Characterization of the AP2/ERF Transcription Factor Family and Expression Profiling of DREB Subfamily under Cold and Osmotic Stresses in Ammopiptanthus nanus.

APETALA2/ethylene-responsive factor (AP2/ERF) is one of the largest transcription factor (TF) families in plants, which play important roles in regulating plant growth, development, and response to environmental stresses. Ammopiptanthus nanus, an unusual evergreen broad-leaved shrub in the arid region in the northern temperate zone, demonstrates a strong tolerance to low temperature and drought stresses, and AP2/ERF transcription factors may contribute to the stress tolerance of A. nanus. In the current study, 174 AP2/ERF family members were identified from the A. nanus genome, and they were divided into five subfamilies, including 92 ERF members, 55 dehydration-responsive element binding (DREB) members, 24 AP2 members, 2 RAV members, and 1 Soloist member. Compared with the other leguminous plants, A. nanus has more members of the DREB subfamily and the B1 group of the ERF subfamily, and gene expansion in the AP2/ERF family is primarily driven by tandem and segmental duplications. Promoter analysis showed that many stress-related cis-acting elements existed in promoter regions of the DREB genes, implying that MYB, ICE1, and WRKY transcription factors regulate the expression of DREB genes in A. nanus. Expression profiling revealed that the majority of DREB members were responsive to osmotic and cold stresses, and several DREB genes such as EVM0023336.1 and EVM0013392.1 were highly induced by cold stress, which may play important roles in cold response in A. nanus. This study provided important data for understanding the evolution and functions of AP2/ERF and DREB transcription factors in A. nanus.


Introduction
Plants are inevitably subjected to various adverse conditions in their life cycle, such as cold stress, drought stress, and salt stress, which negatively affect their growth, development, and yield [1]. Plants respond to environmental stresses at both cellular and molecular levels by regulating the expression of many genes via a variety of signal transduction pathways, and several transcription factor (TF) families, including APETALA2/ethylene-responsive factor (AP2/ERF), WRKY, NAC (NAM, ATAF1, 2, and CUC2), and MYB (myeloblastosis oncogene), play crucial roles in stress signal transduction [2].
AP2/ERF family is one of the largest families in the plant kingdom, and proteins in this family contain at least one AP2/ERF DNA-binding domain that consists of 60~70 conserved amino acids [3]. The AP2 domain is important for the binding activity to the cis-acting elements including the Plants 2020, 9,455 3 of 20
The 174 putative AnAP2/ERF genes encode proteins ranging from 135 to 705 aa in length. The amino acid length of most DREB and ERF subfamily members are between 135 and 360, while those of all AP2 and RAV subfamily members are longer than 360. The pI values of the predicted AP2/ERF TFs range from 4.39 to 10.13, with 114 (65.52%) proteins with pI values > 7. To classify the evolutionary relationships of the AP2/ERF family members in A. nanus, the full-length amino acid sequences of the putative proteins were aligned and the phylogenetic trees were constructed (Figures 2-4). Based on the phylogenetic tree of the AnAP2/ERF family members, all AnAP2/ERF family members were classified into five subfamilies: 92 ERF members, 55 DREB members, 24 AP2 members, 2 RAV members, and 1 Soloist member. The DREB and ERF subfamilies were further divided into six groups (A1 to A6, B1 to B6). Groups A1, A2, A3, A4, A5, and A6 contain 8, 8, 2, 18, 10, and 9 members, respectively, and groups B1, B2, B3, B4, B5, and B6 contain 28,7,29,8,8, and 12 members, respectively [35] (Figures 3 and 4). The 174 putative AnAP2/ERF genes encode proteins ranging from 135 to 705 aa in length. The amino acid length of most DREB and ERF subfamily members are between 135 and 360, while those of all AP2 and RAV subfamily members are longer than 360. The pI values of the predicted AP2/ERF TFs range from 4.39 to 10.13, with 114 (65.52%) proteins with pI values > 7.

Chromosomal Locations of the AnAP2/ERF Family Genes
A total of 172 AnAP2/ERF genes were located on the 9 chromosomes of A. nanus ( Figure 5), and the precise locations of 2 AnAP2/ERF genes, EVM0029153.1 and EVM0032511.1, were not determined. All 172 genes were distributed across all 9 chromosomes but not evenly. More AnAP2/ERF genes were located on chromosomes 2 and 4 (37 and 27, respectively), while only 12 AnAP2/ERF genes were distributed on chromosome 1.

Chromosomal Locations of the AnAP2/ERF Family Genes
A total of 172 AnAP2/ERF genes were located on the 9 chromosomes of A. nanus ( Figure  5), and the precise locations of 2 AnAP2/ERF genes, EVM0029153.1 and EVM0032511.1, were not determined. All 172 genes were distributed across all 9 chromosomes but not evenly. More AnAP2/ERF genes were located on chromosomes 2 and 4 (37 and 27, respectively), while only 12 AnAP2/ERF genes were distributed on chromosome 1.

Gene Structure and Conserved Motifs of the AnAP2/ERF Family Transcription Factors
To analyze the conservation of the predicted AnAP2/ERF family TFs, the deduced amino acid sequences of AP2 domains were aligned with the corresponding sequences in A. thaliana AP2/ERF TFs, which have been identified previously [7] (Figures S1-S4). All of the deduced

Gene Structure and Conserved Motifs of the AnAP2/ERF Family Transcription Factors
To analyze the conservation of the predicted AnAP2/ERF family TFs, the deduced amino acid sequences of AP2 domains were aligned with the corresponding sequences in A. thaliana AP2/ERF TFs, which have been identified previously [7] (Figures S1-S4). All of the deduced amino acid sequences of the AP2 domain of the DREB subfamily proteins have a WLG motif (W33, L34, G35), except for EVM0000157.1, which contains V34. Of the 6 groups of the DREB subfamily, A2 and A6 exhibit a higher level of conservation than other groups ( Figure S1).
The sequence alignments of the ERF B1 to B6 groups showed that most of the AP2 domains of ERF members from A. nanus and A. thaliana contain residues G3, R7, G10, E15, I16, W28, G30, T31, and Y42, and that all the AP2 domains of these ERF members contain residues W28, G30, T31, and Y42. The WLG motif (W28, L29, G30) is also present in almost all the AP2 domains of the ERF subfamily, with the exception of EVM0025962.1, EVM0032511.1, and EVM0032149.1, which contain YLG, WIG, and WIG, respectively. Of the 6 groups of the ERF subfamily, groups B4 and B5 exhibit a high level of conservation, while the AP2 domains of group B6 show a considerable degree of variation ( Figure S2).
EVM0002664.1 and EVM0036388.1 contain the AP2/ERF domain and the B3 domain, thus they were assigned to the RAV subfamily. Twenty-four AnAP2/ERF family members contain two AP2/ERF domains, therefore they are classified into the AP2 subfamily. The AP2 domains of A. nanus RAV and AP2 subfamily members share many conservative amino acid residues with those of A. thaliana. The WLG motif in the RAV subfamily in A. nanus is highly conserved, while in the AP2 subfamily the motif is presented as YLG ( Figures S3 and S4).
The conserved motifs in AnAP2/ERF family proteins were further investigated using MEME [36], and a total of 20 conserved motifs were found ( Figure 6, Figure S5 and S6). Motifs 1, 2, and 5, namely the three motifs located in the AP2 domain regions, were found in nearly all AP2/ERF family members. In the AP2 subfamily of A. nanus, motifs 1~5 were mainly located in the region of two consecutive AP2 domains. It is noteworthy that the AnAP2/ERF members in the same subfamily often have the same set of motifs, and many predicted motifs were distributed in the AnAP2/ERF family in a subfamily-specific manner. For example, motif 20 was only detected in group A1 of subfamily DREB, motif 14 was only detected in group A6 of subfamily DREB, motif 6 was found in groups A1, A4, and A5 of subfamily DREB, motifs 11 and 12 were found in group B3 of subfamily ERF, and motif 9 was specifically detected in groups B5 and B6 of subfamily ERF.
Gene structural analyses revealed that most DREB and ERF subfamily members and all members in the RAV subfamily were intronless ( Figure 6 and Figure S6). Of the 55 members in the DREB subfamily, 53 contained one or no introns (41 members contained no introns); one member contained two introns; and one member contained six introns, which belonged to group A2 of subfamily DREB. In the ERF subfamily, 87 members contained one or no introns (56 members contained no introns), the rest of the 5 members contained 2~4 introns. In contrast, the AP2 and Soloist subfamilies contained more introns than the DREB, ERF, and RAV subfamilies. All the AP2 members contained at least 7 introns and at most 9, and only one member of the Soloist subfamily contained 5 introns.

Duplication and Divergence of AP2/ERF Family Gene in A. nanus
Tandem and segmental duplication play important roles in the evolution of large gene families in plants. To investigate the evolution of the AnAP2/ERF gene family, tandem duplication events in the AnAP2/ERF gene family were analyzed. We found 27 AP2/ERF family genes present in tandem duplication in the A. nanus genome, including 17 ERF subfamily genes (17/92, 18.48%) and 10 DREB subfamily genes (10/55, 18.18%) (Table S2). These AnAP2/ERF genes were located in 11 tandem duplication clusters on 4 chromosomes, i.e., chromosomes 2, 3, 4, and 8 ( Figure 5). Most tandem duplication gene clusters were distributed on chromosomes 2 and 4; specifically, there were 12 AP2/ERF family genes in the 3 tandem duplication gene clusters on chromosome 4 of A. nanus. It is noteworthy that all the genes with tandem duplication relationships were members of groups B1 and B3 in subfamily ERF and groups A1 and A4 in subfamily DREB. member contained two introns; and one member contained six introns, which belonged to group A2 of subfamily DREB. In the ERF subfamily, 87 members contained one or no introns (56 members contained no introns), the rest of the 5 members contained 2~4 introns. In contrast, the AP2 and Soloist subfamilies contained more introns than the DREB, ERF, and RAV subfamilies. All the AP2 members contained at least 7 introns and at most 9, and only one member of the Soloist subfamily contained 5 introns. . Figure 6. Representative motif compositions and gene structures of the AP2/ERF family members in A. nanus. One or two members from Soloist, RAV, and AP2 subfamilies and one member from each ERF and DREB subfamilies were used to show the typical motif A B Figure 6. Representative motif compositions and gene structures of the AP2/ERF family members in A. nanus. One or two members from Soloist, RAV, and AP2 subfamilies and one member from each ERF and DREB subfamilies were used to show the typical motif compositions and gene structures of members in the subfamily or the group. (A) Conserved motif analysis of the AnAP2/ERF family transcription factors. All motifs were identified using the MEME database. (B) Gene structure of AnAP2/ERF family members. Exons and introns are represented by yellow boxes and black lines, respectively, and the untranslated region (UTR) is shown in green.
The rate of non-synonymous (Ka) versus synonymous (Ks) is a common index to judge the existence of selection pressure. Ka/Ks = 1 stands for neutral selection, Ka/Ks >1 indicates accelerated evolution with positive selection, and Ka/Ks <1 means purifying selection [37]. Using the reciprocal best BLAST algorithm, a total of 32 paralogous pairs in AnAP2/ERF family were determined and Ka, Ks, and Ka/Ks values were calculated (Table S4). The Ks distribution peak fell in the 0.50-0.60 range in A. nanus, and the peak of Ka/Ks values fell in the 0.25-0.30 range (Figure 8). A total of 95 orthologous pairs in the AP2/ERF gene family between A. nanus and G. max were identified (Table S5)  3, 4, and 8 ( Figure 5). Most tandem duplication gene clusters were distributed on chromosomes 2 and 4; specifically, there were 12 AP2/ERF family genes in the 3 tandem duplication gene clusters on chromosome 4 of A. nanus. It is noteworthy that all the genes with tandem duplication relationships were members of groups B1 and B3 in subfamily ERF and groups A1 and A4 in subfamily DREB.
We also investigated the segmental duplication of the AP2/ERF gene family in the A. nanus genome using synteny analysis ( Figure 7). A total of 83 AP2/ERF genes were found to be present in tandem duplication in the A. nanus genome, including 30 DREB subfamily genes (30/55, 54.55%), 47 ERF subfamily genes (47/92, 51.09%), 4 AP2 subfamily genes (4/24, 16.67%), and RAV subfamily genes (2/2, 100%) (Table S3). The rate of non-synonymous (Ka) versus synonymous (Ks) is a common index to judge the existence of selection pressure. Ka/Ks = 1 stands for neutral selection, Ka/Ks >1 indicates accelerated evolution with positive selection, and Ka/Ks <1 means purifying selection [37]. Using the reciprocal best BLAST algorithm, a total of 32 paralogous pairs in AnAP2/ERF family were determined and Ka, Ks, and Ka/Ks values were calculated (Table S4). The Ks distribution peak fell in the 0.50-0.60 range in A. nanus, and the peak of Ka/Ks values fell in the 0.25-0.30 range ( Figure 8). A total of 95 orthologous pairs in the AP2/ERF gene family between A. nanus and G. max were identified (Table S5) and the peak of Ks distribution curve was in the 0.4-0.5

ID
Core sequence Cis-acting element Putative Function Figure 9. Cis-acting element prediction in the promoter region of the DREB subfamily genes in A. nanus. The color of the square depicts the quantity of the predicted cis-acting elements in the promoter region.

Expression Profiles of the DREB Subfamily Genes under Osmotic and Cold Stresses in A. nanus
To evaluate the potential function of DREB subfamily genes, specifically, their roles in environmental stress response, gene expression patterns of DREB subfamily genes in A. nanus were investigated using the whole transcriptome sequencing (RNA-seq) datasets and quantitative reverse transcriptase PCR (qRT-PCR) analysis. First, RNA-seq datasets (NCBI SRA accession numbers SRR11089024 to SRR11089029 and SRR11087599 to SRR11087604) were used to quantitate the expression level of DREB subfamily genes in leaves of A. nanus seedlings grown under normal conditions or exposed to delayed short-term treatment (7d) of cold or osmotic stress ( Figure 10). The expression levels of 55 DREB subfamily members in unstressed leaves were highly variable; several genes exhibited high expression levels, such as EVM0014569.1, EVM0034483.1, EVM0017315.1, and EVM0026860.1, whereas some genes displayed very low level or were hard to detect, including EVM0000157.1, EVM0005962.1, EVM0013480.1, EVM0013907.1, EVM0030720.1, and EVM0035886.1.

Expression Profiles of the DREB Subfamily Genes under Osmotic and Cold Stresses in A. nanus
To evaluate the potential function of DREB subfamily genes, specifically, their roles in environmental stress response, gene expression patterns of DREB subfamily genes in A. nanus were investigated using the whole transcriptome sequencing (RNA-seq) datasets and quantitative reverse transcriptase PCR (qRT-PCR) analysis. First, RNA-seq datasets (NCBI SRA accession numbers SRR11089024 to SRR11089029 and SRR11087599 to SRR11087604) were used to quantitate the expression level of DREB subfamily genes in leaves of A. nanus seedlings grown under normal conditions or exposed to delayed short-term treatment (7d) of cold or osmotic stress ( Figure 10). The expression levels of 55 DREB subfamily members in unstressed leaves were highly variable; several genes exhibited high expression levels, such as EVM0014569.1, EVM0034483.1, EVM0017315.1, and EVM0026860.1, whereas some genes displayed very low level or were hard to detect, including  EVM0000157.1, EVM0005962.1, EVM0013480.1, EVM0013907.1, EVM0030720.1, and EVM0035886.1.
After delayed short-term osmotic stress or cold stress (7d), the expression of many DREB subfamily members changed significantly. EVM0026054.1 and EVM0014092.1 were upregulated, and EVM0003995.1 and EVM0016549.1 were downregulated under both stresses, whereas genes such as EVM0023826.1 and EVM0012847.1 were induced under cold stress, and were downregulated or did not exhibit significant change under osmotic stress.
A B Figure 10. Expression profiles of the DREB subfamily genes calculated using RNA-seq datasets. Transcript per million TPM values of all AnDREB subfamily genes in A. nanus leaves under normal conditions (A); fold change values of all AnDREB subfamily genes exposed to delayed short-term After delayed short-term osmotic stress or cold stress (7d), the expression of many DREB subfamily members changed significantly. EVM0026054.1 and EVM0014092.1 were upregulated, and EVM0003995.1 and EVM0016549.1 were downregulated under both stresses, whereas genes such as EVM0023826.1 and EVM0012847.1 were induced under cold stress, and were downregulated or did not exhibit significant change under osmotic stress.
To reveal the dynamic expression patterns of DREB subfamily genes under short-term osmotic and cold stresses, qRT-PCR analyses were conducted using A. nanus seedlings exposed to 3 h, 6 h, 12 h, and 24 h of cold or osmotic stress treatment (Figure 11). A total of 32 pairs of primers were used in qRT-PCR analysis, and we were unable to obtain effective primers for some other DREB genes, including the 6 DREB genes with very low abundance in leaves (i.e., EVM0000157.1, EVM0005962.1, EVM0013480.1, EVM0013907.1, EVM0030720.1, and EVM0035886.1).
Approximately half of the DREB subfamily genes (56.25%) were upregulated more than 2-fold under short-term osmotic stress, whereas under short-term cold stress, the percentage rose to 84.38%. This phenomenon is also observed in the RNA-seq analysis results, indicating that, in general, the response of DREB subfamily genes to cold stress is significantly greater than that of osmotic stress.
Specifically, EVM0026860.1, EVM0005020. These DERB family genes may play essential roles in the response to low-temperature and drought stress in A. nanus.

Discussion
High-quality whole genome sequences of A. nanus enabled the genome-wide identification of AP2-ERF family member genes and a comparison with other plant species. Both the total number of AnAP2/ERF families and the quantity of DREB subfamily members in A. nanus were similar to those of some leguminous plants such as P. vulgaris and C. cajan, and a little higher than those of some other leguminous plants including G. max, M. truncatula, and C. arietinum (Figure 1). Of the five subfamilies of the AP2/ERF family in A. nanus, the ERF subfamily has the largest number of members, followed by the DREB and AP2 subfamilies. Such a distribution pattern was also observed in other plants such as P. vulgaris (Figure 1). It is noteworthy that the number of DREB subfamily members in A. nanus is larger than those of the other leguminous plants (Table S6), which may be associated with the high number of group A1, A2, and A6 of the A. nanus DREB subfamily. In addition, we observed that the quantity of members in the B1 group is significantly higher than those of the other leguminous plants (Table S6). Whether these observations are associated with the abiotic stress tolerance of A. nanus needs further investigation.
Motifs identified in transcription factors can be involved in transcriptional activity, protein-protein interactions, and nuclear localization [7]. In the present study, 20 motifs were identified from the AnAP2/ERF family transcription factors. Motif 1~5 fell in the AP2/ERF domain, motifs 1, 2, and 5 were present in all family members, and motifs 3 and 4 were only found in the AP2 subfamily. The other motifs were located outside the AP2/ERF domain. Of them, the LNFP motif in motifs 1, 6, and 14 was DREB subfamily-specific, which were known to be related to disease resistance [40]. LPR[P/A] motif in motif 6 has been identified as essential signatures for calcineurin B-like (CBL)-interacting serine/threonine-protein kinase 12 [41]. Some motifs were present in a group-specific manner; for example, motif 20 was only detected in the DREB A1 group and motif 14 was only found in the DREB A6 group.
Gene duplications are suggested to be one of the primary driving forces in the evolution of genomes, and segmental duplication, tandem duplication, and transposition events, including retroposition and replicative transposition, are three principal evolutionary pathways to generate duplicated genes. Of the three pathways, segmental and tandem duplications are wide-spread events in plant genomes and play important roles in the evolution of gene families and adaptation to environmental changes in plants [42]. Of the 174 AnAP2/ERF family genes, 63.22% (110/174) were present in segmental or tandem duplication, suggesting that tandem and segmental duplication represent two principal pathways for gene expansion of the AP2/ERF family in A. nanus. Of the 174, 27 (15.52%) AP2/ERF subfamily genes in A. nanus were tandemly duplicated genes, whereas up to 47.70% (83/174) were segmentally duplicated genes. This quantity of segmentally duplicated genes was significantly higher than that of tandem-duplicated genes in the AP2/ERF family in A. nanus, and similar results have been reported in other plant species such as cucumber [43] and common bean [21]. The observation supported the hypothesis that, compared with tandem duplication, segmental duplication made a major contribution to the expansion of the AP2/ERF gene family in A. nanus. However, all tandem-duplicated genes in the AP2/ERF family in A. nanus belonged to the DREB and ERF subfamilies, indicating tandem duplication also played significant roles in the expansion of the DREB and ERF subfamilies.
Promoter analysis provided useful information to understand the upstream transcription factors that govern the tightly controlled regulation of DREB genes. In the past two decades, a large number of research studies on the regulation of DREB genes have revealed a complex network of different transcription factors involved in their regulation, and the DREB promoters have been proposed to be the central hubs that integrate multiple environmental and internal developmental signals [44]. Our results showed that promoter regions of AnAP2/ERF genes contained a certain number of MYB, MYC, and WRKY binding elements, suggesting that certain MYB, MYC, and WRKY transcription factors can regulate the expression of the AP2/ERF family genes in A. nanus.
Among the predicted cis-acting elements in DREB subfamily genes in A. nanus, MYC/ICE1 binding site (CANNTG, the canonical recognition sequence for bHLH family transcription factors) was one of the most frequently occurring ones. Nearly all of the genes carried the MYC/ICE1 binding site except EVM0011666.1, and 44 of the DREB genes (80.0%) carried more than six copies of the MYC/ICE1 binding site, indicating that this cis-acting element may play an important role in the regulation of DREB genes in A. nanus. The expression patterns of DREB genes under cold and osmotic stresses were consistent with the results of promoter analyses, and the large number of MYC/ICE1 binding sites in the promoters of DREB genes may mediate the induction of DREB subfamily genes under low-temperature and drought stress by interacting with ICE1 transcription factors in A. nanus. Previous research has shown that the overexpression of ICE1 enhanced the induction of DREB genes and their target genes, consequently increasing the levels of soluble sugars, and late embryogenesis abundant (LEA) proteins, and finally enhancing tolerance to osmotic and cold stresses in tobacco [45]. Although all cold-induced DREB genes contain multiple copies of the MYC/ICE1 binding site, the copy number of MYC/ICE1 binding sites in the promoters of DREB genes was not positively correlated with the fold change of the expression level.
Another frequently occurring cis-acting element in promoters of DREB subfamily genes in A. nanus is W-box (TTGAC, the binding site of WRKY family transcription factors), which is found in promoters of 50 (90.91%) DREB genes in A. nanus ( Figure 6). Considering the versatile biological function of WRKY family transcription factors [46], A. nanus DREB subfamily genes may respond to environmental signals via regulation by WRKY transcription factors. Some AP2/ERF family genes are arranged in clusters on the chromosome and they may be amplified via tandem duplication. These proteins have similar protein sequences, but we also want to know the evolutionary relationship between them and whether they have similar expression patterns under stress conditions. One of the largest clusters (cluster 10) contained 4 DREB genes, and their order is EVM0035043.2, EVM0026054.1, EVM0017798.1, and EVM0023336.1. The first three genes are members of group A4 and the last one is a member of group A1. Phylogenetic analysis showed that DREB A1 and A4 belong to the same large branch, and A1 is a branch of A4 (Figure 3). EVM0023336.1 is located in the branch of group A1, while EVM0035043.2, EVM0026054.1, and EVM0017798.1 form an independent small branch in branch A4. The multiple sequence alignment of AP2/ERF conservative domains of the DREB subfamily is consistent with the above observations ( Figure S1). We further speculate that the four DREB genes in cluster 10 may have evolved from two ancestor genes, and that EVM0035043.2, EVM0026054.1, and EVM0017798.1 originated from the same ancestor gene. Among the three genes, EVM0026054.1 and EVM0017798.1 are more similar in amino acid sequence, which may be derived from an ancestor gene that appears later. In addition, these four genes exhibited a considerable amount of expression in the leaves, and were upregulated under low temperature.
Both qRT-PCR analysis and RNA-seq datasets were used to evaluate the gene expression level of DREB genes under cold and osmotic stresses, and a collection of cold stress-responsive or osmotic stress-responsive DREB genes were identified. These highly induced DERB genes under cold and osmotic stresses may contribute greatly to the abiotic stress response of A. nanus, and thus deserve to be selected for further functional investigation by ectopic expression in a model plant such as Arabidopsis. Although there is no report on the transgenic functional study of DREB genes in A. nanus, the functional analyses of two DREB genes in A. mongolicus, a relative plant of A. nanus, supported our expression analysis results. AmDREB2C, the ortholog of EVM0017315.1, and AmDREB3, the ortholog of EVM0023826.1, were cloned, and their expression patterns and biological functions were studied [47,48]. AmDREB3 was shown to be induced by drought, salt, heat, cold, and abscisic acid treatments [48], and AmDREB2C exhibited the highest expression levels in winter in seedlings in the field, and its expression was induced by cold, heat, and drought stresses in laboratory-cultured seedlings [47]. The orthologs in A. nanus, EVM0017315.1 and EVM0023826.1, were also highly induced by cold stress (Figures 10 and 11). The transgenic Arabidopsis overexpressing the AmDREB2C and AmDREB3 enhances the tolerance to abiotic stress such as freezing, heat, and drought, indicating that the strategy of mining stress-resistant genes from a stress-resistant plant such as Ammopiptanthus is effective. In brief, the expression profiling of the AP2/ERF family genes under low-temperature and drought stress will be helpful in understanding the biological functions of the individual AP2/ERF transcription factors.

Identification of Putative AP2/ERF Transcription Factors in A. nanus
Genome sequence data were obtained from the A. nanus genome project [32]. The hidden Markov model (HMM) model of the Apetala 2 (AP2) (PF00847) was downloaded from the Pfam database [49], and the AP2/ERF family TFs in A. nanus were identified using HMMER3 (v. 3.0) software [50] with a defined threshold of E < 1e −5 . The NCBI Conserved Domain Search Service (CD Search) [51] was used to confirm manually the predicted AP2/ERF family TFs.

Multiple Alignment and Phylogenetic Analysis
Multiple sequence alignment of the AP2/ERF family TFs was conducted using ClustalW [52]. The phylogenetic trees were constructed with MEGA-X software [53] using the neighbor-joining (NJ) method. The phylogenetic trees were tested using bootstrapping with 2000 replicates. The AP2/ERF family databases of Arabidopsis were downloaded from the plant transcription factor database [54].

Identification of Conserved Domains
Conserved domains were predicted using the tool DIALIGN-PFAM [55]. The deduced amino acid sequences of AP2/ERF genes were analyzed with the MEME (Multiple Em for Motif Elicitation) suite 5.0.1 [36] using the following parameters: optimum width 10~200 amino acids of a motif, and the maximum number of motifs set at 20.

Visualization of AP2/ERF Family Genes on Chromosomes
Diagrams of the chromosome locations of all AP2/ERF genes in A. nanus were visualized using MapChart V2.3 [56].

Prediction of Cis-acting Elements in Promoter Regions of Genes
The software PLACE [57] was used to locate the putative cis-acting elements in the promoter regions of the DREB subfamily genes. Additionally, 2000 bp DNA sequence upstream of the start codon of each DREB gene were used for cis-acting element prediction.

Gene Duplication and Divergence Analysis
Tandem duplication means the generation of tandem gene arrays consisting of identical sequences in close genomic proximity. In the present study, the genes present in tandem duplication were identified using a criterion described previously [58], and tandem-duplicated genes were defined as adjacent homologous genes on a single chromosome, with no more than one intervening gene. The paralogous genes of identified AP2/ERF genes of A. nanus were investigated using MCScanX software [59] and visualized using the program Circos [60]. The divergence between homologous genes and the selective pressure on duplicated genes were estimated by calculating synonymous (Ks) and non-synonymous substitutions (Ka) per site between the duplicated gene pairs using KaKsAnalysis in the PlantTribes collection [61].

Plant Materials and Stress Treatments
The seeds of A. nanus were collected from the arid areas in Wuqia county, Xinjiang Autonomous Region, China. After surface sterilization with 70% (v/v) ethanol, the seeds were planted in 30 cm diameter pots containing a 3:1 (v/v) mixture of perlite and vermiculite. A. nanus seedlings were grown in a plant incubator under 120 µmol m −2 s −1 photosynthetic photon flux density, with a photoperiod of 16 h, at around 25 • C and a relative humidity of 35%. The seedlings were irrigated every four days with a solution of half-strength Hoagland.
Eight weeks after germination, seedlings with similar height were exposed to cold and osmotic stress treatments. The seedlings were randomly split into nine groups. Four osmotic stress-treated groups were watered with 20% PEG-6000 for 3 h, 6 h, 12 h, or 24 h, four cold stress-treated groups were moved to a growth chamber at 4 • C for 3 h, 6 h, 12 h, or 24 h, and the untreated seedlings were used as the control. Leaf samples were collected from seedlings of the control and stress-treated groups, snap-frozen in nitrogen, and stored at −80 • C until use.

qRT-PCR Analysis of the DREB Subfamily Genes in A. nanus
RNA extraction was conducted using the Trizol reagent according to the user's guide (Invitrogen, Carlsbad, CA, USA). Quantitative reverse transcriptase PCR (qRT-PCR) analysis was performed following a method described earlier [62]. Three biological replicates were collected for each group and at least three technical repeats for each biological replicate were assayed. The relative gene expression was calculated using the 2 −∆∆Ct method [63]. The gene expression levels were normalized against an internal reference gene, eukaryotic translation initiation factor 1 (eIF1). Oligo 7 [64] was used to design the primer pairs for qRT-PCR analysis of the DREB subfamily genes, and all primer pairs used in the current study are listed in Table S7.

Statistics
Student t-tests were used to determine significance, and the probability level of p ≤ 0.05 was set for statistical significance. Genes were deemed to be differentially expressed if they demonstrated a fold change of at least 2 and a p-value less than 0.05.

Conclusions
In the current study, 174 AP2/ERF family genes were identified from the genome sequence of A. nanus, and these genes were further divided into five subfamilies, including ERF, DREB, AP2, RAV, and Soloist. Gene expansion occurred in the DREB subfamily and ERF subfamily B1 group, and the major gene amplification mode was segmental duplication, followed by tandem duplication. Multiple sequence alignment validated the conservation of the AP2 domains of AP2/ERF family TFs and revealed the characteristics of the AP2 domains of different subfamilies. Cis-acting element prediction indicated that MYB, ICE1 and WRKY TFs might regulate the expression of DREB genes in A. nanus. Gene profiling identified a collection of cold or osmotic stress-induced genes in the AP2/ERF family. The highly stress-induced DREB genes identified in the present study, including EVM0023336.1, EVM0013392.1, and EVM0026860.1, may contribute to the stress tolerance of A. nanus.
Supplementary Materials: The following are available online at http://www.mdpi.com/2223-7747/9/4/455/s1, Figure S1: Comparison of deduced amino acid sequences of the AP2 domain of the DREB subfamily proteins from A. thaliana (At) and A. nanus (An), Figure S2: Comparison of deduced amino acid sequences of the AP2 domain of the ERF subfamily proteins from A. thaliana (At) and A. nanus (An), Figure S3: Comparison of deduced amino acid sequences of the AP2 domain and B3 domain of the RAV subfamily proteins from A. thaliana (At) and A. nanus (An), Figure S4: Comparison of deduced amino acid sequences of the AP2 domain of the AP2 subfamily proteins from A. thaliana (At) and A. nanus (An), Figure S5: The 20 conserved motifs identified in the AP2/ERF family members, Figure S6. Motif compositions and gene structures of the AP2/ERF family members in A. nanus, Table S1: Physical and chemical characterization of AP2/ERF family from A. nanus , Table S2: The AP2/ERF family genes with tandem duplication relationships in A. nanus, Table S3: AP2/ERF family genes with segmental duplication relationships in A. nanus, Table S4: The Ka/Ks ratios for the paralogous gene pairs in the AP2/ERF gene family in A. nanus, Table S5: The Ka/Ks ratios for the orthologous gene pairs in the AP2/ERF family between A. nanus and G. max. Table S6: Member numbers of each subfamily and group in the AP2/ERF family in A. nanus and other legume species, Table S7: The primers used in qRT-PCR analysis.