INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Proteases play a central role in the mobilization of energy reserves in the germinating barley (Hordeum vulgare) grain, providing the embryo with a supply of amino nitrogen in the form of free amino acids and short peptides. Multiple representatives of the Ser protease, Asp protease, metalloprotease, and Cys protease families have been detected in the germinating barley grain, where the Cys proteases are the most abundant (Zhang and Jones, 1995). Two Cys endoproteases (EPs) EP-A and EP-B are expressed in the scutellum epithelium and aleurone tissue of the germinating grain and are secreted into the starchy endosperm (Marttila et al., 1993). Enzymatic studies of purified EP-A and EP-B showing in vitro digestion of the barley hordein polypeptides provide evidence for their role in storage protein degradation during germination (Koehler and Ho, 1990). The hordein polypeptides, comprising sulfur-poor and sulfur-rich families, share the distinctive properties of solubility in alcohol-water mixtures and domains of Pro/Gln-rich repeated sequences (Shewry and Tatham, 1990). Detailed characterization of EP-A and EP-B has largely been based on the cleavage of non-native substrates (Jones and Poulle, 1990; Zhang and Jones, 1996), which may not provide a direct insight into the specific enzymatic properties associated with degradation of the hordein polypeptides.

As a first step in dissecting the proteolytic events within the germinating grain, we studied the proteolytic cleavage of a single recombinantly expressed member of the sulfur-poor C hordein polypeptide family by EP-A and EP-B (Davy et al., 1998). The two EPs showed cleavage site specificity, with a strong preference for Phe, Leu, or Val in the P₂ position. (The substrate positions are denoted P_i,... . ,P₂, P₁, P₁',P₂',... . ,P_j', in correspondence with the binding subsites S₁, etc., according to Schechter and Berger [1967].)

Conformational studies of members of the hordein polypeptides and the closely related wheat prolamins indicate that intermolecular disulfide bonds between sulfur-rich hordein polypeptides and extensive hydrogen bonding between the repetitive domains of the sulfur-poor hordeins are involved in their assembly into protein bodies in the endosperm vacuole of the developing grain (Tatham and Shewry, 1995; Shimoni and Galili, 1996). Studies in wheat suggest that a NADP/thioredoxin system together with specific proteases are required for the mobilization of the prolamin storage proteins (Kobrehel et al., 1992). We have now investigated the ability of EP-B to hydrolyze the hordein polypeptides in their native state in barley protein bodies and the requirement for a reducing agent.

Synthetic fluorogenic peptides, based on native substrate sequences, provide a powerful tool to examine the cleavage site specificity of EP-B and EP-A. We have exploited this technology by synthesizing a series of peptide substrates to examine systematically the substrate preference of barley Cys EPs at the P₂ and P₁ sites and compare them with papain, a well-characterized plant Cys EP. We then examined the ability of synthetic substrate cleavage rates to predict the cleavage of wild-type and mutated C hordein polypeptides. Since barley Cys EPs are believed to mediate the release of -amylase from the bound form during germination, we have selected this enzyme as an additional native substrate. Our investigations reveal that the amino acid residues at the P₂ and P₁ sites define the cleavage sites of barley Cys EPs and that elimination of these sites from a protein substrate can drastically reduce its rate of hydrolysis.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Purification of EP-A and EP-B

EP-A and EP-B were purified from air-dried (45°C) green malt, provided by Carlsberg A/S, Copenhagen, and titrated as previously described (Davy et al., 1998). Papain (twice recrystallized), Brij 35, and buffers were purchased from Sigma (St. Louis). The purity of papain was established by SDS-PAGE and N-terminal amino acid sequencing, which revealed a single protein species.

Protein Body Isolation and Degradation by EP-B

Protein bodies were isolated from developing endosperm from 10 barley (Hordeum vulgare cv Alexis) spikes harvested 20 DPA using a modification of the published method (Cameron-Mills, 1980). The homogenized endosperms were resuspended in 20 mL of buffer A (20 mM 4-[2-hydroxyethyl]-1-piperazineethanesulfonic acid [HEPES], pH 7.6, 100 mM NaOAc, and 5 mM MgCl₂). The suspension was layered onto a 5-mL 1.75 M Suc cushion made up in buffer A and centrifuged for 2 min at 500g. Material banding on top of the Suc cushion was collected and resuspended in buffer B (20 mM HEPES, pH 7.6, 100 mM NaOAc, 5 mM EDTA, and 0.25 M Suc). This suspension was loaded onto a Percoll step gradient in a 10-mL Corex tube (2 mL each, 1.08 and 1.13 g/m³) and centrifuged at 9,000 rpm in an SS34 rotor. The protein bodies banding on top of the 1.13 g/m³ layer were collected in 250 µL and stored at 20°C.

Protein bodies were diluted 10 times in water, and 6 µL of the diluted suspension was incubated at 25°C with 10 µL of assay buffer (50 mM NaOAc, pH 4.5, and 0.05% Brij 35), 4 µL of EP-B 44 nM solution, and different concentrations of -mercaptoethanol. A 10-µL aliquot was taken after 2 and 24 h and immediately mixed with 3 µL of sample buffer containing SDS kept at 80°C on a heating block and heated for 3 min. The samples were run on 12% acrylamide Tris-Gly gels under reducing conditions, according to the method of Laemmli (1970) on a mini-gel apparatus (Protean II, Bio-Rad Laboratories, Hercules, CA) and stained with Coomassie Brilliant Blue G250 according to the method of Neuhoff et al. (1988).

Synthesis of Internally Quenched Fluorogenic Substrates

Fluorogenic substrates were of the general formula Abz-(Xaa)_n-Tyr(NO₂)-Asp-OH, where Abz (2-aminobenzoyl) is the fluorescent group, 3-nitrotyrosine (Tyr[NO₂]) is the quencher, Xaa is any of the genetically encoded amino acids, and n = 4 to 8 (Meldal and Breddam, 1991). Peptides were synthesized on a solid support (Pega₁₉₀₀) by manual multiple column peptide synthesis (Meldal et al., 1993) in a 20-column library synthesizer (Meldal, 1994). Pega₁₉₀₀ resin was derivatized with a base-labile hydroxymethyl benzoic acid linker using O-benzotriazol-1-yl-N,N,N', IN'-tetramethyl uronium tetrafluoroborate (TBTU) (Knorr et al., 1989).

The C-terminal residue 9-fluorenylmethyloxycarbonyl (Fmoc)-Asp(t-butyl)-OH was attached to the resin by esterification with 1-mesithylenesulfonyl-3-nitro-1,2,4-triazol and N-methylimidazol (Blankemeyer-Menge et al., 1990). The resin was washed with dimethylformamide (DMF) and deprotected with 20% (v/v) piperidine in DMF as previously described (Meldal et al., 1993). The columns were washed with DMF, and Fmoc-Tyr(NO₂)-OH was pre-activated with TBTU, and 4-ethyl morphine for 5 min before addition to each column. The peptide was assembled using Fmoc-Xaa-pentafluorylphenyl esters (3 equivalents) with 3,4-dihydro-4-oxo-1,2,3,-benzotriazo-3-yl-OH catalyst (1 equivalent), and 20% (v/v) piperidine was used for deprotection. The amino acid side chain protections used were trityl (Asn, Cys, Gln, and His), t-butyl (Asp, Glu, Ser, Thr, and Tyr), t-butyloxycarbonyl (Lys), and 2,2,5,7,8-pentamethylchroman-6-sulfonyl (Arg). The last residue added was t-butyloxycarbonyl-Abz-O-3,4-dihydro-4-oxo-1,2,3,-benzotriazo-3-yl (3 equivalents).

The resins were washed with DMF, deprotected, washed with DMF and dichloromethane, and dried by airflow. The protection groups were cleaved off using a mixture containing 87.5% (v/v) trifluoroacetic acid, 2.5% (v/v) ethanedithiol, 5% (v/v) thioanisole, and 5% (v/v) water, and the resins were washed with 95% (v/v) acetic acid, 5% (v/v) diisopropylethylamine in DMF, DMF, and finally dichloromethane. The resins were dried by airflow, and the peptides were cleaved off with 0.1 M NaOH for 2 h and washed out with water. The solutions were neutralized with 0.1 M HCl and lyophilized. The peptides were dissolved in DMF and the insoluble sodium chloride was removed. Purity and identity were confirmed by HPLC, amino acid analysis, and matrix-assisted laser-desorption ionization time of flight mass spectroscopy.

Determination of Enzyme Activity and Kinetic Constants

The enzymatic hydrolysis of the peptide substrates was followed on a luminescence spectrometer (LS 50, Perkin-Elmer, Foster City, CA). The assay, titration, and determination of kinetic constants for EP-A, EP-B, and papain were performed as previously described (Davy et al., 1998). The activity of EP-A and EP-B was assayed with Abz-Phe-Arg-Gln-Gln-Tyr(NO₂)-Asp from pH 2.5 to 7.5 in 100 mM citrate-phosphate containing 2 mM Cys and 2 mM -mercaptoethanol with 5 µL of EP-B (0.24 µM) or 5 µL of EP-A (0.135 µM).

Heterologous Expression and Purification of -Amylase

The entire coding sequence of barley -amylase (accession no. X52321) was amplified by PCR from the plasmid pCC51 (Kreis et al., 1987) with tailed primers (5'-GACGACGACAAGATGGAGGTGAACGTG-3', 5'-GAGGAGAAGCCCGGTTTACATGGTGGC-AGG-3'). The amplified sequence was purified and inserted in-frame into the expression vector pET-32 LIC (Novagen, Madison, WI) by ligation-independent cloning according to the supplier's recommendations. The resulting plasmid (pET32-Amy) contains an open reading frame of 2,079 bp, encoding a fusion protein of 77 kD comprising the 60-kD -amylase fused with an N-terminal 17-kD tag (thioredoxin·Tag, His·Tag, and S·Tag), which facilitates solubilization and purification of the expressed protein.

pET32-Amy was transformed into BL21 (DE3); 0.5-L cultures (Luria-Bertani + 100 mg/L ampicillin) were started from single colonies, grown with vigorous agitation at 37°C until OD₆₀₀ = 0.7, and subsequently transferred to 20°C upon overnight induction of expression with 1 mM isopropylthiogalactoside. The tagged -amylase was purified under native conditions by metal-affinity chromatography using nickel-nitrilotriacetic acid (Ni-NTA) agarose (Qiagen, Valencia, CA) as recommended by the manufacturer. Cells were resuspended in 50 mM phosphate buffer, pH 8.0, 300 mM NaCl, lysed by sonication (3 × 30 s), and the supernatant was loaded onto the Ni-NTA matrix after centrifugation at 10,000g for 20 min. The column was washed with approximately 10 volumes of the loading buffer, followed by 50 mM phosphate buffer, pH 6.0, and 300 mM NaCl until the OD₂₈₀ of the eluate was below 0.1. The -amylase was eluted with 50 mM NaOAc buffer, pH 5.2, and 300 mM NaCl, and the purified fusion protein was treated with enterokinase using the S Tag rEK Purification Kit (Novagen) to remove the N-terminal tag. The purified protein was stored at 20°C following the addition of glycerol to a final concentration of 10%. -amylase activity was assayed using the Betamyl kit (Megazyme International, Bray, County Wicklow, Ireland) according to the manufacturer's instructions.

Effect of EP-B on -Amylase

Four micrograms of -amylase in 20 µL of 150 mM NaCl, 20 mM Tris-HCl, pH 7.5, buffer was incubated overnight with 10 µL of EP-B assay buffer and 20 µL of 44 nM EP-B at 4°C. The samples (10 µL of tag--amylase, -amylase, and -amylase treated with EP-B) were run on 12% acrylamide high-Tris gels under reducing conditions according to the method of Fling and Gregerson (1986). Twenty microliters of -amylase treated with EP-B was subjected to N-terminal sequencing, as previously described (Davy et al., 1998).

Site-Directed Mutagenesis of a C Hordein Gene

Mutations were introduced into a C hordein genomic sequence using an in vitro mutagenesis system (Altered Sites II, Promega, Madison, WI). The C hordein coding sequence in -hor1-17 (Entwistle et al., 1991), cloned in a pET-3d expression vector (Tamas et al., 1994), was excised with BamHI and XbaI restriction enzymes and cloned into the polylinker of the plasmid pALTER-1. The following five antisense mutagenic oligonucleotides, with mutated nucleotides shown in italics and bold, were designed to anneal to the C hordein coding sequence introducing mutations into the encoded protein: 1,5'-GGTGATTGCGACTCTTGGC-3'; 2,5'-TATGGCTGCTGCGAATATGATTGTTGT-3'; 3,5'-ACGGTTGCTGCGATGGAAATGGTTT-3'; 4,5'-GGTTGCTGGGAAATTATTTG-3'; 5,5'-TGTTGGCGGGATGGTTGTTG-3'. Mutant plasmids transformed into JM101 were purified and their nucleotide sequence determined by dideoxy chain termination using the Ampli Cycle Sequencing kit (Perkin-Elmer) and a sequencer (model 373A, Applied Biosystems, Foster City, CA).

Following successive rounds of mutation, a mutated C hordein gene in pALTER-1, encoding a C hordein polypeptide with 12 amino acid substitutions, was selected for expression studies. The mutant C hordein gene was amplified by PCR and cloned into the pET-32 LIC expression vector supplied by Novagen, using the following tailed sense and antisense primers: pET32Chor-s 5'-GACGACGACAAGATGAGGCAACTAAACCCT-3' and pET32Chor-as 5'-GAGGAGAAGCCCGGTCTAGACCATACTCCA-3'. The sequence of the mutant C hordein gene in pET32 LIC was confirmed by sequencing.

C Hordein Expression and Purification

Wild-type C hordein was expressed and purified as previously described (Davy et al., 1998). The mutated C hordein coding sequence cloned in the pET 32 expression vector, was expressed in BL21 (DE3) host cells as a fusion protein having N-terminal His-tag and S-tag domains. Five-hundred milliliters of Luria-Bertani medium containing 100 µg/mL ampicillin was inoculated with a single colony of transformed BL21 cells and incubated at 37°C; when the OD₆₀₀ reached 0.6, cells were induced with isopropylthiogalactoside at 0.4 mM and incubated for 4 h at 37°C. The cells were harvested by centrifugation at 4,000g for 20 min and resuspended in lysis buffer (50 mM NaH₂PO₄, pH 8.0, 300 mM NaCl, and 10 mM imidazole). After incubation with lysozyme (1 mg/mL) on ice for 30 min, the cells were disrupted with a sonicator and centrifuged at 10,000g for 30 min at 4°C. The supernatant was loaded onto Ni-NTA agarose (Qiagen) and eluted under native conditions according to the supplier's instructions. The eluted protein was further purified with the S-Tag rEK Purification Kit (Novagen). The bound C hordein was eluted from the S-Tag agarose column by cleavage of the S-Tag with enterokinase, according to the supplier. The purified mutated C hordein was freeze-dried and stored at 20°C until used.

Assay of C-Hordein and Mutant C-Hordein Degradation by EP-B

Purified wild-type C hordein was dissolved in EP-B assay buffer to a concentration of 1 µg/µL. Twenty-seven microliters of C hordein solution was incubated with 115 µL of assay buffer and 2 µL of 0.27 µM EP-B solution at 40°C. Purified mutant C hordein was dissolved in assay buffer to 0.2 µg/µL and 90 µL of the solution was incubated with 4.6 µL of assay buffer and 1.3 µL of 0.27 µM EP-B solution at 40°C. To follow the degradation process, 16-µL aliquots were taken at different time points and immediately mixed with 4 µL of sample buffer containing SDS kept at 80°C on a heating block and incubated for 3 min. The mutant C hordein degradation was repeated with the same amount of substrate and 15 times more EP-B. The samples were run on 16% (w/v) acrylamide high-Tris gels under reducing conditions according to the method of Fling and Gregerson (1986), and stained with Coomassie Brilliant Blue G250 according to the method of Neuhoff et al. (1988). The level of C hordein was quantified by photographing the gels with a digital camera and using ImageQuant software to integrate pixel densities.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

EP-B Hydrolysis of Native Hordein Polypeptides

Protein bodies were isolated from developing barley endosperm, following homogenization of the tissue and centrifugation on Suc and Percoll step gradients. The prolamin fraction of barley, comprising the alcohol-soluble storage proteins of the grain, is composed of four polypeptide classes localized in the protein bodies. SDS-PAGE of the protein body polypeptides (Fig. 1) revealed sulfur-rich B and -hordeins with M_rs of 36,000 to 42,000, sulfur-poor C hordeins with M_rs of 45,000 to 70,000, and HMW D hordein with a M_r of 90,000. 

View larger version (70K): [in this window] [in a new window]   Figure 1.   SDS-PAGE showing degradation of hordein in isolated protein bodies by purified EP-B. Samples were incubated at 20°C for 2 or 24 h, with varying amounts of added -mercaptoethanol (-ME), the concentration of which (in mM) is shown above each lane. Protein body controls (PB) are with 0 and 80 mM added -mercaptoethanol after 24 h of incubation without EP-B. Molecular mass markers (M) are in kD. The positions of B, C, and D hordeins are indicated on the right.

The polypeptide composition of the protein bodies following a 2-h incubation with purified EP-B showed a rapid hydrolysis of D hordein. The degradation of C, B, and -hordein was strongly enhanced by the addition of the reducing agent -mercaptoethanol at concentrations of 10 mM and above. However, all of the hordein polypeptides in the protein body were accessible to EP-B degradation in the absence of disulfide reduction, judging by the appearance of low-molecular-mass peptides of between 20 and 30 kD after a 2-h incubation and almost complete hordein degradation after 24 h. Two minor polypeptides were detected in the protein body preparation after the 24-h incubation, suggesting that they are poor substrates for EP-B hydrolysis.

Systematic Study of Cys EP Substrate Specificity Using Synthetic Substrates

A series of synthetic fluorogenic substrates were synthesized to make a systematic analysis of the importance of the residue at the P₂ and P₁ positions. Substrates with 19 different amino acid residues at the P₂ position were synthesized with the general sequence Abz-Xaa-Arg-Gln-Gln-Tyr(NO₂)-Asp. Previous studies had shown that this sequence with Phe in the P₂ position, based on an EP-B cleavage site in C hordein, was the best synthetic substrate for barley EPs (Davy et al., 1998). Peptide cleavage by the Cys EPs EP-A, EP-B, and papain was assayed at three substrate concentrations under pseudo-first-order conditions (K_m), and the concentration of active enzyme was determined by titration with trans-epoxysuccinyl-L-leucylamido-(4-guanido)butane. The pH optimum for EP-A and EP-B was 4.0 (Fig. 2), but assays were performed at pH 4.5 to reduce protonation of the amino group of Abz, which decreases its fluorescence (Meldal and Breddam, 1991). The derived second-order kinetic constants k_cat/K_m for each substrate are shown in Table I.

View larger version (18K): [in this window] [in a new window]   Figure 2.   pH optima for EP-A and EP-B were determined by measuring the rate of hydrolysis of Abz-Phe-Arg-Gln-Gln-Tyr(NO2)-Asp from pH 2.5 to 7.5 in 100 mM citrate-phosphate containing 2 mM Cys and 2 mM -mercaptoethanol. , EP-B; , EP-A.

The order of preference for the amino acid residue at the P₂ site in the substrate was very similar for the two barley EPs, EP-A and EP-B, where neutral amino acids with large aliphatic and nonpolar (Leu, Val, Ile, and Met) or aromatic (Phe, Tyr, and Trp) side chains were present in substrates giving the highest k_cat/K_m values. Substrates with small, polar, or charged side chains in the P₂ position had the lowest k_cat/K_m values. While the substrate preference of papain at the P₂ site largely reflected that of EP-A and EP-B, the range of k_cat/K_m values displayed by papain was significantly smaller than that of the barley EPs. The preference of papain for Phe or Tyr at P₂ was much greater than for other aromatic or nonpolar residues. It was not possible to test Pro in the P₂ position of this substrate due to quenching of Abz fluorescence when Pro is immediately adjacent (Ito et al., 1998). Another substrate (Abz-Gln-Pro-Gln-Gln-Pro-Tyr(NO₂)-Asp had previously been synthesized and was a very poor substrate for EP-B (Davy et al., 1998).

A further 20 substrates were synthesized with amino acid substitutions at the P₁ position of the general sequence Abz-Leu-Xaa-Gln-Pro-Tyr(NO₂)-Asp, based on the best substrate from the P₂ series. A Pro residue (instead of Gln) was placed at P₂' to direct cleavage between Xaa and Gln, since it was known that cleavage would not occur with Pro at P₁' (Davy et al., 1998). The cleavage site of the substrates was confirmed by N-terminal sequencing. The k_cat/K_m values of the different substrates for EP-A and EP-B were not strongly influenced by the residue at the P₁ position, with the exception of Pro, where cleavage was not detectable (Table II), and the range of k_cat/K_m values was much less than found for the P₂ substrate series. The best substrates for EP-B contained nonpolar or charged residues at P₁ (Arg, Met, Leu, Phe, Tyr, Gln, Thr, Lys, and Glu), while those with small residues (Ala, Gly, Ser, and Cys) were slightly poorer, and Asn, Asp, Val, and Ile gave the lowest k_cat/K_m values.

The substrate preference profile for EP-A at P₁ was similar to that of EP-B, except that there was a clear preference for positively charged residues (Lys and Arg). For four residues (Lys, Thr, Tyr, and Leu), K_m values were so low that k_cat/K_m values were calculated from Hanes-Woolf plots (Table III). In general, k_cat/K_m values for a particular substrate were lower for EP-A compared with EP-B, particularly for Ala, Cys, Glu, and Ile. In contrast, papain showed a preference for Gly at P₁, followed by Met, Cys, and Arg. When the k_cat/K_m values for EP-A and EP-B were expressed as a ratio of the corresponding values for papain, substrates with Asp, Glu, and Tyr at P₁ were seen to be significantly better substrates for EP-A and EP-B than papain (Table II). Similarly, substrates with Leu at P₂ were better substrates for EP-A and EP-B than papain (Table I).

The substrate preferences seen from the results with synthetic peptides, particularly at P₂, should provide a basis for predicting cleavage sites in native protein substrates. This concept was tested with -amylase and site-specific mutants of C hordein.

C-Terminal Processing of Barley -Amylase by EP-B

The C-terminal cleavage by barley EPs, which releases barley -amylase from a bound form, was investigated with expressed recombinant barley -amylase and EP-B. The barley -amylase coding sequence, cloned in the plasmid pET32-Amy, was expressed as a 77-kD fusion protein, with an N-terminal tag of 17 kD fused to the 60-kD -amylase to facilitate purification.

The expressed -amylase fusion protein (approximately 0.8 µg) was analyzed by SDS-PAGE (Fig. 3A) following purification (lane 1), N-terminal tag cleavage with enterokinase (lane 2), and incubation with EP-B for 16 h at 4°C (lane 3). Following enterokinase treatment, the -amylase migrated as a single 60-kD polypeptide, indicating complete cleavage of the fusion protein. N-terminal amino acid sequencing confirmed the correct N-terminal sequence of the 60-kD -amylase: MEVN (Fig. 3B). The 60-kD -amylase was further cleaved by EP-B, giving polypeptides of approximately 56 to 58 kD. N-terminal amino acid sequence analysis of the peptides released by EP-B digestion revealed that the EP-B cleaved the sequence MGGQ at three sites located near the C terminus of the 60-kD -amylase polypeptide (Fig. 3B), accounting for the digestion products seen by SDS-PAGE. The absence of additional peptide sequences indicates that only the C-terminal domain of the native 60-kD -amylase is accessible to proteolytic attack, and that the three sites cleaved by EP-B are a reflection of EP-B substrate specificity.

View larger version (40K): [in this window] [in a new window]   Figure 3.   A, SDS-PAGE of N-tag--amylase, -amylase, and -amylase treated with EP-B. N-Tag--amylase was expressed and purified from E. coli (lane 1), and the N-terminal tag was cleaved with enterokinase (lane 2) and then incubated with purified EP-B (lane 3). B, Partial sequence of -amylase formatted to emphasize the 11-residue domain, repeated four times in the C-terminal 45 residue extension, showing cleavage sites of enterokinase () and EP-B (). The latter were determined by amino acid sequencing of the products released from -amylase after incubation with EP-B.

Mutant C Hordein Polypeptide with Reduced Susceptibility to EP-B Degradation

The expression of a C hordein polypeptide in Escherichia coli and its purification and folding into a native conformation enables the proteolytic cleavage of a single hordein substrate to be analyzed. EP-B initiates C hordein proteolysis by cleavage at a few primary sites, followed by cleavage at a large number of secondary sites, to yield peptides of four to 15 residues (Davy et al., 1998). To investigate the importance of the primary cleavage sites for C hordein degradation, a mutant C hordein polypeptide was expressed in which the primary EP-B cleavage sites were eliminated by substituting a Ser residue at the P₂ position (see Table I). Five mutagenic primers were used to mutate the C hordein nucleotide sequences encoding the five primary cleavage sites. Because of the highly repetitive C hordein gene sequence, some multiple primer annealing events were unavoidable. The amino acid sequence of the mutant C hordein selected for expression is shown alongside the wild-type protein in Figure 4, with mutations at all five cleavage sites. The additional mutations induced by primer 5 were not localized at primary cleavage sites.

View larger version (38K): [in this window] [in a new window]   Figure 4.   Deduced amino acid sequence of wild-type recombinant C hordein (left), showing primary () and secondary () cleavage sites. The sequence has been formatted to emphasize the repeat domain. The sequence on the right shows the mutant C hordein after removal of the primary cleavage sites (). Point mutations are indicated by shading and include some residues not at primary cleavage sites due to multiple primer annealing events.

The purified mutant and wild-type C hordein polypeptides were incubated with EP-B and samples were taken over 40 min for SDS-PAGE analysis. The wild-type polypeptide was initially cleaved at the primary cleavage sites, releasing a characteristic set of peptide fragments, followed by complete degradation within 40 min. The mutant polypeptide remained largely intact throughout the incubation. Degradation by EP-B was monitored by quantifying the Coomassie-stained full-length C hordein protein in SDS-PAGE (Fig. 5). Digestion of the mutant polypeptide following the addition of 15-fold more units of EP-B indicated that the mutant is susceptible to EP-B hydrolysis. Wild-type C hordein was degraded 7.5 times faster than the mutant form, with 1/15 as much EP-B, so that the mutant was degraded 112 times slower than wild-type C hordein. The absence of detectable discrete cleavage products suggested that the mutant C hordein was randomly cleaved at secondary cleavage sites.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Time course of recombinant C hordein degradation by EP-B. Wild-type C hordein () is degraded rapidly, with intermediate products. Mutant C hordein, lacking the primary cleavage sites, is only slightly degraded after 40 min with the same amount of EP-B as for wild type (), and requires 15 times more EP-B () before significant degradation can be seen. No intermediate degradation products were seen.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The purified barley Cys EP EP-B is able to degrade hordeins in vitro, even when they are presented in the form of insoluble protein bodies. Our results confirm earlier results in planta (Marchylo et al., 1986), which indicated that D hordein is degraded most rapidly during malting. This may be due to its discrete localization in the reticular matrix at the periphery of storage protein bodies (Sørensen et al., 1996), and/or because D hordein is intrinsically a better substrate. The sulfur-rich hordeins, which account for 85% of barley storage proteins, form aggregates that are stabilized by inter- and intramolecular disulfide bonds (Rechinger et al., 1993). The ability of EP-B to degrade both the monomeric C hordein and the aggregated B hordein suggests that protein-to-protein interactions within the protein body are sufficiently weak to allow their rehydration and proteolytic degradation. The addition of -mercaptoethanol increases the rate of hordein degradation, but is not essential. By reducing hordein disulfide bonds, particularly intermolecular disulfide bonds, -mercaptoethanol may increase the solubility of the hordeins and their accessibility to EP-B. This function is probably performed during germination by thioredoxin h and NADPH (Kobrehel et al., 1992) or by glutathione and Cys (Pheifer and Briggs, 1995), which are secreted by the embryo.

Two series of fluorogenic, internally quenched peptide substrates differing at either P₁ or P₂ were synthesized and used to determine the substrate specificity of EP-A and EP-B, with papain as a reference. Papain is one of the best-studied Cys EPs, but its specificity has been primarily determined from the cleavage products of small polypeptides, rather than by measuring k_cat/K_m values for a series of synthetic peptides. The variation in k_cat/K_m values at P₂ was 3,300- to 5,000-fold, which confirms that the substrate cleavage site of Cys EPs is primarily determined by the nature of the residue at P₂. In contrast, the range of k_cat/K_m values at P₁ indicates a broad specificity in this position, with the exception of Pro, which resulted in very slow cleavage. Furthermore, our results show that Tyr and Phe are the best residues at P₂ for papain, while Gly is almost twice as effective as Arg at P₁. For the substrate series used in this work, all three proteases are able to discriminate between Leu and Ile or Val at P₁ by a factor of 10 to 20, and also between Gln and Asn and Glu and Asp (Table II). Papain does not distinguish between Leu, Val, and Ile at P₂ (Table I), unlike EP-A and EP-B. The substrates VRQQ, which is better for EP-A than EP-B, and LEQP, which is better for EP-B than EP-A, could be used to measure the levels of each of these proteases when both are present in crude extracts of malt (germinating) barley.

These data should also enable cleavage sites to be predicted in protein sequences, not taking secondary structure effects into account. Barley -amylase has an extension of 45 residues at the C terminus compared with soybean, and this tail is presumably readily accessible to protease attack, since the crystal structure of soybean -amylase shows the C terminus on the surface of the molecule (Adachi et al., 1998). If the repeat sequence TGGMGGQAEGPT (Fig. 3B) is analyzed for EP-B cleavable sites, taking into account that residues Thr, Glu, Gln, Ala, and Gly at P₂ and Pro at P₁ and P₁', are extremely unfavorable (Table I), then the only predicted cleavage would occur at MGGQ, which corresponds to the experimentally determined site (Fig. 3B). Cleavage was not detected at two other potential cleavage sites within the C-terminal 45 residues (VKGP and IGGE). Other potential EP-B cleavage sites are present in the rest of the -amylase sequence, but are presumably protected by their tertiary structure, as found for native versus heat-treated lipid transfer protein (Davy et al., 1999).

EP-A and EP-B were generally found to have higher k_cat/K_m values for a given synthetic substrate than papain, which is primarily due to lower K_m values, particularly for EP-A (Davy et al., 1998). Four substrates investigated here (LKQP, LTQP, LYQP, and LLQP) have K_m values of 1 µM or less (Table III). The structural basis for this difference is impossible to deduce, since mature EP-A and EP-B are only 50% identical to papain, but they may reflect the way in which EP-A and EP-B have evolved to degrade the Gln- and Pro-rich repeats found in the hordeins. Compared with papain, EP-A and EP-B show a marked preference for Leu at P₂, and for Asp, Glu, and Tyr at P₁, although Asp and Glu are not common in hordeins. An examination of B hordein sequences (Shewry and Tatham, 1990) shows numerous occurrences of potential EP-A and EP-B primary cleavage sites with Leu, Phe, or Val at P₂ and Arg, Gln, Glu, or Gly at P₁ (LRTL, LEAT, VEGV, FQQP, etc.); D hordein (Halford et al., 1992) contains similar sites (LQQG, LGQG, LGSL, VGQL, etc.).

In vitro studies have shown that EP-B initiates the degradation of C hordein by cleaving at five primary cleavage sites whose sequences closely mirror its substrate specificity with synthetic substrates. C hordein is predominantly composed of an octameric repeat enriched in Pro and Gln residues which, in the dry grain, is believed to have a rod-shaped structure stabilized by hydrogen bonding between Gln residues. Upon hydration, the octameric repeat domain is proposed to assume an extended structure, with an equilibrium between a -reverse turn helix and a poly-L-Pro II-like conformation (Tatham and Shewry, 1995). These conformational changes may facilitate access of EP-B to the primary cleavage sites and may explain why hordeins in their native state are so susceptible to proteolytic degradation by EP-B. Four out of the five EP-B primary cleavage sites are located near the N and C termini of the C hordein polypeptide, where the repeat structure is less conserved. Following primary site cleavage, C hordein is rapidly degraded by EP-B at multiple secondary sites. When hydrolysis of the primary sites is prevented, as in our mutated C-hordein, the central domain retains its native conformation and is relatively inaccessible. The absence of discrete fragments upon cleavage of the mutated C-hordein indicates that the initial cleavage occurs at a random secondary site. Our results indicate that the N and C termini stabilize the conformation of the central helical domain, which, upon cleavage of the primary sites, becomes more accessible to cleavage at secondary sites.