions that opened under drought, heat–drought and waterlogging lay at 165 canonical ABA- and dehydration-responsive genes, among them genes encoding ABF2, a 166 HAI-type clade A protein phosphatase 2C (PP2C), the dehydrin RD29B, an HD-Zip I 167 homologue of AtHB7 and AtHB12, and the trehalose-6-phosphate synthase TPS11 (Suppl. 168 Table S5). The accessibility response is therefore largely stress-specific, but drought, heat– 169 drought, and waterlogging converge on a shared set of ABA-responsive regulatory regions. 170 171 Given their different breeding histories, we compared the chromatin landscapes of the two 172 cultivars and their responses to stress. Both had a similar number of accessible regions 173 (127,196 in Désirée and 124,383 in Lady Rosetta; Suppl. Table S1), but the accessibility of 174 these regions differed, separating cultivars more strongly than stress treatments (Fig. 1B). In 175 response to stress, the cultivar differences were most pronounced in heat–drought and much 176 smaller in drought and heat, while waterlogging reshaped almost the same regions in both 177 (Suppl. Table S2). 178 179 We next compared those chromatin changes with the transcriptional response profiled in the 180 same experiment. Among genes that responded on both molecular layers, promoters tended 181 to open at induced genes and close at repressed genes consistently across cultivars and 182 stress treatments (Fig. 1F). However, the magnitude of accessibility changes correlated only 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 183 moderately with the expression changes observed across the genes with a detected promoter 184 peak (Suppl. Fig. S4A), and differentially accessible promoters was accompanied with fewer 185 than one in five of the differentially expressed genes with a promoter peak (Suppl. Fig. S4B). 186 Accessibility changes therefore indicate the direction, but not the extent, of the transcriptional 187 response. 188 189 **Abiotic stress reshapes chromatin accessibility across photosynthesis, growth and** 190 **defence related processes** 191 192 **Figure 2.** **Process level integration analysis of chromatin accessibility and gene expression. A)** 193 Enrichment of MapMan gene sets across all comparisons. Filled and open triangles indicate enrichment 194 in the accessibility and expression datasets, respectively; triangle orientation indicates increased or 195 decreased accessibility or expression, and triangle size represents absolute value of the normalized 196 enrichment score (NES). Opacity of triangles indicates q value: <0.05, 0.05–0.10, and >0.10 correspond 197 to 1, 0.5, and 0.22, respectively. Selected MapMan gene sets are shown, grouped by process and 198 separated by horizontal lines; all 45 gene sets are shown in Suppl. Fig. S5. Full results of enrichment 199 analysis are available in Suppl. Table S3. In parenthesis: number of genes in the gene set for 200 accessibility (A) and expression (R) datasets. **B)** Differential activity of transcription factor motifs (i.e. 201 relative change in chromatin accessibility of ACRs containing the motif) across all comparisons. Motifs 202 were obtained from the JASPAR 2026 plants CORE collection. Cell colour indicates differences in motif 203 activity at FDR < 0.05 (orange – higher activity, violet – lower activity); grey cells are not significant. 204 Rows are ordered according to an average-linkage hierarchical clustering based on 1 − r, where r is the 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 205 Pearson correlation coefficient derived from the RSAT pairwise motif-similarity matrix. Coloured boxes 206 (left column) indicate transcription factor families; grey was used for small families. 207 208 Having established that accessibility changes track the direction of transcriptional responses, 209 we next asked which biological processes these changes affect. Stress reduced accessibility 210 of photosynthetic, plastid and growth-related genes, and increased accessibility of defence211 associated pathways (Fig. 2A, Suppl. Fig. S5, Suppl. Table S3). Under heat and waterlogging, 212 promoter accessibility of photosynthesis genes decreased strongly and coherently in both 213 cultivars, across most photosynthetic subcategories, and gene expression followed the same 214 trend (Fig. 2A, Suppl. Table S3). Under drought, the cultivars responded in opposite directions: 215 in Désirée, both accessibility and expression of photosynthesis genes increased strongly, 216 whereas in Lady Rosetta both decreased, with significant accessibility changes seen only for 217 photosynthesis as a whole, photophosphorylation and for chlororespiration. Under heat– 218 drought, accessibility of photosynthesis-related genes did not change significantly in either 219 cultivar, apart from a decrease in photophosphorylation and photosystem II-related genes in 220 Lady Rosetta, although their expression decreased, less than under heat or waterlogging. 221 Under waterlogging, the response in both cultivars extended beyond photosynthesis to 222 broader plastid-related processes, including chlorophyll metabolism, plastid gene expression 223 and organellar translation, together with reduced accessibility of chloroplast protein import 224 pathways and, in Désirée under waterlogging, of mitochondrial complex I (Suppl. Table S3). 225 Additionally, the enrichments indicate reduced accessibility across several components of the 226 cell-growth machinery, spanning signalling, cytoskeletal organisation, cell-wall organisation 227 and cell division (Fig. 2A, Suppl. Fig. S5). The cell-division response involved both cyclins and 228 cell-cycle inhibitors, indicating reduced accessibility of genes involved in the cell-cycle 229 biological process, as a whole. 230 231 In contrast, defence-associated pathways gained promoter accessibility, most prominently in 232 waterlogging (Fig. 2A). In both cultivars, waterlogging increased accessibility of the pathogen, 233 effector-triggered immunity (ETI), TNL-mediated ETI and defence-mechanism pathways, 234 accompanied by increased transcript abundance. This agrees with flooding- and hypoxia235 induced changes in the transcriptomes, in which WRKY transcription factors and innate236 immunity genes are induced (Hsu et al., 2013; Tamura & Bono, 2022). Waterlogging also 237 increased accessibility of G-type lectin receptor-like kinases, which were concentrated in 238 tandem genomic clusters. 239 240 Transcription-factor families also changed in accessibility, mostly in a stress-specific way (Fig. 241 2A). Under waterlogging, WRKY genes gained accessibility in both cultivars, NAC genes in 242 Désirée, and ZAT genes in Désirée and, although weaker, also in Lady Rosetta (q < 0.10). 243 TCP genes showed a depletion that was consistent in direction across most comparisons but 244 significant only in Lady Rosetta under drought, and under waterlogging. BBX genes lost 245 accessibility in Lady Rosetta under waterlogging, and GATA genes, to a lesser extent, in 246 Désirée under drought and waterlogging (q < 0.10). bZIP genes changed little (q < 0.10 only 247 in Lady Rosetta, under heat and waterlogging; Fig. 2A). Inferred motif activity in ACRs 248 supported the changes in WRKY, NAC, TCP and GATA, whose motifs changed in the same 249 direction in the corresponding stresses (Fig. 2B), but not in ZAT, whose C2H2 zinc-finger 250 motifs did not gain activity under waterlogging. BBX could not be assessed because this family 251 is not represented in the motif library. 252 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 253 Motif activity analysis additionally identified a strong gain of G-box/ABRE regulatory activity 254 under drought and heat–drought, which was not captured among the leading GSEA categories 255 (Fig. 2B, Suppl. Table S4). The ABRE-type bZIP archetype showed the strongest gain in motif 256 activity under water deficit. (Suppl. Fig. S6). This motif was highly enriched in both drought 257 and heat–drought, particularly in Lady Rosetta, but showed little or no response to heat or 258 waterlogging. The ABRE response therefore represents a drought-specific chromatin 259 signature rather than a general response to abiotic stress. We also checked for known heat 260 response transcriptional regulation. Indeed, HSF motif enrichment was observed following 261 heat in Désirée, but not in Lady Rosetta (Fig. 2B, Suppl. Table S4). 262 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 263 **Combined heat and drought stress induces a unique regulatory response** 264 265 **Figure 3.** Combined heat and drought stress induces a largely unique transcriptional and chromatin 266 accessibility response. **A)** Chromatin accessibility (top, one point per peak, n = 56,724) and transcript 267 (bottom, one point per gene, n = 22,134) responses to combined heat and drought (HD), plotted against 268 the individual heat (H) or drought (D) response in Désirée and Lady Rosetta. Grey points respond to 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 269 neither stress (responsive = FDR < 0.05 and |log ₂ FC| > 0.5 for peaks, |log ₂ FC| > 1 for transcripts). 270 Coloured points are classed by how they respond under heat–drought compared with the individual 271 stress: _emergent_ (responsive under heat–drought only), _blunted_ (responsive to the individual stress 272 only), _redirected_ (opposite direction) or, when the direction is the same, _amplified_, _retained_ or _attenuated_ 273 (the heat–drought log ₂ FC is more than 1.2 times, 0.8 to 1.2 times, or less than 0.8 times the individual274 stress log ₂ FC; dashed lines). The bars below the scatter plots give the number of genes in each class 275 and are the colour key for both rows. Pearson r over all peaks or genes is shown in each plot. **B)** Venn 276 diagram showing overlap of differentially expressed genes (FDR < 0.05 and |log 2 FC| > 1) in individual 277 heat and drought, and combined heat–drought stress in cv. Désirée (top) and cv. Lady Rosetta (bottom). 278 Arrows show upregulated (up), downregulated, and disagreeing (up, down) gene overlaps. **C)** 279 Expression and chromatin accessibility responses of 15 drought-induced genes, grouped as ABA 280 signalling and transcriptional regulation and dehydration-protective proteins. Gene short names are 281 those of the Arabidopsis orthologues. Beginning of the arrow indicates responses in drought, end of the 282 arrow indicates responses in heat–drought; filled symbols, FDR < 0.05; open symbols, not significant; 283 'no peak', no tested accessible region. **D)** The _StABF2_ upstream region **.** Top: mean accessibility in 284 control, drought and heat–drought (ATAC-seq coverage normalized to genome-wide depth, where 1 is 285 the genome average; square-root colour scale, capped at 120). Bottom: accessibility log ₂ FC relative to 286 control under drought (open) and heat–drought (filled) at each tested region; asterisks, FDR < 0.05 for 287 the fitted heat–drought versus drought contrast. ATAC-seq, assay for transposase-accessible 288 chromatin with sequencing; FDR, false discovery rate; TSS, transcription start site. 289 290 Combined stresses can elicit responses that are not predictable from the single stresses, but 291 whether this extends to chromatin organization is unknown. We therefore asked how the 292 combined heat–drought response relates to those of heat and drought alone. At the genome293 wide level, heat–drought accessibility changes correlated more strongly with drought than with 294 heat in both cultivars (Fig. 1E, Fig. 3A), and the same held true for expression changes (Fig. 295 3A). Rather than a modified drought response, heat–drought elicited a largely unique 296 response. In the transcriptome, 2,327 genes in Désirée and 1,618 in Lady Rosetta responded 297 specifically to heat–drought, without significant responses to either individual stress, forming 298 the largest treatment-specific gene set in both cultivars (Fig. 3B). Comparison with the 299 individual-stress responses further categorized genes according to how their responses were 300 altered under heat–drought (Fig. 3A). Emergent responses were the largest class in both 301 cultivars. In addition, many drought-responsive genes kept their direction under heat–drought 302 but were amplified or attenuated. 303 304 Functional enrichment supported this interpretation (Suppl. Fig. S7). Protein quality control 305 was the most significant and most consistently intensified response under heat–drought in 306 both cultivars, with the chaperone response approximately doubling in heat–drought 307 compared to drought or heat (Suppl. Fig. S7). This increase was associated with stronger 308 induction of representative HSP90, HSP101/HSA32/CLPB3 and HSP70/HSP40 chaperone 309 components, consistent with an increased protein-folding burden under combined stress. 310 Interestingly, the changes in chromatin accessibility were mild for _HSP_ genes, albeit the HSF 311 motifs activity was increased. Only individual _StHSP90_ and _StHSP70_ genes had increased 312 chromatin accessibility in heat and drought, respectively (Suppl. Table S5). 313 314 In contrast, ABA signalling and dehydration-protective genes were more prominent among 315 drought-induced genes whose response was attenuated or blunted under heat–drought, 316 including potato orthologues of Arabidopsis _ABF2_, _AFP1_, _PP2CA_, _MYB96_ and _RD29B_ (Fig. 317 3C). Changes in their chromatin accessibility were smaller and less consistent under 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 318 combined stress. At the _StABF2_ locus, drought-induced accessibility at distal upstream 319 regions was selectively reduced under heat–drought. Accessibility 1–5 kb upstream of the 320 TSS decreased in the same direction in both cultivars, significantly at all six upstream regions 321 in Désirée, whereas the TSS-proximal region changed only slightly (Fig. 3D). Thus, the 322 reduced StABF2 expression response under heat–drought is associated with attenuation of 323 accessibility mainly at upstream regions rather than at the TSS-proximal region. Independent 324 evidence from cold-acclimated _Solanum commersonii_ further supports StABF2 as an abiotic325 stress regulator (X. Li et al., 2024). We therefore prioritised StABF2 as the leading candidate 326 associated with the water-deficit-induced gain in ABRE accessibility. 327 328 **Abscisic acid synthesis is in a heat–drought stress primed state brought by** 329 **StWRKY33b** 330 We next asked whether chromatin accessibility contributes to stress memory, helping the plant 331 to respond more effectively to a recurring cycle of adverse conditions. We performed an 332 additional independent experiment on the more responsive cv. Désirée, exposing plants to 333 heat, drought and heat–drought for 4 days, then the drought-exposed plants were re-watered 334 and exposed to the second round of drought for 4 days (Fig. 4A). As heat alone did not leave 335 a substantial mark in chromatin accessibility (Fig. 1C, D, Suppl. Fig. S3), we focused on 336 elucidating drought- and heat–drought-inflicted priming (sample description in Suppl. Table 337 S6; soil moisture and plant physiology in Suppl. methods and results, Suppl. Tables S7, S8 338 and Suppl. Fig. S8, S9). We performed transcriptomics analysis at the beginning and end of 339 each stress cycle. 340 341 Because drought alone elicited only a weak transcriptional response in this experiment (Suppl. 342 Table S11), we restricted the memory analysis to the combined heat–drought treatment only. 343 The transcriptional responses of first stress cycle correlated well with our initial experiment 344 (Spearman rho = 0.56, 86% match in direction), and re-watering produced the expected switch 345 from proline synthesis to proline breakdown (Yoshiba et al., 1997). We classified genes by 346 their response to the first drought and its change in the second (Suppl. Table S9), following 347 established classification scheme for transcriptional memory response-patterns (Ding et al., 348 2013). Genes repressed more strongly in the second drought ([-/-]) were enriched in 349 photosynthesis, Calvin cycle, and starch biosynthesis, while genes induced more strongly 350 ([+/+]) were enriched in vesicle trafficking and autophagy, Ca² ⁺ signalling, ethylene action, and 351 cell wall modification, among others (Suppl. Table S10). Genes repressed in the first drought 352 whose response rose in the second ([-/+]) were enriched in NAC and WRKY transcription 353 factors and pathogen response, including salicylic acid immunity regulators, and in further 354 Ca² ⁺ signalling components beyond those amplified in the [+/+] class. The ABA signalling and 355 dehydration response, including _StABF2_, was induced to a similar extent in both cycles ([+/=]). 356 While ABA signalling carried no memory response, ABA metabolism did: the rate-limiting 357 biosynthetic enzyme _StNCED3_ and the ABA 8'-hydroxylase _StCYP707A1_ were classified as 358 [+/+] indicating that ABA metabolism is primed for stronger response in the second stress 359 occurrence. 360 361 Stronger re-induction has been attributed to chromatin that remains accessible after the first 362 stress (Kappel et al., 2023), but has only recently been tested genome-wide using ATAC-seq 363 (Z.-W. Wang et al., 2026). We compared the [+/+] genes with [+/=] genes, considering 364 accessibility from our initial heat–drought experiment. On average, no greater changes in 365 chromatin accessibility were observed around memory genes than around any matched genes 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 366 in both cultivars. Nevertheless, memory genes contained significantly more accessible 367 regions, located upstream of their TSS and in the gene bodies (Fig. 4B, Suppl. Fig. S10). 368 _StNCED3_ illustrates this pattern, with thirteen accessible regions spanning this locus and its 369 TSS region ranking among the top 1.2 % most accessible regions across control and stress370 treated samples. Because accessibility changes varied across regions, enhanced re-induction 371 of _StNCED3_ may depend primarily on the availability of transcriptional regulators rather than 372 additional chromatin opening. To identify potential regulators, we performed a network 373 analysis using the Comprehensive Knowledge Network (CKN) (Bleker et al., 2024). A curated 374 list of potato genes involved in ABA biosynthesis (Suppl. Table S12) was used as seed nodes 375 to identification of their interacting partners and construct a transcriptional regulatory 376 subnetwork. This subnetwork was overlaid with transcriptomics data from the priming 377 experiment showing that _StNCED3_ was mildly upregulated in the first drought cycle, while 378 _StWRKY33b_ was not regulated (Fig. 4C). Regulation of both genes was enhanced in the 379 second round of drought (Fig. 4C, left vs right panel) indicating StWRKY33b may contribute 380 to priming of potato plants. 381 382 To test the hypothesis that StWRKY33b regulates StNCED3, we first examined chromatin 383 accessibility at both loci. Accessibility of the _StWRKY33b_ promoter increased slightly, but 384 changes were not statistically significant (Suppl. Table S5). Three accessible regions were 385 detected within the 4.5 kb _StNCED3_ promoter, none of which changed significantly in either 386 treatment; the strongest change at the locus was a decrease in accessibility 8.7 kb upstream 387 (Fig. 4D, Suppl. Table S5). A binding motif for WRKY transcription factors was identified in the 388 TSS-proximal accessible region (Fig. 4D). To test whether the enhanced induction of 389 _StNCED3_ upon recurring heat–drought stress could be driven by StWRKY33b, we performed 390 transactivation studies in _Nicotiana benthamiana_ . A strong activation of _StNCED3_ gene 391 expression by StWRKY33b was confirmed in these assays, which can explain the plant's 392 response to reoccurring stress (Fig. 4E, Suppl. Table S13). 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 393 394 **Figure 4** . **Priming of plants for additional cycles of drought and heat stress through ABA** 395 **synthesis is modulated by StWRKY33b.** **A)** Experimental design of priming experiment: 18 °C, 28 °C: 396 temperature at which plants were grown; blue drop: time point of rewatering of drought- and heat and 397 drought-stressed plants; black dots: tissue sampling points. **B)** Chromatin accessibility of drought398 memory genes in control Désirée plants. Memory genes (induced in the first drought and more strongly 399 in the second) are compared with non-memory genes induced equally in both droughts. Left: number 400 of accessible regions per gene (test in Suppl. Fig. S10). Right: mean accessibility in control plants at 401 five positions: distal upstream (3–10 kb upstream of the TSS), promoter (0.5–3 kb upstream), TSS402 proximal (TSS ± 500 bp), gene body and 3′ flank (1 kb downstream of the gene end). Accessibility is 403 expressed as the ATAC-seq coverage scaled to the mean coverage of all TSS regions (log ₂ scale). 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 404 Asterisks, Wilcoxon rank-sum test, Holm-adjusted (*, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not 405 significant). **C)** Network analysis of ABA synthesis genes: First neighbours of isoprenoid/ABA synthesis 406 genes (seed list, dark black border) were extracted from a Comprehensive knowledge network (Bleker 407 et al., 2024). The extracted network was further reduced to retain only transcriptional regulation edges 408 to identify the transcription factors involved in regulation of ABA synthesis and gene expression data 409 were mapped onto the network. Node fill colour denotes expression log ₂ fold changes (log ₂ FC) in 410 drought and heat-stressed vs control plants (HD1, left panel) and in re-droughted (after recovery) heat411 stressed vs control plants (HD2, right panel). Red – higher expression in stress compared to control, 412 blue – lower expression in stress compared to control. Arrows indicate direction of interaction. For full 413 names and IDs of all genes in the network see Suppl. Table S12. **D)** Accessibility of _StNCED3_ upstream 414 region. Columns show accessible regions. Top two boxes: mean chromatin accessibility across 415 treatments measured as ATAC-seq coverage normalized to genome-wide depth as reads per genomic 416 content (RPGC). Bottom box: positions of bZIP (blue), bHLH (amber), MYB (orange) and WRKY (pale 417 yellow) transcription factor binding motifs. **E)** Transcriptional regulation of _StNCED3_ by StWRKY33b 418 was confirmed _in planta_ by transactivation assay in _Nicotiana benthamiana_ . The luciferase activity of 419 pStNCED3::iLUC was observed through relative luminescence units (RLU) measured at 30-minute 420 intervals. The pStNCED3::iLUC construct alone served as control of basal activity. Median values with 421 standard error of median for 16 replicates are reported. Grey bands mark the night period. The x-axis 422 represents elapsed time (hh). Promoter region in the construct spanned -4.4kb to 0.2 kb relative to TSS. 423 424 **DISCUSSION** 425 426 Responses of plants to combined and recurrent stresses cannot be predicted from single427 stress responses, and their regulatory basis, including the contribution of chromatin 428 accessibility, remains largely unknown, particularly in crops such as potato. Our integrated 429 chromatin accessibility and transcriptome analysis reveals shared and stress-specific 430 responses to heat, drought and waterlogging and their combination, and shows how a 431 previous stress exposure primes ABA biosynthesis through StWRKY33b. 432 433 Approximately two thirds of the stress-responsive chromatin regions were specific to a single 434 stress (Fig. 1D). Drought, heat–drought and waterlogging shared a substantial set of 435 differentially accessible regions, with regions exhibiting increased accessibility under all three 436 conditions associated with canonical ABA- and dehydration-responsive genes (Fig. 1D; Suppl. 437 Table S5). This shared response signature is consistent with our previous physiology and 438 hormone profiling studies of Désirée, where waterlogging activated ABA responses similar to 439 those induced by drought (Zagorščak et al., 2025). Leaves of waterlogged plants experience 440 water deficit, because oxygen deprivation reduces root water uptake through aquaporins 441 closure (Tournaire-Roux et al., 2003). Our data extend these observations by showing that 442 drought and waterlogging also share regulatory responses at the chromatin level. The overlap 443 was nevertheless partial. The ABRE motif, recognized by group A bZIP transcription factors, 444 gained accessibility genome-wide under drought and heat–drought, but only modestly under 445 waterlogging, where WRKY and NAC motifs predominated (Fig. 2B; Suppl. Fig. S6). 446 Waterlogging thus increases accessibility of subset of ABA-responsive genes, including 447 _StABF2_ itself, without triggering the widespread increase in ABRE accessibility seen under 448 drought. 449 450 Heat had the weakest effect on chromatin accessibility in Désirée, and heat-treated Lady 451 Rosetta plants were barely distinguishable from controls (Fig. 1C; Suppl. Fig. S3). We did not 452 observe the canonical heat-shock signature reported in Arabidopsis, where acute heat shock 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 453 alters accessibility at thousands of regulatory elements for up to two days (Sullivan et al., 454 2014). However, these differences may reflect the experimental conditions. Arabidopsis 455 studies were done on young (7 day-old) seedlings exposed to severe heat shock, whereas 456 our experiments were carried out on fully grown (4-week-old) plants subjected to prolonged, 457 moderate heat stress. Our findings are consistent with the limited changes in DNA methylation 458 reported in two other potato cultivars exposed to prolonged, albeit somewhat more severe, 459 heat stress at a similar developmental stage (Yeo et al., 2026). Nevertheless, strong 460 transcriptional induction of selected heat shock protein ( _HSP_ ) genes was observed after 461 prolonged heat stress exposure, and HSF binding motifs were identified in their regulatory 462 regions (Fig. 2B, Suppl. Table S4). 463 464 Genome-wide changes in chromatin accessibility and gene expression under combined heat– 465 drought stress more closely resembled those induced by drought than by heat, particularly in 466 Lady Rosetta (Fig. 1E, Fig. 3A). Nonetheless, genes exclusively responding to combined 467 heat–drought stress represented the largest response class (Fig. 3B). Thus, despite the 468 stronger influence of drought, combined stress elicited a largely unique transcriptional 469 response, in agreement with previous combined heat and drought stress studies in potato, 470 tobacco and Arabidopsis (Rizhsky et al., 2002, 2004; Zagorščak et al., 2025). 471 472 The sequential stress treatment design may partly explain why the combined heat–drought 473 response more closely resembled drought than heat. Plants were exposed to heat before 474 drought was imposed, which might have already affected their response. Combined stress 475 enhanced the expression of genes involved in protein quality control. While changes in 476 chromatin accessibility were restricted to specific genes, transcriptional induction of 477 chaperones, including _StHSP90,_ roughly doubled under heat–drought compared to drought 478 (Suppl. Fig. S7). This enhanced response may reflect higher leaf temperature: drought479 induced stomatal closure limits transpiration cooling, causing leaves to become warmer than 480 under heat alone (Rizhsky et al., 2002; Zagorščak et al., 2025). In contrast, the ABF group of 481 transcriptional regulators, which act as key regulators downstream of ABA signalling, showed 482 weaker transcriptional induction under heat–drought than under drought alone. Chromatin 483 accessibility at these genes was also reduced in Désirée but not in Lady Rosetta (Fig. 3C). At 484 the _StABF2_ locus, this reduction under heat–drought occurred mainly at distal regions 1–5 kb 485 upstream of the TSS, whereas accessibility at the proximal promoter region changed only 486 slightly (Fig. 3D). Combined stress may therefore alter _StABF2_ regulation through distal 487 regulatory regions rather than through changes in the TSS-proximal promoter. Interestingly, 488 most genes showing combined stress-specific changes in chromatin accessibility in both 489 cultivars were linked to stress signalling, including Ca [2+] signalling, proteasome-mediated 490 protein degradation, and hormonal signalling pathways. Approximately one third of these 491 genes, however, remain functionally uncharacterized (Suppl. Table S5). 492 493 Within an agricultural context, crops frequently experience multiple environmental stresses, 494 either individually or in combination, with repeated cycles of stress and recovery throughout 495 the growing season. We therefore designed a stress-priming experiment to identify novel 496 regulators of heat and drought that provided a stress memory which may enhance responses 497 to recurrent stress. Stronger transcriptional induction following a previous stress has been 498 commonly attributed to sustained chromatin accessibility (Kappel et al., 2023), or the activity 499 of specific transcriptional factors (Crawford et al., 2026; Jiang et al., 2025; Wei et al., 2026). 500 Our data suggest that in potato leaves exposed to prolonged combined stress, pre-existing 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 501 chromatin accessibility, rather than stress-induced chromatin opening, identifies genes 502 exhibiting transcriptional memory. This is consistent with a recent study showing that stress503 responsive regulatory elements are concentrated in constitutively accessible TSS-proximal 504 regions (Morales-Cruz et al., 2026). Thus, pre-existing chromatin accessibility may provide a 505 permissive regulatory state facilitating transcriptional memory upon subsequent stress 506 exposure. 507 508 _StNCED3,_ which encodes an enzyme catalysing the rate-limiting step of ABA synthesis (Vo 509 et al., 2025), supports this model. Its promoter remains constitutively accessible, suggesting 510 that its enhanced re-induction depends primarily on its transcriptional regulators rather than 511 changes in chromatin. By integrating a regulatory knowledge network with the gene 512 expression data, we identified AtWRKY33 as a key regulator of _AtNCED3_ (Fig. 4C). 513 AtWRKY33 was shown to transcriptionally repress _AtNCED3_ in _Arabidopsis_ (S. Liu et al., 514 2015). In potato, we identified two _AtWRKY33_ orthologs, _StWRKY33a_ and _StWRKY33b_ 515 (Suppl. Fig. S11), while only _StWRKY33b_ exhibited enhanced transcriptional induction upon 516 recurrent heat–drought stress in our study (Fig. 4C). Our transactivation assays further 517 demonstrated that StWRKY33b directly activates _StNCED3_ transcription (Fig. 4E), thereby 518 placing stress memory on the upstream regulators of _StNCED3_ . The activating rather than 519 repressive function of StWRKY33b may reflect a functional divergence from AtWRKY33 or a 520 context-dependent switch in its repressor/activator role, as already demonstrated for 521 AtWRKY33 (S. Liu et al., 2015). 522 523 StWRKY33b regulation of heat and drought responses is particularly relevant for a heat524 sensitive crop such as potato (Guillemette et al., 2024), as concurrent heat and drought events 525 are becoming more frequent and severe. Our findings show that prolonged abiotic stress 526 leaves a lasting, largely stress-specific imprint on the chromatin landscape of potato, and that 527 the imprint of combined heat and drought stresses is very different compared to those of single 528 stresses. This imprint allows appropriate response to prolonged stress and encodes a memory 529 of past exposure that primes plants for stress reoccurrence. As heat and drought cycles 530 become increasingly more common within a single growing season, such stress memory may 531 become a decisive determinant of future potato crop performance. Resolving how it is 532 established and maintained will open new routes to breeding crops that are resilient not only 533 to individual stresses but to the compound and reoccurring stress regimes of a changing 534 climate. 535 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 536 **METHODS** 537 538 **Plant growth** 539 Plant growth and stress application followed the protocol detailed in Abdelhakim et al. 540 (Abdelhakim et al., 2024). Potato tissue culture plantlets (cvs. Désirée and Lady Rosetta) were 541 propagated in 90 mm petri dishes containing MS medium (Murashige & Skoog, 1962) 542 supplemented with 20 g/l sucrose and 8 g/l agar at 18 °C ± 4 °C, 16 h light/ 8 h dark, and light 543 intensity 100 µmol m [-] ² s [-] ¹. Four weeks after sub-culture, _in vitro_ plantlets were transferred to 544 250 ml pots containing Klasmann Substrate 2 and kept in a climate-controlled chamber under 545 low light for 1 week. After 10 days, plants were transplanted into 3L pots containing 3:1 546 Klasmann Substrate 2: sand and grown in a growth chamber at 320 µmol m ⁻ ² s ⁻ ¹ (16 h 547 photoperiod) and 22/19°C day/night. Stress treatments were imposed at the early tuberisation 548 stage, 28 days after transplanting the in vitro cuttings. Relative humidity was maintained at 549 55% throughout the experiment, including in the heat-treatment chamber. 550 551 For the initial abiotic stress experiment, plants were exposed to three individual stresses: heat 552 (30/28 °C), drought (20% soil relative water content, SRWC), and waterlogging (160% 553 SRWC), and a combined heat–drought treatment. Control plants were maintained at 22/19 °C 554 and 60% SRWC. 555 556 For the priming experiment, plants were maintained either at 18 or 28 °C. At each temperature, 557 plants were subjected to 4 days of drought or were watered once daily to maintain the field 558 capacity (Suppl. methods and results, Suppl. Fig. S8). After four days of treatment, all plants 559 were sampled, and the drought-treated plants were then watered to field capacity for one day 560 to allow recovery. Plants were sampled again after one day of recovery and then either 561 subjected to a second period of 4 days of drought or watered daily to field capacity prior to the 562 final harvest. Details on sample treatment groups and experimental design are given in Suppl. 563 Table S6. Soil moisture, relative water content, and photosynthetic and stomatal parameters 564 were measured as described in Suppl. methods and results. 565 566 **ATAC-seq sampling and library preparation** 567 At harvest, the second leaf of each plant was collected. Nuclei were isolated straight away 568 from fresh tissue. Plants were processed in batches of four, with treatment groups alternated 569 within each batch. For each sample, approximately 150 mg of leaf tissue with the midvein 570 removed was chopped on ice for 1.5 min (200–250 strokes min ⁻ ¹) with a fresh double-edged 571 razor blade. Chopping was done in 500 µl of ice-cold extraction buffer: 0.25 M sucrose, 10 572 mM Tris-HCl pH 8.0, 10 mM MgCl ₂, 0.5% Triton X-100, 5 mM β-mercaptoethanol and protease 573 inhibitor (Roche). The homogenate was rinsed with 500 µl of buffer and filtered through 100 574 µm and then 40 µm cell strainers (pluriStrainer Mini, pluriSelect). A low-speed spin (50 × g, 5 575 min, 4 °C) removed large debris. Nuclei were then pelleted at 300 × g for 5 min at 4 °C and 576 washed twice in 1 ml of extraction buffer. Tagmentation and library preparation followed the 577 instructions for the ATAC-seq Kit (Diagenode). Briefly, washed nuclei were resuspended in 50 578 µl of tagmentation mix containing 2.5 µl of loaded Tagmentase (Tn5), 0.01% digitonin and 579 0.1% Tween-20, then incubated for 30 min at 37 °C. The reaction was stopped with 300 µl of 580 Buffer ERC. Tagmented DNA was purified with the MinElute Reaction Cleanup Kit (Qiagen) 581 and eluted in 10 µl of Buffer EB. 582 18 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 583 In the preamplification step, tagmented DNA was amplified by PCR with indexed primers 584 (Diagenode) and Phusion High-Fidelity DNA Polymerase (NEB), with an extension time of 30 585 s. To determine the number of additional amplification cycles required, qPCR was performed 586 with an aliquot of first-round PCR product, Phusion High-Fidelity DNA Polymerase (NEB) and 587 1× SYBR Green (Invitrogen). The quantification cycle value corresponding to one-third of the 588 maximum fluorescence was used to determine the number of additional PCR cycles applied 589 to pre-amplified libraries. The pre-amplified libraries were then subjected to the additional PCR 590 cycles. Product size distributions were assessed on a TapeStation system with the D5000 591 ScreenTape Assay (Agilent Technologies). Final PCR products were purified with 1.8x 592 AMPure XP beads (Beckman Coulter). DNA concentrations were measured on a Qubit 593 fluorometer with the Qubit dsDNA High Sensitivity Assay Kit (Thermo Fisher Scientific). 594 595 **ATAC-seq data analysis** 596 Raw reads were quality-checked with FastQC v0.11.9 (Andrews, 2010) and trimmed for Tn5 597 mosaic-end adapters using cutadapt v4.8 (Martin, 2011) in paired-end mode. Reads were 598 aligned to the DM v6.1 genome carrying the UniTato gene models (Zagorščak et al., 2024), 599 supplemented with the complete Désirée organellar genome, comprising the plastid (NCBI 600 DQ386163.2) and the three molecules of the multipartite mitochondrial genome (NCBI 601 MN104801.1, MN104802.1 and MN104803.1), using BWA-MEM v0.7.17-r1188 (H. Li, 2013). 602 Alignments were sorted and indexed with SAMtools v1.20 (Danecek et al., 2021), and PCR 603 duplicates were flagged with samblaster v0.1.26 (Faust & Hall, 2014). Organellar reads were 604 removed before downstream analysis. Unless otherwise stated, analyses used reads with 605 MAPQ ≥ 40 and excluded duplicate reads. 606 607 Dataset quality was assessed using SAMtools stats and deepTools v3.5.4 (Ramírez et al., 608 2016), including fingerprinting, sample correlation and principal component analysis, fraction 609 of reads in peaks (FRiP), and TSS enrichment. Coverage tracks were generated using 610 alignmentSieve (deepTools v3.5.4) with ATAC-seq strand shifts of +4/−5 bp, followed by 611 bamCoverage (deepTools v3.5.4) with 1-bp bins and RPGC normalisation using an effective 612 genome size of 600 Mb. FRiP was calculated relative to the consensus peak set. 613 614 Two libraries were excluded: LR-0501-Dc3, which yielded too few usable reads (0.16 million 615 reads in consensus regions, against at least 1.45 million in every other library; duplicate rate 616 82.1%; TSS enrichment 2.0), and De-0620-C4, whose library required 20 PCR amplification 617 cycles, against 12 or fewer for every other library. The final dataset comprised 54 of 56 618 libraries. Peaks were called separately for each cultivar and treatment group using Genrich 619 v0.6.1 (Gaspar, 2018) in paired-end mode with MAPQ ≥40, duplicate removal, and q ≤0.05. 620 All libraries except LR-0501-Dc3 were used for peak calling. A fixed-width consensus peak 621 universe was generated as non-overlapping 300-bp windows centred on Genrich summits, 622 retaining the higher-signal summit where windows overlapped. The window size was based 623 on the observed peak-width distributions (Suppl. Fig. S2A). Before the consensus was built, 624 peaks whose read counts across libraries tracked the library organellar read fraction (92 to 625 250 peaks per treatment group), indicating plastid-derived reads mapped to nuclear plastid626 like insertions, were removed. 627 628 Reads were quantified with featureCounts v2.0.6 (Liao et al., 2014) in paired-end mode using 629 a minimum fractional overlap of 0.2. Regions were annotated against the UniTato gene models 630 using ChIPseeker v1.48.0 (Yu et al., 2015) with a ±3-kb promoter window. Differential 19 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 631 accessibility analysis was performed in R v4.6.0 using edgeR v4.10.1 (Robinson et al., 2010) 632 _and limma v3.68.4_ (Ritchie et al., 2015) _._ Regions were filtered with filterByExpr (min.count = 633 50). GC-content bias was removed with conditional quantile normalisation (cqn) (Hansen et 634 al., 2012), and the resulting offsets were supplied to voomLmFit (Law et al., 2014) using a 635 one-way design, followed by empirical Bayes moderation (eBayes) with robust estimation and 636 Benjamini–Hochberg correction (Benjamini & Hochberg, 1995). Differentially accessible 637 regions were defined as FDR <0.05 and |log2 fold change| >0.5. Per-cultivar, cultivar638 averaged, stress-by-cultivar, and direct combined-stress versus drought contrasts were 639 extracted from the same model. 640 641 Sample structure was assessed by multidimensional scaling of the 5,000 most variable 642 regions. Response magnitude was summarised from the distribution of absolute log2 fold 643 changes across tested regions, and similarity between contrasts was assessed by Pearson 644 correlation of log 2 fold changes. 645 646 For baseline accessibility analyses, a gene was considered to have an accessible promoter if 647 a consensus region lay within 1 kb or 3 kb of its TSS in at least one of the three control harvest 648 groups of that cultivar. Proportions were computed over all 57,394 annotated genes and over 649 the 22,134 genes retained by the expression filter of the RNA-seq analysis. 650 651 **RNA extraction and sequencing** 652 For the initial experiment RNA-seq analysis, the 3 [rd] leaf of same plants as for the ATAC-seq 653 analysis was sampled. Leaf tissue was homogenised under liquid nitrogen using a mortar and 654 pestle, and total RNA was extracted with MagMAX Plant RNA Isolation Kit (Thermo Fisher 655 Scientific) on a KingFisher Apex purification system (Thermo Fisher Scientific), including 656 DNase I treatment (Thermo Fisher Scientific). Reverse-stranded, polyA-enriched Illumina 657 RNA-seq libraries were prepared and sequenced in paired-end, 150 bp mode on an Illumina 658 NovaSeq X Plus (Novogene). Read adapter trimming was performed by the sequencing 659 service prior to data delivery. 660 661 For the priming experiment RNA-seq analysis, fully expanded leaves were harvested from the 662 fourth node of three independent replicate plants and immediately frozen in liquid nitrogen. 663 Samples were stored at −80 °C prior to RNA extraction. Total RNA was extracted from potato 664 leaves using RNeasy Plant Mini Kit (Qiagen) as described previously (Ducreux et al., 2008). 665 Reverse-stranded, polyA-enriched Illumina RNA-seq libraries were prepared and sequenced 666 in single-end, 100 bp mode on an Illumina HiSeq 2500 by IGA Technology Services (Udine, 667 Italy). 668 669 **Differential expression analysis** 670 For transcript quantification from both RNA-seq experiments, a decoy-aware index using the 671 potato _Solanum tuberosum_ Group Phureja DM genome (Pham et al., 2020) and the UniTato 672 transcriptome (Zagorščak et al., 2024) was built and reads were quantified with Salmon 673 v1.10.3 (Patro et al., 2017). Mapping statistics were aggregated with MultiQC (Ewels et al., 674 2016). Two samples from the initial experiment RNA-seq with outlying extremely low mapping 675 rates were discarded, leaving 246 out of total 248 samples. For the priming experiment, prior 676 to pseudomapping, adapter removal was performed using Trim Galore v2.2.0 given the 677 parameters “--quality 20 --polyA --trim-n”. 678 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 679 Differential expression analysis was performed in R v4.6.0 using edgeR v4.10.1 (Chen et al., 680 2025). Briefly, transcript-level abundances were imported and summarised to gene level 681 counts with tximport v1.40.0 (Soneson et al., 2016), lowly expressed genes were removed 682 with _filterByExpr_ (min.count = 50, min.total.count = 100), counts were TMM-normalised and 683 modelled with _voomLmFit_ (sample.weights = TRUE; Baldoni et al., 2025), and predefined 684 contrasts were tested with _eBayes_ using limma v3.68.4 (Ritchie et al., 2015). Differentially 685 expressed genes were defined as Benjamini–Hochberg-adjusted P <0.05 and, where applied, 686 |log 2 fold change| >1. For the priming experiment, drought-memory classes were assigned to 687 genes in the 28 °C treatment following an established scheme (Ding et al., 2013). Each gene 688 was classified as [HD1/HD2 − HD1], where HD1 is the response to the first drought and HD2 689 − HD1 is the fitted difference between the second and the first drought response, both relative 690 to the well-watered 28 °C control sampled at the same time point. Each term was scored as 691 significantly positive (+), significantly negative (-) or not significant (=) at FDR < 0.05. 692 693 **Integration of RNA-seq and ATAC-seq** 694 Accessibility and expression data were paired by cultivar, stress, and sampling day. Genes 695 were linked to ATAC-seq regions using the ChIPseeker annotation, and one representative 696 promoter region was retained per gene, selected by the lowest accessibility FDR and, for ties, 697 the largest absolute log 2 fold change. 698 699 For genes that were differentially expressed and differentially accessible in both assays (FDR 700 < 0.05), directional concordance was quantified as the proportion showing changes in the 701 same direction. Coupling between accessibility and expression changes was assessed across 702 all genes with a tested promoter peak and an RNA-seq result, without selection on significance 703 or direction, as the Pearson correlation (with 95% confidence interval) and the Spearman rank 704 correlation of the log2 fold changes, separately for each cultivar and stress. The overlap 705 between differentially expressed genes and genes with a differentially accessible promoter 706 was counted over gene-by-comparison tests, pooled across the four stresses and both 707 cultivars. 708 709 Combined-stress responses were compared with each constituent individual stress, 710 separately for transcripts (per gene) and for chromatin accessibility (per accessible region). A 711 gene or region was considered responsive at FDR < 0.05 and |log 2 FC| > 1 for transcripts or 712 |log 2 FC| > 0.5 for accessibility. Genes or regions responsive to either treatment were classified 713 as emergent when responsive only under combined stress, blunted when responsive only 714 under the individual stress, redirected when the direction of change differed, amplified or 715 attenuated when changes were in the same direction but the combined-to-individual log2 fold716 change ratio was >1.2 or <0.8, respectively, and retained otherwise. 717 718 **Gene set enrichment analyses** 719 Gene set enrichment analysis was performed using UniTato Mercator4 v7 (MapMan4) 720 functional bins (Zagorščak et al., 2024). For ATAC-seq, normalised counts of a single 721 promoter-associated region per gene with the highest mean consensus Genrich q-value was 722 used. For RNA-seq, gene-level normalised counts per million were used as input. 723 724 Enrichment was calculated using an R implementation of the Broad’s GSEA (Subramanian et 725 al., 2005) algorithm with signal-to-noise ranking, 1,000 gene-label permutations, gene-set 726 sizes of 15–500, and the unweighted Kolmogorov–Smirnov statistic. A fixed random seed was 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 727 used, and FDR q-values were adjusted separately for each direction of enrichment. Gene sets 728 with FDR <0.05 were considered enriched; FDR 0.05–0.10 was considered suggestive. For 729 display (Fig. 2A, Suppl. Fig. S5), bins with FDR <0.10 in the accessibility data in at least two 730 of the eight comparisons were selected, and nested bins were reduced to the most general 731 qualifying bin. 732 733 Because the ATAC-seq analysis was restricted to genes with annotated promoter regions, the 734 accessibility and expression analyses used different gene universes. Enrichment results were 735 therefore compared by direction rather than magnitude. 736 737 To compare drought and combined-stress responses, genes were ranked using the signed 738 statistic 739 × (log 2 FC HD −log 2 FC D ), sign ( [log] [2] [FC] [D] [ + log] 2 [2] [FC] [HD] ~~)~~ 740 741 where D and HD denote drought and combined heat and drought responses, respectively. 742 MapMan bins were tested using fgsea (Korotkevich et al., 2019) with gene-set sizes of 5–500 743 and 10,000 permutations. The gene universe comprised the union of drought- and combined744 stress-responsive genes, oriented according to their mean response. The same procedure 745 and gene sets were applied to the accessibility data using the fitted combined-stress versus 746 drought contrast at promoter-associated regions. 747 748 Over-representation of MapMan4 bins (levels 2 and 3, 15–600 genes per bin) in the [+/+], [+/-], 749 [-/-] and [-/+] drought-memory classes was tested against all genes changed in the same 750 direction in HD1, using a Cochran–Mantel–Haenszel test stratified by quartiles of absolute 751 HD1 log 2 FC, so that each class is compared with genes of similar first-drought response. Bins 752 with at least five genes in the class were tested, and FDR was controlled with the Benjamini– 753 Hochberg procedure within each class. 754 755 **Motif enrichment** 756 Motif sequences from JASPAR CORE Plants 2026 (Ovek Baydar et al., 2026) were scanned 757 against consensus peak sequences using FIMO (MEME Suite v5.0.5; Grant et al., 2011) with 758 P ≤1 × 10 ⁻ ⁴. The FIMO hits served as the motif annotation for chromVAR and mark the motif 759 positions in Fig. 4D. Redundant JASPAR motif models were grouped into motif archetypes by 760 cutting the precomputed RSAT matrix-clustering hierarchy of the JASPAR 2026 plants CORE 761 collection (Castro-Mondragon et al., 2017) at a normalised correlation (Ncor) of 0.70. 762 763 Differential motif activity was assessed with chromVAR v1.34.1 (Schep et al., 2017) using per764 region GC content as a bias term and 2,000 background iterations, each drawing a GC- and 765 accessibility-matched background region for every peak. Bias-corrected deviation z-scores 766 were quantile-normalised, and differential activity for stress-versus-control comparisons was 767 tested with limma v3.68.4 (Ritchie et al., 2015). A linear model was fitted separately for each 768 cultivar, stress-versus-control contrasts were extracted from it, and standard errors were 769 moderated with eBayes. P-values were Benjamini–Hochberg adjusted across motifs within 770 each cultivar and contrast. Motifs occurring in fewer than 10 regions were excluded. 771 772 **Network-based analyses** 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 773 A list of 72 potato genes involved in isoprenoid biosynthesis (seed genes) was compiled, and 774 their identifiers were unified to DM v6.1 gene identifiers using the UniTato translation table 775 (Zagorščak et al., 2024). Their Arabidopsis orthologues were then identified using PLAZA 5.0 776 (Van Bel et al., 2022), OrthoDB (Kuznetsov et al., 2023), and reciprocal BLAST hits (Benoit & 777 Drost, 2021). Pairs with inconsistent MapMan4 (Schwacke et al., 2019) annotation were 778 excluded, leaving 48 Arabidopsis seed genes. 779 780 Regulatory interactions of ranks 0–2 were taken from AtCKN v2 (Bleker et al., 2024). The first781 order neighbourhood of the seed genes was extracted in R v4.5.2 (R Core Team, 2026) using 782 igraph v2.3.3 (Antonov et al., 2023), and the largest weakly connected component was 783 retained. The network was restricted to transcription factor regulation, retaining edges from 784 transcription factors to seed genes and edges between these transcription factors. Using the 785 priming experiment RNA-seq dataset, each node was mapped to its most strongly differentially 786 expressed potato orthologue (highest absolute log 2 FC in the majority of comparisons) and 787 annotated with its expression changes. The network was visualised in Cytoscape 788 v3.9.1(Shannon et al., 2003). 789 790 **Multiple sequence alignment analysis** 791 To classify the identified novel regulators of ABA synthesis, we performed additional multiple 792 sequence alignment analysis. The orthology tables prepared for the network analyses 793 returned many-to-many mappings between AtWRKY33, its close paralogues (AtWRKY25 and 794 AtWRKY26) and multiple Solanaceae WRKYs. The potato and tomato orthologues of 795 AtWRKY33 were therefore determined by phylogenetic analysis of group I WRKY proteins, 796 which are defined by two WRKY domains, each consisting of a W[R/K]KYG[Q/K/E]K motif 797 followed by an intact C2H2 zinc finger (Eulgem et al., 2000). Group I WRKY proteins were 798 selected from _Arabidopsis thaliana_ Araport11 (Cheng et al., 2017), _Solanum tuberosum_ 799 UniTato (Zagorščak et al., 2024) and _Solanum lycopersicum_ ITAG5.0 (ITAG5.0; Y. Zhou et 800 al., 2022) annotations, using the longest isoform of each gene. Pairwise global alignments of 801 the full-length proteins were computed with pwalign v1.6.0 (Aboyoun & Gentleman, 2025) 802 (BLOSUM62, gap opening 10, gap extension 0.5), and distances were calculated as 1 − 803 sequence identity. A neighbour-joining tree was built with ape v5.8-1 (Paradis & Schliep, 804 2019), rooted on AtWRKY54 (group III), and divided into 10 clusters, equal to the number of 805 Arabidopsis sequences including the outgroup, by repeatedly splitting the deepest clade 806 starting from the root (Suppl. Fig. S11). AtWRKY33 clustered with two genes from each 807 Solanaceae species, corresponding in tomato to SlWRKY33a and SlWRKY33b (J. Zhou et 808 al., 2015). The potato orthologues were named StWRKY33a (Soltu.DM.09G009490) and 809 StWRKY33b (Soltu.DM.06G018840). Analyses were performed in R v4.5.2, and the 810 phylogenetic tree was drawn with ggtree v4.0.5 (Xu et al., 2022). 811 812 **Transient dual-luciferase reporter system** 813 The putative promoter region of the nine-cis-epoxycarotenoid dioxygenase 3 ( _StNCED3_ ; 814 Soltu.DM.07G022620) gene from the _Solanum tuberosum_ group Phureja DM1-3 v6.1 genome 815 was amplified utilizing forward primer: TGGTAGTGCAAATTTTATGTTATGTC and reverse 816 primer: TTTGTAGCATGTGAAGTAGTAGTGGC targeting 4.5 kb upstream of the TSS. The 817 resulting amplified promoter product was then cloned into the GATEWAY cloning vector 818 pDONR207, which was subsequently inserted into the pGWB435 destination vector. The 819 pGWB435 vector harboured a luciferase reporter gene, with incorporation of the potato intron 820 of the L700 gene using the LR clonase reaction (Invitrogen) to complete the construct. 23 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 821 822 Reporter construct _35S::StWRKY33b_ was cloned and transformed into _Agrobacterium_ 823 _tumefaciens_ strain GV3101, alongside the _pStNCED3::iLUC_ construct. The construct was co824 infiltrated into _Nicotiana benthamiana_ (rdr6i) leaves. The Agrobacterium strain carrying the 825 _pStNCED3-iLUC_ fusion in conjunction with the _35S::StWRKY33b_ construct was suspended in 826 a solution containing 10 mM MES (pH 5.8), 10 mM MgSO4, and 150 µM acetosyringone, with 827 an optical density of 0.2 at 600 nm for all cultures utilized in infiltration. 828 829 Two days post-infiltration, leaf discs were incubated in a plant medium composed of MS+2% 830 sucrose, supplemented with 0.02 mg/mL D-luciferin (Promega). Luciferase activity was 831 monitored every 30 minutes using a microplate luminometer Centro LB-960 (Berthold 832 Technologies), and the data were captured using MikroWin 2000 v4.29 software (Mikrotek 833 Laborsysteme). 834 24 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 835 **REFERENCES** 836 837 Abdelhakim, L. O. A., Pleskačová, B., Rodriguez-Granados, N. Y., Sasidharan, R., Perez838 Borroto, L. S., Sonnewald, S., Gruden, K., Vothknecht, U. C., Teige, M., & Panzarová, 839 K. (2024). High Throughput Image-Based Phenotyping for Determining Morphological 840 and Physiological Responses to Single and Combined Stresses in Potato. _Journal of_ 841 _Visualized Experiments_, (208), 66255. https://doi.org/10.3791/66255 842 Aboyoun, P., & Gentleman, R. (2025). _pwalign: Perform pairwise sequence alignments_ 843 (Version 1.6.0) [Computer software]. Bioconductor. 844 https://doi.org/10.18129/B9.bioc.pwalign 845 Andrews, S. (2010). _S-andrews/FastQC: A quality control analysis tool for high throughput_ 846 _sequencing data_ . https://github.com/s-andrews/FastQC 847 Antonov, M., Csárdi, G., Horvát, S., Müller, K., Nepusz, T., Noom, D., Salmon, M., Traag, V., 848 Welles, B. F., & Zanini, F. (2023). _Igraph enables fast and robust network analysis_ 849 _across_ _programming_ _languages_ (Version 1). arXiv. 850 https://doi.org/10.48550/ARXIV.2311.10260 851 Baldoni, P. L., Chen, L., Li, M., Chen, Y., & Smyth, G. K. (2025). Dividing out quantification 852 uncertainty enables assessment of differential transcript usage with limma and edgeR. 853 _Nucleic Acids Research_, _53_ (22), gkaf1305. https://doi.org/10.1093/nar/gkaf1305 854 Benjamini, Y., & Hochberg, Y. (1995). Controlling the False Discovery Rate: A Practical and 855 Powerful Approach to Multiple Testing. _Journal of the Royal Statistical Society Series_ 856 _B:_ _Statistical_ _Methodology_, _57_ (1), 289–300. https://doi.org/10.1111/j.2517 857 6161.1995.tb02031.x 858 Benoit, M., & Drost, H.-G. (2021). A Predictive Approach to Infer the Activity and Natural 859 Variation of Retrotransposon Families in Plants. In J. Cho (Ed.), _Plant Transposable_ 860 _Elements_ (Vol. 2250, pp. 1–14). Springer US. https://doi.org/10.1007/978-1-0716861 1134-0_1 862 Bleker, C., Ramšak, Ž., Bittner, A., Podpečan, V., Zagorščak, M., Wurzinger, B., Baebler, Š., 863 Petek, M., Križnik, M., Van Dieren, A., Gruber, J., Afjehi-Sadat, L., Weckwerth, W., 864 Županič, A., Teige, M., Vothknecht, U. C., & Gruden, K. (2024). Stress Knowledge 865 Map: A knowledge graph resource for systems biology analysis of plant stress 866 responses. _Plant_ _Communications_, _5_ (6), 100920. 867 https://doi.org/10.1016/j.xplc.2024.100920 868 Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y., & Greenleaf, W. J. (2013). 869 Transposition of native chromatin for fast and sensitive epigenomic profiling of open 870 chromatin, DNA-binding proteins and nucleosome position. _Nature Methods_, _10_ (12), 871 Article 12. https://doi.org/10.1038/nmeth.2688 872 Castro-Mondragon, J. A., Jaeger, S., Thieffry, D., Thomas-Chollier, M., & van Helden, J. 873 (2017). RSAT matrix-clustering: Dynamic exploration and redundancy reduction of 874 transcription factor binding motif collections. _Nucleic Acids Research_, _45_ (13), e119– 875 e119. https://doi.org/10.1093/nar/gkx314 876 Chen, Y., Chen, L., Lun, A. T. L., Baldoni, P. L., & Smyth, G. K. (2025). edgeR v4: Powerful 877 differential analysis of sequencing data with expanded functionality and improved 878 support for small counts and larger datasets. _Nucleic Acids Research_, _53_ (2), gkaf018. 879 https://doi.org/10.1093/nar/gkaf018 880 Cheng, C., Krishnakumar, V., Chan, A. P., Thibaud ‐ Nissen, F., Schobel, S., & Town, C. D. 881 (2017). Araport11: A complete reannotation of the _Arabidopsis thaliana_ reference 882 genome. _The Plant Journal_, _89_ (4), 789–804. https://doi.org/10.1111/tpj.13415 25 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 883 Crawford, T., Pratx, L., & Bäurle, I. (2026). Principles and mechanisms of plant acclimation to 884 heat stress. _Nature_ _Reviews_ _Molecular_ _Cell_ _Biology_, _27_ (9), 668–683. 885 https://doi.org/10.1038/s41580-026-00977-x 886 Danecek, P., Bonfield, J. K., Liddle, J., Marshall, J., Ohan, V., Pollard, M. O., Whitwham, A., 887 Keane, T., McCarthy, S. A., Davies, R. M., & Li, H. (2021). Twelve years of SAMtools 888 and BCFtools. _GigaScience_, _10_ (2), giab008. 889 https://doi.org/10.1093/gigascience/giab008 890 Devaux, A., Goffart, J.-P., Petsakos, A., Kromann, P., Gatto, M., Okello, J., Suarez, V., & 891 Hareau, G. (2020). Global Food Security, Contributions from Sustainable Potato Agri892 Food Systems. In H. Campos & O. Ortiz (Eds.), _The Potato Crop: Its Agricultural,_ 893 _Nutritional and Social Contribution to Humankind_ (pp. 3–35). Springer International 894 Publishing. https://doi.org/10.1007/978-3-030-28683-5_1 895 Ding, Y., Liu, N., Virlouvet, L., Riethoven, J.-J., Fromm, M., & Avramova, Z. (2013). Four 896 distinct types of dehydration stress memory genes in Arabidopsis thaliana. _BMC Plant_ 897 _Biology_, _13_ (1), 229. https://doi.org/10.1186/1471-2229-13-229 898 Ducreux, L. J. M., Morris, W. L., Prosser, I. M., Morris, J. A., Beale, M. H., Wright, F., 899 Shepherd, T., Bryan, G. J., Hedley, P. E., & Taylor, M. A. (2008). Expression profiling 900 of potato germplasm differentiated in quality traits leads to the identification of 901 candidate flavour and texture genes. _Journal of Experimental Botany_, _59_ (15), 4219– 902 4231. https://doi.org/10.1093/jxb/ern264 903 Eulgem, T., Rushton, P. J., Robatzek, S., & Somssich, I. E. (2000). The WRKY superfamily of 904 plant transcription factors. _Trends_ _in_ _Plant_ _Science_, _5_ (5), 199–206. 905 https://doi.org/10.1016/S1360-1385(00)01600-9 906 Ewels, P., Magnusson, M., Lundin, S., & Käller, M. (2016). MultiQC: summarize analysis 907 results for multiple tools and samples in a single report. _Bioinformatics_, _32_ (19), 3047– 908 3048. https://doi.org/10.1093/bioinformatics/btw354 909 Faust, G. G., & Hall, I. M. (2014). _SAMBLASTER_ : Fast duplicate marking and structural variant 910 read extraction. _Bioinformatics_, _30_ (17), 2503–2505. 911 https://doi.org/10.1093/bioinformatics/btu314 912 Gallusci, P., Agius, D. R., Moschou, P. N., Dobránszki, J., Kaiserli, E., & Martinelli, F. (2023). 913 Deep inside the epigenetic memories of stressed plants. _Trends in Plant Science_, 914 _28_ (2), 142–153. https://doi.org/10.1016/j.tplants.2022.09.004 915 Gaspar, J. M. (2018). _jsh58/Genrich:_ _Detecting_ _sites_ _of_ _genomic_ _enrichment_ . 916 https://github.com/jsh58/Genrich 917 Grant, C. E., Bailey, T. L., & Noble, W. S. (2011). FIMO: scanning for occurrences of a given 918 motif. _Bioinformatics_, _27_ (7), 1017–1018. https://doi.org/10.1093/bioinformatics/btr064 919 Guillemette, A. M., Hernández Casanova, G., Hamilton, J. P., Pokorná, E., Dobrev, P. I., 920 Motyka, V., Rashotte, A. M., & Leisner, C. P. (2024). The physiological and molecular 921 responses of potato tuberization to projected future elevated temperatures. _Plant_ 922 _Physiology_, _197_ (1), kiae664. https://doi.org/10.1093/plphys/kiae664 923 Guo, M., Yang, F., Zhu, L., Wang, L., Li, Z., Qi, Z., Fotopoulos, V., Yu, J., & Zhou, J. (2024). 924 Loss of cold tolerance is conferred by absence of the WRKY34 promoter fragment 925 during tomato evolution. _Nature_ _Communications_, _15_ (1), 6667. 926 https://doi.org/10.1038/s41467-024-51036-y 927 Hansen, K. D., Irizarry, R. A., & WU, Z. (2012). Removing technical variability in RNA-seq data 928 using conditional quantile normalization. _Biostatistics_, _13_ (2), 204–216. 929 https://doi.org/10.1093/biostatistics/kxr054 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 930 Hsu, F.-C., Chou, M.-Y., Chou, S.-J., Li, Y.-R., Peng, H.-P., & Shih, M.-C. (2013). 931 Submergence Confers Immunity Mediated by the WRKY22 Transcription Factor in 932 Arabidopsis. _The_ _Plant_ _Cell_, _25_ (7), 2699–2713. 933 https://doi.org/10.1105/tpc.113.114447 934 Jiang, Z., van Zanten, M., & Sasidharan, R. (2025). Mechanisms of plant acclimation to 935 multiple abiotic stresses. _Communications_ _Biology_, _8_ (1), 655. 936 https://doi.org/10.1038/s42003-025-08077-w 937 Jiang, Z., Verhoeven, A., Li, Y., Geertsma, R., Sasidharan, R., & Van Zanten, M. (2026). 938 Deciphering acclimation to sublethal combined and sequential abiotic stresses in 939 _Arabidopsis_ _thaliana_ . _Plant_ _Physiology_, _200_ (3), kiae581. 940 https://doi.org/10.1093/plphys/kiae581 941 Kappel, C., Friedrich, T., Oberkofler, V., Jiang, L., Crawford, T., Lenhard, M., & Bäurle, I. 942 (2023). Genomic and epigenomic determinants of heat stress-induced transcriptional 943 memory in Arabidopsis. _Genome Biology_, _24_ (1), 129. https://doi.org/10.1186/s13059 944 023-02970-5 945 Korotkevich, G., Sukhov, V., & Sergushichev, A. (2019). _Fast gene set enrichment analysis_ 946 (p. 060012). bioRxiv. https://doi.org/10.1101/060012 947 Kurtz, J., Andino, R., Boraschi, D., Contreras-Garduño, J., Kachroo, A., Khan, I., Lanz 948 Mendoza, H., Mukherjee, K., Peuß, R., & Ton, J. (2025). Trained immunity and immune 949 priming in plants and invertebrates. _eLife_, _14_, e106597. 950 https://doi.org/10.7554/eLife.106597 951 Kuznetsov, D., Tegenfeldt, F., Manni, M., Seppey, M., Berkeley, M., Kriventseva, E. V., & 952 Zdobnov, E. M. (2023). OrthoDB v11: Annotation of orthologs in the widest sampling 953 of organismal diversity. _Nucleic_ _Acids_ _Research_, _51_ (D1), D445–D451. 954 https://doi.org/10.1093/nar/gkac998 955 Law, C. W., Chen, Y., Shi, W., & Smyth, G. K. (2014). voom: Precision weights unlock linear 956 model analysis tools for RNA-seq read counts. _Genome Biology_, _15_ (2), R29. 957 https://doi.org/10.1186/gb-2014-15-2-r29 958 Li, H. (2013). _Aligning sequence reads, clone sequences and assembly contigs with BWA-_ 959 _MEM_ (Version 2). arXiv. https://doi.org/10.48550/ARXIV.1303.3997 960 Li, X., Chen, L., Liu, T., Chen, Y., Wang, J., & Song, B. (2024). Integrated analysis of ATAC961 seq and transcriptomic reveals the ScDof3 _-ScproC_ molecular module regulating the 962 cold acclimation capacity of potato. _Plant Physiology and Biochemistry_, _210_, 108576. 963 https://doi.org/10.1016/j.plaphy.2024.108576 964 Liang, Z., Zhang, Q., Ji, C., Hu, G., Zhang, P., Wang, Y., Yang, L., & Gu, X. (2021). 965 Reorganization of the 3D chromatin architecture of rice genomes during heat stress. 966 _BMC Biology_, _19_ (1), 53. https://doi.org/10.1186/s12915-021-00996-4 967 Liao, Y., Smyth, G. K., & Shi, W. (2014). featureCounts: An efficient general purpose program 968 for assigning sequence reads to genomic features. _Bioinformatics_, _30_ (7), 923–930. 969 https://doi.org/10.1093/bioinformatics/btt656 970 Liu, H., Able, A. J., & Able, J. A. (2022). Priming crops for the future: Rewiring stress memory. 971 _Trends in Plant Science_, _27_ (7), 699–716. https://doi.org/10.1016/j.tplants.2021.11.015 972 Liu, S., Kracher, B., Ziegler, J., Birkenbihl, R. P., & Somssich, I. E. (2015). Negative regulation 973 of ABA signaling by WRKY33 is critical for Arabidopsis immunity towards Botrytis 974 cinerea 2100. _eLife_, _4_, e07295. https://doi.org/10.7554/eLife.07295 975 Lu, Z., Marand, A. P., Ricci, W. A., Ethridge, C. L., Zhang, X., & Schmitz, R. J. (2019). The 976 prevalence, evolution and chromatin signatures of plant regulatory elements. _Nature_ 977 _Plants_, _5_ (12), Article 12. https://doi.org/10.1038/s41477-019-0548-z 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 978 Martin, M. (2011). Cutadapt removes adapter sequences from high-throughput sequencing 979 reads. _EMBnet.Journal_, _17_ (1), 10. https://doi.org/10.14806/ej.17.1.200 980 Morales, A., De Boer, H. J., Douma, J. C., Elsen, S., Engels, S., Glimmerveen, T., Sajeev, N., 981 Huber, M., Luimes, M., Luitjens, E., Raatjes, K., Hsieh, C., Teapal, J., Wildenbeest, T., 982 Jiang, Z., Pareek, A., Singla-Pareek, S., Yin, X., Evers, J., … Sasidharan, R. (2022). 983 Effects of sublethal single, simultaneous and sequential abiotic stresses on phenotypic 984 traits of _Arabidopsis_ _thaliana_ . _AoB_ _PLANTS_, _14_ (4), plac029. 985 https://doi.org/10.1093/aobpla/plac029 986 Morales-Cruz, A., Greenblum, S. I., Wang, P., Zhang, Y., Yang, L., Daum, C. G., Johnson, J., 987 Baumgart, L. A., & O’Malley, R. C. (2026). Positional grammar of transcription factor 988 binding partitions developmental and stress-response regulation in plants. _Nature_ 989 _Communications_, _17_ (1), 9403. https://doi.org/10.1038/s41467-026-75947-0 990 Murashige, T., & Skoog, F. (1962). A Revised Medium for Rapid Growth and Bio Assays with 991 Tobacco Tissue Cultures. _Physiologia_ _Plantarum_, _15_ (3), 473–497. 992 https://doi.org/10.1111/j.1399-3054.1962.tb08052.x 993 Ovek Baydar, D., Rauluseviciute, I., Aronsen, D. R., Blanc-Mathieu, R., Bonthuis, I., 994 de Beukelaer, H., Ferenc, K., Jegou, A., Kumar, V., Lemma, R. B., Lucas, J., Pochon, 995 M., Yun, C. M., Ramalingam, V., Deshpande, S. S., Patel, A., Marinov, G. K., Wang, 996 A. T., Aguirre, A., … Mathelier, A. (2026). JASPAR 2026: Expansion of transcription 997 factor binding profiles and integration of deep learning models. _Nucleic Acids_ 998 _Research_, _54_ (D1), D184–D193. https://doi.org/10.1093/nar/gkaf1209 999 Paradis, E., & Schliep, K. (2019). ape 5.0: An environment for modern phylogenetics and 1000 evolutionary analyses in R. _Bioinformatics_, _35_ (3), 526–528. 1001 https://doi.org/10.1093/bioinformatics/bty633 1002 Parker, A. H., Zhang, P., Man, K. Y., Tirot, L., Rolfe, S. A., Smith, L. M., Wilkinson, S. W., & 1003 Ton, J. (2025). _Contrasting epigenetic dynamics in chromosome arms and_ 1004 _pericentromeres_ _during_ _ROS1-induced_ _plant_ _immune_ _memory_ . 1005 https://doi.org/10.1101/2025.08.19.671022 1006 Patro, R., Duggal, G., Love, M. I., Irizarry, R. A., & Kingsford, C. (2017). Salmon provides fast 1007 and bias-aware quantification of transcript expression. _Nature Methods_, _14_ (4), 417– 1008 419. https://doi.org/10.1038/nmeth.4197 1009 Pham, G. M., Hamilton, J. P., Wood, J. C., Burke, J. T., Zhao, H., Vaillancourt, B., Ou, S., 1010 Jiang, J., & Buell, C. R. (2020). Construction of a chromosome-scale long-read 1011 reference genome assembly for potato. _GigaScience_, _9_ (9), giaa100. 1012 https://doi.org/10.1093/gigascience/giaa100 1013 R Core Team. (2026). _R: a language and environment for statistical computing_ (Version 4.6.0) 1014 [Computer software]. R Foundation for Statistical Computing. 1015 https://doi.org/10.32614/R.manuals 1016 Ramírez, F., Ryan, D. P., Grüning, B., Bhardwaj, V., Kilpert, F., Richter, A. S., Heyne, S., 1017 Dündar, F., & Manke, T. (2016). deepTools2: A next generation web server for deep1018 sequencing data analysis. _Nucleic Acids Research_, _44_ (W1), W160–W165. 1019 https://doi.org/10.1093/nar/gkw257 1020 Reynoso, M. A., Kajala, K., Bajic, M., West, D. A., Pauluzzi, G., Yao, A. I., Hatch, K., Zumstein, 1021 K., Woodhouse, M., Rodriguez-Medina, J., Sinha, N., Brady, S. M., Deal, R. B., & 1022 Bailey-Serres, J. (2019). Evolutionary flexibility in flooding response circuitry in 1023 angiosperms. _Science_, _365_ (6459), 1291–1295. 1024 https://doi.org/10.1126/science.aax8862 28 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1025 Ritchie, M. E., Phipson, B., Wu, D., Hu, Y., Law, C. W., Shi, W., & Smyth, G. K. (2015). Limma 1026 powers differential expression analyses for RNA-sequencing and microarray studies. 1027 _Nucleic Acids Research_, _43_ (7), e47–e47. https://doi.org/10.1093/nar/gkv007 1028 Rizhsky, L., Liang, H., & Mittler, R. (2002). The Combined Effect of Drought Stress and Heat 1029 Shock on Gene Expression in Tobacco. _Plant Physiology_, _130_ (3), 1143–1151. 1030 https://doi.org/10.1104/pp.006858 1031 Rizhsky, L., Liang, H., Shuman, J., Shulaev, V., Davletova, S., & Mittler, R. (2004). When 1032 Defense Pathways Collide. The Response of Arabidopsis to a Combination of Drought 1033 and Heat Stress. _Plant_ _Physiology_, _134_ (4), 1683–1696. 1034 https://doi.org/10.1104/pp.103.033431 1035 Robinson, M. D., McCarthy, D. J., & Smyth, G. K. (2010). edgeR : a Bioconductor package for 1036 differential expression analysis of digital gene expression data. _Bioinformatics_, _26_ (1), 1037 139–140. https://doi.org/10.1093/bioinformatics/btp616 1038 Schep, A. N., Wu, B., Buenrostro, J. D., & Greenleaf, W. J. (2017). chromVAR: inferring 1039 transcription-factor-associated accessibility from single-cell epigenomic data. _Nature_ 1040 _Methods_, _14_ (10), 975–978. https://doi.org/10.1038/nmeth.4401 1041 Schwacke, R., Ponce-Soto, G. Y., Krause, K., Bolger, A. M., Arsova, B., Hallab, A., Gruden, 1042 K., Stitt, M., Bolger, M. E., & Usadel, B. (2019). MapMan4: A Refined Protein 1043 Classification and Annotation Framework Applicable to Multi-Omics Data Analysis. 1044 _Molecular Plant_, _12_ (6), 879–892. https://doi.org/10.1016/j.molp.2019.01.003 1045 Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D., Amin, N., 1046 Schwikowski, B., & Ideker, T. (2003). Cytoscape: A Software Environment for 1047 Integrated Models of Biomolecular Interaction Networks. _Genome Research_, _13_ (11), 1048 2498–2504. https://doi.org/10.1101/gr.1239303 1049 Soneson, C., Love, M. I., & Robinson, M. D. (2016). Differential analyses for RNA-seq: 1050 Transcript-level estimates improve gene-level inferences. _F1000Research_, _4_, 1521. 1051 https://doi.org/10.12688/f1000research.7563.2 1052 Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S., Ebert, B. L., Gillette, M. A., 1053 Paulovich, A., Pomeroy, S. L., Golub, T. R., Lander, E. S., & Mesirov, J. P. (2005). 1054 Gene set enrichment analysis: A knowledge-based approach for interpreting genome1055 wide expression profiles. _Proceedings of the National Academy of Sciences_, _102_ (43), 1056 15545–15550. https://doi.org/10.1073/pnas.0506580102 1057 Sullivan, A. M., Arsovski, A. A., Lempe, J., Bubb, K. L., Weirauch, M. T., Sabo, P. J., 1058 Sandstrom, R., Thurman, R. E., Neph, S., Reynolds, A. P., Stergachis, A. B., Vernot, 1059 B., Johnson, A. K., Haugen, E., Sullivan, S. T., Thompson, A., Neri, F. V., Weaver, M., 1060 Diegel, M., … Stamatoyannopoulos, J. A. (2014). Mapping and Dynamics of 1061 Regulatory DNA and Transcription Factor Networks in A. thaliana. _Cell Reports_, _8_ (6), 1062 2015–2030. https://doi.org/10.1016/j.celrep.2014.08.019 1063 Tamura, K., & Bono, H. (2022). Meta-Analysis of RNA Sequencing Data of Arabidopsis and 1064 Rice under Hypoxia. _Life_, _12_ (7), 1079. https://doi.org/10.3390/life12071079 1065 Tournaire-Roux, C., Sutka, M., Javot, H., Gout, E., Gerbeau, P., Luu, D.-T., Bligny, R., & 1066 Maurel, C. (2003). Cytosolic pH regulates root water transport during anoxic stress 1067 through gating of aquaporins. _Nature_, _425_ (6956), 393–397. 1068 https://doi.org/10.1038/nature01853 1069 Trapero-Mozos, A., Ducreux, L. J. M., Bita, C. E., Morris, W., Wiese, C., Morris, J. A., 1070 Paterson, C., Hedley, P. E., Hancock, R. D., & Taylor, M. (2018). A reversible light1071 and genotype-dependent acquired thermotolerance response protects the potato plant 29 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1072 from damage due to excessive temperature. _Planta_, _247_ (6), 1377–1392. 1073 https://doi.org/10.1007/s00425-018-2874-1 1074 Van Bel, M., Silvestri, F., Weitz, E. M., Kreft, L., Botzki, A., Coppens, F., & Vandepoele, K. 1075 (2022). PLAZA 5.0: Extending the scope and power of comparative and functional 1076 genomics in plants. _Nucleic_ _Acids_ _Research_, _50_ (D1), D1468–D1474. 1077 https://doi.org/10.1093/nar/gkab1024 1078 van Berloo, R., Hutten, R. C. B., van Eck, H. J., & Visser, R. G. F. (2007). An Online Potato 1079 Pedigree Database Resource. _Potato_ _Research_, _50_ (1), 45–57. 1080 https://doi.org/10.1007/s11540-007-9028-3 1081 Vo, H. M., Charleston, M. A., Brodribb, T. J., & Sussmilch, F. C. (2025). Dehydration rapidly 1082 induces expression of NCED genes from a single subclade in diverse eudicots. _Planta_, 1083 _261_ (2), 46. https://doi.org/10.1007/s00425-025-04626-z 1084 Wang, T.-Q., Peng, X.-R., Gao, H., Wei, L., Wang, S.-M., Wu, Y., & Li, Y. J. (2025). Chromatin 1085 accessibility and transcriptome analyses reveal stress-responsive regulatory elements 1086 in _Zygophyllum_ _xanthoxylum_ . _Plant_ _Physiology_, _199_ (2), kiaf503. 1087 https://doi.org/10.1093/plphys/kiaf503 1088 Wang, Z.-W., Liu, M., Sang, H.-T., Hou, S., Li, G., Tian, R.-M., Yang, Y.-Y., Xie, K., Wang, L., 1089 Li, N.-N., & Jia, K.-H. (2026). Salt priming coordinates transcriptional and epigenetic 1090 states for enhanced salt tolerance in mung bean (Vigna radiata). _Communications_ 1091 _Biology_, _9_ (1), 1384. https://doi.org/10.1038/s42003-026-10574-5 1092 Wei, X., Zhang, H., Ma, R., Zang, J., Wang, Y., & Yu, C. (2026). Heat stress responses of 1093 plants during adaptation to high temperature environments: A review of recent 1094 advances and insights. _Plant_ _Signaling_ _&_ _Behavior_, _21_ (1), 2735211. 1095 https://doi.org/10.1080/15592324.2026.2735211 1096 Wen, S., Jian, H., Shang, L., Kear, P. J., Zhang, M., Li, Y., Yuan, P., & Lyu, D. (2025). 1097 Comprehensive transcriptional regulatory networks in potato through chromatin 1098 accessibility and transcriptome under drought and salt stresses. _The Plant Journal_, 1099 _121_ (6), e70081. https://doi.org/10.1111/tpj.70081 1100 Xu, S., Li, L., Luo, X., Chen, M., Tang, W., Zhan, L., Dai, Z., Lam, T. T., Guan, Y., & Yu, G. 1101 (2022). _Ggtree_ : A serialized data object for visualization of a phylogenetic tree and 1102 annotation data. _iMeta_, _1_ (4), e56. https://doi.org/10.1002/imt2.56 1103 Yeo, D. S. G., Eydam, J., Bruckmüller, J., Kauder, F., Lübeck, J., Kaier, A., Sonnewald, S., & 1104 Sonnewald, U. (2026). Genotype ‐ dependent DNA methylation patterns are negatively 1105 associated with allelic variation rather than heat ‐ induced gene expression in two 1106 contrasting potato genotypes. _The_ _Plant_ _Journal_, _125_ (2), e70690. 1107 https://doi.org/10.1111/tpj.70690 1108 Yoshiba, Y., Kiyosue, T., Nakashima, K., Yamaguchi-Shinozaki, K., & Shinozaki, K. (1997). 1109 Regulation of Levels of Proline as an Osmolyte in Plants under Water Stress. _Plant_ 1110 _and_ _Cell_ _Physiology_, _38_ (10), 1095–1102. 1111 https://doi.org/10.1093/oxfordjournals.pcp.a029093 1112 Yu, G., Wang, L.-G., & He, Q.-Y. (2015). ChIPseeker: An R/Bioconductor package for ChIP 1113 peak annotation, comparison and visualization. _Bioinformatics_, _31_ (14), 2382–2383. 1114 https://doi.org/10.1093/bioinformatics/btv145 1115 Zagorščak, M., Abdelhakim, L., Rodriguez-Granados, N. Y., Široká, J., Ghatak, A., Bleker, C., 1116 Blejec, A., Zrimec, J., Novák, O., Pěnčík, A., Baebler, Š., Perez Borroto, L., Schuy, C., 1117 Županič, A., Afjehi-Sadat, L., Wurzinger, B., Weckwerth, W., Pompe Novak, M., Knight, 1118 M. R., … Teige, M. (2025). Integration of multi-omics data and deep phenotyping 30 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1119 provides insights into responses to single and combined abiotic stress in potato. _Plant_ 1120 _Physiology_, _197_ (4), kiaf126. https://doi.org/10.1093/plphys/kiaf126 1121 Zagorščak, M., Zrimec, J., Bleker, C., Nolte, N., Juteršek, M., Ramšak, Ž., Gruden, K., & Petek, 1122 M. (2024). Evidence-based unification of potato gene models with the UniTato 1123 collaborative genome browser. _Frontiers_ _in_ _Plant_ _Science_, _15_ . 1124 https://doi.org/10.3389/fpls.2024.1352253 1125 Zandalinas, S. I., & Mittler, R. (2022). Plant responses to multifactorial stress combination. 1126 _New Phytologist_, _234_ (4), 1161–1167. https://doi.org/10.1111/nph.18087 1127 Zandalinas, S. I., Sengupta, S., Fritschi, F. B., Azad, R. K., Nechushtai, R., & Mittler, R. (2021). 1128 The impact of multifactorial stress combination on plant growth and survival. _New_ 1129 _Phytologist_, _230_ (3), 1034–1048. https://doi.org/10.1111/nph.17232 1130 Zhang, H., & Sonnewald, U. (2017). Differences and commonalities of plant responses to 1131 single and combined stresses. _The_ _Plant_ _Journal_, _90_ (5), 839–855. 1132 https://doi.org/10.1111/tpj.13557 1133 Zhou, J., Wang, J., Zheng, Z., Fan, B., Yu, J.-Q., & Chen, Z. (2015). Characterization of the 1134 promoter and extended C-terminal domain of Arabidopsis WRKY33 and functional 1135 analysis of tomato WRKY33 homologues in plant stress responses. _Journal of_ 1136 _Experimental Botany_, _66_ (15), 4567–4583. https://doi.org/10.1093/jxb/erv221 1137 Zhou, Y., Zhang, Z., Bao, Z., Li, H., Lyu, Y., Zan, Y., Wu, Y., Cheng, L., Fang, Y., Wu, K., 1138 Zhang, J., Lyu, H., Lin, T., Gao, Q., Saha, S., Mueller, L., Fei, Z., Städler, T., Xu, S., 1139 … Huang, S. (2022). Graph pangenome captures missing heritability and empowers 1140 tomato breeding. _Nature_, _606_ (7914), 527–534. https://doi.org/10.1038/s41586-022 1141 04808-9 1142 31 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.11.765576; this version posted November 13, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.11.765576) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1143 **ACKNOWLEDGEMENTS** 1144 We acknowledge Luka Kranjc and Sandi Hunyadi for the help with data management and 1145 deposition. 1146 1147 **AUTHOR CONTRIBUTIONS** 1148 S.S., R.D.H., M.A.T., S.P., M.T. and K.G. conceived the study. All authors contributed to the 1149 development of the methodology. T.G., L.A., R.C. and M.A.S. performed the experiments. 1150 K.P., M.A.T., S.P. and K.G. provided resources. T.G., M.Z., Ž.R. and M.P. developed the 1151 computational pipelines. T.G., M.Z., M.P. and K.G. performed the formal analysis and curated 1152 the data. T.G., M.Z. and K.G. prepared the figures. T.G., M.Z., M.P. and K.G. wrote the original 1153 draft of the manuscript, and all authors reviewed and edited it. R.D.H., M.A.T., S.P., M.T. and 1154 K.G. supervised the research, administered the project and acquired funding. 1155 1156 **FUNDING** 1157 This work has received funding from the European Commission's Horizon 2020 research and 1158 innovation programme through the project “ADAPT” (grant agreement no. 862858), and by 1159 the Slovenian Research and Innovation Agency (ARIS) through research programmes P41160 0431 and P4-0463 and project Z1-60169. Research infrastructure was supported by ARIS 1161 through the NIB infrastructure programme (I0-0004) and ELIXIR Slovenia (I0-E013). The 1162 James Hutton Institute is supported by the Scottish Government Rural and Environmental 1163 Science and Analytical Services Division. 1164 1165 **DATA AVAILABILITY** 1166 Raw reads from the initial and priming experiment were deposited in the European Nucleotide 1167 Archive (ENA) at EMBL-EBI under accession numbers PRJEB127277 and PRJEB129741, 1168 respectively. 32