Pseudomonas sax Genes Overcome Aliphatic Isothiocyanate–Mediated Non-Host Resistance in Arabidopsis

See allHide authors and affiliations

Science  04 Mar 2011:
Vol. 331, Issue 6021, pp. 1185-1188
DOI: 10.1126/science.1199707


Most plant-microbe interactions do not result in disease; natural products restrict non-host pathogens. We found that sulforaphane (4-methylsulfinylbutyl isothiocyanate), a natural product derived from aliphatic glucosinolates, inhibits growth in Arabidopsis of non-host Pseudomonas bacteria in planta. Multiple sax genes (saxCAB/F/D/G) were identified in Pseudomonas species virulent on Arabidopsis. These sax genes are required to overwhelm isothiocyanate-based defenses and facilitate a disease outcome, especially in the young leaves critical for plant survival. Introduction of saxCAB genes into non-host strains enabled them to overcome these Arabidopsis defenses. Our study shows that aliphatic isothiocyanates, previously shown to limit damage by herbivores, are also crucial, robust, and developmentally regulated defenses that underpin non-host resistance in the Arabidopsis-Pseudomonas pathosystem.

Non-host resistance is the ability of most plant species to resist microbes or viruses that are successful pathogens on other plants. It is the most prevalent form of plant disease resistance, is durable and effective against a broad range of potential pathogens, but our understanding of its molecular basis is still poor (13). Plants generate a huge diversity of natural products, with multiple roles in defense, communication, and development (4). Preformed natural products provide chemical barriers to phytopathogenic fungi (57) and are deterrents in plant-herbivore interactions (8). However, their role in restricting bacterial host range remains obscure, as do the bacterial mechanisms involved in breaching natural product–mediated host defenses. To better understand fundamental host-pathogen biology and to inform the development of sustainable field resistance to major crop diseases (9), we sought to define plant components conferring non-host resistance as well as strategies used by virulent pathogens to overcome resistance barriers.

We observed that extracts from naïve Arabidopsis plants inhibited the growth of most pathovars of Pseudomonas syringae for which Arabidopsis is not a host. In contrast, P. syringae pathovars maculicola (Psm) ES4326 and tomato (Pst) DC3000, which are pathogenic on Arabidopsis, grew well on rich media supplemented with host extract from Col-0 (Table 1) or other accessions (table S1). Using ArabidopsisP. syringae as a model system to dissect plant resistance to non-host pathogens, we screened a Psm ES4326 genomic library for genes conferring resistance in Escherichia coli to Arabidopsis extracts (Table 1) (10). A single operon, designated sax (survival in Arabidopsis extracts), allowed E. coli to grow on Arabidopsis extracts. In E. coli, saxA and saxC are together necessary for resistance: The absence of either resulted in susceptibility, whereas the absence of saxB reduced bacterial growth on extracts (fig. S1A). SaxA has a predicted secretory signal peptide (11) and, although related to class B β-lactamases (12, 13), it is unable to confer resistance to eight representative β-lactam antibiotics (fig. S2A), indicating that SaxA activity is distinct. SaxB is related to isochorismatase. SaxC is a highly conserved member of the AraC/XylS family of transcriptional regulators found in diverse prokaryotes and involved in carbon metabolism, stress response, and pathogenesis (14). Analysis of 35 plant-associated P. syringae genomes showed that only Arabidopsis pathogens have saxA-like genes with ≥90% nucleic acid sequence identity to Psm ES4326 saxA (table S2 and fig. S3).

Table 1

Inhibitory activity of Arabidopsis extracts on bacterial growth in vitro. Bacterial growth was determined 24 hours after inoculation, or as indicated otherwise. Data are means ± SD. Pst strains are from this study; other strain sources are as indicated. Strains ES4326, M1, M2, M3, M5, M6, and DC3000 grow on crucifers. “With ITC” denotes myb mutant plant extract supplemented with authentic sulforaphane (4-MSB ITC) at a concentration of 200 μg ml−1. Statistical significance of growth inhibition (rightmost column): *P < 0.05, **P < 0.01, ***P < 0.001 (two-tailed t test; other values are not significant).

View this table:

Non-host resistance is durable, and hence it is unlikely to be overwhelmed by a single mechanism. Indeed, deletion of saxAB genes in Pst DC3000 or Psm ES4326 had little impact on bacterial growth in Arabidopsis extracts (Table 1). To identify additional protective mechanism(s), we screened for compromised bacterial growth in Arabidopsis extracts after transposon mutagenesis of PstΔsaxAB (10). Two putative multidrug resistance (MDR) efflux genes were identified, with a third similar system revealed in Pst DC3000 by genome analysis (10); these were designated saxF, saxD, and saxG, respectively (fig. S1B). They form a subgroup among the nine resistance-nodulation-division (RND) efflux systems predicted in the Pst DC3000 genome (15), which function to extrude a wide range of substrates including antibiotics and host-derived molecules (16). We sequentially deleted saxF/D/G from the PstΔsaxAB background, which resulted in progressively increased sensitivity (Table 1); hence, these genes were required for robust resistance to Arabidopsis extracts. Deletion of saxAB, saxF/D/G, or saxAB/F/D/G did not impair growth in rich medium (Fig. 1A), indicating that they are not essential. Thus, sax genes have distinct but complementary roles in Pseudomonas resistance to Arabidopsis extracts.

Fig. 1

sax genes in Pst DC3000 protect against Arabidopsis-derived isothiocyanates. (A) sax genes are synergistically required for bacterial resistance to isothiocyanates. Bacterial strains were inoculated into Kings B medium with or without sulforaphane and grown overnight. (B) Bacterial infection released sulforaphane (4-MSB ITC) into apoplastic diffusates. Arabidopsis leaf discs were vacuum-infiltrated with bacteria [optical density at 600 nm (OD600) = 0.1] and incubated at 23°C for 48 hours before analysis. Data are means ± SD; cpl, episomal complementation with the sax operon; gfw, gram fresh weight; nd, not detected.

We purified the Arabidopsis antimicrobial compound restricting non-host Pseudomonas from Col-0 extracts (10) and identified it as sulforaphane [4-methylsulfinylbutyl isothiocyanate (4-MSB ITC)] (fig. S5). Aliphatic isothiocyanates are breakdown products derived from glucosinolates, a class of sulfur- and nitrogen-containing natural products widely distributed in crucifers; Arabidopsis plants produce a variety of glucosinolates, with extensive variation in leaf and root content across different accessions (17). In Col-0, 4-MSB glucosinolate predominantly accumulates in rosette leaves (18); however, Arabidopsis plants with loss-of-function alleles of MYB28 and MYB29 transcription factors, but not Col-0 plants transformed with 35S::saxA, lack aliphatic glucosinolates (19, 20) (table S3). A panel of virulent and non-host bacteria showed a higher median inhibitory concentration (IC50) for sulforaphane in the former (table S5).

Antimicrobial activity of sulforaphane in vitro has been reported (21), but its role in disease resistance in planta is unknown. We tested the protective effect of the sax operon against a panel of isothiocyanates derived from glucosinolates prevalent in Arabidopsis and other crucifers. E. coli carrying the sax operon was more tolerant than the vector control to all five tested compounds (fig. S2B), indicating that saxCAB genes are active against a broad spectrum of isothiocyanates. Deletion of saxAB, saxF/D/G, or saxAB/F/D/G progressively impaired growth in rich medium supplemented with sulforaphane, or in extracts from myb28/29 plants supplemented with sulforaphane, but not in extracts from Col-0 plants transgenic for 35S::saxA (Fig. 1A and Table 1); these findings indicate that SaxA is sufficient for detoxification of sulforaphane and other aliphatic isothiocyanates (Fig. 2A and fig. S2B). Treatments with Arabidopsis extracts or with sulforaphane induced saxA and saxF expression (fig. S4). These findings indicate a broader role for aliphatic isothiocyanates in plant-biotic interactions than previously understood.

Fig. 2

Accumulation of aliphatic glucosinolates and isothiocyanates is necessary for suppressing the quintuple sax mutant in young Arabidopsis leaf tissue. (A) Representative Arabidopsis transgenic plants expressing the bacterial saxA gene fail to produce high levels of sulforaphane (4-MSB ITC) in crude extracts and are fully susceptible to the quintuple sax mutant. (B) Disruption of aliphatic glucosinolate biosynthesis abolished resistance to the quintuple sax mutant in young Arabidopsis leaf tissue. In (A) and (B), discs from young rosette leaves were treated as described (Fig. 1B) and bacterial titer analyzed 3 days after inoculation. Data are means ± SD.

Glucosinolate breakdown is activated by tissue damage after herbivory; the subsequent accumulation of aliphatic isothiocyanates is a major determinant of plant-herbivore interactions (20, 22, 23). However, bacterial infections do not normally cause similar tissue damage. We examined whether infection released isothiocyanates into the apoplastic space that hemibiotrophic bacteria colonize (10). Relative to mock treatments, sulforaphane levels were reduced in diffusate and residue samples treated with wild-type Pst DC3000, whereas infection with the PstΔsaxAB strain increased sulforaphane levels in diffusate (Fig. 1B). Thus, Pst DC3000 infection triggers glucosinolate degradation and release of isothiocyanates into the apoplast, and this Arabidopsis response is countered by Pst DC3000 in a sax-dependent manner.

To determine whether the isothiocyanate resistance of Pst DC3000 plays an important role in pathogenesis on Arabidopsis, we monitored bacterial growth after infiltration of Pst DC3000 or saxAB/F/D/G. Consistent with a previous report (24), we found that sulforaphane levels in apoplastic diffusates of young leaf discs were higher by a factor of ~100 than in old leaf discs infected with the PstΔsaxAB strain (Fig. 3A), which suggests differences in their glucosinolate-mediated defense capacity. Relative to wild-type Pst DC3000, bacterial titer was reduced by six orders of magnitude in the ΔsaxAB/F/D/G strain in the young but not the old leaf discs (Fig. 3B). In planta assays with much lower inoculum density (Fig. 4A) or when inoculated by spraying (fig. S7) produced similarly reduced levels of growth of the ΔsaxAB/F/D/G strain. Thus, sax genes are important for Pst DC3000 to penetrate host barriers and maximize its population in the young leaves.

Fig. 3

sax genes are required for Pst DC3000 survival during infection of young Arabidopsis leaf tissue releasing high levels of sulforaphane (4-MSB ITC). (A) Levels of 4-MSB ITC differed significantly between young and old Arabidopsis leaves. (B) The quintuple sax mutant is unable to maintain infection on young Arabidopsis leaf tissue. The conditions for these bacterial infection experiments were the same as described in Fig. 1B. Data are means ± SD; dpi, days post-inoculation.

Fig. 4

Aliphatic isothiocyanates restrict bacterial growth in planta. (A) Young leaves of Col-0 and myb28/29 plants were syringe-infiltrated with bacteria (OD600 = 0.001). (B) Aliphatic glucosinolate biosynthesis is required for growth suppression of the non-host Pst T1 strain. (C) The saxCAB operon promoted growth of Pst T1 on Col-0 plants. In (B) and (C), young leaves of 5- to 6-week-old plants were infiltrated with bacteria (OD600 = 0.02). Data are means ± SD; EV, empty vector control.

We then examined transgenic Arabidopsis plants overexpressing sax genes to examine whether sax-dependent reactions are sufficient to negate bacterial restriction by the host. The levels of major glucosinolates were unchanged in the saxA-expressing transgenic plants (table S3), suggesting that SaxA targets steps after glucosinolate breakdown. Sulforaphane levels in extracts of young leaves from saxA transgenic plants were lower than those from wild-type Col-0 and saxB transgenic plants by a factor of >20 (Fig. 2A). Growth inhibition of the ΔsaxAB/F/D/G mutant observed in leaves from Col-0 plants was lost in saxA transgenic plants (Fig. 2A) and in the myb28/29 background (Fig. 2B, Fig. 4A, and fig. S7). These observations indicate that aliphatic isothiocyanates are potent host defense effectors that restrict bacterial populations. However, we also observed that the titer of the quintuple sax mutant did not fall until necrotic symptoms had developed, implying that the isothiocyanate-mediated defense might be potentiated by cell death in the host.

To examine the broader role of aliphatic isothiocyanates in restricting non-host pathogens, we tested an expanded panel of crucifer and noncrucifer pathogens (Table 1) with Arabidopsis extracts. In extracts prepared from wild-type Col-0 plants, all the crucifer pathogens tested were able to grow well, whereas the titer of the noncrucifer strains varied between 103.3 and 108.8 CFU/ml and showed a clear trend of growth inhibition. However, when grown in extracts from myb28/29 or 35S::saxA plants, growth inhibition was attenuated in the noncrucifer pathogens. After supplementing extracts of myb28/29 plants with sulforaphane (Table 1) or iberin (3-methylsulfinylpropyl isothiocyanate) (table S4), restriction of noncrucifer pathogens was restored, whereas we observed only small growth differences of the crucifer pathogens between different treatments. Hence, aliphatic isothiocyanates are essential for Arabidopsis extracts to restrict the growth of many noncrucifer bacterial pathogens.

We examined whether saxCAB was sufficient to allow non-host pathogens to overcome Arabidopsis aliphatic isothiocyanate–dependent defenses. A tomato isolate of P. syringae, Pst T1, which is not pathogenic in Arabidopsis (25, 26), does not grow on Arabidopsis extracts (Table 1). saxA is absent from a homologous region of the Pst T1 genome (fig. S6A), whereas its sax-related efflux systems are ≥99% identical to those in Pst DC3000 (fig. S6B). When challenged with Pst T1, young leaves of myb28/29 plants supported ~10 times as much bacterial growth as did Col-0 (Fig. 4B). Transformation of Pst T1 with Psm ES4326 saxCAB rescued Pst T1 growth in Arabidopsis extracts containing sulforaphane (fig. S8) and enhanced growth in young Col-0 leaves by a factor of ~10 (Fig. 4C). Similar results were obtained with the non-host pathovar Ps apii (ATCC 9654) (Table 1 and fig. S9). We conclude that isothiocyanate-mediated defense effectively limits infection by non-host pathogenic Pseudomonas bacteria and that sax genes strongly enhance Pseudomonas virulence in Arabidopsis.

Indolic glucosinolates are required for Arabidopsis to prevent penetration of non-adapted powdery mildew strains (27) and to activate callose deposition as an innate immune response (28). Our findings show that Pseudomonas pathogens require sax gene–dependent mechanisms to overwhelm aliphatic isothiocyanate–mediated non-host resistance in Arabidopsis. This and other work (27, 28) shows how non-host resistance, unlike many examples of major R gene resistance to specific pathogen genotypes, can be resilient, durable, and broadly effective against many potential pathogens, thereby revealing emergent principles for engineering sustainable field resistance for major crop diseases.

Supporting Online Material

Materials and Methods

Figs. S1 to S9

Tables S1 to S6


References and Notes

  1. See supporting material on Science Online.
  2. We thank D. Haas, C. Zipfel, K. Sohn, and D. J. Kliebenstein for providing the pME vectors, bacterial strains, and Arabidopsis myb mutants used in this study; B. Lee and S. Kopriva for help with the glucosinolate assay; and C. Zipfel, C. Dean, A. Maule, and V. Vitart for critical reading of earlier versions of the manuscript. Supported by a grant from the UK Biotechnology and Biological Sciences Research Council (C.L.). J.F., C.C., G.C., L.H., S.F., and P.D. dedicate this paper to the memory of Chris Lamb.
View Abstract

Navigate This Article