Deletions in the repertoire of Pseudomonas syringae pv. tomato DC3000 type III secretion effector genes reveal functional overlap among effectors

Abstract

The gamma-proteobacterial plant pathogen Pseudomonas syringae pv. tomato DC3000 uses the type III secretion system to inject ca. 28 Avr/Hop effector proteins into plants, which enables the bacterium to grow from low inoculum levels to produce bacterial speck symptoms in tomato, Arabidopsis thaliana, and (when lacking hopQ1-1) Nicotiana benthamiana. The effectors are collectively essential but individually dispensable for the ability of the bacteria to defeat defenses, grow, and produce symptoms in plants. Eighteen of the effector genes are clustered in six genomic islands/islets. Combinatorial deletions involving these clusters and two of the remaining effector genes revealed a redundancy-based structure in the effector repertoire, such that some deletions diminished growth in N. benthamiana only in combination with other deletions. Much of the ability of DC3000 to grow in N. benthamiana was found to be due to five effectors in two redundant-effector groups (REGs), which appear to separately target two high-level processes in plant defense: perception of external pathogen signals (AvrPto and AvrPtoB) and deployment of antimicrobial factors (AvrE, HopM1, HopR1). Further support for the membership of HopR1 in the same REG as AvrE was gained through bioinformatic analysis, revealing the existence of an AvrE/DspA/E/HopR effector superfamily, which has representatives in virtually all groups of proteobacterial plant pathogens that deploy type III effectors.

Figure 1. Polymutants lacking clustered effector genes highlight the importance of the CEL in virulence.

(A) The P. syringae pv. tomato DC3000 genome harbors 18 active effectors in six clusters (shaded pink), including the conserved effector locus (CEL). Cluster X is carried on pDC3000A and ΔX strains have been cured of the plasmid. Additional clusters harboring apparently inactive effector genes and pseudogenes are not shown. The hopM1 and avrE genes are shaded yellow to denote that they functionally overlap; the avrPto and avrPtoB genes are similarly shaded blue to denote that they are functionally linked although not genetically clustered; and a line is shown through hopQ1-1 to indicate that all strains used in this study lack this effector because of its avirulence activity in N. benthamiana. (B) Effector gene clusters I and CEL(VI) were deleted using pK18mobsacB. Flanks were amplified by PCR with terminal primer-introduced restriction sites as labeled. Genes are colored using COG function category defaults. B, BamHI; X, XbaI; H, HindIII; E, EcoRI. (C) Deletion of all of the clustered effector genes abolishes the ability of DC3000 to cause symptoms in N. benthamiana leaf areas inoculated via a blunt syringe. The marked leaf areas were infiltrated with the strains indicated in the table below at 3×10⁴ CFU/ml and photographed 9 days post-inoculation. ++++, extensive chlorosis and necrosis; +++ extensive chlorosis and reduced necrosis; ++, reduced chlorosis and highly reduced/delayed necrosis; -, no symptoms; *, limited chlorosis no necrosis. “Q⁻” in the table indicates only hopQ1-1 has been deleted from cluster IV. The T3SS⁻ mutant was ΔhrcQ_B-hrcU CUCPB5113. (D) Deletion of all of the clustered effector genes strongly reduces the ability of DC3000 to grow in N. benthamiana leaf areas inoculated via a blunt syringe. N. benthamiana leaves were infiltrated with the strains indicated in the table at 3×10⁴ CFU/ml (2.5 log CFU/cm² of leaf tissue). Bacterial populations were determined from three 0.8-cm leaf discs 6 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from four leaf samples. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test. This experiment was repeated three times with similar results. (E) The ΔCEL mutant is significantly reduced in symptom production when N. benthamiana leaves are inoculated by dipping. N. benthamiana leaves were inoculated with the strains indicated in the table at 3×10⁵ CFU/ml and photographed 9 days post-inoculation. ++++, extensive chlorosis and necrotic specks; +++, extensive chlorosis and reduced necrotic specks; ++, reduced chlorosis and reduced necrotic specks; -, no symptoms; *, limited chlorosis and no necrotic specks. (F) Deletion of all of the clustered effector genes strongly reduces the ability of DC3000 to grow in N. benthamiana leaves inoculated by dipping. Whole N. benthamiana plants were dipped in 3×10⁵ CFU/ml suspensions of the strains indicated in the table (1.5 log CFU/cm² of leaf tissue). Bacterial populations were determined from three 0.8-cm leaf discs 6 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from four leaf samples. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test. This experiment was repeated two times with similar results. (G) Deletion of all of the clustered effector genes strongly reduces the ability of DC3000 to grow in tomato leaf areas inoculated via a blunt syringe. Tomato leaflets were infiltrated with the strains indicated in the table at 3×10⁴ CFU/ml (2.5 log CFU/cm² leaf tissue) with a blunt syringe. Bacterial populations were determined from three 0.8-cm leaf discs 3 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from four leaflet samples. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test. This experiment was repeated three times with similar results.

Figure 2. Combinatorial deletions reveal effector gene clusters that interplay in promoting virulence in N. benthamiana.

(A) Deleting the CEL cluster from intermediates in the lineage of polymutants used to delete all clustered effector genes reveals functional overlap between the CEL and clusters IV and IX in promoting bacterial growth in N. benthamiana. Leaves were infiltrated with the strains indicated in the table at 3×10⁴ CFU/ml (2.5 log CFU/cm² leaf tissue) with a blunt syringe. “Q⁻” in the table indicates only hopQ1-1 has been deleted from cluster IV. Bacterial populations were determined from three 0.8-cm leaf discs 6 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from four leaf samples. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test. This experiment was repeated three times with similar results. (B) N. benthamiana leaves were infiltrated with the strains indicated at 3×10⁴ CFU/ml with a blunt syringe and photographed 6 days post-inoculation. (C) Additional effector gene cluster polymutants further define functional redundancies with cluster IX in promoting bacterial growth in N. benthamiana. Leaves were infiltrated and assayed as described in (A). This experiment was repeated three times with similar results. (D) Additional effector gene cluster polymutants reveal functional overlap between clusters II and IX in promoting bacterial growth in N. benthamiana. Leaves were infiltrated and assayed as described in (A). This experiment was repeated three times with similar results.

Figure 3. HopR1 is functionally redundant with the CEL in promoting bacterial growth in N. benthamiana.

(A) N. benthamiana leaves were infiltrated with ΔIV or ΔIVΔCEL mutant strains transformed with pBBR P _avrPto hop expression constructs at 3×10⁴ CFU/ml (2.5 log CFU/cm² leaf tissue) with a blunt syringe. The avr/hop genes carried by the constructs are indicated by the final portion of their gene names. Asterisks indicate that the construct also expresses the appropriate effector chaperone. Bacterial populations were determined from three 0.8-cm leaf discs 6 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from four leaf samples. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test. This experiment was repeated four times with similar results. Tomato leaflets were infiltrated with the indicated strains (described further in Table 1) at 3×10⁴ CFU/ml (2.5 log CFU/cm² leaf tissue) with a blunt syringe. Bacterial populations were determined from three 0.8-cm leaf discs 3 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from three separate experiments with four leaflets each. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test.

Figure 4. HopR1 is a member of an effector superfamily that is widespread among phytopathogens.

A distance-based MUSCLE tree shows that HopR1 is in one of three major clades of the AvrE/DspA/E/HopR superfamily. The tree was generated from a structural alignment using the MUSCLE algorithm and displayed using NJplot . The lengths of the branches are measures of structural similarity.

Figure 5. HopR1 suppresses callose formation by the ΔIVΔCEL mutant in N. benthamiana.

N. benthamiana leaves were infiltrated with the ΔIVΔCEL mutant strains transformed with pBBR P _avrPto avr/hop expression constructs at 5×10⁸ CFU/ml with a blunt syringe. Aniline blue-stained callose depositions in N. benthamiana leaves were visualized by epifluorescence microscopy 12 h after inoculation. Numbers are the means and standard deviations of eight 1-cm² microscopic fields of view. The panels in the lower right show enlargements of representative areas from leaf areas inoculated with the ΔIVΔCEL mutant and the mutant expressing hopR1. The experiment was repeated six times with similar results. The scale bar is 100 µm.

Figure 6. Deletion of the flagellin fliC gene rescues growth of ΔavrPtoΔavrPtoB but not ΔIVΔCEL mutants.

N. benthamiana leaves were infiltrated with the indicated strains at 3×10⁴ CFU/ml (2.5 log CFU/cm² leaf tissue) with a blunt syringe. Bacterial populations were determined from three 0.8-cm leaf discs 6 days post-inoculation. Results are the mean and standard deviation of bacterial populations collected from four leaf samples. Means marked with the same letter are not statistically different at the 5% confidence level based on Duncan's multiple range test. This experiment was repeated three times with similar results.

Figure 7. Proposed model for the role of redundant effector groups (REGs) in blocking distinct steps in PTI.

The model depicts flagellin as the major DC3000 PAMP detected by N. benthamiana. A REG comprises effectors (e.g., AvrPto/AvrPtoB) that function redundantly to block a single high-level PTI process (e.g., PAMP perception). There is also intrinsic redundancy in each high-level PTI process (denoted by the red and blue arrows), which prevents the action of a single REG from completely blocking PTI. The existence of candidate minor members of REGs is revealed by observing further reductions in pathogen growth when minor member genes are deleted from mutants lacking the major members, as exemplified by clusters II and IX in the proposed AvrE/HopR1/HopM1 REG. The model predicts that (i) many effectors in the P. syringae pan-genomic effector super-repertoire can be assigned to REGs targeting a limited number of vulnerable high-level processes in PTI, (ii) any virulent strain will contain at least two REGs with at least one active member in each REG, and (iii) REGs can be used to functionally dissect vulnerable processes underlying PTI.