The type III effectors of Xanthomonas

SUMMARY

A review of type III effectors (T3 effectors) from strains of Xanthomonas reveals a growing list of candidate and known effectors based on functional assays and sequence and structural similarity searches of genomic data. We propose that the effectors and suspected effectors should be distributed into 39 so‐called Xop groups reflecting sequence similarity. Some groups have structural motifs for putative enzymatic functions, and recent studies have provided considerable insight into the interaction with host factors in their function as mediators of virulence and elicitors of resistance for a few specific T3 effectors. Many groups are related to T3 effectors of plant and animal pathogenic bacteria, and several groups appear to have been exploited primarily by Xanthomonas species based on available data. At the same time, a relatively large number of candidate effectors remain to be examined in more detail with regard to their function within host cells.

INTRODUCTION

The genus Xanthomonas consists of a large group of plant pathogenic proteobacteria which, as a group, has considerable agricultural impact. Common diseases found in nature include the vascular wilts, cankers, leaf spots, fruit spots and blights. Despite the wide host range of Xanthomonas, individual strains can be highly restricted to particular host plants. Xanthomonas spp. infect hosts encompassing most of the major groups of higher plants, including 124 monocotyledonous and 268 dicotyledonous plant species (Hayward, 1993; Leyns et al., 1984; Starr, 1983). Many xanthomonads require a type III secretion system (T3SS) for pathogenicity on plant hosts, and the requirement reflects the utilization of T3 effectors to mediate the processes of pathogen adaptation to specific host tissues, species and genotypes. T3 effectors are translocated into the host cells where they interfere with host immunity responses or facilitate nutritional or virulence processes of the pathogen (Buttner and He, 2009). The diversity in tissue and host specificity of members of the genus is also undoubtedly reflected in the diversity of T3 effectors in a given pathovar or species. This review provides a catalogue of T3 effectors and candidate effectors that have been identified within Xanthomonas species, and is written as a prelude to a new wave of genomic sequence data and biochemical insight regarding the identification and function of T3 effectors in many plant pathogenic strains, pathovars and species. We aim to highlight the diversity of effectors in Xanthomonas species, recognizing the tremendous advances made in understanding T3 effector function in other bacterial disease complexes, including those of Xanthomonas, which are expertly reviewed in greater detail in this issue and elsewhere (Cunnac et al., 2009; Lindeberg et al., 2009).

CATALOGUE OF XANTHOMONAS EFFECTORS

The number of candidate and known T3 effectors of Xanthomonas stands at nearly 40 groups of related T3 effectors, as proposed here (Table 1). A more extensive list of candidates and updates can be found at http://www.xanthomonas.org/t3e.html, and information regarding updates or corrections to the data can be provided at the site. The effectors are grouped on the basis of sequence relatedness. The candidate status of an effector is based on the lack of functional secretion and/or translocation evidence. We have proposed the adoption of the Xop/Hop/Yop nomenclature, whilst recognizing that some T3 effectors are inexorably linked with the avirulence terminology. Thus, AvrBs1, AvrBs2 and AvrBs3 remain as named. The Xop designation originally was used to refer to proteins secreted in a T3SS‐dependent manner and not strictly in reference to T3 effectors (Noel et al., 2002). XopA was given to the name of the Hpa1 homologue from X. campestris pv. vesicatoria and therefore does not appear here as a T3 effector group (Noel et al., 2002). Hpa1 (also referred to as HpaG) is a harpin‐like protein that is also secreted in a T3SS‐dependent manner, together with the curious pectic lyase‐harpin fusion protein HrpW. All xanthomonads shown to possess a T3SS also have Hpa1/HpaG, but not all strains have a version of HrpW (Kim et al., 2003; Noel et al., 2002; Zhu et al., 2000). Whether harpin‐like proteins (harpin, Hpa1/HpaG and HrpW) function as virulence effectors or as accessory and translocation proteins is an important subject of research. However, the function of the harpin‐like proteins remains unclear and will not be addressed here.

Table 1.

Candidate and known T3 effectors of Xanthomonas.

Torrent of genomic data is likely to expand the repertoire of T3 effectors of Xanthomonas

Candidate T3 effectors, historically, have been identified through the characterization of resistance (R) gene‐mediated responses (2000a, 2000b; Bonas et al., 1989; Ciesiolka et al., 1999; Ronald and Stascawicz, 1988; Rybak et al., 2009; Swords et al., 1996; Whalen et al., 1993; Zhao et al., 2004). Improvements in T3 effector assays that are independent of endogenous R gene responses have greatly facilitated the analysis of T3 effector translocation, and now many of the T3 effector candidates have supporting evidence for T3SS‐dependent secretion and translocation to plant cells based on assays of avirulence reporter fusions (Greenberg and Vinatzer, 2003; Guttman et al., 2002) and/or calmodulin‐dependent adenylate cyclase (Cya) activity assays (Casper‐Lindley et al., 2002; Schechter et al., 2004; Sory et al., 1995). Consequently, additional candidate T3 effectors from Xanthomonas, which do not have associated host hypersensitive reactions (HRs), have come from functional assays of T3SS‐dependent expression, secretion and screens of avirulence protein fusions on a genomic scale (2003, 2002; Roden et al., 2004a; Thieme, 2008). More recently, candidates have been identified from genomic sequence data and subjected to further validation criteria (Collmer et al., 2009). The annotated and assembled genomic sequences of eight strains of Xanthomonas are currently available, representing four pathovars (Lee et al., 2005; Salzberg et al., 2008; da Silva et al., 2002; Thieme et al., 2005). The genomic sequences of three closely related strains of X. oryzae pv. oryzae and X. campestris pv. campestris, respectively, have been determined (Ochiai et al., 2005; Qian et al., 2005; Salzberg et al., 2008; Vorholter et al., 2008). The T3 effector content of the different strains within a pathovar is similar with a few exceptions, and the data from a representative strain for each pathovar are presented (Table 1). In addition, nearly completed or draft sequences are now or will soon be available for nearly double the number of currently sequenced strains. Here, the draft sequences of three distinct strains were examined for candidate T3 effector genes on the basis of similarity searches, and were included as representative of the data expected in the near future. Strain AXO1947 is an atypical X. oryzae pv. oryzae originally isolated in Africa (Gonzalez et al., 2007; Gu et al., 2004). Xanthomonas vasicola pv. musacearum (Xvm and, formerly, X. campestris pv. musacarum) is the causal agent of xanthomonas wilt on banana (Aritua et al., 2008). Xanthomonas perforans 91–118 is one of a group of Xanthomonas species that cause bacterial spot of tomatoes and pepper (Jones et al., 2000). Individual candidate effectors, derived from related sequence data, have been characterized (Jiang et al., 2008; Xu et al., 2008; Wang et al., 2007). Furutani et al. (2009) recently provided T3SS‐dependent translocation evidence for 16 candidate T3 effectors of X. oryzae pv. oryzae, including nine members of new T3 effector groups in Xanthomonas (Table 1, XopK, XopR, XopR‐XopW and XopAB; Furutani et al., 2009).

Subgrouping of T3 effectors

The 40 groups of Xops are likely to underestimate the variety of host physiological functions that are targeted by the T3SS virulence pathway in strains of Xanthomonas, as a number of groups have multiple members. Multimembered groups occur both as groups with relatively divergent members based on sequence relatedness, as found in the XopJ group, and as groups with highly conserved sequence relatedness but apparent proliferation of variants based on repetitive domains, as found in the large AvrBs3 family. The XopJ group, which is part of the broader YopJ family of related effectors in many animal and plant pathogenic proteobacteria with T3SS virulence systems, includes XopJ, AvrRxv, AvrBsT and AvrXv4. The XopJ group may encode members with variant enzymatic activities and target multiple host physiological pathways (Mukherjee et al., 2007). Variations of the large AvrBs3 family, which is also referred to in the literature as the AvrBs3/PthA family, and the transcription activation‐like (TAL) T3 effectors occur both between strains as well as within individual strains of Xanthomonas (Schornack et al., 2006). Individual strains of X. oryzae pv. oryzae are known to contain 19 genes for TAL effectors, and one strain of X. oryzae pv. oryzicola contains 28 TAL effector genes (Ochiai et al., 2005; Salzberg et al., 2008; A. Bogdanove, Iowa State University, Ames, IA). Different TAL effectors target different host genes and are likely to affect different host physiological processes (White and Yang, 2009). Other effector groups with two or three related genes in individual strains include XopE, XopF, XopJ, XopX, XopAD and XopAE (Table 1, http://www.xanthomonas.org/t3e.html). Some of the groups have been further divided into subgroups based on phylogenetic data and, here, these subgroups have been assigned numbers to denote phylogenetic families within a group (e.g. XopF1 and XopF2, Table 1). Phylogenetic analyses remain to be performed on all groups, and additional groups are likely to be proposed. At the opposite end of the spectrum, some groups are defined by members that have been found in relatively few strains or pathovars. Effectors with limited distribution are XopB, XopD, XopT, XopU, XopW, XopAC, XopAF, XopAH, XopAJ, and XopAK.

Genetic flux between pathogenic bacteria

The T3 effectors of Xanthomonas reflect the robust exchange of genetic material between bacteria. Slightly less than one‐half of the groups (15) are related to candidate T3 effectors of Pseudomonas syringae pathovars to varying degrees (Table 1, http://www.xanthomonas.org/t3e.html), and many have related members in Ralstonia solanacearum and Acidovorax avenae ssp. citrulli, the members of which, with some exceptions, are not presented here. Eleven groups have a homologue or, at least, partial relatedness in every strain of Xanthomonas that has been sequenced (Table 1). However, the homologues may not be functional in all strains. For example, the lone XopF open reading frame (ORF) from X. axonopodis pv. citri lacks the N‐terminal coding sequence (Table 1, XAC2785) and is likely to represent a truncated gene. A number of genes for T3 effectors, as predicted from genomic sequences, require re‐annotation because of improper or unknown start sites, as noted for XopE members (Thieme et al., 2007). The T3 effectors of Xanthomonas also include chimeric proteins, which can confuse classification schemes. XAC3230 is the sole member of the proposed XopAI group and contains a 43‐amino‐acid N‐terminal domain that is also found in the N‐termini of effectors in groups XopE and XopJ (Fig. 1). The XAC3230 N‐terminal domain also contains a myristoylation motif, which was previously identified in several T3 effectors of P. syringae, indicating that effectors with this N‐terminal domain are targeted to host cellular membranes. Evidence for the myristoylation of XopJ, XopE1 and XopE2 of X. campestris pv. vesicatoria has been demonstrated by the loss of membrane localization on mutagenesis of the conserved G2 residue in each of the proteins (Thieme et al., 2007). Thus, N‐terminal domain shuffling may not only be involved in the recruitment of new effectors, but also may allow modifications to effector targeting within the host cell (Stavrinides et al., 2006).

Figure 1.

A variety of Xop proteins of different groups share N‐terminal domains. The highlighted text indicates related Xop proteins based on their C‐terminal regions. Yellow, XopAI group; grey, XopE group and related Pseudomonas syringae T3 effector HopPmaB (AAL84240); blue, XopJ‐related proteins. Asterisk indicates the G residue for myristoylation (Nimchuk et al., 2000).

New classes of T3 effectors await biochemical analysis

Functional analyses of candidate T3 effector translocation and effects on host physiology involving Xanthomonas species have not kept pace with the genome sequence analyses, and some effector groups have little or no experimental support other than evidence based on relatedness to effectors of other pathogens. These groups include the aforementioned XopL, XopAD, XopAE and XopAI. XopAI, in addition to the high similarity to the N‐terminal sequences from known T3 effectors, is weakly related to HopO1 (Almeida et al., 2009). However, both XopAI and HopO1 share a VIP2 ADP‐ribosylation motif, indicating that the proteins may have conserved structural features. XopL and XopAE both include members with leucine‐rich repeats (LRRs), a motif also present in XopAC. XopAE represents the HpaF proteins, the genes for which are found within the T3SS (Hrp) gene clusters of most strains, with the exception of X. campestris pv. campestris (Kim et al., 2003). The genes for HpaF are co‐regulated with T3SS genes, and mutations in hpaF from X. axonopodis pv. glycines cause reduced virulence (Kim et al., 2003). No evidence has been reported for T3SS‐dependent secretion of HpaF. A member of XopG was originally identified as part of the putative HrpX regulon in X. campestris pv. campestris ATCC33913 (da Silva et al., 2002). Members are related to the predicted T3 effectors HopH1 and HopAP1 of P. syringae species (Almeida et al., 2009; Feil et al., 2005) and RSp0572 of R. solanacearum GMI1000 (Salanoubat et al., 2002). The XopAD group is noteworthy and solely based on a multigene family of putative T3 effectors from R. solanacearum, which are known as the SKWP proteins because of the presence of a 42‐amino‐acid repeat with the short consensus amino acid segment SKWP (Mukaihara and Tamura, 2009). Related SKWP proteins are encoded by a broad range of plant pathogenic bacteria, including one strain of P. syringae (P. syringae pv. phaseolicola 1448A). All of the strains discussed here have members of XopAD with the exception of X. campestris pv. campestris.

ELICITOR‐MEDIATED IMMUNITY

Many T3 effectors in Xanthomonas undoubtedly function to suppress so‐called pathogen‐associated molecular pattern (PAMP)‐triggered immune (PTI) responses, similar to the suppression observed by the T3 effectors of other plant pathogenic bacteria (Boller and Felix, 2009; Boller and He, 2009; Hotson and Mudgett, 2004; 2009, 2008; Metz et al., 2005). A variety of PTI responses have been characterized, and many are known to be triggered by pathogen recognition receptors (PRRs) on the plant cell surface. PRRs include the receptor‐linked kinase (RLK) family, the best studied of which are the PAMP receptors for flagellin and elongation factor Tu (EF‐Tu), namely FLS2 and EFR. PRRs, on perception of the elicitors, signal a variety of host defence responses, which have been shown to be suppressed by a variety of T3 effectors from P. syringae species at a variety of steps in the signalling pathways (Boller and He, 2009). The RLK family includes Xa21, which was the first R gene to be cloned from rice, and provides broad resistance against many strains of X. oryzae pv. oryzae, the agent of bacterial blight (Song et al., 1995). Evidence indicates that the Xa21 RLK serves as a PAMP signalling function, similar to FLS2 and ERF, with the exception of the specificity of the elicitor. The elicitor for Xa21‐mediated resistance is hypothesized to be an extracellular modified peptide, and, perhaps worth re‐emphasizing here, not the product of the T3SS pathway, despite the fact that the hypothesized elicitor has historically been referred to as AvrXa21 (Lee et al., 2006). A second R gene, Xa26, which gives broad resistance to X. oryzae pv. oryzae, also encodes an RLK (Sun et al., 2004). With regard to PAMPs, flagellin and EF‐Tu, the cognate elicitors for FLS2 and EFR, respectively, are produced by species of Xanthomonas, and pretreatment of plants with the conserved flg22 portion of bacterial flagellin enhances the resistance of Arabidopsis to X. campestris pv. campestris (Sun et al., 2006). Additional compounds have been purified from Xanthomonas that elicit defence‐related responses, including peptidoglycan polymers and lipopolysaccharides (LPSs), and other compounds may have activity in PTI in response to infections by Xanthomonas species, particularly in specific hosts (Erbs et al., 2008; Keshavarzi et al., 2004; Silipo et al., 2005).

Suppression of elicitor‐mediated immunity

Xanthomonas species have been known for some time to suppress host defence responses in a T3SS‐dependent manner and, more recently, specific T3 effectors have been shown to suppress defence responses and defence‐related gene expression (Brown et al., 1995; Keshavarzi et al., 2004). The presence of elicitation‐competent flagellin genes in X. campestris pv. campestris has been shown not to control the degree of virulence, and the absence of FLS2, the RLK receptor for flg22, also has no effect on disease, indicating that X. campestris pv. campestris suppresses PTI as mediated by FLS2 and EFR (Sun et al., 2006). The T3SS of X. oryzae pv. oryzae has been shown to be required for the suppression of rice defence responses that are elicited by the application of extracellular degradative enzymes, indicating that elicitors of PTI include products of microbial enzymes acting on host substrates (Jha et al., 2007). At the same time, Xa21/Xa26‐mediated resistance responses, as broad resistances in rice, are obviously not suppressed by many extant strains of X. oryzae pv. oryzae. If X. oryzae pv. oryzae strains can suppress PRR‐mediated PTI, in general, specific PRRs have apparent signalling pathways that circumvent T3 effector function.

Specific T3 effectors of X. campestris pv. vesicatoria have also been associated with the suppression of various PTI and host defence responses. XopX was identified in a search for a nonhost avirulence gene and was found to increase the susceptibility of Nicotiana benthamiana to both Xanthomonas and Pseudomonas species (Metz et al., 2005). The T3 effectors XopN and XopD have also been shown to reduce PTI and host developmental pathways that may play a role in host defence (2009, 2008). XopD suppresses the expression of a variety of defence‐ and senescence‐related genes in tomato and is associated with lower levels of salicylic acid (SA) and delayed symptom development in infected tissue (Kim et al., 2008). XopN also suppresses defence‐related gene expression in tomato (Kim et al., 2009); in a heterologous assay using P. syringae, XopN suppressed callose deposition in both tomato and Arabidopsis. XopJ has recently been demonstrated to suppress defence‐related callose deposition and, more specifically, host protein extracellular secretion (Bartetzko et al., 2009).

Suppressors of effector‐triggered immunity

T3 effectors are best known for specific T3 effectors that are recognized by host defence surveillance systems and elicit a rapid HR, which has been defined as effector‐triggered immunity (ETI) to distinguish it from PTI‐triggered immunity. T3 effectors are also known to suppress R gene‐mediated defence, and some evidence suggests that some T3 effectors of Xanthomonas are involved in ETI suppression (Rosebrock et al., 2007). AvrBsT, a member of the XopJ/YopJ family, triggers an HR in a single landrace of Arabidopsis with a recessive mutation in a gene for carboxylesterase, which was named SUPPRESSOR OF AVRBST‐ELICITED RESISTANCE1 (SOBER). A model has been proposed that, in the absence of esterase activity, AvrBsT can suppress recognition by a host R gene product. The robust R gene responses to AvrXa10 and AvrXa7 of X. oryzae pv. oryzae in rice cultivars containing the R genes Xa10 and Xa7, respectively, are suppressed when the corresponding genes for the effectors are introduced into the closely related pathogen X. oryzae pv. oryzicola.

VIRULENCE AND DISEASE SYMPTOMATOLOGY ASSOCIATED WITH T3 EFFECTORS

All strains of Xanthomonas that possess a consensus Xanthomonas T3SS pathway are essentially nonpathogenic if the T3SS pathway is inactivated, indicating that, at least in toto, the T3 effectors have critical functions in facilitating host parasitism. At the same time, mutations of individual effector genes often do not result in a change in the virulence phenotype or have a relatively small effect, which may reflect the aspects of the experimental context and/or redundancy of effector function (Castaneda et al., 2005). However, whether suppressing elicitor‐triggered immunity or providing some other beneficial function, a growing number of T3 effectors of Xanthomonas have been associated with enhanced virulence of strains and aspects of disease symptomatology (Table 1). AvrBs2 was one of the first effectors to contribute to virulence, which, in this example, is based on the ability to support increased numbers of bacteria within host tissue (Gassmann et al., 2000; Kearney and Staskawicz, 1990). Curiously, the gene for AvrBs2 is found in all of the strains represented here and has only been shown to contribute to virulence in X. campestris pv. vesicatoria. Four T3 effectors of X. campestris pv. vesicatoria (AvrBs1–4) were all found to have varying degrees of contributions to pathogen fitness in field situations (Wichmann and Bergelson, 2004). An extreme requirement for an individual T3 effector is found in the diseases caused by strains of both X. axonopodis pv. citri and X. oryzae pv. oryzae. In the last two species, mutations in single TAL effector genes (pthA or pthXo1, respectively) can have a debilitating effect on both the ability of the pathogen to spread and to elicit disease symptomatology (Swarup et al., 1991; Yang and White, 2004). The TAL effectors constitute a large family of closely related T3 effectors and many affect virulence and/or disease symptomatology. PthA, which is the critical TAL effector from X. axonopodis pv. citri, is required for host cell hypertrophy and high bacterial populations (Swarup et al., 1991). The phenotype is also observed on in planta expression of the gene and in the absence of bacteria (Duan et al., 1999). A similar behaviour is observed with the gene for AvrBs3 from X. campestris pv. vesicatoria (Marois et al., 2002). Host cell hypertrophy is not observed in rice. However, lesions, which in the rice system, occur as chlorotic and desiccated leaf tissue, are severely limited in the absence of a major TAL effector for virulence (Bai et al., 2000; Yang and White, 2004; Yang et al., 2000). TAL effectors of X. campestris pv. malvacearum and X. oryzae pv. oryzae are associated with increased water soaking, a phenotype often associated with bacterial infections of plants. One benefit of TAL effectors to bacteria may be an enhancement of the release of bacteria from infected tissue. Increased numbers of bacteria in leaf washings were demonstrated for X. campestris pv. malvacearum with Avrb6 in cotton and X. axonopodis pv. citri with PthA in citrus when compared with tissue infected by strains without effectors (Yang et al., 1994). In pepper (AvrBs3) and cotton (Avrb6), significant increases in bacterial populations have not been detected in the presence of effectors. However, enhanced release could have a fitness benefit in field infections, and fitness benefits to AvrBs3 were observed in field experiments (Wichmann and Bergelson, 2004). The importance of TAL effector function in rice is supported by the deployment of recessive resistance (Lee et al., 2003). The R gene xa13 specifically interferes with the function of the TAL effector PthXo1, indicating that interference in TAL effector function has proven to be effective in providing resistance against bacterial blight of rice (Yang et al., 2006).

Both XopD and XopN of X. campestris pv. vesicatoria have been shown to contribute to virulence and disease symptomatology in tomato (2009, 2008; Roden et al., 2004a). Plants infected with strains without XopD show greater leaf tissue necrosis, indicating that XopD is involved in the suppression of host‐mediated tissue degeneration, which may represent a defence reaction to infection (Kim et al., 2008). Strains without XopN, however, possess lower bacterial populations and less necrosis and senescence at later time points after inoculation (Kim et al., 2009). Loss of the XopN homologue in X. campestris pv. campestris also results in reduced disease in radish, and genes for members of the family are widely distributed in Xanthomonas (Jiang et al., 2008) Another widely distributed T3 effector that affects host tissue responses is early chlorosis factor (ECF), which was labelled XopAA (Morales et al., 2005). ECF was isolated by a complementation strategy, whereby strains with a reduced ability to induce chlorosis were converted to a greater ability to induce chlorosis by the transfer of the gene for ECF. ECF had no measurable effect on bacterial populations. Other candidate T3 effector groups with members that have reported effects on virulence and disease symptomatology include XopJ (AvrXv4; Roden et al., 2004b), XopX (Metz et al., 2005), XopAE (HpaF; Kim et al., 2003) and XopAH (AvrXccC; (Wang et al., 2007).

BIOCHEMICAL AND EXPERIMENTAL CLUES TO T3 EFFECTOR FUNCTION

A variety of putative structural motifs contained in the primary structure of T3 effectors have provided insights into the putative biochemical function of T3 effectors. Fifteen of the effector classes of Xanthomonas have structural motifs that have provided clues to the function of some effectors. Twelve of the 15 indicate the possibility of enzymatic function (Table 2). The XopJ (YopJ) group has been reviewed in considerable detail (Mukherjee et al., 2007; Hotson and Mudgett, 2004). Others that have received less focus in Xanthomonas include the two groups (XopAC and XopAH) with the recently described Fido/AvrB domain (Kinch et al., 2009), XopAI (XAC3230) containing a VIP2 actin ADP‐ribosylation domain (Han and Tainer, 2002; Han et al., 2001) and XopG with a zinc protease motif similar to the active site of clostridial neurotoxins (Lalli et al., 2003; Turton et al., 2002) (Table 2). Considerable progress into the specific functions of individual T3 effectors from Xanthomonas and their interaction with host plants has also provided insight into T3 effector function, and highlights from some of the Xop groups are provided below.

Table 2.

Evidence for biochemical and biological functions of T3 effectors from Xanthomonas.

Effector group*	Biochemical or structural motifs†	Host factor‡	Reference
AvrBs2	Glycerolphosphoryl diester phosphodiesterase		Kearney and Staskawicz, 1990; Swords et al., 1996
AvrBs3/PthA	Site‐specific DNA binding; nuclear localization; transcription activation domain	Os8N3 (r); OsTFX1; OsTFIIAγ1; upa20 (p)	Van den Ackerveken et al., 1996; Kay et al., 2007; Romer et al., 2007; Sugio et al., 2007; Yang et al., 2006; Zhu et al., 1998
XopC	Haloacid dehalogenase hydrolase and phosphoribosyl transferase domain		Noel et al., 2003
XopD	Cysteine SUMO C48 protease; DNA binding; nuclear localization; EAR motif		Chosed et al., 2007; Hotson et al., 2003; Kim et al., 2008; Noel et al., 2002
XopE (HopX)	Transglutaminase family		Nimchuk et al., 2007; da Silva et al., 2002; Thieme et al., 2007
XopG (HopH1 and HopAP1)	M27 zinc protease (clostridial toxin)		Ochiai et al., 2005; Salzberg et al., 2008; da Silva et al., 2002
XopJ (HopJ1) [AvrBsT, AvrXv4, AvrRxv]	Acetyltransferase, C55 cysteine ubiquitin‐like protease	14‐3‐3 protein (t)	Bartetzko et al., 2009; Noel et al., 2003; Roden et al., 2004b; Thieme et al., 2007; Whalen et al., 2008
XopH (HopAO1)	Tyrosine phosphatase		Bretz et al., 2003; Espinosa et al., 2003
XopN (HopAU1)	ARM/HEAT repeat	TARK1 (t), 14‐3‐3 (t)	Kim et al., 2009; Roden et al., 2004a
XopQ (HopQ)	Nucleoside hydrolase		Wei et al., 2007
XopAA (HopAE1)	Colicin Ia C8 and C9 domains; ECF		Morales et al., 2005
XopAC	FIDO domain (AMPylation) PF02661; LRR, VHR		Kinch et al., 2009; Xu et al., 2008
XopAH	FIDO (AMPylation) PF02661		Kinch et al., 2009; Wang et al., 2007
XopAI (HopO1)	VIP2; ADP‐ribosyltransferase		da Silva et al., 2002
XopAJ [AvrRx]	Thiol protease, ATP/GTP binding		Zhao et al., 2004

^* Names of related effector groups from Pseudomonas syringae are given in parentheses, and names of other Xanthomonas effectors in the group are given in square brackets.

^† Abbreviations (noted characteristics and inclusion do not imply confirmation of activity or function): ARM/HEAT, armadillo/Huntington, elongation factor 3, PR65/A, TOR domain; EAR, ethylene receptor factor‐associated amphiphilic repression; ECF, early chlorosis factor; FIDO, fic, doc, AvrB domain; LRR, leucine‐rich repeat; SUMO, small ubiquitin modifier protein; VHR, vascular hypersensitivity in Arabidopsis landrace Col‐O.

^‡ Host genes or proteins that are associated with effector by binding, genetic or other experimental evidence. Abbreviations: (p), pepper; (r), rice; TARK, tomato atypical receptor kinase; (t), tomato.

AvrBs3

The AvrBs3 group represents a remarkably large family of closely related T3 effectors of Xanthomonas, which, in toto, are referred to here as the TAL effector family. A variety of TAL effectors are associated with biological effects, including virulence of the pathogen and disease symptoms (Table 3). All TAL effectors have a characteristic overall structure with the most distinguishing feature of the individual effectors being a central repetitive region containing varying numbers of near‐identical repeats of 34 or 35 amino acids. The N‐terminal and C‐terminal portions of the proteins are highly conserved, and differences between effectors, in terms of their biological activity, are related to the particular repetitive region. For example, the type member AvrBs3 has 17.5 repeats, whereas AvrXa7 from X. oryzae pv. oryzae has 26.5 repeats (Bonas et al., 1989; Yang et al., 2000). Shuffling of specific repetitive regions can disrupt avirulence specificity (Herbers et al., 1992; Yang et al., 2005), and, in general, the repetitive regions can be swapped between effectors, and the specific activity associated with the effector will be associated with whichever has the original repetitive region (Yang and White, 2004; Yang et al., 2000). The structural role played by the repetitive region in the function and specificity of individual effectors is unknown, but has been proposed to determine the specific DNA sequence that is bound by the protein within the promoter of targeted host genes (Kay et al., 2009; Romer et al., 2009). The conserved C‐terminal portion of TAL effectors contains potent nuclear localization signal (NLS) motifs, as well as a robust acidic transcription activation domain (AD), both of which are essential for pathogen virulence and associated effects on host disease symptoms (Szurek et al., 2001; Van den Ackerveken et al., 1996; Yang and Gabriel, 1995; Yang et al., 2000; Zhu et al., 1999). In cases in which there is a corresponding R gene, the NLS and AD are also required for R gene‐mediated ETI (Gu et al., 2005; Van den Ackerveken et al., 1996; Yang et al., 2000; Zhu et al., 1998). TAL effectors induce the expression of specific host genes which, in the case of virulence, facilitate pathogen spread and disease (Yang et al., 2006; Sugio et al., 2007; Kay et al., 2007).

Table 3.

Transcription activation‐like (TAL) T3 effectors of Xanthomonas with associated phenotypes.

Effector	Tag*	Species†	Phenotypes‡, host target genes, (host species)§	Reference
Apl1	BAA37119	Xac, NA‐1	H (ci)	Kanamori and Tsuyumu, 1998
Apl2	BAA37120	Xac, NA‐1	Hw (c)	Kanamori and Tsuyumu, 1998
Avrb6	AAB00675	Xcm, XcmH	W, R, b6 (co)	De Feyter et al., 1993; Yang et al., 1994
AvrBs3	P14727	Xcv, 71‐21	H, Bs3 (p), upa20 (p)	Bonas et al., 1989; Kay et al., 2007; Romer et al., 2007
AvrBs4	CAA48680	Xcv, 82‐8	Bs4 (t)	Bonas et al., 1993; Schornack et al., 2004
AvrHah1	ABP97430	Xg XV444	34 and 35¶, W (in pepper in Xcv), Bs3	Schornack et al., 2008
AvrXa7	AAF98343	Xoo, PXO86, T7174, KXO85	Vma, Xa7 (r)	Bai et al., 2000; Hopkins et al., 1992; Ochiai et al., 2005; Yang and White, 2004; Yang et al., 2000
AvrXa10	AAA92974	Xoo, PXO86	Xa10 (r)	Hopkins et al., 1992
AvrXa27	AAY54168	Xoo, PXO99A, others	Xa27 (r)	Gu et al., 2005
Hax2	AAY43358	Xca, 5	35¶, CN	Kay et al., 2005
Hax3	AAY43359	Xca, 5	Bs4	Kay et al., 2005
Hax4	AAY43360	Xca, 5	CN, Bs4	Kay et al., 2005
HssB3.0	BAF46269	Xac, KC21	SHV	Shiotani et al., 2007
PthA‐KC21	BAF46271	Xac, KC21	H	Shiotani et al., 2007
PthA	AAC43587	Xac	Vma,H,R	Swarup et al., 1991
PthA*	ABO77780	Xac, Xc270 (A*)	H	Al‐Saadi et al., 2007
PthAW	ABO77779	Xac, X0053 (AW)	H	Al‐Saadi et al., 2007
PthB	NP_942641	Xcau, B96	H	Al‐Saadi et al., 2007
PthC	ABO77782	Xcau, C340	H	Al‐Saadi et al., 2007
PthN	AAB69865	Xcm, XcmN	W	Chakrabarty et al., 1997
PthXo1	ACD58243	Xoo PXO99A	Vma, Os8N3 (r)	Yang et al., 2006
PthXo2	AAS46026	Xoo PXO71, T7174	Vma	Yang and White, 2004
PthXo3	AAS46027	Xoo, PXO61	Vma	Yang and White, 2004
PthXo6	ACD58920	Xoo, PXO99A	Vm, OsTFX1 (r)	Sugio et al., 2007
PthXo7	ACD57198	Xoo, PXO99A	Vw*, OsTFIIAγ1 (r)	Sugio et al., 2007
PthXo8	ACD60557	Xoo, PXO99A	Vm, small RNA processing (r)	B. Yang and F. F. White, unpublished results

^* Accession number for amino acid sequence.

^† Pathovar designations used at the time of publication. Pathovar abbreviations: Xac, X. axonopodis pv. citri; Xca, X. campestris pv. armoraciae; Xcau, X. citri pv. aurantifolii; Xcm, X. campestris pv. malvacearum; Xcv, X. campestris pv. vesicatoria; Xg, X. gardeneri; Xoo, X. oryzae pv. oryzae.

^‡ Abbreviations: CN, enhanced chlorosis and necrosis of tissue; H, hypertrophy, cell enlargement; Hw, weak hypertrophy; R, increased bacterial numbers released from lesion; SHV, suppression of hypertrophy and virulence; Vm, moderate virulence effect; Vma, major enhancement of virulence on the basis of increased bacterial populations in plant; Vw, weak virulence; W, increase in water‐soaked appearance of tissue. In Xoo, an increase in virulence is also accompanied by increased lengths of leaf lesions.

^§ Host species abbreviations: (c), citrus; (co), cotton; (p), pepper; (r), rice; (ra), radish (Rhapanus sativus); (t), tomato.

^¶ 34, repetitive domain consisting of individual 34 amino acid repeats; 35, individual 35 amino acid repeats.

Based on the recent results with AvrBs3, TAL effectors function as transcription factors that bind specific DNA elements in the host nucleus and induce the expression of targeted host genes (Kay et al., 2007). AvrBs3 specifically binds to a DNA sequence in a number of genes, one of which, upa20 (up‐regulated by AvrBs3 20), encodes a basic helix–loop–helix transcription factor, which is responsible for the hypertrophy of the host cells. Ectopic expression of upa20 is sufficient to cause host cell hypertrophy. The same AvrBs3‐specific sequence is found in the promoter of the Bs3 gene for resistance, resulting in the expression of Bs3 in the presence of AvrBs3 (Romer et al., 2009). Changes in the TAL effector binding element can result in a loss of TAL effector responsiveness. Specifically, the two alleles Bs3 and Bs3‐E of pepper differ by a 13‐bp deletion (Romer et al., 2007). Each element is bound by a corresponding variant, in this case AvrBs3, and a rearranged repetitive domain derivative named AvrBs3Δrep16, respectively, and the binding elements overlap by 11 bases (Romer et al., 2009). Thus, the repetitive domain is probably involved in DNA binding specificity, and other TAL effectors may bind specific DNA elements in the respective host genomes.

The major TAL effector PthXo1 directs the expression of the rice gene Os8N3, a member of the nodulin3 protein family (Yang et al., 2006). Failure to induce Os8N3 results in resistance to infection unless an alternative TAL effector gene is present (Yang et al., 2006). The recessive resistance gene xa13 represents a set of Os8N3 alleles that do not respond to compatible strains of the pathogen and, specifically, to PthXo1 (Chu et al., 2006; Yang et al., 2006; Yuan et al., 2009). Silencing of Os8N3 expression also results in resistance to strains of X. oryzae pv. oryzae that rely, specifically, on PthXo1 for virulence (Chu et al., 2006; Yang et al., 2006). However, Os8N3 has an apparent function in pollen development and, possibly, in other aspects of fertility, and resistant plants produced by gene silencing suffer from low fertility (Chu et al., 2006). Thus, the xa13 alleles represent alterations that prevent or reduce sensitivity to PthXo1, whilst maintaining sufficient Os8N3 expression patterns for development. Os8N3 has been proposed to represent a class of polymorphic loci called susceptibility, or S, genes, with, in this case, PthXo1 as the cognate effector, in analogy with the R genes (Yang et al., 2006). S genes are host loci where variations occur within the targets of the type III effectors, whereas many R genes have been proposed to guard or mimic type III effector targets (Hogenhout et al., 2009). Alternative TAL effectors exist that overcome xa13 resistance, and Os8N3 induction is not involved in their ability to induce a state of host susceptibility (Yang et al., 2006). Three of these effectors are AvrXa7, PthXo2 and PthXo3, and the analyses of the host responses and rice varieties are predicted to reveal an assortment of S genes, as well as R genes, and possible corresponding recessive resistance alleles that have arisen in rice evolution corresponding to TAL effectors (Yang et al., 2006).

Three TAL effectors from a single strain of X. oryzae pv. oryzae, in addition to PthXo1, have been shown to provide contributions to virulence, and two are known to be associated with the elevated expression of two host genes distinct from Os8N3 (Table 3). PthXo6 directs the elevated expression of a host gene named OsTFX1, which encodes a member of the large bZip family of transcription factors in plants (Sugio et al., 2007). A mutant of strain PXO99^A in pthXo6 suffered a reduction of virulence, although not to the same extent as the loss of genes for PthXo1 or any of the other major TAL effectors. In turn, ectopic expression of OsTFX1 in rice abrogated the need for PthXo6 in the pathogen, but did not make the plants more susceptible to PXO99^A (Sugio et al., 2007). The function of OsTFX1 in a normal plant is unknown. bZip transcription factors are involved in the regulation of many cellular processes, and OsTFX1 may induce or repress one or more downstream genes that are involved in host susceptibility or resistance, respectively. Ectopic expression of the gene, other than compensating for the presence of pthXo6 in the pathogen, did not cause any phenotypic abnormalities in the transgenic plants (Sugio et al., 2007). All strains of X. oryzae pv. oryzae that were examined were capable of inducing OsTFX1, supporting the hypothesis that the ability to induce OsTFX1 provides an important component to bacterial fitness in the environment.

PthXo7 is associated with the elevation of OsTFIIAγ1 and has only been found in X. oryzae pv. oryzae strain PXO99^A (Sugio et al., 2007). Loss of PthXo7 from PXO99^A has no measurable effect on strain virulence in susceptible rice plants. A small effect on disease symptomatology was found on transfer to strain PXO86, and the resulting strain was tested on rice containing the recessive resistance gene xa5, suggesting that PthXo7 may be an adaptation to specific host genotypes (Sugio et al., 2007). PXO86 is incompatible on rice containing xa5, which, like OsTFIIAγ1, encodes a form of the small subunit of TFIIAγ (Iyer and McCouch, 2004; Jiang et al., 2006). The xa5 version does not support the growth of PXO86, and the elevation of OsTFIIAγ1 was hypothesized to supply a form of TFIIAγ that facilitates bacterial growth and thus overcomes xa5 resistance. However, pthXo7 does not confer on PXO86 the same level of growth on xa5 plants in comparison with PXO99^A, a strain compatible on xa5, nor do PthXo6‐deficient mutants of PXO99^A suffer appreciable loss of virulence on plants with xa5 (Sugio et al., 2007). Therefore, PthXo7 does not appear to be the sole or even the principal reason for the ability of PXO99^A to cause significant disease in the presence of xa5 (Sugio et al., 2007). A third TAL effector has been identified recently in PXO99^A that provides a significant contribution to virulence and lesion formation, and preliminary findings have indicated that the effector targets aspects of host small RNA biosynthesis (B. Yang and F. F. White, unpublished results). Thus, X. oryzae pv. oryzae appears to have capitalized on the TAL effector family for multiple pathways to alter host physiology for the benefit of the pathogen.

TAL effector‐mediated resistance has revealed novel defence adaptations in plant defence pathways. In addition to the recessive resistance afforded by xa13, which is a change in susceptibility, plants have exploited the function of TAL effectors by evolving promoters of genes that trigger a resistance response that contains TAL effector recognition features. In the case of the rice R gene Xa27, resistance is mediated by the elevated expression of Xa27, resulting in host cell death and resistance to disease (Gu et al., 2005). Xa27 induction is strictly dependent on a TAL effector named AvrXa27, the gene for which is present in many strains of X. oryzae pv. oryzae (Gu et al., 2005). Susceptible rice cultivars have an identical ORF for Xa27, but the allele does not respond to AvrXa27‐mediated induction, presumably as a result of sequence variations that are found in the promoter regions of the respective alleles. Similar to Xa27, the resistance in pepper caused by Bs3, the gene is induced in an AvrBs3‐dependent manner. Production of the Bs3 protein, which has structural features of a flavin oxidase, promotes the collapse and death of infected tissue (Romer et al., 2007). The analysis of the Bs3 promoter revealed that, indeed, the apparent targets of AvrBs3 for virulence (upa20 and possibly others) and Bs3 share AvrBs3 binding sites (Romer et al., 2009). Xa27 and Bs3 are not related on the basis of sequence comparison, and the question remains as to how many host proteins can function similarly to Xa27 and Bs3. The rice R genes Xa7 and Xa10 are known to require intact TAL effector functions, and their analysis will provide more clues with regard to the diversity of host defence responses to this family of T3 effectors (Gu et al., 2008; Porter et al., 2003; Yang et al., 2000; Zhu et al., 1998).

At least one case, a TAL effector, triggers resistance in the presence of a better known type of R gene product. AvrBs4 (also known as AvrBsP and AvrBs3‐2) is the cognate effector for the R gene Bs4 (Schornack et al., 2004). The Bs4 protein is a member of the Toll/interleukin 1 receptor‐nucleotide binding site‐leucine‐rich repeat (TIR‐NBS‐LRR) class of R protein and is predicted to be a cytoplasmic protein. The question arises, therefore, as to whether the detection of AvrBs4 is related to its presence as a foreign protein within the host cell and unrelated to AvrBs4 function per se, or whether AvrBs4, in fact, has a non‐nuclear function in the host (Gurlebeck et al., 2009).

XopD

XopD has a small ubiquitin‐like modifier (SUMO) protease domain, which is a member of the C48 protease family, and XopD has been shown to function in vitro in the release of SUMO from SUMO‐modified plant proteins (Chosed et al., 2007; Hotson et al., 2003). XopD has also been shown to be nuclear localized (Hotson et al., 2003) and, more recently, to have nonspecific DNA binding activity (Kim et al., 2008). Mutagenesis results have indicated that the binding activity is located to a helix–loop–helix region spanning amino acids 113–131. One of the mutants (V118P) has been shown to reduce the ability of XopD to enhance virulence and delay the onset of necrosis at later infection times (12 days after inoculation), indicating that the DNA binding domain is required for maximal XopD effect. Two ethylene response factor‐associated amphiphilic repression (EAR) motifs [L/FDLNL/F(x)P] were also detected within XopD at amino acids 244–249 and 284–289 (designated R1 and R2, respectively; Kim et al., 2008). EAR motifs are found in plant transcription factors that repress defence and stress responses, presumably, as the exact modes of action are unknown, by binding activators of defence or developmental responses. Alterations to the EAR motifs resulted in a reduction in virulence and delayed necrosis effect of XopD, but did not affect the localization or catalytic activity of the proteins. Specific alterations to the catalytic site of XopD also resulted in a loss of virulence and delayed necrosis (Kim et al., 2008). In addition to the biological effects, XopD mutants have also been shown to result in the loss of defence gene expression in comparison with the suppression associated with XopD (Kim et al., 2008). The SUMOylation of host proteins regulates protein stability and, as a consequence, processes in which the proteins are involved, including stress and developmental responses (Ulrich, 2009). Therefore, the hypothesis is that XopD may be directed to specific defence‐ or senescence‐related transcription factors, either through DNA binding or the EAR motifs or both, and the catalytic activity of XopD destabilizes, by deSUMOlaytion, the factors involved in defence responses (Kim et al., 2008).

XopJ

The XopJ group represents a subgroup of the large YopJ family of T3 effectors that are present in diverse pathogenic bacteria (Whalen et al., 2008). These proteins are related to the C55 group of cysteine protease and, more specifically, have been proposed to be ubiquitin‐like proteases and capable of removing SUMO proteins, which regulate the stability and activity of the modified (SUMOlyated) host proteins. YopJ, itself, has been shown to have transacetylase activity and to acetylate serine and threonine residues of several mitogen‐activated kinases (MAPs), preventing MAP kinase phosphorylation and signal transduction during innate immunity responses (Mukherjee et al., 2006). The suppression of HR by AvrBsT in Arabidopsis by recessive mutation in SOBER has been cited as support for transacetylation in a plant system (Cunnac et al., 2007). The results are consistent with a model whereby AvrBsT transacetylase activity is counteracted by the carboxylesterase activity of SOBER in most land races of Arabidopsis on an as yet unidentified host substrate. However, accession Pi‐0 carries a null mutation in SOBER and undergoes an HR in an AvrBsT‐dependent manner. The assumption is made that the carboxylesterase activity of SOBER would be hard to rationalize if AvrBsT was, in fact, a SUMO protease. Another member of the XopJ group, AvrRxv, induces an HR in the tomato cultivar Hawaii 7998, and the ability of AvRxv to induce HR has been shown recently to require the conserved amino acid residues of the catalytic triad (2008, 1993). Furthermore, the ability to elicit an HR was restored by the exchange of the catalytic domain of YopP, which is another member of the YopJ family from the animal pathogen Vibrio parahemolyticus and has previously been shown to transacetylate MAP kinases, with the AvrRxv catalytic domain (Whalen et al., 2008). Assuming that the N‐terminal portion of AvrRxv does not alter the enzymatic function of the catalytic domain, and that the HR‐eliciting activity is related to the virulence activity of AvrRxv, the results provide evidence for the transacetylase activity for AvrRxv. The requirement of ArvRxv for the conserved catalytic residues in the elicitation of HR is shared with AvrBsT (Orth et al., 2000). Regardless, different members of the XopJ group may function as transacetylases, and others, such as AvrXv4, may function as SUMO proteases (Roden et al., 2004b).

The direct host targets of the members of the XopJ group are unknown for both resistance and virulence. XopJ has been shown to localize to the host cell membrane, and localization is at least partially dependent on the myristoylation motif at the N‐terminus (Bartetzko et al., 2009; Thieme et al., 2007). XopJ has also been shown to inhibit host protein secretion and callose deposition, indicating possible interference with PTI signalling (Bartetzko et al., 2009). AvrRxv binds a tomato 14‐3‐3 protein in the yeast two‐hybrid assay (Whalen et al., 2008). Binding is mediated by the N‐terminal segment of AvrRxv (1–141 amino acids), which is proximal to the catalytic C‐terminal portion. Binding has also been shown to be dependent on an apparent 14‐3‐3 binding motif, known as the R18 domain, which is present in AvrRxv. Amino acid substitutions that disrupt the R18 domain also result in the loss of HR elicitation by AvrRxv on delivery by Agrobacterium‐mediated gene transfer (Whalen et al., 2008). 14‐3‐3 proteins function as adapter proteins in a variety of cellular complexes, including MAP kinase pathways, and AvrRxv may be, in part, localized to the site of action by binding to a specific 14‐3‐3 protein involved in host innate immunity signalling, possibly the MAP kinase pathway (Whalen et al., 2008). Once localized, AvrRxv would then transacetylate the as yet unknown host protein and prevent signal transduction.

XopN

Using a yeast two‐hybrid screen, XopN has been shown to interact with a variety of proteins from tomato, including an atypical (phosphorylation‐deficient) receptor‐link kinase named TARK1 and a 14‐3‐3 protein named TFT1. TARK1, similar to other RLK proteins, spans the cell membrane with the LRR domain external to the cell and the kinase (in this case, nonfunctional) domain in the cytoplasm. The interaction with TARK1 has been shown to require a LXXLL sequence in the cytoplasmic portion of TARK1. Alteration of the LXXLL motif resulted in a loss of TARK1 binding to XopN and binding was shown to require a LXXLL motif in the N‐terminal portion of XopN. A nonbinding mutant of XopN was then constructed and tested for effects on the biological phenotypes caused by XopN in X. campestris pv. vesicatoria. The loss of TARK1 binding resulted in an approximately 50% reduction in the virulence effect of wild‐type XopN and did not affect the suppression of callose deposition (Kim et al., 2009). The effect on suppression of defence‐related gene expression was not reported. Therefore, some evidence supports the fact that the interaction of TARK1 and XopN controls, at least in part, the effects of XopN in disease. Furthermore, suppression of TARK1 expression in tomato also resulted in partial suppression of XopN‐mediated susceptibility (Kim et al., 2009). Bacteria without XopN grew to higher populations in plants with reduced TARK1 expression when compared with growth in plants with a wild‐type TARK1. TARK1‐suppressed plants also displayed reduced disease symptoms, presenting the possibility that TARK1, indeed, may be involved in host defence responses (Kim et al., 2009). Atypical kinases have no intrinsic kinase activity and presumably mediate their effects solely through interactions with other components of the RLK pathways (Castells and Casacuberta, 2007). The involvement of the 14‐3‐3 proteins, which are known to be involved in signal transduction in plants and animals, with XopN remains to be shown. XopN contains several anti‐parallel α‐helical repeats, or HEAT (huntingtin, elongation factor 3, PR65/A, TOR1) repeats, the function of which is unknown at present (Roden et al., 2004a). The repeats of XopN may be involved in interactions with TFT1 or other host proteins.

XopAC and XopAH

Two T3 effectors of Xanthomonas contain sequences related to the recently described Fido domain, which is derived from the fic [cyclic adenosine monophosphate (cAMP)‐induced filamentation] and doc (death on curing) domains (Kinch et al., 2009). Structural comparisons resulted in the inclusion of similar segments of the T3 effector AvrB from P. syringae species (Kinch et al., 2009). A fido domain is present in the T3 effector from the animal pathogen Vibrio parahaemolyticus, named VopS, which causes host cell cytotoxicity and covalently modifies Rho GTPases with adenosine monophosphate moieties. The new modification has been called an AMPylation reaction and has been proposed to alter signalling functions of the target proteins, similar to phosphorylation, ribosylation and acetylation modifications (Yarbrough and Orth, 2009). The family contains central motif sequences (HPFx[D/E]GN[G/K]R), with histidine contributing to AMPylation. Although AvrB and related proteins lack the histidine residue, other residues within the structurally related region have been proposed to serve the same function (Kinch et al., 2009). XopAC (XCC2565) of X. campestris pv. campestris contains the consensus fic domain, whereas XopAH (XCC2109) from X. campestris pv. campestris has sequence similarity to AvrB (Table 1). Presumably, T3 effectors in the XopAC and XopAH groups could trans‐AMPylate plant host proteins, as shown for VopS. Interestingly, XopAH from X. campestris pv campestris strain 8004 also contains an identical ORF (XC_1553) to XCC2565 of strain ATCC33913, which was named AvrAC (Xu et al., 2008). Although AvrAC has been shown not to modulate virulence in susceptible plants, the effector elicits a resistance reaction in landrace Col‐0 of Arabidopsis. Curiously, resistance appears to be limited to the vascular system, implicating tissue specificity for resistance and possibly the virulence functions of T3 effectors (Xu et al., 2008).

COGNATE R GENES TO T3 EFFECTORS OF XANTHOMONAS

A greater goal for understanding T3 effector function is to place it in the context of host/pathogen adaptation and to determine whether the knowledge can assist in the design or selection of durable and broad resistance. Host adaptation to disease involves the evolution of modified defence surveillance mechanisms in response to the adaptive processes for virulence within pathogen populations. A variety of R genes that are directed against Xanthomonas species have been cloned and characterized (Table 4). The two largest and best‐characterized classes of protein that function as R genes are the aforementioned RLKs and NBS‐LRR, and, perhaps not surprisingly, over one‐half of the genes are members of these two classes. Bs2, Rxo1, Xa1 and Bs4 encode members of the NBS‐LRR family. Xa21 and Xa26, as members of the RLK class, represent components of the PRR line of defence in rice and, in themselves, are not cognate R genes corresponding to specific T3 effectors. However, the case for suppression of flagellin/FLS2‐mediated PRR signalling by T3 effectors is well advanced in the studies of P. syringae, and interesting questions arise regarding T3 effectors with related functions from Xanthomonas (Boller and He, 2009). One can anticipate the identification of specific T3 effectors in Xanthomonas directed against this class of protein. If species of Xanthomonas, in general, are proficient in RLK signal transduction suppression, strains may adapt to the presence of Xa21, Xa26 or similar R genes through the acquisition of PRR pathway suppressors.

Table 4.

Cloned R genes for resistance to Xanthomonas species.

Gene	Class (plant)	Comments*	Cognate T3 effector	Reference
Bs2	NBS‐LRR (pepper)		AvrBs2	Tai et al., 1999
Bs3	Inducible (pepper)	Flavin oxygenase‐related protein	AvrBs3	Romer et al., 2007
Bs3‐E	Inducible (pepper)	Promoter deletion of Bs3	AvrBs3Δ16	Romer et al., 2009
Bs4	NBS‐LRR (tomato)	Cytoplasm	AvrBs4 (AvrBsP; AvrBs3‐2)	Schornack et al., 2004
Rxo1	NBS‐LRR (maize)	Cytoplasm; broad resistance; transferred to rice from maize	AvrRxo1	Zhao et al., 2005
Xa1	NBS‐LRR (rice)	Cytoplasm	Unknown	Yoshimura et al., 1998
xa13	Promoter mutants of Os8N3; nodulin 3 family (rice)	Membrane protein; unresponsive S gene to TAL effector PthXo1	PthXo1	Chu et al., 2006; Yang et al., 2006
Xa21	RLK (rice)	Extracellular receptor, broad resistance	None	Song et al., 1995
Xa26/Xa3	RLK (rice)	Extracellular receptor, broad resistance	None	Xiang et al., 2006
Xa27	Inducible (rice)	Membrane and cell wall; novel protein; broad resistance	AvrXa27	Gu et al., 2005
xa5	Missense mutant of TFIIAγ5; small subunit of TFIIA transcription factor complex (rice)	Nuclear; broad resistance	Unknown	Jiang et al., 2006; Iyer and McCouch, 2004

^* Comments in relation to R gene product localization, resistance profile, putative function/structure.

In addition to major classes of R genes, studies of Xanthomonas diseases of rice have resulted in the identification of a variety of novel additions to the R gene pantheon, indicating that the struggle continues within the host nucleus (White and Yang, 2009). The host transcriptional machinery is under attack, as revealed by studies of XopD and the various TAL effectors. Xa27 and Bs3 reflect newly discovered strategies to abrogate infection. The induction of self‐inflicted cell damage, as mediated by TAL effector recognition, seems to be remarkably analogous to the phenomenon of host lesion mimics. The host is under apparent selective pressure to evolve toxic gene–promoter combinations that react to TAL effectors. Xa27 and Bs3 products are not related on the basis of sequence or, seemingly, biochemical function. Xa27 is targeted to the host membrane, whereas Bs3 appears to represent a cytoplasmic protein. It will be interesting to discover how many other host proteins exist that can potentially trigger a defensive reaction on expression. The two recessive resistance genes, xa5 and xa13, directed at X. oryzae pv. oryzae, also represent new insights into host adaptations for resistance and will undoubtedly also provide insight into TAL effector function (Chu et al., 2006; Iyer and McCouch, 2004; Jiang et al., 2006; Yang et al., 2006). Both are hypothesized to interfere with effector function, albeit by different mechanisms. xa5, as a component of the transcription complex, may function by interference with general TAL effector interaction with the host transcription machinery. xa13, by contrast, interferes with the action of a specific TAL effector, namely PthXo1 (Yang et al., 2006).

CONCLUSIONS

The experimental analyses of individual T3 effectors of Xanthomonas have provided a unique view of the interactions of plant pathogenic bacteria and their hosts, and genomic sequence analysis has provided an assessment of the similarities with related pathogens and the respective armamentarium of T3 effectors. At the same time, advances in sequencing technology also promise to provide unprecedented access to the diversity of T3 effectors in Xanthomonas as well as other species. Some insights into the aspects of host/pathogen adaptation are already apparent from the recent findings for the TAL effector family. Results with XopJ, AvrRxv, XopD and XopN also implicate a complexity of interactions with host factors. Structural studies of T3 effectors may yield functional similarities not detectable by primary sequence comparisons. Continued advancements in the examination of physiological effects of individual proteins, such as protoplast systems, Agrobacterium‐mediated transfer and T3 effector gene‐deficient strains of the pathogens, will assist in the identification of the individual contributions of each candidate. In some respect, examination of the different effector families is a view of plant–pathogen interactions backwards through time. Groups with highly diverse members shared among plant and animal effectors may reflect the evolutionarily conserved defence pathways of plants and animals. Other families, including the AvrBs3 family, may be newly emerging virulence factors, and have only recently spread throughout the genus. Undoubtedly, the groupings of Xop will be revised, and manual revisions in T3 effector annotation will be required. Additional improvements in T3 effector validation on a genomic scale will also facilitate research. Nevertheless, a plethora of virtually unexamined T3 effectors exist, and the T3 effectors of Xanthomonas will provide fuel for many more studies and insights in the next few years.