2.1. Pepsin Activity and Solubility
The activity of porcine pepsin was characterized in the pH range of 1.1 to 5.0 in both 100% H2O and 100% D2O, using bovine blood hemoglobin as substrate and the INRAE-STLO protocol as described in Salelles, et al. 40. As the porcine pepsin enzyme is known to exhibit maximum activity at pH 2, all measured activities in this study were normalized to the measured activity of pepsin in H2O at pH 2 and are provided as a percentage of this value in Table 1. The characteristic bell shape curve for enzyme activity at different pH values was noted for pepsin in both solvents, with maximum activity observed at pH 2 and a reduction in activity to almost zero at pH 5 (Fig. 1a). We note that the activity of pepsin in H2O was significantly higher than in D2O. For example, at pH 2, the pepsin activity in D2O was ~ 58% compared to that in H2O (Table 1). However, when comparing the relative activity in H2O and D2O at different pH, the shape of the graph (and associated relative change in activity) with increased pH is very similar for both solvents (Table 1), indicating that the reduction in enzyme efficacy in D2O was consistent across the pH range examined.
Table 1
Relative pepsin activity (%) in H2O and D2O at pH 2, 2.6, and 3. All values were obtained by averaging determinations done in triplicate. Different superscript letters indicate statistically significant differences across the first two columns together considering pH 2 in H2O as the 100% activity (1) and the last column alone (2) (p < 0.01). * Relative % of activity in D2O at pH2 was considered as a maximum activity, to allow direct comparison of the rate of change for activity with pH to that of H2O.
| Relative activity (%) |
pH | H2O | D2O | D2O (relative to D2O *) |
2 | 100a (1) | 58 ± 3 c (1) | 100 a (2) |
2.6 | 91 ± 4a (1) | 52 ± 3 c (1) | 90 ± 5 a (2) |
3 | 69 ± 5 b (1) | 38 ± 3 d (1) | 66 ± 5 b (2) |
The decreased activity of pepsin in D2O is not unexpected due to the decrease in the stability of proteins in D2O 38 and the increased strength of hydrophobic interactions 32,41. D2O creates stronger deuterium bonds compared to hydrogen bonds and therefore replacement of H bonds by D bonds in the hydration shell of the enzyme results in stronger, more stable hydrogen bonds both within the enzyme and between the enzyme and solvent. In addition, the strength of hydrophobic interactions in D2O is increased 29,32,37,41,42. The increased strength of the deuterium bonds in D2O reduces the motion of the protein which lowers the activity 42. We note that the decrease in enzyme activity in D2O (compared to H2O) is in agreement with the majority of previous studies 32,41−43, although reaction acceleration was observed in a few cases 44. While the effect of solvent on enzyme activity can be temperature dependent (de Gómez-Puyou, et al. 32), in this study the pepsin activity was measured only at 37°C, of relevance to digestion.
Although the characteristic bell-shaped curve remained the same for casein substrate 40 as other substrates, we note that the activity change with pH observed in this study on hemoglobin was slightly less than that observed in previous studies for different such as chicken or ostrich substrates; for example, at pH 3 the activity of pepsin on hemoglobin was observed to drop by only 30% in our study (compared to a drop of 60–70% observed in previous studies with other substrates) 45–47. However, we note that these previous studies evaluating pepsin activity as a function of pH employed slightly different methodologies and used different substrates 40 and sources of pepsin such as chicken 45 or ostrich 46. Since then, the pepsin enzyme activity assay protocol to assess proteolytic activity has been optimized and standardized by using the INFOGEST protocol commonly in both industrial and academic research. Hemoglobin is the common substrate used in the standardized protocol, due to its stability, rapid digestion, and limited variation between preparations that causes the digestion rate to vary 48. In this study, the standardized substrate hemoglobin was used to determine the action of pepsin. We note, however, that the pH activity profile of pepsin can vary from substrate to substrate as the optimal pH of proteolysis shifts 49,50. Taking into account the wide variety of edible proteins, Salelles, et al. 40 showed that pH activity profiles have differences depending on the substrates. Other substrates, such as wheat gluten or egg white protein, did not produce the bell-shaped curve with maximum activity at pH 2 seen with hemoglobin as a substrate 40, 45–47,51.
On the other hand, the conformation of the substrate could also be affected by the solvent and thus modify the activity of the enzyme. Protein-solvent interactions are one of the most crucial elements stabilising the folded conformation of globular proteins in solution and hence the globular protein hemoglobin may have been significantly impacted by a solvent disruption. The surface of hemoglobin is covered with polar groups and a hydrophobic core within the center. It has an R "relaxed" conformation in the presence of oxygen and a T "tense" shape in the absence of oxygen, with the R conformation exposing a greater hydrophobic surface to the solvent. Water molecules, and hence solvent modification, contribute to the T->R transition of hemoglobin. For instance, alcohol-perturbed solvent stabilizes the R conformation of hemoglobin and exposes a greater hydrophobic surface to solvent. In our scenario, the substitution of H2O with the more hydrophobic D2O may have hindered the T-to-R transition, decreased the exposure of a greater hydrophobic surface and consequently altered the pepsin-hemoglobin ligand interaction.
The solubility of pepsin was measured in D2O and H2O as a function of pH (via determination of protein concentration in the supernatant after centrifugation) to determine if the observed differences in pepsin activity reflect fundamental changes in solubility. The solubility of all pepsin solutions showed a U-shaped curve with the lowest point at pHreading 3 for H2O and D2O respectively (Fig. 1b). We note that for pH values ≥ 5, the protein solubility was approximately 100 percent for both solvents compared to the pepsin solubility in H2O at pH 2. The shape of the solubility curve with a minimum of pH 3 was consistent for both solvents. This is consistent with the relatively low isoelectric point (pI) of porcine pepsin generally between 2.76 and 2.90 52,53.
Protein solubility in solution is correlated with the hydration of charged groups, where over 95% of the charged residues are distributed on the surface of the pepsin (Fig. 1c). We found that the solubility of pepsin in D2O was generally lower than that in H2O, particularly at lower pH values. This is consistent with the more hydrophobic nature of pepsin, which implies that solubility in H2O should be greater than that in D2O (Fig. 1c). In addition, the higher dielectric constant of H2O (81.5 in H2O vs 80.7 in D2O at 273 K) which governs electrostatic interactions, could make H2O a better solvent for pepsin than D2O 28. The reduced solubility of pepsin in D2O compared to H2O may also reflect stronger protein-protein interactions in D2O (as a result of stronger deuterium bonding compared to hydrogen bonding) 38. Proteins in D2O have a stronger tendency to associate and aggregate which decreases solubility. Data reported herein solubility of pepsin are consistent with previous studies which also observed lower solubility in D2O 36,54−56. Additionally, the physical constraints imposed by the viscosity difference impact solubility. D2O’s density (1.1 g/cm3) 57 and viscosity (1.2467 mPa at 20°C) 58 are approximately 10% and 25% greater, respectively than those of H2O. The higher solvent viscosity of D2O indicates a greater stiffness of the protein structure due to the nature and strength of the interaction between the solute and solvent.
2.2. Pepsin Secondary Structure
The far-UV CD spectra of porcine pepsin were measured at pHreading values between 2–7 and in a mixed solvent environment of five different H2O: D2O ratios (100:0, 75:25, 50:50, 25:50, 0:100). Representative spectra are provided in Fig. 2 as molar residue ellipticity (MRE). The negative band observed at 216 nm is characteristic of the presence of β-sheets in the secondary structure of pepsin as expected 59. Dichroweb analysis confirmed that the CD spectra were dominated by β-sheets (37%) similar to previous information on pepsin (44% β-sheets). No significant differences were found in the secondary structure of pepsin over the pH range investigated in H2O (Supporting Information, Table S1). Moreover, changing the solvent from H2O to D2O had no influence on the secondary structure of pepsin at different pH values. This is consistent with previous studies suggesting that the secondary structures of proteins are either unchanged 60 or changed only by a modest degree 61 following the replacement of H by D. In fact, even with far more disruptive solvents such as ethanol and acetonitrile, no difference in secondary structure was observed with 30–60% ethanol or 60% acetonitrile in previous research 22,62. Simon, et al. 22 have reported that porcine pepsin only alters its tertiary structure and results in a functional loss when high concentrations (70%) of organic solvents are present.
Minor changes in secondary structure were observed at a pHreading of 2 with a slight shift in the negative band at 216 nm to lower wavenumbers observed. This suggests some loss of beta-sheet structure, although we note that this was not reflected in a major change to the proportion of β-sheets as calculated by Dichroweb analysis (Table S1). This is consistent with a study by Perlmann and Kerwar 63 who observed a conformational transition of the polypeptide backbone below pHreading 2. We note that the effect is more prominent in H2O compared to D2O (Fig. 2), potentially reflecting the pH difference between D2O and H2O. On the other hand, the effect may be diminished since proteins with predominantly β-sheet type structures are preferentially stabilized in D2O due to the modulation of hydrophobic interactions and stabilization of deuterium bonds 29,32,33,41,42. This is similar to the findings of the study of Dong, et al. 61, where the negative band around 216nm of the CD spectra of thermally denatured pepsin in H2O is more intense and broadened compared to that in D2O. Moreover, deuteration increases the thermal stability of proteins and therefore a smaller change between solvents at 37°C, compared to 25°C, might have been observed due to stronger hydrophobic bonds in D2O 34,35.
2.3. Pepsin Structure with SAXS
SAXS was used to evaluate the size, shape and aggregation of pepsin in phosphate buffer at pH values of 6, 7 and 8. The pH values were selected as pepsin is known to change to a denatured state between pH 6 and 7 13,14,16 and irreversibly denatures at pH 8 10,19. The phosphate buffer has a pKa of 7.20 which allowed SAXS experiments to be performed using only one type of buffer to adjust the pH while reducing any buffer-derived changes to the pepsin structure. Figure 3a shows the SAXS patterns of pepsin D2O and H2O buffer solvents at the three pH values. We note that calculations of the radius of gyration (Rg) and Guinier approximation from SAXS data are typically only valid for monodisperse, and non-aggregated systems, whereas the SAXS patterns presented here how upturns towards low q consistent with the presence of aggregates. We have therefore determined pseudo-Rg values from forced linear regression on the Guinier region (arrowed region, Fig. 3a). These pseudo-Rg values can be related to the average diameter of pepsin molecules, and depict the size, compactness and aggregation of proteins 19,20 64. The pseudo-Rg value would be smaller when the protein is more compact compared to a partially or fully unfolded protein, and larger if there is more aggregation present. The pair distance distribution function, p(r), has been calculated to represent a real space representation of the distances present in the sample, where a gaussian curve is characteristic of spheres, such as globular proteins (Fig. 3b). The maximum dimension of the molecule (Dmax) was additionally estimated from the point where the p(r) function reaches zero (Table S2). Finally, the structural envelope of pepsin was represented by the reconstruction of ab initio structural models using the GASBOR program based on the p(r) data (insert, Fig. 3b). We observed differences in the structural models, and pseudo-Rg and Dmax values, as a function of pH and solvent.
In H2O, calculated values of pseudo-Rg and Dmax are 22.1 Å and 71.0 Å at a pHreading of 6 and 19.9 Å and 63.0 Å at a pHreading of 7, respectively. The p(r) curves present symmetrical shapes with a Dmax of 60–80 Å in both solvents at pH 6 and 7, suggesting monodispersed species present in the systems consistent with a previous study 20. Both porcine pepsin structure studies of Jin, et al. 19 and Konno, et al. 17 showed similar Rg values: ~24 Å and ~ 22 Å for pH 5.6 and 7 respectively. We also found that pseudo-Rg results calculated for pepsin in H2O at pH 7, 19.9 Å for synchrotron were consistent with those reported by Jin, et al. 19 (21.7 Å). Based on the reconstructed ab initio structural models using the GASBOR program, the pepsin structure was more compact at pHreading 7 compared to 6 (insert, Fig. 3b).
While pepsin adopts a more compact form at a pHreading of 7, a more extended form appears at a pHreading of 8, in agreement with a previous study 19. At pH 8, the pseudo-Rg value significantly increased to 25.6 Å with a Dmax of 164 (Table S2). At pH 8, the pepsin structure adopted an elongated shape (Fig. 3d), while the p(r) function became asymmetric with a peak shift indicating loss of its spherical form by either significant unfolding and/or aggregation (Fig. 3b). Due to the asymmetrical p(r) plots at pH 8 suggesting potential aggregation, size-exclusion chromatography (SEC)-SAXS was carried out to investigate the populations of aggregated and denatured forms of pepsin at pH 8 in H2O. Figure 3c shows the average intensity of the SAXS patterns over time of the sample passing the SEC column for pepsin at pH 8. The overlaid red values provide the Rg value for each SAXS pattern, and the values indicate that protein aggregates and monomers were present in the original sample. The different species with different Rg clearly demonstrate the presence of pepsin aggregates at pH 8 in H2O, which possibly also occurred in D2O. Subsequently, SAXS patterns corresponding to two pepsin species with Rg of ~ 40 Å and 17 Å were averaged for further analysis. Figure 3d shows the resulting averaged SAXS patterns corresponding to Rg 40 Å and Rg 17 Å, while the insert shows the p(r) plots and corresponding GASBOR shapes. The Rg 40 Å species of pepsin had an elongated shape with a Dmax over 150 Å, which is also observed in denatured pepsin with highly concentrated urea present20. However, a compacted form was observed for Rg 17 Å species of pepsin, which is attributed to a monomeric form.
Herein, the structure of pepsin in D2O was investigated for the first time using synchrotron SAXS. The pseudo-Rg and Dmax values were found to be smaller than the ones in H2O (Table S2). Based on the pseudo-Rg, Dmax and GASBOR model, the pepsin structure was found to be more compact in D2O compared to H2O (Fig. 3b & Table S2). The more compact structure of pepsin at any pH in D2O aligns with previous observations of proteins in solvents of different hydrophobicity. On the other hand, we suggest that the structure may reflect differences between the actual pH value in D2O (pD) and the pHreading value (pD = pHreading + 0.4). The lower values of pseudo-Rg and Dmax in D2O show less prominent aggregation where the pD value of 7 corresponds to a pHreading of 6.6.
Figure S1 shows the Kratky plots of the SAXS data for pH 6, 7 and 8 in both H2O and D2O. Kratky plots of q2I(q) vs q, are used to distinguish globular and unfolded conformations of proteins. For globular conformations a bell-shaped curve is present but on denaturation, the peak becomes weak or disappears with a plateau forming at larger q values. Here, a globular structure was identified by a pronounced peak pattern in the q region below 0.18 Å−1. Above 0.18 Å−1, the high q Kratky region exhibits linearity with different slopes for each environment, correlated with chain flexibility. The plots with elevated scattering at higher angles suggest that part of the protein is flexible. Pepsin at pHreading 6 showed that pepsin has a globular structure and high chain rigidity for both H2O and D2O. At pHreading 7, the overall size of pepsin in D2O was unchanged but had a weak denaturation process and reduced rigidity and greater chain flexibility (as indicated by its asymmetrical peak) were observed suggesting a weak denaturation process 20. This effect has not been observed in D2O at pD 7 but it was expected that the pHreading would remain at 6.6 due to the 0.4 constant effect 65. A steeper slope with a disrupted peak was observed for pH 8 in H2O and D2O, reflecting the greater chain flexibility of pepsin.