Regulation of phytobacterial T3SS genes
Introduction
Around 20 Gram-negative pathogenic bacteria rely on the needle-like type three secretion system (T3SS) to secret a group of effector proteins that help bacteria to infect eukaryotic host organisms (Jin et al. 2003). The expression of T3SS genes is coordinately regulated by many endogenous regulatory proteins and various external environmental factors (Rahme et al. 1992; Xiao et al. 1992). In phytopathogenic bacteria, the T3SS are encoded by the hrp (hypersensitive response and pathogenicity) genes, which can be divided into two main groups (Tang et al. 2006). The hrp genes of Pseudomonas syringae, Erwinia spp. and Pantoea stewartii belong to group I that is regulated by the HrpL, an ECF (extracytoplasmic factor) family alternate sigma factor (Xiao et al. 1994b). The hrp genes in group II are activated by an AraC-like activator, such as HrpX in Xanthomonas spp. and HrpB in Ralstonia solanacearum (Alfano and Collmer 1997). To date, a large number of T3SS-regulating components acting upstream of hrp genes have been identified and characterized in various phytopathogenic bacteria, including two-component systems (TCS), transcription factors, membrane proteins, quorum-sensing genes, plant-derived compounds and medium components. In order to better understand the molecular basis of bacterial pathogenesis and the microbe-plant interaction, future studies are required to elucidate the nature of T3SS signals, how bacteria sense the signals and the connections between multiple T3SS regulatory genes and hrp/effector genes.
Bacterial type three secretion systems.
Bacteria use at least eight different secretion systems (from type one to type eight) to translocate proteins from the interior cytoplasm cross the membranes to its exterior, which is a very important mechanism for bacteria to adapt surrounding environment and survive (Desvaux et al. 2009). Around 20 Gram-negative pathogenic bacteria rely on the needle-like type three secretion system (T3SS) to secret proteins that help bacteria to infect eukaryotic host organisms (Jin et al. 2003). The T3SS is a sophisticated molecular machinery that is composed of more than 20 different proteins, making it the most complex secretion system. T3SSs are essential for the bacterial virulence, which has been proved by the evidence that T3SS-defecient mutants are nonpathogenic (Zhou and Chai, 2008).
Even though a complete T3SS supramolecular structure has yet to be purified from phytopathogenic bacteria, the elegant work was carried out in the mammalian pathogen Salmonella enterica (Buttner and He, 2009). The T3SS needle structure measures approximately 80 nm in length and 8 nm in width. It starts at the basal body in the bacterial cytoplasm, crosses the two bacterial membranes and extends the needle out of the cell. The basal body is composed of two rings that interact with the inner cytoplasmic and the outer membrane, respectively. An inner rod connects the basal body to the needle, which is made of 100-150 subunits of one single small protein (Kubori et al. 1998). In phytopathogenic bacteria, the T3SS filaments are called hrp (hypersensitive response and pathogenicity) pili (Roine et al. 1997).
The genes encoding the T3SS secretion apparatus are located on the chromosome in some bacteria and on a plasmid in other bacteria. In Pseudomonas syringae, for example, most T3SS genes (called T3 genes hereafter) are located in six operons in a chromosomal hrp island that is responsible for pathogenicity of host plants and the hypersensitive response (HR) of resistant and non-host plants (Collmer et al. 2000). The rest of the T3 genes are scattered in other regions of the genome. On the other hand, Shigella spp. has a large virulence plasmid that encodes all T3SS genes (Buchrieser et al. 2000). The T3SS proteins include regulatory proteins that control the expression of T3SS upon host infection, structure proteins that build the needle complex, effectors that are secreted and are responsible for the virulence and chaperones that help with the secretion of effectors (Buttner and He, 2009).
The expression of T3SS genes is coordinately regulated by many endogenous regulatory proteins and various external environmental factors (Rahme et al. 1992; Xiao et al. 1992). Multiple signal transduction components have been identified in the gene regulation pathways. The mechanism of regulation is highly variable from mammalian pathogens to plant bugs. Most T3SSs in animal pathogens are activated after the bacteria contact with the surface of eukaryotic cells (Hueck, 1998). For example, before cell contact, the channels of Yersinia spp.T3SS are shut by an outer membrane protein YopN, and a negative transcription factor LcrQ that represses yop (Yersinia outer protein) genes. Upon cell contact, YopN is removed, resulting in secretion of LcrQ, which in turn releases the repression of yop genes (Rosqvist et al. 1994). The mechanisms of T3SS gene regulation in the plant bacterial pathogens are divided into two main groups. The hrp genes of Pseudomonas syringae, Erwinia spp. and Pantoea stewartii belong to group I that is regulated by the HrpL, an ECF (extracytoplasmic factor) family alternate sigma factor. On the other hand, the hrp genes in group II are activated by an AraC-like activator, such as HrpX in Xanthomonas spp. and HrpB in Ralstonia solanacearum (Alfano and Collmer 1997). hrp genes are expressed at a very low level in the nutrient rich medium, but activated in the hrp-inducing minimal medium, which is believed to largely mimic the plant apoplast conditions (Jin et al. 2003). The details of the mechanisms that the phytopathogenic bacteria use to activate T3SS are discussed in the following sections.
As soon as the expression of T3 genes is activated, the needle-like T3SS apparatus is constructed. In Pseudomonas syringae, the basal body of the T3SS is encoded by the so-called hrc (hypersensitive response and conserved) genes, which are highly conserved between diverse plant and animal pathogens (Collmer et al. 2000). Eight hrc genes share high similarity with flagellar genes, suggesting that the T3 apparatus is related to a flagellum (He, 1998). P. syringase HrpA is the major structural protein of the hrp pili, which is much longer than the counterparts in mammalian pathogens, suggesting that hrp pili span thick plant cell wall (Jin et al., 2001; Kubori et al., 1998).
Effectors are the T3SS-injected virulence proteins that are responsible for the bacterial pathogenicity. Although no conserved motif can be found in the N-terminal region of multiple effectors, it is believed that the amphipathic nature and the N-terminal amino acid composition of effectors are the secretion signals (Galan and Wolf-Watz, 2006; Arnold et al. 2009; Samudrala et al. 2009). Most effectors of P. syringae reside in the hrp island, where the hrp/hrc genes are flanked by a conserved effector locus (CEL) and an exchangeable effector locus (EEL). The CEL has been shown to be crucial for the bacterial pathogenicity while the EEL only has a minor role (Alfano et al. 2000; Collmer et al. 2000). A handful of effectors are named Avr proteins because they were initially characterized as the proteins that induced the avirulent reaction in the host plants carrying the cognate resistant genes (R genes) (Alfano and Collmer, 2004; Leach and White, 1996). Upon the recognition of the corresponding R genes in the resistant plants, the bacterial Avr proteins elicit the hypersensitive response (HR), which is a rapid local cell death that limits the growth and spread of the pathogen at the infection site. On the other hand, in the susceptible plants without corresponding R genes, the Avr effectors function as virulent determinants by interfering with the host defense mechanisms and manipulating host cellular activities to the benefit of the pathogen (Alfano and Collmer, 2004).
Genome sequencing and bioinformatic analysis have enabled the comprehensive identification of the effector repertoires in various phytobacteria. Many notable studies have characterized the ability of single effectors to suppress plant innate immune defenses (Navarro et al. 2006, 2008, He et al. 2006; Shan et al. 2008), manipulate hormone signaling pathways (Navarro et al. 2006; de Torres-Zabala et al. 2007; Jelenska et al. 2007; Chen et al. 2007), induce cell death (Wei et al. 2007a), act as transcription activators (Yang and White, 2004; Gu et al. 2005; Römer et al. 2007; Kay et al. 2007) and/or display different biochemical activities on plant protein targets (Tang et al. 1994; Fu et al. 2007; Rosebrock et al. 2007; Xiang et al. 2008).
The secretion of many T3SS effectors needs corresponding chaperones, which are small cytoplasmic proteins that specifically bind to individual TTSS effectors. Because the final folding condition for effector proteins is inside the eukaryotic host cell, it is proposed that T3 chaperones prevent the cognate effectors from aggregation or degradation in the bacterial cytoplasm and function as signals to direct the cognate effectors to the T3SS machinery (Losada and Hutcheson, 2005; Feldman and Cornelis, 2003). Most genes encoding T3SS chaperones are linked with cognate effector genes (Guttman et al., 2002). Most T3 chaperones are specifically required for the secretion of their corresponding effectors, whereas some such as HpaB from Xanthomonas campestris pv vesicatoria help to secret multiple effectors (Parsot et al., 2003; Buttner et al., 2004, 2006).
Gene regulation of group I hrp genes in phytobacteria
Expression of the Group I hrp genes in P. syringae, Erwinia spp. and Pantoea stewartii, is regulated by HrpL, an ECF family alternate sigma factor that is essential for the induction of genes carrying an hrp box (GGAACC-N15/16-CCACNNA) in their promoters (Xiao et al. 1994a). HrpL binds directly to the hrp box (Nissan et al. 2005). The hrp box consensus sequence has been used to identify novel candidate T3 effector genes in the genomes of these bacteria via bioinformatic analysis (Fouts et al. 2002; Zwiesler-Vollick et al. 2002; Ferreira et al. 2006). In addition to T3 genes, many of non-T3 genes that also contain the hrp box in the promoters and are induced by hrpL, suggesting relationship between T3SS induction and activation of other biological processes upon bacteria interact with plants (Boch et al. 2002; Lan et al. 2006).
In P. syringae, the hrpL-based induction depends on another alternate sigma factor, RpoN (σ54), and two NtrC-family transcription factors, HrpR and HrpS (Hendrickson et al. 2000; Hutcheson et al. 2001; Xiao et al. 1994b). RpoN controls the transcription of hrpL under a σ54-dependent promoter (Chatterjee et al. 2002; Hendrickson et al. 2000). The hrpR and hrpS genes are in the same operon that is driven by the promoter upstream of hrpR (Grimm et al. 1995; Xiao et al. 1994b). The two highly homologous proteins physically interact and dimerize, which is believed to be essential for the induction of hrpL (Hutcheson et al. 2001). hrpS alone is able to induce hrpL to a very low level, while the full activation of hrpL requires both hrpR and hrpS (Grimm et al. 1995; Hutcheson et al. 2001). Both HrpR and HrpS protein carries an enhancer-binding domain and a motif that interacts with the σ54-RNA polymerase. HrpR and HrpS are proposed to form a heterodimer that binds to the hrpL promoter and induces hrpL transcription by interacting with the RpoN-RNA polymerase under the T3SS-inducing conditions (Hutcheson et al. 2001). In Erwinia spp. and Pantoea stewartii, only hrpS is needed to induce hrpL (Frederick et al. 2003). Another locus rsmA/rsmB in Erwinia carotovora has been demonstrated to control the hrpL expression (Chatterjee et al. 2002). rsmA encodes a small RNA-binding protein and rsmB is an RNA (Chatterjee et al. 1995; Liu et al. 1998). The hrpL transcription is abolished in an rsmB- mutant but is higher in an rsmA- mutant, suggesting that rsmA is a negative regulator and rsmB is positive regulator of T3SS in Erwinia carotovora (Chatterjee et al. 2002).
In P. syringae, HrpS activity is repressed by HrpV, a T3SS negative regulator that physically interacts with HrpS (Preston et al. 1998; Wei et al. 2005). In the inducing medium, an hrpV- mutant displays an elevated level of T3 gene expression, whereas the strain overexpressing hrpV abolishes T3 gene induction. HrpV-mediated repression can be cleared by HrpG, a chaperone-like protein that interacts with HrpV and liberates HrpS from HrpV-mediated repression without affecting the hrpV transcription (Wei et al. 2005). HrpV acts as anti-activator of HrpS, and HrpG is an anti-anti-activator. In Erwinia spp. and Pantoea stewartii, the hrpS-hrpL-hrp cascade is positively regulated by a two-component system hrpXY (Merighi et al. 2003; Nizan-Koren et al. 2003). The phosphorylation of the response regulator HrpY, likely by the cognate histidine kinase HrpX, is required for the activation of the hrpS-hrpL-hrp cascade (Nizan-Koren et al. 2003); however, how hrpXY activates the hrpS expression is still unknown.
The P. syringae HrpR protein is degraded by an ATP-dependent protease, Lon, which degrades unstable or misfolded proteins involved in a variety of biological processes in bacteria (Bretz et al. 2002). HrpR is very unstable in MM but is stabilized in a lon- mutant, leading to elevated expression of T3 genes in KB medium (King et al. 1954; Bretz et al. 2002; Lan et al. 2007). In addition, the lon- mutant hypersecretes T3SS effectors, suggesting a Lon-associated degradation of these effectors. The effectors have been shown to be protected from Lon degradation by their cognate chaperones prior to secretion (Losada and Hutcheson. 2005). In lon- mutant, the expression of hrpL exhibits a dynamic change in MM. hrpL is transcribed at a higher level in the lon- mutant than in the wild-type strain shortly after induction in MM, but it is more abundant in the WT strain at later time points (Lan et al. 2007).
The hrpRS transcription displays a two to four-fold induction in both inducing medium and plant (Rahme et al. 1992; Thwaites et al. 2004; Lan et al. 2006). The expression of hrpRS is regulated by at least two two-component systems (TCS) GacAS and RhpRS (Chatterjee et al. 2003; Lebeau et al. 2008; Xiao et al. 2007). GacA is the cognate response regulator, whereas GacS is the histidine kinase. The GacAS system plays important roles in regulating multiple biological processes in bacteria, such as virulence, production of toxin and antibiotics, quorum-sensing, motility, production of exopolysaccharides, biofilm formation and stress tolerance (Heeb and Haas 2001). In P. syringae pv. tomato DC3000, mutation in gacA attenuates the transcription of hrpRS, rpoN, and hrpL significantly, suggesting that gacA is an important T3SS regulator that is located at the top of the regulatory cascade in DC3000 (Chatterjee et al. 2003). A recent work in Erwinia chrysanthemi 3937 has also demonstrated that GacA is required for the expression of T3 genes (Lebeau et al. 2008). The signal sensed by GacS and the mechanism by which GacA controls the transcription of hrpRS and rpoN have not been identified.
Another TCS mutant has been shown to display diminished induction of the T3 genes in the minimal medium and plant (Xiao et al. 2007; Deng et al. 2009). The mutant carries a transposon insertion in a putative sensor kinase gene rhpS and its pathogenicity is much reduced in host bean plant compared to wild-type bacteria. rhpS is located immediately downstream of a putative response regulator gene rhpR, and the two genes are organized in an operon. rhpS- mutant shows reduced transcriptional induction of hrpR and avrPto, suggesting that rhpS is a key sensor for activating T3 genes of P. syringae in the minimal medium and plant. The deletion mutant of the whole rhpRS locus, ΔrhpRS, and the wild-type strain share similar induction of avrPto and pathogenicity in the host plant, suggesting that RhpR is a negative regulator of T3SS. Overexpression of RhpR in the deletion mutant ΔrhpRS suppresses the induction of T3 genes in a phosphorylation-dependent manner (Xiao et al. 2007). Based on these observations, RhpR is proposed to be phosphorylated by an unknown factor in the rhpS- mutant and the phosphorylated RhpR represses the T3 genes. In wild-type bacteria, RhpS acts as a phosphatase and retains RhpR in a dephosphorylated state when the bacteria are grown in T3 gene-inducing conditions.
In addition to the GacAS and RhpRS systems, HrpA, the major component of the hrp pilus, also has been shown to regulate the hrpRS transcription (Wei et al. 2000). The mutation in hrpA severely compromised the transcription of the hrpRS and hrpL, which can be restored by overexpression of hrpRS. However, the mechanism by which HrpA controls hrpRS is unknown. Besides, mutation in corR, which encodes a response regulator controlling expression of the phytotoxin coronatine in Pseudomonas syringae pv. tomato, shows both a reduction and a delay in the expression of hrpL and a reduction of disease symptom development, compared with wild-type strain. A putative CorR-binding site is located upstream of hrpL, and the gel shift assay confirms the binding of CorR to this DNA motif (Sreedharan et al. 2006).
Gene regulation of group II hrp genes in phytobacteria
The group II hrp genes in Xanthomonas spp. and Ralstonia solanacearum are regulated by the AraC-type transcriptional activator HrpX and HrpB, respectively (Genin et al. 1992; Kamdar et al. 1993; Wengelnik and Bonas, 1996). The protein sequences of HrpX and HrpB proteins are highly conserved. In Xanthomonas spp., HrpX specifically binds to a conserved motif PIP (plant inducible promoter)-box (TTCGC-N15-TTCGC), which is present in the promoter regions of most HrpX-regulated genes (Coebnik et al. 2006) Similarly, many HrpB-regulated genes in R. solanacearum contain an hrpII-box (TTCG-N16-TTCG) in the promoters (Cunnac et al. 2004). Although computational searches for the PIP/hrpII motifs have been successful to identify type III effector genes, some HrpX/HrpB-regulated T3 genes do not have the PIP/hrpII-box, such as avrBs1 and avrBs3 family genes in Xanthomonas spp. (Thieme et al. 2005; Occhialini et al. 2005).
hrpX and hrpB are activated by another key regulator HrpG, an OmpR-type two-component response regulator containing a DNA-binding domain (Brito et al. 1999; Wengenilk et al. 1996). HrpG of Xanthomonas axonopodis pv. citri has been shown to physically interact with an uncharacterized two-component system histidine kinase, suggesting that HrpG may be phosphorylated by it (Alegria et al. 2004). In X. campestris pv. vesicatoria, three point mutations of HrpG are constitutively active in the T3SS repressing medium, suggesting that HrpG may need conformational change to activate T3 gene expression (Wengenilk et al. 1999). Five other loci have been recently demonstrated to be involved in T3SS regulation in Xanthomonas spp. Like Erwinia spp., an rsmA-like gene in X. campestris pv. campestris has reported to play an negative role in pathogenesis, as an rsmA- mutant results in an enhanced bacterial virulence (Chao et al. 2008). A mutation in hpaR, encoding a putative marR family transcriptional regulator, is nonpathogenic to the host cabbage plant. hpaR is regulated by hrpG/hrpX and is repressed in rich medium but induced in T3-inducing medium (Wei et al. 2007b). Zur, the key regulator for zinc homeostasis, positively regulate hrp genes through hrpX, but not hrpG (Huang et al. 2009). Using mutational analysis, a two-component system colRS has been identified as another novel regulator of the pathogenesis of X. campestris (Zhang et al. 2008). Finally, Xanthomonas oryzae pv. oryzae PhoPQ two-component system positively controls hrpG expression and virulence (Lee et al. 2008).
In R. solanacearum, hrpG is constitutively expressed in both rich and minimal media but presumably induced by a plant signal (Brito et al. 1999). Upon sensing plant signal, the expression of hrpG is activated by five upstream signal transduction components, prhA, prhJ, prhI, prhR and phcA, which are discussed in a following section.
Bacterial two-component transduction systems.
Bacteria primarily utilize two-component systems (TCS) to couple environmental signals to adaptive responses (Hoch. 2000). Most bacterial genomes reside more than ten TCSs that play important roles in regulating multiple biological processes, such as metabolism, growth, motility, quorum-sensing and pathogenicity (Gao and Stock. 2009). TCSs generally include a sensor histidine kinase (HK) and a response regulator (RR). Upon sensing specific signals, the HK autophosphorylates the conserved histidine (His) residue of the kinase domain, and the high-energy phosphoryl group is subsequently transferred to the aspartate (Asp) residue of the cognate RR. Phosphorylation of RR induces its conformational change that activates the RR to trigger the response (Stock et al. 2000).
HKs and RRs are modular proteins with variable domains, suggesting that they are versatile in sensing a wide variety of environmental signals. Typical HKs have a diverse transmembrane input domain linked to a conserved cytoplasmic kinase core. The N-terminal diverse input domain enables HKs to perceive a wide variety of stimuli, such as ions, metabolic molecules, light, osmolarity, humility, cell envelop stress, reactive oxygen species and electrochemical gradients (Gao and Stock. 2009). Signals are sensed by different input domains through physical interactions with other signal transduction proteins. Although great advancements have been achieved in understanding the signal sensing mechanisms in a few HKs in recent years, the exact signal for most HKs still remains unknown (Mascher et al. 2006; Szurmant et al. 2007).
The kinase core, where HKs usually autophosphorylate spontaneously, contains an N-terminal dimerization and histidine phosphotransfer (DHp) domain and a C-terminal catalytic and ATP binding (CA) domain. The CA domain has the kinase activity, which transfers a phosphoryl group from ATP to the conserved His residue in the DHp domain (Stock et al. 2000). In many cases, HKs are bifunctional and can have both the kinase and phosphatase activities, which control the level of RR phosphorylation and the response afterwards (Laub and Goulian, 2007). The DHp domain has the phosphatase activity, which is also affected by the interaction between the DHp and CA domains. The conserved His residue in the Escherichia coli osmosensor HK EnvZ has been shown to be essential for the phosphatase activity, suggesting that a reverse transfer of the phosphoryl group from RR to HK maybe account for the dephosphorylation process (Dutta et al. 1996; Zhu et al. 2000). However, some HK mutants that change the conserved His to other residues still remain the phosphatase activity, indicating that the phosphatase activity of HKs involve other mechanisms (Chamnongpol et al. 2003). The level of RR phosphorylation and output response are largely controlled by the HK kinase activity (Fleischer et al. 2007), the HK phosphatase activity (Brandon et al. 2000), or both (Chamnongpol et al. 2003), suggesting the big diversity of mechanisms in HK signal transduction. HKs always function as dimers that are controlled by a trans-phosphorylation mechanism. The CA domain of one dimer subunit phosphorylates the His residue on the DHp domain of the other dimer subunit (Stock et al. 2000).
The typical RR carries an N-terminal REC domain and a C-terminal variable effector domain. The REC domain is responsible for receiving the phosphoryl group from the HK and controlling the effector domain in a phosphorylation-dependent manner. The REC domain is a phosphorylation-activated switch that controls the conformation of RRs. An unphosphorylated REC domain exists in the inactive conformation, whereas the phosphorylation at the conserved Asp residue switches it to the active conformation (Gardino et al. 2007). The REC domain has both phosphoryl transfer and autodephosphorylation activities, which determine the level of RR phosphorylation that controls effector domain activity (Stock et al. 2000).
A great variety of output responses are generated by the diverse of effector domains, which can be categorized into at least five groups (Gao and Stock. 2009). The majority of RRs (63%) contain DNA binding effector domains that can be further grouped into four major subfamilies, including OmpR (30% of all RRs) (Martinez-Hackert et al. 1997), NarL (17%) (Milani et al. 2005), NtrC (10%) (Batchelor et al. 2008) and LytTR (3%) (Sidote et al. 2008). DNA-binding RRs regulate target genes expression through modulation of their own phosphorylation status, inducing dimerization or higher-order oligomerization, thereby controlling its affinity for DNA motifs in the promoter region of downstream genes (Martinez-Hackert et al. 1997). Unlike the prototypical RR structures, nearly 17% of RRs have only REC domains, falling into the second group. Most of these RRs regulate motility by interactions with motor proteins or phosphorylate intermediates in phosphorelay pathways (Varughese et al. 2005). The third group is enzymatic domains that are found in around 13% of RRs. About half of these enzymatic RRs play a role in the regulation of cyclic diguanylate (Romling et al. 2005). The fourth group is consisted of 3% of RRs that reside a small and diverse group of effector domains that interact with other proteins or ligands (Gao and Stock. 2009). Finally, the fifth group includes only 1% of RRs containing RNA binding domains that function as anti-termination factors (O'Hara et al. 1999). An RR can regulate its downstream gene(s) as activator, repressor, or both. DNA recognition sequences or enzymatic domains of RRs are often weakly conserved and need experimental confirmation rather than computational prediction (Gao and Stock. 2009).
Most sequenced bacterial genomes encode dozens of TCS proteins, which makes it possible for cross-phosphorylation between similar DHp and REC domains, resulting in complicated one-to-many, many-to-one, and many-to-many relationships between HK and RR proteins. Besides, approximately 25% of HKs has an additional REC domain that can be phosphorylated by the kinase domain of the HK (Ogino et al. 1998). Phosphorylated REC domain can transfer the phosphoryl group to a His-containing HPt domain and relay the phosphorylation to a cognate RR for output responses, which forms a sophisticated mechanism of His-Asp-His-Asp phosphorelay. This HPt domain can be a part of an HK, a single protein, or part of another membrane protein (Stock et al. 2000).
A handful of TCSs are involved in autoregulation. For example, phoPQ of Salmonella enterica serovar Typhimurium is activated by PhoP by binding to a DNA motif in its own promoter (Soncini et al. 1995; Gupta et al. 2006; Gonzalo-Asensio et al. 2008). Transcription from ompR promoter in the same bacterium requires phosphorylated OmpR for induction (Bang et al. 2002). The Bordetella pertussis response regulator BvgA controls the transcription of many pathogenicity-related genes, and also regulates the transcription of the bvgAS operon (Scarlato et al. 1990). In Mycobacterium tuberculosis, response regulator TcrR autoregulates its own expression by interacting with an AT-rich region (Haydel et al. 2002). Response regulator CovR of the human pathogen Streptococcus mutans binds to its own promoter and represses its own transcription (Chong et al. 2008). HrpXY of Pantoea stewartii also activates both hrpS and its own promoter (Merighi et al. 2003).
Many TCSs play an important role in controlling bacterial pathogenicity. As discussed previously, in phytopathogenic pathogens, a group of TCSs act as critical regulators in controlling the expression of virulence genes, such as GacAS (Chatterjee et al. 2003; Lebeau et al. 2008), RhpRS (Xiao et al. 2007) and CorR (Sreedharan et al. 2006) in P. syringae, ColRS (Zhang et al. 2008) and PhoPQn(Lee et al. 2008) in Xanthomonas spp. and HrpXY in Pantoea stewartii (Merighi et al. 2003; Wei et al. 2005).
Regulation of phytobacterial T3SS genes by host and environmental signals.
Host sensing is essential for the activation of T3 genes of bacterial pathogens, which is responsible for the development of disease (Brencic and Winans 2005). Even though almost nothing is known about the host signals of phytopathogenic bacteria, the most elegant work was carried out in Ralstonina solanacearum. Like many other Gram-negative phytopathogenic pathogens, R. solanacearum. T3SS genes are well induced upon bacteria-plant cell contact (Aldon et al. 2000). A mutation in prhA, a gene encoding the outer membrane protein that is homologous to siderophore receptors, disrupts the induction of T3 genes by the plant, but not by the minimal medium (Marenda et al. 1998). PrhA might sense a yet to identified plant-specific signal, likely a non-diffusible component in plant wall (Aldon et al. 2000).
Two other genes acting downstream of prhA are prhI and prhR, which are organized in the same operon in the hrp gene cluster and encode a sigma factor transmembrane protein and an ECF sigma factor, respectively (Brito et al. 2002). prhIR- mutant shows compromised pathogenicity and HR elicitation. PrhIR are required for the activation of the T3 genes expression in contact with plant, but not with minimal medium. The authors propose that a plant signal sensed by PrhA is transferred to PrhR and passed through the membrane. In the cytoplasm, PrhI is activated by PrhR, and then sequentially activates three transcription factors PrhJ, HrpG and HrpB (Brito et al. 1999; 2002). In addition, a LysR family transcriptional regulator PhcA negatively regulates the protein level but not transcription level of HrpG in rich medium (Genin et al. 2005). It is recently reported that PhcA repress HrpG by binding to the prhIR promoter and repressing its transcription (Yoshimochi et al. 2009).
The T3 genes of P. syringae are strongly induced in the plant. A handful of reports suggest that the perception of plant signals is important for the activation of T3SS genes in Pseudomonas syringe and other bacteria. For example, P. syringae hrpL gene is induced much greater in planta than in the minimal medium, suggesting the presence of a plant-specific signal for T3SS (Rahme et al. 1992). In support of this observation, P. syringae mutants of a conserved hypothetical gene abolish the T3 gene induction in planta, but not in the inducing medium. In comparison to the wild-type bacterium, the mutants show reduced pathogenicity on host plants and the HR on the nonhost plants (Y. Xiao and X. Tang, unpublished results). Similarly, a mutation in the histidine kinase sensor gene rhpS in P. syringae severely compromises T3 gene induction in planta as well as in the minimal medium, suggesting that RhpS senses signals in both conditions (Xiao et al. 2007). It is recently reported that the induction of P. syringae hrpA promoter is enhanced by cell-free exudates from plant cell suspension cultures. Further analysis suggests that some water-soluble plant-cell-derived compounds are the signals that are sensed by bacteria (Haapalainen et al. 2009). Furthermore, a study in Dickeya dadantii (Erwinia chrysanthemi) 3739 has found two plant phenolic acids that induce T3 gene expression, which are the first identified specific T3SS inducers in phytobacteria (Yang et al. 2009).
Some plant signals may act as T3SS repressors that inhibit in planta T3 genes induction. In a study trying to identify host signals for the induction of bacterial T3 genes, an Arabidopsis (a host of P. syringae pv. tomato DC3000) mutant, att1 (aberrant induction of type three genes) has been isolated. The att1- mutant greatly enhances the in planta expression of bacterial T3 genes, suggesting a negative role of ATT1 in regulating T3 gene expression. ATT1 encodes CYP86A2, a cytochrome P450 monooxygenase that catalyzes fatty acid oxidation, which regulates normal cuticle development (Xiao et al. 2004). Certain lipids may reduce T3 gene expression from the intercellular spaces. These lipids might be either cutin monomers or cutin-related fatty acids that CYP86A2 synthesize. In support of this hypothesis, the authors have tested a variety of commercial cutin-related fatty acids and found that some of them, such as oleic acid oxide (cis-9,10-epoxy stearic acid), are capable of repressing hrp promoter activity (Xiao et al. 2004). The negative cutin-related signals may inhibit T3 genes expression during bacterial growth on the leaf surface (Xiao et al. 2004). The chemical nature and how the negative signals are perceived by P. syringae are still unknown. Examples of plant components acting as negative signal for T3 genes have been reported in other phytobacteria. In Dickeya dadantii (Erwinia chrysanthemi), a plant phenolic acid, p-coumaric acid represses the expression of T3 genes, suggesting a plant defense mechanism against bacterial pathogens (Li et al. 2009). Low molecular weight (<10kDa) plant extract also inhibits hrp genes expression in Xanthomonas campestris pv. campestris (Watt et al. 2009).
Bacterial quorum-sensing system has also been recently demonstrated to regulate T3SS in Pseudomonas syringae and Pantoea agglomerans (Chatterjee et al. 2007; Deng et al. 2009; Chalupowicz et al. 2008; 2009). P. syringae produces N-acyl homoserine lactones (AHLs) as the signal of the quorum-sensing system that coordinates multiple bacterial genes expression adaptive to the local population density (Fuqua et al. 1994). A Tn5 insertion mutation in psrA, a Pseudomonas sigma regulator, results in enhanced expression of AHL synthase gene psyI and reduced pathogenicity in host tomato, suggesting the regulatory interaction between quorum-sensing and T3SS (Chatterjee et al. 2007). In support of this observation, AefR, the TetR-type gene known to regulate AHL production in P. syringae pv. syringae (Quinones et al. 2004; 2005), has recently been identified to positively control hrpRS and hrpL in P. syringae pv. phaseolicola (Deng et al. 2009). In gall-forming Pantoea agglomerans, pagI and pagR are responsible for the production and sensing the quorum-sensing signals N-l-homoserine lactones (HSL) (Chalupowicz et al. 2008). The expression of hrpXY, hrpS and hrpL in a pagI mutant or a pagR mutant is significantly repressed compared with wild-type, suggesting that T3SS regulation is subject to quorum-sensing system in P. agglomerans pv. gypsophilae (Chalupowicz et al. 2008, 2009).
Additionally, an iaaH- mutant lacking the auxin biosynthesis and an etz- mutant disrupting the cytokinin pathway display substantially compromised hrpS and hrpL transcription in plant, indicating the involvement of auxin and cytokinin in regulating T3SS in P. agglomerans pv. gypsophilae (Chalupowicz et al. 2009).
Phytobacterial T3 genes are suppressed by nutrient rich media but rapidly induced after being transferred into minimal media (Tang et al. 2006). Even though chemically defined minimal media have been widely used to induce phytobacterial hrp genes, it is hard to identify specific component that is responsible for the induction (Kim et al. 2009). Multiple environmental factors, such as temperature, medium components, osmolaric strength, pH and nutritional conditions, affect T3SS gene expression in liquid media (Huynh et al. 1989; Rahme et al. 1992; van Dijk et al. 1999). The best temperature for the induction of T3 genes in P. syringae is between 20 and 30°C (van Dijk et al. 1999). Complex nutrient sources in rich media and high pH and osmolarity are responsible for T3SS gene repression. On the other hand, the physiological and chemical environment inside plant is thought to be mimicked by the T3SS-inducing media that are low pH and osmotic, nutritionally poor with single carbon source (Huynh et al. 1989). The T3-inducing medium composition varies between different pathogens, which may suggest that the apoplastic conditions of different host plants are different. For example, fructose and sucrose induce P. syringae T3 genes better than other carbon sources tested, while the induction of Xanthomonas T3 genes needs sucrose and sulfur-containing amino acids (Huynh et al. 1989; Schulte et al. 1992). In addition, using a chemostat system, iron has been recently disclosed to induce the transcription of hrpL and an effector gene hopAA1-1 while repress the bacterial growth in the minimal medium (Kim et al. 2009).
Perspectives
To date, tremendous progress has been made in better understanding the phytobacterial T3SS, especially for the novel functions of T3 effectors and the new components controlling hrp genes. Despite studies in the previous years have identified a large number of T3SS-regulating genes in various phytopathogenic bacteria, very little is known about how these genes regulate downstream T3SS pathways. Even though a group of two-component systems have been demonstrated to be involved in T3SS regulation, the key question about the nature of the T3SS signals and how bacteria perceive the signals largely remains to be elucidated in the future.
