Genome-Wide Identification of the Rose SWEET Gene Family and Their Different Expression Profiles in Cold Response between Two Rose Species

Sugars Will Eventually be Exported Transporter (SWEET) gene family plays indispensable roles in plant physiological activities, development processes, and responses to biotic and abiotic stresses, but no information is known for roses. In this study, a total of 25 RcSWEET genes were identified in Rosa chinensis ‘Old Blush’ by genome-wide analysis and clustered into four subgroups based on their phylogenetic relationships. The genomic features, including gene structures, conserved motifs, and gene duplication among the chromosomes of RcSWEET genes, were characterized. Seventeen types of cis-acting elements among the RcSWEET genes were predicted to exhibit their potential regulatory roles during biotic and abiotic stress and hormone responses. Tissue-specific and cold-response expression profiles based on transcriptome data showed that SWEETs play widely varying roles in development and stress tolerance in two rose species. Moreover, the different expression patterns of cold-response SWEET genes were verified by qRT-PCR between the moderately cold-resistant species R. chinensis ‘Old Blush’ and the extremely cold-resistant species R. beggeriana. Especially, SWEET2a and SWEET10c exhibited species differences after cold treatment and were sharply upregulated in the leaves of R. beggeriana but not R. chinensis ‘Old Blush’, indicating that these two genes may be the crucial candidates that participate in cold tolerance in R. beggeriana. Our results provide the foundation for function analysis of the SWEET gene family in roses, and will contribute to the breeding of cold-tolerant varieties of roses.


Introduction
Sugars Will Eventually be Exported Transporter (SWEET) is a new type of sugar transporter that was first identified in Arabidopsis thaliana in 2010 [1]. Comparing with monosaccharide transporters (MST) and sucrose transporters (SUT), SWEET proteins are capable of transporting sugar bidirectionally without energy dependence, and the types of substrates for sugar transport are more extensive [2]. In plants, SWEET proteins usually contain seven transmembrane helices (TMHs); two MtN3/saliva domains with three TMHs are connected by the fourth TMH to form a "3-1-3" structure [3]. The SWEET family is divided into four Clades. Different clades have different preferences for monosaccharides or disaccharides, in which Clades I and II prefer transporting hexose and Clade III prefer transporting sucrose, while SWEET proteins of Clade IV tend to transport fructose on the tonoplast [4][5][6][7].
Due to the advances in whole-genome sequencing in plants, the genome-wide identification of SWEET genes has been reported in many crops, vegetables, and fruits, such as rice (Oryza sativa) [8], soybean (Glycine max) [9], sorghum (Sorghum bicolor) [10], tomato (Solanum lycopersicum) [11], potato (Solanum tuberosum) [12], cabbage (Brassica oleracea) [13], the SWEET gene family in rose and provide valuable information for further functional analysis of RbSWEET genes in response to cold tolerance.

Genome-Wide Identification of the SWEET Family Genes in R. chinensis
A total of 29 SWEET proteins were identified on the basis of HMM profiles of the SWEET domain (PF03083) search results. Meanwhile, 28 SWEET homologous proteins were obtained by using 17 protein sequences of AtSWEETs to BLASTP in the R. chinensis proteome with an e-value threshold of 1 × 10 −5 . After the intersection of these two results, 25 SWEET genes were finally identified in R. chinensis according to the conserved domains in the relevant database (Table 1). These genes were named from RcSWEET1 to RcSWEET17b according to their homology with A. thaliana. Protein characterization analysis showed that the ranges of genomic DNA (gDNA) and coding sequences (CDS) are 717~3289 bp and 708~933 bp, respectively. The amino acid length of RcSWEET family members ranged from 235 (RcSWEET2a, RcSWEET2b, and RcSWEET5c) to 310 (Rc-SWEET12), and the average molecular weight (MW) ranged from 26,284.25 (RcSWEET2b) to 34,702.73 (RcSWEET12). The range of the theoretical isoelectric point (pI) is between 5.93 (RcSWEET15b) and 10.11 (RcSWEET10c). The TMHMM predicted results showed that all 25 RcSWEET proteins contained 7 transmembrane domains, indicating that the transmembrane structure of RcSWEET proteins was relatively conservative ( Figure S1). The gene structures, including untranslated regions (UTRs), exons, and introns, of each RcSWEET gene were predicted based on gDNA and CDS sequences (Table S2). The majority of the RcSWEET genes (RcSWEET1, RcSWEET2a, RcSWEET2b, RcSWEET3, Rc-SWEET4a, RcSWEET5a, RcSWEET5c, RcSWEET5d, RcSWEET9, RcSWEET10a, RcSWEET10b, RcSWEET10c, RcSWEET11a, RcSWEET12, RcSWEET15a, RcSWEET15b, RcSWEET17a, and RcSWEET17b) contained six exons. Two RcSWEET genes (RcSWEET4b and RcSWEET11b) had five exons, whereas five RcSWEET genes (RcSWEET5b, RcSWEET5e, RcSWEET5f, Rc-SWEET5g, and RcSWEET5h) had only one exon and no intron ( Figure 1B). The conserved motifs were investigated based on the protein sequences of the RcSWEET genes. Ten consensus motifs were detected in the RcSWEET genes ( Figure 1C). Each of the 25 members contained two MtN3_slv domains that were composed of three transmembrane helical domains. To investigate the phylogenetic relationship of the SWEET genes in R. chinensis, the identified SWEET family members of rose were constructed alone and combined with the SWEET family members of rice and Arabidopsis. The results showed that 25 RcSWEETs were divided into four clades, including 4 members in Clade I, 10 members in Clade II, 9 members in Clade III, and 2 members in Clade IV ( Figures 1A and 2).
Gene duplication events usually play an indispensable role in gene expansion and the evolution of gene families. A pair of tandemly duplicated genes (RcSWEET10a and RcSWEET12) located on Chr7, and another cluster of five tandemly duplicated genes (RcSWEET9, RcSWEET10b, RcSWEET10c, RcSWEET11a, and RcSWEET11b) located on Chr6 ( Figure 3A). In addition, two segmental duplication events, including RcSWEET4a/ RcSWEET5a and RcSWEET12/RcSWEET15a, were detected on three different chromosomes ( Figure 3B). The selection pressure was further estimated by the Ka (nonsynonymous) and Ks (synonymous) substitution ratios [46]. Only one pair of segmentally duplicated RcSWEETs, namely RcSWEET4a and RcSWEET5a, had a Ka/Ks ratio of 0.15, which was significant, and indicated a synonymous change that has been selected during plant genome evolution (Table S4).

Cis-Elements Analysis in the Promoters of the RcSWEET Genes
To further study the potential regulatory mechanisms of the RcSWEET genes in response to abiotic stress, the 2000 bp upstream sequences from the translation start sites of the RcSWEET genes were submitted into PlantCARE to detect the cis-elements. Seventeen cis-acting elements associated with stress and hormones were identified, including ABRE, ARE, AuxRR core, Box 4, CGTCA-motif, circadian, ERE, GARE-motif, LTR, MBS, P-box, TATC-box, TCA-element, TC-rich, TGACG-motif, TGA-element, and WUN-motif ( Figure 4). The details of the cis-acting elements contained in each RcSWEET gene have been listed in Table S5. All the RcSWEET genes possessed at least one stress-relevant cis-element; ABRE, ARE, and Box 4 were the most abundant genes in the SWEET family, indicating that abscisic acid, light, and anaerobic conditions had a great influence on the expression of RcSWEET genes, and these cis-acting elements were also related to stress response. In particular, 13 elements (52%) contained one or more LTR, suggesting a potential cold-stress response under low temperature conditions.
To further study the potential regulatory mechanisms of the RcSWEET genes in response to abiotic stress, the 2000 bp upstream sequences from the translation start sites of the RcSWEET genes were submitted into PlantCARE to detect the cis-elements. Seventeen cis-acting elements associated with stress and hormones were identified, including ABRE, ARE, AuxRR core, Box 4, CGTCA-motif, circadian, ERE, GARE-motif, LTR, MBS, P-box, TATC-box, TCA-element, TC-rich, TGACG-motif, TGA-element, and WUN-motif ( Figure  4). The details of the cis-acting elements contained in each RcSWEET gene have been listed in Table S5. All the RcSWEET genes possessed at least one stress-relevant cis-element; ABRE, ARE, and Box 4 were the most abundant genes in the SWEET family, indicating that abscisic acid, light, and anaerobic conditions had a great influence on the expression of RcSWEET genes, and these cis-acting elements were also related to stress response. In particular, 13 elements (52%) contained one or more LTR, suggesting a potential coldstress response under low temperature conditions.

Expression Patterns of Cold-Response SWEET Genes in Two Rosa Species
Identically as in R. chinensis, 25 RbSWEET genes have been identified in R. beggeriana based on the transcriptome analysis of leaves and shoots in response to cold stress ( Figure 6). At least 12 genes were statistically detected under the growth temperature of 23/18 • C, including RbSWEET1, RbSWEET2a, RbSWEET2b, RbSWEET3, RbSWEET4b, RbSWEET10c, RbSWEET11a, RbSWEET15a, RbSWEET17a, and RbSWEET17b in both tissues, as well as RbSWEET5f and RbSWEET5h in shoots. The above results presented very similar expression patterns in the leaves and shoots of R. chinensis 'Old Blush'. After cold stress at 4 • C, the differentially expressed genes (DEGs) were screened out of the leaves and shoots of R. beggeriana. In leaves, RbSWEET2a and RbSWEET10c were upregulated, while RbSWEET4b was downregulated during cold stress. RbSWEET1 and RbSWEET15a were downregulated at 1 h and then upregulated until 24 h of cold stress. On the contrary, RbSWEET3 and RbSWEET2b presented the upregulated expression level at 1 h and then the downregulated expression level at 6 h and 24 h during cold stress. In shoots, RbSWEET10c, RbSWEET4b, RbSWEET1, and RbSWEET2b exhibited very similar expression trends as those genes in leaves. In contrast to leaves, there were much higher expression levels of RbSWEET10c and RbSWEET4b in shoots. RbSWEET2a and RbSWEET15a exhibited the downregulated expression levels, but RbSWEET3 fluctuated in shoots during cold stress. In addition, RbSWEET17a did not show a different effect on growth temperature but maintained the relatively higher expression levels in leaves and shoots under both growth temperatures.

Expression Patterns of Cold-Response SWEET Genes in Two Rosa Species
Identically as in R. chinensis, 25 RbSWEET genes have been identified in R. beggeriana based on the transcriptome analysis of leaves and shoots in response to cold stress ( Figure  6). At least 12 genes were statistically detected under the growth temperature of 23/18 °C, including RbSWEET1, RbSWEET2a, RbSWEET2b, RbSWEET3, RbSWEET4b, RbSWEET10c, RbSWEET11a, RbSWEET15a, RbSWEET17a, and RbSWEET17b in both tissues, as well as RbSWEET5f and RbSWEET5h in shoots. The above results presented very similar expression patterns in the leaves and shoots of R. chinensis 'Old Blush'. After cold stress at 4 °C, the differentially expressed genes (DEGs) were screened out of the leaves and shoots of R. beggeriana. In leaves, RbSWEET2a and RbSWEET10c were upregulated, while RbSWEET4b was downregulated during cold stress. RbSWEET1 and RbSWEET15a were downregulated at 1 h and then upregulated until 24 h of cold stress. On the contrary, RbSWEET3 and RbSWEET2b presented the upregulated expression level at 1 h and then the downregulated expression level at 6 h and 24 h during cold stress. In shoots, RbSWEET10c, RbSWEET4b, RbSWEET1, and RbSWEET2b exhibited very similar expression trends as those genes in leaves. In contrast to leaves, there were much higher expression levels of RbSWEET10c and RbSWEET4b in shoots. RbSWEET2a and RbSWEET15a exhibited the downregulated expression levels, but RbSWEET3 fluctuated in shoots during cold stress. In addition, RbSWEET17a did not show a different effect on growth temperature but maintained the relatively higher expression levels in leaves and shoots under both growth temperatures. The expression patterns of DEGs or highly expressed SWEET genes in leaves and shoots of R. chinensis 'Old Blush' and R. beggeriana during cold treatment are shown in Figure 7. In the leaves of R. chinensis 'Old Blush', significantly increased relative expression levels were found in RcSWEET1 at 6 h and RcSWEET15a at 24 h during cold stress. Rc-SWEET2a and RcSWEET10c showed a significant decline at 1 h but recovered at 6 h and 24 h. RcSWEET2b did not significantly change during cold stress. The relative expression levels of RcSWEET3 and RcSWEET17a fluctuated during cold stress, significantly decreasing at 1 h, increasing at 6 h, and declining at 24 h. RcSWEET4b exhibited the exact opposite trend as RcSWEET3, but maintained significantly higher relative expression levels during cold treatment than those at 0 h. In shoots of R. chinensis 'Old Blush', RcSWEET1 was significantly upregulated at 1 h and 24 h, but RcSWEET2a was unchanged during cold treatment. The increased expression levels were observed in RcSWEET2b at 6 h and in RcSWEET10c at 1 h. RcSWEET3 was dramatically upregulated at 1 h, then significantly downregulated at 6 h, but recovered at 24 h. The relative expression levels of RcSWEET4b, RcSWEET15a, and RcSWEET17a significantly increased at 1 h and 6 h but sharply declined at 24 h. The expression patterns of DEGs or highly expressed SWEET genes in leaves and shoots of R. chinensis 'Old Blush' and R. beggeriana during cold treatment are shown in Figure 7. In the leaves of R. chinensis 'Old Blush', significantly increased relative expression levels were found in RcSWEET1 at 6 h and RcSWEET15a at 24 h during cold stress. RcSWEET2a and RcSWEET10c showed a significant decline at 1 h but recovered at 6 h and 24 h. RcSWEET2b did not significantly change during cold stress. The relative expression levels of RcSWEET3 and RcSWEET17a fluctuated during cold stress, significantly decreasing at 1 h, increasing at 6 h, and declining at 24 h. RcSWEET4b exhibited the exact opposite trend as RcSWEET3, but maintained significantly higher relative expression levels during cold treatment than those at 0 h. In shoots of R. chinensis 'Old Blush', RcSWEET1 was significantly upregulated at 1 h and 24 h, but RcSWEET2a was unchanged during cold treatment. The increased expression levels were observed in RcSWEET2b at 6 h and in RcSWEET10c at 1 h. RcSWEET3 was dramatically upregulated at 1 h, then significantly downregulated at 6 h, but recovered at 24 h. The relative expression levels of RcSWEET4b, RcSWEET15a, and RcSWEET17a significantly increased at 1 h and 6 h but sharply declined at 24 h.
As for the leaves and shoots of R. beggeriana, the expression patterns of most RbSWEETs determined using qRT-PCR detection were similar to the transcriptome results. However, the qRT-PCR analysis showed the expression levels of RbSWEET4b increased at 1 h and 24 h in leaves and enhanced at 1 h and 6 h in shoots, instead of the decreased results in the RNA-seq data. In addition, RbSWEET17a did not alter in the transcriptome data but significantly declined in leaves by qRT-PCR analysis during cold stress. As for the leaves and shoots of R. beggeriana, the expression patterns of most RbSWEETs determined using qRT-PCR detection were similar to the transcriptome results. However, the qRT-PCR analysis showed the expression levels of RbSWEET4b increased at 1 h and 24 h in leaves and enhanced at 1 h and 6 h in shoots, instead of the decreased results in the RNA-seq data. In addition, RbSWEET17a did not alter in the transcriptome data but significantly declined in leaves by qRT-PCR analysis during cold stress.
The further comparison analysis disclosed that SWEET genes showed various expression patterns in two species during cold treatment ( Figure 7). For instance, SWEET1 and SWEET4b exhibited upregulated trends in leaves and shoots of both species. The expression patterns of SWEET2a and SWEET10c were quite similar in the shoots of both species. In leaves, RcSWEET2a and RcSWEET10c sharply declined at 1 h, then recovered to the same expression levels as those genes at 0 h, while RbSWEET2a and RbSWEET10c stably enhanced and reached the significant differences after 6 h and 24 h of cold treatment, respectively. Moreover, SWEET2b, SWEET3, and SWEET15a exhibited upregulated trends in the shoots of R. chinensis 'Old Blush' at the beginning of cold stress, but downregulated in the shoots of R. beggeriana during the whole cold treatment.

Subcellular Localization of Three RbSWEET Genes
Based on the expression patterns of the SWEET genes in both species, three RbSWEET genes were selected for subcellular localization. RbSWEET1, RbSWEET2a, and RbSWEET10c were transiently overexpressed in tobacco leaf epidermal cells. As shown in Figure 8, RbSWEET1 and RbSWEET10c were found to be localized to the plasma membrane, and RbSWEET2a was found to be localized to the tonoplast, whereas the control (GFP alone) was observed clearly in the plasma membrane, cytoplasm, and nucleus.  The further comparison analysis disclosed that SWEET genes showed various expression patterns in two species during cold treatment (Figure 7). For instance, SWEET1 and Error bars indicate standard deviation, and asterisks indicate significant differences between the control (0 h) and 4 • C treatment for 1 h, 6 h, 24 h, * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p< 0.0001.

Subcellular Localization of Three RbSWEET Genes
Based on the expression patterns of the SWEET genes in both species, three RbSWEET genes were selected for subcellular localization. RbSWEET1, RbSWEET2a, and RbSWEET10c were transiently overexpressed in tobacco leaf epidermal cells. As shown in Figure 8, RbSWEET1 and RbSWEET10c were found to be localized to the plasma membrane, and RbSWEET2a was found to be localized to the tonoplast, whereas the control (GFP alone) was observed clearly in the plasma membrane, cytoplasm, and nucleus.

Discussion
The SWEET gene family has been identified in many major crops such as rice [8], sorghum [10], and soybean [9], as well as important horticultural plants such as cabbage [13], potato [12], tomato [11], and apple [15]. However, little information has been reported on roses and other ornamental plants, especially their molecular regulatory mechanisms in cold response. The genome sequencing of R. chinensis 'Old Blush' laid the foundation for the whole genome identification and analysis of the rose gene family [48]. In this study, 25 RcSWEET genes were identified and classified into four subfamilies (Clades

Discussion
The SWEET gene family has been identified in many major crops such as rice [8], sorghum [10], and soybean [9], as well as important horticultural plants such as cabbage [13], potato [12], tomato [11], and apple [15]. However, little information has been reported on roses and other ornamental plants, especially their molecular regulatory mechanisms in cold response. The genome sequencing of R. chinensis 'Old Blush' laid the foundation for the whole genome identification and analysis of the rose gene family [48]. In this study, 25 RcSWEET genes were identified and classified into four subfamilies (Clades I-IV) according to their phylogenetic evolutionary relationship, including four members in Clade I, ten members in Clade II, nine members in Clade III, and two members in Clade IV ( Figure 1A). The phylogenetic tree constructed with the SWEET family proteins of Arabidopsis, rice, and rose showed that the SWEET proteins of rose were more closely related to Arabidopsis, which also belongs to dicotyledons (Figure 2). Analysis of the RcSWEET gene structure indicated that most genes contained five or six exons, which was consistent with the analysis results in other species such as cabbage [13] and banana [17]. Moreover, similar intron and extron arrangements were found in the same subfamily members ( Figure 1B), which implied a different function in each clade. In plants, SWEET proteins usually consist of seven transmembrane helices (TMHs), and two MtN3/saliva domains with three TMHs connected by the fourth transmembrane helix to form a "3-1-3" structure [3]. Compared with plants and other eukaryotes, SWEET protein in prokaryotes usually has only three TMHs, which is named SemiSWEET protein, suggesting that eukaryotic SWEET protein is likely to evolve due to gene replication or fusion [6]. Ten consensus motifs were detected in the RcSWEET proteins, and all the RcSWEET proteins contained two MtN3_slv domains. This result suggested that the structure of RcSWEET proteins was highly conserved ( Figure 1C).
Gene duplication events, including whole-genome duplication (WGD), tandem duplication, and segmental duplication events, usually result in functional segregation dur-ing gene expansion and evolution [46]. A total of 11 RcSWEET genes were identified as being tandemly or segmentally duplicated, which included a pair of tandemly duplicated genes (RcSWEET10a and RcSWEET12) located on Chr7, another cluster of five tandemly duplicated genes (RcSWEET9, RcSWEET10b, RcSWEET10c, RcSWEET11a, and RcSWEET11b) located on Chr6, and two pairs of genes (RcSWEET4a-RcSWEET5a and RcSWEET12-RcSWEET15a) were segmentally duplicated (Figure 3). These results indicated that tandem duplication and segmental duplication synergistically contribute to the expansion of the RcSWEET gene family, but the former is more influential than the latter. Although the duplicated RcSWEET genes can be derived from a common ancestor, their expression patterns and functions may have diverged. For example, the expression levels of the tandemly duplicated gene pair of RcSWEET10a-RcSWEET12 identified in this work were very different in the various tissues of R. chinensis 'Old Blush'. In Arabidopsis, the unique AtSWEET10 functions for early flowering [21], but there are three RcSWEET10a/b/c genes in rose. Moreover, RcSWEET10a exhibited higher expression levels in flowers, pistils, and ovaries, suggesting that RcSWEET10a may play an important role during the reproductive development. AtSWEET12 was reported to be responsible for phloem loading and seed filling in Arabidopsis [11,25,28]. However, RcSWEET12 was highly expressed in roots, suggesting that it might be involved in root development in roses.
The regulation of gene expression in higher plants usually occurs at the transcriptional level, which is coordinated by various cis-acting elements and trans-acting factors [49]. However, there are few studies on SWEET transcriptional regulation in plants. For example, the NAC transcription factor ORE1 and three abscisic acid (ABA)-responsive element (ABRE)-binding transcription factors, ABF2 (AREB1), ABF3, and ABF4 (AREB2), were found to be involved in directly regulating senescence-associated genes AtSWEET15 by binding to their promoters in Arabidopsis [27,50]. A waterlogging-responsive ERF (MaRAP2-4) from mint (Mentha spp.) was reported to specifically target the DRE or GCC box in the AtSWEEET10 promoter to regulate soluble sugar availability and waterlogging tolerance [51]. The rice transcription factor OsDOF1 is involved in the regulation of pathogen interactions by targeting the promoters of SWEET11/14. In this study, a wide variety of cis-acting elements have been identified in the promoters of RcSWEET genes, including cis-acting elements associated with stress and hormone response, suggesting that family members are involved in complex signaling pathway regulation (Figure 4). The number of ABRE elements was the largest in the whole family. Previous studies showed that the DREB transcription factor could bind to ABRE in the gene promoter region to respond to cold stress, suggesting that RcSWEETs containing these cis elements might be involved in the regulation of the cold response in roses [52]. Moreover, 13 RcSWEETs contained LTR elements, which might directly participate in the regulation of low temperature response.
Low temperature is one of the typical abiotic stress factors for plants, which has an important effect on the growth and development of plants and their geographical distribution [40]. Cold acclimation at non-freezing temperatures can enhance cold resistance and induce many physiological and biochemical changes in plants, such as the accumulation of osmotic regulatory substances, the removal of reactive oxygen species, and the expression of cold responsive genes (CORs) [53][54][55][56]. It is true that the cold-induced increase of the soluble sugar content should be directly associated with the redistribution and balance of sugar in plants through the regulation of sugar transporters [57]. Previous studies showed that SWEET genes could be induced by low temperatures and participate in cold stress responses in many plants. For example, the overexpression of DlSWEET1 in the tropic fruit plant longan (Dimocarpus longan) improved cold tolerance in transgenic Arabidopsis plants [58]. BoSWEET2b, BoSWEET4a, and BoSWEET15b were significantly upregulated in the cabbage leaves after cold stress [13]. A hexose transporter, CsSWEET2, from cucumber (Cucumis sativus), can improve cold tolerance in Arabidopsis [59]. The double mutant of sweet11 sweet12 exhibited significantly increased freezing tolerance in the leaves of Arabidopsis, indicating that AtSWEET11 and AtSWEET12 function negatively under cold stress [60]. The Arabidopsis transgenic plants with the overexpressed AtSWEET16 gene exhibited higher cold resistance than wild type [29]. Moreover, overexpression of MdSWEET16 in apple 'Orin' calli was able to increase their cold tolerance compared with MdSWEET16 RNA interference calli [30]. Enhancing the expression of CsSWEET16/1a/17 in tea plants could improve cold tolerance in the transgenic Arabidopsis plants [14,31]. In addition, more than 90% of MaSWEETs in banana leaves were induced in response to cold stress in two different varieties [17].
In the present study, all eight candidate genes statistically presented significant differences in the leaves or shoots of two rose species during the different hours of cold treatment. Many of the SWEET genes that have been proven to be activated under cold stress in other species were also found in the two rose species studied. For instance, SWEET1 has been recently reported to play an important role in cold resistance in longan [57], which is also enhanced in two rose species after cold stress. There were increased expression levels of the homologous genes CsSWEET2 in cucumber [59], BoSWEET2b in cabbage [13], and RbSWEET2a in rose after cold treatment. Moreover, BoSWEET4a and BoSWEET15b in cabbage [13], as well as RcSWEET4b and RcSWEET15a in rose, could be activated under the cold growth temperature. We also found that the expression level of SWEET10c was significantly enhanced in the tissues of rose species. As far as we know, this is the first report that the SWEET10 gene could be induced under low temperature in plants.
In Arabidopsis, AtSWEET10 acted downstream of FT to promote plant flowering [21]. PbSWEET10 might also contribute to pollen development in the Chinese white pear [61]. Moreover, StSWEET10b was downregulated in the leaves of potatoes after drought treatment [62]. Recently, GmSWEET10a/b were reported to participate in the domestication of seed development in soybean [63].
Analysis using qRT-PCR further demonstrated that the expression levels of SWEET genes varied between the two species during cold treatment. The similar upregulated expression patterns of SWEET1 and SWEET4b were observed in leaves and shoots of both species, implying that these two genes should be relatively conserved for the cold response in the genus Rosa L. Since RbSWEET1 was localized in the plasma membrane, it may function on the cellular level for the accumulation of soluble sugars after cold stress. RbSWEET2a and RbSWEET10c were significantly upregulated in the leaves of R. beggeriana, while RcSWEET2a and RcSWEET10c sharply declined at 1 h and then recovered in the leaves of R. chinensis 'Old Blush' after cold treatment. The results suggested that RcSWEET2a and RcSWEET10c could be the crucial candidate genes for the cold-induced transportation of soluble sugars in the leaves of the deciduous and extremely cold-tolerant species R. beggeriana. As for R. chinensis 'Old Blush', RcSWEET2a and RcSWEET10c might be more sensitive to cold stress than their homologues in R. beggeriana, where low temperatures could initially impair their activities but then recover after the cold adaptation in this evergreen and moderately cold-tolerant species. Moreover, RcSWEET2a was localized in the tonoplast membrane, but RcSWEET10c was localized in the plasma membrane, implying that RbSWEET2a and RbSWEET10c may have different functions for the acquisition of cold tolerance in R. beggeriana.

Identification of the SWEET Family Genes in R. chinensis
The genome sequence and annotation of R. chinensis 'Old Blush' were downloaded from the NCBI genome database (https://www.ncbi.nlm.nih.gov/genome/?term=rosa+ chinensis, accessed on 18 December 2022) [48]. The Hidden Markov Model (HMM) profiles of the SWEET domain (PF03083) were downloaded from the Pfam database (http://pfam. xfam.org/, accessed on 18 December 2022) and used to search the SWEET proteins of R. chinensis with HMMER (3.3.2) software (e-value < 0.01) [64]. In addition, the protein sequences of 17 AtSWEETs [1] were downloaded from TAIR (http://www.arabidopsis.org/, accessed on 18 December 2022) and used as queries to search the R. chinensis proteome by using BLASTP with an e-value threshold of 1 × 10 −5 . Then the results of the two methods are intersected to obtain the candidate sequences, whose conserved domains were further

Gene Structure, Motif, and Phylogenetic Analysis
Based on the genomic sequence and coding sequence, the gene structure of RcSWEET genes was predicted by the Gene Structure Display Server (GSDS; http://gsds.gao-lab.org/, accessed on 18 December 2022) [66]. The conserved motifs in full-length SWEET proteins were identified by MEME (https://meme-suite.org/meme/tools/meme, accessed on 18 December 2022) [67]. Multiple sequence alignments were carried out considering the full-length SWEET protein sequences from A. thaliana, O. sativa, and R. chinensis (Table S3). Subsequently, an unrooted neighbor-joining (NJ) tree was constructed by MEGA 7.0 with 1000 bootstrap replications [68].

Chromosomal Location and Tandem Duplication Analysis
The information on location for each RcSWEET gene was obtained from the GFF genome annotation of R. chinensis 'Old Blush'. The chromosomal localization of the Rc-SWEET genes was mapped using MapChart (2.3.2) software [69]. The MCscanX software was used to search for tandem and duplicated genes. The Ka and Ks values for duplicated gene pairs (Table S4) were calculated based on the coding sequence alignments using the Ka/Ks calculator in TBtools [46].

Cis-Acting Element Analysis
The upstream sequences (2000 bp) of the 25 RcSWEET coding sequences were retrieved and then submitted to PlantCARE (http://bioinformatics.psb.ugent.be/webtools/ plantcare/html/, accessed on 18 December 2022) to identify cis-acting elements [70]. The prediction results of cis-acting elements were illustrated by GSDS.

RNA-Seq Data Analysis
To investigate the expression patterns of RcSWEET genes in different tissues of R. chinensis 'Old Blush', organ specificity transcriptome data (PRJNA546486) of the rose, including roots, stems, leaves, stamens, prickles, pistils, and ovaries, was extracted from the NCBI SRA database (https://www.ncbi.nlm.nih.gov/sra/, accessed on 18 December 2022) [47]. The raw data have been listed in Table S6. FPKM (Fragments Per Kilobase of transcript per Million mapped reads) was used to examine the expression levels, and the heatmap of RcSWEET genes in different tissues was generated using TBtools based on the log 2 FPKM values.
The grafted seedlings of R. beggeriana 'Old Blush' with the rootstock of R. multiflora were placed in a growth chamber with an air temperature of 23 • C day/18 • C night and a photoperiod of 10 h light/14 h dark. After a week of acclimatization, the grafted seedlings were maintained at 4 • C for 0, 1, 6, and 24 h. Current-year shoots of similar thickness and intact leaves were selected as sampling materials, and all the samples were collected in three biological replicates (the data has not been published yet). The Illumina HiSeq2500 highthroughput sequencing platform was used to sequence the cDNA library and obtain the raw data. All high-quality reads of each sample that passed quality control were mapped to the R. chinensis 'Old Blush' RchiOBHm-V2 genome. Gene expression levels were calculated as FPKM. The RbSWEET genes have been identified, and the transcriptome data are listed in Table S7. The heatmap of RbSWEET gene expression profiles was generated using TBtools based on the log 2 FPKM values. DEGs were screened according to certain criteria (fold change ≥ 2 and p-value ≤ 0.05).

Quantitative Real-Time PCR (qRT-PCR)
The cuttage seedlings of R. chinensis 'Old Blush' and the grafted seedlings of R. beggeriana were used for qRT-PCR, and the cold treatment of the two species is the same as in Section 2.6. Total RNA was extracted using the Quick RNA Isolation Kit (Huayueyang Biotech, Beijing, China). The first-strand cDNA was synthesized using the PrimeScript™ RT reagent Kit with gDNA Eraser (Takara Bio Inc., Dalian, Liaoning Province, China). The qRT-PCR reactions were performed using TB Green™ Premix Ex Taq™ II (TaKaRa, Japan) and carried out on a CFX96 Real-Time PCR System (Bio-Rad Laboratories, Hercules, CA, USA). Conditions for the reaction were as follows: 95 • C for 30 s, followed by 40 cycles of 95 • C for 5 s and 60 • C for 30 s. The specific primers for SWEET genes were designed using the Real-time PCR (TaqMan) Primer and Probes Design Tool (https://www.genscript.com/tools/real-time-pcr-taqman-primer-design-tooland, accessed on 18 December 2022) and have been listed in Table S1. Rba-Tubulin was used as the reference gene. All reactions were performed in three biological replicates, and the 2 −∆∆Ct method was applied to calculate the relative expression. GraphPad Prism 8.4.0 (https://www.graphpad.com, accessed on 18 December 2022) software was used to generate graphs and to perform statistical analyses. A p-value of <0.05 was considered to determine the significance level of the data.

Subcellular Localization in Tobacco
The pCAMBIA1300-c-GFP vector was used for the subcellular localization of the RbSWEET genes in Nicotiana benthamiana. The CDS sequences of three RbSWEET genes were amplified using a forward primer containing a HindIII restriction site and a reverse primer containing a KpnI restriction site. The amplification products were recovered using the FastPure Gel DNA Extraction Mini Kit (Vazyme Bio Co., Nanjing, Jiangsu Province, China) and then recombined with the digested plasmid by using the ClonExpress ® Ultra One Step Cloning Kit (Vazyme). The recombined plasmids were later transformed into Agrobacterium tumefaciens strain GV3101. The infection solution was prepared with sterile water containing 10 mM MgCl 2 , 10 mM 2-Morpholinoethanesulfonic Acid (MES), and 0.1 M Acetosyringone (AS), and then the A. tumefaciens was resuspended to an optical density (OD600) of 1.0. After standing in the dark for 3 h, the infection solution was injected into the leaf of N. benthamiana with a syringe. Two days after infiltration, the transient expression position of the RbSWEET proteins was observed by a Leica TCS SP8 fluorescence microscope (Leica Microsystems, Inc., Buffalo Grove, IL, USA) using filter blocks to select for spectral emission at 488 nm, and the empty vector was used as a control.

Conclusions
In this study, 25 SWEET genes were firstly identified and analyzed in R. chinensis 'Old Blush'. Gene structure and motif analysis indicated that members of the SWEET family in the rose are highly conserved among the species. The prediction of Cis-acting elements suggested that some family members may be involved in complex signaling pathway regulation. The transcriptome data showed that there were at least eight DEGs of SWEETs in the leaves and shoots of R. chinensis 'Old Blush' and R. beggeriana after cold treatment. The qRT-PCR results further disclosed that there were various expression patterns of SWEET genes in two species. SWEET1 and SWEET4b were relatively conserved in Rosa L., with similar increased trends both in tissues and species. However, SWEET2a and SWEET10c exhibited the species difference after cold treatment that sharply upregulated in the leaves of the deciduous and extremely cold-tolerant species R. beggeriana but not the evergreen and moderately cold-tolerant species R. chinensis 'Old Blush'. The results indicated that SWEET2a and SWEET10c might be the crucial candidates to participate in the acquisition of cold tolerance in rose. Our findings provide important information for further research on the function analysis of SWEET genes in cold response and their application in the breeding of rose varieties with strong cold tolerance.