Characterization and Comparative Analysis of RWP-RK Proteins from Arachis duranensis, Arachis ipaensis, and Arachis hypogaea

RWP-RK proteins are important factors involved in nitrate response and gametophyte development in plants, and the functions of RWP-RK proteins have been analyzed in many species. However, the characterization of peanut RWP-RK proteins is limited. In this study, we identified 16, 19, and 32 RWP-RK members from Arachis duranensis, Arachis ipaensis, and Arachis hypogaea, respectively, and investigated their evolution relationships. The RWP-RK proteins were classified into two groups, RWP-RK domain proteins and NODULE-INCEPTION-like proteins. Chromosomal distributions, gene structures, and conserved motifs of RWP-RK genes were compared among wild and cultivated peanuts. In addition, we identified 12 orthologous gene pairs from the two wild peanut species, 13 from A. duranensis and A. hypogaea, and 13 from A. ipaensis and A. hypogaea. One, one, and seventeen duplicated gene pairs were identified within the A. duranensis, A. ipaensis, and A. hypogaea genomes, respectively. Moreover, different numbers of cis-acting elements in the RWP-RK promoters were found in wild and cultivated species (87 in A. duranensis, 89 in A. ipaensis, and 92 in A. hypogaea), and as a result, many RWP-RK genes showed distinct expression patterns in different tissues. Our study will provide useful information for further functional and evolutionary analysis of the RWP-RK genes.


Introduction
Transcription factors are essential components of plant signal transduction pathways. The Arabidopsis genome contains approximately 1500 transcription factors distributed among various signaling pathways [1]. To transfer a signal to its target genes, a transcription factor binds to cis-acting elements in the genes' promoter regions to activate or suppress their expression. Some transcription factors contain similar motifs, have similar functions, and are classified in the same gene family, while others with different conserved domains are placed in different gene families. Many plant transcription factor gene families have been identified and characterized, such as the SQUAMOSA promoter binding protein (SBP) box family, the heat shock transcription factor (Hsf) family, and the RWPXRK motif (RWP-RK) gene family [2][3][4][5][6][7][8].
The RWP-RK gene family is characterized by a conserved motif, a 60 amino acid sequence that is thought to be involved in DNA binding, and has been studied in many species [4,9,10]. For example, there are 14, 8, and 15 RWP-RK members in Arabidopsis, Medicago truncatula, and Oriza sativa, respectively. Moreover, the RWP-RK proteins are classified into two groups, the RWP-RK domain proteins (RKD) and the NODULE-INCEPTION-like proteins (NLP), on the basis of their gene structures. In addition to the RWP-RK motif, all NLPs contain a PB1 (Phox and Bem1) motif, and some members contain a GAF (a new class of cyclic GMP receptor) or GAF-like domain [4,11,12]. The PB1 domain is thought to be involved in protein-protein interactions, and the GAF domain participates in signal transduction or dimerization in plants [4,13,14].
The functions of several RWP-RK proteins have been investigated in the past decades [4]. For example, NLP genes have been shown to function in the plant nitrate response. In Arabidopsis, AtNLP6 and its closest homolog, AtNLP7, are thought to regulate the N starvation response by binding to N metabolism genes and downstream regulatory genes [4,15,16]. In maize, ZmNLP3.1, ZmNLP6, and ZmNLP8 rescue Arabidopsis nlp7 phenotypes and are involved in primary nitrate response by regulating nitrate assimilation under low nitrate conditions [16]. Another NLP gene, NIN, is a key factor in legume nodule formation [4]. NIN was the first transcription factor identified as functioning in Lotus japonicus nodulation, and mutation of L. japonicus NIN blocks infection and prevents nodule organogenesis [17]. In M. truncatula, MtNIN plays a central role in the temporal and spatial regulation of the nodule development process by competitively inhibiting ERF required for nodulation (ERN1) to suppress the expression of Early Nodulin 11 (ENOD11) and increase the expression of Cytokinin Response 1 (CRE1) [4,16]. Moreover, soybean GmNINa, the ortholog of L. japonicus NIN, regulates nodulation through interactions with Nodule Number Control 1 (NNC1) and the activation of miR172c [18].
Many RKD proteins have been shown to function in the gametophyte development [4,19]. The Arabidopsis genome contains five RKD members, designated AtRKD1 to AtRKD5. The expression of AtRKD1, AtRKD2, AtRKD3, and AtRKD4 is high in reproductive organs, whereas the expression of AtRKD5 is ubiquitous [20,21]. AtRKD1 and AtRKD2 have overlapping functions during plant development, and constitutive expression of AtRKD1 or AtRKD2 in sporophyte cells alters gene transcription patterns, causing them to resemble those of egg cells [20]. The rkd4 mutant shows impaired zygotic cell elongation [22], and RKD3 and RKD4 are involved in ensuring polarity and specifying cell identity in the female gametophyte [19].
Peanut is an important legume crop that provides oil and food worldwide. The cultivated peanut, A. hypogaea (AABB genome, 2n = 4x = 40), is thought to be descended from the hybridization and polyploidization of two wild diploids: A. duranensis (AA genome, 2n = 2x = 20) and A. ipaensis (BB genome, 2n = 2x = 20) [23][24][25]. Genome size and gene content have changed during the evolution of the cultivated peanut, and polyploidization has produced many duplicated gene pairs [25][26][27][28][29]. Some duplicated genes have evolved new functions, and some have become pseudogenes [30,31]. The RWP-RK proteins have also changed during evolution, and research on their function in nitrate response and gametophyte development has been carried out in many plant species [4]. However, information on peanut RWP-RK proteins is limited. In this study, we identified and characterized RWP-RK members from wild and cultivated peanuts and investigated the evolutionary relationships among them. Our research will provide essential information for further functional characterization and peanut improvement.

Identification of RWP-RK Genes from Wild and
Cultivated Peanut Genomes. The amino acid sequences of the conserved RWP-RK domain (PF02042) and Arabidopsis RWP-RK proteins downloaded from TAIR (https://www .arabidopsis.org/) were used as BLAST queries against the peanut genome database (https://www.peanutbase.org/) to search for wild and cultivated peanut RWP-RK genes. We also searched for RWP-RK candidate genes using gene annotations in the peanut genome database. The conserved RWP-RK domains in these candidate RWP-RK genes were confirmed by Pfam (http://pfam.xfam.org/search) and NCBI (National Center for Biotechnology Information). The genomic length, CDS length, and number of amino acids for each gene were obtained from the peanut genome database. The characteristics of the peanut RWP-RK genes, including their molecular weights and theoretical isoelectric points, were predicted using ProtParam (https://web.expasy.org/protparam/), and their GC content was determined using DNASTAR (DNASTAR, Madison, WI, USA) [32].

Phylogenetic Relationship Analysis.
The amino acid sequences of the RWP-RK proteins from Arabidopsis, A. duranensis, A. ipaensis, A. hypogaea, M. truncatula, and other plant species described by Chardin et al. [4] were aligned using Clus-talX2 [33]. The resulting alignments were used to construct a phylogenetic tree in MEGA 7.0 with the neighbor-joining method [34]. The RWP-RK proteins from A. duranensis, A. ipaensis, and A. hypogaea were also used to construct a phylogenetic tree in MEGA 7.0 with the neighbor-joining method.

Gene Structure and Conserved Motif
Analyses. The genomic and CDS sequences of the wild and cultivated peanut RWP-RK genes were obtained from the peanut genome database and used to construct exon-intron organization maps with the Gene Structure Display Server program [35]. The full lengths of the RWP-RK proteins were used for conserved motif analysis with MEME tools (http://meme-suite.org/), and the positions of the conserved RWP-RK and PB1 domains in each gene were determined with Pfam.

Analysis of cis-Acting Elements in Promoter
Regions. The promoter region of each RWP-RK gene from A. duranensis, A. ipaensis, and A. hypogaea, 2 kb upstream of the initiation codon, was downloaded from the peanut genome database. cis-Acting elements in these promoter regions were investigated using PlantCARE (http://bioinformatics.psb.ugent.be/ webtools/plantcare/html/) [36] and categorized based on their putative functions.

Chromosomal Locations, Orthologous Gene Pairs, and
Gene Duplication Analyses. The physical positions of RWP-RK members were obtained from the peanut genome database, and chromosomal location maps were generated using MapInspect software (http://www.mybiosoftware.com/mapinspectcompare-display-linkage-maps.html). The evolutionary relationships of orthologous gene pairs were assessed based on phylogenetic trees [37]. The RWP-RK genes of A. duranensis, A. ipaensis, and A. hypogaea were clustered using OrthoMCL 2 International Journal of Genomics software (https://orthomcl.org/orthomcl/) to analyze duplicated gene pairs, and the duplicated gene pairs were drawn using Circos software [5,[38][39][40].

Transcription Pattern Analysis of the RWP-RK Genes.
RNA-seq datasets were downloaded from the peanut genome database (https://peanutbase.org/gene_expression) [41]. Twenty-two different tissues from the cultivated peanut, A. hypogaea, were collected as described by Clevenger et al. [41]. These samples were used for gene expression analysis.  [43], and 10-day-old peanut plants were grown in normal and N-free solution culture. The whole plants were collected for gene expression analysis. Each sample was analyzed using three biological replicates. RNA extraction, quantitative real-time PCR (qRT-PCR), and expression analysis were performed as described by Jin et al. [37]. All the primers used are listed in Table S1.

Identification of RWP-RK Proteins in Wild and Cultivated
Peanuts. To obtain an exhaustive list of the RWP-RK proteins from wild and cultivated peanuts, we used gene annotations and BLAST searches of the peanut genome database. We found 17, 27, and 37 RWP-RK proteins in A. duranensis, A. ipaensis, and A. hypogaea, respectively, using gene annotations. We also used the amino acid sequences of the conserved RWP-RK domain (PF02042) and of 14 RWP-RK proteins from Arabidopsis as BLAST queries, and we found 18, 21, and 34 RWP-RK proteins in A. duranensis, A. ipaensis, and A. hypogaea, respectively, using this method. We used Pfam to determine whether each candidate gene contained the conserved RWP-RK domain, and we ultimately confirmed the presence of 16, 19, and 32 RWP-RK members in A. duranensis, A. ipaensis, and A. hypogaea, respectively (Tables 1 and 2 (Tables 1 and 2). The characteristics of the RWP-RK proteins, including isoelectric point and molecular weight, were also analyzed. The isoelectric points ranged from 5.12 (Araip.KR88K) to 9.51 (Araip.80XBW) in A. ipaensis, 5.09 (Aradu.BBG0S) to 10.6 (Aradu.I1BME) in A. duranensis, and 5.03 (Arahy.DJ079B) to 10 Tables 1 and 2). Because the cultivated peanut AA and BB subgenomes are thought to derive from A. duranensis and A. ipaensis, respectively [25][26][27][28][29], the numbers of the RWP-RK genes in the wild species' chromosomes and the corresponding cultivated peanut chromosomes were analyzed. Six and seven of the ten chromosomes in A. duranensis and A. ipaensis, respectively, had the same RWP-RK gene numbers as their corresponding chromosomes in cultivated peanut. Chromosomes 1, 2, 3, and 7 in A. duranensis and chromosomes 2, 7, and 8 in A. ipaensis had numbers of the RWP-RK genes that differed from those of their corresponding chromosomes in cultivated peanut ( Figure 1).

Classification and Phylogenetic Analysis of the RWP-RK
Genes. The plant RWP-RK genes are mainly classified into two groups: RKD and NLP [4]. To investigate the evolutionary relationships among the peanut RWP-RK genes and to classify them into different subgroups, RWP-RK amino acid sequences from the wild and cultivated peanut genomes were used to construct a phylogenetic tree and analyze the conserved protein domains. The A. duranensis, A. ipaensis, and A. hypogaea genomes contained 8, 11, and 14 RKD members and 8, 8, and 18 NLP members, respectively (Tables 1 and 2, Figure 2). Conserved domain analysis revealed that most of 3 International Journal of Genomics the RKD genes were clustered together in a phylogenetic tree, while several RDK members showed close relationships with the NLP genes, including Arahy.L1HKPT, Araip.YWB61, Araip.377BK, and Araip.5C6JK ( Figure 2). All the RKD members contained a single RWP-RK domain, and most of the NLP members contained one RWP domain and one PB1 domain. However, several NLP proteins contained inconsistent numbers of conserved domains. For example, Arahy.1E9R5B and Aradu.G4SB3 contained two RWP-RK domains and two PB1 domains, Araip.YB35N contained two RWP-RK domains and one PB1 domain, and Araip.73BCB contained one RWP-RK domain and two PB1 domains ( Figure 2).

International Journal of Genomics
Arahy.LH2L98 in A. hypogaea showed close relationships with the nitrate response genes AtNLP6 and AtNLP7, suggesting that they may be involved in the nitrate signal response in peanut [4,15,16]. Aradu.46M2Y in A. duranensis, Araip.38X68 in A. ipaensis, and Arahy.I65W25 and Arahy.V4B-GUX in A. hypogaea showed close relationships with MtNIN, a well-studied gene involved in M. truncatula nodule formation [16], suggesting that these four genes may participate in nodule formation in peanut.

Orthologous Gene Pair Analysis.
Many orthologous gene pairs have been identified among wild and cultivated peanuts [25][26][27][28][29], and we therefore investigated RWP-RK orthologs in the A. duranensis, A. ipaensis, and A. hypogaea genomes. A total of 38 orthologous gene pairs were found in the peanut genomes ( Figure 1 and Figures S1-S3). The two wild peanut species shared twelve orthologous gene pairs, and each wild species shared thirteen orthologous gene pairs with A.
hypogaea ( Figure 1). Among the twelve orthologous pairs in the wild species, seven were found on syntenic loci of the A. duranensis and A. ipaensis chromosomes ( Figure 1). By contrast, Aradu.32F23, Aradu.46M2Y, Aradu.1V6B6, and Aradu.7K2S3 were located on different chromosomes from their A. ipaensis orthologs. Although Aradu.H6JXR and its A. ipaensis ortholog were both located on chromosome 8 in their respective genomes, they were located at different chromosomal positions ( Figure 1). All orthologous gene pairs in A. ipaensis and A. hypogaea were found on the syntenic chromosomal loci, and 11 of the 13 gene pairs in A. duranensis and A. hypogaea were located on the syntenic chromosomal loci. These results suggest that the chromosomal rearrangement may have occurred in the diploid peanut genomes but not in that of cultivated peanut.

Gene Structures and Conserved Motifs of the RWP-RK
Genes. To investigate the structures of the RWP-RK genes, RWP-RK CDS and genomic sequences were downloaded from the peanut genome database and analyzed using the Gene Structure Display Server program [35]. Among the peanut RWP-RK members, 26 of the 33 RKD genes had predicted UTRs (except for Araip.5012K, Araip.X4GVE, Araip.YWB61, Araip.377BK, Aradu.TG0QF, Aradu.1V6B6, and Arahy.GWW51V), and 32 of the 34 NLP members had predicted UTRs (except for Aradu.46M2Y and Araip.YB35N) (Figure 4). For the RKD genes, exon numbers ranged from 1 to 8 and intron numbers ranged from 1 to 7, whereas for the NLP genes, exon numbers ranged from 4 to 13 and intron numbers ranged from 3 to 12. Moreover, 8 of the 12 orthologous pairs from A. duranensis and A. ipaensis, 8 of the 13 from A. duranensis and A. hypogaea, and 5 of the 13 from A. ipaensis and A. hypogaea had similar exon and intron numbers (Figures 1 and 4). MEME tool analysis identified 112 conserved motifs in the RWP-RK proteins ( Figure 5 and Figure S4). Most of the NLP proteins had a greater number of conserved motifs than the RKD proteins, and many genes from the same clades had similar motif structures. For example, Arahy.50HX4L and Araip.J80SY had the same types and numbers of motifs, suggesting that these genes may be derived from a common ancestor.  7 International Journal of Genomics experienced one round of duplication, whereas the cultivated peanut A. hypogaea is thought to have experienced two rounds of duplication [25][26][27][28][29]. Gene duplication often occurs by polyploidization during the plant evolution [44], and therefore, we performed a homology analysis of the individual peanut genomes. One, one, and seventeen duplicated RWP-RK gene pairs were found within the A. duranensis, A. ipaensis, and A. hypogaea genomes, respectively. The duplicated genes Aradu.BBG0S and Aradu.G4SB3 were located close to one another on chromosome 3 in A. duranensis, whereas Araip.5C6JK and Araip.73BCB were located on different chromosomes in A. ipaensis (Figure 6), highlighting the diversity of the two wild peanut genomes. In A. hypogaea, only eight RWP-RK genes had no duplicates, including four RKD genes (Arahy.GWW51V, Arahy.D-D2ABE, Arahy.L1HKPT, and Arahy.552ZQ0) and four NLP genes (Arahy.XY5KEE, Arahy.8R729R, Arahy.1E9R5B, and Arahy.JSL8JQ). Among the duplicated genes in A. hypogaea, fifteen gene pairs were interchromosomal duplications, two duplications were located on the same chromosomes, and no tandem duplications were found. The NLP members formed eleven duplicated gene pairs: Arahy.LH2L98/Arahy.0FWB0U, Arahy.Y03563/Arahy.2T470H, Arahy.DJ079B/Arahy.2T470H, Arahy.EX05TD/Arahy.2T470H,  (Tables 1 and 2, Figure 6).   International Journal of Genomics responses of the RWP-RK genes, we identified cis-acting elements in their promoter regions. Aradu.I1BME and Araip.UMW8F were discarded due to lack of promoter information. A total of 98 kinds of cis-acting elements were found across all peanut RWP-RK gene promoters (87 in A. duranensis, 89 in A. ipaensis, and 92 in A. hypogaea), 60 of which had predicted putative functions, including nine developmentrelated elements (six in A. duranensis, nine in A. ipaensis, and eight in A. hypogaea), five environmental stress-related elements (four in A. duranensis, five in A. ipaensis, and four in A. hypogaea), ten hormone-responsive elements (eight in A. duranensis, eight in A. ipaensis, and ten in A. hypogaea), twenty-six light-responsive elements (twenty-five in A. duranensis, twenty-four in A. ipaensis, and twenty-five in A. hypogaea), four promoter-related elements (four in A. duranensis, four in A. ipaensis, and three in A. hypogaea), and six site-binding-related elements (five in A. duranensis, five in A. ipaensis, and six in A. hypogaea) (Figure 7, Tables S2-S5). Moreover, the numbers and types of cis-acting elements differed among the RWP-RK gene promoters, underscoring

10
International Journal of Genomics the functional diversity of these genes (Figure 7). The number of occurrences of each binding site differed between A. duranensis and A. ipaensis (Figure 8), and the number of occurrences of each binding site in A. hypogaea was close to the sum of its occurrences in A. duranensis and A. ipaensis. For example, A. hypogaea contained 39 methyl jasmonate (MeJA) response elements, and the sum of MeJA response elements in A. duranensis and A. ipaensis was 37 (18 in A. duranensis and 19 in A. ipaensis) (Figure 8). All RWP-RK promoters contained at least one light-responsive element, ranging from 4 to 11 in A. duranensis, 4 to 12 in A. ipaensis, and 1 to 11 in A. hypogaea (Figure 7, Table S2). Moreover, light-responsive elements were the most abundant element in each RWP-RK promoter, with the exception of Arahy.EX05TD. Sixty-one of the sixty-five RWP-RK genes contained the light-responsive element Box 4, suggesting that these genes function in light signaling pathways. The exceptions included Aradu.7K2S3, Araip.P1CBC, Araip.377BK, and Arahy.EX05TD. In addition, all of the RWP-RK promoters contained the promoter-related elements TATA-box and CAAT-box, which are responsible for the promoter function (Table S2).

Transcription Patterns of the Peanut RWP-RK Genes.
To investigate the transcription patterns of the peanut RWP-RK genes, heatmaps were constructed using RNA-seq datasets downloaded from the peanut database [41]. First, we used qRT-PCR analysis to verify the expression levels of several randomly selected RWP-RK genes in several tissues of wild and cultivated peanut. We found that their relative expression levels were similar to those in the published RNA-seq datasets ( Figure S5). The RWP-RK genes were expressed at

13
International Journal of Genomics different levels in different tissues (Figures 9 and 10). Several genes were expressed at high levels in most of the tissues tested, including A. duranensis genes Aradu.YRC2R and Aradu.T4VLF, A. ipaensis genes Araip.R44NW and Araip.Y4AFN, and A. hypogaea genes Arahy.62AJ6F, Arahy.K1SYDF, Arahy.0FWB0U, Arahy.LH2L98, Arahy.1E9 R5B, and Arahy.2T470H (Figures 9 and 10). By contrast, several genes showed extremely low expression levels in all tissues tested; these included Aradu.TG0QF, Araip.UMW8F, Araip.YWB61, Arahy.DD2ABE, Arahy.F3ZC PW, and Arahy.GWW51V (Figures 9 and 10). The MtNIN homologs Aradu.46M2Y, Araip.38X68, Arahy.I65W25, and Arahy.V4BGUX ( Figure 3) were expressed at high levels in roots and nodule roots but at low levels elsewhere (Figures 9 and 10), suggesting that they may function in the nodule formation in peanut. In addition, most orthologous gene pairs showed similar expression levels in the two wild peanut species across many tissues (Figures 1 and 8). For example, the orthologs Aradu.T4VLF and Araip.Y4AFN showed similar expression levels in most tissues, with the exception of pericarp pattee 6. However, most orthologous gene pairs from A. ipaensis and A. hypogaea showed different expression levels, and similar results were found for A. duranensis and A. hypogaea orthologs (Figure 1 and Figure S6), highlighting the differences in gene expression between wild and cultivated peanut.

Discussion
The identification and characterization of the RWP-RK proteins have increased our understanding of nitrogen response and gametophyte development in many plant species [4,15,19]. Peanut is a globally important legume crop, and the characterization and comparative analysis of the RWP-RK pro-teins from wild and cultivated peanuts will increase our understanding of nitrate response and gametophyte development regulation in these species. In the current study, we identified and characterized 67 RWP-RK proteins from the wild species A. duranensis and A. ipaensis and from the cultivated peanut, A. hypogaea.
The A. duranensis, A. ipaensis, and A. hypogaea genomes contain 16,19, and 32 RWP-RK proteins, respectively (Tables 1 and 2). The genome size of A. hypogaea is close to the sum of the A. duranensis and A. ipaensis genome sizes [25,26], and the number of RWP-RK proteins in A. hypogaea is also close to the sum of those in A. duranensis and A. ipaensis (Tables 1 and 2). Although the two wild peanuts are diploid species, their numbers of RWP-RK proteins are different. The genome sizes of A. duranensis, A. ipaensis, and A. hypogaea are 1.25 Gb, 1.56 Gb, and 2.7 Gb, respectively [25][26][27][28][29], indicating that the number of RWP-RK proteins has changed during evolution and is directly related to the genome size in peanuts. By contrast, the Arabidopsis, rice, B. distachyon, and wheat genomes contain 14, 15, 18, and 37 RWP-RK proteins [4,10], respectively, and their genome sizes are 125 Mb [45], 466 Mb [46], 260 Mb [47], and 17 Gb [48], respectively, suggesting that the number of RWP-RK proteins has no direct relationship with the genome size in other plant species. Approximately half of the RWP-RK proteins are classified as NLP members in vascular plants [4], including wild and cultivated peanuts. For example, the Arabidopsis, rice, A. duranensis, A. ipaensis, and A. hypogaea genomes contained nine (64% of all the RWP-RK proteins), six (40%), eight (50%), eight (42.11%), and 18 (56.25%) NLP members, respectively. The two diploid peanuts have the same number of NLP proteins, suggesting that the evolution of the NLP genes is conserved in the wild peanut species. However, the number of NLP proteins is expanded in A. hypogaea and is two greater than the sum of the A. duranensis and A.       ipaensis NLP proteins. By contrast, the number of the RKD proteins in A. hypogaea is five less than the sum of the A. duranensis and A. ipaensis RKD proteins (Tables 1 and 2). These results indicate that the wild species have diverse numbers and types of the RWP-RK proteins and that the numbers and types of the RWP-RK proteins have changed during the evolution in the cultivated peanut.
A. hypogaea chromosomes 1 to 10 are thought to be derived from A. duranensis, and chromosomes 11 to 20 are thought to be derived from A. ipaensis [25][26][27][28][29]. Chromosomes 4 and 6 had no RWP-RK members in both A. duranensis and A. ipaensis, and thus, chromosomes 4, 6, 14, and 16 in A. hypogaea also had no RWP-RK genes (Figure 1). Chromosome 7 in A. duranensis contained one member, Aradu.46M2Y, but no RWP-RK gene was found on chromosome 7 in A. hypogaea. In addition, A. hypogaea chromosome 17 contained 3 RWP-RK members, whereas A. ipaensis chromosome 7 contained only 2 members. Moreover, the Aradu.46M2Y ortholog Arahy.V4BGUX was found on chromosome 17 in A. hypogaea, indicating that the chromosome rearrangement may have occurred between chromosomes 7 and 17 in A. hypogaea.
Wild and cultivated peanuts are thought to have experienced one and two rounds of duplication, respectively [25][26][27][28][29]. Many more orthologous gene pairs from A. duranensis and A. ipaensis (five out of 12) were located on different chromosomes than those from A. ipaensis and A. hypogaea (0 out of 13) or from A. duranensis and A. hypogaea (two out of 13) (Figure 1). In addition, among these orthologous gene pairs, eight of 12 from the two wild peanut species, eight of 13 from A. duranensis and A. hypogaea, and only five of 13 from A. ipaensis and A. hypogaea have similar exon and intron numbers (Figures 1 and 4). These results indicate that the chromosomal rearrangement may have occurred during the first round of duplication, and the gene structure alteration may have occurred during the second round of duplication, especially in the genes derived from A. ipaensis. Moreover, the orthologous gene pairs in the wild species showed similar expression levels in many tissues, whereas orthologous gene pairs from the wild species and cultivated peanut showed distinct expression patterns (Figures 9 and 10, Figure S6), indicating that the expression of the orthologs may have altered during the evolution of cultivated peanut.
cis-Acting elements in promoter regions are responsible for modulating the gene expression. We found different numbers and types of cis-acting elements in the RWP-RK promoters (Figure 7, Tables S2-S5), and these may be responsible for different expression levels of the RWP-RK genes in different tissues (Figures 9 and 10). However, some RWP-RK genes that contained many kinds of cis-acting elements in their promoters nonetheless showed extremely low expression levels in all the tested tissues (e.g., Arahy.DD2ABE and Arahy.F3ZCPW) (Figure 10). The gene expression is affected by many factors in addition to the presence of specific cisacting elements. For example, epigenetic modifications, such as DNA methylation, have a substantial effect on the gene expression [49] and may be responsible for the low expression observed in these RWP-RK genes ( Figure 10). Moreover, the expression of orthologous genes from the two wild peanuts was similar in many tissues, whereas orthologous genes from wild and cultivated peanut showed different expression levels (Figures 9 and 10). The wild peanut experienced one round of whole genome duplication, whereas cultivated peanut experienced two rounds of whole genome duplication, and this may explain their different tissue-specific gene expression patterns. In addition, Aradu.YRC2R and Aradu.T4VLF from A. duranensis, Araip.R44NW and Araip.Y4AFN from A. ipaensis, and Arahy.K1SYDF, Arahy.0FWB0U, Arahy.62AJ6F, and Arahy.LH2L98 from A. hypogaea showed close relationships with the nitrate response genes AtNLP6 and AtNLP7 (Figure 3). However, the expression level of these genes under N-limited conditions was similar to that under normal conditions, suggesting that their gene expression was not regulated by N ( Figure S7).
Gene duplication occurs on various scales during evolution, including segmental, tandem, and whole genome duplications [50]. The cultivated peanut is descended from the hybridization and polyploidization of two wild diploids, which resulted in whole genome duplication and produced many duplicated gene pairs [25][26][27][28][29]. Thus, up to 17 duplicated RWP-RK gene pairs were found in the genome of cultivated peanut. In contrast, only one duplication event, which appears to be a segmental duplication, was found in each of the two wild species (Figure 6). Duplicated genes can be lost, pseudogenized, or become novel genes during evolution [30,51]. The two RWP-RK genes from each duplication event belong to the same subfamily in A. duranensis and A. hypogaea, and these duplicated gene pairs may retain many similar functions because of their shared origin. In contrast, Araip.73BCB belonged to the NLP subgroup, but its duplicate gene Araip.5C6JK was an RKD member. Araip.73BCB contained two PB1 domains whereas Araip.5C6JK had no PB1 domain, and these two genes showed different expression patterns in many tissues (Figure 9), indicating their functional differentiation. In addition, the amino acid numbers of most RKD proteins other than Araip.5C6JK were less than those of NLP proteins (Tables 1 and 2, Figure 2). These results suggest that the PB1 domain of Araip.73BCB may have been lost and the remainder of the gene retained to give rise to a new gene, Araip.5C6JK.

Conclusions
In summary, many characteristics of the RWP-RK proteins were analyzed in wild and cultivated peanuts, including chromosomal locations, gene structures, orthologous gene pairs, conserved motifs, duplications, phylogenetic relationships, cis-acting elements, and transcription patterns. Although the two diploid peanuts A. duranensis and A. ipaensis had some common features, the RWP-RK genes in these two wild species also showed some degree of diversity. In addition, while the RWP-RK genes of cultivated peanut retained some characteristics of those from wild peanuts, they also changed during evolution compared with those of their two diploid ancestors.

Data Availability
The original data of the RWP-RK family genes are available from the peanut genome database (https://www .peanutbase.org/). 16 International Journal of Genomics