Genome-Wide Analysis and Evolution of the Pto-Like Protein Kinase (PLPK) Gene Family in Pepper

The tomato Pto gene, which encodes a serine/threonine kinase (STK) domain-containing protein, confers resistance to bacterial speck disease caused by Pseudomonas syringae pv. tomato (Pst). In this study, in vivo recognition assays using PVX constructs showed that AvrPto was specifically recognized in the pepper genotypes. This AvrPto recognition caused a nonhost hypersensitive response (HR) and localization of the PVX::AvrPto fusion protein to inoculated pepper leaf tissues, which indicates the presence of a similar Pto recognition mechanism in pepper as in tomato. However, genome-wide analysis in pepper revealed no Pto clade corresponding to that in tomato, suggesting an alternative system for Pto recognition in pepper. Nevertheless, 25 Pto-like protein kinases (PLPKs) with a highly conserved STK domain have been identified in the pepper genome. For the majority of the amino acid sites in the STK domain of Ptos and PLPKs, nonsynonymous (dN) to synonymous (dS) nucleotide substitution ratios (ω) were less than one, suggesting that purifying selection played a predominant role in the evolutionary process. However, some amino acid sites were found to be subjected to episodic positive selection in the course of evolution of Pto homologs, and, thus, different evolutionary processes might have shaped the Pto gene family in plants. Based on RNA-seq data, PLPK genes and other Pto pathway genes, such as Prf, Pti1, Pti5, and Pti6 were expressed in all tested pepper genotypes. Therefore, the nonhost HR against Pst in pepper may be due to the recognition of the AvrPto effector by a PLPK homolog, and subsequent action of downstream components of the Pto signaling pathway. However, the possibility remains that the recognition of AvrPto in pepper plants may involve activities of other receptor like kinases (RLKs). The identification of the PLPKs in this study will serve as a foundation for further efforts to understand the roles of PLPKs in nonhost resistance.


Introduction
Plants have developed various defense responses to pathogens. Disease resistance often occurs in a gene-for-gene manner through interaction between proteins encoded by the plant disease resistance (R) genes and the corresponding effector proteins encoded by pathogen avirulence (Avr) genes. Such interactions activate a cascade of defense-related responses in the host to suppress the pathogen attack [1,2]. Several plant R genes have been cloned and progress has been made to understand the host resistance mechanisms [3][4][5][6][7][8].
Plant R genes are classified into eight classes based on the presence of conserved domains [7,9]. The Pto gene in tomato (Solanum pimpinellifolium L.) is among the best characterized R genes [10]. Pto encodes a serine-threonine kinase (STK) and confers resistance to P. syringae strains that express effector protein AvrPto or AvrPtoB [11][12][13]. The Pto kinase interacts directly with AvrPto and AvrPtoB, thereby triggering immunity in the host plant [12,14,15]. Both AvrPto and AvrPtoB have been shown to enhance virulence when expressed in bacterial strains in which they do not occur naturally [16][17][18]. AvrPtoB functions in pathogenesis as an inhibitor of programmed cell death (PCD) disrupting host defense responses [18]. Interestingly, AvrPto is also capable of suppressing PCD induced by nonhost pathogens in Nicotiana benthamiana and tomato [19]. Three classes of downstream effectors have been identified in the Pst resistance system. Pti1 is a protein kinase that enhances the hypersensitive response in vivo; Pti4, Pti5, and Pti6 are defense-related EREBP-like transcription factors [16,20,21] and Prf is a nucleotide-binding leucine-rich repeat (NB-LRR) protein. The phosphorylation of the Pto kinase by AvrPto effector is sensed by Prf, inducing effector-triggered immunity [22][23][24][25].
Despite extensive studies of the Pto pathway genes in plants, particularly in tomato, no functional homologs of Pto that recognize AvrPto have been identified in any genus other than Solanum. There have been efforts to amplify Pto-like sequences using several plant species [26,27]. In these studies, it was suggested that the AvrPto-recognizing molecule may be similar to Pto kinase from tomato. However, mutation analysis has shown that the AvrPto protein is differentially recognized by tomato and tobacco [17], suggesting that the AvrPto-recognizing protein in tobacco may be unrelated to the tomato Pto sequence [17]. Recognition of AvrPto has been reported in soybean [28], suggesting that AvrPto recognition and downstream signaling cascades may be conserved across plant species [28]. The recognition of AvrPto by these plant species also raises a question about the role of the AvrPto-recognizing proteins in nonhost species of Pst.
Recently, Pto-like genes have been identified in several plant species including some Solanum species [27], common bean [26], grapevine [29], cucumber [30], banana [31], strawberry [32], and Citrus [33]. Notably, Pto orthologs are not present in two different S. lycopersicum genotypes [34,35], whereas the AvrPto-specific Pto-mediated resistance has likely been introgressed from S. pimpinellifolium in some cultivated S. lycopersicum varieties [10,36]. Despite advances in the understanding of the molecular genetics of Pto genes, the molecular evolution of the Pto gene remains poorly understood. Understanding the evolution of the Pto genes is important to unravel their functional divergence as well as to understand their role in nonhost disease resistance.
Pepper, an important spice crop, is considered a nonhost of Pst. Several lines of evidence suggest that pepper may have the ability to recognize the AvrPto protein and induce resistance [19]. However, Pto kinase has not been established as the recognizing component. In this study, we tested whether pepper can recognize AvrPto using in vivo recognition assays with a PVX system. In addition, we performed genome-wide analysis of the Pto-like gene family (PLPKs) in pepper and identified 25 full-length PLPK genes, which were classified into eight subclasses (PLPK I to PLPK VIII) based on sequence similarity and phylogenetics. To gain insight into the evolutionary diversity of the plant Pto-like genes, comparative phylogenetic and molecular evolutionary analyses of the Pto gene family were performed, using Pto genes from other Solanaceae family crops, such as tomato, potato, and N. benthamiana, as well as Arabidopsis and rice. Structural characteristics of pepper PLPKs were investigated using various computational tools. Finally, expression profiles of the PLPK genes from various pepper genotypes were investigated using RNA-seq expression data.

Transient expression of AvrPto in pepper using PVX-derived vectors
For transient expression assays using PVX-derived vectors, AvrPto and the mutant AvrPto I96T were provided by G. Martin and X. Tang respectively, and cloned into pPVX201 provided by D. Baulcombe [37]. The 528-bp coding regions of each AvrPto gene was amplified by PCR from the plasmids pPtE6 [28] and pPtE6:I96T [17] using the primers 5'-ATATCGATGGG AAATATATGTGTC-3' and 5'-GAGGTCGACATTATGACGCC-3'. The introduced ClaI and SalI sites in the primers are underlined. The amplified fragments were cloned into pGEM T (Promega, Madison, WI) and confirmed by sequencing. The products were digested with ClaI and SalI and subcloned into corresponding vector sites. The resulting pPVX201 derivatives were designated pPVX201::AvrPto and pPVX201::AvrPto I96T . A derivative of pPVX201 carrying GFP cDNA, pPVX204, was used as a control [37]. For inoculum production, 4-6 week old N. benthamiana plants were mechanically inoculated with 50 μg plasmid in a 15% bentonite suspension in 43 mM sodium phosphate buffer (pH 7.0) [37]. Systemically infected N. benthamiana plants inoculated with pPVX201, pPVX204 or pPVX::AvrPto I96T developed systemic mosaic symptoms 14 days post-inoculation (dpi), whereas pPVX::AvrPto induced bleaching and mottling followed by systemic necrosis [38]. Non-necrotic symptomatic N. benthamiana leaves were homogenized in 50 mM phosphate buffer (pH 7.0) 10-14 dpi and used to mechanically inoculate 6-8 week old pepper seedlings. Mock-inoculated pepper controls were routinely included. Infection was confirmed by hybridization analysis and positive indirect enzyme-linked immunosorbent assay (ELISA) using [39] anti-PVX antibody (Agdia, Elkhart, IN). Total RNA was prepared from uninoculated upper leaves of inoculated plants 14 dpi, blotted onto Hybond N + , and hybridized with radiolabeled PVX-derived and AvrPto sequence. mapping of 75 plants described by Livingstone et al. [40] was used to collect RFLP segregation. The genetic map positions of the cloned PCR fragments were determined relative to the original framework marker data set with MAPMAKER software.
Chromosomal location of the pepper Pto-like genes and other pepper genes related to Pto pathway The chromosomal locations of the PLPK genes and other Pto pathway genes in pepper were obtained from the pepper genome database (http://cab.pepper.snu.ac.kr/). The PLPK genes were mapped to the pepper chromosomes using the MapChart program [41]. Genomic DNA and CDS sequences of the pepper genes were obtained from pepper genome database (http:// cab.pepper.snu.ac.kr/). Exon-intron structure of the PLPK genes and other Pto pathway genes was visualized by comparing genomic DNA sequences with their corresponding CDS sequence with the gene structure display server (http://gsds.cbi.pku.edu.cn/).

Phylogenetic analysis and identification of conserved motifs
To establish evolutionary relationships, protein sequences from various plant species including pepper, tomato, potato, N. benthamiana, Arabidopsis, and rice were included in the phylogenetic tree. To determine phylogenetic relationships between PLPKs in pepper and in other plant species, a multiple sequence alignment was generated including all available PLPK sequences from plant species such as rice, Arabidopsis, tomato, and N. benthamiana, with the predicted pepper PLPKs using the Clustal Omega program with default parameters (http:// www.ebi.ac.uk/Tools/msa/clustalo/). Various classes of Arabidopsis protein kinases [42] were also included in the phylogenetic analyses to distinguish Pto and PLPK members from other classes of protein kinases. Aligned sequences with highly divergent regions or gaps resulting in uncertain alignments were edited using JalView 2.8 [43] and were excluded from further analysis. Phylogenetic analysis of full-length PLPK sequences was performed using the MEGA 6.0 software with the neighbor-joining (NJ) method [44], and the bootstrap analysis was carried out with 1000 iterations. Similarly, a separate phylogenetic analysis on the PLPKs from pepper was performed. The tree was rooted with pepper receptor like kinase (RLK) protein. The pepper PLPK members were classified based on their phylogenetic relationship with PLPK members from other plant species. Subcellular localization of the pepper PLPKs was predicted using the EuLoc web server [45]. The structural and functional domains of the PLPKs were analyzed using the SMART program (http://smart.embl-heidelberg.de/).

Detection of selection pressure
To explore the nature of the selection pressure on the serine/threonine kinase (STK) domain of the Pto and PLPK residues, nonsynonymous (dN) and synonymous (dS) substitutions, and the dN/dS ratio (ω) for each nucleotide codon were estimated using the Datamonkey web server (http://www.datamonkey.org/). To avoid false positives, an automated genetic algorithm recombination detection (GARD) analysis was performed [46]. The site-specific selection pressure was calculated using several codon-based maximum-likelihood methods: fixed-effect likelihood (FEL), internal fixed-effect likelihood (IFEL), and single likelihood ancestor counting (SLAC), as implemented in the Datamonkey web interface [47]. Codon-based nucleotide sites were investigated using the REV nucleotide substitution model. To identify the individual codons that are subject to diversifying selection, the mixed effects model of evolution (MEME) analysis was performed. A ω value <1 indicates sites under negative (purifying) selection, a ω value >1 indicates sites under (diversifying) selection, and a ω value = 1 indicates sites that are not subject to selection pressure.

RNA-seq expression data analysis
The expression patterns of PLPK genes from various pepper genotypes were analysed using RNA-seq data from previous research [48]. The sequence reads were mapped to the pepper transcriptome database (http://cab.pepper.snu.ac.kr) with CLC Genomics Workbench v8.0 (CLC Bio, Aarhus, Denmark) with default parameters. Expression values were measured in RPKM (reads per kilobase of exon model per million mapped reads). The RPKM values were log2-transformed and the heat map was generated using the heatmap.2 function from the ggplots in R-package.

AvrPto specifically induces HR and resistance to PVX in pepper
To determine whether specific recognition of AvrPto occurs in pepper, PVX-derived vectors [49] were used to express AvrPto and AvrPto I96T , a mutant with a substitution at position 96 that prevents the interaction between AvrPto and Pto in vitro and in planta with no effect on virulence [17]. In a preliminary study, six different pepper genotypes, including Perennial, 234, Habanero, Rnaky, ECW, and CM334 were screened for susceptibility to PVX (S1 Fig). When inoculated with the vector alone, 234, Habanero, ECW, and RNaky genotypes developed viral symptoms throughout the plant (S1 Fig). 234 and ECW genotypes were subsequently grown and inoculated with the PVX vector alone or a construct expressing GFP, AvrPto, or AvrPto I96T (Fig 1). Inoculation with the vector alone (S1 Fig) and GFP-containing PVX constructs ( Fig 1A panels I, II) resulted in systemic symptoms and infection was confirmed by the presence of viral antigen and PVX-homologous sequences in both inoculated (data not shown) and upper uninoculated tissue (Fig 1), and in the case of GFP-containing constructs by fluorescence of uninoculated leaves under UV light (data not shown). By contrast, when plants were inoculated with pPVX::AvrPto, symptoms were restricted to lower inoculated leaves, which displayed localized necrotic lesions (Fig 1A panels III, IV). In addition, PVX antigen and PVXand AvrPto-homologous sequences were detected only in inoculated tissue (data not shown), and not in uninoculated tissue ( Fig 1B). In susceptible pepper plants, PVX systematically spreads throughout the plant from the site of infection. Similarly, PVX::AvrPto was also expected to show systemic spread. However, the systemic movement of PVX::AvrPto is restricted in pepper and can only be detected in infected lower leaves. This is may be due to the localization of AvrPto by a Pto homolog in inoculated leaves, and thus resulting in the inhibition of systemic movement of the PVX::AvrPto. Therefore, signal for neither PVX antigen nor AvrPto was detected in the upper uninoculated leaves (Fig 1B, Lane 4).
Because all PVX-susceptible pepper genotypes appeared to recognize AvrPto, based on the appearance of characteristic necrotic local lesions, conventional genetic approaches that require differential host responses are not applicable. The availability of a mutant elicitor, AvrPto I96T that specifically inhibits interaction with Pto, allowed an alternative approach. If a Pto homologous sequence from pepper encodes a functional gene product in vivo that is able to recognize AvrPto and if this recognition event is important in localizing PVX infection, then Pepper seedlings inoculated with PVX vectors for expression of GFP, AvrPto and AvrPto I96T . A: Transient expression of AvrPto and AvrPto I96T using a PVX-derived vector system in C. annuum 'ECW' (upper panels) and C. chinense 'PI159234' (lower panels). Plants were photographed at 14 dpi. Panels I, II: pPVX::GPF; Panels III, IV: pPVX::AvrPto; Panels V, VI: pPVX::AvrPto I96T . B: Gel blot hybridization analysis of PVX RNA accumulation and the presence of AvrPto in uninoculated upper leaves of C. annuum ECW and C. chinense 'PI159234' infected with pPVX vector constructs. Total RNA was prepared from upper uninoculated leaves 14 dpi with the PVX constructs above or from mock-inoculated plants, blotted onto Hybond N + , and hybridized with a radiolabeled PVX-derived sequence (upper panels) and AvrPto (middle panels). Lane 1, mock-inoculated control; Lane 2, pPVX vector alone; Lane 3, pPVX::GFP; Lane 4, pPVX::AvrPto; and Lane 5, pPVX::AvrPto I96T . The lower panel shows ethidium bromide-stained ribosomal RNA prior to blotting.
PVX vectors expressing AvrPto I96T should spread systemically. When pepper seedlings were inoculated with pPVX::AvrPto I96T , chlorotic spots appeared on inoculated leaves that developed necrotic halos of approximately the same size as the necrotic local lesions observed in the incompatible interaction. These symptoms subsequently spread throughout the plant (Fig 1A  panels V, VI). Systemic infection was confirmed by the presence of PVX antigen (data not shown) and hybridization of total RNA from upper uninoculated leaves with PVX-and AvrPto-homologous probes (Fig 1B).
Taken together, only PVX constructs expressing AvrPto were restricted to inoculated tissue in two pepper species. All other PVX constructs moved systemically, which is consistent with the hypothesis that the Pto homologs in pepper may also function to recognize AvrPto in vivo.

Identification, classification and phylogenetics of Ptos and PLPKs
To identify the possible interactor of AvrPto, we performed a genome-wide analysis to search for tomato a Pto homolog in the CM334 pepper cultivar. To distinguish the Pto homologs from other classes of STK-domain containing proteins, a phylogenetic tree was constructed using Pto homologs from various plant species and various RLKs from Arabidopsis [42]. Only the region between subdomain I and XI of the STK domain was included in the phylogenetic tree. Pto homologs, including PLPKs, clearly formed a separate cluster from Arabidopsis RLKs and other protein kinase classes (S2 Fig). Based on sequence similarity and phylogenetic analysis, 25 Pto-like protein kinase genes were identified in the pepper genome (Table 1).
To examine the evolutionary relationships between Pto and PLPK homologs, and to classify the pepper PLPKs, a rooted phylogenetic tree was constructed using the predicted pepper PLPK sequences and full-length Pto and PLPK homologs from various plant species, such as tomato, N. benthamiana, potato, Arabidopsis and rice (Fig 2). Phylogenetic analysis showed the existence of two distinct clusters of Pto homologs, Pto and PLPK (Fig 2). The Pto clade is specific to Solanaceae species, and within this clade homologs of tomato Ptos are clustered together. However, no Pto clade members were found in pepper, suggesting that the origin of Ptos occurred during the early evolution of Solanaceae sublineages. The pepper PLPKs further separated into eight subclasses (PLPK I to PLPK VIII). In subclass PLPK I and PLPK IV, PLPKs from various dicot species were clustered together, whereas in subclasses PLPK II, PLPK III, PLPK V, PLPK VI, and PLPK VII, both dicot and monocot PLPK members were present. However, in subclass PLPK II and PLPK III monocot and dicots formed distinct clades. At least one Arabidopsis PLPK member was observed in all PLPK subclasses, with the exception of subclass PLPK VIII, which is specific to the Solanaceae family. Overall, the phylogenetic analysis showed that pepper PLPKs cluster in a similar pattern to other dicots. Moreover, PLPK members from rice can be found among dicot members throughout the phylogenetic tree, implying that PLPKs have evolved in the common ancestors of monocots and dicots.

Comparative phylogenetic analysis of pepper PLPKs and tomato Ptos and PLPKs
To determine the phylogenetic relationships between pepper and tomato Pto homologs, a phylogenetic tree was generated (S3 Fig), which showed that Pto class members are not present in pepper. All PLPK homologs showed similar clustering patterns in both pepper and tomato with the exception of subclass PLPK VI, which was not represented in tomato. The number of PLPKs observed in the subclasses PLPK I, PLPK II, and PLPK IV was the same in pepper and tomato i.e. 1, 2, and 3, respectively. In each of the PLPK subclasses PLPK III and PLPK V, 2 and 3 members were observed in tomato and pepper, respectively. In subclass PLPK VII, 5 and 6 members were observed in tomato and pepper, respectively. In subclass PLPK VIII, five members were found in pepper, whereas only one PLPK member was found in tomato. These observations suggest that there has been sublineage expansion and diversification of Ptos and PLPKs in plants of the Solanaceae.  Table 1. Seven out of 25 PLPK genes were not assigned to any chromosome. The remaining 18 PLPK genes were localized to six chromosomes. The number of the genes on the chromosomes varied greatly from none on Chr1, Chr4, Chr5, Chr8, and Chr12 to a maximum of seven Evolution of the Pto-Like Protein Kinase (PLPK) Family in Pepper on Chr2. Chr3 and Chr6 each had three PLPK genes. Chr9 and Chr10 each showed two PLPK genes, whereas a single PLPK was localized to Chr7 (S4A Fig). Based on linkage mapping analysis, out of seven PLPK genes, which were not assigned to any chromosome, at least one PLPK gene could be assigned to Chr12, and two PLPKs (HPto-a and HPto-d) which are collinear with the Prf homologs (Prf1.2 (HPrf-c) and Prf1.3 (HPrf-b)), could be assigned to Chr11 (S4B Fig). HPto-a and HPrf-b (Prf1.3) were mapped to the corresponding position of the tomato Pto clade, suggesting that HPto-a is an ortholog of tomato Pto. However, in contrast to tomato Prf, which is embedded in the Pto cluster, HPrf-b was located approximately 2 cM from HPto-a. Among the four Prf homologs, Prf1.1 was located on Chr4, Prf1.2 and Prf1.3 were located on Chr11, and Prf1.4 could not be associated with any chromosome. Among the four Pti1 homologs, Pti1.1, Pti1.2 (HPti1-a), and Pti1.3 were located on Chr3, Chr5, and Chr12, respectively, and Pti1.4 (Hpti1-b) could be located to Chr5 based on the linkage map (S4B Fig). The two Pti1 homologs, HPti1-a and Hpti1-b, that were detected in pepper, showed synteny with tomato Chr5. Pti4, Pti5, and Pti6.1 were located on Chr5, Chr2, and Chr6, respectively. Pti6.2 was not located to any chromosome.
Exon-intron structure analysis showed that most of the pepper PLPK genes are intron less with some exceptions; PLPK3, PLPK5, PLPK6, PLPK7, and PLPK18 genes showed one intron, and PLPK4 gene showed two introns (Fig 3B). In contrast, pepper Prf genes showed three to six introns with different lengths and positions (S5 Fig). Similarly, Pti genes were also found to be highly divergent and showed a complex structure with zero to seven introns having variable distributions (S5 Fig).

Conserved motif analysis of pepper PLPKs
A multiple sequence alignment of the predicted amino acid sequences of the 25 pepper PLPKs and the corresponding region of tomato Pto was performed using the Clustal Omega program. The predicted length of the pepper PLPK sequences varied from 683 to 898 amino acids. Analysis of the predicted amino acid sequences of the PLPKs from pepper showed 54 to 71% sequence identity with tomato Pto. Within the pepper PLPK phylogenic groups, the amino acid sequence identity varied from 35 to 96%. All pepper PLPKs showed a highly conserved STK domain of approximately 275 amino acids (Fig 4). Each of the eight subclasses contains the conserved amino acid residues found in subdomain I through XI of the STK domains that are found in most of the plants. Alignment of amino acid sequences revealed several characteristic features of PLPKs that are highly conserved in the Pto homologs, such as the STK domain, the presence of the activation domain located between subdomains VII and VIII, and the internal P+1 loop site, which is responsible for the specific binding of AvrPto effectors [50]. The STK domain also contained several dispersed variable amino acid sites. In addition, three of the four autophosphorylation sites (Ser or Thr) in the activation domain of Pto [51] are conserved in the corresponding region of all pepper PLPKs (Fig 4). Mutations at the highly conserved Val55 and His94 positions were reported to disrupt the Pto-AvrPto interaction in yeast and inhibit the Pto-mediated resistance [52].
We used the SMART program to confirm the structural features of the PLPKs and found that most of the PLPKs contained four domains; a signal peptide region, a malectin-like region, a transmembrane region, and the STK domain (Fig 3C). By contrast, tomato Pto contained only the STK domain. PLPK18, a member of class PLPK V, harbored an additional kinase domain at its C-terminus. PLPL7, a PLPK VIII class member, included two stress-antifung domains at its N-terminus. In addition, several of the pepper PLPKs showed low complexity regions. The protein subcellular localization prediction tool, EuLoc [45], showed that all pepper PLPKs proteins are predicted to localize to the plasma membrane. All Pti1 proteins were predicted to localize to the plasma membrane as well, whereas Pti4, Pti5, and Pti6 were predicted to localize to the nucleus, and the Prfs were predicted to localize to the cytoplasm.
To identify conserved Pto autophosphorylation sites and other critical residues in the activation domain of the PLPKs from pepper, multiple sequence alignment analysis was performed including the tomato Pto sequence (Fig 4). Three of the five Pto autophosphorylation sites [51], Thr195, Ser198, and Thr199 were highly conserved in pepper PLPKs with a few exceptions; Thr195 was replaced by Gly in PLPK24 and PLPK25, and Ser198 was replaced by Val in PLPK10 and PLPK13, and Thr in PLPK24 and PLPK25 sequences. Ser198 is required for the AvrPto-Pto-mediated hypersensitive response, and was highly conserved in pepper PLPKs. The Pto autophosphorylation site, Thr190, was replaced by a Pro in the majority of the pepper PLPKs. In some of the PLPKs, such as PLPK3, PLPK4, PLPK5, and PLPK7, a Ser was replaced by Thr, creating an alternative phosphorylation site. In PLPK22, Thr190 was replaced by Leu. Two other Pto phosphorylation residues, Thr204 and Tyr207, are crucial for Pto function but are not autophosphorylated in vitro [14,53]. Both these Pto phosphorylation residues are highly conserved; however, the Thr204 was replaced by a Ser in all pepper PLPKs.

Analysis of selection pressure
To examine whether the Pto and PLPK gene classes have been subjected to adaptive evolution following gene duplication, an analysis of variation in selective pressure was carried out. A codon-based multiple sequence alignment was analyzed through the use of various models that were accessed through the Datamonkey web server for the detection of selective pressure [47]. The alignment was also analyzed with GARD to rule out recombination break points, and this analysis showed no evidence of any recombination events. To identify past selection on individual codons, the rate of nonsynonymous (dN) and synonymous (dS) substitutions and the dN/dS (ω) ratio at each codon site were calculated with various Codon-based maximum-likelihood methods (FEL, IFEL, and SLAC). Interestingly, no evidence of positive selection on any site was shown by these methods (Table 2 and S6 Fig). The majority of sites showed evidence of negative/purifying selection, while between 13 and 22 sites (depending on which model was used) were found to be evolving at a neutral rate. Interestingly, the mixed effects model of evolution (MEME) identified 16 amino acid sites (S6 Fig) as being under episodic positive selection (0.1 significance level). Among these, two amino acid sites were located at the N-terminus of the STK domain, one amino acid site each in the II, III, VI-a, VI-b, and VII subdomains, two sites each in the VIII, IX, and X subdomains, and three sites in the XI subdomain. These results suggest that, in addition to purifying selection, positive selection has played an important role in evolution of Pto homologs in plants, and thus in the evolution of adaptive disease resistance.

Expression profile of Pto pathway genes
Expression profiles of Pto pathway genes were generated using previously published RNA-seq data from uniformly grown five different pepper accessions [48]. To analyze the expression profiles of pepper PLPK genes among different accessions, heatmap and hierarchical clustering based on the mean-centered log (base 2) transformed RPKM values of the PLPK genes were generated (Fig 5). Based on expression pattern, pepper PLPK genes clustered into three groups; C1, C2, and C3 (Fig 5). In group C1, PLPK1, PLPK12-PLPK13, and PLPK16-PLPK20, which were relatively highly expressed, clustered together. Interestingly, at least one gene from each pepper PLPK subclass was observed in C1, with the exception of subclass VIII. All of the subclass VIII members (PLPK3-PLPK7), and PLPK22 and PLPK23, which belong to subclass V and VII respectively, were clustered in group C2 and were expressed at moderate levels. The Table 2. Positive selection analysis of the STK domain of Pto-like genes from various plants.

Model
Neutrally evolving sites Positively evolving sites Negatively evolving sites pepper PLPK genes PLPK2, PLPK8-PLPK11, and PLPK14-PLPK15 were clustered in group C3, and were expressed at low levels. Interestingly, at least one gene from each subclass, with the exception of subclasses PLPK II and PLPK VIII, was observed in group C3. Many of the pepper Pto pathway genes, including Pti and Prf, also showed moderate to low expression levels. In general, plant R genes exhibit a low level of constitutive expression in either infected or uninfected plants, which is in accordance with their general role in pathogen recognition [54,55]. Taken together, the expression profiles of Pto pathway genes in the five pepper genotypes (C. annuum 'PI201234' and 'YCM334', C. chinense 'Aji Dulce' and 'PI152225', and C. chacoense 'PI260429') were very similar (Fig 5), indicating a highly conserved gene regulatory mechanism for Pto signaling in pepper species.

Discussion
In tomato, the Pto locus consists of a small cluster of five Pto homologs present on chromosome 5, that possibly evolved through sequential gene duplications and deletions [10,34]. The Pto locus and its gene-for-gene role in AvrPto recognition specificity are believed to be conserved in Solanaceae species [19,54]. However, the recognition specificity of Pto is unknown in a number of Solanaceae species, including pepper and potato. Despite the recognition of AvrPto by many plant species, it is unknown whether the recognition is mediated by Pto-like kinases. In addition, no functional Pto-like proteins that recognize AvrPto have been identified in other plant species, with the exception of tomato and tobacco. Our results using a PVXbased system with pepper provide evidence for in vivo recognition of AvrPto. However, we cannot rule out the possibility that the recognition of AvrPto in pepper plants may be due to Pto-like genes and/or other RLKs, until our results are verified using reverse genetic approaches. It is well established that Pto, through interaction with the NB-LRR Prf, mediates recognition of AvrPto/AvrPtoB in tomato [56]. However, the targets of AvrPto/AvrPtoB appear to be the kinase domains of various RLKs [57][58][59][60] rather than the Pto STK domain; Pto is being used as a molecular bait by Prf to interact with pathogen effector proteins [61,62]. AvrPto/AvrPtoB effectors interact with several plant molecules, such as Arabidopsis GTPase RabE [63,64] and the kinase domains of several RLKs, such as CERK1, BAK1, EFR1, and FLS2 [57][58][59][60]65].
To identify possible AvrPto-interacting tomato Pto homologs in pepper, we carried out a genome-wide analysis, which resulted in the identification of 25 full-length Pto-like (PLPK) genes. Most R genes are members of multigene families, and occur as clusters in plant genomes. Gene duplication and diversification are thought to be responsible for the evolutionary expansion of the R gene clusters [5]. In agreement with this, the present and previous phylogenetic analyses of Pto homologs from a number of plant species revealed several subclasses of Pto homologs [27,30,32]. Interestingly, no Pto clade members were found in pepper. Homology searches against the three different pepper genome databases "CM334", "Chiltepin", and "Zunla-1" (http://cab.pepper.snu.ac.kr/), and pepper expressed sequence tags (ESTs) at NCBI, confirmed the absence of tomato Pto homologs (Pto clade) in the pepper genome. Wan et al. [66] failed to amplify true tomato Pto homologs in pepper when using primers designed against the conserved regions of Pto genes. Furthermore, no homologs of the tomato Pto class were detected R gene mapping studies in pepper [67]. With the exception of the Pto clade, Solanaceae PLPKs from pepper and other plant species are well-distributed among the branches of Evolution of the Pto-Like Protein Kinase (PLPK) Family in Pepper the phylogenetic tree, suggesting that PLPKs arose from common ancestral genes prior to the divergence of monocots and dicots. The absence of tomato Pto homologs in pepper further suggests Solanaceae sublineage-specific expansion of the Pto clade. This observation also suggests that the role of Pto genes in pathogen recognition and disease resistance might have evolved during early evolution of Solanaceae lineages [27,35] or, alternatively, has been lost from pepper.
The underlying mechanism that could result in the loss of the Pto clade in pepper and some tomato genotypes [34,35] is not clearly understood. Although some cultivated tomato, S. lycopersicum, varieties show avrPto-specific Pto-mediated Pst resistance, the resistance was introgressed from S. pimpinellifolium [10,36]. Two possible reasons have been suggested to explain the fact that Pto recognition specificity has been retained in some tomato species and entirely lost in closely related species. First, during the domestication of wild S. lycopersicum, an intense genetic bottleneck might have occurred and resulted in the selection of those S. lycopersicum lines that lacked the Pto gene [35]. Second, fitness costs associated with the Pto locus in the absence of Pst with avrPto gene expression might have resulted in selection against S. lycopersicum lines containing the Pto gene [35]. However, near-isogenic lines of S. lycopersicum introgressed lines with and without the Pto locus showed no evidence of a fitness cost associated with Pst resistance [35,68], although a deleterious effect of the Pto locus may be difficult to determine experimentally [35].
Despite extensive studies in tomato, the functional specificity of Pto homologs in other Solanaceae species is not clearly understood. Pto paralogs share 78 to 91% nucleotide identity with tomato Pto, and despite having functional protein kinase activity, none can interact with AvrPto and AvrPtoB effectors [12,69]. Gene duplication and subsequent diversification of the Pto gene family presumably have led to alternative recognition specificities [70]. Pto and Pto homologs vary only in a few amino acid residues, which could cause protein conformational changes that result in altered ability to physically interact with other proteins [36]. The Fen gene, which is one of the five Pto paralogs, is 87% identical to Pto and participates in the same signaling events leading to HR. However, it is activated by a different signal [70,71]. By contrast, LhirPto shows 97% sequence identity with only 17 amino acid variations that distinguish it from Pto, and yet LhirPto does not abolish Pto disease resistance [35]. These observations further indicate that there has been a selective pressure to retain Pto-AvrPto recognition specificity [72].
Diversifying selection or adaptive selection are presumed to play a key role in the evolution of R genes. Adaptive selection is considered the main force that drives evolution of gene regions that function in host-pathogen recognition [5,73]. Interestingly, purifying selection appeared to be the predominant selective pressure on genes that contain STK domains to maintain their ancestral state [32]. However, in the present study, episodic positive selection was identified in the background, suggesting a possible change in selection pressure during evolution. It has been suggested that adaptive evolution often involves episodic bursts of selection localized to a few sites in a gene, and may affect only a subset of lineages in phylogeny [74]. In agreement with this, MEME analysis showed that some codon sites are under episodic positive selection. The activation domain of Pto, the region between amino acids 182 and 211, plays an important role in AvrPto recognition [27]. Interestingly, our results showed that two positions, Ser198 and Ile208, were subjected to episodic positive selection, implying that they play an important role in adoptive evolution. Recent data show that N. benthamiana plants, which have been used as negative controls for the Pto/AvrPto response, can also recognize AvrPto in a gene-forgene manner, and that this recognition is dependent on Prf [19]. Recognition of AvrPto resulting in plant resistance has also been observed in distant taxa, for example in soybean [28].
These observations further suggest that, in addition to the purifying and balancing selection [72], adaptive selection could also have influenced the evolution of the Pto locus.
In summary, nonhost disease response associated with recognition of the AvrPto effector from Pst was observed in pepper genotypes. Recognition of AvrPto was possibly mediated by Pto homologs or other RLKs. We identified 25 pepper PLPK genes through genome-wide analysis and found that they were divided into eight phylogenetic subclasses. The Pto clade represents a gene family of recent origin in Solanaceae species, but the PLPKs represent diverged genes with an ancient common origin. Tomato Pto paralogs are not found in the pepper genome, suggesting that Pto family genes have undergone different evolutionary processes towards adaptive evolution to Pst resistance. Pepper PLPKs were characterized by the presence of highly conserved STK domains. The presence of carbohydrate binding malectin-like domains in the PLPKs suggests a possible role in sugar sensing and disease resistance signaling pathways [75]. Similar expression profiles of Pto pathway genes was observed in all five pepper genotypes. These observations suggest PLPK genes with highly conserved functional and structural domains share a highly conserved gene regulatory mechanism associated with the Pto signaling pathway among the pepper genotypes. Further studies on Pto and PLPKs, and their interaction with AvrPto or other effectors will provide more insight in the process of plantpathogen coevolution. The present genome-wide analyses of pepper PLPK genes provide an important genetic resource for further functional studies to unravel the molecular basis of nonhost disease resistance in pepper.