Genome-Wide Identiﬁcation and Analysis of the WRKY Gene Family in the Xerophytic Evergreen Ammopiptanthus nanus

: The WRKY family of transcription factors plays important roles in plant growth and responses to biotic and abiotic stresses. Ammopiptanthus nanus , the only evergreen broadleaf shrub endemic to the desert and semi-desert regions of northwestern China, is highly tolerant to various stresses. However, a systematic study of WRKY proteins in A. nanus has not been reported. In the present study, we identiﬁed 63 WRKY genes in the A. nanus genome. Based on the conserved WRKY domains, zinc ﬁnger structures, and phylogenetic relationships in their encoded proteins, we classiﬁed these genes into four groups (group I–IV) and several subgroups (subgroup IIa–IIe). Conserved motif analysis showed that all motifs except those within the WRKY domains had a subfamily-speciﬁc distribution. Expression analysis revealed that the AnWRKY genes had distinct expression patterns, with some being more responsive to herbivory and drought stresses than others. Based on the results of our current study, we speculate that AnWRKY40 and AnWRKY48 are positive regulators of the plant’s response to drought and herbivory stresses, respectively. Our results indicate that AnWRKY genes contribute to the ability of A. nanus plants to withstand harsh, dry conditions.


Introduction
Ammopiptanthus nanus (Popov) Cheng F, a relic plant of the Tertiary period, belongs to the tribe Thermopsideae and the family Fabaceae, and it is a xerophytic leguminous shrub endemic to Central Asian deserts [1]. It is the only broadleaved evergreen that is endemic to the desert areas in northwestern China. It could be a good sand fixing plant with salt alkali, drought, and sand resistances. Moreover, with beautiful tree shape, many butterfly-shaped tubular flowers, golden yellow corolla, and white fluff on the twigs, it could be an ideal greening tree species [2]. In addition, its oil can be used, and some genes in the branches and leaves may produce a medicinal effect [3]. It has the effect of dispelling wind, warming, and dampness, promoting blood circulation and dispersing blood stasis from the perspective of traditional Chinese medical science [3]. In concision, it has ecological, agronomic, and medicinal value. A. nanus occurs only in the southwestern part of the Xinjiang Uygur Autonomous Region and is classified as a II-class nationally protected species in China [3,4]. The genome of A. nanus was sequenced in 2018 and the final size of the assembled this genome was close to the 889 Mb, in which the gene annotation integrality was estimated by BUSCO [1]. Due to the harsh growing environment, A. nanus is subjected to a variety of environmental stresses [4,5], including an arid climate (annual Agronomy 2020, 10 precipitation <200 mm; annual evaporation >2000 mm), extreme temperatures (from −30 • C to 47.6 • C), and serious pests [6]. Drought and pests pose increasingly serious environmental threats to plant growth and revegetation in northwestern China. The main pest of A. nanus is Orgyia ericae Germar, a lepidopteran distributed across northwestern China and parts of Europe [7]. This pest gnaws on leaf buds, leaves, and flower buds, affecting plant growth and fruitage and even causing death.
The WRKY genes are transcriptional regulators that play important roles in plant defense and stress responses [8]. However, the identification of WRKY transcription factors in A. nanus and its roles in stress response are still unclear. The first reported WRKY gene was identified in sweet potato (Ipomoea batatas) in 1994, which could regulate the expressions of sporamin and β-amylase-related genes [9]. WRKY transcription factors were named for the highly conserved 60-amino-acid domain they all contain, which includes a conserved WRKYGQK heptapeptide, followed by a distinctive zinc finger-like motif, C2H2 (C-X 4-5 -C-X 22-23 -H-X-H) or C2HC (C-X 7 -C-X 23 -H-X-C) [8,10]. Based on the number of WRKY domains and the type of zinc finger motifs, WRKY proteins can be divided into four groups (I-IV) [11,12]. Group I proteins usually contain two WRKY domains and C2H2 type zinc finger motifs, whereas groups II-IV proteins contain a single WRKY domain. Group II is characterized by having C2H2 type zinc finger motifs and is further subdivided into five subgroups (IIa, IIb, IIc, IId, and IIe) based on phylogeny [8]. Group III proteins have C2HC zinc fingers and group IV proteins lack a complete zinc finger structure [13].
WRKY transcription factors control a range of physiological and biochemical processes and responds to a variety of abiotic and biotic stresses, such as drought, salinity, herbivory, and pathogens [10,14]. Most WRKY proteins regulate gene expression by specifically binding to a W-box element (TTGAC(C/T)) in the target gene promoter [8]. Previous studies showed that the Siberian crabapple (Malus baccata) MbWRKY5 gene can improve drought and salt tolerance when heterologously expressed in tobacco (Nicotiana tabacum) by increasing proline and ascorbic acid content and enhancing the activities of peroxidase, superoxide dismutase, and catalase [15]. In Arabidopsis thaliana, AtWRKY18 and AtWRKY60, working cooperatively with AtWRKY40, regulate abscisic acid (ABA)-mediated drought and salinity stress responses. These transcription factors bind to the W-box in the promoters of ABI4 and ABI5, ABA signaling pathway genes that affect plant stress resistance [16]. In rice (Oryza sativa), overexpression of OsWRKY11, OsWRKY30, or OsWRKY47 significantly improved drought tolerance by activating the expression of drought-responsive genes [17][18][19]. Many studies show that WRKY transcription factors are also involved in biotic stress responses. For example, overexpression of OsWRKY13 in rice enhanced resistance to blast disease and bacterial blight, increased the accumulation of salicylic acid, and inhibited the expression of jasmonic acid (JA) response genes [20]. Similarly, constitutive overexpression of OsWRKY45 significantly enhanced blast resistance [21]. Wang showed that OsWRKY89 overexpression plants had a reinforced epidermis that enhanced resistance against both rice blast fungus and rice white-back planthopper [22]. In tobacco plants that were attacked by the moth Manduca sexta, NtWRKY3 and NtWRKY6 were strongly upregulated, thereby potentiating and/or maintaining active JA levels during continuous insect attacks [23]. In addition, recent studies showed that VviWRKY40 expression pattern was negatively correlated with that of VviGT14 in grape berries (Vitis vinifera L.) and reported that ABA could downregulate the transcript level of VviWRKY40 in order to control monoterpenoid glycosylation [24].
WRKY genes have been identified in many species, resulting in large gene families that include up to 72 members in Arabidopsis [8], 96 in rice [25], 55 in cucumber (Cucumis sativus) [26], 59 in grapevine [13], and 127 in apple (Malus domestica Borkh.) [27], and the mechanisms by which WRKY proteins regulate some biological process have been explored. However, no studies have examined the WRKY gene family in A. nanus. It is regarded as a plant that can withstand harsh conditions, it is of both scientific and practical interest to comprehensively identify and analyze the WRKY gene family in this plant and to explore the molecular mechanisms underlying its stress tolerance.

Database Search and Identification of WRKY Family Members in A. nanus
The sequences of Arabidopsis WRKY proteins were downloaded from the Arabidopsis database (TAIR, http://www.arabidopsis.org/). The A. nanus genome was downloaded from the GigaScience database (GigaDB, http://gigadb.org/; accession number 100466). First, a local protein database was established containing the proteome of A. nanus using BioEdit 7.2.0 software (Ibis Therapeutics, Carlsbad, USA) [28]. To identify WRKY proteins in A. nanus, a local BlastP search was performed using a hidden Markov model (HMM) profile of the WRKY domain (PF03106) as query sequences. Default parameters were used and the expectation (E) value was less than 0.001. The search results were used to identify candidate genes, and then conserved domain analysis was conducted with the SMART program [29]. After removing low-reliability and redundant sequences, the complete set of A. nanus WRKY family proteins was obtained. To further analyze the zinc finger structure of A. nanus WRKY proteins, candidate protein sequences were aligned in the DNAman program (Lynnon BioSoft, San Ramon, CA, USA). In addition, the theoretical isoelectric point (pI) and molecular weight (MW) were analyzed using the ProtParam program [30].

Exon-Intron Structure and Conserved Motif Analysis
The DNA sequence and coding sequence (CDS) of A. nanus WRKY genes were downloaded from the GigaScience database (accession number 100466), and the exon-intron distribution patterns were analyzed using online GSDS software [33]. The conserved motifs were analyzed using MEME [34], a software tool for predicting conserved sequences of genes, with the following parameters: amino acid width >6 and <200, maximum number of motifs: 15.

Plant and Insect Materials, Growth Conditions, and Stress Treatments
A. nanus seeds were collected from Wuqia County in Xinjiang Province in northwestern China. The seeds were surface sterilized for 7 min in 2% (w/v) KMnO 4 , soaked in water for 12 h, and then sown in sterilized soil (sand: vermiculite: perlite = 4:1:1). Seedlings were grown in a growth chamber under a 14 h light (29 • C)/10 h dark (25 • C) photoperiod and relative humidity of 60-70% [35]. Wild type Arabidopsis (wt, Columbia ecotype) and atwrky48 lines (SALK_066438, kindly provided by Prof. Diqiu Yu; Chinese Academy of Sciences, China) were incubated in the chamber under a 16 h light (22 • C)/8 h dark (20 • C) photoperiod and relative humidity of 70%. The plants used were about five weeks old.
Orgyia ericae cocoons were collected from Alxa Left Banner, China. Plutella xylostella L. eggs were purchased from Jiyuan Baiyun Industry Co., Ltd. (Jiyuan, China). Cocoons or eggs were hatched at 28 • C or conserved at 4 • C. For feeding assays, the insect-inoculated plants were placed in a clear plastic cups covered with plastic net chamber to facilitate the exchange of gases. The O. ericae in the second-or third-instar stage was placed on A. nanus leaves of four-week-old plants without blossoming, and one or two larvae were reared on each of 10 plants per biological replicates (n = 3). One day later, RNA was extracted from the attacked leaves for qRT-PCR analysis. The newly hatched Plutella xylostella L. were placed on rosette leaves of five-week-old Arabidopsis, and three larvae were reared on each of 60 plants per genotype. The larval weights and lengths were measured, and the specimens were photographed after six days.
For drought treatment, RNA was extracted for qRT-PCR analysis from the leaves of five-week-old A. nanus plants that had not been watered for six days.

Expression Analysis
Total RNA was isolated using an RNAprep Pure Plant Kit (TIANGEN, Beijing, China) from A. nanus tender roots, non-lignified stems, young leaves, and cotyledons. RNA was quantified using NanoDrop 2000, and electrophoresis analysis revealed clear bands, indicating intact RNA products. Then 500 ng RNA was reverse transcribed using a PrimeScript TM RT Reagent Kit (Perfect Real Time) (Takara, Dalian, China) for semi-quantitative PCR (semi-qRT-PCR) and quantitative real-time PCR (qRT-PCR) analysis. Specific primers that avoid the conserved regions of WRKY genes were designed using Primer 5.0 software (Premier Biosoft, Palo Alto, CA, USA). For semi-qRT-PCR, a 20 µL reaction system was used containing 10 µL 2 × Premix Taq TM (TaKaRa, Dalian, China), forward and reverse primers diluted to 0.4 mM, and 0.25 µL template mentioned above. The reaction conditions were 29 cycles of 94 • C for 30 s, 60 • C for 30 s, and 72 • C for 30 s. AnACTIN2 was used as a sufficiently stable internal control gene for A. nanus. PCR products were run on 2% agarose gels. For qRT-PCR analysis, the reaction system contained 10 µL of 2 × SYBR Premix Ex Taq II (TaKaRa, Dalian, China) as the enzyme and the other reagents were the same as those used in the semi-qRT-PCR. The reaction conditions were 95 • C for 30 s, followed by 40 cycles of 95 • C for 5 s, and 60 • C for 30 s, and fluorescence was collected after each cycle. The relative expression of genes was calculated using the 2 -∆∆Ct method. Three technical and three biological replicates were performed for each gene.

Identification and Annotation of WRKY Genes in A. nanus
To identify A. nanus WRKY family genes, the HMM (PF03106) profile was applied as a query to search against the local protein database. After further analysis using the SMART program, missing and redundant sequences were removed and 63 WRKY genes were confirmed in the A. nanus genome. In order to better reflect the orthologous relationship between AnWRKY and AtWRKY proteins, we named each AnWRKY gene based on their phylogenies corresponding to individual AtWRKY ( Figure S1).
The characteristics of the AnWRKY genes and their predicted protein products, including subgroup, location, length, PI, MW, and intron number are annotated in Table 2. The PIs ranged from 4.79 (AnWRKY15) to 9.91 (AnWRKY11), and the MWs ranged from 13.3 kDa (AnWRKY45) to 83.5 kDa (AnWRKY25).  We compared the number of WRKY genes in each subfamily in A. nanus to that in Arabidopsis, rice, grapevine, and cucumber ( Table 3). The WRKY gene family in A. nanus has roughly the same characteristics as in other species, but also shows obvious differences. In A. nanus, there are only three genes in subgroup IId, which consequently accounts for a significantly smaller proportion (4.76%) of the total genes than it does in other species.

Phylogenetic Analysis
To examine the phylogenetic relationships among AnWRKY proteins, we constructed a phylogenetic tree based on multiple sequence alignment of 63 AnWRKY and 20 AtWRKY proteins (as references) using MEGA 7.0 software. Consistent with previously published reports [9], the 63 AnWRKY proteins were subdivided into four subgroups, designated groups I to IV (Figure 1; group IV is not shown). Group I was divided into two independent subgroups, group I-N and I-C, based on differences between the N-and C-terminals of the WRKY domains.
To gain insight into the WRKY domain and zinc finger structure, the multiple amino acid sequence of each subgroup member was aligned using DNAman software. As shown in Figure S2, twelve genes with two highly conserved WRKY domains were assigned to group I, which also had C-X 4 -C-X 22 -H-X-H and C-X 4 -C-X 23 -H-X-H zinc fingers at the N-and C-terminus, respectively. Group II was subdivided into IIa (5 members), IIb (10 members), IIc (15 members), IId (3 members), and IIe (8 members), which all had a C-X 5 -C-X 23 -H-X-H zinc finger motif, except for group IIc, which had a C-X 4 -C-X 23 -H-X-H zinc finger. Group III had eight genes, which had C2HC (C-X 7 -C-X 23 -H-X-C) zinc fingers instead of the C2H2 type. Two genes, AnWRKY11 and AnWRKY41, had a WRKY domain but only incomplete zinc finger structures and therefore were assigned to group IV. Although AnWRKY10 clustered with group I-C in the phylogenetic tree, it had only one WRKY motif and should be classified as subgroup IIc based on the presence of this conserved motif. AnWRKY49 clustered with group III genes in the phylogenetic tree, but it was assigned to subgroup IIc because of its C2H2 zinc finger motif. Although most AnWRKYs contain the highly conserved sequence WRKYGQK, this motif was replaced by WRKYGEK in AnWRKY40 and by WRKYGKK in four other genes (AnWRKY50, 51, 56, 59). Notably, the N-terminal WRKY domain of AnWRKY34 was replaced by WKKYGEK, which was a new consensus WRKY motif ( Figure S2). Overall, AnWRKY proteins fell into four groups and had highly conserved C2H2 and C2HC zinc finger motifs, similar to what has been reported for other plants. To gain insight into the WRKY domain and zinc finger structure, the multiple amino acid sequence of each subgroup member was aligned using DNAman software. As shown in Figure S2, twelve genes with two highly conserved WRKY domains were assigned to group I, which also had C-X4-C-X22-H-X-H and C-X4-C-X23-H-X-H zinc fingers at the N-and C-terminus, respectively. Group II was subdivided into IIa (5 members), IIb (10 members), IIc (15 members), IId (3 members), and IIe (8 members), which all had a C-X5-C-X23-H-X-H zinc finger motif, except for group IIc, which had a C-X4-C-X23-H-X-H zinc finger. Group III had eight genes, which had C2HC (C-X7-C-X23-H-X-C) zinc fingers instead of the C2H2 type. Two genes, AnWRKY11 and AnWRKY41, had a WRKY domain but only incomplete zinc finger structures and therefore were assigned to group IV. Although AnWRKY10 clustered with group I-C in the phylogenetic tree, it had only one WRKY motif and should be classified as subgroup IIc based on the presence of this conserved motif. AnWRKY49 clustered with group III genes in the phylogenetic tree, but it was assigned to subgroup IIc because

Analysis of AnWRKY Gene Structure
We explored the structure of the A. nanus WRKY genes by analyzing the positions of conserved motifs, exons, and introns ( Figure 2). Motif 1 indicates a WRKY motif and motif 2 represents a zinc finger motif, which together comprise an integrated WRKY/zinc finger domain. These motifs were present in all of the AnWRKY genes except AnWRKY11 and AnWRKY41, which had WRKYGQK motifs (motif 12) but incomplete zinc finger structures. Motif 3 and motif 7, which were present only in group I genes, are a WRKY motif and a zinc finger motif, respectively, that together form a second, N-terminal WRKY domain in these genes. The distribution of other conserved motifs among the AnWRKY genes exhibited familial specificity. Motif 4 and motif 9 were present only in subgroups IIa and IIb. The highly conserved motif 8 was present only in subgroup IIb, immediately downstream of motif 2. Motif 10 was present only in group I, whereas motif 13 was present in all group III genes except AnWRKY41. Motif 12 was present in most genes in groups IId, IIe, III, and IV.
N-terminal WRKY domain in these genes. The distribution of other conserved motifs among the AnWRKY genes exhibited familial specificity. Motif 4 and motif 9 were present only in subgroups IIa and IIb. The highly conserved motif 8 was present only in subgroup IIb, immediately downstream of motif 2. Motif 10 was present only in group I, whereas motif 13 was present in all group III genes except AnWRKY41. Motif 12 was present in most genes in groups IId, IIe, III, and IV.  Based on our analysis of exon-intron structure, AnWRKY genes that clustered together in the evolutionary tree often had similar exon-intron structures. Genes of subgroup IIb have complex exon-intron structures, whereas those of subgroups IId and IIe are simple. Among the 63 AnWRKY genes we identified, three genes had one intron, eight genes had five or six introns, and the remainder had two to four introns.

Expression of AnWRKY Genes in Different Tissues
To investigate the expression pattern of AnWRKY genes, we used a semi-qRT-PCR method to determine the transcript levels of 63 AnWRKY genes in roots, stems, leaves, and cotyledons. AnACTIN2 was used as an internal control. The specific primer sequences, which did not overlap with any of the conserved WRKY domains, are shown in Table S1.
Most of the AnWRKY genes were more highly expressed in roots than in other tissues. Ten genes, AnWRKY9, 16,22,27,29,35,36,42,48, and 61 were expressed specifically in the roots. Among these genes, five belonged to subgroup IIb and 4 to subgroup IIe. Five genes, AnWRKY37, 38, 40, 56, and 63, showed specific and low expression in leaves. None of the genes were specifically expressed in stems and cotyledons ( Figure S3).
The patterns of gene expression showed subfamily specificity. All of the genes in group I and subgroup IId, except AnWRKY33, were expressed at high or moderate levels in all four tissues. Genes in subgroup IIa and group III were generally expressed at low levels in the four tissues, except for AnWRKY5 and AnWRKY53, which were highly expressed in roots and leaves, respectively. Most genes in subgroups IIb and IIe showed high expression in roots. Thirteen genes, AnWRKY10, 14,18,28,33,41,45,46,47,49,50,55, and 62, showed very low expression in all four tissues ( Figure S3).
Based on their subgroup designation and expression level in the different tissues, 15 AnWRKY genes were selected for further study of their expression patterns with qRT-PCR. These results showed the same general trends as the semi-qRT-PCR results. Transcript levels for AnWRKY61 were 180 times higher in roots than in leaves, and transcript levels for AnWRKY9, 29, and 35 were about 30-40 times higher in roots than in leaves. The expression of AnWRKY5, 7, 21, 23, 26, 27, and 52 was also highest in roots. Expression of AnWRKY13 was highest in the stems, about 14 times higher than in leaves and cotyledons, whereas that of AnWRKY53 was highest in leaves, about 10 times higher than in the other tissues ( Figure 3).

The Effect of Herbivory and Drought on Expression of AnWRKY Genes
To see how the AnWRKY genes respond to stress, we performed a preliminary semi-qRT-PCR analysis of AnWRKY expression in the roots and leaves of plants subjected to herbivory or drought (data not shown). Based on the results, we selected the top nine genes with the most significant expression change for further analysis : AnWRKY9, 18, 27, 29, 43, 45, 48, 62, and 63 for herbivory and  AnWRKY8, 26, 27, 29, 34, 36, 40, 59, and 61 for drought. The effect of herbivory on AnWRKY expression was strongest in leaves of the treated plants, where the expression of all nine genes was upregulated by about 2-to 14-fold (Figure 4). In leaves, Figure 3. Expression analysis of 15 AnWRKY genes in roots, stems, leaves, and cotyledons. AnACTIN2 was used as an internal control. The mean and standard error (SE) were derived from three biological and three technical replicates. The different groups and subgroups are indicated above the graphs. Different letters indicate significant differences at p < 0.05, Student's t-test.

The Effect of Herbivory and Drought on Expression of AnWRKY Genes
To see how the AnWRKY genes respond to stress, we performed a preliminary semi-qRT-PCR analysis of AnWRKY expression in the roots and leaves of plants subjected to herbivory or drought (data not shown). Based on the results, we selected the top nine genes with the most significant expression change for further analysis: AnWRKY9, 18,27,29,43,45,48,62, and 63 for herbivory and AnWRKY8, 26,27,29,34,36,40,59, and 61 for drought.
The effect of herbivory on AnWRKY expression was strongest in leaves of the treated plants, where the expression of all nine genes was upregulated by about 2-to 14-fold (Figure 4). In leaves, the expression of AnWRKY18 was upregulated by about 5.2-fold (statistical significance at *** p < 0.001). The expression of AnWRKY45, 48 and 63 were respectively increased by about 14.1-, 17.6-, and 4.2-fold (** p < 0.01), and the significance of AnWRKY48 not reached to *** p < 0.001, which may be due to more error between biological replicates. The expression of other genes was also upregulated (* p < 0.05) in leaves. In roots, the expression of AnWRKY9, 43, and 48 genes were significantly upregulated. Three genes (AnWRKY18, 45, and 62) were downregulated in roots by about 1.7-to 4-fold, but AnWRKY45 has no statistical significance (Figure 4). Interestingly, the expression levels of these nine genes was low relative to other AnWRKY genes in untreated leaves ( Figure S3). fold, but AnWRKY45 has no statistical significance ( Figure 4). Interestingly, the expression levels of these nine genes was low relative to other AnWRKY genes in untreated leaves ( Figure S3).  18,27,29,43,45,48,62, and 63 in response to herbivory in roots and leaves. Black columns represent data for the one-day herbivory treatment, and gray columns represent the control data. HW: herbivory wounding. AnACTIN2 was used as an internal control. The mean and SE were derived from three biological and three technical replicates. *** p < 0.001, ** p < 0.01, * p < 0.05, Student's t-test.
The effect of drought on AnWRKY expression showed four upregulated genes (AnWRKY8, 26,40,59) and five markedly downregulated genes (AnWRKY27, 29,34,36,61) in both roots and leaves of the treated plants. It should be noted that the expression level of AnWRKY27 and AnWRKY29 was downregulated by about six fold in roots (*** p < 0.001), but there were no obvious changes in leaves. In addition, both AnWRKY34 and AnWRKY59 showed great changes in leaves, but they have no statistical significance, which may be caused by more error between biological replicates. The expression of AnWRKY40 was remarkably upregulated in the roots, but it hardly changed in leaves. AnWRKY8 was strikingly enhanced in leaves, but it did not change significantly in roots ( Figure 5).  18,27,29,43,45,48,62, and 63 in response to herbivory in roots and leaves. Black columns represent data for the one-day herbivory treatment, and gray columns represent the control data. HW: herbivory wounding. AnACTIN2 was used as an internal control. The mean and SE were derived from three biological and three technical replicates. *** p < 0.001, ** p < 0.01, * p < 0.05, Student's t-test.
The effect of drought on AnWRKY expression showed four upregulated genes (AnWRKY8, 26, 40, 59) and five markedly downregulated genes (AnWRKY27, 29, 34, 36, 61) in both roots and leaves of the treated plants. It should be noted that the expression level of AnWRKY27 and AnWRKY29 was downregulated by about six fold in roots (*** p < 0.001), but there were no obvious changes in leaves. In addition, both AnWRKY34 and AnWRKY59 showed great changes in leaves, but they have no statistical significance, which may be caused by more error between biological replicates. The expression of AnWRKY40 was remarkably upregulated in the roots, but it hardly changed in leaves. AnWRKY8 was strikingly enhanced in leaves, but it did not change significantly in roots ( Figure 5). In general, different from herbivory treatment, AnWRKY genes showed an inconsistent variation trend after drought treatment.
Agronomy 2020, 10, x FOR PEER REVIEW 12 of 17 Figure 5. qRT-PCR assays of the expression of AnWRKY8, 26,27,29,34,36,40,59, and 61 in response to drought in roots and leaves. Black columns show data for the six-day drought treatment, and the gray columns show data for the controls. AnACTIN2 was used as an internal control. The mean and SE were derived from three biological and three technical replicates. *** p < 0.001, ** p < 0.01, * p < 0.05, Student's t-test.
In general, the expression of AnWRKY genes showed regular change in the treated plants subjected to herbivory or drought, especially herbivory treatment induced upregulated expression in leaves. These results indicated that AnWRKY genes was involved in the expression regulation under biotic and abiotic stresses.

Insect Feeding Trials in atwrky48
Based on the strong upregulation of AnWRKY48 expression after the herbivory treatment, we hypothesized that this gene might be involved in plant defense responses to insect feeding. AnWRKY48 is homologous to AtWRKY48 in Arabidopsis, but neither gene has been examined for its role in insect resistance. To verify this hypothesis, wild type (WT) and atwrky48 Arabidopsis plants were performed feeding assays simultaneously. As shown in Figure 6a,b, the average weight and length of larvae raised on atwrky48 plants (42.4 ± 0.2 mg, 7.5 ± 0.2 mm) were significantly higher than those of larvae raised on WT plants (37.9 ± 0.2 mg, 7.2 ± 0.2 mm), (* p < 0.05). At the same time, the Plutella xylostella L. larvae reared on atwrky48 were significantly lower than that of larvae reared on WT plants (Figure 6c), indicating that the mutant is less resistant to herbivore attack.  26,27,29,34,36,40,59, and 61 in response to drought in roots and leaves. Black columns show data for the six-day drought treatment, and the gray columns show data for the controls. AnACTIN2 was used as an internal control. The mean and SE were derived from three biological and three technical replicates. *** p < 0.001, ** p < 0.01, * p < 0.05, Student's t-test.
In general, the expression of AnWRKY genes showed regular change in the treated plants subjected to herbivory or drought, especially herbivory treatment induced upregulated expression in leaves. These results indicated that AnWRKY genes was involved in the expression regulation under biotic and abiotic stresses.

Insect Feeding Trials in atwrky48
Based on the strong upregulation of AnWRKY48 expression after the herbivory treatment, we hypothesized that this gene might be involved in plant defense responses to insect feeding. AnWRKY48 is homologous to AtWRKY48 in Arabidopsis, but neither gene has been examined for its role in insect resistance. To verify this hypothesis, wild type (WT) and atwrky48 Arabidopsis plants were performed feeding assays simultaneously. As shown in Figure 6a,b, the average weight and length of larvae raised on atwrky48 plants (42.4 ± 0.2 mg, 7.5 ± 0.2 mm) were significantly higher than those of larvae raised on WT plants (37.9 ± 0.2 mg, 7.2 ± 0.2 mm), (* p < 0.05). At the same time, the Plutella xylostella L. larvae reared on atwrky48 were significantly lower than that of larvae reared on WT plants (Figure 6c), indicating that the mutant is less resistant to herbivore attack.

Discussion
Transcription factors in the WRKY superfamily play important roles in physiological and developmental processes of plants. In this study, we identified 63 novel AnWRKY genes in A. nanus. Only three of these genes (AnWRKY7, 17, and 39) belonged to subgroup IId. As a previous study showed that WRKY genes of subgroup IId could directly interact with calcium-modulated protein [36], we speculate that these three genes may play an important role in the regulation of calcium ions in A. nanus. Furthermore, we analyzed the conserved motifs of AnWRKY genes and established that the distribution of the conserved motifs showed subgroup specificity; most conserved motifs (e.g., motifs 2, 5, 7, 8, 11, and 12) were located near the WRKY region, suggesting that these motifs may be critical for the function of WRKY proteins.
WRKYGQK, a highly conserved domain, was replaced with WRKYGEK in AnWRKY45 and with WRKYGKK in AnWRKY50, 51, 56, and 59. These genes all belong to subgroup IIc. Similar phenomena were reported in Arabidopsis and cucumber [8,26]. As a previous study demonstrated that the WRKYGKK domain failed to bind to the W-box [37], it needs further experimental verification about whether AnWRKY45, 50, 51, 56, and 59 can be combined with the W-box The molecular mechanisms by which these family members function in plants remain to be elucidated. In addition, AnWRKY34 contained motif 3 and 7, which formed the WRKY domain and C2H2 type zinc finger structure of the N-terminal. Although the WRKYGQK domain in the N-terminal was replaced by a new consensus WRKY motif, WKKYGEK [38], AnWRKY34 still belongs to group I. In chrysanthemum (Chrysanthemum morifolium), CmWRKY17 contains the WKKYGEK domain, but belongs to group II. This protein binds strongly to the W-box and weakly to the mutant W-box [39]. Therefore, whether AnWRKY34 will have the same situation as CmWRKY17 also needs further experimental verification.
To explore the expression patterns of the AnWRKY genes, we analyzed the expression level of 63 AnWRKY genes in roots, stems, leaves, and cotyledons by semi-qRT-PCR, and selected 15 genes with representative expression patterns for qRT-PCR detection. Ten genes (AnWRKY4, 7, 11, 17, 23, 26, 32, 39, 52, and 58) were expressed in four different tissues (roots, stems, leaves, and cotyledons). Highly expressed WRKY genes often play key roles in plant growth and development through activating the transcription of downstream target genes [40]. Therefore, we hypothesize that these 10 WRKY genes play fundamental roles in different cell types in A. nanus [41]. By contrast, some AnWRKY genes were only expressed in individual tissues. AnWRKY9, 29,35,36, and 61 were specifically expressed in roots, and all belonged to the IIb or IIe subgroups. Similar to the situation in Populus [40], most PtrWRKY genes were highly expressed in roots, suggesting that these genes play key roles in the metabolism and development of A. nanus roots. Some of the AnWRKY genes were not expressed in the four tissues examined and might be expressed in other tissues that were not examined or only expressed under certain conditions. For example, the expression levels of

Discussion
Transcription factors in the WRKY superfamily play important roles in physiological and developmental processes of plants. In this study, we identified 63 novel AnWRKY genes in A. nanus. Only three of these genes (AnWRKY7, 17, and 39) belonged to subgroup IId. As a previous study showed that WRKY genes of subgroup IId could directly interact with calcium-modulated protein [36], we speculate that these three genes may play an important role in the regulation of calcium ions in A. nanus. Furthermore, we analyzed the conserved motifs of AnWRKY genes and established that the distribution of the conserved motifs showed subgroup specificity; most conserved motifs (e.g., motifs 2, 5, 7, 8, 11, and 12) were located near the WRKY region, suggesting that these motifs may be critical for the function of WRKY proteins.
WRKYGQK, a highly conserved domain, was replaced with WRKYGEK in AnWRKY45 and with WRKYGKK in AnWRKY50, 51, 56, and 59. These genes all belong to subgroup IIc. Similar phenomena were reported in Arabidopsis and cucumber [8,26]. As a previous study demonstrated that the WRKYGKK domain failed to bind to the W-box [37], it needs further experimental verification about whether AnWRKY45, 50, 51, 56, and 59 can be combined with the W-box The molecular mechanisms by which these family members function in plants remain to be elucidated. In addition, AnWRKY34 contained motif 3 and 7, which formed the WRKY domain and C2H2 type zinc finger structure of the N-terminal. Although the WRKYGQK domain in the N-terminal was replaced by a new consensus WRKY motif, WKKYGEK [38], AnWRKY34 still belongs to group I. In chrysanthemum (Chrysanthemum morifolium), CmWRKY17 contains the WKKYGEK domain, but belongs to group II. This protein binds strongly to the W-box and weakly to the mutant W-box [39]. Therefore, whether AnWRKY34 will have the same situation as CmWRKY17 also needs further experimental verification.
To explore the expression patterns of the AnWRKY genes, we analyzed the expression level of 63 AnWRKY genes in roots, stems, leaves, and cotyledons by semi-qRT-PCR, and selected 15 genes with representative expression patterns for qRT-PCR detection. Ten genes (AnWRKY4, 7,11,17,23,26,32,39,52, and 58) were expressed in four different tissues (roots, stems, leaves, and cotyledons). Highly expressed WRKY genes often play key roles in plant growth and development through activating the transcription of downstream target genes [40]. Therefore, we hypothesize that these 10 WRKY genes play fundamental roles in different cell types in A. nanus [41]. By contrast, some AnWRKY genes were only expressed in individual tissues. AnWRKY9, 29, 35, 36, and 61 were specifically expressed in roots, and all belonged to the IIb or IIe subgroups. Similar to the situation in Populus [40], most PtrWRKY genes were highly expressed in roots, suggesting that these genes play key roles in the metabolism and development of A. nanus roots. Some of the AnWRKY genes were not expressed in the four tissues examined and might be expressed in other tissues that were not examined or only expressed under certain conditions. For example, the expression levels of AnWRKY40 in roots and AnWRKY48 in leaves were relatively low, but the levels were significantly increased when A. nanus was subjected to drought or herbivory stress, respectively.
WRKY transcription factors not only play important roles in plant development, but also in the response to abiotic and biotic stresses. In this study, we demonstrated that the expression of AnWRKY genes in leaves changed greatly after herbivory treatment and tended to be upregulated. In other species [40,42,43], the studies on WRKY transcription factors were mainly focused on abiotic or pathogenic treatment, but few studies on herbivory. Only individual WRKY genes have been reported to be related to insect resistance. For instance, Yan et al. [44] revealed the inhibition roles of JAV1-JAZ8-WRKY51 (JJW) complex on expression of JA biosynthesis genes in healthy plants, and insect herbivory rapidly disintegrated JJW complex via inducing Ca 2+ /calmodulindependent phosphorylation of JAV1, derepressing JA biosynthesis genes and giving rise to rapid burst of JA for plant defense. In the current study, nine genes with obvious changes were selected for further analysis by qRT-PCR, of which AnWRKY48 was the most significantly upregulated (about 17.6 times). Insect feeding trials were performed to explore whether AnWRKY48 affects the plant's susceptibility to insect attack. The average weight and length of larvae that fed on atwrky48 seedlings were significantly higher than those raised on WT plants, indicating that these mutants are less resistant to herbivore attack than WT plants and that AtWRKY48 is positively associated with resistance to herbivores ( Figure 6). We speculate that AnWRKY48, a homolog of AtWRKY48, may have similar functions. In order to further verify this function, the exogenous AnWRKY48 may be overexpressed in different model plants for testing the insect resistance ability. In the other hand, establishing a tissue culture system of A. nanus and transient expression technology could be used to analyze the gene functions.
Furthermore, nine genes with marked changes after drought treatment were selected for further analysis by qRT-PCR ( Figure 5). These upregulated or downregulated genes might be positively or negatively correlated with drought resistance, respectively. Consistent with previous research, members in the same subgroup also can have both positive and negative roles in plant defense [45]. This suggested that transcription factor families respond differently to the same stress, even if they belong to the same subgroup. Indeed, AnWRKY40 was the most significantly upregulated in roots, about 5.4 times compared to levels under control conditions. AtWRKY40 (a homolog of AnWRKY40) promotes plant drought stress responses by interfering with the ROS-scavenging pathway and osmolyte accumulation process. The atwrky40 mutant was less tolerant to drought than the wild type, whereas a line that overexpressed AtWRKY40 was more drought tolerant [46]. Therefore, we speculate that AnWRKY40 may be involved in regulating drought tolerance in A. nanus.

Conclusions
We identified 63 AnWRKY genes from the A. nanus genome, and analyzed their gene structures, conserved motifs, phylogenetic relationships, and expression patterns. The AnWRKYs was divided into I-IV groups, and their conserved motifs also exhibited familial specificity. Furthermore, a majority of the AnWRKYs are expressed in more than one tissue, while some genes are expressed only in a specific tissue. Indeed, the AnWRKY genes could make corresponding response according to different stresses. Combined with the results of the study, we found that AnWRKY40 and AnWRKY48 have significantly positive regulatory roles in the plant's tolerance to drought stress and herbivores, respectively. This work provides insight into the WRKY gene family in A. nanus and lays the foundation for deciphering the molecular mechanisms underlying the ability of this relic plant to withstand a range of stresses.