Circadian Clock Gates Verticillium dahliae Defence in Cotton by Modulating JAZs Expression Authors: Ping Wang, Mingming Liu, Chenyu Zhao, Qiaojia Lu, Shiqi Gao, Lijun Guo, Yuzhe Li, Jinping Wan, Li Yuan, Feifei Yi, Chaowei Cai, Jinggao Liu, C. Robertson McClung, Xiao Zhang, Xiao Liu, Xiaodong Xu, Yingfan Cai, Qiguang Xie Published: 2026 ABSTRACT ABSTRACTThe circadian clock plays critical roles in orchestrating temporal regulation of diverse physiological processes. However, its role in mediating root‐associated immunity against soil‐borne pathogens in crops remains poorly understood. Here, we show that cotton circadian clock gates nighttime resistance to vascular fungal pathogen Verticillium dahliae by shaping rhizosphere microbiome composition and activating jasmonic acid (JA)‐dependent immune signalling. Pathogen infection, in turn, accelerates clock pace and perturbs nighttime microbial community structure. We identify GhLUX, an evening complex component, as a key regulator of this defence program. GhLUX enhances immunity by repressing GhJAZ expression, thereby amplifying JA signalling at night. Overexpression of GhLUX in field trials confers enhanced resistance to V. dahliae while simultaneously improving cotton fibre yield and quality. These findings uncover a mechanistic link between the circadian clock, rhizosphere microbiota and root immunity in cotton and suggest that circadian regulators can be harnessed to optimize both disease resistance and agronomic performance. Circadian clock regulatory networks enable plants to integrate environmental temporal cues and respond rapidly to stress signals, greatly enhancing their environmental adaptation (Wang et al. 2011; Xu, Yuan, Yang, et al. 2022). Circadian gating mechanisms help to synergize interrelated or isolate conflicting physiological and biochemical activities throughout the day. In short, plant and crop circadian clocks can allocate energy and resources more efficiently during the trade‐off between growth and stress resistance (Greenham and McClung 2015; Xu, Yuan, and Xie 2022). The circadian clock regulates many important agronomic traits, which have good prospects for agricultural production applications (Steed et al. 2021). Most of the core components in the plant circadian working model established so far are transcription factors or transcriptional cofactors, and thus interlocked TTFLs (Transcriptional‐Translational Feedback Loops) constitute the core mechanism of the circadian regulation network (Nohales and Kay 2016). LUX ARRHYTHMO/PHYTOCLOCK1 (LUX) as a transcription factor recruits EARLY FLOWERING4 (ELF4) and EARLY FLOWERING3 (ELF3) to form the circadian evening complex (EC) that represses the expression of core circadian genes such as CIRCADIAN CLOCK ASSOCIATED1 (CCA1), NIGHT LIGHT–INDUCIBLE AND CLOCK‐REGULATED1/2 (LNK1/2), PSEUDO‐RESPONSE REGULATOR9/7 (PRR9/7), GIGANTEA (GI) and LUX itself at the end of the day and early night (Ezer et al. 2017; Helfer et al. 2011; Mizuno et al. 2014; Nusinow et al. 2011; Zhang et al. 2019). In turn, the circadian morning components, CCA1/LHY, repress LUX expression (Hazen et al. 2005), whereas RVE8, a related single Myb domain subfamily member, antagonistically activates LUX expression (Hsu et al. 2013). In addition, expression of LUX is also affected by PRR5 and TIMING OF CAB EXPRESSION 1 (TOC1, also known as PRR1; Huang et al. 2012; Nakamichi et al. 2012). The study of circadian clock‐gated rhizosphere microbial community diversity and function is an emerging interdisciplinary field. The rhizosphere microbiome exhibits dynamic rhythmic changes over a 24‐h period, and the host plant's internal clock or chronotype is the decisive factor in regulating these dynamics (Bending et al. 2025; Xu and Dodd 2022). The endogenous plant circadian clock may affect the composition and abundance of root exudates by regulating the rhythm of plant metabolism, further driving the rhythmicity of plant‐microbe carbon metabolism and exchange (Staley et al. 2017). Genetic perturbation of the plant circadian clock in lines including lhy‐11, LHY‐ox, CCA1‐ox, toc1‐21 or ztl‐30 (inhibiting TOC1 degradation) alters the rhythmicity of specific rhizosphere microbial communities under diurnal conditions (Hubbard et al. 2018; Newman et al. 2022; Staley et al. 2017). In rice, there were huge differences in the composition and activity of the rhizosphere microbial communities between a light–dark cycling regimen and continuous darkness (Zhao et al. 2021). This likely reflects the dramatic environmental changes in the rhizosphere induced by the light–dark cycle. Photosynthetic activity in the light induced an increase in oxygen and dissolved organic carbon leading to unique microbial assemblies in light versus dark. Similarly, oxygen released into the otherwise anaerobic rhizosphere from rice roots during photosynthesis alters the redox and electron supply status in the rhizosphere. However, whether this situation is unique to rice needs further study in other plant species (Zhao et al. 2021). Cotton (Gossypium spp.) is the most important natural fibre crop worldwide, yet its fibre yield and quality have been seriously restricted by Verticillium wilt, caused by Verticillium dahliae (V. dahliae) (Umer et al. 2023). The plant hormone jasmonates (JAs) balance immune and growth responses and improve plant survival in harsh environments (Howe et al. 2018; Li et al. 2023). Previous studies have indicated that the JA biosynthesis and signal transduction pathways in cotton play major roles in V. dahliae resistance (Shaban et al. 2021; Sun et al. 2017; Xiong et al. 2021). JAZs serve as a key regulatory node, and different family members are regulated at both transcriptional and post‐transcriptional levels and may serve as a switch that balances resistance and growth pathways. For example, GhWRKY1 and GhWRKY70 regulate the abundance of JAZs at the transcriptional level (Li et al. 2014; Zhang et al. 2023), whereas GhRVE2 acts at the post‐transcriptional level (Liu et al. 2023), ultimately regulating resistance to V. dahliae through the JAZ‐MYC module. Increasing evidence shows that the circadian clock governs numerous elements of JA signalling and response including the timing of biosynthetic gene expression and sensitivity to JA (Covington et al. 2008; Shin et al. 2012; Zhang et al. 2019). However, how the circadian system interacts with the JA pathway during V. dahliae infection in cotton remains poorly understood. Although studies in several crop species have elucidated circadian control of agronomic traits, equivalent research in cotton has lagged, largely due to its complex polyploid genome. In Arabidopsis, the core evening complex component LUX not only maintains circadian rhythmicity but also regulates salicylic acid (SA) and JA signalling, contributing to resistance against foliar pathogens such as Pseudomonas syringae (Zhang et al. 2019). Despite this, it is unknown whether LUX also mediates clock‐dependent regulation of root‐associated microbial communities. Here we investigated how the endogenous circadian clock of cotton gates the composition of the rhizosphere microbial community, with a specific focus on the role of the LUX ortholog GhLUX. Using genetic manipulation of GhLUX expression, we assess its impact on microbial community structure and resistance to V. dahliae. Additionally, we explore the dynamic interactions between V. dahliae and other rhizosphere microbes, uncovering a previously unrecognized role for V. dahliae as a Zeitgeber—a temporal cue capable of resetting the host circadian system. Our data show that V. dahliae can modulate clock function via GhLUX, leading to the activation of JA‐dependent defence pathways. This study provides the first direct evidence that GhLUX enhances cotton defence against a soil‐borne vascular pathogen through circadian regulation. Importantly, our findings bridge a critical gap between model systems and crop species, offering translational insights into circadian‐mediated immunity in economically important plants. As highlighted in recent literature, the validation of model‐derived regulatory mechanisms in crop systems is essential for advancing their practical application in agriculture (Roeder et al. 2025). Results The Cotton Circadian Clock Maintains the Stability of the Rhizosphere Microbiome at Nighttime The optimization of root‐associated bacterial and fungal communities, the rhizosphere microbiome, can improve soil properties, increase plant nutrient utilization efficiency, and inhibit pathogen infestation of plants through roots, promoting plant growth, health and resistance (Lee et al. 2021; Wang et al. 2024). The rhizosphere microbiome shows temporal dynamics throughout the 24‐h day, and plant‐microbial interactions are influenced both by external periodic environmental cues and the plant's endogenous circadian clock (Newman et al. 2022; Zhao et al. 2021). However, how the plant circadian clock affects the rhizosphere microbial communities is incompletely understood. To investigate the crosstalk between the cotton circadian clock and the rhizosphere microbiome, we disrupted the cotton circadian clock by artificially disrupting the environment's normal photoperiod, and then compared the diel changes of the rhizosphere microbiome in disrupted clock (DC) plants with that in normal circadian clock (NC) plants (Figure 1A). We disrupted the cotton's circadian clock by reversing the light–dark (LD) cycles every 2 days. This protocol has significant advantages because the cotton's clock needs to be reset every 2 days, which will have an extremely adverse effect on growth, development and environmental adaptability. qRT‐PCR results showed that, the transcriptional peak phases of GhLUX and GhCCA1 were significantly shifted, and their amplitude fluctuations became more irregular under DC conditions compared with NC (Figure 1B). After DC treatment, significant phase shifts were observed in the other core circadian genes. The robustness of the morning gene GhPRR9 was significantly reduced, and the evening genes GhLUX, GhTOC1, GhELF3 and GhELF4 also exhibited double expression peaks in the second cycle (Figure S1). Leaf movement rhythm can directly reflect changes in the endogenous circadian clock, and we also examined the effects of DC treatment on leaf movement rhythm. The result show that the phase of leaf movement in DC‐treated plants was significantly shifted, indicating that the DC experiment had achieved the goal of disrupting the circadian clock (Figure S2). That suggesting that DC treatment resulted in circadian clock disruption or uncoupling of the feedback loops. The cotton circadian clock maintains the stability of rhizosphere microbiome at nighttime. (A) Illustration of LD cycles designed to disrupt circadian rhythms. Normal circadian clock (NC), the 12‐h light/12‐h dark (LD) cycles is not disturbed; Disrupted circadian clock (DC): Initially a 27‐days LD cycle, which was then set to a 2‐days DL + 2‐days LD + 2‐days DL combination, with sampling at ZT1 and ZT13 under the next LD photoperiod condition. White or black boxes represent light or dark, respectively. (B) qRT‐PCR assay of the transcriptional patterns of the representative circadian evening gene LUX and the early morning gene CCA1 in leaves under NC and DC conditions within 2 days from the start of sampling. (C) Rhizosphere microbiome α‐diversity (Chao1, Shannon and Simpson index) in the soils of each treatment (n = 8 soil samples). The box plots indicate the median (center line), the 25th and 75th percentiles (box), and the range of non‐outlier values (whiskers). Letters indicate significant difference among the samples as determined by one‐way ANOVA analysis using the Tukey's honestly significant difference (HSD) test (p < 0.05). (D–F) Principal coordinates analysis (PCoA) based on Bray–Curtis distances displaying a significant difference in the rhizosphere microbiomes (PERMANOVA). (G, H) Differences in the relative abundances of fungal phyla in the top 10 genus categories between disrupted and normal circadian clock in ZT1 and ZT13. Data are means ± SE of eight biological replicates; p‐values were calculated using pairwise Student's t‐test (*p < 0.05; **p < 0.01; ***p < 0.001). After the cotton seedlings were cultured under a light–dark cycle (12L/12D) for 27 days, they were cultured for another 6 days or transferred to jet lag conditions (altering the photoperiod, which caused DC) for 6 days. Then at time points ZT1 or ZT13 (Zeitgeber Time 1 or ZT 13, where ZT 0 corresponds to lights on), root zone soil samples were collected at the indicated time points (Figure 1A). Both the experimental group and the control group were subjected to the same light–dark cycle at the same sampling time points. Therefore, through experimental design, when analysing and comparing the data later, the possibility that light might influence the rhizosphere microbial community through the vascular bundles can be ruled out. Sequencing raw reads were processed through quality control, yielding high‐quality clean reads, followed by denoising to obtain denoised reads. These were merged and non‐chimeric reads were retained for final analysis (Table S1). The results showed that, DC exhibited significant effects on fungal α‐diversity (p < 0.05) at ZT13. Although there was no significant difference in the Chao1 index (species richness), both the Simpson index (dominance and stability, higher values indicating greater dominance) and the Shannon index (integrating richness and evenness, higher values representing higher diversity and even distribution) increased significantly at ZT13 between DC and NC (Figure 1C). Principal coordinate analysis (PCoA) based on Bray–Curtis distances revealed high similarity in β‐diversity of the cotton rhizosphere fungi diversity at ZT1 and ZT13 in NC. Specifically, permutational multivariate analysis of variance (PERMANOVA) analysis indicated no significant difference in fungal β‐diversity between ZT1 and ZT13 (R 2 = 0.073, p‐value = 0.202) (Figure 1D). We then compared the differences in the rhizosphere microbiome at ZT1 and ZT13 under conditions of DC and NC. PCoA results at ZT1 showed no significant differences between two conditions (R 2 = 0.083, p‐value = 0.088) (Figure 1E). However, at ZT13, compared with the NC, the eight biological replicates of the DC experiment were clustered in relatively independent areas, and the degree of dispersion between samples was large (R 2 = 0.132, p‐value = 0.001) (Figure 1F). We also sampled non‐planted soil as a control group to rule out the possibility that light and dark cues directly affect rhizosphere microbiome. Analysis using the Chao1 index, Simpson index, and Shannon index revealed no significant differences, particularly at ZT13, where there was no significant difference between NC and DC (Figure S3). These results indicate that even if light and dark cues can be directly transmitted to soil microorganisms (e.g., through the plant's vascular bundle), they do not cause differences in rhizosphere microbiome. Therefore, the differences in rhizosphere microbiota at ZT13 are due to the plant's endogenous circadian clock. Using Metastats, we compared microbial communities between various conditions: ZT1 of NC versus ZT13 of NC, ZT1 of NC versus ZT1 of DC and ZT13 of DC versus ZT13 of NC (Table S2). Unexpectedly, there were no significant changes in the relative abundance of the top 10 fungal genera between in‐phase ZT1 and in‐phase ZT13 samples (Figure S4). However, the decrease in Penicillium abundance (DC vs. NC) was more significant in ZT13 compared with ZT1, whereas unclassified Ascomycota increased significantly in ZT1 (DC vs. NC), which was even not ranked in the top 10 in ZT13 (Figure 1G,H). Furthermore, at ZT13, the disrupted circadian rhythm affected multiple fungal taxa, including Fusarium, Candida, Penicillium, Trichoderma and Mortierella, showing significant changes (Figure 1H). In summary, disruption of the cotton circadian clock had less significant effect on the rhizosphere microbiome at ZT1, but it caused significant changes at ZT13, suggesting that the cotton circadian clock plays a key role in maintaining the stability of the rhizosphere microbiome at nighttime. Verticillium dahliae Inoculation Accelerates the Circadian Clock and Alters the Rhizosphere Microbiome of Cotton Verticillium dahliae, a widespread soil‐borne vascular pathogen, is the main pathogen causing the devastating Verticillium wilt of cotton. In this study, we found that although the overall abundance of Verticillium was low (~0.1%) and showed no significant difference between ZT1 and ZT13 in NC, its abundance significantly decreased at ZT1 in DC versus NC (Figure S5). Further analysis of cotton roots inoculated with V. dahliae revealed a significant increase in Verticillium abundance at both ZT1 and ZT13, confirming successful inoculation (Figure 2A). PCoA analysis showed that microbial communities in eight inoculated biological replicates clustered separately from mock at both time points (R 2 = 0.096, p‐value = 0.008 at ZT1; R 2 = 0.103, p‐value = 0.002 at ZT13), indicating that V. dahliae inoculation significantly altered the composition of fungal communities in the cotton rhizosphere (Figure 2B,C). Further analysis showed that V. dahliae inoculation at ZT1 suppressed Penicillium abundance, whereas at ZT13, significant differences were observed in Fusarium, Penicillium and Mortierella abundances (Figure 2D,E and Table S3), indicating that the endogenous circadian clock of cotton gates the response of microbial communities to V. dahliae treatment. Verticillium dahliae inoculation accelerates the circadian clock and alters the rhizosphere microbiome of cotton. (A) Differences in the relative abundances of Verticillium in the genus categories between ZT1 and ZT13 without or with V991 infection in rhizosphere microbiome. Data are means ± SE of eight biological replicates; p‐values were calculated using pairwise Student's t‐test (**p < 0.01). The photoperiodic conditions, inoculation and sampling time points are shown above. (B, C) Principal coordinates analysis (PCoA) based on Bray–Curtis distances displaying a significant difference in the rhizosphere microbiomes (PERMANOVA) between normal circadian clock in ZT1 and ZT13. (D, E) Differences in the relative abundances of fungal phyla in the top 10 genus categories in ZT1 and ZT13 after V991 infection. Data are means ± SE of eight biological replicates; p‐values were calculated using pairwise Student's t‐test (*p < 0.05; **p < 0.01; ***p < 0.001). (F, G) Verticillium dahliae V991 displayed morphological circadian rhythms, and cotton root extract in the daytime (ZT1) had a stronger inhibitory effect on hyphal growth of V. dahliae. The red line represents the hyphal growth diameter after pasting the round filter paper. Bar = 1 cm. For statistical diagram of hyphal growth diameter, data are means ± SD of three biological replicates; Letters indicate significant difference among the samples as determined by a one‐way ANOVA analysis using the Tukey's honestly significant difference (HSD) test (p < 0.05). (H) Verticillium dahliae infection altered the endogenous circadian rhythm of cotton cotyledon movement. Circadian period length with V991 treatment around ZT1 or ZT13 was statistically analysed. White or grey boxes represent subjective day or night. To further investigate the feedback regulation of V. dahliae on the cotton circadian clock, we first detected the diurnal (light–dark cycles) and circadian (constant conditions) phenomenon of V. dahliae growth patterns. We cultured V. dahliae under light–dark cycles (12L/12D) and then transferred to continuous darkness (DD). In DD, the cultures showed a uniform growth pattern, whereas under 12L/12D cycles, concentric rings of growth appeared, suggesting that the growth of V. dahliae has a diurnal rhythm but fails to display an overt circadian rhythm under DD (Figure 2F), which is consistent with earlier reports (Cascant‐Lopez et al. 2020). We then extracted root exudates from cotton at ZT1 and ZT13 to explore their effects on V. dahliae growth. Both ZT1 and ZT13 root exudates inhibited V. dahliae growth, with ZT1 showing a stronger inhibitory effect (Figure 2G). Pathogenesis‐related (PR) genes, PR1 and PR4, were significantly induced after V. dahliae inoculation, with a stronger induction at ZT1 (Figure S6), suggesting that the cotton circadian clock gates resistance to V. dahliae at the genetic and metabolic levels. Furthermore, we used rhythmic cotyledon movement to analyse the effect of V. dahliae infection on the cotton circadian clock. V. dahliae isolate V991 inoculation at ZT1 and ZT13 resulted in a short‐period length of leaf circadian movement phenotype (Figure 2H and Table S4), indicating that V. dahliae infection indeed affects the cotton circadian clock, consistent with previously reported observations that bacterial infection causes acceleration of the circadian pace in Arabidopsis (Zhang et al. 2013). The Circadian Clock Gates Cotton Resistance to Verticillium dahliae Infection After V. dahliae infects susceptible cotton plants, symptoms of Verticillium wilt gradually worsen, progressively causing the yellowing of leaves, stem dryness and root rot, ultimately severely affecting the yield and quality of cotton. Although damage to cotton is a long‐term process, our experimental results show that cotton's resistance to V. dahliae is gated by the circadian clock such that resistance is stronger in the early morning than in the evening. To determine whether the cotton circadian clock gates resistance to V. dahliae infection under more realistic conditions, we transferred cotton (Cultivar, ‘BM‐1’) seedlings cultured in 12L/12D for 27 days under strict greenhouse conditions to NC and DC photoperiods for continuous growth. Three days later, V. dahliae infection was performed at ZT1 and ZT13 (Figure 3A). Compared to wild‐type plants, cotton seedlings with disrupted circadian rhythms did not show a significant reduction in vegetative growth phenotype before inoculation (Figure S7A). If the circadian clock gates cotton's resistance to V. dahliae, then there will be temporal differences in infection efficiency. After 16 days post‐inoculation (dpi), cotton grown under DC showed more disease symptoms compared with NC (Figure 3B), indicating that the circadian clock is involved in regulating resistance to V. dahliae. We harvested tissues at ZT1 and ZT13 and further compared the fungal biomass of V991, disease index and fungal recovery rate in DC and NC (Figure 3C–E). Notably, all three assays consistently showed that disrupting the circadian clock significantly reduced V. dahliae resistance, and the resistance was stronger at ZT1 than at ZT13. The evening‐phased oscillator GhLUX responds to induction of Verticillium dahliae. (A–E) Disrupted circadian clock altered disease symptoms of cotton plants with V. dahliae inoculation and biomass of V. dahliae V991 in cotton roots. Schematic representation of LD cycle entrainment and of disrupted LD cycles used in this study (A). Disease symptoms of cotton plants at 16 dpi with V. dahliae (B). Cultivars, ‘BM‐1’. dpi, days post inoculated. Relative biomass of V991 in cotton roots (C). Roots were harvested at 3 dpi with V. dahliae V991 at ZT1 or ZT13. Genomic V. dahliae DNA was extracted for qPCR quantify the biomass of V991. GhUBQ7 was used as a normalization control. Data are shown as mean ± SD of Three technical replicates from one of three independent biological experiments (n = 3; ≥ 6 plants/biological replicate); all experiments yield congruent results. Relative disease index and fungal renewal cultivation levels in cotton plants shown in (D, E). For statistical diagram of relative disease index, data are means ± SD of three biological replicates (n = 3; ≥ 30 plants/biological replicate); the disease index of plants V. dahliae‐inoculated at ZT1 under normal circadian clock was normalized to 1. Letters indicate significant difference among the samples as determined by a one‐way ANOVA analysis using the Tukey's honestly significant difference (HSD) test (p < 0.05). The disease index = [(∑grade n × the number of diseased plants at the corresponding grade)/(total cotton plants × 4)] × 100. (F, G) Relative biomass of V. dahliae in cotton roots. 12L/12D cycle entrainment shown in (F), black arrow with ZT1, 13 represents timing of roots inoculation with V991. Roots were harvested at 3, 6 and 9 h post inoculated (hpi). Data are shown as mean ± SD of three technical replicates from one of three independent biological experiments (n = 3; ≥ 6 plants/biological replicate); all experiments yield congruent results. (H) Fold change of transcript accumulation of the evening‐phased core oscillators in different cotton tissues after V. dahliae inoculation. Cotton roots were inoculated with V991 at ZT13, then the first true leaves, stems and roots (4 plants mixed for one sample) were harvested at 0, 3, 6 and 9 h post inoculation. GhUBQ7 was used as a normalization control. Data are shown as means ± SD of three technical replicates from one of three independent biological experiments; all experiments yield congruent results. Although V. dahliae V991 exhibited growth rhythmicity only under LD in culture medium (driven by the external photoperiod), related studies have not determined whether V. dahliae exhibits circadian rhythmicity (regulated by the plant's endogenous metabolic circadian rhythms) in host plants or under specific environmental conditions (Cascant‐Lopez et al. 2020). To further test whether cotton's endogenous clock enhances defence against V. dahliae, we compared the relative biomass of V. dahliae V991 in cotton roots at 3 h, 6 h and 9 h post inoculated (hpi) entrained either in‐phase or out‐of‐phase (Figure 3F). The results showed that after inoculation with V. dahliae at ZT1, there was no significant difference in the accumulation of V991 biomass in cotton roots, regardless of whether the photoperiods of cotton and V. dahliae were synchronized or not. However, at ZT13, the accumulation of V991 biomass in cotton roots significantly increased under the opposite photoperiod condition compared to the synchronized condition (Figure 3G). These data suggest that when cotton and V. dahliae share the same diurnal cycle, cotton is able to mount relatively effective defences against V. dahliae in both the morning and evening. However, when their photoperiods are not synchronized, cotton's nighttime defence capability is more strongly diminished. Since cotton's time‐specific defence responses are gated by the circadian clock, we hypothesized that cotton core circadian evening genes may play an important role. Therefore, we inoculated cotton with V. dahliae spore suspensions or water (mock) at ZT13 and investigated the transcript abundance of GhLUX, GhELF3 and GhTOC1 evening gene copies in different cotton tissues (Figure 3H and Figure S7A–E). The results showed that these genes responded differently to V. dahliae, with GhLUX showing the strongest response, particularly GhLUX07. Furthermore, the expression patterns of these genes varied across tissues, being primarily induced in roots and stems (Figure 3H and Figure S8). These findings suggest that the core circadian clock genes GhLUX, GhELF3 and GhTOC1 respond to V. dahliae infection at night and may be involved in cotton's resistance to the pathogen. GhLUX, Component of the Evening Complex (EC), Positively Regulates Cotton Resistance to Verticillium dahliae and Increases Cotton Yield and Fibre Quality Next, we focused on the role of the core circadian evening genes GhLUX, GhELF3 and GhTOC1 in cotton resistance to V. dahliae. Through sequence blast and alignment, we identified a total of six LUX homologous genes in cotton. Based on the chromosomes where these LUX genes are located, we named them GhLUXA03, GhLUXD03, GhLUXA05, GhLUXD05, GhLUXA07 and GhLUXD07 (Figures S9F and S10). We selected fragments that specifically target GhLUX05 (GhLUXA05 and GhLUXD05) and GhLUX07 (GhLUXA07 and GhLUXD07) (Figure S11). GhELF3 and GhTOC1 each have two copies in the cotton genome, and their expression is spatiotemporally specific, both being nighttime genes (Figure S1 and Figures S9D and S12). The CDS sequences of the two GhELF3s (Gh_A09G097800 and Gh_D09G093100) are 96.79% similar and highly conserved (Figure S9A). The CDS sequences of the two GhTOC1s (Gh_A03G084800 and Gh_D03G106900) are 97.95% similar and also highly conserved (Figure S13A). We selected fragments that specifically target GhELF3s and GhTOC1s to silence these clock genes through Virus‐induced gene silencing (VIGS) (Figure S14A). Since previous results have demonstrated that LUX, a core component of the circadian clock, is an important node that mediates circadian clock function and defence responses (Zhang et al. 2019), we first investigated the role of GhLUXs in resistance to V. dahliae. Preliminary experiments showed that silencing the GhLUXs gene decreased cotton (TM‐1) resistance to V. dahliae. To test the importance of GhLUX more rigorously in cotton resistance to V. dahliae, we used VIGS to knockdown GhLUX expression in the resistant cultivar Zhongzhimian 2 (ZZM2). Compared to mock treatment, VIGS knockdown cotton seedlings did not show a significant reduction in vegetative growth phenotype before inoculation (Figure S7B). After inoculating V. dahliae, the silenced TRV:GhLUX05 or TRV:GhLUX07 plants showed more severe symptoms and higher disease index than the control (TRV:00) plants (Figure 4A,D). We found that TRV:GhLUX05 and TRV:GhLUX07 could specifically and effectively inhibit the corresponding endogenous genes without affecting other GhLUX copies (Figure 4B,C and Figure S15), indicating that both of them are involved in regulating resistance to V. dahliae (Figure 4D,E). Consistent with this, observed browning of vascular tissue, and fungal recovery assay, all supported that reduced expression of GhLUXs reduced cotton resistance to V. dahliae (Figure S16A,B). However, silencing of GhELF3s and GhTOC1s had little effect on cotton resistance to V. dahliae (Figures S13B and S14B). Although the qRT‐PCR results demonstrated that we had effectively inhibited the expression of GhELF3s and GhTOC1s (Figures S13C and S14C), further disease index, observed browning of vascular tissue, and fungal recovery assay all indicated that both GhELF3s and GhTOC1s made little contribution to V. dahliae resistance (Figures S13D–F and S14D–F). This suggests that not all core evening genes of the circadian clock are involved in V. dahliae resistance, and perhaps only GhLUXs play a dominant role in defence against V. dahliae. Cotton GhLUXs positively regulates the resistance to Verticillium dahliae. (A–E) Silencing GhLUXs reduced resistance to V. dahliae in cotton. Disease symptoms of TRV:00 and GhLUXs‐silenced seedlings (resistant cultivars ‘ZZM2’) at 16 dpi with V991 (A). TRV:00: Negative control, with no targeted gene fragment; TRV:GhLUX05: Targeted to GhLUXA05 and GhLUXD05. TRV:GhLUX07: Targeted to GhLUXA07 and GhLUXD07; Expression of GhLUX05 and GhLUX07 in roots of TRV:00 and GhLUXs‐silenced cotton plants (B, C). GhLUX05 and GhLUX07 transcripts level was analysed by qRT‐PCR. Disease index (D) in TRV:00 and GhLUXs‐silenced cotton plants shown in (A). Data are means ± SD of three biological replicates (n = 3; ≥ 30 plants/biological replicate, **p < 0.01, Student's t‐test). Relative biomass of V991 in the roots of TRV:00 and GhLUXs‐silence (TRV:GhLUXs) plants after inoculated in ZT1 or ZT13 (E). (F–H) Phenotype of BM‐1 and GhLUX‐OE4 (Overexpression of GhLUXA07) at adult stage in Verticillium wilt artificial disease nursery. Scale bar = 10 cm. Statistics of disease incidence (G) and disease index (H) of plants in (F). Jimian11, susceptible control; ZZM2, resistant control. Data are the mean ± SD of three biological replicates. The statistical analysis method is one‐way ANOVA. Different letters represent p < 0.05. (I, J) GhLUX overexpression improved the fibre quality of cotton. Fibre quality phenotypes (I), length, strength (cN•tex−1), and spinning consistency index (J) of wild type (BM‐1) and GhLUX‐OE plants. Data are the mean ± SD of three biological replicates. (**p < 0.01, ***p < 0.001, Student's t‐test); Bars = 1 cm. To further validate the role of GhLUX in Verticillium wilt resistance and its potential application in cotton production, we constructed a cotton GhLUXA07 overexpression vector driven by the CaMV35S promoter. Stable transgenic plants were obtained through genetic transformation (Figure S17). We first assessed the V. dahliae resistance of these GhLUXA07 overexpression plants. After inoculation with V. dahliae V991, the vascular browning and mycelium formation in excised cotton stems were reduced in GhLUX‐OE3 and GhLUX‐OE4 plants, consistent with their fungal renewal cultivation levels (Figure S16C,D). G. hirsutum cultivar BM‐1 exhibited severe wilting and chlorosis, whereas the GhLUX‐OE3 and GhLUX‐OE4 plants showed milder symptoms and lower disease index (Figure 4F–H and Figure S18A,B). To investigate whether GhLUX participates in the early defence response against V. dahliae, we inoculated the plants with V. dahliae V991 at ZT13. The results showed that at 3, 6 and 9 hpi, the biomass of V. dahliae in GhLUX‐OE plants was reduced, and fewer V991‐GFP spores were attached to the root surface (Figure S18C,D), indicating that GhLUX also regulates the early immune response to V. dahliae. These results indicate that overexpression of GhLUXA07 significantly enhances cotton's resistance to V. dahliae. We further investigated the conservation of LUX function in different plant species, and the results showed that overexpression of GhLUXA07 in Arabidopsis lux‐4 mutant and wild‐type could rescue and enhance the resistance phenotype to Verticillium wilt (Figure S19). Analysis of cotyledon movement showed that overexpression of GhLUX led to a gradual decrease in the robustness of the circadian rhythm (Figure S20), which is similar to previously published results in Arabidopsis. Cotton is an important fibre crop, but there is often a trade‐off between disease resistance and growth/development traits. Typically, enhanced disease resistance negatively affects growth traits including fibre yield and quality. Therefore, we evaluated the fibre yield and quality of the GhLUX‐OE plants. The results showed that overexpression of GhLUX increased fibre length, strength and spinning consistency index (Figure 4I,J). Additionally, GhLUX‐OE plants exhibited increased unginned cotton, ginned cotton and boll weight, but no significant changes were observed in lint percentage or the number of bolls per plant. The overall improvement in these agronomic traits led to an increase in fibre yield (Tables S5 and S6). These results suggest that overexpression of GhLUX not only enhances resistance to V. dahliae but also improves fibre yield and quality without compromising cotton growth and development. GhLUX Inhibits GhJAZs of JA Pathway to Enhance Resistance Against Verticillium Wilt Given that the overexpression of GhLUX enhances resistance to V. dahliae and improves cotton fibre yield and quality, we aimed to uncover the regulatory mechanisms underlying this phenomenon, providing a foundation for molecular genetic breeding. We hypothesized that GhLUXs, as part of the circadian EC, may directly regulate downstream target genes to confer resistance to Verticillium wilt or indirectly modulate resistance through the circadian clock output pathways. First, we confirmed that under LD conditions, the transcript abundance of GhLUXA07 in the overexpressing lines (GhLUX‐OE3 and GhLUX‐OE4) was significantly higher than those of the endogenous GhLUXA07 in the BM‐1, and lost rhythmicity (Figure S21A). We also analysed the expression patterns of other core circadian genes in the overexpression lines and BM‐1. In contrast to GhLUXA07, the expression of GhCCA1 was reduced in GhLUX‐OE3 and GhLUX‐OE4, but its expression phase was not significantly altered, with GhCCA1 still peaking in the morning (Figure S21B). Similarly, the expression of GhPRRs (GhPRR9, GhPRR7, GhPRR5 and GhTOC1) and other EC complex genes (GhELF4, GhLUX03 and GhLUX05) was also suppressed in the GhLUX‐OE lines. Although the transcriptional robustness was reduced, the expression phase remained unaffected (Figure S21C–I). These results suggest that sustained overexpression of GhLUXA07 inhibits the transcription of core circadian genes. To explore the potential mechanism by which GhLUX regulates cotton resistance to V. dahliae, we performed RNA‐Seq experiments using the GhLUX overexpression lines inoculated with V. dahliae V991. By comparing the differentially expressed genes (DEGs) between GhLUX‐OE4 and BM‐1 plants inoculated with V. dahliae V991 (Table S7), we found that the upregulated DEGs were primarily enriched in biological processes related to cell wall modification, pectin catabolic processes and xyloglucan metabolic processes. These processes are associated with cell wall biogenesis, as well as immune responses such as chitin catabolic processes and systemic acquired resistance (Table S8). Notably, eight downregulated DEGs in the jasmonic acid (JA)‐mediated immune response pathway were all GhJAZs (Figure 5A). We then analysed the expression of all members of the GhJAZ family (30 genes in total) in upland cotton ( Gossypium hirsutum ), and as expected, most GhJAZ genes were downregulated in the GhLUX‐OE4 plants (Figure 5B). Further analysis of the promoter regions (3000 bp upstream of the ATG codon) of these GhJAZ genes revealed that 27 out of 30 GhJAZ promoters contained LUX Binding Site (LBS) elements. We further examined the diurnal expression patterns of representative GhJAZ2, GhJAZ11 and GhJAZ26 genes in BM‐1 and GhLUX‐OE plants. Each of these three genes exhibited a 24‐h diurnal rhythm in BM‐1, but their expression was significantly suppressed in the GhLUX‐OE plants (Figure 5C). We inoculated plants with V. dahliae V991 at ZT13 and then compared the transcriptional response between GhLUX overexpressing lines and BM‐1. The basal expression levels of GhJAZ2 and GhJAZ11 were already significantly reduced in the GhLUX overexpression plants. We hypothesized that GhJAZ2 and GhJAZ11 may be circadian target genes, which transcription peak is approximately ZT16 (= ZT13 + 3), so there showed a turning point at 3 hpi. Presumably GhLUX is a negative regulator of GhJAZ2 and GhJAZ11 expression, and thus the most significant differences in GhJAZ2 and GhJAZ11 transcript abundance were found between the GhLUX‐OE and BM‐1 strains at 3 hpi. In addition, we found that GhJAZ26 was slightly upregulated at 3 hpi in GhLUX‐OE lines compared to BM‐1, although the abundance of JAZ26 transcripts was consistently low in the GhLUX‐OE lines. In contrast, as a control, the expression of the JA signalling pathway marker gene GhPR4 was significantly upregulated 6 and 9 h after V. dahliae infection in both BM‐1 and in GhLUX‐OE lines (Figure 5D). GhLUX inhibited the rhythmic expression of GhJAZs to enhance cotton's response to Verticillium dahliae infection. (A) GO analysis of DEGs genes between BM‐1 and GhLUX‐OE plants after V. dahliae infection. For RNA‐sequencing, cotton BM‐1 and GhLUX‐OE seedlings were subjected to 23°C, 12L/12D cycles for 3 weeks. The roots were inoculated with V991 at ZT13 and sampled at 3 h post inoculation for RNA isolation. (B) Heatmap of GhJAZs expression in cotton. BMM, BM‐1 plants treated with water (Mock); OEM, GhLUX‐OE4 plants treated with water (Mock); BMV, BM‐1 plants inoculated with V991; OEV, GhLUX‐OE plants inoculated with V. dahliae. The FPKM values were used to draw the heatmap. Clustered and normalized (0–1) by row. (C) The diel enrichment of GhJAZs transcript in BM‐1 and GhLUX overexpression plants. Seedlings were entrained at 23°C, 12L/12D cycles for 3 weeks before roots were harvested every 4 h for RNA isolation. (D) The GhJAZs and GhPR4 transcripts' accumulation in roots after V. dahliae inoculation. (E, F) LUC imaging in tobacco leaves was used to analyse the regulatory effects of GhLUXA07 on GhJAZ2, GhJAZ11 and GhJAZ26 promoters. (G) Y1H system assay showing GhLUXA07 interacting with GhJAZ2, GhJAZ11 and GhJAZ26 promoters. (H) ChIP‐qPCR assay enrichment of p35S::GhLUXA07‐Flag with GhJAZ2, GhJAZ11 and GhJAZ26 promoters in roots. GhUBQ was analysed as a negative control. Data are the mean ± SD of three biological replicates (**p < 0.01, Student's t‐test). (I) The cotton circadian clock exhibits enhanced nighttime communication with the rhizosphere microbiome, which in turn modulates immune signalling. We used a transient expression system in the model plant Nicotiana benthamiana to further verify that overexpression of GhLUXA07 could significantly inhibit the bioluminescence intensity of luciferase driven by the GhJAZ2, GhJAZ11 and GhJAZ26 promoters (Figure 5E,F). Yeast one‐hybrid analysis also demonstrated that GhLUXA07 could directly bind to the promoters of GhJAZ2, GhJAZ11 and GhJAZ26 (Figure 5G). Finally, we performed the ChIP experiments to confirm that GhLUXA07 can indeed bind to the specific LBS cis‐elements of the GhJAZ2 and GhJAZ11 promoters in vivo, but no immediate enrichment signal was detected in the LBS region of the GhJAZ26 promoter. However, GhLUXA07 was still detected to be enriched to a certain extent in the non‐LBS region of the GhJAZ26 promoter, suggesting that there may be other potential binding sites or other reasons (Figure 5H). Since GhLUX regulates the transcription of almost all GhJAZ family members, we further investigated genes, such as GhPDF1.2, GhVSP, GhVSP2‐2, GhTHI2‐1, GhTHI2‐2 and GhTHI2‐4, which are downstream of the JA signalling pathway that may be involved in V. dahliae resistance. The results showed these JA downstream genes were generally downregulated in GhLUX knockdown VIGS lines and significantly upregulated in GhLUX overexpression lines. These results indicate that GhLUX enhances V. dahliae resistance by inhibiting the GhJAZ family and thereby activating the JA signalling pathway (Figure S22). Taken together, these results support a model in which the cotton circadian clock can shape rhizosphere community structure and microbe–microbe interactions, thereby affecting the temporal‐specific infection and resistance to V. dahliae. The core component of the circadian clock, GhLUX, can quickly respond to the infection of V. dahliae. As a transcriptional repressor of the circadian clock in the evening, it can inhibit the expression of core genes GhJAZ2, GhJAZ11 and GhJAZ26 in the JA signalling pathway, ultimately gating resistance to Verticillium wilt (Figure 5I). Discussion Endogenous circadian clocks are timekeeping mechanisms found broadly among eukaryotes and some prokaryotes. Although their core components are different, most eukaryotic clocks share the mechanism of TTFLs, which gate multiple physiological and biochemical activities in output pathways at different levels, ranging from transcription to metabolism. Cotton may have more complex circadian clock TTFLs due to the genome‐wide gene multicopy phenomenon (Hao et al. 2021; Lou et al. 2012; Wang, Du, et al. 2020). Conversely, V. dahliae, whereas it has the key components of the fungal clock, only exhibits diurnal rhythms under the conditions studied so far (Figure 2F; Cascant‐Lopez et al. 2020), but the presence or absence of endogenous circadian rhythms under the post‐infestation parasitic stage is not clear. Here we show that disrupting the cotton circadian clock by repeatedly reversing the photoperiod had a significant effect on the composition and structure of rhizosphere community at ZT13 (Figure 1C–H). Further experimental results of inoculating V. dahliae at different time points also support that ZT13 has a significantly higher impact on rhizosphere microorganisms than ZT1 (Figure 2B–E). The results suggest that the cross‐talk between the cotton circadian clock and rhizosphere microorganisms may mainly occur at night. Previous studies on exudate profiles have shown that specific soluble sugars, lipids and flavonoids, secondary metabolites closely related to plant‐microbe interactions, exhibit diurnal fluctuations in their secretion patterns (Badri et al. 2010; Brauner et al. 2018; McLaughlin et al. 2023; Robert et al. 2025). The cotton LUX involved in this study belongs to the MYB family transcription factor (type SHAQKYF) expressed in the evening, and temporal recruits ELF3 and ELF4 to form the circadian evening complex, exerting transcriptional repression at night. Our study strongly suggests that circadian core components, including LUX, may participate in regulating plant‐microbe interactions by directly or indirectly modulating the synthesis, transport, and root exudation of specific metabolites. LUX may affect resistance to V. dahliae by mediating the nighttime rhizosphere microbiome of cotton, which requires further confirmation in future studies. More interestingly, we found that V. dahliae infection can accelerate the pace of cotton's circadian clock (Figure 2H). A similar phenotype, in which bacterial infection leads to a shortened circadian period, has also been reported in Arabidopsis (Zhang et al. 2013), and our results suggest that certain specific rhizosphere microorganisms might function as zeitgebers. To date, there has been little research evidence as well as speculation about how plants actively regulate rhizosphere microbial communities on time scales aligned with diurnal or circadian cycles (Hubbard et al. 2018; Newman et al. 2022; Zhao et al. 2021). Based on previous studies, we make two reasonable speculations. First, since the circadian clock gates many physiological and biochemical processes, the mechanisms of root secretion and water uptake are both likely to be gated by the circadian clock (McLaughlin et al. 2023; Staley et al. 2017; Takase et al. 2011). Second, both diurnal rhythms of plant rhizosphere exudates and root water use efficiency cause daily fluctuations in carbon sources or other nutrients, water, and even signalling molecules in the root zone, all of which will ultimately affect the structural and functional diversity of microbial communities in the rhizosphere. In this study, we found that compared with mock, inoculation of V. dahliae V991 during the day or at night could significantly increase the abundance of Verticillium, which is also very reasonable. Interestingly, a significant decrease in the relative abundance of Fusarium and Penicillium, along with a significant increase in Mortierella, was observed in the rhizosphere microbial communities of V. dahliae V991 inoculated samples at ZT13 compared to the mock (Figure 2E). Mortierella species are generally non‐pathogenic to plants and appear to help promote plant development and suppress soil diseases (Wang et al. 2022). Fusarium is a widespread fungal plant pathogen and can infect numerous crops (Beccari et al. 2022). For example, our recent study shows that the circadian clock in Fusarium oxysporum regulates its pathogenicity toward tomato by controlling the zinc starvation response and secondary metabolism (Lu et al. 2025). Penicillium is somewhere in between the above two, including beneficial, neutral and harmful, and its impact on plants may be related to the specific species and the corresponding host plants (Srinivasan et al. 2020). Our results suggest that the cotton circadian clock gates the interaction between V. dahliae and other rhizosphere microbial communities. Since the relative abundance of Fusarium, Penicillium and Mortierella also changed in disrupted cotton clock samples at ZT13 compared to the normal clock samples (Figure 1H), we are more inclined to believe that alterations in the microbiome are due to active recruitment of plants mediated by the root circadian clock rather than passive results of V. dahliae inoculation. However, this speculation still needs further investigation. Recent studies have shown that biotic stressors such as nematode and pathogen can directly regulate the host's circadian clock through effectors (Fu et al. 2025; Wang et al. 2026). The results showed that inoculation with V. dahliae also accelerates the circadian rhythm of cotton (Figure 2H). This suggests that circadian clocks not only regulate stress pathways, but that stress pathways can also regulate circadian rhythms via feedback. However, subsequent work is required to determine whether this period shortening affects resistance or is merely a side effect. Using wild‐type with disrupted circadian clock, GhLUX knockdown and overexpressed cotton lines, we demonstrated that the maintenance of normal circadian clock function is essential for cotton Verticillium wilt resistance. The evening complex component GhLUX gates defence response to V. dahliae (Figures 3A–E and 4A–H), which belongs to an output pathway of the circadian clock. At the same time, we identified cotton GhLUX05 and GhLUX07 copies that could directly respond to V. dahliae invasion (Figure 3F–H), suggesting that the input pathway of V. dahliae infection to the circadian clock is also mediated by GhLUX. The simultaneous involvement of GhLUX in both input and output pathways suggests that GhLUX acts as a key node between the Verticillium wilt signalling pathway and the circadian clock regulatory network. When we analysed the transcriptomic data of GhLUX regulating cotton resistance, we found that almost all JAZ family members in the overexpressing GhLUX lines were significantly down‐regulated after V. dahliae infection (Figure 5B,C). Subsequently, we combined tobacco leaf transient transfection, yeast one‐hybrid and ChIP methods to demonstrate that GhLUX directly inhibits the expression of cotton JAZs at the transcriptional level (Figure 5E–H). This is an unexpected but very reasonable result, because studies have shown that the JA pathway is involved in cotton's resistance to V. dahliae (He et al. 2018; Ji et al. 2023; Li et al. 2014; Liu et al. 2023). A major challenge in the breeding of cotton and other crops is the integration of desirable traits. Due to the widespread trade‐off between growth and defence, improved disease resistance is often accompanied by a decline in yield or fibre quality. However, we investigated that cotton plants overexpressing LUX do not have any negative effect on cotton fibre yield and quality, and even enhance them with increasing the boll weight. Knockout or overexpression of core clock genes is often accompanied by photoperiod‐dependent flowering phenotypes. Therefore, in breeding, the strategy of overexpressing circadian genes throughout the entire plant may affect cultivation regionality. Previous studies suggest that overexpressing circadian genes using tissue‐specific promoters in roots, mesophyll cells, or leaf epidermal cells, which do not significantly affect flowering time (Endo et al. 2014), which may be a good option for applying LUX in cotton breeding in the future. This implies that we have good prospects for application in cotton agricultural production. Perhaps we can achieve the breeding goals of increasing V. dahliae resistance as well as improving cotton yield and quality at the same time by overexpressing GhLUX specifically in whole plants or specific tissues and organs. Materials and Methods Plant Materials and Growth Conditions Cotton ( Gossypium hirsutum L.) materials used in this study were cultivars ‘Texas Marker‐1’ (TM‐1), a genetic standard for G. hirsutum and cultivars ‘zhongzhimian 2’ (ZZM2), also the resistant cultivars in G. hirsutum . Cultivars ‘Baimian‐1’ (BM‐1) were used as background to create GhLUXA07‐overexpression transgenic lines using the vector of MWV066 modified with aminoglycoside 3′‐adenylyltransferase (AADA) selection marker and 3×Flag target, and T3 generation seedlings were used for Verticillium resistance verification. Cotton plants were cultivated with sterilized soil (a mix of peat and sawdust) in growth chamber at 75% relative humidity, 23°C, and a 12‐h light/12‐h dark (12L/12D) photoperiod, light intensity is approximately 100 μmol m−2 s−1. DNA Extraction, PCR Amplification and Bioinformatic Analysis The soil samples were collected from the field of Jinming Campus of Henan University, Kaifeng City, Henan Province, China (114°19′ E, 34°49′ N). The average annual temperature was 15.9°C, and the average precipitation was 983.0 mm. The soil was screened with a 2 mm sieve, and the nutrient soil (PINDSTRUP) was mixed in a 2:1 ratio, then 150 g mixed soil was packed in a 7 × 7 × 8 cm box for cotton cultivation. The cotton seeds were soaked in water for 12 h, 30°C, and then transferred to a humid environment for further germination for 12 h. When the radicle length of the seeds reached about 1 cm, the seeds were planted in a box with mixed soil, and cultivate for 30 days. The DNA was extracted from rhizosphere and bulk soils (0.5 g soil samples) using the TGuide S96 Magnetic Soil/Stool DNA Kit (TIANGEN, DP812) according to the manufacturer's instructions. The quantity and quality of DNA were determined using a BioTek Synergy HTX before PCR. The ITS1 and ITS2 region of fungal were amplified through primer pairs ITS1F (5′‐CTTGGTCATTTAGAGGAAGTAA‐3′) and ITS2 (5′‐GCTGCGTTCTTCATCGATGC‐3′) and ITS2F (5′‐GCATCGATGAAGAACGCAGC‐3′) and ITS2R (5′‐TCCTCCGCTTATTGATATGC‐3′), respectively. Subsequently, the library was inspected by Qsep‐400 method, and sequencing was performed by Biomarker Technologies Co. Ltd. (Beijing, China, Illumina NovaSeq 6000). According to the quality of single nucleotide, raw data were primarily filtered by Trimmomatic (version 0.33). Identification and removal of the primer sequence were processed by Cutadapt (version 1.9.1). PE reads obtained from previous steps were assembled by USEARCH (version 10) and followed by chimera removal using UCHIME (version 8.1). The high‐quality reads generated from the above steps were used in the following analysis. OTU Clustering: USEARCH (version 10.0) was applied to cluster sequences with similarity over 97% (default). Conservative threshold for OTU filtration is 0.005%. ASV Analysis: DADA2 method in QIIME2 (version 2020.06) was applied to de‐noise sequences, generating ASVs. Conservative threshold for ASV filtration is 0.005%. Using SILVA (Release138, http://www.arb‐silva.de) as reference database, the Naive Bayes classifier was used to annotate the feature sequences, and the species classification information corresponding to each feature, and then the sample community composition could be counted at various levels (phylum, class, order, family, genus, species). Alpha (Chao1, Shannon and Simpson index) and Beta Diversity (PCoA) analyses were conducted using QIIME2 (https://qiime2.org/). Permutational multivariate analysis of variance (PERMANOVA) was employed to evaluate significant differences between different groups. Cultivation of Verticillium dahliae and Cotton Root Inoculation The highly aggressive V. dahliae strain ‘V991’ (and ‘V991‐GFP’, kindly provided by Cotton Institute of the Chinese Academy of Agriculture Sciences) was stored at −80°C as conidial suspensions preserved in 50% glycerol. To obtain V. dahliae plates, 1 μL conidial suspensions were incubated on every potato‐dextrose agar (Solarbio, Cat#P8931) plate. For plate assays, the plates were cultured in 25°C incubators with different light cycles: constant darkness, constant light, or 12L/12D cycles. To get a spore suspension, cultures on the plate were washed with 1 mL sterile water; the liquid was collected and inoculated into 50 mL potato dextrose broth (Solarbio, Cat#P9240). Then the broth was placed in a shaker at 120 rpm and 25°C for about 5 days to obtain a liquid culture. To infect cotton roots, the Czapek broth in which V. dahliae was cultured was filtered through gauze to remove hyphae and solid impurities, leaving the spore suspension liquid. The spore suspension liquid needed to be diluted to 1 × 107 conidia/mL with sterile distilled water for further infection experiment. When cotton plants were at the state of two true leaves, the seedlings were carefully removed from the soil; then the roots were dipped in the spore suspension liquid for 2 min to inoculate V. dahliae, and finally the seedlings were replanted into new soil (Chang et al. 2020). Root inoculation treatment of cotton plants in ZT13 was operated under a green light source (applies to all experiments conducted in dark conditions in this study). Extraction of Metabolites From Cotton Root and Fungi Inhibition Assay Cotton seedlings were grown to two true leaves states under conditions of 23°C, 75% relative humidity, and photoperiodic of 12‐h light/12‐h dark, and samples of cotton roots were collected at ZT1 and ZT13 (operated under a green light source), respectively, then each sample was fully ground with liquid nitrogen and 200 mg powder was taken in a 1.5 mL centrifuge tube for total metabolite extraction. Add 750 μL extract A (ddH2O) and ultrasonic crushing at 0°C for 30 min; all supernatant was collected by centrifugation at 15 000 rpm, 4°C, for 10 min. Then add 750 μL extract B (methanol: acetonitrile = 1:1) and ultrasonic crushing at 0°C for another 30 min, centrifugation at 15 000 rpm, 4°C, for 10 min. Combined the two supernatants, dried with nitrogen, and dissolved with 100 μL 50% methanol solution. Dropped the conidia suspension of V. dahliae V991 at the center of the PDA plate and then cultured it under 25°C for about 1 week. 50 μL metabolites solution dropped on a filter paper which was placed on the edge of the V991 cultures, equal volume of 50% methanol solution as mock. After that, the PDA plate was continued to be cultured with the previous conditions, and observed the appearance of the fungus inhibition zone around the metabolites and mock. Circadian Rhythm of Cotton Cotyledon Movement Leaf movement was measured as described in soybean previously with minor modifications (Wang, Yuan, et al. 2020). Briefly, cotton (cultivar ‘BM‐1’) seedlings were grown under 75% relative humidity, 23°C, and a 12L/12D cycle for about 14 days. Poured 30 mL 1 × 107 conidia/mL V. dahliae V991 conidia suspension or mock (water) into each pot (2 plants/pot) at ZT1 or ZT13. Cotton plants then were released into free‐running conditions (constant light). Images of the tip of the unifoliolate leaf and the first true leaf position were recorded 30 min apart for 4 days using a Hikvision digital camera (DS‐2CD3T45F (D)‐I5S). The leaf position was assessed from the time series images using Image J and quantified with BRASS software. Circadian in Phase or Out of Phase Inoculation Cotton (cultivar ‘BM‐1’) seedlings were grown under 75% relative humidity, 23°C, and a 12‐h light/12‐h dark photoperiod for about 21 days. The 200 mL potato dextrose broth (PDB) was divided into four 100 mL conical bottles, autoclaved at 121°C for 20 min, each of which was inoculated with 1 mL 1 × 107 conidia/mL V. dahliae V991 conidia suspension. All the conical bottles were placed in a shaker with 12L/12D light cycles, two of which are covered (or removed) with tin foil to create 12D/12L photoperiodic culture conditions, 120 rpm, 25°C, for 5 days. The cultures were prepared into 1 × 107 conidia/mL conidia suspension and infected cotton roots at ZT1/ZT13. Biomass Detection of Verticillium dahliae V991 in Cotton Roots Specific amplified primers were designed (primer 5.0) according to the ITS (Internal Transcribed Space) region of the V. dahliae V991 ribosome gene (Table S9). Fresh cotton tissue inoculated with V. dahliae V991 was sampled, all tissues in each sample were fully ground with liquid nitrogen, then 100 mg powder was taken for total genomic DNA extraction, using a FastPure Plant DNA Isolation Mini Kit (Vazyme, Nanjing, China). Using GhUBQ7 gene as an internal reference gene, we measured the relative expression of V991 DNA by qPCR, thus reflecting the relative biomass of V991 in Cotton (Luo et al. 2021). Virus‐Mediated Gene Silencing (VIGS) Assay The Virus‐mediated Gene Silencing (VIGS) assay was performed as previous described. Briefly, pTRV2 vector inserted with target gene's specific fragment or GhCLA (positive control) gene, pTRV2 (empty) and pTRV1 vectors were transformed into Agrobacterium GV3101 strain, respectively. When the OD600 value of GV3101 culture reached 0.7–0.8, collect the bacteria and resuspended with VIGS infection solution (10 mM MES, 10 mM MgCl2, 200 μM acetosyringone), then co‐cultured at 25°C, 150 rpm, in the dark for 2–3 h. The cotton seedlings were cultured at 23°C, 75% humidity, 12 h light and 12 h dark light cycle until the cotyledon was flat. The infection solution of TRV:target, TRV:GhCLA and TRV:00 vectors were mixed with pTRV1 solution in equal volume, respectively, and then injected the mixed solution into the young cotyledon of cotton through a syringe (Gao et al. 2013). Cotton seedlings were continue cultivated under the former conditions until TRV:GhCLA plants showed albino phenotype, the expression level of target gene was detected by RT‐PCR for interference efficiency monitoring. All primer sequences are listed in Table S12. RNA‐Seq Analysis Seedlings of ‘BM‐1’ and GhLUX‐OE4 were grown under 12L/12D cycles at 23°C for 21 days, and then were inoculated with 1 × 107 conidia/mL V. dahliae V991 conidia suspension or mock (water) at ZT1/ZT13. Then roots were collected at 3 hpi (hours post inoculated) and were frozen immediately in liquid nitrogen for total RNA extraction. RNA‐seq libraries were prepared using the Illumina Directional mRNA‐Seq Library Preparation Kit and sequenced on an Illumina NovaSeq 6000 platform (Biomarker, Beijing, China), resulting in double‐end 150‐bp reads in each sample. Sequenced reads were aligned to the Gossypium hirsutum CRI_v1 genome (https://cottonfgd.net/) and analysed using BMKCloud (www.biocloud.net). DEGs were identified by a significance analysis when Fold Change ≥ 2 and FDR (False Discovery Rate) < 0.01. Gene ontology (GO) term enrichment analysis for the DEGs and GhJAZs gene expression heatmap were performed using BMKCloud (www.biocloud.net). ChIP Assays ChIP assays were performed as previously described (Saleh et al. 2008). Seedlings of 35S:GhLUX‐OE‐Flag were grown under 12L/12D cycles at 23°C for 21 days at ZT13. The materials were harvested and crosslinked with 1% formaldehyde (Sigma‐Aldrich, F8775). Protein G‐Agarose beads (Roche, Cat# 11243233001) and an anti‐Flag antibody (Sigma‐Aldrich, Cat# F1804) were used for ChIP analysis. Primers amplifying a fragment in GhCCA1 LBS (LUX binding site) or GhUBQ7 were used for the positive control or negative control, respectively. All primer sequences are listed in Table S11. Luciferase Reporter Assay The promoter sequences of GhJAZ2 (−841 to −2902 bp), GhJAZ11 (−58 to −3000 bp) and GhJAZ26 (−197 to −2856 bp) were cloned into pGreenII 0800‐LUC vector, respectively, serving as reporters. Full‐length GhLUXA07 CDS (coding sequences) were cloned into pGreenII 62‐SK vector, serving as the effector. The constructs were then transformed into Agrobacterium strain GV3101 to transform tobacco. Agrobacterium strain GV3101 cells with reporter and effector were diluted to OD600 = 0.8 and co‐infiltrated into leaves of N. benthamiana. The infection solution with reporter and effector was mixed in equal volume. After 48 h of culture in the dark, fluorescein substrate (0.2 mg mL−1) was sprayed on the surface of the leaves, and the fluorescein signal in the leaves was detected by Fusion‐FX chemiluminescence instrument (Vilber, Paris, France). All primer sequences for vector construction are listed in Table S13. Y1H Assay The Matchmaker Gold Yeast One‐Hybrid Library Screening System was used for detecting interaction between GhLUXA07 and GhJAZ2, GhJAZ11 or GhJAZ26 promoters. The promoter sequences of GhJAZ2 (−841 to −2902 bp), GhJAZ11 (−58 to −3000 bp) and GhJAZ26 (−197 to −2856 bp) were cloned into pAbAi vector to obtain pBait‐AbAi constructs, respectively. The pBait‐AbAi recombinants were linearized with BbsI endonuclease, and integrated into the Y1HGold yeast genome by homologous recombination method. Specific Y1HGold reporters containing GhJAZs promoters serve as reporters. Full‐length GhLUXA07 CDS (coding sequences) were cloned into pGADT7 vector, and then transformed into Y1HGold strain as the effector. The recombinant yeast was inoculated on SD/‐Ura or SD/‐Leu plates containing aureobasidin A (AbA) at different concentrations and cultured in incubator at 30°C. All primer sequences are listed in Table S13. Accession Numbers Sequence data for the genes described in this article can be found in CottonFGD (https://cottonfgd.net/) under the following accession numbers: GhCCA1 (Gh_A11G109400), GhPRR9 (Gh_A11G108700), GhPRR5 (Gh_A11G015800), GhTOC1 (Gh_A03G084800), GhELFA3 (Gh_A09G097800), GhELFD3 (Gh_D09G093100), GhLUXA03 (Gh_A03G059200), GhLUXD03 (Gh_D03G137700), GhLUXA05 (Gh_A05G080400), GhLUXD05 (Gh_D05G091600), GhLUXA07 (Gh_A07G109300), GhLUXD07 (Gh_D07G106900), GhPR4 (Gh_D11G284200), GhJAZ2 (Gh_D06G087500), GhJAZ11 (Gh_A07G018400), GhJAZ26 (Gh_D08G272600) and GhUBQ7 (Gh_Contig00740G000200). Author Contributions Conceptualization: Q.X., Y.C., X.X. and X.L. Methodology: P.W., M.L., C.Z., Q.L., L.G. and S.G. Visualization: P.W., C.Z. and L.Y. Software: P.W., M.L., C.Z. and Q.L. Formal analysis: Q.X., Y.C., X.X., X.L., P.W., M.L., C.Z., Q.L., Y.L., J.W., F.Y. and C.C. Data curation: Q.X., Y.C., X.X., P.W. and M.L. Validation: P.W., M.L., C.Z. and S.G. Funding acquisition: Y.C., X.X., Q.X., X.L. and X.Z. Project administration: X.X., Q.X. and Y.C. Resources: Y.C., P.W., Q.X. and X.X. Supervision: X.X., Q.X., Y.C. and X.Z. Writing – original draft: Q.X., P.W., M.L., X.X. and Y.C. Writing – review and editing: Q.X., Y.C., X.X., C.R.M., J.L., P.W., M.L. and L.G. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1: qRT‐PCR assay of the transcriptional patterns of the other core circadian genes under NC and DC conditions. The data presented above represents the results of one representative biological replicate. The experiment was performed twice with consistent results. Figure S2: Analysis of the circadian rhythm of cotton cotyledon movement under NC and DC conditions. The data presented above represents the results of one representative biological replicate (N = 5). The experiment was performed twice with consistent results. White or grey boxes represent subjective day or night. p‐values were calculated using pairwise Student's t‐test (*p < 0.05; **p < 0.01). Figure S3: Differential analysis of rhizosphere microorganisms in samples from non‐planted soil controls. (A–C) Rhizosphere microbiome α‐diversity (Chao1, Shannon and Simpson index) in the non‐planted soil controls (n = 8 soil samples). The box plots indicate the median (center line), the 25th and 75th percentiles (box) and the range of non‐outlier values (whiskers). Letters indicate significant difference among the samples as determined by one‐way ANOVA analysis using the Tukey's honestly significant difference (HSD) test (p < 0.05). (D) Principal coordinates analysis (PCoA) based on Bray–Curtis distances displaying a significant difference in the rhizosphere microbiomes (PERMANOVA). Figure S4: Differences in the relative abundances of fungal phyla in the top 10 genus categories between ZT1 and ZT13. The materials and methods are shown in Figure 1G,H. Data are means ± SE of eight biological replicates, p‐values were calculated using pairwise Student's t‐test (*p < 0.05; **p < 0.01; ***p < 0.001). Figure S5: Differences in the relative abundances of Verticillium in the genus categories between disrupted or normal circadian clock in ZT1 and ZT13. Data are means ± SE of eight biological replicates, p‐values were calculated using pairwise Student's t‐test (***p < 0.001). Figure S6: Transcript level of GhPR1 and GhPR4 after Verticillium dahliae infection at ZT1 or ZT13 in the cotton roots. Roots were harvested at 0, 3, 6 and 9 h post inoculated (hpi) with V. dahliae V991 conidia suspension or water (Mock), then for RNA isolation. Data are shown as mean ± SD of three technical replicates from one of three independent biological experiments, all experiments yield congruent results. Figure S7: Pre‐treatment phenotypes in cotton seedlings. (A) Pre‐treatment phenotypes of cotton seedlings with normal and disrupted clock. (B) Pre‐treatment phenotypes of control and GhLUX‐silenced cotton seedlings. Figure S8: Transcriptional abundance of GhLUX03, GhLUX05 and GhLUX07 genes in different tissues after inoculation. BM‐1 cotton seedlings were grown for 21‐days under LD conditions. Root, stem and leaf samples were collected at two time points (ZT1 and ZT13) throughout the day, and total RNA was extracted. Using GhUBQ7 as an internal control and the root data at time point ZT1 as 1, the data were normalized. The data represent one result from three biological replicates, all experiments yield congruent results. Figure S9: Identification of evening‐phased clock genes in cotton. (A–E) Circadian rhythmic enrichment of evening‐phased clock gene transcripts in free‐running conditions. Cotton seedlings (Cultivars, ‘TM‐1’) were subjected to 12L/12D cycles, 23°C, for 3 weeks before transferred into constant light (LL) conditions. After three days in LL, leaves were harvested every 3 h for RNA isolation. Transcripts levels of GhLUX03 (A), GhLUX05 (B) GhLUX07 (C), GhTOC1 (D) and GhELF3 (E) were calculated by qRT‐PCR assay. GhUBQ7 was used as a normalization control. Data are shown as means ± SD of three technical replicates from one of three independent biological experiments. White or grey boxes represent subjective day or subjective night, respectively. (F) The phylogenetic tree (left), motif of protein (center) and gene structure (right) of GhLUXs, AtLUX and AtBOA. The capital letters A and D stand for the sub‐genomes originated from two diploid progenitor species Gossypium arboreum (A‐genome donor) and G. raimondii (D‐genome donor), respectively. The numbers next to the capital A/D letters stand for chromosome numbers. Figure S10: Sequence alignment of GhLUX03, GhLUX05 and GhLUX07 genes. The dark blue region means 100% of identity. The pink, blue and yellow regions mean over 75%, 50% and 33% the identities, respectively. Figure S11: Multiple sequence alignment among GhLUXA05, GhLUXD05, GhLUXA07, GhLUXD07 and GhLUXs VIGS fragment. GhLUX05 VIGS fragment target to GhLUXA05 and GhLUXD05; similarly, GhLUX07 VIGS fragment target to GhLUXA07 and GhLUXD07. Multiple sequence alignment was performed by DNAMAN software, and the sequence data used to do the comparisons are available at CottonFGD (https://cottonfgd.net/). The dark region means 100% of identity. The pink, blue and yellow regions mean over 75%, 50% and 33% the identities, respectively. Figure S12: qRT‐PCR analysis of tissue‐specific expression of GhELF3 and GhTOC1. BM‐1 cotton seedlings were grown for 21‐days under LD conditions. Root, stem and leaf samples were collected at two time points (ZT1 and ZT13) throughout the day, and total RNA was extracted. Using GhUBQ7 as an internal control and the root data at time point ZT1 as 1, the data were normalized. The data represent one result from three biological replicates, and the trends across the three replicates were consistent. Figure S13: Silencing the evening‐phased clock gene GhELF3 did not alter cotton resistance to Verticillium dahliae. (A) Multiple sequence alignment among GhELF3A09, GhELF3D09 and GhELF3 VIGS fragment. The dark blue region means 100% of identity. The blue and yellow regions mean over 50% and 33% the identities, respectively. (B) Disease symptoms of GhELF3‐silenced seedlings at 16 dpi with V. dahliae. TRV:00: negative control, no targeted gene fragment; TRV:GhELF3: targeted to GhELF3A09 and GhELF3D09. The albino phenotypes of TRV:GhCLA (positive control, targeted to GhCLA) seedlings showed the VIGS experiment was effective. (C) Expression level of GhELF3 in roots of TRV:GhELF3 and TRV:00 cotton plants. GhUBQ7 was used as a normalization control. Data are means ± SD of three biological replicates (n = 3; 4 plants/biological replicate). The expression value of TRV:00 was normalized as 1 (**p < 0.01, Student's t‐test). (D–F) Disease index (D), stem anatomy (E) and fungal renewal cultivation levels (F) in TRV:00 and TRV:GhELF3 cotton plants shown in (A). Bar = 1 mm (stem anatomy); uniform segments of cotton stems were selected from the same internodal region and manually sectioned into cross‐sections using a razor blade. Data are means ± SD of three biological replicates (n = 3; ≥ 30 plants/biological replicate; ns, no significance difference; Student's t‐test). dpi, days post inoculation. Figure S14: Silencing the evening‐phased clock gene GhTOC1 did not alter cotton resistance to Verticillium dahliae. (A) Multiple sequence alignment among GhTOC1A03, GhTOC1D03 and GhTOC1 VIGS fragment. The dark blue region means 100% of identity. The blue and yellow regions mean over 50% and 33% the identities, respectively. (B) Disease symptoms of GhTOC1‐silenced seedlings at 16 dpi with V. dahliae. TRV:00: negative control, no targeted gene fragment; TRV:GhTOC1: targeted to GhTOC1A03 and GhTOC1D03. The albino phenotypes of TRV:GhCLA (positive control, targeted to GhCLA) seedlings showed the VIGS experiment was effective. (C) Expression level of GhTOC1 in roots of TRV:GhTOC1 and TRV:00 cotton plants. GhUBQ7 was used as a normalization control. Data are means ± SD of three biological replicates (n = 3; 4 plants/biological replicate). The expression value of TRV:00 was normalized as 1 (**p < 0.01, Student's t‐test). (D–F) Disease index (D), stem anatomy (E) and fungal renewal cultivation levels (F) in TRV:00 and TRV:GhTOC1 cotton plants shown in (A, left and center). Bar = 1 mm (stem anatomy); Uniform segments of cotton stems were selected from the same internodal region and manually sectioned into cross‐sections using a razor blade. data are means ± SD of three biological replicates (n = 3; ≥ 30 plants/biological replicate; ns, no significance difference; Student's t‐test). Figure S15: Expression level of GhLUX03 (A), GhLUX05 (B) and GhLUX07 (C) in roots of TRV:GhLUX05 and TRV:GhLUX07 cotton plants. GhUBQ7 used as a normalization control. Data are shown as means ± SD of three technical replicates from one of three independent biological experiments. Figure S16: Stem anatomy and fungal renewal cultivation levels in cotton plants shown in Figure 4 and Figure S10A. (A, B) Stem anatomy (A) and fungal renewal cultivation levels (B) in TRV:00 and TRV:GhLUX07 cotton plants shown in Figure 4A. Bars = 1 mm (stem anatomy); Uniform segments of cotton stems were selected from the same internodal region and manually sectioned into longitudinal sections using a razor blade. (C, D) Stem anatomy (C) and fungal renewal cultivation levels (D) in BM‐1 and GhLUX‐OE cotton plants shown in and Figure S10A. Bars = 1 mm (stem anatomy); 9‐cm plates were used in fungal renewal cultivation experiment. Figure S17: Identification of GhLUX overexpressing transgenic lines of cotton. (A) The construction diagram of p35S::GhLUXA07‐Flag (GhLUX‐OE). AADA: aminoglycoside 3′‐adenylyltransferase; 35Spro: 35S promotor; F1/R1 and F2/R2: primers were used to identify the successfully transformed plants by PCR. The black arrows indicate the position and direction of the primers. Primers for construction are listed in Table S13. (B) Direct‐PCR verification of p35S::GhLUXA07‐Flag using genomic DNA from BM‐1 and GhLUX overexpression plants. BM‐1, the background varieties for genetic transformation. #1–#4: four GhLUX overexpression transgenic lines. F1/R1 or F2/R2 primers were used for PCR in left or right, respectively. M: Marker of 2k plus. All primer sequences are listed in Table S10. (C) Western blotting was used to detect p35S::GhLUXA07‐Flag in BM‐1 and transgenic plants in (B). Primary antibody of anti‐Flag mouse monoclonal antibody and the secondary antibody of HRP‐conjugated goat anti‐mouse IgG were used. Coomassie Brilliant Blue (CBB) was used as a total‐protein loading control. Figure S18: Overexpression of GhLUXA07 enhanced Verticillium dahliae resistance in cotton. (A, B) GhLUX‐OE3 and GhLUX‐OE4 seedlings at 15 dpi with V. dahliae. Disease index (B) in BM‐1 and GhLUX‐OE plants shown in (A). Data are the mean ± SD of three biological replicates. The statistical analysis method is one‐way ANOVA. Different letters represent p < 0.05. (C) Spore distribution of V. dahliae on lateral root surface. V. dahliae V991‐GFP strain (kindly provided by Cotton Institute of the Chinese Academy of Agriculture Sciences) was used to infection the cotton roots in ZT13, then lateral roots were harvested at 12 hpi (light on) and imaged by a laser confocal microscope. Red fluorescence shows the cell wall after binding to Propidium Iodide (PI), and green fluorescence shows the spores of V. dahliae V991‐GFP. Spores attached on root surface are indicated by white arrow. An enlarged view of area in white box is placed on the right. Bars = 100 μm. (D) Relative biomass of V. dahliae V991 in the roots of BM‐1 and GhLUX‐OE plants with V. dahliae inoculation at ZT13. The expression value of BM‐1 was normalized as 1. Data are shown as means ± SD of three technical replicates from one of three independent biological experiments. Figure S19: Overexpression of GhLUXA07 enhanced Verticillium dahliae resistance in Arabidopsis. (A) Disease symptoms of Col‐0; lux‐4 mutant and p35S::GhLUXA07/lx‐4 Compensation line; p35S::GhLUXA07/Col‐0 over expression line. Seedlings were subjected to 23°C, 12L/12D light cycles for 3 weeks in growth chamber, then inoculated with V. dahliae (isolate V991) by pouring conidia suspension around the roots. The main stems were removed at 14 dpi, then photos were taken. (B) Expression of GhLUXA07 in different lines shown in (A). The reference gene GhUBQ7 was used as a normalization control. Data are means ± SD of three biological replicates (n = 3; 4 plants replicate). The expression value of Col‐0 was normalized as 1. (C) Disease index of different lines shown in (A). Data are means ± SD of three biological replicates (n = 3; ≥ 60 plants/biological replicate). Letters indicate significant difference among the samples as determined by a one‐way ANOVA analysis using the Tukey's honestly significant difference (HSD) test (p < 0.05). Figure S20: Analysis of the circadian rhythm of cotyledon movement in GhLUX overexpressing plant lines. The data presented above represents the results of one representative biological replicate (N = 5). White or grey boxes represent subjective day or night. Figure S21: The effects of GhLUX overexpression on circadian rhythm and clock genes expression. The diel enrichment of clock gene transcripts in BM‐1 and GhLUX overexpression lines. Cotton seedlings were subjected to 12L/12D cycles, 23°C, for 3 weeks, then roots (4 plants mixed for one sample) were harvested at 4‐h intervals RNA isolation. GhLUXA07 (A), GhCCA1 (B), GhPRR9 (C), GhPRR7 (D), GhPRR5 (E), GhTOC1 (F), GhELF4 (G), GhLUX03 (H) and GhLUX05 (I) primers were used for qRT‐PCR assay. GhUBQ7 was set as a normalization control. Data are shown as means ± SD of three technical replicates from one of three independent biological experiments; all experiments yield congruent results. White or black boxes represent day or night, respectively. Figure S22: Analysis of JA signal downstream responsive gene expression levels in GhLUX VIGS lines and overexpression lines. Transcript level of GhPDF1.2, GhVSP, GhVSP2‐2, GhTHI2‐1, GhTHI2‐2 and GhTHI2‐4 in GhLUX knockdown plants (A) and GhLUX Overexpression plants (B). The reference gene GhUBQ7 was used as a normalization control. Data are shown as means ± SD of three technical replicates from one of three independent biological experiments. Table S1: Quality assessment of sequencing data in Figures 1 and 2. Table S2: Analysis of differences in the relative abundances of fungal phyla at the genus level as depicted in Figure S1. Table S3: Analysis of differences in the relative abundances of fungal phyla at the genus level as depicted in Figure 2E. Table S4: Parameters of leaves rhythmic movement in Figure 2H. Table S5: Comparison of yield traits and yield from growing field between GhLUX‐OE plants and BM‐1. Table S6: Yield of BM‐1 and GhLUX‐OE4 plants in the disease nursery. Table S7: DEGs between BM‐1 and GhLUX‐OX plants after Verticillium dahliae infection. Table S8: GO analysis of DEGs between BM‐1 and GhLUX‐OX plants after Verticillium dahliae infection in Figure 5A. Table S9: Primers for Verticillium dahliae V991 biomass quantification. Table S10: Primers for direct‐PCR verification of transgenic lines. Table S11: Primers for ChIP‐qPCR. Table S12: Primers for qRT‐PCR. Table S13: Primers for construction.