Physiological changes in tobacco leaves at different harvesting times
To record the appearance changes of leaves (middle position of seedling), we compared samples harvested at 35 d (T1), 42 d (T2), 49 (T3) and 56 d (T4) after tobacco topping. T1 samples showed a green color which gradually declined along with the yellow color increasing, especially in T3 and T4 samples (Fig. 2a). The visible white color appeared in the main branch vein from T1 samples to T4 samples (Fig. 2a). These appearance changes were more obvious in pseudo-color processing (Fig. 2b). It is noted that yellow-white mature spots appeared in T3 and T4 samples. To understand the physiological dynamics in these samples, pigment contents and photosynthesis parameters were measured. Contents of chlorophyll a and chlorophyll b increased slightly in the T2 samples followed by a decrease in the T3 and T4 samples compared to the T1 samples (Fig. 2c-d). For the content of carotenoid, only a significant reduction was observed in the T4 samples relative to other samples (Fig. 2e).
Compared with T1 samples, T2 samples had an increase in the stomatal conductance (Gs) and transpiration rate (Tr) by 18.65% and 11.06% (Fig. 2g-h). In contrast, Gs were sharply reduced by 78.32% and 75.51% in the T3 and T4 samples, respectively and Tr were largely reduced by 86.60% and 74.20% in the T3 and T4 samples, respectively (Fig. 2g-h). The Net photosynthetic (Pn) level showed a decline of 26.92% compared with T1 samples, and a further decline of 65.80% and 54.32% in the T3 and T4 samples, respectively (Fig. 2f). These results suggest that tobacco leaves might undergo a physiological mature transition from the T2 to the T3 stage.
Differences in the nutritional quality of tobacco samples
Yang et al. [19] reported that chemical composition indexes are tightly correlated with tobacco leaf nutritional quality. We thus determined the chemical composition indexes in our samples (Table 1). Compared with T1, the contents of nicotine, reducing sugar, total sugar, and total nitrogen in tobacco leaves increased in T2 and decreased in T3 and T4. Among them, reducing sugar content and total sugar content had a significant increase of 2.5-fold and 1-fold respectively, in comparison with T1. The contents of potassium and chlorine in tobacco leaves showed an increasing trend. The contents of potassium and chlorine in T4 samples were the highest, increasing by 126.67% and 86.05% compared with T1 respectively. The ratio of potassium to chlorine was positively correlated with tobacco leaf flammability. However, there was no significant difference in the potassium-chlorine ratio among all samples. Sugar-nicotine ratio is related to the aroma and taste of tobacco leaves. Among samples, the sugar-nicotine ratio reached the maximum value in the T2 sample. As a result of harvesting tobacco leaves too late, their quality will decline.
Table 1
Nutritional quality of cigar tobacco leaves after curing under different treatments
Groups | Nicotine (% DW) | Reducing sugar (% DW) | Total Sugar (% DW) | Total Nitrogen (% DW) | Potassium (% DW) | Chlorine (% DW) | Sugar-nicotine ratio | K/Cl |
T1 | 5.39 ± 0.22a | 0.32 ± 0.12bc | 0.78 ± 0.15bc | 3.50 ± 0.15ab | 1.65 ± 0.32b | 0.43 ± 0.03b | 0.14 ± 0.02b | 3.79 ± 0.49a |
T2 | 6.29 ± 0.17a | 1.12 ± 0.41a | 1.56 ± 0.38a | 3.94 ± 0.33a | 2.03 ± 0.17b | 0.47 ± 0.05b | 0.25 ± 0.06a | 4.35 ± 0.10a |
T3 | 6.93 ± 0.84a | 0.93 ± 0.57ab | 1.31 ± 0.57ab | 3.54 ± 0.27ab | 2.63 ± 1.25ab | 0.75 ± 0.04a | 0.19 ± 0.06ab | 3.48 ± 1.50a |
T4 | 3.20 ± 1.65b | 0.18 ± 0.04c | 0.42 ± 0.06c | 3.42 ± 0.26b | 3.74 ± 0.50a | 0.80 ± 0.17a | 0.15 ± 0.05b | 4.77 ± 0.42a |
Metabolomics analysis of tobacco leaves after different treatments
Principal component analysis
To investigate the changes in cigar tobacco metabolic processes under varying harvesting times, we analyzed the metabolites of the above samples using LC-MS. Eventually, 1078 valid peaks were detected in cation mode and 1075 valid peaks in anion mode for different treatment groups (Fig. 3c). PCA analysis of metabolic profiles showed T1 and T2 were clustered and T3 was close to T4 samples either in positive ion modes (POS) or negative ion modes (NEG) (Fig. 3a-b). According to this result, a distinct physiological metabolic process occurred between T2 and T3 samples. In T3 and T4 samples, this process slowly changed.
Screening of different metabolites and KEGG analysis of functional pathways in samples
Based on the metabolic data of T2 vs T1, T3 vs T2 and T4 vs T3, we set the VIP > 1.0 and P < 0.05 to identify differential metabolites (DMs). A total of 2016 significant DMs were obtained, among which the T2 vs T1 group screened out 180 down-DMs and 95 up-DMs; the T3 vs T2 group screened out 914 down-DMs and 562 up-DMs; and the T4 vs T3 group checked out 571 down-DMs and 168 up-DMs (Fig. 4a-c). Therefore, the T3 vs T2 group represents the key physiological changes since its DMs account for 73.21% of all DMs. In addition, there were more down-DMs than up-DMs in all groups.
We next focused on the T3 vs T2 group’s DMs to obtain the hierarchical rank of physiological functions involved in metabolites. According to annotation in the KEGG compound database, the T3 vs T2 group’s DMs were mainly clustered in three categories: compounds with biological roles, phytochemical compounds, and lipids, respectively (Fig. S1a-c). The most prominent three metabolites are amino acids、fatty acids、monosaccharides and phospholipids in the category of “compounds with biological roles” (Fig. S1a). The most prominent three metabolites are alkaloids derived from tryptophan and anthranilic acid、monoterpenoids (C10)、diterpenoids (C20)、flavonoids and alkaloids derived from ornithine in the category of “phytochemical compounds” (Fig. S1c). The top three metabolites are FA01 fatty acids and conjugates, PR01 isoprenoids, and PK12 flavonoids in the category of “lipid metabolites” (Fig.S1b). The metabolic pathways for these DMs were established using annotation of the KEGG pathway database. Most DMs involved three pathways, metabolism, environmental information processing and genetic information processing (Fig. 4d). The metabolism pathway had the most abundant secondary metabolic pathways such as biosynthesis of other secondary metabolites (67 DMs), amino acid metabolism (67 DMs) and metabolism of cofactors and vitamins (34 DMs).
KEGG pathway enrichment analysis of differential metabolites
We further screened the significantly enriched metabolic pathways from the pathways involved in T3vsT2 DMs using P-value ≤ 0.05 as the threshold. There were 27 enriched metabolic pathways (the top 20 were displayed in Fig. 5), of which 6 were highly significant enriched metabolic pathways with P < 0.001. Among them are tryptophan metabolism, secondary metabolism in plants, alanine, aspartate and glutamate metabolism, phenylpropanoid biosynthesis, zeatin biosynthesis, and aminoacyl-tRNA biosynthesis (Fig. 5). In the top 20 pathways, a large portion of DMs were enriched in amino acid synthesis and metabolism pathways including alanine, aspartate, lysine glutamate, tyrosine and phenylalanine. In addition, the enrichment ratios of betalain biosynthesis, cyanoamino acid metabolism, linoleic acid metabolism, and plant hormone signal transduction were over 0.15.
Key differential metabolites in samples
Combined analyses of KEGG enrichment in DMs of T3vsT2, T3vsT1, T4vsT1 and T4vsT2 found tryptophan metabolism pathway and zeatin biosynthesis pathway were the most significantly enriched pathways shared in all groups with P < 0.001 (Fig. 5 and Fig. S2). In zeatin biosynthesis pathway, 4-up DMs and 5-down DMs were found from T1 to T4 (Fig. 6a). In particular, isopentenyl adenosine was the most significant with upregulation of 128.17% in T3 and T4 samples compared with T1 and T2 samples (Fig. 6b). In tryptophan metabolic pathway, 5-up DMs and 11-down DMs were found (Fig. 6c). Among them, indolepyruvic acid was the most significantly differentially expressed, which was down-regulated by 34.29% in T4 compared with T1(Fig. 6d).
Effect of lower leaf's growth on metabolism of cutter leaf
In production practice, lower leaves are usually harvested one week earlier than T2 cutter leaves. However, whether removal time of lower leaf affects T2 cutter leave’s physiological maturity remains unclear. Here, we removed lower leaves from seedlings at 3 weeks, 2 weeks or 1 week before T2 harvest and named these T2 samples as F1, F2 and F3, respectively. Samples were then analyzed with LC-MS for metabolite changes.
PCA analysis
PCA analyses of POS modes and NEG modes clearly concentrated each sample point (Fig. 7a-b), indicating that metabolites within each sample are highly consistent. F1 and F2 had very similar distribution patterns with 207 (differential ion peaks) DIPs in POS modes and 220 DIPs in NEG modes (Fig. 7c), suggesting that their metabolites are not much different. F3 samples, on the other hand, differed fromF1 and F2 samples. F3 vs F2 samples had 390 DIPs in POS modes and 358 DIPs in NEG modes, and F3 vs F1 samples had 528 DIPs in POS modes and 486 DIPs in NEG modes (Fig. 7c). This result suggests that F3 samples had differential physiological processes from F1 and F2. It supports the concept that lower leaf growth can affect cutter leaves' physiological maturity.
Different metabolites analysis and functional pathways of KEGG compounds
A total of 1072 DMs were screened out in F2vsF1 and F3vsF2 groups, including 222 down-DMs and 205 up-DMs in F2vsF1 group(Fig. 8a), and 486 down-DMs and 262 up-DMs in F3vsF2 group (Fig. 8b).This result confirmed that harvesting time of lower leaf can affect the physiological process of cutter leaves. Based on annotation of the KEGG compound database, DMs from the F3vsF2 groups grouped into three categories: compounds with biological roles, phytochemical compounds, and lipids (Fig.S3), which is consistent with T1-T4 results (Fig.S1). In the category of “compounds with biological roles”, lipids were the largest compounds and the number of lipids was 10 (Fig.S3a). Ranked second were steroids followed by peptides and antibiotics, nucleic acids and organic acids, steroid hormones, carbohydrates and vitamins in order of numbers. Category of “lipids” included 37 fatty acyls, 21 sterol lipids, 19 prenol lipids, 14 polyketides, 6 glycerophospholipids and 2 sphingolipids (Fig. S3b). Category of “phytochemical compounds” included 19 terpenoids, 16 alkaloids, 14 phenylpropanoids, 9 flavonoids, 6 amino acid related compounds, 1 polyketide and 1 fatty acid (Fig. S3c). In the KEGG PATHWAY database, DMs of the F3 vs F2 groups were primarily involved in "metabolism, environmental information processing, and genetic information processing" (Fig. 8c). There was a high amount of metabolism, including amino acid metabolism, secondary metabolite biosynthesis, and lipid metabolism.
KEGG pathway enrichment of differential metabolites
Significantly enriched metabolic pathways were further screened in F3 vs F2 group using P-value ≤ 0.05. as the threshold. Arachidonic acid metabolism was only one enriched metabolic pathway with P < 0.001 (Fig. 9). In addition, amino acid metabolic pathways (e.g., tyrosine metabolism) as well as tricarboxylic acid cycle-related pathways (e.g., citrate cycle) were also significantly enriched. These results indicate that DMs of F3 vs F2 group mainly enriched in the amino acid metabolic pathway and lipid metabolic pathway.
Key differential metabolites in enriched pathways
We also found that arachidonic acid metabolic was also the highest enriched pathway in all F3vsF1 KEGG enriched pathways (Fig. S4). Subsequently, the DMs of arachidonic acid metabolic pathways in the F2vs F1 and F3 vs F1 groups were extracted. No up-DMs were found and total 15 metabolites showed down regulation from F1 to F3 samples (Fig. 10a). The most down-regulated was (19S)-Hydroxyeicosatetraenoic acid (19(S)-HETE), with a 15.07% decrease in F3 over F1 (Fig. 10b), followed by 20-Hydroxyeicosatetraenoic acid (20-HETE) and Prostaglandin A2 (PGA2), with a 7.89% decrease and 7.65% decrease, respectively (Fig. 10c-d).