Fluorogenic Peptide Substrate for Quantification of Bacterial Enzyme Activities

A novel peptide substrate (A G G P L G P P G P G G) was developed for quantifying the activities of bacterial enzymes using a highly sensitive Fluorescence Resonance Energy Transfer (FRET) based assay. The peptide substrate was cleaved by collagenase class I, II, Liberase MTF C/T, collagenase NB1, and thermolysin/neutral protease, which was significantly enhanced in the presence of CaCl2. However, the activities of these enzymes were significantly decreased in the presence of ZnSO4 or ZnCl2. Collagenase I, II, Liberase MTF C/T, thermolysin/neutral protease share similar cleavage sites, L↓G and P↓G. However, collagenase NB1 cleaves the peptide substrate at G↓P and P↓L, in addition to P↓G. The enzyme activity is pH dependent, within a range of 6.8 to 7.5, but was significantly diminished at pH 8.0. Interestingly, the peptide substrate was not cleaved by endogenous pancreatic protease such as trypsin, chymotrypsin, and elastase. In conclusion, the novel peptide substrate is collagenase, thermolysin/neutral protease specific and can be applied to quantify enzyme activities from different microbes. Furthermore, the assay can be used for fine-tuning reaction mixtures of various agents to enhance the overall activity of a cocktail of multiple enzymes and achieve optimal organ/tissue digestion, while protecting the integrity of the target cells.

Scientific RepoRts | 7:44321 | DOI: 10.1038/srep44321 approach is costly; on the other hand, the variability in donor characteristics per se introduces more variables that affect enzyme evaluation. Currently, there are several methods available to measure collagenase activity 9,16 . In particular, the Wünsch assay has been used to measure class II collagenase 16 , but was found to be unsuitable for measuring class I collagenase activity. These measurements are carried out using each manufacture's own protocol, which make it difficult to standardize the assay. Furthermore, the cleavage sites of pancreas ECM by varies enzymes have not been investigated before, since this will help to further refine the characteristics of enzyme(s) used for pancreas digestion. Hence, a more accurate and reliable method is needed to quantify the enzyme activity. In this study, we first synthesized and characterized a novel peptide substrate that can be used to assess the collagenolytic activity of the enzymes including collagenase class I, II and thermolysin/neutral protease using Fluorescence Resonance Energy Transfer (FRET). The method is sensitive to determine the kinetic parameters of various enzymes using micromolar concentrations of substrate [17][18][19][20][21] . The substrate contains 5-CarboxyFluorescein-Aminohexyl Amidite (5-Fam) fluorogenic groups. The principle of the assay is based on the fact the donor fluorophore and the acceptor fluorophore (quencher) are in sufficient proximity in the substrate to allow a resonance energy transfer between the fluorophore and the quencher. The change in resonance energy transfer and increase in fluorescence intensity upon substrate cleavage are associated with higher intensity of enzymatic reaction. Furthermore, we examined how the enzymes cleave the peptide substrate under different conditions that exist in actual process of pancreatic tissue digestion: (1) effect of reaction components on enzyme activity; (2) effect of pH; and (3) effect of endogenous pancreatic proteases on the peptide substrate.

Results
Characterization of synthesized peptides. The pancreas has been used as a surrogate for enzyme evaluation. Although, this has been relatively helpful in understanding pancreas digestion and ultimately freeing islets, the procedure is expensive, lengthy, and challenging. Hence, the need for a simple and effective assay to evaluate and standardize enzyme activity is highly desirable to overcome these problems 11 . The unique characteristic of the enzyme substrate developed is described in Table 1. Our new substrate is 5Fam-AGGPLGPPGPGGK-dabcyl, which has different donor/acceptor fluorophores group localization, which can be cleaved by thermolysin as compared to a previous study 18 . Complete amino-acid sequences of the ECM: collagens (I, II, III, IV, V), laminin, and fibronectin were obtained in FASTA format from UniProtKB 22 . The amino-acid sequence of our newly synthesized peptide substrate was compared to the ECM using BLAST (basic local alignment search tool) 23 . Interestingly, our peptide has a uniquely similar sequence to collagen bands (I, II, III, IV, V) with identities of 80%, 80%, 80%, 90%, and 80%, respectively, as shown in Table 1.
In our study, the cleavage site was determined based on the results of mass spectrometric analysis (Supplementary Information) of collagenase peptide fragments obtained from an enzymatic reaction with various enzymes tested. As noted in Fig. 1, the results show that Roche collagenase class I and II have similar cleavage sites on the peptide substrate, L ↓ G, and P ↓ G. However, Serva collagenase NB1 has three different cleavage sites: G ↓ P, P ↓ L, and P ↓ G. The neutral protease, which is produced by Clostridium histolysticum, has three cleavage sites: P ↓ L, L ↓ G, and P ↓ G, while thermolysin prepared from Bacillus thermoproteolyticus rokko has four cleavage sites: P ↓ L, L ↓ G, P ↓ G, and G ↓ G. Interestingly, beside collagenase NB1, all other enzymes share the same cleavage sites, P ↓ G and L ↓ G.
However, as shown in Fig. 2c,d,e,f, the reaction curves of both 0.3 μ g/ml of collagenase II (n = 3, Fig. 2c,d) and 17 μ g/ml of thermolysin (n = 3, Fig. 2e,f) did not significantly change when CaCl 2 was added. There were no significantly differences in terms of kinetic parameters V max and K m (collagenase II, p = 0.490; thermolysin, Effect of zinc ion on enzyme activity. This experiment was conducted since the chemical containing zinc ions are supplemented in culture media used for islet culture post isolation. The effect of zinc ions on enzyme activity was investigated using ZnSO 4 (16.7 μ M) and ZnCl 2 (16.7 μ M) solution. NaSO 4 (16.7 μ M) was used as a control to compare the Zn 2+ effect to the influence of Na + or SO 4 2− ions. The reaction was conducted in the presence and absence of CaCl 2 (4.14 mM). This concentration was found to be optimal for pancreas digestion 11 . The substrate used in this experiment was at a concentration of 20 μ M. As shown in Fig. 3, for all the four enzymes tested, the measured RFU was significantly decreased by adding either ZnSO 4 or ZnCl 2, regardless of presence or absence of CaCl 2 , whereas, no significant difference was found by adding , and 1790 ± 60 (100 mM). There were significant differences of RFU levels between the controls and various concentrations of EGTA (25, 50, and 100 mM) for collagenase I (0.45 μ g/ml), collagenase II (0.3 μ g/ml), and thermolysin (17 μ g/ml) (Fig. 4, p < 0.0001). EGTA is a mono calcium-chelating agent and therefore the presence of EGTA chelated Ca 2+ further substantiates the importance of CaCl 2 during digestion process.
Effect of pancreatic proteases on the peptide substrate. All the enzymes (collagenase I, collagenase II, Liberase MTF C/T, collagenase NB1, thermolysin, and neutral protease) cleaved the peptide substrate. In this study, thermolysin was used to compare with the pancreatic endogenous proteases. The results showed that thermolysin demonstrated a dose-dependent cleavage of peptide substrate both in the presence and absence of 4.14 mM CaCl 2.

Discussion
We describe a novel peptide substrate for measuring enzyme activities of collagenase I, II, Liberase MTF C/T, collagenase NB1, thermolysin, and neutral protease. Our substrate is unique since it cannot be cleaved by pancreatic endogenous proteases such as trypsin, chymotrypsin, and elastase. It is highly specific for bacterial enzymes, sensitive, simple, and fast to monitor enzyme activities. Currently, there is no standard single assay to measure activities of all enzymes available to digest the pancreas and to free the islets. Although the Wünsch assay has contributed tremendously to the development of collagenase, it can only measure collagenase class II. It cannot be used for measuring collagenase class I, thermolysin or neutral protease 9,16,24-28 . On the other hand, our peptide substrate can be used to measure the activities of these enzymes including thermolysin and neutral protease. Roche Liberase MTF C/T and Serva collagenase NB1/neutral protease are GMP products and currently being used for human islet isolation. In this study, we compared Liberase MTF collagenase and Serva collagenase NB1. It was clearly showed that enzyme activities from various manufacturers are variable as shown in unit per ml, indicating that these enzymes though produced from Clostridium histolyticum, the strain variable might influence the product and stability. Indeed, Liberase MTF collagenase and Serva collagenase NB1enzyme activities were different as shown in Table 3. The enzyme activities were calculated using the modification of previously described formulation 29 .  Adult human pancreata have strong collagen bands, which can be degraded by Liberase MTF C/T Roche enzymes 11 . However, enzyme digestion of pancreata from younger donors is challenging because islets are not free from acinar cells resulted into embedded/mantled islets 10,11,30,31 . Therefore, Serva collagenase NB1 and neutral protease were used for pancreas digestion to free the islets from younger donors. Furthermore, with the known sequence of the peptide substrate, analysis of the cleavage sites on peptide substrate upon enzymatic reaction became possible. Therefore, characterization of ECM of the pancreata from different age groups may help to develop a cocktail of multiple enzyme(s) using our peptide substrate, overcoming the variability among age groups. Interestingly, our peptide substrate (A G G P L G P P G P G G) has 80-90% similar structure of collagen bands: I, II, III, IV, and V as shown in Table 1, using BLAST and the amino acid sequences of collagens as described in the results section. Islet mass and quality are current index for determining enzyme efficacy. Thus, our peptide substrate may replace current expensive and challenging procedure for pancreas digestion to free the islets. It would be of immense benefit to evaluate multiple enzymes using this peptide substrate so that appropriate enzyme cocktails could be prepared for tissue dissociation with minimal cell death 24,32 .
Interestingly, the cleavage sites of our peptide substrate by different enzymes that currently being used by multiple centers showed that Liberase MTF C/T was quite different from Serva collagenase NB1 (Fig. 1). It is well known that collagenase is Ca 2+ dependent 11 . In this study, we performed enzyme kinetic study and used two important kinetic parameters V max and K m to compare the enzyme activity in the presence and absence of CaCl 2 . V max reflects how fast the enzyme can catalyze the reaction. In this specific study, the higher V max , the faster the Figure 5. Effects of pH on enzymatic activity. All the enzymes except collagenase I, there were significant differences between different levels of pH. For collagenase II, the reaction was significantly diminished at pH 8.0, compared with pH 6.8 and pH 7.5 (**p < 0.01). With regard to thermolysin, the reaction was significantly diminished at pH 8.0 when compared to all other pH conditions (****p < 0.0001). For neutral protease, there was significant difference between pH 7.0 and pH 8.0 (*p < 0.05). The assay was conducted using a substrate concentration of 20 μ M (n = 2).

Figure 6. Reaction of endogenous pancreatic proteases and peptide substrate.
Thermolysin demonstrated a dose dependent cleavage of peptide substrate both in the presence and absence of 4.14 mM CaCl 2 . All three pancreatic proteases: trypsin, chymotrypsin, and elastase showed no activity compared to thermolysin (2.5, 5, 10 μ M), regardless of the presence or absence of CaCl 2 (****p < 0.0001). The peptide substrate was used at a concentration of 50 μ M (n = 3). enzymes cleave the peptide substrate. Our data clearly showed that only collagenase I and neutral protease, the cleavage of the peptide was enhanced significantly in the presence of CaCl 2 reflected by significantly higher V max values, underlying that CaCl 2 should always be used regularly during the digestion process for islet isolation. The results also showed that the enzymatic cleavage was significantly inhibited or diminished in the presence of zinc. Previous studies have shown that collagenase enzymes bind to islets post culture and attributed this as the cause of the deteriorating effects resulting in islet loss 33 . However, collagenase activity was undetectable in culture media taken from islet samples post culture using our highly sensitive peptide substrate. The pH value of the reaction buffer affects enzyme activity. Our results indicated that the enzyme activity was not affected at pH 6.8, pH 7.0, and pH 7.5. Thus, it is suggesting that the pH should be kept within the physiological range during digestion process for optimum islet isolation, cell function and survival post transplantation. Within the four sets of pH conditions in this experiment, only the pH 8.0 diminished enzyme activity, therefore the pH 8.0 condition is not recommend to perform enzyme assay when using our peptide substrate.
There are multiple endogenous pancreatic proteases such as trypsin, chymotrypsin, and elastase, which play an important role in hydrolysis of complex nutrients 34,35 . These proteases could be activated during the digestion process and thus may influence islet quality and yields [36][37][38] . However, our results showed that our peptide substrate cannot be cleaved by the three predominantly endogenous pancreatic proteases using similar conditions for evaluating microbial enzymes.
In conclusion, we synthesized a novel peptide that can be utilized for quantifying bacterial enzyme activity using a FRET assay. The peptide is microbial collagenase, thermolysin/neutral protease specific, and could not be degraded by endogenous pancreatic proteases such as trypsin, chymotrypsin, and elastase. The method can be applied to measure enzyme activities of the products from various sources. Furthermore, this method can be used to monitor collagenase activity and examine the effects that any additional activator or inhibitor may have on enzymatic function, particularly during the tissue digestion process. Synthesis of Peptide Substrate. Peptide synthesis was designed and performed according to the general methods outlined in Kaplan et al. 39 with some modifications using Fmoc-Lys (dde)-OH and K (dabcyl). After removal of the terminal Fmoc group, 5(6)-carboxyfluorescein was activated. The excess reagents were washed out with piperidine: DMF (1:4). The Lys (Dde) was deprotected using 2% hydrazine in DMF, washed with DMF and DCM and 4-dimethylaminoazobenzene-4′ -carboxylic acid (dabcyl) was activated in the usual manner and coupled as any standard amino acid. The resin was washed and the peptide was cleaved from the resin by standard methods. The peptide was purified by HPLC as described previously 39 and the sequence was confirmed by mass spectrometry. The stock peptide substrate was dissolved in DMSO and stored at − 80 °C (light protected) until used.  FRET Assay for Enzymatic Reaction. The reaction kinetics between enzymes and newly synthesized peptide substrate were evaluated using FRET assay. In brief, fluorogenic peptide substrate was synthesized in which donor and quencher (dabcyl) molecules were attached to corresponding amino acids. After cleaving by the enzymes, fluorophore quenching was diminished due to the separation of donor and quencher moieties [40][41][42] . As a result, the donor fluorescence increases dramatically, which can be measured with excitation 485 nm and emission at 535 nm 43 . The enzymatic reaction was performed in a 96-well round bottom black plate (Costar ID#3915, Corning, NY). Briefly, 190 μ l of substrate with a final concentration of 5-80 μ M and 10 μ l of specific enzyme samples were added to each well, creating a final volume of 200 μ l/well. A blank sample was used containing only substrate with no enzyme. In addition, enzyme samples were boiled for 5 min at 80 °C and used as a control. The plate was incubated (light protected) for 1 hr at 22 °C and the reaction was stopped by adding 50 μ l of 40 mM EDTA solution (pH 8.0). The fluorescence was read at excitation 485 nm and emission 535 nm using the Tecan Magellan V 6.5 Genios plate reader (Tecan Systems, Inc., San Jose, CA, USA). Microsoft excel was used to further extrapolate the data.

Chemicals
Enzyme Kinetic Assay. The Michaelis-Menten Model of enzyme kinetics and Lineweaver-Burk plot and equation were used to establish the rate of the enzymatic reaction in relationship to the substrate using known substrate concentrations 43 . The Michaelis-Menten model was chosen for analysis as it allows for the comparison of different enzyme enhancers and the effect of these activators on the enzymatic reaction 44 . The V max and K m values were calculated using GraphPad Prism. The substrate concentrations used in this kinetics study were: 5, 10, 20, 40, 60 and 80 μ M. The concentrations of the tested enzyme were: 0.45 μ g/ml for collagenase I, 0.3 μ g/ml for collagenase II, 17 μ g/ml for thermolysin, and 20 μ g/ml for neutral protease. In this study, using the newly synthesized peptide substrates and enzymes, we tested following influential factors on enzyme activity: (i) effect of CaCl 2 (4.14 mM) on enzymatic kinetics; (ii) effect of ZnSO 4 (16.7 μ M), ZnCl 2 (16.7 μ M), NaSO 4 (16.7 μ M); (iii) effect of chelating agents EGTA (25, 50, 100 mM); (iv) effect of pH (pH 6.0-8.0); (v) scaling-up concentrations (2.5, 5, 10 μ g/ml) of pancreatic endogenous proteases (trypsin, chymotrypsin and elastase) were also used to examine the efficiency to cleave peptide substrate.
Statistical Analysis. GraphPad Prism (GraphPad Software 6.0, La Jolla, CA, USA) was used for analyzing the data and generating the graphs. Both non-linear Michaelis-Menten and linear Lineweaver-Burk plots were used for the enzyme kinetics study. Kinetic parameters V max and K m were obtained from the Michaelis-Menten model by plotting the reaction velocity at different concentration of peptide substrate. Non-linear Michaelis-Menten curves for the enzymatic reaction in the presence and absence of CaCl 2 were compared using Prism. All samples were run in duplicates, and results were reported as average ± standard error of the mean (SEM). One-way or two-way ANOVA analysis followed by Tukey's multiple comparisons test was used to conduct multiple variable comparisons when applicable. Differences in data were considered significant when P values were less than 0.05. Coefficient of Variation (CV) was calculated using Microsoft Office Excel 2011.