Physiological and transcriptome analysis reveals the differences in nitrate content between lamina and midrib of flue-cured tobacco

Nitrate is an important precursor of tobacco-specific nitrosamines (TSNAs) and a remarkable difference in nitrate accumulation between lamina and midrib of flue-cured tobacco has long been observed. However, the physiological and molecular mechanisms underpinning this difference remain poorly understood. In this study, physiological and genetic factors impacting nitrate accumulation were identified in pot experiments using flue-cured tobacco K326 with contrasting nitrate content between lamina and midrib. The results showed that three times higher NO3-N content was observed in midrib than that in the lamina, along with lower pigment, NH4-N content, nitrate reductase activity (NRA), sucrose synthetase activity (SSA), and glutamine synthetase activity (GSA) in midrib. Transcriptome analysis revealed that expression of genes involved in porphyrin and chlorophyll metabolism, carotenoid biosynthesis, photosynthesis-antenna proteins, photosynthesis, carbon fixation in photosynthetic organisms, starch and sucrose metabolism, nitrogen metabolism, and biosynthesis of amino acids were significantly lower in midrib than in lamina. qRT-PCR results showed that the expression level of nitrate transporter genes LOC107782967, LOC107806749, LOC107775674, LOC107829632, LOC107799198, LOC107768465 decreased by 2.74, 1.81, 49.5, 3.5, 2.64 and 2.96-folds while LOC107789301 increased by 8.23-folds in midrib but not in lamina. Reduced chlorophyll content might result in low carbohydrate formation which is the source of energy and carbon skeleton supply, then the low capacity of nitrogen reduction, assimilation and transportation, and the poor ability of nitrate reallocation but the high capacity of accumulation might lead to nitrate accumulation in midrib. The results laid the foundation for reducing nitrate content and TSNA formation in tobacco midribs and their products.

Tobacco is an industrial crop that is widely grown throughout the world. Tobacco leaf consists of lamina and midrib, with midrib accounting for about 25-30% of the leaf weight. Not only lamina but also midrib is widely used as raw materials for cigarette production through the making of reconstituted tobacco sheets or midrib cut. Therefore, tobacco midribs have great value when scientifically processed. Midrib has a lower tar level, so it plays a significant part in reducing the hazards of cigarettes 1 . The usage of midrib is also beneficial to cost cutting, thus improving the utilization efficiency of tobacco raw materials. However, the disadvantage of midrib is also obvious, among which is substantial higher levels of nitrate content and subsequent higher formation and accumulation of tobacco-specific nitrosamines (TSNAs) 2,3 than that in the lamina. Nitrate content in midrib of cured tobacco leaf is usually more than 10 times higher than that in the lamina of the same cured leaf 2 .
TSNA is prone to induce malignant tumors in animals and was classified as the first class carcinogen by the International Agency for Research on Cancer 4 . It is well recognized that nitrate is an important precursor of tobacco-specific nitrosamines (TSNAs). Nitrate may easily be reduced to nitrite by microbial activity during leaf curing 5 or produce gaseous NOx during leaf storage under warm or hot conditions 6 , and the subsequent nitrosation of tobacco alkaloids by these nitrosating agents may lead to much increased levels of TSNA formation and accumulation in midrib 1 . Therefore, the reduction of nitrate content is a key for reducing TSNA formation, and the investigation of the mechanisms of nitrate accumulation in midrib is essential, so as to lay the foundation for reducing nitrate content and TSNA formation in tobacco midribs and their products.

Differences in enzymes activities and nitrogen compounds between lamina and midrib.
The results showed that pigment content, enzyme activities, and nitrogen compounds were different between lamina and midrib ( Fig. 1a-l). Chlorophyll a content, chlorophyll b, and carotenoid contents were significantly lower in midrib than those in the lamina. Also, SSA was always lower in midrib than that in the lamina. Lower pigment content may have an influence on carbon fixation and lead to low carbohydrate accumulation in midrib. Also the nitrate reductase activity (NRA) and glutamine synthetase activity (GSA) were lower in midrib than in the lamina. In addition, NH 4 -N, NO 2 -N, total nitrogen content (TN), and soluble protein content in midrib were dramatically lower than those in midrib while the NO 3 -N content and the ratio of NO 3 -N/total nitrogen content (TN) were significantly higher, indicating that the ability of nitrate reduction and assimilation in lamina was higher than midrib. It is noteworthy that the NO 3 -N content accumulated to 25.96 mg g −1 in midrib and was 3.1 times than that in the lamina, which might be due to the weak ability of nitrogen reutilization, leading to nitrate accumulation in midrib.
Quality control, gene expression, and correlation analysis between samples. After filtering the raw reads, a high rate of clean reads from each sample was achieved. In short, the mapping rates of all the samples to the reference genome were above 93%, the GC content of all samples was stable with the distribution ranging from 43.16 to 44.09% and the QC30 value of all samples was above 91% (Table 1), implying successful library construction and RNA sequencing. As shown in Fig. 2a, the FPKM expression levels for each sample were calculated. In addition, the range of correlation coefficients among intra-class was distributed between 0.98 and 1.00 (Fig. 2b). And principal component analysis (PCA) of the data profiles from all 6 samples revealed a high correlation among all samples (Fig. 2c). These results demonstrated that the sequencing data in the present study were adequately representative and valid.

Differentially expressed gene (DEG) selection, Gene Ontology (Go) enrichment, and Kyoto
Encyclopedia of Genes and Genomes (KEGG) pathway analysis of DEGs. The fold change (FC) > 2 or FC < 0.5, and a P.adjust < 0.05, were used thresholds to determine the DEGs. A total of 7560 DEGs (3446 upregulated and 4114 downregulated) were identified between the lamina and the midrib groups (Fig. 3a). And the volcano of differentially expressed genes between the lamina and the midrib was achieved (Fig. 3b).  Comparative analysis of DEGs correlated with carbon and nitrogen metabolism. Transcriptome sequencing technology provides a large amount of information regarding the DEGs that are involved in specific biological responses. Figure 4 showed that porphyrin and chlorophyll metabolism, carotenoid biosynthesis, photosynthesis-antenna proteins, photosynthesis, carbon fixation in photosynthetic organisms, starch and sucrose metabolism, nitrogen metabolism, and biosynthesis of amino acids were lower in midrib than in lamina. In addition, we searched the genes involved in porphyrin and chlorophyll metabolism (LOC107777980, LOC107786828, and LOC107788874) (  (Fig. 4i), which might also be the cause for higher nitrate content in the midrib.
Expression levels of genes related to nitrate transport. Nitrate transporters play an essential role in nitrogen metabolism. The expression level of genes involved in nitrate transporting (LOC107782967, LOC107806749, LOC107775674, LOC107829632, LOC107799198, LOC107768465) was down-regulated while LOC107789301 and LOC107770138 were up-regulated in midrib compared to that in the lamina (Fig. 5). And the qRT-PCR results showed that the expression patterns of the eight genes were identical to those detected by transcriptome sequencing, which confirmed the reliability of RNA-seq data and explained the reason why nitrate content was higher in midrib than that in the lamina.  -TKF  TCA GAC ATG GGT TCC GTG TG   LOC107782967-TKR  GGG GGT CAG CAA CAT AGC AA   LOC107806749-TKF  CAA CAC GAC AGG CAA AGC AG   LOC107806749-TKR  CAA ATC ATC GGC AGC AGC AT   LOC107775674-TKF  TGG AGG GCT ATG CCT TAT GTT   LOC107775674-TKR  AAG CAC CGA GCA ATG GTA TGA   LOC107829632-TKF  CAG TGG TCG TTG ATG GTG ATG   LOC107829632-TKR  TTG ATA GGC TGG CAG GAG GTA   LOC107799198-TKF  GTT CCG ATT TGT CGT CGT TTC   LOC107799198-TKR  GTG GCA TTT GCA TCA TTG GTC   LOC107768465-TKF  GGA TGA AGG AAT GTG GGC TCT   LOC107768465-TKR  TCT TCG GTT TCT GGT GTC

Discussion
In recent year, the midrib has been widely used in cigarette production in the form of tobacco sheets. However, our study showed that the midrib had higher NO 3 -N content of more than 3 times than lamina (Fig. 1a), which is not beneficial to tobacco safety and harm reduction. One strategy to decrease the content of nitrate is to identify the physiological and molecular mechanisms contributing to nitrate accumulation in the midrib. In the studies presented here, the pots experiment was employed to study the physiological and transcriptome differences between lamina and midrib. Overall, the present study demonstrated that the expression of genes involved in porphyrin and chlorophyll metabolism, carotenoid biosynthesis, photosynthesis-antenna proteins, photosynthesis, carbon fixation in photosynthetic organisms, starch and sucrose metabolism, nitrogen metabolism, and biosynthesis of amino acids were significantly lower in midrib than in the lamina (Fig. 4a-h), which might be the cause for higher nitrate accumulation in the midrib. It has long been recognized that chlorophyll content is used as an indicator of photosynthetic capacity and photosynthesis and C metabolism functions to provide both energy and C skeletons for plant growth and N assimilation 8,10 . Our results showed that the midrib had lower chlophyll a chlophyll b, carotenoid, and SSA than that of the lamina (Fig. 1i-l). The previous study has shown that the midribs tend to have fewer chloroplasts in C 3 and CAM plants, which might be the reason for lower chlophyll content in the midrib 24 . More than 30 genes are involved in the chlophyll biosynthesis pathway and any genetic mutation may affect the synthesis of chlorophyll 25 . HEMA1 is considered to play the major role in tetrapyrrole biosynthesis and antisense HEMA1 Arabidopsis plants showed decreased levels of chlophyll 26 . In Arabidopsis thaliana, Alexey et al. 27 showed that the chlorophyll biosynthesis pathway was suppressed in this ChlI mutant. In accordance with this, our results found that some key genes related to pigment biosynthetic process and C metabolism were significantly down-regulated in the midrib, including LOC107777980 (MgPME) 28 , LOC107783891 (CHLI) 29 , LOC107763283 (hemA), and LOC107783257 (CHLP) 30 , which play crucial roles in chlophyll biosynthesis, LOC107772713 (PSY2) 31,32 , which encodes phytoene   www.nature.com/scientificreports/ synthase and controls the carbon flux through the carotenoid biosynthetic pathway, LOC107785687 (SPS2) edcoding sucrose-phosphate synthase that plays the role of rate-limiting steps in sucrose synthesis in higher plant 33 . The down-regulation of these genes might decrease the chlorophyll formation and photosynthesis efficiency in the midrib. Besides, the lower chlorophyll content resulted in a decrease of the chlorophyll a/b binding proteins in midrib. LOC107772663 (LHCb1), which is one of the most abundant chloroplast proteins in plants and mainly functions to collect and transfer light energy to photosynthetic reaction centers 34 , was significantly repressed in midrib. Previous studies showed that that in CAM plants the photochemical parameters describing the performance of PSII were significantly lower in the midribs than in the interveinal leaf area, which reduced the photosynthesis 24 . In the present study, PSI, PSII, and photosynthetic electron transport are key components in the photosynthetic pathway. While LOC107810205 (PsbR), LOC107784985 (PsaO) and LOC107803171 (petF), which were involved in PSI, PSII, and photosynthetic electron transport were down regulated in midrib. LOC107771723 (rbcS), which encodes a key enzyme in the calvin cycle and assimilates atmospheric CO 2 into the biosphere 35 , was also down regulated in midrib. This is consistent with the physiological differences between lamina and midrib. Carbon metabolism is closely related to nitrogen metabolism. The lower capacity of photosynthesis and carbon fixation might influence the nitrogen metabolism and resulted in higher level nitrate in the midrib. NR and GS are two of the most important enzymes in N assimilation 9 . The ammonium taken up by AMTs or derived from nitrate is used to produce a variety of amino acids via the GS/GOGAT cycle 8 . Lu et al. 36 showed that expression of a constitutively activated nitrate reductase (NR) enzyme dramatically decreases leaf nitrate levels in burley tobacco. Meanwhile, recent literature also suggests that the overexpression of GS is able to increase the activity of GS and promote N assimilation efficiency 37 . NLP7 is a primary regulator in nitrate response and regulates the expression of several nitrate-responsive genes including NIA1, NIA2, NRT2.1, and NRT2.2 19,20 . And OsNLP4 transactivats the NRE motif at the promoter of OsNiR encoding nitrite reductase in rice 21 . Xiang et al. 38 has demonstrated that NLP7-overexpressing plants showed lower nitrate accumulation. In this study, NLP7 and NLP4 were down-regulated in the midrib, which was inconducive to the decrease of nitrate accumulation in the midrib. Further investigation of the expression of genes encoding nitrate response, transport, and assimilation led to the discovery of nitrate response genes (NPF6.3, NLP4, and NLP7), nitrate transporters (NPF2.13, NPF3.1, NPF7.3, NPF1.2, and NPF7.2), and nitrate assimilation genes (NIA, GS and GOGAT ) with contrasting transcriptional responses in lamina and midrib. And our results showed that midrib was lower in NR activity, GS activity, NH 4 -N, and soluble protein content while higher in NO 3 -N and NO 3 -N/TN than midrib, suggesting that midrib might retain a weaker capacity of nitrate assimilation. In plants, NO 3 − accumulation depends on its absorption, transport, and metabolism, among which there is a close interdependency that facilitates the coordinated regu- In Conclusion, significant differences were observed in nitrate accumulation between lamina and midrib of flue-cured tobacco. Pigment content and SSA in midrib were significantly lower than that in the lamina, which resulted in insufficient C skeleton for nitrogen metabolism. Meanwhile, the greater nitrate accumulation was probably conferred by more disadvantageous aspects such as weak nitrogen reduction, weak nitrogen assimilation, poor ability in reallocation, and high capacity of accumulating nitrate in midrib than in the lamina. The above insights to the physiological and molecular basis of carbon and nitrogen differences in lamina and midrib would be helpful for providing direction for decreasing nitrate accumulation in the midrib.

Materials and methods
Plant material and study design. The flue-cured tobacco variety K326 was used in this study. Seeds were sterilized with 2% (v/v) sodium hypochlorite for 5 min twice and then were sown in a floating system. Forty days after sowing, seedlings were transplanted in 7.1 cm × 7.8 cm (diameter × depth) plastic pots and cultivated with Hoagland solution. Pot experiments were conducted on substrate culture in the greenhouse that maintained a temperature of 25 ± 2 °C, an average photosynthetic photon flux density of 400 μmol m −2 s −1 , and relative humidity of 80%. Laminas and midribs were collected separately 15 days after seedlings being transplanted. Fully expanded leaves (length > 5 cm, up to down, the fourth leaf from top) from the same position in three pots of each treatment was sampled in an ice box. Half of the samples were frozen in liquid nitrogen and stored in a freezer at − 80 °C, while the other half were deactivated at 105 °C for 20 min and then dried at 60 °C for 48 h. Frozen samples were used for transcriptome analysis, enzyme activity determination, soluble protein and NH 4 -N content investigation. Dried samples were used for determination of nitrate content. Every treatment had three biological replicates. The K326 seeds used in this study were provided by Yunnan Tobacco Company and the collection of the plant material complied with relevant institutional, national and international guidelines and legislation. In preliminary tests, laminas and midribs of seedlings were collected on the 7th, 15th, and 21st days after seedlings being transplanted to determine the difference in nitrate content. The results showed that the nitrate content of midrib was significantly higher than that of the lamina on the 15th day. So laminas and midribs were collected separately 15 days after seedlings being transplanted.