Interlaboratory validation of an optimized protocol for measuring α-amylase activity by the INFOGEST international research network

Participating laboratories

Coordinating laboratory: Teagasc Food Research Centre, Moorepark, Fermoy, Co Cork P61 C996, Ireland.

Participating laboratories:

• Laboratory of Food Chemistry and Biochemistry, Department of Food Science and Technology, School of Agriculture, Aristotle University of Thessaloniki, P.O. Box 235, 54124, Thessaloniki, Greece

• Global Oatly Science and Innovation Centre, Rydbergs Torg 11, Space Building, Science Village, 22 484 Lund, Sweden

• Laboratory of Food Technology, Department of Microbial and Molecular Systems (M2S), KU Leuven, Kasteelpark Arenberg 23, PB 2457, 3001, Leuven, Belgium

• INRAE, Institut Agro, STLO, 35042 Rennes, France

• School of Biosciences, Faculty of Health and Medical Sciences, University of Surrey, Guildford, GU2 7XH, United Kingdom

• Nofima AS, Norwegian Institute of Food, Fisheries and Aquaculture Research, PB 210, N-1433, Ås, Norway

• Center for Innovative Food (CiFOOD), Department of Food Science, Aarhus University, Agro Food Park 48, Aarhus N 8200, Denmark

• Department of Horticulture, Martin-Gatton College of Agriculture, Food and Environment, University of Kentucky, Lexington, Kentucky, USA

• Department of Agricultural, Food, Environmental and Animal Sciences, University of Udine, Italy

• Wageningen Food & Biobased Research, Wageningen University & Research, 6708 WG Wageningen, The Netherlands

• Quadram Institute Bioscience, Rosalind Franklin Road, Norwich Research Park, Norwich, NR4 7UQ, United Kingdom

• Department of Food Engineering, Faculty of Engineering, Ege University, 35100, İzmir, Türkiye

Materials

The chemicals and four test products used in the ring study are presented in (Table 3). They were ordered by the coordinating laboratory, aliquoted and shipped to each of the participating laboratories. All laboratories received aliquots from the same batch of each product, with the exception of 3,5-dinitrosalicylic acid (DNSA) which came from two different lots. Prior to shipping, calibration curves established with solutions prepared from both of these lots were compared, and showed nearly equivalent results (Figure S1 in Supplementary material-Section “Protocol implementation at each laboratory”).

Equipment needed

The list of equipment required is provided as guidance below.

Preparation of reagents and enzyme solutions

Vortex mixer, pH meter with glass electrode, heating/stirring plate, incubator.

Enzyme assay

Water-bath or thermal shaker (e.g. PCMT Thermoshaker, Grant Instruments, United Kingdom) for enzyme–substrate incubations at 37 °C. Boiling bath (e.g. SBB Aqua 5 Plus, Grant Instruments, United Kingdom) or thermal shaker (e.g. PCMT Thermoshaker, Grant Instruments, United Kingdom) suitable for use at 100 °C. Spectrophotometer (e.g. Shimadzu UV-1800 Spectrophotometer, Shimadzu Corporation, Japan) or plate reader (e.g. BMG Labtech CLARIOstar Plus, BMG Labtech, Germany).

Basic materials

Volumetric flasks, heatproof bottle with lid (e.g. Duran bottle), magnetic stirrer, timer, thermocouple, safe lock microtubes (2 or 1.5 mL), heat (and water) resistant pen or labels for the microtubes, disposable standard cuvette or disposable polystyrene 96-well plate.

Preparation of reagents and enzymes

20 mM Sodium phosphate buffer (with 6.7 mM sodium chloride, pH 6.9 ± 0.3)

Prepare a stock solution by dissolving 1.22 g NaH₂PO₄ (anhydrous form), 1.38 g Na₂HPO₄ (anhydrous form) and 0.39 g NaCl in 90 mL purified water and make up the volume to 100 mL. Before use, dilute 10 mL of stock solution to 95 mL with purified water. Confirm that the pH of the buffer, when heated to 37 °C, is within the required working range (pH 6.9 ± 0.3). If needed, adjust the pH, using 1 M NaOH or HCl as required, before making up the volume to 100 mL.

Maltose calibrators

Prepare a 2% (w/v) maltose stock solution in phosphate buffer. Prepare a calibrator series by diluting the maltose stock solution in phosphate buffer as indicated in Table S2 (Supplementary Material – Section “Protocol implementation at each laboratory”). Store in the fridge (or freezer if not for use during the same day).

Colour reagent (96 mM DNSA with 1.06 M sodium potassium tartrate)

Dissolve 1.10 g of DNSA in 80 mL of 0.50 M NaOH at 70 °C in a glass beaker or bottle (partly covered to limit evaporation) on a pre-heated heat/stir plate with continuous stirring and temperature monitoring (e.g. using a thermocouple). Once the DNSA is fully dissolved, add 30 g of sodium potassium tartrate and continue stirring until it dissolves. Remove from heat and wait until the solution cools to room temperature. Bring to 100 mL with purified water. Store at room temperature protected from light for up to 6 months. If precipitation occurs during storage, re-heat to 45 °C while stirring on a heat-stir plate.

Starch solution

Potato starch pre-gelatinized in sodium phosphate buffer (1.0% w/v) is used as substrate. Pre-heat a heat-stir plate (setting it to 250 °C—300 °C is suggested) and pre-heat an incubator (or water bath) to 37 °C. Weigh 250 mg of potato starch into a heatproof bottle and add 750 μL of ethanol (80% v/v). Stir on a vortex mixer to wet all the starch powder (this is a critical step for the complete solubilisation of the starch). Add 20 mL of sodium phosphate buffer and mix again using a vortex mixer making sure that the powder is fully dispersed and there are no lumps in the solution. Cover the bottle with the lid to minimize evaporation (but making sure it is loose enough to let out excess steam) and place on the pre-heated heat-stir plate stirring at 180 rpm. When the solution starts bubbling, start the timer and boil on the heat-stir plate stirring continuously for exactly 15 min. Cool in the incubator/water bath for 15 min (or until it is safe to handle). Make up the volume of the starch solution to 25 mL in a volumetric flask by adding purified water. Store the solution in a closed bottle in an incubator (or water bath set to 37 °C) and use within 2 h. If the starch solution does not clarify significantly a new solution needs to be prepared, as this may indicate poor solubilisation and or gelatinization of the starch. Prepare a fresh solution each time as storing or freezing can cause starch retrogradation and influence the results of the assay.

α-amylase solutions

The preparation of the enzyme solutions is a critical step. Solutions prepared from enzyme powders should be carefully prepared following the same protocol each time to ensure adequate powder hydration and dispersion. After weighing the enzyme powder and adding the adequate amount of sodium phosphate buffer, stock solutions should be stirred in an ice bath (at around 250 rpm) for 20 min before any further dilutions (Graphical protocol in Fig. 6 and Picture S1 in the Supplementary Material). Subsequent dilution(s) of the stock solution(s) should be performed using sodium phosphate buffer to reach the recommended enzyme concentration of 1.0 ± 0.2 U/mL. For the four products tested in the ring trial, recommended concentrations are provided as reference in Table S7 (Supplementary material). For enzyme preparations, it is recommended to start from a stock solution prepared by adding 20 – 100 mg of enzyme powder to 25 mL of sodium phosphate buffer. For human saliva, a stock solution can be prepared by mixing 80 µL of saliva with 920 µL of buffer.

Each enzyme should be tested at three different concentrations prepared by diluting 0.65 mL, 1.00 and 1.50 mL of enzyme stock solution with 1.35, 1.00 and 0.50 mL of buffer, respectively (Table S3). These diluted enzyme solutions are referred to as solutions C1, C2 and C3. Enzyme solutions should always be kept on ice and used within 30 min of preparation.

Enzymatic assay

An overview of the enzyme assay is presented in (Fig. 6).

Preparative procedures

Before starting, the following preparations are recommended: set the heating-block (water bath) as required to ensure 37 °C inside the microtubes (see troubleshooting advice, Table 2); pre-warm the starch solution to 37 °C; prepare a polystyrene container with ice.

Sample collection tubes

For each incubation that will be carried out, label and pre-fill four microtubes with 75 μL of DNSA colour reagent.

Incubations

Set three microtubes (one for each diluted enzyme solution C1, C2 and C3) in the preheated thermal shaker and let the temperature equilibrate before adding 500 µL of pre-warmed potato starch solution to each tube (maintain the tubes closed until the enzyme is added to prevent evaporation). Add 500 µL of diluted enzyme solution C1, C2 and C3 to the corresponding tubes at regular intervals. It is recommended to start the timer immediately when the α-amylase solution is added to the first tube and leave a 30 s interval before each subsequent addition.

Sample collection

Take a 150 μL aliquot from each tube after 3, 6, 9 and 12 min of incubation (respecting the order and intervals at which the incubations were initiated) and transfer it immediately to the corresponding sample collection tube pre-filled with DNSA to stop the reaction. Each aliquot should be taken as closely as possible to its respective sample collection time, within a maximum of ± 5 s.

Absorbance measurements

Prepare the maltose calibrators by mixing 150 µL of each maltose calibrator with 75 µL of DNSA reagent. Centrifuge the samples and calibrators (1000 g, 2 min) so that all droplets are brought back into solution. Place the samples and calibrators in the thermal shaker (or boiling bath) (100 °C, 15 min) and then transfer them to an icebox to cool for 15 min. Add 675 µL of purified water to each tube and mix by inversion. Transfer the samples and calibrators to a cuvette or pipette to a microtiter plate (300 µL per well) and record the absorbance at 540 nm (A_540nm).

Ring trial organization

Preliminary testing

Throughout the protocol optimization phase, the assay was repeated multiple times by the coordinating laboratory to define practical aspects. Each of the four test products has been assayed at different concentrations. The final test concentrations were defined by choosing a test concentration that allowed for an adequate distribution of the endpoint measure (spectrophotometry absorbance) and communicated to the participating laboratories.

Protocol transference

A detailed written protocol (Supplementary material) was transferred to each participating laboratory including the recommendations for concentrations of the test products. All laboratories were invited to an online training session that included a video of the assay followed by a Q&A session to clarify any doubts. All labs carried out the assay and reported their results on a standard Excel file between May and November 2023.

Incubation temperatures

All laboratories tested the four enzyme preparations at 37 °C as described above. A subgroup of five laboratories also repeated the assays at 20 °C with the purpose of trying to establish a correlation between the results obtained at both temperatures.

For incubations at 20 °C protocol adaptations were performed as follows. A different recipe was used to prepare the 200 mM sodium phosphate buffer stock solution. It consisted of 1.26 g NaH₂PO₄, 1.29 g Na₂HPO₄ and 0.39 g NaCl. The dilutions (10 mL stock diluted to 95 mL with purified water) and pH (6.9 ± 0.3) were the same as those for the buffer used at 37 °C. All reagents and solutions requiring the use of buffer were freshly prepared using this buffer recipe. The recommended concentrations of the α-amylase stock solutions were adjusted to ensure that enough enzymatic activity was present.

Calculations

Calibration curve

The A_540nm of the colour reagent blank was subtracted from the readings of all maltose calibrators and their concentration (mg/mL) was plotted against the corresponding ΔA_540nm. For reference purposes, using a 96 well plate, the absorbance at 540 nm should increase linearly from approximately 0.05 (for the colour reagent blank) to 1.5 for the highest maltose concentration. The calibration blank should not be included as a data point in the calibration curve.

Enzyme activity definition

The definition of α-amylase activity resulting from the application of the newly developed protocol is the following:

• Based on the definition originally proposed by Bernfeld: one unit liberates 1.0 mg of maltose equivalents from potato starch in 3 min at pH 6.9 at 37 °C.

• Based on the international enzyme unit definition standards: one unit liberates 1.0 μmol of maltose equivalents from potato starch in 1 min at pH 6.9 at 37 °C.

Amylase activity units based on the definition originally proposed by Bernfeld were multiplied by the conversion factor 0.97 to convert the result into IU.

Enzyme activity calculation

The first step was to subtract A_540nm of the colour reagent blank from all readings. The calibration curve was then used to calculate the maltose concentrations (mg/mL) reached with each diluted enzyme solution (C1, C2 and C3) at each sampling point during incubations. Enzyme concentrations during incubations were then calculated as mg/mL for enzyme powders, or µL/mL for liquid (saliva) samples.

For each diluted enzyme solution (C1, C2 or C3), maltose concentrations (mg/mL) were plotted against time (t_min) and the corresponding linear regression was established to determine the reaction kinetics’ slope ((text{m}t{text{min}})). For each enzyme concentration, units of enzyme were calculated using the following equation.

A template Excel file is provided for calculations in the Supplementary Material.

Statistical analysis and assessment of method’s performance

Data visualization and statistical analyses have been performed in R (version 4.3.2)²⁹. The packages ggplot2³⁰ and ggdist³¹ have been used in the preparation of the plots presented in the manuscript.

Outlier analysis was conducted on non-transformed data to preserve the original variability and scale of the datasets. First, Cochran’s test (outliers package in R³²) was used to assess intralaboratory variability and did not reveal any outliers. Subsequently, for interlaboratory comparisons, boxplot analysis, Bias Z-scores and Grubbs’ test³² were employed complementarily. The results reported by one lab for three test products (pancreatin, α-amylase M and α-amylase S) assayed at 37 °C were more than 1.5 interquartile ranges below the 25^th or above the 75^th percentiles, consistent with unsatisfactory Bias Z-scores (|z|> 3). Grubb’s test confirmed these as outliers and they have been excluded from the statistical analysis. All results in the 20 °C dataset fell within 1.5 interquartile ranges of the 25^th and 75^th percentiles (Fig. 5), consistent with satisfactory Bias Z-scores (|z|< 2) (Supplementary Figure S4). While Grubbs’ test identified two potential outliers (Lab A for pancreatin and Lab D for α-amylase M), this outcome was considered less reliable due to the small sample size (n = 5) and lack of corroboration from boxplot and Bias Z-score analyses, and so these results were retained.

Statistical analysis of the dataset resulting from the implementation of the protocol at 37 °C has been carried out to investigate the effects of the tested products, concentrations and incubation conditions (thermal shaker vs. water bath with or without shaking) as well as the two-way and three-way interactions between these factors. Normality of this dataset has been confirmed through the Shapiro–Wilk test (p > 0.05). The homogeneity of variances, as assessed using Levene’s test in the Rstatix package version 0.7.2³³, was not confirmed (p < 0.001). Due to the limited availability of suitable non-parametric alternatives, a logarithm transformation was performed on this data set enabling homogenisation of the variances and application of a three-way ANOVA (Rstatix package). Statistically significant effects were further examined using Pairwise T-Test comparisons, applying Bonferroni adjustments for multiple comparisons as required. The results obtained when implementing the protocol at 20 °C were normally distributed, but homogeneity of variances was not confirmed for this dataset either. The corresponding logarithm transformed data frame did not conform to normality, hence the Kruskal–Wallis test was applied to examine the significance of the differences between the four products, followed by the Bonferroni-corrected Wilcoxon test for pairwise comparisons (all tests performed using the Rstatix package). Statistically significant effects have been accepted at the 95% level.

For each laboratory and product, an individual ratio of α-amylase activity at 37 °C to 20 °C was calculated, and the mean of these ratios across all laboratories was determined for each product. The 95% confidence interval for this mean ratio was computed using the t-distribution. Normal distribution and homogeneity of variances have been confirmed for this dataset, hence one-way ANOVA was used to investigate whether the ratios obtained for each product were significantly different.

For a thorough understanding of the method’s reliability, precision, and transferability across different laboratory settings three complementary metrics have been used: Z-scores based on bias scores for a standardized evaluation of systematic errors, repeatability and reproducibility.

Z-scores were calculated to standardize the comparison of bias scores across laboratories and products enabling to assess the overall agreement between individual laboratory results and the mean for each product. For each product, bias scores were first calculated for each laboratory using the mean of all laboratories as the reference value and then converted to z-scores:

x is the individual laboratory result, X is the mean of all laboratories, and SD is the standard deviation. Z-scores interpretation followed standard criteria with |z|≤ 2 as satisfactory, 2 <|z|< 3 as questionable, and |z|≥ 3 as potentially unsatisfactory.

Repeatability (measured as intralaboratory coefficient of variation, CV_r), which quantifies method precision within each laboratory, reflecting consistency under identical conditions, was calculated as the root mean square of the individual laboratory’s CVs:

CV_r is the coefficient of variation under repeatability conditions (intralaboratory); (i) indexes each laboratory, ({CV}_{i}) is the coefficient of variation for laboratory (i); L is the number of participating laboratories.

Reproducibility (measured as coefficient of variation, CV_R), a measurement of method’s consistency across different laboratories indicates its robustness to varying environments and operators, was calculated for each tested product as:

CV_R is the coefficient of variation under reproducibility conditions (interlaboratory); SD and X correspond to the standard deviation and mean values calculated from interlaboratory data.

Coefficients of variation below 30%^15,16 are frequently considered to be indicators of small intra- and interlaboratory variability. In some cases, critical thresholds for repeatability (intralaboratory CV) are set at 20%³⁴.