Genome-Wide Identiﬁcation and Expression Analysis of SWEET Family Genes in Sweet Potato and Its Two Diploid Relatives

: Sugar Will Eventually be Exported Transporter (SWEET) proteins are key transporters in sugar transportation. They are involved in the regulation of plant growth and development, hormone crosstalk, and biotic and abiotic stress responses. However, SWEET family genes have not been explored in the sweet potato. In this study, we identiﬁed 27, 27, and 25 SWEETs in cultivated hexaploid sweet potato ( Ipomoea batatas , 2n = 6x = 90) and its two diploid relatives, Ipomoea triﬁda (2n = 2x = 30) and Ipomoea triloba (2n = 2x = 30), respectively. These SWEETs were divided into four subgroups according to their phylogenetic relationships with Arabidopsis . The protein physiological properties, chromosome localization, phylogenetic relationships, gene structures, promoter cis -elements, protein interaction networks, and expression patterns of these 79 SWEETs were systematically investigated. The results suggested that homologous SWEETs are differentiated in sweet potato and its two diploid relatives and play various vital roles in plant growth, tuberous root development, carotenoid accumulation, hormone crosstalk, and abiotic stress response. This work provides a comprehensive comparison and furthers our understanding of the SWEET genes in the sweet potato and its two diploid relatives, thereby supplying a theoretical foundation for their functional study and further facilitating the molecular breeding of sweet potato.


Introduction
Sugar Will Eventually be Exported Transporters (SWEETs) play key roles in sugar transport across plasma and intracellular membranes in both prokaryotes and eukaryotes [1]. Almost all SWEETs are present in the membrane structure, such as the plasma membrane and Golgi membrane [2]. As membrane proteins, SWEETs have three transmembrane domains (3TMs) in bacteria but have seven transmembrane domains (7TMs) in eukaryotes [3]. The 3TMs are encoded by a PQ-loop called the Mtn3 domain, which carries conserved proline and glutamine motifs [4,5]. The 7TM helices are folded into two parallel three-helix bundles connected by one central TM [1,6,7]. Since the 7TMs in SWEETs may not be sufficient for creating a functional pore as other types of sugar transporters carrying 12TMs, two SWEETs usually form a functional pore that permits sugar substrate transportation by oligomerization [1,3,7,8]. Accumulating evidence has revealed that SWEETs could homo-or hetero-oligomerize. The co-expression of a mutated and non-functional

Identification and Characterization of SWEETs in the Sweet Potato and Two Diploid Relatives
The plant morphology of the cultivated hexaploid sweet potato is different from that of its diploid relatives, especially since the diploid relatives cannot form tuberous roots ( Figure 1). To comprehensively identify all SWEETs in the sweet potato and its two diploid relatives, we employed three typical strategies (i.e., blastp search, hmmersearch, and the CD-search database). A total of 79 SWEETs were identified in I. batatas (27), I. trifida (27), and I. triloba (25), which were named "Ib", "Itf ", and "Itb", respectively. The physicochemical properties were analyzed using the sequence of IbSWEETs (Table 1). The genomic length of the 27 IbSWEETs ranged from 1052 bp (IbSWEET8.1) to 5747 bp (IbSWEET15.7), and the CDS length varied from 823 bp (IbSWEET9.1) to 1557 bp (IbSWEET2.3). The amino acid lengths of IbSWEETs ranged from 153 aa (IbSWEET15.7) to 321 aa (IbSWEET15.1), with the molecular weight (MW) varying from 17.64 kDa (IbSWEET15.7) to 35.41 kDa (IbSWEET15.1). The isoelectric point (pI) of IbSWEET15.6 (5.81) was the lowest among all the IbSWEETs, indicating that it is an acidic protein. The pI of the other SWEETs was distributed from 7.61 (IbSWEET15.1) to 9.98 (IbSWEET8. 3), suggesting that they are basic proteins. All the IbSWEETs contained Ser, Thr, and Tyr phosphorylation sites. All the IbSWEETs were stable with an aliphatic index of more than 100, except for IbSWEET3.1, which obtained an aliphatic index of 98. 25. The grand average of the hydropathicity (GRAVY) value of all the IbSWEET proteins varied from 0.281 (IbSWEET3.1) to 1.070 (IbSWEET2. 3), indicating that they are hydrophobic. The subcellular localization prediction assay showed that most of IbSWEETs were located in the cell membrane, except three IbSWEETs: IbSWEET15.6 and IbSWEET15.7, which were located in the cell membrane and chloroplasts, and IbSWEET1.1, which was located in the cell membrane and Golgi apparatus. Most of the IbSWEETs have seven transmembrane helical segments (TMHs); several (i.e., IbSWEET6.3, -8.1, -8.3, -9.2, -9.3, -15.2, -15.3, -15.4, and -15.7) have six TMHs; a few (i.e., IbSWEET2.3, -3.1, -6.2, and -10.5) have five TMHs, and IbSWEET15.6 has four TMHs. The three-dimensional structural models showed that there are three conserved α-helices in both N-terminal and C-terminal of all IbSWEETs ( Figure S1).   The SWEETs were distributed across 11, 10, and 11 chromosomes of I. batatas, I. trifida, and I. triloba, respectively ( Figure 2). In I. batatas, five IbSWEETs were detected on LG4 and LG10; three on LG11; two on LG1, LG2, LG8, LG9, LG13, and LG15; and one on LG5 and LG12, whereas no genes were detected on LG3, LG6, LG7, or LG14 (Figure 2a). In I. trifida and I. triloba, the distribution of SWEETs on Chr01 (3), Chr04 (2), Chr11 (2), Chr12 (2), Chr13 (2), and Chr06 (1) was similar, but their distribution on other chromosomes was different (Figure 2b,c). The results indicated a variation and loss of SWEETs during evolution, causing the difference between the distribution and disproportion of SWEETs on the chromosomes in sweet potato and its two diploid relatives.    Table S1.

Phylogenetic Relationship of SWEETs in the Sweet Potato and Its Two Diploid Relatives
To study the evolutionary relationship of SWEETs in I. batatas, I. trifida, I. triloba, and Arabidopsis, we constructed a phylogenetic tree for 96 SWEETs of these four species (i.e., 27 in I. batatas, 27 in I. trifida, 25 in I. triloba, and 17 in Arabidopsis) ( Figure 3). All the SWEETs were unevenly distributed on each branch of the phylogenetic tree. Interestingly, the SWEETs in I. trifida, I. triloba, and Arabidopsis were divided into four subgroups (Groups I to IV), but in I. batatas, they were divided into three subgroups (Groups I to III) according to the evolutionary distance ( Figure 3). The specific distribution of the SWEETs was as follows (total: I.batatas, I. trifida, I. triloba, and Arabidopsis): Group I (22:6, 5, 6, 5), Group II (23:5, 8, 7, 3), Group III (43:16, 10, 10, 7), and Group IV (8:0, 4, 2, 2) ( Figure 3; Table S1). We named IbSWEETs, ItfSWEETs, and ItbSWEETs based on their homology with homologs in Arabidopsis, and only AtSWEET1/2/3/5/6/7/8/9/10/15/16 from Arabidopsis had homologous proteins in I. batatas, I. trifida, and I. triloba. These results indicate that the number and type of SWEETs distributed in each subgroup in the sweet potato differed from those of its two diploid relatives and Arabidopsis.

Phylogenetic Relationship of SWEETs in the Sweet Potato and Its Two Diploid Relatives
To study the evolutionary relationship of SWEETs in I. batatas, I. trifida, I. triloba, and Arabidopsis, we constructed a phylogenetic tree for 96 SWEETs of these four species (i.e., 27 in I. batatas, 27 in I. trifida, 25 in I. triloba, and 17 in Arabidopsis) ( Figure 3). All the SWEETs were unevenly distributed on each branch of the phylogenetic tree. Interestingly, the SWEETs in I. trifida, I. triloba, and Arabidopsis were divided into four subgroups (Groups Ⅰ to Ⅳ), but in I. batatas, they were divided into three subgroups (Groups Ⅰ to Ⅲ) according to the evolutionary distance ( Figure 3). The specific distribution of the SWEETs was as follows (total: I.batatas, I. trifida, I. triloba, and Arabidopsis): ), Group Ⅲ (43:16, 10, 10, 7), and Group Ⅳ (8:0, 4, 2, 2) ( Figure 3; Table S1). We named IbSWEETs, ItfSWEETs, and ItbSWEETs based on their homology with homologs in Arabidopsis, and only AtSWEET1/2/3/5/6/7/8/9/10/15/16 from Arabidopsis had homologous proteins in I. batatas, I. trifida, and I. triloba. These results indicate that the number and type of SWEETs distributed in each subgroup in the sweet potato differed from those of its two diploid relatives and Arabidopsis. Furthermore, a total of 142 SWEET proteins from six plant species (i.e., 27 in I.batatas, 27 in I. trifida, 25 in I.triloba, 17 in Arabidopsis, 21 in rice, and 24 in maize) were used for the phylogenetic analysis. They were divided into four subgroups (Groups I to IV) (Figure 3), which indicated that the evolutionary relationship of the SWEETs was relatively conserved in the plant.

Conserved Motif and Exon-Intron Structure Analysis of SWEETs in the Sweet Potato and Two Diploid Relatives
Furthermore, sequence motifs in the 27 IbSWEETs, 27 ItfSWEETs, and 25 ItbSWEETs were analyzed using the MEME website, and the five most conserved motifs were identified (Figure 4a and Figure S2). Most of the SWEETs contained these five conserved motifs, except for a few SWEETs that were differentiated in the number and species of motifs in I.batatas, I.trifida, and I.triloba, such as IbSWEET15.2 (containing motifs 2-5), ItfSWEET15.2 (containing motifs 1-5), and ItbSWEET15.2 (containing motifs 1-5) (Figure 4a). The PQloop acts as a key structure for the helix of the SWEETs

Cis-Element Analysis in the Promoter of IbSWEETs in Sweet Potato
Promoter cis-elements in plants initiate the gene functions related to plant development, hormone regulation, and stress response. Therefore, we performed a cis-element analysis using the 1500 bp promoter region of IbSWEETs. According to the predicted functions, we divided the elements into five categories: core elements, development regulation elements, hormone-responsive elements, abiotic/biotic stress-responsive elements, and light-responsive elements ( Figure 5). A large number of core elements were identified in the 27 IbSWEETs (CAAT-box and TATA-box) ( Figure 5). Most of the IbSWEETs contained several development elements, such as the O2-site, which was a zein metabolism regula- To better understand the structural diversity among SWEETs, the exon-intron structures were analyzed ( Figure 4c). The number of exons in the SWEETs ranged from two to eight. In more detail, the SWEETs of Group I contained two to six exons; the SWEETs of Group II contained five or six exons; the SWEETs of Group III contained four to six exons; and the SWEETs of Group IV contained five to eight exons ( Figure 4c). The exon-intron structures of some homologous SWEETs were different in I. batatas compared to those in I. trifida and I. triloba, such as IbSWEET8.1 (containing two exons), ItfSWEET8.1 (containing six exons), and ItbSWEET8.1 (containing six exons) in Group I, IbSWEET9.2 (containing five exons) and ItbSWEET9.2 (containing six exons) in Group III, and ItfSWEET16.1 (containing six exons), and ItbSWEET16.1 (containing eight exons) in Group IV (Figure 4c). These results indicated that the SWEET family may have undergone a lineage-specific differentiation event in the sweet potato genome.

Cis-Element Analysis in the Promoter of IbSWEETs in Sweet Potato
Promoter cis-elements in plants initiate the gene functions related to plant development, hormone regulation, and stress response. Therefore, we performed a cis-element analysis using the 1500 bp promoter region of IbSWEETs. According to the predicted functions, we divided the elements into five categories: core elements, development regulation elements, hormone-responsive elements, abiotic/biotic stress-responsive elements, and light-responsive elements ( Figure 5). A large number of core elements were identified in the 27 IbSWEETs (CAAT-box and TATA-box) ( Figure 5). Most of the IbSWEETs contained several development elements, such as the O2-site, which was a zein metabolism regulatory element (found in IbSWEET3.1, -6.2, -8.1, -9.3, -10.1, -10.4, and -15.1); the CAT-box, which was associated with meristem formation (found in IbSWEET2.2, -2.3, -6.2, -8.2, -8.3, -9.2, -10.2, and -15.3); and the GCN4 motif, which was involved in controlling seed-specific expression (found in IbSWEET3.1 and IbSWEET6.1) ( Figure 5). However, no developmentrelated elements were found in the 1500 bp promoter region of IbSWEET15.2, IbSWEET15.6, and IbSWEET15.7. Moreover, light-responsive elements such as the G-box, BOX4, and AE-box were abundant in the promoters of IbSWEETs ( Figure 5).
Additionally, some abiotic elements, such as the drought-responsive elements DREcore, MYB, and MYC; the salt-responsive elements LTR, MBS, and W-box; the lightresponsive elements ERE and LTR; and biotic elements, such as WRE3, W-box, and the WUN motif, were identified in most IbSWEETs ( Figure 5). All the IbSWEETs possessed several hormone elements, including ABRE for ABA-responsive elements, TGA-element for IAA-responsive elements, TATC-box for GA-responsive elements, the CGTCA and TGACG motifs for MeJA-responsive elements, and the TCA motif for SA-responsive elements (Figure 5). These results suggest that IbSWEETs are involved in the regulation of plant growth and development, hormone crosstalk, and abiotic stress adaption in the sweet potato.

Protein Interaction Network of IbSWEETs in the Sweet Potato
To explore the potential regulatory network of IbSWEETs, we constructed an IbSWEET interaction network based on Arabidopsis orthologous proteins ( Figure 6). Protein interaction predictions indicated that some IbSWEETs (IbSWEET1, 6, 8, 9, and 10) could interact with other IbSWEETs to form heterodimers. In addition, SWEETs can interact with pollen development-related protein DEX1 [38], circadian rhythm-related protein FKF1 [39,40], and pathogen responsive-related protein RIN4 and RPM1 [41,42]. IbSWEET2, IbSWEET3, and IbSWEET9 can interact with translation regulation-related protein PUM23 [43]. IbSWEET15 can interact with plant senescence regulatory-related protein SAG12 [44]. These results indicate that IbSWEETs are involved in the regulation of plant growth and development and biotic stress adaption in the sweet potato. core, MYB, and MYC; the salt-responsive elements LTR, MBS, and W-box; the light-re-sponsive elements ERE and LTR; and biotic elements, such as WRE3, W-box, and the WUN motif, were identified in most IbSWEETs ( Figure 5). All the IbSWEETs possessed several hormone elements, including ABRE for ABA-responsive elements, TGA-element for IAA-responsive elements, TATC-box for GA-responsive elements, the CGTCA and TGACG motifs for MeJA-responsive elements, and the TCA motif for SA-responsive elements ( Figure 5). These results suggest that IbSWEETs are involved in the regulation of plant growth and development, hormone crosstalk, and abiotic stress adaption in the sweet potato.

Protein Interaction Network of IbSWEETs in the Sweet Potato
To explore the potential regulatory network of IbSWEETs, we constructed an IbSWEET interaction network based on Arabidopsis orthologous proteins ( Figure 6). Protein interaction predictions indicated that some IbSWEETs (IbSWEET1, 6, 8, 9, and 10) could interact with other IbSWEETs to form heterodimers. In addition, SWEETs can interact with pollen development-related protein DEX1 [38], circadian rhythm-related protein FKF1 [39,40], and pathogen responsive-related protein RIN4 and RPM1 [41,42]. IbSWEET2, IbSWEET3, and IbSWEET9 can interact with translation regulation-related protein PUM23 [43]. IbSWEET15 can interact with plant senescence regulatory-related protein SAG12 [44]. These results indicate that IbSWEETs are involved in the regulation of plant growth and development and biotic stress adaption in the sweet potato.

. Expression Analysis in Various Tissues
To investigate the potential biological function of IbSWEETs in plant growth and development, the expression levels in six representative tissues (i.e., bud, petiole, leaf, stem, pencil root, and tuberous root) of I. batatas were analyzed using real-time quantitative PCR (qRT-PCR) (Figure 7). Nonetheless, different subgroups showed diversified expression patterns in six tissues. IbSWEETs in Group II showed higher expression levels in all the tissues as compared to the other subgroups. Among all the IbSWEETs, six IbSWEETs (i.e., IbSWEET1.1, −2.1, −2.2, −2.3, −9.2, and −10.2) were highly expressed in all the tissues, especially IbSWEET10.2, which was highly expressed by more than 1000-fold in all the tissues. Interestingly, all the IbSWEETs showed high expression levels in the petiole.

Expression Analysis in Different Varieties
We analyzed the expression levels of IbSWEETs in sweet potato varieties with different flesh colors (white flesh: Jiyuan3 and Shangshu19; yellow flesh: Longshu9 and Yanshu32; purple flesh: Luozi5 and Qin12-20-11) (Figure 9). Interestingly, the expression levels of most IbSWEETs in the yellow-fleshed varieties were higher than those in the white-and purple-fleshed varieties. This data indicates that IbSWEETs may be involved in carotenoid accumulation in sweet potato tuberous roots.

Expression Analysis in Different Varieties
We analyzed the expression levels of IbSWEETs in sweet potato varieties with different flesh colors (white flesh: Jiyuan3 and Shangshu19; yellow flesh: Longshu9 and Yanshu32; purple flesh: Luozi5 and Qin12-20-11) (Figure 9). Interestingly, the expression levels of most IbSWEETs in the yellow-fleshed varieties were higher than those in the white-and purple-fleshed varieties. This data indicates that IbSWEETs may be involved in carotenoid accumulation in sweet potato tuberous roots.

Expression Analysis of Hormone Response
To investigate the potential biological functions of IbSWEETs in the hormone signal transduction and crosstalk of plants, we investigated the expressions of SWEETs under various hormonal treatments in order to explore the relationships between SWEETs and hormones. We performed qRT-PCR to evaluate the expression levels of IbSWEETs in response to hormones, including ABA, GA, IAA, MeJA, and SA ( Figure 10). Under ABA

Expression Analysis of Hormone Response
To investigate the potential biological functions of IbSWEETs in the hormone signal transduction and crosstalk of plants, we investigated the expressions of SWEETs under various hormonal treatments in order to explore the relationships between SWEETs and hormones. We performed qRT-PCR to evaluate the expression levels of IbSWEETs in response to hormones, including ABA, GA, IAA, MeJA, and SA ( Figure 10). Under ABA treatment, IbSWEET6.3 (10.30-fold), IbSWEET10.4 (3.76-fold), and IbSWEET15.7 (4.59-fold) were highly induced (Figure 10a). Under GA treatment, all of the IbSWEETs were strongly induced at 0.5 or 1 h (Figure 10b). Under IAA treatment, most of the IbSWEETs were repressed, except IbSWEET9.2, -10.5, and -15.2 (Figure 10c). Under MeJA, most of the IbSWEETs were induced after 24 h. IbSWEET2.1, -2.2, and -2.3 were induced by MeJA at all of the time points (Figure 10d). Under SA treatment, most of the IbSWEETs were sharply repressed at 0.5 h but induced at other time points (Figure 10e). These results indicate that IbSWEETs are differentially expressed in response to various types of hormone induction and that they participate in the crosstalk between various hormones.   (Figure 11). In I.triloba, the ItbSWEETs showed expression patterns that differed from the homologous gene in I. trifida. ItbSWEET2.2, -5.1, -6.1, and -15.3 were induced by ABA. ItbSWEET1.1, -1.2,  -3.1, -6.1, -8.1, -10.3, -15.1, and -15.3 were induced by GA3. ItbSWEET1.1, -2.1, -8.1, -10

Expression Analysis under Abiotic Stresses
To explore the possible roles of IbSWEETs in an abiotic stress response, we analyzed the expression patterns of IbSWEETs using the RNA-seq data of a drought-tolerant variety (Xu55-2) under drought stress, and the RNA-seq data of a salt-sensitive variety (Lizixiang) and a salt-tolerant line (ND98) under salt stress [45,46]. IbSWEET2.1, -10.4, -15.1, and -15.7 were induced by both PEG and NaCl treatments in Xu55-2 and ND98 ( Figure 12).
In addition, we also analyzed the expression patterns of SWEETs using the RNA-seq data of I. trifida and I. triloba under drought and salt treatments [36].  Figure S3). Taken together, these results indicate that SWEETs are differentially expressed in response to various abiotic stresses in the sweet potato and its two diploid relatives.

Expression Analysis under Abiotic Stresses
To explore the possible roles of IbSWEETs in an abiotic stress response, we analyzed the expression patterns of IbSWEETs using the RNA-seq data of a drought-tolerant variety (Xu55-2) under drought stress, and the RNA-seq data of a salt-sensitive variety (Lizixiang) and a salt-tolerant line (ND98) under salt stress [45,46]. IbSWEET2.1, -10.4, -15.1, and -15.7 were induced by both PEG and NaCl treatments in Xu55-2 and ND98 ( Figure 12).

Discussion
Sugar transporters are major players in the distribution of photo-assimilates to various heterotrophic sink organs. SWEETs act as key sugar transporters and play a role in crop yield and quality formation, especially in tuberous-root crops [1][2][3][4][5][6][7][8]. However, the functions and transcriptional regulatory mechanisms of SWEETs remain largely unknown in sweet potato. Tuberous roots are the main tissues harvested from sweet potato, but sweet potato's probable progenitor diploids I.trifida and I. triloba cannot form tuberous roots. Due to the complex genetic background of cultivated sweet potato, recent studies on its gene families have mainly focused on I.trifida and I. triloba [36,[47][48][49]. In this study, we systematically identified SWEETs and compared their characteristics between cultivated hexaploidy sweet potato and its two diploid relatives based on their genome sequences. A genome-wide study of SWEETs is necessary to gain a better understanding of their functions and the molecular breeding of sweet potato.

Evolution of the SWEET Gene Family in the Sweet Potato and Its Two Diploid Relatives
In this study, a total of 79 SWEETs were identified in sweet potato and its two diploid relatives. The number of SWEETs identified in I. batatas (27) was the same as that in I. trifida (27), but there were two fewer in I. triloba (25) (Figure 2; Table S1). Genomic alignment revealed the differentiation and evolution of chromosomes [50]. The chromosome localization and distribution of the SWEETs in each chromosome differed between I. batatas, I. trifida, and I. triloba; 11 chromosomes contained SWEET genes in I.batatas and I. triloba, but 10 chromosomes contained SWEET genes in I.trifida (Figure 2). Based on the phylogenetic relationship, the SWEETs were divided into four subgroups (Group Ⅰ to Ⅳ). There were no IbSWEETs in Group Ⅲ (Figure 3). Moreover, the number and type of In addition, we also analyzed the expression patterns of SWEETs using the RNA-seq data of I. trifida and I. triloba under drought and salt treatments [36].  Figure S3). Taken together, these results indicate that SWEETs are differentially expressed in response to various abiotic stresses in the sweet potato and its two diploid relatives.

Discussion
Sugar transporters are major players in the distribution of photo-assimilates to various heterotrophic sink organs. SWEETs act as key sugar transporters and play a role in crop yield and quality formation, especially in tuberous-root crops [1][2][3][4][5][6][7][8]. However, the functions and transcriptional regulatory mechanisms of SWEETs remain largely unknown in sweet potato. Tuberous roots are the main tissues harvested from sweet potato, but sweet potato's probable progenitor diploids I.trifida and I. triloba cannot form tuberous roots. Due to the complex genetic background of cultivated sweet potato, recent studies on its gene families have mainly focused on I.trifida and I. triloba [36,[47][48][49]. In this study, we systematically identified SWEETs and compared their characteristics between cultivated hexaploidy sweet potato and its two diploid relatives based on their genome sequences. A genome-wide study of SWEETs is necessary to gain a better understanding of their functions and the molecular breeding of sweet potato.

Evolution of the SWEET Gene Family in the Sweet Potato and Its Two Diploid Relatives
In this study, a total of 79 SWEETs were identified in sweet potato and its two diploid relatives. The number of SWEETs identified in I. batatas (27) was the same as that in I. trifida (27), but there were two fewer in I. triloba (25) (Figure 2; Table S1). Genomic alignment revealed the differentiation and evolution of chromosomes [50]. The chromosome localization and distribution of the SWEETs in each chromosome differed between I. batatas, I. trifida, and I. triloba; 11 chromosomes contained SWEET genes in I.batatas and I. triloba, but 10 chromosomes contained SWEET genes in I.trifida (Figure 2). Based on the phylogenetic relationship, the SWEETs were divided into four subgroups (Group I to IV). There were no IbSWEETs in Group III (Figure 3). Moreover, the number and type of SWEETs distributed in each subgroup of the sweet potato and its two diploid relatives were different from those in Arabidopsis and other plants (Figure 3). These results reveal that the SWEET gene family might have undergone a lineage-specific differentiation event in the terrestrial plant genome.
Five conserved motifs were identified in all 79 SWEETs, and all the SWEETs were found to contain a PQ-loop, indicating that these motifs are evolutionarily conserved among the sweet potato and its two diploid relatives. In Arabidopsis, four SANT-domain proteins (SANT1-4) were found to form a complex with HDA6 to regulate flowering [37]. Only ItfSWEET9.1 and ItbSWEET9.1, which were highly expressed in the flower and flower bud, were found to contain a SANT domain (Figure 4b). Introns usually act as buffer zones or mutation-resistant fragments that reduce adverse mutations and insertions. Moreover, introns also play essential roles in mRNA export, transcriptional coupling, alternative splicing, gene expression regulation, and other biological processes [50,51]. Here, the exonintron distributions of some homologous SWEETs were different in I. batatas compared with those in I. trifida and I. triloba (Figure 4c). For example, in Group I, IbSWEET8.1 contained one intron, but its homologous genes, ItfSWEET8.1 and ItbSWEET8.1, contained five introns. In Group III, IbSWEET15.1, ItfSWEET15.1, and ItbSWEET15.1 contained six, four, and six exons, respectively. In the sweet potato and the two diploids, these differences in the exon-intron structure may result in the different functions carried out by SWEETs in plant development [52][53][54].

Different Functions of SWEETs in Tuberous Root Development in Sweet Potato
In plants, SWEETs have been reported to be involved in root development and assimilate accumulation. The atsweet11 and atsweet12 double mutants exhibited delayed root development and severe modifications to the chemical composition of the xylem cell wall [19]. The knockout of OsSWEET11 significantly decreased the sucrose concentration in mutant embryo sacs and led to defective grain filling [27,55]. For the sweet potato, the formation and development of tuberous roots is critical to the roots' yield and quality. Storage-root formation has been considered to be a process of assimilate accumulation [56]. As major transporters governing long-distance transport and sugar accumulation in sink cells, SWEETs may play vital roles in tuberous root development in the sweet potato [12,57]. In this study, most IbSWEETs peaked during the initial development stage (20 d) and the rapid expansion stage (50 d) of the tuberous roots, respectively ( Figure 8). These results indicate that IbSWEETs may participate in tuberous root formation by regulating assimilate accumulation in sweet potato.
The flesh color of the tuberous root is one of the most important quality characteristics of the sweet potato. Most of the IbSWEETs were highly expressed in the yellow-fleshed varieties, which are rich in carotenoids (Figure 9). Carotenoids are derived from two isoprene isomers, isopentenyl diphosphate (IPP) and its allylic isomer, dimethylallyl diphosphate (DMAPP). IPP and DMAPP come from the Calvin-Benson cycle by fixed carbon [58,59]. Additionally, SWEETs' transport of sucrose is a key step for fixed-carbon transport in the phloem; thus, they may provide a sufficient precursor substance for carotenoid production in the sweet potato [11,60,61]. These data indicate that IbSWEETs may be involved in carotenoid accumulation in sweet potato tuberous roots by transporting photo-assimilates.
However, further study is required to underlie the regulatory mechanisms of SWEETs on tuberous root development and carotenoids accumulation.

Different Functions of SWEETs in Hormone Crosstalk in the Sweet Potato and Its Two Diploid Relatives
SWEETs have been reported to participate in the regulation of multiple hormones. The interaction between SWEETs and CWINV (cell wall invertase), which encodes an enzyme that catalyzes the hydrolysis of sucrose into glucose and fructose, may lead to the loss of apical dominance and the appearance of multiple shoots under cytokinins [62]. The atsweet13 and atsweet14 double mutant line showed function loss in transporting exogenous GA [24][25][26]. OsSWEET13a was found to be involved in the transport of GA to young leaves during the early developmental stage [24]. The overexpression of OsSWEET5 inhibited auxin concentration, signaling, and translocation in rice [25]. In this study, each IbSWEET gene could be induced by at least two hormones. IbSWEET2.1, which contained an ABAresponsive element (i.e., ABRE, and an SA-responsive element, or the TCA motif), was induced by ABA, GA, and MeJA but repressed by IAA and SA. However, ItbSWEET2.1 was induced by IAA, and there was no significant change in ItfSWEET2.1 under IAA treatment. IbSWEET8.1, which contained a TCA motif, was induced by GA, MeJA, and SA but repressed by ABA and IAA treatments ( Figure 10). However, ItbSWEET8.1 was induced by IAA. IbSWEET15.5, which contained a GA-responsive element (i.e., the TATC-box, and JA-responsive elements, or a TGACG motif, an ABRE, and a TCA motif), was significantly induced by GA and SA. IbSWEET15.3, which contained a TGACG motif and an ABRE was repressed under ABA treatment, but ItbSWEET15.3 was induced by ABA, GA, and IAA. ItbSWEET16.1 was repressed under ABA treatment, but ItfSWEET16.1 was induced by ABA ( Figure 11). These results indicate that SWEETs are involved in the crosstalk of multiple hormones and that homologous SWEET genes participate in different hormone pathways in sweet potato and its two diploid relatives (Tables S2 and S3). However, the roles of SWEETs in the regulation of hormone crosstalk still need further investigation.

Different Functions of SWEETs in Abiotic Stress Response in the Sweet Potato and Its Two Diploid Relatives
SWEETs have been reported to participate in the abiotic stress response in plants. In grapes, VvSWEET11 and VvSWEET15 were found to be significantly induced by heat treatment [63]. In Arabidopsis, AtSWEET15 was highly expressed under cold and salinity treatments [64]. Here, SWEETs were differentially expressed in response to various abiotic stresses in the sweet potato and its two diploid relatives. In the sweet potato, IbSWEET2.1, -10.4, -15.1, and -15.7 were induced by both PEG and NaCl treatments in Xu55-2 and ND98 ( Figure 12). Moreover, the diploids I. trifida and I. triloba could be used to discover functional genes, particularly genes conferring resistance or tolerance to biotic and abiotic stress, which were possibly lost in the cultivated sweet potato during its domestication [57]. In the two diploid relatives, ItfSWEET2.  Figure S3). These SWEETs may serve as candidate genes for use in improving abiotic stress tolerance in sweet potato.

Phylogenetic Analysis of SWEETs
Multiple sequence alignment of the deduced amino acid sequences of the SWEETs from I. batatas, I. trifida, I. triloba, Arabidopsis, Zea mays, and Oryza sativa were aligned with Clustal X, and the alignment was imported into MEGA11 to create a phylogenetic tree using the neighbor-joining method with 1000 bootstrap replicates (www.megasoftware.net, accessed on 3 December 2022) [68]. Then, the phylogenetic tree was constructed using iTOL (http://itol.embl.de/, accessed on 3 December 2022).

Domain Identification and Conserved Motif Analysis of SWEETs
The conserved motifs of the SWEETs were analyzed using MEME software (https: //meme-suite.org/meme/, accessed on 5 August 2022). The MEME parameters were set to search for a maximum of 15 motifs with a motif width comprised between 5 and 50 residues [69].

Protein Interaction Network of SWEETs
The protein interaction networks of the SWEETs were predicted using STRING (https: //cn.string-db.org/, accessed on 7 August 2022) based on Arabidopsis homologous proteins. The network map was built using Cytoscape software [71].

qRT-PCR Analysis of SWEETs
The salt-tolerant sweet potato (I. batatas) line ND98 was used for qRT-PCR analysis in this study [45]. In vitro grown ND98 plants were cultured on Murashige and Skoog (MS) medium at 27 ± 1 • C under a photoperiod consisting of 13 h of cool-white fluorescent light at 54 µmol m −2 s −1 and 11 h of darkness. The sweet potato plants were cultivated in a field in the campus of China Agricultural University, Beijing, China.
For expression analysis in various tissues, the total RNA was extracted from the buds, leaves, petioles, stems, pencil roots, and tuberous root tissues of 3-month-old field-grown ND98 plants; the different development stage of the tuberous root tissues of Y25 (3 d, 10 d, 20 d, 30 d, 40 d, 50 d, 60 d, 70 d, 80 d, and 90 d) and the tuberous root tissues of different field-grown plants at 90 d (Jiyuan3, Shangshu19, Longshu9, Yanshu32, Luozi5, and Qin12-20-11) were analyzed using the TRIzol method (Invitrogen). For the expression analysis of the hormone treatment, the leaves were sampled at 0, 0.5, 1, 3, 6, 12, 24, and 48 h after being treated with 100 µM ABA, 100 µM GA, 100 µM IAA, 100 µM MeJA, and 100 µM SA, respectively. Three independent biological replicates were taken, each with three plants. qRT-PCR was conducted using the SYBR detection protocol (TaKaRa, Kyoto, Japan) on a 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). The reaction mixture was composed of first-strand cDNA, a primer mix, and an SYBR Green M Mix (TaKaRa; code RR420A) with a final volume of 20 µL. A sweet potato actin gene (GenBank AY905538) was used as an internal control. The relative gene expression levels were quantified using the comparative C T method [72]. The specific primers used for the qRT-PCR analysis are listed in Table S4. The heat maps of the gene expression profiles were constructed using the TBtools software (v.1.098696) [65].

Transcriptome Analysis
The RNA-seq data of ItfSWEETs and ItbSWEETs in I. trifida and I. triloba were downloaded from the Sweetpotato Genomics Resource (http://sweetpotato.plantbiology.msu. edu/, accessed on 10 August 2022). The RNA-seq data of IbSWEETs in I. batatas were obtained from the NCBI SRA repository under the accession number SRP092215 [45,46]. The expression levels of the SWEETs were calculated as fragments per kilobase of exon per million fragments mapped (FPKM). The heat maps were constructed using the Tbtools software (v.1.098696) [65].

Conclusions
In this study, we identified and characterized 27, 27, and 25 SWEETs in cultivated hexaploidy sweet potato (I. batatas, 2n = 6x = 90) and its two diploid relatives, I. trifida (2n = 2x = 30) and I. triloba (2n = 2x = 30), respectively, based on genome and transcriptome data. The protein physicochemical properties, chromosome localization, phylogenetic relationships, gene structures, promoter cis-elements, and protein interaction networks of these 79 SWEETs were systematically investigated. Moreover, the tissue specificity and expression patterns of the SWEETs in tuberous root development, hormone responses, and abiotic stress responses were analyzed using qRT-PCR and RNA-seq. The results indicated that there was a differentiation in the functions of homologous SWEETs in the sweet potato and its two diploid relatives, and each SWEET gene played different vital roles in the plants' growth and development, carotenoid accumulation, hormone crosstalk, and abiotic stress response. This study provides valuable insights into the structure and function of SWEET genes in the sweet potato and its two diploid relatives.