Integrated multi-omic analyses uncover a regulatory link between photosynthesis and drought tolerance in field-grown sorghum

Integrated multi-omic analyses uncover a regulatory link between photosynthesis and drought tolerance in field-grown sorghum | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 2 February 2026 V1 Latest version Share on Integrated multi-omic analyses uncover a regulatory link between photosynthesis and drought tolerance in field-grown sorghum Authors : Li’ang Yu 0000-0002-9556-011X , Giovanni Melandri , Anna C. Nelson Dittrich , Hamada AbdElgawad 0000-0001-9764-9006 , Gerrit Beemster 0000-0001-6014-053X , Ciara Denise Garcia , A. Elizabeth Arnold , Brian D. Gregory , Duke Pauli , and Andrew Nelson 0000-0001-9896-1739 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.177001054.41782127/v1 221 views 101 downloads Contents Abstract Sorghum growth and phenotypic measurement Sorghum oxidative stress-related biochemical measurements RNA extraction and library construction Transcriptome data processing Construction of co-expression networks Functional and regulatory network analyses Gene ontology and pathway enrichment analyses Diverse phenotypic response to drought among sorghum accessions binding motif associated genes exhibit a drought stress response pattern Data availability Acknowledgments Author contributions Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Global warming and increasing water scarcity pose major challenges to agriculture, emphasizing the need to translate stress-biology insights into the development of drought-resilient cultivars. In this study, we evaluated a panel of six diverse sorghum ( Sorghum bicolor ) accessions grown under field-imposed drought conditions in central Arizona, integrating physiological, transcriptomic, and oxidative stress measurements over a seven-week period. Using network analyses informed by co-expression correlations, transcription factor (TF) binding motif signatures, and protein–protein interactions, we identified a drought-associated module strongly correlated with photosynthetic capacity. This module contained a stress-responsive TF, SbDof8 (referred to here as SbCDF2/3-like, or SbCDF2/3L) , as a highly connected hub gene, and CDF2/3-associated binding motifs were over-represented in the promoters of co-expressed module members. These co-expressed members were enriched for stress response, metabolic, and photosynthesis-related processes, and consistently maintain higher expression under drought in tolerant compared to sensitive accessions. Analysis of an independent sorghum drought time-course dataset comprised of two unique accessions revealed concordant expression patterns of SbCDF2/3L and photosynthesis-associated module genes between drought-tolerant and drought-sensitive genotypes, reinforcing the robustness of this regulatory module. Together, our results highlight a subset of photosystem I (PSI)–related genes, including light-harvesting proteins, PSI subunits, and importantly, a potential drought responsive transcriptional regulator that are collectively upregulated in resilient accessions as a means of coping with drought response. These data highlight promising breeding targets for improving drought resilience and biomass productivity in sorghum under field conditions. Integrated multi-omic analyses uncover a regulatory link between photosynthesis and drought tolerance in field-grown sorghum Li’ang Yu 1 , Giovanni Melandri 2,3 , Anna C. Nelson Dittrich 1 , Hamada AbdElgawad 4 , Gerrit T.S. Beemster 4 , Ciara Denise Garcia 2 , A. Elizabeth Arnold 2,3 , Brian D. Gregory 5 *, Duke Pauli 2,3 *, Andrew D. L. Nelson 1 * 1 Boyce Thompson Institute, Cornell University, Ithaca, NY 14850, USA 2 School of Plant Sciences, University of Arizona, Tucson, AZ, 85721, USA 3 Center for Agroecosystem Research in the Desert (ARID), University of Arizona, Tucson, AZ 85721, USA 4 Integrated Molecular Plant Physiology Research, University of Antwerp, 2020 Antwerp, Belgium 5 Department of Biology, University of Pennsylvania, Philadelphia, PA, USA *Address correspondence to: Dr. Andrew D. L. Nelson The Boyce Thompson Institute, Cornell University 533 Tower Road, Ithaca, NY, 14850. [email protected] or Dr. Duke Pauli The University of Arizona, School of Plant Sciences 1140 E. South Campus Drive, Tucson, AZ, 85721 [email protected] or Dr. Brian D. Gregory University of Pennsylvania, Department of Biology, 103 H Lynch Laboratory 433 South University Avenue, Philadelphia, PA, 19104 [email protected] Current Email addresses: LY: [email protected] , GM: [email protected] , ACND: [email protected] , HAE: [email protected] , GTSB: [email protected] , CDG: [email protected] , AEA: [email protected] Abstract Global warming and increasing water scarcity pose major challenges to agriculture, emphasizing the need to translate stress-biology insights into the development of drought-resilient cultivars. In this study, we evaluated a panel of six diverse sorghum ( Sorghum bicolor ) accessions grown under field-imposed drought conditions in central Arizona, integrating physiological, transcriptomic, and oxidative stress measurements over a seven-week period. Using network analyses informed by co-expression correlations, transcription factor (TF) binding motif signatures, and protein–protein interactions, we identified a drought-associated module strongly correlated with photosynthetic capacity. This module contained a stress-responsive TF, SbDof8 (referred to here as SbCDF2/3-like, or SbCDF2/3L) , as a highly connected hub gene, and CDF2/3-associated binding motifs were over-represented in the promoters of co-expressed module members. These co-expressed members were enriched for stress response, metabolic, and photosynthesis-related processes, and consistently maintain higher expression under drought in tolerant compared to sensitive accessions. Analysis of an independent sorghum drought time-course dataset comprised of two unique accessions revealed concordant expression patterns of SbCDF2/3L and photosynthesis-associated module genes between drought-tolerant and drought-sensitive genotypes, reinforcing the robustness of this regulatory module. Together, our results highlight a subset of photosystem I (PSI)–related genes, including light-harvesting proteins, PSI subunits, and importantly, a potential drought responsive transcriptional regulator that are collectively upregulated in resilient accessions as a means of coping with drought response. These data highlight promising breeding targets for improving drought resilience and biomass productivity in sorghum under field conditions. Keywords: Sorghum, Drought response, Systems biology, Photosynthesis, Oxidative stress metabolites Introduction Sorghum bicolor (sorghum) is the fifth most important industrial crop worldwide and is grown for grain, brewing, and cellulosic biomass production. As a member of the Poaceae and a close relative of maize, sorghum is believed to have been domesticated around 4,000–6,000 years ago in Africa, and to have then diverged into a variety of landraces with diversified morphological variation due to selection across broad geographical locations (Morris et al. 2013). Most sorghum cultivars were historically, and are currently, grown in arid and semi-arid areas due to their tolerance to limited irrigation during growth. However, when exposed to prolonged and severe (Chen et al. 2025) water deficit sorghum experiences a reduction in a number of agronomic traits such as biomass and grain yield (Abreha et al. 2021). These traits are believed to be controlled by multiple genes and influenced by a number of developmental and environmental cues (Singh et al. 2025). Thus, a better understanding of the genetic mechanisms allowing sorghum to interpret and coordinate responses to environmental cues is essential for maintaining yield under stress conditions (Alizadeh et al. 2020). Drought induces numerous unfavorable phenotypic and physiological changes, such as reduced yield, altered plant growth, and increased susceptibility to biotic stressors. Plant phenotypic responses to reduced water availability are associated with physiological and molecular responses. More drought-tolerant plants typically exhibit increased water use efficiency (WUE), which is often achieved through reduced transpiration and consequently leads to lower net CO₂ assimilation. Metabolite, and oxidative stress enzyme activity levels also change in response to drought, as commonly seen by the increased levels of specific soluble sugars and amino acids able to act as osmoprotectants (Hasanuzzaman et al. 2013; Zargar et al. 2017; Fàbregas and Fernie 2019). These phenotypic, physiological, and metabolic traits, when combined with additional molecular data, can facilitate the identification of drought tolerance mechanisms linked to enhanced performance. Drought responses are coordinated at the cellular level through a complex set of regulatory control points, guided by the activation and repression of transcription factors (TFs) that in turn regulate the expression of their target genes (Zhu 2016; Kim et al. 2024). These regulatory networks receive and interpret cues from the environment and govern the transcriptional cascades that link stress signaling to downstream defense and adaptation mechanisms that ultimately determine the effect on phenotypes such as yield (Sato et al. 2024). While large population-level studies of sorghum have identified strong drought tolerance phenotypes useful for breeding and selection (Ortiz and Salas-Fernandez 2022; Tsehaye et al. 2024), the crosstalk between many upstream regulatory genes and downstream defense genes and yield-related traits remains unclear and largely unexplored. Systems-level approaches have been quite successful at identifying the elements that influence stress and yield in several crop species (Bang et al. 2019; Manna et al. 2021; Yu et al. 2024), although much of this work is often done in a single crop accession. One family of plant-specific transcription factors, the DNA binding with one finger (Dof) family, has been observed in a number of species to directly couple environmental signaling with the transcriptional regulation of the photosynthetic machinery. Unlike general stress regulators, Dofs appear to exert specific control over both the light reactions and carbon fixation pathways. For instance, the Maize Dof1 TF has been observed to activate expression of the phosphoenolpyruvate carboxylase (PEPC) genes in both rice and Arabidopsis via heterologous expression (Kurai et al. 2011). This induction was more pronounced under limited nitrogen conditions, and resulted in increased biomass and carbon assimilation. The rice atypical Dof, OsDes1 was recently shown to directly bind the promoter of a nuclear encoded member of the electron transport chain machinery, Rieske FeS ( OsPetC ), activating its expression and thereby conferring a “stay-green” phenotype fundamental to yield maintenance (Qiu et al. 2024). However, this Dof regulatory capacity is nuanced. In wheat, TaDof7.6 inhibits key C4-photosynthesis associated enzymes (e.g., PPDK); suppression of TaDof7.6 led to enhanced photosynthetic capacity and increased grain filling (Zhang et al. 2025). Furthermore, Dofs have been reported to be upregulated during mild drought stress in maize (Nelissen et al. 2018). Finally, a specific subclass of Dofs, the cycling Dofs (CDFs), have also been shown modulate light-regulated genes in a circadian fashion (Saibo et al. 2009). Interestingly, while CDFs, such as Arabidopsis CDF1 and CDF5 (Henriques et al. 2017; Furihata et al. 2025) have primarily been reported as regulators of flowering time, Arabidopsis CDF3 has been observed to enhance photosynthetic capacity under abiotic stress in Arabidopsis (Corrales et al. 2017), with ectopic expression resulting in enhanced biomass, yield, and carbon/nitrogen assimilation in tomato. While the Dof family of TFs are clearly important for integrating all manner of environmental cues into how plants regulate photosynthesis and core metabolism, this family is large, and determining which Dofs perform these important functions is still poorly understood. To identify the genetic basis by which sorghum responds to drought stress, we leveraged a systems biology approach to study six phylogenetically distinct sorghum accessions from the Sorghum Association Panel (SAP) with different levels of drought tolerance (Boatwright et al. 2022). These accessions were grown and evaluated in the field in the hot and arid environment of the Arizona low desert and exposed to two different watering regimes after flowering. We collected matched transcriptomic, oxidative stress status, photosynthetic, and whole plant phenotypic data at multiple points throughout the season to obtain an integrated understanding of the responses to soil water conditions. We uncovered a suite of genes in a co-expression network that associated with genotypic factors where the tolerant and sensitive accessions performed differently. Based on co-expression and promoter binding element enrichment, we identified a homolog of the cycling DoF (CDF) subfamily acting as a likely key regulator of this gene network. Based on phylogenetic assessment and promoter enrichment, we refer to this putative sorghum CDF as a CDF2/3-Like transcription factor (SbCDF2/3L). SbCDF2/3L target genes exhibited a consistent drought-response pattern across both our field experiment and an independent sorghum drought time-course dataset (Varoquaux et al. 2019). Drought-tolerant genotypes maintained elevated expression under water stress relative to control conditions of both SbCDF2/3L and a functionally coupled group of photosynthetic genes coordinating carbon fixation, photosystem I assembly, and electron transport. Elevated expression of these genes in drought tolerant accessions facilitated higher photosynthetic capacity, whereas sensitive genotypes showed pronounced downregulation along with strongly reduced photosynthetic performance. Overall, these data support this cohort of genes as strong candidates for targeted breeding aimed at sustaining biomass accumulation and productivity under drought conditions. Methods and Materials Sorghum growth and phenotypic measurement Six selected sorghum accessions (PI 533871: M1, PI 533961: DL/59/1530, PI 656041: 80M, PI 656053: N290-B, PI 656076: SC1271, and PI 656096: SC391) were grown at the University of Arizona Maricopa Agricultural Center (MAC) in Maricopa, AZ, US. All plants were well-watered (WW) at the same rate (~24% soil volume metric water content, SVWC), corresponding approximately to the soil field capacity) until flag leaf appearance in the whorl (~47 days after planting), after which irrigation was reduced to a subset of plots to mimic drought treatment (Water-limited, WL; 16% SVWC). Destructive plant phenotyping was conducted at the end of the season (maturity stage). Measured traits included above ground biomass, root biomass, and three-leaf weights. For other time-series molecular phenotypes, samples were collected weekly (on Thursdays) for seven weeks (week 1-week 7, from August 13 th to September 24 th , 2020) on a specific subset of a large field experiment conducted at the University of Arizona’s Maricopa Agricultural Center (MAC), Maricopa, AZ, US. For transcriptomics and analysis of other biochemical traits, samples consisting of six leaf disks from five different plants per accession, treatment, and time point were collected (from 10:30 to 11:30) in three different 1.5 mL tubes and immediately snap-frozen in liquid nitrogen. LI-6800 derived photosynthesis-related traits ( N = 29), (https://www.licor.com, LI-COR Lincoln, NE) were measured in the morning (from 9:00 to 10:00) and afternoon (from 13:00 to 14:00) with the following settings: Flow rate = 600 µmol s -1 ; Mixing fan speed =12,000 rpm; CO2 = 420 ppm; Light= 2,000 µmol m -2 s -1 . The youngest fully expanded leaf on the main stem of each plant was selected for the measurements and data were logged as soon as assimilation and stomatal conductance curves (monitored in real-time on the instrument display) reached a plateau (on average 30 seconds after a leaf was clamped on to). Sorghum oxidative stress-related biochemical measurements Levels of oxidative stress markers and molecular antioxidants, and antioxidant enzyme and photorespiration enzyme activities were analyzed following established protocols in previous study (Melandri et al. 2020). Oxidative stress markers, such as malondialdehyde (MDA) and protein carbonyl content, were measured to assess lipid peroxidation and protein oxidation (Hodges et al. 1999). Photorespiration-related enzyme activities, glycolate oxidase (GOX) and hydroxypyruvate reductase (HPR), were assessed based on established methodologies (Schwitzguebel and Siegenthaler 1984). Lipoxygenase (LOX) was extracted in potassium phosphate buffer (pH 7.0), and its activity was measured by changes in conjugated dienes (Steczko et al. 1991). Total antioxidant capacity (TAC), total polyphenol content (Poly), and flavonoids were evaluated using standard approaches (Benzie and Strain 1999). Enzyme activities of key antioxidants, including ascorbate peroxidase (APX), dehydroascorbate reductase (DHAR), monodehydroascorbate reductase (MDHAR), glutathione reductase (GR), peroxidase (POX), superoxide dismutase (SOD), glutathione peroxidase (GPX), catalase (CAT), ascorbate oxidase (AO), and glutathione S -transferase (GST), were determined using microplate-based kinetic assays (Dhindsa et al. 1982). RNA extraction and library construction Approximately five flash-frozen sorghum ¼ leaf disks per sample were used as starting material for RNA extractions. RNA extraction was performed using the RNeasy Plant Mini kit (Qiagen #74904) and RNase-free DNase set (Qiagen #79254) according to manufacturer instructions. Depending on total RNA yield for each sample, 2.5-10 ug total RNA was used as starting material for mRNA enrichment using the Dynabeads mRNA purification kit (Invitrogen #61006) according to manufacturer instructions, except with the addition of a second binding and elution step to further enrich mRNA. RNA-seq libraries were then constructed from 10-60 ng mRNA per sample using the Amaryllis Nucleics YourSeq Duet Full Transcript library prep kit with combinatorial dual indexes (now Active Motif #23001 or #23002) according to manufacturer instructions. Prepared libraries were sequenced by Novogene with Illumina NextSeq 500 (PE × 150bp). Transcriptome data processing Sorghum RNA-reads of six accessions were aligned to sorghum reference genome using Hisat2 with default settings (Kim et al. 2019). FeatureCounts was used for quantification using the gene ID as the meta-level to count reads (v2.0.1, parameter: -s 1 -p -t mRNA -g ID -O). Read counts of 34,130 primary transcripts were normalized using DESeq2 after the removal of transcripts with less than an average of one count across 96 samples (Liao et al. 2013). We incorporated the time point (week 1, week 3, week 5, and week 7), accession IDs ( A: M1, B: DL/59/1530, C: 80M, D: N290-B, E: SC1271, and F: SC391) , and watering conditions (control and drought) as the experimental design factors. Condition-wise differentially expressed genes (DEGs) over each time point per accession were identified using the cutoff of F DR < 0.5 (Love et al. 2014). A log 2 -transformed fold change (FC) greater than 1 was used to filter up-regulated genes and a less than -1 log 2 FC filtered the down-regulated genes. RNA reads of a published large-scale sorghum drought transcriptome dataset covering 8-week post-flowering drought experiment on two cultivars (Varoquaux et al. 2019, GEO: http://www.ncbi.nlm.nih.gov/geo/) under accession no. GSE128441 was processed using the same pipeline above. For both public data and RNA-seq in the current experiment, TPM (Transcripts Per Million) score was calculated for read counts normalization for comparing relative expression of genes among samples in each experiment. Construction of co-expression networks The co-expression network was constructed using the WGCNA package with variance stabilizing transformed (VST) normalized read counts with 34,130 transcripts across 96 samples (Love et al. 2014). Initially, a soft threshold (Beta score) was applied to fit the scale-free topology model assumption which generates a minimum number to make the index curve flatten out reaching a high value (> 0.85, (Yu et al. 2024). Further, the adjacency matrix represented by gene expression was transformed into a topological overlap matrix (TOM) to classify genes using a dynamic tree-cut approach. Finally, genes that shared high-confidence correlation (abs(coefficient) > 0.8) were merged into the same module (parameter: deepsplit = 4, tree cut height = 0.2). The pairwise Pearson correlation between each gene from modules and the module-wise eigengene values was tested to characterize each gene’s module membership (MM) under the same module. The weight score, a parameter to denote the robustness of connections of each edge, was filtered by taking genes associated with the top 10% high-confidence weight ranking connections for downstream functional analysis. Lastly, the metabolite profiles, oxidative stress-related enzyme activities, and LI-6800-derived traits were integrated into the network using Pearson correlation (PC) between gene expression and these molecular phenotypes. Trait-correlated modules were selected based on their correlation level between MM scores and phenotypes (cutoff: r > 0.5, or r < -0.5, p-value < 0.05). The correlation between gene expression and phenotypes was estimated as gene significance (GS) to interpret the importance of genes related to each molecular phenotype. Functional and regulatory network analyses To study the TFs and their downstream target genes (protein-DNA interactions), TFs and protein kinases from each co-expression module were predicted using the iTAK tool based on conserved protein-coding domains (Zheng et al. 2016). To identify potential upstream regulators of genes in each module, we performed a motif enrichment analysis for the 2kb upstream promoter regions of genes in the same module using AME (–scoring avg –method fisher –hit-lo-fraction 0.25) and the Arabidopsis DAPseq consensus motif database (Bailey et al. 2015; O’Malley et al. 2016). The protein-protein interactions (PPI), PPIs among genes present in the co-expression network were acquired from the STRING database as the third layer of evidence to infer the functions or pathways of modules (parameter: confidence level: > 0.7, source: experiments and database source(von Mering et al. 2003). As a result, co-expression patterns, PPIs, and TF-binding interactions were merged into an integrated network. Gene ontology and pathway enrichment analyses Functional enrichment analysis of pathways (Kyoto Encyclopedia of Genes and Genomes (KEGG) database) and Gene Ontology (GO) analysis was performed to infer the biological function of genes associated with subset genes assigned in co-expression modules and differentially expressed genes between control and drought conditions. ShinyGO (https://bioinformatics.sdstate.edu/go77) was used to perform GO and KEGG enrichment analyses based on the sorghum gene functional annotations provided within the website. (Ge et al. 2019). Phylogenetic assessment of SbCDF2/3-like genes Sorghum Dof homologs were identified by reciprocal best BLAST using CoGe (genomevolution.org; Nelson et al. 2018)) with Arabidopsis CDFs 1, 2, and 3 as query (AT5G62430, AT5G39660, and AT3G47500, respectively) with E-value of 1e-10. Sorghum bicolor v3.1 (Phytozome,McCormick et al. 2018)), Oryza sativa japonica (Ensembl Release 36;Kawahara et al. 2013)), and TAIR10 genomes and annotations were used for searching. Rice and Arabidopsis Dofs were named according to nomenclature derived from Lijavetzky et al., 2003, with Sorghum labeling derived from Kushwaha et al., 2011. To infer relationships, multiple sequence alignments were generated using the Geneious multiple sequence alignment algorithm within Geneious v2023.0 (www.geneious.com) with default settings on both the core Dof domain (plus 10AA on either side) as well as full length protein sequences. RAxML was used to infer phylogenies (GTR-Gamma with rapid bootstrapping and 100 bootstraps). A larger Dof subfamily, comprised of both D 1 +d 1 and D 2 +d 2 subfamilies was used to place sorghum Dofs within the CDF-like family, with a secondary analysis on D 1 +d 1 -specific genes (which contains the Arabidopsis CDF-specific genes). Data visualization and usage of databases All statistical visualizations were generated in R (v4.3.2) with multiple packages. Briefly, plots include the line, boxplot, PCA, and enrichment graphs generated by the ggplot2 package, and heatmaps using the pheatmap package. R Scripts used for plots were deposited in (https://rpubs.com/LeonYu/Sorghum-comparative-transcriptomics)The integrated gene networks containing TF-binding, PPIs, and co-expression derived from WGCNA were generated using Cytoscape. Results Diverse phenotypic response to drought among sorghum accessions All six sorghum accessions were selected from the sorghum diversity panel (SAP) and a previous trial study (Boatwright et al. 2022; Yu et al. 2025) based on seed viability, genetic distinctiveness, and phenotypic variation to drought treatment (shoot and root biomass, plant height, and yield) Among the selected lines, four accessions (A, B, E, and F) were African accessions, while the remainder (C and D) were U.S. breeding accessions. To assess plant performance of these accessions under drought conditions, plants were grown in replicated plots as part of a large field trial. All plants were irrigated at normal rates (control conditions) until the reproductive stage (appearance of the flag leaf), after which a drought treatment was implemented by reducing irrigation to a subset of plots, as determined by continuous monitoring of soil water content ( Figure 1a ). Phenotypic measurements and molecular sampling (transcriptomics and oxidative stress biochemical traits) occurred during the 1 st , 3 rd , 5 th , and 7 th week after the beginning of the drought treatment ( Figure 1a ) , with destructive measurements (e.g., biomass and root microbiome composition) occurring at the 7 th week of sampling. Soil and plant-level phenotypic measurements confirmed that the reduction in irrigation was sufficient to observe drought stress responses in the panel. Overall, treatment effects were readily apparent and particularly pronounced at the end of season, with the M1 (“A”) accession showing the most pronounced leaf necrosis ( Figure 1b ). Endophytic root microbiome bacterial and fungal composition and distribution were impacted by water treatment, with clear separation between drought-treated and non-treated samples ( Figure 1c ). We found that the “A” accession exhibited less degrees of separation between control and drought conditions than other accessions ( Figure 1c ). We further used biomass measurements and two root morphological measurements to estimate the level of drought tolerance ( Figure 1d ). When comparing drought and control data separately, we observed the significant variation among accessions over four traits ( p < 0.05). In particular, the “A” (M1) and “D” (N290-B) accessions had the lowest total biomass under both control and drought conditions ( p < 0.05), potentially indicating either poor adaptation to both heat and drought stress or inherently lower biomass associated with these genetic backgrounds ( Figure 1d ). For root morphological phenotypes, “A” exhibited significantly lower length and width compared with other five accession under control conditions ( Figure 1d, p < 0.05), but these differences became less pronounced under drought stress ( Figure 1d ). When drought tolerance was evaluated using drought/control biomass ratios, “A” (M1) and “F” (SC 391) displayed a pronounced reduction of two distinct indicators, total biomass and leaf weight ( Figure S1a ), whereas “A” showed an increase in root related morphological features. To develop a clearer understanding of the leaf-level responses to treatment, we examined photosynthetic efficiency and capacity ( Figures 1d and S1b ). Here, photosynthetic efficiency and capacity were estimated by calculating maximum photosystem II efficiency in light conditions (Fv’/Fm’) and net CO 2 assimilation (A net ; Figure 1e ). These data indicate that the “A” accession displayed hallmarks of drought stress and heat stress, particularly at the later time points (week 5 and week 7, Figure 1e and Figure S1b ), whereas other accessions, such as “B” and “C”, showed a more robust response to drought conditions. These combined morphological and physiological data suggest that these six accessions display phenotypic diversity to drought stress, with B and C being more tolerant, and A and F being more sensitive. Thus, these phenotypic classifications should allow for the uncovering of molecular mechanisms by which sorghum responds to water limiting conditions. Figure 1. Drought response diversity of selected sorghum genotypes (a) A field drought experiment was conducted in Maricopa, AZ during the summer of 2020 (left). Drought treatment levels were monitored using soil moisture sensors at different soil depths over time (right). (b) Morphological responses of sorghum accessions under drought, showing plant architecture at the terminal field sampling stage. (c) Microbiome features, represented by 515 classified bacterial taxa and 46 fungal taxa, showing intra-accession variation based on principal component analysis (PCA) at the terminal sampling stage. (d) Physiological measurements of whole-plant, leaf, and root traits under two watering conditions ( N = 5 per accession) were compared using Kruskal test. (e) Bar plots showing photosynthetic efficiency and capacity under drought conditions, measured as Fv′/Fm′ and net CO₂ assimilation, respectively, at each time point for drought-treated plants ( N = 10 per accession). Pairwise Wilcoxon tests were used to compare each accession relative to “A” ( p < 0.05). Sorghum accessions displayed molecular response variation during drought stress To characterize the molecular responses of drought stress in our panel, we next assessed the change in transcript abundance over the two months of the treatment (week 1, 3, 5, and 7) using RNA-seq. Initial quality control revealed high correlation between replicates (Pearson correlation coefficient; r > 0.95). Using hierarchical clustering and Pearson correlation analysis of global transcript abundance ( N = 22,314 , mean TPM > 1), the 96 samples clustered into subgroups that fit their genotypic classification ( Figure 2a ). Separation due to treatment effect was observed for accession “B”, whereas accession “A” separated by time (week 1 clustered separately from weeks 3-7), suggesting a response to prolonged exposure to the environment for this accession ( Figure 2a ). In general, these data point to genotype being the largest driver of separation, followed by treatment, for most accessions. Figure 2. Molecular response to drought among sorghum accessions (a) A hierarchical clustering (Pearson correlation) of 34,130 transcripts normalized by DESeq2 was used to group transcriptomic features among the 96 samples taken over the course of the experiment (Weeks 1, 3, 5, 7; drought vs control). The scale indicates the Pearson correlation coefficient (PCC) score. (b) A line graph representing the number of up (circle) or down (triangle) regulated genes over time. (c) The enrichment levels of GO terms that were shared between DEGs from at least two accessions are shown. GO terms must have been enriched to at least a FDR < 0.05 in order to be shown. The scale of enrichment is shown to the left. (d) Pairwise Pearson correlation coefficient (PCC) of 24 chemicals associated with antioxidant metabolites, antioxidant enzymes, and oxidative stress markers were displayed by heatmap with the three chemical classes denoted. Red dashed boxes show correlation of metabolites belonging to same class and red solid line boxes highlight correlation between two classes. (e) Representative two oxidative makers (HPR and GOX) from ( d ) were selected to display ratio between drought/control across time for each of the six genotypes. To unpack the transcriptomic responses to treatment in the six accessions, pairwise comparisons of differentially expressed genes (DEGs; FDR < 0.05) between control and drought for each timepoint and genotype were performed ( Supplemental Table 1 ). Accession “C” showed the lowest response to treatment, whereas accessions “E” and “A” had the most DEGs across the four weeks ( Figure 2b ). Gene ontology (GO) term enrichment of these DEGs (up and down-regulated, FDR < 0.05) reinforced prior observations that each of these genotypes was responding to the treatment by modulating unique pools of transcripts ( Figure 2b ). Interestingly, when comparing DEGs across all timepoints between drought and control in the “A” and “F” accessions, we see no significant enrichment of “stress response” related GO terms ( Figure 2c ). However, we did observe enrichment of these terms when comparing timepoints under stress (e.g., week 7 vs week 1; Supplemental Table 2 ). Thus, these accessions are experiencing and responding to the imposed stress conditions, but not to the degree observed for the other four accessions. Highlighting the unique responses to treatment across these accessions, when examining the overlap of treatment-induced DEGs among the six accessions, minimal overlap was observed for either up- or down-regulated DEGs, consistent with primarily genotype-driven global transcriptomic variation observed across the six accessions ( Figure 2a and Figure S2 ). To gather additional clues as to how these accessions responded to drought treatment at the molecular level, we next examined the stress-induced changes in components of oxidative stress response or resolution pathways. Samples for oxidative stress-related metabolite and transcriptome measurements levels were collected simultaneously to enable stronger associations. This metabolite profiling dataset comprised of oxidative stress markers, antioxidant metabolites, and antioxidant-related enzymes ( N = 24). Pairwise correlation analysis revealed strong within-class correlations ( Figure 2d ), with oxidative stress markers (MDA, LOX, GOX, and HPR) showing partial negative correlations with several antioxidant-related enzymes (e.g., GPX, Trxs, and Grxs). In contrast, most antioxidant metabolites were positively correlated with a subset of antioxidant enzymes, including SOD, AO, POX, CAT, and APX. Together, correlation in metabolites and related enzymes reveal the expected coordinated oxidative stress response across accessions. We then compared oxidative stress-induced response levels, quantified as drought/control ratios, across several representative metabolites. Among oxidative stress markers, we focused on two enzymes, HPR and GOX, which play key roles in the photorespiratory pathway and link photosynthetic carbon metabolism with reactive oxygen species (ROS) homeostasis ( Figure 2e ; Keech et al. 2017). In addition, four antioxidant components (glutathione, ascorbate, Frxs, and GPX) were included, as they protect photosynthetic and photorespiratory processes by maintaining redox balance and scavenging ROS ( Figure S3 ; Sakhno et al. 2019; González et al. 2021). At weeks 5 and 7, we observed pronounced stress responses, with accessions “A” and “F” showing higher accumulation of two oxidative stress markers ( Figure 2e ) and generally reduced levels of antioxidant enzymes and metabolites compared with the other four accessions ( Figure S3 ). These patterns indicate elevated oxidative stress in “A” and “F”, consistent with the absence of stress-related GO terms among DEGs ( Figure 2b ). Overall, transcriptomic and oxidative stress profiles align with photosynthetic traits and other physiological parameters, highlighting that “A” and “F” are drought-sensitive, with accession “A” showing the poorest performance across multiple criteria. Identification of a drought-associated cohort of co-expressed transcripts With each of these molecular and physiological traits examined and the relative level of drought sensitivity among accessions assessed, we next took a comparative systems-level approach to connect these traits to the molecular pathways responsible for controlling them. To do so, we constructed a co-expression network starting with the top 10% most variable transcripts, as measured by median absolute deviation (MAD) score. We then recursively explored parameter space, increasing the number of included transcripts until we achieved an optimized network with the fewest number of unassigned transcripts and the highest scale-free topology model (Yu et al. 2024). The resulting network contained 22,314 transcripts (65.5% of annotated genes in sorghum) classified into 44 modules with a low portion of unclassified transcripts ( N = 2,646) and high scale-free model correlation ( r = 0.91, slope = -1.93, Supplemental Table 3 ). Using this framework, we incorporated phenotypic data into modules using trait-module correlations. Briefly, the first principal component value (eigenvalue) per module was calculated using their transcript abundance PCA. Leaf level physiological measurements and biochemical traits related to oxidative stress status were then correlated with the eigenvalues of each module ( Supplemental Tables 4 and 5 ), leading to 13 modules with significant trait-module correlations ( p < 0.05, Figure 3a , left panel). Several modules (e.g. “Lightyellow”, “Darkturquiose”, and “Darkmagenta”) were strongly correlated with both physiological and biochemical traits. To further prioritize modules, we examined pairwise correlation values between each gene within a module and the associated traits (referred to as gene significance, GS), to uncover modules harboring genes with the strongest correlation to traits of interest. This approach identified a suite of genes with high GS (|GS| > 0.65, denoted in Figure 3a ; N = 1,167). In particular, we observed the largest number of genes with strong positive correlation ( N = 213; Darkmagenta) with photosynthetic parameters and oxidative stress markers. In addition, “Darkmagenta” had the highest proportion of genes that passed the GS threshold relative to the overall module, with the exception of the “magenta” module which was correlated with a small number of traits. Thus, the “Darkmagenta” module was prioritized for identifying potential drivers of the drought stress in this panel. Figure 3. Selection of co-expression modules contributing to drought tolerance (a) The number of transcripts per module, correlation between traits and the module ( blue and red scale ), levels of KEGG and GO enrichment ( purple and yellow scale ), and the number of TFs with enriched motifs for each module are displayed. Scales for module-trait correlation values, as well as GO-enrichment levels, are shown to the bottom left. Only those module-trait correlations with a p -value < 0.05 are shown. (b) The eigenvalue of each “Darkmagenta” module, divided by accession, treatment, and time, was plotted. (c) Scatter plot shows module membership (MM) and gene significance (GS) of genes in “Darkmagenta” module, with genes colored based on their strong expression-trait correlation, GO enrichment, or annotation as a stress-responsive transcription factor. To gain insights into the biological processes associated with each of these modules, GO and KEGG pathway enrichment analyses were performed (modules with FDR < 0.05 are displayed, Figure 3a , middle panel). This approach revealed modules with biological processes of interest, such as stress response, translation, metabolic pathways terms. In addition, we performed a motif enrichment analysis using the Arabidopsis DAP-seq motif database (O’Malley et al. 2016) to identify binding motifs for orthologous TFs over-represented in the promoters of genes in the 13 modules ( FDR < 0.05). We then determined which of the enriched motifs matched to TFs found within the corresponding module, suggesting those TFs may target co-expressed genes and potentially serve as a hub TF ( Figure 3a , right panel). Using the Analysis of Motif Enrichment (AME, Bailey et al. 2015), we found the binding motifs of 10 TFs that were significantly enriched in the 2 kb upstream promoter regions of genes in the “Darkmagenta” module. Of these TFs, 4 are annotated as regulators of abiotic stress (e.g. bZIP , ERF , and CDF family TFs; Supplemental Table 6 ), agreeing with our inference that this represents a module of stress-associated genes. We further characterized the “Darkmagenta” module to determine if it exhibited distinct patterns that can be attributed to genotype, condition, or time point. To do so, we plotted the eigenvalue of each accession under each treatment. Using this approach, we observed strong genotypic effects where the eigenvalue for accession “A” separated from the other accession, particularly at weeks 3-7 ( Figure 3b ). This pattern agrees with the distinct leaf-level physiological response of accession “A”, suggesting that the suite of genes in “Darkmagenta” may contribute strongly to drought tolerance ( Figure 1d ), lower photosynthetic capacity/efficiency ( Figure 1e ), and higher level of oxidative stress perceived ( Figure 2d ) under stressed conditions in this genotype. Finally, we examined module membership (MM) of the transcripts in this module relative to their GS to evaluate whether genes strongly associated with physiological traits also occupy central positions within the module ( Figure 3c ). We observed a positive correlation between MM and GS for this module, with genes that were correlated with photosynthesis traits (GS > 0.7) also displaying high module membership (MM > 0.75), indicating that trait-relevant genes tend to be highly connected within the module. In addition, genes annotated with stress response–related GO terms also exhibited elevated MM values ( green dots ), supporting the functional relevance of this module. Among the four AME-enriched stress-response TFs, Sb CDF2/3L showed the highest MM and GS, suggesting a prominent role as a regulatory hub linking transcriptional and photosynthetic trait variation ( Figure 3c ). Collectively, these integrated network analyses identify a core set of highly connected genes with strong trait associations and highlight CDF2/3L as a likely a hub regulator coordinating stress-responsive photosynthetic capacity. binding motif associated genes exhibit a drought stress response pattern Based on the framework described above for identifying core genes within the “Darkmagenta” module, we further examined the hub CDF2/3L TF and its relationship with other trait-associated genes. This gene was previously annotated as SbDof8 (Kushwaha et al. 2011), however, phylogenetic analysis and motif similarity based on Arabidopsis and rice nomenclature places it in the “Cycling Dof” (CDF) subfamily (Lijavetzky et al. 2003), sister to Arabidopsis CDF2 (AT5G39660) and CDF3 (AT3G47500; Figure S4 ). Interestingly, this SbCDF2/3L homolog is the only sorghum CDF/Dof in the “darkmagenta” module ( Figure S5 ). The Arabidopsis CDF3 (Arabidopsis DAP-seq database ID: C2C2dof_tnt.CDF3, TAIR ID: AT3G47500, sorghum ID: Sobic.002G421900, https://phytozome-next.jgi.doe.gov/) has previously been reported in both Arabidopsis and, via heterologous experiments in tomato, as a transcriptional activator that enhances drought, oxidative stress tolerance, carbon and nitrogen metabolism (Noguero et al. 2013; Corrales et al. 2014, 2017; Renau-Morata et al. 2024). Out of the 1,093 genes within the “Darkmagenta” module, 681 (62% in total, e-value < 0.005) were predicted to be bound by the putative SbCDF2/3L TF in their proximal promoters ( Figure 4a ). Careful examination of the binding motifs of those 681 putative CDF2/3L -bound and co-expressed genes revealed that the consensus motif contains the core CDF2/CDF3 binding motif (5’-T/AAAAG-3’; Figure 4a ) reported in previous molecular and biochemical studies from maize (Yanagisawa and Schmidt 1999) and in the Arabidopsis cistrome database (O’Malley et al., 2016). Interestingly, an examination of Sb CDF2/3L ’s expression pattern across the 96 sorghum samples revealed a dramatic decrease in the “A” accession in weeks 3-7 under both control and drought conditions ( Figure 4b ). Additionally, we examined the correlation between Sb CDF2/3L and each of its predicted targets ( N = 412 , selected by random extraction) in the “Darkmagenta” module, a similar number of non-CDF2/3L targeted transcripts within the module, as well as randomly selected transcripts from outside of the “Darkmagenta” module ( N = 412). A significantly higher correlation was observed between CDF2/3L and within-module targets compared with the other two groups ( p = 0.0018 and p < 2.2e-16, Figure 4c ). Lastly, we compared the log 2 FC score (Control/Drought) of CDF2/3L putative targets and non-CDF2/3L targets across all accessions and time points. This comparison revealed a significantly higher log 2 FC ( p = 0.01) of CDF3 putative targets ( Figure 4d) , suggesting these putative target genes are induced under drought conditions in a SbCDF2/3L-dependent manner. Figure 4. Examination of the photosynthesis and stress response related network (a) Proximal (2 kb upstream) promoter elements for the 1,093 genes in the “Darkmagenta” module were examined for motif enrichment by Analysis of Motif Enrichment (AME). The consensus motif for CDF2/3L in these promoters (bottom left), and the number of CDF2/3L targets within the “Darkmagenta” module, are shown. The core motif for the Arabidopsis CDF2/3L is shown to the right. (b) SbCDF2/3L relative abundance (TPM) among genotypes, time points, and treatments. (c) The correlation between SbCDF2/3L and predicted (in-module) target transcripts, non-SbCDF2/3L target in-module transcripts, and randomly selected transcripts from outside the module ( N = 412 for each set) was compared. The number in each group was subset to match the maximum number of non-SbCDF2/3L predicted targets within the “Darkmagenta” module. Significance between groups is shown (Wilcox test, p < 0.05). (d) The log 2 FC(Drought/Control) value of all samples for the CDF2/3L-associated genes and others from ( c ) is shown. Significance (shown) calculated with a Wilcox test. (e) A child network of 230 genes with high module membership was visualized using Cytoscape. Solid grey lines indicate co-expression only, dashed red lines indicate protein-protein interactions derived from the STRING database, and the blue dashed and arrowed lines highlight DAP-seq-derived putative targets of SbCDF2/3L. Nodes (genes) are colored based on trait-association and grouped based on GO-term enrichment. Red stars indicate genes overlapping net CO 2 assimilation QTLs from previous GWAS (Ortiz et al. 2017). The above CDF2/3L DAP-seq motif enrichment data, along with the traits correlated with the “Darkmagenta” module, suggest that SbCDF2/3L is acting as a potential central regulator connecting the drought response in these accessions to downstream photosynthetic traits. To further examine this relationship, we more closely examined the CDF2/3L-motif enriched genes and the physiological and molecular trait associated genes with the “Darkmagenta” module. To do so, we selected the transcripts ( N = 82) with the highest GS score (cut-off: GS > 0.70) and then integrated GO terms, TF-binding, and protein-protein interaction data (PPI, Figure 4e; Supplementary Table 7, STRING: https://string-db.org/ ) resulting in a sub-network of 186 out of the 1,093 transcripts in the original “Darkmagenta” module. Within this sub-network, 129 transcripts were predicted to be CDF2/3L targets based on Analysis of Motif Enrichment (AME; blue dashed lines, Figure 4e ). Of these, ten genes were strongly correlated with CO 2 net assimilation (Blue boxes, Figure 4e ), eleven with flavonoid content (grey boxes, Figure 4e ), four with leaf temperature (orange boxes, Figure 4e ), and six with total antioxidant capacity (peach boxes, Figure 4e ). Based on GO enrichment, ~ 200 genes in the module were associated with GO terms for abiotic stress response and metabolic processes. In addition, several genes with strong trait-transcript abundance correlation in this module coincide with genes previously identified through QTL mapping of net CO 2 assimilation (Ortiz et al., 2017; marked with red stars, Figure 4e, Supplementary Table 6 in the “QTL” column). Thus, this integrated network suggests that SbCDF2/3L is regulating the drought response variation of a number of novel and previously reported stress associated genes in sorghum. Validating the SbCDF2/3L regulatory response in publicly available field stress data We next aimed to determine the robustness of the observed relationship between SbCDF2/3L, its suite of putative target genes, and performance under drought conditions in sorghum. To do so, we analyzed a publicly available sorghum drought time-course RNA-seq dataset (Varoquaux et al. 2019), quantifying relative gene expression (TPM) between a post-flowering drought-tolerant (BT×642) and a drought-sensitive (RT×430) genotype, focusing on core genes identified within the “Darkmagenta” module in our field-derived data. Initially, we grouped the GO term-enriched genes into three classes (Photosystem I, stress response, and metabolic process; Figure 5a; left ) and three groups of genes associated with oxidative stress levels and photosynthetic traits, including flavonoid level, net CO 2 assimilation, and leaf temperature (surface; Figure 5a; right ). We then used the relationship between abundance under treatment and control conditions [log 2 (Drought/Control)] to represent the variation in the treatment response of these six groups of genes over the timespan of the experiment (7 weeks and 8 weeks in the current study and the public data, respectively). This comparative analysis demonstrated that in both experiments, the drought resilient accessions are more likely to harbor elevated expression levels of these trait and stress-response associated transcripts under drought conditions. In contrast, both poor performing accessions (“A” and “RTx430”) displayed reduced levels of transcripts in all six groups of genes with high GS or MM in the “Darkmagenta” module ( Figure 5a ). These data suggests that poor performing sorghum accessions are unable to elicit a stress response and cannot maintain optimal photosynthetic efficiency under drought conditions. Additionally, given the overlap between these two experiments, these genes are likely important markers of a robust drought response in sorghum. Figure 5. Transcriptomic response of key “Darkmagenta” genes in two datasets (a ) Time-course transcriptomic profiles of drought responses, shown as log₂(Drought/Control), derived from a large-scale public sorghum drought dataset (Varoquaux et al. , 2019) and the current study. Genes were grouped based on GO term enrichment (left panels) or trait association (r > 0.7, right panels). For the public dataset, temporal trends were compared between a drought tolerant (BT×642) and sensitive genotype (RT×430) using smoothed spline curves with mean values (bold lines) and standard error (shaded area) displayed. For the current dataset, gene expression distributions at each time point are shown as boxplots, with statistical comparison between accession “A” and the other accessions using the Wilcoxon rank-sum test ( p < 0.05). (b) Relative expression (TPM) of SbCDF2/3L in the public dataset over the 8 weeks of the post-flowering drought experiment between two genotypes under control and drought conditions. (c) Comparison of log₂(Drought/Control) expression changes between the drought-tolerant genotype BT×642 and the drought-sensitive genotype RT×430 across three gene categories: genes lacking predicted CDF3 binding motifs, genes containing predicted CDF2/3L binding motifs, and genes that both contain CDF2/3L binding motifs and show significant correlations with photosynthetic traits using Wilcoxon rank-sum test ( p < 0.05). Given the connection between these trait-associated genes and SbCDF2/3L in the “Darkmagenta” module, we next examined the expression profile of SbCDF2/3L in the drought-tolerant genotype BT×642 and the drought-sensitive genotype RT×430. Although SbCDF2/3L expression was down-regulated in response to drought in both genotypes, the drought-sensitive genotype exhibited significantly lower expression levels compared with the tolerant genotype over the time ( Figure 5b, p < 0.05). This pattern is consistent with the reduced SbCDF2/3L expression observed in accession “A” relative to the other five genotypes ( Figure 4b ), supporting a critical role for SbCDF2/3L in drought-responsive regulation in sorghum. We further assessed drought-responsive expression differences in this public dataset between genes predicted to contain, versus those lacking CDF2/3 binding motifs within the “Darkmagenta” module ( Figure 5c ). Genes harboring putative CDF2/3 binding motifs displayed significantly higher log₂(Drought/Control) ratios in the drought-tolerant BTx642 genotype compared with the sensitive genotype ( p = 0.003), whereas no significant difference was observed for genes without CDF2/3 binding motifs in their 2-kb promoter regions ( p = 0.84). Interestingly, this drought response pattern was further amplified when restricting the analysis to genes that both contain predicted CDF2/3 binding motifs and are correlated with LI-6800–derived photosynthetic traits ( p < 2.2e-16). This pattern is consistent with the transcriptomic response in our field data ( Figure 4c and 4d ). Collectively, these cross-experiment analyses using two independent datasets corroborate SbCDF2/3L’s role as a potential positive regulator of photosynthetic and stress response genes that contribute to drought tolerance in sorghum. Identification of photosynthetic process candidates as targets for enhancing drought tolerance in sorghum Through systems-level data integration and interrogation of public datasets, we uncovered a tight coupling between drought stress responses and a suite of photosynthesis-associated traits and genes. Building on this framework, we next asked if any of these high-confidence candidate genes could be prioritized for functional validation in sorghum. Here, we identified 129 genes by combining genes from the “Darkmagenta” module with high Photosynthesis trait correlation, including net CO 2 assimilation and leaf surface temperature ( Figure 6a, GS > 0.65, left), for GO enrichment. As expected, the results revealed several terms with relevance to light intensity response and photosynthesis ( Figure 6a , right). We focused on the 10 photosynthesis-associated genes as we hypothesized that they might act as a molecular link between drought response and plant performance. Among these 10 photosynthesis term enriched genes ( Figure 6b ), five encode for proteins with numerous reported physical interactions between each other and other proteins within the photosynthetic pathway (Mehari et al. 2021; Cheng et al. 2025; Huang et al. 2025; von Mering et al. 2003, Figure 6b; orange lines ). Furthermore, the enriched set contained metabolic regulators that links photorespiration, carbon metabolism, and ABA signaling to sustain photosynthesis and growth under low CO 2 . (e.g., PPC2, You et al. 2020), as well as downstream signaling proteins that facilitate appropriate responses under stress (e.g., STN7 and PGR5, (Guangxin et al. 2025; Lee et al. 2025). Taking these ten photosynthesis-associated genes, we next compared their expression profiles across both the publicly available large-scale drought dataset and the current field experiment-derived RNA-seq data. These 10 genes showed uniformly lower expression under drought conditions in poor-performing genotypes (RT×430 and accession “A”; Figure 6c ). In contrast, expression of these genes was largely maintained at control-like levels, or only modestly reduced, in the drought-tolerant genotypes ( Figure 6c, bottom right ). Collectively, these data highlight a subset of photosynthesis-associated transcripts that are tightly linked to plant performance under drought stress in sorghum. Figure 6. Multi-trait and biological process integration identifies candidate genes underlying photosynthetic capacity under drought in sorghum (a ) Venn diagram shows genes with high GS (r > 0.65) with net CO₂ assimilation and leaf temperature. Genes associated with these traits were subjected to GO enrichment analysis, revealing significant photosynthesis-related and light-response biological process-enriched genes. (b) Functional categorization of photosynthesis-enriched genes ( N = 10) based on their roles in metabolic (CO₂/HCO₃⁻ fixation), photosynthetic complex assembly, electron transport, and assimilating/regulating photosynthetic performance with downstream stress responses. Orange dashed lines highlight PPIs derived from the STRING database. (c) Drought-responsive expression patterns of photosynthesis-enriched genes from ( a ). Upper panel shows a time-course heatmap (weeks 1–7) of gene expression (Z-score) under control and drought conditions across six genotypes (A–F). Lower left panels depict expression dynamics in the BT×642 and RT×430 background from Varoquaux et al., 2019. Lower right panel shows the mean drought-induced expression change (Z-score) of photosynthesis-enriched genes (Wilcoxon test, p < 0.05). (d) Conceptual model illustrating the integration of field phenotyping and multi-omics analyses to identify drought-responsive networks. Differences between drought-tolerant and drought-sensitive genotypes are reflected at phenotypic (biomass, architecture, photosynthetic capacity, antioxidant levels, oxidative stress status) and transcriptomic level, culminating in the identification of candidate stress response regulators. Discussion Drought stress typically induces complex phenotypic and molecular responses that often lead to reduced biomass, seed set, and photosynthesis, thereby adversely affecting the overall productivity of agricultural crops (Lesk et al. 2022). Sorghum is generally well adapted to hot and dry environments, although there is variation in how different sorghum accessions perform, variation that can be leveraged to shed light on the exact regulatory mechanisms by which sorghum achieves maximal growth under these conditions. Given the complex nature of plant responses to water deprivation, a variety of stress-responsive genes have been identified in sorghum and other plant species (Johnson et al. 2015; Pardo and VanBuren 2021; Pardo et al. 2023; Marks et al. 2024). However, induction of a transcriptional regulator does not imply that its downstream response is beneficial to plant growth and yield. Thus, we still lack an understanding of which of these regulatory mechanisms and their targets contribute to improved agricultural traits in a field setting and therefore could serve as breeding markers or targets for manipulation. We used a representative subset of the genetic and phenotypic diversity found within sorghum to identify molecular mechanisms that might explain the reported resilience to hot and dry environments in this species. Our panel of six distantly related sorghum accessions, grown under hot and arid field conditions, revealed distinct physiological and molecular responses to hot and arid conditions, with some accessions exhibiting minimal reductions in biomass and photosynthetic efficiency, whereas others (e.g. “A”) performed quite poorly. Leaf-level physiological traits tracked closely with oxidative stress status across the six accessions, allowing us to make inferences as to which molecular mechanisms were contributing to phenotypic success ( Figure 6d ). Using a comparative systems-level approach, we identified cohorts of transcripts whose expression patterns were most closely linked to molecular and physiological responses to drought stress ( Figure 4 ). One co-expression module in particular exhibited variation due to both treatment and genotype, signified by a contrasting response between poor and strong performers ( Figure 3b ). This module was enriched for genes involved in photosynthetic processes and abiotic stress responses, consistent with the central role of photosynthetic status as a key physiological indicator of drought stress (Pinheiro and Chaves 2011; Zargar et al. 2017; Zahra et al. 2023). At the regulatory level, we identified one stress-induced TF in particular that was positively associated with photosynthetic traits and plant performance ( SbCDF2/3L; Figure 3 and 4 ). While originally annotated as SbDof8 (Kushwaha et al. 2011), we found that this TF clusters with, and harbors similar Dof motifs as, Arabidopsis and rice CDF2/3 orthologs. Indeed, promoter enrichment and correlation between SbCDF2/3L and target genes across two independent experiments, as well as the lack of other CDF or Dof homologs in the module suggest that SbCDF2/3L is likely a key regulator at the junction between photosynthesis and drought stress responses in sorghum. This assumption is further supported by extensive evidence from previous studies demonstrating coordinated regulation between drought-responsive TFs and photosynthesis-associated target genes under drought conditions (Zhao et al. 2017; Muhammad et al. 2020; Bai et al. 2025). Further functional validation will of course be critical in confirming this model. Of note, a number of Dofs have been found to modulate responses to the environment across the land plant lineage. However, the circadian nature of the Arabidopsis CDFs has only been demonstrated for a few grass Dof TFs, such as RDD1 (Iwamoto et al. 2009; Zhang et al. 2023), and whether this is a common feature of CDF homologs such as SbCDF2/3L and is a crucial aspect of their ability to integrate plant growth under stress conditions remains to be determined. By closely examining photosynthesis-related responses to water limitation in two independent experiments, we delineated the transcriptomic patterns underlying differential drought responses in drought-tolerant and sensitive genotypes ( Figure 5a ). We identified a subset of genes that were both strongly correlated with LI-6800-derived photosynthetic traits and significantly enriched for photosynthesis-related GO terms. Notably, these genes span multiple layers of the photosynthetic process, indicating that drought-induced reductions in photosynthesis are not confined to a single pathway but involve coordinated perturbations across the photosynthetic machinery. Many of these photosynthetic trait-correlated genes are known to be responsive to abiotic stress ( Fig. 6b ). For instance, LHCA proteins involved in capturing and transferring solar energy have been shown to influence drought tolerance in barley, fava bean, cotton, and other species. (Mehari et al. 2021; Cheng et al. 2025; Huang et al. 2025). Furthermore, Psao , Psak , and Psal encode PSI subunits that cooperate with LHCA proteins to stabilize PSI machinery and sustain electron transport under drought stress (Seok et al. 2014; Dudhate et al. 2018). In drought-tolerant plants, PETC generally exhibits a delayed and attenuated downregulation, likely due to improved ROS homeostasis and sustained D1 protein turnover (Silva et al. 2024). At the regulatory level, STN7 links photosynthetic complex formation and leaf senescence in Arabidopsis (Lee et al. 2025) and PGR5 is a key regulator of PSI cyclic electron transport and supports photosystem protection under drought stress (Guangxin et al. 2025). Thus, these genes represent a concerted effort to combat drought stress, one that appears to be regulated largely by a plant-specific transcription factor, SbCDF2/3L, and one that could be promising for further investigation in the context of drought-induced photosynthesis improvement in sorghum. We therefore propose a conceptual model in which modulating a key stress response TF, SbCDF2/3L, specifically under drought conditions, may have positive effects on both drought tolerance and photosynthetic capacity ( Figure 6d ). Dissecting and potentially enhancing this transcriptional regulation in the future offers a promising avenue to sustain photosynthetic performance under water-limited conditions, thereby providing a translational framework for improving drought resilience, biomass accumulation, and yield stability in sorghum production. Data availability Transcriptome sequencing reads are available for download from the National Center for Biotechnology Information (NCBI) BioProject database under the accession number: PRJNA1119650 . Computational pipelines used for data analysis were deposited under the GitHub page (https://rpubs.com/LeonYu/Sorghum-comparative-transcriptomics). The R markdown files (R code) for data visualization in figures were deposited in Rpubs workspace (https://rpubs.com/LeonYu/). Acknowledgments We would like to thank the members within the Nelson, Pauli, and Gregory labs for their insightful comments. We would like to thank Dr. Cliff Weil for insightful comments about the SAP. This work was supported by NSF IOS PGRP 2023310 (to ADLN, DP, and BDG), NSF IOS PGRP 1849708 (to BDG), NSF MCB 2427729 (to BDG), NSF PGRP 2102120 (to ADLN and DP), Department of Energy (DOE) DE-AR0001101 (to DP), DE-SC0023305 (to DP), DE-SC0020401 (to DP), Cotton Incorporated 18-384, 20720, and 21-830 (to DP), NSF DBI-2019674 (to DP and ADLN). Author contributions LY, GM, BDG, ADLN, and DP developed the experiment. GM, and DP designed the field experiment and collected field-related data. ACND extracted sorghum RNA for sequencing. CDG and AEA worked on the microbiome-related experiments. HA and GB generated the biochemical data related to oxidative stress status. LY analyzed all RNA-seq and developed co-expression networks and data integration. LY, BDG, ADLN, and DP wrote the manuscript. Figure Legends Main Figures Figure 1. Drought response diversity of selected sorghum genotypes (a) A field drought experiment was conducted in Maricopa, AZ during the summer of 2020 (left). Drought treatment levels were monitored using soil moisture sensors at different soil depths over time (right). (b) Morphological responses of sorghum accessions under drought, showing plant architecture at the terminal field sampling stage. (c) Microbiome features, represented by 515 classified bacterial taxa and 46 fungal taxa, showing intra-accession variation based on principal component analysis (PCA) at the terminal sampling stage. (d) Physiological measurements of whole-plant, leaf, and root traits under two watering conditions ( N = 5 per accession) were compared using Kruskal test. (e) Bar plots showing photosynthetic efficiency and capacity under drought conditions, measured as Fv′/Fm′ and net CO₂ assimilation, respectively, at each time point for drought-treated plants ( N = 10 per accession). Pairwise Wilcoxon tests were used to compare each accession relative to “A” ( p < 0.05). Figure 2. Molecular response to drought among sorghum accessions (a) A hierarchical clustering (Pearson correlation) of 34,130 transcripts normalized by DESeq2 was used to group transcriptomic features among the 96 samples taken over the course of the experiment (Weeks 1, 3, 5, 7; drought vs control). The scale indicates the Pearson correlation coefficient (PCC) score. (b) A line graph representing the number of up (circle) or down (triangle) regulated genes over time. (c) The enrichment levels of GO terms that were shared between DEGs from at least two accessions are shown. GO terms must have been enriched to at least a FDR < 0.05 in order to be shown. The scale of enrichment is shown to the left. (d) Pairwise Pearson correlation coefficient (PCC) of 24 chemicals associated with antioxidant metabolites, antioxidant enzymes, and oxidative stress markers were displayed by heatmap with the three chemical classes denoted. Red dashed boxes show correlation of metabolites belonging to same class and red solid line boxes highlight correlation between two classes. (e) Representative two oxidative makers (HPR and GOX) from ( d ) were selected to display ratio between drought/control across time for each of the six genotypes. Figure 3. Selection of co-expression modules contributing to drought tolerance (a) The number of transcripts per module, correlation between traits and the module ( blue and red scale ), levels of KEGG and GO enrichment ( purple and yellow scale ), and the number of TFs with enriched motifs for each module are displayed. Scales for module-trait correlation values, as well as GO-enrichment levels, are shown to the bottom left. Only those module-trait correlations with a p -value < 0.05 are shown. (b) The eigenvalue of each “Darkmagenta” module, divided by accession, treatment, and time, was plotted. (c) Scatter plot shows module membership (MM) and gene significance (GS) of genes in “Darkmagenta” module, with genes colored based on their strong expression-trait correlation, GO enrichment, or annotation as a stress-responsive transcription factor. Figure 4. Examination of the photosynthesis and stress response related network (a) Proximal (2 kb upstream) promoter elements for the 1,093 genes in the “Darkmagenta” module were examined for motif enrichment by AME. The consensus motif for CDF3 in these promoters (bottom left), and the number of CDF3 targets within the “Darkmagenta” module, are shown. The core motif for the Arabidopsis CDF3 is shown to the right. (b) CDF3 relative abundance (TPM) among genotypes, time points, and treatments. (c) The correlation between CDF3 and predicted (in-module) target transcripts, non-CDF3 target in-module transcripts, and randomly selected transcripts from outside the module (N = 412 for each set) was compared. The number in each group was subset to match the maximum number of non-CDF3 predicted targets within the “Darkmagenta” module. Significance between groups is shown (Wilcox test, p < 0.05). (d) The log 2 FC(Drought/Control) value of all samples for the CDF3-associated genes and others from ( c ) is shown. Significance (shown) calculated with a Wilcox test. (e) A child network of 230 genes with high module membership was visualized using Cytoscape. Solid grey lines indicate co-expression only, dashed red lines indicate protein-protein interactions derived from the STRING database, and the blue dashed and arrowed lines highlight DAP-seq-derived putative targets of CDF3. Nodes (genes) are colored based on trait-association and grouped based on GO-term enrichment. Red stars indicate genes overlapping net CO 2 assimilation QTLs from previous GWAS (Ortiz et al. 2017). Figure 5. Transcriptomic response of key “Darkmagenta” genes in two datasets (a ) Time-course transcriptomic profiles of drought responses, shown as log₂(Drought/Control), derived from a large-scale public sorghum drought dataset (Varoquaux et al. , 2019) and the current study. Genes were grouped based on GO term enrichment (left panels) or trait association (r > 0.7, right panels). For the public dataset, temporal trends were compared between a drought tolerant (BT×642) and sensitive genotype (RT×430) using smoothed spline curves with mean values (bold lines) and standard error (shaded area) displayed. For the current dataset, gene expression distributions at each time point are shown as boxplots, with statistical comparison between accession “A” and the other accessions using the Wilcoxon rank-sum test ( p < 0.05). (b) Relative expression (TPM) of SbCDF2/3L in the public dataset over the 8 weeks of the post-flowering drought experiment between two genotypes under control and drought conditions. (c) Comparison of log₂(Drought/Control) expression changes between the drought-tolerant genotype BT×642 and the drought-sensitive genotype RT×430 across three gene categories: genes lacking predicted CDF3 binding motifs, genes containing predicted CDF2/3L binding motifs, and genes that both contain CDF2/3L binding motifs and show significant correlations with photosynthetic traits using Wilcoxon rank-sum test ( p 0.65) with net CO₂ assimilation, leaf temperature, and chamber temperature. Genes associated with these traits were subjected to GO enrichment analysis, revealing significant photosynthesis-related and light-response biological processes-enriched genes. (b) Functional categorization of photosynthesis-enriched genes ( N = 10) based on their roles in metabolic (CO₂/HCO₃⁻ fixation), photosynthetic complex assembly, electron transport, and regulatory layers. Orange dash lines highlight PPIs derived from STRING database . (c) Drought-responsive expression patterns of photosynthesis-enriched genes. Upper panel shows a time-course heatmap (weeks 1–7) of gene expression (Z-score) under control and drought conditions across six genotypes (A–F). Lower left panels depict expression dynamics in the BT×642 and RT×430. Lower right panel shows the mean drought-induced expression change (Z-score) of the 10 photosynthesis-enriched genes (Wilcoxon test, p < 0.05). (d) Conceptual model illustrating the integration of field phenotyping and multi-omics analyses to identify drought-responsive networks. Differences between drought-tolerant and drought-sensitive genotypes are reflected at phenotypic (biomass, architecture, photosynthetic capacity, oxidative stress status) and transcriptomic level, culminating in the identification of a stress-response gene network and candidate regulators. Supplementary Figures Figure S1. Physiological response of drought tolerance in sorghum accessions (a) Boxplots show changes of biomass and root phenotypes (Drought/Control) among six accessions using Kruskal test. (b) Bar plots showing photosynthetic efficiency and capacity under control conditions, measured as Fv′/Fm′ and net CO₂ assimilation, respectively, at each time point for drought-treated plants ( N = 10 per accession). Pairwise Wilcoxon tests were used to compare each accession relative to “A” ( p < 0.05). Figure S2. Upset plots of DEGs between two watering conditions (a, b) Comparisons of the up-regulated (left) and down-regulated (right) genes among samples per genotype across each time point, identified in a pair-wise manner. Numbers of DEGs were displayed using the upset plots. The top bar charts denote the shared (or unique) DEGs corresponding to the inclusion of samples below, and bottom-left bars denote the total numbers of DEGs derived from each pair-wise comparison. Figure S3. Stress response profiling of four antioxidant products Boxplots show changes of antioxidant metabolites and enzymes (Drought/Control) over four weeks among six accessions using Kruskal test. Figure S4. Phylogenetic tree of Dof gene family across species Maximum-likelihood (ML) tree was constructed using RAxML with Dof domain genes identified in sorghum, rice, and Arabidopsis. Dof nomenclatural assignment comes from prior publications (see Results). Figure S5. Expression profile of selected Dof genes in sorghum Relative expression (TPM) of four CDF-like Dof genes in sorghum across six accessions over four timepoints under control and drought conditions. References Abreha KB, Enyew M, Carlsson AS, Vetukuri RR, Feyissa T, Motlhaodi T, Ng’uni D, and Geleta M . Sorghum in dryland: morphological, physiological, and molecular responses of sorghum under drought stress. Planta. 2021: 255 (1):20. https://doi.org/10.1007/s00425-021-03799-7 Alizadeh MR, Adamowski J, Nikoo MR, AghaKouchak A, Dennison P, and Sadegh M . A century of observations reveals increasing likelihood of continental-scale compound dry-hot extremes. Sci Adv. 2020: 6 (39):eaaz4571. Bai W, Salih H, Yang R, Yang Q, Jin P, Liang Y, Zhang D, and Li X . ScDREBA5 enhances cold tolerance by regulating photosynthetic and antioxidant genes in the desert moss Syntrichia caninervis. Plant Cell Environ. 2025: 48 (5):3293–3313. https://doi.org/10.1111/pce.15336 Bailey TL, Johnson J, Grant CE, and Noble WS . The MEME suite. Nucleic Acids Res. 2015: 43 (W1):W39-49. https://doi.org/10.1093/nar/gkv416 Bang SW, Lee D-K, Jung H, Chung PJ, Kim YS, Choi YD, Suh J-W, and Kim J-K . Overexpression of OsTF1L, a rice HD-Zip transcription factor, promotes lignin biosynthesis and stomatal closure that improves drought tolerance. Plant Biotechnol J. 2019: 17 (1):118–131. https://doi.org/10.1111/pbi.12951 Benzie I and Strain J . Ferric reducing/antioxidant power assay: direct measure of total antioxidant activity of biological fluids and modified version for simultaneous measurement of total antioxidant power and ascorbic acid concentration. Methods Enzymol. 1999: 299 :15–27. https://doi.org/10.1016/S0076-6879(99)99005-5 Boatwright JL, Sapkota S, Jin H, Schnable JC, Brenton Z, Boyles R, and Kresovich S . Sorghum Association Panel whole-genome sequencing establishes cornerstone resource for dissecting genomic diversity. Plant J. 2022: 111 (3):888–904. https://doi.org/10.1111/tpj.15853 Chen C, Ge F, Du H, Sun Y, Sui Y, Tang S, Shen Z, Li X, Zhang H, Mei C, et al. A comprehensive omics resource and genetic tools for functional genomics research and genetic improvement of sorghum. Mol Plant. 2025: 18 (4):703–719. https://doi.org/10.1016/j.molp.2025.03.005 Cheng J, Wu T, Zhou Y, Al-Saud NBS, Cheng B, Admas T, Zhang W, and Pan R . The alternative splicing of HvLHCA4.2 enhances drought tolerance in barley by regulating ROS scavenging and stomatal closure. Int J Biol Macromol. 2025: 307 (Pt 4):142384. https://doi.org/10.1016/j.ijbiomac.2025.142384 Corrales A-R, Carrillo L, Lasierra P, Nebauer SG, Dominguez-Figueroa J, Renau-Morata B, Pollmann S, Granell A, Molina R-V, Vicente-Carbajosa J, et al. Multifaceted role of cycling DOF factor 3 (CDF3) in the regulation of flowering time and abiotic stress responses in Arabidopsis. Plant Cell Environ. 2017: 40 (5):748–764. https://doi.org/10.1111/pce.12894 Corrales A-R, Nebauer SG, Carrillo L, Fernández-Nohales P, Marqués J, Renau-Morata B, Granell A, Pollmann S, Vicente-Carbajosa J, Molina R-V, et al. Characterization of tomato Cycling Dof Factors reveals conserved and new functions in the control of flowering time and abiotic stress responses. J Exp Bot. 2014: 65 (4):995–1012. https://doi.org/10.1093/jxb/ert451 Dhindsa RS, Plumb-Dhindsa PL, and Reid DM . Leaf senescence and lipid peroxidation: Effects of some phytohormones, and scavengers of free radicals and singlet oxygen. Physiol Plant. 1982: 56 (4):453–457. https://doi.org/10.1111/j.1399-3054.1982.tb04539.x Dudhate A, Shinde H, Tsugama D, Liu S, and Takano T . Transcriptomic analysis reveals the differentially expressed genes and pathways involved in drought tolerance in pearl millet [Pennisetum glaucum (L.) R. Br]. PLoS One. 2018: 13 (4):e0195908. https://doi.org/10.1371/journal.pone.0195908 Fàbregas N and Fernie AR . The metabolic response to drought. J Exp Bot. 2019: 70 (4):1077–1085. https://doi.org/10.1093/jxb/ery437 Furihata H, Zhu Z, Nishida K, Sakuraba Y, Tsuji A, Yamashita H, Nosaki S, Tachibana R, Yamagami A, Ikeda Y, et al. Structural insights into CDF1 accumulation on the CONSTANS promoter via a plant-specific DNA-binding domain. Nat Plants. 2025: 11 (4):836–848. https://doi.org/10.1038/s41477-025-01946-6 Ge SX, Jung D, and Yao R . ShinyGO: a graphical gene-set enrichment tool for animals and plants. Bioinformatics. 2019. https://doi.org/10.1093/bioinformatics/btz931 González M-C, Cejudo FJ, Sahrawy M, and Serrato AJ . Current knowledge on mechanisms preventing photosynthesis redox imbalance in plants. Antioxidants (Basel). 2021: 10 (11):1789. https://doi.org/10.3390/antiox10111789 Guangxin J, Zheyuan W, Jiaqi S, Hongrui Z, Kexin W, Jingjing X, Nan S, Tanhang Z, Siyue Q, Changjun D, et al. The Trx-Prx redox pathway and PGR5/PGRL1-dependent cyclic electron transfer play key regulatory roles in poplar drought stress. Tree Physiol. 2025: 45 (1):tpaf004. https://doi.org/10.1093/treephys/tpaf004 Hasanuzzaman M, Nahar K, Gill SS, and Fujita M . Drought stress responses in plants, oxidative stress, and antioxidant defense. Climate change and plant abiotic stress tolerance. 2013:209–250. Henriques R, Wang H, Liu J, Boix M, Huang L-F, and Chua N-H . The antiphasic regulatory module comprising CDF5 and its antisense RNA FLORE links the circadian clock to photoperiodic flowering. New Phytol. 2017: 216 (3):854–867. https://doi.org/10.1111/nph.14703 Hodges DM, DeLong JM, Forney CF, and Prange RK . Improving the thiobarbituric acid-reactive-substances assay for estimating lipid peroxidation in plant tissues containing anthocyanin and other interfering compounds. Planta. 1999: 207 (4):604–611. https://doi.org/10.1007/s004250050524 Huang T, Dong L, Han S, Han X, Yin J, Hou L, and Liu Y . Comprehensive analysis of Vicia faba light-harvesting chlorophyll a/b binding protein (Lhc) revealed the roles of VfLhcb1.5, VfLhcb3.3, VfLhcb4, and VfLhca4 in photosynthesis and stress tolerance. Plant Stress. 2025: 18 (101052):101052. https://doi.org/10.1016/j.stress.2025.101052 Iwamoto M, Higo K, and Takano M . Circadian clock- and phytochrome-regulated Dof-like gene, Rdd1, is associated with grain size in rice. Plant Cell Environ. 2009: 32 (5):592–603. https://doi.org/10.1111/j.1365-3040.2009.01954.x Johnson SM, Cummins I, Lim FL, Slabas AR, and Knight MR . Transcriptomic analysis comparing stay-green and senescent Sorghum bicolor lines identifies a role for proline biosynthesis in the stay-green trait. J Exp Bot. 2015: 66 (22):7061–7073. https://doi.org/10.1093/jxb/erv405 Kawahara Y, de la Bastide M, Hamilton JP, Kanamori H, McCombie WR, Ouyang S, Schwartz DC, Tanaka T, Wu J, Zhou S, et al. Improvement of the Oryza sativa Nipponbare reference genome using next generation sequence and optical map data. Rice (N Y). 2013: 6 (1):4. https://doi.org/10.1186/1939-8433-6-4 Keech O, Gardeström P, Kleczkowski LA, and Rouhier N . The redox control of photorespiration: from biochemical and physiological aspects to biotechnological considerations. Plant Cell Environ. 2017: 40 (4):553–569. https://doi.org/10.1111/pce.12713 Kim D, Paggi JM, Park C, Bennett C, and Salzberg S . Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol. 2019: 37 :907–915. https://doi.org/10.1038/s41587-019-0201-4 Kim J-S, Kidokoro S, Yamaguchi-Shinozaki K, and Shinozaki K . Regulatory networks in plant responses to drought and cold stress. Plant Physiol. 2024: 195 (1):170–189. https://doi.org/10.1093/plphys/kiae105 Kurai T, Wakayama M, Abiko T, Yanagisawa S, Aoki N, and Ohsugi R . Introduction of the ZmDof1 gene into rice enhances carbon and nitrogen assimilation under low-nitrogen conditions: ZmDof1 enhances C and N assimilation in rice. Plant Biotechnol J. 2011: 9 (8):826–837. https://doi.org/10.1111/j.1467-7652.2011.00592.x Kushwaha H, Gupta S, Singh VK, Rastogi S, and Yadav D . Genome wide identification of Dof transcription factor gene family in sorghum and its comparative phylogenetic analysis with rice and Arabidopsis. Mol Biol Rep. 2011: 38 (8):5037–5053. https://doi.org/10.1007/s11033-010-0650-9 Lee S, Park J, Back S, Poudyal RS, Mun D-G, Bae J, Gwak N, Lee S-W, Hwang D, and Kim Y . Identification of STN7-associated leaf senescence networks in thylakoid membranes of Arabidopsis. iScience. 2025: 28 (6):112577. https://doi.org/10.1016/j.isci.2025.112577 Lesk C, Anderson W, Rigden A, Coast O, Jägermeyr J, McDermid S, Davis KF, and Konar M . Compound heat and moisture extreme impacts on global crop yields under climate change. Nat Rev Earth Environ. 2022: 3 (12):872–889. https://doi.org/10.1038/s43017-022-00368-8 Liao Y, Smyth G, and Shi W . featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2013: 30 (7):923–930. https://doi.org/10.1093/bioinformatics/btt656 Lijavetzky D, Carbonero P, and Vicente-Carbajosa J . Genome-wide comparative phylogenetic analysis of the rice and Arabidopsis Dof gene families. BMC Evol Biol. 2003: 3 (1):17. https://doi.org/10.1186/1471-2148-3-17 Love MI, Huber W, and Anders S . Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014: 15 (12):550. https://doi.org/10.1186/s13059-014-0550-8 Manna M, Thakur T, Chirom O, Mandlik R, Deshmukh R, and Salvi P . Transcription factors as key molecular target to strengthen the drought stress tolerance in plants. Physiol Plant. 2021: 172 (2):847–868. Marks RA, Van Der Pas L, Schuster J, Gilman IS, and VanBuren R . Convergent evolution of desiccation tolerance in grasses. Nat Plants. 2024: 10 (7):1112–1125. https://doi.org/10.1038/s41477-024-01729-5 McCormick RF, Truong SK, Sreedasyam A, Jenkins J, Shu S, Sims D, Kennedy M, Amirebrahimi M, Weers BD, McKinley B, et al. The Sorghum bicolor reference genome: improved assembly, gene annotations, a transcriptome atlas, and signatures of genome organization. Plant J. 2018: 93 (2):338–354. https://doi.org/10.1111/tpj.13781 Mehari TG, Xu Y, Magwanga RO, Umer MJ, Kirungu JN, Cai X, Hou Y, Wang Y, Yu S, Wang K, et al. Genome wide identification and characterization of light-harvesting Chloro a/b binding (LHC) genes reveals their potential role in enhancing drought tolerance in Gossypium hirsutum. J Cott Res. 2021: 4 (1). https://doi.org/10.1186/s42397-021-00090-8 Melandri G, AbdElgawad H, Riewe D, Hageman JA, Asard H, Beemster GTS, Kadam N, Jagadish K, Altmann T, Ruyter-Spira C, et al. Biomarkers for grain yield stability in rice under drought stress. J Exp Bot. 2020: 71 (2):669–683. https://doi.org/10.1093/jxb/erz221 von Mering C, Huynen M, Jaeggi D, Schmidt S, Bork P, and Snel B . STRING: a database of predicted functional associations between proteins. Nucleic Acids Res. 2003: 31 (1):258–261. https://doi.org/10.1093/nar/gkg034 Morris GP, Ramu P, Deshpande SP, Hash CT, Shah T, Upadhyaya HD, Riera-Lizarazu O, Brown PJ, Acharya CB, Mitchell SE, et al. Population genomic and genome-wide association studies of agroclimatic traits in sorghum. Proc Natl Acad Sci U S A. 2013: 110 (2):453–458. https://doi.org/10.1073/pnas.1215985110 Muhammad I, Shalmani A, Ali M, Yang Q-H, Ahmad H, and Li FB . Mechanisms regulating the dynamics of photosynthesis under abiotic stresses. Front Plant Sci. 2020: 11 :615942. https://doi.org/10.3389/fpls.2020.615942 Nelissen H, Sun X-H, Rymen B, Jikumaru Y, Kojima M, Takebayashi Y, Abbeloos R, Demuynck K, Storme V, Vuylsteke M, et al. The reduction in maize leaf growth under mild drought affects the transition between cell division and cell expansion and cannot be restored by elevated gibberellic acid levels. Plant Biotechnol J. 2018: 16 (2):615–627. https://doi.org/10.1111/pbi.12801 Nelson ADL, Haug-Baltzell AK, Davey S, Gregory BD, and Lyons E . EPIC-CoGe: managing and analyzing genomic data. Bioinformatics. 2018: 34 (15):2651–2653. https://doi.org/10.1093/bioinformatics/bty106 Noguero M, Atif RM, Ochatt S, and Thompson RD . The role of the DNA-binding One Zinc Finger (DOF) transcription factor family in plants. Plant Sci. 2013: 209 :32–45. https://doi.org/10.1016/j.plantsci.2013.03.016 O’Malley RC, Huang S-SC, Song L, Lewsey MG, Bartlett A, Nery JR, Galli M, Gallavotti A, and Ecker JR . Cistrome and epicistrome features shape the regulatory DNA landscape. Cell. 2016: 165 (5):1280–1292. https://doi.org/10.1016/j.cell.2016.04.038 Ortiz D, Hu J, and Salas Fernandez MG . Genetic architecture of photosynthesis in Sorghum bicolor under non-stress and cold stress conditions. J Exp Bot. 2017: 68 (16):4545–4557. https://doi.org/10.1093/jxb/erx276 Ortiz D and Salas-Fernandez MG . Dissecting the genetic control of natural variation in sorghum photosynthetic response to drought stress. J Exp Bot. 2022: 73 (10):3251–3267. https://doi.org/10.1093/jxb/erab502 Pardo J and VanBuren R . Evolutionary innovations driving abiotic stress tolerance in C4 grasses and cereals. Plant Cell. 2021: 33 :3391–3401. https://doi.org/10.1093/plcell/koab205 Pardo J, Wai CM, Harman M, Nguyen A, Kremling KA, Romay MC, Lepak N, Bauerle TL, Buckler ES, Thompson AM, et al. Cross-species predictive modeling reveals conserved drought responses between maize and sorghum. Proc Natl Acad Sci U S A. 2023: 120 (10):e2216894120. https://doi.org/10.1073/pnas.2216894120 Pinheiro C and Chaves MM . Photosynthesis and drought: can we make metabolic connections from available data? J Exp Bot. 2011: 62 (3):869–882. https://doi.org/10.1093/jxb/erq340 Qiu T, Wei S, Fang K, Zhang M, Li Y, Feng Y, Cheng Y, Zhang S, Tian J, Gao A, et al. The atypical Dof transcriptional factor OsDes1 contributes to stay-green, grain yield, and disease resistance in rice. Sci Adv. 2024: 10 (34):eadp0345. https://doi.org/10.1126/sciadv.adp0345 Renau-Morata B, Jiménez-Benavente E, Gil-Villar D, Cebolla-Cornejo J, Romero-Hernández G, Carrillo L, Vicente-Carbajosa J, Medina J, Molina RV, and Nebauer SG . Arabidopsis CDF3 transcription factor increases carbon and nitrogen assimilation and yield in trans-grafted tomato plants. Plant Physiol Biochem. 2024: 210 (108607):108607. https://doi.org/10.1016/j.plaphy.2024.108607 Saibo NJM, Lourenço T, and Oliveira MM . Transcription factors and regulation of photosynthetic and related metabolism under environmental stresses. Ann Bot. 2009: 103 (4):609–623. https://doi.org/10.1093/aob/mcn227 Sakhno LO, Yemets AI, and Blume YB . The role of ascorbate‐glutathione pathway in reactive oxygen species balance under abiotic stresses. Reactive Oxygen, Nitrogen and Sulfur Species in Plants. 2019:89–111. https://doi.org/10.1002/9781119468677.ch4 Sato H, Mizoi J, Shinozaki K, and Yamaguchi-Shinozaki K . Complex plant responses to drought and heat stress under climate change. Plant J. 2024: 117 (6):1873–1892. https://doi.org/10.1111/tpj.16612 Schwitzguebel JP and Siegenthaler PA . Purification of peroxisomes and mitochondria from spinach leaf by percoll gradient centrifugation. Plant Physiol. 1984: 75 (3):670–674. https://doi.org/10.1104/pp.75.3.670 Seok MS, You YN, Park HJ, Lee SS, Aigen F, Luan S, Ahn JC, and Cho HS . AtFKBP16-1, a chloroplast lumenal immunophilin, mediates response to photosynthetic stress by regulating PsaL stability. Physiol Plant. 2014: 150 (4):620–631. https://doi.org/10.1111/ppl.12116 Silva CRL, de Souza CB, dos Santos CM, Floreste BS, Zani NC, Hoshino-Bezerra AA, Bueno GC, Chagas EBR, and Menossi M . The sugarcane ScPetC gene improves water-deficit and oxidative stress tolerance in transgenic tobacco plants. Agronomy (Basel). 2024: 14 (7):1371. https://doi.org/10.3390/agronomy14071371 Singh A, Newton L, Schnable JC, and Thompson AM . Unveiling shared genetic regulators of plant architectural and biomass yield traits in the Sorghum Association Panel. J Exp Bot. 2025: 76 (6):1625–1643. https://doi.org/10.1093/jxb/eraf012 Steczko J, Donoho GA, Dixon JE, Sugimoto T, and Axelrod B . Effect of ethanol and low-temperature culture on expression of soybean lipoxygenase L-1 in Escherichia coli. Protein Expr Purif. 1991: 2 (2–3):221–227. https://doi.org/10.1016/1046-5928(91)90075-t Tsehaye Y, Menamo TM, Abay F, Tadesse T, and Bantte K . Multi-locus genome-wide association study for grain yield and drought tolerance indices in sorghum accessions. Plant Genome. 2024: 17 (4):e20505. https://doi.org/10.1002/tpg2.20505 Varoquaux N, Cole B, Gao C, Pierroz G, Baker CR, Patel D, Madera M, Jeffers T, Hollingsworth J, Sievert J, et al. Transcriptomic analysis of field-droughted sorghum from seedling to maturity reveals biotic and metabolic responses. Proc Natl Acad Sci U S A. 2019: 116 (52):27124–27132. https://doi.org/10.1073/pnas.1907500116 Yanagisawa S and Schmidt RJ . Diversity and similarity among recognition sequences of Dof transcription factors. Plant J. 1999: 17 (2):209–214. https://doi.org/10.1046/j.1365-313x.1999.00363.x You L, Zhang J, Li L, Xiao C, Feng X, Chen S, and Hu . Involvement of abscisic acid, ABI5, and PPC2 in plant acclimation to low CO2. Journal of Experimental Botany. 2020: 71 (14):4093–4108. Yu L, Dittrich ACN, Zhang X, Brock JR, Thirumalaikumar VP, Melandri G, Skirycz A, Edger PP, Thorp KR, Hinze L, et al. Regulation of a single inositol 1-phosphate synthase homeologue by HSFA6B contributes to fibre yield maintenance under drought conditions in upland cotton. Plant Biotechnol J. 2024: 22 (10):2756–2772. https://doi.org/10.1111/pbi.14402 Yu L, Melandri G, Dittrich AC, Calleja S, Rozzi B, Ganguly DR, Palos K, Srinivasan A, Brewer EK, Fischer H, et al. The mRNA covalent modification dihydrouridine regulates transcript turnover and photosynthetic capacity during plant abiotic stress. bioRxiv. 2025:2025.01. 17.633510. https://doi.org/10.1101/2025.01.17.633510 Zahra N, Hafeez MB, Kausar A, Al Zeidi M, Asekova S, Siddique KHM, and Farooq M . Plant photosynthetic responses under drought stress: Effects and management. J Agron Crop Sci. 2023: 209 (5):651–672. https://doi.org/10.1111/jac.12652 Zargar SM, Gupta N, Nazir M, Mahajan R, Malik FA, Sofi NR, Shikari AB, and Salgotra RK . Impact of drought on photosynthesis: Molecular perspective. Plant Gene. 2017: 11 :154–159. Zhang P, Gao S, Jin J, Li H, Zhang Y, Nie X, Ma T, Wang H, Nadeem B, Shi P, et al. Suppression of the transcription factor TaDOF7.6 enhances photosynthesis and energy metabolism to boost wheat yield. J Agric Food Chem. 2025: 73 (26):16660–16671. https://doi.org/10.1021/acs.jafc.5c03339 Zhang Y, Chen G, Deng L, Gao B, Yang J, Ding C, Zhang Q, Ouyang W, Guo M, Wang W, et al. Integrated 3D genome, epigenome and transcriptome analyses reveal transcriptional coordination of circadian rhythm in rice. Nucleic Acids Res. 2023: 51 (17):9001–9018. https://doi.org/10.1093/nar/gkad658 Zhao P, Cui R, Xu P, Wu J, Mao J-L, Chen Y, Zhou C-Z, Yu L-H, and Xiang C-B . ATHB17 enhances stress tolerance by coordinating photosynthesis associated nuclear gene and ATSIG5 expression in response to abiotic stress. Sci Rep. 2017: 7 (1):45492. https://doi.org/10.1038/srep45492 Zheng Y, Jiao C, Sun H, Rosli HG, Pombo MA, Zhang P, Banf M, Dai X, Martin GB, Giovannoni JJ, et al. ITAK: A program for genome-wide prediction and classification of plant transcription factors, transcriptional regulators, and protein kinases. Mol Plant. 2016: 9 (12):1667–1670. https://doi.org/10.1016/j.molp.2016.09.014 Zhu J-K . Abiotic stress signaling and responses in plants. Cell. 2016: 167 (2):313–324. https://doi.org/10.1016/j.cell.2016.08.029 Information & Authors Information Version history V1 Version 1 02 February 2026 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords drought response photosynthesis photosynthesis: carbon reactions sorghum systems biology transcriptome Authors Affiliations Li’ang Yu 0000-0002-9556-011X Boyce Thompson Institute for Plant Research View all articles by this author Giovanni Melandri The University of Arizona School of Plant Sciences View all articles by this author Anna C. Nelson Dittrich Boyce Thompson Institute for Plant Research View all articles by this author Hamada AbdElgawad 0000-0001-9764-9006 Universiteit Antwerpen Geintegreerd Moleculair Plantenfysiologisch Research View all articles by this author Gerrit Beemster 0000-0001-6014-053X Universiteit Antwerpen Geintegreerd Moleculair Plantenfysiologisch Research View all articles by this author Ciara Denise Garcia The University of Arizona School of Plant Sciences View all articles by this author A. Elizabeth Arnold The University of Arizona School of Plant Sciences View all articles by this author Brian D. Gregory University of Pennsylvania Department of Biology View all articles by this author Duke Pauli The University of Arizona School of Plant Sciences View all articles by this author Andrew Nelson 0000-0001-9896-1739 [email protected] Boyce Thompson Institute for Plant Research View all articles by this author Metrics & Citations Metrics Article Usage 221 views 101 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Li’ang Yu, Giovanni Melandri, Anna C. Nelson Dittrich, et al. Integrated multi-omic analyses uncover a regulatory link between photosynthesis and drought tolerance in field-grown sorghum. Authorea . 02 February 2026. DOI: https://doi.org/10.22541/au.177001054.41782127/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.177001054.41782127/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a00c047879a60db4',t:'MTc3OTYyMjc0MQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();