Determination of the influence of the pH of hydrolysis on enzyme selectivity of Bacillus licheniformis protease towards whey protein isolate

Determination of the influence of the pH of hydrolysis on enzyme selectivity of Bacillus licheniformis protease towards whey protein isolate

International Dairy Journal 44 (2015) 44e53 Contents lists available at ScienceDirect International Dairy Journal journal homepage: www.elsevier.com...

1MB Sizes 0 Downloads 36 Views

International Dairy Journal 44 (2015) 44e53

Contents lists available at ScienceDirect

International Dairy Journal journal homepage: www.elsevier.com/locate/idairyj

Determination of the influence of the pH of hydrolysis on enzyme selectivity of Bacillus licheniformis protease towards whey protein isolate  a, Stefano Sforza a, b, Peter A. Wierenga a, Harry Gruppen a, * Claire I. Butre a b

Laboratory of Food Chemistry, Wageningen University, Postbus 17, 6700 AA Wageningen, The Netherlands Department of Food Science, University of Parma, Parco Area delle Scienze 59/A, 43124 Parma, Italy

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 September 2014 Received in revised form 25 November 2014 Accepted 8 December 2014 Available online 30 December 2014

Enzymatic protein hydrolysis is typically described by the degree of hydrolysis and by the enzyme specificity. While the specificity describes which cleavage sites can potentially be cleaved, it does not describe which are preferred. To identify the relative rate at which each individual cleavage site is hydrolysed, enzyme selectivity has recently been introduced. To test the effect of pH on selectivity, whey protein isolate (WPI) was hydrolysed with Bacillus licheniformis protease (BLP) at pH 7.0, 8.0 or 9.0. At all pH values, large differences in the enzyme selectivity (from <0.004% to 27%) towards the different cleavage sites of b-lactoglobulin were observed. The changes in selectivity as a function of pH were significant (up to 80% increase or decrease). This significant variation in selectivity shows that this parameter may be useful in understanding peptide release kinetics in enzymatic hydrolysis under different conditions. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction While enzymatic hydrolysis of proteins is often studied, little is known about the mechanism of hydrolysis (i.e., which peptides are formed and when). The characterisation of protease activity, for instance, often starts by identifying the optimal pH and temperature for the hydrolysis. The optimum is then defined as the conditions under which the highest conversion after a set amount of time is observed. In this way, the optimal pH of Bacillus licheniformis protease (BLP) was determined to be pH 8.0 (Svendsen & Breddam, 1992). The activity can also be determined from the rate of protein hydrolysis using plots of the degree of hydrolysis (DH) versus time. In this way, an increase in activity of a protease from Bacillus subtilisis with increasing pH (from 8 to 10) towards the hydrolysis of lez-Tello, & whey proteins was determined (Camacho, Gonza Guadix, 1998). Neither method, however, yields information on the mechanism of hydrolysis or on the peptides that are released. A first indication of the mechanism of hydrolysis is obtained from the decrease of the amount of intact protein as a function of

* Corresponding author. Tel.: þ31 317 483211. E-mail address: [email protected] (H. Gruppen). http://dx.doi.org/10.1016/j.idairyj.2014.12.007 0958-6946/© 2014 Elsevier Ltd. All rights reserved.

the DH. It has, for instance, been shown that, during hydrolysis of haemoglobin by pepsin, the proportion of remaining intact protein (at any DH) was higher at pH 4.5 than at pH 3.5 (Dubois, NedjarArroume, & Guillochon, 2005). A slow degradation of the intact protein (as function of DH) implies that the enzyme has more difficulty in hydrolysing the intact protein than in hydrolysing the intermediate peptides formed. On the other hand, a rapid degradation of the intact protein indicates that no such hindrance is present. These two extreme cases have been referred to as one-byone and zipper mechanisms, respectively (Adler-Nissen, 1986). The accessibility of the intact protein is important, but still does not provide information on which peptide will be released preferentially. For such information, the selectivity of the enzyme towards the different cleavage sites needs to be determined. The selectivity was recently defined as the relative rate at which the enzyme cleaves each individual cleavage site compared with the , total rate of hydrolysis of all cleavage sites in the protein (Butre Sforza, Gruppen, & Wierenga, 2014a). In this way, the selectivity is more detailed than the enzyme specificity, which only describes the amino acids after which an enzyme will cleave a peptide bond. An indication that the pH of hydrolysis influences the enzyme selectivity was found in the tryptic hydrolysis of b-casein (Vorob'ev, , 2000). It was shown that the Dalgalarrondo, Chobert, & Haertle kinetics of peptide formation and release of peptides with Arg

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

residues were not affected by the changes in pH, while, for peptides after Lys, the pH of hydrolysis did influence the peptide kinetics. In this example, the change in selectivity is explained by changes in the charge state of the amino acid on the cleavage site. For other enzymes, the changes in selectivity may be due to the charges of amino acids in the neighbourhood of the cleavage site. Moreover, changes in pH can also affect the charge state of amino acids in the active site of the enzyme, although it is currently not well understood if this would indeed change the enzyme selectivity. To determine the influence of the pH on the selectivity of the enzyme, whey protein isolate (WPI) was hydrolysed by BLP at different pH values. In addition to the selectivity, the hydrolysis was characterised by three other descriptors: (i) the rate of hydrolysis; (ii) the proportion of remaining intact b-lactoglobulin as a function of the DH; and (iii) peptide molecular mass distribution. For the latter, the peptides obtained at the different DH values were annotated and quantified. The selectivity was then determined from the change in concentration of each individual peptide with time, i.e., peptide release kinetics. 2. Materials and methods

45

 et al., 2014). The degree of described in a previous study (Butre hydrolysis (DH) was calculated based on the added volume of NaOH using 1/a ¼ 1.20 at 40  C and pH 8.0, 3.0 at pH 7.0 and 1.02 at pH 9.0  et al., 2012). The and htot(WPI) ¼ 8.5, as described previously (Butre experiments were performed once. It has been previously shown that the data from repeated hydrolysis experiments (6 times) at several conditions showed that the error in DH (standard deviation/  et al., mean  100%) at each point in time is lower than 10% (Butre 2012). Samples were taken during the hydrolysis at different degrees of hydrolysis (1.5, 3, 4.5, 6% for all conditions and at 7% for the hydrolysis performed at pH 8.0 and pH 9.0). The samples were first centrifuged (19,000  g, 5 min, 15  C), supernatant and pellet were separated, and the supernatants were inactivated by adjusting the pH to 2.0 with 5 M HCl. The pH was readjusted to 8.0 with NaOH after at least 10 min of inactivation before storage of the samples at 20  C. The initial rate of hydrolysis of BLP towards WPI was determined from the slope of the linear portion of the DH versus time curves. The initial rate of hydrolysis is expressed here as the concentration of bonds cleaved per mg of enzyme per time in seconds (mol L1 mg1 s1).

2.1. Materials Bipro, a commercial whey protein isolate (WPI) was obtained from Davisco Foods International Inc. (Le Sueur, MN, USA). The WPI contained (by weight) 74.0% b-lactoglobulin, 12.5% a-lactalbumin, 5.5% bovine serum albumin and 5.5% immunoglobulin, according to , specifications of the supplier. The protein content (N  6.32) (Butre Wierenga, & Gruppen, 2012) of the powder was 93.4% (w/w) as determined by the Dumas method. BLP (4.5%, w/w, protein, N  6.25, 0.3 AU mg1 min1 as determined by the azocasein assay of Akpinar & Penner, 2001), specific for Glu-X bonds and for Asp-X bonds (Breddam & Meldal, 1992), was obtained from Novozymes (NS-37005) (Bagsvaerd, Denmark). The BLP was partly insoluble and was fractionated by first solubilising the BLP powder in 10 mM NaH2PO4 pH 5.8, and stirring overnight at 4  C. The suspension was centrifuged (4000  g, 10 min, 25  C) and the supernatant obtained was extensively dialysed against 150 mM NaCl, followed by dialysis against demineralised water using cellulose dialysis membranes (cut-off 12e14 kDa). The freeze dried material was found to contain 60% (w/w) protein (N  6.25), of which 78% is the enzyme and 14%  et al., is the pro-peptide (6.9 kDa) as determined previously (Butre 2014a). An activity of 3.9 AU mg1 min1 was determined by the azocasein assay. All other chemicals were of analytical grade and purchased from Sigma (St. Louis, MO, USA) or Merck (Darmstadt, Germany). 2.2. Hydrolysis Protein solutions were prepared by dispersing WPI powder at a concentration of 45% (w/v) in Millipore water, followed by stirring overnight at 4  C. This sample preparation was chosen to have the , Sforza, Gruppen, same conditions as previous experiments (Butre & Wierenga, 2014b). Insoluble parts were removed by centrifugation (4000  g, 30 min, 20  C) and the supernatant obtained (30%,  et al., w/v) was diluted to 1% (w/v) as determined by UV280 (Butre 2012). The 1% (w/v) protein solutions were obtained from separate dispersions. Hydrolyses were performed using a pH-stat. Solutions of WPI at 1% (w/v) were hydrolysed at 40  C and pH 7.0 or 9.0 using 0.2 M NaOH to keep the pH constant. Protein solutions (10 mL) were equilibrated for least 15 min at 40  C and adjusted to pH 7.0 or 9.0 before addition of BLP dissolved at 5% (w/v) in Millipore water (0.30 mL of enzyme mg1 protein). The results of hydrolysis at pH 8.0 were obtained in the same conditions and

2.3. Reversed-phase ultra-high performance liquid chromatography Samples obtained during hydrolysis were analysed on an Acquity ultra-high performance liquid chromatography (UPLC) System (Waters, Milford, MA, USA) using an reversed-phase (RP) Acquity UPLC BEH 300C18 column (2.1  150 mm, 1.7 mm particle size) with an Acquity BEH C18 Vanguard pre-column (2.1  50 mm, 1.7 mm particle size). Eluent A was 1% (v/v) acetonitrile (ACN) containing 0.1% (v/v) trifluoroacetic acid (TFA) in Millipore water and eluent B was 100% ACN containing 0.1% (v/v) TFA. To reduce disulphide bridges and to facilitate peptide annotation, samples were first incubated for 2 h with 100 mM dithiothreitol (DTT) in 50 mM Tris HCl, pH 8.0, at a concentration of 0.5% (w/v). After incubation, samples were further diluted in eluent A to a final concentration of 0.1% (w/v) and centrifuged (19,000  g, 10 min, 20  C). Samples (4 mL) were injected into the column, which was thermostated at  et al., 2014a). 40  C (Butre Two different elution profiles were used in this study as  et al., 2014a), one was used to deterdescribed previously (Butre mine the amount of remaining b-lactoglobulin (A þ B) and a-lactalbumin and the other one was used for the separation of peptides. Detection was performed using a PDA, which was scanning the absorbance from 200 to 400 nm at a 1.2 nm resolution, with 20 points per second. 2.4. Electron spray ionisation time of flight mass spectrometry The mass spectra of the hydrolysates were determined with an online Synapt high definition mass spectrometer (Waters, Milford, MA, USA), coupled to the RP-UPLC system, equipped with a z-spray electrospray ionisation (ESI) source, a hybrid quadrupole and an orthogonal time-of-flight (Q-TOF) using the same parameters as in  et al., 2014a). UV and mass spectrometry a previous study (Butre MS data were acquired using MassLynx software v4.1 (Waters). 2.5. Peptide identification and quantification The proportion of remaining intact a-lactalbumin and b-lactoglobulin (A and B) was determined by dividing the area of the UV214 peak obtained for each protein at different DH of hydrolysis by the peak area obtained for each intact protein before hydrolysis.

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

46

For peptide identification, internal lock mass was applied on every chromatogram for MS and MS/MS data with two previously identified masses chosen among the most abundant masses, of which the sequences were identified manually. The identification of MS and MSe ion peaks was done by Biopharmalynx 1.3 software  et al., 2014a). For the with the parameters described before (Butre total peptide quantification, peaks in the UV214 chromatogram were included until 90e95% of the total UV214 area was included. Peptides from b-lactoglobulin and a-lactalbumin, as well as the intact protein, were quantified using UV214. In case of co-eluting peptides from either b-lactoglobulin or a-lactalbumin, the UV peak area was divided over the co-eluting peptides in a ratio based on the MS  et al., 2014a). Only peptides from b-lactoglobulin intensity (Butre were considered for the determination of the selectivity. The quantification of the peptides was based on the UV signal at 214 nm, using Equation (1): 6

Cpeptide ¼ 1*10

! A214 Q ε214 lVinj kcell

(1)

where Cpeptide (mM) is the concentration of peptide, A214 (AU min) is the UV peak area at 214 nm, Vinj (mL) is the volume of sample injected, Q the flow rate in mL min1, and l is the path length of the UV cell, which is 1 cm according to the manufacturer. The value for the cell constant kcell was previously determined to be 0.66 using pure peptides, but depends on the geometry of the UV detector used (Kosters, Wierenga, De Vries, & Gruppen, 2011). The molar extinction coefficient at 214 nm of each peptide (ε214) was calculated as described before (Kuipers & Gruppen, 2007).

The peptide sequence coverage was calculated to be 96 ± 3% on average. To determine the completeness of the quantification, the molar sequence coverage was calculated. For that, the concentration of each amino acid Cn was calculated using the molar concentration of each individual (annotated) peptide Cpeptide. With this, each individual amino acid of the primary sequence of blactoglobulin was quantified in the hydrolysate. The molar sequence coverage is calculated using Equation (4):

0 B B Molar sequence coverageð%Þ ¼ B B1  @

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi1 P ðCn C0 Þ2 C ð#AA1Þ C C  100 C C0 A

(4)

where Cn (mM) is the concentration of each individual amino acid n in the protein obtained from quantification of the peptides; C0 (mM) is the initial concentration of the protein and #AA is the number of  et al., amino acids in the sequence of the parental protein (Butre 2014a). In addition, the degree of hydrolysis (DH) was calculated from the molar quantification of all peptides using Equation (5):

P DH ð%Þ ¼

Cpeptides  C0  100 Co  #aa  Co

(5)

where Cpeptide and C0 are the concentrations of all peptides in the hydrolysates and the initial concentration of protein (mM), respectively, and #aa is the number of amino acids in the protein. 2.8. Determination of the enzyme selectivity

2.6. Molecular mass distribution profile The weight-based concentration of each peptide resulting from hydrolysis of b-lactoglobulin was calculated from the molar concentration. The peptides were divided into molecular mass ranges (<1 kDa, 1e1.5 kDa, 1.5e2 kDa, 2e2.5 kDa, 2.5e3 kDa, >3 kDa and intact protein). The weight-based concentrations of the different peptides were summed in each molecular mass range. 2.7. Quality of the annotation and of the quantification The quality of the peptide identification and of the quantification was described by three different parameters. First, the amino acid sequence coverage was determined by calculating the number (#) of unique annotated amino acid divided by the total number (#) of amino acids present in the parental protein sequence:

Amino acid sequence coverage ¼

# unique annotated amino acids # amino acids in protein sequence (2)

This was found to be 100% for all samples. To take into account the fact that each part of the sequence is covered by different peptides, the peptide sequence coverage was determined as a second quality parameter on the peptide annota et al., 2014a). This is defined as the number of amino tion (Butre acids (#AA) in annotated peptides divided by the total number of amino acids (#AA) present in all peptides (annotated peptides and missing peptides) as in Equation (3)

Peptide sequence coverage ¼

# AA ðannotated peptidesÞ # AAðannotated peptidesÞ þ # AA ðmissing peptidesÞ

(3)

To describe the selectivity of the enzyme, the rate at which individual peptide bonds are hydrolysed (e.g., between Glu-45 and Leu-46), was determined. To do this, first the concentrations of all peptides originating from hydrolysis of the same peptide bond (Ci,t) in the parent molecule are calculated using Equation (6):

Ci;t ¼

Xn

 o  Cpeptide ½x  yt i ¼ x  1∪i ¼ y

(6)

In other words, the concentration of cleavage products formed at each time point t, after hydrolysing peptide bond no. i equals the sum of all peptides of sequence [xy], for which i ¼ (x1) (i.e., starts on position P10 ) or i ¼ y (i.e., ends on position P1). This means it is the sum of the peptides that are formed from hydrolysis after the amino acid i or which end with the amino acid i. For each cleavage site, the apparent cleavage rate constant (ki,app in s1) was then calculated by fitting Equation (7) to the experimental data.

  Ci;t ¼ 2 C0  C0 eki;app t

(7)

where Ci,t is the concentration of cleavage products determined for each time point t and C0 is the initial concentration of protein (mM). The rate constant of selective hydrolysis ki (s1 mg1 enzyme) is defined as:

ki ¼

ki;app mE

(8)

where mE is the mass of enzyme added for the hydrolysis (in mg). Subsequently, the selectivityb-Lg (%) is calculated by dividing the rate constant of selective hydrolysis ki of each individual cleavage site by the sum of the rate constant of selective hydrolysis of all cleavage sites at each protein concentration as in Equation (9):

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

k SelectivitybLg ð%Þ ¼ P i  100 kj

(9)

The relative selectivity was calculated using the highest selectivity for each cleavage site as 100%. 2.9. Reproducibility of annotation and quantification To test the reproducibility of the quantification method, one sample (1% WPI, DH 4.5%) was injected 3 times. The concentration determined for individual peptides showed a typical standard error of 6%. In addition, the reproducibility of the complete analysis was tested. For this, two separate hydrolysis experiments were performed at the same protein concentration (0.5%, w/v, WPI). During each hydrolysis, samples were taken at DH 1.5, 3, and 6%. Peptides were annotated and quantified for all hydrolysates. The error in the sequence coverage was calculated for the hydrolysates at each DH. The values of errors were averaged to determine the analysis error. The quality of the annotation as described by the peptide sequence coverage showed a standard error of 2.5%. The molar sequence coverage showed a standard error of 18%. The determined selectivity showed a typical standard error of 15%. 3. Results and discussion 3.1. Protein hydrolysis at different pH values The initial rate of hydrolysis of a 1% WPI with BLP, calculated from the initial part of the DH versus time curves, was 2 times higher at pH 9.0 than at pH 7.0 (Fig. 1A). In addition, by increasing the pH from 7.0 to 9.0, an increase of the DH reached at 10,000 s of hydrolysis was observed, with DH ¼ 4% being reached at pH 7.0 and DH ¼ 8% at pH 9.0 (Fig. 1B). These results agree with reported activity measurements of BLP towards whey protein concentrate (Madsen & Qvist, 1997). However, as mentioned in the introduction, an optimum activity of BLP at pH 8.0 has also been described (Svendsen & Breddam, 1992). In that study, the activity showed a

47

bell-shaped curve, with an optimum at pH 8.0 and lower activities at pH 7.0 and 9.0. Since this was determined from the release of the fluorescent part of a synthetic peptide, and not on WPI, it is concluded that such observed effects depend on the substrate used. In addition to a change in the initial rate and DH reached, a change in activity towards intact protein was observed. The proportion of remaining intact b-lactoglobulin at any DH increased in the order pH 7.0 > pH 8.0 > pH 9.0 (Fig. 2A). For example, at DH ¼ 3% there remained 10% b-lactoglobulin at pH 9.0, 20% at pH 8.0 and 50% at pH 7.0. The affinity of the enzyme for the intact protein (compared with the affinity for derived peptides) can be estimated from the slope of the linear decrease of intact b-lactoglobulin. The affinity of BLP towards b-lactoglobulin was 3 times higher at pH 9.0 than at pH 7.0. A similar effect was observed for the hydrolysis of WPI by trypsin, where the concentration of intact blactoglobulin was decreased faster at pH 8.5 than at pH 7.0 (150 mg min1 at pH 8.5 and 50 mg min1 at pH 7.0; Cheison, Leeb, Toro-Sierra, & Kulozik, 2011). This change in activity towards intact proteins may be related to the stability of the protein. It has been suggested that the rate of hydrolysis increases with the decrease in thermal stability (Huang, Catignani, & Swaisgood, 1994). The stability of b-lactoglobulin, determined by the denaturation temperature, decreases from 76  C at pH 7.0 to 58  C at pH 9.0, as determined by DSC at 100 mM NaCl (Haug, Skar, Vegarud, Langsrud, & Draget, 2009). This is an indication that the structural stability of proteins is closely related to their susceptibility to enzymatic hydrolysis. For a-lactalbumin, in contrast to b-lactoglobulin, the data for all pH values fell on the same curve (Fig. 2B). The proportion of remaining intact a-lactalbumin decreased almost linearly with the DH. This shows a low affinity of BLP for intact a-lactalbumin. The fact that the degradation profiles of the intact a-lactalbumin were similar at all pH values correlates with the fact that the stability of a-lactalbumin does not change as a function of pH as shown by differential scanning calorimetry (Boye, Alli, & Ismail, 1997). b-Lactoglobulin is preferred over a-lactalbumin as expected, based on the higher proportion of b-lactoglobulin than a-lactalbumin in

1 Fig. 1. Initial rate of hydrolysis (mol L1bonds s1 mgenz ) of BLP (A) and hydrolysis curves of 1% WPI hydrolysed by BLP at different pH values and at 40  C (B).

Fig. 2. Proportion of remaining (A) b-lactoglobulin and (B) a-lactalbumin as a function of DH as determined by pH-stat, for the different pH values of hydrolysis: -, pH 7.0; :, pH 8.0; A, pH 9.0.

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

48

Fig. 3. Elution profiles (UV214) from 1% (w/v) WPI hydrolysed at pH 7.0 to DH 3% using (A) intact protein gradient and (B) peptide gradient.

the WPI used. Furthermore, b-lactoglobulin contains relatively more Glu residues than a-lactalbumin. The lower concentration of a-lactalbumin and the lower content of Glu residues result in a lower rate of hydrolysis of this protein compared to b-lactoglobulin.

sample preparation due to lower solubility of certain peptides in the eluent might also result in a decrease of total UV area. The proportion of UV area attributed to b-lactoglobulin (72% on average) was quite similar for all samples, and corresponds to the proportion of b-lactoglobulin in the WPI used (74%). The variation in the concentration of b-lactoglobulin at DH ¼ 0% for the three pH values of hydrolysis is due to daily variations in the sample preparation. Based on the molar concentration of peptide and intact protein concentrations determined during analysis of each hydrolysate, the degree of hydrolysis of b-lactoglobulin was calculated (Equation (5)). It was found that the calculated DH values were slightly underestimated compared with the values obtained with the pH-stat at high DH (Table 1). This can be partly explained by the fact that only peptides from b-lactoglobulin are included in this calculation of the DH, while the value calculated by the pH-stat represents the hydrolysis of all proteins in WPI. The concentration of the 16 aspecific peptides increased with DH and, at certain DH, with the time of incubation needed to reach that DH (which is longer at lower pH) (Table 2). Preliminary experiments on the stability of a synthetic peptide (b-lg[56e62]) at pH 8.0 and 9.0 showed little effect of pH (data not shown). The increase in concentration is thus correlated to increasing time of incubation rather than conditions of hydrolysis. It was assumed that certain peptides formed as a result of the enzyme action are not stable in solution; these peptides auto-hydrolyse to form the a-specific peptides.

3.2. Identification and quantification of the peptides Only peptides resulting from hydrolysis of b-lactoglobulin were analysed in detail to determine the influence of pH on the mechanism of hydrolysis. In total, by combining all hydrolysates, 84 peptides resulting from hydrolysis of b-lactoglobulin were annotated and quantified. Examples of chromatograms obtained with both elution profile are shown in Fig. 3. While most peptides formed agreed with the known specificity of BLP (i.e., for Glu and Asp residues), a total of 16 a-specific peptides were annotated. This means that, on one or both sides of these peptides, cleavages have not occurred after a glutamic or aspartic acid residue. The quality of the quantification, described by the molar sequence coverage, was found to be on average 76 ± 10% over all hydrolysates (Equation (4)). To further describe the quality of the quantification, the weight-based concentration of each peptide was calculated and summed at each DH (Table 1). The total weightbased concentration decreased with increasing DH. This is in part explained by the decrease in total area annotated. Still, a decrease in the total UV214 area is expected at increasing DH, due to a decrease in the number of peptide bonds. In addition, some losses in the

Table 1 Total concentration of the peptides annotated as a function of DH for the three pH of hydrolysis. Parameter

DH (%) 0 pH7

1.5 pH8

pH9

647 556 582 Concentration intact b-Lg (mg L1) Concentration specific peptides (mg L1) Concentration aspecific peptides (mg L1)a Total concentration 647 556 582 (mg L1) 35 30 32 Total concentration (mmol L1) 0 0 0 Calculated DH (b-Lg) (%) Annotated UV area 9.89 8.00 9.38 (AU min  104) 7.82 6.71 7.03 Annotated b-Lg UV area (AU min  104)

3

4.5

6

7

pH7

pH8

pH9

pH7

pH8

pH9

pH7

pH8

pH9

pH7

pH8

pH9

540

313

159

352

66

21

275

0

0

48

0

0

nd

0

0

165

316

311

290

430

441

381

420

520

332

330

486

nd

272

460

0

0

0

705

629

471

645

498

463

678

423

525

479

355

507

nd

324

490

101

156

126

172

230

217

263

290

298

343

303

355

nd

314

391

2.8

1.3

1.0

22.6

3.0

4.9

98.6

25.6

21.8

pH7 pH8

nd

51.7

pH9

29.9

1.2 9.93

2.2 9.37

1.6 7.89

2.4 8.92

3.5 7.63

3.3 7.69

4.1 8.53

4.6 7.34

4.7 7.99

5.5 7.65

4.8 6.60

5.7 nd 7.70 nd

5.0 5.56

6.4 7.82

8.38

7.30

5.47

7.43

5.26

5.21

6.50

5.06

5.77

5.28

4.40

5.58 nd

3.62

5.32

a Aspecific peptides correspond to peptides obtained from cleavages on one or both sides that did not occur after the expected amino acids based on the enzyme specificity (i.e., Glu and Asp residues); nd, not determined.

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

49

Table 2 Concentration of aspecific peptides as a function of DH for the different pH of hydrolysis. Peptide

[56e59] [56e62]a [60e62] [75e80] [75e85]a [75e89]a [81e89] [86e89]a [135e141] [135e145] [135e150] [135e157]a [135e158]a [135e162]a [138e145] [138e157]a [138e158]a [142e157] [142e158] [146e150] [146e157] [146e158] [151e157] [151e158] [151e162] a

DH (%) at pH 7.0

DH (%) at pH 8.0

DH (%) at pH 9.0

1.5

3

4.5

6

1.5

3

4.5

6

7

5.38

11.2

14.3

18.4

15.67

2.22 27.01 0.457

5.02 0.519 20.9 4.34 0.926 0.709 2.52 0.74 4.94 7.80 1.50

30.7

3.22 32.96 1.18 1.35

15.0

30.2

8.46 26.20 7.08 9.24 0.995 23.7 5.34 3.86

14.83 14.97 11.69 21.59 5.82 12.54 3.82 6.73 1.65

4.77 7.07 0.702

2.49 1.88

0.451 1.23 1.29

0.720 0.655 1.04 2.09

2.53

3.82 4.96

3.71

11.6 2.65

1.57 3.93 6.94

0.671 0.297 3.05 9.14 4.66

0.350 0.540

0.269 0.220 0.803 0.757 0.193 1.01 1.51 0.743 1.29

6.59 23.14 2.42 5.38 8.94 1.24 3.37 1.77 0.702 2.02 0.263 0.292 3.74 1.55 4.17 4.21 3.17 5.41 3.93 0.294

5.86

0.792 6.99 13.9

0.565 14.1 8.51

0.783 21.2 1.29

1.5 9.71

4.76

1.39 4.84 17.2

3

4.5

6

7

2.15 17.0

6.50 14.8 3.82

20.9 4.46 14.3

29.4

32.8

0.515 36.5

1.42 34.3

2.75

2.80

4.12 22.5 2.68

5.12 19.4 1.66

19.8

1.78 13.2 12.7

2.68 21.0 5.58

0.380

17.3

0.188 1.71

0.485 3.70

0.233

0.292

1.13

0.244 0.279 2.41

1.07

Assumed parental peptide.

As an indication of changes in the peptide profile due to changes in the pH of hydrolysis, the weight-based concentration of all peptides was summed in each molecular mass class (Fig. 4). At DH ¼ 3%, less intact b-lactoglobulin and a larger amount of intermediate peptides (>3000 Da) was found at pH 8.0 and 9.0 than at pH 7.0 (Fig. 4A). At DH ¼ 6%, only at pH 7 was a small amount of intact b-lactoglobulin still present. This resulted in a lower quantity of peptides in the mass range from 1500 to 3000 Da at pH 7.0 than at pH 8.0 or pH 9.0 (Fig. 4B).

3.3. Influence of the pH on the enzyme selectivity 3.3.1. Kinetics of peptide formation To characterise the influence of the pH of hydrolysis on the mechanism of hydrolysis, the rates of formation and further breakdown of the peptides as a function of the DH for the different pH values of hydrolysis were firstly determined. Four peptides representative of the different types of peptide release kinetics are shown (Fig. 5). As described in a previous study, peptides b-lg[1e45] and b-lg[135e162] are intermediate peptides that can be formed from a single cleavage on the intact protein.

Peptide b-lg(A)[90e127] is an intermediate peptide that can be formed from two cleavages on the intact protein and peptide b et al., 2014a). The conlg(A)[115e127] is a final peptide (Butre centration of the peptides as a function of DH shows comparable trends for each pH, while different maximum concentrations are reached. For instance, the maximum concentration of peptide b-lg [1e45] was 18 mM at pH 9.0, 14 mM at pH 8.0 and 10 mM at pH 7.0, and so the maximum concentration is 50% lower at pH 7.0 than at pH 9.0 (Fig. 5A). Similarly, for peptide b-lg[135e162], a difference of 50% is observed in the maximum concentration of the peptide at pH 7.0 and 9.0 (Fig. 5D). The concentration of peptide b-lg-(A) [115e127], a final peptide, increases with increasing DH in the same way for all pH (Fig. 5C). This shows significant differences in the mechanism of hydrolysis as a function of pH. For some peptides, there is a larger accumulation of intermediate peptides at pH 9.0 than at pH 7.0; for other peptides, the change in pH has no significant effect.

3.3.2. Kinetics of the cleavage site product formation After annotation and quantification of all peptides, the concentration of cleavage products Ci,t was calculated using Equation

Fig. 4. Molecular mass distribution for three different pH of hydrolysis for (A) DH 3% and (B) DH 6% of 1% WPI hydrolysed by BLP at pH 7.0 ( ),pH 8.0 ( )and pH 9.0 ( ).

50

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53





Fig. 5. Peptide concentration as a function of the degree of hydrolysis (DH) for pH 7.0 (◊), pH 8.0 ( ) and pH 9.0 ( ): panel A, b-lg[1e45] LIVTQTMKGLDIQKVAGTWYSLAMAASDISLLDAQSAPLRVYVEE; panel B, b-lg(A)[90e127] (E)-NKVLLVLDTDYKKYLLFCMENSAEPEQSLVCQCLVRTPE; panel C, b-lg(A)[115e127] (E)-QSLVCQCLVRTPE; panel D, b-lg [135e162] (E)-KFDKALKALPMHIRLSFNPTQLEEQCHI.

(6) to determine the apparent cleavage rate constant ki,app (Equation (7)). The concentration of cleavage products Ci,t was compared for the three pH values of hydrolysis for four typical examples (Fig. 6). As for the peptides, there were large differences in the effect of pH on the concentration of cleavage site products for different cleavage sites. The concentration of cleavage products of cleavage site 74 is the same for all pH values of hydrolysis tested as a function of DH (Fig. 6B). The concentration of cleavage products for cleavage site 157 was higher at pH 7.0 than at pH 8.0 or 9.0 for all DH values (Fig. 6C). For cleavage site 158, a lower concentration of the cleavage products was obtained at pH 7.0 than at pH 8.0 or 9.0 (Fig. 6D). The apparent cleavage rate constant, ki,app, and the rate constant of selective hydrolysis, ki, were determined from the concentration

of cleavage products as a function of the time of hydrolysis (Equation (6); Fig. 7). The cleavage sites are divided into five groups based on a high or low rate of selective hydrolysis. The first group has an average rate constant of selective hydrolysis of 1.6  103 mg1 enzyme s1 (Fig. 7A). The second group has an average rate constant of selective hydrolysis of 3.0  104 mg1 enzyme s1 (Fig. 7B). This second group has 2 subclasses, with average rate constants of 6.1  104 mg1 enzyme s1 and 1.1  104 mg1 enzyme s1 (Fig. 7B). The fourth group, with the lowest average rate constant of selective hydrolysis (4.9  105 mg1 enzyme s1), corresponds to aspartic acid residues and one glutamic acid residue (Fig. 7C). Finally, a fifth group contains cleavage site Glu112 and four cleavage sites next to aspartic acid residues, 53, b-lg(A)64, 98 and 130, for





Fig. 6. Concentration of cleavage products as a function of DH for b-lactoglobulin cleavage sites (A) 45; (B) 74; (C) 157 and (D) 158, at pH 7.0 (◊), pH 8.0 ( ) and pH 9.0 ( ).

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

51

Table 3 Effect of pH on the selectivity (%) of the cleavage sites.a Effect

Increase Up and stable Up and down

Down and stable

Down and up (Glu)

Down and up (Asp)

Cleavage sites

45 134 158 55 62 96 74 127 131 157 129 44 89 108 51 65 114 11 28 33 85 137

pH 7.0

8.0

9.0

11.53 3.68 2.67 12.56 7.71 n.d. 21.97 11.02 0.783 1.98 0.474 0.055 14.78 5.62 2.25 1.40 0.783 0.011 0.064 0.227 0.210 0.241

13.14 10.81 6.52 27.00 13.51 0.022 11.99 4.90 0.269 0.050 n.d. n.d. 7.69 3.23 0.418 0.161 0.125 0.007 0.010 0.061 0.047 0.034

15.25 12.83 5.96 14.02 7.80 0.004 14.13 3.49 0.490 0.281 n.d. n.d. 13.41 5.49 4.24 1.03 0.972 0.042 0.061 0.159 0.164 0.165

a Peptides were analysed until DH of 6% for pH 7.0 and DH of 7% for pH 8.0 and 9.0; for certain cleavage sites no corresponding peptides were found (n.d.), indicating that the selectivity is <0.003%.

Fig. 7. Rate constant of selective hydrolysis of the different cleavage sites for the different pH of hydrolysis of 1% WPI by BLP divided into four groups with ( ) pH 7.0, ( ) pH 8.0 and ( ) pH 9.0. The fifth group is not illustrated as the cleavage sites belonging to this group are those for which no cleavages were annotated.

which no cleavages were found. Based on the enzyme specificity, cleavages are expected on these cleavages sites. For the first four groups, an increased rate constant of selective hydrolysis was found with increasing pH (Fig. 7). This is expected based on the overall rate of hydrolysis determined based on the DH curves for the different pH of hydrolysis (Fig. 1). For the cleavage sites after glutamic acid residues, the average rate is 5.9  105 mg1 enzyme s1 at pH 7.0, 1.2  103 mg1 enzyme s1 at pH 8.0, and 1.4  103 mg1 enzyme s1 at pH 9.0. 3.3.3. Determination of the selectivity Using the rate constant of selective hydrolysis, the enzyme selectivity towards the cleavage sites was calculated (Equation (9)). For all hydrolysis condition, it is clear that there are large differences in the selectivity towards the different cleavage sites. The enzyme selectivity varied from 27% to 0.003% and to 0 for some cleavage sites (Table 3). The cleavage sites are divided into 5 groups based on the behaviour of the enzyme selectivity as a function of

the pH (Table 3). The selectivity of the enzyme towards cleavage site 45 is increasing with increasing pH (Fig. 8A; Table 3). Furthermore, 10 out of 22 cleavage sites show a clear minimum with low selectivity at pH 8.0 (Fig. 8E and F) and high selectivity at pH 7.0 and 9.0. For 3 cleavage sites, a clear optimum was observed at pH 8, and low selectivity at pH 7.0 and 9.0 (e.g., cleavage site 55; Fig. 8C). To try to understand the different types of behaviour and the changes in selectivity, the primary structure of the substrate was considered. On three parts of the b-lactoglobulin sequence, a combination of two sequential cleavage sites is found (44Ee45E; 129De130D and 157Ee158E). At pH 8.0 and 9.0, the first cleavage site of each sequence (i.e., 44E, 129D and 157E) is not cleaved or cleaved with a lower selectivity than at pH 7.0 (Fig. 8D; Table 3). This seems to be due to the fact that the BLP is hindered by the presence of negatively charged amino acid residues at the P10 position (Fig. 9). This has been previously suggested based on the rate of hydrolysis of synthetic peptides (Breddam & Meldal, 1992). However, in the pH range 7.0e9.0, the charge states of Glu and Asp are not expected to change. The reason for this observation may then be sought in the active site of the enzyme. The active site of BLP contains the three residues Ser, His and Asp (Mil'gotina, Voyushina, & Chestukhina, 2003). The pKa of the side chain of the histidine residue is 6.0. Consequently, there is still a significant proportion (9%) of His with a positive charge at pH 7.0, while there are 10 and 100 times less charged His side chains at pH 8.0 and pH 9.0, respectively. This decreased protonation of His may affect the interaction with the negatively charged substrate amino acids (Glu, Asp). While the changes in the charge state of His in the active site of the enzyme is the same for all peptides, it probably has an influence in the presence of a negative charge on the substrate. In this way, the negative charge on position P10 might be more favoured at pH 7.0 than at pH 8.0 and pH 9.0. Consequently, the cleavage sites (44, 129 and 157) with a negative residue on P10 have a higher selectivity at pH 7.0 than at pH 8.0 or 9.0. For the other cleavage sites, by looking at 3 amino acid residues on both side of the cleavage site, there seems to be no correlation between primary sequence and trend for

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

52

Fig. 8. Relative selectivity as a function of pH for the different cleavage sites in b-lactoglobulin during hydrolysis of 1% WPI by BLP at three pH values: panel A, site 45; panel B, sites 134 (◊) and 158 ( ); panel C, sites 55 (A), 62 (-) and 96 (✕); panel D, sites 44 (B), 74 (◊), 127 ( ), 129 (✶), 131 (:) and 157 (✕); panel E, sites 51 ( ), 65 (✕), 89 (A), 108 ( ) and 114 (✶); panel F, sites 11 (◊), 28 ( ), 33 ( ) 85 (✕) and 137 (þ). The symbols under the panel identifier letters refer to Fig. 9.





selectivity as a function of pH (Fig. 9). Moreover, the diversity of selectivity behaviour suggests that the charge state of the subsite is not the main influence on the selectivity. Some of the changes in selectivity as a function of pH observed in this study can be explained by changes of the charge state in the active site of the enzyme. Another study on the hydrolysis of casein by trypsin also showed effects of the pH. In that case, the pH (7e10) affects the charge state of the amino acids on the substrate (Lys, Arg). This also resulted in significant changes in the hydrolysis patterns (Vorob'ev et al., 2000). Different behaviour in the formation of peptides was also observed with optimum pH for the formation of some of the peptides. Finally, it should also be noted that the mechanism is different from pH 7.0 to 9.0 with respect to the intact protein. At pH 7.0, the intact protein is less favoured and the intermediate peptides are not accumulated. In this type of mechanism, which can be correlated with the one-by-one mechanism, all putative cleavages sites are used at an earlier DH than in the zipper mechanism. This might also partly explain why cleavage sites 44 and 129 are annotated at pH 7.0 and not at pH 8.0 or 9.0.







4. Conclusions During hydrolysis, an enzyme will cleave (in principle) all peptide bonds that fit the enzyme specificity. However, not all cleavage sites are hydrolysed at the same rate. This selectivity for different cleavage sites determines the final composition of the hydrolysate. By changing the pH of the hydrolysis of WPI by BLP, not only the rate of hydrolysis, but also the mechanism of hydrolysis was changed. The rate of hydrolysis of b-lactoglobulin as a function of DH decreased at pH values closer to the iso-electric point, where the protein stability is highest. The selectivity towards the different cleavage sites varied from 27% to 0.004% or 0 for some cleavages sites. With increasing pH, the selectivity towards certain cleavage sites increased, while for others it decreased. The changes in selectivity as a function of pH were significant (up to 80% increase or decrease). These changes are difficult to predict, but are important since they determine the composition of the final hydrolysate. Although the current work does not provide direct tools to direct the selectivity, it does show how selectivity is determined and how the pH changes the selectivity towards certain cleavage sites.

Acknowledgements

Fig. 9. Amino acid sequence of b-lactoglobulin; the symbols refer to the different type of behaviour of the selectivity as a function of the pH of hydrolysis in Fig. 8.

The research leading to these results was conducted within the EU-ITN LEANGREENFOOD network, funded by the European Community's Seventh Framework Program [FP7/2007-2013] under grant agreement n⁰ 238084.

C.I. Butre et al. / International Dairy Journal 44 (2015) 44e53

References Adler-Nissen, J. (1986). Enzymic hydrolysis of food proteins. London, UK: Elsevier Applied Science Publishers. Akpinar, O., & Penner, M. H. (2001). Peptidase activity assays using protein substrates. In R. E. Wrolstad, et al. (Eds.), Current protocols in food analytical chemistry. Corvallis, OR, USA: John Wiley & Sons. Boye, J. I., Alli, I., & Ismail, A. A. (1997). Use of differential scanning calorimetry and infrared spectroscopy in the study of thermal and structural stability of alactalbumin. Journal of Agricultural and Food Chemistry, 45, 1116e1125. Breddam, K., & Meldal, M. (1992). Substrate preferences of glutamic-acid-specific endopeptidases assessed by synthetic peptide substrates based on intramolecular fluorescence quenching. European Journal of Biochemistry, 206, 103e107. , C. I., Sforza, S., Gruppen, H., & Wierenga, P. A. (2014a). Introducing enzyme Butre selectivity: a quantitative parameter to describe enzymatic protein hydrolysis. Analytical and Bioanalytical Chemistry, 406, 5827e5841. , C. I., Sforza, S., Gruppen, H., & Wierenga, P. A. (2014b). Determination of the Butre influence of substrate concentration on the enzyme selectivity using whey protein isolate and Bacillus Licheniformis protease. Journal of Agricultural and Food Chemistry, 62, 10230e10239. , C. I., Wierenga, P. A., & Gruppen, H. (2012). Effects of ionic strength on the Butre enzymatic hydrolysis of diluted and concentrated whey protein isolate. Journal of Agricultural and Food Chemistry, 60, 5644e5651. lez-Tello, P., & Guadix, E. M. (1998). Influence of enzymes, pH Camacho, F., Gonza and temperature on the kinetics of whey protein hydrolysis. Food Science and Technology International, 4, 79e84. Cheison, S. C., Leeb, E., Toro-Sierra, J., & Kulozik, U. (2011). Influence of hydrolysis temperature and pH on the selective hydrolysis of whey proteins by trypsin and potential recovery of native alpha-lactalbumin. International Dairy Journal, 21, 166e171.

53

Dubois, V., Nedjar-Arroume, N., & Guillochon, D. (2005). Influence of pH on the appearance of active peptides in the course of peptic hydrolysis of bovine haemoglobin. Preparative Biochemistry and Biotechnology, 35, 85e102. Haug, I. J., Skar, H. M., Vegarud, G. E., Langsrud, T., & Draget, K. I. (2009). Electrostatic effects on b-lactoglobulin transitions during heat denaturation as studied by differential scanning calorimetry. Food Hydrocolloids, 23, 2287e2293. Huang, X. L., Catignani, G. L., & Swaisgood, H. E. (1994). Relative structural stabilities of b-lactoglobulins A and B as determined by proteolytic susceptibility and differential scanning calorimetry. Journal of Agricultural and Food Chemistry, 42, 1276e1280. Kosters, H. A., Wierenga, P. A., De Vries, R., & Gruppen, H. (2011). Characteristics and effects of specific peptides on heat-induced aggregation of b-lactoglobulin. Biomacromolecules, 12, 2159e2170. Kuipers, B. J. H., & Gruppen, H. (2007). Prediction of molar extinction coefficients of proteins and peptides using UV absorption of the constituent amino acids at 214 nm to enable quantitative reverse phase high-performance liquid chromatographyemass spectrometry analysis. Journal of Agricultural and Food Chemistry, 55, 5445e5451. Madsen, J. S., & Qvist, K. B. (1997). Hydrolysis of milk protein by a Bacillus licheniformis protease specific for acidic amino acid residues. Journal of Food Science, 62, 579e582. Mil'gotina, E. I., Voyushina, T. L., & Chestukhina, G. G. (2003). Glutamyl endopeptidases: structure, function, and practical application. Russian Journal of Bioorganic Chemistry, 29, 511e522. Svendsen, I., & Breddam, K. (1992). Isolation and amino acid sequence of a glutamic acid specific endopeptidase from Bacillus licheniformis. European Journal of Biochemistry, 204, 165e171. , T. (2000). Kinetics of bVorob'ev, M. M., Dalgalarrondo, M., Chobert, J. M., & Haertle casein hydrolysis by wild-type and engineered trypsin. Biopolymers, 54, 355e364.