SWEET Transporters and the Potential Functions of These Sequences in Tea (Camellia sinensis)

Tea (Camellia sinensis) is an important economic beverage crop. Its flowers and leaves could be used as healthcare tea for its medicinal value. SWEET proteins were recently identified in plants as sugar transporters, which participate in diverse physiological processes, including pathogen nutrition, seed filling, nectar secretion, and phloem loading. Although SWEET genes have been characterized and identified in model plants, such as Arabidopsis thaliana and Oryza sativa, there is very little knowledge of these genes in C. sinensis. In this study, 28 CsSWEETs were identified in C. sinensis and further phylogenetically divided into four subfamilies with A. thaliana. These identified CsSWEETs contained seven transmembrane helixes (TMHs) which were generated by an ancestral three-TMH unit with an internal duplication experience. Microsynteny analysis revealed that the large-scale duplication events were the main driving forces for members from CsSWEET family expansion in C. sinensis. The expression profiles of the 28 CsSWEETs revealed that some genes were highly expressed in reproductive tissues. Among them, CsSWEET1a might play crucial roles in the efflux of sucrose, and CsSWEET17b could control fructose content as a hexose transporter in C. sinensis. Remarkably, CsSWEET12 and CsSWEET17c were specifically expressed in flowers, indicating that these two genes might be involved in sugar transport during flower development. The expression patterns of all CsSWEETs were differentially regulated under cold and drought treatments. This work provided a systematic understanding of the members from the CsSWEET gene family, which would be helpful for further functional studies of CsSWEETs in C. sinensis.


INTRODUCTION
SWEET is a novel kind of low-affinity sugar transporter, which does not depend on the pH value of the environment and transports in two directions along the concentration gradient (Yuan and Wang, 2013). The members of the SWEET family are widely distributed, which are ubiquitous in higher eukaryotes and also distributed in protozoa, metazoa, fungi, bacteria, and archaea (Hamada et al., 2005;Saier et al., 2006;Xuan et al., 2013). The membrane proteins encoded by the SWEET have a certain number of conserved transmembrane domains, which are named MtN3/saliva (Hamada et al., 2005). This domain was first found in nodulin from the roots of Medicago sativa (Gamas et al., 1996). Recently, phylogenetic analysis has been performed for the main protein members of the SWEET family, which are divided into three branches: (i) the plant-like SWEET, most of which contain two MtN3/saliva transmembrane domains, (ii) the animal-like SWEET, which has two specific MtN3/saliva domains, and (iii) some SWEET proteins from bacteria to Archaea (cocci) and nematodes, which consist of an MtN3/saliva domain containing three transmembrane alpha helices (Yuan and Wang, 2013). A large number of SWEET family members have been found in Arabidopsis thaliana (Wipf et al., 2020), Vitis vinifera (Chong et al., 2014), Solanum lycopersicum (Shammai et al., 2018), Manihot esculenta (Cao et al., 2019), Eucalyptus grandis (Yin et al., 2020), etc., and are involved in many physiological processes. For instance, RPG1 (ATSWEET8) plays a vital role in microspore outer wall formation in Arabidopsis (Guan et al., 2008); in Oryza sativa, the SWEET family members are also involved in reproductive development (Wang et al., 2010); SAG29 (AtSWEETl5), located on the plasma membrane in Arabidopsis, can regulate cell activity in a hypersaline environment (Seo et al., 2011). The SWEETs also participate in the regulation of the aging process. The overexpression of OsSWEET5 can cause growth delay and premature aging in rice seedlings, while no phenotypic changes were observed in the knockout ones (Zhou et al., 2014). Chu et al. (2006) found that the rice harboring mutant OsSWEET11 (Xa13) can resist powdery mildew, suggesting that SWEETs not only served as a glucose transporter but also participated in the interaction between host and pathogen. However, the members of the CsSWEET gene family have not yet been identified, and the roles of CsSWEETs remain unclear in tea plant (Camellia sinensis).
Camellia sinensis, which originated from the tropical area of Southwest China, is an important economic beverage crop in China (Wei et al., 2018;Zhang et al., 2020b). Sugar transport and homeostasis contribute to plant growth and development (May et al., 1998;Lastdrager et al., 2014;Rahimi et al., 2019;Pan et al., 2020;Saddhe et al., 2020). Some published papers focused on the genes that participated in sugar metabolism, such as hexose kinase, invertase, and galactinol synthase, in C. sinensis (Yue et al., 2015;Zhou et al., 2017;Samarina et al., 2020). Due to the economic interest in C. sinensis as a beverage crop, its genome, proteome, and transcriptome were recently sequenced and released (Wei et al., 2018;Wu et al., 2018;Liu et al., 2020;Xia et al., 2020;Zhang et al., 2020b), which help us to analyze the SWEET genes in C. sinensis systematically. In the present study, 26 CsSWEETs in C. sinensis were identified, and their gene structures, phylogenetic, microsynteny, and expression were analyzed. Our study revealed the functions of these CsSWEETs and provided candidate genes for further research.

MATERIALS AND METHODS
Database Search for C. sinensis Genome The C. sinensis genome with GFF annotation, CDS, and protein files were downloaded from the Tea Plant Information Archive 1 (TPIA) database (Xia et al., 2019). The AtSWEETs of A. thaliana and HsSWEET1 of Homo sapiens were obtained from TAIR 2 and Genbank 3 , respectively. The HMM profile of MtN3/saliva domain (PF03083) was obtained from the Pfam database 4 (Mistry et al., 2020), and the HMMER (version 3.1) software (Mistry et al., 2013) was used to identify CsSWEET proteins in C. sinensis genome (E-value ≤ 1e−3). The Pfam (version 33.1), SMART (version 9), and InterPro (version 5) were used to confirm the above-mentioned identified sequences with E-value ≤ 1e−3 (Jones et al., 2014;Letunic and Bork, 2018;Mistry et al., 2020). Finally, we manually discarded these sequences with a lack of complete or core MtN3/saliva domain for further analyses.

Phylogenetic Analyses
The MAFFT (version 7) software was used to execute the multiple alignments of all SWEET proteins with default parameters (Katoh et al., 2005). The best substitution model of these SWEET proteins was determined by ModelFinder software (Kalyaanamoorthy et al., 2017). The maximum likelihood (ML) tree was determined by IQ-tree (version 2.1.2) software (Nguyen et al., 2015) with an SH-aLRT test for 1,000 random addition replicates and a bootstrap test for 10,000 replicates as described by Cao et al. (2020a). The Figtree software was used to visualize this ML tree.

Internal Repeats and Topological Analyses
The ClustalX (version 2) software (Thompson et al., 2003) was used to create the multiple alignments of SWEETs from C. sinensis and A. thaliana. The HHrepID was used to identify the internal repeats in SWEET proteins (Söding et al., 2006). In addition, we also used the AveHAS to calculate the topological conservation (Zhai and Saier, 2001a,b) as well as create the hydropathy, amphipathicity, and similarity plots, respectively (Zhai and Saier, 2001a). The GFF annotation file was obtained from the TPIA (see text footnote 1) database (Xia et al., 2019). The TBtools (version 1) software was used to generate this gene structure map based on the GFF annotation file . The MEME online tool was used to predict the conserved motifs (Bailey et al., 2015).

Microsynteny Analysis
The MicroSyn (version 1) software was used to detect the microsynteny of CsSWEET genes in C. sinensis genome with a threshold E-value of <1e−5 (Cai et al., 2011). In this study, we determined a syntenic block, a region which include three or more conserved homolog genes that were distributed within 15 genes upstream and downstream of CsSWEET genes as described by Cao et al. (2020b).

RNA-Seq Expression Analysis
In our study, transcriptome data from various tissues in the public database NCBI contained three biological replicates. The RNA-seq reads, including seed, flower, stem, root, two and a bud, one and a bud, old leaf, mature leaf, the second leaf, the first leaf, lateral bud, apical bud, early stage lateral bud, CP24 (24 h after pollination; CP, cross-pollinated), CP48, CP72, SP24 (SP, self-pollinated), SP48, and SP72, were obtained from NCBI (PRJNA291116 and PRJNA230752). The Trimmomatic (version 0.33) was used to carried out the quality-based trimming (Bolger et al., 2014). The HISAT2 was used to map the paired reads to the C. sinensis genome with default parameters (Kim et al., 2019). The StringTie was used to calculate the fragments per kilobase of exon model per million reads mapped values of differently expressed genes (Pertea et al., 2016).

RESULTS AND DISCUSSION
The Phylogenetic Analysis Divides CsSWEET Genes Into Four Subfamilies in C. sinensis In the present study, we identified 26 CsSWEET genes in C. sinensis genome by using HMMER 3.0 software, similarly as in Manihot esculenta, Jatropha curcas, Ricinus communis, Vernicia fordii, Malus × domestica, Oryza sativa, Zea mays, Glycine max, and Pyrus bretschneideri Yuan and Wang, 2013;Chong et al., 2014;Cao et al., 2019;Yin et al., 2020). Subsequently, we renamed the newly identified CsSWEETs based on the nomenclature of the AtSWEETs of A. thaliana. The detailed information of each CsSWEET, including chromosome position, gene identifiers, isoelectric point, molecular weight, and gene name, is shown in Table 1.
To gain insight into the evolutionary relationship of CsSWEETs in C. sinensis, the SWEETs from C. sinensis and A. thaliana were aligned by MAFFT software. Remarkably, HsSWEET1 from Homo sapiens was an outgroup, which has only one copy and could transport glucose in H. sapiens. In V. fordii, Cao et al. (2019) found that VfSWEET1 contributes to the balance of sucrose levels, and other VfSWEETs may play key roles in the growth and development of plants (Cao et al., 2019). In A. thaliana, AtSWEETs were identified with versatile functions, such as control fructose content and sucrose efflux (Chen et al., , 2012. In our study, the SWEETs from C. sinensis and A. thaliana were clustered into four subfamilies, including subfamily I, subfamily II, subfamily III, and subfamily IV (Figure 1). Subsequently, there was at least one AtSWEET from A. thaliana in each subfamily. Previously published papers have confirmed that members from different SWEET subfamilies may have multiple biological functions, such that the members from subfamily I and subfamily II were efficient hexose transporters (Figure 1), the members from subfamily III may be responsible for sucrose transporters, and the members from subfamily IV appear to be predominantly fructose transporters (Chen et al., , 2012Hu et al., 2018;Cao et al., 2019).

The Structure Analyses Reveal the Distribution of Exon-Intron and the Conserved Composition of Motif in C. sinensis
According to the predicted sequences and GFF annotation files, we determined the distribution of exon-intron of the 26 CsSWEET genes in C. sinensis. As shown in Supplementary  Figure 1, most CsSWEET genes contained five introns; however, several of the members from the CsSWEET gene family contained two, three, or four introns. Remarkably, there are extreme differences in the number of introns from the CsSWEET subfamily II, ranging from 4 to 14, indicating that these CsSWEETs may have experienced intron loss or gain some during evolution (Supplementary Figure 1). These results were also confirmed by previous studies. For example, Cao et al. (2019) found that the number of introns from VfSWEETs ranged from four to 10. Additionally, we found that, although the lengths of members from CsSWEET gene family vary, introns stepped into almost the same position of the gene open reading frame. The MEME website was used to identify the sequence characteristics of CsSWEET genes. Subsequently, we determined 20 distinct motifs for CsSWEETs and annotated these motifs by Pfam and SMART database (Letunic and Bork, 2018;Mistry et al., 2020). The SWEET domain (MtN3/saliva) was encoded by motifs 1, 2, 3, and 4 (Supplementary Figure 2). However, the remaining motifs were not scanned for function annotations in Pfam and SMART database. Overall, the CsSWEET proteins within the same subfamily exhibited similar conserved motifs, especially in paralogs gene pairs, suggesting that these sequences might contain similar functions in C. sinensis.

Internal Repeat and Transmembrane Domain Analysis of the CsSWEETs Reveal Insight Into Key Functional Residues
In general, SWEET proteins have been confirmed to contain seven transmembrane helixes (TMHs) in eukaryotes (Feng and Frommer, 2015;Jia et al., 2017). As expected, the membrane topology of CsSWEETs and AtSWEETs revealed that these proteins contained a moderately amphipathic nature with seven well-conserved hydrophobicity peaks (TMH1-TMH7) as determined by AveHAS plot (Figures 2A,B). Compared with TMHs, the loop area between two TMHs is not conservative in this study ( Figure 2C). Previously published manuscripts have shown that the internal fusion and duplication of small fragments play an important role in the SWEETs evolution, which has been the focus of discussion by researchers (Hu et al., 2018;Cao et al., 2019). Researchers have studied the internal duplication of the SWEET proteins (Li et al., 2017;Hu et al., 2018;Cao et al., 2019), but direct evidence of the CsSWEETs from C. sinensis has not been reported. To confirm the presence of internal duplication, a repeat analysis of CsSWEETs and AtSWEETs was performed by the HHrepID program (Söding et al., 2006). The results confirmed that TMH1-TMH3 and TMH5-TMH7 were located in the duplicated regions, implying that SWEETs originate from an ancestral three-TMH unit which experienced an internal repeat duplication, which is in agreement with previously published papers (Li et al., 2017;Hu et al., 2018;Cao et al., 2019). Li et al. (2017) revealed that Y, P, and D from TMH2, TMH5, and TMH7 were fully conserved residues in pear, respectively (Li et al., 2017). Among these three residues, Y mutation into A will lead to the loss of transport function in A. thaliana (Xuan et al., 2013). Proline (P) is a key component of TMHs and allows for dynamic processes during the transport cycle as described by Deber and Therien (2002). In the present study, we suggested that these three residues (Y, P, and D) might contain important functions within the activity of CsSWEETs in C. sinensis. In addition, the mutation of 58 amino acid residues G into D can significantly reduce AtSWEET1 activity in A. thaliana (Xuan et al., 2013). However, this amino acid is not completely conserved in C. sinensis, including the residue G in CsSWEET12 that had changed to N and in CsSWEET15a that had changed to K, but these proteins also have transport activity (Figure 2C), which has been verified by RNA-seq. We also noted that the first and the second MtN3/saliva domain included TMH1-TMH3 and TMH5-TMH7, respectively.

Microsynteny Analysis Indicates That Large-Scale Duplication Events Contribute to the Expansion for
CsSWEET Genes in C. sinensis The genome of C. sinensis experienced two rounds of wholegenome duplications (WGDs), including an ancient WGD event (∼90 to 100 Mya) and a recent WGD event (∼30 to 40 Mya) (Wei et al., 2018;Wu et al., 2018;Liu et al., 2020;Xia et al., 2020;Zhang et al., 2020b). To gain insight into the expansion mechanism of the members from CsSWEET gene family, we carried out a microsynteny analysis with E-value 10−5 as described by FIGURE 4 | Expression profiles of the CsSWEETs in different tissues, including seed, flower, stem, root, two and a bud, one and a bud, old leaf, mature leaf, the second leaf, the first leaf, lateral bud, apical bud, and early stage lateral bud. CP24 = 24 h after pollination; CP, cross-pollinated -CP48, CP72; SP24 (SP, self-pollinated), SP48, and SP72. Cao et al. (2020b). In our study, CsSWEET3c/CsSWEET3a, CsSWEET10c/CsSWEET12, CsSWEET10c/CsSWEET10b, CsSW EET7b/CsSWEET7c, CsSWEET3c/CsSWEET3b, CsSWEET3a/C sSWEET3b, CsSWEET7a/CsSWEET5c, and CsSWEET2b/Cs SWEET2a contained more than three pairs of conserved flanking genes, suggesting that significant microsynteny might occur in the CsSWEET genes (Figure 3). These results indicated that the large-scale duplication events contribute to the expansion for CsSWEET genes in C. sinensis.
Generally, the evolutionary data of WGD events were estimated by Ks values (Wang et al., 2011;Tiley et al., 2018;Zwaenepoel and Van De Peer, 2019). As shown in Table 1, we found that the Ks values of CsSWEET duplication pairs ranged from 0.0103 to 2.1158 ( Supplementary  Table 1 and Supplementary Figure 3). Subsequently, we found that CsSWEET3c/CsSWEET3a, CsSWEET3a/CsSWEET3b, and CsSWEET2b/CsSWEET2a might be generated through a recent WGD event, while CsSWEET3c/CsSWEET3b and CsSWEET10c/CsSWEET12 might be generated through an ancient WGD event, suggesting that these two rounds of WGD events might play key roles in the expansion for CsSWEET genes in C. sinensis.

The Expression of CsSWEETs Reveals Possible Functions in C. sinensis
Previously published work provided transcriptome data for CsSWEET genes in C. sinensis (Xia et al., 2019;Xia et al., 2020). The members from SWEET gene family are found to play diverse functional roles in various tissues and contribute to different sugar transport mechanisms in plants (Chen, 2014;Hedrich et al., 2015;Li et al., 2017;Cao et al., 2019). Differential expression analysis of CsSWEETs in C. sinensis is helpful for us to find out the specialized functions of these CsSWEETs in sugar transport from the practical application point of view. In the present study, we collected transcriptome data from 19 different tissues, including seed, flower, stem, root, two and a bud, one and a bud, old leaf, mature leaf, the second leaf, the first leaf, lateral bud, apical bud, early stage lateral bud, CP24 (24 h after pollination; CP, cross-pollinated), CP48, CP72, SP24 (SP, self-pollinated), SP48, and SP72. The published papers indicated that the expression of SWEETs in seeds pollen, flower, and embryo sacs in Sorghum bicolor, V. fordii, Litchi chinensis, O. sativa, and A. thaliana was higher (Yuan and Wang, 2013;Mizuno et al., 2016;Cao et al., 2019;Xie et al., 2019), implying that these genes may contribute to reproductive development. Here CsSWEET1a that was clustered into the same subfamily with VfSWEET1 from V. fordii, PbSWEET14 from P. bretschneideri, and AtSWEET1 from A. thaliana was extremely highly expressed in the leaf of C. sinensis. PbSWEET14 and VfSWEET1 contribute to the efflux of sucrose in leaves (Li et al., 2017;Cao et al., 2019), while AtSWEET1 plays an important role as a single glucose transporter in multiple systems. Therefore, the CsSWEET1a might play crucial roles in the efflux of sucrose and act as a glucose uniporter in the leaves of C. sinensis (Figure 4). AtSWEET16 and AtSWEET17 were shown to be vacuolar hexose transporters that controlled fructose content . The CsSWEET17b, which is the orthologous gene of AtSWEET16 and AtSWEET17, was extremely highly expressed in the stems of C. sinensis, suggesting that this gene could control fructose content as a hexose transporter. The expression patterns of CsSWEET2a, CsSWEET5b, and CsSWEET7a were diverse in several tissues, implying that these CsSWEETs might play a role in these tested tissues of C. sinensis (Figure 4).
In C. sinensis, the expression of CsSWEETs was examined in pollen development. We found that 10 CsSWEETs, including CsSWEET5a, CsSWEET3a, CsSWEET9a, CsSWEET7b, CsSWEET17a, CsSWEET9b, CsSWEET15c, CsSWEET10c, CsSWEET1b, and CsSWEET5c, were highly expressed in one or several pollen stages of C. sinensis (Figure 4). As previously reported, Chen et al. (2012) identified AtSWEET11 and AtSWEET12 that contributed to phloem loading and pollen nutrition. Gao et al. (2018) found that the OsSWEET11 from O. sativa can greatly affect the starch pollen content. In our study, CsSWEET9b, CsSWEET9a, and CsSWEET10c are the orthologous genes of AtSWEET11 and AtSWEET12, indicating that these highly expressed genes may share similar roles and functions.

Stress-Induced Expression Profiles of CsSWEETs in C. sinensis
Plants often suffer from a variety of abiotic stresses in the process of growth and development (Franco et al., 2011;Keunen et al., 2013;Etesami, 2018). Previous studies have shown that SWEETs help to control the responses to environmental stresses in plants (Li et al., 2018;Zhang et al., 2020a). Therefore, the expression patterns of CsSWEETs were investigated in response to different stresses, including cold stress and drought stress. In our study, the RNA-seq data for cold stress and drought stress were obtained from the TPIA (see text footnote 1) database (Xia et al., 2019). In response to drought, seven genes (CsSWEET3a, CsSWEET7a, CsSWEET1a, CsSWEET10a, CsSWEET2a, CsSWEET7b, and CsSWEET2c) were up-regulated under polyethylene glycol (PEG) for 72 h. Compared with control, nine genes (CsSWEET2b, CsSWEET5b, CsSWEET15c, CsSWEET1b, CsSWEET9b, CsSWEET17a, CsSWEET17c, CsSWEET7c, and CsSWEET15b) were down-regulated under all PEG treatments (Figure 5). In response to cold, four genes (CsSWEET3b, CsSWEET1a, CsSWEET1b, and CsSWEET15c) were up-regulated under all cold treatments. Compared with control, nine genes (CsSWEET5b, CsSWEET2a, CsSWEET17b, CsSWEET15b, CsSWEET17a, CsSWEET7a, CsSWEET2b, CsSWEET3a, and CsSWEET17c) were down-regulated under all cold treatments (Figure 6). Among them, the expression of CsSWEET1a was significantly up-regulated in all chosen two treatments, indicating that this gene might help tea resist environmental stresses. This phenomenon was different from previous studies. For example, the banana MaSWEET1a was not induced by osmotic, cold, and salt stresses (Miao et al., 2017). The expression level of GhSWEET1 was only upregulated at 6 h under cold stress, but it was almost not expressed when induced by drought stress (Li et al., 2018). Taken together, we believed that tea plants might have developed specialized regulatory mechanisms for different abiotic stresses.

CONCLUSION
In the present study, we identified 28 CsSWEETs in the C. sinensis genome. Further analyses for ML tree, intron-exon, and duplication suggested that the conservation of CsSWEETs was accompanied by a certain degree of divergence. The expression profiles of all CsSWEETs suggested that several genes were highly expressed in reproductive tissues, indicating that these genes played important roles in sugar transport. Additionally, 28 CsSWEETs were differently expressed for RNA-seq data under cold and drought stresses, indicating that tea plants might have developed specialized regulatory mechanisms for different abiotic stresses.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
LJ and XZ performed the experiments, analyzed the data, and wrote the manuscript. LJ, CS, XZ, and JY conceived the research and revised the manuscript. All the authors read and approved the final manuscript.