Genome-Wide Identification of the A20/AN1 Zinc Finger Protein Family Genes in Ipomoea batatas and Its Two Relatives and Function Analysis of IbSAP16 in Salinity Tolerance

Stress-associated protein (SAP) genes—encoding A20/AN1 zinc-finger domain-containing proteins—play pivotal roles in regulating stress responses, growth, and development in plants. They are considered suitable candidates to improve abiotic stress tolerance in plants. However, the SAP gene family in sweetpotato (Ipomoea batatas) and its relatives is yet to be investigated. In this study, 20 SAPs in sweetpotato, and 23 and 26 SAPs in its wild diploid relatives Ipomoea triloba and Ipomoea trifida were identified. The chromosome locations, gene structures, protein physiological properties, conserved domains, and phylogenetic relationships of these SAPs were analyzed systematically. Binding motif analysis of IbSAPs indicated that hormone and stress responsive cis-acting elements were distributed in their promoters. RT-qPCR or RNA-seq data revealed that the expression patterns of IbSAP, ItbSAP, and ItfSAP genes varied in different organs and responded to salinity, drought, or ABA (abscisic acid) treatments differently. Moreover, we found that IbSAP16 driven by the 35 S promoter conferred salinity tolerance in transgenic Arabidopsis. These results provided a genome-wide characterization of SAP genes in sweetpotato and its two relatives and suggested that IbSAP16 is involved in salinity stress responses. Our research laid the groundwork for studying SAP-mediated stress response mechanisms in sweetpotato.


Introduction
Stress-associated protein (SAP) genes, encoding A20/AN1 zinc-finger domain-containing proteins, are considered good candidates to improve abiotic stress tolerance in crops [1,2]. The first SAP gene identified in plants was OsiSAP1 from elite indica rice (Oryza sativa L. var. Pusa Basmati-1) [3]. Members of this family contain one or two types of special zinc-finger domains, A20 and/or AN1. The A20 domain was first found in the A20 protein from human umbilical vein endothelial cells, and its conserved structure is CX 2-4 CX 11 CX 2 C (C stands for cysteine residue and X stands for other amino acid residues) [4]. The AN1 domain was first identified in the protein encoded by animal hemisphere 1 (AN1) from Xenopus laevis [5], and its conserved structure in plants is CX 2/4 CX 9-12 CX 1-2 CX 4 CX 2 HX 5 HXC (H stands for histidine residue) [6].
It has been shown that many SAP genes are involved in abiotic stress responses in plants. Overexpression or ectopic expression of most SAPs enhanced stress tolerance in

Identification of SAPs in Sweetpotato and Its Two Diploid Wild Relatives
To identify SAP genes in sweetpotato, a scan was carried out using the hidden Markov models of A20/AN1 zinc fingers against the sweetpotato genome [29] and transcriptome databases (unpublished data). A total of 38 candidate sequences were obtained. Based on CD-search results, 20 of them were validated as IbSAP members and named as IbSAP1 through IbSAP20. Most IbSAP genes were identified in both the genome and transcriptome, but IbSAP13 and IbSAP17 were only identified in the transcriptome. IbSAP4, IbSAP11, and IbSAP18 had two copies, while the others had only one copy. Coding sequence (CDS) lengths of IbSAPs ranged from 429 bp (IbSAP8) to 885 bp (IbSAP17), while the molecular weight (Mw) of IbSAP proteins ranged from 15.40 kDa (IbSAP20) to 32.56 kDa (IbSAP17). The predicted grand average of hydropathicity (GRAVY) index of IbSAPs ranged from −0.935 (IbSAP9) to −0.346 (IbSAP2), and the isoelectric point (pI) distribution of IbSAPs was from 7.49 (IbSAP4 and IbSAP5) to 9.48 (IbSAP14), which indicated that IbSAPs are hydrophilic alkaline proteins (GRAVY < 0 and pI > 7). LOCTREE 3 website analysis suggested that most IbSAPs were located in the cytoplasm (IbSAP1-14, IbSAP18-20); however, IbSAP15 and IbSAP16 might be located in the endoplasmic reticulum and IbSAP17 in the nucleus (Table 1). We also screened the genomes of Ipomoea triloba and Ipomoea trifida, two relative species of sweetpotato [30] and identified 23 and 26 SAP genes, respectively. They were named ItbSAP/ItfSAP1-22(L) (Like) according to their phylogenetic relationship with IbSAPs (Table S1). Among these SAP genes, ItbSAP7, ItbSAP18, ItbSAP22, ItfSAP7L, ItfSAP18, and ItfSAP22 were multicopy genes. Each IbSAP and their orthologs in I. triloba and I. trifida share similar amino acid sequences with a coverage ranging from 98.01% to 100% and a gap percentage ranging from 0.00% to 3.98%. Orthologs from the three Ipomoea species share a high degree of identity (90.73% to 100%) and positive amino acids (92.72% to 100%) (Table S2). These results indicated that SAPs are highly conserved in I. batatas and its wild relatives, I. triloba and I. trifida. We did not find any orthologs of two ItbSAPs (ItbSAP21/22) and two ItfSAPs (ItfSAP21/22) in the sweetpotato genome and/or transcriptome. IbSAP18, ItbSAP18, and ItfSAP18 had two gene copies each (Tables 1 and S1), which indicated that SAP18 duplicated in the two relatives or their ancestor(s).

Phylogenetic Analysis of IbSAPs and SAPs from Some Other Plants
To investigate the phylogenetic relationship between IbSAPs and their orthologs in the two relative species of sweetpotato as well as other plants, a total of 125 SAP protein sequences were used to build an unrooted neighbor-joining phylogenetic tree ( Figure 2). Similar to the phylogenetic relationship of 14 AtSAPs and 18 OsSAPs reported before [2], all SAPs examined fall into four clades. In each clade, numbers of SAPs in each species were as follows (total: I. batatas, I. triloba, I. trifida, A. thaliana, O. sativa, S. lycopersicum, Z. mays): clade I (7, 8, 9, 1, 6, 4, 4); clade II (5, 7, 7, 8, 4, 1, 4); clade III (5, 5, 7, 1, 3, 4, 2); clade IV (3,4,3,4,5,2,3). Each specific clade included SAPs of both monocotyledons (O. sativa and Z. mays) and dicotyledons (I. batatas, I. triloba, I. trifida, A. thaliana, and S. lycopersicum) plants. However, some subclades were monocot specific (such as ZmAN11 to OsSAP8 in clade I or ZmAN15 to OsSAP15 in clade III) or dicot specific (such as SlSAP4 to ItfSAP1 in clade I or ItfSAP10 to ItbSAP12 in clade III). These results indicated that SAP proteins appeared before the divergence of monocot and dicot plants and have been evolving after species divergence. The number and distribution of IbSAPs was most similar to the two wild relative species. However, there were more SAP members in I. triloba and I. trifida than in I. batatas. As mentioned above, no SAPs in I. batatas were orthologous to ItfSAP21/22 and ItbSAP21/22. Moreover, two or more sequences in I. trifida or I. triloba were orthologous to certain IbSAP members. Two sequences (ItfSAP3 and ItfSAP3L) in I. trifida were orthologous to IbSAP3. Two sequences (ItfSAP7 and ItfSAP7L) in I. trifida and two sequences (ItbSAP7 and ItbSAP7L) in I. triloba were orthologous to IbSAP7. Three sequences (ItfSAP9, ItfSAP9L1, and ItfSAP9L2) in I. trifida were orthologous to IbSAP9 ( Figure 2). That fewer SAPs were identified in sweetpotato than in its two diploid relatives might be due to some alteration in chromosome features that occurred during or after the evolution of sweetpotato, or due to the incomplete sweetpotato genome sequence.

Gene and Protein Structures of IbSAPs, ItbSAPs, and ItfSAPs
The gene structures of Ib/Itb/ItfSAPs ( Figure 3b) and conserved domains of their encoding proteins (Figures 3c and S1) were analyzed sorted based on the phylogenetic tree of IbSAPs, ItbSAPs, and ItfSAPs ( Figure 3a). The gene structures of IbSAPs were similar to those of ItbSAPs and ItfSAPs, and were conserved in each clade. In clade I, most SAPs have one or two introns in their 5′-untranslated regions (5′-UTRs), except that IbSAP1, ItfSAP3L, IbSAP7, and ItbSAP7 have no intron, and that ItfSAP7L has two introns in the

Gene and Protein Structures of IbSAPs, ItbSAPs, and ItfSAPs
The gene structures of Ib/Itb/ItfSAPs ( Figure 3b) and conserved domains of their encoding proteins (Figure 3c and Figure S1) were analyzed sorted based on the phylogenetic tree of IbSAPs, ItbSAPs, and ItfSAPs ( Figure 3a). The gene structures of IbSAPs were similar to those of ItbSAPs and ItfSAPs, and were conserved in each clade. In clade I, most SAPs have one or two introns in their 5 -untranslated regions (5 -UTRs), except that IbSAP1, ItfSAP3L, IbSAP7, and ItbSAP7 have no intron, and that ItfSAP7L has two introns in the coding sequence. In clades II and III, most SAPs have no intron, except that IbSAP6 has one intron in the 3 -UTR, Ib/Itb/ItfSAP19 and IbSAP12 have one intron in the 5 -UTR, and that ItfSAP9L1/2 have one intron in the coding sequence. In clade IV, Ib/Itb/ItfSAP15 have no intron, while Ib/Itb/ItfSAP16/17 have one intron in coding sequences.  Protein structure analysis showed that SAPs from I. batatas, I. triloba, and I. trifida in each clade were conserved. Most Ib/Itb/ItfSAPs in clades I, II, and III had one A20 zinc finger and one AN1 zinc finger (Figures 3c and S1), which is considered a "typical SAP" [1]. Ib/Itb/ItfSAP8, and Ib/Itb/ItfSAP12, only had an AN1 domain (Figure 3c). We found an atypical A20 domain in Ib/Itb/ItfSAP6 (CX3CX8CX2C) (Figure S1), while a typical A20 domain should be CX2-4CX11CX2C [4]. This kind of variation only occurred in Ib/Itb/ItfSAP6 out of all SAPs mentioned in this study. Moreover, we found just one con- Protein structure analysis showed that SAPs from I. batatas, I. triloba, and I. trifida in each clade were conserved. Most Ib/Itb/ItfSAPs in clades I, II, and III had one A20 zinc finger and one AN1 zinc finger (Figures 3c and S1), which is considered a "typical SAP" [1]. Ib/Itb/ItfSAP8, and Ib/Itb/ItfSAP12, only had an AN1 domain ( Figure 3c). We found an atypical A20 domain in Ib/Itb/ItfSAP6 (CX 3 CX 8 CX 2 C) ( Figure S1), while a typical A20 domain should be CX 2-4 CX 11 CX 2 C [4]. This kind of variation only occurred in Ib/Itb/ItfSAP6 out of all SAPs mentioned in this study. Moreover, we found just one conserved cysteine residue instead of an A20 domain in Ib/Itb/ItfSAP8 and Ib/Itb/ItfSAP12 ( Figure S1). In addition, some SAP proteins in I. triloba and I. trifida only had an A20 domain and no AN1 domain, such as ItbSAP7L, ItfSAP3L, and ItfSAP9L1/2 ( Figure 3c). These SAPLs might have evolved due to mutations after gene duplication. In addition, clade IV Ib/Itb/ItfSAPs had only two AN1 domains as well as 0/2 C 2 H 2 domains (Figure 3c). The similar gene structures and conserved protein domains in each clade indicate that Ib/Itb/ItfSAPs in each clade might have expanded due to gene duplication.
A20 domains in SAPs of the three Ipomoea species analyzed were conserved (except for the incomplete A20 domain in Ib/Itb/ItfSAP6), while AN1 domains differed among them. The AN1 domain sequence in typical Ib/Itb/ItfSAPs with both A20 and AN1 domains was CX 2 CX 10 CX 1 CX 4 CX 2 HX 5 HXC, which is considered a typical AN1 domain pattern (CX 2/4 CX 9-12 CX 1-2 CX 4 CX 2 HX 5 HXC). The AN1 domain sequence in Ib/Itb/ItfSAP8 and Ib/Itb/ItfSAP12 (containing only one AN1 domain) was CX 4 CX 10 CX 1 CX 4 CX 2 HX 5 HXC, and in Ib/Itb/ItfSAP15-17 (containing only two AN1 domains and 0/2 C 2 H 2 domains) was CX 4 CX 12 CX 2 CX 4 CX 2 HX 5 HXC ( Figure S1), which is considered an expanded pattern of AN1 [6]. A similar variant AN1 domain was also found in the SlSAP10 protein of tomato, SX 4 CX 10 CX 1 CX 4 CX 2 HX 5 HXC (S stands for serine residue), in which the first cysteine residue is substituted by a serine residue ( Figure S1). Considering sweetpotato and tomato belong to the Solanales, this result suggests that a specific evolution event occurred in SAP families of Solanales.

Cis-Acting Elements in the Promoter of IbSAPs
To predict regulatory factor binding sites in promoters of IbSAP genes, 2-kb upstream sequences of IbSAPs were analyzed. The cis-acting elements identified in IbSAP promoters can be divided into three groups, i.e., hormone response elements, stress response elements, and light signal response elements (Table 2, Figure S2). Hormone response elements in the promoters included ABRE, as-1, ERE, GARE-motif, P-box, TCA-element, and TGAbox, which mainly respond to ABA, auxin, salicylic acid, methyl jasmonate, ethene, or gibberellin ( Table 2). Stress response elements included DRE, LTR, MBS, MYB, STRE, TC-rich repeats, W box, WRE3, and WUN-motif, which are mainly involved in defense and stress responses. Light signal response elements included Box 4, TCT-motif, Sp1, chs-CMA1a/2a, GATA-motif, G-Box, and GT1-motif, which are involved in light responses. Distribution of cis-acting elements in IbSAP promoters are shown in Figure S2. This result indicated that IbSAP genes are primarily regulated by hormones, stresses, and light signals.  To investigate expression patterns of IbSAPs in different organs, IbSAP mRNA levels in leaf, stem, tuberous root, pencil root, and fibrous root of 90-day-old sweetpotato plants were analyzed (Figure 4). Based on their mRNA profiles, the 20 IbSAPs were divided into three groups. The first group-including IbSAP1, 3, 4, and 13-had the highest or second-highest expression levels in tuberous roots, indicating that these IbSAPs might function during the development of tuberous root. The second group-including IbSAP2, 5-8, 16, 17, and 20-had the highest or second-highest expression levels in leaves, indicating that these IbSAPs might function in photosynthesis, immunology, or leaf development. The third group, including the other IbSAPs, had the highest expression levels in pencil roots or fibrous roots, indicating that these IbSAPs might function in absorption and utilization of nutrients and water. Besides, most IbSAPs had the lowest or second-lowest mRNA levels in stems.  Organ-specific expression patterns of SAPs in the two diploid relatives of potato were also analyzed using the RNA-seq results of root 1, root 2, stem, leaf, and flower bud of I. triloba and I. trifida [30]. The Itb/ItfSAPs were divided into two according to their organ-specific expression patterns ( Figure 5). The first group in Itb/ItfSAPs highly expressed in roots. In I. triloba, the first group included ItbSAP and 14-16; while in I. trifida, the first group included ItfSAP1, 4, 6, 7, 9-12, 14-16, The second group included Itb/ItfSAPs highly expressed in aerial parts (i.e., ste flower, or flower bud) or were not expressed in the organs examined. ItbSAP4, Itb and ItfSAP21 had the highest expression level in stems while ItbSAP21 and ItfSAP only detected in stems. Two ItbSAPs (ItbSAP17 and 20) and four ItfSAPs (ItfSAP8 and 20) had highest expression levels in flower buds, while five ItbSAP4s (i.e., ItbS 5,13,and 19) and four ItfSAPs (ItfSAP2, 3, 5, and 13) had the highest expression l flowers. Here, ItbSAP8 (expressed more in root than in leaf) and ItfSAP8 (expresse in leaf than in root) had different expression patterns, and the expression pa ItfSAP8 was more similar to that of IbSAP8, showing highest expression levels in (Figure 4). In addition, three ItbSAPs (ItbSAP7L, 18, and 22) and five ItfSAPs (ItfS 9L1, and 9L2) were not detected in any organs examined. Organ-specific expression patterns of SAPs in the two diploid relatives of sweetpotato were also analyzed using the RNA-seq results of root 1, root 2, stem, leaf, flower, and flower bud of I. triloba and I. trifida [30]. The Itb/ItfSAPs were divided into two groups according to their organ-specific expression patterns ( Figure 5). The first group included Itb/ItfSAPs highly expressed in roots. In I. triloba, the first group included ItbSAP1, 6-12, and 14-16; while in I. trifida, the first group included ItfSAP1, 4, 6, 7, 9-12, 14-16, and 18. The second group included Itb/ItfSAPs highly expressed in aerial parts (i.e., stem, leaf, flower, or flower bud) or were not expressed in the organs examined. ItbSAP4, ItbSAP21, and ItfSAP21 had the highest expression level in stems while ItbSAP21 and ItfSAP21 were only detected in stems. Two ItbSAPs (ItbSAP17 and 20) and four ItfSAPs (ItfSAP8, 17, 19, and 20) had highest expression levels in flower buds, while five ItbSAP4s (i.e., ItbSAP2, 3, 5, 13, and 19) and four ItfSAPs (ItfSAP2, 3, 5, and 13) had the highest expression levels in flowers. Here, ItbSAP8 (expressed more in root than in leaf) and ItfSAP8 (expressed more in leaf than in root) had different expression patterns, and the expression pattern of ItfSAP8 was more similar to that of IbSAP8, showing highest expression levels in leaves ( Figure 4). In addition, three ItbSAPs (ItbSAP7L, 18, and 22) and five ItfSAPs (ItfSAP3, 7, 9L1, and 9L2) were not detected in any organs examined.

Expression Patterns of IbSAPs, ItbSAPs, and ItfSAPs in Response to Salinity, Drought, or ABA Treatments
Since the promoter regions of IbSAPs had stress response and hormone response cisacting elements, we determined whether IbSAPs responded to salinity, drought, or ABA treatments ( Figure 6). Although IbSAP4 was insensitive to salinity and IbSAP17 insensitive to salinity and ABA treatments, most IbSAPs were induced within 6 h after salinity, drought, or ABA treatments, while IbSAP5 was rapidly induced after 1 h of salinity treatment. The expression of IbSAPs increased up to 20-fold (IbSAP16 at 12 h salinity treatment) compared to control conditions. The mRNA levels of two IbSAPs (IbSAP18 and IbSAP20) under salinity treatment, five IbSAPs (IbSAP5, 16, and 18-20) under drought treatment, and three IbSAPs (IbSAP3, 5, and 20) under ABA treatment kept increasing during the 24h treatment. These results indicated that almost all of the IbSAPs can be induced by salinity, drought, or ABA treatment.  Since the promoter regions of IbSAPs had stress response and hormone response cis-acting elements, we determined whether IbSAPs responded to salinity, drought, or ABA treatments ( Figure 6). Although IbSAP4 was insensitive to salinity and IbSAP17 insensitive to salinity and ABA treatments, most IbSAPs were induced within 6 h after salinity, drought, or ABA treatments, while IbSAP5 was rapidly induced after 1 h of salinity treatment. The expression of IbSAPs increased up to 20-fold (IbSAP16 at 12 h salinity treatment) compared to control conditions. The mRNA levels of two IbSAPs (IbSAP18 and IbSAP20) under salinity treatment, five IbSAPs (IbSAP5, 16, and 18-20) under drought treatment, and three IbSAPs (IbSAP3, 5, and 20) under ABA treatment kept increasing during the 24-h treatment. These results indicated that almost all of the IbSAPs can be induced by salinity, drought, or ABA treatment.
In addition, the expression patterns of I. triloba and I. trifida SAPs in response to ABA, salinity, and drought treatments were also analyzed using the RNA-seq data [30]. Under ABA treatment, 14 ItbSAPs and 16 ItfSAPs were induced, while 5 ItbSAPs and 6 ItfSAPs were repressed. Under salinity treatment, 10 ItbSAPs and 17 ItfSAPs were induced, while 9 ItbSAPs and 3 ItfSAPs were repressed. Under drought treatment, 10 ItbSAPs and 16 ItfSAPs were induced, while 9 ItbSAPs and 4 ItfSAPs were repressed. Four ItbSAPs and six ItfSAPs were not detected under those treatments. These results indicated that most ItbSAPs and ItfSAPs respond to salinity, drought, or ABA treatments, and they might play important roles in abiotic stress responses in I. triloba and I. trifida. Expression patterns of SAP orthologs in I. triloba and I. trifida were similar (Figure 7). However, most IbSAPs were induced by stress treatments and no IbSAPs were repressed ( Figure 6). compared to control conditions. The mRNA levels of two IbSAPs (IbSAP18 and IbSAP20) under salinity treatment, five IbSAPs (IbSAP5, 16, and 18-20) under drought treatment, and three IbSAPs (IbSAP3, 5, and 20) under ABA treatment kept increasing during the 24h treatment. These results indicated that almost all of the IbSAPs can be induced by salinity, drought, or ABA treatment.  In addition, the expression patterns of I. triloba and I. trifida SAPs in response to ABA, salinity, and drought treatments were also analyzed using the RNA-seq data [30]. Under ABA treatment, 14 ItbSAPs and 16 ItfSAPs were induced, while 5 ItbSAPs and 6 ItfSAPs were repressed. Under salinity treatment, 10 ItbSAPs and 17 ItfSAPs were induced, while 9 ItbSAPs and 3 ItfSAPs were repressed. Under drought treatment, 10 ItbSAPs and 16 ItfSAPs were induced, while 9 ItbSAPs and 4 ItfSAPs were repressed. Four ItbSAPs and six ItfSAPs were not detected under those treatments. These results indicated that most ItbSAPs and ItfSAPs respond to salinity, drought, or ABA treatments, and they might play important roles in abiotic stress responses in I. triloba and I. trifida. Expression patterns of SAP orthologs in I. triloba and I. trifida were similar (Figure 7). However, most IbSAPs were induced by stress treatments and no IbSAPs were repressed ( Figure 6).

Ectopic Expression of IbSAP16 Enhanced Salinity Tolerance of Transgenic Arabidopsis Plants
Many SAP genes regulate stress tolerance in plants [3,[7][8][9][10][11][12]. However, the mechanism of SAP proteins containing only AN1 domains or AN1 and C2H2 domains in the stress response is not well-studied. IbSAP16 is one of the two AN1-AN1-C2H2-C2H2 type SAPs in sweetpotato, and the other one is IbSAP17 (Figure 3). IbSAP16 was upregulated more than 20-fold after salinity stress, however the expression level of IbSAP17 changed less than 2-fold ( Figure 6). Therefore, to test the hypothesis that IbSAP16 confers salinity tolerance, we ectopically expressed IbSAP16 driven by the 35 S promoter in Arabidopsis thaliana Col-0 (WT). Two independent IbSAP16-overexpressing transgenic Arabidopsis

Ectopic Expression of IbSAP16 Enhanced Salinity Tolerance of Transgenic Arabidopsis Plants
Many SAP genes regulate stress tolerance in plants [3,[7][8][9][10][11][12]. However, the mechanism of SAP proteins containing only AN1 domains or AN1 and C 2 H 2 domains in the stress response is not well-studied. IbSAP16 is one of the two AN1-AN1-C 2 H 2 -C 2 H 2 type SAPs in sweetpotato, and the other one is IbSAP17 (Figure 3). IbSAP16 was upregulated more than 20-fold after salinity stress, however the expression level of IbSAP17 changed less than 2-fold ( Figure 6). Therefore, to test the hypothesis that IbSAP16 confers salinity tolerance, we ectopically expressed IbSAP16 driven by the 35 S promoter in Arabidopsis thaliana Col-0 (WT). Two independent IbSAP16-overexpressing transgenic Arabidopsis lines (L5 and L22) with different IbSAP16 mRNA levels ( Figure S3) were selected for further study.
To investigate the function of IbSAP16 in the salinity stress response, germination rates of L5, L22, and WT under salinity stress were investigated (Figure 8). Without NaCl treatment, or under 50 and 100 mmol·L −1 NaCl treatment, germination rates of IbSAP16transgenic lines were similar to WT. Only on day 3 on media with 100 mmol·L −1 NaCl, the germination rate of L22 was significantly higher than that of WT. Finally, approximately 100% of L5, L22, and WT seeds germinated on the 1/2 Murashige and Skoog (MS) media with 0, 50, or 100 mmol·L −1 NaCl. When NaCl concentration was increased to 150 or 200 mmol·L −1 , germination rates of L5 and L22 seeds were significantly higher than those of WT seeds (Figure 8a-e). These results suggested that ectopic expression of IbSAP16 enhanced salinity tolerance of transgenic Arabidopsis at the germinating stage. germination rate of L22 was significantly higher than that of WT. Finally, approximately 100% of L5, L22, and WT seeds germinated on the 1/2 Murashige and Skoog (MS) media with 0, 50, or 100 mmol·L −1 NaCl. When NaCl concentration was increased to 150 or 200 mmol·L −1 , germination rates of L5 and L22 seeds were significantly higher than those of WT seeds (Figure 8a-e). These results suggested that ectopic expression of IbSAP16 enhanced salinity tolerance of transgenic Arabidopsis at the germinating stage.  replicates. "*" or "**" indicate a significant difference from that of WT at p < 0.05 or 0.01, by Student's t-test.
Furthermore, survival rates of L5, L22, and WT seedlings under salinity stress were determined. While all seedlings survived under control conditions, only 19.3% of WT seedlings survived under 250 mmol·L −1 NaCl treatment, but 81.0% and 74.1% of L5 and L22 seedlings, respectively, survived the 250 mmol·L −1 NaCl treatment (Figure 9). These results suggested that ectopic expression of IbSAP16 enhanced salinity stress tolerance of transgenic Arabidopsis at the seedlings stage. Furthermore, survival rates of L5, L22, and WT seedlings under salinity stress wer determined. While all seedlings survived under control conditions, only 19.3% of W seedlings survived under 250 mmol·L −1 NaCl treatment, but 81.0% and 74.1% of L5 an L22 seedlings, respectively, survived the 250 mmol·L −1 NaCl treatment (Figure 9). Thes results suggested that ectopic expression of IbSAP16 enhanced salinity stress tolerance o transgenic Arabidopsis at the seedlings stage.

Discussion
A20/AN1 zinc finger domain containing proteins play a pivotal role in regulatin immune responses in animals and abiotic stress responses in plants. They have bee widely studied for their participation in physiological and biochemical reactions such a immune responses and apoptosis. Human ZNF216, containing both A20 and AN1 do mains, regulates immune responses by inhibiting the activation of nuclear factor-kappa (NF-κB). Consequently, ZNF216 reduces allergic reactions and cell apoptosis [31,32]. Pro tein A20 inhibits the activity of NF-κB as well, which reduces cell apoptosis caused by th tumor necrosis factor (TNF) [4,33,34]. Investigations of SAP proteins in plants were con ducted much later than those in animals. The first A20/AN1 domain containing protei encoding gene identified in plants was OsiSAP1, which is induced by multiple abioti stresses. Ectopic expression of OsiSAP1 in tobacco conferred tolerance to abiotic stresse [3]. Up to now, SAPs in rice, Arabidopsis, maize, polar, tomato, cotton, Medicago truncatula apple, Brassica napus, soybean, and cucumber were identified [2,6,[17][18][19][20][21][22][23]. However, th function of the SAP family in sweetpotato is still poorly understood.

Evolution of the SAP Gene Family in Sweetpotato and Its Two Diploid Relatives
In this study, a total of 69 SAPs were identified in hexaploid sweetpotato and its tw diploid relatives, including 20 IbSAPs, 23 ItbSAPs, and 26 ItfSAPs (Tables 1 and S1, an Figure 1), which were divided into four clades (clade I to IV) according to their structure ( Figure 2). The number of IbSAP genes (20) is more than that found in diploid plants, suc as A. thaliana (14), rice (18), maize (11), and tomato (13) [2,6,17]. However, in some poly ploidy crops, the number of A20/AN1 zinc finger encoding genes is more than that i Data are presented as mean ± SE (n = 3). "**" indicates a significant difference from that of WT at p < 0.01, by Student's t-test.

Discussion
A20/AN1 zinc finger domain containing proteins play a pivotal role in regulating immune responses in animals and abiotic stress responses in plants. They have been widely studied for their participation in physiological and biochemical reactions such as immune responses and apoptosis. Human ZNF216, containing both A20 and AN1 domains, regulates immune responses by inhibiting the activation of nuclear factor-kappa β (NF-κB). Consequently, ZNF216 reduces allergic reactions and cell apoptosis [31,32]. Protein A20 inhibits the activity of NF-κB as well, which reduces cell apoptosis caused by the tumor necrosis factor (TNF) [4,33,34]. Investigations of SAP proteins in plants were conducted much later than those in animals. The first A20/AN1 domain containing protein encoding gene identified in plants was OsiSAP1, which is induced by multiple abiotic stresses. Ectopic expression of OsiSAP1 in tobacco conferred tolerance to abiotic stresses [3]. Up to now, SAPs in rice, Arabidopsis, maize, polar, tomato, cotton, Medicago truncatula, apple, Brassica napus, soybean, and cucumber were identified [2,6,[17][18][19][20][21][22][23]. However, the function of the SAP family in sweetpotato is still poorly understood.

Evolution of the SAP Gene Family in Sweetpotato and Its Two Diploid Relatives
In this study, a total of 69 SAPs were identified in hexaploid sweetpotato and its two diploid relatives, including 20 IbSAPs, 23 ItbSAPs, and 26 ItfSAPs (Tables 1 and S1, and Figure 1), which were divided into four clades (clade I to IV) according to their structures ( Figure 2). The number of IbSAP genes (20) is more than that found in diploid plants, such as A. thaliana (14), rice (18), maize (11), and tomato (13) [2,6,17]. However, in some polyploidy crops, the number of A20/AN1 zinc finger encoding genes is more than that in sweetpotato. For example, 37 and 57 SAP genes were identified in tetraploid crops, cotton [18] and B. napus [21], respectively. In addition, in two diploid wild relatives of sweetpotato, more SAP genes were identified. No orthologs of two ItbSAPs (ItbSAP21 and ItbSAP22) and two ItfSAPs (ItfSAP21 and ItfSAP22) were identified in the sweetpotato. Sweetpotato is a hexaploid crop (2x = 6n = 90); however, the genome sequence was assembled into only 15 pairs rather than 45 pairs of chromosomes [29]. Thus, there should additional IbSAP gene members or gene copies in the sweetpotato genome. Moreover, only 18 of 20 IbSAPs were identified from the sweetpotato genome database [29], while IbSAP13 and IbSAP17 were identified from our unpublished transcriptome database. One of the reasons of this result might be that genome sequences vary between different sweetpotato varieties, and some new SAPs emerged after sweetpotato evolved from the ancestor species. The other reason might be that some IbSAPs are located in chromosomes with relatively poor sequencing quality. Nevertheless, our results supplement the sweetpotato genome database, but further improvement of the sweetpotato genome sequence is urgently needed to promote research on sweetpotato gene functions.
The gene structures of most orthologous SAPs were similar between sweetpotato and its wild relatives-I. triloba and I. trifida-except that IbSAP1, 6, 7, and 12 were different from their orthologs in I. triloba and I. trifida. The differences were mainly intron insertion or deletion. Meanwhile, the gene structures of these orthologous SAPs in ItbSAPs and ItfSAPs were almost the same (Figure 3b). This result suggested that the SAP gene family is relatively conserved, but still some changes occurred during or after formation of the hexaploid species from its diploid relatives.
Many SAPs in plants contain a typical SAP structure (an AN1 domain at the C-terminus and an A20 domain at the N-terminus) [1]. Ten out of 14 SAP proteins in Arabidopsis, 11 out of 18 in rice, 9 out of 13 in tomato, and 8 out of 11 in maize are typical SAPs [2,6,17]. Similarly, we found that 14 out of 20 IbSAPs, 14 out of 23 ItbSAPs, and 15 out of 26 ItfSAPs contain the typical SAP structure (Figures 3c and S1). The other Ib/Itb/ItfSAP proteins contain 1 or 2 AN1 domains with 0 or 2 C 2 H 2 zinc finger domains, but no A20 domains (Ib/Itb/ItfSAP6 were exceptions, which contain one incomplete A20 domain). We found that the AN1 domain in the five IbSAPs lacking A20 domains was CX 4 CX 9-12 CX 1-2 CX 4 CX 2 HX 5 HXC, which is different from the typical AN1 (CX 2 CX 9-12 CX 1-2 CX 4 CX 2 HX 5 HXC) in those SAPs harboring A20 domains. There is a two-amino-acid addition at the N-terminus in the former AN1, which is classified as an expanded AN1 [6]. This observation was in agreement with previously published results [6]. However, we identified a few differences. All "Type I" SAP proteins (clades I, II, and III in this study) other than SlSAP10 had the typical AN1 domain, while "Type II" SAPs (clade IV in this study) had the expanded AN1 domain. The A20-lacking SAPs, SAP15, SAP16, and SAP17 of sweetpotato, I. triloba, and I. trifida belong to clade IV, while SAP8 and SAP12 belong to clade III. However, SAP8 and SAP12 of these three Ipomoea plants and their ortholog in tomato (SlSAP10) had the expanded pattern of the AN1 domain. In addition, only Ib/Itb/ItfSAP6 in clade II had an incomplete A20 domain (Figures 3 and S1). These results suggest that special evolutionary events have occurred in SAP genes in Ipomoea or Solanales species.
To date, the evolutionary history of the sweetpotato is still debated. It is, however, commonly accepted that I. triloba and I. trifida are closely related to sweetpotato [35,36]. In this study, we identified IbSAPs in sweetpotato and their orthologs in I. triloba and I. trifida. These three SAP gene families share high similarities, which provided further evidence for the theory that I. triloba and I. trifida might be the ancestral species of sweetpotato.

Different Functions of SAPs on Growth and Development between Sweetpotato and Its Two Diploid Relatives
SAPs play pivotal roles in growth and development of plants. OsSAP4 in rice is constitutively expressed and negatively regulates the synthesis of gibberellin, thus controlling plant height, leaf blade size, and grain size [37]. AtSAP5 is expressed throughout A. thaliana plants and positively regulates plant growth by mediating the degradation of AtMBP-1 [38]. Similarly, most of the IbSAPs, ItbSAPs, and ItfSAPs were expressed in all organs tested. Meanwhile, mRNA levels of SAPs varied greatly among different organs of sweetpotato, as well as in its two diploid relatives (Figures 4 and 5). The expression level of IbSAP10 in pencil roots was 100-fold higher than the IbSAP18 level in tuberous roots. Similar results were previously reported for expression patterns of tomato SAPs, for instance, the mRNA level of SlSAP1 was approximately 120-fold higher than SlSAP13 mRNA levels in unstressed 9-day-old tomato seedlings [17].
Most IbSAPs shared similar expression patterns in different organs with their orthologs in I. triloba and I. trifida. However, there were some exceptions, and these distinguished expression patterns of SAP orthologs in I. batatas, I. triloba, and I. trifida might indicate different functions. IbSAP16 in sweetpotato might mainly regulate leaf development, while its orthologs in I. triloba and I. trifida might mainly regulate root development. This is because IbSAP16 has higher expression levels in leaves than in other organs (Figure 4), while ItbSAP16 and ItfSAP16 has higher expression levels in roots ( Figure 5). Moreover, IbSAP3 might play an important role in the development of tuberous roots in sweetpotato, because it is highly expressed in tuberous roots (Figure 4). I. triloba and I. trifida do not have tuberous roots, while ItbSAP3 and ItfSAP3 has higher expression levels in stems and leaves than in root ( Figure 5). These results suggest that most orthologs of SAPs in I. triloba and I. trifida might play similar roles in growth and development, but that they have derived new functions when the two ancestors evolved into sweetpotato.

Different Functions of SAPs in Multiple Abiotic Stress Responses between Sweetpotato and Its Two Diploid Relatives
Many SAP genes are induced by various abiotic stresses, and overexpression or ectopic expression of certain SAP genes confers stress tolerance in transgenic plants. For example, the first SAP gene identified in plants, OsiSAP1, is induced by abiotic stresses such as salinity, desiccation, submergence, heavy metals, mechanical wounding, and ABA treatment [3]. Ectopic expression or overexpression of OsiSAP1 confers tolerance to abiotic stress in transgenic tobacco [3], Arabidopsis [7], and rice [8]. An Arabidopsis SAP gene, AtSAP12, is responsive to cold, drought, genotoxic, osmotic, oxidative, salinity, and wounding stresses [39]. Similar to SAPs in other plants, IbSAPs are induced by salinity, dehydration, or ABA treatment ( Figure 6), which most likely is due to the presence of stress response and ABA response cis-acting elements in promoter regions of IbSAPs (Table 2, Figure S2). Many SAPs from other plants have been reported to be induced rapidly by multiple abiotic stresses. OsSAP1, AtSAP10, and MusaSAP1 (SAP1 in banana) are induced within 15 or 30 min after exposure to various stresses [3,40,41]. We found that some IbSAPs could also rapidly respond to abiotic stresses. IbSAP1, IbSAP3, IbSAP4, and IbSAP5 are upregulated just 1 h after salinity, drought, or ABA treatments, and IbSAP5 even had the highest mRNA levels just 1 h after salinity treatment ( Figure 6). Most IbSAPs reached their peak value 6 h after stress treatments. By contrast, other IbSAPs reached their highest mRNA levels 12 h after treatments, such as IbSAP14 and IbSAP16 in response to salinity and ABA treatments, and IbSAP8 in response to salinity and drought treatments. These results indicated that IbSAPs might function at different stages of stress responses in sweetpotato.
Some A20-lacking SAPs have functional redundancy in response to abiotic stresses. Compared to wild type, homozygous loss-of-function Arabidopsis mutants of SAP genes encoding two AN1 domains have no visible phenotypic differences under various stress conditions [6]. In our research, IbSAP1 encoding a protein containing two AN1 and two C 2 H 2 domains, seemed to be insensitive to abiotic stresses. The expression level of IbSAP17 changed less than 2-fold after salinity, drought, and ABA treatments ( Figure 6). This result suggested functional redundancy of A20-lacking SAPs in response to abiotic stresses as well. However, further investigations are needed to confirm this assumption.
In addition, hormone response cis-acting elements for auxin, ethylene, gibberellin, salicylic acid, and methyl jasmonate, and light signal response cis-acting elements were found in promoter regions of IbSAPs (Table 2 and Figure S2). Some SAPs were previously reported to be regulators of biotic stress responses or plant growth and development, such as OsDOG (OsSAP11), which is induced by gibberellins and negatively regulates rice cell elongation [42]. AtSAP9-overexpressing plants showed delayed flowering time and were more susceptible to P. syringae pv. phaseolicola infections [13]. Moreover, AtSAP5, Pha13, and SlSAP3 are involved in regulating resistance to viral and bacterial infections [15,16]. IbSAPs might be involved in biotic stress responses and regulation of plant growth and development as well. This is worth investigating, because sweetpotato virus disease (SPVD) has been seriously affecting the production of sweetpotato [43].
In addition, expression patterns of orthologous SAPs in diploid I. triloba and I. trifida were different from those in sweetpotato (Figure 7). Some Itb/ItfSAPs were induced by salinity and drought stresses and ABA treatment as well; however, some Itb/ItfSAPs were repressed. These stress-induced or different expression patterns of Ib/Itb/ItfSAPs homologs in response to abiotic stress could both provide potential candidate genes for further functional research and for modifying abiotic stress tolerance of plants.

AN1-AN1-C 2 H 2 -C 2 H 2 Type SAPs Can Be Candidates to Improve Salinity Stress Tolerance of Plants
Most functional analyses of SAPs in plants have focused on typical SAP proteins, which have one AN1 domain and one A20 domain, while only a few studies have investigated SAPs containing only AN1 domains [22,39,44,45] and AN1-AN1-C 2 H 2 -C 2 H 2 type SAPs [11]. IbSAP16 is one of the two AN1-AN1-C 2 H 2 -C 2 H 2 encoding genes in sweetpotato. Its expression was induced by salinity treatment. The ortholog of IbSAP16 in A. thaliana-AtSAP13-positively regulates tolerance to multiple abiotic stresses [11]. In our study, we found that IbSAP16-transgenic Arabidopsis lines showed stronger tolerance to salinity stress in both germination and vegetative growth stages (Figures 8 and 9). These results indicated that IbSAP16 from sweetpotato also functions as a positive regulator in response to abiotic stresses, and AN1-AN1-C 2 H 2 -C 2 H 2 type SAPs should be considered as candidates for improving salinity stress tolerance in plants. These results also laid the foundation for IbSAP16-mediated abiotic stress response mechanism studies.

Chromosomal Distribution of IbSAPs, ItbSAPs, and ItfSAPs
The location information of IbSAPs were obtained from the sweetpotato gff3 annotation file downloaded from the Ipomoea Genome Hub (https://sweetpotao.com/, accessed on 2 December 2019). The location information of ItbSAPs and ItfSAPs were obtained from gff3 annotation downloaded from Sweetpotato Genomics Resource (http: //sweetpotato.plantbiology.msu.edu/, accessed on 2 December 2019). The SAP genes in I. batatas, I. triloba, and I. trifida with location information were mapped to chromosomes by TBtools software [50].

Analyses of Gene Structure and Conserved Domains of SAP Members
Gene structure information was obtained from gff3 annotation of I. batatas, I. triloba, and I. trifida. Conserved domains were analyzed by the NCBI branch web CD-search tool as mentioned above. Gene structures and conserved domains were generated using TBtools software [50].

Cis-Acting Regulatory Elements Analyses of IbSAPs
For promoter analyses, genomic DNA sequences 2 kb upstream of the translation initiation site of IbSAPs were extracted from the sweetpotato genome database. Cis-acting regulatory elements in the upstream sequences were then analyzed using the online database PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/, accessed on 26 May 2020) [54].

Plant Materials
Sweetpotato (Ipomoea batatas) cultivar Xushu 28, a variety developed in our institute with high yield and salinity tolerance, was used in this study, and the sampling permission was obtained. Leaf, stem, tuberous root, pencil root, and fibrous root were sampled 90 d after planting for quantitative real-time PCR analyses of IbSAPs in different organs. Shoots of 20 cm length were cultured in a one-quarter Hoagland solution and kept in an MLR-352 Panasonic illumination incubator at 25 ± 1 • C with 16 h light/8 h dark cycles. The light intensity of the illumination incubator was set at LS5 (21637 lx). Arabidopsis plants were kept in a growth chamber (22 • C, 16 h light/8 h dark cycle) for functional analysis of IbSAP16.

Stress Treatments of Sweetpotato Shoots
Sweetpotato shoots were treated with 100 mmol·L −1 NaCl (salinity stress), 15% polyethylene glycol (PEG) 6000 (simulated drought stress), or 10 µmol·L −1 ABA in onequarter Hoagland solution after their roots grew to 3-5 cm. Shoots treated with one-quarter Hoagland solution was used as control (CK). The roots were sampled at 0, 1, 6, 12, and 24 h after treatments for quantitative real-time PCR analysis of IbSAPs. All samples were frozen in liquid nitrogen right away and stored at −70 • C before RNA isolation.

RNA Isolation and Quantitative Real-Time PCR Analysis
RNA was isolated using the Generay total RNA isolation kit (Shanghai Generay Biotech Co., Ltd., Shanghai, China, code no. GK3016) and treated with DNase I following the instructions of the manufacturer. RNA samples were checked by denaturing agarose gel electrophoresis and ND-1000 to determine RNA quality and concentration. First strand cDNA was synthesized from 2 µg total RNA using the ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo, Japan, code no. FSQ-301). The primers used for RT-qPCR were designed using Primer 3 Plus (http://www.bioinformatics.nl/cgi-bin/primer3 plus/primer3plus.cgi, accessed on 3 December 2019) [55] and were listed in Table S3. Real-time qPCR was performed using the Toyobo SYBR Green Realtime PCR Master Mix (TOYOBO Co., Ltd. Japan, code no. QPK-201). The reaction cycles were set as follows: 95 • C (2 min)/(95 • C (15 s)/60 • C (15 s)/72 • C (20 s) × 40 cycles); the melt curve of each reaction was detected per 0.3 • C from 60 • C to 95 • C. These reactions were put in 96-well optical reaction plates (Applied Biosystems, Foster City, CA, USA) and carried out in ABI StepOne Plus real-time PCR system. Each reaction was repeated three times. IbARF (Ipomoea batatas ADP-ribosylation factor, GenBank accession no. JX177359) and AtActin2 were used as the internal control to normalize the variation among sweetpotato [56] and Arabidopsis templates, respectively. The relative mRNA levels were calculated using the 2 −∆∆Ct method [57].

Transcriptome Analysis of ItbSAPs and ItfSAPs Expression Patterns
The RNA-seq data of I. triloba and I. trifida were obtained from the Sweetpotato Genomics Resource (http://sweetpotato.plantbiology.msu.edu/, accessed on 2 December 2019) [30]. The value of fragments per kilobase of exon per million fragments mapped (FPKM) were used to represent the relative expression level of ItbSAPs and ItfSAPs. The heat maps were visualized by TBtools software [50].

Plant Expression Vector Construction of IbSAP16 and Arabidopsis Transformation
Full-length CDS of IbSAP16 was cloned from cDNA of Xushu 28 using primers shown in Supplementary Table S1 and ligated into vector pCR8/GW/TOPO (Invitrogen, Carlsbad, CA, USA, code no. K252020) to obtain pCR8/GW/TOPO-IbSAP16. pGWB12-IbSAP16 (35S::IbSAP16) was constructed by LR reaction between pGWB12 and pCR8/GW/TOPO-IbSAP16. Agrobacterium tumefaciens strain GV3101 was used to transfer 35S::IbSAP16 into Arabidopsis by using the floral dipping method [58]. T1 generation Arabidopsis lines were selected by germinating the seeds on half MS media with 50 mg·L −1 hygromycin and confirmed by PCR and RT-qPCR. T3 homozygous lines were screened and used for further experiments.

Assay for Salinity Tolerance of Transgenic Arabidopsis
For germination-rate assays, sterilized Arabidopsis seeds from IbSAP16-transgenic lines and WT plants were planted on 1/2 MS media with 0, 50, 100, 150, and 200 mmol·L −1 NaCl and kept at 4 • C for 2 days before they were cultured in a growth chamber (22 • C, 16 h light/8 h dark cycle). Germination rates were recorded daily within a week. Seven-day-old seedlings with similar growth rates were transplanted to pots. Three weeks later, the plants were watered with 250 mM NaCl solution or pure water (300 mL for each pot) once a week for 3 weeks to determine their survival rates under salinity stress.

SPSS 22
Statistical software was used in statistical analysis. Means and standard errors were calculated and analyzed with ANOVA, and the probabilities for significance were estimated with Student's t-test.

Conclusions
In this study, 20, 23, and 26 SAP genes encoding A20/AN1 zinc finger proteins were identified in I. batatas, I. triloba, and I. trifida. Two of the 20 IbSAPs were only identified in our unpublished transcriptomics database, which supplements the sweetpotato genome database. The gene structures and conserved protein domains of SAP family members in sweetpotato and their orthologs in I. triloba and I. batatas were similar. Sequences alignment of SAP in sweetpotato, I. triloba, I. trifida, and other plants revealed Solanales-specific AN1 zinc finger domain in SAPs containing only one AN1 zinc finger domain, which indicated specific evolution events occurred in SAP gene families of Solanales plants. The different organ-specific and stress response expression patterns of IbSAPs, ItbSAPs, and ItfSAPs suggested that these SAP genes are involved in plant growth and development, as well as abiotic stress responses. In addition, we showed that IbSAP16-an AN1-AN1-C 2 H 2 -C 2 H 2 type IbSAP-is a positive regulator in response to salinity stress in transgenic A. thaliana. These results provided a characterization of SAP genes in I. batatas, I. triloba, and I. trifida and laid the groundwork for studying IbSAP-mediated stress response mechanisms. The function and regulatory network of certain SAPs in sweetpotato, I. triloba, and I. trifida merits further investigation. Experiments and publication costs are funded by these programs. The funder has no role in study design, data collection analysis and interpretation, decision to publish, or preparation of the manuscript.

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

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

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