Evaporative cooling signals for wound healing in plants

A CBF–SA module links wound-induced evaporative cooling to tissue repair in plants | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results A CBF–SA module links wound-induced evaporative cooling to tissue repair in plants Joseph Balem , Chenjiao Tan , Nathália Cássia Ferreira Dias , Mariah Arnold , Sorrel Tran , Paul M. Severns , Paulo J.P.L. Teixeira , Changying Li , View ORCID Profile Li Yang doi: https://doi.org/10.1101/2025.05.23.655667 Joseph Balem 1 Department of Plant Pathology, College of Agricultural & Environmental Sciences, University of Georgia , Athens 30602, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Chenjiao Tan 2 Department of Agricultural and Biological Engineering, Institute of Food and Agricultural Sciences, University of Florida , Gainesville, FL, 32611, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Nathália Cássia Ferreira Dias 3 Luiz de Queiroz College of Agriculture, University of São Paulo Find this author on Google Scholar Find this author on PubMed Search for this author on this site Mariah Arnold 1 Department of Plant Pathology, College of Agricultural & Environmental Sciences, University of Georgia , Athens 30602, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Sorrel Tran 1 Department of Plant Pathology, College of Agricultural & Environmental Sciences, University of Georgia , Athens 30602, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Paul M. Severns 1 Department of Plant Pathology, College of Agricultural & Environmental Sciences, University of Georgia , Athens 30602, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Paulo J.P.L. Teixeira 3 Luiz de Queiroz College of Agriculture, University of São Paulo Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: li.yang1{at}uga.edu Changying Li 2 Department of Agricultural and Biological Engineering, Institute of Food and Agricultural Sciences, University of Florida , Gainesville, FL, 32611, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: li.yang1{at}uga.edu Li Yang 1 Department of Plant Pathology, College of Agricultural & Environmental Sciences, University of Georgia , Athens 30602, USA 4 Department of Plant Biology, Franklin College of Arts and Sciences, University of Georgia , Athens 30602, USA 5 The Plant Center, University of Georgia , Athens 30602, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Li Yang For correspondence: li.yang1{at}uga.edu Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract Repairing damaged tissues is essential for the survival of all organisms. In plants, tissue injury rapidly triggers defense and repair programs. However, the molecular mechanisms linking early injury cue to the later stages of wound repair remain unclear. Here, we show that wounding of Arabidopsis leaves induces localized low temperature at the injury site, likely caused by evaporative cooling, which is accompanied by an activation of cold-responsive genes. Using thermal imaging combined with computer vision and deep learning, we developed a workflow to monitor the dynamics of wound healing in a quantitative, non-invasive and real-time manner. Mechanistically, we show that C-repeat Binding Factor (CBF) transcription factors are required for the activation of injury-associated cold response and downstream salicylic acid (SA) signaling. The CBF–SA module promotes lignin deposition and wound repair. Together, these findings reveal a link between a wound-induced biophysical cue and the tissue repair program. Introduction Responding to and repairing tissue damage is essential for an organism’s survival. In nature, wounds to aboveground plant tissues can be caused by abiotic sources such as hail and strong winds, as well as biotic interactions such as herbivore feeding. Natural developmental processes, such as lateral root emergence and leaf shedding, also leave injured tissues. In addition, wounds are purposefully created in agricultural and horticultural practices such as grafting, pruning, harvesting and tissue regeneration. Before wounds are fully healed, their presence may not only impair proper organ functioning, but can become entry points for microbial pathogens, resulting in plant disease. Therefore, wound healing is an important adaptive response to tissue damage that occurs throughout a plant’s lifespan. The wound response in Arabidopsis thaliana involves a rapid and dynamic network of signaling pathways that limits damage, suppresses pathogen invasion, and promotes tissue repair ( Schilmiller & Howe, 2005 ; Ikeuchi et al ., 2020 ; Vega-Muñoz et al ., 2020 ). When injured, damaged cells immediately release damage-associated molecular patterns (DAMPs), such as cell wall fragments and cytosolic content (e.g. ATP, Glutathione) ( Tanaka & Heil, 2021 ). The perception of DAMPs by neighboring cells triggers a cascade of intracellular signaling events, including calcium ion influx and the production of reactive oxygen species (ROS) ( De Lorenzo et al ., 2018 ; Tanaka & Heil, 2021 ). Mitogen-activated protein kinase (MAPK) cascades, such as MPK3 and MPK6, transmit damage signals to downstream responses ( Beckers et al ., 2009 ; Zhang & Zhang, 2022 ). Phytohormones, particularly jasmonic acid (JA), play key roles in modulating wound responses ( Koo & Howe, 2009 ). The role of mechanical barriers, such as cuticle repair and suberization, contributes to sealing the wound site and preventing water loss or pathogen entry ( Serrano et al ., 2014 ; Rains et al ., 2022 ). Thus, wound-triggered deposition of callose and lignin are usually considered as hallmarks of the healing process ( Jacobs et al ., 2003 ; Xu et al ., 2024 ), although the threshold levels of callose and lignin required for effective wound sealing remain unclear. Sensing and responding to temperature fluctuations is a crucial ability for plants adapting to a changing environment. C-repeat Binding Factor (CBF) transcription factors act as central regulator of cold response in Arabidopsis ( Cook et al ., 2004 ; Jia et al ., 2016 ; Hwarari et al ., 2022 ). Mutations in CBF genes lead to various defects related to sensing cold temperature including reduced freezing tolerance, altered flowering time, and reduced accumulation of osmoprotectants ( Gilmour et al ., 1998 ; Jia et al ., 2016 ; Hwarari et al ., 2022 ), which are acclimatization and adaptive responses to low temperature. CBFs bind to a C-repeat/Dehydration Responsive Element (CRT/DRE) motif in the promoters of genes responsive to cold stresses including a variety of proteins annotated as Cold Regulated genes (COR genes) ( Stockinger et al ., 1997 ; Liu et al ., 1998 ; Park et al ., 2015 ; Hwarari et al ., 2022 ). In different temperature ranges, COR genes have multiple functions in stabilizing membrane phospholipids, maintaining hydrophobic interactions, ion homeostasis, and scavenging ROS ( Artus et al ., 1996 ; Steponkus et al ., 1998 ; Puhakainen et al ., 2004 ; Bozovic et al ., 2013 ; Wang et al ., 2017 ; Ortega et al ., 2024 ). While significant progress has been made in understanding the wound response in plants, a critical knowledge gap remains in elucidating the mechanisms that drive the later stages, particularly wound healing. Currently, there is a lack of robust, quantitative methods to measure and track wound healing dynamics, limiting our ability to connect early wound signaling events to the final healing step. Studies have demonstrated the potential of Deep Learning (DL) in diverse applications, including plant phenotyping ( Murphy et al ., 2024 ), stress detection ( Islam et al ., 2024b ), and yield prediction ( Jhajharia et al ., 2023 ). Advances in DL, particularly convolutional neural networks (CNNs), have enabled precise and efficient segmentation and detection of leaves from complex images. For instance, U-Net and its variants have been widely used for leaf segmentation, achieving high accuracy in splitting leaves from complex backgrounds ( Bhagat et al ., 2022 ; Deng et al ., 2023 ; Fu et al ., 2023 ; Yi et al ., 2023 ; Zhang et al ., 2026 Zhang, 2026 #6200). Additionally, instance segmentation models such as Mask R-CNN have proven effective in parsing regions within individual leaves ( Chowdhury et al ., 2024 ; Huang et al ., 2024 ; Islam et al ., 2024a ; Alfred et al ., 2025 ), allowing for accurate quantification of leaf traits. These methods have significantly reduced the manual effort to extract features and improved the accuracy of image-based monitoring, making high-throughput phenotyping feasible. For instance, mask R-CNN was employed to segment plants in the thermal images, achieving an error ≤ 0.5 ℃ ( Jiang et al ., 2018 ). Similarly, Mask R-CNN was utilized to segment apples in RGB images. The segmentation masks were then mapped onto the thermal images to measure apple temperatures, achieving an error < 0.5 ℃ ( Amogi et al ., 2024 ). Here we report that a DL-based workflow was capable of accurately monitoring wound-induced temperature differences. We present results demonstrating that wound-induced evaporative cooling is necessary and sufficient to induce a localized cold response. The key players in cold response, CBF transcription factors, were required for the activation of wound-induced cold response locally. In addition, CBFs also contribute to the wound-induced SA response. Mutations in both CBFs and SA biosynthesis/signaling genes led to delayed healing and reduced lignin deposition at wound sites. Based on these observations, we propose that CBF-mediated temperature sensing pathway may relay the wound-associated evaporative cooling signal to SA-governed wound repair in Arabidopsis. Results Cold-response genes were activated following wounding To investigate the early transcriptional response to wounding, we analyzed the transcriptomic changes at 30 minutes and 60 minutes after cutting in detached Arabidopsis leaves. Compared to the control condition (time 0), a total of 6132 genes were differentially expressed at one or both time points. Specifically, 4522 genes were differentially expressed at 30 minutes, 5107 at 60 minutes, and 3497 genes were shared between the two time points. Hierarchical clustering grouped these wound-responsive genes into eight distinct clusters based on their expression profiles ( Figure 1A , Supplementary Table 1). Clusters 1-6 comprised 3154 genes that were up-regulated in response to wounding, while Clusters 7 and -8 included 2978 genes that were down-regulated ( Figure 1B ). As expected, JA-responsive genes were prominently activated upon wounding ( Supplementary Figure 1A ). Interestingly, genes associated with cold response were significantly enriched among the up-regulated genes, particularly in Clusters 1, 2 and 6 ( Supplementary Figure 1 ). Notably, Cluster 2 and 6 included well-known marker genes of the cold response (i.e., COR6.6 , RD29A , KIN1 , COR47A , COR15A , and COR15B ) ( Figure 1A and 1C ), indicating that wounding activated a cold-response-like program in plants. To validate these findings, we cross-referenced our results with published transcriptomes from wounded Arabidopsis hypocotyls ( Ikeuchi et al ., 2017 ), leaf blade ( Fiorucci et al ., 2022 ) and petiole ( Kilian et al ., 2007 ; Liu et al ., 2022 ). Consistent with our observations, cold-responsive marker genes were also up-regulated in these datasets ( Supplementary Figure 1B ). Together, these results suggest that activation of the cold-response pathway is associated with the Arabidopsis wound response. Download figure Open in new tab Figure 1: Cold-responsive genes are up-regulated in wounded tissues. A) Hierarchical clustering of the 6132 genes that responded to wounding at one or more time points. 1.5 fold-changes for the comparisons ‘T30 vs. T0’ and ‘T60 vs. T0’ are shown on the right. Statistically significant differentially expressed genes (DEG) are marked with blue stripes. Notably, cold-responsive marker genes are enriched in Cluster 2 and 6, indicating their up-regulation following wounding. B) Expression patterns of the eight gene clusters identified in the heatmap. Clusters 1–6 represent genes up-regulated in response to wounding, while Clusters 7 and 8 comprise down-regulated genes. C) Expression profiles of six cold-responsive marker genes: COR6.6 ( KIN2 ), COR15A , COR15B , COR47A , KIN1 , and RD29A . Red bars indicate the mean, and error bars represent standard deviation. FDR (False Discovery Rate) values are indicated for each comparison. Wound sites had low temperature compared to leaf blade Given the activation of cold-responsive genes in wounded tissues, we directly measured the temperature of a wounded leaf to determine if there was a causal reduction in temperature after wounding. We cut the tip of Arabidopsis leaves grown under a short-day condition in soil. The real-time temperature profile was monitored using an infra-red, thermal imaging camera ( Figure 2A ). Consistent with the gene expression results ( Figure 1 ), infra-red thermal imaging revealed an immediate temperature reduction (within seconds) at the cut sites ( Figure 2B , Supplementary Video 1). On average, the local cooling at 5 minutes following wounding was ∼ 1.5 °C lower than the rest of the leaf ( Figure 2B ). We also observed similar immediate temperature reductions following wounding in Solanum lycopersicum ( Sl ), N. benthamina ( Nb ), Epipremnum aureum ( Ea ) and Kalanchoe daigremontiana ( Kd ) ( Figure 2C ). In addition to these room temperature experiments (22°C), wound-induced cooling was also observed when the environmental temperature was at 4°C and 40°C ( Supplementary Figure 2 ). To confirm the cooling effect at wound sites was not due to thermal imaging error, we used a fine-tip thermocouple to measure temperature at wound sites ( Figure 2D ) and found that the hand measured thermocouple temperatures also displayed an average of ∼ 1.5 °C lower at the wound site ( Figure 2E ). Thus, two independent methods confirmed that an immediate cooling occurs at the wound site on a leaf, which coincided with the activation of cold-responsive genes ( Figure 1 ). Download figure Open in new tab Figure 2: Localized temperature reduction at leaf wounding sites. A) Arabidopsis plants were imaged in a photobooth using Fotric’s 616C R&D thermal camera. B) Within 2-5 seconds after mechanical wounding, cutting sites exhibited a temperature reduction of 1.5-2.0°C, as indicated by the dark purple hue of the wounding sites in the thermal image. C) Temperature reduction at cutting sites were observed in Epipremnum aureum ( Ea ), Nicotiana benthamiana ( Nb ), Kalanchoe daigremontiana ( Kd ), and Solanum lycopersicum ( Sl ). Upper panels: RGB images; Lower panels: thermal images. D) Measuring wounding site and leaf blade temperature using a fine-tip thermocouple. E) Reduced temperature at wound sites were detected in various plants using a thermocouple. *:p<0.05 using unpaired t-test with a two-tailed hypothesis. W: wound; L: leaf. Each dot represents a measurement from one leaf. Since the temperature reduction was only observed at the cutting site (∼ within 1-2 mm of the wound), we further monitored the spatiotemporal pattern of a cold responsive promoter ( proCOR15A::GUS ) after wounding ( Baker et al ., 1994 ). COR15A is induced by freezing ( Steponkus et al ., 1998 ) and low ambient temperatures ( Franklin & Whitelam, 2007 ). Notably, it was also among the wound-induced genes in our RNA-seq experiment ( Figure 1A and 1C ). We observed an induction of COR15A promoter activity at as early as 10 mins after cutting, and the staining signal was preferentially detected at the cutting site ( Figure 3A ). The localized expression pattern became evident after 1 hour and was maintained in an upregulated state for 2-3 days ( Figure 3A ). In contrast, an auxin-responsive promoter ( proDR5::GUS ) and an indicator of cell division ( proCYCLINB1;1::GUS ) were not activated at the cutting site of a leaf residue within the 7-day period we monitored ( Figure 3A ), indicating that limited auxin response and cell proliferation occurred at these wound sites. Our data showed that wound-induced cooling coincided with the temporospatial expression pattern of cold-responsive genes. Download figure Open in new tab Figure 3: Evaporative cooling activated COR15A. A) Expression pattern of COR15A , DR5 and CYCLIN B1;1 promoters after cutting. Staining of GUS reporters driven by proCOR15A (top), proDR5 (middle) and proCYCLIN B1;1 (bottom) after wounding. proCOR15A::GUS was induced as early as 1 hour after cutting. proDR5 and proCYCLIN B1;1 were not induced at wound sites on leaf residue within 7DAC. B) and C) RGB (B) and thermal (C) images showing a wounded leaf with lanolin applied on the right side. Note the loss of cooling on the lanolin applied side. D) Staining of wounded leaves carrying proCOR15A::GUS (left) and proSPL3::GUS (right). Lanolin was applied on the right cutting edge of each leaf. E) qPCR of COR15A and KIN1 in wounded tissues with or without lanolin. Bars represent standard error. *: p<0.05 using an unpaired (Welch’s) t-test with a one-tailed hypothesis. Each dot represents a biological replicate. F) and G) RGB (F) and thermal (G) images showing leaves with water droplets. A reduced temperature was associated with water droplets. H) Staining of proCOR15A::GUS in leaves with water droplets. Arrowheads indicate the location of water droplets. of x/y: x leaves showed observed pattern in a total of y leaves. I) Wound-induced activation of COR15A depended on CBFs. Wound samples were harvested at 1 hour or 1 DAC. Bars represent standard deviation. *: p<0.05 using an unpaired (Welch’s) t-test with a one-tailed hypothesis. Each dot represents a biological replicate. To test a hypothesis that the induction of cold-responsive gene was due to evaporative cooling, we applied lanolin, a waxy water-repellent substance, onto half of a wound site to selectively block water evaporation ( Figure 3B ). We observed that the region with lanolin had a higher temperature than the exposed side of wound ( Figure 3C ). Consistently, the induction of COR15A promoter was compromised at lanolin-treated half ( Figure 3D ). As a control for lanolin treatment, proSPL3:GUS , a gene unresponsive to cold, did not show a reduction in the region applied with lanolin ( Figure 3D ). Using qPCR, we also observed that wound induction of COR15A and KIN1 transcripts was restored to basal level in the presence of lanolin ( Figure 3E ). As a complementary approach, we generated wounds on a seedling submerged under water ( Supplementary Figure 3 ). After two hours, the seedlings were stained for the pattern of proCOR15A::GUS ( Supplementary Figure 3 ). Compared to control plants with leaves cut in the open air, the submerged plants did not show activation of COR15A promoter at its cutting sites ( Supplementary Figure 3A and B ). These results suggest that wound associated evaporative cooling may contribute to the transcriptional activation of cold responsive genes. COR15A are often shown to be induced under freezing stress ( Steponkus et al ., 1998 ; Franklin & Whitelam, 2007 ). Compared with freezing, evaporative cooling only mildly reduced leaf temperature in our studies ( Figure 2 ). To determine whether evaporative cooling was sufficient to induce COR15A promoter activity, we applied a drop of room-temperature water on leaf surface for several hours and monitored proCOR15A::GUS activity ( Figure 3F , 3G). We observed a low temperature associated with the site of water droplet and a preferential activation of proCOR15A::GUS underneath the droplet ( Figure 3H ). Since no wounding was involved in this assay, the results suggest that a persistent evaporative cooling is sufficient to locally induce a subset of cold-responsive gene expression. Next, we explored the genetic requirement of CBF transcription factors in wound-induced activation of COR15A . We compared the mRNA level of COR15A at the wound site from Col-0 and a triple mutant of CBFs , cbf1/2/3 (CS73345) ( Xie et al ., 2021 ), and a transgenic line overexpressing CBF1 ( 35S::CBF1 , CS69498) ( Gilmour et al ., 2004 ). Using qPCR, we observed an induction of COR15A in Col-0 after wounding, consistent with the RNA-seq results and the GUS reporter ( Figure 3I ). Such induction was abolished in the cbf1/2/3 triple mutant, implicating a CBF-dependent mechanism ( Figure 3I ). COR15A level was generally higher in the CBF1-overexpression line even in the absence of wounding ( Figure 3I ). These observations indicate that CBFs contribute to the wound-associated activation of COR15A . A deep learning-based workflow to monitor wound temperature Thermal imaging of evaporative cooling at the wound site provides a real-time, non-invasive approach to monitor wound healing ( Figure 2 ). With an assumption that healed tissue would restore wound site temperature due to the gradual loss of evaporative cooling ( Figure 4A ), we proposed a DL-based workflow to measure the temperatures of the wound and blade sites ( Figure 4B ). RGB and thermal images were first registered based on the corner points of the checkboards ( Figure 4C ). Next, a DL-based segmentation model, YOLO-seg, was employed to segment the blade and wound sites in RGB images ( Figure 4D ). The resulting segmentation masks were then mapped onto the raw thermal images to extract the temperatures of the blade and wound areas ( Figure 4E ). Finally, the temperature differences were calculated based on the extracted temperatures. Download figure Open in new tab Figure 4: DL-based workflow for temperature difference measurement between blade and wound extraction. A) A diagram showing the method to measure temperature difference between wound site and leaf blade. Representative images of leaves at 30 mins and 3 days after wounding. Leaf margins were outlined. Note the lack of evaporative cooling in leaves at 3 DAC. B) The overview of the DL-based method for temperature difference extraction. C) Thermal and RGB image registration. (i-ii): Copper grid detection (red bounding boxes) and corner extraction (blue dots) results on both RGB and thermal images, respectively. (iii): The aligned results by overlaying the transformed RGB image onto the thermal image. D) Segmentation results of the blade (blue masks) and wound (yellow masks). E) Temperature extraction of the blade and wound based on the segmentation results. F) Linear regression between DL-extracted temperatures and the ground truth of blade sites. G) Linear regression between DL-extracted temperatures and the ground truth of wound sites. The performance of the proposed method on the testing datasets of 22 images demonstrated high accuracy in detecting copper grids of the checkerboards for both RGB and thermal images. Specifically, the Precision, Recall, [email protected] , and [email protected] :0.95 were 99.9%, 99.6%, 99.5%, and 97.3%, respectively, for RGB images, and 99.9%, 100.0%, 99.5%, and 97.0% for thermal images. The segmentation performance for the leaf blade and wound sites on RGB images was also strong, with overall Precision, Recall, [email protected] , and [email protected] :0.95 reaching 97.3%, 96.7%, 98.0%, and 70.9%, respectively. Furthermore, the extracted temperatures for the leaf blade and wound sites using the proposed method closely correlated with ground truth measurements obtained manually from thermal images and the thermocouple measurements ( Figure 4F and 4G ). These results indicated that the proposed DL-based workflow accurately estimated temperatures, with leaf blade predictions showing slightly higher accuracy than those for wound sites. Specifically, the mean absolute errors (MAE) were 0.09°C for blade sites and 0.15°C for wound sites, with R² values of 0.976 and 0.931, respectively. In summary, the DL-based workflow can automatically collect both leaf blade and wound site temperatures using RGB and thermal imagery. CBFs were required for normal wound healing We then used the DL-based workflow to monitor the temperature differences between wound site and leaf blade over a 4-day period in wild type Col-0 and the triple mutant of cbf1/2/3 ( Xie et al ., 2021 ). We defined “ΔTm” as the difference between wound and leaf blade temperature and expected it to approach zero upon wound sealing ( Figure 4A ). In Col-0, the cooling effect was readily detected immediately after cutting with a temperature difference around 1.5°C ( Figure 5A ). The temperature difference between the wound site and leaf blade decreased between one and three days after cutting (DAC) ( Figure 5A ), but by 4 DAC no temperature difference was detected, indicating that the wound was healed ( Figure 5A ). At both 2 and 3 DAC, the cbf1/2/3 mutant showed a greater temperature difference between the leaf blade and wound site than Col-0, suggesting a delayed healing process ( Figure 5A and 5B , Supplementary Figure 5 ). We also used lignin deposition as a proxy for wound healing ( Xu et al ., 2024 ) and observed compromised lignin deposition in the cbf1/2/3 mutants, indicating that the CBF transcription factors are required for wound-induced lignin deposition ( Figure 5D and 5E ). Together, these results suggested that the CBFs contribute to wound healing, probably by sensing the evaporative cooling signal. Download figure Open in new tab Figure 5: CBFs are required for normal wound healing. A) Healing curve of Col-0 and cbf1/2/3 in 4 DAC. cbf1/2/3 mutant showed a delayed healing process. The insertion shows temperature differences at 3 DAC in Col-0 and cbf1/2/3 . *: p<0.05 using unpaired t-test with a two-tailed hypothesis. B) Representative thermal images from Col-0 and cbf1/2/3 at 1 DAC and 4 DAC. Note the reduced temperature at cutting sites at 2DAC. Dashed lines indicate leaf margin. C) Representative images of lignin deposition at wound site at 1 DAC. Note the reduced lignin deposition in cbf1/2/3 mutant. D) Quantification of lignin deposition at wound sites in Col-0 and cbf1/2/3 . *: p<0.05 using unpaired t-test with a two-tailed hypothesis. Each dot represents a measurement from one wounded leaf. SA responses were compromised in cbf1/2/3 after wounding To further explore the downstream factors connecting CBFs to wound healing, we compared wound-induced transcriptomes between Col-0 and the cbf1/2/3 mutant at 0, 10 min, 60 min, and 24 h after wounding. Because cooling occurs only at the wound sites, we harvested tissues within approximately 5 mm of the cutting site ( Figure 6A and Supplementary Figure 6A ). At each time point, we detected 99, 46, 186, and 151 differentially expressed genes (DEGs) between Col-0 and cbf1/2/3 , respectively ( Supplementary Figure 6B ). We then clustered the 335 DEGs that were differentially expressed between Col-0 and cbf1/2/3 in at least one time point ( Figure 6A ). As expected, we observed the activation of cold response following wounding ( Figure 6A and 6B ). Notably, CBFs were rapidly activated within 10 min after wounding, and their expression returned to steady-state levels within 1 h (cluster 1) ( Supplementary Figure 6C ). This transient activation preceded a wave of CBF-activated genes observed at 60 min in cluster 7 ( Figure 6A ). In the cbf1/2/3 mutants, genes identified as CBF transcriptional targets, based on published ChIP-seq data ( Song et al ., 2021 ), were overrepresented among the downregulated DEGs at 10 min, further supporting that these genes are regulated by CBFs upon wounding ( Supplementary Figure 6D ). Download figure Open in new tab Figure 6. CBF-mediated regulation of wound-induced SA response and tissue healing A) Heatmap showing hierarchical clustering of 335 DEGs between Col-0 and the cbf1/2/3 triple mutant following wounding. CBF1, CBF2, and CBF3 were rapidly induced within 10 min after wounding, followed by activation of cold-responsive genes after 1 hour. B) Gene Ontology (GO) enrichment analysis of DEGs reveals significant overrepresentation of salicylic acid (SA)-responsive pathways. C) Venn diagram illustrating the overlap between SA-, JA-, and CBF-regulated genes at 24 hrs after wounding. Among CBF-dependent DEGs at 24 hr, 111 genes overlap with SA-responsive genes, whereas no overlap is observed with JA-responsive genes. SA and JA responsive genes were defined as in ( Yang et al ., 2017 ) D) Wound healing kinetics in the NahG mutant, monitored by thermal imaging. The temperature difference (ΔT) between wounded and unwounded tissue was significantly greater at 2 and 3 days post-cutting compared to Col-0. E) Representative images of lignin deposition at wound sites in Col-0 and SA-related mutants. Sid2 , NahG and npr1 exhibited reduced lignin accumulation compared to Col-0. Lignin deposition was detected by its autofluorescence under UV excitation using fluorescence microscopy. Coi1 was used as a positive control ( Xu et al ., 2024 ). F) quantification of lignin deposition for Col-0 and mutants showed in E. *: p<0.05 using unpaired t-test with a two-tailed hypothesis. G) Representative images of lignin deposition at wound sites in Col-0 and cbf1/2/3 with mock or BTH treatment. H) Quantification of lignin deposition in G. *: p<0.05 using unpaired t-test with a two-tailed hypothesis.. Ns: not significant. Each dot represents a measurement from one wounded leaf. As part of wound response, the SA response was activated following wounding and inhibited wound-induced regeneration ( Hernández-Coronado et al ., 2022 ; Tran et al ., 2023 ). Noteworthy, gene ontology (GO) analysis of CBF-dependent DEGs revealed significant enrichment of salicylic acid (SA)–related processes, including response to SA, systemic acquired resistance (SAR), and SA signaling pathways ( Figure 6B ). These GO terms were enriched in clusters 1, 4, 5, and 8, which include genes that were failed to be activated in the cbf1/2/3 mutant upon wounding ( Figure 6B ). These results indicate that CBFs are required for the wound-induced SA response. Notably, at 24 hours after wounding, 111 of 151 CBF-regulated genes overlapped with the SA regulon ( Yang et al ., 2017 ). In contrast, although activation of JA and ABA pathways are important for wound repair ( Koo & Howe, 2009 ; Xu et al ., 2024 ), among the 335 CBF-regulated genes, only 22 and 30 overlapped with previously reported ABA- and JA-responsive genes, respectively ( Wang et al ., 2011 ; Yang et al ., 2017 ). Wound-induced activation of the LOX2 promoter, a JA-responsive marker, was not blocked by lanolin treatment ( Supplementary Figure 6E ), suggesting that CBFs may not be essential for wound-induced JA activation. . At 24 hours after wounding, none of the 151 DEGs overlapped with JA-responsive genes ( Figure 6C ) ( Yang et al ., 2017 ), further confirming that the CBFs have limited influence on wound-induced JA response within this time frame. On the other hand, COR15A induction was not altered in mutants defective in ABA biosynthesis ( aba3-1 ) ( Supplementary Figure 6F ), so wound-induced activation of COR15A is likely independent of JA and ABA signaling. These results strongly support that CBFs regulate a wound-induced SA regulon. We therefore tested whether blocking SA pathway affects wound healing. In NahG transgenic plants, which exhibit reduced SA accumulation due to overexpression of a Pseudomonas -derived salicylate hydroxylase, wound healing was delayed, as indicated by a larger temperature difference between the wound site and surrounding leaf tissue at 2 days after cutting ( Figure 6D ). Furthermore, wound-induced lignin deposition was decreased NahG , sid2 (mutant defective in Isochorismate Synthase 1-controlled SA biosynthesis), and npr1 (an SA receptor), as observed in cbf1/2/3 and coi1 ( Figure 6E and 6F ). To test the capacity of exogenous SA to rescue wound-induced lignin deposition in the cbf1/2/3 mutant, we pre-treated Col-0 and cbf1/2/3 plants with BTH, an SA analog, 24 hours prior to cutting. At 1-day post-cutting, BTH treatment significantly enhanced lignin deposition at the cutting site in mutant plants and restored it to the level in wild type ( Figure 6G and 6H ), suggesting that SA may promotes lignin deposition downstream of CBF function. Together with the gene expression profile, we conclude that SA response acts downstream of CBFs to promote wound repair. Discussion The healing of wounded plant tissues is an essential process that protects plants from subsequent damage caused by opportunistic microbial invaders, herbivorous insects, and pathogenic organisms. Despite its importance, a major challenge in understanding wound healing lies in defining the point at which a wound can be considered fully healed. We provide experimental evidence that wound-induced evaporative cooling, driven by abrupt yet subtle temperature changes, can trigger the activation of cold response pathways. Thermal imaging and the deep learning workflow provided a reliable and sensitive method for us to track wound healing through the gradual increase in wound site temperatures. Presumably, when the temperature at the wound site is indistinguishable from the remaining undamaged leaf blade, the wound is sealed. It is possible that the wound is not entirely healed at this point because the sensitivity of the camera may limit our ability to measure relatively small differences in temperature loss. Regardless, thermal imaging combined with molecular changes (e.g. induction of COR15A and lignin deposition) offers an important method to evaluate physical and molecular responses related to the healing end point. This is a key step towards a more refined understanding of the plant wound healing responses. In this study, we found that evaporative cooling following injury was sufficient to activate a cold response dependent on CBF transcription factors. Loss of CBFs resulted in deficient lignin deposition at the wound site and impaired wound healing, highlighting the biological importance of seemingly slight temperature reductions. The observation that COR15A was induced by evaporative cooling from a persistent water droplet further support that plants are highly sensitive to temperature changes across different temporal scales. While many studies on thermal sensing have focused on plant responses to physiologically extreme temperatures, such as freezing and heat stress, ambient temperature fluctuations also profoundly impact development and stress responses ( Dong et al ., 2020 ; Kerbler & Wigge, 2023 ). At the cellular level, molecular thermosensors respond quantitatively to fine temperature variations. For instance, in Arabidopsis , hypocotyl elongation is regulated by ambient temperatures ranging from 17°C to 37°C ( Hahm et al ., 2020 ; Qiu et al ., 2021 ). The number of PhyB-containing nuclear bodies gradually decreases as the temperature rises from 12°C to 27°C, with a small 4°C difference being sufficient to induce PhyB dissociation and hypocotyl elongation ( Qiu et al ., 2019 ). These responses to subtle temperature difference is similar to our observations that sudden evaporative cooling of ∼ 1.5 °C at wound site appears sufficient to elicit a physiological response. Several lines of evidence indicate connections between low temperatures and callose or lignin accumulation ( Schütte et al ., 2025 ). In pepper, two lignin biosynthesis genes, CaCAD1 and CaPOA1 , are induced by prolonged cold exposure ( Xiao et al ., 2025 ). Our results showed that CBF promotes the wound-induced SA response. It is therefore plausible that SA serves as a link between sensing evaporative cooling and the deposition of callose and lignin, given its established role in pathogen-induced callose accumulation ( Tateda et al ., 2014 ) and its crosstalk with lignin biosynthesis ( Delaney et al ., 1994 ; Wang et al ., 2025 ). Notably, DEGs between Col-0 and cbf1/2/3 mutants at 1 h and 24 h after cutting are not enriched for CBF-binding motifs, suggesting that CBFs likely regulate SA-responsive genes indirectly through yet unidentified intermediates. Interestingly, previous studies have shown that cbf1/2/3 mutants did not exhibit altered activation of SA-responsive genes ( PR1 and PR5 ) in response to bacterial infection or exogenous SA treatment, indicating that CBFs are not required for pathogen-induced SA signaling ( Li et al ., 2020 ; Li et al ., 2024 ). We also assessed bacterial pathogen growth and the induction of PR1 and PR5 following BTH treatment in our cbf1/2/3 mutant and observed no significant differences compared to the wild type ( Supplementary Figure 7 ). Together, these findings suggest that additional wound-associated factors that are not present during pathogen infection, may act synergistically with CBFs to activate wound-induced SA signaling. Overall, these data support the existence of an intermediate regulatory component linking CBF transcription factors to the wound-induced SA response. The deep learning approach developed in our study combined multiple imaging modalities, i.e., RGB and thermal images, to detect the wound sites with precision. This integrated approach not only identified wounds but also quantified localized temperature variations—a key indicator of physiological stress, disease progression, and healing efficiency ( Wen et al ., 2023 ). By automating the analysis of these multimodal datasets, the method significantly reduces reliance on labor-intensive manual assessments, reduces observer bias and enables high-throughput phenotyping. The workflow’s capability to track thermal profiles at specific injury sites offers deeper insights into plant resilience mechanisms. Overall, these capabilities can accelerate fundamental research, improve crop management strategies, and enhance data-driven decisions to boost agriculture productivity and long-term sustainability. Materials and methods Plant materials and growth conditions Arabidopsis wild type and mutants used in this study were summarized in supplementary Table 2. Seeds were sown on Sungro Propagation Mix Question soil and placed in 4°C cold room for 3 days prior to being transferred to a short-day growth room with 23°C/19°C day/night temperatures and 45% humidity. Plants were watered weekly and placed in a random block design to reduce the impact of environmental factors. Plants for thermal imaging and GUS staining were 4-6 weeks old. GUS staining Leaf tissues were submerged in a GUS staining solution [100 mM sodium phosphate (pH 7.0), 1 mM EDTA (pH 8), 1% Triton-X-100, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide and 1 mg/ml X-Gluc] and vacuum infiltrated for 10 min. Following incubation at 37°C (3 hours for proLOX2:GUS, 5 hours for proCOR15A:GUS, proDR5:GUS, and proCYCB2:GUS. The staining solution was then replaced with 70% ethanol and washed several times until the tissue was clear. Once cleared, the samples were blotted dry, placed face-down on a CanoScan LiDE 400 PDF scanner, and scanned at 1200 dpi with all color correction and blemish removal settings disabled. qPCR RNA extraction was performed using TRIzol reagent (Invitrogen). The qPCR was performed in the Applied Biosystems QuantStudio 1 Real-Time PCR system with SYBR Green master mix (Applied Biosystems). PCR conditions were set as follows: 95°C for 5 mins, 40 cycles of 95°C for 15s, 56°C for 30s and 72°C for 20-30s. TUBULIN 2 ( AT5G62690 ) was used as a reference gene. We confirmed that TUBULIN 2 was not a wound-induced DEG. The delta-delta CT was used for fold-change calculation after confirming the PCR efficiency was at least more than 95%. The oligonucleotides used are COR15A-F: 5’-AACTCAGTTCGTCGTCGTTTCTC-3’; COR15A-R: 5’-CCCAATGTATCTGCGGTTTCAC-3’; TUB-F: 5’-AGCAATACCAAGATGCAACTGCG-3’ and TUB-FR: 5’-TAACTAAATTATTCTCAGTACTCTTCC-3’; PR1-F: 5’GCCTTCTCGCTAACCCACAT-3’ PR1-R: 5’-CGGAGCTACGCAGAACAACT-3’; PR5-F: 5’-ACTGTGGCGGTCTAAGATGTA-3’ PR5-R: 5’-CGTATTTGCAATCTCCCGATCC-3’ KIN1-F: 5’-GCTGGAGCTTCCGCGCA-3’ KIN1-R: 5’-TTATAACTCCCAAAGTTGACTCGG-3’ Acquisition of thermal imaging Fotric’s 616C R&D station equipped with a 40.5 mm thermal lens and AnalyzIR software was used to acquire thermal images (FOTRIC). The station was situated in NEEWER’s 20”x20”x20” Studio Light Box with LED lights strung in the canopy and black fabric lining the bottom of the booth. The height of the camera mount was adjusted to create a 15 cm object distance and emissivity is designated as 1. A 3x6 cm 2 checkboard with alternating copper and paper cells was used as an RGB image alignment marker. The checkboard was mounted on a 30° rubber wedge to avoid directing the heat signature from the camera body back towards the camera. Arabidopsis seeds were grown between four and six weeks prior to cutting. On the day of cutting, half of the leaf blade was removed with ceramic scissors. A plastic stake was placed below the cutting site with forceps to minimize soil temperature affecting image quality. At 30 minutes post-cutting, the plant was transferred to the photo booth. Two pre-chilled checkboard alignment markers were placed within the camera’s field of view and away from any leaves. After adjusting the focus until the cutting site was well defined, snapshots were taken with the AnalyzIR workstation. A report of experimental parameters and a thermal data CSV file were generated for each image. Before removing the plant from the chamber, RGB images were taken using a Samsung Galaxy S24 phone mounted on a metal stand 15 cm away from pots. Color correction and image processing function was disabled. Manual temperature measurement from thermal images In AnalyzIR, the average temperature from a line tracing the cutting site was recorded as the cutting site temperature. The line was then moved to the blade area of the same leaf to measure leaf blade temperature. The temperature difference between the wounding site and leaf blade was used as a quantitative readout of cooling effect. In the time-lapse experiments of 24 hrs, the thermal camera was positioned in a growth chamber in order to monitor plants in constant growth conditions. An imaging rate of 6 seconds per frame was used to capture the time lapse, and AnalyzIR segmentation and graphing functions were used to extract temperature values. Manual temperature measurement from thermocouple measurements Using an Omega HH374 Data Logger Thermometer equipped with a Type K beaded probe (Omega), temperatures on the leaf blade and the cutting site were measured. All measurements with the thermocouple were performed on the same timeline as the thermal image measurements. Lignin deposition quantification Cut leaf samples were transferred to glass slides and imaged with the Zeiss Axio Imager 2 (Zeiss). To capture the lignin autofluorescence, an excitation wavelength of 488 nm and a filter for 535 nm light were used. All exposure times, gamma settings, and fields of view were standardized across all samples. ImageJ was used to remove background autofluorescence and quantify the average lesion area per micron of wounding site length. RNA-seq analysis For the RNA-seq gene expression analysis following wounding, total RNA was extracted from two-week-old Arabidopsis (Col-0 ecotype) leaf samples subjected to wounding stress. Samples were harvested at three time points: prior to treatment (time 0), 30 minutes post-wounding, and 60 minutes post-wounding. The experiment included three biological replicates per condition. For the RNA-seq experiment comparing wound responses in Col-0 and in the cbf1/2/3 mutant, tissues within 1 cm of wounding sites were harvested following cutting off 1/3 of a leaf using scissors. Libraries were prepared using the TruSeq RNA library prep kit (Illumina) according to the manufacturer’s protocol and sequenced with the Illumina Nextseq platform. Read quality was assessed using FastQC version 0.11.8 (Babraham Bioinformatics, Cambridge, UK), and low-quality reads and adapter sequences were removed with Trimmomatic version 0.36 ( Bolger et al ., 2014 ). The resulting high-quality reads were then aligned to the TAIR10 reference genome using HISAT2 version 2.2.0 ( Kim et al ., 2015 ), and raw read counts were quantified using the featureCounts function from the Sub-read package version 2.0.0 ( Liao et al ., 2013 ). Differential gene expression analysis was conducted using the edgeR package in R ( Robinson et al ., 2010 ). Genes with low expression levels were filtered by excluding those with fewer than 1 count per million in at least five libraries. Normalization of read counts was performed using the trimmed mean of M-values (TMM) method, implemented with the calcNormFactors function in edgeR ( Robinson & Oshlack, 2010 ). To control for multiple testing, p-values were adjusted using the Benjamini-Hochberg procedure (Benjamini and Hochberg, 1995). Genes were considered differentially expressed if they had a false discovery rate (FDR) ≤ 0.05 and exhibited a minimum 1.5-fold change in expression. Comparisons were made between gene expression levels at 30 and 60 minutes after wounding versus unwounded samples (time 0). Gene Ontology (GO) term enrichment analysis was performed using Gene Set Enrichment Analysis (GSEA) (Yi et al. 2013) to identify biological processes associated with the differentially expressed genes. For the Col-0 vs cbf1/2/3 mutant comparison, GO enrichment analysis was performed using the compareCluster function in clusterProfiler ( Yu et al ., 2012 ).Hierarchical clustering was performed using the ComplexHeatmap package in R ( Gu et al ., 2016 ), based on the Euclidean distance and the complete-linkage method. Gene expression values in TPM (transcript per million) were normalized to z-scores for clustering analyses. Raw RNA-seq reads and the count matrix generated in this study are available at the NCBI Gene Expression Omnibus under the accession number GSE293328. Additionally, public gene expression datasets ( Kilian et al ., 2007 ; Ikeuchi et al ., 2017 ; Fiorucci et al ., 2022 ; Liu et al ., 2022 ) were utilized to investigate whether cold-responsive marker genes were induced in independent experiments involving wounded Arabidopsis tissues, using a threshold of |log2 fold change| ≥ 1 and FDR ≤ 0.05. Method for testing binding sites enrichments Enrichment analysis was performed using a hypergeometric test (phyper function in R) to determine whether genes identified as targets of CBF transcription factors, based on published ChIP-seq data ( Song et al ., 2021 ), were overrepresented among the differentially expressed genes in our RNA-seq experiment (DEGs). The set of expressed genes was used as the background universe and p-values were adjusted for multiple testing using the Benjamini–Hochberg method. Development of DL-based workflow for wound detection The DL-based workflow consists of four main steps ( Figure 4B ): (1) copper grid detection and corner point extraction; (2) RGB image and thermal image registration based on affine matrix; (3) blade and wound segmentation on RGB images; and (4) mapping the segmented blade and wound to aligned thermal images and temperature difference calculation. Two checkerboards made of foam grids and copper sheets were placed around the wounded area, and an RGB camera (Samsung Galaxy S24) and a thermal camera (600 R&D, FOTRIC) were used to collect the RGB and thermal images from the top view. Due to the different resolutions (3000x4000 for RGB and 640x480 for thermal) and varying parameters such as the focal lengths of the cameras, the RGB and thermal images were not aligned. Therefore, we extracted corners on the checkerboards for alignment. As the temperatures of the foam and copper are different, the copper is lighter, and the foam is darker in the thermal images. To detect copper grids and extract the corner points, the detection model, YOLOv8-obb ( Jocher et al ., 2023 ) was trained to detect the copper grids with orientated bounding boxes in RGB and thermal images, respectively. A total of 71 RGB images and 38 thermal images were annotated using Roboflow ( Dwyer et al ., 2024 ) with oriented bounding boxes representing copper grids. Images were then divided into training and testing datasets with a ratio of 8:2. All training images were further augmented three times, including horizontal and vertical flipping, 90-degree rotations, and additional adjustments to saturation, brightness, and exposure for the RGB images. This resulted in a total of 171 RGB images and 87 thermal images for training. The detection model was trained for 100 epochs with an input resolution of 2048×2048 pixels for RGB images and 640×640 pixels for thermal images. An initial learning rate of 0.002 and an AdamW optimizer were used on the HiPerGator high-performance computing cluster with 8 AMD EPYC ROME CPU cores, one NVIDIA DGX A100 GPU node (80GB), and 32 GB of memory. After detecting all copper grids, the four corners of each grid were extracted and the distances between each corner were calculated. Then, the closest two corners were obtained, and the center point was calculated as the corner point of these two grids. A sorting algorithm that sorts the corner points from left to right and top to down to align the same corner points on RGB and thermal images. The affine matrix was calculated based on the sorted points and the affine transformation was applied to the RGB image to register with the thermal image ( Figure 4C ). To detect the blade and wound sites, a segmentation model, YOLOv8-seg ( Jocher et al ., 2023 ), was trained. A total of 103 images were annotated using Roboflow with polygons representing blade and wound sites. Images were then divided into training and testing datasets with a ratio of 7:3. All training images were further augmented three times, including horizontal and vertical flipping, 90-degree rotation, and rotation between ±15 degrees, resulting in 216 images for training. The segmentation model was trained for 200 epochs with an input resolution of 640×640 pixels, an initial learning rate of 0.0017, and an AdamW optimizer on the HiPerGator high-performance computing cluster. After obtaining the masks of the blade and wound sites ( Figure 4D ), the centroids of the two closest masks were calculated. Additionally, the mask of the wound site was eroded with a kernel size of 3×3 pixels and moved along the direction formed by the centroids to the centroid of the blade with 10 pixels. Then, the mask of the wound site and the intersection over union between the moved mask and blade mask were mapped to the raw thermal data. The mean temperature values were calculated within two masks to represent the wound and blade temperatures, respectively ( Figure 4E ). The temperature difference was calculated by subtracting the temperature of the wound from the temperature of the blade. Bacterial growth assay Pseudomonas syringae DC3000 strain was grown under 28°C on King’s B (KB) solid medium (40 g/L proteose, 20 g/L glycerol and 15 g/L agar). The medium contained rifamycin for selection and cycloheximide to inhibit fungi growth. Glycerol stock of the bacterial strains stored under -80°C. Bacterial stock was streaked on plate for a 2-day growth and was cultured in liquid KB medium for an additional day. After pelleting and resuspending bacteria in 10 mM MgCl 2 =:solution, the solution was diluted until the optical density at 600 nm was adjusted to 0.025 as measured with a spectrophotometer (Eppendorf BioPhotometer). The final volume of inoculating solution was further supplemented with 0.005% Silwet L-77 surfactant to ensure even inoculation. Arabidopsis plants grown in the same conditions as those grown for thermal imaging were dipped for several seconds into the inoculating solution. After inoculation, plants were covered by transparent lids for one hour. Leaf samples were harvested 48 hours post-inoculation using a corer to collect four tissue discs from different leaves per sample. Leaf samples were ground with a homogenizer (OMNI International) and serially diluted. KB plates with 10 μL of bacteria suspension per sample were grown at 28°C for two days. Colony forming units were counted manually and normalized according to inside area of the corer. Download figure Open in new tab Figure S1: GO terms associated with wound-induced genes. A) Gene Ontology (GO) enrichment analysis indicating biological processes associated with each cluster. Each dot represents a significantly enriched GO term. Highlighted processes indicate “Response to cold”, “Cold acclimatation” and “Jasmonic acid biosynthetic process”, which are all enriched among up-regulated genes. B) Cold-responsive genes are consistently up-regulated in response to wounding across independent experiments. Gene expression is presented as z-scores, and differentially expressed genes are indicated by a thick border in each treatment. C) Up-regulation of cold-responsive genes in response to wounding. Expression of six marker genes of the cold response, represented as a cartoon-like heatmap. The image is derived from the Arabidopsis eFP Browser (Waese et al., 2017), based on AtGenExpress dataset from ( Kilian et al ., 2007 ). Expression levels are indicated by a gradient color scale, ranging from yellow (low expression) to red (high expression), across a time-course of 0, 0.25h, 0.5h, 1h, 3h, 6h, 12h and 24h after wounding. Download figure Open in new tab Figure S2: Evaporative cooling at wound site observed in cold and heat environments. Arabidopsis plants were imaged in a growth chamber at 37°C (A) or in a cold room at 6°C (B). Temperature differences between wound sites and leaf blades were observed under both conditions, with a greater difference detected in leaves at 37°C. Video S1: Temperature reduction occurred at wound sites immediately after cutting. Download figure Open in new tab Figure s3: Wound-induced COR15A activation was not observed when plants were cut under water. A) An image showing a leaf being cut in water. B) Staining of proCOR15A::GUS in leaves wounded under water (right) or in air (left). The red circles indicate cutting sites. Download figure Open in new tab Figure S5: Healing curve of Col-0 and cbf1/2/3 in the same frame. A) Thermal image snapshot of Col-0 and cbf1/2/3 at 2 Days after cutting. B) Time course recording of temperature difference (Wound temperature – blade temperature) from a wounded Col-0 and cbf1/2/3 leaf. Download figure Open in new tab Figure s6. Temporal transcriptomic dynamics in response to wounding. A) Principal component analysis (PCA) of RNA-seq samples showing that time post-wounding is the primary factor driving variation in gene expression across samples. B) Number of differentially expressed genes (DEGs) identified at each time point following wounding. C) Expression dynamics of CBF1 , CBF2 , and CBF3 . D) Enrichment analysis of CBF-binding motifs among DEGs at each time point. CBF-binding sites were defined based on data from ( Song et al ., 2021 ), and enrichment significance was assessed using a hypergeometric test, with p-values adjusted by the Benjamini–Hochberg method. E) proLOX2::GUS staining following wounding and lanolin treatment, showing that this induction of GUS activity was not blocked by lanolin treatment. ns: not significant with unpaired (Welch’s) t-test. Each dot represents a measurement from one wounded leaf. F) Expression of COR15A in aba3-1 mutant. ns: not significant with unpaired (Welch’s) t-test. Each dot represents a biological replicate. Download figure Open in new tab Figure s7. CBFs mutations did not alter resistance to bacterial pathogens and BTH-induced SA activation. A) Bacterial growth in Col-0 and the cbf1/2/3 mutant at 3 days post inoculation. B–C) Induction of PR1 and PR5 in Col-0 and cbf1/2/3 mutants under mock or BTH treatment. BTH was applied by foliar spray, and no wounding occurred during the treatment. Ns: not significant using unpaired t-test with a two-tailed hypothesis. Each dot represents a biological replicate. Supplementary Table 1: List of genes Figure 1A and Figure 6A . Supplementary Table 2: Numeric raw data. Acknowledgements L.Y. is supported by the National Science Foundation (IOS-2039313) and the National Institutes of Health (R35GM143067). P.J.P.L.T. is funded by the Serrapilheira Institute (grant G-1811-25705), the National Council for Scientific and Technological Development (CNPq, grant 308349/2022-9), and the São Paulo Research Foundation (FAPESP; grant 2018/24432-0). N.C.F.D. is supported by a fellowship from FAPESP (2023/10167-1). C.L. is supported by the Research Capacity Fund (Hatch Multistate) program, project award no. 7009022, from the U.S. Department of Agriculture’s National Institute of Food and Agriculture. Funder Information Declared National Science Foundation, https://ror.org/021nxhr62 , IOS-2039313 National Institutes of Health , R35GM143067 Instituto Serrapilheira , G-1811-25705 National Council for Scientific and Technological Development , 308349/2022-9 São Paulo Research Foundation , 2018/24432-0 FAPESP , 2023/10167-1 National Institute of Food and Agriculture, https://ror.org/05qx3fv49 , 7009022 Footnotes We included new data on wound-induced transcriptomic changes in wild-type and cbf1/2/3 mutant plants. Reference ↵ Alfred R , Leo J , Kaijage SF . 2025 . Detectron2-enhanced mask R-CNN for precise instance segmentation of rice blast disease in Tanzania: Supporting timely intervention and data-driven severity assessment . Smart agricultural technology : 101301 . ↵ Amogi BR , Ranjan R , Khot LR . 2024 . Mask R-CNN aided fruit surface temperature monitoring algorithm with edge compute enabled internet of things system for automated apple heat stress management . Information Processing in Agriculture 11 ( 4 ): 603 – 611 . OpenUrl CrossRef ↵ Artus NN , Uemura M , Steponkus PL , Gilmour SJ , Lin C , Thomashow MF . 1996 . Constitutive expression of the cold-regulated Arabidopsis thaliana COR15a gene affects both chloroplast and protoplast freezing tolerance . Proceedings of the National Academy of Sciences 93 ( 23 ): 13404 – 13409 . OpenUrl Abstract / FREE Full Text ↵ Baker SS , Wilhelm KS , Thomashow MF . 1994 . The 5′-region of Arabidopsis thaliana cor15a has cis-acting elements that confer cold-, drought-and ABA-regulated gene expression . Plant molecular biology 24 : 701 – 713 . OpenUrl CrossRef PubMed Web of Science ↵ Beckers GJ , Jaskiewicz M , Liu Y , Underwood WR , He SY , Zhang S , Conrath U . 2009 . Mitogen-activated protein kinases 3 and 6 are required for full priming of stress responses in Arabidopsis thaliana . The Plant Cell 21 ( 3 ): 944 – 953 . OpenUrl Abstract / FREE Full Text ↵ Bhagat S , Kokare M , Haswani V , Hambarde P , Kamble R . 2022 . Eff-UNet++: A novel architecture for plant leaf segmentation and counting . Ecological Informatics 68 : 101583 . OpenUrl ↵ Bolger AM , Lohse M , Usadel B . 2014 . Trimmomatic: a flexible trimmer for Illumina sequence data . bioinformatics 30 ( 15 ): 2114 – 2120 . OpenUrl CrossRef PubMed Web of Science ↵ Bozovic V , Svensson J , Schmitt J , Kohn C . 2013 . Dehydrins (LTI29, LTI30, COR47) from Arabidopsis thaliana expressed in Escherichia coli protect thylakoid membrane during freezing . Journal of the Serbian Chemical Society 78 ( 8 ): 1149 – 1160 . OpenUrl CrossRef ↵ Chowdhury M , Reza MN , Jin H , Islam S , Lee G-J , Chung S-O . 2024 . Defective Pennywort Leaf Detection Using Machine Vision and Mask R-CNN Model . Agronomy 14 ( 10 ): 2313 . OpenUrl CrossRef ↵ Cook D , Fowler S , Fiehn O , Thomashow MF . 2004 . A prominent role for the CBF cold response pathway in configuring the low-temperature metabolome of Arabidopsis . Proceedings of the National Academy of Sciences 101 ( 42 ): 15243 – 15248 . OpenUrl Abstract / FREE Full Text ↵ De Lorenzo G , Ferrari S , Cervone F , Okun E. 2018 . Extracellular DAMPs in plants and mammals: immunity, tissue damage and repair . Trends in immunology 39 ( 11 ): 937 – 950 . OpenUrl CrossRef PubMed ↵ Delaney TP , Uknes S , Vernooij B , Friedrich L , Weymann K , Negrotto D , Gaffney T , Gut-Rella M , Kessmann H , Ward E , et al. 1994 . A central role of salicylic Acid in plant disease resistance . Science 266 ( 5188 ): 1247 – 1250 . OpenUrl Abstract / FREE Full Text ↵ Deng Y , Xi H , Zhou G , Chen A , Wang Y , Li L , Hu Y . 2023 . An effective image-based tomato leaf disease segmentation method using MC-UNet . Plant Phenomics 5 : 0049 . OpenUrl PubMed ↵ Dong X , Yan Y , Jiang B , Shi Y , Jia Y , Cheng J , Shi Y , Kang J , Li H , Zhang D , et al. 2020 . The cold response regulator CBF1 promotes Arabidopsis hypocotyl growth at ambient temperatures . EMBO J 39 ( 13 ): e103630 . OpenUrl CrossRef PubMed ↵ Dwyer B , Nelson J , Hansen T 2024 . Roboflow (version 1.0)[software] . ↵ Fiorucci A-S , Michaud O , Schmid-Siegert E , Trevisan M , Allenbach Petrolati L , Çaka Ince Y , Fankhauser C . 2022 . Shade suppresses wound-induced leaf repositioning through a mechanism involving PHYTOCHROME KINASE SUBSTRATE (PKS) genes . PLoS Genetics 18 ( 5 ): e1010213 . OpenUrl ↵ Franklin KA , Whitelam GC . 2007 . Light-quality regulation of freezing tolerance in Arabidopsis thaliana . Nat Genet 39 ( 11 ): 1410 – 1413 . OpenUrl CrossRef PubMed Web of Science ↵ Fu J , Zhao Y , Wu G . 2023 . Potato leaf disease segmentation method based on improved UNet . Applied Sciences 13 ( 20 ): 11179 . OpenUrl ↵ Gilmour SJ , Fowler SG , Thomashow MF . 2004 . Arabidopsis transcriptional activators CBF1, CBF2, and CBF3 have matching functional activities . Plant molecular biology 54 ( 5 ): 767 – 781 . OpenUrl CrossRef PubMed Web of Science ↵ Gilmour SJ , Zarka DG , Stockinger EJ , Salazar MP , Houghton JM , Thomashow MF . 1998 . Low temperature regulation of the Arabidopsis CBF family of AP2 transcriptional activators as an early step in cold-induced COR gene expression . The Plant Journal 16 ( 4 ): 433 – 442 . OpenUrl CrossRef PubMed Web of Science ↵ Gu Z , Eils R , Schlesner M . 2016 . Complex heatmaps reveal patterns and correlations in multidimensional genomic data . bioinformatics 32 ( 18 ): 2847 – 2849 . OpenUrl CrossRef PubMed ↵ Hahm J , Kim K , Qiu Y , Chen M . 2020 . Increasing ambient temperature progressively disassembles Arabidopsis phytochrome B from individual photobodies with distinct thermostabilities . Nat Commun 11 ( 1 ): 1660 . OpenUrl CrossRef PubMed ↵ Hernández-Coronado M , Araujo PCD , Ip P-L , Nunes CO , Rahni R , Wudick MM , Lizzio MA , Feijó JA , Birnbaum KD . 2022 . Plant glutamate receptors mediate a bet-hedging strategy between regeneration and defense . Developmental cell 57 ( 4 ): 451 – 465 . e456. OpenUrl CrossRef PubMed ↵ Huang F , Li Y , Liu Z , Gong L , Liu C . 2024 . A method for calculating the leaf area of pak choi based on an improved mask R-CNN . Agriculture 14 ( 1 ): 101 . OpenUrl ↵ Hwarari D , Guan Y , Ahmad B , Movahedi A , Min T , Hao Z , Lu Y , Chen J , Yang L . 2022 . ICE-CBF-COR signaling cascade and its regulation in plants responding to cold stress . International journal of molecular sciences 23 ( 3 ): 1549 . OpenUrl PubMed ↵ Ikeuchi M , Iwase A , Rymen B , Lambolez A , Kojima M , Takebayashi Y , Heyman J , Watanabe S , Seo M , De Veylder L. 2017 . Wounding triggers callus formation via dynamic hormonal and transcriptional changes . Plant physiology 175 ( 3 ): 1158 – 1174 . OpenUrl Abstract / FREE Full Text ↵ Ikeuchi M , Rymen B , Sugimoto K . 2020 . How do plants transduce wound signals to induce tissue repair and organ regeneration? Current opinion in plant biology 57 : 72 – 77 . OpenUrl CrossRef PubMed ↵ Islam S , Reza MN , Chowdhury M , Ahmed S , Lee K-H , Ali M , Cho YJ , Noh DH , Chung S-O . 2024a . Detection and segmentation of lettuce seedlings from seedling-growing tray imagery using an improved mask R-CNN method . Smart agricultural technology 8 : 100455 . OpenUrl CrossRef ↵ Islam S , Reza MN , Samsuzzaman SA , Cho YJ , Noh DH , Chung S-O , Hong SJ . 2024b . Machine vision and artificial intelligence for plant growth stress detection and monitoring: A review . Precision Agriculture 6 ( 1 ): 34 . OpenUrl ↵ Jacobs AK , Lipka V , Burton RA , Panstruga R , Strizhov N , Schulze-Lefert P , Fincher GB . 2003 . An Arabidopsis callose synthase, GSL5, is required for wound and papillary callose formation . The Plant Cell 15 ( 11 ): 2503 – 2513 . OpenUrl Abstract / FREE Full Text ↵ Jhajharia K , Mathur P , Jain S , Nijhawan S . 2023 . Crop yield prediction using machine learning and deep learning techniques . Procedia Computer Science 218 : 406 – 417 . OpenUrl CrossRef ↵ Jia Y , Ding Y , Shi Y , Zhang X , Gong Z , Yang S . 2016 . The cbfs triple mutants reveal the essential functions of CBF s in cold acclimation and allow the definition of CBF regulons in Arabidopsis . New Phytologist 212 ( 2 ): 345 – 353 . OpenUrl CrossRef PubMed ↵ Jiang Y , Shuang L , Li C , Paterson AH , Robertson J 2018 . Deep learning for thermal image segmentation to measure canopy temperature of Brassica oleracea in the field . 2018 ASABE Annual International Meeting : American Society of Agricultural and Biological Engineers. 1 . ↵ Jocher G , Chaurasia A , Qiu J 2023 . YOLO by Ultralytics . ↵ Kerbler SM , Wigge PA . 2023 . Temperature Sensing in Plants . Annu Rev Plant Biol 74 : 341 – 366 . OpenUrl CrossRef PubMed ↵ Kilian J , Whitehead D , Horak J , Wanke D , Weinl S , Batistic O , D’Angelo C , Bornberg-Bauer E , Kudla J , Harter K . 2007 . The AtGenExpress global stress expression data set: protocols, evaluation and model data analysis of UV-B light, drought and cold stress responses . The Plant Journal 50 ( 2 ): 347 – 363 . OpenUrl CrossRef PubMed Web of Science ↵ Kim D , Langmead B , Salzberg SL . 2015 . HISAT: a fast spliced aligner with low memory requirements . Nature methods 12 ( 4 ): 357 – 360 . OpenUrl PubMed ↵ Koo AJ , Howe GA . 2009 . The wound hormone jasmonate . Phytochemistry 70 ( 13-14 ): 1571 – 1580 . OpenUrl CrossRef PubMed Web of Science ↵ Li S , He L , Yang Y , Zhang Y , Han X , Hu Y , Jiang Y . 2024 . INDUCER OF CBF EXPRESSION 1 promotes cold-enhanced immunity by directly activating salicylic acid signaling . Plant Cell 36 ( 7 ): 2587 – 2606 . OpenUrl CrossRef PubMed ↵ Li Z , Liu H , Ding Z , Yan J , Yu H , Pan R , Hu J , Guan Y , Hua J . 2020 . Low Temperature Enhances Plant Immunity via Salicylic Acid Pathway Genes That Are Repressed by Ethylene . Plant Physiol 182 ( 1 ): 626 – 639 . OpenUrl Abstract / FREE Full Text ↵ Liao Y , Smyth GK , Shi W . 2013 . The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote . Nucleic acids research 41 ( 10 ): e108 – e108 . OpenUrl CrossRef PubMed ↵ Liu Q , Kasuga M , Sakuma Y , Abe H , Miura S , Yamaguchi-Shinozaki K , Shinozaki K . 1998 . Two transcription factors, DREB1 and DREB2, with an EREBP/AP2 DNA binding domain separate two cellular signal transduction pathways in drought-and low-temperature-responsive gene expression, respectively, in Arabidopsis . The Plant Cell 10 ( 8 ): 1391 – 1406 . OpenUrl Abstract / FREE Full Text ↵ Liu W , Zhang Y , Fang X , Tran S , Zhai N , Yang Z , Guo F , Chen L , Yu J , Ison MS . 2022 . Transcriptional landscapes of de novo root regeneration from detached Arabidopsis leaves revealed by time-lapse and single-cell RNA sequencing analyses . Plant Communications 3 ( 4 ). OpenUrl ↵ Murphy KM , Ludwig E , Gutierrez J , Gehan MA . 2024 . Deep learning in image-based plant phenotyping . Annual review of plant biology 75 . ↵ Ortega MA , Celoy RM , Chacon F , Yuan Y , Xue L-J , Pandey SP , Drowns MR , Kvitko BH , Tsai C-J . 2024 . Altering cold-regulated gene expression decouples the salicylic acid–growth trade-off in Arabidopsis . The Plant Cell 36 ( 10 ): 4293 – 4308 . OpenUrl CrossRef PubMed ↵ Park S , Lee CM , Doherty CJ , Gilmour SJ , Kim Y , Thomashow MF . 2015 . Regulation of the Arabidopsis CBF regulon by a complex low-temperature regulatory network . The Plant Journal 82 ( 2 ): 193 – 207 . OpenUrl CrossRef PubMed ↵ Puhakainen T , Hess MW , Mäkelä P , Svensson J , Heino P , Palva ET . 2004 . Overexpression of multiple dehydrin genes enhances tolerance to freezing stress in Arabidopsis . Plant molecular biology 54 : 743 – 753 . OpenUrl CrossRef PubMed Web of Science ↵ Qiu Y , Li M , Kim RJ-A , Moore CM , Chen M . 2019 . Daytime temperature is sensed by phytochrome B in Arabidopsis through a transcriptional activator HEMERA . Nature Communications 10 ( 1 ): 140 . OpenUrl PubMed ↵ Qiu Y , Pasoreck EK , Yoo CY , He J , Wang H , Bajracharya A , Li M , Larsen HD , Cheung S , Chen M . 2021 . RCB initiates Arabidopsis thermomorphogenesis by stabilizing the thermoregulator PIF4 in the daytime . Nat Commun 12 ( 1 ): 2042 . OpenUrl CrossRef PubMed ↵ Rains MK , Caron C , Regan S , Molina I . 2022 . Chemical and molecular characterization of wound-induced suberization in poplar (Populus alba× P. tremula) stem bark . Plants 11 ( 9 ): 1143 . OpenUrl PubMed ↵ Robinson MD , McCarthy DJ , Smyth GK . 2010 . edgeR: a Bioconductor package for differential expression analysis of digital gene expression data . bioinformatics 26 ( 1 ): 139 – 140 . OpenUrl CrossRef PubMed Web of Science ↵ Robinson MD , Oshlack A . 2010 . A scaling normalization method for differential expression analysis of RNA-seq data . Genome biology 11 : 1 – 9 . OpenUrl CrossRef ↵ Schilmiller AL , Howe GA . 2005 . Systemic signaling in the wound response . Current opinion in plant biology 8 ( 4 ): 369 – 377 . OpenUrl CrossRef PubMed Web of Science ↵ Schütte D , Remmo A , Baier M , Griebel T . 2025 . Cold exposure transiently increases resistance of Arabidopsis thaliana against the fungal pathogen Botrytis cinerea . Physiological and Molecular Plant Pathology 136 : 102579 . OpenUrl ↵ Serrano M , Coluccia F , Torres M , L’Haridon F , Métraux J-P . 2014 . The cuticle and plant defense to pathogens . Frontiers in Plant Science 5 : 274 . OpenUrl PubMed ↵ Song Y , Zhang X , Li M , Yang H , Fu D , Lv J , Ding Y , Gong Z , Shi Y , Yang S . 2021 . The direct targets of CBFs: In cold stress response and beyond . J Integr Plant Biol 63 ( 11 ): 1874 – 1887 . OpenUrl CrossRef PubMed ↵ Steponkus PL , Uemura M , Joseph RA , Gilmour SJ , Thomashow MF . 1998 . Mode of action of the COR15a gene on the freezing tolerance of Arabidopsis thaliana . Proceedings of the National Academy of Sciences 95 ( 24 ): 14570 – 14575 . OpenUrl Abstract / FREE Full Text ↵ Stockinger EJ , Gilmour SJ , Thomashow MF . 1997 . Arabidopsis thaliana CBF1 encodes an AP2 domain-containing transcriptional activator that binds to the C-repeat/DRE, a cis-acting DNA regulatory element that stimulates transcription in response to low temperature and water deficit . Proceedings of the National Academy of Sciences 94 ( 3 ): 1035 – 1040 . OpenUrl Abstract / FREE Full Text ↵ Tanaka K , Heil M . 2021 . Damage-associated molecular patterns (DAMPs) in plant innate immunity: applying the danger model and evolutionary perspectives . Annual review of phytopathology 59 ( 1 ): 53 – 75 . OpenUrl CrossRef PubMed ↵ Tateda C , Zhang Z , Shrestha J , Jelenska J , Chinchilla D , Greenberg JT . 2014 . Salicylic acid regulates Arabidopsis microbial pattern receptor kinase levels and signaling . Plant Cell 26 ( 10 ): 4171 – 4187 . OpenUrl Abstract / FREE Full Text ↵ Tran S , Ison M , Ferreira Dias NC , Ortega MA , Chen Y-FS , Peper A , Hu L , Xu D , Mozaffari K , Severns PM . 2023 . Endogenous salicylic acid suppresses de novo root regeneration from leaf explants . PLoS Genetics 19 ( 3 ): e1010636 . OpenUrl PubMed ↵ Vega-Muñoz I , Duran-Flores D , Fernández-Fernández ÁD , Heyman J , Ritter A , Stael S . 2020 . Breaking bad news: dynamic molecular mechanisms of wound response in plants . Frontiers in Plant Science 11 : 610445 . OpenUrl PubMed ↵ Wang P , Cui X , Zhao C , Shi L , Zhang G , Sun F , Cao X , Yuan L , Xie Q , Xu X . 2017 . COR27 and COR28 encode nighttime repressors integrating Arabidopsis circadian clock and cold response . Journal of integrative plant biology 59 ( 2 ): 78 – 85 . OpenUrl PubMed ↵ Wang R-S , Pandey S , Li S , Gookin TE , Zhao Z , Albert R , Assmann SM . 2011 . Common and unique elements of the ABA-regulated transcriptome of Arabidopsis guard cells . BMC genomics 12 ( 1 ): 216 . OpenUrl CrossRef PubMed ↵ Wang Z , Guo J , Luo W , Niu S , Qu L , Li J , Chen Y , Li G , Yang H , Lu D . 2025 . Salicylic Acid Cooperates With Lignin and Sucrose Signals to Alleviate Waxy Maize Leaf Senescence Under Heat Stress . Plant Cell Environ 48 ( 6 ): 4341 – 4355 . OpenUrl ↵ Wen T , Li J-H , Wang Q , Gao Y-Y , Hao G-F , Song B-A . 2023 . Thermal imaging: The digital eye facilitates high-throughput phenotyping traits of plant growth and stress responses . Science of The Total Environment 899 : 165626 . OpenUrl CrossRef PubMed ↵ Xiao J , Cao B , Tang W , Sui X , Tang Y , Lai Y , Sun B , Huang Z , Zheng Y , Li H . 2025 . The CaCAD1-CaPOA1 module positively regulates pepper resistance to cold stress by increasing lignin accumulation . International Journal of Biological Macromolecules 290 : 139979 . OpenUrl CrossRef PubMed ↵ Xie Q , Yu Q , Jobe TO , Pham A , Ge C , Guo Q , Liu J , Liu H , Zhang H , Zhao Y . 2021 . An amiRNA screen uncovers redundant CBF and ERF34/35 transcription factors that differentially regulate arsenite and cadmium responses . Plant, cell & environment 44 ( 5 ): 1692 – 1706 . OpenUrl CrossRef ↵ Xu H , Dong C , Wu Y , Fu S , Tauqeer A , Gu X , Li Q , Niu X , Liu P , Zhang X . 2024 . The JA-to-ABA signaling relay promotes lignin deposition for wound healing in Arabidopsis . Molecular plant 17 ( 10 ): 1594 – 1605 . OpenUrl CrossRef PubMed ↵ Yang L , Teixeira PJ , Biswas S , Finkel OM , He Y , Salas-Gonzalez I , English ME , Epple P , Mieczkowski P , Dangl JL . 2017 . Pseudomonas syringae Type III Effector HopBB1 Promotes Host Transcriptional Repressor Degradation to Regulate Phytohormone Responses and Virulence . Cell Host Microbe 21 ( 2 ): 156 – 168 . OpenUrl CrossRef PubMed ↵ Yi X , Wang J , Wu P , Wang G , Mo L , Lou X , Liang H , Huang H , Lin E , Maponde BT . 2023 . AC-UNet: an improved UNet-based method for stem and leaf segmentation in Betula luminifera . Frontiers in Plant Science 14 : 1268098 . OpenUrl PubMed ↵ Yu G , Wang LG , Han Y , He QY . 2012 . clusterProfiler: an R package for comparing biological themes among gene clusters . OMICS 16 ( 5 ): 284 – 287 . OpenUrl CrossRef PubMed Web of Science ↵ Zhang K , Gao R , Shen A , Wang Y , Pang G , Yang L , Liu X , Liu X , Sun J . 2026 . NJ-Unet: A novel Unet-based greenhouse tomato leaf segmentation neural network and its application . Information Processing in Agriculture . ↵ Zhang M , Zhang S . 2022 . Mitogen-activated protein kinase cascades in plant signaling . Journal of integrative plant biology 64 ( 2 ): 301 – 341 . OpenUrl PubMed View the discussion thread. Back to top Previous Next Posted May 12, 2026. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following A CBF–SA module links wound-induced evaporative cooling to tissue repair in plants Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share A CBF–SA module links wound-induced evaporative cooling to tissue repair in plants Joseph Balem , Chenjiao Tan , Nathália Cássia Ferreira Dias , Mariah Arnold , Sorrel Tran , Paul M. Severns , Paulo J.P.L. Teixeira , Changying Li , Li Yang bioRxiv 2025.05.23.655667; doi: https://doi.org/10.1101/2025.05.23.655667 Share This Article: Copy Citation Tools A CBF–SA module links wound-induced evaporative cooling to tissue repair in plants Joseph Balem , Chenjiao Tan , Nathália Cássia Ferreira Dias , Mariah Arnold , Sorrel Tran , Paul M. Severns , Paulo J.P.L. Teixeira , Changying Li , Li Yang bioRxiv 2025.05.23.655667; doi: https://doi.org/10.1101/2025.05.23.655667 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Plant Biology Subject Areas All Articles Animal Behavior and Cognition (7635) Biochemistry (17691) Bioengineering (13892) Bioinformatics (41937) Biophysics (21452) Cancer Biology (18588) Cell Biology (25504) Clinical Trials (138) Developmental Biology (13378) Ecology (19899) Epidemiology (2067) Evolutionary Biology (24320) Genetics (15609) Genomics (22506) Immunology (17736) Microbiology (40394) Molecular Biology (17181) Neuroscience (88605) Paleontology (666) Pathology (2832) Pharmacology and Toxicology (4824) Physiology (7641) Plant Biology (15156) Scientific Communication and Education (2045) Synthetic Biology (4294) Systems Biology (9825) Zoology (2271)