Genome-Wide Expression Profiling Analysis of Kiwifruit GolS and RFS Genes and Identification of AcRFS4 Function in Raffinose Accumulation

The raffinose synthetase (RFS) and galactinol synthase (GolS) are two critical enzymes for raffinose biosynthesis, which play an important role in modulating plant growth and in response to a variety of biotic or abiotic stresses. Here, we comprehensively analyzed the RFS and GolS gene families and their involvement in abiotic and biotic stresses responses at the genome-wide scale in kiwifruit. A total of 22 GolS and 24 RFS genes were identified in Actinidia chinensis and Actinidia eriantha genomes. Phylogenetic analysis showed that the GolS and RFS genes were clustered into four and six groups, respectively. Transcriptomic analysis revealed that abiotic stresses strongly induced some crucial genes members including AcGolS1/2/4/8 and AcRFS2/4/8/11 and their expression levels were further confirmed by qRT-PCR. The GUS staining of AcRFS4Pro::GUS transgenic plants revealed that the transcriptionlevel of AcRFS4 was significantly increased by salt stress. Overexpression of AcRFS4 in Arabidopsis demonstrated that this gene enhanced the raffinose accumulation and the tolerance to salt stress. The co-expression networks analysis of hub transcription factors targeting key AcRFS4 genes indicated that there was a strong correlation between AcNAC30 and AcRFS4 expression under salt stress. Furthermore, the yeast one-hybrid assays showed that AcNAC30 could bind the AcRFS4 promoter directly. These results may provide insights into the evolutionary and functional mechanisms of GolS and RFS genes in kiwifruit.


Introduction
The productivity, development and growth of plants can be seriously affected by diverse environmental stresses, such as low or high temperatures, drought, salinity, pests and diseases [1]. Under stress conditions, plants perform a series of physiological and biochemical responses, cellular and molecular mechanisms by activating many stress-responsive genes or transcription factors and synthesizing various functional proteins [2]. At the cellular level, plants produce many compatible molecules solutes, such as mannitol, proline and oligosaccharides (galactinol, trehalose, raffinose and stachyose), which serve as regulatory compounds to deal with these abiotic stresses [3]. Raffinose family of oligosaccharides (RFOs), especially raffinose, plays an essential role in increasing the osmotic pressure in cells. Furthermore, raffinose also acts as antioxidants, signals, transport and storage of carbon, and membrane stabilizer to protect the plant cell from dehydration [4].
It is well known that raffinose and galactinol are two major oligosaccharides. Galactinol synthase (GolS; EC: 2.4.1.123) is responsible for raffinose biosynthesis in the initial stage, which is catalyzed by the reaction of Myo-inositol and UDP-galactose to generate galactinol [5]. Galactinol synthase is an important enzyme regulating the osmoprotectant

Identification, Characterization, and Phylogenetic Analysis of the GolS and RFS Genes in Kiwifruit
Using seven AtGolS proteins from A. thaliana as query sequences, Twenty-two GolS genes in kiwifruit were identified (Additional file 1: Table S1). We named AcGolS1-AcGolS9 in A. chinensis and AeGolS1-AeGolS13 in A. eriantha based on the gene names in A. thaliana. Twenty-four RFS genes were also identified in the kiwifruit genome, and designated as AcRFS1-AcRFS12 in A. chinensis and AeRFS1-AeRFS12 in A. eriantha (Additional file 1: Table S1). The characterization information of the GolS and RFS genes in kiwifruit, including locus ID, linkage group distribution, the length of coding sequences, molecular weight (MW), theoretical isoelectric point (pI), and subcellular localization prediction were listed in Additional file 1: Table S1.
Phylogenetic trees based on four species were created to describe the evolutionary of the RFS and GolS gene families. As shown in Figure 1A, four and two GolS genes from A. eriantha and A. chinensis, respectively, were found in GolS-I. In the clade GolS-II, only AeGolS6 was found. Six members of GolS-III were found, including three GolS genes from

Structure, Protein Motif, and Cis-Element in Promoter Regions of the GolS and RFS Genes
To investigate the exon-intron organization of AcGolS and AeGolS genes, five and four introns were observed in most GolS genes in kiwifruit ( Figure 2A). We saw only one exon in AcGolS6 and AeGolS13 (Figure 2A). AeGolS5 contained the five introns, which had the most significant introns in GolS genes from A. eriantha (Figure 2A). Many GolS genes in the same groups had similar exon-intron structure, which was highly conservative in kiwifruit ( Figure 2A). The numbers of RFS introns varied from 3 to 14 ( Figure 2B). The RFS genes belonging to RFS-III and RFS-IV had more introns than other ( Figure 2B). AeRFS5 contained 14 introns in RFS-III, while other genes contained 13 exons ( Figure 2B). The distribution of exons indicated that RFS genes belonging to the same group had similar gene structures and exon numbers (Figure 2A,B). A total of 10 motifs named motif1 to motif10 were predicted in the kiwifruit RFS and GolS genes. Except for AeGolS5/6/7/8, all GolS genes contained motif2 (Figure 2A). Two motifs were found in AcGolS6 and AcGolS8 (Figure 2A). Motif3 was distributed on GolS-II-IV ( Figure 2A). All motifs were observed in AeRFS1/4/6/10 ( Figure 2B). There were more than seven motifs in most RFS genes in kiwifruit ( Figure 2B). The 2000 bp sequences upstream of the start codon of AcGolS and

Structure, Protein Motif, and Cis-Element in Promoter Regions of the GolS and RFS Genes
To investigate the exon-intron organization of AcGolS and AeGolS genes, five and four introns were observed in most GolS genes in kiwifruit ( Figure 2A). We saw only one exon in AcGolS6 and AeGolS13 (Figure 2A). AeGolS5 contained the five introns, which had the most significant introns in GolS genes from A. eriantha (Figure 2A). Many GolS genes in the same groups had similar exon-intron structure, which was highly conservative in kiwifruit ( Figure 2A). The numbers of RFS introns varied from 3 to 14 ( Figure 2B). The RFS genes belonging to RFS-III and RFS-IV had more introns than other ( Figure 2B). AeRFS5 contained 14 introns in RFS-III, while other genes contained 13 exons ( Figure 2B). The distribution of exons indicated that RFS genes belonging to the same group had similar gene structures and exon numbers (Figure 2A,B). A total of 10 motifs named motif1 to motif10 were predicted in the kiwifruit RFS and GolS genes. Except for AeGolS5/6/7/8, all GolS genes contained motif2 (Figure 2A). Two motifs were found in AcGolS6 and AcGolS8 (Figure 2A). Motif3 was distributed on GolS-II-IV ( Figure 2A). All motifs were observed in AeRFS1/4/6/10 ( Figure 2B). There were more than seven motifs in most RFS genes in kiwifruit ( Figure 2B). The 2000 bp sequences upstream of the start codon of AcGolS and AcRFS genes were used to analyze cis-elements in their promoter regions by PlantCARE. Sixteen putative cis-elements responsive to biotic stresses [including W-box, TC-rich repeats (defense and stress-responsive element), and WUN-motif (wound-responsive element), LTR (low-temperature responsive element), MBS (MYB binding site), and phytohormones such as ABA [abscisic acid, ABRE (ABA-responsive element)], auxin (TGA-element, AuxRRcore), GA (gibberellin, GARE), JA (jasmonic acid, CGTCA-motif), SA (salicylic acid, TCAelement) and endosperm development (P-box) were present in the promoters of GolS and RFS genes in kiwifruit (Additional file 1: Supplementary Figure S1). The same conserved motifs in homologous RFS genes might have similar functions and correlations between evolutionary relationships and conserved motifs.
AcRFS genes were used to analyze cis-elements in their promoter regions by PlantCARE. Sixteen putative cis-elements responsive to biotic stresses [including W-box, TC-rich repeats (defense and stress-responsive element), and WUN-motif (wound-responsive element), LTR (low-temperature responsive element), MBS (MYB binding site), and phytohormones such as ABA [abscisic acid, ABRE (ABA-responsive element)], auxin (TGA-element, AuxRR-core), GA (gibberellin, GARE), JA (jasmonic acid, CGTCA-motif), SA (salicylic acid, TCA-element) and endosperm development (P-box) were present in the promoters of GolS and RFS genes in kiwifruit (Additional file 1: Supplementary Figure S1). The same conserved motifs in homologous RFS genes might have similar functions and correlations between evolutionary relationships and conserved motifs.

Duplication, Synteny, and Chromosomal Distribution of GolS and RFS Genes
The segmental and whole genome duplications of GolS and RFS genes were determined in A. chinensis and A. eriantha. A total of 18 pairs of GolS genes segmental duplicates were observed in the kiwifruit genome ( Figure 3A). All RFS genes within the kiwifruit genome had segmental duplicates, suggesting a high level of conservation of the RFS gene family ( Figure 3B). A comparative syntenic map of A. chinensis, A. eriantha and A. thaliana was constructed (Additional file 2: Figure S2A). Interestingly, we found that 6 AcGol and 9 AeGolS genes had a syntenic relationship with AtGolS genes and located on chr1 (Additional file 2: Figure S2A,B). 7 AcRFS and 8 AeRFS genes had a syntenic relationship with AtRFS genes and located on chr1, chr3, chr4 and chr5 (Additional file 2: Figure S2C,D). All GolS and RFS in kiwifruit were mapped to the 37 chromosomes. The chromosomal distribution information of GolS and RFS genes was shown in Additional file 2: Figure S3.

Duplication, Synteny, and Chromosomal Distribution of GolS and RFS Genes
The segmental and whole genome duplications of GolS and RFS genes were determined in A. chinensis and A. eriantha. A total of 18 pairs of GolS genes segmental duplicates were observed in the kiwifruit genome ( Figure 3A). All RFS genes within the kiwifruit genome had segmental duplicates, suggesting a high level of conservation of the RFS gene family ( Figure 3B). A comparative syntenic map of A. chinensis, A. eriantha and A. thaliana was constructed (Additional file 2: Figure S2A). Interestingly, we found that 6 AcGol and 9 AeGolS genes had a syntenic relationship with AtGolS genes and located on chr1 (Additional file 2: Figure S2A,B). 7 AcRFS and 8 AeRFS genes had a syntenic relationship with AtRFS genes and located on chr1, chr3, chr4 and chr5 (Additional file 2: Figure S2C,D). All GolS and RFS in kiwifruit were mapped to the 37 chromosomes. The chromosomal distribution information of GolS and RFS genes was shown in Additional file 2: Figure S3.

Expression Patterns Analysis of AcGolS and AcRFS Genes in Different Plant Tissues, during the Fruit Development and under Hormone-Induced Conditions
As shown in Figure 4A, six of nine AcGolS g enes (except AcGolS1/3/5) were highly expressed in the leaf. AcGolS3 expression was higher in the stem than in other tissues. AcGolS1 was found to be highly expressed in the shoot. AcGolS2/5/8/9 were mildly expressed in the shoot ( Figure 4A). Only three AcGolS genes (AcGolS3/5/8) had low expression during the fruit development. Six of nine AcGolS genes were induced by abscisic acid ( Figure 4A). AcGolS1/2/5/7 were induced in response to GA (gibberellins). Only AcGolS9 was strongly expressed under salicylic acid (SA) treatments ( Figure 4A).

Expression Patterns Analysis of AcGolS and AcRFS Genes in Different Plant Tissues, during the Fruit Development and under Hormone-Induced Conditions
As shown in Figure 4A, six of nine AcGolS g enes (except AcGolS1/3/5) were highly expressed in the leaf. AcGolS3 expression was higher in the stem than in other tissues. AcGolS1 was found to be highly expressed in the shoot. AcGolS2/5/8/9 were mildly expressed in the shoot ( Figure 4A). Only three AcGolS genes (AcGolS3/5/8) had low expression during the fruit development. Six of nine AcGolS genes were induced by abscisic acid ( Figure 4A). AcGolS1/2/5/7 were induced in response to GA (gibberellins). Only AcGolS9 was strongly expressed under salicylic acid (SA) treatments ( Figure 4A).

Expression Profiles Analysis of AcGolS and AcRFS Genes in Response to Abiotic and Biotic Stresses
To further explore the potential function of AcGolS and AcRFS genes under abiotic and biotic stresses, the expression pattern of 9 AcGolS and 12 AcRFS genes was analyzed through the transcriptome data (Additional file 1: Table S4). As indicated in Figure 5A, AcGolS4/6/7 exhibited similar expression patterns in response to cold stress between coldsensitive A. arguta variety 'Kuilv male' (KL) and cold-tolerant A. arguta variety 'Ruby-3' (RB), and their expression in RB was significantly higher in KL after cold stress, whereas AcGolS2/5/8 exhibited no significant change. Under salt stress, AcGolS2 and AcGolS4 were up-regulated and peaked at 10d in salt-tolerant A. deliciosa variety 'Guichang' (GC), and more highly in salt-tolerant than in salt-sensitive A. chinensis variety 'Hongyang' (HY) ( Figure 5A). Transcript levels of AcGolS1 and AcGolS6/7 in HY were instantaneously upregulated at 10d, but these genes in GC were not significantly affected across all the time points ( Figure 5A). AcGolS3/4/8 were significantly induced at 24 h after Pseudomonas syringae pv. actinidiae (Psa) treatment in resistant A. eriantha variety 'Huate' (HT), and were significantly higher than susceptible A. chinensis variety 'Hongyang' (HY) at all-time points ( Figure 5A). AcGolS6 and AcGolS7 were significantly up-regulated (fold change > 2) from 48 h to 96 h in A. chinensis variety 'Hongyang' (HY), while were relatively low at whole time points in A. eriantha variety 'Huate' (HT) under Psa treatment ( Figure 5A).
AcRFS11 and AcRFS12 were down-regulated under cold stress at whole time points in the A. arguta variety 'Ruby-3' (RB) ( Figure 5B). AcRFS1 and AcRFS8 were significantly induced at 1h in RB while they were no significant changes in A. arguta variety 'Kuilv

Expression Profiles Analysis of AcGolS and AcRFS Genes in Response to Abiotic and Biotic Stresses
To further explore the potential function of AcGolS and AcRFS genes under abiotic and biotic stresses, the expression pattern of 9 AcGolS and 12 AcRFS genes was analyzed through the transcriptome data (Additional file 1: Table S4). As indicated in Figure 5A, AcGolS4/6/7 exhibited similar expression patterns in response to cold stress between cold-sensitive A. arguta variety 'Kuilv male' (KL) and cold-tolerant A. arguta variety 'Ruby-3' (RB), and their expression in RB was significantly higher in KL after cold stress, whereas AcGolS2/5/8 exhibited no significant change. Under salt stress, AcGolS2 and AcGolS4 were up-regulated and peaked at 10d in salt-tolerant A. deliciosa variety 'Guichang' (GC), and more highly in salt-tolerant than in salt-sensitive A. chinensis variety 'Hongyang' (HY) ( Figure 5A). Transcript levels of AcGolS1 and AcGolS6/7 in HY were instantaneously up-regulated at 10d, but these genes in GC were not significantly affected across all the time points ( Figure 5A). AcGolS3/4/8 were significantly induced at 24 h after Pseudomonas syringae pv. actinidiae (Psa) treatment in resistant A. eriantha variety 'Huate' (HT), and were significantly higher than susceptible A. chinensis variety 'Hongyang' (HY) at all-time points ( Figure 5A). AcGolS6 and AcGolS7 were significantly up-regulated (fold change >2) from 48 h to 96 h in A. chinensis variety 'Hongyang' (HY), while were relatively low at whole time points in A. eriantha variety 'Huate' (HT) under Psa treatment ( Figure 5A).
AcRFS11 and AcRFS12 were down-regulated under cold stress at whole time points in the A. arguta variety 'Ruby-3' (RB) ( Figure 5B). AcRFS1 and AcRFS8 were significantly induced at 1h in RB while they were no significant changes in A. arguta variety 'Kuilv male' (KL) after cold stress treatment ( Figure 5B). 80% of the AcRFS genes had the highest expression level at 10d in A. deliciosa variety 'Guichang' (GC) with salt treatment ( Figure 5B). AcRFS1/5/8 were down-regulated under salt stress in both GC and HY varieties. At 24 h, the transcriptional levels of AcRFS5 and AcRFS8 in the A. eriantha variety 'Huate' (HT) were considerably higher than those in the A. chinensis variety 'Hongyang' (HY) in response to Psa ( Figure 5B). The expression levels of AcRFS11 in HT increased from 0 h to 24 h and then decreased ( Figure 5B). The expression of AcRFS11 was 2-fold lower in HY during the whole Pas treated time than in HT at 24 h ( Figure 5B).

Network Analysis of Hub TF Targeting to Key AcGolS and AcRFS Genes
Co-expression network of hub transcription factors targeting key AcGolS and AcRFS genes was constructed under cold, salt and Psa stresses. AcGolS4 under cold treatment, AcGolS2 under salt treatment and AcGolS3 under Psa treatment were selected as key Ac-GolS genes and all transcription factors (2000 TFs) for co-expression analysis. AcRFS4 under cold and salt treatment and AcRFS7 under Psa treatment were selected as key AcRFS gene and all transcription factors (2000 TFs) for co-expression analysis. As shown in Figure 6, the hub AcNAC, AcMYB, AcERF, AcWRKY, AcbHLH and so on transcription factors related to abiotic and biotic stresses were observed. There was a strong correlation between AcNAC30 and AcRFS4 with high expression levels under salt treatment ( Figure 6B). Co-expression network analysis indicated that these hub transcription factors genes in the network might play an important role in responding to the abiotic (cold and salt) and biotic stresses (Psa).

Network Analysis of Hub TF Targeting to Key AcGolS and AcRFS Genes
Co-expression network of hub transcription factors targeting key AcGolS and AcRFS genes was constructed under cold, salt and Psa stresses. AcGolS4 under cold treatment, AcGolS2 under salt treatment and AcGolS3 under Psa treatment were selected as key AcGolS genes and all transcription factors (2000 TFs) for co-expression analysis. AcRFS4 under cold and salt treatment and AcRFS7 under Psa treatment were selected as key AcRFS gene and all transcription factors (2000 TFs) for co-expression analysis. As shown in Figure 6, the hub AcNAC, AcMYB, AcERF, AcWRKY, AcbHLH and so on transcription factors related to abiotic and biotic stresses were observed. There was a strong correlation between AcNAC30 and AcRFS4 with high expression levels under salt treatment ( Figure 6B). Co-expression network analysis indicated that these hub transcription factors genes in the network might play an important role in responding to the abiotic (cold and salt) and biotic stresses (Psa).

Verification of Key AcGolS and AcRFS Genes Expression under Abiotic Stresses
To further validate the expression pattern of AcGolS and AcRFS genes under abiotic stresses, the stress-responsive AcGolS (AcGolS1/2/4/8) and AcRFS (AcRFS2/4/8/11) genes were selected for qRT-PCR analysis. As shown in Figure 7A, AcGolS1 and AcGolS2 were significantly induced after NaCl, drought (DT), and waterlogging (WT) stresses. AcGolS4 showed the highest expression level in response to cold and heat stresses. AcRFS2/4/11were significantly up-regulated in response to cold and NaCl treatments. Only AcRFS8 was strongly expressed under heat and cold treatments. The results confirmed that these genes were really induced by abiotic stresses.

Verification of Key AcGolS and AcRFS Genes Expression under Abiotic Stresses
To further validate the expression pattern of AcGolS and AcRFS genes under abiotic stresses, the stress-responsive AcGolS (AcGolS1/2/4/8) and AcRFS (AcRFS2/4/8/11) genes were selected for qRT-PCR analysis. As shown in Figure 7A, AcGolS1 and AcGolS2 were significantly induced after NaCl, drought (DT), and waterlogging (WT) stresses. AcGolS4 showed the highest expression level in response to cold and heat stresses. AcRFS2/4/11 were significantly up-regulated in response to cold and NaCl treatments. Only AcRFS8 was strongly expressed under heat and cold treatments. The results confirmed that these genes were really induced by abiotic stresses.

Salt Stress Regulating AcRFS4 Promoter
To further investigate salt-induced AcRFS4 expression, the 2000 bp promoter sequence of AcRFS4 was fused to the β-glucuronidase (GUS) coding region and transformed into wild-type Arabidopsis (WT). Analysis of three independent transgenic plants (AcRFS4Pro::GUS-1, -2, and -3) revealed that salt treatment strongly induced GUS expression in the transgenic lines relative to controls (Figure 8). This finding indicated that the transcription of AcRFS4 was enhanced by salt stress.
were selected for qRT-PCR analysis. As shown in Figure 7A, AcGolS1 and AcGolS2 were significantly induced after NaCl, drought (DT), and waterlogging (WT) stresses. AcGolS4 showed the highest expression level in response to cold and heat stresses. AcRFS2/4/11were significantly up-regulated in response to cold and NaCl treatments. Only AcRFS8 was strongly expressed under heat and cold treatments. The results confirmed that these genes were really induced by abiotic stresses.

Salt Stress Regulating AcRFS4 Promoter
To further investigate salt-induced AcRFS4 expression, the 2000 bp promoter sequence of AcRFS4 was fused to the β-glucuronidase (GUS) coding region and transformed into wild-type Arabidopsis (WT). Analysis of three independent transgenic plants (AcRFS4Pro::GUS-1, -2, and -3) revealed that salt treatment strongly induced GUS expression in the transgenic lines relative to controls (Figure 8). This finding indicated that the transcription of AcRFS4 was enhanced by salt stress.

Functional Characterization of AcRFS4 in Response to Salt Stress
To assess the role of AcRFS4 under salt stress, three transgenic homozygous T3generation Arabidopsis (OE1, OE2 and OE3) were obtained. As shown in Figure 9A, wildtype (WT) and transgenic plants grew well, without obvious difference, under normal conditions (control, CK). Under 150 mmol L −1 NaCl, transgenic plants survived but WT were very small and their leaves turned yellow ( Figure 9B). Quantitative measurements indicated that the root lengths of transgenic plants were significantly longer than wildtype under salt stress ( Figure 9C). These results indicated that AcRFS4 over-expression indeed increased the tolerance to salt stress in transgenic Arabidopsis. Moreover, the raffinose contents in OE plants were significantly higher than that of WT under control

Functional Characterization of AcRFS4 in Response to Salt Stress
To assess the role of AcRFS4 under salt stress, three transgenic homozygous T3generation Arabidopsis (OE1, OE2 and OE3) were obtained. As shown in Figure 9A, wildtype (WT) and transgenic plants grew well, without obvious difference, under normal conditions (control, CK). Under 150 mmol L −1 NaCl, transgenic plants survived but WT were very small and their leaves turned yellow ( Figure 9B). Quantitative measurements indicated that the root lengths of transgenic plants were significantly longer than wild-type under salt stress ( Figure 9C). These results indicated that AcRFS4 over-expression indeed increased the tolerance to salt stress in transgenic Arabidopsis. Moreover, the raffinose contents in OE plants were significantly higher than that of WT under control condition ( Figure 9D), suggesting that AcRFS4 has a conserved role in raffinose biosynthesis. Under salt stress, OE plants accumulated even more raffinose than WT ( Figure 9D), indicating that AcRFS4 increased the salt tolerance possibly by enhancing raffinose accumulation.

Subcellular Localization of the AcNAC30 Protein
The transient expression of 35S::GFP plasmid showed that GFP fluorescence was served throughout the cells, while the green fluorescence signals of the 35S::AcNAC GFP plasmid were specifically found in the nucleus ( Figure 10). This result indicated AcNAC30 might encode a nuclear-localized protein, which was consistent with its fu tional characteristics in regulating gene transcription.

Subcellular Localization of the AcNAC30 Protein
The transient expression of 35S::GFP plasmid showed that GFP fluorescence was observed throughout the cells, while the green fluorescence signals of the 35S::AcNAC30-GFP plasmid were specifically found in the nucleus ( Figure 10). This result indicated that AcNAC30 might encode a nuclear-localized protein, which was consistent with its functional characteristics in regulating gene transcription.

AcNAC30 a Key Transcription Factor Regulating AcRFS4
The transcript level of AcNAC30 was significantly up-regulated upon salt application ( Figure 11A). Compared with control, the transcript level of AcNAC30 was nearly 25 times higher after salt treatment ( Figure 11A). Promoter self-activation of Y1H Gold showed that ≥150 ng mL −1 aureobasidin A (AbA) inhibited the growth of Y1H Gold containing the pABAi-AcRFS4-Pro recombinant plasmid ( Figure 11B). Only the positive control strain and transformed Y1H Gold harboring both pABAi-AcRFS4-Pro and pGADT7-AcNAC30 could grow in a medium without leucine (-Leu) (150 ng mL −1 AbA), validating protein-DNA interaction of AcNAC30 and the AcRFS4 promoter ( Figure 11B).

AcNAC30 a Key Transcription Factor Regulating AcRFS4
The transcript level of AcNAC30 was significantly up-regulated upon salt application ( Figure 11A). Compared with control, the transcript level of AcNAC30 was nearly 25 times higher after salt treatment ( Figure 11A). Promoter self-activation of Y1H Gold showed that ≥150 ng mL −1 aureobasidin A (AbA) inhibited the growth of Y1H Gold containing the pABAi-AcRFS4-Pro recombinant plasmid ( Figure 11B). Only the positive control strain and transformed Y1H Gold harboring both pABAi-AcRFS4-Pro and pGADT7-AcNAC30 could grow in a medium without leucine (-Leu) (150 ng mL −1 AbA), validating protein-DNA interaction of AcNAC30 and the AcRFS4 promoter ( Figure 11B).

AcNAC30 a Key Transcription Factor Regulating AcRFS4
The transcript level of AcNAC30 was significantly up-regulated upon salt application ( Figure 11A). Compared with control, the transcript level of AcNAC30 was nearly 25 times higher after salt treatment ( Figure 11A). Promoter self-activation of Y1H Gold showed that ≥150 ng mL −1 aureobasidin A (AbA) inhibited the growth of Y1H Gold containing the pABAi-AcRFS4-Pro recombinant plasmid ( Figure 11B). Only the positive control strain and transformed Y1H Gold harboring both pABAi-AcRFS4-Pro and pGADT7-AcNAC30 could grow in a medium without leucine (-Leu) (150 ng mL −1 AbA), validating protein-DNA interaction of AcNAC30 and the AcRFS4 promoter ( Figure 11B).

Discussion
Raffinose, the smallest member of RFOs, is widely found in the leaves, roots, seeds and tubers of plants, for instance, Arabidopsis thaliana [36], Brassica napus [37], Nicotiana tabacum [13], Zea mays [14], Cicer arietinum [38], Gossypium hirsutum [39], Sesamum indicum [17] and so on. As two significant oligosaccharides in the raffinose family, raffinose and galactinol play a vital role in responding to abiotic and biotic stresses in the plant [4]. Galactitol and raffinose synthase enzymes are critical for raffinose synthesis [5]. To date, there were a few reports regarding GolS and RFS genes in horticultural crops, such as cucumber [40], peas [41], apple [12], banana [42], tea [7] and grapevine [24]. Kiwifruit is an important horticultural crop with enriching in vitamin C [26]. Here, we identified 22 GolS and 24 RFS genes from the genome sequence of A. chinensis variety 'Hongyang' and A. eriantha variety 'White'. As shown in Figure 1A, phylogenetic analysis showed that 21 GolS genes were classified into four subgroups ( Figure 1A), which is consistent with that previously reported in other species, such as cotton [19], tomato [16], tobacco [13], and sesame [17]. As shown in Figure 1B, phylogenetic analysis showed that 24 RFS genes were classified into six subgroups, consistent with cotton [19]. A similar exon/intron structure with 1 to 5 and conserved motifs of most RFS gene members shared in the same group was found in kiwifruit, which was also supported by phylogenetic relationships ( Figure 2B). In short, these results showed that RFS gene family in kiwifruit might be relatively conservative during evolution.
The previous studies showed that gene family expansion and evolution of new functions frequently occurred during gene duplication, particularly in adaptation to abiotic and biotic stresses [43]. The segmental and tandem gene duplications as two significant factors in gene family generation and maintenance may correspond to functional differences among gene family members during gene evolution [44]. In the current study, only one tandem duplication event of GolS gene family was found in A. chinensis, while it was not found in A. eriantha (Additional file 2: Figure S3). This result is consistent with that previously reported in Casava [15]. High segmental duplications (more than 80% of genes) were observed in kiwifruit GolS and RFS gene families (Figure 3), as suggested by previous findings in Solanum lycopersicum and Brachypodium distachyon [16]. Segmental duplication exhibited exceptionally different expression patterns. There was often a strong link between the function and expression pattern of segmental repeat genes. AcGolS4/6/7 genes belonging to segmentally duplicated gene pairs in the same subgroup ( Figure 1) respond to cold, salt and Psa stresses ( Figure 5A). Together, these results indicated that gene duplication might play potential roles in providing genetic sources with novel biological functions during the evolution of the kiwifruit GolS and RFS gene family.
It has been reported that GolS and RFS genes may play a key role in synthesizing galactinol or raffinose and regulating the response to abiotic and biotic stresses [4]. There are still no papers investigating GolS and RFS gene-mediated tolerance to abiotic and biotic stresses in kiwifruit despite extensive researches on other plant species. Comprehensive expression profiles of AcGolS and AcRFS genes under different hormone and abiotic stresses treatments were analyzed via transcriptome data from this study and previous studies (Additional file 1, Table S4). AcGolS2 and AcGolS3 were slightly induced by JA treatments (Figure 4A), which is consistent with the report by Li et [15], Under high concentrations of ABA treatment, galactitol, raffinose contents and galactitol synthase activities were significantly higher than those in control, indicating that exogenous ABA induces the accumulation of RFO in somatic embryos of alfalfa (Medicago sativa L.) [45]. In this study, we found that most AcGolS and AcRFS genes were significantly activated by ABA ( Figure 4A), suggesting that ABA signaling may regulate these genes. Seki et al. (2002) suggested that AtGolS1 and AtGolS2 slightly induced ABA, while ABA does not induce AtGolS3 in A. thaliana [46]. Promoter analysis of the Chickpea CaGolS gene revealed that CaGolS2 was more sensitive to ABA treatment, which might be related to the ABRE core sequence (ACGT) enriched in its promoter sequence [47]. Under ABA treatment, ZmVP1 and ZmABI5 interacted to regulate ZmGolS2 expression to promote raffinose accumulation in maize seeds [48]. So far, few articles have shown that ABA treatment can promote the expression of RFS genes. Notably, the gene expression up-regulation of six AcRFS genes (AcRFS1/4/6/7/10) was induced transcriptionally in ABA treatment ( Figure 4B), suggesting that activation of AcRFS gene might be positively correlated with ABA levels. Exogenous ABA-treated improved levels of VvRafs1 gene expression in grapevine buds [49]. Unsurprisingly, abiotic stresses induced GolS and RFS genes in plants [50]. Over-expression of CsGolS1 in cucumber enhanced the assimilate translocation efficiency and accelerated the growth rates of sink leaves, fruits and whole plants under cold stress [51]. The normal growth and physiological processes in kiwifruit were seriously affected under salt stress [28]. Under salt stress, AcGolS2 and AcGolS4 up-regulated and peaked at 10d in salt-tolerant A. deliciosa variety 'Guichang' (GC) and more highly than in salt-sensitive A. chinensis variety 'Hongyang' (HY) (Figure 5A), which suggested that they were saltresponsive AcGolS genes. As above-mentioned, MeGolS1 reached a rapid peak expression at 12 h in response to salt conditions [15]. TsGolS2 was up-regulated under salt stress, and overexpression of TsGolS2 enhanced tolerance to salt in Arabidopsis [52]. In addition, overexpression of PtrGolS3 resulted in higher RFO content and other stress-related metabolites (proline, salicylic acid, amino acids, etc.) compared with wild type, which may increase RFO in woody plants under short-term salt treatment understanding of metabolism [53]. In Arabidopsis, AtRS5 (At5g40390) was reported to participate in abiotic stresses using a reverse genetic approach [22]. ZmRFS overexpressing Arabidopsis plants displayed a significantly tolerance to drought stress [54]. In this study, the expression levels of AcRFS2/4/8/11 were higher in abiotic stress-tolerant (cold and salt) kiwifruit cultivars and significantly higher than those in abiotic stress-insensitive kiwifruit cultivars ( Figure 5B), which suggests that these AcRFS genes might be positively involved in cold and salt tolerances of kiwifruit. A previous study indicated that Na + is toxic for kiwifruit [55]. Therefore, we can't exclude the possibility that the increased expression of AcRFS genes under salt stress might be caused by the toxicity of Na + , although there is no report indicating that Na + is toxic to all kiwifruit varieties. Besides, our study found that overexpression of AcRFS4 in Arabidopsis enhanced the salt tolerance of transgenic plants. More raffinose was significantly accumulated in AcRFS4-OE Arabidopsis lines than in WT under salt stress ( Figure 9D). Therefore, we speculate that overexpression of AcRFS4 would accumulate more raffinose for combating with the salt stress in kiwifruit. So far, a few studies have shown how GolS and RFS genes are transcriptionally regulated under abiotic and biotic stresses. The BhWRKY1 transcription factor was involved in dehydration tolerance by regulating BhGolS1 gene in Boea hygrometrica [56]. ZmDREB1A directly regulated ZmRFS to improve raffinose biosynthesis and enhance plant tolerance to cold stress [57]. Many elements related to phytohormones, biotic and abiotic stresses were observed in the upstream 2000 bp promoter of the GolS and RFS genes in A. chinensis and A. eriantha (Additional file 2: Figure S1). There was a strong correlation between AcNAC30 and AcRFS4 by WGCNA analysis under salt ( Figure 6). AvNAC030 can increase plants' salt tolerance by improving ROS removal efficiency and maintaining the intracellular and extracellular osmotic balance to protect the integrity of the membrane [58]. By qRT-PCR analysis, it was shown that the transcription of both AcNAC30 and AcRFS4 was significantly induced under salt stress (Figures 7B and 11A). Apart from this, GUS staining result of AcRFS4 promoter revealed that the transcription of AcRFS4 was induced by salt stress (Figure 8). Considering the results of Y1H, we believe that the transcription factor AcNAC30 may activate AcRFS4 expression by directly binding its promoter ( Figure 11B). This work provides a foundation for future investigation of AcGolS, AcRFS and AcNAC30 gene functions for kiwifruit's abiotic and biotic stresses tolerance.

Identification and Phylogenetic Analysis of GolS and RFS Genes
The A. chinensis variety 'Hongyang' v3.0 and A. eriantha variety 'White' genome sequences were obtained from the Kiwifruit Genome Database (http://kiwifruitgenome. org/ (accessed on 4 May 2021)) [59]. First, the BLASTP tool was used to screen AtGolS and AtRFS genes in Arabidopsis against all of the protein sequences in each kiwifruit genome, using an E-value threshold of 1.0. Secondly, the profile hidden Markov model (PF05691 and PF01501) of the Pfam 32.0 database (http://pfam.xfam.org/ (accessed on 2 June 2021)) was further applied to confirm the hits under an E-value <1.0 [60]. We removed the genes, which did not include conserved domains of GolS and RFS genes. Finally, these remaining candidate protein sequences were used to calculate physicochemical properties by the online tool ProtParam, including molecular weight (MW), theoretical isoelectric point (pI), and hydrophilic mean (GRAVY). The possible subcellular location of GolS and RFS genes in kiwifruit was predicted using the WolfPSORT tool [61]. Multiple protein sequences were aligned by using MUSCLE [62]. The Neighbor-Joining (NJ) phylogenetic trees were constructed using MEGA software with 1000 bootstrap replicates [63]. AtGolS and AtRFS amino acid sequences were downloaded from TAIR (https://www.arabidopsis. org/ (accessed on 2 August 2021)) in this study.

Analysis of Gene Structure, Protein Motif and Cis-Element in Promoter Regions
The gene structures were presented using the online Gene Structure Display Service (http://gsds.cbi.pku.edu.cn (accessed on 10 August 2021)) [64]. The conserved motifs were displayed by using the online tool MEME Suite5.1.1 (version 5.1.1; http://meme-suite.org/ (accessed on 16 August 2021)) [65]. The gene promoters from the initiation codon (2000 bp before ATG) were retrieved from the kiwifruit genome. Then, the putative cis-regulatory elements in the promoter region sequences were predicted via the PLACE database (http: //www.dna.affrc.go.jp/PLACE/ (accessed on 30 August 2021)).

Analysis of Gene Duplication, Synteny and Chromosomal Locations
The potential gene duplication and events synteny relationships between kiwifruit and A. thaliana of GolS and RFS gene families were determined using MSCanX software [66]. The physical location maps of chromosomes were drawn using MapInspect 1.0 software.

Expression Analysis of AcGolS and AcRFS Genes
The RNAseq raw sequence data of 78 samples covering diverse tissues at different developmental time points were downloaded from the NCBI website (Bioproject ID PR-JNA324539) [67]. The raw transcriptome data of two A. arguta genotypes, 'Kuilv male' (KL) and 'Ruby-3' (RB), with high and low freezing tolerance, respectively, at −25 • C for 0 h, 1 h, and 4 h were obtained from the NCBI database, with project number PRJNA248163 [68]. We also downloaded the RNA sequencing raw data of resistant A. eriantha variety 'Huate' and susceptible A. chinensis variety 'Hongyang' at 0, 12, 24, 48, and 96 h after inoculation with Psa from the NCBI (Bioproject ID PRJNA514180) [31]. The rest of the transcriptome raw data (unpublished) came from our own research group. Trimmomatic was performed to filter the raw sequence datas [69]. The clean datas were adopted to map the reference A. chinensis variety 'Hongyang' v3.0 genome using HISAT2 [70]. The genes were quantified with the featureCounts package in R. The heatmaps were constructed using the TBtools software [71].

Co-Expression Network Analysis of Hub Transcription Factors
Co-expression network hub transcription factors targeting key AcRFS and AcGolS genes were analyzed by using the 'cor' function of the R package with Pearson's correlation coefficients (r ≥ 0.9 or r ≤ −0.9 and p-values ≤ 0.05) under cold, salt and Psa stresses conditions. Cytoscape 3.4.0 software (http://www.cytoscape.org/ (accessed on 30 August 2021)) was used to visually describe co-expression network results [72]. The degree >N50 was used to identify the genes of hub transcription factors.

Plant Material and Abiotic Stresses Treatments
A. chinensis variety 'Hongyang' (HY) tissue culture seedlings were transferred to a soil mixture of perlite and sand (3:1, v/v). All seedlings were grown in a growth chamber at a temperature of 18 • C (night) and 24 • C (day), relative humidity of 60-80%, and a 14/10 h photoperiod (daytime, 06:00-20:00). The seedlings were irrigated with water once every two days. After two months, they were randomly divided into six groups for stresses treatments. For heat and cold stress treatment, the seedlings were transferred into two chambers with the temperature set at 48 • C and 4 • C, respectively [73]. Treated seedlings were harvested at 6 h after treatment. For salt, drought and waterlogging stresses, the seedlings were soaked in 0.6% NaCl for 6 days [28]. the seedlings were flooded for 7 days [74], and the seedlings were dried for 14 days [75]. Non-treated seedlings were used as the control (CK). All samples were immediately frozen in liquid nitrogen and stored at −80 • C.

RNA Isolation and Quantitative Real-Time PCR (qRT-PCR) Analysis
According to the manufacturer's instructions, total RNA was extracted from all samples using the RNAprep pure Plant Kit (TIANGEN, Beijing, China). For qRT-PCR analysis, first-strand cDNAs were synthesized from DNaseI-treated total RNA using the Hifair®III 1st Strand cDNA Synthesis kit (Yeasen, Shanghai, China) according to the manufacturer's instructions. qRT-PCR was performed on the Biorad CFX96 real-time PCR system using the ChamQ SYBR qPCR Master Mix (Vazyme, Nanjing, China). The relative expression levels were calculated by using 2 −∆∆Ct method. The qRT-PCR assays were performed with three biological and technical replicates. The gene-specific primers are listed in Supplementary Additional file 1: Table S2.

Isolation of the AcNAC30 and AcRFS4 Full-Length CDS, and AcRFS4 Promoter Sequence
The CTAB method was implemented to extract the total DNA of A. chinensis variety 'Hongyang' (HY)leaves [76]. Transcription factor AcNAC30 was identified based on the sequence in the kiwifruit database (http://kiwifruitgenome.org/ (accessed on 6 October 2021), Accession Nos. Actinidia09980). AcNAC30 and AcRFS4 full-length CDS were cloned from reverse-transcribed cDNA with specific primers (Additional file 1: Table S3). The AcRFS4 promoter was cloned from extracted DNA with AcRFS4-Pro primers (Additional file 2: Supplementary Table S3). The cloning conditions were as follows: predenaturation for 3 min at 98 • C; 30 cycles of template denaturation for 10 s at 98 • C, primer annealing for the 60 s at 58 • C, and 30 s at 72 • C; and a final extension step of an additional 5 min at 72 • C. The amplified products were connected to the pESI-T vector and transferred into DH5α Chemically Competent Cell, and the recombinant plasmids were obtained through white spot screening.

Analysis of AcRFS4 GUS Histochemical Staining and AcNAC30 Protein Subcellular Localization Assay
The 20,000 bp AcRFS4 promoter sequence was inserted into pCAMBIA1301 to drive the GUS reporter gene by using BglII and BamHI. AcRFS4Pro::GUS transgenic Arabidopsis of 10-day-old seedlings in MS/2 media. Then, transgenic Arabidopsis were treated with 150 mM NaCl for 24 h. Finally, transgenic Arabidopsis were used for histochemical staining. GUS staining was performed according to Jefferson et al. [77]. Primers for construction of the AcRFS4Pro::GUS vector were provided in Additional file 1: Table S3.
The CDS of AcNAC30 without the stop codon was constructed into the vector pCAMBI A1305-35S-GFP vector using double enzyme digestion (cut with Xbal and BamHI). Then, the plasmid (35S::AcNAC30-GFP) was transferred into Agrobacterium strain EHA105. This strain was injected into 1-month-old tobacco leaves. Finally, the GFP fluorescence was detected by confocal laser scanning microscopy. Primers used to construct the 35S::AcNAC30-GFP vector were provided in Additional file 1: Table S3.

Transcription Activation Assay in Yeast
The yeast one-hybrid (Y1H) assays were conducted based on the method previously described [78]. The AcRFS4 promoter sequence containing three core cis-elements (CATGT binding site) was inserted into a pAbAi vector via the HindII and SalI sites (pABAi-AcRFS4-Pro primers in Additional file 1: Table S3). The recombinant vector of pABAi-AcRFS4-Pro was linearized with BstBI and transferred into Y1H Gold through PEG/LiAc. The fulllength region of AcNAC30 was embedded into the pGADT7 vector (AD) through EcoRI and SalI sites (pGADT7-AcNAC30 primers in Additional file 1: Table S3). Transformed Y1H Gold harboring pABAi-AcRFS4-Pro and pGADT7-AcNAC30 were cultured to test the interaction on SD/-Leu with aureobasidin A (150 ng mL −1 AbA) for three days at 30 • C. Finally, an autoactivation analysis was conducted according to the manufacturer's protocol. The p53-promoter fragment and pGADT7-Rec (AD-Rec-P53) co-transform Y1H Gold as a positive control. Empty pGADT7 was transformed in Y1H Gold with pABAi-AcRFS4-Pro as a negative control.

Plasmid Construction, Transgenic Arabidopsis Generation and Salt Stress Tolerance
Complete CDS of AcRFS4 was cloned between the NcoI and BstEII sites of the pCAM-BIA1302 vector with the In-Fusion and a pair of primers (Additional file 1: Table S3). The AcRFS4-1302 vector was initially transformed to Agrobacterium strain GV3101. Finally, it was transformed to Arabidopsis using the floral dip method [79]. The AcRFS4-1302 transgenic Arabidopsis of T1, T2, and T3 generations was selected on 1/2 MS medium with 25 mg·L −1 Hygromycin B (HYG). T3 homozygous of these transgenic lines were used for salt stress tolerance assays. The seeds of WT and homozygous T3 transgenic Arabidopsis (OE1, OE2 and OE3) were separately placed on 1/2 MS medium containing 150 mmol·L −1 NaCl. After 7 days, the root lengths of WT and transgenic Arabidopsis was measured by using Image J 1.8.0. The leaves of WT and transgenic Arabidopsis under salt stress treatments were collected, then raffinose were extracted from WT and transgenic Arabidopsis leaves according to Li et al. [80]. Raffinose was determined by high performance liquid chromatography (HPLC) with amide column (4.6 mm × 150 mm id., 5 µm;). A Waters X-bridge amide column (Waters, USA) was washed by methanol/H 2 O (90:10) as the mobile phase at speed of 0.5 mL/min for separation of soluble sugar components. An evaporative light-scattering detector (ELSD; Waters 2424) was applied to monitor the sugar signal.

Conclusions
The galactinol synthase (GolS) and raffinose synthetase (RFS) are two critical enzymes synthesizing raffinose, which play an important role in modulating plant growth and a variety of biotic or abiotic stresses. A total of 22 GolS and 24 RFS genes were identified in A. chinensis and A. eriantha, respectively. We comprehensively analyzed the GolS and RFS gene families and their involvement in kiwifruit's abiotic and biotic stresses responses at the genome-wide scale. RNA-seq and qRT-PCR analyses revealed that AcGolS1 and AcGolS2 were significantly induced after NaCl, DT, and WT stresses. AcRFS4 was significantly up-regulated in response to NaCl treatments. The GUS staining result revealed that the transcriptional regulation of AcRFS4 was influenced by salt stress. Overexpression of AcRFS4 in Arabidopsis proved that this gene enhanced the salt tolerance of transgenic plants at physiologic (higher raffinose content). A yeast one-hybrid assay demonstrated that AcNAC30 could interact with the AcRFS4 promoter. Therefore, we speculate that AcNAC30 promotes the expression of AcRFS4 gene to increase the accumulation of raffinose, thereby improving the salt stress tolerance of kiwifruit.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The datasets generated for this study can be found in SRA, the accession number: PRJNA691387, PRJNA681641 and PRJNA514180.

Conflicts of Interest:
The authors declare no conflict of interest.