Functional characterization and analysis of transcriptional regulation of sugar transporter SWEET13c in sugarcane Saccharum spontaneum

Sugarcane is an important crop for sugar production worldwide. The Sugars Will Eventually be Exported Transporters (SWEETs) are a group of sugar transporters recently identified in sugarcane. In Saccharum spontaneum, SsSWEET13c played a role in the sucrose transportation from the source to the sink tissues, which was found to be mainly active in the mature leaf. However, the function and regulation of SWEETs in sugarcane remain elusive despite extensive studies performed on sugar metabolism. In this study, we showed that SsSWEET13c is a member of SWEET gene family in S. spontaneum, constituting highest circadian rhythm-dependent expression. It is a functional gene that facilitates plant root elongation and increase fresh weight of Arabidopsis thaliana, when overexpressed. Furthermore, yeast one-hybrid assays indicate that 20 potential transcription factors (TFs) could bind to the SsSWEET13c promoter in S. spontaneum. We combined transcriptome data from developmental gradient leaf with distinct times during circadian cycles and stems/leaves at different growth stages. We have uncovered that 14 out of 20 TFs exhibited positive/negative gene expression patterns relative to SsSWEET13c. In the source tissues, SsSWEET13c was mainly positively regulated by SsbHLH34, SsTFIIIA-a, SsMYR2, SsRAP2.4 and SsbHLH035, while negatively regulated by SsABS5, SsTFIIIA-b and SsERF4. During the circadian rhythm, it was noticed that SsSWEET13c was more active in the morning than in the afternoon. It was likely due to the high level of sugar accumulation at night, which was negatively regulated by SsbZIP44, and positively regulated by SsbHLH34. Furthermore, in the sink tissues, SsSWEET13c was also active for sugar accumulation, which was positively regulated by SsbZIP44, SsTFIIIA-b, SsbHLH34 and SsTFIIIA-a, and negatively regulated by SsERF4, SsHB36, SsDEL1 and SsABS5. Our results were further supported by one-to-one yeast hybridization assay which verified that 12 potential TFs could bind to the promoter of SsSWEET13c. A module of the regulatory network was proposed for the SsSWEET13c in the developmental gradient of leaf and circadian rhythm in S. spontaneum. These results provide a novel understanding of the function and regulation of SWEET13c during the sugar transport and biomass production in S. spontaneum.


Background
High crop yields are beneficial for food production, yet the underlying mechanism that determine crop yields is still not fully understood. Through absorbing light energy, plants leaf can fix carbon to form photosynthetic products (carbohydrates) [1]. Sugars, is one of the primary carbohydrate form [2]. In plants, 35% ~ 40% of sugars are consumed to provide energy for various growth activities including cell elongation, division, differentiation, nutrient absorption, and plant development. Majority of sugars function as metabolic intermediates in cells (monosaccharides, amino acids, organic acids, etc.) and the remaining sugars are stored in the vacuole as the original formation, fixed to a mobile polymer (plasmid starch), or added to structural biomass (cellulose, hemicellulose and lignin) [3].
Sucrose is synthesized in mesophyll cells and diffuses from cell to cell via plasmodesmata or a long-distance transportation. Hereafter sucrose could be transferred from the phloem parenchyma cells to the apoplast and loaded into the phloem sieve element/companion cells (SEs/CCs) via sucrose transporters [4][5][6][7][8]. Sugars Will Eventually be Exported Transporter (SWEETs)/sucrose transporters (SUTs) are a series of crucial transporters that play essential roles in sucrose transport pathways [9,10]. SWEET proteins with seven transmembrane helices are identified as a novel group of sugar transporters that facilitate the transport of sugars like glucose, fructose and/or sucrose bi-directionally. SWEETs participate in numerous metabolic and physiological processes in plants, including phloem loading, nectar secretion, pollen nutrition, bacterial infection, senescence, seed filling and copper transport [11]. In Arabidopsis thaliana, researchers have identified 17 members of SWEET gene family [12]. AtSWEET9 expresses at the nectary and possesses a negative correlation with the starch content in plants, which is required for nectar secretion [13,14]. AtSWEET13 and AtSWEET14 mediate cellular gibberellin uptake to regulate anther dehiscence, as well as seeds and seedling phenotypes [15]. AtSWEET11 and AtSWEET12, presenting on the plasma membrane of phloem parenchyma cells, mediate phloem loading of sucrose [16]. As the closest homolog gene of AtSWEET11 and AtSWEET12, AtSWEET13 has higher expression in Atsweet11;12 double mutant plants. Moreover, Atsweet11;12 double mutants have lower levels of starch accumulation in leaves and radioactive tracer efflux from the petioles, mirroring SWEET11 and SWEET12 cooperatively transport sucrose out of leaves [16]. In rice (Oryza sativa L.), the SWEET gene family consists of 21 members [17]. OsSWEET11 can not only transport sucrose but contributes to grain filling. OsSWEET13 functions as the targeted susceptible gene for PthXo2 of X. oryzae pv. oryzae in the bacterial blight disease. Additionally, OsSWEET11, 12 and 14 are verified as the targets of Xoo effectors [18,19]. Osdof11 (DNA binding with one finger) mutation exerts a stunted phenotype due to reduced sucrose transported by low expression of both SUT and SWEET genes [20]. In maize (Zea mays L.), ZmSWEET13s paralog (a, b, c) triple knock-out mutants exhibit severely reduced photosynthesis abilities and dysregulated glucose metabolism [21].
Sugarcane (Saccharum spp., Poaceae), a perennial and tropical/subtropical crop, contributes to approximately 80% sugar and 40% of global ethanol production for humans [22]. As a typical C4 crop, sugarcane displays high efficiency on the utilization of water, nitrogen, and light compared to C3 crops, leading to greater yields when confronted with intense light, low carbon dioxide concentration and drought conditions. The sugarcane's hyperploid and large genome have slowed down the research within sugar-related fields.
Recently, we identified 22 members of SWEET gene family in S. spontaneum [23], among which SWEET13s (a, b, c) were found to be duplicated genes. Noticeably, SWEET13c is increased by 1001-fold expression from the leaf base to tip, which was consistent with photosynthesis intensity (from weak to strong) through the blade. During the diurnal cycle, SWEET13c had maximal and minimum expression at 6:00 and 18:00, respectively. Additionally, 2.15-fold change of SWEET13c transcript abundance was detected,,suggesting that it is involved in diurnal rhythms and functions as a strong candidate gene which is supporting sucrose transport from source to sink in S. spontaneum [23]. Here, we focus on the SsSWEET13c to explore its ability to transport sugars and facilitate plant growth. We also aimed to identify the potential transcription factors (TFs) binding to the upstream of SsSWEET13c. Furthermore, we assessed the expression pattern of all identified TFs. These findings could help us further reveal the physiological effects and regulatory mechanisms of SWEET13c in S. spontaneum.

Conclusions:
A module of the regulatory network was proposed for the SsSWEET13c in the developmental gradient of leaf and circadian rhythm in S. spontaneum. These results provide a novel understanding of the function and regulation of SWEET13c during the sugar transport and biomass production in S. spontaneum. Keywords: Saccharum spontaneum, Sugar transport, SWEET13c, Transcription factor

SsSWEET13c overexpression reduces contents of sugars in leaves
Given that the transformation system of sugarcane is still unsolved, to examine whether SsSWEET13c can transport sugars, we conducted experiments in a transgenic A. thaliana over-expressing SsSWEET13c driven by the constitutive Cauliflower mosaic virus (CaMV) 35S promoter. Transgenic lines were screened to the homozygous T3 generation. Results of RT-qPCR showed that six lines of 35S pro : SsSWEET13c presented higher expression levels relative to the A. thaliana Columbia ecotype (wildtype, WT) (Additional file 1). Considering the variation of overexpression folds, the average expressional value of these six 35S pro : SsSWEET13c lines was utilized through consecutive measures (Additional file 1).
The leaves of WT and overexpression plants were collected to allow the determination of the contents of glucose, sucrose, fructose and total soluble sugar after 21 days of growth. The results suggested that SsSWEET13c overexpression plants possessed a significant alteration in the sugar-transporting ability relative to WT. For glucose content, the WT plants contained 76.3 mg g −1 , while the SsSWEET13c overexpression plants contained 19.3 mg g −1 , which was significantly reduced (p < 0.0001) (Fig. 1a). Compared with 18.0 mg g −1 of the WT plants, the average sucrose level of SsSWEET13c overexpression plants was 13.1 mg g −1 (Fig. 1a). Regarding the fructose content, the average level of 5.2 mg g −1 in SsSWEET13c overexpression plants was lower than that of 7.2 mg g −1 in the WT (Fig. 1a). Regarding the total soluble sugar contents, WT plants contained 510.2 mg g −1 , whereas a significant reduction to 348.1 mg g −1 was observed in SsSWEET13c overexpression plants (p < 0.05) (Fig. 1a). These results indicated that SsSWEET13c could assist in transporting the produced sugar,lead to reduction of sugar concentration.

SsSWEET13c overexpression enhances the root length and fresh weight
For analysis of sugar content, we took eight seeds from WT and six over-expression SsSWEET13c lines and observed for 7 d for any changes in root elongation. The results suggested that the average root length of SsS-WEET13c overexpression lines was 3.41 cm, with the longest root being 5.54 cm which is significantly longer than the 2.22 cm of the WT (Fig. 1b, c and Additional file 2). Furthermore, we transferred the A. thaliana plants into nutrient soil to cultivate for additional 7 d (at 23 °C and 16 h/8 h light/dark), We measured average fresh weight of these WT and SsSWEET13c overexpression lines including the whole plant weight, root weight and shoot weight. In the Arabidopsis WT, the average fresh weight of the whole plants, shoots and roots were 9.53 mg, 7.73 mg and 1.90 mg, respectively ( Fig. 1d and Additional file 2). However, in SsSWEET13c overexpression lines, the average fresh weights were 58.0 mg (whole plants), 54.7 mg (shoots) and 3.2 mg (roots) (Fig. 1d, e). These results demonstrated that SsSWEET13c is responsible for increase increased in root length and fresh weight (Fig. 1b-e).
Next, we correlated the differential accumulation of sugar (glucose, sucrose, fructose and total soluble sugar) in the leaves with biomass accumulation (length of root and weight of root, shoot and whole plant). We found out that, in comparison with WT, sugar content in leaves of overexpression SsSWEET13c plants decreased by more than 1.38-fold change, while biomass accumulation increased by more than 1.65-fold change ( Table 1). This indicated that sugar content in leaves were negativey correlated with biomass accumulation.

Yeast one-hybrid identifies 20 TFs regulating SWEET13c
Although the ability of SsSWEET13c to transport sugars and enhanced plant growth had been preliminarily determined. Further, we interested in the TFs which are regulating SsSWEET13c and their gradual increase in expression level from the leaf base to tip, and a decrease in expression during the day followed by rising expression level at night. The prediction of TF binding sites was initially performed by the PlantCare database. In total, we predicted 27 potential TF binding sites in a 2 kb region upstream from the promotor of SsSWEET13c (Table 2). Among them, 15 TFs binding sites, including three Box 4, two chs-CMA1a, four G-Box, TCCC-motif, GT1-motif, MRE, TCT-motif, ATC-motif, and 3-AF1, are predicted to be responsive to light. Three binding sites, CGTCAmotif, TGACG-motif and ABRE, were predicted as hormone-responsiveness elements, and two MBS binding sites were predicted as drought-inducibility elements ( Table 2).
To further determine TFs binding role at the promoter of SsSWEET13c, a yeast one-hybrid (Y1H) experiment was performed. To begin, the promoter sequence was divided into five short fragments, named F0 (-1999 to -1721), F1 (-1580 to -1250), F2 (-1218 to -707), F3 (-660 to -310) and F4 (-309 to -1) (Fig. 2, Additional file 3). Except for F2, the remaining four segments containing the core motifs in promoters of the target gene were cloned upstream of the Aureobasidin (AbA) resistance gene (Additional file 4) and transformed intoY1H Gold strains as a bait. As a result, the self-activation of three bait fragments was successfully inhibited. The inhibition concentration was 200 ng/µL for F0, F1 and F3. Next, yeast competent cells with the promoter fragments of SsSWEET13c were transformed with S. spontaneum Fig. 1 The measurement of sugar contents, root length and fresh weight for wild-type and 35S:SsSWEET13c of Arabidopsis plants. a sugar contents b root length (7-days old). c root phenotype (7-days old). d fresh weight (15-days old). e plant phenotype (15-days old). *, p < 0.05. **, p < 0.01. ***, p < 0.001. ****, p < 0.0001 cDNA library plasmids, as well as the positive (pAbAi-53) and negative controls, which were cultivated on the basic SD medium lacking the Leu with a corresponding AbA level. Yeast cells grown were screened for a second time under the same cultivated conditions. Finally, using the universal primers of the vector pGADT7 for DNA sequence, we obtained sequencing results of segments interacting with promoters of SsSWEET13c and 97 gene annotation results based on the BLAST search (https:// www. ncbi. nlm. nih. gov/). Of those, 20 genes were summarized as potential TFs regulating SsS-WEET13c via aligning using PlantRegMap/PlantTFDB

Expressional patterns of the TFs regulating SsSWEET13c
To test the expression patterns of TFs summarized from the Y1H experiment on SsSWEET13c, we conducted RNA-seq analyses in S. spontaneum at three RNA-seq datasets: 1) different developmental stages, 2) leaf developmental gradient and 3) diurnal cycles (Additional file 7).

Expression pattern of TFs at different developmental stages
To explore the regulation of TFs on SsSWEET13c at different developmental stages, we performed RNA-seq at the seedling and mature stages, including the seedling leaf, seedling stem, mature leaf, and mature stems (from top to bottom: stem 3, stem 6 and stem 9) (Fig. 3). SsS-WEET13c displayed 3.6-fold enhanced expression in the seedling leaf compared with that in the lowest tissue of mature leaf,. Expressional levels in the stem 3, 6 and 9 gradually increased by 20.81-fold change expression, which implies SsSWEET13c could be associated with leaf development and secondary cell wall cellulose synthesis. Among these TFs, SsHB36 was mainly expressed in the leaves which was 3.92-fold higher than that in the stems, SsERF4 was mainly expressed in stem 3 of mature stages which is 3.84-fold higher than its expression in the leaves and 3.46-fold higher than that in the stem 9. SsbZIP44 was mainly expressed in the stem. and it increased 3.12-fold and 4.90-fold expression in seedling stem and mature stem compared with that in seedling leaves and mature leaves individually (Fig. 3). SsRAP2.4, SsbHLH34, SsILR3 and SsbZIP23 exhibited a high expression level in both seedling and mature leaves.

Expression pattern of TFs at leaf developmental gradient
To investigate the regulation of TFs on SsSWEET13c expressions in the leaves of source tissues, the transcriptome analyses and expression profiles were performed during the leaf developmental gradient (Fig. 4a). Corresponding to the maize leaf developmental gradient [25], we divided the second leaf at 11-day old with 15 cm length of S. spontaneum into 15 identical segments (each segment of 1 cm). Simultaneously, the S. spontaneum leaf was parted into four zones, namely a basal zone (base, the sink tissue), a transitional zone (6 cm, the sinksource transition), a maturing zone (10 cm), and a mature zone (tip, active C4 photosynthesis). Four out of 20 TFs, including SsABS5, SsERF4, SsTFIIIA-b and SsbZIP23 (Pearson Correlation, PC ≥|0.5|, p < 0.05) displayed negative expression patterns relative to those in SsSWEET13c across the leaf developmental gradient, and the gene expressional levels gradually decreased from the leaf base with low photosynthesis to the leaf tip with active photosynthesis (Fig. 4b, Additional file 8). Nine out of 26 TFs, including SsMYR2, SsWRKY18, SsNID1, SsRAP2.4, SsbHLH035, SsbHLH34, SsILR3, SsKUA1, SsTFIIIA-a and SsWRKY123 (PC ≥|0.5|, p < 0.05), displayed positive expression patterns relative to those in SsSWEET13c and the gene expression levels gradually increased from the foliar base to the tip (Fig. 4c). Expression patterns of SsMYR2, SsHB36, SsbZIP44, SsDEL1, SsMYBS1, SsNID1, SsRAP2.4, SsbHLH34 and SsKUA1 were verified by reverse transcription quantitative PCR (RT-qPCR) in three leaf segments from leaf gradients (Additional file 9).
One-to-one yeast hybridization verification of 12 potential TFs binding to the promoter of SsSWEET13c To verify the most significant TFs determined by Y1H and transcriptome analysis, constructs containing the TFs fragments were extracted from the yeast strains screened by the Y1H secondary library and proliferated via e.coli competent propagation. Then,constructs were transformed into yeast competent cells with pBait-AbAi bait vector containing the corresponding promoter fragments and cultivated them for 2 ~ 3d. The yeast solution was serially diluted (10 times, 100 times and 1,000 times) for spotting, and p53-AbAi was used as a positive control. One-to-one verification results of Y1H successfully confirmed 12 TFs, named SsDEL1, SsTFIIIA-b, SsHHO2, SsbHLH035, SsHB36, SsMYR2, SsbZIP44, SsABS5, Ssb-HLH34, SsRAP2.4, SsERF4 and SsTFIIIA-a. In addition, the SsWEET13c promoter was assessed for TF binding sites (TFBS) in F0, F1 and F3 to which TFs might bind (Fig. 2a). We have found 12 TFs with their predicted TFBS in the corresponding fragment, indicating that these 12 TFs could bind to the promoter and regulate the expression of SsSWEET13c (Fig. 2b).

Discussion
Sugarcane possesses complex genetic backgrounds due to its high polyploidy level. This high polyploidy is problematic due to unavailability of transgenic seedlings with characterized phenotypes. Therefore, in this study, we over-expressed SsSWEET13c in the model plant A. thaliana to identify the sugar transportability of SsSWEET13c. In S. spontaneum, SsSWEET13c dramatically increased from the leaf maturing to tip zones (Fig. 4b). A recent study has found that S. spontaneum SsSWEET13c homolog ZmSWEET13s played primary roles in sucrose transport in maize [21]and as a possible key player had been attributed to C4 photosynthesis [26]. Noticeably, ZmSWEET13s showed the highest expression in leaf tips [21]. We thus deduced that SWEET13c was mainly involved in photosynthesis for sugar transport. Sugars are produced in the leaves, and which are then transported to other organs and act as signals that play crucial roles in plant growth and development [27]. Relative to wild type, sugar content in leaves of overexpressed SsS-WEET13c A. thaliana discernibly decreased, while the root length/weight, shoot weight and the fresh weight increased significantly (Fig. 1). This may have been due to the enhanced sugar efflux from the leaf driven by SsS-WEET13c, and transported sugar to the root and shoot [16]. Our results indicated that SsSWEET13c increased the photosynthesis ability for the growth and development. There were similar conclusions in previous studies, where overexpression pear PbSWEET4 caused sugar reduction and early senescence in leaves [28]. Moreover, overexpression of A. thaliana AtSWEET16 exhibited increased growth efficiency [29]. Overexpression Grapevine VvSWEET4 was consistent with SsSWEET13c in A. thaliana, displaying enhanced root growth but distinct from those observed in SsSWEET13c in A. thaliana, e.g., higher glucose and fructose contents and higher radiolabeled glucose passive uptake in Grapevine [30].
Some crop yield depends on the efficient allocation of sucrose from leaves to seeds [21]. But the information on plant SWEET proteins acting as sugar transporters in seed development is limited. Seed of sweet mutants in rice has been shown to accumulated more sucrose, glucose or fructose but less starch [31]. This is in the agreement of our findings where we found that the overexpression of SsSWEET13c resulted in reduced sugar content. It was observed that SsSWEET13c in SWEETs clade III [23], while AtSWEET13 and AtSWEET14 were in clade III. Noticeably, sweet13;14 double mutants increased seedling and seed size [15], which was distinct from those observed in the triple zmsweet13 knock out mutants in maize. Moreover, OsSWEET11 and OsS-WEET15 in clade III are expressed preferentially in the caryopsis and played a key role in seed filling with sucrose efflux functions in rice [18,32]. Similarly, OsS-WEET11 is a sugar transporter with a key role in seed development [31]. Their knockout mutants plant of OsSWEET11 showed severely shrunken mature caryopses [18,32,33]. SWEET in the same clade may perform divergent functions in monocots and dicots. We hypothesized that overexpression of SsSWEET13c in A. thaliana might involve in increased seed size by sugar efflux and starch production. Thus, SsSWEET13c might be one of the important candidates for the high biomass of S. spontaneum.
TFs can bind to promoter regions of the target gene at conserved sites, in turn regulating different physiological activities in plants. Cis-element prediction revealed that 15 light-responsive sites existed in the upstream 2 kb promoter of SsSWEET13c ( Table 2). The prediction could be supported by the evidence of transcriptome dynamics and the expression patterns of TFs for the SsS-WEET13c. As the expression of SsSWEET13c increased from the leaf base (fewer lights) to the tip (more lights), the conjecture is that the leaf segments with sufficient light create abundant glycogen through photosynthesis. Thus, a higher level of sugar transporter SsSWEET13c is essential for assisting the sugar transfer in S. spontaneum leaves. During circadian rhythm, the expression level of SsSWEET13c shows a gradual decrease and increase, This led us to speculate that SsSWEET13c will transport the carbohydrates, produced at the day-time, to the other tissues and organs at night to provide the necessary glycogen for the growth and respiration in S. spontaneum. In addition, 20 TFs were identified as potential regulators for the SsSWEET13c based on the Y1H, and at least eight of them are in response to circadian rhythm. The verification of the Y1H self-activation experiment provided the optimal AbA concentration for the two screenings of the subsequent Y1H experiments and avoid false positives in the present experiment (Fig. 2b), Additionally 20 TFs were verified to be regulators for SsSWEET13c. Based on the expression patterns in the leaf segments (Fig. 4), four TFs show a negative regulation pattern, whereas ten TFs show a positive regulation pattern for SsSWEET13c expression. The negative TFs for the SsSWEET13c have a much lower expression level than the positive TFs (Fig. 4). Therefore, we suggested that the positive TFs are the main regulators regulating SsSWEET13c.
Based on a series of transcriptome analyses for potential TFs binding to the promoter of SsSWEET13c, 12 TFs (SsABS5, SsbHLH035, SsbHLH34, SsbZIP44, SsDEL1,  SsERF4, SsHB36, SsHHO2, SsMYR2, SsRAP2.4, SsTFI-IIA-a and SsTFIIIA-b) were further identified by oneto-one Y1H. The SsbHLH035 is the closest ortholog to the AT5G57150 in A. thaliana. The AT5G57150 is expressed in roots and shoot, and showed strong differences in expression levels between the shoot and root (root/shoot > 20-fold) [34]. In this study, overexpression of SsSWEET13c increased the root length in A. thaliana. In S. spontaneum, the expression pattern of SsbHLH035 along with the leaf gradient was similar to that of SsS-WEET13c (PC = 0.89, p < 0.0001) (Fig. 3). Therefore, the regulation of SsbHLH035 for SsSWEET13c may cause an increase in the root length. But further experiments are necessary for verification, such as detecting the expression of SsbHLH035 in the roots, and further confirming the effect of SsbHLH035 on SsSWEET13c for growth and development.
In S. spontaneum, the expression level of SsbZIP44 peaked at 16:00 h, and the diurnal expression pattern was negative to that of SsSWEET13c (PC = -0.61, p = 0.0059). The close homologous gene of SsbZIP44 is AtbZIP44 in A. thaliana. AtbZIP44 belonged to the S1 subgroup, which is translationally repressed by sucrose signaling [35]. Therefore, we assumed that SsbZIP44 in S. spontaneum responded to sucrose signals and displayed negative regulation patterns to SsSWEET13c.
In S. spontaneum, the expression level of SsRAP2.4 gradually increased from the leaf base to the tip and maximum at 8:00 h then decline at 16:00 h, It was found that expression level in the leaves being 1.62-3.70-fold higher than that in the stems (Figs. 3, 4, and 5). SsRAP2.4 is a close homolog to RAP2.4 in A. thaliana, which belongs to the DREB (dehydration-responsive element-binding proteins) subfamily of AP2/ERF proteins [36]. RAP2.4 is involved in senescence control, as well as ethylene and cytokinin-related developmental processes [37]. Overexpression of RAP2.4 results in root hair formation, hypocotyl and root elongation [37]. SsSWEET13c may also play a function associated with cytokinesis as the overexpression of the SsSWEET13c in plant increase the root length and the fresh weight. Therefore, SsRAP2.4 was assumed to play a role in promoting cytokinesis by regulating SsS-WEET13c in the S. spontaneum in different tissues.
SsHB36 TF is the closest ortholog of A. thaliana ATHB16, belonging to the HD-ZIP TF family. It is reported that ATHB16 mRNA was detected in all organs examined and acted as a growth regulator. However, reduced levels of ATHB16 expression in transgenic Arabidopsis plants caused an increase in leaf cell expansion [38]. In S. spontaneum, SsHB36 was highly expressed in the leaf base which was 2.40-fold higher than its expression in the leaf tip, and the expression level in the stem 3 was 3.48-fold higher than that in the stem 9. Leaf base and stem 3 are active parts for the growth and development in S. spontaneum (Figs. 3-4). Therefore, SsHB36 might be able to participate in shoot growth by regulating SsSWEET13c, and providing the necessary energy (sucrose) for cell elongation.
Overall, with regards to the function for SsSWEET13c and its potential TFs, based on the gene expression profiles we proposed a regulatory network for the SsS-WEET13c (Fig. 6). In S. spontaneum, SsSWEET13c played a role in the sucrose transportation from the source to the sink tissues and was mainly active in the mature leaf. In the source tissues, SsSWEET13c was mainly positively regulated by SsbHLH34, SsTFIIIA-a, SsMYR2, SsRAP2.4 and SsbHLH035, while negatively regulated by SsABS5, SsTFIIIA-b and SsERF4. During the circadian rhythm, SsSWEET13c was more active in the sunset which was likely due to the high level of sugar accumulation at night. While, SsSWEET13c was negatively regulated by SsbZIP44, and positively regulated by SsbHLH34. In the sink tissues, SsSWEET13c was also active for sugar accumulation, which was positively regulated by SsbZIP44, SsTFIIIA-b, SsbHLH34 and SsTFIIIA-a, and negatively regulated by SsERF4, SsHB36, SsDEL1 and SsABS5. Nevertheless, this speculation still needs to be verified by tissue-specific expression and localization of these genes. Furthermore, interactions between the TFs and SsSWEET13c promoter could be further verified by other approaches as well such as electrophoretic mobility shift assay (EMSA).

Conclusions
In conclusion, our study demonstrated that excessive SsSWEET13c will decrease sugar contents in leaves and boost the root length and the fresh weight at the same time in plants. Additionally, 20 TFs with the ability to bind to the SsSWEET13c promoter were identified through Y1H. Furthermore, our transcriptome analyses of TFs on developmental gradient leaf, distinct times during circadian cycles and stems/leaves at different growth stages generate potential expression patterns and regulatory network between these TFs and SsS-WEET13c. Together, these results help us to explain the underlying processes of SsSWEET13c and its impact on breeding and the generation of high-quality sugar and biomass.

Plant material and treatment
The founding Saccharum species, S. spontaneum (SES 208, 2n = 8x = 64, originated in the USA) were utilized in this study. The plant material was identified by Irvine JE [39]. The plants were cultivated in the greenhouse (14 h light 30 °C/10 h darkness 22 °C, 60% relative humidity), at the Center for Genomics and Biotechnology, Fujian Agriculture and Forestry University (FAFU, Fuzhou, China). The mature plants were planted in a field on the FAFU campus [40]. The collection of S. spontaneum and the performance of experimental research on such plant were complied with the national guidelines of China. We declared that S. spontaneum used for this study do not require application for special permissions.
The expression analyses for S. spontaneum, including developmental leaf segments, different times, leaves and stems at different periods, were performed as described in previous studies [23,41,42].
Seeds of A. thaliana Columbia ecotype were stored in an incubator (26 °C, 80% relative humidity) in the Center for Genomics and Biotechnology (FAFU, Fuzhou, China); transgenic lines 35Spro:SsSWEET13c were acquired by two consecutive screenings. All the seedlings were incubated on the MS medium at 4 °C for one day, then transferred to the greenhouse (24 °C, 16 h/8 h light/dark, and 80% relative humidity) for seven days. The seedlings were then transferred to nutrient soil (PINDSTRUP 0-10 mm) for growth (24 °C, 16 h/8 h light/dark, and 80% relative humidity).

RNA isolation, reverse transcription and cloning of SsSWEET13c
Total RNA was extracted using TRIzol reagent (Invitrogen, Shanghai, China). Acquired RNA quality was detected by the agarose gel electrophoresis and the NanoDrop 2000 spectrophotometer (keep OD260/ OD280 = 1.8 ~ 2.0, OD260/OD230 ≈ 2.0). Good quality RNA was reverse-transcribed to first strand cDNA using PrimeScript ™ 1st Strand cDNA Synthesis Kit (Takara, Dalian, China). The primers of SsWEET13c (Additional file 10) were designed via both Snap Gene Viewer and Primer Premier 5.0 and the PCR was performed with PrimeSTAR ® Max DNA Polymerase (Takara, Dalian, China).

Vector construction and plant transformation of SsSWEET13c
Using the Gateway method, the SsSWEET13c full-length cDNA was recombined into the pDONR207 vector by the Gateway entry clone. The pMDC140 vector was then used for the expression clone of SsSWEET13c and the CaMV35S promoter. Hygromycin B was the selectable marker employed. Constructs were transformed into Agrobacterium tumefaciens GV3101 and then introduced into WT Arabidopsis by the floral dip method [43]. At the T3 generation, homozygous transgenic lines were used for further assays.

Quantitative reverse transcription PCR (RT-qPCR)
RNA extraction and the first-strand cDNA synthesis were performed in accordance with the above methods. Integrated DNA Technologies (http:// www. idtdna. com/ Prime rquest/ Home/ Index) were used to design genespecific primer pairs (Additional file 11). SYBR Green method was utilized for RT-qPCR. The product was purchased from the company of GenStar (A310-02). The 25S ribosomal RNA and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were taken as housekeeping genes. The reaction system as follows: 10µL of 2 × Real-Star Green Fast Mixture, 0.5 µL of 50 ng/µL cDNA template, 0.5 µL of 10 mM forward primer, 0.5 µL of 10 mM reverse primer 0.5 µL, 8.5 µL of RNase-free H2O. BioRad CFX Connect Real-Time system was used with the following program: 2 min of denaturation at 95 ℃, followed by 40 cycles of 15 s at 95 ℃, 15 s at 60 ℃ and 30 s at 72 ℃. The relative expression levels for SsSWEET13c gene were calculated using the 2-ΔΔCt method [44].

DNA isolation, sequence analysis and molecular cloning of the promoter
Total genomic DNA was extracted by CTAB method. The upstream 2,000 bp promoter sequence of SsSWEET13c was selected for the cis-regulatory element online prediction via PlantCare [45]. The promoter sequence was divided into 5 shorter sequences to amplify them from the sugarcane genomic DNA.

Measurement of sugar, root length and fresh weight content
Measurement kits for glucose, sucrose fructose and the total soluble sugar contents were purchased from Nanjing Jiancheng Bioengineering Research Institute (Nanjing, China). The glucose content was determined by the glucose oxidase method [46]. Glucose measurement kit includes standard solution, calibration solution and glucose solution. For calculation of experimental results: Fructose content (mg/mL) = (Absorbance of sample mixed solution-Absorbance of empty)/ (Absorbance of sample mixed solution-Absorbance of empty) * Concentration of standard solution (mg/mL) *fold of dilution. The total soluble sugar content was defined using the colored reaction between sugar and enthrone. Measurement kit includes 100 µg/mL standard solution and substrate solution, and concentrated sulfuric acid didn't supply. 200 µL sample and standard solution were mixed with 100 µL substrate solution and 1 mL concentrated sulfuric acid and boiled for 10 min. Absorbance was observed under O.D. value of 620 nm. For calculation of experimental results: The total soluble sugar content (µg/g) = (Absorbance of sample mixed solution-Absorbance of empty)/ (Absorbance of sample mixed solution-Absorbance of empty) * Concentration of standard solution (mg/mL)/fresh weight of samples (g)/ volume of water * fold of dilution.
The root length of the plants was measured using a ruler, and then the average root length was calculated. Fresh weight was measured with a balance and the average weight was calculated.

Yeast One-Hybrid
For one-Hybrid Library: The sugarcane cDNA yeast library materials were collected from the 60-day-old (seedling stage) S. spontaneum (SES 208), including mature leaf 1, roll leaf 1, stem 3, stem 6, stem 9, root and flower. The total RNA was extracted and the cDNA fragment was synthesized by reverse transcription after mRNA isolation. Using the Gateway method, cDNA was recombined into the pDONR222 vector by Gateway entry clone (BP reactions) to generate the primary cDNA library. The pGADT7 vector was then used for the expression clone (LR recombination reaction) to obtain the secondary library required for promoter binding examination. The yeast library was constructed by Ouyi Biomedical Technology Co., Ltd (Shanghai, China).
The procedures are followed by Matchmaker ® Gold Yeast One-Hybrid Library Screening System User Manual (Clontech, Cat. Nos. 630491, 630,466, 630,499). Briefly, the 1999 bp SsSWEET13c upstream region (starting from ATG) was divided into five fragments (F0, F1, F2, F3 and F4). Primers used to amplify these fragments are listed (Additional file 3). The isolated shorter fragments of SsSWEET13c promoter were inserted into pAbAi vectors to construct the bait vector. After confirmation by sequencing and reconstructed vectors were linearized with restriction enzymes and create a bait/reporter strain by integrating the pBait-AbAi plasmid into the yeast competent cells Y1HGold yeast genome. Growth on the dropout medium (SD) lacking Uracil (-Ura) with AbA enables the determination of the correct concentrations to be used for screening the cDNA library. Construction and screening were conducted using a cDNA library by cotransformation and in vivo homologous recombination, grown on SD lacking leucine (-Leu) with AbA. The positive colonies were sequenced using the universal primer of pGADT7-F/R and then blasted against NCBI and PlantTFDB datasets.

Analysis of the expression profiles of SsSWEET13c and potential TFs
The cDNA libraries for each sample were established according to the manufacturer's protocol from Illumina TruSeq ® RNA. Applied 100-bp paired-end sequencing via Illumina HiSeq2500 equipment at the Center for Genomics and Biotechnology (FAFU, Fuzhou, China). The raw data were aligned on the sorghum gene models using TRINITY (https:// github. com/ trini tyrna seq/ trini tyrna seq/ wiki) and the transcripts per million (TPM) value was calculated via the RSEM method.