Comparative effects of superheated steam and hot air roasting on the in vitro phospholipase A2 inhibitory activity and polyphenol composition of cocoa beans

TL;DR

This study compares how two roasting methods affect cocoa beans' anti-inflammatory properties and polyphenol content.

Contribution

The study reveals that superheated steam roasting enhances anti-inflammatory activity and specific polyphenol levels more effectively than hot air roasting.

Findings

Superheated steam roasting at 175 °C reduced PLA2 inhibitory IC50 by 50% compared to hot air roasting.

SHS-roasted beans had higher proanthocyanidin tetramer and pentamer levels at 175 °C for 8 min.

Hot air roasting increased total phenolic content but reduced specific PAC levels.

Abstract

Cocoa, derived from Theobroma cacao, is a popular food ingredient used to produce chocolate. Cocoa is rich in polyphenols and human and laboratory model studies have indicated that cocoa and chocolate can mitigate chronic inflammatory conditions. Cocoa is subjected to post‐harvest processing operations including fermentation and roasting prior to use. These steps, which are essential for flavor development, have been shown to lead to changes in the content and profile of polyphenols in cocoa. Less is known about the effect of these operations on the bioactivity of cocoa. The present study compared the impact of using superheated steam (SHS) versus hot air (HA) as a roasting medium for cocoa beans on in vitro phospholipase A2 (PLA2) inhibitory activity, a surrogate for anti‐inflammatory activity, and polyphenol chemistry.

Beans were roasted at 150 °C (10, 20, 30 min), 175 °C (8, 15 min) and 200 °C (7 min) using SHS or HA. The median inhibitory concentration (IC50) against PLA2 of extracts from SHS beans at 175 °C was 50% lower than comparable HA beans. Although total phenolic content in HA beans for all roasting treatments was 20% to 33% higher than SHS beans, the levels of proanthocyanidin (PAC) tetramer and pentamer in beans roasted at 175 °C for 8 min were 23% and 36% higher, respectively, than those roasted using HA.

The findings of this study suggest that SHS roasting could be used to produce beans with greater anti‐inflammatory activity and higher levels of key PACs in comparison with HA. © 2026 The Author(s). Journal of the Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Figures

In vitro phospholipase A2 (PLA2) inhibitory activity of superheated steam (SHS) or hot air (HA) roasted cocoa beans. Polyphenol‐enriched extracts were prepared from cocoa beans roasted at 150 °C for (A) 10, (B) 20, or (C) 30 min; at 175 °C for (D) 8, or (E) 15 min; or at 200 °C for (F) 7 min. Black squares represent HA and white squares represent SHS. Data represent the means of three determinations. Error bars represent the standard deviations.

Effect of superheated steam (SHS) or hot air (HA) roasting on polyphenol content of cocoa beans. Folin–Ciocâlteu values were used as a surrogate for total phenolic content. Catechin, epicatechin (EC), PAC dimer, PAC trimer, PAC tetramer, and PAC pentamer were quantified by liquid chromatography–mass spectrometry (LC–MS). Black bars represent HA and grey bars represent SHS. Data represent the means of three determinations. Error bars represent the standard deviations. Within each time and temperature combination, values with different superscript letters are significantly different from one‐way analysis of variance (ANOVA) with post‐hoc Tukey's test (P < 0.05).

Tables

Table 1. Processing temperatures and times in hot air and superheated steam ovens

Temperature (°C)	Target process time, without come‐up time (CUT) correction (min)	Corrected hold time (min) a
150	10	7.5
150	20	15
150	30	25
175	8	4.5
175	15	11.5
200	7	3.5

Table 2. In vitro phospholipase A2 (PLA2) median inhibitory concentrations (IC50) of superheated steam (SHS) or hot air (HA) roasted cocoa beans a

Temp (°C)	Time (min)	IC50 (μg mL−1)		Area under the curve (AUC)0→100μg mL−1 (% ×μg mL−1)
		SHS	HA	SHS	HA
150
	10	30.8 ± 4.8a	19.7 ± 4.2a	4486 ± 339.9a	3870 ± 145.1a
	20	34.9 ± 2.2a	34.8 ± 7.1a	4748 ± 141.0a	4739 ± 548.8a
	30	51.6 ± 8.0a	63.4 ± 9.4a	5685 ± 305.6a	5908 ± 120a
175
	8	8.6 ± 4.8l	27.7 ± 5.7m	3199 ± 129.0l	4930 ± 81.1m
	15	30.8 ± 2.2m	67.6 ± 18.5m	4248 ± 278n	5964 ± 497.0o
200
	7	19.8 ± 6.4x	23.8 ± 5.2x	3413 ± 478.2x	4171 ± 308.8x

Full text

INTRODUCTION

Cocoa, a food ingredient derived from the seeds of Theobroma cacao, is a rich source of phenolic compounds, which may be present at up to 15% on a dry weight basis.1 Monomeric flavan‐3‐ols, which are the most abundant polyphenols found in cocoa, can polymerize to form oligomers and polymers called B‐type proanthocyanidins (PACs).2 The phenolic content in cocoa beans depends on plant genetics, geographical origin, abiotic and biotic stress, degree of ripeness at harvest, agronomic factors, and post‐harvest storage and processing conditions.3 Consumption of cocoa and chocolate has been associated with mitigation of obesity and chronic inflammation.4

Positive cardiometabolic effects have been demonstrated by studies in animal models.5 It has been reported that the blood glucose and fructosamine levels of diabetic obese mice decreased significantly following a 3 week consumption of 0.5% or 1% cocoa liquor PACs.6 Yamashita et al.7 found that high‐fat (HF) diet‐fed C57BL/6 mice treated with cocoa liquor PAC extract for 13 weeks had less glucose intolerance, less hyperglycemia, and less white adipose tissue mass than controls. Dorenkott et al.8 compared the effects of 12‐week dietary supplementation of cocoa flavan‐3‐ol monomers, oligomers, and polymers in HF‐fed mice and found that dietary PAC oligomer supplementation reduced weight gain and increased glucose and insulin tolerance more effectively than the monomer and polymer fractions in comparison with HF‐fed controls.

It was previously reported that cocoa extracts and cocoa PACs can inhibit pancreatic lipase, amylase, and phospholipase A_2_ (PLA_2_), enzymes that play key roles in macronutrient digestibility and inflammation.9 The inhibitory potency of individual PACs increased proportionally with their degree of polymerization (DP). Cocoa extracts enriched in polymeric PACs also inhibited interleukin (IL)‐8 production by tumor necrosis factor (TNF)α‐stimulated HT‐29 colon cancer cells more effectively than fractions enriched in monomeric flavan‐3‐ols or oligomeric PACs.10

During post‐harvest processing, cocoa beans are typically roasted in a hot air (HA) oven at 120–250 °C for 10–50 min depending on the processor, application, and desired sensory characteristics. The purpose of roasting is to develop the color and flavor of the cocoa beans after fermentation.11 Roasting can lead to polyphenol loss due to thermal and oxidative degradation.12 Bordiga et al. found that the Folin–Ciocâlteu (FC) value, a surrogate for total phenolic content, of chocolate prepared with cocoa beans from different geographical origins decreased by 10% to 30% after a ‘typical’ roasting process but they did not specify the roasting conditions.12 Stanley et al. observed a decrease in the FC value and in the levels of PACs with DP below 3 but an increase of more than 200% in the levels PAC hexamer and heptamer after roasting at 170 and 190 °C.13 Moreover, there was an observed increase in the in vitro pancreatic lipase inhibitory activity with roasting time at 170 °C, even though FC values decreased. These results indicate that polyphenol composition, rather than FC, may be a better predictor of some bioactivity. Consistent with this hypothesis, another study found that the mean degree of PAC polymerization, but not FC values, correlated well with in vitro α‐glucosidase enzyme inhibitory activity.14 A correlation between the mean DP and the inhibition of key digestive enzymes such as PLA_2_, α‐amylase, and pancreatic lipase by cocoa PACs has also been reported.9

Superheated steam (SHS) as a roasting and drying medium for foods has recently attracted attention due to its high heat transfer coefficient, generation of a low oxygen environment, and possibility of energy reuse.15 It is obtained by adding sensible heat to saturated steam at a given pressure so that its temperature increases beyond the saturation point.16 At atmospheric pressure, SHS is created by adding more heat to saturated steam so that its temperature rises beyond 100 °C. A study by Yodkaew et al. compared the quality of fatty acids in rice roasted using HA and SHS in a fluidized bed dryer at 170, 190, and 210 °C.17 The SHS‐roasted rice (210 °C) had reduced lipid oxidation and lower levels of free fatty acids compared to HA‐roasted rice. It has been reported that SHS treatment at 175 °C for 15 min increased the FC value and radical scavenging activity of Baccaurea pubera, a red‐colored fruit found in Borneo, in comparison with fresh fruit.18 It was also shown that SHS roasting of Robusta coffee beans at 190–250 °C resulted in greater moisture loss, lower lightness, increased sugar levels, and reduced acetic acid levels than HA roasting.19

To date, a limited number of studies have focused on the effect of SHS roasting on cocoa bean chemistry. Zzaman et al.20 compared the effect of SHS roasting with HA roasting at 150–250 °C for 10–50 min and found that FC, total flavonoid content, and antioxidant capacity were higher in the SHS roasted beans. Using a response surface study design, the same group found that roasting at 192 °C and 10 min resulted in the highest FC value, flavonoid content, and antioxidant capacity.21 These studies measured general markers of phenolic content but they did not examine changes in the levels of individual polyphenols nor did they assess changes in any markers of potential biological activity.

The overall objective of this study was to compare HA and SHS as methods for roasting in terms of in vitro phospholipase PLA_2_ inhibitory activity, FC values, and the catechin, epicatechin (EC), and PAC dimer, trimer, tetramer, and pentamer content of the resulting cocoa. The hypothesis was that the SHS roasted cocoa beans would have a greater in vitro PLA_2_ enzyme inhibitory activity and higher concentration of PAC oligomers than HA roasted beans.

MATERIALS AND METHODS

Chemicals and reagents

(−)‐Epicatechin, (+)‐catechin, and gallic acid were purchased from Millipore‐Sigma (St Louis, MO, USA) and their purity was greater than 98%. B‐type PAC dimers, trimers, tetramers, and pentamers were purchased from Planta Analytica (New Milford, CT, USA), with purities greater than 95% pure. All solvents were purchased from Fisher Scientific (Pittsburgh, PA, USA) and were high‐performance liquid chromatography (HPLC) grade. Dimethyl sulfoxide (DMSO) (99.9% pure) was purchased from Millipore‐Sigma. All other reagents were of the highest commercially available grade.

Roasting protocol

Raw, unfermented cocoa beans were obtained from Cargill Inc. (Wayzata, MN, USA), fermented using simulated pulp media in a thoroughly agitated bench‐scale system for 168 h (fermentation temperature increased 3.5 °C/24 h to a final temperature of 45 °C), and dried as previously described.14 Intact cocoa beans weighing 0.7–1 g were chosen for the experiments. Hot air roasting experiments were carried out in a Fisher Scientific Isotemp Oven Model 630F (Thermo Fisher Scientific, Waltham, MA, USA) and the SHS experiments were carried out in a Sharp SHS countertop oven model SSC0586DS (Sharp Corporation, Osaka, Japan). The heating rate in both ovens was determined by inserting J‐type thermocouples at the center of individual cocoa beans and securing them with thermally conductive glue (Omegabond 200 two‐part epoxy with high thermal conductivity; Omega Engineering Inc., Norwalk, CT, USA) (Supporting Information, Fig. S1), and connecting the thermocouples to a CR3000 micrologger (Campbell Scientific, Logan, UT, USA). To compensate for the differing internal oven volumes, the temperature settings on each oven were selected so that the temperature at the center of a cocoa bean was the same in both ovens during roasting. A preliminary study was conducted to determine the upper limits of the temperature and time combinations beyond which the cocoa beans started to burn. Roasting experiments were carried out at 150 °C (10, 20, 30 min), 175 °C (8, 15 min) and 200 °C (7 min) using SHS or HA (Table 1). Roasted cocoa beans were winnowed by hand, ground in a coffee grinder (Model CG715S, Changsha Shardor, Changsha, China) with liquid nitrogen to produce a powder with particle size of 500–700 μm, and stored under refrigeration at 4 °C until further treatment. Roasting experiments were performed in triplicate (n = 3).

Come‐up time measurement

Come‐up times (CUTs) for both ovens used for roasting experiments were measured, for all roasting temperatures, using a J‐type thermocouple inserted into individual cocoa beans as described above. Corrected processing times were calculated by subtracting 50% of the come‐up time from the target processing time; they were designated as corrected hold times.22

Polyphenol extraction

Ground cocoa was defatted four times with hexane (4:1, solvent:cocoa, v:w). Polyphenols were extracted three times from the defatted cocoa powder using 80% aqueous acetone containing 0.1% acetic acid (4:1, solvent:cocoa, v:w). The acetone in the extracts was removed under vacuum and the water was removed by freeze‐drying (VirTis Genesis 35 XL freeze dryer, ATS Scientific, Warminster, PA, USA). Dried extracts were dissolved in DMSO at a final concentration of 50 mg mL−1 and stored at −80 °C.

PLA2

inhibition

Inhibition of PLA_2_ by cocoa extracts was examined using a commercially available fluorometric method (EnzChek PLA_2_ Assay kit, Thermo Fisher Scientific). Buffered PLA_2_ solution (1 U/mL, pH 8.9, 50 μL) and cocoa extracts (0–100 μg/mL, 50 μL) were combined in a 96‐well plate. A fluorogenic PLA_2_ substrate (red/green BODIPY PC‐A2, 1.67 μM, 30 μL) was added to each well. After incubation at room temperature for 10 min in dark conditions, fluorescence was measured at an excitation wavelength (λ ex) of 485 nm and at an emission wavelength (λ em) of 538 nm (Fluoroskan Ascent FL, Thermo Fisher Scientific). The fluorescence of the reaction in the presence of each extract concentration was normalized to the fluorescence of the reaction containing a vehicle control and the relative PLA_2_ was expressed as ‘%control activity’. The vehicle control was normalized to 100%. Assays were performed in duplicate. The median inhibitory concentration (IC_50_) of each cocoa extract was calculated using regression analysis of the dose–response curves. To quantify the overall inhibitory activity of each HA or SHS extract, the area under the dose response curves from 0–100 μg mL^−1^ was calculated using the trapezoid rule.

Folin–Ciocâlteu assay

Folin–Ciocâlteu values were measured as a proxy for total phenolic content using a modification of an existing colorimetric method.23 Briefly, cocoa extracts (50 mg mL^−1^ in DMSO) were diluted 1:100 with deionized water. Diluted cocoa extract (40 μL) was combined with 1560 μL of deionized water, and 100 μL of FC reagent (Sigma Aldrich, St Louis, MO, USA). This mixture was vortexed and incubated at room temperature for 5 min in dark conditions. The reaction was stopped by the addition of 300 μL of sodium carbonate (200 g L^−1^) to the reaction mixture. The solution was incubated at 37 °C for 30 min. Absorbance was measured at λ = 765 nm using a Thermo Scientific Multiskan Go ultraviolet‐visible (UV‐visible) spectrophotometer (Thermo Fisher Scientific). The FC value was quantified by comparison with a standard curve of gallic acid and expressed as mg gallic acid equivalent (GAE) g^−1^ fat‐free cocoa. Analyses were performed in triplicate.

Reversed‐phase liquid chromatography mass spectrometry analysis

Cocoa extracts (50 mg mL^−1^ in DMSO) were diluted to 10 mg mL^−1^ with water and filtered through 0.45 μm polytetrafluoroethylene (PTFE) filters. Liquid chromatography–mass spectrometry (LC–MS) analysis was conducted using an Agilent 1290 Infinity II LC system coupled to an Agilent 6460 Triple Quad MS (Agilent Technologies, Santa Clara, CA, USA). The analytes were separated using an Acquity ultra‐performance liquid chromatography high strength silica (UPLC HSS) (Waters Corp, Milford, MA, USA) column (2.1 mm × 100 mm, 1.8 μm pore). The mobile phase solvents were 0.1% aqueous formic acid (A) and acetonitrile containing 0.1% formic acid (B). The gradient program involved an initial isocratic phase at 100% A. From 0.5 to 6 min, the proportion of B was increased to 31.6% and from 6 to 7.5 min to 90%. The mobile phase was held at 90% B for 3.5 min. The mobile phase was then switched to 100% A and held for 1.5 min. The flow rate was 0.4 mL min^−1^ and the column temperature was 35 °C. Multi‐reaction monitoring (MRM) parameters were optimized for each compound (Supporting Information, Table S1) using the Agilent optimizer system with a unique fragmentor (between 135–350 V) and collision energy (between 13–41 V) for each compound. Negative ionization mode was used for all compounds. Compounds were quantified by comparison with standard curves of authentic standards (1.5–10 μg mL^−1^).

Statistical analysis

The IC_50_ was calculated by regression analysis of the PLA_2_ assay dose response curves. Overall inhibitory activity was examined by calculating the area under the curve of the dose–response curves from 0–100 μg mL^−1^ using the trapezoid rule. Both analyses were performed using GraphPad Prism (GraphPad Software, Boston, MA, USA). To compare the effect of HA or SHS roasting on cocoa bean polyphenol chemistry and PLA_2_ inhibitory activity, a one‐way analysis of variance (ANOVA) with Tukey's post‐hoc test for measurements with equal variances was performed separately for each temperature using Minitab (Minitab Inc., State College, PA, USA). For values with unequal variances, a one‐way ANOVA with Welch's correction and the Brown–Forsythe test was performed using GraphPad Prism.

RESULTS AND DISCUSSION

Come‐up time measurement

The heating curves for HA and SHS at each roasting temperature were very similar (Supporting Information, Fig. S1). The CUT for the 150 °C, 175 °C, and 200 °C were determined as 10 min, 7 min, and 7 min, respectively. As the z values for flavor development and polyphenol content changes during the roasting process were not known, 50% of the CUT was subtracted from the target processing time to arrive at the ‘corrected hold time’ (Table 1).

In vitro phospholipase A2

inhibition

The in vitro PLA_2_ inhibitory efficacy of cocoa extracts was examined as a marker of anti‐inflammatory bioactivity. At 150 °C, the dose–response curve for extracts from cocoas roasted for 10 min shows that HA‐roasted beans had greater inhibitory activity than SHS‐roasted beans at 10 and 25 μg mL^−1^ of defatted polyphenol extract (Fig. 1(A)). These differences were lost in beans roasted for 20 or 30 min (Fig. 1(B),(C)). Differences were observed in the PLA_2_ inhibitory activity of extracts from SHS and HA beans roasted for 8 and 15 min at 175 °C, with the extracts from SHS roasted bean having greater inhibitory activity than those of HA at concentrations greater than 25 μg mL^−1^ (Fig. 1(D),(E)). Small differences were observed in the inhibitory activity of extracts prepared from beans roasted at 200 °C at 50 and 100 μg mL^−1^, extracts from SHS‐roasted beans had greater inhibitory activity than those from HA‐roasted beans (Fig. 1(F)). The IC_50_ values of the extracts from beans that had been roasted with SHS at 175 °C, 8 min were significantly lower than those from beans roasted with HA (Table 2). The areas under the curve (AUCs) of the dose response curves for extracts from cocoa beans roasted with SHS at 175 °C for 8 or 15 min were approximately 35% or 29% smaller, respectively, than those of extracts from cocoa beans roasted with HA (Table 2). No such differences were observed for extracts from beans roasted at 150 °C or 200 °C. These results are in line with a report by Stanley et al., who observed an increase in the in vitro pancreatic lipase inhibitory potency of extracts prepared from beans HA roasted for 20 min at 170 °C but not at lower temperatures or shorter roasting times.13

In vitro phospholipase A2 (PLA2) inhibitory activity of superheated steam (SHS) or hot air (HA) roasted cocoa beans. Polyphenol‐enriched extracts were prepared from cocoa beans roasted at 150 °C for (A) 10, (B) 20, or (C) 30 min; at 175 °C for (D) 8, or (E) 15 min; or at 200 °C for (F) 7 min. Black squares represent HA and white squares represent SHS. Data represent the means of three determinations. Error bars represent the standard deviations.

Table 2: In vitro phospholipase A2 (PLA2) median inhibitory concentrations (IC50) of superheated steam (SHS) or hot air (HA) roasted cocoa beans a

Folin–Ciocâlteu values

The FC values varied significantly with roasting time and medium (HA or SHS) at each temperature (Fig. 2). Extracts from HA roasted beans at 150 °C for 10 and 20 min exhibited significantly higher FC values than extracts from SHS roasted beans. The difference in FC values of extracts from HA and SHS decreased with increasing roasting time at 150 °C. At 175 °C, the FC values of the extracts from SHS and HA roasted beans were not changed significantly after 8 min of roasting, but the FC value in the extract from SHS roasted beans was 26% lower than that from HA roasted beans. At 200 °C, the FC of the extract from the SHS roasted beans was 17% lower than that from HA roasted beans. These results differ from those of Zzaman and Al‐din Sifat who reported that roasting cocoa beans for 16 min at 184 °C using SHS had 36% higher FC values than those roasted with HA.24 Folin–Ciocâlteu values have some limitations due to interference caused by the reaction of the FC reagent with non‐phenolic reducing compounds, which can include Maillard browning products.25 For example, the more aggressive roasting conditions employed by Zzaman and Al‐din Sifat could have resulted in the production of greater amounts of Maillard browning products, leading to higher FC values.24 Phenolic compounds with different chain lengths can also exhibit different reactivity with the FC reagent, further confounding the measurements.26 Further studies using full‐factorial designs may be necessary to gain a complete picture of the effects of SHS or HA on total phenolic content in cocoa.

Effect of superheated steam (SHS) or hot air (HA) roasting on polyphenol content of cocoa beans. Folin–Ciocâlteu values were used as a surrogate for total phenolic content. Catechin, epicatechin (EC), PAC dimer, PAC trimer, PAC tetramer, and PAC pentamer were quantified by liquid chromatography–mass spectrometry (LC–MS). Black bars represent HA and grey bars represent SHS. Data represent the means of three determinations. Error bars represent the standard deviations. Within each time and temperature combination, values with different superscript letters are significantly different from one‐way analysis of variance (ANOVA) with post‐hoc Tukey's test (P < 0.05).

Individual polyphenol levels

The concentrations of catechin, EC, and PAC dimer, trimer, tetramer, and pentamer in a subset of samples representing different levels of PLA_2_ inhibitory activity were quantified by LC–MS (Fig. 2). Cocoa beans roasted under HA at 150 °C for 30 min had significantly lower concentrations of catechin than beans roasted using SHS, whereas EC concentrations did not differ between the two roasting media (Fig. 2). Cocoa beans roasted under HA for 30 min at 150 °C had significantly higher concentrations of PAC dimer, trimer, tetramer, and pentamer than beans roasted with SHS for the same amount of time at 150 °C (Fig. 2). At 175 °C, no significant differences in catechin, EC, or PAC dimer and trimer concentrations were observed between roasting media after 8 or 15 min roasting time (Fig. 2). The SHS‐roasted beans did, however, have significantly higher concentrations of PAC tetramer and pentamer than HA‐roasted beans after 8 min, but not 15 min, roasting time (Fig. 2). The concentration of individual flavan‐3‐ols and PACs were not significantly different between the media when beans were roasted at 200 °C for 7 min (Fig. 2).

The results from the PLA_2_ enzyme inhibition assay cannot be completely explained by the individual PACs measured by direct quantification using LCMS. It was expected that polyphenol extracts that showed a higher in vitro PLA_2_ inhibitory activity would have a higher concentration of PAC oligomers. Thus, polyphenol extracts obtained from cocoa beans roasted using SHS at 175 °C were expected to have higher concentrations of PAC oligomers than those obtained from the corresponding HA roasted beans. Higher PAC tetramer and pentamer levels were found in the polyphenol extracts from beans roasted at 175 °C for 8 min using SHS compared with those roasted with HA but no differences were found in beans roasted at 175 °C for 15 min even though differences were observed in PLA_2_ inhibitory activity. One limitation of the current study is that PACs with a DP below 5 were not quantified due to a lack of commercially available analytical standards. This information could be the key to explaining the differences between the two roasting treatments. For example, it is possible that beans roasted with SHS for 15 min at 175 °C have higher levels of PACs with DP greater than 5 compared with beans roasted using HA for an equivalent amount of time at the same temperature. These differences would explain the greater PLA_2_ inhibitory activity of the SHS beans. A higher proportion of PAC oligomers (DP ~6 or 7) could be produced due to the accelerated oxidation of monomeric flavan‐3‐ols within a temperature range of 170–180 °C. In a HA roasting environment, this oxidation could also lead to the formation of large PAC polymers (DP > 10), which have a tendency to bind to proteins or polysaccharides in the cocoa bean, becoming less available for extraction from the food matrix.27 Modeling studies of PLA_2_ enzyme inhibition by PACs had found that linkages between the PAC molecule and the functional groups on the tunnel‐like structure in which the active site of the PLA_2_ enzyme lies is the cause for the inhibition of PLA_2_ enzyme.28 The increased affinity of larger PAC polymers to bind to other components in the cocoa bean like proteins or polysaccharides can make these large PAC polymers unavailable for binding to the enzyme molecule, thus leading to a lower inhibition potential. The reduced oxygen during SHS roasting could prevent the formation of large PAC molecules and the retention of a higher concentration of PAC oligomers, which would be free to bind to the enzyme and inhibit it. This could explain the differences in the in vitro PLA_2_ enzyme inhibition potential of the extracts prepared from HA‐ and SHS‐roasted cocoa beans at 175 °C.

CONCLUSION

Superheated steam was studied as an alternative to HA as a roasting medium for cocoa beans. Under specific roasting conditions, SHS roasting at 175 °C resulted in cocoa extracts with significantly greater in vitro PLA_2_ inhibitory activity than those of HA roasted beans. The FC values of HA roasted beans were generally higher than those of SHS roasted beans but the SHS roasted beans with the greatest PLA_2_ inhibitory activity contained higher levels of PAC tetramer and pentamer than comparable HA roasted beans. These differences may explain the increased in vitro PLA_2_ inhibitory activity. Further studies are needed to elucidate the underlying chemical mechanisms that explain the differences between HA and SHS as media for roasting cocoa beans and to extend the results of the in vitro PLA_2_ inhibition assay to an in vivo model.

CONFLICT OF INTEREST

The authors have no conflicts of interest to declare.

Supporting information

Figure S1. Time‐dependent changes in internal temperature of cocoa beans roasted with superheated steam or hot air. (A) J‐type thermocouple was inserted into the center of cocoa bean and secured with thermally conductive glue. Internal temperatures were determined in beans roasted at (B) 150°C, (C) 175°C, or (D) 200°C using superheated steam or hot air. Data represent the mean of two determinations.

Table S1. MRM transitions, fragmentor voltage and collision energy values used for individual polyphenols measured using LCMS.