Incorporation of 15N-Labeled Ammonia into Glutamine Amide Groups

Nov 7, 2011 - ABSTRACT: Protein-glutaminase (PG) is an enzyme that catalyzes the deamidation of protein-bound glutamine residues. We found that an ...
0 downloads 0 Views 4MB Size
ARTICLE pubs.acs.org/JAFC

Incorporation of 15N-Labeled Ammonia into Glutamine Amide Groups by Protein-Glutaminase and Analysis of the Reactivity for α-Lactalbumin Noriko Miwa,*,† Nobuhisa Shimba,† Mina Nakamura,† Keiichi Yokoyama,† Noriki Nio,† Eiichiro Suzuki,† and Kenji Sonomoto§ †

Institute of Life Science, Ajinomoto, Company, Inc., 1-1 Suzuki-cho, Kawasaki-ku, Kawasaki-shi 210-8681, Kanagawa, Japan Department of Bioscience and Biotechnology, Faculty of Agriculture, Graduate School, Kyushu University, Fukuoka 812-8581, Japan

§

bS Supporting Information ABSTRACT: Protein-glutaminase (PG) is an enzyme that catalyzes the deamidation of protein-bound glutamine residues. We found that an enzyme labeling technique (ELT), which is a stable isotope labeling method based on transglutaminase (TGase) reaction, is applicable for PG. PG catalyzed incorporation of 15N-labeled ammonium ions into reactive glutamine amide groups in αlactalbumin similarly to TGase and deamidated the most reactive glutamine amide group once labeled with 15N. Furthermore, we investigated the effect of ammonium ions on the PG activity by peptide mapping, and more reactive glutamine residues were detected than were detected by the ELT in the presence of ammonium ions. This is probably because ammonium ions are competitive inhibitors, causing decreased reactivity for glutamine residues. We propose the reaction scheme of PG in the presence of the 15N-labeled ammonium ions and show that the ELT method with PG is useful for evaluating the activity of PG. KEYWORDS: ammonium ion, deamidation, NMR, protein-glutaminase

’ INTRODUCTION Protein deamidation, which hydrolyzes the amide groups on glutamine or asparagine residues in proteins, has attracted a great deal of attention in the food industry.1,2 In particular, enzymatic protein deamidation has been desired because of the advantages offered by the high specificity and mild reaction conditions of the process. A new enzyme, protein-glutaminase (PG, EC 3.5.1.44), which catalyzes the deamidation of proteinbound glutamines, was discovered in culture supernatant from Chryseobacterium proteolyticum.3,4 Deamidation generally decreases the isoelectric point of proteins due to the increased number of negatively charged carboxyl groups and enhances protein solubility in many mildly acidic food systems. Also, deamidation of proteins can improve various protein functionalities, such as emulsification, foaming, and gelation properties.4 The effect of deamidation by PG on several food proteins has been investigated,5 8 and as an example of this research, Gu et al. have reported the influence of PG on the structure of bovine α-lactalbumin, one of the major components of milk whey proteins.7 Transglutaminase (TGase; protein-glutamine γ-glutamyltransferase, EC 2.3.2.13) catalyzes an acyl transfer reaction between the γ-carboxyamide groups of peptide-bound glutamine residues (acyl donors) and a variety of primary amines (acyl acceptors), including the ε-amino groups of lysine residues.9 11 In the absence of amines, TGase catalyzes the deamidation of glutamine residues, in which water acts as the acyl acceptor instead of an amine. Yamaguchi et al. showed that PG does not catalyze an acyl transfer reaction between N-carbobenzoxy-L-glutaminyl-glycine and hydroxylamine.4 r 2011 American Chemical Society

The reaction mechanism of TGase has been proposed.12,13 First, γ-carboxyamide groups in glutamine residue interact with TGase and, subsequently, the primary amino groups of a variety of amines or the ε-amino groups of lysine residues are involved in the formation of new covalent bonds. On the basis of such a reaction mechanism, a nuclear magnetic resonance (NMR)based method using an enzymatic labeling technique (ELT) has been proposed by Shimba et al.14 16 In the ELT, the 15N nuclei are incorporated into the γ-carboxyamide groups of glutamine residues in arbitrary proteins.12 The ELT enables not only the preparation of stable isotope-labeled proteins but also the analysis of both the site specificities and reaction rates of TGase. In this study, we found that 15N-labeled ammonium ion reacts with γ-carboxyamide groups of glutamine residues in the presence of PG as well as TGase. The main purpose of this paper is to describe how PG catalyzes the incorporation of 15 N-labeled ammonium ions into reactive glutamine amide groups similarly to TGase by the ELT. The substrate used here was α-lactalbumin, as information on its primary structure is available.7,17 The ELT has made it possible to investigate the PG reactivity, that is, the number of PG-reactive glutamine amide groups and each reaction rate in the presence of a large excess of ammonium ions. Another purpose of this paper is to examine Received: July 19, 2011 Revised: November 3, 2011 Accepted: November 7, 2011 Published: November 07, 2011 12752

dx.doi.org/10.1021/jf2028895 | J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry

ARTICLE

Figure 1. Schematic drawing of ELT by using TGase (in the solid-line box). The labeled ammonium ion reacts with the γ-carboxyamide groups of glutamine residues and exchanges with their amide groups. The reaction scheme of PG in the presence of labeled ammonium ions is shown in the dottedline box. PG catalyzes exchanges of the amide groups of glutamine residues with the labeled ammonium ions and deamidation of both labeled and unlabeled glutamine residues.

the effect of ammonium ions on PG reactivity. We also discuss the difference of the reaction mechanism in the presence of ammonium ions between TGase and PG.

’ MATERIALS AND METHODS Materials. PG (EC.3.5.1.44) derived from C. proteolyticum was produced by Corynebacterium glutamicum, according to the method previously described by Kikuchi et al.18 Microbial TGase was prepared from the culture supernatant of Streptomyces mobaraensis as described previously.19 15NH4Cl and D2O were purchased from Nippon Sanso (Tokyo, Japan) and ISOTEC (Tokyo, Japan), respectively. The bovine milk α-lactalbumin (catalog no. L-5385, type I) was purchased from Sigma (St. Louis, MO). Other chemicals were purchased from Wako Pure Chemical Co. (Osaka, Japan) and Nakarai Tesque (Kyoto, Japan) and were of analytical reagent grade. Sample Preparation. A 10 mg/mL aliquot of α-lactalbumin solution was prepared by adding 200 mM 15NH4Cl and PG at the enzyme/substrate ratio of 1:5000 (w/w) in 20 mM Tris-HCl buffer, pH 7.5, in 95% H2O/5% D2O. After incubation at 37 °C for 24 h, the protein solution was placed into a 5 mm NMR tube for NMR measurements. For comparison, microbial TGase was added to the same substrate solution instead of PG at the enzyme/substrate ratio of 1:250 (w/w). As a control experiment, the substrate solution was reacted under the same conditions, except there was no addition of enzyme. To estimate the reaction rate of PG, we started consecutive measurements of 1H 15N heteronuclear single-quantum coherence (HSQC) after the addition of the enzyme. The enzyme concentration was measured by a Bradford assay using Coomassie brilliant blue 5 solution (Nakarai Tesque) and bovine serum albumin (Pierce, Rockford, IL) as a standard. NMR Measurements. NMR experiments were performed on a Bruker DMX600 spectrometer equipped with a triple-resonance probe head with XYZ triple-axis gradient coils. All spectra were recorded at 37 °C. The 1H and 15N chemical shifts were relative to the solvent H2O as 4.66 ppm.20 We used the XWINNMR software package (Bruker Co.21) for data processing and analysis.

The HSQC spectra were recorded with spectral widths of 8400 Hz for H and 1400 Hz for 15N. The pulse sequence for the HSQC spectra was as described by Bodenhausen and Ruben.22 The WATERGATE water suppression scheme with the 3 9 19 refocusing pulse was incorporated into the reverse INEPT step.23 A total of 2048 data points were used in the t2 dimension, and 200 transients were acquired for the t1 points. Prior to 2D Fourier transformation, the acquired data were multiplied by Gaussian functions in t2 and t1 and were zero-filled to yield a 1024 (F2)  512 (F1) matrix of the real data points. After samples were incubated for 1 or 3 h, the 1H 15N HSQC spectra in the presence of PG were measured 20 times for the estimation of the reaction rate. Consecutive measurements were started within 20 min after the addition of the enzyme, and each measuring time was about 135 min. A series of peak intensities was extracted in each set of 2D data. 1

Degree of Deamidation Rate of α-Lactalbumin by PG.

A 10 mg/mL aliquot of α-lactalbumin solution was incubated with PG at the enzyme/substrate ratios of 1:5000 (w/w) and 1:1000 (w/w) in 20 mM Tris-HCl buffer, pH 7.5. After incubation at 37 °C for 5 or 24 h, the reaction mixture was heated at 100 °C for 6 min with a heating block to stop the reaction. Aliquots of the reaction mixture (100 μL) were placed in microcentrifuge tubes, followed by the addition of 100 μL of 12% trichloroacetic acid to stop the enzymatic reaction. After centrifugation at 12000 rpm at 5 °C for 5 min, the ammonia released by the PG reaction in the supernatant was measured following an NADH-glutaminate dehydrogenase using an F-kit (Roche Diagnostics GmbH, Mannheim, Germany). The degree of deamidation was expressed as the ratio of deamidated residues to total glutamine residues. The total number of glutamine residues in α-lactalbumin was assumed to be six according to previous data of the primary sequence of α-lactalbumin.7,17 Protein Digestion by V8 Protease or Trypsin. An aliquot of the reaction mixture (final concentration of α-lactalbumin, 2.5 mg/mL) was suspended in 0.1% (w/v) RapiGestSF (Waters Corp., Milford, MA), 50 mM NH4HCO3 (pH 7.8) and 5 mM EDTA equilibrated at 37 °C for 2 min. Disulfide bonds were reduced by adding 10 mM DTT and incubating the samples at 60 °C for 30 min. Subsequent alkylation was achieved through the addition of 30 mM iodoacetic acid, followed by incubation at room temperature (approximately 25 °C) for 45 min in the 12753

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry

ARTICLE

Figure 2. 1H 15N HSQC spectra of α-lactalbumin incubated with or without enzymes in the presence of 15NH4Cl: (a) TGase; (b) following TGase, PG; (c) PG; (d) no enzyme. The enzyme reaction was conducted at 37 °C for 24 h. Observed cross-peaks labeled with TGase or PG are marked A, B, C, and D. dark. Proteolytic digestion was performed with V8 protease (enzyme/ substrate = 1:50) or trypsin (enzyme/substrate = 1:40) at 37 °C for 6 h. Protein digests were centrifuged at 12000 rpm for 5 min, and the supernatant peptides were filtered for LC-MS analysis. LC-MS Peptide Mapping. An HPLC-ESI-MS system of the LCMS-2010 (Shimadzu, Kyoto, Japan) was used for identification of the peptides obtained by V8 protease or trypsin digestion of αlactalbumin. A YMC-Pack Pro C18 column (150  4.6 mm, particles size = 5 μm) was used for separation. The solvents were 0.1% (v/v) trifluoroacetic acid (TFA) in water (solvent A) and 0.1% (v/v) TFA in 80% acetonitrile (solvent B). Elution was performed with a linear gradient of 5 55% solvent B over 120 min at a flow rate of 0.5 mL/min at 30 °C. The elution of peptides was monitored at 215 nm, and the sample injection volume was 60 μL. Mass spectrometric data were collected over the mass range of 50 2000, in a positive ion mode with a single-quadrupole mass analyzer. Analysis of proteolytic digest peptide molecular mass data was assisted by the Internet-based PeptideMass at EXPASY.24

’ RESULTS AND DISCUSSION Incorporation of 15N from 15NH4Cl into γ-Carboxyamide Groups of Glutamine Residues in α-Lactalbumin by PG. The

ELT developed by Shimba et al. is a stable isotope labeling method based on TGase reaction. The reaction in the solid-line box in Figure 1 shows the reaction mechanism for ELT. In the ELT with TGase, labeled ammonium ions react with the γcarboxyamide groups of glutamine residues and exchange with their amide groups. A large excess of 15N-labeled ammonium ions increases the incorporation of the 15N nucleus. The 15Nlabeled glutamine residues are observed in the 15N edited spectra using NMR. PG is an enzyme that catalyzes deamidation of protein-bound glutamine residues. We hypothesized that if the glutamine residues once labeled with 15N by TGase are also substrates for PG, the cross-peaks observed in the 1H 15N HSQC spectrum would vanish due to deamidation. To test this

Figure 3. Integrated intensity of the signals indicated by A, B, C, and D in Figure 2 and those for the corresponding signals in α-lactalbumin that reacted with TGase (a) and PG (b) were plotted against the reaction time.

hypothesis, we incubated the substrate solution containing αlactalbumin in 5% D2O/95% H2O in the presence of 0.2 M 15 NH4Cl and TGase at 37 °C. Figure 2a shows the 1H 15N HSQC spectrum of 15N-labeled α-lactalbumin after 24 h of reaction. Three pairs of cross-peaks were observed, showing that three glutamine residues were labeled with 15N. After that, PG was added to this reaction mixture, which was then incubated for an additional 24 h. There was a newly formed cross-peak, so the total number of cross-peaks was four, as shown in Figure 2b. This result was different from what we expected would occur. The possible reasons why glutamine residues labeled with 15N increased by adding PG were considered to be the following: (1) TGase reaction increased the number of 15N-labeled glutamine amide residues by prolonged reaction time by PG-induced conformational change in the substrate; (2) the signal positions of the three existing glutamine residues labeled with TGase were moved by PG-induced conformational change; or (3) PG increased the 15N-labeled glutamine residue. 12754

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry The integrated intensity of the existing three glutamine residues (signals A, B, and C) labeled with 15N by TGase gradually increased as the reaction proceeded (Figure 3a). This is because many more 15N-labeled ammonium ions exist than unlabeled ammonium ions, and the reaction system shifts the right direction, producing 15N-labeled glutamine residues (as shown in the solid-line box of Figure 1). Further incubation with TGase did not change signal A, B, or C (data not shown). By the addition of PG after TGase reaction, signal A increased, whereas signals B and C hardly changed. The integrated intensity of a newly labeled glutamine residue (signal D) gradually increased (data not shown). To confirm the possibility of reason 3, that is, that PG increased the 15N-labeled glutamine residue, we added PG without TGase. Three cross-peaks were observed in the 1 H 15N HSQC spectrum of α-lactalbumin after 24 h of reaction (Figure 2c), which means that three glutamine residues were labeled by PG in the presence of ammonium ions. In the absence of any enzymes, no cross-peaks were observed in the 1H 15N HSQC spectrum (Figure 2d), indicating that no glutamine residues were labeled. The result of Figure 2b is consistent with what occurs when PG labeling (Figure 2c) is added to TGase labeling (Figure 2a). Signals A and C were labeled by both TGase and PG. This suggests that little conformational change in the substrate was induced by PG. There is little probability of reason 1 or 2 being why glutamine residues labeled with 15N increased by the addition of PG. This15N-labeling with PG was considered to be the incorporation of 15N into γ-carboxyamide groups of glutamine residues, not to be the reverse reaction of deamidation in the presence of 15 NH4Cl. The reverse reaction of deamidation did not occur, which was confirmed by the following experiment. When Z-GluGly, the deamidation product of Z-Gln-Gly, was incubated with PG in the presence of a large excess of ammonium ions, the formation of Z-Gln-Gly was not detected at all (data not shown). As shown in Figure 3b, the integrated intensity of each signal observed in Figure 2c increased from the beginning of the reaction. This result strongly suggests not the reverse reaction of deamidation but the incorporation of 15N into γ-carboxyamide groups of glutamine residues as well as TGase. This observation can be explained as follows. In the absence of NH4+, a nucleophilic attack of H3O+ for the carbonyl groups in aqueous solution occurs after PG and α-lactalbumin form the intermediate in the transition between glutamine residues. However, in the presence of 15N-labeled ammonium ions, the nucleophilic attack of NH4+ for carbonyl groups dominantly occurs instead of the attack of H3O+. On the bsis of this consideration, we suggest the reaction scheme of PG in the presence of the 15N-labeled ammonium ions as shown in Figure 1. In the presence of 15N-labeled ammonium ions, PG catalyzes the incorporation of 15N into γ-carboxyamide groups of glutamine residues similarly to TGase. Besides, the deamidation reaction takes place. Both 15N-labeled and unlabeled glutamine residues can be deamidated by PG; however, the 15Nlabeled glutamine residues are predominantly deamidated because there exists an excess of 15N-labeled ammonium ions in the reaction system and the reaction shifts in the forward direction. Analysis of the PG Reactivity for Glutamine Residues by the ELT. The time course of the integrated intensity of signal D in Figure 3b is the important feature in the PG reaction under an excess of 15N-labeled ammonium ions. Although the integrated intensity of two signals, A and C, showed a steady increase as the PG reaction proceeded, the profile of signal D showed a different

ARTICLE

Figure 4. 1H 15N HSQC spectra of α-lactalbumin incubated at 37 °C for 24 h with 15NH4Cl in the presence of PG: (a) enzyme/substrate ratio = 1:1000 (w/w); (b) enzyme/substrate ratio = 1:10000 (w/w). Observed cross-peaks labeled with PG are marked A E.

pattern. The integrated intensity of signal D increased more rapidly than any other signal up to 6 h immediately after the beginning of the reaction and then gradually decreased. This result is considered to be a result of deamidation of glutamine residues once labeled with 15N by PG. From this result, we inferred that the glutamine residue D is the most reactive and that eventually 15N-labeled glutamine residues are converted to glutamic acid residues, accompanying the NMR signal disappearance. We confirmed that signal D actually disappeared in the 1 H 15N HSQC spectrum after 24 h of incubation at 37 °C when a 5-fold amount of the enzyme was added (Figure 4a). The reactivity of glutamine residue C seems to be the lowest, because its integrated intensity increased much more slowly than that of the other two signals (Figure 3b). In response to the addition of a 5-fold amount of the enzyme, as shown in Figure 4a, the intensity of signal C in the 1H 15N HSQC spectrum increased, whereas, in contrast, signal C was not observed after the addition of half the amount of the enzyme (Figure 4b). These results show that the appearance of the signal on the 1H 15N HSQC spectrum is dependent on the enzyme dosage. When α-lactalbumin was incubated with PG at the enzyme/substrate ratios of 1:10000, 1:5000, and 1:1000 (w/w) at 37 °C for 24 h, the number of PGreactive glutamine residues was estimated as 2, 3, and 5, respectively. With regard to signal C, it had seemed possible that the deamidation reaction occurred simultaneously and there was not much 15N amide left; however, the above results denied such a possibility. By combining information from both the number of signals and the 15N labeling speed calculated from the signal intensity in the NMR spectrum, it is possible to infer the PG reactivity of each glutamine residue, although the integrated signal intensity indicates the sum of incorporation of 15N nuclei into glutamine residue and deamidation. Reactivity of Glutamine Residues for PG and TGase. We found that the ELT can be applied to PG, so that it became possible to infer not only the reactivity of each glutamine residue for PG but to compare the substrate specificities and reaction rates between PG and TGase. Signals A and C shown in Figure 2c were also observed in Figure 2a, whereas signal D was undetected in Figure 2a. Signal B, shown in Figure 2a, was not observed in Figure 2c. These results suggest that glutamine residues B and D can react with just TGase and PG, respectively. We can easily find that there are some differences in the site specificity for the glutamine residue between TGase and PG. With regard to the reaction rate of glutamine residues, in the PG reaction, the integrated intensity, as shown in Figure 3b, 12755

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry

ARTICLE

Table 1. Reaction Samples Prepared for Peptide Mapping Using LC-MSa sample

S/E ratio

reaction time (h)

released ammonia (mmol/L)

2

5000/1

5

0.84 (0.09)

20.1

3

5000/1

24

2.03 (0.31)

48.3

4

1000/1

24

3.76 (0.15)

89.4

1

a

0

deamidation degree (%) 0

Reaction mixtures containing 10 mg/mL α-lactalbumin were incubated at 37 °C.

Figure 5. Primary sequence of bovine α-lactalbumin of 123 amino acid residues. Solid lines show the amino acid sequences determined by V8 digestion. The down arrows indicate the cleavage sites of native αlactalbumin, and the up arrows indicate the cleavage sites of PGmodified α-lactalbumin. The amino acid sequence marked by a broken underline is from tryptic digestion.

indicates the sum of incorporation of 15N into γ-carboxyamide groups of glutamine residues and deamidation. Therefore, it is difficult to compare directly the reaction rate of each glutamine residue between TGase and PG. However, we at least assert that the reactivity of glutamine A for PG is much higher than that for TGase, judging from the integrated intensity of signal A in Figure 3. Reactivity of PG for Glutamine Residues in α-Lactalbumin in the Absence of Ammonium Ions. We can examine the number of PG-reactive glutamines and each reaction rate using the ELT, provided there exists a large excess of ammonium ions in the reaction system. However, in actual food systems in which there are few or no ammonium ions, the PG reactivity observed in the ELT is not necessarily similar. Hence, we investigated the effect of ammonium ions on the PG reactivity by peptide mapping using LC-MS. The samples for peptide mapping analysis were prepared as shown in Table 1. The reaction mixtures containing α-lactalbumin were incubated with PG at the enzyme/substrate ratio of 1:1000 or 1:5000 (w/w) at 37 °C in the absence of ammonium ions. At an enzyme/substrate ratio of 1:5000 (w/w), with the same dosage level of PG as described for the ELT experiment (Figure 3), the deamidation degrees were 20 and 48% after 5 and 24 h of incubation, respectively. At an enzyme/substrate ratio of 1:1000 (w/w), the deamidation degree reached approximately 90% after 24 h of incubation, suggesting that almost all six glutamine residues in α-lactalbumin finally could be modified by PG. The reaction mixtures were subjected to enzymatic digestion after the reduction of disulfide bonds with DTT and the following carboxymethylation. Figure 5 shows the primary structure of α-lactalbumin, consisting of 123 amino acid residues. Solid lines below the sequence in Figure 5 show the sequenced peptides obtained from V8 protease digests, and the broken line shows the sequenced peptides obtained from trypsin digests. Figure 6 shows the elution patterns of the V8 protease digest of α-lactalbumin untreated (upper panel) and treated with PG at the enzyme/substrate ratio of 1:1000 (w/w) for 24 h at 37 °C (lower panel). Many more peaks were observed in the

chromatograms of PG-modified α-lactalbumin than in that of the untreated sample, which indicates that glutamine residues of the substrate were converted to glutamic acids by the PG-catalyzed deamidation reaction, resulting in the appearance of new peaks, because V8 protease cleaves to the carboxyl side of glutamic acid. Table 2 summarizes the results of LC-MS analysis of nine peaks detected in chromatograms, and peaks 1 9 correspond to the numbers shown in Figure 6. Considering the form of the MS ion peaks and retention time shifts on reverse-phased chromatography (Supporting Information), we see that peak 2 is changed to peak 5, which includes deamidated Gln2. Peaks 8, 3, and 9 include deamidated Gln39, Gln43, and Gln54, respectively. Peaks 6 and another peak 9 (peak 9 contains two fragments) are formed from deamidated Gln117. Therefore, it was confirmed that five glutamine residues (Gln2, Gln39, Gln43, Gln54, and Gln117) among six in α-lactalbumin were deamidated by V8 protease digestion. Because the cleavage peptide containing Gln65 could not be found from V8 protease digests, trypsin digestion was then performed. Peak T1, containing Gln65, was obtained in the sample treated with PG at the enzyme/substrate ratio of 1:5000 (w/w) for 5 h at 37 °C (Figure 7a). Peak T2, containing deamidated Gln65, was obtained by the sample treated with PG at the enzyme/substrate ratio of 1:1000 (w/w) for 24 h at 37 °C (Figure 7c). Thus, all glutamine residues, that is, Gln2, Gln39, Gln43, Gln54, Gln65, and Gln117, in α-lactalbumin could be modified by PG. This result agrees with that of the deamidation degree described above. PG Reactivity of Each Glutamine Residue in α-Lactalbumin. From the result of the LC-MS analysis of two reaction samples, which were treated with PG at the enzyme/substrate ratio of 1:5000 for 5 and 24 h, we inferred the PG reactivity of each glutamine residue in α-lactalbumin. In these samples, peaks 6, 8, and 9 were clearly detected, suggesting that Gln39, Gln54, and Gln117 could be modified with PG. Among them, Gln117 was considered to be the most reactive, because a remarkable increase in peaks 6 and 9 was detected, even in the sample having a lower deamidation degree (enzyme/substrate = 1:5000, for 5 h at 37 °C). The mass spectrum of peak 9 showed mainly m/z 849.5 and 425.4 ion peaks, indicating the peptide WLCEKL (positions 118 123) was contained. This result means that a peptide containing Gln54 was hardly detected. The reactivity of Gln39 was thought to be relatively high, because the definite increase of peak 8 in Figure 6 was observed with the increase in reaction time at the enzyme/substrate ratio of 1:5000. On the other hand, peaks 3 and 5 were not observed clearly in these two samples, indicating that Gln2 and Gln43 were little modified by PG. Figure 7b shows the formation of the small peak T2 from the sample treated with PG at the enzyme/substrate ratio of 1:5000 for 24 h at 37 °C, indicating that Gln65 was slightly deamidated by PG, although it was scarcely modified in the 5 h reaction sample. 12756

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry

ARTICLE

Figure 6. Elution patterns of the V8-digest of α-lactalbumin, treated without enzyme and with PG at 37 °C. The enzyme/substrate ratio and reaction time are indicated on each chromatogram. The retention time range of the chromatograms on the left side shows 30 50 min, and that on the right side is 60 80 min. Typical peaks were numbered in order of elution.

Table 2. Assignment of Glutamine Residues Deamidated by PG in α-Lactalbumin Using LC-MS Analysis

a Peaks were extracted from V8 protease digests of deamidated α-lactalbumin by RP-HPLC as shown in Figure 6. b The m/z of each peak was identified by analysis of the mass spectrum of the peak. Mass spectra were acquired in positive ion mode.

Although it is difficult to assign the 15NH signals observed in the 1H 15N HSQC spectrum to specific glutamine residues in the substrate protein for TGases or PG, some 15NH signals corresponding to glutamine residues modified by PG were deduced from the result of LC-MS peptide mapping. Among three glutamine residues (signals A, C, and D in Figures 2c and 3), signal D appears to be the most reactive glutamine residue,

namely, Gln117. Gln39 seems to have the second-highest reactivity for PG, so we assume that signal A is Gln39. Because Gln43, Gln54, and Gln65 showed a low reactivity for PG in the absence of ammonium ions, it is difficult to distinguish which was glutamine residue C. It appeared that Gln117 was the most reactive residue for PG, which can be explained as follows. Gln117 is located in the 12757

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry

ARTICLE

Figure 7. Mass chromatograms (m/z 1004 and 670) of a peptide obtained by tryptic digestion of deamidated α-lactalbumin incubated with PG at 37 °C at the enzyme/substrate ratios of 1:5000 (w/w) for 5 h (a), 1:5000 (w/w) for 24 h (b), and 1:1000 (w/w) for 24 h (c). Mass spectra of peaks eluting at 48.8 and 49.6 min (peaks T1 and T2, respectively) are shown on the right sides of panels a and c, respectively.

C-terminus of α-lactalbumin and is also a part of aromatic cluster I, which consists of residues Phe31, His32, Gln117, and Trp118.25 This region has been noted for its conformational variability and enhanced mobility.26 The side-chain environments of Phe31 and Gln117 are relatively exposed (40 70% of surface area accessible).26 Also, Trp118 adjacent to Gln117 is reasonably exposed to the solvent.27 Therefore, we consider Gln117 to be located in a region more accessible to PG than other glutamine residues. Deamidation can cause a change in the functional properties of proteins. α-Lactalbumin has an important function, as it is one of the components of the lactose synthase system.28 In aromatic cluster I, changes at Gln117 or Trp118 were found to specifically reduce the affinity for β-galactosyltransferase.29 We suppose that deamidation of Gln117 by PG causes a reduction in binding ability to β-galactosyltransferase. Further investigations are needed to confirm the effects of PG on the functionalities of α-lactalbumin.

Gu et al. identified PG-modified glutamine residues of αlactalbumin in the molten globule state to PG.7 In their experimental data, there is a lack of reliable data for Gln117. They showed that no modification of Gln117 was confirmed by the C-terminal sequence of the whole molecule of deamidated α-lactalbumin. Nevertheless, it is uncertain how they can tell the difference between Glu117 and Glu121 (both are located at the C-terminus of α-lactalbumin) when Glu117 was partly formed by deamidation of Gln117 by PG reaction. In their experiment, α-lactalbumin was incubated with carboxypeptidase Y without prior disulfide reduction. Under such conditions, we think that the yield of C-terminal amino acids would be very low, and consequently quantitative data of Gln/Glu117, located at the seventh amino acid from the C-terminus, would scarcely be obtainable. 12758

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry Effect of Ammonium Ions on the Reactivity of PG for Glutamine Residues of α-Lactalbumin. We examined the PG

reactivity of glutamine residues in α-lactalbumin in the presence and absence of ammonium ions. In the ELT, three glutamine residues were modified by PG in the presence of ammonium ions. Under the same dosage of PG, in the absence of ammonium ions, five glutamine residues, not including Gln2, were thought to be attacked by PG. The result of peptide mapping showed that all six glutamine residues were modified when the enzyme to substrate ratio was 1:1000, whereas the results of the ELT showed that only five signals were observed in the 1H 15N HSQC spectrum. Thus, we found that the number of glutamine residues attacked by the enzyme was affected by the existence of ammonium ions. This can be explained in terms of both structural changes induced by the deamidation and competitive inhibition. In the absence of ammonium ions, only the deamidation reaction occurs. Then, once the glutamine residue is deamidated, the newly formed carboxyl group enhances the unfolding of the protein via electrostatic repulsion, thereby increasing the possibility of modification of glutamine residues by the enzyme. However, when ammonium ions as a competitive inhibitor are present in abundance, they cause a substantial delay of the deamidation reaction, because the nucleophilic attack of NH4+ for carbonyl groups dominantly occurs instead of H3O+. In this case, the exchange reaction with ammonium ions has no influence upon the substrate conformation, because it retains the same chemical structure; therefore, the accessibility of glutamine residues toward the enzyme is not changed. This may be why the glutamine residues attacked by PG were limited under the high concentration of ammonium ions.

’ ASSOCIATED CONTENT

bS Supporting Information. Form of MS ion peaks and retention time shifts on reverse-phased chromatography. This material is available free of charge via the Internet at http:// pubs.acs.org. ’ AUTHOR INFORMATION Corresponding Author

*Phone: +81-44-223-4174. Fax: +81-44-246-6081. E-mail: [email protected].

’ ACKNOWLEDGMENT We are grateful to Maiko Shinagawa for NMR measurements. ’ ABBREVIATIONS USED DTT, dithiothreitol; EDTA, ethylenediaminetetraacetic acid; ELT, enzymatic labeling technique; HPLC-ESI, high-performance liquid chromatography electrospray ionization; HSQC, heteronuclear single-quantum coherence; NADH, reduced nicotinamide adenine dinucleotide; NMR, nuclear magnetic resonance; PG, proteinglutaminase; TFA, trifluoroacetic acid; TGase, transglutaminase. ’ REFERENCES (1) Hamada, J. S. Deamidation of food proteins to improve functionality. Crit. Rev. Food Sci. Nutr. 1994, 34, 283–292.

ARTICLE

(2) Schwenke, K. D. Enzyme and chemical modification of proteins. In Food Proteins and Their Applications; Damodaran, S., Paraf, A., Eds.; Dekker: New York, 1997; pp 393 424. (3) Yamaguchi, S.; Yokoe, M. A novel protein-deamidating enzyme from Chryseobacterium proteolyticum sp. nov., a newly isolated bacterium from soil. Appl. Environ. Microbiol. 2000, 66, 3337–3343. (4) Yamaguchi, S.; Jeenes, D. J.; Archer, D. B. Protein-glutaminase from Chryseobacterium proteolyticum, an enzyme that deamidates glutaminyl residues in proteins. Purification, characterization and gene cloning. Eur. J. Biochem. 2001, 268, 1410–1421. (5) Yong, Y. H.; Yamaguchi, S.; Gu, Y. S.; Mori, T.; Matsumura, Y. Effects of enzymatic deamidation by protein-glutaminase on structure and functional properties of α-zein. J. Agric. Food Chem. 2004, 52, 7094–7100. (6) Yong, Y. H.; Yamaguchi, S.; Matsumura, Y. Effects of enzymatic deamidation by protein-glutaminase on structure and functional properties of wheat gluten. J. Agric. Food Chem. 2006, 54, 6034–6040. (7) Gu, Y. S.; Matsumura, Y.; Yamaguchi, S.; Mori, T. Action of protein-glutaminase on α-lactalbumin in the native and molten globule states. J. Agric. Food Chem. 2001, 49, 5999–6005. (8) Miwa, N.; Yokoyama, K.; Wakabayashi, H.; Nio, N. Effect of deamidation by protein-glutaminase on physicochemical and functional properties of skim milk. Int. Dairy J. 2010, 20, 393–399. (9) Folk, J. E.; Finlayson, J. S. The ε-(γ-glutamyl) lysine cross-link and catalytic role of transglutaminase. Adv. Protein Chem. 1977, 31, 1–133. (10) Folk, J. E. Transglutaminase. Annu. Rev. Biochem. 1980, 49, 517–531. (11) Lorand, L.; Losowsky, M. S.; Miloszewski, K. J. M. Human factor XIII: Fibrin-stabilizing factor. Prog. Hemostasis Thromb. 1980, 5, 245–290. (12) Yee, V. C.; Pedersen, L. C.; Le Trong, I.; Bishop, P. D.; Stenkamp, R. E.; Teller, D. C. Three-dimensional structure of a transglutaminase: human blood coagulation factor XIII. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 7296–7300. (13) Pedersen, L. C.; Yee, V. C.; Bishop, P. D.; Le Trong, I.; Teller, D. C.; Stenkamp, R. E. Transglutaminase factor XIII uses proteinase-like catalytic triad to cross-link macromolecules. Protein Sci. 1994, 3, 1131–1135. (14) Shimba, N.; Yamada, N.; Yokoyama, K.; Suzuki, E. Enzymatic labeling of arbitrary proteins. Anal. Biochem. 2002, 301, 123–127. (15) Shimba, N.; Yokoyama, K.; Suzuki, E. NMR-based screening method for transglutaminase: rapid analysis of their substrate specificities and reaction rates. J. Agric. Food Chem. 2002, 50, 1330–1334. (16) Shimba, N.; Shinohara, M.; Yokoyama, K.; Kashiwagi, T.; Ishikawa, K.; Ejima, D.; Suzuki, E. Enhancement of transglutaminase activity by NMR identification of its flexible residues affecting the active sites. FEBS Lett. 2002, 517, 175–179. (17) Acharya, K. A.; Stuart, D. I.; Walker, N. P. C.; Lewis, M.; Phillips, D. C. Structure of baboon α-lactalbumin at 1.7 Å resolution. J. Mol. Biol. 1989, 208, 99–127. (18) Kikuchi, Y.; Itaya, H.; Date, M.; Matsui, K.; Wu, L.-F. Production of Chryseobacterium proteolyticum protein-glutaminase using the twin-arginine translocation pathway in Corynebacterium glutamicum. Appl. Microbiol. Biotechnol. 2008, 78, 67–74. (19) Ando, H.; Adachi, M.; Umeda, K.; Matsuura, A.; Nonaka, M.; Uchino, R.; Tanaka, H.; Motoki, M. Purification and characterization of a novel transglutaminase derived from microorganisms. Agric. Biol. Chem. 1989, 53, 2613–2617. (20) Wishart, D. S.; Bigam, C. G.; Yao, J.; Abildgaard, F.; Dyson, H. J.; Oldfield, E.; Markley, J. L.; Skyes, B. D. 1H, 13C and 15N chemical shift referencing in biomolecular NMR. J. Biomol. NMR 1995, 6, 135–140. (21) XWINNMR software Manual (Bruker Co); http://www.bruker.com. (22) Bodenhausen, G.; Ruben, D. J. Natural abundance nitrogen-15 NMR by enhanced heteronuclear spectroscopy. Chem. Phys. Lett. 1980, 69, 185–189. (23) Sklenar, V.; Piotto, M.; Leppik, R.; Saudek, V. Gradient-tailored water suppression for 1H-15N HSQC experiments optimized to retain full sensitivity. J. Magn. Reson. Ser. A 1993, 102, 241–245. 12759

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760

Journal of Agricultural and Food Chemistry

ARTICLE

(24) http://tw.expasy.org/tools/peptidemass.html. (25) Ramakrishnan, B.; Qasba, P. K. Crystal structure of lactose synthase reveals a large conformational change in its catalytic component, the β-1,4-galactosyltransferase-1. J. Mol. Biol. 2001, 310, 205–218. (26) Pike, A. C.; Brew, K.; Acharya, K. R. Crystal structures of guinea-pig, goat and bovine α-lactalbumin highlight the enhanced conformational flexibility of regions that are significant for its action in lactose synthase. Structure 1996, 4, 691–703. (27) Chakraborty, S.; Ittah, V.; Bai, P.; Luo, L.; Haas, E.; Peng, Z. Structure and dynamics of the α-lactalbumin molten globule: fluorescence studies using proteins containing a single tryptophan residue. Biochemistry 2001, 40, 7228–7238. (28) Hill, R. L.; Brew, K. Lactose synthetase. Adv. Enzymol. 1975, 43, 411–490. (29) Grobler, J. A.; Wang, M.; Pike, A. C.; Brew, K. Study by mutagenesis of the roles of two aromatic clusters of α-lactalbumin in aspects of its action in the lactose synthase system. J. Biol. Chem. 1994, 269, 5106–5114.

12760

dx.doi.org/10.1021/jf2028895 |J. Agric. Food Chem. 2011, 59, 12752–12760