Genome-wide analysis of sugar transporter genes in maize (Zea mays L.): identification, characterization and their expression profiles during kernel development

Plant materials and growth condition

The maize inbred line B73 was used in this study. B73 seeds were sterilized with mercuric chloride and cultured in ddH₂O at 28 °C (light)/23 °C (dark) with a 16 h light/8 h dark photoperiod (Li et al., 2013). After germination, the seedlings with uniformed growth were selected and moved into the field. Subsequently, various tissues and kernels at different days after pollination were collected for the analysis of ZmST expression levels.

Identification and characterization of ST proteins in maize

To investigate putative ST genes in Zea mays, two methods were employed. First, 62 AtST sequences in Arabidopsis were downloaded and used to perform a BLAST search against the Zea mays genome obtained from maizeGDB (https://www.maizegdb.org/) with default parameters (Long et al., 2021). Additionally, the Hidden Markov Model (HMM) profiles of the Sugar_tr domain (PF00083), MFS-1 (PF07690) and MFS-2 (PF13347) were obtained from Pfam (http://pfam.xfam.org/) and utilized for HMMER 3.0 searches against the potential ST proteins in maizeGDB (Prakash et al., 2017; Mistry et al., 2021). All potential ZmST proteins were determined on NCBI (https://www.ncbi.nlm.nih.gov/cdd/) and SMART (https://smart.embl.de/).

Comparison of the numbers of ST gene families in different plants

ST genes from various plants, including Arabidopsis (Büttner, 2007), rice (Deng et al., 2019), tomato (Reuscher et al., 2014), pear (Li et al., 2015), strawberry (Liu et al., 2020), grape (Afoufa-Bastien et al., 2010), Longan (Fang et al., 2020), and apple (Wei et al., 2014) were analyzed to compare the number of STs across different plant species.

Chromosomal location, collinearity and duplication event analyses

The chromosomal locations of ZmST genes on chromosomes and chromosome synteny were performed by TBtools (Chen et al., 2020). Gene duplication analyses were conducted as previously described. The ratio of non-synonymous substitution rate (Ka) to synonymous substitution rate (Ks) was calculated by TBtools.

Phylogenetic tree analysis of STs from different plants

Amino acid sequences of STs from Zea mays, Arabidopsis thaliana and Oryza sativa were used to create a phylogenetic tree. The phylogenetic tree was constructed by MEGA 7.0 software using neighbor-joining (NJ) phylogenetic method with 1,000 bootstrap replications (Kumar, Stecher & Tamura, 2016).

Gene structure, conserved motif and domain analyses

The gene structure of ZmST genes was analyzed by TBtools software. The conserved motifs of ZmST proteins were analyzed with MEME (https://meme-suite.org/meme/) (Bailey et al., 2009). The maximum number of predicted motifs was set to 15. The final graph was presented by TBtools.

Expression heatmap of transcriptome

RNA-Seq datasets from different tissues were acquired from maizeGDB to analyze the expression profiles of the ZmST genes (Stelpflug et al., 2015). Ten tissues from maize vegetative development to reproductive development stages were used to identify tissue specificity of ZmST genes. The expression data of STs was visualized using the TBtools.

RNA extraction and qRT-PCR

RNAs from B73 materials were extracted by RNAprep Pure Plant Kit (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer’s instruction. About 1-2 µg of RNA was using to reverse transcribe with HiScript® III All-in-one RT SuperMix reverse kit reagents (Vazyme Biotech Co., Ltd.). In order to eraser gDNA and synthesize cDNA, 4 µl of 5 × All-in-one RT SuperMix, 1 µl of Enzyme Mix, 1–2 µg of RNA and appropriate volume of RNase-free ddH₂O were mixed. Subsequently, the mixture was incubated at 50 °C for 15 min, followed by a temperature increase to 85 °C for 5 s. The qRT-PCR was performed with primers listed in Table S1, with ZmACTIN1 as an internal reference. qPCR was run on CFX96TM real-time system (Bio-Rad, Hercules, CA, USA), with ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd., Nanjing, China), as previously described (Fang et al., 2023). Finally, the calculation method for ZmST genes expression levels was proposed by Livak and Schmittgen (Livak & Schmittgen, 2001).

Cis-acting regulatory elements analysis in the ZmST gene promoters

The promoter regions of ZmSTs were obtained with TBtools software for promoter analysis. The cis-acting regulatory elements were identified by PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) and presented with TBtools.

Sixty-eight ZmST genes are identified in maize genome

In this study, a total of 68 gene sequences encoding putative ST proteins were identified. Their physicochemical properties, including gene ID, protein size, molecular weight (MW), isoelectric point (p I), the grand averages of hydropathicity (GRAVY), and localization prediction, were characterized (Table 1). The molecular weight of ZmST proteins ranged from 41.83 kDa (ZmSFP1) to 80.96 kDa (ZmTST3), while the isoelectric points ranged from 4.72 (ZmTST3) to 9.82 (ZmSTP12) (Table 1). The grand averages of hydropathicity for all ST proteins indicated their hydrophobic nature. Subcellular localization analysis revealed that all ZmST proteins were located in the cell membrane (Table 1, Table S2).

ZmST proteins are divided into eight groups

A neighbor-joining tree with 199 STs, including 68 ZmSTs from maize, 62 AtSTs from Arabidopsis, and 69 OsSTs from rice, was constructed (Fig. 1). The phylogenetic tree suggested that the sugar transporters in maize were classified into eight groups. Among them, the VGT clade consisted of ZmVGT1 and ZmVGT2, while the STP clade contained ZmSTP1 to ZmSTP20. Additionally, the PMT, SFP, SUT, INT, pGlcT and TST clades included 16, 11, 7, 4, 4, and 4 members, respectively, and were annotated as ZmPMT1 to ZmPMT16, ZmSFP1 to ZmSFP11, ZmSUT1 to ZmSUT7, ZmINT1 to ZmINT4, ZmpGlcT1 to ZmpGlcT4, ZmTST1 to ZmTST4 (Fig. 1). Furthermore, phylogenetic analysis showed that there were some closely related orthologous STs between maize and rice, implying the existence of a set of ancestral ST genes before the divergence of the two species (Fig. 1).

Additionally, a comparison of the numbers of different ST groups among maize, Arabidopsis, rice, tomato, pear, woodland strawberry, grape, longan, and apple was carried out (Table 2). The results revealed that STP and PMT were the largest clades in maize, consistent with previous findings in rice, pear, and apple. Conversely, in Arabidopsis, tomato, strawberry, grape, and longan, the largest clades were STP and SFP.

Segmental duplication events are observed in ZmSTs

To investigate features of the ZmSTs gene family, we analyzed the chromosome distribution of each ZmST gene. Our findings revealed that the ZmST genes were located on all 10 chromosomes of maize (Fig. 2A). Chromosome 1 had the highest number of ZmST genes, including ZmPMT1, ZmPMT2, ZmPMT5, ZmPMT7, ZmSFP1, ZmSFP2, ZmSTP2, ZmSTP3, ZmSTP8, ZmSTP13, ZmTST1, ZmSUT1, and ZmSUT3. Chromosome 2 contained ZmPMT3, ZmPMT4, ZmPMT8, ZmPMT10, ZmSTP11, ZmSTP15, and ZmSTP20. ZmpGlcT1, ZmSFP3, ZmSTP10, and ZmSUT2 were located on chromosome 3. Chromosome 4 harbored ZmpGlcT4, ZmPMT14, ZmPMT15, ZmPMT16, ZmSTP5, ZmSTP6, ZmSTP9, ZmTST4, and ZmSUT6. Chromosome 5 contained ZmSTP18, ZmTST2, ZmTST3, ZmVGT1, ZmVGT2, ZmSUT4, and ZmSUT5. Chromosome 6 had ZmSFP6, ZmSFP7, and ZmSFP8 genes, while chromosome 7 contained ZmINT2, ZmINT3, ZmpGlcT3, ZmPMT6, ZmPMT9, ZmPMT11, ZmSTP1, ZmSTP4, ZmSTP14, ZmSTP16, and ZmSTP17. Chromosome 8 harbored ZmpGlcT2, ZmSFP4, ZmSFP5, ZmSFP9, ZmSFP10, and ZmSFP11. Chromosome 9 contained ZmSTP7, ZmSTP12, and ZmSU7, while chromosome 10 harbored ZmINT1, ZmINT4, ZmPMT12, ZmPMT13, and ZmSTP19 (Fig. 2A). Notably, all members of the ZmVGT group were located on chromosome 5, and four members of the ZmINT group were evenly distributed on chromosomes 7 and 10. Chromosome 6 only contained three ZmSFP members. Aside from these observations, the distribution of other ZmST genes on maize chromosomes was uneven.

Gene duplications are important for the expansion of gene families (Cannon et al., 2004; Konrad et al., 2011). Collinear analysis showed that 15 gene pairs, ZmpGlcT1 and ZmpGlcT2, ZmPMT7 and ZmPMT9, ZmPMT7 and ZmPMT10, ZmPMT8 and ZmPMT9, ZmSFP3 and ZmSFP4, ZmSFP6 and ZmSFP9, ZmSTP5 and ZmSTP18, ZmSTP5 and ZmSTP19, ZmSTP5 and ZmSTP20, ZmSTP18 and ZmSTP19, ZmSTP18 and ZmSTP20, ZmSTP19 and ZmSTP20, ZmSUT1 and ZmSUT3, ZmSUT1 and ZmSUT7, ZmSUT3 and ZmSUT7, have undergone segmental duplication events (Fig. 2B). Further analysis revealed that all the Ka/Ks values of the ST gene pairs were less than 1, indicating that the duplication events occurred under purifying selection (Fig. 2B, Table 3).

ZmST gene structures are highly conserved

The gene structures play crucial roles in the evolution and functional diversification of multiple gene families (Lei et al., 2020). The structural analysis of ZmSTs indicated that all ZmST genes, except ZmSTP6, harbored at least one intron. Several genes, namely ZmSFP1, ZmSFP2, ZmSFP3, ZmSFP4, ZmSFP5, ZmSFP6, ZmSFP7, ZmSFP8, ZmSFP9, ZmSFP10, ZmSFP11, ZmpGlcT1, ZmpGlcT2, ZmpGlcT3, ZmpGlcT4, ZmVGT1, ZmVGT2, ZmSUT1, ZmSUT4 and ZmSUT7 contained more than ten introns. However, genes within the same group usually had a similar number of exons (Figs. 3A and 3B). ZmSFP1 gene had 19 exons, whereas ZmSTP6 had only one exon, which indicated that the exons gain and loss occurred during the evolution of ZmST gene family. The structures of ZmSTs in duplication pairs, such as ZmpGlcT1 and ZmpGlcT2, ZmPMT7 and ZmPMT10, ZmSFP3 and ZmSFP4, ZmSTP5 and ZmSTP19, ZmSTP5 and ZmSTP20, ZmSUT1 and ZmSUT7, were highly similar.

We used the MEME online tool to predict the potentially conserved motifs of 68 ZmSTs. Among the 15 distinct motifs identified, only motif 6 was present in all ZmST proteins (Figs. 3C and 3D, Table S3). Notably, significant differences were observed in the conserved motifs between the SUT subfamily and MST subfamilies, despite their similar function in sugar transport (Figs. 3C and 3D). Motifs 1, 2, 4 and 5 were present in all 61 MST proteins, but were absent in SUT proteins, while motifs 13 and 15 existed in all 7 SUT proteins but not in MST proteins. These findings indicated that motifs 1, 2, 4 and 5 may be critical for the function of MST subfamilies, while motifs 13 and 15 may be necessary for the function of SUT subfamily. This may be due to functional differences between MST subfamilies (monosaccharide transport) and SUT subfamily (sucrose transport). Although all MST proteins contained motifs 1, 2, 4 and 5, the specific types and numbers of motifs varied among each subfamily. Motif 3 was absent in VGT subfamily and ZmSTP3 protein. Motif 7 was not present in any members of the SFP subfamily but was observed in the VGT subfamily. Motif 12 was absent only in ZmINT3 and ZmpGlcT4. Motif 14 was only observed in STP subfamily. Similar motif structures were observed in gene pairs such as ZmpGlcT1 and ZmpGlcT2, ZmPMT7 and ZmPMT9, ZmPMT7 and ZmPMT10, ZmPMT8 and ZmPMT9, ZmSFP3 and ZmSFP4, ZmSFP6 and ZmSFP9, ZmSTP5 and ZmSTP18, ZmSTP5 and ZmSTP20, ZmSTP18 and ZmSTP20, as well as ZmSUT1 and ZmSUT7.

ZmST proteins were predicted to form a compact helix bundle

In order to explore the potential roles of ZmST proteins, we predicted their conserved domains and 3D models of all ZmSTs with NCBI-CDD and Swiss-model, respectively. All ZmST proteins contained a conserved MFS domain, which facilitated the transportation of various substrates (including sugars, ions, nucleosides, amino acids and so on) through the cytoplasm or inner membrane, except SUT clade. This SUT clade comprised a conserved GPH_sucrose superfamily domain, which might export sucrose from photosynthetic sources to the phloem or import sucrose into sucrose sinks (Fig. 4). Furthermore, 3D prediction demonstrated that all the ZmSTs were folded into 8–13 transmembrane domains, and then formed a compact helix bundle (Figs. 5, S1). However, it’s worth noting that most members within the same group exhibited a similar 3D structure. For instance, the ZmTST subfamily members displayed a unique central loop, composed of approximately 320 amino acids, which connected to the predicted transmembrane domains, a feature absent in all other sugar transporters (Fig. 5).

ZmST promoters contain the cis-acting elements for light, phytohormone, stress and development

The regulatory cis-elements are the binding sites for transcription factors, carrying information to regulate the gene expression in biological pathways. Thus, we extracted the promoter regions of 68 ZmST genes and examined their cis-acting elements using PlantCARE database 5.0. A total of sixty-six cis-elements were identified and categorized into four main groups: photoresponse, hormonal response, stress response and development. Among these cis-elements, thirty were related to light-responsive pathway, indicating their role in responding to light, which aligned with the function of sugar transporters in the distribution of photosynthetic products. Additionally, fourteen were associated with hormone response, eleven with abiotic and biotic stress response, and eleven with plant growth and development (Fig. 6). All promoters of ZmSTs had the same number of CGTCA-motif and TGACG-motif, and most ZmSTs were regulated by abscisic acid and methyl-jasmonate. ZmSUT4 and ZmSFP9 had seven and nine CCGTCC-boxes in their promoters, respectively, indicating their potential involvement in meristem-specific activation. The GCN4_motif was present in the promoters of ZmINT2, ZmPMT8, ZmPMT16, ZmSFP2, ZmSFP6, ZmSFP8, ZmSTP4, ZmSTP6, ZmSTP9, ZmSTP17, ZmTST3, ZmSUT4 and ZmSUT5, suggesting their potential role in endosperm expression. The RY-element was found in the promoters of ZmINT2, ZmPMT8, ZmPMT16, ZmSFP2, ZmSFP6, ZmSFP8, ZmSTP4, ZmSTP6, ZmSTP9, ZmSTP17, ZmTST3, ZmSUT5 and ZmSUT7, indicating their potential involvement in seed-specific regulation. The most abundant cis-elements in ZmST promoter regions were G-box, ABRE and STRE, implying that ZmSTs may participate in maize’s growth, development and response to light, hormones and stress.

ZmSTs exhibit distinctive expression profiles across ten different tissues in maize

We downloaded the transcriptome data for ZmST genes and generated an expression pattern map using data from ten different tissues: young leaves, mature leaves, old leaves, roots, stems, tassels, cobs, embryos, endosperms, and seeds. Our analysis revealed that nineteen ZmST genes (ZmINT1, ZmpGlcT1, ZmpGlcT2, ZmpGlcT3, ZmpGlcT4, ZmPMT4, ZmPMT13, ZmSFP4, ZmSFP5, ZmSFP7, ZmSFP9, ZmSTP16, ZmTST1, ZmTST2, ZmTST4, ZmVGT1, ZmVGT2, ZmSUT2 and ZmSUT4) showed constitutive expression (log2 (FPKM+1) ≥ 1). Among these genes, ZmpGlcT1, ZmpGlcT2, ZmpGlcT4, ZmSFP5, ZmTST1, ZmTST2, ZmTST4, ZmSUT2 and ZmSUT4 showed the highest expression levels across all tested tissues and organs (log2(FPKM+1) ≥ 3) (Fig. 7). On the other hand, eleven ZmST genes (ZmINT3, ZmPMT14, ZmPMT15, ZmSFP2, ZmSFP3, ZmSTP6, ZmSTP9, ZmSTP17, ZmSTP18, ZmSTP19 and ZmSUT3) showed no expression across all tested tissues and organs (Fig. 7). Furthermore, we observed that two ZmST genes (ZmPMT1 and ZmPMT7), four genes (ZmPMT3, ZmPMT6, ZmSTP5, and ZmSTP14), and four genes (ZmSTP10, ZmSTP11, ZmTST3, and ZmSUT6) exhibited specific expression only in old leaves, roots, and tassels, respectively, with hardly detected expression in other organs. Additionally, we found that ZmSTP and ZmPMT groups displayed high expression in vegetative organs, whereas ZmPMT group members were scarcely detected in developing seeds. However, ZmTST group members showed higher transcription level in the developing seeds (Fig. 7).

Most sugar transporter genes showed a notable upregulation during the early stage of grain filling

As shown in Fig. 7, the diverse expression patterns of ZmST genes during stages of embryo, endosperm, and seed development suggested a significant role of ZmSTs in maize kernel development. To explore the probable functions of ZmST genes, we randomly selected 24 ZmST genes representing eight groups and executed qRT-PCR analysis across embryo and endosperm developmental stages as well as seed maturity stages. ZmpGlcT2, ZmSFP5, ZmSTP3, ZmTST1, ZmTST2, ZmVGT1, ZmVGT2 and ZmSUT2 showed up-regulation during embryo development and down-regulation during endosperm development. Conversely, ZmSTP7 were down-regulated during embryo development but up-regulated during endosperm development. Some genes like ZmINT4, ZmPMT4, ZmPMT8, ZmSFP10, ZmSTP15 and ZmSUT1 were down-regulated, while ZmSFP7 was up-regulated, during both embryo and endosperm development (Fig. 8). In particular, the expression levels of most tested ZmST genes, such as ZmINT4, ZmpGlcT4, ZmPMT4, ZmPMT5, ZmPMT8, ZmPMT9, ZmPMT13, ZmSFP5, ZmSFP7, ZmSFP9, ZmSFP10, ZmSTP7, ZmSTP15, ZmVGT1 and ZmSUT2, increased gradually in the early stage of grain filling, reaching a peak at 8-12 days after pollination, and then declined gradually (Fig. 9A). As supported by previous studies emphasizing the significance of the grain filling stage for seed quality and yield (Ji et al., 2022), we investigated the grain-filling rate of B73 over four days from 4 DAP to 28 DAP. Our findings revealed a substantial increase in the grain-filling rate during the developmental stages of 8-12 DAP and 16-20 DAP, aligning with the observed expression pattern of seed maturation (Fig. 9B). These results underlined the crucial role of ZmST genes in the embryonic, endospermic, and seed development stages.