Original papers
Selective Elimination of Bitter Peptides by Adsorption to Heat-treated Porous Silica Gel
2019 Volume 25 Issue 2 Pages 179-186
Details
2019 Volume 25 Issue 2 Pages 179-186
Protein hydrolysates are used for various purposes such as hypoallergenic food, clinical applications, and functional food. However, hydrolysed peptides usually present intense bitter tastes. We previously created a new type of silica gel without chemical modification and found that this material highly adsorbed hydrophobic or positively charged peptides under neutral conditions. Bitter peptides possess bulky basic groups or hydrophobic groups, and we examined the applicability of these as a new bitterness removal carrier. As a result, we confirmed that bitter peptides were more selectively adsorbed compared to other commercial materials and the bitter taste was significantly reduced. These results indicate the utility of our material as a pretreatment carrier. Moreover, silica gel is approved as a food additive and we created this material by subjecting to heat treatment only; thus, it has high food safety and is applicable as a bitterness-masking material.
Protein hydrolysates have a number of interesting features from an industrial point of view (André et al., 2017). Several studies have demonstrated their solubility (Chiang et al., 1999; Lamsal et al., 2007; Sarita et al., 2011), emulsifying capability (Zhang et al., 2015) and oil retention capacity (Santos et al., 2011), which are important in preventing phase separation in certain types of food. Because of such characteristics, protein hydrolysates are used in a number of applications such as salad dressings, spreads, ice cream, coffee whitener, and emulsified meat products like sausages or luncheon meats (Saha and Hayashi, 2001).
Protein hydrolysates are also used in hypoallergenic food, clinical applications, and functional food. The most common means to reduce allergenicity is enzymatic hydrolysis of principal proteins, thereby eliminating antigenic epitopes (Ena et al., 1995). Therefore, the enzymatic degradation of whey protein is widely employed in the food industry and various proteinases have been tested (Kim et al., 2007; Schmidt et al., 1995). Protein absorption can occur as peptides as well as amino acids. Indeed, absorption as short-chain peptides (mainly, di- and tripeptides) is considered to be a more efficient method of amino acid absorption compared to an equivalent amount of free amino acid (Ito et al., 2013; Siemensma et al., 1993). Therefore, protein hydrolysates are also used for dietary applications such as energy drinks, geriatric products, sports nutrition, and weight-control diets (Clemente, 2000). In some studies, many types of peptides showing physiologically active effects have been identified from protein hydrolysates and are known as bioactive peptides. Bioactive peptides are released during enzymatic proteolysis (gastrointestinal digestion, in vitro hydrolysis using proteolytic enzymes) of proteins and also during food processing (cooking, fermentation, ripening) (Daliri et al., 2017). Currently, functional food ingredients based on bioactive peptides derived from whey proteins are sold commercially (Dullius et al., 2018).
As mentioned above, protein hydrolysates are used for various purposes. However, despite their advantages, hydrolysed proteins usually present intense bitter tastes. To date, there have been many reports on the identification of bitter taste peptides from several kinds of food protein hydrolysis (Lemieux and Simard, 1992; Liu et al., 2016). A summary of the characteristics of bitter peptides indicated that peptides containing hydrophobic or basic amino acid residues exhibit bitter tastes. Detection of bitter tastes is understood as follows: bitter peptides possess two determinant sites, the binding unit (BU) and the stimulating unit (SU). The bulky hydrophobic group functions as the BU, which is the primary site that determines the bitterness of the peptide. The bulky basic group including an α-amino group or hydrophobic group, functions as the SU, a secondary site that also contributes to a bitter taste (Maehashi and Huang, 2009). Therefore, the reason why hydrolysed peptides present intense bitter tastes is attributable to the exposure of apolar groups initially within the protein tridimensional structure, leading to sensation of a bitter taste (André et al., 2017). This is the most serious problem in the practical use of food protein hydrolysates.
Many methods with the aim of debittering protein hydrolysates have been studied (Saha and Hayashi, 2001). Selective separation is when bitter components are removed by adsorbing on materials (Murray and Baker, 1952) or by extracting with azeotropic secondary butyl alcohol, aqueous ethanol, or aqueous isopropyl alcohol (Lalasidis and Sjoberg, 1978; Lalasidis, 1978). Another method is the masking method, which employs monosodium glutamate and several glutamyl oligopeptides, acidic phospholipids, lysophospholipids and cyclodextrins. Cyclodextrins may reduce unpleasant tastes in two ways: by forming chemical complexes with disagreeable taste compounds and by preventing their interaction with taste bud carrier proteins (Szejtli and Szente, 2005). Degradation methods including further hydrolysis of bitter peptides with enzymes have also been reported. For example, the bitterness was removed by using aminopeptidases such as Aeromonas caviae T-64 and Aspergillus acid carboxypeptidase (Arai et al., 1970; Izawa et al., 1997).
However, these debittering methods can have unintended consequences. Activated carbon can adsorb excess amounts of peptides non-selectively. Therefore, although debittering was achieved, activated carbon treatment involved a loss of protein nitrogen. Meanwhile, masking method, for example cyclodextrin, was necessary to use a large excess to obtain effects. Also, degradation methods gave rise to significant amounts of primarily hydrophobic amino acids, which may inevitably affect the taste qualities and have a high osmolality (Pedersen, 1994).
In our previous study, we created a new type of porous silica gel without chemical modification, and investigated its adsorption properties using a tri-, penta- and hepta-peptide library. We found that the adsorption properties were summarized by pI and the hydrophobicity of peptides, and under neutral conditions, the remaining silanol groups are deprotonated and negatively charged. Thus, hydrophobic or positively charged peptides were highly adsorbed due to hydrophobic and electrical interaction (Imai et al., 2018).
Most bitter peptides possess hydrophobic or basic properties; thus, we hypothesized that our silica gel could be employed as a new material to selectively remove bitter peptides. Next, we plotted tri-, penta- and hepta-taste peptides (in this study, umami and bitter peptides) that were reported previously in color maps, and only bitter peptides were plotted in the high adsorption region (Fig. 1). The objective of the present study was to investigate whether our previously developed silica gel could be used as a new bitterness removal carrier to solve the problems described above, that is, as a material with high bitter selectivity and high removal efficiency. Notably, the silica gel created by us showed higher removal selectivity and efficacy of bitterness for umami than that of commercially available materials. Silica gel is approved as a food additive and our material was only subjected to heat treatment; thus, it shows high food safety and is applicable for use as a bitterness-masking material.
Color maps of peptide adsorption ratios to heat treated silica gel under pH 7.4 condition (3-mer, 5-mer and 7-mer peptides). The vertical axis indicates hydrophobicity and the horizontal axis indicates pI of peptides. ● represents the point of each bitter peptide and ○ represents that of umami peptide. The higher score indicates the better adsorbing peptide. The detailed method of creating color map is referred by our previous report (Imai et al., 2018).
Materials Porous silica gel, SMB-100-5, was supplied by Fuji Silysia Chemical LTD., Aichi, Japan. SMB-100-5 was granulated after the addition of NaOH to adjust the isoelectric point to pH 9. Heat-treated silica gel was created by calcination at 600 °C for 2 h. Activated carbon was supplied by Futamura Chemical CO., LTD., Aichi, Japan. Polystyrene beads were supplied by Mitsubishi Chemical Corporation, Tokyo, Japan (SEPABEADS™ SP850, SP825L) and Organo Corporation, Tokyo, Japan (Amberlite™ XAD2000) (Table 1). Casein hydrolysate was donated by a manufacturer as a powder. The degree of hydrolysis was 25%, which was estimated by formol titration. The average molecular weight (MW) was 355 Da. Molecular weight distribution was as follows: < 200 Da (10%), 200–500 Da (53%), 500–1,000 Da (24%), 1,000–2,000 Da (12%), 2,000 Da < (1%). Peptides (Sequence; IPAVF, SPE) were purchased from GL Biochem Ltd., Shanghai, China.
| Adsorbent | Silica gel | Polystyrene beads SP850 | Polystyrene beads SP825L | Polystyrene beads XAD2000 | Activated carbon | Heat Treated silica gel |
|---|---|---|---|---|---|---|
| Size (µm) | 5 | 250≧ | 250≧ | 320–440 | 75–180 | 5 |
| Surface Area (m2/g) | 299 | 930 | 930 | 550 | – | 255 |
| Pore Volume (mL/g) | 0.78 | 1.1 | 1.4 | 0.6 | – | 0.71 |
Peptide library synthesis Peptide arrays were synthesized using a cellulose membrane and a spot synthesizer (Intervis, ASP222, Cologne, Germany) as previously described (Imai et al., 2018). After punching, the individual resulting peptide-containing disks (peptide spots) were placed in the single wells of a 96-well plate filter (MSRLN0410; Merck Millipore, Burlington, MA, USA) and 180 µL of PBS (phosphate buffered saline) was added to each well. After 1 h incubation at room temperature, the solution containing peptides was released from the disk and filtered into a 96-well plate by vacuum filtration. Each filtrate was used for the peptide adsorption experiments.
Peptide adsorption experiments Adsorbents were suspended in PBS at 100 mg/mL (in the case of activated carbon; 1 mg/mL). A mixture of 50 µL of the suspensions and 150 µL of the peptide solution were vigorously shaken and left to equilibrate for 5 min at room temperature. After centrifugation at 9,300 × g for 1 min, the amount of adsorbed peptide was determined by measuring the amount of peptide remaining in the supernatant using a fluorimetric assay (Puddu and Perry, 2012). In the assay, 10 µL of fluorescamine (5 mg/ mL in acetone) was added to a 150-µL aliquot of the supernatant in a 96-well plate, and the fluorescence intensity was measured at 355 nm excitation and 460 nm emission (Fluoroskan Ascent™ Microplate; Thermo Fisher Scientific, Waltham, MA, USA). All assays were repeated three times, and the data are presented as mean values and standard deviation (SD). We calculated the average adsorption ratios for all umami peptides (white bars in Fig. 2) and bitter peptides (gray bars in Fig. 2). Selectivity was calculated by dividing the average bitter peptides adsorption ratio by that of umami peptides.
Umami and bitter peptide adsorption assay with 50 peptides (21 umami peptides and 29 bitter peptides). White bars are umami peptides and gray bars are bitter peptides. (A) Silica gel, (B) Hydrophobic resin (SP850), (C) Hydrophobic resin (SP825L), (D) Hydrophobic resin (XAD2000), (E) Activated carbon, (F) Heat treated silica gel. Adsorbents were suspended in PBS at 100 mg/mL, except activated carbon. Activated carbon was added at 1 mg/mL.
Selectivity measurements of umami and bitter peptide mixtures To illustrate the selective adsorption properties of the heat-treated silica gel, the adsorption ability for two types of peptides and mixed peptides was tested. Herein, IPAVF and SPE were selected as model peptides of bitter peptides and umami peptides, respectively. A 250-µL aliquot of a solution containing a silica suspension (50 mg/mL) and peptide solutions (0.4 mM) was vigorously shaken. After centrifugation (9,300 × g, 1 min) to separate the supernatant and silica gel, 200 µL of solution was injected into the HPLC to quantify the peptide adsorption. Chromatographic analysis was performed with an HPLC system (JASCO, Tokyo, Japan) equipped with a pump (model PU 2086 Plus), UV detector (model MD 4017), and C18 column (d = 4.6 mm, L = 250 mm; Kanto Chemical Co., Inc., Tokyo, Japan). Solvent A contained 0.1% TFA in Milli Q water, and solvent B contained 0.1% TFA in acetonitrile. Separation of peptides was obtained using a linear gradient from 0% to 100% of solvent B for 30 min. The column was maintained at 30 °C, the flow rate was 1 mL/min, and the eluted peaks were detected by UV absorbance at 220 nm.
Selectivity measurements of casein hydrolysate mixtures The mixture of 150 µL of silica suspension (240 mg/mL), 200 µL of the SPE and IPAVF peptide solution (1 mM) and casein hydrolysates (30 mg/mL) was prepared, shaken vigorously and left to equilibrate for 5 min at room temperature. After centrifugation (9,300 × g, 10 min) to separate the supernatant and silica gel, 200 µL of solution was injected into the HPLC. Separation of peptides was obtained using a linear gradient from 0% to 100% of solvent B for 90 min. The other conditions were the same as described above.
Sensory test A sensory test was carried out in reference to a previous report with some modifications (Nakano et al., 2007). Each sample was prepared as follows: the control sample was prepared by dissolving casein hydrolysate powder in distilled water at 5 mg/mL. The “before treatment” sample was prepared by dissolving 50 mg/mL of casein hydrolysates. The “after treatment” sample was prepared by dissolving 50 mg/mL of casein hydrolysates and adding 90 mg/mL silica beads. Then, the mixture was vigorously shaken and centrifuged (9,300 × g, 10 min) to separate the supernatant and silica gel. The supernatant was used as the sample for the sensory test. Each sample was diluted in distilled water at an ambient temperature. The sensory test was conducted in individual booths with 11 panelists (4 males and 7 females, aged 21–25 years). Panelists were made up of individuals capable of sensing bitterness when control samples were tested.
The sensory test consisted of 2 steps: first, panelists tasted the control sample and rinsed their palates. Next, panelists tasted the before or after treatment samples and recorded the bitter intensity of the sample tasted last compared to the control sample (1 = slightly bitter, 2 = a little bitter, 3 = bitter, 4 = very bitter and 5 = extremely bitter). Taking into account the dulling of sensitivity due to the accumulation of bitterness, the taste order consisted of the control sample first. In one sensory test, only one set of a control sample and a tasting sample was tasted, and another set was performed the other day.
Characterization of silica gels and other materials used as absorbents in this study We employed two types of silica gels: normal silica gel and heat-treated silica gel, in which the normal silica gel was subjected to calcination at 600 °C for 2 h. Table 1 shows the structural properties of the two types of silica gels and supplied materials, which are commercially available.
Adsorption of umami and bitter peptides on various adsorbents We firstly investigated the adsorption ratios of the 50 peptides (21 umami peptides and 29 bitter peptides) listed in Table 2 using normal silica gel, polystyrene beads, and activated carbon (Fig. 2 A–E). The normal silica gel (Fig. 2A) and polystyrene beads (Fig. 2 B–D) adsorbed bitter peptides selectively. In contrast, the activated carbon (Fig. 2E) absorbed the peptides non-selectively, although the maximum amount of peptides adsorbed. The polystyrene beads were mainly adsorbed by hydrophobic interactions and π-π stacking (Qiang et al., 2017).
| Sequence | Reference |
|---|---|
| EQ, EG, VP, GF, AP | (Yamamoto and Hukusaki, 2014) |
| EGS | (Kim et al., 2015) |
| ED, TE, ES, EE, EDE, DES, EEE, EV, ADE, AED, SPE | (Nishimura, 2001) |
| DA, DV, EEN, EPAD | (Maehashi et al., 1999) |
| Sequence | Reference |
|---|---|
| GL, LF, LK, RL, RLL, SKGL, | (Imamura et al., 1978) |
| MI, LPQE, EIVPN, VRGPFP | (Toelstede and Hofmann, 2008) |
| LW, LRF, EVLN, NENLL, VPPFLQ, APFPEVF, KAVPYPQ | (Lemieux and Simard, 1992) |
| LLF, YGLF, IPAVF | (Liu et al., 2014) |
| FPQ, ILQ, LPQ, LPPFS, GYYPT, IPFVHPS, IALRT | (Liu et al., 2016) |
| QLFNPS | (Ito, and Miyake, 2015) |
| RPPPFFF | (Temussi, 2012) |
Silica gels mainly adsorb by electrostatic interactions between negatively charged silanol groups and positively charged adsorbates (Patwardhan et al., 2012; Rimola et al., 2013). Bitter peptides exhibit high hydrophobicity, with some being positively charged. In contrast, umami peptides are negatively charged short peptides that include glutamic acid and aspartic acid (Temussi, 2012). Therefore, the silica gel adsorbed bitter peptides by electrostatic interaction, while umami peptides were not highly adsorbed because of electrostatic repulsion. Activated carbon showed maximum binding capacity when the solution pH was near the protein's isoelectric point, where there is a minimum effective charge on the protein (Stone and Kozlov, 2014). Bitter peptides do not have any particular charge characteristic, although they are highly hydrophobic. Therefore, it is generally thought that the selective adsorption of bitter peptides using activated carbon is very difficult.
Silica gels can undergo a change in surface hydrophobicity by heat treatment (Rimola et al., 2013; Zhuravlev, 2000). During calcination, some of the silanol groups on the silica surface obtain a siloxane structure, resulting in a more hydrophobic surface. Then, we subjected the silica gels to calcination at 600 °C for 2 h and performed the same experiment. As a result, only the adsorption ratios of bitter peptides were drastically improved (Fig. 2F).
We calculated selectivity by dividing the average bitter peptides adsorption ratio by that of umami peptides (Table 3). Compared with the silica gel, selectivity was 2.6 times greater following calcination. Calcination at 600 °C did not completely eliminate all silanol groups; however, about half of the silanol groups displayed an altered bridge structure. Therefore, bitter peptides were adsorbed between the negatively charged residual silanol groups and siloxane bridges by both electrostatic interaction and hydrophobic interaction. Conversely, the negatively charged umami peptides were repulsed by the negatively charged silanol groups due to electrostatic repulsion.
| Adsorbent | Umami | Bitter | Selectivity (Bitter/Umami) |
|---|---|---|---|
| Silica gel | 9.4 | 32.4 | 3.4 |
| Hydrophobic resin (SP850) | 4.2 | 27.5 | 6.4 |
| Hydrophobic resin (SP825L) | 11.3 | 34.7 | 3.1 |
| Hydrophobic resin (XAD2000) | 5.8 | 23.0 | 4.0 |
| Activated carbon | 16.6 | 14.3 | 0.9 |
| Heat treated silica gel | 7.5 | 66.7 | 8.8 |
We expected that all bitter peptides would be highly adsorbed and that all umami peptides would show limited adsorption. However, some peptides did not conform to these expectations. VP, GF and AP were relatively highly adsorbed compared to the other umami peptides. This was because these peptides did not contain negatively charged amino acids, eliminating electrostatic repulsion as a factor. GL, MI, GYYPT, EIVPN, NENLL and EVILN showed relatively low adsorption compared to the other bitter peptides. In general, the longer the peptide, the greater its hydrophobicity. Therefore, although GL and MI consisted of hydrophobic amino acids, they did not exhibit high hydrophobicity. Although GYYPT, EIVPN and NENLL were long peptides, they contained large numbers of hydrophilic amino acids or negatively charged amino acids. The pI and hydrophobicity were as follows: GYYPT (5.85, −1.06), EIVPN (5.38, 0.02), NENLL (5.20, −0.58), respectively, and the average values of the other penta-peptides plotted in Fig. 1 were 6.26 for pI and 1.28 for hydrophobicity. These peptides were plotted low or left compared to the other bitter peptides. This characteristic resulted in weak driving forces. Although there were some exceptions, the heat-treated silica gel adsorbed bitter peptides more selectively.
Selective removal of bitter peptides Bitter peptides are contained in protein hydrolysates, making it necessary to consider peptide-peptide interactions. In this study, we evaluated adsorption properties in the presence of umami peptides. We used SPE and IPAVF as representatives of umami peptides and bitter peptides, respectively, and evaluated the amount of peptide adsorption by HPLC. In the chromatogram, the SPE peak was not decreased after adsorption; in contrast, the IPAVF peak was drastically decreased, indicating that the heat-treated silica gel exhibited selective removal even in mixed solutions of umami peptides (Fig. 3). The amount of adsorbed peptides was calculated using a calibration curve. The adsorbed amount of IPAVF was 7.28 ± 0.09 nmol/mg, which was 3.11 times greater than SPE. This result indicated that even in the presence of umami peptides, the adsorption behavior was the same as when alone. Next, in order to investigate the influence of other factors besides umami peptides, we investigated adsorption properties in the presence of casein hydrolysates (Fig. 4). SPE and IPAVF were added to casein hydrolysates and selectivity was evaluated. We found that even in the presence of hydrolysates, our silica gel could selectively adsorb bitter peptides.
The chromatogram of SPE and IPAVF mixtures. Gray line is before treatment and black line is after treatment. The amount of adsorption peptides were calculated by calibration curve.
The chromatogram of casein hydrolysate mixtures. Gray line is before treatment and black line is after treatment.
Sensory test of casein hydrolysates treated with or without heated silica gel Finally, we examined the ability of the heat-treated silica gel to remove bitter components from casein hydrolysates. The degree of bitterness reduction was evaluated using a sensory test. The bitterness scores of before- and after-treatment were determined. The score of after-treatment (2.3 ± 1.2) was smaller than that of before-treatment (4.4 ± 0.8), as shown in Fig. 5. Eight of 11 panelists scored the after-treatment 2 points lower, two scored 1 point lower and one panelist scored 1 point greater than the before-treatment. Panelists who gave the before-treatment a high score tended to give the after-treatment a low score. Deviation of the score for after-treatment was somewhat greater than that of before-treatment; however, no significant difference was observed. These results indicated that the heat-treated silica gel could be employed to remove the bitter taste from protein hydrolysates.
Bitterness score of before and after treatment. Black line represents the average score. Line width represents the number of panelists. Thick line is 2 panelists and thin line is 1 panelist.
We examined the applicability of heat-treated silica gel for the removal of bitter compounds from protein hydrolysates. Future studies will focus on the optimization of removal conditions for more selective and efficient removal of bitter peptides. In addition, we aim to investigate further applications of this material. Our results showed that peptides with similar properties could be selectively adsorbed on the basis of physicochemical properties. Thus, we would like to investigate the utility of this material for selective condensation (e.g., affinity chromatography) on the basis of physicochemical characteristics.
Acknowledgements This work was partially supported by JSPS KAKENHI (Grant Number: JP16H04575) and Aichi Science and Technology Foundation (17J6401b). We would like to thank Mr. Kamimura (Fuji Silysia Chemical LTD.) for creating the silica gel.
