Type III secretion chaperones of Pseudomonas syringae protect effectors from Lon-associated degradation


E-mail hutcheso@umd.edu; Tel. (+1) 301 405 5498; Fax (+1) 301 314 9489.


The hrp type III secretion system (TTSS) of Pseudomonas syringae translocates effector proteins into the cytoplasm of host cells. Proteolysis of HrpR by Lon has been shown to negatively regulate the hrp TTSS. The inability to bypass Lon-associated effects on the regulatory system by ectopic expression of the known regulators suggested a second site of action for Lon in TTSS-dependent effector secretion. In this study we report that TTSS-dependent effectors are subject to the proteolytic degradation that appears to be rate-limiting to secretion. The half-lives of the effectors AvrPto, AvrRpt2, HopPsyA, HopPsyB1, HopPtoB2, HopPsyV1, HopPtoG and HopPtoM were substantially higher in bacteria lacking Lon. TTSS-dependent secretion of several effectors was enhanced from Lon mutants. A primary role for chaperones  appears  to  be  protection  of  effectors  from Lon-associated degradation prior to secretion. When coexpressed with their cognate chaperone, HopPsyB1, HopPsyV1 and HopPtoM were at least 10 times more stable in strains expressing Lon. Distinct Lon-targeting and chaperone-binding domains were identified in HopPtoM. The results imply that Lon is involved at two distinct levels in the regulation of the P. syringae TTSS: regulation of assembly of the secreton and modulation of effector secretion.


Type III secretion systems (TTSS) are used by many pathogenic Gram-negative bacteria to translocate virulence proteins, known as effectors, into host cells (Cornelis and Van Gijsegem, 2000). TTSS consist of a needle-like translocation apparatus (also known as a secreton) assembled from polypeptides conserved between diverse bacterial species. The array of effectors translocated by the TTSS of each pathogen is distinct but function analogously to suppress cellular defences or enhance parasitism of host cells (Cornelis and Van Gijsegem, 2000; Collmer et al., 2002; Greenberg and Vinatzer, 2003; Jin et al., 2003; Bretz and Hutcheson, 2004). The number of effectors produced by pathogenic strains varies from as few as nine in Yersinia (Cornelis, 2002) to as many as 60 for individual Pseudomonas syringae strains (Collmer et al., 2002; Greenberg and Vinatzer, 2003; Jin et al., 2003).

The mechanism of effector secretion is thought to be similar for all TTSS. A cryptic N-terminal secretion signal directs effectors to the secretion port (Cornelis, 2003; Ramamurthi and Schneewind, 2003). Both mRNA-linked and peptide sequence-dependent mechanisms for secretion of effectors have been proposed (Anderson and Schneewind, 1997; Anderson et al., 1999; Lloyd et al., 2001; 2002; Ramamurthi and Schneewind, 2003). Irrespective of the mechanism, effectors appear to be introduced and directed through the central channel of the secreton in an adenosine triphosphate (ATP)-dependent process (Woestyn et al., 1994; Pozidis et al., 2003). The size restriction of the secreton channel (∼20 Å) (Kubori et al., 2000; Blocker et al., 2001) predicts that effectors are likely to be ‘unfolded’ during the secretion process. Consistent with this model, peptide structures with diameters greater than the ∼20–30 Å obstructed the secretion channel (Feldman et al., 2002).

Many TTSS effectors have been shown to interact in the bacterial cytoplasm with small acidic proteins called type III secretion chaperones that are proposed to act as secretion pilots or as factors that maintain effectors in a ‘secretion-competent’ state (Page and Parsot, 2002; Feldman and Cornelis, 2003). The crystal structure of SptP or YopE interacting with their respective chaperones SicP or SycE, revealed that effectors were maintained in an ‘unfolded’ state by wrapping around the chaperone (Stebbins and Galan, 2001; Birtalan et al., 2002). Another proposed role of chaperones has been to block enzymatic or toxic activity of effectors within the bacterial cytoplasm. More recently, chaperones such as SicA, FlgN, SycD/LcrH and IpgC have been proposed to control the timing of effector secretion, therefore establishing a possible secretion hierarchy (Boyd et al., 2000; Karlinsey et al., 2000; Tucker and Galan, 2000; Anderson et al., 2002; Mavris et al., 2002).

Pseudomonas syringae strains express a TTSS encoded by the hrp pathogenicity island that is closely related to its counterpart found in strains of Yersinia spp. (Hutcheson, 1999; Collmer et al., 2000). Analysis of the regulatory network associated with the hrp TTSS showed that HrpR and HrpS are truncated enhancer binding proteins that interact to form an activation complex for the hrpL promoter (Hutcheson et al., 2001). HrpL, an alternative sigma factor, uses a conserved promoter sequence to direct the expression of all known genes encoding components of the hrp translocation apparatus as well as the translocated effectors and their cognate chaperones (Xiao and Hutcheson, 1994; Xiao et al., 1994; Fouts et al., 2002). Expression of the TTSS of P. syringae is minimal under nutritionally rich growth conditions, but is activated in planta or under conditions that mimic the host environment, such as acidic minimal salts media (i.e. inducing conditions) (Xiao et al., 1992). During a search for a postulated negative regulator of the hrp TTSS, Lon protease was linked to the degradation of HrpR under repressive conditions to modulate expression of hrpL (Bretz et al., 2002). Regulated proteolysis by Lon functions in several regulatory networks in eubacteria, and Lon has also been associated with the degradation of abnormally folded proteins (Gottesman, 1996). Interestingly, lon null mutants of P. syringae exhibited substantially higher secretion of an effector and induced plant defence responses faster than the wild-type strain (Bretz et al., 2002). Ectopic expression of HrpRS, however, has not reproduced this phenotype, suggesting that Lon could have an additional role in the regulation of type III secretion in P. syringae strains.

Here we report that Lon-associated regulation of the P. syringae TTSS also involves regulated proteolysis of effectors prior to secretion that can be suppressed by the cognate chaperone. Lon-associated degradation involved a specific region the effector that was distinct from its chaperone-binding domain. The data advance our understanding of the factors impacting TTSS and the essential role of chaperones in protecting effectors from degradation prior to secretion.


Secretion through the hrp TTSS is regulated beyond HrpL

To determine whether the role of Lon in the regulation of the hrp TTSS is limited to the control of hrpL expression, we attempted to bypass the known regulatory activity of Lon by ectopic expression of HrpL. If the role of Lon were solely to regulate hrpL expression by degrading HrpR during repressive conditions, then vector-directed ectopic expression of hrpL should result in the same phenotype as a Δlon mutant. Secretion of natively expressed AvrPto could be detected at low levels in immunoblots of concentrated culture filtrates from Lon+ DC3000 grown under inductive conditions for hrp expression (Fig. 1A) when Lon-associated degradation of HrpR has been shown to be minimal (Bretz et al., 2002). Ectopic expression of hrpL from pMLL600 did not affect AvrPto secretion from the wild-type cells or the lon null mutant. AvrPto levels in culture filtrates of strains carrying the Plac-hrpL expression system were indistinguishable from their counterparts lacking the hrpL construct. The expression system for hrpL was confirmed to be active and insensitive to Lon. The hrpL expression construct could activate the expression of hopPsyA::lacZ in an Escherichia coli Lon+ background (data not shown), indicating that HrpL was active in Lon+ cells.

Figure 1.

Ectopic expression of hrpL does not bypass Lon-associated regulation of the hrp cluster. Levels of natively expressed AvrPto (A) or β-galactosidase expressed from pLLlac600 (B) were monitored in culture filtrates of DC3000 (Lon+), JB7 (lon::Tn) and a secretion mutant, A9 (hrpA::Tn) expressing hrpL from pMLL600, or carrying the expression vector pDSK600 (Empty vector). Overnight cultures were diluted into M63Fructose with 1% casein hydrolysate and grown to an OD600 of 1.0. The cells were collected and washed in M63Fructose and diluted to an OD600 of 0.6 in hrp-inductive media (M63Fructose, pH 5.5) and grown for 6 h. The supernatants were concentrated 50× in centrifugal devices and 10 µg of total protein from each sample were fractionated by electrophoresis in a 12% SDS-PAGE. Levels of anti-AvrPto or anti-β-galactosidase reactive proteins were monitored using immunoblots. AvrPto migrated as a 18 kDa polypeptide and β-galactosidase as a 115 kDa polypeptide.

As before (Bretz et al., 2002), up to 50-fold higher levels were detected in culture filtrates of the Lon JB7 mutant. The apparent secretion of AvrPto did not appear to be due to lysis. A similar construct constitutively expressing the lacZYA operon (pLLlac600) exhibited equal activities in both DC3000 and JB7 as steady state levels of LacZ activity were nearly identical (approximately 350 Miller Units) and similar levels of LacZ were detected in immunoblots of whole cell lysates (Fig. 1B). In parallel experiments, LacZ was not detected in concentrated culture filtrates of DC3000 and JB7 carrying the aforementioned constitutively expressed Lac construct (Fig. 1B).

Consistent with the absence of an effect on AvrPto secretion, ectopic expression of HrpL also did not affect the development of the hypersensitive response (HR) in leaves of Nicotiana tabacum (data not shown) whereas the Lon mutant elicited a more rapid HR as reported previously (Bretz et al., 2002). The observation that ectopic expression of HrpL could not bypass the Lon-mediated repression of the hrp TTSS activity coupled with the previous observations of enhanced secretion of ectopically expressed AvrPto from a Lon mutant (Bretz et al., 2002) implied that Lon had a second activity in the regulation of the hrp TTSS.

Pseudomonas syringae effectors are subject to Lon-associated degradation

The role of Lon in the proteolysis of abnormally folded proteins (Gottesman, 1996) raised the possibility that Lon could affect the stability of AvrPto and other TTSS effectors as they are predicted to be in an unfolded state prior to secretion (Stebbins and Galan, 2001; Feldman et al., 2002). The stability of AvrPto was compared in the Lon+ DC3000 (pMLAvrPto600) and Lon JB7 (pMLAvrPto600) ectopically expressing AvrPto from the Lon-insensitive pDSK600 expression system described above. AvrPto levels were monitored by probing whole-cell extracts with anti-AvrPto polyclonal sera after inhibition of translation. The apparent stability of AvrPto in translationally inhibited cells was substantially higher in the lon mutant. The observed half-life of AvrPto was 8 min in DC3000, but was greater than 30 min in JB7 (Fig. 2). No degradation products were evident (data not shown). Notable differences in the initial levels of AvrPto were observed between Lon+ and Lon strains. Roughly five times more AvrPto could be detected in JB7 when compared with DC3000 (Fig. 2). As the rate of synthesis would be expected to be equivalent in the two strains, the difference in accumulation of AvrPto is likely to be attributable to faster degradation of AvrPto in the presence of Lon. In contrast, the levels of a similarly expressed bacterial alkaline phosphatase (BAP-FLAG) were indistinguishable between the two strains (Fig. 2) and the half-life of BAP-FLAG was found to be greater than 30 min in either DC3000 or JB7 (Fig. 2).

Figure 2.

Effect of Lon protease on stability of effectors in P. syringae. Overnight cultures of DC3000 (wild-type) or JB7 (lon::Tn) transformants carrying pMLAvrPto600, pSGAS, pAvrRpt2D or pBAPFLAG600 ectopically expressing the effectors AvrPto, HopPsyA, AvrRpt2 or BAP-FLAG respectively, were diluted into M63 fructose media and grown at 25°C to an OD600 of 0.6. Translation was stopped by the addition of tetracycline to a final concentration of 200 µg ml−1 and 250 µl samples were taken at the times (min) indicated. The cells were collected by centrifugation and resuspended in 50 µl of SDS-PAGE loading buffer to lyse the cells. Whole-cell lysates were separated by 12% SDS-PAGE, blotted onto supported nitrocellulose and immunoprobed with polyclonal anti-AvrPto, anti-HopPsyA, anti-AvrRpt2 or monoclonal anti-FLAG sera. Levels of effectors were estimated in scanned images using NIH Image 1.59 and normalized to total cells. The abundance of effectors (DC3000, open circles; JB7, filled squares) was calculated relative to their initial amounts at time 0. Estimated levels from single experiment are shown, but similar results were obtained in at least three additional experiments.

To assess whether other effectors might also be targets for Lon-associated degradation, the stability of additional effectors was monitored in P. syringae strains. Using polyclonal antibodies raised against the purified polypeptides (gifts of S. Heu, and J. Greenberg respectively), HopPsyA from Psy61 and AvrRpt2 from P. syringae pv. maculicola ES4326 were also found to be unstable in Lon+P. syringae cells when ectopically expressed from vector Plac promoters. As with AvrPto, the initial levels of both HopPsyA and AvrRpt2 were 2–5 times higher in the Lon mutant than in Lon+ DC3000 (Fig. 2B and C) and Pss61 (data not shown). The observed half-lives of both effectors were considerably longer in the Lon mutant. HopPsyA exhibited an apparent half-life of 6 min in the wild-type strains that increased to >30 min in the Lon mutant. Similarly, the calculated half-life of AvrRpt2 increased threefold in the Lon mutant. Each of these effectors could be detected in culture filtrates of Lon Pss61-KL11 transformants carrying the above ectopic expression constructs but only minimal levels were detected in culture filtrates from Lon+ Pss61 transformants (data not shown).

Lon modulates hrp TTSS in an E. coli background

Previous studies had shown that E. coli MC4100 expressing the hrp cluster carried by pHIR11 could elicit effector-dependent responses in tobacco leaves (Huang et al., 1988; Pirhonen et al., 1996), and therefore, assembles a functional TTSS. Because a P. syringae lon allele could complement a mutation in its E. coli counterpart (Bretz et al., 2002), inactivation of lon in E. coli should modulate effector levels and stability similarly to that observed in P. syringae. SG22622 (wt) and SG22623 (Δlon) were transformed with pHIR11 and their ability to elicit an effector-dependent response in tobacco plants was tested. As observed in P. syringae, Lon SG22623 (pHIR11) elicited a visible programmed cell death response in 14 h, whereas Lon+ SG22622 (pHIR11) took over 30 h to produce a similar response (data not shown). These results suggested that the role of Lon in the regulation of the hrp TTSS in E. coli could be similar to that observed in P. syringae.

To determine if inactivation of lon also caused enhanced secretion of effectors from E. coli transformants expressing the P. syringae hrp TTSS, levels of AvrPto, AvrRpt2 and HopPsyA, ectopically expressed from the vector's Lac promoters, were monitored in culture filtrates. Only small amounts of AvrPto, AvrRpt2 and HopPsyA were detected in culture filtrates of Lon+ SG22622 (pHIR11) transformants (Fig. 3). AvrPto, AvrRpt2 and HopPsyA, however, were more abundant in culture filtrates of the Lon SG22623 (pHIR11) and were not obvious in culture filtrates of Lon TTSS SG22623 (pLAFR3) (Fig. 3; -hrp) consistent with minimal lysis of cells. These results suggest that the role of Lon in the regulation of effector secretion is similar in the E. coli and P. syringae systems.

Figure 3.

P. syringae effectors are secreted from an E. coli lon background. AvrPto, HopPsyA and AvrRpt2 were ectopically expressed from pDSK600, pYXSS or pDSK519 in E. coli SG22622 (wild-type) or SG22623 (Δlon) carrying either pLAFR3 or pHIR11 expressing a functional hrp TTSS. Overnight cultures were diluted into M63Fructose and cells were grown 5 h at 37°C. The supernatants were concentrated as in Fig. 1 and protein levels determined using the micro-BSA protein assay kit (Pierce, Ca). Total proteins, 10 µg, were loaded and separated by 12% SDS-PAGE and anti-AvrPto, anti-HopPsyA or anti-AvrRpt2 reactive proteins were identified in immunoblots. AvrPto migrated at an 18 kDa polypeptide; HopPsyA migrated at a 39 kDa polypeptide; and AvrRpt2 migrated at a 28 kDa polypeptide. Equivalent results were obtained in three different experiments.

Lon affects stability of effectors in E. coli

To assess whether Lon plays a role in the stability of effectors in E. coli, the half-lives of AvrPto, HopPsyA and AvrRpt2 were determined in translationally inhibited Lon+ SG22622 and Lon SG22623 as before. For all three of these effectors, a two- to fivefold increase in their observed half-life was detected in the lon null mutant as observed in P. syringae(Fig. 4). Similar to what was observed in P. syringae, the initial amounts of the polypeptides were lower in the wild-type than the lon mutant cells for most effectors. The abundance of effectors in the wild-type cells was between 20% and 40% of that observed in the Δlon cells (Fig. 4).

Figure 4.

Lon affects the stability of P. syringae effectors in E. coli. E. coli SG22622 (WT) or SG22623 (Lon) carrying AvrPto, AvrRpt2 or HopPsyA expressed as in Fig. 3, and HopPtoG, or HopPsyE expressed from pTrcHis2 as His fusions were grown to an OD600 of 0.6 and translation was stopped by the addition of excess chloramphenicol (200 µg ml−1). Whole cell lysates were probed in immunoblots using corresponding polyclonal or anti-His antibodies and half lives calculated as in Fig. 2. HopPsyE migrated as a 50 kDa polypeptide and HopPsyG resolved as a 35 kDa polypeptide. Equivalent results were obtained in three different experiments.

The stability of effectors from other P. syringae strains was also affected by Lon. HopPtoG from P. syringae pv. tomato 5846 and HopPsyE1 from P. syringae pv syringae DH015 were cloned into pTrcHis2 to generate a C-terminal 6xHis epitope tag and the stability of each fusion was monitored in SG22622 and SG22623 using anti-His antibodies. A comparison between the half-life of native AvrPto and 6xHis-tagged AvrPto showed that the fusion did not alter the stability of AvrPto (data not shown). In Lon+ strains, some effectors had half-lives less than 2 min whereas other where longer than 6 min (Fig. 4). In all cases, the measured stability of the epitope-tagged effector was two- to fourfold higher in Δlon mutants. These results suggest that effectors are generally susceptible to degradation by Lon prior to secretion, irrespective of the source and host strain.

Chaperones reduce Lon-associated degradation of effectors

In P. syringae three chaperones, ShcA, ShcM and ShcV, have been demonstrated to be essential for secretion of their substrate effector, HopPsyA, HopPtoM and HopPtoV respectively (van Dijk et al., 2002; Badel et al., 2003; Wehling et al., 2004). ShcB1, a partial homologue of ShcA, has been reported to be the chaperone for HopPsyB1, a partial homologue of HopPsyA (Charity et al., 2003). In these cases, P. syringae effectors were secreted from wild-type cells only in the presence of their chaperone. To investigate whether their secretion was possibly due to a stabilizing effect from their chaperones, the half-lives of HopPsyB1, HopPtoM and HopPsyV1, produced by an allele of hopPsyV (Charity et al., 2003), were determined in Lon+ SG22622 or Lon SG22623 in the presence or absence of their cognate chaperones, ShcB1, ShcM and ShcV1 respectively. Each chaperone was cloned individually into pDSK519 such that expression was directed from the vector's lac promoter and the construct was transformed into the E. coli strains carrying one of the previously mentioned effector constructs. In each case, the coexpression of the chaperone had the expected stabilizing effect on its cognate effector. The apparent half-life of HopPsyB1 increased from under 1 min in SG22622 (pLLhopB1-trc)(pDSK519) to 5 min in SG22622 (pLLhopB1-trc)(pLLshcB1D) (Fig. 5). This effect was only observed in Lon+ SG22622. HopPsyV1 was also stabilized by its apparent chaperone. Expression of ShcV1 increased the half-life of HopPsyV1 to >30 min which is comparable to its half-life in Lon SG22623. Likewise, the half-life of HopPtoM increased from 5 min in the absence of its chaperone to >20 min when ShcM was expressed (Fig. 5). The effects of these chaperones were specific, as expression of a heterologous chaperone did not affect the stability of the effectors HopPsyB1 or HopPsyV1 (data not shown). The results suggest that a primary role for chaperones in TTSS could be protection of effectors from Lon-associated degradation, therefore augmenting their cytoplasmic abundance and potential for secretion.

Figure 5.

Effectors are stabilized by their chaperone. Overnight cultures of E. coli ectopically expressing HopPsyB1, HopPsyV1, or HopPtoM carrying their cognate chaperone (Chap) or empty vector in SG22622 (WT) or SG22623 (Lon) were diluted and grown in LB at 37°C until they reached an OD600 of 0.6 and the half-life of each effector was estimated as in Fig. 2. The estimated half-lives (min) are shown below the corresponding immunoblot. HopPsyB1, HopPsyV1 and HopPtoM migrated as a 40 kDa, 45 kDa and 75 kDa polypeptides respectively. The results presented are representative of at least three experiments.

The Lon-targeting domain of HopPtoM does not overlap its chaperone-binding domain

The manner in which chaperones protect effectors from degradation is unknown. Other Lon-degraded proteins, such as SulA and UmuD, are targeted to Lon by specific amino or carboxy terminal motifs (Gonzalez et al., 1998; Ishii and Amano, 2001). It is possible that a TTSS chaperone protects its cognate effector by masking a Lon-targeting motif. Alternatively, the chaperone could be forming a complex with the effector that is immune to degradation or directly inhibiting Lon activity. To test the possibility that the chaperone directly inhibits Lon, the activity of RcsA was monitored in strains expressing the chaperones ShcB1, ShcV1 or ShcM from pDSK519. RcsA is a transcriptional activator of cpsB whose activity is regulated by Lon (Gottesman et al., 1985). A chromosomal cpsB::lacZ reporter that is only expressed when RcsA is not degraded by Lon was used to monitor Lon activity. Consistent with the specificity of activity for the chaperones, none of the chaperones had an effect on the Lac phenotype of the indicator strains, indicating that the chaperones do not generally inhibit Lon activity (data not shown).

The apparent chaperone-binding domain (CBD) of HopPtoM had previously been localized to an internal 300 aa domain between residues 100–400 (Badel et al., 2003). To determine if the Lon-targeting domain of HopPtoM overlaps this apparent CBD, His-tagged derivatives of HopPtoM were created that carried either amino or carboxy terminal deletions that mirrored those used in the previous analyses (Fig. 6A). One of the constructs (Δ400) expressed the C-terminal region of HopPtoM that did not interact with ShcM whereas the other two constructs (Δ200 and N200) contained coding sequence for regions that did interact (Badel et al., 2003). When the constructs were expressed in Lon+ cells, the truncations expressing the C-terminus of the peptide were degraded (Fig. 6B). Among these, the Δ400 truncation lacking any of the CBD exhibited a short half-life (3 min). In contrast, the truncation that expressed only the N-terminal 200 amino acids and included a portion of the CBD did not appear to be degraded by Lon as the half-lives were equal in both strains (>30 min)(Fig. 6B). These results suggest that a Lon-targeting domain is located in the C-terminal 312 amino acids of HopPtoM and does not overlap with the CBD.

Figure 6.

The C-terminus of HopPtoM targets the peptide for degradation.
A. Schematic diagram of HopPtoM truncations used in this study. CBD, chaperone-binding domain.
B. N-terminal truncations lacking the first 200 or 400 amino acids (HopPtoMΔ200, and HopPtoMΔ400 respectively) or a C-terminal truncation expressing only the first 200 amino acids (HopPtoMN200) were expressed in SG22622 (WT) and SG22623 (Lon) or SG22622 carrying pShcMD (Chap). The half-lives of each derivative shown below the corresponding immunoblot were estimated as in Fig. 2, using polyclonal antibodies raised against HopPtoM. Equivalent results were obtained in three different experiments. The reduced level of HopPtoMΔ400 observed in the 8 min sample from the lon mutant was not observed in the other replicates.

To verify that the C-terminus of HopPtoM could target the peptide for Lon-associated degradation, fusions between HopPtoMΔ400 and the maltose-binding protein (MBP) were constructed and tested for sensitivity to Lon. MBP–LacZ has previously been shown to be immune to degradation by Lon, and a similar assay has been used to determine the Lon-targeting motif in SulA (Ishii and Amano, 2001). A MBP:LacZ fusion expressed in Lon+ SG22622 or Lon SG22623 had half-lives of roughly 35 min. In contrast, when HopPtoMΔ400 was fused to MBP, it was degraded in the Lon+ background with a half-life of 8.5 min (Fig. 7), consistent with the presence of a C-terminal targeting domain. This fusion was only slowly degraded in the Lon mutant background as its half-life was 24 min (data not shown). Further studies showed that the targeting motif localized to the C-terminal 100 residues of the peptide. This region fused to MBP (MBP-HopPtoMΔ612) was degraded with a half-life of roughly 6 min in Lon+ cells whereas its half-life was greater than 30 min in a Lon background (data not shown). Taken together these results support the hypothesis that the Lon-targeting domain of HopPtoM is located in the C-terminal region distinct from the CBD.

Figure 7.

An MBP fusion to the C-terminus of HopPtoM is degraded.
A. Schematic diagram of MBP fusions used in this study.
B. The Δ400, 612 or N200 truncations of HopPtoM (Fig. 6) were amplified by PCR using specific primers and used to create fusions with the C-terminus of MBP in the vector pMAL-p2x (New England Biolabs) following manufacturer's instructions. Cells expressing MBP::LacZ or the different fusion proteins were induced for 2.5 h. Translation was stopped using chloramphenicol (200 µg ml−1) and levels of the fusions estimated in 250 µl samples at the indicated times. The relative amounts of the fusion proteins were monitored in immunoblots using anti-MBP antibodies, and the half-lives were estimated from densitometry analyses as in Fig. 2.


Lon has been shown previously to regulate assembly of the hrp-encoded TTSS in P. syringae strains through its effects on the activity of the hrpL promoter (Bretz et al., 2002). Regulated proteolysis of the transcriptional activator HrpR controlled the activity of the hrpL promoter, thereby modulating expression of the Hrp regulon. Interestingly, the lon mutants also secreted several effectors consistent with a TTS+ phenotype that could not be reproduced by ectopic expression of hrpRS or hrpL. This suggested that Lon had an additional activity in the regulation of hrp TTSS. Consistent with this hypothesis, eight distinct effectors isolated from several P. syringae strains were found to be unstable in Lon+P. syringae and/or E. coli strains, but were relatively stable in the corresponding Lon mutants. Consistent with directed degradation of effectors, a C-terminal 100 aa region was identified in one of these effectors that targeted the effector for Lon-associated degradation and several chaperones were shown to suppress the degradation of the cognate effector only in Lon+ cells. Thus it appears that Lon has a dual function in the regulation of the hrp TTSS of P. syringae strains by: (i) regulating the assembly of the TTSS and expression of effectors through regulated proteolysis of HrpR (Bretz et al., 2002); and (ii) as reported here controlling the accumulation of effectors prior to secretion through turnover.

Lon belongs to a family of cytosolic ATP-dependent proteases that are highly conserved among prokaryotes and Archea, and are also found in some organelles of eukaryotic cells (reviewed in Wickner and Maurizi, 1999). Like other bacterial energy-dependent proteases, Lon forms a multimeric complex that couples ATP-dependent protein unfolding with an endopeptidase activity (Van Melderen et al., 1996; Wickner and Maurizi, 1999) to rapidly degrade targeted proteins without forming partially degraded intermediates (Nishii et al., 2002). Each of these activities is assigned to independent domains separated by a ‘sensor and substrate-discrimination’ domain (SSD) which functions in substrate selection. Each of these functional domains is conserved in the P. syringae Lon homologue and the P. syringae Lon can complement its E. coli counterpart (Bretz et al., 2002). P. syringae lon::Tn mutants exhibited increased cell length and greater UV sensitivity relative to wild-type similar to Lon null mutants of other bacteria. Thus, the P. syringae Lon appears to function similarly to its homologues of other bacteria.

Lon is well known for its involvement in the degradation of unstable regulatory proteins, such as FlhC/FlhD, SulA and RcsA/RcsB (Stout et al., 1991; Higashitani et al., 1997; Claret and Hughes, 2000). Another principal activity of Lon is the degradation of misfolded or abnormal proteins, such as those that occur because of the incorporation of irregular amino acids or temperature-sensitive mutations (Gottesman, 1996). Because TTSS-linked effectors are predicted to be in an unfolded state prior to secretion (Stebbins and Galan, 2001; Page and Parsot, 2002; Feldman and Cornelis, 2003), they would be strong candidates for degradation by proteases such as Lon. Among the effectors studied, all had relatively short half-lives in Lon+ cells. Consistent with the observed degradation rates, steady state levels of effectors expressed from a Lon-insensitive lacUV5 promoter were 20–50% of the levels detected in Lon mutants. Since levels of LacZ and BAP expressed from the same promoter were indistinguishable between Lon+ and Lon cells, the rate of synthesis for these ectopically expressed effectors is likely to be equivalent in Lon+ and Lon cells. The differences in the steady state levels, thus, are consistent with the observed differences in the rate of decay. The absence of proteolytic degradation products in immunoblots of Lon+ cells suggests that Lon directly functions in the degradation of the effectors. It is possible, however, that another protease is activated by Lon that then functions in the degradation of the effectors.

Our results are consistent with other studies proposing that chaperones act as stabilizing agents for their effectors. In Yersinia and Salmonella strains, cytoplasmic levels of some effectors are higher in presence of their cognate chaperone consistent with protection from proteolytic degradation (Frithz-Lindsten et al., 1995; Fu and Galan, 1998). Our studies indicated that Lon-associated degradation of the tested P. syringae effectors could be suppressed by their cognate chaperones. The effect was specific to a chaperone–effector interaction as only the cognate chaperone stabilized the effector. It is likely, then, that one of the major roles for chaperones is in fact to protect effectors from Lon-associated degradation. Consistent with this hypothesis, the chaperones had little effect on the half-lives of effectors in strains lacking Lon. In contrast to many effectors of mammalian pathogens, most effectors expressed by P. syringae strains do not have an obvious cognate chaperone. In P. syringae DC3000, only seven out of the 40–58 postulated effectors have been associated with a chaperone, as identified by physical interaction studies or their genetic proximity. Why some effectors have a chaperone, and most do not is an enigma. It is possible that chaperones in P. syringae do not fit the general notion that they must be proximal to their corresponding effector (Page and Parsot, 2002; Feldman and Cornelis, 2003), and hence the complete array of chaperones has not been characterized. However, there are few, if any, chaperone-like gene products reported among the currently known HrpL-regulated genes in P. syringae strains (Fouts et al., 2002; Guttman et al., 2002; Zwiesler-Vollick et al., 2002). Interestingly, E. coli strains carrying the cloned P. syringae hrp cluster had enhanced levels of the ectopically expressed effectors consistent with the expression of a presently unidentified stability factor.

The mechanism by which P. syringae chaperones protect effectors from Lon-associated degradation does not appear to be due to the masking of the targeting motif in the case of HopPtoM. ShcM has been reported to interact with an internal domain of HopPtoM located between residues 100–300 (Badel et al., 2003) whereas the Lon-targeting domain in HopPtoM was shown here to be located in a carboxy terminal domain that does not overlap with the previously established CBD. Most likely, this chaperone–effector complex has a Lon-insensitive conformation. Crystal structures of other chaperone–effector interactions have revealed a high degree of secondary structure (Birtalan and Ghosh, 2001; Stebbins and Galan, 2001; Evdokimov et al., 2002). As unfolding of the substrate is an essential part of Lon proteolysis (Van Melderen et al., 1996), the formation of a stable complex structure with the chaperone would inhibit its degradation by Lon.

The sequence responsible for targeting HopPtoM for degradation was contained within the carboxy-terminal 100 residues. Other studies have found that a motif formed by the carboxy terminal eight residues in SulA targets that protein for degradation by Lon (Ishii and Amano, 2001). There was no sequence similarity between the C-terminus of HopPtoM and SulA, and motif scans using various algorithms did not detect in any other similarities. CcdA, another substate targeted for Lon-mediated degradation by a C-terminal signal did not share any sequence similarity with SulA (Van Melderen et al., 1996), indicating that these targeting sequences can be quite diverse. In addition, other Lon-degraded substrates can be targeted for degradation by signals contained at the N-terminus (Gonzalez et al., 1998; Smith et al., 1999). Indeed, preliminary results suggest that at least one other effector is targeted for degradation by an N-terminal motif (L.C. Losada and S.W. Hutcheson, unpubl. results). A conserved sequence could not be discerned among the studied effectors at either terminus, implying that the mechanism for targeting effectors for degradation by Lon is more complex than a conserved amino acid sequence and is currently being examined further.

Ordered secretion of TTSS-dependent effectors has been proposed for several mammalian pathogens, such as Yersiniae and Salmonella (Boyd et al., 2000; Darwin and Miller, 2001; Page and Parsot, 2002; Wulff-Strobel et al., 2002; Feldman and Cornelis, 2003). This hierarchy appears to link secretion of an effector with a stage of infection. Chaperones have been demonstrated to play a crucial role in determining the hierarchy of effector secretion by either indirectly controlling the transcription of specific effector genes or the affinity of the effectors for secretion apparatus (Boyd et al., 2000; Karlinsey et al., 2000; Tucker and Galan, 2000; Anderson et al., 2002; Mavris et al., 2002). As suppression of Lon activity increased the secretion of effectors from the hrp TTSS, it appears that Lon-associated degradation could be rate-limiting to effector secretion [but other effects on the secretion system itself (Boddicker and Jones, 2004) can not be completely excluded]. Chaperones may help establish effector secretion order by providing differential levels of protection from degradation. Whether ordered secretion of effectors occurs in P. syringae, however, has not been established because of the minimal secretion of effectors from  Lon+ cells  and  the  lability  of  translocated  effectors in plant cells (see Nimchuk et al., 2003; Bretz and Hutcheson, 2004).

In conclusion, Lon plays an important role not only in the regulation of assembly of the TTSS, but also in controlling the secretion of effectors through that system. The impact of Lon on effector stability may not be unique to P. syringae effectors as the effect is also observed in E. coli strains. Effectors of many bacteria are secreted when coexpressed with their cognate chaperone (Cheng et al., 1997; Fu and Galan, 1998; Tucker and Galan, 2000; van Dijk et al., 2002; Badel et al., 2003; Wehling et al., 2004). YopE and SptP, for example, are readily secreted in the presence of their corresponding chaperones, SycE and SicP respectively, but only a small fraction is present in the absence of the chaperone (Cheng et al., 1997; Fu and Galan, 1998). The mechanism for effector degradation and chaperone stabilization has not been established in these systems but seems likely to be due to Lon-associated degradation. Lon has been shown to influence the assembly and activity of the SPI TTSS of Salmonella (Takaya et al., 2002; Boddicker and Jones, 2004). Here we propose a model in which effectors would have two distinct fates in the bacterial cell: secretion through the TTSS or degradation by Lon. Differential turnover of effectors could provide an alternative and novel mechanism for the hierarchical secretion of effectors.

Experimental procedures

Bacterial stains and media

Strains and plasmids used in this study are described in Table 1. Bacteria were routinely grown on King's B medium (Atlas, 1993). Plasmids were propagated in E. coli DH5α. E. coli strains were grown at 37°C, and P. syringae strains were grown at 25°C. Luria–Bertani and M63 minimal salts media were used as described previously (Sambrook and Russell, 2001). M63 medium was supplemented with 1 mM MgSO4 and 1% fructose (M63F). The following antibiotics were added where needed at the indicated concentrations (in micrograms per millilitre): ampicillin, 200; kanamycin, 50; spectinomycin, 100; tetracyline, 25; nalidixic acid, 50; rifampicin, 200; and chloramphenicol, 30.

Table 1.  Strains, plasmids and primers used in this study.
Strain or plasmidGenotype or phenotypeReference or source
E. coli
 DH5αend A1 hsd R17 (rk- mk-) sup E44 thi-1 rec A1 gyr A96 rel A1 Δ(arg R-lac ZYA) U169φ80dlacZΔM15Invitrogen
 SG22622cpsB::lacZΔara malP::lacIqS. Gottesman
 SG22623SG22622 Δlon-510S. Gottesman
 SLR400araD139Δ(ara leu)7697 derivative of MC4100S. Benson
 TOP10FmcrA Δ(mrr-hsd RMS-mcr BC)φ80lac ZΔM15 Δlac X74 rec A1 araΔ139 Δ(ara-leu)7697 gal U gal K rps L (StrR) end A1 nup GInvitrogen
Pseudomonas syringae
 A9DC3000 hrpA mutant, HR, Rifr, KanrWei et al. (2000)
 DC3000Wild-type, Rifr, HR+ Tomato and Arabidopsis pathogenCuppels (1986)
 JB7DC3000 lon::Tn, Rifr KanrBretz et al. (2002)
 Psy61Wild-type, Nalr, HR+ Weak bean pathogenBaker et al. (1987)
 Psy61-KL11Nalr Kanr, lon::TnBretz et al. (2002)
 pAvrRpt2D1.0 kb PCR product containing avrRpt2 cloned into pDSK519This work
 pBAPFLAG600Bacterial Alkaline Phosphatase amplified from pFLAG-CTS-BAP cloned into pDSK600This work
 pDSK519Broad-host range vector, IncQ KanrKeen et al. (1988)
 pDSK600Broad-host range vector, IncQ SprMurillo et al. (1994)
 pFLAG-CTS-BAPBacterial Alkaline Phosphatase in pFLAG-CTCSigma-Aldrich
 pHIR11pLAFR3 derivative carrying P. syringae pv. syringae 61 hrp/hrc cluster, TcrHuang et al. (1988)
 pHopPsyB1trchopPsyB1 cloned into pTrcHis2This work
 pHopPsyEtrchopPsyE cloned into pTrcHis2This work
 pHopPtoGtrchopPtoG cloned into pTrcHis2This work
 pHopPtoMtrchopPtoM cloned into pTrcHis2This work
 pHopPtoMΔ400trcC-terminal 936 bp of hopPtoM cloned into pTrcHis2This work
 pHopPtoMΔ200trcC-terminal 1536 bp of hopPtoM cloned into pTrcHis2This work
 pHopPtoM200trcN-terminal 600 bp of hopPtoM cloned into pTrcHis2This work
 pHopPtoMΔ400MALMBP fusion to the C-terminal 936 bp of hopPtoM in pMAL-p2xThis work
 pHopPtoMΔ612MALMBP fusion to the C-terminal 300 bp of hopPtoM in pMAL-p2xThis work
 pHopPsyV1trchopPsyV1 cloned into pTrcHis2This work
 pLAFR3Tcr, IncP1Staskawicz et al. (1987)
HindIII-PstI fragment from pRG970, subcloned into pDSK519 and then the lac operon excised with EcoRI and cloned into pDSK600, Spr, Lac+This work
 pMAL-p2xMBP::lacZ fusionNew England Biolabs
 pMLAvrPto6000.45 kb PCR cloned as EcoR1-HindIII into pDSK600Bretz et al. (2002)
 pMLL6000.4 kb PCR product containing hrpL cloned into pDSK600This work
 pSGAS3.6 kb fragment containing hopPsyA in pYXSSHeu and Hutcheson (1993)
 pSHAMBhopPsyA cloned into pMLB1034Heu and Hutcheson (1993)
 pShcB1D0.5 kb PCR product cloned into pDSK519This work
 pShcMD0.5 kb PCR product cloned into pDSK519This work
 pShcV1D0.5 kb PCR product cloned into pDSK519This work

General DNA manipulations

Restriction enzymes were purchased from Invitrogen (Bethesda, MD, USA). T4 DNA ligase was acquired from New England Biolabs (Beverly, MA, USA) and used according to the manufacturer's recommendations. Basic manipulations were done using standard procedures (Sambrook and Russell, 2001). Polymerase chain reactions (PCRs) were performed using a PCRSprint thermal cycler (Hybaid, Ashford, UK) with 50 µl reaction volumes. Unless indicated otherwise, ProofPro polymerase (Continental Laboratory Products, San Diego) was used to amplify fragments for cloning.

Construction of 6xHis protein fusions

The genes for individual effectors were amplified from chromosomal DNA from the indicated strain using the following primers: hopPsyB1 (Psy B5) ATGAACCCGATACAAACG, TTCCAACCTGAATGCCGG; hopPtoM (Pto DC3000) ATGATCAGTTCGCGGATC, ACGCGGGTCAAGCAAGCC; hopPsyE1 (Psy W4N15) ATGAGACCTGTCGGTGGG, GACCTTATAAGACAGGAC; hopPsyV1 (Psy B728a) ATGAATATCTCAGGTCCG, AGGCTTGGCCCGGACCCT; and hopPtoG (Pto DH015) ATGAGACCCGTCGGTGGA, ATCAGCGCCAACAATCGG. The products were cloned into pTrcHis2 using the TOPO TA cloning kit (Invitrogen, Carlsbad, NJ, USA) following manufacturer's instructions. The fusion was verified by sequencing and by immunodetection as described by the manufacturer.

Cloning of effector chaperones

shcPsyB1 and shcPsyV1 were amplified from genomic DNA using the following primers: GCTCTAGACAGTTCGGG ATTGACAGG, CGGAATTCCGACAGACGCTGGAATACGG; and CTCTAGAACTGGACATGACGCTGGA, GCTCTAGA ATCGAATAGTCCCCGCCA respectively. The PCR products were cloned into pDSK519 as XbaI or BamHI fragments. shcM was amplified and cloned into pTrcHis2 using the primers ATGACCAACAATGACCAG and CTGGAATCTCCCAG GAGC. The primers GCTCTAGAGAATAAACCATGGCCCTT and GCTCTAGAGATTTAATCTGTATCAGG were used for subcloning shcM into pDSK519 or pBAD33 as an XbaI fragment. Inserts were verified by restriction analysis and sequencing.

Cloning of HopPtoM truncations

N-terminal truncations of HopPtoM were amplified using the forward primers MΔ200 GCCGGTCGTGCAAGCAAG, MΔ400 CTGAAAAGCGAACACGGT, and the reverse primer ACGCGGGTCAAGCAAGCC. C-terminal truncations were generated using the forward primer ATGATCAGTTCGCG GATC, and the reverse primer M200 aa GTATTCGC CAAGGGCAGT. All amplification products were cloned into pTrcHis2 and inserts were verified by sequencing and by immunodetection following manufacturer's directions.

Maltose-binding protein (MBP) fusions

The 3′ 936 bp of hopPtoM were amplified using the primers GCTCTAGATTAACGCGGGTCAAGCAA and GCTCTAGA AAAAGCGAACACGGTGAG, digested with XbaI and ligated to pMAL-p2x (New England Biolabs, Beverly, MA, USA) to create a C-terminal fusion to MBP following manufacturer's instructions.

β-Galactosidase assays

β-Galactosidase activity in bacterial cells was estimated by the procedures of Miller (1971).

Plant assays

Pseudomonas syringae DC3000, JB7 and A9 were transformed with pMLL600 or pDSK600. Overnight cultures grown at 25°C were harvested and diluted in M63 minimal media. N. tabacum cv Samsun leaves were infiltrated with a 108 cfu ml−1 suspension of the indicated strains as described previously (Huang et al., 1988). Infiltrated leaf panels were scored for responses beginning 2 h after inoculation and monitored for 24 h. For E. coli, SG22622 and SG22623 were transformed with pHIR11 or pLAFR3 and inoculated into N. tabacum plants in the same manner as P. syringae.

Secretion of effectors from bacterial cultures

Culture supernatants were collected as described previously (Bretz et al., 2002). Briefly, P. syringae DC3000, JB7 and A9 carrying pMLL600 or pDSK600 cells were grown in KB overnight. Cells were harvested, diluted into fresh KB medium containing selecting antibiotics and 2 mM IPTG and grown to an OD600 0.6. Cells were harvested, washed once with M63 and transferred to 50 ml of M63F at an OD600 of 0.6. After 6 h, cells in a 500 µl sample were harvested and cell pellets were resuspended in SDS-PAGE sample buffer for analysis of proteins in whole cell lysates. An identical sample was collected and it was resuspended in 10% SDS to calculate protein concentrations. Culture filtrates were obtained by centrifugation and concentrated 50× using MilliPore Ultra-free centrifugal filter devices with a 5 KDa exclusion limit. Total protein concentration in whole cell lysates was measured using the BCA Total Protein Assay kit (Pierce, Rockford, IL, USA). Total protein concentration in supernatant samples were measured using the MicroBCA Assay Kit (Pierce).

Escherichia coli SG22622 or SG22623 carrying pHIR11 or pLAFR3 and ectopically expressing AvrPto or HopPsyA from pDKS600 or pYXSS respectively, were grown overnight in KB medium with proper antibiotics. One millilitre of overnight culture was used to inoculate 50 ml of M63 media containing fructose and cultures were grown to an OD600 of 1.0. Supernatants and whole cell lysates were collected in the same manner as P. syringae.


Total proteins, 10 µg from every sample, were separated by SDS-PAGE and transferred to nitrocellulose membranes. Immunoblotting was carried out using a polyclonal antibody raised against AvrPto, AvrRpt2, HopPsyA or HopPtoM at 1:3000 dilution. Commercial antibodies, anti-His (Novagen, San Diego), anti-FLAG (Sigma, St Louis, MS, USA), anti-MBP (New England Biolabs) and secondary antibodies conjugated to horseradish peroxidase were used following manufacturer's recommendations. Cross-reactive proteins were visualized using the ECL chemiluminesence kit (Amersham-Pharmacia, Piscataway, NJ, USA).

Stability of effectors

Overnight cultures of P. syringae strains expressing AvrPto, HopPsyA or AvrRpt2 were diluted into M63 media containing fructose as the carbon source to an OD600 of 0.6 and incubated for 4 h at 25°C. After incubation, tetracycline (200 µg ml−1) was added to inhibit translation. Cells were harvested at specified times, lysed in sample buffer and fractionated in 12% SDS-PAGE gels. For E. coli, SG22622 or SG22623 were grown overnight at 37°C, diluted into fresh M63F containing 1% casein hydrolysate and grown to an OD600 of 0.6. After incubation, chloramphenicol (200 µg ml−1) was added and samples were collected as in P. syringae. The half-life effectors were calculated from the exponential decay in levels quantified in scanned images using NIH Image 1.59. Each experiment was repeated at least three times.


This work was funded by award MCB0215417 from the National Science Foundation. We thank J. Bretz, N. Ekborg, M. Howard and R. Stewart for their helpful comments on the manuscript.