Saccharum species, E. rufipilus and S. officinarum, studied as an experimental material. While, two species from the Andropogonae tribe, S. spontaneum and S. bicolor, are used as a reference species.

ST genes in E. rufipilus and S. officinarum were initially identified by performing BLASTP (with a cutoff e-value 1e⁻⁵) search with already reported ST genes from S. spontaneum (Zhang et al., 2018). Distinguished ST genes were further annotated using the CDD batch search (Marchler-Bauer et al., 2015). Then, the candidate ST genes were confirmed through the PFAM (with e value <1e⁻⁵) database (Mistry et al., 2021), using HMMER software v3.2.1.

ST gene sequences of E. rufipilus and S. officinarum were aligned with S. spontaneum and S. bicolor ST gene sequences using the maximum likelihood (ML) method. The ML tree was generated using MEGA, version 7.0, with a bootstrap value of 1,000 replicates and a “Poisson correction” model (Tamura et al., 2021). The results were then visualized in the interactive tree of life (iTOL) program for generating a phylogenetic tree (Letunic and Bork, 2019).

To analyze the conserved motifs among all the ST genes of E. rufipilus and S. officinarum, protein sequences were submitted to the online MEME (motif-based sequence analysis tools) suite 4.11.1 program (http://meme.nbcr.net/mem/cgi-bin/mem.cgi) (Bailey et al., 2015). The following parameters were adjusted: maximum number of motifs 20, minimum and maximum length between 15 and 60, number of repetitions, any. The results were then visualized in the TBtools program (Chen et al., 2018). Gene structures were identified using DNA sequences and exons of ST genes displayed using the Gene Structure Display Server (GSDS) (Hu et al., 2015). Molecular weight (MW) and isoelectric point (pI) of each ST gene were determined using the ExPASy online tool (http://web.expasy.org/compute_pi/). Subcellular localization of ST proteins was predicted through WoLF PSORT (http://wolfpsort.hgc.jp) (Horton et al., 2007). The transmembrane domain from the amino acid sequences of ST proteins were predicted using TMHMM Server v.2.0 (http://www.cbs.dtu.dk/services/TMHMM/).

The promoter region upstream of 2500 bp of transcription start site of each ST gene was extracted using TBtools. cis-regulatory elements (CRE) located in ST promoters were predicted using PlantCARE online database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/). The identified cis-regulatory elements were analyzed using the Simple Biosequence viewer function on TBtools.

ST gene positions on chromosomes were detected from the General Feature Format Files (GFF3), and karyotyping was executed in TBtools (Chen et al., 2018). The duplication pattern for ST genes and protein-coding genes was investigated by an all-vs-all local BLAST with an E-value <1e⁻⁵. BLAST results were imported into MCScanX (v0.8) software (http://chibba.pgml.uga.edu/mcscan2/). Whole Genome Duplication (WGD)/Segmental Duplication (SD), and Tandem Duplication (TD) were identified with default parameters (Wang et al., 2012). Orthologous ST genes between S. spontaneum, E. rufipilus, S. officinarum and S. bicolor were identified using the Dual Synteny Plot tool in TBtools. The coding sequences of ST gene pairs were subjected to synonymous (ks) and non-synonymous (ka) substitution ratio according to the Nei-Gojobori method (Kumar et al., 2016). If the ratio of Ka/Ks is greater than 1, it shows positive selection. When the value of Ka/Ks is equal to 1, it represents neutral selection. However, when the ratio of Ka/Ks is less than 1, it suggests negative or purifying selection. Additionally, we used the Ks values for estimation of duplication event time (T) in MYA (Million Years Ago). Here, T = Ks/(2 × 6.1 × 10⁻⁹) × 10⁻⁶ Mya (Gaut et al., 1996).

We utilized different samples from 4–6 months old E. rufipilus and S. officinarum (B-48) tissues, including mature leaf, seedling stem, pre-mature stem, and mature stem. Total RNA from these harvested samples was extracted using a Quick-RNA™ Miniprep kit (Zymo Research, USA), according to the manufacturer’s recommendations.

We used RNA-seq data for E. rufipilus and S. officinarum, which included transcript abundance [Transcript Per Million (TPM)] from various developmental stages from our previous studies. Additionally, we incorporated RNA-seq data for S. officinarum that analyzed circadian rhythms, gradients of leaf development, and the effects of different hormones such as abscisic acid (ABA), gibberellin (GA), indole acetic acid (IAA), and ethylene (ETH) (Zhang et al., 2018, 2021; Hua et al., 2022; Wang et al., 2023).

Trimmed reads were aligned with the reference gene model of the S. spontaneum genome by Trinity with default settings (Grabherr et al., 2011). Transcript assembly was performed by Stringtie (Pertea et al., 2015), and TPM values were estimated by RSEM method (Li and Dewey, 2011).

E. rufipilus and S. officinarum stalks were grown in a greenhouse under growth conditions of 16h 30°C, 8h/22°C, and a relative humidity of 75%. Three sugar treatments (1% Sucrose, 1% Glucose, and 1% Fructose) were applied uniformly to the 6 months old seedlings. Each experiment was performed in triplicate. Leaf samples were collected after 8h of treatments and stored at −80°C for further processing.

For quantitative RT-qPCR, about ≤1µg RNA was obtained from seedling stem, pre-mature stem, mature stem, and mature leaf samples. 5xPrimescript RT master mix (Takara Bio) was used for cDNA synthesis. RT-qPCR reaction comprising of the following reaction mixture in 20 µL solution: 10 µL of Master Mix (SYBR Green; Roche, Germany), 1 µL of template cDNA, 1 µL of forward and reverse primers each (10 µM), and water up to the final volume. Thermal cycling was as follows: initial denaturation at 95°C for 10 min, subjected by 40 cycles of denaturation at 95°C for 15 s, further annealing, and extension at 60°C for 60 s. For normalization of expression data, Actin and eEF-1a were used as reference genes. Each experiment was performed in three technical replicates. The relative expression level of each sample was calculated with the 2^-ΔΔCq method, as mentioned earlier (Akbar et al., 2021). A list of primers is mentioned in Supplementary Table S1 .

Full length CDS (SUT1-T1) without stop codon was amplified from S. officinarum cDNA ( Supplementary Table S1 ). Amplified fragment of SUT1-T1 was inserted into pCAMBIA1300-GFP vector. The vectors were then transformed into the Agrobacterium tumefaciens GV3101 strain. The bacterial suspensions with OD600 = 0.5 were inoculated into Nicotiana benthamiana leaves. FM™4-64 (N-(3-Triethylammoniumpropyl) was used as fluorescent dye. After 2–3 days of inoculation, leaf portions were excised from the injected area and observed under GFP fluorescence signals using a laser confocal microscope (LEICA TCS SP8).

Yeast One-hybrid Assay (Y1H) screening was conducted using the Matchmaker Gold Yeast One-Hybrid Screening System (Clontech, 630489). The targeted sequences (bait) were cloned into the pAbAi vector. SUT1-T1 promoter region (2000 bp upstream region from start codon) was amplified in four fragments [−1 to −500 (cis-1), −501 to −1000 (cis-2), −1001 to −1500 (cis-3), −1501 to −2000 (cis-4)] and cloned into pAbAi vector. The pAbAi-bait plasmids were then linearized with BstB1 restriction enzyme and transformed into Y1H Gold yeast-competent cells. The colonies were screened on synthetic dextrose medium lacking uracil. The bait strains were then tested for Aureobasidin A (AbA) resistance, and the minimal inhibitory concentration of AbA was determined for bait strains. A cDNA library (prey) was generated by Ouyi Biomedical Technology Co., Ltd. (Shangai, China). The AD-prey vectors were co-transformed with bait-pAbAi plasmids and screened on SD/-Leu/AbA media. Potential binding partners were confirmed through sequencing. TFs were predicted through the Plant transcription factor database (PlantTFDB) (https://planttfdb.gao-lab.org/prediction.php). The primers used in Y1H assay are mentioned in Supplementary Table S1 .

For the one-to-one interaction of SUT1-T1 with LSD and NAC TFs, we extracted yeast plasmids from colonies confirmed by sequencing. The colonies were grown in YPDA medium, and plasmids were extracted using a yeast plasmid extraction kit (Solarbio, D1160). Due to the low copy number of yeast cells, we initially transformed 5 μL of the extracted yeast plasmids into Escherichia coli (E.coli) DH5α competent cells. Next, we transformed 100 ng of prey plasmid into yeast strain containing the corresponding bait plasmid to verify one-to-one interaction. The growth of the transformed yeast was analyzed on the SD/-Leu/AbA* selection medium.

A genome-wide search was performed using E. rufipilus and S. officinarum amino acid sequences as a query with BLASTp analysis, against the previously published ST genes in S. spontaneum (Zhang et al., 2021; Xiao et al., 2022). Query sequences were further confirmed through batch CDD search and PFAM. Subsequently, we identified 78 reliable ST genes from E. rufipilus and 86 ST genes from S. officinarum. These ST genes were further divided into eight subfamilies, including seven monosaccharide transporter families (VGT, INT, SFP, PLT, STP, MST, and pGlcT) and one sucrose transporter family (SUT). Consistent with earlier findings, phylogenetic analysis distributed the ST genes into eight distinct subfamilies ( Figure 1 ) (Zhang et al., 2021; Xiao et al., 2022). All these ST genes were numbered according to their designated position in the evolutionary tree, with respective of S. spontaneum and S. bicolor ST genes. Using S. spontaneum ST genes as a reference, phylogenetic tree distinguished the 4 members in INT, 5 members each in MST and SUT, 2 members in pGlcT and VGT each, 25 members in PLT, 8 members in SFP, and E. rufipilus genome. In S. officinarum genome, 4 members were found in INT and VGT each, 7 members each in MST and SUT, 3 members in pGlcT, 38 members in PLT, and 8 members in SFP. Noticeably, the highest number of ST genes were found in the STP subfamily in E. rufipilus, which were 27 in numbers while in S. officinarum PLT subfamily contained the highest gene numbers, with 38 members.

ST gene families in E. rufipilus encode proteins with a length ranging from 348 amino acids (Erufi. 02G037360-PLT23) to 761 amino acids (Erufi.10G030260-MST2). Their molecular weight ranges from 37.14 KDa (Erufi.02G037360-PLT23) to 80.91 KDa (Erufi.10G030260-MST2). Isoelectric point (pI) predicted in the range of 4.74 pH (Erufi.10G030260-MST2) to 10 pH (Erufi.09G014150-STP27). Several trans-membrane domains are found to be between 6 and 12. Subcellular localization predicted by WoLF PSORT identified 40 ST genes located in plasmalemma, 27 in vacuole, 6 in chloroplast, and 5 in cytoplasm ( Supplementary Table S2A ). In S. officinarum, ST proteins range in length from 310 amino acids (Soffic.09G0019430-6P-SFP7) to 760 amino acids (Soffic.06G0004520-3E-MST2). Their molecular weight ranges from 33.109 KDa (Soffic.09G0019430-6P-SFP7) to 80.687 KDa (Soffic.06G0004520-3E-MST2). Iso-electric point (pI) is in the range of 4.76 pH (Soffic.06G0004520-3E-MST2) to 10.3 pH (LAp.01H0033380-SFP4). The number of transmembrane domain ranges between 6 and 12. Subcellular localization detected 45 ST genes in plasmalemma, 27 in vacuole, 7 in chloroplast, and 7 in cytoplasm ( Supplementary Table S2B ).

Each member of the ST gene family has a unique sequence and distinct structural features. We identified conserved motifs to comprehend the structural diversity and evolutionary relationship. Importantly, few motifs were detected only in an individual family, which depicts that they are associated with specific functions. We detected 15 conserved motifs in ST genes of both species using the MEME suite ( Figures 2A, B ). In E. rufipilus and S. officinarum, commonly found motif at the N terminal is motif 7, except in SUT, while at C-terminal motif 2 is most common. Evidently, in the monosaccharide family number, the specificity of conserved motifs is highly comparable. However, some specific motifs are present or absent in each subfamily. Moreover, the number of conserved motifs in STP genes is 13 in both E. rufipilus and S. officinarum. Similarly, in the PLT subfamily conserved motifs are also found to be 13 in number in most of ST genes in both species. Furthermore, the number of conserved motifs in SUT genes ranges from 3 to 6 in E. rufipilus and 3 to 5 in S. officinarum. Conserved motifs depicted the origination of subfamilies from a common ancestor. However, slight differences showed the specificity and functional divergence of each subfamily.

Next, we detected seven superfamilies (MFS superfamily, MFS-GLUT10-12_Class3_like, MFS_GLUT_like, MFS_GLUT6_8_Class3-like, MFS_HMIT_like, MFS_STP, GPH_sucrose superfamily) in ST genes ( Figures 2C, D ). Each subfamily possesses a distinct superfamily. The PLT family was found to have MFS superfamily, VGT has MFS-GLUT10-12_Class3_like superfamily, MST also contained MFS superfamily, pGlcT is composed of MFS_GLUT_like superfamily, SFP has MFS_GLUT6_8_Class3-like superfamily, and INT encodes MFS_HMIT_like superfamily. Moreover, STP contained MFS_STP and SUT composed of the GPH_sucrose superfamily. The occurrence of these superfamilies contributed to the uniqueness and specific functionality of ST genes.

Regarding ST gene conservation, gene structures were illustrated with Gene Structure Display Server (GSDS). Each gene family was observed to display conserved exon and intron positions relative to domain position and structure ( Figures 2E, F ). To comprehend the evolutionary relationships and characteristics of ST sequences, the cDNA sequence of each ST gene was aligned with the genomic sequence. The majority of ST genes comprises only exons. In E. rufipilus, a total of 23 ST genes contain only exons, while 55 genes are composed of exons and untranslated regions (UTRs) as well. The number of exons per gene ranges from 2 to 18. Among 55 genes containing UTRs, only 16 genes have 3′ UTR and 39 have both 5′ and 3′ UTRs. In S. officinarum, 43 ST genes composed of only exons, with the number of exons per gene ranging from one to as many as 20 exons per gene.

For identification of cis-regulatory elements in the region, upstream 2.5 kb sequences from the start codon of each gene were submitted to PlantCARE. Promoter regions from ST genes revealed a number of cis-acting elements related to different functions ( Supplementary Figures S1A, B ). The majority of identified cis-acting elements were grouped into cellular function, stress response, light response, and hormonal regulation categories ( Table 1 ). A number of cis elements are related to light responses such as ACE and AE box. Box4, GT1 motif, I-box, and Sp1 are present in both S. officinarum and E. rufipilus.

We investigated the chromosomal distribution of ST genes on 10 chromosomes of E. rufipilus and S. officinarum ( Figures 3A, B ). We found that there is no substantial co-relation between the number of genes and chromosome length. The majority of the genes were found to be concentrated on Chr01 and Chr02 in S. officinarum while on Chr02 in E. rufipilus. All genes were indiscriminately distributed on chromosomes, with most genes located at the lower telomere.

Genomic sequence duplications, including WGD, TD, and SD, provide a genetic link for evolution (Xiao et al., 2022). The ST gene family duplication event was performed by MCScanX. Duplication analysis identified dispersed duplication (DD), WGD, or SD events in both genomes. WGD or SDs was found to be the primary reason for ST gene family expansion ( Supplementary Tables S3A, B ). In E. rufipilus, a total of 52 genes displayed WGD, or SD, and 26 genes were found to be DD. However, in S. officinarum, 58 genes depicted WGD or SD and 29 genes showed SD. This indicates that WGD or SD might be the major driving force in the evolution of the ST gene family. Additionally, 25 ST gene pairs were identified in E. rufipilus while 24 gene pairs were found in S. officinarum. These gene pairs were distributed on all 10 chromosomes ( Supplementary Figure S2 ).

To further infer the evolutionary rate of ST genes, we calculated Ka/Ks (the ratio of synonymous and non-synonymous substitution rate). Ks values investigated positive (Darwinian) selection or negative (purifying) selection and duplication dates. ST gene pairs estimated to have undergone purifying selection. Duplication time was estimated to be in the range between 8.802 and 277.38 Mya in E. rufipilus and 0.182-90.444 Mya in S. officinarum ( Supplementary Tables S4A, B ). We performed collinearity analysis with S. spontaneum and S. bicolor to further identify homology between related species. All the ST genes belonging to E. rufipilus and S. officinarum showed collinearity with the syntenic region in S. spontaneum and S. bicolor ( Supplementary Figure S3 ).

To gain insights into the role of ST genes in sugar transport and mobilization, we analyzed the transcriptome profiles of ST gene expression in various tissues of E. rufipilus and S. officinarum. We found that the ST genes exhibited different expression patterns in different tissues. TPM values of selected ST genes were visualized using heatmaps. Specifically, we focused on the highly expressing genes in the mature leaf zone. In E. rufipilus, the following genes were up-regulated in the mature leaf zone: SUT2, STP18, STP15, STP4, STP17, STP1, STP8, PLT2, PLT7, PLT18, PLT15, PLT9, PLT8, PLT14, PLT22, PLT19, SFP3, INT1, INT2, and MST4 ( Figure 4A ). However, in S. officinarum, the undermentioned genes were enhanced in the mature leaf zone: SUT1-T1, SUT2, STP16, STP11-T1, STP17, STP2, STP1, STP13, STP8-2, PLT9, PLT18, PLT3-2, PLT17-T1, PLT7, PLT12-T2, PLT12-T1, PLT17, PLT12, SFP4-T1, INT3, VGT2, VGT3, VGT3-T1 ( Figure 4B ).

To verify the accuracy of the RNA seq data, we conducted RT-qPCR on different tissue samples from E. rufipilus and S. officinarum. The RT-qPCR results showed that SUT1 is highly expressed in the mature stem of E. rufipilus and mature leaf of S. officinarum. However, INT1 exhibits high expression in the pre-mature stem of E. rufipilus and mature leaf of S. officinarum. Additionally, in S. officinarum, there is higher expression of STP4 and PLT11 in pre-mature stem. In E. rufipilus, PLT15 and STP23 depicted enhanced expression in mature leaf. Conclusively, we observed consistent trends in the relative expression of selected ST genes and their corresponding FPKM values ( Supplementary Figure S4 ).

To better understand the role of ST genes, we treated the S. officinarum and E. rufipilus leaves with different sugar solutions and analyze their expression pattern through RT-qPCR. Expression data was obtained from 4 to 6 months old leaves of S. officinarum and E. rufipilus, treated with 1% solution of sucrose, glucose, and fructose. The expression level of ST genes showed variation during different sugar treatment ( Figure 5 ).

During 1% sucrose treatment, SUT1 and SUT2 showed enhanced expression in both E. rufipilus and S. officinarum leaves. Additionally, SFP expression was also found to be relatively higher in sucrose-treated S. officinarum leaves. Furthermore, the 1% glucose treatment regulates expression of MST, VGT, INT, and pGlcT gene levels in both studied species as compared to their wild-type counterparts. Interestingly, the 1% fructose treatment resulted in increased expression of MST, VGT, and INT in both E. rufipilus and S. officinarum. We also observed moderately higher expression levels of PLT, SFP, pGlcT, and STP in both species compared to their wild-type counterparts. Notably, the expression levels of SUT1 and SUT2 were remarkably low during the glucose and fructose treatments.

Based on the analysis of expression data, we identified SUT1-T1 with high expression in the mature leaf zone of S. officinarum. We selected this gene to investigate its subcellular location. Confocal microscopy determines that SUT1-T1 is located in the plasma membrane ( Figures 6A, B ). This finding is consistent with the prediction of the WoLF PSORT tool, which also identified the plasma membrane as the location of the SUT1-T1 gene.

In order to identify the potential TFs that bind to the SUT1-T1 promoter region, we divided the 2000 bp promoter region into smaller fragments of 500 bp ( Figure 6C ). These fragments were referred to as bait and were used to screen sugarcane cDNA libraries using the Y1H system. The activity of the bait fragments was repressed using 100 ng ml⁻¹ of AbA ( Figure 6D ). We determined that the minimal inhibitory concentration of AbA for screening the sugarcane cDNA library was 100 ng ml⁻¹. A total of 100 clones for each fragment were grown on SD/-Leu/+AbA (150 ng ml⁻¹) medium and screened for TFs prediction ( Figure 6E ). Of the total clones screened, only four showed potential TFs. According to Plant TFDB, NAC (Soffic.10G0023830-1A) and LSD (Soffic.07G0003300-4D) were identified as potential TFs ( Figure 6F ). Next, the one-to-one interaction between NAC and LSD TFs with the SUT-T1-cis2 promoter region confirms that both TFs interact with the gene promoter region ( Figure 6G ).

To investigate the expression patterns of TFs identified in the Y1H experiment, RNA-seq analyses were performed in S. officinarum, focusing on various developmental stages, diurnal cycles, and leaf development. We compared the expression patterns of NAC and LSD TFs with SUT1-T1 by analyzing the seedling leaves. Seedling leaves, measuring 15 cm in length, were divided into 15 equal segments. The expression of SUT1-T1 consistently increased from the basal to the mature zone. LSD exhibited a positive correlation with SUT1-T1, while NAC depicted a negative expression pattern ( Figure 7 ).

To investigate how TFs regulate SUT1-T1 at various developmental stages, we measured their transcript abundance in mature tissues, specifically in leaves and stems. Both TFs, NAC and LSD, exhibited higher expression levels in mature stem-internode 9. In contrast, SUT1-T1 was expressed in seedling leaves. This indicates that NAC and LSD TFs negatively regulate the SUT1-T1 gene ( Supplementary Figure S5A ). Next, we compared the expression pattern of TFs on SUT1-T1 during day-night rhythm in S. officinarum ( Supplementary Figure S5B ). SUT1-T1 showed higher expression levels during the daytime (from 6:00 to 10:00) and decreased expression for the reminder of day. LSD displayed enhanced expression in the morning (at 10:00) while NAC showed increased expression from 14:00 until midnight.

The expression patterns of NAC in the leaves and stems of S. officinarum were assessed following treatments with different hormones (IAA, ETH, GA, ABA) at various time points ( Supplementary Figure S5C ). NAC was found to be highly expressed in the leaves after ABA treatment and in the stem after 24h. However, NAC expression was higher in the stem at 48h and 96h after ETH treatment. GA increased NAC expression at all time points in the stem, while in the leaves, it was only elevated after 96h. IAA induced a strong NAC expression in the leaves after 48h of treatment. However, ETH and ABA had no effect on LSD expression, whereas GA and IAA enhanced LSD expression in all treated samples. SUT1-T1 expression is high in stems treated with ABA for 48 and 96 h. ETH increases SUT1-T1 expression in the 48h leaf, 24h stem, and 96h stem samples. GA enhances SUT1-T1 expression in the 24h and 96h leaf samples, as well as the 24h stem. Conversely, IAA enhanced the expression of SUT1-T1 in all treated leaf samples. Overall, there is no correlation observed in the expression patterns of NAC, LSD, and SUT1-T1 under various hormone treatments. Additionally, we have found that the expression pattern of the selected genes depicted a consistent trend between transcriptome data and RT-qPCR values ( Supplementary Figure S6 ).