Conceived and designed the experiments: JE BM EM MTB GMG CK. Performed the experiments: JE BM EM. Analyzed the data: JE BM EM MTB GMG CK. Contributed reagents/materials/analysis tools: JE BM EM MTB. Wrote the paper: JE BM MTB GMG CK.
C.K. is a Director & Stockholder in Alvine Pharmaceuticals, a company that is developing an oral enzyme drug for celiac disease. None of the other authors have any financial interests to disclose.
Celiac sprue is a life-long disease characterized by an intestinal inflammatory response to dietary gluten. A gluten-free diet is an effective treatment for most patients, but accidental ingestion of gluten is common, leading to incomplete recovery or relapse. Food-grade proteases capable of detoxifying moderate quantities of dietary gluten could mitigate this problem.
We evaluated the gluten detoxification properties of two food-grade enzymes, aspergillopepsin (ASP) from
ASP markedly enhanced gluten digestion relative to pepsin, and cleaved recombinant α2-gliadin at multiple sites in a non-specific manner. When used alone, neither ASP nor DPPIV efficiently cleaved synthetic immunotoxic gluten peptides. This lack of specificity for gluten was especially evident in the presence of casein, a competing dietary protein. However, supplementation of ASP with DPPIV enabled detoxification of moderate amounts of gluten in the presence of excess casein and in whole-wheat bread. ASP was also effective at enhancing the gluten-detoxifying efficacy of cysteine endoprotease EP-B2 under simulated gastric conditions.
Clinical studies are warranted to evaluate whether a fixed dose ratio combination of ASP and DPPIV can provide near-term relief for celiac patients suffering from inadvertent gluten exposure. Due to its markedly greater hydrolytic activity against gluten than endogenous pepsin, food-grade ASP may also augment the activity of therapeutically relevant doses of glutenases such as EP-B2 and certain prolyl endopeptidases.
Celiac sprue is an inheritable, life-long disease that is characterized by an inflammatory reaction to dietary gluten in the human small intestine. The hallmark of the disease is a characteristic flattening of intestinal villi along with crypt hypertrophy. As a consequence, there is tremendous loss of surface area and malabsorption of nutrients, vitamins and minerals. If untreated, celiac sprue is associated with complications such as anemia, bone diseases, infertility, neurological problems, cancer and other complications due to persistent inflammation and micronutrient deficiencies. Screening studies predict that approximately 1% of the United States' population has the disease, yet only ca. 10% of affected individuals have been diagnosed thus far
A potential cost-effective solution to the aforementioned problem is an oral protease or protease mixture that is composed solely of commercially available food-grade enzymes. In this study we have evaluated the gluten detoxification properties of one such enzyme cocktail. It includes at least two proteases, aspergillopepsin (ASP) from
The Materials Safety Data Sheet of the commercial ASP powder from Bio-Cat Inc. designates this enzyme as aspergillopepsin from
The Materials Safety Data Sheet of the commercial Peptidase P powder from Bio-Cat Inc. designates this enzyme as dipeptidyl peptidase IV from
The total protein content and specific activity of food-grade ASP and DPPIV evaluated in this study is summarized in
ASP | DPPIV | |
Total protein (t = 0) | 14±0.6% | 16±0.6% |
Total protein (t = 6 mo) | 13±0.4% | 15±0.7% |
Specific activity (t = 0) | 15.2±0.2 units/µg | 2.3±0.1 units/g |
Specific activity (t = 6 mo) | 13.4±0.4 units/µg | 2.7±0.1 units/g |
Each measurement is an average of four enzyme samples, stored and analyzed independently.
It is well established that mammalian pepsin exerts only minimal proteolytic action on dietary gluten. Our search for food-grade gluten detoxifying enzymes therefore led us to aspergillopepsin as a potentially more potent analog of mammalian pepsin. We first wished to investigate the specificity of ASP towards recombinant gluten proteins and synthetic gluten peptides. Available literature suggests that ASP cleaves proteins relatively non-specifically. For example, ASP cleaves ribonuclease A at Tyr-X, Phe-X, His-X, Asn-X, Asp-X, Gln-X and Glu-X bonds
(A) Reverse-phase HPLC traces of the 33-mer peptide incubated with ASP in a 20∶1 (w/w) peptide:enzyme ratio at pH 4.5. (B) Reverse-phase HPLC traces of the 33-mer peptide mixed with gluten digested by pepsin, and incubated with ASP in a 20∶80∶1 (w/w/w) peptide:gluten:enzyme ratio at pH 4.5. The ordinate scale is identical in all traces. The peak at 15 min corresponds to an internal standard (TAME).
Although ASP is not specific for immunotoxic gluten epitopes, it may potentiate gluten detoxification by cleaving larger proteins into short peptides that are accessible substrates for more specific endopeptidases and exopeptidases
The purified, recombinant protein substrate was incubated for 15 min with food-grade ASP in a 50∶1 (w/w) substrate:enzyme ratio at pH 4.5. Cleavage sites were identified by LC/MS/MS, and are indicated by arrows.
Based on these findings with structurally defined substrates, we hypothesized that, although ASP was considerably more effective than mammalian pepsin at cleaving gluten into short peptides, it would have limited ability to detoxify gluten in the context of a complex meal. To test this hypothesis, we compared the activity of pepsin versus ASP against 15 mg/ml whole gluten or 15 mg/ml gluten mixed with 50 mg/ml casein under simulated gastric conditions (0.03 M HCl). The markedly superior activity of ASP against whole gluten is illustrated in
Whole gluten powder (15 mg/ml) was incubated with either 0.35 mg/ml ASP, 0.6 mg/ml pepsin, or both enzymes at 37°C for 60 min at pH 2. The resulting product mixture was analyzed by reverse-phase HPLC, as described in the text. The breakdown of longer gluten peptides into shorter ones is generally indicated by a decrease in absorbance at higher retention times and a concomitant increase in absorbance at lower retention times. The peak at 15 min corresponds to an internal standard (TAME).
(A) Whole gluten powder (15 mg/ml) or (B) whole gluten powder (15 mg/ml) mixed with casein (50 mg/ml) was incubated with either 0.6 mg/ml pepsin or 0.35 mg/ml ASP at 37°C for 15 or 60 min in 0.03 M HCl. Gliadin peptides present in the digests were analyzed by competitive ELISA using the G12 monoclonal antibody, which is specific for the immunotoxic gliadin epitope QPQLPY.
Preliminary studies with a variety of other over-the-counter enzymes led us to conduct more detailed evaluations of peptidase P from
Whole gluten powder (50 mg/ml) was incubated for 60 min at 37°C with 0.35 mg/ml ASP in the presence or absence of 0.7 mg/ml DPPIV. Because this enzyme preparation is inactive at pH <4, the reaction was conducted in 0.03 M HCl in the presence of 1.5 mg/ml CaCO3 to simulate addition of a standard over-the-counter antacid/calcium supplement. The pH of the resulting mixture varies between 4–5 for the duration of the reaction. The peak at 15 min corresponds to an internal standard (TAME).
We first wished to verify the proteolytic action of ASP and DPPIV on baked gluten in whole wheat bread. For this purpose digestions and sample analyses were performed under simulated gastric conditions developed earlier for bread
The reverse phase HPLC traces show the residual peptide content after whole wheat bread was incubated with pepsin (0.6 mg/ml) under simulated gastric conditions, followed by trypsin (0.375 mg/ml) and chymotrypsin (0.375 mg/ml) under simulated duodenal conditions (PTC) or PTC+ASP (0.35 mg/ml)+DPPIV (0.7 mg/ml). The peak at 15 min corresponds to an internal standard (TAME). For experimental details, see text.
To assess the dependence of gluten proteolysis on ASP and DPPIV dose, we digested 50 mg/ml gluten in the presence of 0.03 M HCl and 3 mg/ml CaCO3 by varying the concentration of one enzyme while keeping the other constant. Representative data is shown in
Whole gluten powder (50 mg/ml) was incubated for 60 min at 37°C with 0.7 mg/ml DPPIV and variable ASP concentration (A) or 0.35 mg/ml ASP and variable DPPIV concentration (B). The reactions were conducted in 0.03 M HCl in the presence of 3 mg/ml CaCO3. The peak at 15 min corresponds to an internal standard (TAME).
Additionally, to assess the potential of the above ASP+DPPIV cocktail for detoxifying gluten under conditions that more accurately mimic the continuously changing environment of the post-prandial stomach, gluten (alone or in combination with casein) was digested with either pepsin or ASP+DPPIV. This experiment was started by adding gluten (or gluten plus casein), an appropriate quantity of each enzyme and 3 mg/ml CaCO3 to 0.03 M HCl to simulate the entry of food and a physiological quantity of antacid in the empty stomach. At this point the pH of the mixture was 3 (in the case of gluten alone) or 4 (in the case of gluten+casein). Every 5 min thereafter, an additional 0.01 M equivalent of concentrated HCl was added in order to simulate the periodic squirting of acid into the fed stomach. After 1 hr, the pH was 1 (in the case of gluten alone) or 2.5 (in the case of gluten+casein). At this time-point, samples were withdrawn from each reaction mixture and analyzed via a competitive ELISA assay that is specific for a representative immunotoxic epitope. Treatment of gluten or gluten plus casein with ASP+DPPIV reduced the epitope concentration by 7- or 4-fold, respectively, compared to the samples treated with pepsin only (
(A) Whole gluten powder (15 mg/ml) or (B) whole gluten powder (15 mg/ml) mixed with casein (50 mg/ml) was incubated with either 0.6 mg/ml pepsin or 0.35 mg/ml ASP and 0.7 mg/ml DPPIV at 37°C for 60 min. The digestions were initiated in 0.03 M HCl and 3 mg/ml CaCO3, and every 5 min additional HCl was added. Gliadin peptides present in the digests were analyzed by competitive ELISA using the G12 monoclonal antibody, which is specific for the immunotoxic gliadin epitope QPQLPY.
Finally, to confirm that the reduction in QPQLPY epitope measured by competitive ELISA (
Whole gluten powder (15 mg/ml) mixed with casein (50 mg/ml) was incubated with either 0.6 mg/ml pepsin or 0.35 mg/ml ASP and 0.7 mg/ml DPPIV. The digestions were initiated in 0.03 M HCl and 3 mg/ml CaCO3, and every 5 min additional HCl was added. After digestion for 60 min at 37°C, the samples were further treated with 0.375 mg/ml trypsin and 0.375 mg/ml chymotrypsin for 30 min at pH 6.0. The immunotoxic peptide content in the water-soluble fraction of the digests was measured by a T-cell proliferation assay. A stimulation index of 1 indicates background levels of T-cell proliferation and is denoted with a horizontal line. The name of the individual T-cell lines is indicated in the graph.
As discussed above, an attractive feature of aspergillopepsin is that, unlike mammalian pepsin, aspergillopepsin is able to extensively hydrolyze dietary gluten into short peptides. This finding suggests that ASP should be able to complement the glutenase activities of other promising enzymes such as the glutamine-specific cysteine endoprotease EP-B2 from barley
Whole gluten powder (15 mg/ml) and casein (50 mg/ml) were digested with 0.6 mg/ml pepsin for 60 min at pH 4 (gastric conditions, GC), followed by treatment with 0.375 mg/ml trypsin and 0.375 mg/ml chymotrypsin for 30 min at pH 6.0 (duodenal conditions, DC). Where indicated, 0.35 mg/ml ASP, and/or 0.375 mg/ml proEP-B2 were added at the beginning of the gastric digestion phase. Gliadin peptides present in the digests were analyzed by competitive ELISA using the G12 monoclonal antibody, which is specific for the immunotoxic gliadin epitope QPQLPY.
There is a cogent need for some form of near-term supportive therapy for celiac sprue, because dietary gluten exclusion is often less than complete. Patients on a restricted diet frequently suffer nutrient and mineral malabsorption, resulting in deficiencies of iron or calcium, and consequently in anemia and osteoporosis, respectively. This has motivated us to evaluate commercially available food-grade enzymes for their ability to detoxify low-to-moderate doses of gluten. By way of calibration, one slice of wheat bread contains ca. 3–4 g gluten, whereas the threshold of “safe” gluten for celiac patients is estimated to be in the 10–100 mg range
Aspergillopepsin, unlike mammalian pepsin, is able to extensively hydrolyze dietary gluten into short peptides. However, in contrast to glutenases such as prolyl endopeptidases or barley EP-B2, which have high specificity for immunotoxic Pro- and Gln-rich gluten peptides such as the 33-mer from α2-gliadin, ASP lacks specificity for these peptides. By virtue of its relatively low sequence specificity, it can hydrolyze these peptides, but the presence of competing substrates slows this hydrolysis considerably. The extensive exoproteolytic action of a second enzyme such as the fungal DPPIV accelerates clearance of short peptides, thereby facilitating the access of longer gluten peptides to the ASP active site and their eventual detoxification
Food-grade aspergillopepsin (ASP) from
Gluten peptides were synthesized on solid-phase using Boc/HBTU chemistry, purified by reverse phase HPLC, and lyophilized as described
The identity of aspergillopepsin was confirmed via N-terminal sequence analysis and mass spectrometry of a trypsin digest of the major protein observed at 41 kD by SDS-PAGE. Due to its low abundance in the commercial enzyme powder from
The specificity of ASP was assessed by LC/MS/MS analysis of α2-gliadin digested with the commercial enzyme powder from
The protein concentration in each commercial enzyme preparation was determined by the Bradford protein assay. A standard calibration curve was generated using bovine serum albumin in the concentration range of 2–12 µg/ml.
ASP activity was measured using the spectrophotometric hemoglobin units of tyrosine (HUT) assay. The amount of tyrosine liberated as trichloroacetic acid-soluble peptides upon hemoglobin digestion was quantified by monitoring absorbance at 280 nm. In a total reaction volume of 1.5 ml, 1.3% (w/v) of bovine hemoglobin was reacted at 37°C with three separate enzyme concentrations (final concentrations of 1.7 µg/ml, 5 µg/ml, and 8 µg/ml on a total protein basis). After 10 min, the reaction was quenched using trichloroacetic acid (TCA, Sigma 490–10) added to a final concentration of 3.2% (w/v). Samples were centrifuged and the A280 was recorded. One HUT unit of protease activity is defined as that amount of enzyme that produces an A280 of 0.001 per min at pH 2.0 and 37°C, measured as TCA-soluble products using hemoglobin as a substrate (final volume = 4 ml, light path = 1 cm).
DPPIV activity was measured via a standard kinetic assay, using the chromogenic substrate Gly-Pro-p-nitroanilide dissolved in phosphate buffered saline, pH 7.4. Absorbance was monitored at 410 nm. One unit of DPPIV activity is defined as the amount of enzyme that produces 1 µmol of p-nitroaniline per min at pH 4.5 and room temperature.
To evaluate the gluten detoxifying activity of a given enzyme or enzyme cocktail, an experimental protocol developed earlier was used to mimic the gastric digestion of either whole gluten or whole wheat bread
To simulate duodenal digestion, we adjusted the pH to 6.0 at the end of the gastric phase. Pancreatic enzymes (trypsin and chymotrypsin) were added to yield final concentrations of 0.375 mg/ml each. Addition of elastase and carboxypeptidase A have minimal incremental effect on gluten digestion and thus were not added
Digested gluten or bread samples were centrifuged for 10 min at 13,400•g and filtered through a 0.2 µm low protein binding filter. The filtrate was chromatographically separated on a 4.6×150 mm reverse phase C18 protein and peptide column (Vydac, Hesperia, CA) using Dynamax SD-200 pumps (Varian, Palo Alto, CA) (1 ml/min), a Varian 340 UV detector set at 215 nm and a Varian Prostar 430 autosampler. Solvent A was water with 5.0% acetonitrile and 0.1% TFA. Solvent B was acetonitrile with 5.0% water and 0.1% TFA. Samples were analyzed using a gradient described previously
Relative amounts of immunotoxic gliadin epitopes in any gluten-containing sample were quantified by a competitive enzyme-linked immunoabsorbent assay (ELISA) using the horseradish peroxidase-conjugated G12 monoclonal antibody (Biomedal, Seville, Spain)
The content of 33-mer peptide present in gluten digests was determined by triple quadrupole LC-MS. Whole gluten (15 mg/ml) and casein (50 mg/ml) were digested with pepsin for 60 min at pH 4.0 followed by treatment with 0.375 mg/ml trypsin and 0.375 mg/ml chymotrypsin for 30 min at pH 6.0. In designated digests, pepsin was supplemented with 0.35 mg/ml ASP and/or 0.375 mg/ml EP-B2. Heat-quenched digests were centrifuged (16,100 x g) and the supernatants were diluted in an equal volume of cold acetonitrile containing 0.1% formic acid to precipitate larger proteins. Samples were vortexed, incubated for 2 h at 4°C, and centrifuged (16,100 x g) for 10 min at 4°C. Supernatants were mixed with an equal volume of 0.1% formic acid in water to dilute the acetonitrile concentration to 25%, and injected on a Micromass Quattro Premier triple quadrupole LC-MS system. The concentration in each sample was determined by comparison to a 33-mer standard curve. The limit of quantification was 2 nM.
Digests were treated with 100 µg/ml recombinant human transglutaminase 2 in 200 mM MOPS, pH 7.2, and 15 mM CaCl2 for 2 h at 37°C. The samples were heated at >95°C for at least 5 min to inactivate the enzyme, centrifuged for 2 min at 13,000 rpm to pellet insoluble material, and frozen at −20°C until use in T cell proliferation assays.
T-cell proliferation assays were performed using DQ2 homozygous 9088 cells and T-cell lines P28 TCL1 and P34 TCL1 as described in Siegel et al. (
We thank Jonathan Gass for helpful advice during the initial stages of this research.