Portal de Periódicos da CAPES

The specificity of polygalacturonase-inhibiting protein (PGIP): a single amino acid substitution in the solvent-exposed β-strand/β-turn region of the leucine-rich repeats (LRRs) confers a new recognition capability

1999; Springer Nature; Volume: 18; Issue: 9 Linguagem: Inglês

10.1093/emboj/18.9.2352

1460-2075

Fiona Leckie, Benedetta Mattei, Cristina Capodicasa, Andrew M. Hemmings, L. Nuss, B. Aracri, Giulia De Lorenzo, Felice Cervone,

Polyamine Metabolism and Applications

Article4 May 1999free access The specificity of polygalacturonase-inhibiting protein (PGIP): a single amino acid substitution in the solvent-exposed β-strand/β-turn region of the leucine-rich repeats (LRRs) confers a new recognition capability F. Leckie F. Leckie Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author B. Mattei B. Mattei Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author C. Capodicasa C. Capodicasa Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author A. Hemmings A. Hemmings Schools of Biological and Chemical Sciences, University of East Anglia, Norwich, NR4 7TJ UK Search for more papers by this author L. Nuss L. Nuss Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author B. Aracri B. Aracri Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author G. De Lorenzo G. De Lorenzo Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author F. Cervone Corresponding Author F. Cervone Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author F. Leckie F. Leckie Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author B. Mattei B. Mattei Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author C. Capodicasa C. Capodicasa Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author A. Hemmings A. Hemmings Schools of Biological and Chemical Sciences, University of East Anglia, Norwich, NR4 7TJ UK Search for more papers by this author L. Nuss L. Nuss Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author B. Aracri B. Aracri Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author G. De Lorenzo G. De Lorenzo Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author F. Cervone Corresponding Author F. Cervone Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy Search for more papers by this author Author Information F. Leckie1, B. Mattei1, C. Capodicasa1, A. Hemmings2, L. Nuss1, B. Aracri1, G. De Lorenzo1 and F. Cervone 1 1Dipartimento di Biologia Vegetale, Università di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy 2Schools of Biological and Chemical Sciences, University of East Anglia, Norwich, NR4 7TJ UK *Corresponding author. E-mail: [email protected] The EMBO Journal (1999)18:2352-2363https://doi.org/10.1093/emboj/18.9.2352 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Two members of the pgip gene family (pgip-1 and pgip-2) of Phaseolus vulgaris L. were expressed separately in Nicotiana benthamiana and the ligand specificity of their products was analysed by surface plasmon resonance (SPR). Polygalacturonase-inhibiting protein-1 (PGIP-1) was unable to interact with PG from Fusarium moniliforme and interacted with PG from Aspergillus niger; PGIP-2 interacted with both PGs. Only eight amino acid variations distinguish the two proteins: five of them are confined within the β-sheet/β-turn structure and two of them are contiguous to this region. By site-directed mutagenesis, each of the variant amino acids of PGIP-2 was replaced with the corresponding amino acid of PGIP-1, in a loss-of-function approach. The mutated PGIP-2s were expressed individually in N.benthamiana, purified and subjected to SPR analysis. Each single mutation caused a decrease in affinity for PG from F.moniliforme; residue Q253 made a major contribution, and its replacement with a lysine led to a dramatic reduction in the binding energy of the complex. Conversely, in a gain-of-function approach, amino acid K253 of PGIP-1 was mutated into the corresponding amino acid of PGIP-2, a glutamine. With this single mutation, PGIP-1 acquired the ability to interact with F.moniliforme PG. Introduction Polygalacturonase-inhibiting proteins (PGIPs), present in the cell walls of many plants (Cervone et al., 1997), belong to the large family of leucine-rich repeat (LRR) proteins. At present, >100 LRR proteins of diverse origin (microbia, animals and plants) have been described. The LRR is a versatile structural motif responsible for many protein–protein interactions and involved in many different cell functions such as receptor dimerization, domain repulsion, regulation of adhesion and binding events (Buchanan and Gay, 1996). In the few cases investigated so far, the importance of LRRs for interaction with other molecules has been clearly assessed. For example, the specific binding sites of decorin, a protein belonging to the proteoglycan family, for collagen type I have been localized within the sixth LRR, where a single mutation, E180K, is responsible for a major diminution in binding (Kresse et al., 1997). Hormone binding of the lutropin–choriogonadotropin receptor (LH/CG-R) (G-protein-coupled receptors) was localized within the LRRs 1–6 of the receptor (Puett et al., 1996; Thomas et al., 1996). A significant advance in understanding the structural basis of LRR-mediated molecular interactions comes from crystallographic studies of the ribonuclease inhibitors (RI). Co-crystallization of porcine RI (pRI) and human placental RI (hRI) with RNase A and angiogenin (Ang), respectively, has been achieved (Kobe and Deisenhofer, 1996; Papageorgiou et al., 1997). In these proteins, a repeated β-strand/β-turn structure is determined by the presence in each LRR module of the motif xxLxLxx, where the leucine residues form a hydrophobic core, while the side chains of the amino acids flanking the leucines are solvent exposed and interact with the ligands (Kobe and Deisenhofer, 1993, 1995). Twenty six out of 28 contact points between pRI and RNase A occur in the β-strand or β-turn region of the LRRs; similarly, 25 of the 26 contact points between hRI and Ang are located in the β-strand or β-turn region. The majority of the hydrogen bonds and van der Waals contacts in the two complexes are distinctive, indicating that the ability of the inhibitor to recognize different ligands is based on its ability to interact with a number of features unique to each of them (Papageorgiou et al., 1997). However, a thorough analysis of the contribution of the single amino acids in the RI–RNase interactions is difficult, due the high number of contacts established in the complexes. In plants, LRR proteins play a relevant role in both development and defence, where specificity of recognition is a fundamental prerequisite. These proteins include PGIPs, the products of the resistance (R) genes Cf of tomato, which confer resistance to different races of the fungus Cladosporium fulvum (Hammond-Kosack and Jones, 1997), and Xa21 of rice, which confers resistance to Xanthomonas oryzae pv. oryzae (Wang et al., 1996), as well as several orphan receptor kinases involved in Arabidopsis development, such as ERECTA (Torii et al., 1996), CLAVATA1 (Clark et al., 1997) and a putative receptor for brassinosteroids (Li and Chory, 1997). All these proteins share LRRs of the extracellular or extracytoplasmic type, characterized by the consensus sequence LxxLxxLxLxxNxLT/SGxIPxxLGx (Kajava, 1998), and a similarity not only in the LRR region but also in the regions outside the LRR domain (Bent, 1996; De Lorenzo and Cervone, 1997). Although to a lesser extent, PGIPs also share similarity with R gene products characterized by LRRs of the intracellular type (Hammond-Kosack and Jones, 1997). Because R gene products are thought to function as receptors for pathogen-encoded avirulence (Avr) proteins, it has been hypothesized that sequence variation within LRRs influences recognition specificity. Comparison of members of the Cf family has identified the β-sheet/β-turn region as a ‘hypervariable’ region, probably responsible for the ligand specificity in this class of proteins (Parniske et al., 1997). However, analysis of the molecular basis of recognition specificity either in the R gene products or in the development-related LRR receptors is not yet possible because the nature of the ligands for the latter is still unknown, while the evidence for a direct molecular interaction between an LRR R protein and an avr product is still awaited. Given their close structural relationship to these proteins, PGIPs and their ligands, polygalacturonases (PGs), represent a unique model system for studying the structural bases of recognition specificity of plant LRR proteins. This knowledge can be exploited for designed manipulation of the LRR structure to generate new specific molecular interactions for the control of developmental processes or the creation of new resistance traits in plants. PGIPs interact with fungal endopolygalacturonases and inhibit their enzymatic activity in vitro (De Lorenzo and Cervone, 1997). The proteins isolated from bean (Cervone et al., 1987), pear (Stotz et al., 1993), raspberry (Johnston et al., 1993), tomato (Stotz et al., 1994) and soybean (Favaron et al., 1994) have differential inhibition spectra towards a range of PGs from phytopathogenic fungi. Different inhibitory activities against PGs have also been observed in PGIPs from a single plant source (Desiderio et al., 1997), indicating that pgip genes have undergone diversification during evolution. Like many plant R genes, pgip genes are organized into complex multigene families. In Phaseolus vulgaris, the pgip gene family consists of at least five members and perhaps as many as 15 (Frediani et al., 1993). Previous data suggest that different members of the family encode PGIPs with nearly identical biochemical characteristics but distinct specificity, i.e. the ability to interact with different fungal PGs (Desiderio et al., 1997). We have now searched for and isolated, from a cDNA library of P.vulgaris cv. Pinto, two members of the pgip gene family (pgip-1 and pgip-2), which, within the coding regions, differ by only 26 nucleotides. Upon expression in Nicotiana benthamiana, we have investigated by surface plasmon resonance (SPR) the ligand specificity of the encoded proteins before and after site-directed mutagenesis to evaluate the energetic parameters of individual interface contributions, which cannot be gained from crystallographic data. Here, we report the distinct ability of PGIP-1 and PGIP-2 to recognize fungal PGs, and on the role and contribution of the single amino acids that distinguish PGIP-1 and PGIP-2 in the specific interactions with PGs from Aspergillus niger and Fusarium moniliforme. Our results show that the residues determining the recognition specificity of PGIP reside in the region flanking the predicted β-sheet/β-turn structure of the protein, and that a single amino acid variation in the β-strand/β-turn motif can confer to PGIP a new recognition capability. Results Isolation and characteristics of the pgip clones A cDNA library from suspension-cultured cells of P.vulgaris cv. Pinto was screened using as a probe the genomic pgip clone previously isolated from a library of P.vulgaris cv. Saxa (Toubart et al., 1992). Seventeen clones were purified and subjected to restriction enzyme digestion and Southern blot analysis; the 10 longest inserts were subcloned in the SalI–EcoRI site of the pBlueScript SK+ plasmid and sequenced. Since previous data had shown that PGIPs with different specificities have indistinguishable biochemical characteristics, suggesting that different pgip genes might be highly similar (Desiderio et al., 1997), even a few nucleotide differences among the cDNAs were not neglected and were confirmed carefully. All the 10 sequenced cDNAs exhibited a poly(A) tail and could be grouped into two classes, each corresponding to a distinct pgip gene. Within each class, cDNAs had completely matching nucleotide sequences but different lengths, because, being partial cDNAs, they differed at their 5′ ends. One class included four cDNAs and corresponded to a gene, pgip-1, identical and co-linear with the genomic pgip clone from P.vulgaris cv. Saxa (Toubart et al., 1992); within this class, only one clone contained the two in-frame ATGs described in the Saxa pgip clone, while the other three cDNAs were shorter. The other class (six cDNAs) corresponded to a different gene, named pgip-2, that shares a high degree of identity with pgip-1 (99.1% in the coding region). The longest cDNA started with an ATG, corresponding to the second ATG of pgip-1 (Figure 1). Figure 1.Top: nucleotide sequence alignments of the coding sequences of pgip-1 and pgip-2 isolated from a P.vulgaris cv. Pinto cDNA library. For pgip-2, only nucleotides that differ from pgip-1 are indicated. Dots indicate identity. Start and stop codons are indicated in bold. HindIII and MluI endonuclease restriction sites are underlined and indicated by H and M, respectively. Bottom: PGIP-2 LRR structure. (A), signal peptide; (B), presumed N-terminus of the mature protein; (C), 10.5 LRRs; and (D), C-terminus. Putative glycosylation sites are indicated by an asterisk. The box indicates the area of the protein predicted to form the β-sheet/β-turn structural motifs. Based on the comparison between pgip-1 and pgip-2, amino acid residues corresponding to synonymous nucleotide changes are indicated in bold, those corresponding to non-synonymous variations are highlighted. Amino acids are numbered according to the corresponding residues of PGIP-1 (see also the footnotes to Table I). Download figure Download PowerPoint Nucleotide changes (a total of 26, resulting in 10 amino acid changes between PGIP-1 and PGIP-2) are more frequent in the region encoding the C-terminal half of the LRR domain (Table I). A high number and a particular distribution of non-synonymous (11/26) substitutions compared with synonymous (15/26) substitutions can be observed. Furthermore, as shown in Figure 1, seven of the 11 non-synonymous substitutions lead to amino acid differences which are located in the LRR domain: five of these are internal to the xxLxLxx motif predicted to form the solvent-exposed β-sheet/β-turn structure of the protein, while two other amino acid substitutions are very close to this region and their side chains presumably are also solvent exposed. Each LRR contains only one variation, and some LRRs (the second, third, fourth, seventh and ninth) are invariant between the two proteins. The remaining three variant amino acids are outside the LRR domain: two are located in the signal peptide of the protein and therefore do not affect the structure of the mature protein, and one resides in the C-terminal region of the protein. Instead, most synonymous nucleotide changes correspond to residues located outside the β-sheet/β-turn structural motif. Table 1. Amino acids which distinguish PGIP-1 from PGIP-2a Amino acid position PGIP-1 PGIP-2 26 R S 29 L H 89 H L 181 G V 207 A S 253 K Q 300 Q H 320 K Q 326 S A 340 S A The amino acid position refers to the residue in PGIP-1. Since the cDNA clone coding for PGIP-2 contains only one methionine codon corresponding to amino acid position 10 in PGIP-1, the amino acid position 26 in PGIP-1 corresponds to the amino acid position 17 in PGIP-2. In order to evaluate the possible functional significance of the point mutations and the subsequent amino acid changes, the relative mutability of the variant amino acids between PGIP-1 and PGIP-2 was analysed according to Dayhoff et al. (1972) by comparing their rate of acceptance within protein families which display point mutations. The more acceptable an interchange is between two amino acids, the more frequent it is, and this depends on the chemical and physical similarities between the amino acids. Conversely, a low rate of acceptance should be a rare event and is indicative of a selection pressure in favour of diversification. Substitutions with a low rate of acceptance are typical of proteins involved in recognition functions. Some of the variations between PGIP-1 and PGIP-2 have a low rate of acceptance: in particular, the variations G181V and K253Q, occurring within the putative β-sheet/β-turn motif, and H89L and K320Q, contiguous to this structural motif, are low or moderately accepted. Instead, variations S326A and S340A, as well as variations Q300H and A207S, both have a high rate of acceptance, and are not expected to determine major functional differences. Inhibition specificity of PGIP-1 and PGIP-2 The inhibitory activity of bulk bean PGIP is a composite of the activities of several PGIPs (Desiderio et al., 1997). In order to analyse the individual contribution of PGIP-1 and PGIP-2 to the total inhibition spectrum, the complete coding sequences of the genes pgip-1 and pgip-2 were introduced separately into the expression cassette of the virus vector based on potato virus X (PVX) (Baulcombe et al., 1995) to create PVX.PGIP-1 and PVX.PGIP-2 for transient expression in N.benthamiana as previously described (Desiderio et al., 1997). The putative leader or signal peptide sequence from nucleotide +1 to +27 of pgip-1 was also added to pgip-2 as this sequence may be important for high level expression of the protein (Devoto et al., 1998). Western blot analysis of crude protein extracts from symptomatic plants inoculated with PVX.PGIP-1 and PVX.PGIP-2 demonstrated the presence of a PGIP-specific signal with a molecular mass of 39 kDa which was absent in wild-type extracts (data not shown). After purification to homogeneity from N.benthamiana extracts, the ability of increasing amounts of PGIP-1 and PGIP-2 to inhibit PGs from different fungi was investigated. PGIP-1 showed a specificity spectrum very similar to that reported for PGIP-1 from cv. Saxa (Desiderio et al., 1997): 30 ng inhibited the homogeneous PG from A.niger at almost 100%, but did not inhibit a homogeneous PG of F.moniliforme expressed in Saccharomyces cerevisae (Caprari et al., 1996), and had a reduced ability to inhibit crude preparations of PG from Fusarium oxysporum f.sp. lycopersici and Botrytis cinerea. PGIP-2 (30 ng) almost completely inhibited all PGs used, with the exception of PG from F.oxysporum f.sp. lycopersici which was only partially inhibited (60%) (Figure 2). Therefore, PGIP-2 exhibits a broader spectrum of interactions than PGIP-1, and shows an interaction feature absent in PGIP-1, i.e. the ability to recognize F.moniliforme PG. Figure 2.Inhibition of various fungal PGs by increasing amounts of PGIP-1 (♦) and PGIP-2 (▪). The enzymes used were: 0.015 U of homogeneous A.niger PG (A), 0.008 U of homogeneous F.moniliforme PG (B), 0.01 U of a crude PG preparation from F.oxysporum f.sp. lycopersici (C) and 0.004 U of a crude PG preparation from B.cinerea (D). Download figure Download PowerPoint The interaction between purified PGIP-1 and PGIP-2, and homogeneous A.niger and F.moniliforme PGs, was examined by using a biosensor based on SPR (Granzow and Reed, 1992; Schuster et al., 1993). PGIP-1 and PGIP-2 were immobilized as ligands on sensor surfaces while PGs were passed in solution as analytes over the surface. The interactions between PGIP-1 or PGIP-2 with increasing amounts of the two PGs were analysed kinetically. PGIP-1 was unable to interact with F.moniliforme PG, and showed an affinity towards A.niger PG (KD = 62.1 nM) comparable with that of PGIP-1 from cv. Saxa (KD = 40 nM) (Desiderio et al., 1997). Instead, PGIP-2 interacted with both enzymes. In comparison with PGIP-1, PGIP-2 showed a much higher affinity for A.niger PG (KD = 0.96 nM) and had the capacity to interact with F.moniliforme PG (KD= 47.7 nM) (Figure 3; Tables II and III). Figure 3.Interactions between PGIP-1, PGIP-2 or domain swaps PGIP-12 and PGIP-21, and A.niger or F.moniliforme PG. The different PGIPs were immobilized separately as ligands on sensor surfaces, and the increasing concentration of PGs indicated below were passed in solution as analytes over the surface. The different panels show the SPR sensorgrams. (Concentrations listed from bottom to top curve.) PGIP-1, concentrations of A.niger PG: 6.5, 13, 23, 27, 46, 91, 274, 548 and 822 nM, and 1.1 μM; concentrations of F.moniliforme PG: 160, 400 and 800 nM, and 1.6, 3.2 and 4.8 μM. PGIP-2, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 46, 114, 228 and 456 nM, and 1.1, 2.3 and 4.6 μM; concentrations of F.moniliforme PG: 40, 80, 160, 400 and 800 nM, and 1.6, 3.2, 4.8 and 6.4 μM. PGIP-12, concentrations of A.niger PG: 6.5, 13, 23, 26, 46, 91, 274 and 548 nM, and 1.1 μM; concentrations of F.moniliforme PG: 40, 80, 160, 240, 320, 440, 640 and 882 nM, and 1.7, 2.6 and 3.5 μM. PGIP-21, concentrations of A.niger PG: 6.8, 16, 34, 68, 137, 228 and 685 nM, and 1.1 and 2.3 μM; concentrations of F.moniliforme PG: 24, 48, 80, 240, 480 and 800 nM, and 1.2 and 1.6 μM. RU, resonance units. Download figure Download PowerPoint Table 2. Kinetics and equilibrium of the interactiona between F.moniliforme PG and different PGIPs kon (per Ms) koff (per s) KD (nM) ΔG (kcal/mol) ΔΔG = ΔGmut − ΔGw (kcal/mol) PGIP-2 (wt) 4.86×104 2.32×10−3 47.7 −9.98 0.00 PGIP-1 −b −b −b PGIP-12 1.76×104 3.78×10−3 215 −9.09 0.89 PGIP-21 −b −b −b PGIP-2 L89H 3.0×104 2.48×10−3 82.6 −9.66 0.33 PGIP-2 V181G 4.44×104 4.26×10−3 96 −9.57 0.41 PGIP-2 S207A 3.43×104 2.9×10−3 84.5 −9.64 0.34 PGIP-2 Q253K 8.2×103 2.9×10−2 3536.6 −7.43 2.55 PGIP-2 H300Q 4.22×104 2.78×10−3 65.8 −9.79 0.19 PGIP-2 Q320K 1.71×104 1.58×10−3 92.4 −9.59 0.39 PGIP-2 A326S 1.57×104 2.61×10−3 166.2 −9.24 0.74 PGIP-2 A340S 3.43×104 2.23×10−3 65 −9.80 0.18 Kinetic parameters were determined by SPR analysis. KD values were calculated as koff/kon. The free energy of the formation of the complex was calculated from the equation ΔG = RTlnKD. ΔΔG values were calculated from the equation ΔΔG = −RTln(KDwt/KDmut). b No interaction. Table 3. Kinetics and equilibrium of the interactiona between different PGIPs and A.niger PG kon (per Ms) koff (per s) KD (nM) ΔG (kcal/mol) ΔΔG = ΔGmut − ΔGw (kcal/mol) PGIP-2 (wt) 3.89×105 3.74×10−4 0.96 −12.30 0.00 PGIP-1 1.01×105 6.27×10−3 62.1 −9.83 PGIP-12 1.07×105 1.4×10−3 13.1 −10.75 1.55 PGIP-21 4.41×105 4.6×10−3 10.4 −10.89 1.41 PGIP-2 L89H 3.69×105 2.9×10−4 0.78 −12.42 −0.12 PGIP-2 V181G 2.38×105 1.36×10−3 5.73 −11.24 1.06 PGIP-2 S207A 5.13×105 3.2×10−4 0.62 −12.56 −0.26 PGIP-2 Q253K 5.25×105 2.79×10−3 5.31 −11.28 1.01 PGIP-2 H300Q 3.6×105 2.34×10−4 0.65 −12.53 −0.23 PGIP-2 Q320K 2.35×105 3.75×10−4 1.59 −12.00 0.30 PGIP-2 A326S 4.24×105 3.74×10−4 0.88 −12.35 −0.05 PGIP-2 A340S 2.43×105 3.58×10−4 1.47 −12.04 0.25 Kinetic parameters were determined by SPR analysis. KD values were calculated as koff/kon. The free energy of the formation of the complex was calculated from the equation ΔG = RTlnKD. ΔΔG values were calculated from the equation ΔΔG = −RTln(KDwt/KDmut). Analysis of domain-swapped PGIPs In order to understand which region of PGIP-2 is responsible for recognition of F.moniliforme PG, we swapped domains between PGIP-1 and PGIP-2, exploiting the presence of a restriction enzyme site, MluI, in the coding region of both genes (Figure 1). The region of pgip-2 encoding the N-terminal portion (nucleotides 1–722, corresponding to amino acids 1–241) was replaced with the corresponding portion of pgip-1 to create pgip-12, and vice versa to create pgip-21. The sequences encoding the swapped PGIPs were introduced separately into the PVX vector. Following expression in N.benthamiana, the chimeric proteins were purified and analysed by SPR using A.niger and F.moniliforme PGs as analytes. SPR analysis showed that PGIP-21 does not interact with F.moniliforme PG (Figure 3; Tables II and III) and, in comparison with PGIP-2, exhibits an affinity 10-fold lower for the A.niger PG. Instead, PGIP-12 interacts with both enzymes, and exhibits, in comparison with PGIP-2, affinities 4- and 13-fold lower towards F.moniliforme and A.niger PGs, respectively. Increased KD values are due to changes in both kon and koff for PGIP-12, and in koff alone for PGIP-21. We concluded that although residues crucial for the interaction with the F.moniliforme PG are located in the C-terminal half of the LRR domain, residues in the N-terminal region also contribute, albeit weakly, to the interaction. Site-directed mutagenesis of PGIP-2: analysis with PG of F.moniliforme The contribution of each single amino acid to the interaction with F.moniliforme PG was studied by mutating, in a loss-of-function approach, each of the variant amino acids of PGIP-2 into the corresponding residue of PGIP-1. Thus a series of eight mutated pgip-2 genes were created and expressed in N.benthamiana. The encoded proteins were purified and individually immobilized on separate sensor chips for analysis by SPR with F.moniliforme PG as analyte. The SPR data (Figure 4; Table II) showed that mutation Q253K strongly affects the interaction. The kinetic constants could not be calculated due to the very weak interaction between the mutated protein and the enzyme: the KD is at least 70-fold higher than that calculated for the wild-type PGIP-2, with both a decrease in the kon value and an increase in the koff value of the dissociation of the complex. PGIP-2 carrying the mutation A326S also showed a decreased affinity: the KD is 3.5-fold higher with respect to PGIP-2, mainly due to a decrease in the kon. These data are consistent with the data obtained with the domain-swapped PGIPs, and in particular with the observation that PGIP-21, which contains both mutations Q253K and A326S, does not interact with F.moniliforme PG. Other mutations had minor effects: mutation V181G produced a 2-fold decrease in the affinity, and mutations L89H and Q320K caused very little variation in the affinity for the enzyme. Mutations H300Q and A340S had little or no effect on the interaction between PGIP-2 and F.moniliforme PG (Figure 4; Table II). Figure 4.Interaction between different PGIP-2 mutants and A.niger or F.moniliforme PG. The different panels show the SPR sensorgrams. (Concentrations listed from bottom to top curve.) L89H, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 46, 114, 228 and 456 nM, and 1.1, 2.3 and 4.6 μM; concentrations of F.moniliforme PG: 40, 80, 160, 400 and 800 nM, and 1.6, 3.2, 4.8 and 6.4 μM. V181G, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 46, 114, 228 and 456 nM, and 1.1, 2.3 and 3.4 μM; concentrations of F.moniliforme PG: 14, 40, 80, 240 and 800 nM, and 1.2 and 1.6 μM. S207A, concentrations of A.niger PG: 4.6, 11, 23, 46, 114 and 228 nM, and 1.1 and 2.3 μM; concentrations of F.moniliforme PG: 16, 40, 80, 160, 400 and 800 nM, and 1.6 and 3.2 μM. Q253K, concentrations of A.niger PG: 2.3, 11, 23, 114 and 228 nM, and 1.1 and 2.3 μM; concentrations of F.moniliforme PG: 14, 40, 80, 240 and 800 nM, and 1.2 and 1.6 μM. H300Q, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 114, 228 and 456 nM, and 1.1, 2.3 and 3.4 μM; concentrations of F.moniliforme PG: 14, 40, 80, 240 and 800 nM, and 1.2 and 1.6 μM. Q320K, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 114, 228 and 456 nM, and 1.1, 2.3 and 3.4 μM; concentrations of F.moniliforme PG: 14, 40, 80, 240, 400 and 800 nM, and 1.2 and 1.6 μM. A326S, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 114, 228 and 456 nM, and 1.1, 2.3 and 3.4 μM; concentrations of F.moniliforme PG: 14, 40, 80, 240, 400 and 802 nM, and 1.2 and 1.6 μM. A340S, concentrations of A.niger PG: 2.3, 4.6, 11, 23, 114, 228 and 456 nM, and 1.1, 2.3 and 3.4 μM; concentrations of F.moniliforme PG: 14, 40, 80, 160, 240, 400 and 800 nM, and 1.2 and 1.6 μM. RU, resonance units. Download figure Download PowerPoint Differences in binding free energies between the F.moniliforme PG–wild-type PGIP-2 complex and those of each single mutated protein interacting with F.moniliforme PG were calculated from the equation . The results are summarized in Table II. The replacement of Q253 with K decreases the binding energy by 2.55 kcal/mol, accounting for much of the binding energy of the complex. The residues A326 and V181 make lesser contributions, and all the other residues do not appear to play important roles. The importance of the amino acids distinguishing PGIP-2 from PGIP-1 was also studied in inhibition assays using PGIP-2s mutated in one or two amino acids. Sixty nanograms of all PGIP-2 single mutants were able to inhibit F.moniliforme PG, with the exception of mutant Q253K, which had a much reduced inhibition activity (30%). The double mutant V181G/Q253K as well as the double mutant Q253K/A326S did not inhibit the enzyme; instead, the double mutant V181G/A326S was able to inh