Genome-Wide Identification and Expression Profiling of Monosaccharide Transporter Genes Associated with High Harvest Index Values in Rapeseed (Brassica napus L.)

Sugars are important throughout a plant’s lifecycle. Monosaccharide transporters (MST) are essential sugar transporters that have been identified in many plants, but little is known about the evolution or functions of MST genes in rapeseed (Brassica napus). In this study, we identified 175 MST genes in B. napus, 87 in Brassica oleracea, and 83 in Brassica rapa. These genes were separated into the sugar transport protein (STP), polyol transporter (PLT), vacuolar glucose transporter (VGT), tonoplast monosaccharide transporter (TMT), inositol transporter (INT), plastidic glucose transporter (pGlcT), and ERD6-like subfamilies, respectively. Phylogenetic and syntenic analysis indicated that gene redundancy and gene elimination have commonly occurred in Brassica species during polyploidization. Changes in exon-intron structures during evolution likely resulted in the differences in coding regions, expression patterns, and functions seen among BnMST genes. In total, 31 differentially expressed genes (DEGs) were identified through RNA-seq among materials with high and low harvest index (HI) values, which were divided into two categories based on the qRT-PCR results, expressed more highly in source or sink organs. We finally identified four genes, including BnSTP5, BnSTP13, BnPLT5, and BnERD6-like14, which might be involved in monosaccharide uptake or unloading and further affect the HI of rapeseed. These findings provide fundamental information about MST genes in Brassica and reveal the importance of BnMST genes to high HI in B. napus.


Introduction
Photosynthetic products play essential roles in plant growth and development. These products, particularly sugars (including polyols, monosaccharides, and sucrose), are synthesized in photosynthetic organs (source tissues) [1,2] and transported to heterotrophic cells (sink tissues) [3]. The transport and distribution of sugars are important for maintaining the balance between source and sink tissues [4,5]. Plants contain two major types of sugar transporters: sucrose transporters and monosaccharide transporters (MSTs) [6]. MSTs are important transmembrane transporters that have been identified in many land plants and that function in carbohydrate flux [7,8].
queries. The HMMsearch program (HMMER 3.0, http://hmmer.org/) was then used to further confirm the presence of protein domains using the AtMST Pfam numbers as queries.

Multiple Sequence Alignment and Phylogenetic Analysis
Multiple sequence alignment and phylogenetic analysis were performed to investigate the evolutionary relationships of the MSTs among the four species. Multiple sequence alignment was performed with MEGA 7.0 (Molecular Evolutionary Genetics Analysis) with default parameters. A phylogenetic tree was constructed using the neighbor-joining (NJ) method, with bootstrap analysis of 1000 replicates [46]. The tree was further visualized with FigureTree software.

Analysis of the Chromosomal Locations, Gene Structures, and Conserved Motifs of the BnMST Genes
Detailed information about the locations of the BnMST genes, their deduced protein sequences, and the relationships between Brassica and A. thaliana genes was obtained from the Brassica napus Genome Database. The chromosomal distribution of the genes was visualized with Map-Chart2.2 [47]. The gene structures were mapped using the Gene Structure Display Server (GSDS) [48]. The conserved motifs were analyzed using MEME online (http://meme-suite.org/tools/meme) [49,50]; all parameters were set to default settings, except the maximum number of predicted motifs, which was set to 20. In addition, we calculated the molecular weights (Mw) and isoelectric points (PI) of the BnMSTs using the ExPASy proteomics server database (https://www.expasy.org/tools/) [51]. The subcellular localizations of the proteins were predicted with MultiLoc2 (https://abi-services.informatik.unituebingen.de/multiloc2/webloc.cgi) [52].

RNA-seq Analysis
Sugar transporters may influence the HI through increased seed yields. To identify the DEGs from the set of 175 BnMST genes identified in high HI materials, we download and analyzed RNA-seq data for B. napus from the NCBI database (ID Number SRP072900). We extracted the transcriptome data for the 175 BnMST genes and constructed a heatmap using R-Studio. The data were generated from seven tissues (stems, mature leaves, buds on the main branch, seeds from the main branch, seeds from lateral branches, silique pericarps from the main branch, and silique pericarps from lateral branches) from four materials with a high or low HI and biological yield (BY) [32].

Plant Materials
To identify the expression patterns of the differentially expressed BnMST genes, which expressed differently for materials with a high and low HI, seeds from the materials CQ24 (SWU47, High HI) and CQ46 (Ning You 12, Low HI) were obtained from the Chongqing Rapeseed Technology Research Center, China. The plants were cultivated under field conditions in Chongqing for 2 years (2016 and 2017). Ten plants of rapeseed were planted in two experimental plots, and each plot contained five individual plants. After the mature stage, the two main characters, consisting of the dry weight of the above-ground biomass yield and seed yield per plant, were investigated. The harvest index was then calculated with the ratio of seed yield to above-ground biomass. Finally, we obtained the HI data of 10 individual plants from two replications respectively. The phenotypes of CQ24 and CQ46 are shown in Table 1. Samples of different tissues during various growth periods (stems at 0 and 30 days after flowering; leaves at the beginning of flowering; and flowers, buds, seeds, and silique pericarps at 7, 14, 21, 30, and 40 d after flowering) were collected from the two materials, immediately frozen in liquid nitrogen, and stored at −80 • C until use.
2.6. RNA Extraction and Validation of RNA-seq Data by qRT-PCR RNA was extracted from the samples using an RNeasy Extraction Kit (Invitrogen, Carlsbad, CA, USA) and employed as a template to produce cDNA with a Reverse Transcription Kit (TaKaRa Biotechnology, Dalian, China). As described by Qu [53], quantitative reverse-transcription PCR (qRT-PCR) was performed to identify the expression patterns of the DEGs identified by RNA-seq using specific primers designed with Premier 5.0 [54]. The primers, which were confirmed using BLAST online tools (BRAD, http://brassicadb.org/brad/blastPage.php), are shown in Table S1. The UBC21 gene was used as an endogenous reference gene [55]. The relative expression level of each gene was calculated using the 2 − Ct method [56]. The qRT-PCR was completed with three technical repetitions. The results were visualized with GraphPad Prism5.0 software [57,58].
To explore the evolutionary relationships of the MSTs, we performed a phylogenetic analysis. The 295 MST genes were divided into seven subfamilies in this phylogenetic tree. This result is consistent with previous findings in Arabidopsis [8]. We named the seven subfamilies based on the types of monosaccharides they transport [8]: the sugar transport protein (STP), vacuolar glucose transporter (VGT), tonoplast monosaccharide transporter (TMT), plastidic glucose transporter (pGlcT), polyol transporter (PLT), inositol transporter (INT) and ERD6-like subfamilies ( Figure 1).

Chromosome Locations of BnMST Genes and Duplication Analysis
The 175 BnMST genes are unevenly distributed on the B. napus chromosomes (Figure 2A), including 86 genes in the A subgenome and 89 in the C subgenome. Of these, 153 BnMST genes were identified on the 19 B. napus chromosomes, and the 22 remaining BnMST genes were mapped to pseudo-molecule chromosomes Ann and Cnn or to Unn (unknown). Chromosomes A01, A03, A05, A06, C01, and C05 each contain more than 10 BnMST genes, whereas chromosomes A02, A04, A10, C02, C04, C06, and C09 contain fewer than six BnMST genes ( Figure 2A). The syntenic relationships of the 175 BnMST genes are indicated by lines in Figure 2B. Sixty-nine homologous pairs were identified among the 175 BnMST genes ( Figure 2B). The syntenic relationships among the four species are shown in Figure 2C. The copy numbers of the MST genes in A. thaliana and B. napus ranged from one to eight (Table 2), indicating that some genes were lost or duplicated during evolution. For example, the syntenic homologs of AtERD6-like15, AtERD6-like16, and AtpGlcT4 have been lost in Brassica species. However, three homologous syntenic genes of AtERD6-like6 are present in B. rapa and B. oleracea and six are present in B. napus (three each in the A and C subgenomes) ( Figure 2C,

Chromosome Locations of BnMST Genes and Duplication Analysis
The 175 BnMST genes are unevenly distributed on the B. napus chromosomes (Figure 2A), including 86 genes in the A subgenome and 89 in the C subgenome. Of these, 153 BnMST genes were identified on the 19 B. napus chromosomes, and the 22 remaining BnMST genes were mapped to pseudo-molecule chromosomes Ann and Cnn or to Unn (unknown). Chromosomes A01, A03, A05, A06, C01, and C05 each contain more than 10 BnMST genes, whereas chromosomes A02, A04, A10, C02, C04, C06, and C09 contain fewer than six BnMST genes ( Figure 2A). The syntenic relationships of the 175 BnMST genes are indicated by lines in Figure 2B. Sixty-nine homologous pairs were identified among the 175 BnMST genes ( Figure 2B). The syntenic relationships among the four species are shown in Figure 2C. The copy numbers of the MST genes in A. thaliana and B. napus ranged from one to eight (Table 2), indicating that some genes were lost or duplicated during evolution. For example, the syntenic homologs of AtERD6-like15, AtERD6-like16, and AtpGlcT4 have been lost in Brassica species. However, three homologous syntenic genes of AtERD6-like6 are present in B. rapa and B. oleracea and six are present in B. napus (three each in the A and C subgenomes) ( Figure 2C, Table 2).  Figure 2C means chromosomes from diverse species or the same species chromosome from diverse sub-genomes.

Exon-Intron Structures and Conserved Motif Analysis of the BnMSTs
We analyzed the exon-intron structures and conserved motifs of the BnMSTs to obtain additional information about their protein profiles (Figure 3). The number of exons ranged from 2 (BnSTP4-2) to 33 (BnVGT1-1). Genes in the BnERD6-like subfamily (except BnERD6-like1, BnERD6-like3, and BnERD6-like6) and BnSTP subfamily (BnSTP2, BnSTP4, BnSTP6, BnSTP8, BnSTP9, BnSTP10, and BnSTP11) contained significantly fewer exons than genes in the other subfamilies. This analysis, combined with phylogenetic analysis of the BnMSTs, indicated that the higher the homology of the sequences, the more similar their genetic structures. For example, BnSTP9-1 and BnSTP9-4 have highly similar gene structures, as do BnSTP10-2 and BnSTP10-6, whereas the exon-intron structures of BnSTP9-1, BnSTP9-4 vs. BnSTP10-2, BnSTP10-6 are significantly different. Twenty conserved motifs were predicted using the MEME server (Figure 4 and Figure S1). Motif 10 is present in all BnMST family members except STP, and motif 13 is in all BnMST family members except INT and PLT. Motifs 9 and 20 are only present in the BnSTP subfamily and motif 19 is only present in the BnSTP, BnINT, and BnERD6-like subfamilies. The BnSTP and BnpGlcT subfamilies contain unique motif 15 and 17, respectively.
We predicted the subcellular localizations of the proteins using the online tool MultiLoc2 (Table S3). Most of the BnMST proteins were predicted to be located in the cytoplasm (45 BnMST proteins), chloroplast (34 BnMST proteins), or nucleus (12 BnMST proteins), and the others were predicted to be in secretory pathways.

qRT-PCR Analysis of DEGs in Different Tissues between Diverse Materials
We performed qRT-PCR to examine the expression patterns of the 31 DEGs in diverse tissues of plants at different stages of growth between materials with extremely high (CQ24) and low (CQ46) HI values ( Figure 6, Table 1). The phenotype of plants and siliques between CQ24 and CQ46 are shown in Figure 6A,B, respectively. Phenotype analysis indicated that the traits of the harvest index (HI), seed yields (SY), and biological yields (BY) were significantly different between the two materials ( Figure 6C). These 31 genes were specifically expressed in different tissues, growth stages, and materials ( Figure 7 and Figure S2, Table S5). We divided the DEGs into two major categories based on the qRT-PCR results: BnSTP1, BnSTP7, BnSTP13, BnTMT1, BnTMT2, BnINT2, BnpGlcT2, BnPLT5, BnERD6, and BnERD6-like5 were highly expressed in silique pericarps ( Figure 7B), whereas BnSTP5, BnSTP12, BnERD6-like10, and BnERD6-like14 were highly expressed in seeds ( Figure 7A). In addition, BnVGT2-2 was only expressed in line CQ24 ( Figure S2). The expression levels of several DEGs increased rapidly in seeds and silique pericarps during the later period of plant growth and development (30-40 d after flowering), including BnSTP5, BnSTP12, BnINT2, BnpGlcT1, BnpGlcT2, BnPLT3, BnPLT5, BnERD6-like5, and BnERD6-like10. Perhaps the genes that are expressed at higher levels in silique pericarps (source organs) play roles in photosynthate loading into source organs, whereas genes expressed specifically in seeds (sink organs) might play important roles in the unloading of carbohydrates. These results suggest that these BnMST genes might increase the HI of B. napus by enhancing the translocation of assimilates to grains (seeds) and that the BnMST genes likely play essential roles in the transport of monosaccharides during the later stages of seed growth and development.

Discussion
B. napus (AACC, 2n = 38) is an allotetraploid plant derived from two diploid species (B. rapa, n = 10, and B. oleracea, n = 9) [59], as confirmed by numerous experimental crosses. The genetic relationships of these species are described by the "U s triangle" model. Comparative genomic analysis between A. thaliana and B. rapa has clearly confirmed the occurrence of a whole genome triplication (WGT) event [60] millions of years ago [61,62]. In the current study, we identified 175 MST genes from B. napus. However, we identified one to eight homologs of each AtMST gene in B. napus, rather than six for every gene ( Figure 2C, Table 2), perhaps due to genome shrinkage and redundancy. Indeed, Mun et al. reported that genome shrinkage and the differential loss of duplicated genes occurred after the WGT event [63], and other studies have provided insights into the processes of genome duplication, loss, or retention [64,65]. In addition, we identified 87 and 83 MST genes from B. oleracea and B. rapa, respectively. The number of BnMST genes in B. napus (175) is greater than the sum of BoMST and BrMST genes ( Table 2), pointing to possible redundancy among the BnMST genes. When we constructed a phylogenetic tree for the four plant species, all MST genes were assigned to seven subfamilies (Figure 1) based on the nomenclature used in A. thaliana [8]. Interestingly, the seven subfamilies did not all follow the same patterns of gene redundancy. The INT subfamily was fully compliant with the WGT event, whereas the number of B. napus genes in the PLT, STP, and pGlcT subgroups was less than the sum of genes in B. rapa and B. oleracea, suggesting that gene elimination occurred in these three subgroup ( Figure 2C, Table 2).
By combining the results of the gene structure, conserved motif, and protein analysis of the BnMSTs, we determined that genes from the same subgroup have similar features. For example, BnSTP and BnERD6-like subfamily members have significantly fewer exons than genes in the five other subfamilies (Figure 3). Similarly, only BnSTP subfamily members contain motif 9 and 20, and most BnMST family members (except BnSTP subfamily proteins) contain motif 10. Furthermore, motif 19 is only present in the BnSTP, BnERD6-like, and BnINT subfamilies ( Figure 4). Interestingly, BnSTP (especially BnSTP10) and BnERD6-like subfamily members were expressed at much lower levels than members of the five other subgroups, as shown in the heatmap ( Figure 5, genes marked with yellow triangles). In conclusion, gene redundancy and gene elimination occurred during the genomic evolution of BnMST, which further drove the diversification of homologous genes in B. napus. In addition, during the genomic evolution of the BnMST genes, changes in exon-intron structures might have resulted in different coding regions, thereby altering the expression patterns and functions of these genes [66].
Among BnMST family members, we were interested in identifying genes with vital functions in seed development and ripening. We therefore analyzed RNA-seq data for the phenotypic characters related to the HI (Table S4) and used the gene expression patterns to construct a heatmap ( Figure 5). Most, but not all, BnMST genes in the same subfamilies had similar expression patterns. For example, genes from the BnpGlcT subfamily were ubiquitously expressed in all tissues, whereas genes in the six other subfamilies were expressed in specific tissues ( Figure 5). However, BnSTP5 and BnSTP13, which belong to the same subfamily, showed completely different expression profiles: BnSTP5 expression strongly increased during the later stage of seed development, whereas BnSTP13 was primarily expressed in leaf tissue at the beginning of flowering and in stems at 30 d after flowering (DAF; Figure 7B). These results suggest that homology does not necessarily reflect a similar function. There were some discrepancies between the heatmap and qRT-PCR results, perhaps due to the different materials and diverse RNA samples examined.
Whereas the expression of BnSTP5 strongly increased in seeds at 30 DAF, BnSTP1 was mainly expressed in silique pericarps at 14-30 DAF ( Figure 7A), suggesting that STP1 might contribute to sugar uptake and is primarily expressed in the early stage of seed development [38]. BnSTP13 was only strongly expressed in stems at 30 DAF and in leaves at the beginning of flowering. This expression pattern is in complete agreement with reports on Arabidopsis, which suggest that overexpressing AtSTP13 could improve the glucose uptake capacity and increase the plant biomass or enhance disease resistance [36,37,67]. STP8 is highly expressed in reproductive organs, whereas STP7 plays an important role in sugar uptake and recycling in the cell wall [68]. STP6 and STP13 are the only genes known to function in fructose transport among STP subfamily members [67].
AtPLT5 encodes another plasma membrane transporter responsible for fructose uptake; this gene is primarily expressed in sink tissues [69]. In contrast, in the current study, BnPLT5 was primarily expressed in leaves at the beginning of flowering and in silique pericarps at 30-40 DAF (Figure 7B), perhaps because the silique pericarp can act as a source or sink tissue at different stages of plant development. The TMT subfamily members TMT1 and TMT2 are highly expressed in various tissues and might be important for the seed yield, whereas TMT3 is barely expressed throughout a plant's lifecycle [27,70]. The expression patterns of pGlct, INT, and VGT subfamily members were different from those identified in previous studies, perhaps due to the different species analyzed.
Interestingly, several genes were more strongly expressed in materials with low, as opposed to high, HI values, such as BnSTP5, BnSTP12, BnERD6-like10, and BnERD6-like14 ( Figure 7A). Among BnMST family members, only these genes were expressed at significantly higher levels in seeds than in silique pericarps. Luo et al. suggested that the HI is primarily influenced by the relationship between "source", "flow", and "sink" and that "flow" is the crucial limiting factor when the "source" is sufficient and the "sink" is not fully utilized [34]. Based on this finding, we hypothesize that BnSTP5, BnSTP12, and BnERD6-like14 take part in the sugar unloading process in sink organs and that a negative feedback effect might occur among these BnMST genes when the "sink" is full. Perhaps the roles of these BnMST genes could be functionally verified in the future.
To better explore the localizations and interactions of the BnMST proteins, we predicted their subcellular localizations using an online tool. The 175 BnMST proteins are localized to diverse compartments, such as the cytoplasm, chloroplast, mitochondria, and nucleus (Table S3). In A. thaliana, pGlcT and MEX1 function in the export of starch degradation products from chloroplasts [71]. However, we predicted that BnpGlcT1 is localized to chloroplasts, whereas BnpGlcT2 is localized to the cytoplasm (Table S3), suggesting that functional segregation might occur in the same subfamily. Furthermore, BnSTP5, BnSTP12, and BnERD6-like14, which are encoded by genes with similar expression patterns, are all located in secretory pathways, indicating that some BnMST proteins likely interact with other proteins from different subfamilies.

Conclusions
In the current study, we performed a systematic study of the BnMST gene family. Genome-wide identification, phylogenetic analysis, and syntenic analysis among B. napus, B. oleracea, B. rapa, and A. thaliana indicated that gene redundancy and elimination occurred during the evolution of BnMST family members. Our analysis of RNA-seq data, gene structures, and conserved motifs indicated that changes in the exon-intron structure could lead to the presence of different coding regions and alter gene expression patterns and functions. Based on subcellular localization and expression analysis of DEGs between materials with two different HI values, BnSTP5, BnSTP13, BnPLT5, and BnERD6-like14, which are specifically expressed in seeds, silique pericarps, or stems, represent excellent candidate genes for further functional studies. Our findings provide basic information about MST genes in Brassica napus and uncover several candidate genes related to the HI for further analysis.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4425/11/6/653/s1: Figure S1: Weblogo plots of the 20 conserved motifs; Figure S2: Expression pattern of the other five DEGs in different tissues and growth periods between two plant materials with extremely high (CQ24) and low (CQ46) HI values; Table S1: Specific primers used for qRT-PCR analysis of the 31 DEGs; Table S2: Homologous genes among the A and C sub-genomes of B. napus, B. rapa, B. oleracea, and their homologs in Arabidopsis; Table S3: Complete list of the 175 BnMST genes identified in our study; Table S4: RNA-seq data for the 175 BnMST genes; Table S5: Expression analysis and significance tests among the 31 DEGs.