Physicochemical characteristics, ATR-FTIR molecular interactions and in vitro starch and protein digestion of thermally-treated whole pulse flours

Food Research International 105 (2018) 371–383

Contents lists available at ScienceDirect

Food Research International journal homepage: www.elsevier.com/locate/foodres

Physicochemical characteristics, ATR-FTIR molecular interactions and in vitro starch and protein digestion of thermally-treated whole pulse flours

T

Chávez-Murillo Carolina Estefaníab, Veyna-Torres Jorge Ivanb, Cavazos-Tamez Luisa Maríaa, ⁎ de la Rosa-Millán Juliána, , Serna-Saldívar Sergio Othona a

Tecnologico de Monterrey, Escuela de Ingeniería y Ciencias, Av. Eugenio Garza Sada 2501 Sur, CP 64849 Monterrey, NL, Mexico Instituto Politécnico Nacional, Unidad Profesional Interdisciplinaria de Ingeniería campus Zacatecas (UPIIZ-IPN), Blvd. del Bote S/N Cerro del Gato Ejido La Escondida, Col. Ciudad Administrativa, C.P. 98160, Zacatecas, Zac, Mexico b

A R T I C L E I N F O

A B S T R A C T

Keywords: Pulse flours In vitro starch digestion Annealing Heat moisture treatment ATR-FTIR

Hydrothermal treatments, annealing (ANN) and heat moisture treatment (HMT) were applied to four whole pulse flours (black bean, broad bean, chickpea and lentil) with the aim to increase their slow digestible (SDS) and resistant starch (RS) fractions. In order to assess differences in their molecular interactions, they were analyzed and compared by ATR-FTIR before and after in vitro digestion. Both hydrothermal treatments promoted changes on starch granular architecture, being reflected on their thermal and pasting properties, that where positively correlated with their amylose and protein contents (R = 0.96, P < 0.01). Overall, the proposed hydrothermal treatments increased their SDS and RS fractions, but they had different effect on their in vitro protein digestion. The ATR-FTIR analysis of cooked flours before and after digestion showed that thermal treatments promoted new physical interactions at molecular scale between starch and proteins, that were correlated with the amount of RS fraction. The outcomes of this study could help to understand the slow digestion properties and possible interactions of the flour components in these four pulses.

1. Introduction In recent years, pulses have received great attention mainly because they present a large array of nutraceuticals (Azarpazhooh & Boye, 2012; Du, Jiang, Yu, & Jane, 2014; Ma, Boye, Azarnia, & Simpson, 2016). From the nutritional point of view, they contain a high proportion of complex carbohydrates, being starch the main fraction. The starches associated to pulses have been related with higher thermal stability during pasting, as well as significant amounts of soluble and insoluble fiber that delay its starch hydrolysis, resulting in low in vivo and in vitro digestion rate by amylolytic enzymes (Brummer, Kaviani, & Tosh, 2015; Chung et al., 2008; Jenkins et al., 1982; Maaran, Hoover, Vamadevan, Waduge, & Liu, 2016). The other nutritionally relevant macromolecule of pulses is their proteins, which presents good essential amino acids balance (Divekar et al., 2016; Ghumman, Kaur, Singh, & Singh, 2016). In Mexico there is a great diversity of pulses cultivars, that is cultivated for farmers' self-consumption and represents an important food staple with different energy value, depending on their protein and starch bioavailability. In addition, the natural occurrence of phenolic compounds has been a recurrent subject in their study as functional ingredients (Megías, Cortés-Giraldo, Alaiz, Vioque, & Girón-Calle, 2016;

⁎

Moussou et al., 2016). From the technological perspective, the functional and digestion properties of starches and proteins can be modified with the use of chemical or enzymatic modifications, resulting in ingredients with improved characteristics, that are often related with their functional and digestion properties (Ahn, Kim, & Ng, 2005; Romano Giosafatto, Masi, & Mariniello, 2015). Nevertheless, one of the current trends in food processing is to avoid the use of chemical reagents that may negatively affect the consumer health and cause environmental issues due to inadequate residue disposal. For this reason, the use of thermal treatments like annealing (ANN) and heat moisture treatment (HMT) are some of the alternatives to modify some of the physicochemical as well as the starch and protein digestion characteristics in whole flours (Tester & Morrison, 1990; Tosh & Yada, 2010). Previous studies have shown that pulses seed matrix composition and their molecular arrangement delay starch hydrolysis; thus, potentially decreasing the predicted glycemic index (pGI) which may have an impact on the protein availability of the cooked flours (Torres, Rutherford, Muñoz, Peters, & Montoya, 2016). However, little attention has been paid to the possible molecular interactions that may occur among the different seed components, and their effect on viscosity and in vitro digestion properties. Such molecular interactions could be

Corresponding author. E-mail address: [email protected] (J. de la Rosa-Millán).

https://doi.org/10.1016/j.foodres.2017.11.029 Received 15 December 2016; Received in revised form 9 November 2017; Accepted 19 November 2017 Available online 21 November 2017 0963-9969/ © 2017 Elsevier Ltd. All rights reserved.

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

2.5. Color characteristics

qualitatively analyzed by the non-invasive ATR-FTIR, which has proven to be useful in quantify the changes on secondary structures of proteins and crystalline starch conformations (Cai & Singh, 2004; Carbonaro, Maselli, Dore, & Nucara, 2008; Manning, 2005). The aims of this investigation were to characterize the physicochemical and digestion properties of native and thermally treated whole pulse flours and to analyze their molecular interactions as during its typical consumption as food, before and after in vitro digestion on both the protein and carbohydrate fractions by ATR-FITR.

Color characteristics of native and thermally treated flours were measured on the CIE 1976 color scale using a Minolta CM-600, Chroma Meter (Konica Minolta Co., Osaka, Japan). For each sample, 5 readings were taken from powder samples and recorded as lightness (L ∗), redness-greenness (a∗) and yellowness-blueness (b ∗) values. To establish color differences of the different whole pulse flour treatments and to observe the effect of thermal treatments, the color index value (Δ E) was calculated with the following equation.

2. Materials and methods

ΔE = ((ΔL)2 + (Δa)2 + (Δb)2)1/2

2.1. Whole flour preparation

(1)

2.6. Colorimetric total phenolic determination

Food-grade pulse seeds from black bean (Phaseolus vulgaris), broad bean (Vicia faba), chickpea (Cicer arietinum) and lentil (Lens culinaris) were acquired with a local seed distributor in Zacatecas, México. The seeds were ground to produce whole flours in a Wiley mill (Arthur Thomas, Philadelphia, PA) equipped with a 2 mm diameter screen. All flours were sieved to pass the US 80 mesh. The overs or coarse particles were further milled to pass this sieve. Afterwards, whole flours were stored in sealed plastic bags at − 20 °C until further processing.

Determination of total phenolic content was performed using the Folin–Ciocalteu method as described by Singleton, Orthofer, LamuelaRaventós, and Lester (1999). The samples and gallic acid standard curve (0.016–0.1 mg gallic acid mL− 1; y = 3.0161x − 0.0216; R2 = 0.9915) were read at 725 nm. Results were expressed as mg of gallic acid equivalents per gram of flour sample (d.s.b.). Absorbance values of colorimetric assays were read in a microplate reader Synergy HT, Biotek (Winooski, USA); with Gen5™2.0 data analysis software spectrophotometer.

2.2. Hydrothermal treatments

2.7. Rapid viscosity analysis

For ANN, 100 g of each pulse flour (d.s.b.) were weighed into hermetic glass containers, blended with 400 mL of distilled water and homogenized during 30 min. The containers then were tight sealed and kept for 24 h at 10 °C. Afterwards the containers were incubated at 65 °C for 24 h in a convection oven. After the incubation time, the containers were opened and the flour samples vacuum dried at 40 °C for 24 h, followed by milling to pass through a US 80 mesh screen and kept into hermetic containers until analysis. For HMT treatment, the flours (100 g) were weighted into a plastic bag, where the appropriate amount of distilled water was added until reach a 30% moisture content. The bags were tight closed and their contents were homogenized until evenly wetted without the presence of dry lumps. Afterwards, the bags were stored at 4 °C for 4 h to stabilize the moisture content within the flour; then, their contents were poured into hermetic glass containers which were immediately incubated at 120 °C for 24 h in a forced air convection oven. After the time, the containers were cooled down to 25 °C and opened, its content spread on an aluminum tray followed by 24 h of a vacuum dry process at 40 °C, thereafter, the dried flours were grounded until pass the US 80 mesh screen and stored in hermetic containers in a tempered room (≈25 °C) until further use. The processing conditions (both for ANN and HMT) were based on previous pilot plant trials and were selected due to pilot plant process efficiency.

The pasting profiles of flours were measured with a Rapid Visco Analyzer (RVA Model 1170, Newport Scientific, Warriewood, NSW, Australia) using a constant 25.5 g suspension with 12% solids. The heating profile was: hold at 50 °C for 2 min, heat to 90 °C at 15 °C/min (heating cycle), hold at 90 °C for 4 min (high temperature hold), cool to 50 °C at −15 °C/min (cooling cycle) and hold at 30 °C for 4 min (low temperature hold). 2.8. Differential scanning calorimetry features of whole pulse flours For this analysis, 2 mg (based on total starch content) of whole pulse flours were placed in semi-hermetic anodized aluminum capsules (Perkin Elmer, B02190062, US), hydrated with the appropriate amount of distilled water (3 volumes, based on total sample weight) and containers carefully sealed. Once hydrated, samples were kept for 30 min at room temperature (≈25 °C) and subsequently heated from 30 to 120 °C at a temperature rate of 10 °C/min in a Diamond DSC (Perkin Elmer, NorTcolk, VA, USA) apparatus which was calibrated with an Indium reference cell before experimental measurements were performed. An empty capsule was used as reference for each determination. The transition parameters, onset (To), peak (Tp), and conclusion (Tc) temperatures of gelatinization as well as the endothermic enthalpy (ΔH) were calculated using the Pyris manager software (Perkin Elmer, Nortfolk, VA, USA).

2.3. Physicochemical characteristics Native and heat-treated pulse flours were analyzed for moisture, protein, fat and ash according to approved AACCI methods 44-01.01, 46-13.01, 30-20.01 and 08-01.01, respectively (AACCI, 2000) and their carbohydrates calculated by difference. The total and damaged starch (TS and DS, respectively), as well as the total amylose and total dietary fiber contents (TDF) were determined with commercial available kits KTSTA, K-SDAM, K-AMYL and K-TDFR respectively following the procedures described by Megazyme (Wicklow, Ireland).

2.9. Starch digestion fractions The in vitro starch digestibility of whole pulse flours was determined according the Englyst, Kingman, and Cummings (1992) protocol with slight modifications. Pulse flours (400 mg) were hydrated with 10 mL of deionized water and heated in a boiling water bath for 20 min with vortexing every 5 min. The tubes were cooled at 37 °C and 8 mL of pepsin dispersion (5.21 mg/mL) added and incubated in a shaking water bath at 37 °C and 200 strokes/min for 30 min. Then, 8 mL of 0.5 M sodium acetate buffer (pH 5.2) were added and homogenized. To each reaction tube, 4 mL of an enzyme solution (pancreatin, amyloglucosidase and invertase) and 7 glass beads (7 mm diameter) were added. After 20 and 120 min of reaction, 1 mL aliquots were taken and immediately mixed with 2 mL of 80% ethanol, its glucose content was

2.4. Starch granule morphology The starch morphology and birefringence patterns were observed with a Motic BA-210 digital microscope (Hong Kong, China). The images were acquired at 40 × magnification (×40) under normal and polarized light. 372

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

generated in absorption mode with mid-IR (ca. 4000–800 cm− 1) with a resolution of 4 cm− 1 by 20 scans. A half-band width of 15 cm− 1 and a resolution enhancement factor of 1.5 with Bessel apodization were employed. Intensity measurements were performed on both the original and the deconvoluted spectra by calculating the height of the absorbance bands from their baseline. The regions of specific interest studied included the protein Amide I, II and III groups, as well as the secondary protein structures, α-helix and β-sheet in the IR regions of approximately 1715 to 1480 cm− 1. Moreover, the starch crystalline regions from 1300 to 800 cm− 1 were analyzed to estimate the changes on their relative intensities. All chemical functional groups were identified using published reports (Susi & Byler, 1983; Vernon-Carter et al., 2015; Yu, McKinnon, Christensen, & Christensen, 2004).

quantified with the glucose oxidase-peroxidase reagent. Starch classifications based on the rate of hydrolysis were: rapidly digestible starch (RDS) (digested within 20 min), slowly digestible starch (SDS) (digested between 20 and 120 min) and resistant starch (RS) (undigested after 120 min). With the aim to study the molecular characteristics of cooked and digested flours by FTIR, an additional 5 mL aliquot was taken at 120 min of reaction, and poured into a test tube which contained 20 mL of absolute ethanol (200 proof), the tube was centrifuged at 9000 × g for 30 min, the supernatant was discarded and the resulting pellet was freeze dried, milled and sieved to pass a 100 US mesh. 2.10. In vitro starch digestion rate and predicted glycemic index The method described by Granfeldt, Björck, Drews, and Tovar (1992) was employed to evaluate the in vitro rate of starch hydrolysis in cooked flours dispersions, which were incubated in a boiling water bath for 30 min with constant magnetic stirring (300 rpm) to promote starch gelatinization, after the time they were cooled down to 37 °C following an enzymatic digestion. The percentage of hydrolyzed starch by Porcine pancreatic α-amylase at 30, 60, 90, 120 and 180 min was estimated. The hydrolysis index (HI) was calculated from the ratio between the area under the hydrolysis curve compared with a reference sample (white bread). The predicted GI (pGI) was estimated from the HI and relative values calculated using the equation established by Goñi, García-Alonso, and Saura-Calixto (1997), with a reported correlation coefficient of R = 0.89, P < 0.05.

pGI = 39.71 + 0.549 (HI)

2.12. Statistical analyses A one-way variance analysis was performed, and when significant differences were found at a significance level of 0.05, a Tukey test of multiple comparisons was used. With the aim to understand the possible interactions within whole flours with their functional and digestion properties, a Pearson correlation analysis was performed at both P < 0.05 and P < 0.01 levels of significance. With the aim to evaluate the influences due to composition and treatment a principal component analysis (PCA). All statistical analyses were performed with the Minitab software 17 (version 17.3). All experiments and procedures described in this research were performed in triplicate unless otherwise specified.

(2)

3. Results and discussion 2.11. In vitro protein digestion analysis 3.1. Chemical composition characteristics The protocol of Hsu, Vavak, Satterlee, and Miller (1977) was employed to estimate the protein digestibility in cooked pulse flours, this with the aim to emulate the cooking steps prior to pulses consumption, this methodology presents a good correlation with in vivo models (R = 0.90) and the recovered information is helpful to infer some of the possible molecular interactions occurred due to thermal treatments as shown in previous studies (De La Rosa-Millan et al., 2015; de la RosaMillán, Orona-Padilla, Flores-Moreno, & Serna-Saldívar, 2017). Briefly, 50 mL of an aqueous suspension of the flours at a concentration of 6.25 mg protein/mL was prepared and incubated in a boiling water bath with constant magnetic stirring for 30 min. After, the flour suspensions were cooled down to 37 °C and its pH immediately adjusted to pH 8.0 with either 0.1 N HCl or NaOH. Once the pH was reached, 5 mL of a multi-enzyme solution consisting of 1.6 mg trypsin (15 units/mg), 3.1 mg chymotrypsin (60 units/mg) and 1.3 mg peptidase (40 units/ mg) per/mL (adjusted to pH 8.0) was added to each reaction recipient and incubated for 10 min at 37 °C. After the time, the Δ pH was used to calculate the protein digestibility values, that were compared against a 99.9% pure casein standard (which is assumed to be 100% digestible). 2.11. Analysis of the molecular interactions during in vitro digestion by ATR-FTIR. The 25-mL aliquots recovered after the Englyst protocol were centrifuged at 11,000 ×g during 30 min. The supernatants were discarded and the pellets were re-suspended in 10 mL of absolute ethanol to remove all the hydrolyzed molecules, this step was repeated twice. The washed pellets were vacuum-dried for 12 h at 45 °C. To establish comparisons with undigested samples, whole pulse flours dispersions (20% solids content) were prepared and incubated in a boiling water bath for 30 min under constant stirring. Afterwards, the samples were quickly frozen with liquid nitrogen and lyophilized. Once dried, all samples were grounded to pass a No. 70 US Mesh, and stored until analysis in an ATR-FTIR apparatus (Spectrum 1, Perkin Elmer, Norwalk, VA). The molecular spectral data of whole and digested cooked pulse flours were collected and corrected with the air background, and further analyzed with the Spectrum software (ver. 5.3.0). The spectra were

The chemical composition of flours is determinant on their use as food ingredients due to its direct impact in the final product characteristics. The studied pulse flours showed protein contents ranged from 19.53 to 29.56% for chickpea and broad bean, respectively; that were online with previous reports for the same sources (Aryee & Boye, 2016; Du et al., 2014). A similar trend regarding lipid amounts was found on the same sources, as well as relative high ash contents, in our case, the lentils and black beans contained ≤4.63% of lipids and 4.06 to 5.90% of ash, respectively (Table 1). 3.2. Starch and fiber composition of pulse flours The carbohydrates represent the largest fraction of legume flours, from which starch is their major component. In this regard, the whole pulse flours TS content ranged from 43.14 to 47.25% for black bean and lentil flour, respectively (Table 2). As result of the milling steps during flour preparation, a fraction of DS was present, with values of 4.16 to 6.32% for black bean and broad bean, respectively. When thermal treatments were applied, the DS increased up to 9.43% for the broad bean + ANN and to 11.63% for the chickpea + HMT. Previous reports have related the DS fraction with denser starch-protein matrices, that are susceptible to higher friction damage (Singh, Singh Sandhu, & Kaur, 2004; Yu et al., 2004). The amount of DS may promote detrimental effects on pasting formation and overall performance in food systems (Brummer et al., 2015; Tosh & Yada, 2010). These increments could be related with the partial disruption of the granular surface, promoted by fractures and fissures generated during the hydration-dehydration processes during thermal treatments; such features also formed void areas within the starch granules (Fig. 1) which reflect their granular rearrangement, as reported by others (Chung et al., 2008; O'Brien & Wang, 2008). From the nutritional perspective, the whole pulse flours contained good amounts of both soluble and insoluble dietary fibers (Tosh & Yada, 2010). Other studies have found that these fractions may change due to the application of high temperatures combined with high 373

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Table 1 Chemical composition of thermal treated whole pulse flours. Sample

Moisture (%)

Protein (%)

Lipids (%)

Ash (%)

Black bean Broad bean Chickpea Lentil Black bean + ANN Broad bean + ANN Chickpea + ANN Lentil + ANN Black bean + HMT Broad bean + HMT Chickpea + HMT Lentil + HMT

8.20 ± 0.53b 10.11 ± 0.31a 8.10 ± 0.74b 9.50 ± 0.73a 8.66 ± 0.64b 9.21 ± 0.49a 8.77 ± 0.90b 9.36 ± 0.44a 10.16 ± 0.62a 8.39 ± 0.06b 9.86 ± 0.10a 7.54 ± 0.07c

22.17 29.56 19.53 22.15 22.31 29.46 20.15 21.89 22.91 30.20 20.61 22.91

2.20 2.03 4.63 3.63 2.17 2.06 4.37 3.68 2.12 2.00 4.61 3.60

5.57 5.15 4.83 5.10 5.33 5.01 4.87 5.10 5.90 5.06 4.99 5.19

± ± ± ± ± ± ± ± ± ± ± ±

0.09b 0.72a 0.69d 0.51b 0.38b 0.59a 0.95c 0.11b 0.27c 0.74a 0.68d 0.74b

± ± ± ± ± ± ± ± ± ± ± ±

0.07b 0.40b 0.81a 0.15b 0.58c 0.52b 0.11a 0.25b 0.29b 0.20c 0.13a 0.21b

± ± ± ± ± ± ± ± ± ± ± ±

Carbohydrates (%)a 0.81a 0.30a 0.51a 0.37a 0.61a 0.97a 0.02a 0.85a 0.55a 0.58a 0.14a 0.92a

70.06 63.26 71.01 69.12 70.19 63.47 70.61 69.33 69.07 62.74 69.79 68.30

± ± ± ± ± ± ± ± ± ± ± ±

0.47b 0.75d 0.52a 0.21b 0.27b 0.00d 0.08a 0.73b 0.79b 0.85d 0.98b 0.02c

Data are the average of three replicates ± standard deviation. Different letters in the same column indicate significant differences (P < 0.05). ANN: annealing, HMT: heat moisture treatment. a Calculated by difference.

HMT counterparts were from 31.21 to 60.52. We hypothesized that the excess water, combined with mild temperatures promoted partial solubilization of pigments, thus favoring their leaching from testa and cotyledon tissues, resulting in colored flours. In the same sense, the analyzed legumes present phenotypical differences among different varieties that could impact their color characteristics. The chickpea variety in this study was a commercial available light-colored hybrid, without the presence of notorious or significant amount of pigments in the testa layers, that on the other hand were present on the other legumes. On the other hand, during HMT the restricted amount of water, high temperatures and the chemical composition of flours may generate Maillard reaction products, as reported by Moussou et al. (2016). When observed under the microscope, the untreated flours showed intact starch granules with well-defined birefringence pattern when observed under polarized light (Fig. 1D). On the other hand, the ANN and HMT flours presented apparently void granules as well as diffused birefringence (Fig. 1G and M) that could have resulted from the rearrangement of the granular crystalline architecture. When thermal treatments were applied, the formation of dense like particles was evident, that prevail after cooking and 120 min of in vitro digestion. These features may be related with their physicochemical and its protein and starch in vitro digestion changes, as reported by other authors (Chung et al., 2008; De La Rosa-Millan et al., 2015).

shear during mechanical processing (Alam, Pathania, & Sharma, 2016; Ma, Boye, Swallow, Malcolmson, & Simpson, 2016). Despite this, in this study, the processing conditions applied for the different treatments did not significantly affect TDF content (Table 2). 3.3. Phenolic content of pulse flours In pulses, the bioactive compounds, like polyphenols, are often concentrated in the seed testa with lower but still significant quantities on their cotyledons (Jukanti, Gaur, Gowda, & Chibbar, 2012; Kumar, Rani, & Hussain, 2016). However, these molecules are sensitive to heat, for this reason when pulses are cooked, there a significant decrease on their amounts and potential bioactivity (Ma, Boye, Azarnia, et al., 2016). The total phenolic content of raw flours ranged from 1.32 to 1.96 mg/g of gallic acid eq. for broad bean and black bean, respectively (Table 3). After ANN, the flours showed similar values compared to their native counterparts, whilst when HMT was applied, the phenolic content of flours significantly decreased, this could be related with the high temperatures use in this process, which are known to decrease the bioactivity of these molecules (Moussou et al., 2016). 3.4. Color characteristics of whole pulse flours and starch morphology One of the main characteristics when choosing food ingredients is their color, which has a direct impact on the overall acceptance of the final products (Shariati-Ievari et al., 2016). The use of hydrothermal treatments promoted significant color changes on flours (Table 3). Regarding Δ E, the ANN flours ranged from 45.58 to 86.85, whereas the

3.5. Rapid viscosity analysis of whole pulse flours From the agronomic point of view, the pulse seeds present different cotyledon arrangement, composition and thus physicochemical

Table 2 Starch and fiber composition of whole pulse flours. Sample

TS (%)

Black bean Broad bean Chickpea Lentil Black bean + ANN Broad bean + ANN Chickpea + ANN Lentil + ANN Black bean + HMT Broad bean + HMT Chickpea + HMT Lentil + HMT

43.14 43.43 45.32 47.25 43.06 42.26 44.95 47.61 43.65 42.11 44.99 47.65

± ± ± ± ± ± ± ± ± ± ± ±

0.14c 0.57c 0.29b 0.11a 0.33c 0.67d 0.87b 0.98a 0.31c 0.76d 0.55b 0.54a

Amylose (%)

DS (%)

TDF (%)

23.19 23.91 20.19 22.11 20.06 22.14 20.24 20.16 24.06 24.11 20.20 22.26

4.16 ± 0.18e 6.32 ± 0.54d 5.55 ± 0.59d 4.67 ± 0.27e 8.26 ± 0.11c 9.43 ± 0.90b 8.82 ± 0.21c 8.95 ± 0.34c 11.15 ± 0.35a 10.26 ± 0.16b 11.63 ± 0.84a 11.43 ± 0.07a

14.19 12.61 13.28 13.02 14.62 13.20 13.42 13.62 14.86 13.33 13.05 13.76

± ± ± ± ± ± ± ± ± ± ± ±

0.53a 0.67a 0.56c 0.75b 0.63c 0.46d 0.18c 0.33c 0.39a 0.39a 0.57c 0.46b

± ± ± ± ± ± ± ± ± ± ± ±

0.22a 0.30c 0.80b 0.38b 0.36a 0.61b 0.59b 0.21b 0.93a 0.56b 0.85b 0.98b

IDF (%)

SDF (%)

10.61 ± 0.93a 9.83 ± 0.64b 9.66 ± 0.77b 10.06 ± 0.78a 11.31 ± 0.87a 10.27 ± 0.55a 9.88 ± 0.47b 10.85 ± 0.27a 11.35 ± 0.22a 10.60 ± 0.32a 10.42 ± 0.55a 10.11 ± 0.47a

3.58 2.78 3.62 2.96 3.31 2.93 3.54 2.77 3.50 2.73 2.63 3.65

Data are the average of three replicates ± standard deviation. Different letters in the same column indicate significant differences (P < 0.05). ANN: annealing, HMT: heat moisture treatment, TS: total starch, DS: damaged starch, TDF: total dietary fiber, IDF: insoluble dietary fiber, SDF: soluble dietary fiber.

374

± ± ± ± ± ± ± ± ± ± ± ±

0.19a 0.35b 0.80a 0.69b 0.24a 0.05b 0.08a 0.81b 0.02a 0.10b 0.14b 0.69a

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Fig. 1. Morphological and birefringence characteristics of pulse flours with different thermal treatments. A, B and C broad bean flour under normal light, C, D and F broad bean flours under polarized light. G, H and I under normal light; J, K and L: annealed chickpea flours under polarized light. M, N and O: heat moisture treated lentil flours under normal light; P, Q and R: heat moisture treated lentil flours under polarized light. Left to right columns: raw flours, cooked (30 min under constant stirring at boiling temperature) and after 120 min of digestion of the Englyst in vitro digestion procedure.

375

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Table 3 Total Phenolics and color changes on thermal treated whole pulse flours. Sample

Total phenolics

L

Black bean Broad bean Chickpea Lentil Black bean + ANN Broad bean + ANN Chickpea + ANN Lentil + ANN Black bean + HMT Broad bean + HMT Chickpea + HMT Lentil + HMT

1.96 1.43 1.32 1.61 1.85 1.33 1.26 1.48 1.06 0.79 0.73 0.86

79.86 85.32 86.80 83.42 45.33 55.26 84.66 50.55 30.19 35.33 56.54 39.42

± ± ± ± ± ± ± ± ± ± ± ±

0.24a 0.13b 0.17b 0.12b 0.16a 0.18b 0.11c 0.10b 0.13d 0.13e 0.02e 0.14e

a ± ± ± ± ± ± ± ± ± ± ± ±

0.01e 0.06b 0.10a 0.11d 0.04i 0.10g 0.14c 0.05h 0.02l 0.01k 0.12f 0.20j

0.88 0.43 2.36 0.55 4.26 3.63 1.26 3.33 6.66 5.55 4.23 7.01

± ± ± ± ± ± ± ± ± ± ± ±

0.03i 0.05k 0.06g 0.07j 0.04d 0.07e 0.08h 0.01f 0.08b 0.04c 0.02d 0.05a

b

ΔE

6.76 ± 0.07e 7.26 ± 0.08d 22.13 ± 0.06a 6.39 ± 0.09f 2.21 ± 0.04l 2.41 ± 0.09j 19.33 ± 0.02c 3.01 ± 0.05j 4.26 ± 0.07h 3.93 ± 0.08i 21.17 ± 0.09b 4.43 ± 0.04g

80.15 85.63 89.61 83.67 45.58 55.43 86.85 50.75 31.21 35.98 60.52 40.28

± ± ± ± ± ± ± ± ± ± ± ±

0.09e 0.05c 0.05a 0.06d 0.02i 0.07g 0.09b 0.09h 0.02l 0.10k 0.02f 0.07j

Data are the average of three replicates ± standard deviation. Different letters in the same column indicate significant differences (P < 0.05). ANN: annealing, HMT: heat moisture treatment, L = luminosity, a = blueness to greenness, b = yellowness to redness. ΔE = color index. Fig. 2. Rapid viscosity profiles of thermal treated flours. A) Native whole pulse flours, B) ANN whole pulse flours and C) HMT whole pulse flours. BB: black bean; BrB: broad bean; Ch: chickpea; L: lentil.

376

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Table 4 Rapid viscosity parameters of whole pulse thermal treated flours. Sample

Peak time (min)

Pasting temperature (°C)

Peak viscosity (cP)

Breakdown (cP)

Setback (cP)

Final viscosity (cP)

Black bean Broad bean Chickpea Lentil Black bean + ANN Broad bean + ANN Chickpea + ANN Lentil + ANN Black bean + HMT Broad bean + HMT Chickpea + HMT Lentil + HMT

7.00 5.93 6.66 7.00 7.00 7.00 5.53 7.00 6.26 6.86 6.40 6.46

82.2 ± 0.23c 86.25 ± 0.70b 79.6 ± 0.64d 81.9 ± 0.54c 85.51 ± 0.35b 93.2 ± 0.47a 79.5 ± 0.50d 93.9 ± 0.64a 85.67 ± 0.41b 82.11 ± 0.29c 92.91 ± 0.27a 90.93 ± 0.67b

229 ± 12.19g 450 ± 11.19f 616 ± 8.29c 830 ± 5.69b 116 ± 7.90h 569 ± 7.56d 856 ± 5.94a 542 ± 4.30e 14 ± 3.33i 13 ± 1.41i 15 ± 1.37i 12 ± 1.92i

58 ± 0.99b 15 ± 1.40g 23 ± 1.29e 10 ± 1.09h 19 ± 1.52f 40 ± 0.95d 69 ± 0.99a 53 ± 1.29c 1 ± 0.52i 2 ± 0.93i 1 ± 0.55i 3 ± 0.49i

439 ± 0.21b 442 ± 0.68a 138 ± 0.50h 238 ± 0.77e 141 ± 0.50g 196 ± 0.35f 417 ± 0.90c 304 ± 0.39d 9 ± 0.06l 12 ± 0.83i 10 ± 0.67j 8 ± 0.36k

610 ± 0.84g 877 ± 2.73c 731 ± 2.02e 1188 ± 3.08b 238 ± 1.98h 725 ± 1.39f 1204 ± 3.58a 793 ± 1.54d 22 ± 0.25i 23 ± 3.32i 24 ± 2.68i 17 ± 1.43j

± ± ± ± ± ± ± ± ± ± ± ±

0.97a 0.30a 0.57a 0.17a 0.05a 0.29a 0.18b 0.57a 0.14b 0.45a 0.09a 0.38a

Data are the average of three replicates ± standard deviation. Different letters in the same column indicate significant differences (P < 0.05). ANN: annealing, HMT: heat moisture treatment. Table 5 Thermal parameters of thermal treated whole pulse flours. Sample

To (°C)

Black bean Broad bean Chickpea Lentil Black bean + ANN Broad bean + ANN Chickpea + ANN Lentil + ANN Black bean + HMT Broad bean + HMT Chickpea + HMT Lentil + HMT

67.36 67.96 68.43 65.11 67.16 68.89 66.75 64.36 65.01 63.17 63.10 62.19

± ± ± ± ± ± ± ± ± ± ± ±

Tp (°C) 0.26a 0.29a 0.67a 0.67b 0.74a 0.17a 0.42a 0.38b 0.83b 0.80c 0.76c 0.38c

70.26 71.18 69.14 67.26 68.43 69.36 67.33 65.43 65.26 65.11 66.26 65.64

± ± ± ± ± ± ± ± ± ± ± ±

Tc (°C) 0.34a 0.73a 0.16b 0.38d 0.41c 0.30b 0.23d 0.11e 0.10e 0.60e 0.35d 0.29e

72.41 73.19 70.28 70.43 69.31 70.15 68.06 66.00 67.42 68.26 68.43 67.02

± ± ± ± ± ± ± ± ± ± ± ±

0.69a 0.19a 0.61b 0.78b 0.26c 0.33b 0.41c 0.71e 0.49d 0.58c 0.16c 0.16d

ΔH (J/g)

Tc–To (°C)

9.64 8.43 8.88 9.76 6.51 5.29 6.33 6.11 4.33 5.17 6.00 4.89

5.05 5.23 1.85 5.32 2.15 1.26 1.31 1.64 2.41 5.09 5.33 4.83

± ± ± ± ± ± ± ± ± ± ± ±

0.12a 0.55b 0.41b 0.03a 0.02c 0.31e 0.38d 0.08d 0.52e 0.32e 0.26d 0.90e

± ± ± ± ± ± ± ± ± ± ± ±

0.79a 0.86a 0.09c 0.15a 0.45b 0.72c 0.91c 0.30c 0.33b 0.49a 0.05a 0.12a

Data are the average of three replicates ± standard deviation. Different letters in the same column indicate significant differences (P < 0.05). ANN: annealing, HMT: heat moisture treatment, To: onset gelatinization temperature, Tp: peak gelatinization temperature, Tc: conclusion gelatinization temperature, ΔH: enthalpy, Tc–To: gelatinization range. Table 6 Starch and protein in vitro digestion fractions of cooked thermal treated whole pulse flours. Sample

RDS (%)

Black bean Broad bean Chickpea Lentil Black bean + ANN Broad bean + ANN Chickpea + ANN Lentil + ANN Black bean + HMT Broad Bean + HMT Chickpea + HMT Lentil + HMT

82.11 80.26 85.26 80.06 77.93 73.75 77.10 70.60 59.63 60.52 61.10 59.60

± ± ± ± ± ± ± ± ± ± ± ±

0.43b 0.22c 0.77a 0.34c 0.68d 0.33e 0.19d 0.44f 0.65g 0.68g 0.37g 0.97g

SDS (%)

RS (%)

HI

10.25 ± 0.45i 11.14 ± 0.44h 9.05 ± 0.76j 12.68 ± 0.65g 11.11 ± 0.54h 16.61 ± 0.55e 14.26 ± 0.20f 19.26 ± 0.39d 25.11 ± 0.96c 26.19 ± 0.76c 27.11 ± 0.89b 30.14 ± 0.65a

7.64 ± 0.58f 8.60 ± 0.32e 5.69 ± 0.46g 7.26 ± 0.61f 10.96 ± 0.19c 9.64 ± 0.26d 8.64 ± 0.09e 10.14 ± 0.65c 15.26 ± 0.43a 13.29 ± 0.43b 11.79 ± 0.78b 10.26 ± 0.17c

87.71 84.66 90.15 82.16 78.11 76.42 76.16 75.33 64.67 68.42 67.21 66.36

pGI ± ± ± ± ± ± ± ± ± ± ± ±

0.44b 0.42c 0.39a 0.49d 0.59e 0.03f 0.85f 0.55f 0.27j 0.80g 0.16h 0.47i

87.86 86.18 89.20 84.81 82.59 81.66 81.52 81.06 75.21 77.27 76.60 76.14

Prot. Dig. (%) ± ± ± ± ± ± ± ± ± ± ± ±

0.60a 0.58b 0.37a 0.83c 0.49d 0.04d 0.27d 0.65d 0.11e 0.08e 0.49e 0.69e

65.43 76.23 80.61 79.44 78.17 82.43 87.15 84.26 70.19 73.26 79.63 77.54

± ± ± ± ± ± ± ± ± ± ± ±

0.47i 0.61f 0.29d 0.49d 0.16e 0.02c 0.56a 0.02b 0.64h 0.23g 0.63d 0.03e

Data are the average of three replicates ± standard deviation. Different letters in the same column indicate significant differences (P < 0.05). ANN: annealing, HMT: heat moisture treatment, RDS: rapidly digestible starch, SDS: slowly digestible starch, RS: resistant starch, HI: hydrolysis index, pGI: predicted glycemic index, Prot. Dig.: In vitro protein digestibility.

despite their similarities in starch and amylose contents (Table 2). On the other hand, when ANN was applied, there was a significant decrease on the viscosity parameters (Fig. 2B), which was even lower in HMT treated samples (Fig. 2C). Other studies have shown that thermal treatments on starchy foods significant affects pasting characteristics due to the partial disruption of the starch granular architecture, however previous reports have shown that the density of the cotyledon matrix along with a significant amount of soluble and insoluble fiber

properties. Regarding pasting properties several authors have found significant correlations with starch content, in which the amylose molecules are of significant relevance (de la Rosa-Millán et al., 2017; Tester & Morrison, 1990). Additionally, the protein content has been found to be negatively correlated with starch pasting properties because it affects granule swelling, breakdown and setback of starch granules (Azarpazhooh & Boye, 2012; Ghumman et al., 2016). In our study, the RVA profiles of native flours differed among the four pulses (Fig. 2A), 377

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Fig. 3. FTIR spectra of thermal treated flours. A) Black bean, B) black bean after 120 min of digestion, C) AI:AII ratios, before and after digestion. AI: Amide I, AII: amide II. BB: black bean, BrB: broad bean, Ch: chickpea, L: lentil, ANN: annealing, HMT: heat moisture treatment.

ANN was applied, that were even lower after HMT (Table 5). Previous studies have related this phenomenon, showing a reduced granular stability due to the partial disruption of the crystalline structure, which often reflects on lower transition temperatures and Δ H values (Chung et al., 2008; Jacobs & Delcour, 1998). In this aspect, significant positive correlations of Δ H vs. Breakdown (R = 0.58, P < 0.05), Final viscosity (R = 0.63, P < 0.05), DS (R = 0.85, P < 0.01) were found, which reflects the structural influence of crystal arrangement in starch granules. In the same sense, the gelatinization range (Tc–To) presented significant differences in ANN flours (from 1.26 to 2.15 °C broad bean and black bean, respectively). This variations were more evident after HMT treatment (from 2.41 to 5.33 °C for black bean and lentil, respectively) and native flours (from 1.85 to 5.32 °C for chickpea and lentil, respectively); these differences could be related with the molecular rearrangement occurred in these thermal treated flours, which due to an increased crystalline structure could delay the heat transfer into starch granules, conducing to changes on gelatinization transition temperatures (De La Rosa-Millan et al., 2015). The gelatinization range was strongly correlated with the final pasting viscosity value (R = 0.96, P < 0.01) that reflects the importance of the crystallite heterogenicity

have a negative effect on viscosity, mainly due to they may restrict starch granules swelling, increasing the starch thermal stability during heat processing (Brummer et al., 2015; Du et al., 2014). Additionally, the breakdown viscosity was significantly lower when thermal was applied (R = −0.78, P < 0.01), when compared with native flours (Table 4). These differences could be related with the rearrangement of the internal granular architecture, as well as the flour protein content which have been reported to decrease swelling of starch granules, hence promoting lower water absorption of flours, rendering low viscosities (Du et al., 2014). 3.6. Thermal characteristics of pulse flours Whole flours are complex systems in which starch molecules may interact with fiber, protein and other minor components that overall can affect their heat capacity (De La Rosa-Millan et al., 2015; Divekar et al., 2016; Ma, Boye, Swallow, et al., 2016). Our native pulse flours showed starch gelatinization temperatures similar to previous reports for similar sources (Chung et al., 2008; Singh et al., 2004). On the other hand, there was a significant decrease on all thermal parameters when 378

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Fig. 4. Deconvoluted spectra of thermal treated whole pulse flours. A) Untreated black bean flour spectra of secondary protein structures, B) α helix to β sheet ratio, C) intermolecular to Intramolecular associations. BB: black bean, BrB: broad bean, Ch: chickpea, L: lentil, ANN: annealing, HMT: heat moisture treatment.

Millan et al., 2015; O'Brien & Wang, 2008). In this study, cooked native flours showed high amounts of RDS (from 80.06 to 85.26% for lentil and chickpea, respectively) with RS contents varying from 5.69 to 8.60% for chickpea and broad bean, respectively (Table 6). When ANN was applied, we found a significant increase on the SDS (up to 19.26% for lentil) and RS fraction (up to 10.96% black bean). On the other hand, the HMT treatment promoted a decrease of the RDS fraction which ranged from 59.60 to 61.10% for lentil and chickpea, respectively. Overall, the thermally processed flours showed significantly lower pGI, from 81.06 to 82.58 for lentil + ANN and black bean + ANN; and from 75.21 to 77.27 for black bean and broad bean + HMT. The pGI showed positive correlations with the amounts of RDS (R = 0.96, P < 0.01), DS (R = 0.75, P < 0.01), Tp (R = 0.82, P < 0.01) and Δ H (R = 0.82, P < 0.01) and negative correlation with the Tc–To value (R = −0.94, P < 0.01), the correlation with this parameters indicates the dependence of granule integrity during cooking processes, as the one used prior to the in vitro digestion that affects the way in which starch is digested, and also the influence of thermal treatments that promoted damage on the surface of starch

during the viscosity development and its behavior during cooling as well as the amylose content (R = 0.694, P < 0.01) that show the influence of this molecule on granular stability (Brummer et al., 2015; Ma, Boye, Azarnia, et al., 2016; Shariati-Ievari et al., 2016). 3.7. Starch in vitro digestion properties of whole pulse flours Pulse seeds are one of the main staple in eastern cultures mainly due to their high amount of complex carbohydrates and proteins that complements the amino acid profile of cereals. However, the nutritional availability of these nutrients is greatly affected by the type and extent of cooking method. In a general way, cooking in excess water increases the amount of rapidly digestible starch (RDS); nevertheless, various studies have shown that pulses starch presents low digestion properties and therefore low glycemic indexes (Jenkins et al., 1982; Hoover & Zhou, 2003; Jukanti et al., 2012). In the same sense, previous research has shown that the use of hydrothermal treatments, tends to decrease its digestion rate by amylolytic enzymes, resulting in relatively high amounts of SDS and RS after in vitro digestion protocols (De La Rosa379

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Fig. 5. Deconvoluted spectra of starch crystal characteristics. A) Effect of thermal treatment, B) effect of digestion, C)1047/1022 ratio. BB: black bean, BrB: broad bean, Ch: chickpea, L: lentil, ANN: annealing, HMT: heat moisture treatment.

showed in vitro protein digestion rates from 84.81 to 89.20%, for lentil and chickpea, respectively. When ANN and HMT were applied, the protein digestion decreased to 81.06 to 81.52% for lentil and chickpea and 75.21 to 77.27% for black bean and broad bean, respectively. This behavior could be related with the protein rearrangement due to higher temperatures during HMT, as well as the reorganization of the other flour components, mainly the amylose molecules which showed a strong negative correlation (R = − 0.850, P < 0.01) which resulted in a lower protein hydrolysis, despite the differences in the hydrolysis protocols. 3.8. Molecular characteristics of digested thermal treated pulse flours by ATR-FTIR. Due to the applied thermal treatments, there were notable differences in spectral intensity among the studied cooked pulses (Fig. 3A). These can be related with the amount of water employed during thermal processes, as shown by other studies (Chung et al., 2008; O'Brien & Wang, 2008). In terms of the ANN treatments, the excess water utilized resulted in greater unfolding of molecular protein structures, that along with the mild temperature probably increased the motion range of molecules. On the other hand, the restricted amount of water and high temperatures employed to produce the HMT flours

granules (Fig. 1H). 3.8. In vitro protein digestion of whole pulse flours Regarding protein digestion, other studies involving thermal processing (as cooking or autoclaving) have shown to facilitate their hydrolysis, mainly due to molecular rearrangements occurred during thermal denaturation, that involves torsion and flexibility of protein secondary structures α-helix and β-sheet conformations (1635 and 1650 cm− 1, respectively). In the case of ANN the digestion is facilitated due to the relative mild temperatures and to the amount of water, which promotes molecular plasticity, enabling a higher range of rearrangement. In the case of HMT the high temperatures combined with the restrictive amount of water promoted the interaction with starch and protein molecules promoting the formation of dense particles which may present slower digestion patterns. (De La Rosa-Millan et al., 2015; Divekar et al., 2016; Torres et al., 2016). On the other hand, the use of high temperatures in low moisture environments may lead to irreversible changes on amino acids accessibility; thus, decreasing they overall quality, that is reflected in poor solubility and lower bio-availability (Torres et al., 2016). In this study, the cooked whole flours 380

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Fig. 6. Principal component analysis of thermal treated pulse flours. A) Loading plots, B) cluster analysis.

Christensen, 2009; De La Rosa-Millan et al., 2015). Results herein indicated that the cooked pulse flours presented high intensities on the 1650 cm− 1 wavelength, that is related with the α–helix protein structures (Fig. 4A) whereas in the digested fractions, the β–sheets (at 1635 cm− 1) prevailed, and showed a positive correlation with the Δ H value (R = 0.65, P < 0.05). In thermal treated cooked flours, there was a higher ratio of the β–sheets structures, compared with the untreated flours, which ranged as broad bean < chickpea < black bean < lentil. On the other hand, after 120 min of digestion, there was an increasing trend on the α: β ratios on all pulse flours when ANN and HMT were applied, showing that a granular rearrangement occurred in the starch fraction which also affected the protein fraction of these flours. In this study the ratio of α-helix/β-sheet presented negative correlation with SDS (R = − 0.761, P < 0.05), as well as with pGI (R = − 0.680, P < 0.05) and DS content (R = − 0.615, P < 0.05). A similar trend was found during the analysis of their molecular inter (1625 cm− 1) and intra (1616 cm− 1) associations within the Amide III band (Fig. 4C), that has been previously related with the torsion and

promoted protein damage and rearrangement of the structures that lead to a shrinkage of their tertiary structure (De La Rosa-Millan et al., 2015). After 120 min of digestion, the thermally treated broad bean pulse flours still showed a protein fraction with high intensity (Fig. 3B). Such differences were reflected on their AI: AII ratios (Fig. 3C) which other authors have related with the extent of protein denaturation (Carbonaro et al., 2008; Moussou et al., 2016). Regarding the protein zone there were evident changes after digestion, in this aspect the HMT flours showed lower AI: AII ratio intensities compared with their ANN or untreated counterparts, that negatively correlated with the RS fraction (R = − 0.73, P < 0.01). The deconvolution analysis of the spectra provided valuable information about protein secondary structures as well as the presence of crystalline and amorphous starch structures (Chung et al., 2008; Vernon-Carter et al., 2015). Previous studies have shown that the α–helix and β–sheet protein secondary structures are responsible for modulating the extent of protein digestion and consequently affect functional properties like gel strength and emulsion and foaming capacities (Doiron, Yu, McKinnon, & 381

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

flexion characteristics of molecules (Carbonaro et al., 2008; Manning, 2005). In a previous study, the molecular characteristics of this band were related with thermal properties derived from the interaction among proteins, starch and fiber fractions (Cai & Singh, 2004; Manning, 2005). Regarding the spectral signals from starch molecules, the cooked whole flours presented a higher intensity on their (1049/1022 cm− 1 ratio) (Crystalline to amorphous) signals before digestion (Fig. 5A) whereas in the case of ANN and HMT flours (Fig. 5B) lower spectral intensities when compared with the untreated samples. On the other hand, after 120 min of in vitro digestion, the intensity ratio of 1049/ 1022 cm− 1 signals decreased, but they were higher in ANN and HMT flours, when compared with the untreated flour (Fig. 5C). Such trend was observed in all flours, and could be related with a higher proportion of ordered structures promoted by the applied hydro thermal treatments; as shown by the correlation of crystalline structures signal (1047 cm− 1) with the RS fraction (R = 0.70, P < 0.05). These results agree with previous studies, in which the chemical composition of pulse flours presented restrictions to enzymatic hydrolysis, and that such resistance was enhanced when hydrothermal treatments were applied. (Aryee & Boye, 2016; Chung et al., 2008; Moussou et al., 2016). Further research is currently carried on about the molecular weight and amino acids sequence of the protein fraction as well as the molecular weight distribution of starches in the digested flours.

References AACCI International (Ed.). (2000). Approved methods of analysis(11th ed.). AACCI International: St. Paul, MN. Ahn, H. J., Kim, J. H., & Ng, P. K. W. (2005). Functional and thermal properties of wheat, barley and soy flours and their blends treated with a microbial transglutaminase. Journal of Food Science, 70, 380–386. Alam, M. S., Pathania, S., & Sharma, A. (2016). Optimization of the extrusion process for development of high fibre soybean-rice ready-to-eat snacks using carrot pomace and cauliflower trimmings. LWT - Food Science and Technology, 74, 135–144. Aryee, A. N., & Boye, J. I. (2016). Comparative study of the effects of processing on the nutritional, physicochemical and functional properties of lentil. Journal of Food Processing and Preservation. http://dx.doi.org/10.1111/jfpp.12824 (Article in Press). Azarpazhooh, E., & Boye, J. I. (2012). Composition of processed dry beans and pulses. In M. Siddiq, & M. A. Uebersax (Eds.). Dry beans and pulses production, processing and nutrition (pp. 101–128). (1st ed.). Oxford, UK: Blackwell Publishing Ltd. Brummer, Y., Kaviani, M., & Tosh, S. M. (2015). Structural and functional characteristics of dietary fibre in beans, lentils, peas and chickpeas. Food Research International, 67, 117–125. Cai, S., & Singh, B. R. (2004). A distinct utility of the amide III infrared band for secondary structure estimation of aqueous protein solutions using partial least squares methods. Biochemistry, 43, 2541–2549. Carbonaro, M., Maselli, P., Dore, P., & Nucara, A. (2008). Application of Fourier transform infrared spectroscopy to legume seed flour analysis. Food Chemistry, 108, 361–368. Chung, H. J., Liu, Q., Donner, E., Hoover, R., Warkentin, T. D., & Vandenberg, B. (2008). Composition, molecular structure, properties, and in vitro digestibility of starches from newly released Canadian pulse cultivars. Cereal Chemistry, 85, 471–479. De La Rosa-Millan, J., Lin, A. H.-M., Osorio-Díaz, P., Agama-Acevedo, E., Hamaker, B. R., & Bello-Perez, L. A. (2015). Influence of annealing flours from raw and pre-cooked plantain fruit on cooked starch digestion rates. Starch-Starke, 67, 139–146. de la Rosa-Millán, J., Orona-Padilla, J. L., Flores-Moreno, V. M., & Serna-Saldívar, S. O. (2017). Physicochemical and digestion characteristics of starch and fiber-rich subfractions from four pulse bagasses. Cereal Chemistry, 94, 524–531. Divekar, M. T., Karunakaran, C., Lahlali, R., Kumar, S., Chelladurai, V., Liu, X., ... Jayas, D. S. (2016). Effect of microwave treatment on the cooking and macronutrient qualities of pulses. International Journal of Food Properties (Article in Press). Doiron, K., Yu, P., McKinnon, J. J., & Christensen, D. A. (2009). Heat-induced protein structure and subfractions in relation to protein degradation kinetics and intestinal availability in dairy cattle. Journal of Dairy Science, 92, 3319–3330. Du, S. K., Jiang, H., Yu, X., & Jane, J. (2014). Physicochemical and functional properties of whole legume flour. LWT- Food Science and Technology, 55, 308–313. Englyst, H. N., Kingman, S. M., & Cummings, J. H. (1992). Classification and measurement of nutritionally important starch fractions. European Journal of Clinical Nutrition, 4, 33–50. Ghumman, A., Kaur, A., Singh, N., & Singh, B. (2016). Effect of feed moisture and extrusion temperature on protein digestibility and extrusion behaviour of lentil and horsegram. LWT - Food Science and Technology, 70, 349–357. Goñi, I., García-Alonso, A., & Saura-Calixto, F. (1997). A starch hydrolysis procedure to estimate glycemic index. Nutrition Research, 17, 427–437. Granfeldt, Y., Björck, I., Drews, A., & Tovar, J. (1992). An in vitro procedure based on chewing to predict metabolic responses to starch in cereal and legume products. European Journal of Clinical Nutrition, 46, 649–660. Hoover, R., & Zhou, Y. (2003). In vitro and in vivo hydrolysis of legume starches by αamylase and resistant starch formation in legumes - A review. Carbohydrate Polymers, 54(4), 401–417. http://dx.doi.org/10.1016/S0144-8617(03)00180-2. Hsu, H. W., Vavak, D. L., Satterlee, L. D., & Miller, G. A. (1977). A multienzyme technique for estimating protein digestibility. Journal of Food Science, 42, 1269–1273. Jacobs, H., & Delcour, J. A. (1998). Hydrothermal modifications of granular starch, with retention of the granular structure: A review. Journal of Agricultural and Food Chemistry, 46, 2895–2905. Jenkins, D. J. A., Thome, M. J., Camelon, K., Jenkins, A. L., Rao, A. V., Taylor, R. H., ... Francis, T. (1982). Effect of processing on digestibility and the blood glucose responses: A study of lentils. American Journal of Clinical Nutrition, 36, 1093–1101. Jukanti, A. K., Gaur, P. M., Gowda, C. L. L., & Chibbar, R. N. (2012). Nutritional quality and health benefits of chickpea (Cicer arietinum L.): A review. British Journal of Nutrition, 108, S11–S26. Kumar, V., Rani, A., & Hussain, L. (2016). Essential amino acids profile of differentially processed soy products and their efficiency in meeting daily requirement. Nutrition and Food Science, 46(2), 237–245. http://dx.doi.org/10.1108/NFS-07-2015-0082. Ma, Z., Boye, J. I., Azarnia, S., & Simpson, B. K. (2016). Volatile flavor profile of Saskatchewan grown pulses as affected by different thermal processing treatments. International Journal of Food Properties, 19, 2251–2271. Ma, Z., Boye, J. I., Swallow, K., Malcolmson, L., & Simpson, B. K. (2016). Techno-functional characterization of salad dressing emulsions supplemented with pea, lentil and chickpea flours. Journal of the Science of Food and Agriculture, 96, 837–847. Maaran, S., Hoover, R., Vamadevan, V., Waduge, R. N., & Liu, Q. (2016). In vitro amylolysis of pulse and hylon VII starches explained in terms of their composition, morphology, granule architecture and interaction between hydrolyzed starch chains. Food Chemistry, 192, 1098–1108. Manning, M. C. (2005). Use of infrared spectroscopy to monitor protein structure and stability. Expert Review of Proteomics, 2, 731–743. Megías, C., Cortés-Giraldo, I., Alaiz, M., Vioque, J., & Girón-Calle, J. (2016). Isoflavones in chickpea (Cicer arietinum) protein concentrates. Journal of Functional Foods, 21, 186–192.

3.8.1. Principal component analysis In this study we use PCA to overview of the interrelationships between the starch structural properties due to thermal treatments (Fig. 6A). The first and second components (PC1 and PC2) accounted for an accumulative variance of 77.40%. The loading plot (Fig. 1A) of PCA contains the correlations among some physicochemical and thermal properties and the starch digestion fractions. PC1 accounted 54.80 of the accumulated variance, in which the main contributor factors were the SDS, Damaged starch and the amount of RS, along with the spectral characteristics related with the inter/intra molecular association of proteins (1625/1616 cm− 1) and the overall secondary structures (Amide I/Amide II ratio). Meanwhile, the digestion properties RDS and pGI where related with the total phenolic contents, and crystalline organization of starch, related with DH and Tp (Gelatinization enthalpy and peak gelatinization, respectively), which is influenced by the crystalline/amorphous ratio of starches (1047/ 1022 cm− 1). The cluster analyses of PC1 and PC2 formed three groups according to similarities among the flours and the evaluated thermal treatments, which reflects that the pulse source has a significant effect on the studied parameters when thermal treatments are applied as in the case of native chickpea and broad bean with HTM which present substantial differences in their respective groups (Fig. 6B).

4. Conclusions The use of hydrothermal treatments on whole pulse flours could be a feasible alternative to obtain functional ingredients and foods with low glycemic index. The pulse flours presented physical interactions that occurred at molecular level, mostly between starch and proteins, which led to decrease the starch hydrolysis but higher protein digestion. This study showed that it was feasible to promote such interactions in food systems aimed towards the modulation of both starch and protein digestion rates and therefore to impart unique functional and nutraceutical characteristics.

Acknowledgements CECHM and JIVT appreciate the financial support from SIP-IPN and EDI-IPN. 382

Food Research International 105 (2018) 371–383

C.E. Chávez-Murillo et al.

Susi, H., & Byler, D. M. (1983). Protein structure by Fourier transform infrared spectroscopy: Second derivative spectra. Biochemical and Biophysical Research Communications, 115, 391–397. Tester, R. F., & Morrison, W. R. (1990). Swelling and gelatinization of cereal starches. I. Effects of amylopectin, amylose, and lipids. Cereal Chemistry, 67, 551–557. Torres, J., Rutherford, S. M., Muñoz, L. S., Peters, M., & Montoya, C. A. (2016). The impact of heating and soaking on the in vitro enzymatic hydrolysis of protein varies in different species of tropical legumes. Food Chemistry, 194, 377–382. Tosh, S. M., & Yada, S. (2010). Dietary fibres in pulse seeds and fractions: Characterization, functional attributes, and applications. Food Research International, 43, 450–460. Vernon-Carter, E. J., Hernandez-Jaimes, C., Meraz, M., Lara, V. H., Lobato-Calleros, C., & Alvarez-Ramirez, J. (2015). Physico-chemical characterization and in vitro digestibility of gelatinized corn starch dispersion fractions obtained by centrifugation. Starch/Staerke, 67, 701–708. Yu, P., McKinnon, J. J., Christensen, C. R., & Christensen, D. A. (2004). Using synchrotron-based FTIR microspectroscopy to reveal chemical features of feather protein secondary structure: Comparison with other feed protein sources. Journal of Agricultural and Food Chemistry, 52, 7353–7361.

Moussou, N., Corzo-Martínez, M., Sanz, M. L., Zaidi, F., Montilla, A., & Villamiel, M. (2016). Assessment of Maillard reaction evolution, prebiotic carbohydrates, antioxidant activity and α-amylase inhibition in pulse flours. Journal of Food Science and Technology, 1, 1–11. O'Brien, S., & Wang, Y. J. (2008). Susceptibility of annealed starches to hydrolysis by αamylase and glucoamylase. Carbohydrate Polymers, 72, 597–607. Romano, A., Giosafatto, V., Masi, P., & Mariniello, L. (2015). Impact of dehulling on the physico-chemical properties and in vitro protein digestion of common beans (Phaseolus vulgaris L.). Food & Function, 6, 1345–1351. Shariati-Ievari, S., Ryland, D., Edel, A., Nicholson, T., Suh, M., & Aliani, M. (2016). Sensory and physicochemical studies of thermally micronized chickpea (Cicer arietinum) and green lentil (Lens culinaris) flours as binders in low-fat beef burgers. Journal of Food Science, 81, S1230–S1242. Singh, N., Singh Sandhu, K., & Kaur, M. (2004). Characterization of starches separated from Indian chickpea (Cicer arietinum L.) cultivars. Journal of Food Engineering, 63, 441–449. Singleton, V. L., Orthofer, R., Lamuela-Raventós, R. M., & Lester, P. (1999). Analysis of total phenols and other oxidation substrates and antioxidants by means of Folin–Ciocalteu reagent. Methods in Enzymology. 299. Methods in Enzymology (pp. 152–178). U.S.A.: Academic Press.

383