Heterologous Expression of PKPI and Pin1 Proteinase Inhibitors Enhances Plant Fitness and Broad-Spectrum Resistance to Biotic Threats

Kunitz-type (PKPI) and Potato type I (Pin1) protease inhibitors (PIs) are two families of serine proteinase inhibitors often associated to plant storage organs and with well known insecticidal and nematicidal activities. Noteworthy, their ability to limit fungal and bacterial pathogenesis in vivo or to influence plant physiology has not been investigated in detail. To this aim, we generated a set of PVX-based viral constructs to transiently and heterologously express two potato PKPI (PKI1, PKI2) and three potato Pin1 (PPI3A2, PPI3B2, PPI2C4) genes in Nicotiana benthamiana plants, a widely used model for plant-pathogen interaction studies. Interestingly, transgenic plants expressing most of the tested PIs showed to be highly resistant against two economically important necrotrophic fungal pathogens, Botrytis cinerea and Alternaria alternata. Unexpectedly, overexpression of the PKI2 Kunitz-type or of the PPI2C4 and PPI3A2 Potato type I inhibitor genes also lead to a dramatic reduction in the propagation and symptom development produced by the bacterial pathogen Pseudomonas syringae. We further found that localized expression of PPI2C4 and PKI2 in N. benthamiana leaves caused an increase in cell expansion and proliferation which lead to tissue hypertrophy and trichome accumulation. In line with this, the systemic expression of these proteins resulted in plants with enhanced shoot and root biomass. Collectively, our results indicate that PKPI and Pin1 PIs might represent valuable tools to simultaneously increase plant fitness and broad-spectrum resistance toward phytopathogens.


INTRODUCTION
PKPI and Pin1 are among the most abundant naturally occurring plant serine proteinase inhibitors (PIs). Large amounts of these inhibitors accumulate in plant reproductive and storage organs, as in the case of Solanum tuberosum (Heibges et al., 2003;van den Broek et al., 2004). Nevertheless, their genes are also transcribed, however at lower levels, in all other plant tissues (Kuo et al., 1984;Lincoln et al., 1987;Jofuku and Goldberg, 1989;Heitz et al., 1993;Wang et al., 2003Wang et al., , 2008Turrà et al., 2009). Serine PIs expression apart from being regulated at developmental, spatial and speciesspecific level (Lee et al., 1986;Rosahl et al., 1986;Balandin et al., 1995;Singh et al., 2009;Tamhane et al., 2009;Turrà et al., 2009), is boosted up by various external stimuli including wounding, insect feeding and microbial infections (Mello et al., 2001;van Loon et al., 2006;Turrà and Lorito, 2011), being one of the bestcharacterized defense reactions activated by the plant in response to pathogen and insect attack. Many studies on the effect of serine PIs, either artificially introduced into diets or heterologously expressed in transgenic plants, have shown the ability of these proteins to reduce the growth and development of a wide range of herbivorous insects and pathogenic nematodes mainly by interfering with nutrient digestibility and fertility (Jongsma et al., 1995;Urwin et al., 1995;Altpeter et al., 1999;Andrade et al., 2003;Cai et al., 2003;Srinivasan et al., 2005).
Besides, few serine PIs have also shown inhibitory activity against bacteria and fungi in vitro by reducing their growth or conidial germination and hyphal swelling, respectively (Lorito et al., 1994;Dunaevsky et al., 1996;Chen et al., 1999;Soares-Costa et al., 2002;Kim et al., 2005;Hermosa et al., 2006;Kim et al., 2006;Di Cera, 2009). However, while serine PIs insecticidal and nematicidal activities have been efficiently proven in planta, their ability to alter plant resistance against fungal or bacterial pathogens in vivo has remained elusive.
Proteases and PIs play important roles in plant-pathogen interactions; nevertheless, testimony for their endogenous role in plants is relatively recent and our current understanding of the diverse physiological processes regulated by PIs is rapidly expanding (van der Hoorn, 2008;Turrà and Lorito, 2011;Grosse-Holz and van der Hoorn, 2016). To date, protease-PI interactions have been shown to regulate many diverse aspects of the plant life cycle including senescence and programmed cell death (PCD), leaf trichome density and branching, seed and flower development and sieve element maturation (Solomon et al., 1999;Xu et al., 2001;Sin and Chye, 2004;Pak and Van Doorn, 2005;Liu et al., 2006;Xie et al., 2007;Luo et al., 2009;Boex-Fontvieille et al., 2015;Rustgi et al., 2017).
As proteolysis is a fundamental process in all living beings, in order to avoid undesired side-effects plants must carefully control endogenous protease activity in both a timely and a spatial manner (van der Hoorn, 2008;Turrà and Lorito, 2011). In a previous work from our group, we have shown that PI gene members of the Pin1 and PKPI families are differentially expressed in Solanum tuberosum var. Desireè plants upon abiotic or biotic insults, or in a tissue-dependent manner, thus indicating a possible role for these PIs as both endogenous-and defenserelated plant regulators (Turrà et al., 2009).
In this study, the effect of transient expression of different members of the Pin1 and PKPI families on plant resistance toward fungal and bacterial pathogens and on plant physiology is reported. When heterologously expressed in Nicotiana benthamiana, different potato PKPI and Pin1 genes confer protection against B. cinerea and A. alternata, two agronomically important pathogens. Moreover, in vivo assays designed to challenge PIs-expressing plants with Pseudomonas syringae pv. tabaci also revealed enhanced plant resistance to bacterial attack. In addition to this, overexpression of two of these serine PI genes also caused severe developmental effects on N. benthamiana plants, including over-accumulation of trichomes and growth enhancement. These phenotypes were accompanied by the high inhibitory activity of total soluble proteins (TSP) extracted from transformed leaf patches toward yet unknown proteases present in the N. benthamiana leaf apoplast. Based on these results, we propose that Pin1 and PKPIs are critically involved in host resistance and modulation of plant physiology.

Microbial Strains, Plants, and Culture Conditions
Nicotiana benthamiana plants were cultivated and maintained at 25 • C in a phytocabinet under 16/8 h light-dark photoperiod.

Construction of PVX::PI Gene Fusions
To amplify full-length cDNAs of PKI1, PKI2, PPI3A2, PPI3B2, and PPI2C4 genes (Hermosa et al., 2006;Turrà et al., 2009), total RNA was extracted from 100 mg of Solanum tuberosum var. Desireè sprouts using the TRI Reagent (Ambion, Austin, TX, United States). First-strand cDNA was synthesized using the Reverse Transcription System kit (Promega, Madison, WI, United States) and 1 µg random primers for every 2 µg of total RNA, following the supplier's instructions. The PKPI and Pin1 derivatives were amplified by PCR using the oligonucleotide combinations indicated in Supplementary Table S1, subcloned into pGEM-T Easy vector (Promega), and ligated into the ClaI and SalI sites of the A. tumefaciens binary PVX vector pGR106 (Lu et al., 2003).
Transient Expression of PIs Genes in N. benthamiana and in vivo Resistance Assay on PVX-Infected Plants N. benthamiana seedlings at the fourth true-leaf stage were used for A. tumefaciens infiltration. To evaluate local effects of PI overexpression, A. tumefaciens overnight cultures diluted to an OD600 of 0.25 with sterile distilled water were used to infiltrate the abaxial side of the leaf (using a needleless 5 ml syringe). Alternatively, to achieve systemic transformation of plants, third and fourth leaves of 2-3-week-old N. benthamiana seedlings were wounded twice along the midvein with a sterile wooden toothpick previously streaked over an A. tumefaciens culture grown on solid agar medium.
For P. syringae infection, 5 days after A. tumefaciens syringeinfiltration, the same leaf areas were infiltrated (the abaxial side of the leaf) with 40 µl of a 1 × 10 8 P. syringae cells/ml culture (adjusted with sterile distilled H 2 O to OD600 = 0.24). After 2, 4, 7, and 9 days, the necrotic zone around the inoculation site was imaged. Every treatment was repeated at least 6 times and on at least three different plants. A separate round of experiments was used to quantify P. syringae growth in the agroinfiltrated leaves. Ten µl of a P. syringae cell suspension (OD600 = 0.24) was applied to needle-pricked leaves. Plants were covered with clear polyethylene bags and sealed around the base using elastic bands, to keep humidity. After 1, 3, and 6 days, six leaf discs (0.8 cm diameter) per treatment were excised from inoculated areas, pooled and ground (Ultra-Turrax T25 basic, IKA Labortechnik, Germany) in 10 mM MgSO 4 (1 ml/disc), by keeping the tube in an ice bath. An aliquot of the homogenate was plated on LB-rifampicin agar at three different dilutions (10 −3 , 10 −4 , and 10 −5 ) and colonies were counted 1 and 2 days after incubation at 28 • C. Another aliquot (15 µl) of the homogenate was mixed with 110 µl of LB-rifampicin and incubated in 96-well microtiter plates at 28 • C (120 rpm for 16 h). Optical density at 550 nm was then measured with a Bio-Rad microplate reader (Bio-Rad, Richmond, CA, United States). P. syringae inoculation was independently repeated on at least six leaves of at least three different plants.
For B. cinerea and A. alternata infections, upper leaves of systemically transformed N. benthamiana plants (11 days after A. tumefaciens toothpick inoculation) were challenged with 10 µl of germination solution (20 mM glucose, 20 mM potassium phosphate) containing 10 5 or 10 7 conidia/ml of B. cinerea or A. alternata, respectively. All plants were covered with transparent polyethylene bags and sealed around the base using elastic bands, to keep humidity. The appearance of necrotic spots was assessed 2, 4, and 6 days after inoculation and disease incidence recorded.
Each pathogen-PVX construct combination was assayed on at least four different leaves of at least three different plants. All infection assays were repeated at least twice. Data are presented as mean values ± SD of different experiments. To assess statistical differences between control (PVX::gfp) and PI expressing samples a Yates' corrected chi-squared test (twosided) was used. Statistical differences between treatments at each time point were assessed by one-way ANOVA with post hoc Tukey HSD Tests.

RT-PCR Analysis of PIs Expression
For RT-PCR validation of transient PI expression, total RNA was isolated from 100 mg of control (PVX::gfp) and PIs transformed N. benthamiana leaves (5 days post agroinfiltration) using the TRI Reagent (Ambion). RT-PCR was performed on equal amounts of total RNA using the Reverse Transcription System kit (Promega) and of 1 µg random primers for each 2 µg total RNA, following the supplier's instructions. Two microliters of first-strand cDNA solution were used as a template for RT-PCR experiments. Amplifications of Pin1, PKPI, and β-tubulin gene transcripts were performed as indicated earlier (Turrà et al., 2009). Plasmid DNA of the cloned cDNAs PPI3B2 and PKI1 (Turrà et al., 2009) were used as positive controls and PCR products of the constitutively expressed β-tubulin gene used as a quantitative control. Amplifications were repeated in independent occasions on neo-synthesized cDNA from at least three independently repeated experiments.

Protein Extractions
For the extraction of the TSP, ten grams of untransformed or of locally transformed N. benthamiana leaf areas (7 days after A. tumefaciens syringe-inoculation) were ground with an Ultra-Turrax Homogenizer in 30 ml of 0.1 M sodium phosphate buffer (pH 6.8) by keeping the tube in an ice bath. The slurry was incubated 1h on ice under occasional shaking, filtered through four layers of cheesecloth, and cleared by centrifugation at 50,000 g for 1 h at 4 • C in a Beckman L7-65 Ultracentrifuge (Beckman, Milan, Italy).
Apoplastic fluids (AF) were prepared from N. benthamiana leaves according to the method of Moehnke et al. (2008), with minor modifications. Briefly, 10 g of leaf material was vacuuminfiltrated for 2 min with 100 ml infiltration buffer [100 mM Tris/HCl (pH 7.5), 10 mM MgCl 2 ]. Leaves were then dried with sterile paper towels and placed into the barrel of a 50 ml syringe. The syringe was subsequently inserted into a 50 ml falcon tube with the needle hub facing downwards and spun at 2,000 g for 10 min at 4 • C. After centrifugation, AF was collected from the bottom of the centrifuge tube.
All protein extracts were subjected to filtration and dialysis by using Centriprep YM-3 devices (Amicon Corporation, Danvers, MA, United States), filter sterilized (0.22 µm) and stored at −20 • C if not immediately used. Protein concentrations were determined by a Bradford DC protein assay (Bio-Rad) using bovine serum albumin as a standard.

In vitro Evaluation of Plant Crude Extracts Activity
Inhibition assays of N. benthamiana TSP and AF proteolytic activities by TSP extracted from PVX::gfp, PVX::PKI2, and PVX::PPI2C4 transformed N. benthamiana leaf areas were carried out in microtiter plates by using azocasein (Sigma-Aldrich) as chromogenic substrate and in-gel protease assays using the Bio-Rad zymogram buffer system, following the previously described procedures (Tian et al., 2004;Hermosa et al., 2006). For the first method, a total of 20 µg of TSP and AF from untransformed plants were preincubated with 20 µg of TSP from PVX-transformed plants in a volume of 250 µl for 30 min at room temperature, and followed by incubation with 200 µl of 1% azocasein (w/v) at 37 • C for 1 h. The reaction was halted by adding an equal volume of 10% (w/v) TCA. After 10 min on ice, the reaction mixture was centrifuged for 10 min at 13,000 g and the supernatant mixed with an equal volume of 1 M NaOH. The optical density at 450 nm was then determined with a Bio-Rad microplate reader. The percentage of the remaining protease activity was therefore plotted relative to that of TSP and AF samples from untransformed plants incubated with TSP extracted from PVX:gfp transformed plants. Experiments were performed in triplicate and repeated in at least two independent occasions. Data represent the mean value [±SD (standard deviation)] across all experiments. Statistical differences between treatments were assessed by one-way ANOVA with post hoc Tukey HSD Tests.
For zymogen in-gel protease assays, 20 µg of TSP from mock-(H 2 O) and PVX-infiltrated N. benthamiana leaf areas were mixed with zymogram sample buffer and loaded on a 10% SDS-polyacrylamide gel without boiling or addition of reducing reagents. Following electrophoresis, the gel was incubated in 1x zymogram renaturation buffer for 30 min. Then the gel was incubated in 1x zymogram development buffer for 18 h at 37 • C before staining with 0.5% Coomassie Brilliant Blue. Areas of protease activity were revealed as cleared bands on a blue background.
To assess the antifungal activity of plant crude extracts, 10 µl of a solution of 10 7 conidia/ml of B. cinerea or A. alternata were mixed with 40 µl of plant crude extracts (50 µg/ml) and 40 µl of Potato Dextrose Broth (PDB). After 48 h of incubation at 28 • C in a 96-well microtiter plate, the change in optical density at 550 nm was determined using a Bio-Rad microplate reader. Each experiment was repeated at least three times and data presented as the percentage of growth inhibition relative to that of the PVX:gfp transformed plants. Data correspond to mean values (± SD) across all experiments. Statistical differences between treatments were assessed by one-way ANOVA with post hoc Tukey HSD Tests.

Determination of Plant Growth and Analysis of Leaf Surface Expansion
The effect of PI transient expression on N. benthamiana shoot and root growth was evaluated in pot experiments. Seedlings were grown in vitro at 23 • C under 24 h fluorescent lighting (3,500-6,000 lux) on Murashige and Skoog (MS) salt medium (ICN Pharmaceuticals Inc., Cleveland, OH, United States), and 1% bacto-agar (Difco, Detroit, MI, United States) first and then transferred to soil in 10 cm diameter pots and left to grow in a phytocabinet as described above. Plantlets at the fourth-leaf stage were A. tumefaciens toothpick inoculated, as described above. Complete shoots and roots were collected separately (21 days after agroinfection). Roots were briefly rinsed to remove attached sand, quickly dried with a paper towel, incubated for 72 h at 75 • C and weighted to estimate the dry weight. Experiments were repeated twice (n = 10). Data correspond to mean values (± SD) across all experiments. Statistical differences between treatments were assessed by one-way ANOVA with post hoc Tukey HSD Tests.
To determine leaf disc surface expansion and leaf strip curvature of PI transformed leaf areas, the methods were adapted from those of Gevaudant and coworkers (Gevaudant et al., 2007). Briefly, fully developed PVX-transformed N. benthamiana leaf areas (7 days after A. tumefaciens syringe-inoculation) were used. Leaf discs (1 cm diameter) and strips (2 × 10 mm) were cut from the interveinal region and incubated for 24 h at room temperature in 10 mM Sucrose, 10 mM KCl, and 0.5 mM 2-(N-Morpholino)ethanesulfonic acid hemisodium salt (MES), pH 6.0. Leaf discs were photographed with a digital camera before and after the treatment and their surface estimated as their pixel content, by the use of the ImageJ software (Collins, 2007). Differences in surface increase between treatments and controls (PVX::gfp) were expressed as a percentage of the initial disc area. Leaf strip curvature was estimated as the angle made by the two tangents to the two terminal parts of each strip. Each experiment was repeated three times on at least 15 leaf discs or strips per treatment. Statistical differences between treatments were assessed by one-way ANOVA with post hoc Tukey HSD Tests.
For microscopic analysis of cell size and nuclei density, adaxial epidermis from agroinfiltrated (7, 13, and 20 days after A. tumefaciens syringe-infiltration) N. benthamiana leaf areas was used. Briefly, entire leaves were detached and immersed for 1 h in 1% (v/v) Tween 20 before peeling off the adaxial epidermis from the agro-infiltrated area. Tissues were mounted in water and observed with an Axioskop2 Plus microscope (Zeiss, Milan, Italy). Cell size measurement was performed by using the ImageJ software (Collins, 2007). For nuclear visualization, leaf epidermis was stained with 4 ,6-diamidino-2-phenylindole (DAPI; 1 µg/ml) for 20 min, and mounted in 50% (v/v) phosphatebuffered saline (PBS)-glycerol for observation. Quantification of epidermial cell size and nuclear density was repeated at least four times on 500 cells/leaf per treatment. To assess statistical differences between control (PVX and PVX:gfp) and PI expressing samples a Yates' corrected chi-squared test (two-sided) was used.

Sequence Data and Bioinformatic Analysis
BLAST searches were performed in the NCBI database 1 . Signal peptide and extracellular localization predictions were performed by using the SignalP and WoLF PSORT softwares, respectively (Emanuelsson et al., 2007;Horton et al., 2007).

Potato Kunitz and Pin1 PIs Are Efficiently Expressed in Nicotiana benthamiana Plants
Five PCR fragments corresponding to three previously characterized (PPI3A2, PPI3B2, and PPI2C4) full-length Pin1 cDNAs and two full-length Kunitz cDNAs, one of them previously characterized (PKI1) and a novel one (PKI2) were amplified from potato sprouts (Turrà et al., 2009). BlastX analysis of the latter gene indicated about 95% identity with the S. tuberosum P1H5 gene (AAM10743) and 94% identity with PKI1. All amplified PIs were predicted to be secreted to the apoplast according to the SignalP and WoLF PSORT prediction tools.
Recombinant PVX plasmids containing the full-length cDNA of all of the above mentioned PI genes were used for transient transformation of N. benthamiana plants. For local expression of PI genes, the method of transient Agrobacterium-mediated expression by leaf infiltration was chosen. Transgene expression in agroinfiltrated leaf patches was confirmed 5 days after agroinfiltration by extracting total RNA and performing RT-PCR using PI specific oligonucleotides (Supplementary Table S1) followed by DNA sequence analysis (Figures 1A,B). For systemic transgene expression, a toothpick-inoculation system was used. In this case transgene expression in the upper non-inoculated parts of the plant was verified by monitoring GFP fluorescence 5, 7, 11, 14, and 21 DPI with the PVX::gfp viral construct. All inoculated plants displayed systemic GFP expression starting from 11 DPI (Figure 1C).

Transient Expression of Kunitz and Pin1
Inhibitor Genes Increases Plant Resistance Toward the Fungal Pathogens Botrytis cinerea and Alternaria alternata To evaluate the effect of Kunitz and Pin1 expression on plant resistance toward fungal pathogens, plants systemically transformed with PVX::gfp and PVX::PIs were challengeinoculated with suspensions of B. cinerea and A. alternata spores. Necrotic symptoms 6 days postinoculation are shown in Figures 2A, 3A. The size of the disease lesions was also measured after 2, 4, and 6 days in the case of B. cinerea and after 4 and 6 days in the case of A. alternata infection (Figures 2B, 3B).
While B. cinera symptoms developed markedly on PVX::PPI3A2 transformed plants, those expressing the PKI1 or PKI2, PPI3B2, and PPI2C4 genes exhibited partial or almost complete disease resistance, respectively (Figure 2A). Noteworthy, the reduction in the severity of necrotic lesions ranged from 87% (PVX::PKI2) to 100% (PVX::PPI3B2) 6 days after spore inoculation when compared to the control (PVX::gfp) treated plants (Figure 2B).  Different letters indicate significant differences (P < 0.01) among treatments at each DPI according to One-way ANOVA with post hoc Tukey HSD Test. (C) In vitro inhibitory activity of total soluble proteins (TSP; 50 µg/ml) extracted from PVX-transformed N. benthamiana leaf areas toward B. cinerea. Fungal growth inhibition was measured as the reduction of the optical density at 550 nm relative to the control treatment (TSP extracted from PVX::gfp transformed plants). Values are the means of triplicate determinations; error bars represent the standard deviation. Different letters indicate significant differences (P < 0.05) among the different treatments according to One-way ANOVA with post-hoc Tukey HSD Test.
To understand if the inhibition of B. cinerea and A. alternata growth on PI-expressing N. benthamiana plants depended on the chemical composition of leaf TSP, the in vitro inhibitory activity of TSP extracted from PIs-transformed N. benthamiana leaf areas was compared to that of PVX::gfp transformed ones. Data reported in Figures 2C, 3C show that all TSP extracted from PItransformed plants were able to reduce the growth of both fungal pathogens, except those from PPI3A2-expressing plants that selectively inhibited A. alternata but not B. cinerea proliferation.
Transient Expression of PKI2, PPI3A2, and PPI2C4 PI Genes Increases Plant Resistance Toward the Bacterial Pathogen Pseudomonas syringae pv. tabaci To determine whether Kunitz and Pin1 gene overexpression confers protection against bacterial phytopathogens in N. benthamiana plants, a cell suspension of P. syringae pv. tabaci was syringe-infiltrated into the abaxial side of PVX::PIs or PVX::gfp transformed leaf areas. Chlorotic and necrotic symptoms markedly developed on PVX::gfp and PVX::PKI1 transformed plants 7 days post-infection ( Figure 4A). Interestingly, while a mild reduction of symptoms was observed on PVX::PPI3A2-and PVX::PPI3B2-infected plants, in those expressing the PKI2 or the PPI2C4 gene only a small necrosis surrounding the point of inoculation, comparable to that observed in the water inoculated leaves (data not shown), was visible ( Figure 4A).
As necrotic lesions showed an irregular outline and it was difficult to exactly determine the area of necrosis, the number of surviving P. syringae cells in infected leaves was quantified (Figures 4B,C). Leaf discs were collected from distinct infiltrated areas from each treated plant, homogenized and plated on LB-rifampicin agar at three different dilutions (10 −3 , 10 −4 , and 10 −5 ). Alternatively, a 10 −3 dilution was mixed with LB-rifampicin in 96 well-plates and incubated with constant shaking at 28 • C for 16 h. Colony counting and optical density readings retrieved similar results, reported in Figures 4B,C. No inhibition of bacterial growth was observed 6 days after inoculation in PVX::PKI1-transformed plants when compared to control ones (PVX::gfp). Strikingly, a strong reduction of P. syringae population was detected in plants transformed with the PVX::PPI3B2 and PVX::PPI3A2 vectors (∼40-60%), and complete resistance was observed in those transformed with the PKI2 and PPI2C4 genes. As expected, mock infiltrated leaves showed no necrotic symptoms and no bacterial colonies grew after plating their homogenates.   (Figures 5A-C). Because these macroscopic phenotypes are indicative of a role of PKI2 and PPI2C4 serine PIs in the regulation of cell division and plant development, we tested whether locally transformed N. benthamiana leaves showed higher rates of cell expansion. Interestingly, leaf discs from PVX::PKI2 and PVX::PPI2C4 transformed plants expanded twice as much as the control ones (PVX::gfp) in a 24 h incubation period ( Figure 5D). In an additional set of experiments, leaf strip curvature or epinasty, a phenotype often related to cell expansion (Keller and Van Volkenburgh, 1997), was analyzed by measuring the curvature of leaf strips excised from PVX-transformed patches 24 h after incubation. Notably, PPI2C4-and PKI2-expressing leaf strips bent 3-4-fold more than GFP-expressing ones (Figure 5E), indicating leaf asymmetrical expansion, with the adaxial surface growing faster than the abaxial one.

days after syringe-infiltration
To confirm the relevance of PKI2 and PPI2C4 proteins in the regulation of plant development, N. benthamiana seedlings were systemically transformed with the PVX::PKI2 and PVX::PPI2C4 viral constructs. Morphological analyses showed that PKI2-and PPI2C4-overexpressing plants grew faster, developed bigger root systems and leaves and exhibited an increase in root and shoot dry weight of more than three and two times, respectively, when compared to controls (Figures 5F-J).

Effects of PI Expression on Nicotiana benthamiana Endogenous Protease Activity, Epidermal Cell Expansion, and Division
To ask whether the hypertrophic phenotype observed in PKI2 and PPI2C4 expressing leaves reflects induction of cell division, we performed a detailed microscopic analysis of the adaxial epidermis of PVX::PPI2C4 and PVX::gfp agroinfiltrated patches 7, 13, and 21 DPI. While no alterations of cell morphology could be observed, an increase in the average cell size before (7 DPI) and in nuclear density after (13 DPI) was detected in PPI2C4 expressing plants (Figures 6A-C,F). Accordingly, PPI2C4 expression also lead to the appearance of tight clusters of small-sized cells, some of which being trichomes (Figures 6D,E), 13 DPI. Similar results were obtained in PVX::PKI2 expressing leaves (data not shown). As expected, none of these changes increase of leaf discs. Leaf strips and discs (1 × 1 cm) from transiently transformed leaf areas (7 days after A. tumefaciens syringe-infiltration) were incubated for 24 h as indicated in section "Materials and Methods". Images were taken before and after incubation and the strip curvature (E) and the increase of leaf disc area (D) (expressed as a percentage of the initial leaf disc area) were calculated. Values are the means of 15 different leaf strips or discs measurements repeated in three independent experiments; error bars represent the standard deviation. Different letters indicate significant differences (P < 0.01) according to One-way ANOVA with post hoc Tukey HSD Test. (F-J) Effects of systemic PKI2 and PPI2C4 over-expression on plant growth. Third and fourth leaves of 2-3-week-old N. benthamiana seedlings were tooth-pick incoculated with the indicated viral constructs. PVX or PVX::gfp transformed plants were used as controls. Representative pictures of entire plants (F), root systems (G) and newly formed leaves (H), and measurement of root (I) and shoot (J) dry weight 21 days after A. tumefaciens inoculation. Values are the means of 10 different measurements repeated in three independent experiments; error bars represent the standard deviation. Different letters indicate significant differences (P < 0.01) according to One-way ANOVA with post hoc Tukey HSD Test.
was observed in control PVX (data not shown) or PVX::gfp transformed leaves.
To understand if misregulation of cell division in these plants was associated with an alteration of the endogenous protease activity, TSP from mock-(H 2 O) and PVX-infiltrated N. benthamiana leaf areas were either directly used in a in-gel protease assay (Figure 7A) or the first extracts mixed with the latter's to measure the residual protease activity ( Figure 7B). Similar results were obtained in these two assays. In the first case, an additional protease band was clearly visible in the GFP-expressing sample and in the mockinoculated control, but not in the TSP from PVX::PKI2 and PVX::PPI2C4 transformed plants ( Figure 7A). In the second case, TSP from PKI2-and PPI2C4-expressing samples showed 100 and 60% higher inhibitory activities, respectively, than those from control plants (GFP expressing). Interestingly, TSP from PKI2-and PPI2C4-expressing samples also showed a high degree of inhibition (90%) toward N. benthamiana apoplastic protease activity, when compared to the GFPexpressing controls (Figure 7C). (B,C) TSP (B) or intercellular fluids (C) isolated from untransformed N. benthamiana plant leaves were incubated with TSP extracted from PVX-transformed N. benthamiana leaf areas (7 days after syringe-infiltration) and the remaining protease activity was analyzed by measuring the change in optical density at 450 nm after 1 h of incubation in the presence of a 1% (w/v) azocasein solution. Values are the means of triplicate determinations; error bars represent the standard deviation. Different letters indicate significant differences (P < 0.05) according to One-way ANOVA with post hoc Tukey HSD Test.

DISCUSSION
Hydrolysis and protein synthesis, as well as the regulation of these physiological processes, are fundamental phenomena impacting both plant development and susceptibility/resistance to pathogens (Rawlings et al., 2004(Rawlings et al., , 2014Turrà and Lorito, 2011;Grosse-Holz and van der Hoorn, 2016). Indeed, several members of the serine protease group, a widely distributed set of extracellular and intracellular proteolytic enzymes, act as pathogenicity factors in different plant pathogens including fungi, oomycetes, bacteria, insects, and nematodes (Ryan, 1990;Urwin et al., 1995;Atkinson et al., 2003;Gvozdeva et al., 2004;Lopez-Solanilla et al., 2004;Hermosa et al., 2006;Pekkarinen et al., 2007;Luo et al., 2009;Thomas and van der Hoorn, 2018). Besides, serine proteases also regulate a panoply of endogenous processes in plants including innate immunity, cell death and nitrogen uptake (Kohli et al., 2012;Grosse-Holz and van der Hoorn, 2016;Salvesen et al., 2016;Balakireva and Zamyatnin, 2018). Interestingly, inhibitors of these enzymes, especially those belonging to the multigene PI families PKPI and Pin1 have been shown to accumulate in plant tissues in a highly precise spatial and temporal manner and following both abiotic and biotic threats (Hermosa et al., 2006;Wang et al., 2008;Singh et al., 2009;Turrà et al., 2009;Boex-Fontvieille et al., 2015;Rustgi et al., 2017). These findings together with the increasing evidence of their inhibitory activity both in vitro and in vivo toward insects and nematodes, and in vitro toward fungi and bacteria is indicative of a possible multitasking activity of these PI families in both the regulation of plant physiology and biochemical defense responses (Duan et al., 1996;Cai et al., 2003;Vila et al., 2005;Turrà and Lorito, 2011;Quilis et al., 2013). However, a clear correlation between PKPI and Pin1 expression and in planta modulation of developmental processes or resistance toward fungal or bacterial pathogens is currently missing. We have previously shown that extracellular proteases secreted by the fungal pathogen B. cinerea mainly belong to the serine protease class (Hermosa et al., 2006). We further identified, from the complex set of plant-produced PIs, different protein products belonging to the PKPI and Pin1 serine proteinase inhibitor families and showing high inhibitory activity on both fungal growth and disease development when exogenously supplemented to the fungal inoculum source (Hermosa et al., 2006). Now, to demonstrate their efficacy in vivo toward fungal plant pathogens, we have heterologously expressed different potato PKPI and Pin1 genes in N. benthamiana, a model system to study the effect of transgene expression on plant-pathogen interactions (Goodin et al., 2008). Consistent with our previous findings, all tested PIs genes (except the Pin1 PPI3A2) efficiently reduced the in vitro growth and increased plant resistance toward two fungal pathogens, B. cinerea and A. alternata. Noteworthy, no B. cinerea symptoms developed at all, even 6 days postinoculation, on N. benthamiana plants transformed with PPI3B2, the potato Pin1 inhibitor formerly identified for its strong antifungal activity on B. cinerea in vitro (Hermosa et al., 2006). To understand if in planta over-expression of PKPI and Pin1 genes could also alter plant resistance toward bacterial phytopathogens, we challenged N. benthamiana transformed plants with the bacterial pathogen P. syringae pv. tabaci. Interestingly, bacteria survival in PKI2-, PPI3A2-, PPI3B2-, and PPI2C4-expressing plants dropped over time and necrotic symptoms barely developed on PKI2 and PPI2C4 transformed leaves, suggesting that PKPI and Pin1 PIs might act by exerting either a direct and/or indirect antiproliferative activity on Pseudomonas. This hypothesis is supported by two lines of evidence. First, several studies have already shown the importance of Pseudomonas spp. cysteine and serine proteases in the degradation of both structural and soluble host plant proteins and virulence (Engel et al., 1998;Axtell et al., 2003;Hotson and Mudgett, 2004). Second, uninfected PKI2-, and PPI2C4-transformed areas from fully developed leaves enlarged abruptly and accumulated higher amounts of trichomes, epidermal cells specialized in defending the plant from both biotic and abiotic stresses (Wagner, 1991;Amme et al., 2005). Importantly, this phenotype was accompanied by an increase in epidermial cell growth and division in PVX:PPI2C4 agroinfiltrated leaves and an overall shoot and root size in systemically transformed plants, indicative for an endogenous inhibitory activity of these PIs, as also shown by TSP activity on apoplastic proteases.
Overall our findings show for the first time that specific members of the Pin1 and PKPI PI families might act as multifunctional proteins playing fundamental roles in both the regulation of important plant physiological processes such as cell development and differentiation as well as wide-spectrum disease resistance against fungal and bacterial pathogens. These results might represent a framework for the future selection of Pin1 and PKPI genes to be used either individually or in gene pyramiding approaches to obtain fast-growing trees or crops with broad-resistance to biotic threats.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
DT, SV, SW, and ML designed and conceived the study, and wrote the manuscript. DT, SV, and RM performed the experiments and analyzed the data. All authors read and approved the final version of the manuscript for publication.