INTRODUCTION

Infection of plants with pathogens results in the induction of numerous host-specific biochemical responses, some of which are critical for the ability of the plant to withstand diseases (1). Physiological and pathological studies have examined diseased plants in the hope of uncovering the cause(s) of the pathogen-induced distress, and there is a plethora of studies describing the dramatic effect of pathogen infection upon different aspects of plant metabolism and catabolic disturbances (2-4). What is lacking, however, is a clear understanding of how different pathogens promote the often deleterious symptoms observed by switching on a common cascade of cellular events resulting in the disease syndrome and the accompanying resistant character to subsequent pathogenic attacks.

Viroids are the smallest known plant infectious agents, made up of nude, circular, single-stranded RNA molecules of a few hundred nucleotides which do not code for any protein (5, 6), and thus they are an adequate model to analyze the physiological and molecular basis of plant responses to pathogen infection. This is more relevant if we consider that the viroid elicited responses resemble those resulting from infection by more complex type of pathogens or different kind of stresses (7, 8).

It has been shown previously (8) that plants infected with viroids produce de novo synthesis of a set of host-encoded proteins termed pathogenesis-related (PR)¹ proteins. The function of some PR proteins as hydrolytic enzymes (e.g. chitinases or -1,3-glucanases) has been demonstrated in many plant species irrespective of the nature of the attacking pathogen, and they appear to play a role in the induced defense response of the plant to combat pathogens (9, 10).

Protein degradation as well as protein processing and maturation are believed to be important events in the plant defense response (11-13). In this regard, one of the viroid-induced PR proteins has been identified as a protease that was termed PR-P69 (14, 15). Recent molecular cloning (11) indicated that P69 is a protein structurally related to the yeast processing protease Kex2, the prototypic member of the eukaryotic subtilisin-like protease family. This new finding broadens our understanding of defense responses activated by microbes, and opens a new perspective to unravel how plants perceive pathogenic insults and activate signaling processes that results in a resistant character to subsequent pathogenic challenges.

In this paper we have have explored the existence of additional proteolytic activities activated in tomato plants following viroid infection. We performed comparative chromatographic fractionation of cell homogenates and determined differences in proteolytic profiles between healthy and infected plants. Using specific antibodies raised against the previously identified P69 protease, we identified a distinct, but immunologically related P69-like proteinase, that accumulates also in infected tissues. These studies were followed by the cloning of an inducible gene encoding a new subtilisin-like proteinase member related to the previously identified P69 protease (11). We also discuss possible functions for this inducible protease family during pathogenesis in plants.

EXPERIMENTAL PROCEDURES

Conditions for growth of tomato plants (Lycopersicon esculentum), and method for inoculation with citrus exocortis viroid have been described (8). Tissues were harvested 4 weeks after inoculation and stored at 80 °C.

Restriction enzymes and modification enzymes were obtained from Boehringer. DEAE Sepharose CL-6B, T7-sequencing, and Ready-to-go DNA labeling kits were obtained from Pharmacia. Radioactive compounds were from Amersham Corp. All other commonly available reagents were of analytical grade.

All operations were carried out on ice or at 4 °C. Leaf tissue (10 g) showing symptoms of viroid infection was frozen in liquid nitrogen, thawed, and homogenized in a mortar and pestle with 30 ml of buffer A (50 mM Tris-HCl, pH 7.21, 2 mM dithiothreitol, 1 mM CaCl₂, 1 mM MgCl₂). The homogenate was filtered through cheesecloth and centrifuged at 15,000 × g for 20 min. Supernatants were immediately used or stored at 70 °C.

The supernatant was adjusted to 20% saturation with solid (NH₄)²SO₄ and centrifuged at 15,000 × g for 10 min. The supernatant was then adjusted to 70% saturation with (NH₄)²SO₄ and centrifuged as described above. The resulting pellet was dissolved in 2.5 ml of buffer B (50 mM Tris-HCl, pH 7.5, 0.1 mM dithiothreitol) and desalted in Sephadex G-25 (PD10 columns, Pharmacia) equilibrated in buffer A. The eluted proteins were then applied to a DEAE-Sepharose CL-6B column (1.5 × 10 cm) equilibrated in buffer A. After washing with equilibration buffer, the column was eluted with 250 ml of a linear gradient of 0-0.3 M NaCl in the same buffer. Fractions were assayed for proteolytic activity using fluorescein isothiocyanate-casein as substrates as described previously (14) and the protein content analyzed by SDS-PAGE.

SDS-PAGE analyses were carried out in 14% polyacrylamide gels as described previously (14). Gels were stained with Coomassie Brilliant Blue. M_r markers used were bovine serum albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), soybean trypsin inhibitor (20 kDa), and lactalbumin (14 kDa). Blotting of proteins onto nitrocellulose membranes from SDS-PAGE gels was performed as described previously (15). The blots were then processed for immunological staining as described previously using anti-P69 antiserum (11) and the antigen-antibody sandwich revealed with peroxidase-conjugated goat anti-rabbit IgG and 4-chloro-1-naphthol/H₂O₂ staining.

A -ZAP cDNA library was constructed from poly(A) mRNA isolated from viroid-infected tomato leaves (16). The library was screened at 52 °C as described by Church and Gilbert (17) with a radiolabeled p26 cDNA encoding the P69A protein (11), and plaques were isolated by standard techniques (16). cDNA inserts were excised from phage DNA, cloned into pBluescript (Stratagene, San Diego, CA), and sequenced by using a T7 polymerase kit (Pharmacia). DNA sequence analysis was performed on both strands by the dideoxy chain termination method (18). Nested deletions of the cDNA cloned in pBluescript SK+ were generated using an exonuclease III-S1 nuclease kit (Promega). Sequence searches and analyses were done using FASTA, MAP, and BESTFIT routines of the University of Wisconsin Genetic Computer Group package (19).

RNA was purified from different tomato plant tissues as described (20). For RNA gel blot analysis, 15 µg of total RNA were electrophoresed on 1% agarose gels containing formaldehyde and blotted onto Nytran membranes (Schleicher & Schuell). Equal loading of RNA was verified by ethidium bromide staining of the gel before transfer to the membrane. Alternatively, the RNA blots were hybridized with a radiolabeled cDNA probe for the constitutively expressed pentaubiquitin (5xUBI) gene to verify equal loading. DNA was isolated from leaves as described (21). For DNA gel blot analysis, 10 µg of DNA was digested with restriction enzymes and electrophoresed in a 0.7% agarose gel and blotted onto Nytran membranes. RNA and DNA gel blots were probed with the entire cDNA insert or with the first 300 base pairs of cDNA (internal PstI fragment), which was radiolabeled by random priming using T7 polymerase (Pharmacia). Hybridization and washing conditions of filters were done at high stringency (70 °C) as described (17).

cDNA synthesis, quantification of the products, and reverse transcriptase-mediated PCR were conducted as described (22). The oligonucleotide primers (50 pmol each), op9-PCR1 (5-CTCGGCCATGTAGCCAAT-3) and op9-PCR2 (5-TCTTCAAACAACTGTG-3) were used to PCR amplify an internal region of the in vitro synthesized single-stranded cDNA from the different mRNA sources in a Perkin-Elmer DNA Cycler. PCR amplification was programmed for 30 cycles, with each cycle consisting of 94 °C for 1 min, 50 °C for 2 min, and 70 °C for 1.5 min. The amplified DNA fragments were visualized in agarose gels or alternatively hybridized with a radiolabeled p9 cDNA probe. The inability of these primer pairs to amplify the closely related P69A protease sequence was confirmed in control PCR reactions that included 10 ng of plasmid DNA containing the P69A cDNA as template. Positive controls were carried out in parallel reactions containing 10 ng of plasmid DNA containing the P69B cDNA.

RESULTS

Aiming to resolve proteolytic activities induced in tomato plants upon infection with citrus exocortis viroid, leaf tissue homogenates were fractionated by ion-exchange chromatography on DEAE-Sepharose CL-6B columns. Fig. 1 (left panel) shows chromatographic elution profiles derived from tissue homogenates from healthy and citrus exocortis viroid-infected plants after elution with a linear gradient of 0-0.3 M NaCl. Two remarkable differences in proteolytic activities were observed between both of them when fluorescein isothiocyanate-casein was used as substrate for the determination of proteolytic activities. In profiles from viroid-infected plants (Fig. 1B, left), the appearance of two different peaks of proteolytic activity were observed when compared with the proteolytic profile from healthy plants (Fig. 1A, left): (a) peak I (eluting with the washing buffer) and (b) peak II (eluting at 0.08 M NaCl) (Fig. 1B, left panel). Analysis by SDS-PAGE of proteins present in the corresponding column fractions revealed differences in the protein profiles. In the case of viroid-infected plants (Fig. 1D, right panel) the appearance of a major set of proteins corresponding to the previously identified inducible PR proteins (15) was observed. These proteins eluted in the void volume fractions, matched peak I of proteolytic activity, and contained the characteristic 69-kDa protein band previously identified as a source of proteolytic activity (11, 14, 15). SDS-PAGE of proteins in column fractions corresponding to peak II from infected plants did not revealed major differences when compared with healthy controls, except in an additional 69-kDa protein band which was not detected in the equivalent column fractions derived from healthy plants. The inducible nature and close similarity in molecular mass of the detected 69-kDa proteins in peak I and II prompted us to search for immunological relationship between each other. Pooled proteins recovered in peak I and II, and the corresponding pooled proteins from healthy controls, were subjected to Western blot analysis using an anti-P69 antisera (Fig. 2). This revealed that the antibodies specifically immunodecorated the 69-kDa protein bands present in both peak I and II fractions from infected plants, while no inmunodecoration was observed in the corresponding fractions from healthy plants. Since the chromatographic separation was based on the net surface charge of the proteins, the partitioning of these two distinct immnunoreactive 69-kDa polypeptides favored the idea that the previously identified P69 proteinase present in crude homogenates from viroid-infected plants (14) was in fact a mixture of two different isoenzymes with similar biochemical and immunological properties but with different chromatographic behavior.

Based on the criteria of close immunological relationship between these two viroid-inducible 69-kDa proteinases, we attempted cloning of the new P69 member by screening with antibodies and also by screening at low stringency hybridization with the p26 cDNA clone previously described (11) as a probe, in cDNA libraries from viroid-infected tomato. Immunoscreening of a -ZAP cDNA library from viroid-infected tomato plants (16) did not render any good result (not shown). Conversely, hybridization-based screening of the same cDNA library with a radiolabeled p26 cDNA probe rendered a total of nine cross-hybridizing cDNA clones. Restriction endonuclease mapping and partial sequence analysis of all these cDNA clones indicated that 6 of them represented the previously identified P69 gene, while the remaining 3 cDNA clones represented a different gene. One of these latter cDNAs (p9 cDNA clone) was selected and analyzed in detail. The nucleotide sequence and the derived amino acid sequence predicted from the only open reading frame present in p9 are shown in Fig. 3. The cDNA insert contains 2407 nucleotides (excluding the poly(A) tail), and RNA gel blot using the p9 cDNA as a probe gave a band of 2.5 kilobases in size (see below), indicating that the p9 cDNA clone is an almost full-length clone. The open reading frame starts with an ATG codon at position 11-13 that is surrounded by nucleotide sequences that matches the consensus sequence conserved for translation initiation in eukaryotes (23), and ends at the TAG stop codon at position 2246-2248 of the cDNA. The open reading frame is preceded by an in-frame stop codon reinforcing the assumption that the most likely 5-Met initiation codon is that located at position 11-13. The encoded protein has a predicted molecular weight of 78,992 and a total of 745 amino acids, which is larger that the 69-kDa proteins recognized by antibodies (see Fig. 2). This was further verified by in vitro transcription-translation experiments of the p9 cDNA in rabbit reticulocytes which render the synthesis of an about 80-kDa protein (data not shown). However, the hydropathy profile of the derived sequence (data not shown) and computer based comparison of amino acid sequence of the NH₂ terminus indicated the existence of a preprosequence. This consists of a hydrophobic signal peptide at the extreme NH₂ terminus. The "(3,1) rule," as proposed by von Heijne (24), predicts cleavage of this signal peptide after Ser-22. This is followed by a 92-amino acid prosequence which is a typical feature of proteases of the subtilisin family and for which, the proteolytic removal of the prosequence is an important step in the generation of the active protease from the inactive zymogen (25). The putative NH₂-terminal amino acid of the mature protein is Thr, identified by comparison with other plant subtilisin-like proteases (11, 26, 27), and is indicated in Fig. 3. Thus, the predicted mature enzyme contains 631 amino acids with a predicted molecular weight of 66,138. The amino acid composition predicted for this mature protein is consistent with that determined previously for the P69 proteinase (86% identical), and this structural similarity explains the observed cross-reacting of each other when using anti-P69 antibodies upon ion-exchange chromatography fractionation. For the sake of uniformity we will refer to this newly identified P69 proteinase as P69B, while the previously identified one will be designed as P69A. The amino acid sequence comparison of P69B with that of P69A is illustrated in Fig. 4A.

Within the mature P69B protein the amino acid residues Asp-146, His-203, and Ser-531, common to all subtilisin-like proteases, were identified. The sequences surrounding the catalytic sites (catalytic triad) are similar to the catalytic triad essential for all of the subtilisin-like members to function as proteases, and highest when compared with P69A (Fig. 4, A and B). Also the protein has an Asn residue (Asn-305) that has been found to be highly conserved in this position and is catalytically important in the subtilisins (28, 29). Despite the similarities around the amino acids of the active center, there is an insertion of a long sequence (226 amino acids) between the stabilizing Asn-305 and the reactive Ser-531 relative to all other subtilisin-like proteases, which are separated by much shorter distances. This displacement has also been observed in the three other subtilisin-like proteinases recently identified from plants (11, 26, 27). The meaning of such a displacement remains unknown, but at face value it could represent a characteristic signature of the subtilisin enzymes from plants.

The differential expression pattern of P69B was determined in different tissues from healthy and viroid-infected plants by Northern blot hybridization. The filters were hybridized at high stringency with the radiolabeled p9 cDNA insert. The results revealed that the level of P69B mRNA, which has a size of 2.5 kilobases, is increased markedly in leaf, stem, and less in root tissue from infected plants, while the corresponding mRNA preparations derived from healthy plants (mock-inoculated plants) showed a very weak cross-hybridization with the radiolabeled probe and only after prolonged exposure of the filter with the film (Fig. 5).

The possibility that the observed increases in P69B gene expression in infected tissues could be masked by cross-hybridization with the mRNA for the previously identified P69A gene was also entertained. Gene-specific RT-PCR reactions provided an alternative approach to verify that the expression of the newly identify P69B gene was indeed induced in infected plants (Fig. 6). In vitro synthesized single-stranded cDNAs from mRNA samples of leaves, stems, and roots from either healthy or viroid-infected plants were assayed by PCR in a 2-fold dilution series. Amplification with primers designed to render a 1202-base pair PCR product (oligonucleotide op9-PCR1 and op9-PCR2), and specific for internal sequences of P69B but not for the P69A gene, were used in these reactions to specifically detect P69B gene expression. Fig. 6 shows an example of the RT-PCR data. RNA preparations from leaf, stem, and root tissues of viroid-infected tomato plants, but not the equivalent RNA preparations of healthy plants, produced the expected 1200-base pair size product. The derived PCR products were also confirmed by Southern blot analysis with a radiolabeled p9 cDNA probe (data not shown). As a control, plasmids containing the P69B cDNA or the P69A cDNA were amplified under similar conditions and only the plasmid containing the P69B cDNA render the expected size PCR product (Fig. 6). This further sustains that the RT-PCR products observed using these primers were P69B-specific and also support that P69B is induced during pathogenesis. The absence of amplified PCR products in samples from healthy plants suggests that the weak cross-hybridizing mRNA species observed in Northern blots from healthy plants may represent expression of different subtilisin-like members with other housekeeping functions.

To further characterize the existence of other P69-related genes in tomato plants, more extensive analysis was completed using genomic DNA and cDNA probes. Genomic DNA gel blot analysis, shown in Fig. 7, was performed to determine the extent of complexity of the P69-encoding genes. A blot containing tomato genomic DNA digested with several restriction enzymes was probed with the entire p9 cDNA and demonstrated the presence of a very large number of cross-hybridizing fragments and of different hybridization intensities (Fig. 7, left), suggesting that P69 may conform a large multigene family of high complexity. This high complexity was also observed, but at a reduced scale, when a similar digested DNA was hybridized with a cDNA probe encompassing only the preprodomain sequence, excluding the catalytic domain (Fig. 7, right). Since none of the restriction enzymes used in this experiment cut within the preproregion, this result enables us to likely estimate that the P69 gene family is composed of at least seven different, but closely related members, of which only two of these genes (P69A and P69B) have been identified so far.

DISCUSSION

In this work, we provide structural and functional information on P69B, a second member of a family of plant proteases induced during the response of tomato plants to pathogen attack. P69B represents a new plant subtilisin-like protease (EC 3.4.21.14) based on amino acid sequence conservation and structural organization (28, 29) which is highly related (86% identical amino acids residues) to the previously identified PR-P69 protease (11) (here renamed as P69A). The predicted primary structure of P69B indicates that the protein is synthesized as a precursor (preproenzyme) composed of three distinct domains: a 22-amino acid signal peptide, a 92-amino acid propolypeptide, and a 631-amino acid mature polypeptide that is the active form of the enzyme that accumulates in vivo. Within the mature polypeptide, the amino acid sequences surrounding Asp-146, His-203, and Ser-531 is the most salient feature of P69B and are closely related to those of the catalytic sites (catalytic triad) of the subtilisin-like serine proteases. Thus, from analysis of the different domains along with the conserved amino acid sequences surrounding the catalytic site, we have ascribed P69B as a new member of the subtilisin-like family of serine protease.

In eukaryotes, this class of serine proteases is also referred to as subtilisin-related proprotein convertases. These are involved in the proteolytic processing of peptide hormones and other precursor proteins, cleaving at sites comprised of pairs of basic amino acid residues (dibasic sites) (28). Since the identification of kexin (Kex2) as the subtilisin-like serine protease responsible for processing pro--mating factor in the yeast Saccharomyces cerevisiae (30), seven mammalian homologues which participate in the post-translational processing of protein and protein hormones were identified and named furin, PC1 (also called PC3), PC2, PC4, PACE4, PC5 (also called PC6), and PC7 (see Seidah et al. (31) for one of the last updates). More recently, genes that encode Kex2/subtilisin-like endoproteases have also been isolated from Mollusca, insects, and nematodes. Two genes isolated from Drosophila melanogaster called Dfur-1 and Dfur-2 (32-34) encode propeptide convertases with sequence similarity to human furin. Likewise, genes isolated from Caenorhabditis elegans include celpc2, which shows sequence similarity to PC2 (35). In addition, molecular characterization of the bli-4 gene from C. elegans, which was shown by genetic analysis of lethal mutants to be essential for the normal development of this organism, has revealed that bli-4 also encodes gene products related to the Kex2/subtilisin-like protease (36). Moreover, sequence analysis of bli-4 predicts four protein products, which have been designated blisterases A, B, C, and D. These proteins share a common amino terminus, but differ at the carboxyl termini, and are most likely produced from alternatively spliced transcripts.

In plants, only indirect evidence was available to substantiate the existence of such proteases (37) until a subtilisin-like protease was sequenced from melon fruits (26). Subsequently, cDNA clones encoding subtilisin-like proteases have been identified in other plant species including the previously identified P69A from tomato plants (11). Also a cDNA encoding a subtilisin-like protease (ag12) from Alnus glutinosa was shown to be expressed during early stages of actinorhizal nodule development (27). Likewise, a gene from Arabidopsis thaliana encoding a closely related subtilisin-like protease (ara12) has been identified and showed constitutive expression (27). Collectively, all these recent data provides further evidence that the subtilisin gene family has been conserved throughout evolution of multicellular organisms including higher plants.

At variance with the developmentally regulated pattern of expression observed in plants for the different members of this family of proteases, the tomato plant homologues under consideration are expressed during pathogenesis. However, Southern analysis reveals that many more members are present in the genome of tomato plants, and suggest that very likely additional members, still to be identified, might be expressed at certain stages of development. From an speculative point of view, the degree of homology between P69A and P69B, along with the conservation of related genomic sequences, likely indicates that these genes arose through gene duplication events, thus suggesting that subtilisin-like enzymes could display redundant biochemical functions, some of which are apparently important during pathogenesis. Biochemical redundancy is an argument that poses a problem when one tries to define the precise biological significance of individual subtilisin-like proteases. In animal system, natural substrates for subtilisin-like proteases have been identified (28, 29), but it has been observed that when the different subtilisin-related proprotein convertase members are removed from their biological context and assayed in vitro, many of these proteases are able to process the same substrates. This fact raises the question of whether such a functional redundancy exists among the family members in vivo and how this might be regulated. One insight into defining functionality has been provided by examining the expression and localization of the individual members, which appears to be a function of the structural differences between each family member (38). From such analysis, it has been shown that substrate specificity in vivo can be influenced by both restricting expression to particular tissues and also compartmentalization of the individual enzymes to specific intracellular locations (39), and is a likely explanation for delimiting redundant functions which may serve backup roles to a principal protease. Thus, it appears likely that similar mechanism(s) operates in plants for controlling the specific action of the different subtilisin-like members either under normal or under pathogenic or other stress-related situations.

Besides the conservation of this family of proteases along evolution, all plant subtilisin-like enzymes described so far have in common some signatures that differentiate them from the rest of other eukaryotic enzymes. In particular, P69B, as occurs with the rest of plant subtilases, shows the insertion of a long sequence (226 amino acids) between the conserved Asn residue and the reactive Ser residues of the catalytic triad, relative to all other subtilisin-like proteases. The meaning of such a conserved displacement remains unknown, but its conservation may suggests it subserves important functions in regulating the properties of this subgroup of subtilases. Furthermore, all subtilisin-like enzymes are initially synthesized as inactive proenzymes containing a proregion which is believed to function as an intramolecular chaperone, guiding the correct folding of the protease domain and preventing it from being active until the proregion is removed in the appropriate compartment (25). However, in all other eukaryotic subtilases the proregion and the catalytic domains are separated by the RXXR motif (positions 1 to 4 from the cleavage site) and have an additional conserved basic residue at the 6 position (25). In the case of plant subtilases, only the basic residue at 6 is conserved, whereas the RXXR motif is not. This may suggest that the way plant subtilases mature and become active may also have diverged with respect to the others, presumably as a prerequisite to reach their final cellular destination through the secretory pathway. In this regard, all other eukaryotic subtilases function intracellularly within specific regions of the secretory pathway, whereas the plant enzymes apparently function after secretion to the outside of the cell. This final consideration also holds true for subtilisins from bacteria. Moreover, previous immunocytochemical localization studies of the P69 protease (40) showed that the enzyme(s) locates in the apoplast, and reconciles with the biochemical evidence that most, if not all, of the soluble enzyme can be recovered in an active form within the intercellular washing fluid fraction (40). Thus, the site of activity of plant subtilases could be the extracellular matrix. This conclusion gains more acceptance if we consider that the extracellular matrix (including the cell wall) has long been thought to be part of the lytic compartment of the plant cell as defined by Matile (41).

The fact that the P69 proteolytic system presently described is induced during pathogenesis in plants is also in marked contrast with the rest of subtilases members. This peculiarity, together with the extracellular localization of the mature enzymes, open new perspectives for the interpretation of possible biological roles of subtilisin-related enzymes, and points toward a role of these proteases to modulate the interaction of the plant cell surface with the extracellular environment.

There is a wealth of examples in animal systems that document pivotal connections between extracellular matrix proteases and important protein substrates during biological processes related to signal transduction, such as morphogenesis, tissue repair, wound healing, or in disease states (42, 43). In most of these systems, alteration in specific proteases significantly alters availability and activity of membrane-anchored growth factors and other important molecules through the effects on their specific processing. In this respect, we have recently identified a developmentally regulated extracellular matrix-associated protein (named LRP) that is post-translationally processed by the P69 proteolytic system in viroid-diseased plants (44). LRP belongs to the conserved leucine-rich repeat family of proteins that mediate molecular recognition and/or interaction processes in the extracellular matrix of eukaryotic cells to initiate different signaling processes (45). Although the physiological ramifications derived from the processing of LRP remains unknown, this observation is indicative that members of the subtilisin family are participating in proteolytic processes occurring in the plant cell surface. Also a membrane-bound 60-kDa Kex2-like protease has been proposed to act in the recognition and processing of systemin, the traveling peptide hormone mediating signaling processes during wound response in plants (37), and thus constitutes another example of involvement of plant subtilisin-like proteases in pericellular processing of important signaling molecules. Thus, finely tuned proteolytic degradation of the extracellular matrix, by the concerted activation of genes encoding distinct proteolytic enzymes, including the subtilisin-like proteases here described, that recognize and degrade pericellular substrates can be postulated as a mechanism by which plant cells can initiate phenotypic changes during different developmental processes and in pathological conditions, and which ultimately can influence a number of different cellular processes important for the survival of the entire organism.

However, we cannot disregard the possibility that these proteases could also participate as active defensive tools directed toward the attacking pathogens or by hydrolyzing proteins secreted by the intruder. In fact, genes encoding a leucine aminopeptidase or an aspartic protease are also activated and secreted during the defense response of tomato plants (12, 13). Thus, it is conceivable that coordinate expression of genes encoding different proteases, like leucine aminopeptidase, aspartic protease, or subtilisins, with different substrate specificities, may reflect common strategies evolved by plants to defend against pathogens, either acting as a first defense barrier or through the post-translational modification of proteins that participate in the activation of defense responses. Furthermore, the previous observation that a concurring extracellular aspartic protease controls the overaccumulation of P69 proteases and related proteins (46, 47), suggests that a tightly controlled mechanism of interplay between different proteases, acting either as proteolytic enzymes or as substrates for the action of other proteases, operates in plants. This consideration adds new elements of discussion that should be entertained in the future for a deeper understanding of how pericellular proteolysis is controlled in plants. Further studies on the role of pathogen-induced extracellular proteases in transgenic plants, as well as identification and cloning of additional family members and determination of interrelationship and substrate specificity, are one of our next challenges for the future.