Type III effector diversification via both pathoadaptation and horizontal transfer in response to a coevolutionary arms race

Abstract

The concept of the coevolutionary arms race holds a central position in our understanding of pathogen-host interactions. Here we identify the molecular mechanisms and follow the stepwise progression of an arms race in a natural system. We show how the evolution and function of the HopZ family of type III secreted effector proteins carried by the plant pathogen Pseudomonas syringae are influenced by a coevolutionary arms race between pathogen and host. We surveyed 96 isolates of P. syringae and identified three homologs (HopZ1, HopZ2, and HopZ3) distributed among approximately 45% of the strains. All alleles were sequenced and their expression was confirmed. Evolutionary analyses determined that the diverse HopZ1 homologs are ancestral to P. syringae, and have diverged via pathoadaptive mutational changes into three functional and two degenerate forms, while HopZ2 and HopZ3 have been brought into P. syringae via horizontal transfer from other ecologically similar bacteria. A PAML selection analysis revealed that the C terminus of HopZ1 is under strong positive selection. Despite the extensive genetic variation observed in this family, all three homologs have cysteine-protease activity, although their substrate specificity may vary. The introduction of the ancestral hopZ1 allele into strains harboring alternate alleles results in a resistance protein-mediated defense response in their respective hosts, which is not observed with the endogenous allele. These data indicate that the P. syringae HopZ family has undergone allelic diversification via both pathoadaptive mutational changes and horizontal transfer in response to selection imposed by the host defense system. This genetic diversity permits the pathogen to avoid host defenses while still maintaining a virulence-associated protease, thereby allowing it to thrive on its current host, while simultaneously impacting its host range.

Figure 1. Phylogenetic Analyses

(A) Neighbor-joining tree of YopJ family of T3SEs proteins. Bootstrap support is indicated above each node, with only values >60% being shown. T3SEs from P. syringae are highlighted. Accession numbers for each protein are presented in parentheses following the protein name and species. (B) Neighbor-joining gene genealogy of the hopZ1 T3SEs alleles with bootstrap analysis as above and hopZ2_Ppi895A used as an out-group. The genetic organization of the three functional allele classes (hopZ1a, hopZ1b, and hopZ1c) and two degenerate hopZ1 alleles (ψhopZ1a and ψhopZ1b) are presented to the right of the gene genealogy. The large gray rectangle represents the region of shared similarity among the alleles, and corresponds to the coding sequence of hopZ1a. The dark vertical rectangle to the left of each gene represents the T3SS promoter element known as the hrp box. The solid black vertical line represents the stop codon for each allele. Arrows at the 5′ end of the coding sequences indicate which alleles are functional. Triangles above each coding sequence indicate insertions, with the insertion size indicated within. The ψhopZ1a allele has a nonsense mutation at nucleotide 171, and a degenerate hrp box in which the conserved GGAACC sequence has mutated to CCAACC. It is not transcribed under T3SS induction conditions. The ψhopZ1b allele is disrupted by an insertion sequence.

Figure 2. Distribution of HopZ T3SEs in the P. syringae Species Complex

The distribution of HopZ T3SEs in relation to the core genome phylogeny of 96 natural P. syringae isolates. The phylogenetic tree (neighbor-joining, 1,000 bootstrap) was generated by MLST of four housekeeping genes. The host of isolation is presented to the right of each strain. See [40,65] and Table S1 for details. Black squares represent presence of the corresponding gene, whereas white ones represent absence. hopZ1 alleles are annotated. The full sequence of hopZ1a_PsyB76 was not obtained although its expression was confirmed.

Figure 3. Congruence between the HopZ1 Gene Genealogy and the MLST Core Genome Phylogeny of P. syringae

See Figures 2 and S1 for details on the phylogenetic methods used. Gray areas labeled g2–g5 in the center of the figure indicate the major phylogroups as determined by MLST. Bootstrap support values (>60%) are indicated above nodes.

Figure 4. PAML Selection Analysis

(A) Amino-acid alignment of the HopZ1 alleles. See Figure S3 for details. The three amino acids in the catalytic triad are indicated by arrows presented above the alignment. (B) Stacked histogram representing the posterior probabilities for the three site classes identified by codeml model M3 (discrete) along the HopZ1 sequence for all functional alleles. The dN/dS (ω) values for each site class are given below the figure. (C) Same as above but using only hopZ1a and hopZ1b to remove the signal from the frameshift mutation in hopZ1c.

Figure 5. Protease Activity of HopZ T3SEs in P. syringae

Protease activity of HopZ1a_PsyA2, HopZ1c_PmaES4326, HopZ2_Ppi895A, and HopZ3_PsyB728a measured using Invitrogen RediPlate 96 EnzChek Protease Assay Kit green fluorescence is shown as fluorescence production (OD 535) after a 1-h incubation of the purified His-tag fusion proteins with a green fluorescence–labeled substrate at room temperature. Trypsin was used as a positive control. The grid along the bottom of the figure indicates if the wild-type protein was used, if plant extract was added to the reaction, or if the cysteine-to-alanine mutant protein was used. The error bars indicate standard errors between two replicates. Each experiment was repeated three times with similar results except for HopZ3_PsyB728a, which gave inconsistent results due to instability of the protein. A gelatine in-gel protease assay was performed to confirm the enzymatic activity of HopZ3_PsyB728a. Letters above the histogram bars indicate significance classes from a Fisher's Least Significant Difference test. wt, wild-type.

Figure 6. In Vivo Defense Induction Assays

The plant host, infiltrated strain, and hopZ allele carried by the infiltrated strain are indicated below each leaf image. (A–E) Z1a, Z1b, Z1c, and Z2 represent the infiltration zones of HopZ1a, HopZ1b, HopZ1c, and HopZ2, respectively. Z1a-C/A and Z2-C/A identify the infiltration zone of the HopZ1a and HopZ2 alleles carrying the cysteine-to-alanine replacement in the catalytic core, respectively. C represents the vector control. In all panels HopZ1a induces an HR, while the other infiltrations either result in disease symptoms or no response. (A) Transient expression assay using N. benthamiana leaves infiltrated with A. tumefaciens C58C1(pCH32), carrying HopZ1 or HopZ2 T3SEs in pMDD1 vector. (B) Soybean (Glycine max cv. OAC Bayfield) leaves infiltrated with PgyBR1, carrying pUCP20 or pUCP20HopZ1a _PsyA2. PgyBR1 carries hopZ1b natively. (C) A. thaliana (column 1) leaves infiltrated with PmaES4326, carrying pUCP20 or pUCP20HopZ1a _PsyA2. PmaES4326 carries hopZ1c natively. (D) Rice (O. sativa cv. Jefferson) leaves infiltrated with Por36–1 carrying pUCP20 or pUCP20HopZ1a _PsyA2. Por36–1 carries ψhopZ1a natively. (E) Sesame (S. indicum) leaves infiltrated with PseHC-1 carrying pUCP20 or pUCP20HopZ1a _PsyA2. PseHC-1 carries ψ hopZ1b natively.

Figure 7. Arms-Race Model

(A–C) Star-shaped figures in bacteria represent T3SEs. T₁ and T₂ are two virulence targets in the host. R is a resistance protein that induces the host-defense response upon recognition of a plant-virulence target modified by a bacterial effector. (A) The pathogen secretes the ancestral HopZ1a T3SE which modifies a host-virulence target and contributes to the disease process. (B) The plant host evolves or acquires an R protein that recognizes the modified target of HopZ1a, and uses this as a cue to induce the defense response, making the plant host resistant to the pathogen. (C) Strong positive selection imposed by the host R protein results in the evolution of modified HopZ T3SEs (e.g., HopZ1b and HopZ1c) or the acquisition of homologs from ecologically similar plant pathogens (e.g., HopZ2 and HopZ3). These T3SEs retain cysteine-protease function, but are not recognized by the R protein, and therefore do not induce a defense response. They may avoid R-protein recognition by either attacking a new virulence target in the host, or by modifying the original target so that it is not recognized by the R protein.