Recognition of corn defense chitinases by fungal polyglycine hydrolases

Abstract

Polyglycine hydrolases (PGH)s are secreted fungal endoproteases that cleave peptide bonds in the polyglycine interdomain linker of ChitA chitinase, an antifungal protein from domesticated corn (Zea mays ssp. mays). These target‐specific endoproteases are unusual because they do not cut a specific peptide bond but select one of many Gly‐Gly bonds within the polyglycine region. Some Gly‐Gly bonds are cleaved frequently while others are never cleaved. Moreover, we have previously shown that PGHs from different fungal pathogens prefer to cleave different Gly‐Gly peptide bonds. It is not understood how PGHs selectively cleave the ChitA linker, especially because its polyglycine structure lacks peptide sidechains. To gain insights into this process we synthesized several peptide analogs of ChitA to evaluate them as potential substrates and inhibitors of Es‐cmp, a PGH from the plant pathogenic fungus Epicoccum sorghi. Our results showed that part of the PGH recognition site for substrate chitinases is adjacent to the polyglycine linker on the carboxy side. More specifically, four amino acid residues were implicated, each spaced four residues apart on an alpha helix. Moreover, analogous peptides with selective Gly‐>sarcosine (N‐methylglycine) mutations or a specific Ser‐>Thr mutation retained inhibitor activity but were no longer cleaved by PGH. Additonally, our findings suggest that peptide analogs of ChitA that inhibit PGH activity could be used to strengthen plant defenses.

Introduction

Polyglycine hydrolases (PGH)s are highly selective, secreted fungal proteases that truncate specific plant chitinases involved in plant‐fungal defense. PGH activity was discovered by proteomic analysis of hybrid field corn infected with the ear rot pathogen Cochliobolus carbonum (syn Bipolaris zeicola), when it was observed that two corn chitinases are modified as a result of disease.1 The modified chitinases were identified as two alloforms of ChitA, an abundant protein in corn seeds that had been implicated in anti‐fungal defense.2 The ChitA modification was an amino‐terminal truncation due to proteolytic cleavage by a fungal protease that was subsequently named Bipolaris zeicola chitinase modifying protein (Bz‐cmp). Bz‐cmp was found to cleave ChitA, a two domain chitinase, in an inter‐domain region composed almost entirely of glycine residues.3 The reaction produces a 24 kDa chitinase protein from the C‐terminus and several small, cysteine‐rich chitin‐binding peptides derived from the N‐terminus. Analysis of these product peptides by MALDI‐TOF/MS showed that several different peptides are produced, resulting from cleavage at one of eight different peptide bonds within polyglycine4 (Fig. 1).

Figure 1.

Protease activity of PGHs. PGHs cleave the glycine rich linker region of maize chitinases ChitA and ChitB (middle). Conversion of full‐length ChitA to truncated protein is visualized by SDS‐PAGE gel analysis (top). This enables activity to be compared under standard conditions. Release of amino‐terminal peptide products is visualized by MALDI‐TOF/MS (bottom). This shows a series of ions separated by the mass of glycine (57 mass units). The data reveal which peptide bonds are cleaved, and the product ratios.

We have subsequently purified and characterized Es‐cmp, a second PGH from a related sorghum pathogen, Epicoccum sorghi (syn. Phoma sorghina),5 and have shown it to be a 640 amino acid glycoprotein with a catalytic region characteristic of serine‐type protease family S12. In addition to the catalytic region, PGHs have a larger, N‐terminal region that has an as yet unidentified function. Moreover, BLAST searches6 of fungal genomes7, 8 have identified additional PGH homologs in fungal plant pathogens in the Dothidiomycetes and the related class Sordariomycetes, and in the wood‐decaying mushroom Galerina marginata.5

Full length PGH homologs appear to be absent in the plant and animal kingdoms, but the general enzyme family, designated S12 proteins (also called penicillin‐recognizing proteins (PRP)s) exist in most organisms, including humans.9, 10 PRPs differ structurally from the alpha/beta hydrolase superfamily,11 and are identified by their conserved S‐X‐X‐K motif, which contains an essential catalytic serine. The type enzyme is a carboxypeptidase/transpeptidase involved in synthesis of bacterial cell walls12. This protease and related homologs are the targets of penicillin and related beta‐lactam antibiotics.13 Other types of PRPs include beta‐lactamases that cleave and inactivate beta‐lactam antibiotics,14 a peptidase that cleaves dipeptides of D amino acids,15 an esterase16 and D‐aminopeptidases.17, 18, 19, 20 Additionally, PRPs of unknown function are involved in production of the genotoxin colibactin and the antibiotic zwittermicin A.21, 22, 23 Despite the ubiquity of PRPs those from plant pathogenic fungi are the first reported as specific endopeptidases. The described fungal PGHs do not have beta‐lactamase or aminopeptidase activity and also are not inhibited by beta‐lactams or clavulinic acid (beta‐lactamase inhibitor),5 but this could be due to an inability of compounds to access the active site rather than an absence of affinity, as shown for esterase EstB.16

In the present work we have identified several determinants for the recognition and proteolytic cleavage of maize ChitA by Es‐cmp, a PGH from E. sorghi. Several peptides were synthesized that mimic the polyglycine‐rich region of ChitA, and their interactions with Es‐cmp has been evaluated by SDS‐PAGE and MALDI‐TOF/MS. We observed that certain Gly‐Gly peptide bonds were selectively hydrolyzed by Es‐cmp analogous to the cleavage of full length ChitA. Some peptides were also found that inhibit the Es‐cmp‐catalyzed truncation of ChitA. Our findings demonstrate that Es‐cmp recognizes four amino acid residues located outside of the polyglycine linker region of ChitA, and that this binding interaction may align the conserved S‐X‐X‐K active site motif for selective hydrolysis of specific Gly‐Gly peptide bonds within the linker domain.

Results

Polyglycine‐containing peptides as substrates

To determine the substrate specificity of the Es‐cmp protease we designed a series of synthetic peptides with overlapping coverage of the hydrolytic site on ChitA (Table 1). An initial set of eight peptides consisted of an N‐terminal pentaglycine motif and increasing chain length at the C‐terminus corresponding to the amino acid residues in ChitA. Incubation of these peptides with Es‐cmp was followed by MALDI‐TOF/MS. Noticeably, the 20‐mer G5‐20 (1739.80 Da, [M + H]⁺ = 1741.11) is readily cleaved in the polyglycine region to give ions at [M 2Gly]⁺ (m/z 1625.81) and M ‐ 4Gly (m/z 1511.89) with a minor cleavage occurring at M ‐ 5Gly (m/z 1454.8; Fig. 2). This peptide has the C‐terminal DAFF motif that includes adjacent Phe residues that are highly conserved in ChitA homologs. The 19‐mer peptide G5‐19 lacking the C‐terminal Phe is also cleaved by Es‐cmp but to a lesser degree. Hence, small ions are observed for M ‐ 2Gly and M ‐ 4Gly but with ∼5% of the activity observed on G5‐20. Peptides shorter than this at the C‐terminus, such as G5‐18 (Fig. 2), are apparently resistant to hydrolysis by Es‐cmp. The FF motif, therefore, is implicated in substrate binding by these experiments.

Table 1.

Peptides from Figure 2 and Figure 3. _→ indicates site of peptide bond cleavage. *% based on [M + H]⁺ peak areas

Peptide Name	Sequence	Products	%*	Calculated mass
G5‐20	GGGGGSGGANVANVVSDAFF	GGGG→GS.FF GG→GGGS.FF GGGGG→S.FF	71.3 21.9 6.8	1739.80
G5‐19	GGGGGSGGANVANVVSDAF	GGGG→GS.AF GG→GGGS.AF	75.2 24.8	1592.62
G5‐18	GGGGGSGGANVANVVSDA	None		1445.18
G4‐19	GGGGSGGANVANVVSDAFF	GGG→GS.FF		1682.75
G4‐20‐FD	GGGGSGGANVANVVSDAFFD	GGG→GS.FD		1797.84
SAR5‐20	GGGG[SAR]SGGANVANVVSDAFF	GGG→G[SAR].FF GG→GG[SAR].FF	68.8 31.2	1753.85
SAR3‐20	GG[SAR]GGSGGANVANVVSDAFF	None		1753.85
SAR2‐20	G[SAR]GGGSGGANVANVVSDAFF	G[SAR]GG→GS.FF		1753.85
SAR3,5‐20	GG[SAR]G[SAR]SGGANVANVVSDAFF	None		1767.88
T6‐20	GGGGGTGGANVANVVSDAFF	GGGG→GT.FF GG→GGGT.FF GGG→GGT.FF	63.4 20.4 16.2	1753.86
T16‐20	GGGGGSGGANVANVVTDAFF	GG→GGGS.FF GGGG→GS.FF	62.2 37.8	1753.86

Figure 2.

MALDI‐TOF/MS analysis of PGH peptide reactions. Each of the first five panels represents a different peptide, identified in the upper left hand corner. Sequences are found in Table I. Each peptide was incubated without enzyme (top half) or with Es‐cmp (bottom half). Peptide bond cleavage positions deduced from the data are represented as >. The bottom right panel shows products from PGH reactions with G5‐20 peptides that contain amino acid mutations at the terminal Phe; hence FF‐>FA, FF‐>FE, FF‐>FK, or FF‐>FW.

This is confirmed by the finding that the peptides G4‐19 and G4‐20‐FD, both of which contain the Phe‐Phe motif, are also cleaved by Es‐cmp (Fig. 2). The three N‐terminal glycines are removed from G4‐19 (1682.75 Da) by the Es‐cmp treatment to generate [M ‐ 3Gly + H]⁺ and [M ‐ 3Gly + Na]⁺ ions, corresponding to the M ‐ 4Gly cleavage of G5‐20 (Fig. 2). Similarly, treatment of G4‐20‐FD (m/z 1799.15) also results in the cleavage of three glycines to give [M – 3Gly + H]⁺ at m/z 1627.26. To confirm the importance of the terminal Phe, mutant G5‐20 peptides with other terminal amino acids were tested as substrates (Fig. 2, bottom right). Compared to G5‐20, all four peptides were worse substrates as indicated by reduced intensity of product peaks. This was especially evident when Ala was the terminal amino acid. Together these data show that Es‐cmp preferentially cleaves the last Gly‐Gly bond before the N‐terminal Ser, and that substrate binding is at least partially mediated by the C‐terminal Phe.

N‐Methylated Peptide Bonds and Ser‐to‐Thr Modifications Prevent Proteolysis by Es‐Cmp

In‐vitro polyglycine hydrolase activity of recombinant Es‐cmp and Bz‐cmp

Based on the above results a second set of 20‐mer peptides was designed with specific N‐terminal glycines replaced by sarcosine (Sar, N‐methylglycine). An N‐methyl peptide bond is therefore introduced at each Sar position (Table 1). The C‐terminal Phe‐Phe binding site is retained for all of these peptides. They differ from G5‐20 only in that each sarcosine introduces an additional methyl group. Noticeably, the activity of Es‐cmp is reduced on these peptides compared to the corresponding polyglycine series based on the end‐product intensities (Fig. 3). For example, SAR5‐20, with a sarcosine at the 5‐position, is cleaved at M‐2Gly and M‐3Gly, but product peak intensities are reduced compared to G5‐20 or G4‐19 substrates, and the N‐methylated peptide bond is not cleaved (Fig. 3). Introduction of Sar at position 3 (SAR3‐20) or two sarcosines (SAR3,5‐20) completely blocks the hydrolytic activity (Table 1), although SAR2‐20 loses the N‐terminal Gly‐Sar‐Gly‐Gly to generate a [M + H]⁺ ion at m/z 1512.12 (Fig. 3).

Figure 3.

MALDI‐TOF/MS analysis of PGH reactions on peptides with added methyl groups. Each panel represents a different peptide, identified in the upper left hand corner. Sequences are found in Table I and noted in the bottom half of the panel. Each peptide was incubated alone (top half) or with Es‐cmp (bottom half). Peptide bond cleavages deduced from the data are represented as >.

Two further synthetic peptides were tested with Thr replacing the Ser residues at position 6 (T6‐20) or position 16 (T16‐20). Like the sarcosine peptides, these peptides differ from G5‐20 only in that they have an additional methylene group, this time on the Thr sidechain. The T16‐20 peptide was expected to behave similar to G5‐20 because Thr and Ser are both present at this position in alloforms of ChitA that are susceptible to Es‐cmp activity. As expected, Es‐cmp showed similar activity with T16‐20 substrate as with G5‐20, to generate M ‐ 4Gly (m/z 1641.40) and M ‐ 2Gly (m/z 1526.67) ions (Fig. 3). Smaller ions resulting from cleavage between the fifth Gly and the first Ser was also observed. With T6‐20, however, the reaction rate is much slower as a result of the Ser‐to‐Thr replacement introduced adjacent to polyglycine.

Polyglycine‐containing peptides as protease inhibitors

The initial set of eight peptides, composed of an N‐terminal pentaglycine and increasing in chain length at the C‐terminus based on the sequence of ChitA, was tested to determine if any of the peptides inhibited Es‐cmp [Fig. 4(A)]. We reasoned that since some of the peptides were cleaved by Es‐cmp, they might also function as competitive inhibitors to prevent cleavage of ChitA substrate. As an initial test, peptides were added to reactions at a ratio of 1,000:1 (1 mM peptide and 1 micro M ChitA). Following incubation, reactions were analyzed by SDS‐PAGE and protein staining [Fig. 4(A)]. The three shortest peptides, G5‐13 and G5‐14 and G5‐15, did not inhibit Es‐cmp. Equal amounts of truncated ChitA were produced in these reactions compared to reactions without added peptide. Peptides of sixteen through nineteen amino acids (G5‐16 to G5‐19) mildly inhibited the reaction as indicated by less truncated ChitA products and more unreacted substrates. Reactions containing these four peptides showed similar amounts of protease activity [Fig. 4(A)], suggesting that a minimum length of sixteen amino acids is required for inhibition. The twenty amino‐acid peptide (G5‐20) was a more potent inhibitor. Most of the ChitA in this reaction remained full‐length while only a faint truncated product band was observed [Fig. 4(A)20]. Since G5‐20 inhibits more strongly than G5‐19 it suggests that the additional Phe residue enhances the interaction with the ES‐cmp protease.

Figure 4.

SDS‐PAGE analysis of peptide inhibitors. (A) Pentaglycine peptides with increasing length were added to Es‐cmp/ChitA reactions at 1 mM. Inhibition is indicated by an increase in the full‐length protein and a decrease in truncated protein, compared to a control (No peptide). (B) Same as above, but peptides contained either S6T mutation (left series) or G5[Sar] mutation (right series). (C) Summary comparison of peptide inhibitors. Each peptide was tested at 1 mM, 0.1 mM, and 0.01 mM. (D) Inhibition by G5‐20 peptides, at 1 mM, containing the indicated amino acid in place of the terminal Phe.

Because the above peptides functioned as protease inhibitors and because we previously observed that addition of N‐methyl groups to peptides made them resistant to proteolysis, we tested peptides with added N‐methyl groups to see if they also inhibited Es‐cmp. If these peptides did not inhibit it would suggest that these peptides do not bind Es‐cmp, explaining the lack of reaction products. If these peptides did inhibit it would suggest that these peptides specifically inhibit active site function. The results showed that neither the Ser to Thr modification nor the Gly to sarcosine change altered the ability of the synthetic peptide to function as an inhibitor [Fig. 4(B)]. In each case the peptides of fourteen and fifteen amino acids did not inhibit, while peptides of sixteen through nineteen amino acids where mildly inhibitory. And like with wild‐type peptides, reactions with either of the two twenty amino acid peptides had only faint product bands on SDS‐PAGE gels, indicating that they function as stronger inhibitors than the nineteen amino acid peptides due to addition of the second Phe. These results show that N‐methylation or the Ser‐Thr modification of the peptides does not inhibit binding to the protease, but disrupts cleavage of bound peptides. The observed equivalent inhibition for the sarcosine series and the Ser‐Thr mutant series compared to those in Figure 4(A), however, could result from a somewhat lower binding affinity combined with a lower cleavage rate.

To further compare how length of the polyglycine region may influence the degree of inhibition, we compared the longest peptide that did not inhibit (G5‐15) and the shortest peptide that showed inhibition (G5‐16) with the strongest inhibitor (G5‐20) at three different concentrations [Fig. 4(C)]. Additionally, we tested wild‐type (G6‐22) and serine‐to‐threonine mutant peptides with an additional amino acid on each end (G6‐22S6T). Results from the first three peptides at 1 mM concentration were identical to those above, showing no inhibition (G5‐15) some inhibition (G5‐16) or strong inhibition (G5‐20). None of these peptides inhibited at 0.1 mM concentration or 0.01 mM. Addition of amino acids to each end resulted in a substantial increase in inhibition [Fig. 4(C), G6‐22 and G6‐22S6T]. These peptides strongly inhibited at 0.1 mM, but not at 0.01 mM.

As a final test of polyglycine‐containing peptides as Es‐cmp inhibitors, we tested G5‐20 mutants with the final Phe mutated to either Trp, Lys, Glu, or Ala. When added to Es‐cmp and ChitA at a concentration of 1 mM, all four peptides inhibited less than G5‐20 [Fig. 4(D)]. The loss of inhibition was most significant when Lys was the terminal amino acid. This peptide did not inhibit Es‐cmp activity on substrate ChitA, while the other mutant peptides were weaker inhibitors than G5‐20. These results are consistent with those observed when these mutant peptides were tested as substrates (Fig. 1, lower right panel) and confirm the importance of the terminal Phe.

Polyglycine‐free peptides as protease inhibitors

We next tested an overlapping peptide library to determine if peptides could inhibit without the amino‐terminal glycines. This library was comprised of eleven 20 amino acid peptides based on G5‐20. Each peptide in the library differed from its predecessor by having one amino acid removed from the amino terminus and one amino acid added to the carboxy terminus [Fig. 5(A)]. To determine if any of these peptides can be cleaved by Es‐cmp, they were each tested in protease assays. Only the initial G5‐20 peptide gave products after prolonged incubation (not shown). As the first peptide, G5‐20, was already known to inhibit strongly at 1 mM, these peptides were tested at a lower concentration (0.25 mM) in assays with ChitA. At this concentration, G5‐20 strongly inhibited the reaction, with only a faint product band visible on the SDS‐PAGE gels [Fig. 5(A), lane 1]. The next three peptides in the series gave increasingly weaker inhibition, visualized as progressively less ChitA substrate and more truncated product in these reactions (lanes 2‐4). The next peptide, G1‐20, was a potent inhibitor, completely inhibiting the reaction (lane 5). The subsequent three peptides showed increasingly less inhibition (lanes 6‐8) followed by another strong inhibitor, G(3)‐20 (lane 9). The last two peptides were less effective (lanes 10, 11). Overall the results show a periodicity of four, with every fourth peptide functioning as a strong inhibitor of the proteolytic activity. As G5‐20, G1‐20, and G(3)‐20 all show equivalent inhibition, the benefit of adding the Lys and Gly residues appears to be offset by the loss of residues on the amino‐terminal side. Strong inhibition exhibited by peptides in lanes 5 and 9 demonstrate polyglycine is not required for inhibition, but that it is the adjacent downstream sequence that is important. They also show that addition of Lys four amino acids after Phe and Gly another four residues further away increases the strength of inhibition relative to the previous peptide. While the inhibition observed for G5‐20 demonstrates that these amino acids are not required, they may be important if the peptide extends past the FF sequence. These results combined with those from the previous section suggest the importance of a series of four amino acids, Ser, Phe, Lys, and Gly, in the interaction with Es‐cmp.

Figure 5.

SDS‐PAGE analysis of overlapping peptide library. (A) Each peptide (right) was added to Es‐cmp/ChitA reactions at 0.25 mM. Inhibition is indicated by an increase in full‐length protein and a decrease in trucated protein when compared to a control reaction without peptide (No peptide). Strong inhibitors occur with a periodicity of four (1, 5, 9). (B) Inhibitor activity of G(3)‐20 mutant peptides. A total of sixteen peptides, each with a single amino acid change, were tested as inhibitors of Es‐cmp activity on ChitA when added at 0.25 mM. On the top gel, mutations occur at the indicated Phe (left) or Ser (right). On the bottom gel, mutations occur at the indicated Lys (left) or Gly (right). Each gel contains control reactions with G(3)‐20 (leftmost lane), and without peptide (next to last lane) for comparison.

The importance of the identified amino acids was tested by performing inhibitor assays with sixteen mutants of G(3)‐20. Each mutant peptide contained a single amino acid change at one of the four positions [Fig. 5(B)]. All four peptides with mutations to the Phe residue were weaker inhibitors than G(3)‐20. These results are similar to those found for the same mutations when tested in the G5‐20 background [Fig. 3(D)]. The biggest difference is that in the previous assays the conservative substitution Trp had a milder effect than observed in the G(3)‐20 background. Similar results were obtained with G(3)‐20 mutants with mutations to the Ser residue. Three of the four mutants were weaker inhibitors with only the conservative Ser to Ala change tolerated. G(3)‐20 mutants with changes at the identified Lys were all weaker inhibitors [Fig. 5B, bottom gel], with two of the four peptides not showing any inhibitor activity (Arg and Gln). The final four mutants, with changes at the terminal Gly, each had a different effect on inhibition. The Ala mutant did not change the inhibition activity of the peptide, the Ser mutant reduced inhibition moderately, the Asp mutant did not inhibit, and the His mutant inhibited more strongly than G(3)‐20. Of the sixteen mutant peptides, thirteen resulted in weakening or loss of inhibition activity, one improved inhibition activity, and only two showed no change, indicating that these four amino acids are important for binding of G(3)‐20 to Es‐cmp.

Discussion

The Es‐cmp endoprotease specificity resulting from active site interactions with the polyglycine linker domain of ChitA substrate are revealed by analyzing its cleavage products. Previous MALDI‐TOF/MS analysis of peptides cleaved from ChitA by Es‐cmp suggested that glycines were selected for on both sides of the cleaved bond. Product peptides resulted when at least three glycines were adjacent to the cleaved peptide bond on both sides.4 But analysis of peptides cleaved from ChitB, an isoform in which polyglycine is interrupted by two consecutive serines, showed that only the amino acids on each side of the cleaved peptide bond were important. On the amino side the selectivity was absolute—glycine was the terminal amino acid on all product peptides. On the carboxy side glycine was almost always present but rare cleavage of a Gly‐Ser bond was also observed.4

Analysis of Es‐cmp products from peptide substrates in the present work solidifies this model of limited active site selectivity. The first non‐glycine amino acid in pentaglycine peptides, a serine, was usually two or four amino acids on the carboxy side of the cleaved bond, as indicated by loss of four or two glycines from pentaglycine peptides (Fig. 2). As with ChitB, some cleavage of a Gly‐Ser bond was also observed—G5‐20 and T16‐20 products with all five glycines removed were detected (Fig. 2, G5‐20; Fig. 3, T16‐20). Endoproteases that exhibit active site specificity for only one substrate amino acid, usually the one immediately on the amino side of the targeted bond, are the most common type.24

Addition of fourteen amino acids to the carboxy side of pentaglycine was necessary to observe minimal proteolysis (Fig. 3, G5‐19). Addition of one more amino acid, Phe 20, led to a substantial increase in product formation (Fig. 3, G5‐20). In inhibition assays, Phe 20 was again shown to be important for increased inhibition (Fig. 4). The inhibitor results also suggest the importance of Ser 16, which was needed for minimal inhibition (Fig. 4). When an overlapping peptide library was tested as potential PGH inhibitors, two additional amino acid residues, Lys 24 and Gly 28, were found to increase inhibition (Fig. 5). When combined, the substrate and inhibitor data implicate four amino acids in Es‐cmp recognition of ChitA: Ser 16, Phe 20, Lys 24, and Gly 28. When modeled on a structure of truncated ChitA25 (PDB = 4MCK) these four amino acids form one face containing an alpha helix (Fig. 6). We propose that these four amino acids interact with Es‐cmp to bring the polyglycine linker in proximity to the active site, allowing cleavage of one of the Gly‐Gly peptide bonds through limited active site selectivity.

Figure 6.

Structure of the Es‐cmp binding site. The four amino acids, identified as important for protease recognition in this study, were modeled onto the truncated ChitA structure (red cartoon) as blue sticks.

But this interaction near the active site combined with limited active site selectivity does not fully explain the specificity of polyglycine hydrolases. Multiple lines of evidence support the existence of at least one additional interaction between PGHs and substrate chitinases. The first is that peptides only inhibited Es‐cmp when added in 100 to 1000‐fold excess of substrate ChitA (Figs. 4 and 5). When added at a 10‐fold excess, even the best inhibitors failed to affect the reaction (Fig. 4C). Additional evidence of unknown interactions comes from a previous study. When ChitA was pre‐cleaved by Fv‐cmp, a protease from Fusarium verticillioides that cleaves the amino‐terminal domain leaving only eight amino acids on the amino‐terminal side of the first linker glycine, Es‐cmp did not cleave polyglycine.4 This indicated that part of PGH‐ChitA recognition involves the ChitA N‐terminal domain. Furthermore, the catalytic S12 protease region makes up less than half of PGH proteins. The larger amino‐terminal region is novel and has no predicted function. It is tempting to assume that this larger region of PGHs is involved in specific protein‐protein interactions that are involved in recognition of ChitA, possibly through interactions with the ChitA amino‐terminal domain.

In summary, we have shown that synthetic peptides modeled after ChitA are recognized by PGHs. These peptides are susceptible to cleavage and also can inhibit protease activity on ChitA, presumably through competitive binding. Four neighboring amino acids, with a periodicity of four in the primary sequence, form a recognition surface that enables binding of the adjacent polyglycine into the active site for cleavage of a Gly‐Gly peptide bond. The results suggest that peptide analogs of ChitA could be used to inhibit PGH activity, strengthening plant defenses against fungal pathogens.

Materials and Methods

Peptide synthesis

Synthesis of peptides was carried out using well‐established protocols for Fmoc solid phase peptide synthesis, with HBTU, HoBt and DIEA as coupling reagents.26 DMF (Fisher Scientific, Waltham, MA) was used as the synthesis solvent. A rink amide low‐loading resin (Aapptec, Louisville, KY) was used and syntheses were executed on a 0.15 mmol scale. Peptides were cleaved and deprotected with a mixture of TFA/triisopropanolsilane/water/anisole (18:0.7:0.7:0.7 by volume) for 3‐4 hours. The cleaved mixture was evaporated using a rotary evaporator until a small volume was obtained, and the residual TFA solution was precipitated with cold (–20^°C) diethyl ether (Fisher Scientific, Waltham, MA) and dried under vacuum for 24 hours before further purification. Crude peptides were characterized by MALDI‐TOF/MS as described previously.4

Peptide purification

Peptides were partially purified via dialysis. Crude peptides were dissolved at 10 mg/mL (or less, depending on solubility limit) in MilliQ water, followed by pH adjustment with NaOH to achieve a pH of 7. The peptide solution was then dialyzed through a SpectraPor dialysis membrane in MilliQ water for 48‐72 hours. During this time the water was changed 2 times per day. The solution containing the peptide was then frozen and lyophilized to collect the pure peptide.

Protein purification

Recombinant Es‐cmp protease and ChitA chitinase substrate were produced by heterologous strains of Pichia pastoris and purified from expression cultures as described previously.3, 5

Peptide substrate assays

Each peptide (1 mM) was combined with Es‐cmp (200 nM) in sodium acetate buffer (10 mM, pH 5.2). Reactions were incubated for 16 hours at 25^°C. After incubation, reactions were combined with an equal volume of matrix (saturated 2,5‐dihydroxybenzoic acid in acetonitrile) and spotted onto the Matrix‐assisted laser desorption/ionization (MALDI) MS target. The instrument was a Bruker‐Daltonics Microflex LRF (Bruker‐Daltonics, Billerica, MA) with a pulsed N2 laser (3000 shots, 337 Hz, 60 Hz pulse), and reflectron aquisition.

Peptide inhibitor assays

Peptides were combined with ChitA (1 µM) in sodium acetate buffer (10 mM, pH 5.2). Reactions were initiated by addition of protease (30 pM). Reactions were incubated for 1 hour at 30^°C and stopped by adding SDS‐PAGE loading buffer (0.001\% bromophenol blue, 6\% glycerol, 2\% sodium dodecyl sulfate, 50 mM Tris‐Cl, pH 6.8, 100 mM DTT final) followed by incubation in boiling water for 1 minute. Reactions were analyzed by SDS‐PAGE using a Criterion Vertical Electrophoresis Cell, 12.5% Criterion Tris‐HCL gels (18 well, 30 µL). Gels were run for 1 hour at 200 V, stained with Oriole florescent protein stain for 90 minutes, and photographed with a Gel Doc EZ System (Bio‐Rad, Hercules CA).

Conflict of Interest

The authors would like to state that they have no conflict of interest regarding this manuscript.