Phytophthora infestans effector AVRblb2 prevents secretion of a plant immune protease at the haustorial interface

To gain insight into AVRblb2 virulence activities inside host cells, we constructed a functional N-terminal GFP fusion to mature AVRblb2 (lacking the signal peptide) (Fig. S1 A and B). The GFP:AVRblb2 fusion protein accumulated mainly at the cell periphery when expressed in Nicotiana benthamiana (Fig. 1A). The fluorescence signal remained associated with the plasma membrane after salt-induced plasmolysis of the epidermal cells (Fig. 1A). The AVRblb2 signal largely overlapped with the red fluorescence of a coexpressed plasma membrane-localized RFP (pm-RK), confirming that AVRblb2 accumulates at the host plasma membrane (Fig. 1A). Similar membrane localization of AVRblb2 was observed with an N-terminal RFP fusion (Fig. 1A) and in stable transgenic GFP:AVRblb2 lines of N. benthamiana (Fig. S1C). To determine the extent to which localization of AVRblb2 is altered during infection, we inoculated the GFP:AVRblb2 N. benthamiana lines with several P. infestans strains, including 88069td, a transgenic strain expressing the red fluorescent marker tandem dimer RFP (known as tdTomato) (13). The AVRblb2 signal preferentially accumulated around haustoria inside infected plant cells, whereas its even distribution at the plasma membrane remained unaltered in cells without haustoria (Fig. 1B). Haustorial accumulation of AVRblb2 first occurred at discrete focal sites (one per haustorium) before covering the entire surface of the haustoria (Fig. 1B and Fig. S1C). Perihaustorial accumulation of AVRblb2 only partially overlapped with callose deposited around haustoria and was not associated with callose encasements (Fig. S1D) (14).

To determine the degree to which AVRblb2 affects P. infestans infection, we performed pathogen assays with the transgenic GFP:AVRblb2 N. benthamiana plants. The GFP:AVRblb2 plants showed enhanced susceptibility to P. infestans, resulting in increased pathogen colonization and sporulation relative to control lines (Fig. 1 C and D and Fig. S2). This finding indicates that AVRblb2 has a virulence activity and suggests that it causes significant impairment of host defense responses.

To identify the host targets of AVRblb2, we used in planta coimmunoprecipitation (co-IP) followed by liquid chromatography–tandem mass spectrometry (LC-MS/MS). In total, we detected five N. benthamiana proteins that specifically associated with AVRblb2 (Table S1). One of these targets is the PLCP C14, a conserved solanaceous protein orthologous to Arabidopsis RD21, rice Oryzain, and maize Mir3 cysteine proteases (15–20). C14 is a complex modular protein featuring a predicted N-terminal secretion signal and a self-inhibitory prodomain, which is followed by peptidase, proline-rich, and granulin domains (Fig. 2A). In plant cells, C14/RD21 converts into immature (iC14) and mature (mC14) isoforms (Fig. 2A) that accumulate into various subcellular compartments and the apoplast (16–18). C14 and other PLCPs have been implicated in plant immunity, including pathogen perception and disease resistance (5, 21). Like some other PLCPs, C14 is also targeted by apoplastic protease inhibitor effectors of P. infestans and other filamentous pathogens (5, 6, 18), and the expression of the potato C14 gene is rapidly induced during P. infestans infection (15). Thus, we decided to initially focus on C14, and we will study other AVRblb2-associated proteins in the future.

We validated the specific in planta association between AVRblb2 and tomato C14 (LeC14) with co-IP (Fig. 2B). AVRblb2-purified immunocomplexes contained iC14 (Fig. 2B), and, occasionally, the less abundant mC14 isoform of the protease could be detected (Fig. S3). We next addressed whether C14 associates with AVRblb2 in planta by using confocal microscopy. As previously reported for the C14 ortholog RD21 (16, 17), fluorescently tagged full-length C14 and a C14 construct lacking the granulin domain (C14pep) localized to the endoplasmic reticulum (ER), endomembrane compartments, vacuoles, and apoplast (Fig. 2C and Fig. S4 B–E) and accumulated around haustoria in infected cells (Fig. S4F). Notably, in the presence of AVRblb2, the localization of C14 dramatically shifted toward the cell periphery, resulting in a marked overlap with the fluorescence signals of AVRblb2 and a plasma membrane marker (Fig. 2C and Fig. S5A). A similar shift in C14 distribution was observed in plant cells containing haustoria (haustoriated plant cells), resulting in colocalization and focal accumulation of AVRblb2 and C14 around haustoria (Fig. 2D). Altogether, these results indicate that AVRblb2 associates with C14 and alters its subcellular distribution.

The increased peripheral accumulation of C14 triggered by AVRblb2 prompted us to address whether AVRblb2 affects secretion of C14. AVRblb2 significantly reduced apoplastic levels of C14 without affecting its intracellular accumulation (Fig. 3A and Fig. S5B). AVRblb2 did not inhibit secretion in general because it did not affect secretion of the pathogenesis-related serine protease P69B nor did it reduce overall protein levels in the apoplast (Fig. S5C). We confirmed these observations by microscopy. AVRblb2 significantly reduced the apoplastic accumulation of a C14pep:RFP fusion protein but increased its levels in the cytoplasm, mainly in the cell periphery and vacuoles (Fig. 3B). In contrast, AVRblb2 did not alter the apoplastic levels of a fusion of basic chitinase signal peptide to GFP (SP:GFP), confirming again that AVRblb2 does not generally prevent secretion (Fig. S5D). In summary, these results indicate that AVRblb2 inhibits secretion of the host protease C14.

Consistent with these observations, we noted a gradual decrease in C14 levels in the tomato apoplast during P. infestans infection starting at 24 h after inoculation (Fig. 3C). Intracellular levels of C14 were also altered during infection with an increase in iC14 levels (Fig. 3C). These changes contrast sharply with the well-known increase in apoplastic levels of pathogenesis-related proteins, such as P69B, over the course of infection (Fig. 3C) (22). The decrease in apoplastic C14 during infection could be attributable to the preferential targeting of this protein to haustorial sites, where it is prevented from secretion. However, based on this experiment, we cannot exclude that the observed effect might involve additional effectors to AVRblb2.

To address the link between AVRblb2 haustorial localization and its function, we tested several deletion mutants and identified an 8-aa C-terminal deletion mutant (AVRblb2^47–92) that did not exclusively accumulate at the cell periphery or around haustoria (Fig. 4 A–C). Notably, this mutant lost its ability to enhance P. infestans growth (Fig. 4D), weakly associated with C14 protease, and failed to attenuate apoplastic C14 accumulation (Fig. 4 E and F). Thus, localization of AVRblb2 and C14 secretion inhibition are genetically linked to the enhanced virulence activity of this effector. Conversely, the mutant retained its activity to trigger an Rpi-blb2–dependent hypersensitive response (Fig. 4G). These results indicate that AVRblb2 localization is essential for its virulence function but not for its avirulence activity.

To determine the extent to which C14 contributes to plant immunity, we altered C14 expression in N. benthamiana. Stable RNAi N. benthamiana lines, carrying two independent C14 hairpin constructs (5), showed significantly enhanced susceptibility to P. infestans compared with control lines as assessed by increased disease symptoms (Fig. S6). Conversely, overexpression of both C14 and C14pep in N. benthamiana resulted in reduced P. infestans colonization relative to controls (Fig. 5 and Fig. S7 A and B). This enhanced immunity conferred by C14 could be partially reversed by simultaneous overexpression of AVRblb2 (Fig. S7 C and D). In the course of these experiments, we also determined that C14 silencing did not affect Rpi-blb2 recognition of AVRblb2, indicating that C14 is not “guarded” by Rpi-blb2 (Fig. S8) (23, 24). These results indicate that C14 plays a positive role in plant immunity. We propose a model in which C14 is targeted by two classes of effectors in separate plant cell compartments: whereas AVRblb2 prevents secretion of C14 within haustoriated plant cells, other P. infestans effectors inhibit C14 in the apoplast as shown by Kaschani et al. (5) (Fig. 6).

In this study, we showed that the P. infestans effector AVRblb2 interferes with defense-related secretion in haustoriated plant cells. AVRblb2 can enhance susceptibility by reducing overall C14 levels in the host apoplast. The execution of plant cell-autonomous immunity requires cytoskeleton reorganization and polarized secretion (25, 26). Although some bacterial effectors are known to target secretory pathways, little is known about the underlying cellular and molecular processes (27–29). The role of focal secretion in immunity has been difficult to dissect with standard genetic approaches because mutants often show pleiotropic effects that perturb plant development (26, 30). Our work indicates that effectors can be used as molecular probes to unravel unknown facets of focal immunity and potentially dissect the diversity of secretory vesicles and their cargo.

Our results implicate focal secretion of a plant defense protease in plant immunity. C14 could contribute to immunity by degrading non-self molecules or by playing a signaling role. C14 is known to accumulate during senescence and dehydration stress, but whether it also contributes to these processes is unclear (31, 32). During abiotic stress, C14 undergoes complex changes in subcellular localization. It accumulates in atypical ER-derived vesicles and, upon dehydration stress, can directly traffic into the vacuole in a nonclassical way bypassing the Golgi pathway (16, 17). Also, during desiccation, C14 is released to the cytosol and nucleus (32). We showed that perturbation of C14 trafficking by P. infestans AVRblb2 limits its role in plant immunity. AVRblb2 association with C14 might occur directly or indirectly through an intermediate transmembrane protein or molecule. AVRblb2 could intercept C14 at various subcellular sites, including during release or fusion of secretory vesicles to the plasma membrane at the haustorial interface. Finally, our findings that C14 overexpression overcomes AVRblb2 activity and leads to enhanced resistance to P. infestans point to immediate biotechnological applications for engineering late blight-resistant potato and tomato crops.

Effector expression constructs were designed to have N-terminal epitope or fluorescent tags replacing the native signal peptide to enable intracellular expression in plant cells. Detailed information about the plasmid constructs, transient gene-expression assays, and production of transgenic plants is described in SI Materials and Methods.

FLAG:AVRblb2 and its plant interactors were coimmunoprecipitated with anti-FLAG resins under nondenaturing conditions. Co-IP experiments, preparation of peptides for LC-MS/MS, and Western blot analysis are described in SI Materials and Methods.

Secretion inhibition assays were performed by using Agrobacterium-mediated transient gene expression. Detailed procedures are provided in SI Materials and Methods.

Hypersensitive response assays on C14 silencing are described in SI Materials and Methods.

Detailed RT-PCR procedures are explained in SI Materials and Methods.

Confocal microscopy analysis was performed on Leica DM6000B/TCS SP5 confocal microscope (Leica Microsystems). Detailed procedures of confocal microscopy and callose/aniline blue staining of infected material is described in SI Materials and Methods.