S. Yamaguchi, Gifu R & D Center, Amano Enzyme Inc., Kagamigahara, Gifu 509-0108, Japan. Fax: + 81 583 79 1232, Tel.:. + 81 583 79 1220, E-mail: LDV01447@nifty.ne.jp
A novel protein-deamidating enzyme was purified to homogeneity from Chryseobacterium proteolyticum and the gene encoding it was cloned. The enzyme is a monomer with a pI of 10.0, a measured Mr of ≈ 20 000 and a calculated Mr of 19 860. Extensive comparison with Streptoverticillium transglutaminase showed that the protein-deamidating enzyme lacked transglutaminase activity in terms of hydroxamate-formation between benzyloxycarbonyl-Gln-Gly and hydroxylamine, or monodansylcadaverine incorporation into casein. The enzyme deamidated the two glutaminyl residues in the oxidized insulin A chain and deamidated both casein and the oxidized insulin B chain with higher catalytic efficiencies (kcat/Km) than with short peptides. The enzyme was active against several proteins, including insoluble wheat gluten, but did not deamidate asparaginyl residues in peptides, free glutamine or other amides. The enzyme was therefore named protein-glutaminase (EC 3.5.1). The gene encoding the protein was cloned and, when expressed in Escherichia coli, the protein product had protein-glutaminase activity and cross-reacted with antiserum raised against the purified enzyme. The protein-glutaminase was shown to be expressed as a prepro-protein with a putative signal peptide of 21 amino acids and a pro-sequence of 114 amino acids. The amino-acid sequence had no obvious homology to any published sequence and is therefore a novel protein-glutaminase.
Protein deamidation, the hydrolysis of side chain amido groups of protein-bound glutaminyl or asparaginyl residues to release ammonia, has received attention in terms of biochemistry and industrial applications. Many proteins, including enzymes and hormones, can be deamidated in vivo[1,2], leading to alterations in higher-structure and function. The physiological role of protein deamidation in cells, which is usually considered to be nonenzymatic, is speculated to be a post-translational modification process related to the regulation of protein folding, protein breakdown, and aging . A hypothesis has been proposed  that protein deamidation is a molecular clock. Eye lens crystallins are very stable proteins with long half-lives in vivo, and are proteins that have been investigated in this regard [4,5]. Transglutaminase may catalyze protein deamidation  as well as catalyzing protein cross-linking and amine-incorporation . Recent reports indicate the significance of protein deamidation by transglutaminase in vivo[8–11].
Most plant storage proteins in seeds, such as cereal prolamins and legume globulins, contain a high level of glutamine residues. The abundant amido groups in these proteins are thought to serve as a nitrogen source for seed germination. Deamidation of proteins preceding their proteolytic degradation was observed in germinating seeds [12,13] and increased susceptibilities of deamidated proteins to proteases have been inferred in plant storage proteins [13,14]. Vaintraub and colleagues reported the possible involvement of enzymatic action in this process and the partial purification of the protein-deamidating enzymes from germinating wheat grain , kidney beans and squash .
Enzymes that catalyze the deamidation of glutaminyl residues in peptides have been found in Bacillus circulans[17–19]. The isolated enzymes consisted of two distinct enzymes; peptidoglutaminase I (EC 126.96.36.199) and II (EC 188.8.131.52). The former deamidated a C-terminal glutaminyl residue, and the latter deamidated internal glutaminyl residues in peptides as well as a C-terminal glutaminyl residue. Both enzymes were active against short peptides but not against proteins.
In the food protein industry, protein deamidation is regarded as a promising method to improve protein functionality in food systems [20–22]. In general, deamidated proteins have a decreased isoelectric point due to increased negatively charged carboxyl groups, resulting in a protein with good solubility at more acidic pH values. This is particularly important for food proteins because the pI values of most food proteins are acidic or semi-acidic and the pH values of many food systems are also in the same pH range. Deamidation of proteins could also cause the alteration of tertiary structure. An unfolding of the protein would take place because of the electrostatic repulsion of newly formed negatively charged carboxyl groups. This unfolding leads to the exposure of hydrophobic regions, previously buried in the interior of the protein, to the aqueous surroundings. This alteration results in a protein with an improved amphiphilic character that could be used as an emulsifier or foaming agent. Deamidations of food proteins have been investigated by various methods (reviewed in [20,22]) such as mild acid treatment, anion-catalyzed deamidation, dry heating under mild alkaline conditions, and thermal treatment. Although deamidations by these treatments improved protein functionalities, there were undesired side-effects, such as concomitant peptide bond cleavage, unavoidably brought about by the chemical/physical treatments. Therefore, enzymatic methods have been desired because of their advantages of being selective and mild. The possibilities of the use of transglutaminase [23–25] peptidoglutaminases [26–28], and proteases [29–31] for this purpose have been explored. These enzymes, however, were unsuitable for protein deamidation because the primary catalytic reactions of transglutaminase and protease are not deamidation, and the primary substrate of peptidoglutaminases is not protein. A new enzyme is required that catalyzes the deamidation of protein as its primary reaction [22,29].
After screening microorganisms from many environmental soils, we discovered a novel protein-deamidating enzyme from a bacterium . This extracellular enzyme catalyzed the deamidation of intact protein without transglutaminase and protease activities. The enzyme-producing bacterium was taxonomically identified as a new species belonging to the genus Chryseobacterium. This nonpathogenic bacterium was named C. proteolyticum. We report here the purification and characterization of the enzyme and the encoding gene. The enzyme was named as protein-glutaminase according to its characteristics elucidated in this study. The deduced amino-acid sequence of the enzyme indicates that the enzyme has not previously been described.
Materials and methods
Bacterial strain and culture conditions
C. proteolyticum strain 9670 used in this work was isolated from soil in Tsukuba, Japan, as a protein-deamidating enzyme producer as described previously . For the production of protein-glutaminase, cultures were grown with shaking (200 r.p.m.) in Luria–Bertani basal medium (Oxoid, Basingstoke, UK) at 25 °C for 40 h.
Purification of protein-glutaminase
A partially purified protein-glutaminase was obtained according to the method described previously (phenyl-Sepharose chromatography)  with slight modifications. Addition of 2 mm EDTA to the culture supernatant was omitted. The ultrafilter-concentrate was freeze dried and resuspended in 2.0 m NaCl in 10 mm sodium phosphate buffer (pH 6.5) instead of dialysis against this buffer. The partially purified enzyme fraction obtained by phenyl-Sepharose chromatography was dialyzed against 10 mm sodium phosphate buffer (pH 6.5) and then freeze dried. The powder was dissolved in 0.6 m NaCl, 0.025% Tween 20 in 10 mm sodium phosphate buffer (pH 6.5) and subjected to a Sephacryl S-100 High Resolution HiPrep 26/60 column (Amersham Pharmacia Biotech, Uppsala, Sweden) pre-equilibrated with the same buffer. Proteins were eluted with the same buffer at a flow rate of 1.0 mL·min−1 using a fast protein liquid chromatography system (Amersham Pharmacia Biotech). The active fraction, which was eluted from 211 to 282 mL, was concentrated by ultrafiltration using Macrosep, a centrifugal membrane concentrator (Mr 1000 cut, Filtron Technology Co., Northborough, MA, USA). The concentrate was stored at −20 °C.
Protein-glutaminase assay and protein determination
The standard method for protein-glutaminase assay using Cbz-Gln-Gly (Peptide Laboratory, Osaka, Japan) as a substrate and the definition of an enzyme unit were as described previously . The protein concentration was measured with the BCA assay (Pierce, Rochford, IL, USA) using bovine serum albumin as a standard.
Effects of pH and temperature on the activity and stability of protein-glutaminase
The effect of pH on protein-glutaminase activity was determined using 40 mm boric acid/acetic acid/NaOH buffers with the pH range from 3 to 12. The pH stability was determined by incubating the enzyme at 30 °C for 18 h in the above buffers, and measuring the remaining activities. The effect of temperature on the protein-glutaminase activity was determined by measuring the activity at the indicated temperature with a reaction time of 10 min. The temperature stability was determined by incubating the enzyme at the indicated temperatures in 50 mm sodium phosphate buffer (pH 7.0). Samples were removed at 10 min or 60 min, and the remaining activity was measured.
Gel electrophoresis and Western blotting
SDS/PAGE was performed with precast 4–12% Bis-Tris gradient gels and Mes running buffer using a minicell apparatus (Novex/Invitrogen, Groningen, the Netherlands) according to the manufacturer’s instructions. Proteins were stained with Coomassie brilliant blue R-250 or silver-stained using a kit from Wako Pure Chemicals, Osaka. For Western blotting, proteins in SDS/PAGE gels were transferred onto a poly(vinylidene difluoride) membrane (Immobilon-P, Millipore, Bedford, MA, USA) by electrophoretic blotting with Sartoblot II-S (Sartorius, Göttingen, Germany). The membrane was incubated in Tris/borate/EDTA/Tween (10 mm Tris/HCl, pH 7.0, 0.15 m NaCl, 0.1% Tween 20) containing 1% bovine serum albumin for 60 min, and then in Tris/borate/EDTA/Tween containing mouse polyclonal anti-(protein-glutaminase) serum (1 : 10 000) for 60 min. Color development was performed by using a ProtoBlot Western Blot AP System for mouse (Promega, Madison, WI, USA) according to the manufacturer’s instructions.
Molecular mass determination
The apparent molecular mass of the purified, denatured protein-glutaminase was determined by SDS/PAGE and comparison with a broad range molecular mass standard mixture (New England Biolabs, Hitchin, UK). The apparent molecular mass of native enzyme was determined by size-exclusion chromatography with a TSK-gel G2000SWXL (Tosoh, Tokyo, Japan) or a Superose 12 column (Amersham Pharmacia Biotech). The column was run with 10 mm sodium phosphate buffer, pH 6.5, 0.2 m NaCl at a flow rate of 1.0 mL·min−1 for TSK-gel G2000SWXL or run with 10 mm sodium phosphate buffer, pH 6.5, 0.6 m NaCl at a flow rate of 1.0 mL·min−1 for Superose 12. Molecular mass markers were: bovine serum albumin (Mr 66 000), carbonic anhydrase (Mr 29 000), ribonuclease A (Mr 13 700), aprotinin (Mr 6500) from Sigma, and ovalbumin (Mr 45 000), chymotrypsinogen A (Mr 25 000) from Amersham Pharmacia Biotech.
Isoelectric focusing was performed according to the method of Vesterberg and Svensson . A 0–50% glycerol density gradient containing 1.36% ampholine pH 8–10.5 (Amersham Pharmacia Biotech) and the enzyme solution was made in a 110-mL column. A dialyzed enzyme fraction, containing 15 U of the Cbz-Gln-Gly-deamidating activity, obtained from phenyl-Sepharose chromatography was applied. The column was maintained under a constant potential of 600 V for 48 h at 4 °C. Fractions (0.9 mL) were collected, and the enzyme activity and pH were measured for each fraction.
Deamidation of oxidized insulin chains
To the reaction mixture (final volume: 300 µL) containing 1.58 mm (4 mg·mL−1) oxidized insulin A chain (bovine, Sigma) or 1.14 mm (4 mg·mL−1) oxidized insulin B chain (bovine, Sigma), 50 mm sodium phosphate, pH 6.5, which was preheated at 37 °C for 5 min, 6 µg (11.1 µL) of protein-glutaminase was added, and the mixture was incubated at 37 °C. At the indicated incubation time, 0.08 vol. (24 µL) of 80% trichloroacetic acid was added to the mixture to terminate the reaction. After centrifugation at 18 000 g for 5 min, released ammonia in the supernatant was determined by a NADH/glutamate dehydrogenase method as described previously . For the blank reaction, enzyme was added after addition of trichloroacetic acid. For amino-acid sequencing, deamidated oxidized insulin A chain obtained by 120-min incubation was subjected to RP-HPLC using a µBondasphere 5µ C18-300 Å column (Millipore/Waters). The column was pre-equilibrated with 0.1% trifluoroacetic acid and the peptide was eluted by two steps of linear acetonitrile gradient in 0.1% trifluoroacetic acid; from 0 to 20% acetonitrile for 20 min and 20–40% acetonitrile for 10 min at a flow rate of 1.0 mL·min−1. The deamidated A chain that eluted at a time of 23 min was subjected to amino-acid sequencing, which was performed using an Applied Biosystems Procise Sequencer 610 A at the Protein Sequencing and Peptide Synthesis Facility, John Innes Centre (Norwich, UK).
Hydroxamate formation reaction
To the mixture (100 µL) containing 30 mm Cbz-Gln-Gly, 100 mm hydroxylamine, 10 mm glutathione (reduced form) and 200 mm Tris/HCl, pH 7.0, which was preheated at 37 °C for 5 min, 10 µL of protein-glutaminase (5.4 µg) or purified Streptoverticillium transglutaminase (18.8 µg) was added, and the reaction mixture was incubated at 37 °C. Streptoverticillium transglutaminase from strain S-8112, a variant of Streptoverticillium mobaraense, was purified according to the method of Ando et al. . For the determination of hydroxamate formed, 100 µL of the stop solution (4% trichloroacetic acid, 1 m HCl, 1.67% FeCl3) was added to the reaction mixture at the indicated incubation time to terminate the reaction. After centrifugation at 18 000 g for 5 min, the absorbance at 540 nm in the supernatant was measured and the hydroxamate content was determined against the standard curve prepared using l-glutamic acid γ-monohydroxamate. For the blank reaction, enzyme was added after addition of the stop solution. For the determination of ammonia released, the termination of reaction and the measurement of ammonia were performed as described previously . The experiments in the absence of hydroxylamine were also performed for both enzymes, in which 100 mm hydroxylamine was removed from the reaction mixture.
Incorporation of monodansylcadaverine into N′,N′-dimethylcasein
The reaction was carried out according to Gorman and Folk  with some modifications. The reaction mixture (40 µL) containing 0.4% N′,N′-dimethylcasein (Sigma), 0.25 mm monodansylcadaverine, 3 mm dithiothreitol, and 10.8 µg of protein-glutaminase or Streptoverticillium transglutaminase, was incubated at 37 °C for 15 h. The buffer used was 100 mm sodium phosphate, pH 6.5 for protein-glutaminase or 100 mm Tris/HCl, pH 7.0 for transglutaminase. After SDS/PAGE, dansylcadaverine compounds in the gel were visualized under ultra-violet light, and the proteins were then stained by Coomassie brilliant blue R-250.
Determinations of the kinetic parameters and substrate specificity
The activity of protein-glutaminase toward various substrates was determined by measuring the formation of ammonia under the reaction conditions described above with some modifications. For the determination of kinetic parameters, various concentrations of substrate dissolved in 200 mm sodium phosphate buffer, pH 6.5, were incubated with 1–20 µg of protein-glutaminase for 15–60 min. The Km and Vmax values were calculated from double-reciprocal Lineweaver–Burk plots evaluated by linear regression analyses. The kcat value was calculated based on the molecular mass of the monomeric protein-glutaminase deduced from the gene sequence. For examination of the protein substrate specificity, protein substrates (10 mg·mL−1) were treated with protein-glutaminase (2, 20 or 200 µg·mL−1) for 60 min. Suspensions were used for some insoluble proteins. Casein, Hammarsten, was from Merck/BDH, Poole, United Kingdom. Cbz-Gln-O-methyl, Gly-Gln-Gly, Phe-Gln-Gly-Pro, and Gly-Gln-Pro-Arg were from Backem, Saffron Walden, UK. Gly-Gln, Cbz-Asn-Gly, and Cbz-Asn were from Kokusan Chemical, Tokyo, Japan. All other sustrates including peptides, proteins, and amine compounds were from Sigma.
The N-terminal amino-acid sequence of the purified protein-glutaminase was determined using a Applied Biosystems Procise Sequencer 492 at the Research Institute of Food Science, Kyoto University (Uji, Japan). Analyses of an internal terminal amino-acid sequence was performed at the Protein Sequencing and Peptide Synthesis Facility, John Innes Centre, as follows. Purified protein-glutaminase was alkylated by performic acid under reducing conditions, and then digested by trypsin. A peptide purified by RP-HPLC of the trypsin-digestion mixture was then subjected to amino-acid sequencing.
Generation of a DNA probe for the protein-glutaminase gene cloning
Chromosomal DNA of C. proteolyticum strain 9670 was isolated . Two degenerate oligonucleotide mixes were synthesized using a Applied Biosystems model 394 DNA synthesizer (PerkinElmer) for use in PCR: sense primer 5′-GCI(TA)(CG)IGTIAT(TCA)CC(TACG)GA(TC)GT-3′ and antisense primer 5′-A(AG)(AGTC)AC(AG)CA(AG)TT(AGTC)GT(AG)TT(AGT)AT-3′. PCR reaction mixture (50 µL) contained: 1 × PCR reaction buffer (PerkinElmer, Branchburg, NJ, USA), 200 pmol of sense primer, 200 pmol of antisense primer, 200 µm each dNTP (Promega), 1.25 U of Taq polymerase (PerkinElmer), and 0.1 µg of C. proteolyticum strain 9670 chromosomal DNA as a template. The reaction mixture was incubated in a Hybaid Omnigene™ Thermal Cycler (Teddington, Middlesex, UK) under the following conditions: stage 1,1 cycle of 94 °C for 5 min; stage 2, 30 cycles of 94 °C for 1 min, 44 °C for 1 min, and 72 °C for 1 min; stage 3, 1 cycle of 72 °C for 10 min. The reaction product was purified from a 1% agarose gel by using a QIAquick gel extraction kit (Qiagen GmbH, Hilden, Germany), and cloned into a pCRII vector (Invitrogen, Groningen, the Netherlands) according to the manufacturer’s instructions. This PCR fragment was used as a DNA probe for gene cloning.
Cloning of the protein-glutaminase gene
Chromosomal DNA of C. proteolyticum strain 9670 was digested with EcoRI and ligated into an EcoRI-cut λ ZAPII vector (Stratagene, Cambridge, UK). The mixtures were packaged using Gigapack III Gold (Stratagene) to obtain the gene library. The library was screened by plaque hybridization with the PCR-generated probe according to Stratagene’s instructions. The probe was labeled with a Megaprime DNA Labeling System and [α-32P]dCTP (Amersham Pharmacia Biotech). Phage particles were recovered from the positive plaques, and plasmid p9T1-2, originating from one of the positive plaques, was obtained by in vivo excision according to Stratagene’s instructions. Genomic Southern hybridization was performed using ECL direct nucleic acid labeling and detection system (Amersham, Pharmacia Biotech) with a stringency wash of 0.5 × NaCl/Cit, 0.4% SDS, and 6 m urea at 42 °C.
Plasmid p9T1-2, which contained a 4.9-kb chromosomal DNA fragment, was restriction-mapped and the region containing the protein-glutaminase gene was identified by Southern analysis (ECL system) using the PCR probe. Several restriction fragments for sequencing were subcloned into the Escherichia coli vector pBluescript (Stratagene). DNA sequencing was performed with an Applied Biosystems automated DNA sequencer model 373 (PerkinElmer) using a Taq DyeDeoxy Terminator Cycle sequencing kit (PerkinElmer) according to the manufacturer’s instructions. The nucleotide sequence reported in this paper will appear in the GenBank/EMBL/DDBJ nucleotide sequence data bank with accession number AB046594.
Construction of Escherichia coli expression plasmid
PCR primers containing EcoRI (underlined below) sites were designed to amplify a DNA fragments coding for the mature form of the protein-glutaminase: a sense primer 5′-CCGAATTCTTGGCGAGTGTAATTCCTGATG-3′ (corresponding to nucleotide numbers 586–607 in Fig. 6); an antisense primer, 5′-TCGAATTCTTAAAATCCACAGCTGGATAC-3′ (nucleotide number 1123–1143). PCR conditions were as described above except for the following changes: 25 pmol of each primer was used, 25 ng of plasmid p9T1-2 was used, and the amplification condition stage 2 was 30 cycles of 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 1 min. Amplified fragment was initially cloned into EcoRI-cut pCRII, and verified by sequencing. The EcoRI fragment was then cloned into EcoRI site of the expression vector pGEX-1λT (Amersham Pharmacia Biotech), downstream of the glutathione S-transferase gene. The resultant plasmid, pN7-9 was used to transform E. coli BL21 (Amersham Pharmacia Biotech) to express the fusion protein.
Expression of the protein-glutaminase in E. coli
E. coli BL21 harboring pN7-9 or pGE-1λT was precultured at 37 °C for 16 h in 10 mL of Luria–Bertani medium supplemented with 100 µg·mL−1 of ampicillin. The preculture (5 mL) was inoculated into a 250-mL conical flask containing 45 mL of Luria–Bertani medium supplemented with 100 µg·mL−1 of ampicillin. After 80 min incubation (D600 = 0.97–1.14) at 37 °C with rotary-shaking at 200 r.p.m., 10 µL of 0.5 m isopropyl thio-β-d-galactoside was added, and the culture was incubated for a further 4 h under the same conditions. Cells were harvested by centrifugation at 5000 g for 5 min at 4 °C, and re-suspended in 5 mL of 50 mm Tris/HCl (pH 8.0), 2 mm EDTA. Cells were lysed by the addition of 50 µL of 10 mg·mL−1 egg white lysozyme (Sigma), and subsequent mild ultra-sonication (three cycles of 10-s on and 30-s off) on ice to shear the DNA. The lysate was then centrifuged at 12 000 g for 15 min at 4 °C. A portion (100 µL) of the supernatant was incubated with 4 µL of thrombin solution [1000 U thrombin (Amersham Pharmacia Biotech) per mL in 9 mm sodium phosphate (pH 6.5), 140 mm NaCl] at 25 °C for 16 h. A control without thrombin was included. Samples were then subjected to enzyme assay, SDS/PAGE, and Western blotting.
Purification of the protein-glutaminase
The protein-deamidating enzyme was purified from the culture supernatant of C. proteolyticum strain 9670 to a homogeneous state (purification summary described in Table 1). The enzyme eluted from the Sephacryl S-100 chromatography in a very late fraction, corresponding to an elution fraction for low molecular mass compounds < 5000. This indicates that the enzyme had some affinity for the Sephacryl S-100 resin and that therefore this step was effective in separating the enzyme from other protein components. After a final purification step, the enzyme was enriched 153-fold with a yield of 30% (Table 1). The purified enzyme was judged to be homogeneous by SDS/PAGE stained with Coomassie brilliant blue R-250 (Fig. 1) and silver (not shown).
Table 1. Purification of the protein-glutaminase from Chryseobacterium proteolyticum.
Total protein (mg)
Total activity (U)
Specific activity (U·mg−1)
The apparent molecular mass of the purified, denatured enzyme was estimated to be 20 000 (Fig. 1) and 19 000 (not shown) on SDS/PAGE under reducing and nonreducing conditions, respectively. By size-exclusion chromatography, the molecular mass of the native enzyme was determined to be 16 000 and 9000 by using TSk-gel S2000SWXL and Superose 12 columns, respectively. Lower estimations of molecular mass on the size-exclusion chromatographies might be caused by the affinity of the enzyme for the gel matrices. It was concluded that the protein-deamidating enzyme exists as a monomeric single polypeptide with an apparent molecular mass of 20 000. The isoelectric point of the enzyme was 10.0 as determined by density gradient isoelectric focusing.
Deamidation of oxidized insulin chains
To determine which amino acids in protein are deamidated by the enzyme, oxidized insulin A and B chains were incubated with the enzyme. The released ammonia during the reactions was monitored and the complete amino-acid sequence of the deamidated A chain was determined. Figure 2 shows the ammonia-releasing patterns from A and B chains by the enzyme. Two moles and one mole of ammonia per mole of A and B chains were released, respectively. As insulin A chain has two residues of glutamine (Gln5 and Gln15) and two residues of asparagine (Asn18 and Asn21) in its 21 amino-acid sequence, and the B chain has one of each residue (Gln4 and Asn3) in its 30-amino-acid sequence, the enzyme seemed likely to deamidate either glutamine or asparagine with a deamidation degree of 100%. The amino-acid sequence of the deamidated A chain was determined to be G-I-V-E-(E5)-(–)-(–)-A-S-V-(–)-S-L-Y-(E15)-L-E-N-Y-(–)-N, where (–) indicates no peak observed on the cycle of protein sequencing analysis and all four (–)s correspond to the positions of cysteic acid in the oxidized insulin A chain. This result indicates that the enzyme deamidates glutaminyl residues and not asparaginyl residues in protein.
Comparison with transglutaminase
In a previous paper, we reported that the partially purified protein-deamidating enzyme showed no transglutaminase activity . To examine this in more detail, the purified enzyme was compared with transglutaminase from S. mobaraense in terms of hydroxamate-formation between Cbz-Gln-Gly and hydroxylamine, deamidation of Cbz-Gln-Gly, and amine-incorporation into casein. Figure 3 shows the ammonia-releasing and hydroxamate-formation patterns from Cbz-Gln-Gly, in the presence or absence of hydroxylamine, by both enzymes. The protein-deamidating enzyme released one mole of ammonia from one mole of Cbz-Gln-Gly with no formation of hydroxamate in the presence of hydroxylamine (Fig. 3A). This ammonia-releasing pattern was the same as that of the reaction in the absence of hydroxylamine (Fig. 3C). Streptoverticillium transglutaminase produced one mole of hydroxamate as well as one mole of ammonia from one mole of Cbz-Gln-Gly in the presence of hydroxylamine as expected (Fig. 3B). In the absence of hydroxylamine, transglutaminase released very little ammonia from Cdz-Gln-Gly. The initial rate for deamidation of the protein-deamidating enzyme was 25.01 µmol·min−1·mg−1 (Fig. 3A,C), whereas those for transglutamination (formation of hydroxamate) and deamidation of Streptoverticillium transglutaminase were 10.64 µmol·min−1·mg−1 (Fig. 3B) and 0.03 µmol·min−1·mg−1 (Fig. 3D), respectivety.
The amine-incorporation reaction of the protein-deamidating enzyme was also examined. N′,N′-dimethylcasein and monodansylcadaverine, a fluorescent primary amine, were incubated with the enzyme and the reaction products were analyzed by SDS/PAGE. As shown in Fig. 4A, no incorporation of the fluorescent monodansylcadaverine into N′,N′-dimethylcasein was observed in the reaction products by the protein-deamidating enzyme (lane 3), where the unreacted monodansylcadaverine was observed at the bottom of the gel as also seen in the control (without enzyme, lane 2). Streptoverticillium transglutaminase catalyzed the monodansylcadaverine-incorporation into N′,N′-dimethylcasein as expected (lane 4). Most of the fluorescence was observed in the bands corresponding to the N′,N′-dimethylcasein protein bands (lanes 4 of Fig. 4A,B). The protein bands of protein-deamidating enzyme (Mr 20 000, Fig. 4A, lane 5) and Streptoverticillium transglutaminase (Mr 40 000, Fig. 4A, lane 6) can be seen in each reaction mixture (Fig. 4A, lane 3 and 4, respectively).
Substrate specificity and kinetic parameters
To examine the substrate specificity of the enzyme, the kinetic parameters toward casein, oxidized insulin B chain, and several short peptides were compared (Table 2). The protein-deamidating enzyme showed higher catalytic efficiencies (kcat/Km) toward protein (casein) and a long chain peptide (oxidized insulin B chain) than toward the short peptides. Among the short peptides containing an internal glutaminyl residue tested, Cbz-Gln-Gly and Cbz-Gln-O-methyl were good substrates but three others (Gly-Gln-Gly, Phe-Gln-Gly-Pro, and Gly-Gln-Pro-Arg) were very poor. The enzyme also showed activity towards short peptides containing a C-terminal glutaminyl residue but the efficiencies were low. Amide compounds such as free glutamine (the substrate of glutaminase, EC184.108.40.206), C-terminal amido-containing short peptides, Pro-Leu-Gly-NH2 and Cbz-Gly-NH2 (the substrates for peptide amidase  or peptidase ), propionamide, acrylamide, and acetamide (the substrates for amidase, EC220.127.116.11) were hardly deamidated (less than 0.5 of kcat/Km). No activity was detected toward Cbz-Asn-Gly, Cbz-Asn, or free asparagine. Table 3 shows the specific activities of protein-deamidating enzyme toward several proteins. The enzyme deamidated most of the proteins tested but the efficiencies varied. Milk caseins and wheat gluten were good substrates, whereas bovine serum albumin and hen egg ovalbumin were very poor.
Table 2. Kinetic parameters of the protein-glutaminase determined for casein, oxidized insulin B chain, and several short peptides containing glutaminyl residue. The Km value for casein was expressed in terms of a concentration of glutaminyl residues in the protein based on an average molar content of glutaminyl residue in casein of 17 mol per mol of protein and an average molecular mass for casein of 23261. These figures were calculated based on the number of glutaminyl residues in the four components of casein (αS1-, αS2-, β, and κ-casein) and the relative content of each component in the casein preparation .
Insulin B, oxidized
Table 3. Specificity on protein substrates of the protein-glutaminase.
These results, together with those of the insulin chain deamidation experiments, indicate that the protein-deamidating enzyme from C. proteolyticum strain 9670 is a protein-glutaminase.
Effects of pH and temperature
The protein-glutaminase activity exhibited a broad pH optimum between pH 5 and 7 (Fig. 5A). The enzyme showed more than 90% of the remaining activity at pH 5–8.7 after incubation of the enzyme in the buffers at various pHs for 18 h (Fig. 5A). The optimal temperature for the activity was between 50 and 60 °C when the activity was assayed in sodium phosphate buffers for 10 min at various temperatures (Fig. 5B). The heat stability studies indicated the protein-glutaminase retained more than 93% of its activity after incubation at up to 55 °C for 60 min. The enzyme lost 21% and 90% of its activity after incubation at 60 °C for 10 min and 60 min, respectively (Fig. 5B).
The effects of various metal ions and chemical reagents on the protein-glutaminase activity were examined. The activities were assayed under the standard conditions using Cbz-Gln-Gly as a substrate in the presence of metal ions (2 mm) or chemical reagents (1 mm). HgCl2 and AgNO3 completely inhibited the activity. Partial inhibitions were observed with CuCl2 (74.9%), ZnCl2 (69.9%), and FeCl2 (26.8%). Other metal ions including CaCl2, MgCl2, CoCl2, MnSO3, NiCl2, and FeCl3 did not show any significant inhibitory or stimulatory effects on the activity. Among the chemical reagents tested, only iodoacetamide caused an inhibition (47.1%), but other thiol-reagents, p-chloromercuribenzoic acid and N-ethylmaleimide, did not significantly affect the activity. Serine enzyme inhibitors (phenylmethanesulfonyl fluoride and diisopropyl fluorophosphates), the chelating reagent (EDTA), and reducing agent (dithiothreitol) had no significant effects.
The first 20 N-terminal amino acids of the purified protein-glutaminase were determined to be L-A-S-V-I-P-D-V-A-T-L-N-S-L-F-N-(E,Q17)-I-K-N, where the seventeenth amino acid (E,Q) indicates that glutamic acid and glutamine were almost equally observed. The first 20 N-terminal amino acids of a tryptic digest of the enzyme were determined to be S-P-S-N-S-Y-L-Y-D-N-N-L-I-N-T-N-C-V-L-T. Regions used for design of PCR primers employed in gene cloning are underlined. PCR amplification using the degenerate primer pair resulted in a 0.48-kb DNA fragment. DNA sequencing of the fragment revealed an open reading frame that completely matched the remaining amino-acid sequences in the two determined protein sequences (from A9 to N20 in the N-terminal sequence of the purified enzyme and from S1 to L12 in the sequence of the tryptic digest), confirming that the amplified PCR fragment could encode part of the protein-glutaminase gene. Southern hybridization analysis of C. proteolyticum strain 9670 genomic DNA, digested with various restriction enzymes, confirmed that the gene exists as a single copy in the genome. Screening of the genomic library with the 0.48-kb PCR probe resulted in 57 positive signals from a library of 50 000 independent phages. Analyses of phages from five independent plaques revealed that all phages contained a 4.9-kb EcoRI fragment, which size corresponded to that observed in Southern blot of EcoRI-digested chromosomal DNA. From one clone, a plasmid, designated p9T1-2, was isolated by in vivo excision. The region containing the putative protein-glutaminase gene in the 4.9-kb EcoRI fragment was localized by Southern analysis. DNA sequencing of this region revealed a 963-bp open reading frame, which included the 0.48-kb PCR probe sequence in-frame (Fig. 6). The open reading frame encoded a predicted protein of 320 amino acids.
The determined N-terminal amino-acid sequence of the purified protein-glutaminase was found to start at position 136 in the deduced amino-acid sequence of the open reading frame, suggesting that the first 135 amino acids could comprise a prepro-sequence. The first ≈ 20 amino acids of the putative prepro-sequence had characteristics of a typical signal peptide. A signal peptide prediction program, signalp, indicated the C-terminus of A21 as a potential signal peptide cleavage site for Gram-negative bacteria. Five internal Met residues existing in the prepro region were not considered to indicate translation initiation codons by analyses with the signalp program. It was predicted therefore that the protein-glutaminase from C. proteolyticum was synthesized as a prepro-form with a 21-amino-acid signal peptide and a 114 amino-acid pro-region. The mature enzyme was predicted to contain 185 amino-acid residues with a calculated Mr of 19 860, matching the 20 000 estimated by SDS/PAGE. The amino-acid sequences of the prepro-and mature form were entered in the Genbank and Swiss-Prot protein sequence databases to check homology with known proteins. No matches were found, indicating that the protein-glutaminase from C. proteolyticum is a novel protein. No amino-acid sequence motifs and domains defined in the PROSITE Dictionary of Protein Sites and Patterns  were found. The amino-acid composition of the mature enzyme is characterized by a high content of cysteine residues (11 out of 185 residues). The seventeenth amino acid of the mature form, where Glu and Gln were almost equally observed in the direct amino-acid sequencing of the purified enzyme, was coded for Gln by the gene. This indicates partial self-deamidation in the enzyme.
Expression of the protein-glutaminase
To confirm the identity of the cloned gene, an expression plasmid pN7-9 was constructed, in which a PCR-amplified DNA fragment encoding the mature form of protein-glutaminase was cloned behind the C-terminal region of the E. coli glutathione S-transferase gene. SDS/PAGE analyses of E. coli cell lysates showed that the transformant expressed a large amount of a fusion protein of apparent Mr≈ 42 000 (Fig. 7, lane 3), whereas the E. coli harboring pGE-1λT expressed a protein of Mr≈ 25 000 (lane 2), which was considered to be gluthatione S-transferase protein. Most of the fusion protein appeared to be insoluble as it was absent from the supernatant of the cell lysate as judged by SDS/PAGE (lane 5). Western blot analysis showed that the fusion protein was recognized by the anti-(protein-glutaminase) serum (Fig. 7B). A small amount of fusion protein existed as a soluble protein as seen in lane 5 of Fig. 7B. Nothing was observed in the control (the supernatant of the E. coli harboring pGE-1λT, lane 4). After thrombin treatment of the supernatants, a cross-reacting band with Mr 20 000, the size of the mature protein-glutaminase, appeared in the supernatant of E. coli harboring pN7-9 and the band corresponding to the fusion protein with Mr 42 000 diminished (Fig. 7B, lane 7). Nothing was observed in the control again (lane 6). The protein-glutaminase activities in the supernatant of the cell lysate of the E. coli harboring pN7-9 were detected as 19.02 mU·mL−1 toward Cbz-Gln-Gly and 12.06 mU·mL−1 toward casein. The ratio of casein-deamidating activities to Cbz-Gln-Gly-deamidating activity was 0.63, which was similar to that observed for the purified C. proteolyticum enzyme. No activities toward either substrate were detected in the E. coli harboring pGE-1λT. From the results of the analyses of Western blotting and enzyme activities, it was concluded that the cloned gene encoded the protein-glutaminase.
We have reported here the purification and characterization of a protein-glutaminase (EC 3.5.1) from C. proteolyticum strain 9670. The protein-glutaminase catalyzed the deamidation of glutaminyl residues in the substrate polypeptide, resulting in the conversion of glutaminyl residues to glutamyl residues and release of ammonia. The enzyme prefers proteins or long peptides to short peptides as its substrate. The general reaction scheme is:
Peptidoglutaminases I and II, from Bacillus circulans have been purified and characterized [17–19]. The protein-glutaminase differs from the peptidoglutaminases in many respects. Most importantly, the peptidoglutaminases are inactive against higher molecular mass peptides and proteins. For example, casein is not deamidated by peptidoglutaminases . Even after denaturation of protein substrates, the enzymes were inactive towards casein and whey proteins  and only slight enhancement of deamidation was observed with the heat/alkaline-treated proteins . To deamidate proteins substantially by peptidoglutaminases, pretreatment of proteins with a protease is essential [17,26,28,42]. In contrast, protein-glutaminase deamidated intact casein with a higher efficiency than short peptides. The peptidoglutaminases catalyzed hydroxamate-formation between Cbz-Gln-Gly and hydroxylamine  whereas protein-glutaminase did not. Physico-chemical properties differ between peptidoglutaminase I and II  and the protein-glutaminase. Peptidoglutaminases are dimers with subunit sizes of Mr 42 000 and 51 000 and pI values of 4.1 and 4.0, whereas protein-glutaminase has a Mr of 20 000 and a pI of 10.0. Peptidoglutaminases are located intracellularly but the protein-glutaminase is a secreted enzyme.
Protein-deamidating enzymes, described as protein deamidase, in germinating plant seeds were reported as enzymes possibly involved in the seed germination process [15,16]. The wheat grain enzyme was partially purified and characterized . Although comparison between this enzyme and the protein-glutaminase is limited by the available data for the wheat enzyme, the reactivity to free glutamine was clearly different. The wheat enzyme deamidated free glutamine significantly, i.e. the relative activity for free glutamine was at least 25% of those of proteins . Protein-glutaminase does not deamidate free glutamine. The sensitivity of these enzymes to CaCl2 and SH-reagents is also different.
Transglutaminase (EC 18.104.22.168) is an enzyme that catalyzes the acyl transfer reaction in which the γ-carboxyamido groups of glutaminyl residues in proteins or peptides are the acyl donors. The acyl acceptors are amine compounds, including protein (peptide)-bound lysine. In the absence of amines, transglutaminase catalyzes the deamidation of glutaminyl residues, in which water acts as the acyl acceptor instead of an amine. Previously, we reported that protein-glutaminase lacked the typical transglutaminase reactions such as hydroxamate formation between Cbz-Gln-Gly and hydroxylamine, and formation of cross-linked higher molecular mass products from casein . In this study, we added further evidence that the protein-glutaminase is not a transglutaminase by comparison with a transglutaminase purified from a variant strain of S. mobaraense. The deamidation rate toward Cbz-Gln-Gly of Streptoverticillium transglutaminase was slow (0.03 µmol·min−1·mg−1) compared to that of the protein-glutaminase (25.01 µmol·min−1·mg−1). The deamidating activity of the transglutaminase was also much less than its transglutamination activity (10.64 µmol·min−1·mg−1). The amine-incorporation activity into casein, which is another typical transglutaminase reaction, was not detected for the protein-glutaminase.
Among the proteins tested, casein was the best substrate for the protein-glutaminase. Albumins such as bovine serum albumin and ovalbumin were less susceptible to the enzyme. Casein is well known to be a disordered, flexible protein , which could aid access to the target glutaminyl residues. In contrast, compact proteins like albumin have a poor reactivity. Wheat gluten (and its component, gliadin), which has a high content of glutaminyl residues (≈ 34%) , was another good substrate for the protein-glutaminase. The protein-glutaminase can also deamidate insoluble proteins such as gluten and collagen in an aqueous suspension. Activities were low against the short peptides tested except for Cbz-Gln-Gly and Cbz-Gln-O-methyl. A possible dependency of the protein-glutaminase reaction on the amino-acid sequence surrounding the target glutaminyl residues, as well as the dependencies on the molecular mass and higher structure of the protein substrates, are issues which require further elucidation.
We also report here the gene cloning and deduced amino-acid sequence of protein-glutaminase from C. proteolyticum strain 9670. Data suggest that the protein-glutaminase is synthesized as a preproenzyme with a putative signal peptide of 21 amino acids and a pro-region of 114 amino acids. The length of the putative signal peptide falls into the range (18–26 amino acids) of a typical Sec-type signal peptide  although longer signal peptides, ranging from 39 to 50 amino acids, were predicted for most of the extracellular enzymes (five out of seven enzymes reported) from a bacterium belonging to the genus Chryseobacterium (C. menigosepticum) [46–51]. Many bacterial extracellular enzymes, mainly proteases, are synthesized as pro-enzymes. A protease (zinc metalloendopeptidase, flavastacin) from C. menigosepticum was also presumed to be synthesized as a pro-enzyme . Because the primary substrate of protein-glutaminase is protein, one of the possible roles of the pro-sequence might be the protection of the cell from the protein-deamidating activity. Transglutaminase, for which the primary substrate is also protein, was recently reported to be synthesized in S. mobaraence as an inactive pro-enzyme probably in order to protect the cells from its activity .
Mature protein-glutaminase consisted of 185 amino acids. The amino-acid sequence had no obvious homology to any sequences in the public databases analyzed by the blast program and no short sequences matched to the known protein motifs and domains analyzed by the prosite program. Inhibitor studies showed that some SH-reagents had inhibitory effects on the protein-glutaminase activity. Although the calculated pI based on the amino-acid sequence of the protein-glutaminase was 8.02 (by the peptidesort program of Wisconsin Package, GCG10), deviating from the experimental value of 10.0, a predicted pI value of 9.99 could be obtained provided that 10 out of 11 cysteine residues present in the amino-acid sequence had no free thiol, i.e. were disulfide-bonded. Further analyses on the cysteine residues in the protein may provide insight into the identification of the enzyme active site.
A hypothesized physiological role of the protein-glutaminase, an involvement in the degradation of protein to be utilized as energy or nutritional sources in cooperation with proteases, was previously mentioned . An increased susceptibility of casein deamidated by the protein-glutaminase to protease was observed (data not shown). This result, together with the higher reactivity of the protein-glutaminase to proteins rather than short peptides reported here, may support this hypothesis.
The enzymatic modification of proteins is a promising method to improve protein functionality desired for the use of proteins as food ingredients, being superior to chemical modification because of its specificity of reaction and environmental friendliness. Protease  and transglutaminase  are widely used for this purpose. The protein-glutaminase is a new type of enzyme with significant potential for the enzymatic modification of food proteins.
We thank Dr Mike J. Naldrett at the John Innes Centre and Dr Yasuki Matsumura at the Research Institute of Food Research, Kyoto University, for amino-acid sequencing. We also thank Mr Yoshiaki Kurono, Amano Enzyme Inc., and Dr Liliana G. Santiago at Facultad Ingenieria Quimica, Universidad Nacional del Litoral, Santa Fe, Argentina, for providing the purified Streptoverticillium transglutaminase and soy protein isolate, respectively.
*Present address: School of Life and Environmental Sciences, University of Nottingham, University Park, Nottingham NG7 2RD, UK.
Note: the nucleotide sequence reported in this paper will appear in the GenBank/EMBL/DDBJ nucleotide sequence data bank under accession number AB046594.