In order to investigate the impacts of humidity on chlorophyll degradation during flue-curing, three types of tobacco with distinct chlorophyll contents resulting from various nitrogen fertilization regimes were selected (Table 1). These included control tobacco, Laohan tobacco, and Fanqing tobacco. Leaves from each type were subjected to two different humidity (RH) conditions during curing (Treatment 1: 79% RH; Treatment 2: 93% RH), while temperature was maintained constant (Supplementary Figure S1).

As shown in Figure 1A, all tobacco leaves exhibited a typical green coloration at the onset of curing (0 h), followed by progressive yellowing as curing processed. Compared with the control leaves, both Laohan and Fanqing tobacco leaves displayed a visible darker green color at 0 h, consistent with their higher initial chlorophyll contents. Quantitative analysis confirmed that both chlorophyll a and chlorophyll b contents were significantly higher in Laohan and Fanqing leaves than in the control group at the initial time point (Figure 1B). Specially, the chlorophyll a content in control leaves was approximately 4000 μg/g at 0 h, whereas values of 7380 μg/g and 8140 μg/g were observed in Laohan and Fanqing leaves, respectively (Figure 1B). Moreover, the content of chlorophyll b is approximately 45% and 49% of that measured in Laohan and Fanqing leaves, respectively.

Chlorophyll content declined continuously in all three tobacco types throughout the curing process, regardless of humidity treatment. However, the rate of chlorophyll loss differed markedly between the two humidity conditions. Under high humidity (Treatment 2, 93% RH), chlorophyll degradation occurred more rapidly than under moderate humidity (Treatment 1, 79% RH) (Figure 1B). In the control group, chlorophyll a and chlorophyll b contents decreased to approximately 40% of their initial levels after 18 h under Treatment 1, whereas they declined to about 17% under Treatment 2. Similarly, in Laohan tobacco, chlorophyll a content decreased from 7380 μg/g at 0 h to 3108 μg/g under Treatment 1 and to 2184 μg/g under Treatment 2 at 18 h. In Fanqing tobacco leaves, chlorophyll a content reached 3354 μg/g under Treatment 1 and 2672 μg/g under Treatment 2 at the same time point. These findings indicate that elevated relative humidity is closely associated with a faster decline in chlorophyll content during tobacco flue-curing, particularly in leaves with high chlorophyll levels.

To characterize transcriptome-wide responses of tobacco leaves to different humidity conditions during flue-curing, RNA sequencing was performed on samples collected throughout the curing process. A total of 63 RNA-seq libraries were generated, yielding approximately 62 million clean reads and 586.6 Gb of high-quality sequencing data (Supplementary Table S1). Samples at 0 h three tobacco types prior to curing and were therefore excluded from humidity-related. Samples collected at 0 h represented three tobacco types prior to curing and were therefore excluded from humidity-related transcriptomic comparisons. The remaining 54 samples were subjected to downstream analyses and grouped according to humidity treatment, with temperature maintained constant at 38 °C during curing.

Principal component analysis (PCA) revealed clear separation between samples exposed to moderate humidity (Treatment 1,79% RH) and high humidity (Treatment 2, 93% RH) at corresponding curing time points, indicating distinct global transcriptional profiles in response to humidity differences during curing (Figure 2A). Differential expression analysis further showed that the number of differentially expressed genes (DEGs) between the two humidity treatments increased progressively with curing duration, with 373, 703, and 911 DEGs at 6 h, 12 h, and 18 h, respectively (Figure 2B).

To gain insight into the functional categories associated with humidity-responsive transcriptional changes, gene ontology (GO) enrichment analysis was performed separately for up-regulated and down-regulated DEGs at each time point (Figure 2B). At 6 h, up-regulated DEGs were predominantly enriched in GO terms related to “response to stimulus”, “response to wounding”, and “defense response”. At 12 h, enrichment shifted toward terms including “chloroplast envelope”, “response to salt”, and “response to water”, while at 18 h, up-regulated DEGs were mainly associated with “response to heat”, “chloroplast organization”, and “response to temperature stimulus”.

In contrast, down - regulated DEGs at 6 h were primarily enriched in GO terms related to “protein folding”, “response to unfolded protein”, and “chaperone - mediated protein folding”. At 12 h, these genes were enriched in terms associated with chromatin structure, including “nucleosome”, “nucleosome assembly”, and “chromatin remodeling”. By 18 h, enrichment of down-regulated DEGs was mainly observed in hormone-related categories, such as “jasmonic acid hydrolase”, “regulation of jasmonic acid - mediated signaling pathway”. Overall, down - regulated DEGs were largely associated with protein folding, chromatin structure, and hormone-related processes, whereas up - regulated DEGs were primarily linked to stress responses and chloroplast-related functions during curing. Notably, enrichment of GO terms related to chloroplast structure and organization was first observed at 12 h in response to differences in humidity conditions.

To relate transcriptional changes to physiological outcomes, chlorophyll contents were compared between the two humidity treatments during curing. At both 12 and 18 h, the contents of chlorophyll a and chlorophyll b in tobacco leaves subjected to high humidity (Treatment 2, 93% RH) were significantly lower than those under moderate humidity (Treatment 1, 79% RH) (Figure 2C). For instance, chlorophyll a content under high humidity was approximately 62% and 51% of that under moderate humidity at 12 h and 18 h, respectively. These results indicate that high humidity during curing is associated with enhanced chlorophyll loss in tobacco leaves. Based on these transcriptomic and physiological observations, subsequent analysis focused on transcriptional features associated with chlorophyll degradation under humidity conditions.

Chlorophyll degradation is a stepwise biochemical process in which green chlorophyll molecules are converted into non-colored fluorescent catabolites (pFCC) through the coordinated action of multiple enzymes (Figure 3A). To examine transcriptional changes associated with chlorophyll degradation during tobacco flue-curing, we analyzed the expression profiles of 18 structural genes involved in the chlorophyll degradation pathway 54 tobacco samples, including control (CK), Fanqing (FQ), Laohan (LH) leaves under different humidity conditions (Figure 3B).

In order to investigate the impact of the single factor of different humidity conditions on the degradation of chlorophyll during the tobacco drying process, we selected the control group for detailed analysis. At the early curing stage (6 h), no significant differences in the expression of chlorophyll degradation-related structural genes were detected between humidity treatments. In contrast, at 12 h, coinciding with accelerated chlorophyll loss, a significant decrease in expression was observed for a PAO family gene (Nitab4.5_0003134g0120, hereafter referred to as PAO4). At 18 h, PAO4 expression did not differ significantly between treatment, whereas a CLH family (Nitab4.5_0004743g0040, hereafter referred to as CHL6) exhibited pronounced upregulation under high-humidity conditions (Figure 3C). These stage specific expression patterns were temporally consistent with the observed dynamics of chlorophyll degradation during curing.

To further characterize transcriptional features associated with PAO4 and CLH6 expression under different humidity conditions, correlation-based transcriptional association analyses were performed (Figure 3D). For PAO4, three transcription factors showed significant correlations with its expression across curing samples (Supplementary Table S1). Among them, a heat shock protein (HSP) family gene (Nitab4.5_0003546g0070) displayed the strongest negative correlation with PAO4 expression (R = -0.91, P = 0.01), followed by an ethylene-responsive factor (ERF) gene (R = -0.87, P = 0.02).

In the case of CLH6, 12 transcription factors were significantly correlated with its expression (Supplementary Table S2). Notably, nine of these belonged to the HSP gene family. The HSP gene Nitab4.5_0003546g0070 exhibited a strong negative correlation with CLH6 expression in response to humidity treatments (Figure 3D, Supplementary Table S2). Significant correlations were also observed for transcription factors from the MYB, ERF, and NYC families. Among these, the MYB gene Nitab4.5_0000068g0400 showed the highest positive correlation with the CLH6 expression (R = 0.95, P = 0.003) (Supplementary Table S2).

Collectively, these results indicate that PAO4 and CLH6 display distinct, stage-dependent expression patterns during tobacco curing and are associated with humidity-responsive transcriptional changes. The correlation-based networks constructed here highlight transcription factor families—including ERF, HSF, MYB, NAC, bZIP, MADS, and WRKY (Figure 3D)—that are transcriptionly associated with key chlorophyll degradation genes during flue-curing.

In addition to humidity, temperature is also a crucial environmental factor influencing chlorophyll degradation during tobacco flue-curing. PCA showed that transcriptome profiles of fresh tobacco leaves collected at 0 h were clearly separated from those of leaves subjected to curing at 38 °C for 6 h, 12 h, 18 h, indicating substantial transcriptional reprogramming during the curing process (Figure 4A).

Differentially expression analysis between fresh tobacco leaves and cured leaves identified a large number of DEGs, the number of which increased with curing duration (Supplementary Table S3). GO enrichment analysis revealed that these DEGs were significantly enriched in biological processes related to “pigment binding”, “chlorophyll binding”, “chloroplast membrane”, “response to abiotic stress”, as well as “drought - related response” (Figure 4B). These results indicate that high-temperature curing is associated with pronounced transcriptional changes in genes related to chloroplast function, pigment metabolism, and stress resistance.

Among the genes involved in the chlorophyll degradation pathway (Figure 3A), two structural genes — Stay-Green 6 (SGR) (Nitab4.5_0000140g0110) and Chlorophyllase 5 (CLH5, Nitab4.5_0000252g0240) — exhibited markedly increased expression levels in cured leaves compared with fresh leaves (Figure 4C). Notably, the expression of SGR6 increased by approximately 150-fold during curing relative to the 0 h samples, highlighting a strong association between SGR6 transcription and high-temperature curing conditions.

To further characterize transcriptional features associated with SGR6 and CLH5 expression during high-temperature curing, correlation analysis was conducted between the expression levels of these structural genes and transcription factors. A total of 123 TFs and 82 TFs showed significant correlations with the expression of SGR6 and CLH5, respectively (Supplementary Tables S3, S4). Subsequently, a VENN analysis revealed that 69 TFs were significantly correlated with both SGR6 and CLH5 expression, whereas 54 TFs and 14 TFs were uniquely associated with SGR6 or CLH5, respectively (Figure 4D).

Among the TFs associated with SGR6 and CLH5 expression, 53 displayed negative correlations and 16 showed positive correlations. These TFs belong to multiple families, including ERF, bZIP, HSF, MADS, MYB, NAC, and WRKY (Figure 4D). The gene IDs of TFs positively correlated with SGR6 and CLH5 expression are shown in Figure 4D, and include representative from the ERF, bZIP, MYB, and NAC families. Together, these correlation-based analyses identify candidate transcription factors whose expression patterns are associated with key chlorophyll degradation genes during high-temperature flue - curing.

Flue-cured tobacco cultivar Hongda (Nicotiana tabacum L.), a widely cultivated variety in China with stable agronomic performance and well-characterized curing characteristics, was grown at the Yanhe Experimental Station in Hongta District, Yuxi City, Yunnan Province, China (24.14°N, 102.29°E, 1,635 m altitude). The field experiment was conducted from May to September 2024 under local conventional management practice.

Three nitrogen fertilization regimes were applied in the field. The control treatment received a pure nitrogen application rate of 82.5 kg/ha, corresponding to the standard fertilization level recommended for flue-cured tobacco production. Two high-nitrogen treatments were established to simulate common farmer practices associated with excessive nitrogen input: the Laohan (LH) and Fanqin (FQ) treatments, both of which received a pure nitrogen application rate of 142.5 kg/ha (Table 1). Other agronomic practices were kept consistent across treatments.

At physiological maturity, fresh tobacco leaves were harvested from the middle position of plants to minimize variation associated with leaf developmental stage. The harvested leaves were subjected to postharvest curing treatments under controlled environmental conditions. Specifically, leaves were incubated at a constant temperature of 38 °C, corresponding to the initial yellowing stage commonly applied in commercial flue-curing, during which chlorophyll degradation is most active. Two relative humidity (RH) regimes were imposed: Treatment 1 (79% RH) and Treatment 2 (93% RH).

Postharvest treatments were carried out in a programmable environmental test chamber (Memmert CTC256, Kunming, Yunnan, China). Leaf samples were collected at 6 h intervals during curing. Upon sampling, tissues were immediately frozen in liquid nitrogen and stored at -80 °C until further physiological and transcriptomic analyses. The experiment was conducted with three independent biological replicates, each replicate consisting of 15 tobacco leaves collected from different plants within the same treatment.

The chemical composition and chlorophyll content were analyzed for samples collected. Chlorophyll content were assessed using the methods outlined in the Chinese tobacco industry standards “Tobacco and tobacco products—Determination of plastid pigments—High performance liquid chromatography method (YC/T382—2010)”. There are three biological replicates for chemical analysis.

Total RNA was extracted from tobacco leaf samples using the Trizol reagent kit (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol. RNA quality and integrity was assessed by 1.0% agarose gel electrophoresi, RNA purity was evaluated based on the absorbance ratio of A260/A280. Only RNA samples with high integrity and acceptable purity were used for subsequent analyses. After total RNA was extracted, eukaryotic mRNA was enriched by Oligo (dT) beads. RNA library construction and RNA-sequencing were performed by Novogene (Beijing, China).

Raw sequencing reads were subjected to quality control using fastp (v1.21) to remove adaptor sequences, low−quality reads, and reads containing excessive ambiguous bases. Clean reads were then aligned to the Nicotiana tabacum genome (https://solgenomics.net/ftp/genomes/Nicotiana_tabacum/edwards_et_al_2017/assembly/) using Hisat2 (2.0.5), with the reference genome index built by the same software. The mapped reads from each sample were assembled in a reference-guided manner using StringTie (1.3.3b). Gene-level read counts were generated FeatureCounts (v1.5.0-p3). Gene expression levels were normalized and expressed as fragments per kilobase of transcript per million mapped reads (FPKM), calculated based on gene length and the number of mapped reads.

The clean reads were aligned to the Nicotiana tabacum reference genome (solgenomics_nicotiana_tabacum_edwards2017_Ch) with Hisat2 (2.0.5). Gene-level read counts were generated using featureCounts, and only uniquely mapped reads were retained for downstream analyses. Differential expression analysis was performed using the DESeq2 R package, which is based on raw read counts. P-values were adjusted for multiple testing using the Benjamini-Hochberg false discovery rate (FDR) method. Genes with an absolute log₂fold change ≥ 1 and an FDR-adjusted P < 0.05 were defined as differentially expressed genes (DEGs).

For expression visualization and comparative analyses, transcript abundance was additionally estimated as FPKM using StringTie (v1.3.3b). Principal component analysis (PCA) was conducted based on normalized gene expression values to evaluate sample clustering and biological replicate consistency. Correlation analysis among biological replicates was performed using the Hmisc R package (https://hbiostat.org/R/Hmisc/). Gene Ontology (GO) enrichment analysis of DEGs was carried out using TBtools-II (Chen et al., 2023), with all expressed genes used as the background set. GO terms with an adjusted P < 0.05 were considered significantly enriched. All transcriptomic analyses were conducted with three independent biological replicates per treatment.

Statistical analyses between two experimental groups were performed using two-tailed Student’s t-test in SPSS 19.0 (SPSS Inc., Chicago, IL, USA). Differences with P < 0.05 were considered statistically significant. Correlation analysis, differential expression analysis and PCA were conducted using R software.