HopAS1 recognition significantly contributes to Arabidopsis nonhost resistance to Pseudomonas syringae pathogens


Author for correspondence:
Jonathan D. G. Jones
Tel: +44 01603 450400
Email: jonathan.jones@tsl.ac.uk


  • Plant immunity is activated by sensing either conserved microbial signatures, called pathogen/microbe-associated molecular patterns (P/MAMPs), or specific effectors secreted by pathogens. However, it is not known why most microbes are nonpathogenic in most plant species.
  • Nonhost resistance (NHR) consists of multiple layers of innate immunity and protects plants from the vast majority of potentially pathogenic microbes. Effector-triggered immunity (ETI) has been implicated in race-specific disease resistance. However, the role of ETI in NHR is unclear.
  • Pseudomonas syringae pv. tomato (Pto) T1 is pathogenic in tomato (Solanum lycopersicum) yet nonpathogenic in Arabidopsis. Here, we show that, in addition to the type III secretion system (T3SS)-dependent effector (T3SE) avrRpt2, a second T3SE of Pto T1, hopAS1, triggers ETI in nonhost Arabidopsis.
  • hopAS1 is broadly present in P. syringae strains, contributes to virulence in tomato, and is quantitatively required for Arabidopsis NHR to Pto T1. Strikingly, all tested P. syringae strains that are pathogenic in Arabidopsis carry truncated hopAS1 variants of forms, demonstrating that HopAS1-triggered immunity plays an important role in Arabidopsis NHR to a broad-range of P. syringae strains.


The plant innate immune system provides resistance to most microbes. Effector-triggered immunity (ETI) and pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) are major components of plant innate immunity (Jones & Dangl, 2006). PTI is activated via recognition of conserved microbial patterns by cell surface-localized pattern recognition receptors (PRRs) (Zipfel et al., 2004; Jones & Dangl, 2006; Segonzac & Zipfel, 2011). Successful pathogens secrete virulence factors and suppress PTI (Jones & Dangl, 2006). Pathogenic microbes differ from each other in the range of plant species on which they can grow and reproduce. Several studies have suggested that PTI is involved in nonhost resistance (NHR) and that pathogen effector-mediated suppression of PTI partially compromises NHR (Ham et al., 2007; Ferrante et al., 2009; Lacombe et al., 2010; Zhang et al., 2010). NHR requires multiple defense layers and therefore is considered to be more durable than race-specific resistance, which is usually dependent on the recognition of a single effector (Holub & Cooper, 2004; Nurnberger & Lipka, 2005; Schulze-Lefert & Panstruga, 2011).

Although ETI is a major component of host/pathogen race specificity, little is known of the extent to which ETI plays a role in NHR. In some cases, avirulence (Avr) effectors trigger ETI in nonhost plants, suggesting a role for ETI in determining pathogen host range (Staskawicz et al., 1987; Kobayashi et al., 1989; Wei et al., 2007; Wroblewski et al., 2009). However, because Avr effectors contribute to pathogen virulence on host plants, they can be maintained in pathogen populations even though they trigger ETI on some plants (Kearney & Staskawicz, 1990).

The Gram-negative plant pathogenic bacterium Pseudomonas syringae infects and causes disease in a wide range of plant hosts (Hirano & Upper, 2000). Pseudomonas syringae pv. tomato (Pto) T1 causes bacterial speck disease in tomato (Solanum lycopersicum) but is nonpathogenic in Arabidopsis, whereas a different strain, Pto DC3000, is pathogenic in tomato and Arabidopsis. The genomes of Pto DC3000 and T1 are fully sequenced and provide a useful tool to dissect the roles of ETI in NHR in the model plant Arabidopsis (Buell et al., 2003; Almeida et al., 2009).

Materials and Methods

Plasmid constructions and mobilizations to Pseudomonas and Agrobacterium tumefaciens strains

All Pto T1 T3SS-dependent effectors (T3SEs) and hopAS1 were cloned in the pENTR-SD-D-TOPO vector (Invitrogen), sequenced and maintained in Escherichia coli DH5α. Subsequently, an LR reaction was performed using pBS46 (Swingle et al., 2008) and pER8:N-HS (a binary vector to express estradiol-induced N-terminally HA (haemagglutinin)-tagged protein in plant cells) as destination vectors for expression in Pseudomonas and Arabidopsis thaliana (L.) Heynh., respectively. Plasmid mobilizations from E. coli DH5α to Pseudomonas strains were performed using standard triparental mating as described previously (Sohn et al., 2007). Agrobacterium tumefaciens AGL1 was transformed with pER8:N-HS:hopAS1 by electroporation. Primer sequences are available upon request.

Generation of Pto T1 ΔhopAS1 and ΔavrRpt2ΔhopAS1 mutant strains

To generate Pto T1 ΔhopAS1, a central 660-bp-long fragment of the hopAS1PtoT1 gene was amplified. Stop codons were added in-frame to the 5′ end of the primers and an ApaI site was added to the 5′ end of the forward primer and a ClaI site to the 5′ end of the reverse primer. The fragment was cloned into pBAV208 and the resulting plasmid was introduced into PtoT1 as previously described (Mohr et al., 2008), giving Pto T1 ΔhopAS1, which was confirmed by sequencing to have two fragments of the hopAS1 gene, one 5′ fragment with a stop codon after bp 174 and one 3′ fragment starting with a stop codon at bp 838. Pto T1 ΔavrRpt2ΔhopAS1 was generated as follows: a 1-kb upstream fragment of hopAS1PtoT1 was PCR-amplified, digested with BamHI and XbaI, and cloned in pRK415 (Keen et al., 1988) to create pRK415AS1A. A 1.5-kb downstream fragment of hopAS1PtoT1 was PCR-amplified, digested with XbaI and HindIII, and cloned in pRK415AS1A to create pRK415AS1AB. Subsequently, a spectinomycin resistance gene was amplified, digested with HindIII and cloned in pRK415AS1AB to create pRK415AS1ASPB. pRK415AS1ASPB was mobilized into a Pto T1 strain that already had an insertion of plasmid pBA208 between the avrRpt2 hypersensitive response and pathogenicity (hrp) box and the avrRpt2 start codon and that does not trigger resistance to P. syringae 2 (RPS2)-dependent hypersensitive response (HR) (data not shown). Transformants were selected on King’s B agar media containing kanamycin (50 μg ml−1), tetracycline (10 μg ml−1) and spectinomycin (100 μg ml−1). A single colony was used to inoculate 10 ml of NYG (nutrient broth, yeast extract and glucose) broth and cultured for 2 d without antibiotics. After three subcultures, cells were spread on King’s B agar media containing kanamycin (50 μg ml−1) and spectinomycin (100 μg ml−1). After 3 d of incubation in 28°C, selected colonies were tested for tetracycline resistance. The colonies that were resistant to kanamycin (50 μg ml−1) and spectinomycin (100 μg ml−1) but sensitive to tetracycline (10 μg ml−1) were selected and DNA was isolated for PCR analysis. Replacement of the hopAS1 open reading frame (ORF) by the spectinomycin resistance gene was confirmed by sequencing the PCR product.

Generation of transgenic Arabidopsis plants

A transgenic Arabidopsis line was generated by following the protocol described previously (Clough & Bent, 1998).

In planta bacterial growth and HR assays

Leaves of Arabidopsis, turnip (Brassica rapa) or tomato were infiltrated with bacterial suspensions using a 1-mL needleless syringe for the HR assay or the bacterial growth assay. For spraying infection, Arabidopsis plants were sprayed with bacterial suspensions (0.05% silwet L-77) and covered with a transparent plastic cover for 48 h. The infected leaf samples were collected at 3 or 4 d post infection (dpi), ground in sterilized 10 mM MgCl2, serially diluted and spotted on NYG or low-salt LB (Luria-Bertani) agar medium containing appropriate antibiotics. Numbers of colonies were counted after 2 d of incubation at 28°C. The infected plants were kept in a growth chamber up to 7 dpi to observe disease symptoms. For the HR assay, the infected leaves were observed up to 48 h post infection (hpi).

Immunoblot analysis

Protein extraction and immunoblot analyses were performed as previously described (Sohn et al., 2009). In short, PFO1-T3SS strains carrying pBS46:T3SE constructs were harvested in lysis buffer (140 mM NaCl, 2.7 mM KCL, 10 mM Na2HPO4 and 1.8 mM KH2PO4, pH 7.3) after growing on King’s B agar medium (30 ug ml−1 chloramphenicol, 5 ug ml−1 tetracycline and 20 μg ml−1 gentamycin) for 2 d at 28°C. After sonication (3 × 10 s) and brief centrifugation, the supernatant was mixed with SDS-loading buffer and boiled for 4 min before being loaded in a sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) gel for immunoblot analysis. Transgenic Arabidopsis T2 lines expressing HA-HopAS1PtoT1 were grown on Murashige and Skoog (MS) agar medium (hygromycin 40 μg ml−1) for 10–14 d and then transferred to the same medium containing 50 μM estradiol. After 24 h of incubation, two to three seedlings were collected and frozen in liquid nitrogen for protein extraction. Immunoblot analysis was performed using anti-HA antibody (Roche; 3F10).

Results and Discussion

The Pto T1 T3SE hopAS1 triggers ETI in Arabidopsis

To better understand the role of ETI in host range determination, we compared the T3SE repertoires of Pto DC3000 and T1 (Almeida et al., 2009). To proliferate and cause disease in tomato, both strains require a functional type III secretion system (T3SS) (Fig. 1a). However, on Arabidopsis, wild-type Pto T1 shows as little growth as Pto T1 or Pto DC3000 ΔhrcC (hrp-conserved) mutant lacking a functional T3SS, whereas wild-type Pto DC3000 is highly virulent (Fig. 1b). We set out to investigate why Pto T1 is nonpathogenic in Arabidopsis whereas Pto DC3000 is highly pathogenic.

Figure 1.

The Pseudomonas syringae pv. tomato (Pto) T1 T3SE hopAS1 triggers effector-triggered immunity (ETI) in nonhost Arabidopsis plants. (a, b) Pto T1 is a nonhost pathogen in Arabidopsis. In planta growth was measured for Pto DC3000 and T1 wild type or ΔhrcC strains in tomato (cv Moneymaker) (a) or Arabidopsis (b) leaves. Five-wk-old leaves were hand-infiltrated using a 1-ml needless syringe with bacterial suspensions (optical density (OD)600 = 0.0001 for tomato and OD600 = 0.001 for Arabidopsis) and samples were taken at 3 d post infection (dpi) to measure bacterial number in infected leaves. Results are the mean ± SE of bacterial colonies recovered from nine leaf samples each containing four leaf discs (1 cm2). Means labeled with the same letter are not statistically different at the 5% confidence level based on Tukey’s test (a, b). This experiment was repeated twice with similar results. (c) hopAS1 triggers P. syringae 2 (RPS2)-independent hypersensitive response (HR) in Arabidopsis. Arabidopsis leaves were infiltrated with PFO1-T3SS carrying empty vector (pBBR 1MCS-5), pBS46:avrRpt2PtoT1-HA or pBS46:hopAS1PtoT1-HA (OD600 = 0.4). Photographs were taken at 36 h post infection (hpi). This experiment was repeated 4 times with similar results. (d) hopAS1-triggered immunity does not require RPS2. Wild-type or mutant Arabidopsis accession Columbia (Col-0) leaves were infected with Pto DC3000 carrying empty vector (pBBR 1MCS-5) or pBS46:hopAS1PtoT1-HA as explained in (a). Four-wk-old plants were infected by spraying with bacterial suspensions (OD600 = 0.1) and kept in a humidity chamber for 2 d, and samples were taken at 4 d post infection (dpi) to measure bacterial number in infected leaves. Results are the mean ± SE of bacterial colonies recovered from eight leaf samples each containing four leaf discs (1 cm2). Means labeled with the same letter are not statistically different at the 5% confidence level based on Tukey’s test. This experiment was repeated twice with similar results. (e) Transient expression of HA-HopAS1 in Arabidopsis cells causes cell death. Photographs were taken at 3 d after estradiol treatment. (f) hopAS1-triggered immunity does not require EDS1, NDR1, SID2 and RAR1. Experimental conditions and statistical analysis were same as in (d).

Pto T1 carries the T3SE avrRpt2, which triggers RPS2-dependent resistance (Bent et al., 1994; Mindrinos et al., 1994; Almeida et al., 2009). Strain T1 also shows slightly enhanced virulence in an Arabidopsis mutant lacking functional RPS2 (rps2-101c) (Mindrinos et al., 1994), yet still does not cause disease (Fig. 1b). Based on these results, we hypothesized that the inability of Pto T1 to cause disease in rps2-101c is attributable to inefficient suppression of PTI by Pto T1 T3SEs and/or additional ETI triggered by a Pto T1 T3SE other than avrRpt2. Thus, we decided to search for additional Pto T1 T3SEs that are recognized in an RPS2-independent manner.

To identify an additional Avr T3SE from Pto T1, we cloned all 13 predicted T3SEs of Pto T1 that are absent or significantly different (< 90% amino acid identity) from those of Pto DC3000 (Supporting Information Table S1) (Almeida et al., 2009). Each of the 13 Pto T1-specific T3SEs was expressed under the constitutive nptII (neomycin phosphotransferase II) promoter in the nonpathogenic Pseudomonas fluorescens PFO1 strain carrying a functional T3SS (PFO1-T3SS) (Thomas et al., 2009). All Pto T1-specific T3SEs were expressed well in PFO1-T3SS except HopS1-HA, which we could not detect (Supporting Information Fig. S1). PFO1-T3SS expressing AvrRpt2PtoT1 triggered an RPS2-dependent HR within 20 hpi (Fig. 1c), while the Pto T1 T3SE HopAS1PtoT1 triggered a weak HR in Arabidopsis Columbia (Col-0) and rps2-101c at 36–48 hpi (Fig. 1c). No other Pto T1-specific T3SE triggered an HR in Arabidopsis Col-0 within 48 hpi (data not shown).

As HR is often, but not always, associated with ETI, we tested if any Pto T1 T3SE can restrict virulent bacterial growth in Arabidopsis Col-0, by constructing Pto DC3000 strains carrying individual Pto T1-specific T3SEs and measuring bacterial populations in infected leaves at 3 dpi. Only Pto DC3000 carrying avrRpt2PtoT1 or hopAS1PtoT1 reproducibly showed significantly reduced growth compared with Pto DC3000 (empty vector) (Fig. S2). Moreover, Pto DC3000 carrying hopAS1PtoT1 triggered HR and resistance in rps2-101c, verifying that hopAS1PtoT1-triggered immunity is independent of RPS2 (Fig. 1c,d).

hopAS1PtoT1 is preceded by an hrp box in the 5′ regulatory region and HopAS1PtoT1 protein is delivered to plant cells in a T3SS-dependent manner (Almeida et al., 2009). A conserved domain search (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) using the HopAS1PtoT1 protein (1362 amino acids) sequence predicted that the N-terminal region of HopAS1PtoT1 shares 10% similarity over 17% of the sequence with the bacterial chromosome segregation protein SMC (structural maintenance of chromosomes); no other homologies were detected.

Transient expression of Avr proteins in resistant plant cells often causes cell death responses (McNellis et al., 1998). To investigate whether HopAS1PtoT1 also triggers cell death when transiently expressed in plant cells, we generated a stable Arabidopsis Col-0 line (T2 generation) conditionally expressing HA-HopAS1PtoT1 (hence, At-EST:HA:HopAS1PtoT1) upon estradiol treatment. The HA-HopAS1PtoT1 protein was well expressed in Arabidopsis cells after 24 h of estradiol treatment of 2-wk old At-EST:HA:HopAS1PtoT1 seedlings grown on MS agar plates (Fig. 1e). Unlike several P. syringae T3SEs (e.g. AvrRpt2 and AvrRps4) (Mudgett & Staskawicz, 1999; Sohn et al., 2009), HA-HopAS1PtoT1 protein expressed in At-EST:HA:HopAS1PtoT1 seedlings migrated similarly to HopAS1PtoT1-HA protein from PFO1-T3SS, indicating the post-translational processing of HopAS1PtoT1 into a smaller form does not occur in plant cells (Fig. 1e). At-EST:HA:HopAS1PtoT1 seedlings showed severe cell death after 72–96 h of estradiol treatment (Fig. 1e). This result is in accordance with the previous HR (Fig. 1c) and T3SS-dependent secretion (Almeida et al., 2009) data when HopAS1PtoT1 was secreted from PFO1-T3SS or Pto DC3000, and demonstrates that HopAS1PtoT1 triggers ETI from inside plant cells.

Previously, it was shown that Pto T1 virulence was not significantly enhanced in an Arabidopsis mutant lacking RAR1 (required for Mla12 resistance), which is required for some ETI (Muskett et al., 2002; Tornero et al., 2002; Almeida et al., 2009). In many cases, signaling components downstream of R genes are well conserved (i.e. EDS1 (enhanced disease susceptibility 1)- or NDR1 (non race-specific disease resistance 1)-dependent ETI). To investigate the genetic requirements of hopAS1PtoT1-triggered immunity in more detail, we compared the growth of Pto DC3000 carrying empty vector (pBBR 1MCS-5) or hopAS1PtoT1 in several mutants impaired in ETI signaling. Pto DC3000 carrying empty vector or hopAS1PtoT1 showed enhanced virulence in eds1, ndr1, sid2 (salicylic acid induction deficient 2) and rar1 mutants compared with Col-0 wild-type plants (with the exception of Pto DC3000 carrying hopAS1PtoT1 in the rar1 mutant) (Fig. 1f). Nonetheless, Pto DC3000 carrying hopAS1PtoT1 showed significantly reduced virulence and disease symptom development compared with Pto DC3000 (empty vector) in all four tested mutants, demonstrating that hopAS1PtoT1-triggered immunity is, at least partially, independent of these signaling components (Figs 1f, S3). Several Arabidopsis R genes (e.g. RPP7 (resistance to peronospora parasitica 7), RPP8 and RPP13) conferring resistance to downy mildew (Hyaloperonospora arabidopsidis) function mostly independently of EDS1, NDR1 and salicylic acid (SA) pathways (McDowell et al., 2000; Bittner-Eddy & Beynon, 2001). Recently, it was shown that P. syringae effector hopZ1a-triggered immunity also does not require known ETI-signaling components (Lewis et al., 2010). hopAS1PtoT1- and hopZ1a-triggered Arabidopsis immunity thus uses EDS1- and NDR1-independent signaling.

A mutation in hopAS1 confers enhanced and reduced virulence of Pto T1 in Arabidopsis and tomato, respectively

As hopAS1 triggers immunity in Arabidopsis mutants lacking RPS2, NDR1, SID2 or RAR1, which are required for avrRpt2-triggered immunity, we hypothesized that Pto T1 ΔhopAS1 might confer enhanced virulence compared with wild-type Pto T1 in rps2-101c. To test this hypothesis, Pto T1 ΔhopAS1 was generated and in planta growth was measured. Pto T1 ΔhopAS1 showed slightly but significantly enhanced virulence compared with Pto T1 wild-type or the ΔhrcC mutant in wild-type Col-0 plants, indicating that hopAS1-triggered immunity plays a role in Arabidopsis NHR to Pto T1 (Fig. 2a). Moreover, Pto T1 ΔhopAS1 showed further enhanced virulence compared with Pto T1 wild-type in rps2-101c, indicating that hopAS1- and avrRpt2-triggered immunity quantitatively contribute to Arabidopsis NHR to Pto T1 (Fig. 2a).

Figure 2.

Pseudomonas syringae pv. tomato (Pto) T1 ΔhopAS1 shows enhanced virulence in Arabidopsis and reduced virulence in tomato plants. (a) Pto T1 ΔhopAS1 shows enhanced growth in Arabidopsis. Leaves of 5-wk-old Arabidopsis accession Columbia (Col-0) wild type or rps2-101c mutant plants were infected with Pto T1 wild type or its derivatives as described in Fig. 1(b). Results were obtained from eight leaf samples each. Means labeled with the same letter are not statistically different at the 5% confidence level based on Tukey’s test. This experiment was repeated three times with similar results. (b) hopAS1 is required for full virulence of Pto T1 in tomato leaves. The bacterial growth assay was performed as described in Fig. 1(a), except that results were obtained from four leaf samples. Means labeled with the same letter are not statistically different at the 5% confidence level based on Tukey’s test. This experiment was repeated twice with similar results. (c) Application of coronatine enhances survival of nonhost bacteria in Arabidopsis. Bacterial infection and the growth assay were performed as described in (a) with bacterial suspensions (OD600 = 0.0002) with (+Cor) or without (−Cor) coronatine. Samples were taken at 4 d post infection (dpi). Results were obtained from four leaf samples. Means labeled with the same letter are not statistically different at the 5% confidence level based on Tukey’s test. Error bars (a–c) represent standard error (±SE). This experiment was repeated twice with similar results.

Why does Pto T1 carry functional hopAS1PtoT1 which can restrict its host range? One plausible scenario is that hopAS1 is an important virulence factor for growth of Pto T1 on its natural host plants. Therefore, we investigated whether hopAS1 is involved in Pto T1 virulence on tomato. Pto T1 ΔhopAS1 showed slightly, but significantly, less growth compared with Pto T1 wild type or ΔhopAS1 complemented with a plasmid-borne hopAS1PtoT1-HA in tomato (cv Moneymaker) leaves (Fig. 2b). This suggests that hopAS1 is required for full virulence of Pto T1 during colonization of tomato leaves. We conclude that Pto T1 maintains a functional hopAS1 because of its function in virulence during colonization of host plants despite the resulting reduction in host range.

The phytotoxin coronatine is one of the major virulence determinants of Pto DC3000 (Brooks et al., 2004). Comparative genomic studies have shown that Pto T1 lacks biosynthetic genes for coronatine (Almeida et al., 2009). Thus, we investigated whether exogenous application of coronatine (purified from P. syringae pv. glycinea; Sigma) can suppress Arabidopsis NHR, which would result in enhanced Pto T1 virulence. We also generated a Pto T1 ΔavrRpt2ΔhopAS1 double mutant strain and measured its virulence with or without coronatine application. Infiltration of Arabidopsis Col-0 leaves with coronatine enhanced the growth of Pto DC3000 Cor- (a Pto DC3000 AK87 mutant which carries mutations in cmaA (coronamic acid A) and cfa6 (coronafacic acid 6)) (Brooks et al., 2004) compared with wild-type Pto DC3000 in Arabidopsis Col-0 (Fig. 2c). In addition, coronatine treatment enabled the Pto T1 wild type and the ΔavrRpt2ΔhopAS1 mutant to show slightly enhanced growth in Arabidopsis Col-0 plants, indicating that coronatine may play a role in suppressing NHR (Fig. 2c). By contrast, we could not detect any enhanced growth on Arabidopsis of the Pto T1 ΔhrcC mutant or wild-type Pto DC3000 when treated with coronatine (Fig. 2c). As coronatine treatment enhanced the growth of Pto T1 wild type, T1 ΔavrRpt2ΔhopAS1 and DC3000 Cor- strains but had no effect on Pto T1 ΔhrcC growth, coronatine is not sufficient for suppression of NHR and may require the action of one or more T3SEs. Alternatively, coronatine may suppress ETI triggered by another T3SE from Pto T1. It is interesting to note that, even though avrRpt2-triggered ETI, hopAS1-triggered ETI and lack of coronatine additively contribute to NHR (10–50 times reduced bacterial growth), Pto T1 ΔavrRpt2ΔhopAS1 (+ coronatine) still showed up to 100 times less growth compared with Pto DC3000 wild type (Fig. 2a,c). However, we cannot exclude the possibility that exogenous application of coronatine may not fully compensate for the coronatine deficiency of Pto DC3000 Cor- strains. Recently, it was shown that P. syringae sax (survival in Arabidopsis extracts) genes play an essential role in overcoming isothiocyanate-based defenses, resulting in enhanced virulence of nonhost Pseudomonas strains, including Pto T1, in Arabidopsis (Fan et al., 2011). However, introduction of saxCAB genes in Pto T1 ΔhopAS1 or Pto T1 ΔavrRpt2ΔhopAS1 did not result in enhanced bacterial growth in the Col-0 wild type or the rps2-101c mutant (Fig. S4). These results suggest the involvement of other factors in Arabidopsis NHR to Pto T1. Conceivably, Pto T1 T3SEs are not as efficient at suppressing PTI in Arabidopsis as those of Pto DC3000. In this case, Pto DC3000 T3SEs that are not present in Pto T1 would be good candidates for testing whether they enhance the virulence of Pto T1 ΔhopAS1 in rps2-101c.

Arabidopsis-infecting P. syringae strains carry truncated hopAS1 alleles

We investigated hopAS1 alleles from a diverse array of P. syringae strains to assess the diversity in this T3SE family and determine its impact on plant immunity. In particular, we analyzed hopAS1 nucleotide (NT) sequences from 28 P. syringae strains belonging to 11 pathovars and isolated from many different plant species (Table 1). Among the 28 strains, all strains pathogenic on Arabidopsis (Debener et al., 1991; Whalen et al., 1991; Rohmer et al., 2003; Yan et al., 2008) carry truncated hopAS1 alleles (Table 1), including Pto DC3000, which has a nonsense mutation (CAT to TAG) at nucleotide position 838–840, and P. syringae pv. maculicola (Pma) ES4326, which has a premature stop brought about by a 2-bp deletion at nucleotide position 3803–3804 (Fig. 3a). All hopAS1 alleles from the 19 strains nonpathogenic on Arabidopsis carry full-length hopAS1 alleles, indicating that hopAS1-triggered immunity is strongly associated with Arabidopsis NHR to these strains (Table 1).

Table 1. Pseudomonas syringae strains virulent in Arabidopsis carry a truncated hopAS1 allele
P. syringae strainaVirulence in ArabidopsisHopAS1 polymorphism (aa length/mutationb)HopAS1-triggered HR in ArabidopsisPto DC3000 (hopAS1) virulence in Arabidopsisc
  1. aPto, Pseudomonas syringae pv. tomato; Pph, Pseudomonas syringae pv. phaseolicola; Pan, Pseudomonas syringae pv. antirrhini; Pca, Pseudomonas cannabina; Pta, Pseudomonas syringae pv. tabaci; Pav, Pseudomonas syringae pv. avellanae; Pse, Pseudomonas syringae pv. sesami; Pla, Pseudomonas syringae pv. lachrymans; Pgy, Pseudomonas syringae pv. glycinea; Pma, Pseudomonas syringae pv. maculicola; Pal, Pseudomonas cannabina pv. alisalensis.

  2. bMutations were caused by nucleotide change, insertion or deletion of one or two nucleotides.

  3. chopAS1-triggered immunity was determined by comparing in planta growth of Pto DC3000 (pBBR 1MCS-5:hopAS1) with that of Pto DC3000 (pBBR 1MCS-5) in Arabidopsis accession Columbia (Col-0).

  4. dHR, hypersensitive response. HR was triggered by infiltrating 5- or 6-wk-old Arabidopsis Col-0 leaves with PFO1-T3SS carrying hopAS1 (OD600 = 0.4). HR was scored at 36–48 h post infection (hpi).

  5. eNA, not available.

  6. fAll pathogenic strains analyzed cause disease in at least one Arabidopsis accession (data not shown).

  7. gPma ES4326 belongs to P. cannabina pv. alisalensis based on multilocus sequence typing (Bull et al., 2010).

Pto T1NonpathogenicFull-length (1362)HRdDecreased
Pto JL1065NonpathogenicFull-length (1362)HRDecreased
Pph 1448ANonpathogenicFull-length (1361)HRDecreased
Pan 126NonpathogenicFull-length (1371)HRDecreased
Pca CFBP2341NonpathogenicFull-length (1369)HRDecreased
Pta 11528NonpathogenicFull-length (1361)NAeNA
Pav BPIC631NonpathogenicFull-length (1361)NANA
Pph HB10Y1NonpathogenicFull-length (1361)NANA
Pph 1302ANonpathogenicFull-length (1361)NANA
Pta 6606NonpathogenicFull-length (1361)NANA
Pse HC_1NonpathogenicFull-length (1361)NANA
Pph NPS3121NonpathogenicFull-length (1361)NANA
Pph Y5_2NonpathogenicFull-length (1361)NANA
Pla 107NonpathogenicFull-length (1361)NANA
Pla YM7902NonpathogenicFull-length (1361)NANA
Pgy KN44NonpathogenicFull-length (1361)NANA
Pgy LN10NonpathogenicFull-length (1361)NANA
Pgy UnB647NonpathogenicFull-length (1361)NANA
Pgy BR1NonpathogenicFull-length (1361)NANA
Pma M6PathogenicfTruncated (402/frame shift)NANA
Pto DC3000PathogenicTruncated (279/frame shift)No HRSimilar
Pto ICMP3443PathogenicTruncated (400/frame shift)NANA
Pma ES4326gPathogenicTruncated (1329/frame shift)No HRSimilar
Pma F1PathogenicTruncated (216/frame shift)NANA
Pma F9PathogenicTruncated (279/frame shift)NANA
Pma M3PathogenicTruncated (279/frame shift)NANA
Pal CFBP6866PathogenicTruncated (1329/frame shift)NANA
Pal T3CPathogenicTruncated (1329/frame shift)NANA
Figure 3.

Virulent Pseudomonas syringae strains, Pseudomonas syringae pv. tomato (Pto) DC3000 and Pseudomonas syringae pv. maculicola (Pma) ES4326, carry nonfunctional hopAS1. (a) Comparison of amino acid sequences of HopAS1 from seven Pseudomonas strains. Full names for the strains are listed in Table 1. Numbers indicate the amino acid identity (%) compared with HopAS1PtoT1. For HopAS1PtoDC3000, 279 aa of HopAS1PtoT1 was used to determine identity level. Asterisks indicate the positions of mutation. (b) hopAS1PtoDC3000 and hopAS1PmaES4326 do not trigger HR. Five-wk-old Arabidopsis leaves were infiltrated with PFO1-T3SS carrying hopAS1 (OD600 = 0.4). Photographs were taken at 36 h post infection (hpi). This experiment was repeated three times with similar results. (c) hopAS1PtoDC3000 and hopAS1PmaES4326 do not trigger ETI. Arabidopsis Col-0 leaves were hand-infiltrated with Pto DC3000 strains carrying hopAS1-HA (OD600 = 0.001). Results were obtained from nine leaf samples. Means labeled with the same letter are not statistically different at the 5% confidence level based on Tukey’s test. This experiment was repeated twice with similar results. Error bars (c) represent standard error (±SE). (d) hopAS1PtoT1 and hopAS1PtoJL1065 trigger cell death response in turnip. Leaves of 4-wk-old turnip (cv Just Right) plants were hand-inoculated with PFO1-T3SS carrying hopAS1-HA (OD600 = 0.01). Photographs were taken at 10 d post infection (dpi).

To test the avirulence activity of hopAS1 alleles from strains that do not cause disease in Arabidopsis Col-0, we cloned hopAS1 from five nonpathogenic and two pathogenic P. syringae strains (Fig. 3a). When expressed in PFO1-T3SS, the truncated hopAS1PtoDC3000 and hopAS1PmaES4326 alleles did not trigger any visible HR, whereas the full-length hopAS1 from the five nonpathogenic strains triggered macroscopic HR at 36–48 hpi in Arabidopsis Col-0 (Fig. 3b). In addition, Pto DC3000 carrying hopAS1PmaES4326 or empty vector (pBBR 1MCS-5) showed similar levels of growth, while Pto DC3000 carrying hopAS1 alleles from the nonpathogenic strains showed significantly reduced growth as compared with Pto DC3000 (empty vector) in Arabidopsis at 4 dpi (Fig. 3c).

To assess the role of hopAS1 alleles in alternative hosts, we used PFO1-T3SS to deliver Pto T1, Pto JL1065 and Pto DC3000 hopAS1 alleles into turnip (cv Just Right) leaves. Consistent with the Arabidopsis results, the full-length hopAS1PtoT1 and hopAS1PtoJL1065 alleles, but not the truncated hopAS1PtoDC3000 allele, triggered HR-like symptoms at 8–10 dpi, suggesting that hopAS1-triggered immunity is conserved in brassicaceous hosts (Fig. 3d). These results indicate that hopAS1-triggered immunity plays an important role in the NHR of Arabidopsis and turnip to diverse P. syringae strains and emphasize the importance of ETI not only in pathogen race determination but also in host range and NHR. Interestingly, several closely related strains, for example Pto DC3000/Pto T1 or Pca CFBP2341/Pal CFBP6866, show different pathogenicity levels in Arabidopsis. Because full-length hopAS1 alleles must necessarily have evolved before disrupted hopAS1 alleles, one could infer that the most recent common ancestor of each of these groups of strains was not pathogenic in Brassicaceae. Alternatively, the most recent common ancestor was pathogenic on ancestors of extant Brassicaceae species that had not yet evolved hopAS1-triggered immunity. Identifying the genetic basis of hopAS1-triggered immunity in Arabidopsis and other Brassicaceae will enable further insights to be obtained into the dynamics of the evolutionary arms race between the Brassicaceae family and P. syringae.

However, hopAS1 is clearly not the only factor determining host specificity in Arabidopsis, as some P. syringae strains that are not pathogenic in Arabidopsis do not carry any hopAS1 allele (for example Psy B728a) (Feil et al., 2005). In addition, the Pto T1 ΔavrRpt2ΔhopAS1 mutant grows significantly less well than Pto DC3000 in Arabidopsis Col-0 (Fig. 2c). Consequently, the absence of hopAS1 and avrRpt2 is necessary but not sufficient for P. syringae compatibility with Arabidopsis. Other virulence factors (e.g. coronatine and virulence-promoting T3SEs) absent from Pto T1 are needed to overcome Arabidopsis NHR, and these await identification.

By definition, we would expect negligible genetic variation within a species for nonhost resistance. As the absence of full-length and functional hopAS1 in the P. syringae strains we tested is completely associated with the ability to colonize Arabidopsis accession Col-0, we wondered how well hopAS1-triggered immunity is conserved among Arabidopsis accessions. Therefore, we tested the responses of 24 diverse Arabidopsis accessions to PFO1-T3SS-delivered HopAS1. We also tested avrRpm1, avrRpt2 or avrPphB-triggered HR as controls to see if other Avr genes isolated from nonpathogenic P. syringae strains of Arabidopsis are broadly recognized or not. All 24 accessions showed HR to PFO1-delivered HopAS1, whereas several accessions did not show HR in response to avrRpm1, avrRpt2 or avrPphB (Table S2). This suggests that hopAS1-triggered immunity more significantly contributes to Arabidopsis NHR to P. syringae pathogens than other more well-known ETIs.


We are grateful to Jeff Dangl (University of North Carolina, USA) and Marc Nishimura (University of North Carolina) for sharing unpublished sequences with us and providing helpful comments on our manuscript. We also thank Jeff Chang and Bill Thomas (Oregon State University) for Pf. PFO1, Bryan Swingle (USDA) for pBS46 plasmid, Jane Parker (MPIZ, Cologne, Germany) for pER8:N-HS plasmid, Alan Collmer (Cornell University, USA) for pRK415 plasmid, Barbara Kunkel (Washington University, USA) for the Pto DC3000 Cor- mutant, Jun Fan (John Innes Centre, UK) for pME6012:saxCAB plasmid, and Matthew Smoker, Jodie Pike and JIC horticultural staff for generating and maintaining Arabidopsis transgenic lines. This work was carried out with the support of the Gatsby Foundation (UK) and the ‘Cooperative Research Program for Agriculture Science & Technology Development (Project No. PJ007850201006)’ Rural Development Administration (Republic of Korea). Research in the Vinatzer lab was funded by the National Science Foundation (Award IOS 0746501).