Polygalacturonase-Inhibiting Protein Interacts with Pectin through a Binding Site Formed by Four Clustered Residues of Arginine and Lysine

Abstract

Polygalacturonase-inhibiting protein (PGIP) is a cell wall protein that inhibits fungal polygalacturonases (PGs) and retards the invasion of plant tissues by phytopathogenic fungi. Here, we report the interaction of two PGIP isoforms from Phaseolus vulgaris (PvPGIP1 and PvPGIP2) with both polygalacturonic acid and cell wall fractions containing uronic acids. We identify in the three-dimensional structure of PvPGIP2 a motif of four clustered arginine and lysine residues (R183, R206, K230, and R252) responsible for this binding. The four residues were mutated and the protein variants were expressed in Pichia pastoris. The ability of both wild-type and mutated proteins to bind pectins was investigated by affinity chromatography. Single mutations impaired the binding and double mutations abolished the interaction, thus indicating that the four clustered residues form the pectin-binding site. Remarkably, the binding of PGIP to pectin is displaced in vitro by PGs, suggesting that PGIP interacts with pectin and PGs through overlapping although not identical regions. The specific interaction of PGIP with polygalacturonic acid may be strategic to protect pectins from the degrading activity of fungal PGs.

Many recognition events between plants and pathogens take place in the apoplast. The plant cell wall is a repository of oligosaccharides, which act as signal molecules and have regulatory properties (Bellincampi et al., 1996; Ridley et al., 2001). Local wall modifications are produced upon pathogen attack, including the degradation of wall components, weakening of its structure, deposition of new material, and stiffening (Willats et al., 2001). Changes in the structure and composition of the wall are finely regulated during developmental stages and defense responses (Steele et al., 1997; Ridley et al., 2001).

Since pectin is a major constituent of the wall, enzymatic modification of this component has received considerable attention. Pectins form either ionic or covalent interactions with different classes of proteins located in the cell wall. For example, extensins are associated either ionically to acidic pectins or covalently to a feruloylated sugar of pectin (Qi et al., 1995). Pectin also establishes a strong, probably covalent, interaction with the extracellular domain of wall-associated kinases (He et al., 1999; Anderson et al., 2001; Decreux and Messiaen, 2005). On the other hand, the charged homogalacturonan (HGA) forms ionic interactions with proteins exposing positive charges in a favorable orientation, like in the case of isoperoxidases (Carpin et al., 2001), or binds proteins through a Ca²⁺-dependent cross-link, like in the case of arabinogalactan proteins (Baldwin et al., 1993). The analysis of the interaction between pectin and proteins is sometimes complicated by the structural complexity of pectin and by the different nature of interactions established by a single protein with different pectic domains.

Polygalacturonase-inhibiting protein (PGIP) is an extracellular protein known to interact with fungal endopolygalacturonases (PGs) and is involved in plant defense. The inhibition of PGs is thought to be an important event during plant defense against fungi (Federici et al., 2001; Mattei et al., 2001b; Ferrari et al., 2003; Manfredini et al., 2005). Fungal PGs are the first cell wall-degrading enzymes produced when fungi are grown on cell wall material and are a prerequisite for wall degradation by other cell wall-degrading enzymes. The inhibition of PGs by PGIP is also thought to cause the accumulation in the plant apoplast of oligogalacturonides, which act as elicitors of a wide range of defense responses (De Lorenzo et al., 2001). The occurrence of PGIPs has been reported in the cell walls of a variety of dicotyledonous plants and in several monocotyledonous plants (De Lorenzo et al., 2001; Kemp et al., 2003). The mature PGIP is characterized by the presence of a Leu-rich repeat (LRR) domain that represents more than two-thirds of the protein; this motif forms a solvent exposed parallel β-sheet surface that mediates protein-protein interactions (De Lorenzo et al., 2001; Mattei et al., 2001a; Federici et al., 2006). The amino acids of PGIP that determine recognition capability and affinity for fungal PGs are internal to the β-sheet region (Leckie et al., 1999; Sicilia et al., 2005).

In this work we show that PGIP establishes, through a binding site present on the protein surface, specific interactions with stretches of unesterified HGA. The binding of PGIP to the cell wall is instrumental to ensure the presence of the inhibitor where pathogen infection initiates.

RESULTS

PGIP Interacts with Cell Walls both in Planta and in Vitro

PGIP is tightly bound to the cell wall and is extracted only with high ionic strength buffers. Stem segments of transformed tomato (Lycopersicon esculentum) plants overexpressing the isoform 1 of PGIP from Phaseolus vulgaris (PvPGIP1; Desiderio et al., 1997) were vacuum infiltrated with 50 mm sodium acetate buffer, pH 5.0, in the presence or in the absence of 0.3 m NaCl to obtain intercellular washing fluids (IWFs) according to the procedure described in “Materials and Methods.” Contamination of IWFs by cytoplasmic components was ruled out by measuring Glc-6-P dehydrogenase activity, which accounted for less than 1% of total extractable activity.

Western-blotting analysis showed that PvPGIP1 is present in IWFs obtained with buffer containing 0.3 m NaCl but not in IWFs obtained with buffer lacking NaCl, demonstrating that PvPGIP1 is ionically bound to the cell wall (Fig. 1).

Figure 1.

Western-blot analysis of PvPGIP1 extracted from 35S∷Pvpgip1 tomato stems. Lane 1, Purified PvPGIP1 (10 ng); lane 2, IWF of tomato stems obtained with 50 mm sodium acetate buffer containing 0.3 m NaCl (4 μg of proteins); lane 3, IWF obtained with 50 mm sodium acetate buffer (7 μg of proteins).

To single out which component of the wall interacts with PvPGIP1, tomato cell walls were fractionated as described in “Materials and Methods.” The cell wall material, i.e. the alcohol-insoluble solid (AIS), was suspended in acetate buffer and subsequently fractionated into a pellet and a buffer-soluble solid (BSS) fraction. After deproteination, the pellet was fractionated into the chelating agent-soluble solid (ChASS) and “Residue.” The sugar composition and degree of methylesterification (DM) of each fraction are reported in Table I. Each fraction is characterized by a distinctive distribution of sugars and contains various amounts of uronic acids, which are potentially involved in binding the positively charged PGIP (pI approximately 9).

Table I.

Sugar composition (% of moles) and DM of the uronic acid in the cell wall fractions from tomato stems

The total amount of uronic acids in AIS may be underestimated as it is not possible to obtain a complete polysaccharide hydrolysis by acid. Therefore, the value of DM in AIS may be overestimated. UrAc, Uronic acid.

Cell Wall Fraction	Rha	Ara	Xyl	Man	Gal	Glc	UrAc	DM	Total Sugar
								%	mg/10 mg DW of each fraction
AIS	6.0	10.2	15.3	2.2	16.8	24.1	25.4	89.7	7.2
BSS	5.7	14.6	1.9	3.7	14.5	46.0	13.5	89.3	9.7
ChASS	5.8	6.8	14.7	7.2	7.5	8.3	49.7	49.8	10.0
Residue	3.0	6.4	15.6	5.6	6.8	56.2	6.3	50.0	8.2

Using an affinity chromatography approach, the different wall fractions were tested for their ability to bind PvPGIP1. Columns containing polyacrylamide gel entrapping AIS, BSS, ChASS, and Residue, in the presence of CaCl₂, were prepared according to Penel and Greppin's procedure (Penel and Greppin, 1996). Homogeneous PvPGIP1 was loaded onto each column, and the nonbound and bound proteins were recovered by washing the column with 20 mm sodium acetate buffer, pH 5.0, containing 100 μm CaCl₂ and then eluting the bound PvPGIP1 with phosphate-buffered saline (PBS) containing 0.3 m NaCl. The relative amount of bound PGIP with respect to the amount of loaded PGIP (100%) was determined by inhibition assay against polygalacturonase II (PGII) from Aspergillus niger (AnPGII). Control columns containing only polyacrylamide did not retain PvPGIP1. PvPGIP1 bound differentially to different cell wall fractions; the level of bound PvPGIP1 was maximal (63%) on columns containing the uronic acid-rich fraction ChASS and minimal (7%) on the fraction Residue, which contains a low amount of uronic acids (Table II).

Table II.

Amount of PvPGIP1 bound to polyacrylamide gel columns containing different cell wall fractions

The Ca²⁺-wall fraction/polyacrylamide columns equilibrated with 20 mm sodium acetate buffer, pH 5.0, containing 100 μm CaCl₂ were loaded with PvPGIP1, extensively washed with the equilibration buffer, and eluted with PBS containing 0.3 m NaCl. The control was performed by loading PvPGIP1 onto a polyacrylamide gel column; n.d. stands for not detectable. The percentage of bound PGIP with respect to the amount of loaded PGIP (100%) was determined by inhibition assays against AnPGII. The percentage reported is the average (±sd) of three different experiments.

Column Containing:	Bound PGIP
	%
AIS	58 (±4)
BSS	38 (±3)
ChASS	63 (±6)
Residue	7 (±2)
Control	n.d.

The interaction with the pectic components of the cell wall also was observed for the isoform 2 of PGIP from P. vulgaris (PvPGIP2; data not shown), which differs from PvPGIP1 by only eight amino acids in the LRR motif (Leckie et al., 1999).

The Three-Dimensional Structure of PGIP Reveals a Binding Site for Pectins

The three-dimensional structure of PvPGIP2 (Di Matteo et al., 2003; Fig. 2A) revealed a positively charged patch located between the two β-sheets B1 and B2, directly below the negative pocket putatively involved in PG binding (Fig. 2B). The positive patch consists of a cluster of regularly spaced Arg and Lys residues (R183, R206, K230, and R252, according to the numbering reported in Di Matteo et al. [2003]) that protrude into the solvent, creating a regular distribution of charges (Fig. 2). We investigated whether the patch of Arg and Lys acts as a pectin-binding site. The residues were substituted with the polar amino acid Gln to generate variant proteins. Two single mutants in which R183 and K230 were mutated into Gln (QRKR and RRQR) were obtained. Also, the double mutant QRQR having both R183 and K230 mutated into Q; the triple mutants QRQQ and QQQR with the mutations R183Q, K230Q, and R252Q or R183Q, R206Q, and K230Q, respectively; and the quadruple mutant QQQQ with the mutations R183Q, R206Q, K230Q, and R252Q were constructed. Both the wild-type PvPGIP2 (RRKR) and mutants were individually expressed in Pichia pastoris, and their capability to interact with polygalacturonic acid (PGA) was analyzed at pH 5.0 and 7.0 either in the presence or in the absence of calcium. The wild-type PGIP showed maximum binding capacity in the presence of 100 μm Ca²⁺ at pH 5, whereas in the absence of calcium the binding was reduced by approximately 50%. In contrast with the wild type, the variants with a single mutation exhibited reduced capacity of binding PGA in the same conditions; remarkably, the decrease of affinity toward PGA was more evident in the presence of Ca²⁺. Moreover, the variants with double, triple, and quadruple mutations did not bind PGA (Table III). At pH 7.0 the binding of PvPGIP2 to PGA was not observed. Alginate (ALG) and polyacrylamide alone (data not shown) also were used as controls, both in the presence or absence of calcium at pH 5.0 and pH 7.0. The binding of the inhibitor to ALG and polyacrylamide was not observed in all the conditions used (data not shown).

Figure 2.

A, View of the pectin-binding motif composed by four positively charged residues (R183, R206, K230, and R252), protruding into the solvent and located between sheet B1 and sheet B2. The residues are shown in ball-and-stick representation. B, GRASP (Graphical Representation and Analysis of Structural Properties) electrostatic potential surface of PvPGIP2. Regions of negative and positive potential are shown in red and blue, respectively. A wide negative pocket, putatively involved in binding PG, is located in the middle of the inner concave surface of the protein (Di Matteo et al., 2003). A positively charged cluster involved in pectin binding is located directly below the negative pocket.

Table III.

Amount of PvPGIP2s bound to polyacrylamide gel columns containing PGA at pH 5.0

The percentage of bound PvPGIP2 with respect to the amount of loaded PvPGIP2 (100%) was determined by SDS-PAGE. At pH 7.0, binding of PvPGIP2 to PGA was not observed either in the absence or in the presence of Ca²⁺. Wild-type (in italic) and mutated PGIP variants are indicated in the first column.

	PGIP Bound %
	PGA
	EGTA	Ca2+
RRKR	47	100
QRKR	43	32
RRQR	7	17
QRQR	0	0
QRQQ	0	0
QQQR	0	0
QQQQ	0	0

The interaction of the PvPGIP2 variants with AnPGII was tested by inhibition assays and analyzed by surface plasmon resonance. All variants maintained inhibitory activity against AnPGII. The parameters of the interaction obtained by surface plasmon resonance analysis of mutated and wild-type proteins with AnPGII were similar, indicating that mutation of the pectin-binding site does not affect binding of the protein to PG (data not shown). The proximity of the pectin binding site to the region known to interact with PG (Leckie et al., 1999) suggested that, upon binding the enzyme, PvPGIP2 may be removed from the pectic matrix. Consequently, AnPGII and PG of Fusarium moniliforme (FmPG) were tested for their ability to compete with pectin for the interaction of PvPGIP1 and PvPGIP2. Columns of polyacrylamide entrapping pectin were separately loaded at pH 5.0 with the two inhibitors and, after washing, were subsequently loaded with buffer containing AnPGII or FmPG. It was shown that while PvPGIP1 interacts with AnPGII and is unable to interact with FmPG, PvPGIP2 interacts with both fungal PGs, though with different affinity (K_D = 0.96 nm for the interaction with AnPGII and 47.7 nm for the interaction with FmPG; Leckie et al., 1999). Figures 3A and 4A show that, in the absence of PGs (buffer alone), the inhibitor was not present in the flow through and was totally recovered in the bound fraction; the amount of both PGIPs initially loaded was completely retained by the column. Both enzymes were able, with different efficiency, to displace the bound PvPGIP2 from the column (Fig. 3, B–D), whereas only AnPGII displaced PvPGIP1 (Fig. 4, B and C) consistently with the different specificities of the inhibitors.

Figure 3.

Effect of AnPGII or FmPG on the binding of PvPGIP2 to Ca²⁺-PGA/polyacrylamide columns. A, Silver-stained SDS-PAGE showing that eluates with washing buffer alone do not release PGIP2. B, Silver-stained SDS-PAGE showing the effect of the FmPG on the binding of PvPGIP2 to Ca²⁺-PGA/polyacrylamide column; numbers at the bottom of the figure indicate the relative amounts of PGIP determined by densitometer. C, Silver-stained SDS-PAGE showing the effect of AnPGII on the binding of PvPGIP2 to Ca²⁺-PGA/polyacrylamide column. D, Western blot obtained from the gel showed in C using an antibody against PvPGIP2; numbers at the bottom of the figure indicate the relative amounts of PGIP as determined by densitometer. E, Western blot obtained from the gel showed in C using an antibody against AnPGII. Flow through (Ft) refers to the fraction collected from the column after loading PG, and “Bound” is the fraction collected after elution with PBS containing 0.3 m NaCl.

Figure 4.

Effect of AnPGII or FmPG on the binding of PvPGIP1 to Ca²⁺-PGA/polyacrylamide columns. A, Silver-stained SDS-PAGE showing that eluates with washing buffer alone do not contain PGIP1. B, Silver-stained SDS-PAGE showing the effect of AnPGII on the binding of PvPGIP1 to Ca²⁺-PGA/polyacrylamide column. C, Silver-stained SDS-PAGE showing the effect of FmPG on the binding of PvPGIP1 to Ca²⁺-PGA/polyacrylamide column. Flow through (Ft) refers to the fraction collected from the column after loading PG, and “Bound” is the fraction collected after elution with PBS containing 0.3 m NaCl.

The Pattern of Methylesterification Influences the Interaction of PGIP with Pectin

PvPGIP1 was used for binding experiments either with partially de-esterified pectins with a DM of 41% (P41), obtained from lime (Citrus aurantifolia) pectin by blockwise de-esterification using a plant pectin methylesterase (PME), and nonblockwise pectins with a DM of 43% (F43), obtained by de-esterification using a fungal PME. When PvPGIP1 was loaded on a P41 column in the presence of 100 μm Ca²⁺ at pH 5.0, the interaction was strong and comparable to that observed with PGA, indicating that the presence of blockwise de-esterified stretches is sufficient to confer to pectin the capability to interact with PGIP. On the other hand, when PvPGIP1 was loaded on F43 column at pH 5.0 in the presence of calcium, only 15% of the inhibitor was retained (Fig. 5). The binding of PGIP to P41 and F43 did not occur when at least two residues of the positively charged cluster were mutated, indicating that also in this case the binding involves the interaction of the same domain of PGIP with the stretches of unesterified HGA of the pectin samples.

Figure 5.

SDS-PAGE analysis of fractions obtained from affinity chromatography of PvPGIP1 loaded onto polyacrylamide columns polymerized in the presence of 100 μm CaCl₂ at pH 5.0 entrapping blockwise de-esterified pectin with a DM of 41% (Ca²⁺-P41/polyacrylamide) and nonblockwise de-esterified pectin with a DM of 43% (Ca²⁺-F43/polyacrylamide columns). The columns equilibrated with 20 mm acetate buffer, pH 5.0, containing 100 μm CaCl₂ were loaded with PvPGIP1, extensively washed with the equilibration buffer, and eluted with PBS containing 0.3 m NaCl. Bound and not-bound (Ft) fractions were analyzed by SDS-PAGE. Numbers at the bottom of the figure indicate the relative amounts of bound and not-bound PvPGIP1, as determined by densitometer and inhibitory activity against PGII from A. niger. Uronic acid assay was carried out in the eluate to exclude that pectins entrapped in the gel were not released. Bound and not-bound (FT) fractions were analyzed by SDS-PAGE. Numbers at the bottom of the figure indicate the relative amounts of bound and not-bound PvPGIP1, as determined by densitometer and inhibitory activity against PGII from A. niger. Same relative amounts of PGIP were revealed with the two different procedures.

DISCUSSION

Certain pectic domains consist of PGA bearing evenly distributed negative charges, which may interact with proteins exposing positive charges in favorable orientation. The crystal structure of PvPGIP2 revealed a cluster of regularly spaced Arg and Lys residues (R183, R206, K230, and R252) protruding into the solvent and creating a regular distribution of positive charges (Fig. 2A). A pectin-binding site with similar arrangement has been predicted previously by homology modeling for the zucchini (Cucurbita pepo) peroxidase APRX (Carpin et al., 2001). Site-directed mutagenesis of these residues in PvPGIP2 demonstrates that this array defines a specific pectin-binding site close to the LRR motif of the inhibitor. Indeed, the lack of binding of PGIP to ALG reinforces the notion that a specific interaction arises from the exact spatial fitting of Arg and Lys and the negatively charged motif of HGA. Comparable interactions are established between several animal proteins and anionic polymers of the extracellular matrix, such as heparin (Esko and Selleck, 2002).

Inhibitors mutated in the Arg-Lys cluster show altered or no affinity for PGA. Furthermore, these mutants maintain their ability of interacting with PG; the specificity of PGIP-PG interaction correlates with the capability of PG to displace the PGIP binding to pectin. This indicates that PGIP interacts with PG and pectin by engaging the same region but distinct binding epitopes. The observation that PGIP binds both fungal PGs and pectins, with no obvious structural similarities, further demonstrates its versatile capability of recognition and interaction. It is of note that a specific region of decorin, an animal LRR protein that presents intriguing similarities with PGIP, is involved in binding either the collagen type I or the transforming growth factor-β (Santra et al., 2002). LRR sequences also are involved in steroid binding, as demonstrated for the brassinosteroid receptor BR1 (Kinoshita et al., 2005). Possibly, some LRR repeats can adopt atypical folds that are suitable to form binding domains also for nonprotein molecules.

Our results suggest that stretches of unmethylated pectin are the anchorage place of the protein in the wall in physiological conditions: Unesterified pectins are able to bind PGIP at acidic pH, close to the physiological apoplastic environment, and Ca²⁺ can modulate the binding. In fact, Ca²⁺ is required for optimal binding to PGA, indicating that the distribution of negative charges of HGA in the egg-box conformation best matches the regular arrangement of positive charges on PGIP. This optimal matching is strongly affected by the mutation of any of the residues on PGIP.

The interaction of PGIP2 with PGA in the absence of Ca²⁺ is reduced by approximately 50%; the mutation of K230, which disrupts the positive cluster, affects the binding more than the mutation of R183, in which the charge density created by the stretch of three positive charges is sufficient to maintain the binding almost at the same level. Interaction at pH 7.0 is not observed either in the absence or in the presence of Ca²⁺. The lack of interaction at neutral pH cannot be ascribed to different protonation of positive residues (Arg and Lys) of PGIP or carboxylic groups of PGA. It is possible instead that a conformational transition of the PGA structure occurs and increases the pH, creating a charge distribution not suitable for PGIP binding (Cesaro et al., 1982). Indeed, we have observed by circular dichroism analysis that PGA at pH 7.0 in the presence of 100 μm calcium is subjected to a conformational change with respect to the form present at pH 5.0 (data not shown). A change of the apoplastic pH to values higher than 5 occurs in response to pathogen infection (Bolwell et al., 2002), and this may cause the release of PGIP from pectin to initiate the defense response.

Remarkably, PGIP does not bind exclusively Ca²⁺-pectate as observed for several peroxidases, for which EGTA is sufficient to abolish the binding. PGIP may also bind HGA chains without egg-box conformation, and this indicates that the protein is anchored not only to the egg-box-rich middle lamella and cell corners but also to different domains of the cell wall. Among the different cell wall fractions used in this study, PGIP binds preferentially ChASS that has the maximum content in uronic acids and a low DM. However, the affinity of PGIP for pectin fractions is not simply dependent on their DM; AIS and BSS, for example, have a similar DM but display a differential binding. It also is relevant that PGIP interacts strongly with pectin P41 having a blockwise pattern of de-esterification and less with pectin F43 with a more random pattern of methylation. In both cases, no interaction occurs when the positive residues of PGIP are mutated. We conclude that a requirement for binding is the presence in the pectin sample of several adjacent charged residues and is independent of its overall DM. Therefore, the ability of PGIP to discriminate between AIS and BSS may be due to the presence in the former fraction of longer stretches of unesterified HGA. The specific binding of PGIP to HGA reflects the way by which the inhibitor is localized in the cell wall and has the physiological significance of protecting the substrate of fungal PG in many places of the wall. Moreover, PGIP is strategically located in the wall to favor the release of elicitor-active oligogalacturonides, which may further strengthen the plant's defense against invading fungi.

MATERIALS AND METHODS

Plant Material

Tomato (Lycopersicon esculentum cv Moneymaker) and 35S∷Pvpgip1-transformed tomato plants (Desiderio et al., 1997) were grown in a growth chamber at 25°C under a 16-/8-h photoperiod for 1 month and subsequently transferred in greenhouse.

IWF Isolation

IWFs were collected from tomato stems by centrifugation as previously described (Salvi et al., 1990) and according to Lohaus et al. (2001). Two-centimeter-long stem sections were excised from stems isolated from 3-month-old 35S∷Pvpgip1 plants. Segments were stacked upright on a 105-μm nylon mesh in the bottom of a 20-mL plastic syringe. The packed sections were washed with 20 mm sodium acetate buffer, pH 5.0, for 5 min. Afterward, they were vacuum infiltrated for 5 min with 20 mm sodium acetate buffer, pH 5.0, in the presence or in the absence of 0.3 m NaCl. IWF was recovered by centrifuging the vacuum-infiltrated stems at 500g for 5 min at 4°C. The amount of IWF obtained from 1 g of tissue (fresh weight) was 0.2 to 0.3 mL. The degree of cytosolic contamination of IWF was determined by measuring the Glc-6-P dehydrogenase activity according to Takahama (1993). Total extracts were obtained by homogenizing fresh stem tissue with cold 20 mm sodium acetate buffer, pH 5.0, containing 1 m NaCl (1 mL per g of tissue) in a Waring blender at 4°C. The homogenate was than shaken for 1 h, centrifuged at 10,000g for 20 min at 4°C, and the supernatant collected. Protein concentration was determined according to Bradford (1976) using Bio-Rad reagents and bovine serum albumin as standard. SDS-PAGE was performed as described by Laemmli (1970).

Immunoblotting experiments were performed as described previously (Desiderio et al., 1997).

Materials and Strains

The following materials and reagents were obtained commercially as indicated: Pichia pastoris wild-type strain X-33, Escherichia coli TOP10F′, P. pastoris expression vector pGAPZαA, Zeocin, and agarose, all from Invitrogen; QuikChange site-directed mutagenesis kit from Stratagene; VivaFlow 200 from Sartorius; EcoRI and XbaI from Promega; AvrII from Roche; QIAprep spin plasmid miniprep kit from Qiagen; and PGA from citrus from Sigma.

Construction of an Expression Vector Encoding the PvPGIP2 Gene

The wild-type PvPGIP2 gene was cloned in pGAPZαA between the EcoRI and XbaI sites to generate a construct with the PGIP2 native signal sequence replaced by the yeast α-factor signal sequence. The construct was amplified by transforming E. coli TOP10F′ competent cells. Transformants were selected on low-salt Luria-Bertani plates containing 25 μg/mL Zeocin and confirmed by direct PCR amplification to have the gene of interest using two primers, one on the gene and one on the pGAPZαA. One PCR-positive colony was picked with a sterile tip and used to inoculate 3 mL of low-salt Luria-Bertani (1% tryptone, 0.5% yeast extract, 0.5% NaCl, pH 7.4) liquid medium containing 25 μg/mL Zeocin, and the culture was grown overnight at 37°C at 250 rpm. The plasmid was extracted from the cells using a plasmid mini prep kit. It was quantified by agarose gel, linearized with AvrII, and concentrated by isopropanol precipitation to obtain 3 μg of the plasmid in a total volume of 5 μL. Before P. pastoris transformation, the construct was analyzed by digestion with EcoRI and XbaI restriction enzymes, followed by agarose gel, to see the presence of the gene in the plasmid.

PvPGIP2 Site-Directed Mutagenesis

Site-directed mutagenesis was performed to modify the Arg and Lys predicted to be critical for the binding of the PGIP to pectins. Four mutants were designed: a double mutant in which Arg-183 and Lys-230 were replaced by Gln (QRQR); two triple mutants, one with Arg-183, Lys-230, and Arg-252 replaced by Gln (QRQQ) and one with Arg-183, Lys-230, and Arg-206 replaced by Gln (QQQR); and one quadruple mutant with Arg-183, Lys-230, Arg-252, and Arg-206 replaced by Gln (QQQQ).

Mutageneses of the PvPGIP2 gene were made through PCR-based site-directed mutagenesis using the QuikChange site-directed mutagenesis kit. PCR was carried out directly on PGIP2/pGAPZαA construct. Internal overlap primers were designed that hybridize at the site of the desired mutation and contain the relevant mismatched bases. The primers used were as follows: 5′-GCGTTCGTTGACTTGTCTCAAAACATGCTGGAGGGTGAT-3′ and 5′-GTCACCCTCCAGCATGTTTTGAGACAAGTCAACGAACGC-3′ for R206Q, 5′-TTGAACGGGTTGGATCTGCAGAACAACCGTATCTATGGG-3′ and 5′-CCCATAGATACGGTTGTTCTGCAGATCCAACCCGTTCAA-3′ for R252Q, 5′-ACGTCGATGACCATCTCCGAGAACCGCCTCACCGGGAAG-3′ and 5′-CTTCCCGGTGAGGCGGTTCTGGGAGATGGTCATCGACGT-3′ for R183Q, and 5′-CAGAAGATACATCTGGCGCAGAACTCTCTTGCCTTTGAT-3′ and 5′-ATCAAAGGCAAGAGAGTTCTGCGCCAGATGTATCTTCTG-3′ for K230Q.

The mutations were confirmed by sequencing of the gene (GeneLab).

Transformation of P. pastoris and Selections of Transformants

The construct was used to transform wild-type P. pastoris X-33 competent cells. The transformants were selected on YPDS (1% yeast extract, 2% peptone, 2% dextrose, and 1 m sorbitol) plates containing 100 μg/mL Zeocin. Ten or more transformant colonies were picked with a sterile tip and used to inoculate fresh YPDS plates containing 100 μg/mL Zeocin and 3 mL of liquid BMMY-modified medium (0.4% yeast extract, 0.6% tryptone, 50 mm phosphate buffer, pH 6, 1.34% YNB, 4 × 10^−5% biotin, 2% Glc) containing 100 μg/mL Zeocin. After 5 d of incubation at 28°C at 300 rpm, the cultures were centrifuged and the supernatants used to perform an agar diffusion assay (cup-plate) to choose the higher-level expression clones. The colony that expressed the highest level of protein was used for the production of PGIP or mutants proteins. The selected colony on the YPDS plate was first picked with a sterile tip to 1 mL of liquid BMMY modified containing 100 μg/mL Zeocin, and, after 3 d of incubation at 28°C at 300 rpm, the culture was used to inoculate 100 mL of fresh BMGY medium containing 100 μg/mL Zeocin. When the new culture reached an OD₆₀₀ = 20 to 30, it was used to inoculate 2 L of fresh BMGY (1% yeast extract, 2% peptone, 1% glycerol, 100 mm potassium phosphate, pH 6) medium and was incubated at 28°C at 300 rpm for 5 d. At the end of this time, the culture was centrifuged and the supernatant was concentrated by VivaFlow 200 and dialyzed against 20 mm sodium acetate, pH 4.7.

Agar Diffusion Assay

Agar diffusion assay (cup-plate) was performed according to Taylor and Secor (1988). Fifteen microliters (or 40 μL) of culture medium containing PGIP and PG from Colletotrichum acutatum was added to 0.5-cm wells on plates containing 100 mm sodium acetate, pH 4.7, 0.5% PGA, and 0.8% agarose. Plates were incubated for 12 h at 30°C, and the halo caused by enzyme activity was visualized after 5 min of treatment with 6 n HCl. The smaller halo corresponds to the higher amount of PGIP.

Purification of the Proteins

After dialysis, wild-type or mutated PGIP2 expressed in P. pastoris was purified by ion-exchange chromatography as described previously (Desiderio et al., 1997). Fractions containing PGIP2 wild type or mutants, detected by SDS-PAGE analysis, were pooled, dialyzed against 10 mm sodium acetate, pH 4.7, and subjected to isoelectrofocusing in a IEF Rotofor system (Bio-Rad) in a pH 3 to 10 gradient. Fractions containing PGIP were collected and dialyzed against 20 mm sodium acetate, pH 4.7, and finally concentrated to 1 mg/mL by using a VIVASPIN-2 concentrator with a cutoff of 5 kD.

Preparation and Purification of PGIP and PG

PvPGIP1 was purified from 35S∷Pvpgip1 transgenic tomato plants by a nonaffinity-based purification procedure as described previously (Desiderio et al., 1997). PvPGIP2 was purified from Potato virus X-infected tissues of Nicotiana benthamiana plants by an affinity-based procedure on a Sepharose-Aspergillus niger PG column as described previously (Leckie et al., 1999). PGII from A. niger was a kind gift of Dr. J. Benen, and PG of Fusarium moniliforme expressed in Saccharomyces cerevisiae was prepared as described by Caprari et al. (1996). PG and PGIP activities were measured with the standard parahydroxybenzoic acid hydrazide assay as already described (Leckie et al., 1999).

Isolation of Cell Wall Fractions

Cell walls were isolated from tomato stems of 3-month-old plants grown in greenhouse and fractionated according to the protocol of Stolle-Smits et al. (1997) with some modifications, as described previously (Capodicasa et al., 2004).

Sugar Composition

Cell wall fractions were prehydrolyzed by dispersing the dried sample in 7.3 m sulfuric acid for 1 h at 30°C, followed by hydrolysis in 1 m sulfuric acid for 3 h at 100°C under continuous stirring. The hydrolyzed fractions were then converted in alditol acetates and analyzed on a glass column (15 m × 0.53 mm i.d.), coated with DB-225 (film thickness 1.0 μm) in a Carlo Erba Fractovap 2300 GC. Uronic acids were determined as described by Ahmed and Labavitch (1977), and neutral sugars were measured using the phenol-sulfuric colorimetric assay according to DuBois et al. (1956).

Determination of DM

The amount of methyl groups was determined by HPLC as described by Voragen et al. (1986).

Matrix Polysaccharide/Polyacrylamide Affinity Chromatography

Affinity chromatography gels were prepared according to Penel and Greppin (1996) with some modifications. A total of 800 μg of cell wall fractions, PGA, or pectins with different DM were mixed with 1 mL of 12% polyacrylamide solution in the presence or absence of 10 mm CaCl₂. Ten microliters of APS and 12.5 μL of Temed were added and the polymerization was allowed overnight. Lime (Citrus aurantifolia) pectins with different patterns (blockwise and nonblockwise) and defined DM were prepared by enzymatic treatment of highly methylesterified lime pectin (E81; GRINDSTED Pectin URS 1200) as described previously (Limberg et al., 2000; Willats et al., 2001). Briefly, one series was produced by blockwise de-esterification of E81 with plant PME isolated from orange (Citrus sinensis) peel (P series), while another series was produced by nonblockwise de-esterification of E81 with a fungal PME from A. niger (F series). After polymerization, the gels were fragmented in small particles and packed into chromatography columns (50 × 5 mm) that were equilibrated with 20 mm sodium acetate buffer, pH 5.0, or with 20 mm HEPES, pH 7.0, in the presence or absence of CaCl₂. Columns prepared without CaCl₂ were washed with buffer containing 1 mm EGTA. Each column was equilibrated with 20 mm sodium acetate buffer, pH 5.0, or 20 mm HEPES, pH 7.0, in the presence or in the absence CaCl₂, loaded with PGIP2 (10 μg), and extensively washed with the buffer used for column equilibration. We have excluded that pectins entrapped in the polyacrylamide columns were eluted by high salt concentration by performing, several times, uronic acid assay of the eluate (York et al., 1985; Leckie et al., 1999). The bound protein was eluted with PBS containing 1 m NaCl and 1 mm EGTA. To determine the amount of PGIP eluted or retained by the columns, all the collected fractions were analyzed by SDS-PAGE and assayed for PGIP inhibitory activity.

Effect of PGII from A. niger and PG from F. moniliforme on the Binding of PvPGIP2 and PvPGIP1 to Pectins

PvPGIP2 (15 μg) and PvPGIP1 (15 μg) were separately loaded onto two Ca²⁺-PGA/polyacrylamide columns equilibrated in 20 mm sodium acetate buffer containing 100 μm CaCl₂. The columns were washed with 20 mm sodium acetate buffer containing 100 μm CaCl₂. Subsequently, AnPGII (15 μg) or FmPG (15 μg) dissolved in the same buffer was loaded onto each column. Flow through, the fraction collected from the column after loading PG, was recovered; the columns were washed again with the buffer and the bound proteins were eluted with PBS containing 0.3 m NaCl. To determine the amount of PGIP eluted or retained by the columns, all the collected fractions were analyzed by SDS-PAGE and western blot, and a densitometric analysis was done.