Papain-like immune proteases (PLCPs) are promising engineering targets for crop protection, given their significant roles in plant immunity for key crops such as tomato, maize and citrus (Misas-Villamil et al., 2016). The wide range of pathogen-secreted PLCP inhibitors highlights the importance of these proteases in defending against various pathogens. Depletion of the apoplastic immune PLCP Phytophthora-inhibited protease 1 (Pip1) from tomato, for instance, causes hyper-susceptibility to bacterial, fungal and oomycete tomato pathogens (Ilyas et al., 2015). Immunity by Pip1 in wild-type tomato is, however, suboptimal since Pip1 is suppressed during infection by diverse pathogensecreted inhibitors, such as the cystatin-like EpiC2B from the oomycete late blight pathogen Phytophthora infestans (Tian et al., 2007). Here, we tested whether we could increase Pip1-based immunity against late blight by engineering Pip1 into an EpiC2B-insensitive protease. To guide Pip1 mutagenesis, we generated a structural model of the EpiC2B-Pip1 complex using AlphaFold-Multimer (Evans et al., 2022). This structural model represents a classic interaction between the tripartite wedge of cystatin (EpiC2B) in the substrate binding groove of papain (Pip1). This model indicated that engineering Pip1 to prevent inhibition without affecting Pip1 substrate specificity is possible because the interaction surface of Pip1 with EpiC2B is larger than the substrate binding groove (Figure 1a). We selected nine residues for targeted mutagenesis of Pip1 that are predicted to directly interact with EpiC2B but are not in the substrate binding groove (Figure 1a). To generate the greatest disruption in protease-inhibitor interaction, we substituted these residues into bulky amino acids of the opposite charge. Following this strategy, we generated four double mutants and one single mutant of Pip1 and produced these proteins in planta via agroinfiltration (Table S1, S2). All proteins were detected in the apoplast of agroinfiltrated Nicotiana benthamiana plants using the anti-Pip1 antibody (Figure 1b), and all are active proteases because they reacted with TK011, a new fluorescent activitybased probe for PLCPs (Figure S1). Preincubation with 3 μM purified EpiC2B reduced labelling of Pip1 and some of the Pip1 mutants, but the D147R/V150E mutant of Pip1 seems insensitive to EpiC2B inhibition (Figure 1b) and was hence called engineered Pip1 (ePip1). Preincubation with increasing EpiC2B concentrations revealed that ePip1 is insensitive to up to 8 μM EpiC2B, whereas wild-type Pip1 can be inhibited by 1 μM EpiC2B (Figure 1c). We next tested whether ePip1 enhances immunity against late blight. Taking advantage of the fact that N. benthamiana is a natural null mutant for Pip1 (Kourelis et al., 2020) and that infections of agroinfiltrated leaves are well-established assays for P. infestans, we first tested whether transient expression of wildtype Pip1 enhances late blight immunity. Wild-type Pip1 and the catalytic mutant of (C153A/C154A, Pip1*) were transiently expressed side-by-side in N. benthamiana leaves, and the agroinfiltrated area was infected with zoospores of transgenic P. infestans expressing fluorescent tDtomato (Chaparro-Garcia et al., 2011). Tissue expressing Pip1 supported reduced P. infestans growth when compared to the Pip1* negative control (Figure 1d), confirming that Pip1 is an immune protease, also when expressed in N. benthamiana. Importantly, growth of fluorescent P. infestans was further reduced in tissues expressing ePip1 when compared to wild-type Pip1 (Figure 1d), demonstrating that EpiC2B-insensitive ePip1 increases immunity against P. infestans.

Thus, we …