Identification and expression analysis of the CqSnRK2 gene family and a functional study of the CqSnRK2.12 gene in quinoa (Chenopodium quinoa Willd.)

Sucrose non-fermenting 1 (SNF1)-associated protein kinase 2 (SnRK2) proteins belong to a relatively small family of plant-specific serine/threonine (Ser/Thr) protein kinases. SnRK2s participate in the abscisic acid (ABA) signaling pathway and play important roles in many biotic and abiotic stresses. At present, no SnRK2 gene has been reported in quinoa, and the recently published genome for this species provides an opportunity to identify and characterize the SnRK2 gene family. We identified 13 SnRK2 genes in the C. quinoa genome by bioinformatics analysis. Based on their phylogenetic relationships, these genes were divided into three subfamilies, similar to the situation in other plant species. Gene duplication analysis showed that there were seven pairs of homologous genes in the CqSnRK2 family, and that purifying selection played an important role in the evolution of SnRK2 genes. Gene structure analysis showed that the first exon in the SnRK2 family genes has the same length as the last exon, and that CqSnRK2 genes in the same subfamily have similar gene structures. Sequence analysis showed that the N-terminal region contains three highly conserved motifs. In addition, many kinds of cis-elements were identified in the promoter region of CqSnRK2, including those for hormone responses, stress responses, and tissue-specific expression. Transcription data analysis and qRT-PCR results showed that CqSnRK2 has different expression patterns in roots, stems, and leaves, and responded to biotic and abiotic stresses such as low temperature, salt, drought, and abscisic acid (ABA). In addition, we found that the protein encoded by CqSnRK2.12 was localized to the cytoplasm and nucleus, and there was no self-activation. The results of CqSnRK2.12 overexpression showed that transgenic Arabidopsis thaliana lines had increased drought tolerance compared to the controls. The results of our study provide references for further studies on the evolution, function, and expression of the SnRK2 gene family in quinoa.

themselves against a variety of environmental stresses that allow them respond to adverse conditions by regulating the production of beneficial substances in their cells or by initiating the expression of stress-related genes [2,3]. Among these mechanisms are protein kinases and phosphatases that are the main components of intracellular signal transduction pathways which play key roles in the plant response to stresses [4]. Sucrose non-fermenting-1 (SNF1)-related protein kinases (SnRKs) are a family of serine/threonine (Ser/Thr) protein kinases widely found in plants. SnRKs are highly conserved and participate in various physiological processes [5,6]. Based on the conserved kinase activity domains, SnRKs can be divided into three subfamilies; SnRK1, SnRK2, and SnRK3. SnRK1s are key components of cell signal transduction pathways, and the N-terminal region has a conserved catalytic domain, which regulates the energy balance of the cell by interacting with different forms of sugar in the cell. In addition, SnRK1s also respond to biological and abiotic stresses to regulates the growth and development of plants [6,7]. SnRK1 proteins are highly homologous in structure and function to the AMPK protein in animals and the SNF1 protein in yeast. SnRK2 and SnRK3 are plant-specific protein kinases, which play key roles in plant resistance signaling [5].
The SnRK2 (Sucrose non-fermenting 1-related protein kinases 2) proteins are a relatively small plant-specific protein kinase family with typical N-terminal and C-terminal functional domains [8]. The N-terminal catalytic domain is highly conserved. The kinase domains present in AMPK and SNF1 are homologous and share 42%-46% similarity; the C-terminus is the regulatory region which is mainly composed of two subdomains called Domain I and Domain II [8,9]. Domain I is composed of 30 amino acids and is present in all SnRK2 family members. It is mainly activated by adverse conditions, but does not depend on ABA. Domain II is composed of 40 amino acids, and is located close to the C-terminus, is unique to ABA-dependent SnRK2s, and is an essential structure in the response to ABA [8][9][10][11]. Compared with SnRK1, the C-terminus of SnRK2 lacks 140-160 amino acids and is relatively short. The distinctive feature of the C-terminus is an acidic patch structure rich in glutamic acid or aspartic acid (D/E) residues. Based on this, the SnRK2 family can be divided into the SnRK2a (Subclass I and Subclass II) and SnRK2b (Subclass III) subfamilies. The C-terminus of SnRK2a is rich in aspartic acid, while SnRK2b is rich in glutamic acid [9,11,12]. Studies have found that the C-terminal domain of SnRK2 is related to enzyme activation, ABA signal transmission, and protein interaction [13]. In addition, SnRK2 genes participate in ABA-dependent and ABA-independent signaling pathways to regulate osmotic stress, stomata opening and closing, seed germination, seedling development, and other plant growth processes [5,14,15]. Therefore, based on whether it is induced by ABA, the SnRK2 family can be divided into Groups I, II, III. Group I is not induced by ABA, Group II is slightly induced by ABA, and Group III is strongly induced by ABA. AtSnRK2. 1,AtSnRK2.4,AtSnRK2.5,AtSnRK2.9, and AtSnRK2.10 belong to Group I; AtSnRK2. 7 and AtSnRK2.8 belong to Group II; and AtSnRK2.2, AtSnRK2.3, and AtSnRK2.6 belong to Group III [16].
Quinoa (Chenopodium quinoa Willd.) is an annual dicotyledonous plant in the family Amaranthaceae that originated in the Andes Mountains of South America, quinoa seeds are rich in high-quality protein that contains a balance of essential amino acids that are required in the the human diet [33]. The grain is also gluten-free and is rich in vitamins, polyphenols, flavonoids, saponins and phytosterols, making it a complete nutritious food that can meet the basic nutritional needs of the human body [34]. The cultivation of quinoa outside of the Andean region has been heavily promoted by the United Nations Food and Agriculture Organization. Quinoa not only has extremely high nutritional value, but is also genetically diverse species with strong drought and salt tolerance that can be grown in marginal soils with pH values from 4.5 to 9.5 [35]. Exploring the biological basis of its excellent nutritional and agronomic characteristics has become a major focus of current crop research. In 2017, a high-quality reference genome of quinoa was released, which not only sparked a boom in quinoa research, but also promoted the breeding and cultivation of quinoa [34]. However, studies of the SnRK2 gene family in quinoa have not been reported at present. Therefore, in this study, we identified 13 SnRK2 family members from quinoa, and analyzed their basic physical and chemical properties, gene structure, and promoter cis-elements. In addition, the study of the SnRK2 gene expression patterns in response to different stress treatments has given us a deeper understanding of this gene family gene in quinoa. The results of our study will provide a basis for future research on the functions of this gene family.

Identification and basic properties analysis of quinoa SnRK2 Gene
Based on protein homology searches, we identified 13 SnRKs genes from the quinoa genome, all of which encode proteins with complete serine/threonine protein kinase catalytic domains. These genes were named CqSnRK2.1-CqSnRK2.13 (Table 1). The lengths of the proteins predicted to be encoded by the CqSnRK2 gene family range from 133 aa (CqSnRK2.5) to 401aa (CqS-nRK2. 1/CqSnRK2.13). The predicted molecular weights are between 15,129.46 Da (CqSnRK2.5) and 45,652.03 Da (CqSnRK2.5). Ten of the proteins encoded by the CqS-nRK2 genes have a theoretical isoelectric point below 7, and are acidic proteins. The instability index is between 26.15 (CqSnRK2.10) and 48.06 (CqSnRK2.1). The average fatty acid index ranges from 79.42 (CqSnRK2.10) to 94.74 (CqSnRK2.10). The hydrophobicity index of all CqSnRK2 proteins is < 0, indicating that they are all hydrophilic protein. In addition, the subcellular localization prediction showed that nine of the proteins localize to the cytoplasm, and two of them (CqSnRK2.6 and CqSnRK2.9) are predicted to be localized to the nucleus.

Phylogenetic analysis of the SnRK2 family in quinoa
In order to understand the evolutionary relationships of the proteins encoded by the CqSnRK2 genes, we used 53 SnRK2 protein sequences from five different plant species to construct a phylogenetic tree (Fig. 1, Table  S1). Of these, 13 CqSnRK2 proteins were from quinoa, 10 ZmSnRK2 proteins were from maize, 10 AtSnRK2 proteins were from Arabidopsis, 8 StSnRK2 proteins were from potato, and 12 VrSnRK2 proteins were from mung bean. As shown in Fig. 1, the phylogenetic

Chromosomal locations and duplication of SnRK2 genes in quinoa
In order to further study the genetic divergence and gene duplication events in the CqSnRK2 gene family, we determined the chromosomal locations of the 13 CqS-nRK2 genes. The results (Fig.S1) showed that except for two CqSnRK2 genes located on Chr05 (B), the remaining 11 CqSnRK2 genes were evenly distributed on all 11 chromosomes.
In addition, gene duplication is considered to be an important force that drives evolution. Studies have shown that duplicate genes are produced at a very high rate, about 0.01 times per gene per million years. There are generally two types of replication, tandem repetition and fragment repetition. Tandem duplications occur on the same chromosome, while segmental duplications can occur on the same or different chromosomes. In this study, we found seven pairs of duplicated genes ( Table 2) that can be classified as segmental duplication events. In addition, the non-synonymous (K A ) and synonymous (K S ) substitution rates between the seven repeated gene pairs was calculated. Ka/Ks = 1 indicates "neutral or no selection", Ka/Ks < 1 indicates "negative or purifying selection"; and Ka/Ks > 1 indicates "positive selection". In this study, the Ka/Ks ratios of the seven pairs of repeated genes were all < 1, indicating that these gene pairs have undergone purifying selection. The intron-exon structure can provide a more systematic and comprehensive understanding of the conserved features and the evolution of the gene family. In this study, we found that most CqSnRK2 genes have a highly conserved distribution of exons and introns. In eight of the CqSnRK2 genes (except for CqSnRK2.1, CqSnRK2.4, CqS-nRK2.5, CqSnRK2.10, and CqSnRK2.13) the first exon has the same length as the last exon, and the CqSnRK2 genes in the same subfamilies have similar gene structures ( Fig. 2A, Table S2). The 13 CqSnRK2 genes all have between 3 and 12 introns and 4-13 exons. Ten of the CqSnRK2 genes have 7-9 exons. In addition, eight CqS-nRK2 genes have untranslated regions (UTRs) at the 3' and 5' ends, three genes (CqSnRK2.4, CqSnRK2.10, and CqSnRK2.13) have no terminal UTRs, and two genes (CqSnRK2. 1 and CqSnRK2.8) have no UTR at the 5' end.

Construction of a protein interaction network and cis-acting element analysis
The study of protein interaction networks not only helps to understand the functions of the individual proteins that comprise the network, but also helps to explain the biological functions of most proteins in the cell, and identifies key proteins in the protein interaction network to enable the study of proteins related to stress resistance. Therefore, in order to further understand the interaction of SnRK2 genes in quinoa, in this study, we constructed a protein interaction network diagram based on the homologous proteins from Arabidopsis ( Fig. 3). Among them, we found that AtSnRK2.5 (CqSnRK2.4 and CqSnRK2.5) and AtSnRK2.10 (CqS-nRK2.6, CqSnRK2.9 and CqSnRK2.13) participate in ABA-independent abiotic stress resistance and regulate the osmotic stress response [36]. AtSnRK2.8 (CqS-nRK2.2, CqSnRK2.11, and CqSnRK2.12) is involved in hormone signal transduction (salicylic acid) [37]. AtOST1 (Open Stomatal1) regulates ABA-mediated stomatal closure and is a protein kinase that functions upstream of the ROS (reactive oxygen species) formation pathway. OST1 domain 2 interacts with PP2C-type phosphatases (ABI1 and ABI2). In Arabidopsis, this interaction between PP2C and OST1 integrates ABA and osmotic stress signals and regulates the closure of stomata [38]. Therefore, CqSnRK2.1, CqSnRK2.3, CqS-nRK2.7, CqSnRK2.8, and CqSnRK2.10 may also have similar functions in this study. In addition, it has been shown that SnRK2 can directly phosphorylate ABF proteins and the ABI1 protein [39]. In this study, SnRK2 protein can interact with proteins such as ABF, ABI, and HABI to perform the corresponding functions.
Cis-acting elements located in the promoter region of a gene have multiple functions and play key roles in regulating the function of the gene. Therefore, we analyzed the promoter sequence upstream of the CqS-nRK2 genes. Eighteen cis-acting element elements with different sequences were identified upstream of the 13 CqSnRK2 genes (Fig. 4, Table S3). Further analysis revealed that these elements are mainly involved in hormone responses, the stress response, and tissuespecific expression. In terms of hormone responses, eight cis-elements respond to ABA (ABRE, 3-AF1 binding site, AAGAA-motif, A-box, AT-TATA-box,  CAAT-box, motif-sequence, and TATA-box). One cis element responds to methyl jasmonate (CAAT-box), one cis element responds to auxin (CAAT-box), and two cis elements respond to gibberellin and salicylic acid (CAAT-box and TATA-box). The promoters of 12 of the CqSnRK2 genes (all except CqSnRK2.3) contain ABA response elements. Three cis-elements (AAGAAmotif, A-box, and CAAT-box) are involved in inducing the drought response, the CAAT-box element responds to low temperature, and the CAAT-box and TATA-box  In terms of tissue-specific expression, the 3-AF1 binding site, TATA-box, AAGAA-motif, and CAAT-box elements are related to endosperm expression, motifsequence is related to anaerobic induction and zein metabolism, and the TATA-box starts in the root effect.

Expression patterns of CqSnRK2 genes in different tissues and stress treatments
Using transcriptome data, we evaluated the expression patterns of the CqSnRK2 genes in the different stress treatments (drought, low temperature, heat, and salt) and in the different tissues and organs (  6, and CqSnRK2.9) at 5 and 7 days after drought stress were significantly lower than in the control. In addition, the relative expression of CqS-nRK2.10 gene was significantly higher than in the control (50-fold) at 5 days after drought stress. Expression of quinoa SnRK2 family genes is induced by PEG6000, low temperature, ABA, and NaCl treatments (Fig. 6, Table S7). Analysis of SnRK2 gene expression at different time points over a 48 h period revealed that the relative expression levels of SnRK2 genes were significantly different at different treatment times. The results of qRT-PCR analysis showed that after the 200 mmol/L NaCl treatment, the relative expression of five genes (CqSnRK2.4, CqSnRK2. 10,CqSnRK2.11,CqSnRK2.12,and CqSnRK2.13) reached the maximum at 3 h after treatment, and the expression of CqSnRK2.10 was upregulated to 60-fold that of the 0 h control. Five genes (CqSNRK2.1, CqSnRK2.2, CqSnRK2.5, CqSnRK2.7, and CqSnRK2.8) showed the highest relative expression levels at 24 h after treatment. The expression level of CqSnRK2.5 was 75 times higher than at 0 h. In the 200 μmol/L ABA treatment, we found that the relative expression of most CqSnRK2 genes changed little after ABA treatment, and the relative expression of seven genes (CqSnRK2.1-CqS-nRK2.4, CqSnRK2.6, CqSnRK2.9, and CqSnRK2. 13) was tenfold that of the control at all time points after treatment. The relative expression of nine genes (CqSnRK2.1-CqSnRK2. 5,CqSnRK2.7,CqSnRK2.8,CqSnRK2.11,and CqSnRK2.13) peaked at 48 h after treatment, and there were significant differences compared with expression at 0 h. CqSnRK2.10 and CqSnRK2.12 are a duplicated gene pair, and the relative expression level of both reached the maximum at 3 h after treatment. CqSnRK2.10 had the largest relative up-regulation, which was 34.14-fold higher than expression at 0 h; CqSnRK2.6 expression peaked at 9 h after treatment, and was 1.98-fold higher that at 0 h. CqSnRK2.9 was down-regulated at all time points after ABA treatment, which implies that this gene plays a negative regulatory role in plant abiotic stress. After 20% PEG treatment, there were significant differences in the expression of the CqSnRK2 family genes and the control. We found that the expression of seven genes (CqSnRK2.1, CqSnRK2.2, CqSnRK2.6, CqSnRK2.9, CqS-nRK2. 10,CqSnRK2.11,and CqSnRK2.12) reached their maximum at 48 h after treatment. Among these genes, the relative expression of CqSnRK2.11 was the most highly up-regulated, and was 80-fold higher than expression at 0 h. In addition, we found that the expression of specific genes was significantly different from the control The data are the mean ± SE of three independent biological samples, the vertical line represents the standard deviation, and the primer information is in Table S4 at certain time points after PEG treatment (CqSnRK2.1 gene was significantly different from the control at 12 h and 48 h after the treatment, and CqSnRK2.2 expression was only significantly different from the control at 48 h after treatment), and expression of some genes, such as CqSnRK2.8, CqSnRK2.10, CqSnRK2.11, CqSnRK2.12, and CqSnRK2.13 was significantly different from the control at multiple time points after PEG treatment. In the 4℃ Fig. 6 qRT-PCR analysis of 13 CqSnRK2 gene expression patterns under different stress (200 mmol/L NaCl, 200umol/L ABA, 20% PEG and 4℃ low temperature). TUB-9 of quinoa as an internal reference gene, the relative expression level was calculated by 2 −ΔΔCt , and different lowercase letters represented significance at 0.05 level (P < 0.05). The data are the mean ± SE of three independent biological samples, the vertical line represents the standard deviation low temperature treatment, the relative expression levels of 12 genes reached the maximum at 9 h (CqSnRK2.2, CqSnRK2.4, CqSnRK2.7, CqSnRK2.8, CqSnRK2.9, CqS-nRK2.10, and CqSnRK2.13) or 24 h (CqSnRK2.1, CqS-nRK2.3, CqSnRK2.6, CqSnRK2.11, and CqSnRK2.12) after treatment, and the expression of CqSnRK2.5 at 9 h and 24 h was also relatively high compared to the control. Thus, our results show that the relative expression almost all of the CqSnRK2 genes showed an increasing expression pattern from 0 h-9 h, and the relative expression of genes between 9 and 48 h showed a decrease-increasedecrease expression pattern. Expression levels of 10 CqSnRK2 genes (CqSnRK2.1, CqSnRK2.4, CqSnRK2.5, CqSnRK2.7-CqSnRK2.13) were significantly different from the controls at all time points after low temperature treatment. The relative expression level of CqSnRK2.8 at each time point after low temperature treatment was very high (at least 40-fold that of the control), leading us to hypothesize that this gene plays a key role in the response of plants to low temperature stress.

Subcellular localization of CqSnRK2.12 and detection of self-activation in yeast
The DNA construct carrying the chimeric CqSnRK2.12-GFP gene in Agrobacterium was infiltrated into tobacco leaves and the fluorescent signal was observed in the epidermal cells by laser scanning confocal microscopy. As shown in Fig. 7A, the CqSnRK2.12-GFP fusion protein was expressed in the cytoplasm and the nucleus, which was consistent with the predicted localization for the CqSnRK2.12 protein. In addition, transcriptional self-activation of the protein was detected. The results of the bait vector self-activation test (Fig. 7B) showed that yeast cells containing the positive control vector could grow on all three media. The negative control grew only on SD/-Trp/-Leu medium, and yeast cells containing PGBKT7-CqSnRK2.12 + pGADT7 were able to grow on SD/-Trp/-Leu medium. However, there was no colony growth on SD/-Leu/-Trp/-His/X-α-Gal and SD-Leu/-Trp/-His/-Ade/X-α-Gal media. These results showed that the CqSnRK2.12 protein did not have transcriptional self-activation.

Overexpression of CqSnRK2.12 increased salt and drought stress tolerance in Arabidopsis
In order to further study the function of CqSNRK2.12 gene, we introduced CqSnRK2.12 into Arabidopsis thaliana by Agrobacterium tumefaciens-mediated transformation, and obtained homozygous T 3 -generation over-expression lines (OE-1, OE-2, OE-3). Under normal growth conditions, we found that there was no significant difference in growth between the transgenic OE lines and the WT (Fig. 8a, b). However, in the 50 mM NaCl and 20% PEG treatments, we observed significant differences in growth between the transgenic and WT plants. In the 50 mM NaCl treatment (Fig. 8c, d), the root lengths of transgenic and WT plants were inhibited compared with a Sterile water for 7 days. b Root length statistics in Sterile water. c 50 mM NaCl for seven days. d Root length statistics in 50 mM NaCl for seven days. e 20% PEG for 7 days. f Root length statistics in 20% PEG for 7 days. Different lowercase letters represented significance at 0.05 level (P < 0.05). The data are the mean ± SE of three independent biological samples, the vertical line represents the standard deviation the CK, (no salt) but the root lengths of plants overexpressing CqSnRK2.12 were significantly longer than in WT plants; the average root length in the transgenic plants was 1.718-fold that of wild plants. In the 20% PEG treatment (Fig. 8e, f ), the average root length of the OE transgenic plants was 2.21-fold that of WT plants. These results indicate that CqSnRK2.12 participates in the regulation of drought and salt stress responses, and that overexpression of CqSNRK2.12 could induce drought and salt tolerance in transgenic Arabidopsis plants.

Discussion
When plants encounter harmful external environments, they can adapt to stresses through various morphological, physiological, and molecular reactions. In these processes, the phosphorylation of protein kinases induced by stress plays an important role in plant sensing and the response to environmental stress [40]. The SnRK2 gene family encodes plant-specific Ser/Thr protein kinases that play important roles in a variety of plant signal transduction and stress responses (drought resistance, cold resistance, and salt-alkali resistance) [22]. In this study, we identified 13 SnRK2 genes the quinoa (Chenopodium quinoa) genome. The number of SnRK2 gene family members in the monocots Brachypodium (10), rice (10), wheat (10), maize (11), and sugarcane (10) are between 10 and 11, indicating that this gene family is highly conserved in monocotyledonous plants. The number of SnRK2 family genes is more variable in dicotyledons, ranging between 6 and 22; examples are Arabidopsis (10), soybean (22), cotton (20), cherry (6), and Populus pilosa (12) [9,[19][20][21][22][23][24][25][26][27][28]. The higher numbers in some species indicate that this gene family is prone to gene duplication in dicots. The number of SnRK2 gene family members in the moss species Physcomitrium patens is four. Compared with bryophytes, the number of SnRK2 protein in angiosperms has increased significantly, which may be related to polyploidization and gene duplication in angiosperms, and may also be because the genes in the SnRK2 family expanded due to the whole genome duplication events that occurred after these plant lineages diverged. At the same time, we can clearly see that the number of SnRK2 genes shows an evolutionary trend from aquatic algae to terrestrial mosses and ferns to angiosperms, which is consistent with a previous report [5]. A phylogenetic analysis showed that the proteins encoded by the SnRK2 genes group in three clades in this study. Previous studies showed that the SnRK2 proteins from multiple plants species also cluster into three discrete groups [20,21,27,41]. Sequence alignments showed that the N-terminus of the SnRK2 proteins contains a relatively conserved protein kinase catalytic domain, while the C-terminus is highly specific and consists of an activation domain that responds to abiotic stress and an ABA acidic amino acid regulatory domain. Subcellular localization prediction of the quinoa SnRK2 proteins indicates that the members of this family are mainly localized to the cytoplasm and nucleus, while CqSnRK2.4 is found in the microbody, and may be involved in the glycolysis pathway. CqSnRK2.13 is localized to the plasma membrane, and its function could be related to the synthesis of membrane lipoproteins.
Gene structure prediction shows that most members of the quinoa SnRK2 gene family have 7-9 exons, and the distribution of exons within each subfamily is similar. There are large differences in the sequences of SnRK2 genes between the different subfamilies which is mainly due to the differences in the lengths of introns and other non-coding regions, which is consistent with findings in cotton and rice [10,24]. Motif analysis showed that the order of the motifs is conserved in each subfamily, and that most of the motif positions within the same subfamily are also basically the same. This may indicate that genes in the same subfamily have similar functions. Only motif 9 is quite different, indicating that the motifs of the same subfamily have similar functions. Wheat SnRK2 proteins show a high degree of evolutionary conservation, which is consistent with research results in mung bean [23]. These results point out the high-level evolutionary conservation of the SnRK2 protein family in plants. During evolution, most gene families expand through continuous duplication and differentiation. On the one hand, gene duplication and differentiation provide abundant original genetic material, but on the other hand, these processes can also produce most of the genetic variation which makes it possible to adapt to changing environmental conditions through natural selection. Gene duplication analysis in this study identified seven pairs of duplicated genes in the CqSnRK2 family, which led to an expansion of the number of gene family members. Duplication events including segmental duplication and tandem repeats have played critical roles in the expansion of gene families in plants [42,43]. In addition, based on the chromosomal location and changes,such as mutations in gene structure, new members may show diverse functions [44,45]. Our findings show that some duplicated CqSnRK2 genes in quinoa experienced purifying selection and have evolved new functions.
When plants are subjected to abiotic stresses such as drought, high temperature, low temperature, and salinity, the signals related to the perception of these stresses will be converted into responses through a series of signal transduction pathways [1]. These responses will activate the transcription of genes that encode stress-related transcription factors that then bind to the cis-acting regulatory elements in the promoters of the corresponding genes, initiating the expression of these genes in the specific response to the stress [1,2,46]. Therefore, the functional study of cis-acting elements present in stress resistance-related genes is particularly important for revealing the mechanisms that control plant stress resistance. The promoter regions of SnRK2 genes usually contain cis-acting elements related to hormones and stress responses, such as the ABA response element ABRE, the gibberellic acid response element GARE-motif, the methyl jasmonate response element CGTCA-motif, the heat shock response element HSE, the low temperature stress response element LTRE, the drought response element DRE, the stress response element TC-rich repeats, and the MYB transcription factor binding element MBS. These cis-actig elements are related to hormone signal transduction and stress responses [36], and the SnRK2 gene family plays a crucial role in the plant response to abiotic stress and ABA-dependent plant development [47,48]. In this study, we found that CqSnRK2.11 (19 cis-acting elements) and CqSnRK2.12 (37 cis-acting elements) contain the AAGAA-motif, AT-TATAbox, CAAT-box, and motif-sequence elements, and most CqSnRK2 genes contain 6-12 cis-acting elements (between 3-6 element types). We can speculate that most of the genes in this gene family are involved in the stress response mechanisms to cold, drought, and salt tolerance. Among them, the CqSnRK2.12 gene may be involved in a variety of stress responses. In addition, we found that 12 CqSnRK2 genes (all except CqSnRK2.3) contain ABA response elements, which may be an important reason why the transcription of genes in this family is strongly induced by ABA.
Our study found that the relative expression levels of the CqSnRK2.2, CqSnRK2.5 and CqSnRK2.11 genes were high in stems, flowers, and mature seeds, and very low in leaves and developing seeds. Expression of CqS-nRK2.6 and CqSnRK2.9 was almost undetectable in inflorescences and apical meristems, while CqSnRK2.5 and CqSnRK2.13 were highly expressed in both the inflorescence and apical meristem. The expression levels of most CqSnRK2 genes were higher in white quinoa flowers than in yellow quinoa flowers, and the expression levels of all CqSnRK2 genes in white quinoa fruits were higher than those in yellow quinoa fruits. It can be seen that the SnRK2 gene family members are expressed in various tissues of quinoa, but that the expression levels in vary in the different tissues t, showing a certain degree of tissue/organ specificity. In addition, the experimental results also reflect the differences between varieties. Kobayashi [10] used Northern blot hybridization to detect the expression of the SnRK2 gene family SAPK in rice tissues such as leaves, leaf sheaths, and roots under different stress conditions, and showed that the expression of members of this gene family was different in different tissues. The conclusions of this study are consistent with the results of our study. The specificity of the expression of SnRK2 gene family members in different tissues under normal conditions may be because they mediate different physiological metabolic processes in plants under normal growth conditions. Studies have shown that SnRK2 proteins regulate seed dormancy and germination, sex differentiation, root system morphogenesis, flowering, fruit ripening, yield formation, plant height, and other processes that affect the growth and development of plants [49]. This may be the reason for the differential expression of SnRK2 genes in different tissues found in this study.
The SnRK2 proteins comprise a family of protein kinases that participate in the responses to abiotic stresses in plants. SnRK2 proteins play important roles in the normal growth and development of plants, responding to ABA signals as well as abiotic stress caused by drought, high salt, low temperature, and salinity [5]. Previous studies have shown that overexpression of the wheat TaSnRK2.4 gene can significantly improve plant tolerance to salt, drought, and freezing damage [50]. Overexpression of AtSnRK2.8 significantly improved drought resistance [51]. In this study, the expression of quinoa genes homologous to AtSNRK2.8 (CqSnRK2.2, CqSnRK2.11, and CqSnRK2.12) increased significantly in response to drought and PEG treatments. These three quinoa CqSnRK2 genes belong to the Group II subfamily, and their expression is slightly induced by ABA. Therefore, their function in the drought response may be mediated through the ABA signaling pathway. At the same time, the promoter regions of these three genes contain multiple ABA response elements, especially CqSnRK2.12, which contains 21 ABA response elements. However, it is interesting that the expression of CqSnRK2.12 under drought stress was not as high as that of CqSnRK2.2 and CqSnRK2.11, suggesting that there may be other uncharacterized cis-elements or unknown mechanisms associated with stress that are involved in the regulation of these genes. AtSnRK2.6, also known as OST1, belongs to the Group III subfamily and is strongly induced by ABA. This gene is the main positive regulator of ABA signal transduction and stomatal control [49]. Silencing of AtSnRK2.6 expression leads to damage to stomatal closure and serious water loss in Arabidopsis leaves, indicating that AtSnRK2.6 mainly plays a positive regulatory role in drought stress. In this study, six quinoa genes (CqSnRK2.1, CqSnRK2.3, and CqSnRK2.6-CqS-nRK2.9) belonging to Group III were strongly expressed in response to drought stress. The high expression levels of these genes may induce the ABA signaling pathway and result in stomatal closure, thus enhancing drought resistance in quinoa. In addition, in Arabidopsis, AtSnRK2.2, AtSnRK2.3, and AtSnRK2.6 are usually activated by ABA and the proteins encoded by these gene can phosphorylate ABRE binding factors. Therefore, they are important for the activation of ABA-responsive genes. The AAPK gene in broad bean is also a member of the SnRK2 family, and it is induced by ABA in guard cells to regulate stomatal closure [52]. The overexpression of SPK3 in soybean is induced by exogenous ABA, which increases the plant response to hyperosmotic stress [53]. SnRK2.10 in Populus tomentosa interacts with ABI1, AHG1, and AHG3 to participate in ABA signal transduction, and responds to drought and salt stress [27]. The protein interaction network predicted in this study shows that quinoa SnRK2 family proteins can also interact with ABI1, ABI2, and ABF2 and interact with HAB1 to respond to abiotic stresses such as drought, salinity, and low temperature. Studies have shown that in Arabidopsis, except for AtSnRK2.9, other SnRK2 protein kinase genes can be activated by substances that regulate osmotic pressure, such as sucrose, mannitol, sorbitol, or NaCl [5]. This result was also confirmed in our study, and the expression of the gene family members changed significantly in response to the three abiotic stresses. Studies have shown that transgenic plants overexpressing the maize SAPK8 and rice SAPK4 genes have significantly enhanced salt tolerance compared to WT controls [31]. Previous studies have shown that SnRK2 Group III genes are important components of ABA signaling pathways and can be induced by exogenous ABA [10]. In this study, we found that transcription of most of the Group I, Group II, and Group III genes can be induced by ABA, a result that is inconsistent with previous studies in Arabidopsis and rice [10], which indicates that quinoa has a complex ABA-dependent signal transduction pathway. Studies have shown that all AtSnRK2 family members in Arabidopsis are not induced by low temperature, but some SnRK2 genes have a strong response to cold, including PKABA1, TaSnRK2. 3,TaSnRK2.4,ZmSnRK2.3,ZmSnRK2.7,and GsAPK [25]. In contrast, our results also show that expression of all CqSnRK2s in quinoa can be induced by low temperature and are significantly up-regulated, but it is not entirely clear whether their expression contributes to conferring frost resistance and enhanced plant survival; therefore, this finding provides a new perspective on the involvement of SnRK2 genes in the response to low temperature stress. At the same time, we found an interesting phenomenon that the expression patterns of partially duplicated genes were slightly different or significantly different, such as the duplicated gene pair CqSnRK2.3/CqSnRK2.4, which presented different expression patterns after PEG stress. It indicated that these two genes Functional divergence occurred during evolution, and this functional divergence may be related to homeopathic elements contained in the promoter region, which was also reported in Wang's study [46]. Although the CqSnRK2 family genes are involved in the responses to a variety of abiotic stresses, the effects of these stresses on expression do not fully reflect their function, because their activation is more directly and rapidly regulated after translation. Therefore, more studies on the functional characterization of CqSnRK2 genes are needed.

Conclusion
In this study, we systematically identified the SnRK2 family from the quinoa genome. We used bioinformatics method to describe the physical and chemical properties, gene structure, protein interactions and promoter elements of 13 SnRK2 genes. Gene structure and motif analyses corroborated the phylogenetic analysis results. The cis-elements found in the promoters of the CqS-nRK2 gene were found to be associated with phytohormones and abiotic stresses. RT-qPCR analysis revealed that CqSnRK2. CqSnRK2.8, were significantly up-regulated in response to low temperature, salt and drought stress. In addition, the overexpression of CqSnRK2.12 increased salt and drought stress tolerance in Arabidopsis.

Materials and treatments
The test material used was the quinoa cultivar 'Longliu No. 2 that was obtained as seed from the Gansu Academy of Agricultural Sciences. Seed was sown in the greenhouse of the College of Life Science and Technology of Gansu Agricultural University in March 2021. Quinoa seeds of the same size and fullness that were free of pests and diseases were selected and soaked in warm water for 20 min, after which they were sown in pots containing 2 kg of sandy loam soil. Thirty seeds were sown per pot and were covered with vermiculite, after which the pots were transferred to a greenhouse and grown at a temperature of 20-25℃ with front and back ventilation, natural light, and normal watering. When the seedlings had grown to the 6-8 leaf stage (about one months after sowing), they were thinned, leaving 10 seedlings in each pot. After the seedlings are colonized, they were subjected to five stress treatments: (1) PEG stress: 20% PEG; (2) Salt stress: 200 mmol/L NaCl; (3) Low temperature stress: 4℃; (4) Hormone treatment: 200 μmol/L ABA. The treatments were performed at 0 h when the light period started, and quinoa leaves were sampled at 0, 3, 6, 9, 12, 24, and 48 h after the initiation of the treatments. (5) Drought treatment: plants were not watered, and quinoa leaves and roots were sampled at 0, 3, 5, and 7 days after watering ceased. The collected materials were placed at -80 °C prior to their use in the subsequent experiments. There were three replicates for each treatment.

RNA extraction and cDNA synthesis
Total RNA was extracted using AG RNA e x Pro Reagent (Beijing, China). The extracted RNA was the template for cDNA synthesis using the Evo M-MLV RT Kit (Beijing, China) to reverse transcribe 0.5-2 μg of purified total RNA into first-strand cDNA for gene amplification and real-time quantitative PCR (qRT-PCR) assays.

Identification and cloning of the Quinoa SnRK2 protein kinase family
In this study, the deduced amino acid sequence of the Arabidopsis SnRK2 protein was used as the query sequence. Genes encoding homologous proteins were identified in the quinoa genome [34], and the quinoa SnRK2 gene family members were initially screened as candidate genes. We then used HMMER (https:// www. ebi. ac. uk/ Tools/ hmmer/) to search for conserved sequences and to eliminate candidate genes for proteins that do not contain the specific SnRK2 kinase domain (registration number: PF00069.25). The conserved CqSnRK2 domains were determined using the tools NCBI-CDD (https:// www. ncbi. nlm. nih. gov/ cdd/), SMART(http:// smart. embl. de/) and Pfam (http:// pfam. xfam. org/). Finally, 13 CqSnRK2 genes were identified.

Construction of a phylogenetic tree of the Quinoa SnRK2 protein family and related information analysis
After screening, the quinoa gene family that encodes SnRK2 proteins was analyzed using the Expasy (https:// web. expasy. org/) online tool ProtParam to compute the physical and chemical properties such as the number of amino acids, theoretical isoelectric point, molecular weight, and hydrophobicity index. The PSORT Prediction (http:// psort1. hgc. jp/ form. html) tool was used to predict the subcellular locations of the SnRK2 proteins [54]. A phylogenetic tree was then constructed with the Maximum Likelihood method as implemented in MEGA X using the default values [55]. The confidence levels of the individual branches were estimated by bootstrapping using 1,000 replicates. The quinoa GFF3 annotation file and genome sequence were imported into the software TBtools [56] to draw the exon-intron structural diagrams. The conserved protein motifs were analyzed with MEME (http:// meme-suite. org/ tools/ meme) [57]. TBtools was used to locate genes on chromosomes from the GFF annotation files [56]. We detected duplicate gene pairs using the plant genome replication database server http:// chibba. agtec. uga. edu/ dupli cation/ index/ locket). Clustal W software was used to predict the amino acid sequences of the partially repeated CqSnRK2 genes [55]. TBtools software calculates the Ka and Ks values of duplicated genes [56]. The protein interaction network was constructed based on orthologous proteins from the model plant Arabidopsis thaliana, and the network diagram was drawn by String (https:// cn. string-db. org/) [58]. The 2,000 bp of DNA sequence upstream of the CqSnRK2 genes was extracted by TBtools [56] as the promoter regions, and PlantCARE (http:// bioin for-matics. psb. ugent. be/ webto ols/ plant care/ html/) [59] was used to analyze the cis-elements in these promoter regions.

Expression pattern analysis and quantitative real-time fluorescence PCR analysis
The expression data for the different tissues and organs of quinoa and the different stress treatments (drought, salt, and low temperature) were downloaded from the NCBI BioProject database (accession numbers PRJNA394651 and PRJNA306026, respectively).
Fragments per kilobase of exon model per million mapped reads (FPKM) values were used to represent the expression levels of the SnRK2 genes. We selected the transcriptome data of genes belonging to the SnRK2 transcription factor family. Using TBtools software (https:// github. com/ CJChen/ TBtoo ls) [56], log 10 FPKM values were calculated, a heat map was drawn, and the SnRK2 genes in quinoa tissues and the different treatments were then analyzed [60]. Quantitative real-time fluorescence PCR (qRT-PCR) was performed on an Mx3005P Real-Time PCR System (Stratagene, USA) and the SYBR Green Premix Pro Taq HS qPCR Kit (Beijing, China) were used to determine gene expression levels. The primers for qRT-PCR were designed using the Takara online tool (https:// www. takar abio. com/). The internal reference gene control was TUB-9 (Table S4). The amplification system consisted of 1 μL cDNA, 0.5 μl upstream and downstream primers (10umol/L), 10 μl reaction mix, and 8 μl ddH 2 O in a total volume of 20 μL. The reaction conditions consisted of a single denaturation step at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. The fluorescence curve and the melting curve were analyzed after the reactions were completed. The relative expression of each gene was calculated using the 2 −∆∆CT method [61].

Subcellular localization of the CqSnRK2.12 protein and transcriptional activation of CqSnRK2.12
Using homologous recombination, the cloned CqS-nRK2.12 gene was transferred to an expression vector containing the gene for green fluorescent protein (GFP) to construct GV1300::CqSnRK2.12-GFP, and the vector plasmid was transferred to Agrobacterium GV1300 by electroporation. Positive clones were selected and cultured in liquid medium. The plasmid-containing Agrobacterium strains were infiltrated into tobacco (Nicotiana benthamiana) leaves. Expression of the CqSnRK2-12: GFP fusion protein in leaf epidermal cells was observed and imaged using a confocal laser scanning microscope (LSCM).
The CqSnRK2.12 gene was cloned into the plasmid pGBKT7 at the Nco I site in frame with the yeast GAL4 bonding domain (BD). The bacterial colonies were screened by PCR using 3' BD and T7 primers, and the amplified PCR products were sequenced. A plasmid extraction kit was used to extract the plasmid DNA from an E. coli culture containing the correct recombinant plasmid. The pGBKT7-CqSNRK2.12 vector was transferred into yeast strain AH109 and cultured on an SD/TRP plate at 30 °C. Colonies that grew on the plate after 2-3 d were screened by PCR. The recombinant strain was grown in liquid medium, and 10 μl aliquots of a dilution series (10 -1 , 10 -2 , 10 -3 , 10 -4 , and 10 -5 ) were spotted onto different media and the growth of the yeast was observed at 30 °C.

Generation of transgenic Arabidopsis thaliana plants and determination of related physiological parameters
The CqSnRK2.12 gene (without the terminator sequence) was transferred into the pCAMBIA-1302-EGFP vector by homologous recombination, and the recombinant plasmid was transformed into E. coli DH5α. After screening on solid medium containing kanamycin resistance, positive clones were selected and sequenced. The recombinant Plasmid pCAMBIA-1302: CqSnRK2.12 was transferred into Agrobacterium tumefaciens GV3101. Positive clones were identified, and plants of the wild-type (WT) A. thaliana ecotype Columbia (Col-0; seeds obtained from the Gansu Academy of Agricultural Sciences) were transformed by the floral dip method. The T 0 -generation transgenic seeds were sterilized in 6.25% NaClO solution for 15 min, washed with sterile ddH 2 O three times, and sown on MS (Murashige-Skoog) medium containing 50 μg/mL hygromycin B for selection of T 1 -generation transformants. Genomic DNA was extracted from the T 1 -generation plants to determine whether the CqS-nRK2.12 gene construct was inserted into the Arabidopsis genome. Total RNA was also extracted from leaves and used in qRT-PCR assays.
PCR detection: Genomic DNA was extracted from leaves of kanamycin resistant transgenic plants and WT A. thaliana Col-0 plants; the CqSnRK2.12-OE recombinant plasmid DNA was used as the positive control and Col-0 DNA was the negative control. Reference primer pairs HYG-417BP (F: 5'-AAA TCC GCG TGC ACG AGG  T-3'; R: 5'-TCC TTG TTG TTC CTT CGG TTC TGT GTG  TGC A-3') and HYG-501BP (F: 5'-GAG ATA CCG CCA GTC-3'; R: 5'-CAA GAC CCT GCG AAC CGA -3') were used to identify transgenic plants by PCR amplification. T 3 -generation transgenic plants were identified by hygromycin (150 mg/L) selection for subsequent experimental analysis. For the salt stress and drought treatments, seeds of the transgenic and WT lines were sterilized and cultured on 20% PEG and 50 mM NaCl, respectively. The plant phenotypes were observed and the root length was measured after 7 days of growth on MS medium at 22 • C.