Total phenolic content, antioxidant activity, shelf-life and reconstitutability of okra seeds powder: influence of milling and sieving processes

Abstract

This study aimed at relating phytochemical properties (total phenolic content and antioxidant activity), reconstitutability (water absorption capacity and water solubility index) and preservation ability (water activity) of okra seed powders to physicochemical properties (particle size distribution and moisture content). Okra seed powders were produced at three different milling frequencies, and then, each obtained sample was sieved with 180, 315 and 500 μm sieves. An increased milling frequency resulted in lower median particle size, water activity and moisture content. The maximal water activity and moisture content were 0.583 ± 0.010 and 12.07 ± 0.05% (w/w), respectively, showing the good preservation ability of okra seed powders. Total phenolic content and antioxidant activity were raised at higher milling frequency. Reconstitutability was significantly enhanced at higher milling frequency and/or for smaller median particle size. Thus, the successive milling and sieving process was successful in improving physicochemical and functional properties of okra seed powders, especially for smaller particles.

Graphical Abstract

Antioxidant activity, okra seed powder, particle size, polyphenols, powder reconstitution, preservation

Introduction

Okra fruit (Abelmoschus esculentus (L.) Moench) is widely consumed around the world and it is cultivated in tropical and subtropical regions (Torkpo et al., 2009; Jarret et al., 2011). Okra is renowned for its nutritional potential and health benefits: chemical composition analyses of okra revealed its richness in proteins, fat, minerals (Mn, Ca, Mg, Fe, Cu) and vitamins (B2, B3, B6, B9, C) (Moyin-Jesu et al., 2007; Xia et al., 2015). Mature okra seed is a good source of oil rich in unsaturated fatty acids such linolenic and linoleic acids, which are both essential for humans (Ndangui et al., 2010; Kumar et al., 2013). It has been reported in the literature that okra pods have high contents in polysaccharides and phenolic compounds including flavonoids (Arapitsas, 2008). Phenolic compounds are known to possess strong antioxidant and anti-stress effects (Ahmed et al., 2016; Sabitha et al., 2011). Although the high phytochemical content of okra pods has recently been proved (Hu et al., 2014; Xia et al., 2015, 2018), little is known about the distribution of bioactive compounds within granulometric classes of okra seeds powdered by milling.

Nowadays, milling is commonly used for the production of new food ingredients (Karam et al., 2016; Becker et al., 2017; Deli et al., 2019). Powdered plant parts are used as ingredients by many industries in the food, pharmaceutical, and cosmetics fields (Leuenberger and Lanz, 2005). Moreover, food ingredients and pharmaceutical excipients are often manufactured in powder form (Siliveru et al., 2017; Deli et al., 2019). Advantages of the powder technology reside in the convenience offered by powdered products, especially regarding higher shelf-life and easier transportability (Forny et al., 2011). Successive drying, milling and sieving processes have shown a great potential in the production of food powders with enhanced functional properties (Brewer et al., 2014; Becker et al., 2017; Deli et al., 2019; Waiss et al., 2020). It leads to the production of microparticles from plant material (whole plant, aerial parts, stems, roots, seeds and/or fruits). It is worth noting that drying, milling and sieving processes have a great impact on physicochemical and functional characteristics of produced powders (Landillon et al., 2008), such as particle size and shape distributions, surface roughness, density, cohesion, compressibility, water solubility, wettability, surface reactivity and hygroscopicity (Becker et al., 2017; Deli et al., 2019). A previous study performed with the same okra seed powders as in the current study showed a strong correlation between median particle size and proximate composition: smaller granulometric classes contained more proteins and fat, but less water, carbohydrates and minerals (Waiss et al., 2020). Also, the influence of milling frequency and median particle size of okra seed powder on flow properties was highlighted (Waiss et al., 2020): powders with large median particle sizes were not cohesive and had better flowability. Functional properties of powders (flowability, preservation ability, reconstitutability, etc.) are very important from an industrial point of view (Räsänen et al., 2003), as they determine the conditions by which they should be manufactured, handled, stored and reconstituted with a view to protect bioactive molecules (Deli et al., 2019a; Meghwal & Goswami, 2014; Petit et al., 2017). Powder physicochemical characteristics are often strongly influenced by particle size and physical structure (Meghwal & Goswami, 2014; Deli et al., 2019). Thus, it is of great importance to determine the physical and hydration characteristics of food powders in order to predict their behaviour during storage and reconstitution for an efficient use as functional food ingredients (Benković et al., 2017). The current study was intended to increase knowledge about the influence of particle size on phytochemical properties, reconstitutability and preservation ability of okra seeds powder. Okra seeds were ground at three milling frequencies (6000, 12 000 and 18 000 rpm); then, each powder sample was fractionated in four granulometric sizes by sieving: ≤180 μm, 180–315 µm, 315–500 µm and ≥500 µm. Last, physicochemical properties (particle size distribution and moisture content), ability to interact with water (water activity, water absorption capacity and water solubility index) and phytochemical properties (total phenolic content and antioxidant activity) of powders were evaluated.

Material and methods

Material: okra seeds

Okra seeds were purchased from farmers of Impfondo locality, in the north of the Republic of the Congo (1° 32′28″ N; 18°07′10″ E). Then, the okra seeds were removed from the pods and sun-dried in the open air at 30 °C. Dry seeds were stored at 10 °C until use for powder production.

Drying and milling of okra seeds

500 g dried okra seeds were milled with an ultra-centrifugal mill ZM 200 (Retsch, France) using a 10-cm sieve drilled with 1-mm trapezoidal holes at three milling frequencies (6000, 12 000 and 18 000 rpm). Then, obtained powder was collected in sealable polyethylene bags and stored at 10 °C until sieving.

Sieving of okra seed powders

Each okra seed powder issued from the milling step was separated into two aliquots. The first aliquot was immediately put in sealed polyethylene bags and stored at 10 °C until analysis; these powders were not sieved; thus, they will be named unsieved powders hereafter. The other aliquot was sieved with an Analyzer 3 Spartan sieve shaker (Fritsch, Idar-Oberstein, Germany) with three superimposed sieves of 180, 315 and 500 µm mesh sizes, at 0.5 mm vibration amplitude during 24 min. Hence, ≤180 μm, 180–315 µm, 315–500 µm and ≥500 µm granulometric classes were obtained by sieving. These granulometric classes were finally put in sealed polyethylene bags and stored at 10 °C until analyses.

Physical properties

Particle size distribution

Powder granulometric characteristics define the granular state of a powder. In the pharmaceutical industry, particle size analysis is of utmost importance as it partly conditions the bioactivity of medicines (Carmona et al., 2003). The particle size distribution of unsieved powders and granulometric classes of okra seed powders was determined by laser diffraction using dry dispersion (Mastersizer 3000, Malvern Instruments France, Orsay, France). Dispersion conditions for all powders except for the ≤180 µm powder samples were as follows: 100% air pressure at 4 bar, 3 mm hopper gap and 40% feed rate. For the ≤180 µm powder samples, the following dispersion conditions were employed: 100% air pressure at 4 bar, 4 mm hopper gap and 70% feed rate. The chosen size estimator was the equivalent diameter in volume. Classical granulometric parameters were determined: d10, d50 and d90, which correspond to the first decile, the median and the ninth decile of the particle size distribution and the span, which quantifies the width of the particle size distribution (eqn 1):

$$\text{Span}=\frac{\mathrm{d}90-\mathrm{d}10}{\mathrm{d}50}$$

(1)

Moisture content and water activity

Moisture content was measured by weight loss after drying 3 g powder sample at 103 °C for 5 h (AFNOR, Agence Française de Normalisation, 1976).

Water activity of okra powders was evaluated with a water activity meter (HygroPalm23-AW-A, Rotronic, France): 15 g powder was put in a WP-40 sample holder and the quick analysis mode was employed at 22 ± 3 °C, leading to water activity stabilisation in about 5 min.

Phytochemical properties

Preparation of hydromethanolic extracts

Hydromethanolic extracts of okra seed powders were prepared by maceration at room temperature (22 °C) in order to preserve the bioactivity of extracted molecules by minimising their thermal alteration (Becker et al., 2016; Ćujić et al., 2016). It is worth noting that this method permitted to obtain a solution of extractable phenolic compounds, but it was not able to extract chemically bound phenolic compounds. 2 g powder was put in 25 mL methanol-distilled water mixture (70/30% (v/v)) during 24 h under stirring at 300 rpm. Then, the resulting extract was centrifuged (Thermo Scientific, Heraeus Megafuge 8R Centrifuge) at 3 460 g for 20 min at 20 °C and the obtained supernatant was filtrated using Whatman filter paper n° 1 (Sigma Aldrich, France). The filter paper was washed with 20 mL methanol-distilled water mixture (70/30% (v/v)), and the filtrate was also centrifuged at 3460 g for 20 min at 20 °C. Last, the resulting supernatant was collected and stored at 4 °C until analysis.

Total phenolic content

In test tubes were added 30 µL okra powder hydromethanolic extract, 2.67 mL distilled water, 500 µL Folin–Ciocalteu reagent (1 N) and 300 µL sodium carbonate (Na₂CO₃, 20% (w/v)). The mixture was vortexed and dark-incubated during 30 min at room temperature (22 °C). Finally, absorbance of obtained solutions was measured at 750 nm with a Cary 50 Scan UV/visible spectrophotometer (Agilent, Santa Clara, CA, USA). Standard solutions of gallic acid (from 0.01 to 0.03 g L⁻¹ by 0.005 mg mL⁻¹ steps) were used for calibration (R² = 0.99), and total phenolic content was expressed in terms of milligrams of gallic acid equivalents per gram of dry matter (mg GAE/g DM).

Antioxidant activity evaluated by DPPH scavenging

The 1,1-diphenyl-2-picrylhydrazyl (DPPH) free radical scavenging method offers a first approach for evaluating the antioxidant potential of a compound, an extract or other biological sources (Kitagaki and Tsugawa, 1999). The DPPH^• scavenging method was performed according to the method of Masuda et al. (1999). 500 µL of 0.1 mM DPPH solution was mixed with 500 µL okra seed powder hydromethanolic extract. This mixture was vortexed, then dark-incubated at 30 °C for 30 min, and finally, its absorbance was measured at 517 nm against a blank solution composed of 500 µL DPPH solution mixed with 350 µL methanol and 150 µL distilled water that was left 30 min at rest. The antioxidant activity was evaluated as the percentage of DPPH radical inhibition and calculated from the absorbance decrease using (eqn 2) (Buijnsters et al., 2001):

$$\text{\% Inhibition of DPPH radical}=\left[\frac{\text{ABS}-\text{abs}}{\text{ABS}}\right]\times 100$$

(2)

with ABS (-) representing the absorbance of the blank solution and abs (-) the absorbance of the solution obtained from okra seeds powder extract after the DPPH assay.

Antioxidant activity evaluated by ABTS scavenging

ABTS^+• radical was produced by a reaction between an aqueous solution of ABTS at 7 mM and a potassium persulfate K₂S₂O₈ solution at 4.9 mM. This mixture was stirred for 16 h in the dark at room temperature (22 °C), leading to a blue green-coloured solution containing the ABTS^+• cation radical (Cai et al., 2015). Afterwards, the ABTS^+• solution was then diluted with ethanol until reaching between 0.65 and 0.70 absorbance at 734 nm. For analysis, 100 µL powder sample hydromethanolic extract was mixed with 900 µL ABTS^+• radical solution and the mixture was vortexed. After 10-min incubation at room temperature (22 °C), the absorbance of the mixture was measured at 734 nm against a blank solution (mixture of 100 µL phosphate buffer at pH 7.4 and 900 µL ABTS^+• radical solution). The results were expressed in terms of inhibition percentage according to (eqn 3):

$$\%\text{Inhibition of ABTS radical}=\left[\frac{A_0-A_1}{A_0}\right]\times 100$$

(3)

with A₀ (-) representing the absorbance of the blank solution and A₁ (-) the absorbance of the solution obtained from okra seeds powder extract after the ABTS^+• assay.

Hydration properties

Water absorption capacity (WAC) and water solubility index (WSI) were evaluated according to Deli et al. (2019). Water absorption capacity (WAC) was determined as follows: first, 1 g okra seeds powder and 10 mL distilled water were mixed in a centrifuge tube, then the mixture was stirred at 300 rpm (Variomag Poly) for 30 min at 20 °C and centrifuged (Thermo Scientific, Heraeus Megafuge 8R Centrifuge) at 1500 g for 30 min at 20 °C. The supernatant was carefully removed. The pellet was weighed, dried in an oven at 103 ± 2 °C for 6 h, cooled in a desiccator and weighed again. WAC was evaluated as follows (eqn 4):

$$\text{WAC}=\frac{\text{mass of wet sediments}}{\text{mass of dried sediments}}$$

(4)

Where masses of wet and dry sediments were expressed in grams.

The following protocol was used to determine the water solubility index of powder samples: 2.5 g (M₁) okra powder and 20 mL distilled water were mixed in a centrifuge tube, then the mixture was stirred at 300 rpm (Variomag Poly) for 30 min at 20 °C. After centrifugation (Thermo Scientific, Heraeus Megafuge 8R Centrifuge) at 1500 g for 20 min at 20 °C, the supernatant was carefully removed and the pellet was dried in an oven at 103 ± 2 °C for 6 h, cooled in a desiccator, and weighed (M₂). WSI was calculated according to eqn 5:

$$\text{WSI}\left(\%\right)=\left(\frac{M_1-M_2}{M_1}\right)\times 100$$

(5)

Statistical analysis

One-factor analysis of variance, the calculation of Pearson’s correlation coefficients, and principal components analysis (PCA) were performed by using Microsoft Excel 2016 and significant differences were assessed with Tukey's HSD at a 5% confidence level with the DSAASTAT add-on (Andrea ONOFRI, Department of Agricultural, Food and Environmental Sciences, University of Perugia, Italy). The determination of Pearson’s correlation coefficients and PCA were performed to examine correlations between characterised properties of okra seed powders: median particle size, span, total phenolic content, results of DPPH and ABTS assays, water activity, moisture content, WAC, and WSI. Results were expressed as means ± standard deviations.

Results and discussion

Particle size distribution

The analysis of the particle size distribution was performed in order to check whether the successive milling and sieving processes were efficient in micronising okra seeds and separating unsieved powders into sufficiently different granulometric classes. Table 1 shows that the median particle size decreased when the milling frequency was increased: for instance, for unsieved powders, median particle size was lowered value, from 723 µm at 6000 rpm to 478 µm at 12 000 rpm and 229 µm at 18 000 rpm. The median particle size of granulometric classes was found between the mesh sizes of their respective upper and lower sieves as expected when sieving is completed, confirming the efficiency of the sieving process to separate powders in granulometric classes differing in median particle size.

Table 1

Granulometric parameters of granulometric classes (≤180 µm, 180–315 µm, 315–500 µm, ≥500 µm) and unsieved samples of okra seed powders obtained at 6000, 12 000 and 18 000 rpm milling frequencies (Waiss et al., 2020)

Milling frequency (rpm)	Granulometric classes	d10 (µm)	d50 (µm)	d90 (µm)	Span (-)
6000	≤180 µm	23.1 ± 0.2ab	85.3 ± 0.6a	181.3 ± 0.5a	1.87 ± 0.01f
	180–315 µm	152.0 ± 1.0de	257.0 ± 1.0c	406.0 ± 1.0b	0.97 ± 0.01d
	315–500 µm	326.6 ± 6.0g	498.3 ± 2.5e	745.3 ± 2.5c	0.87 ± 0.03bcd
	≥500 µm	589.3 ± 18.5i	1100.0 ± 43.5h	2173.3 ± 73.7g	1.44 ± 0.02e
	Unsieved powder	141.0 ± 7.8d	723.3 ± 29.9f	1670.0 ± 65.5f	2.11 ± 0.02h
12 000	≤180 µm	21.9 ± 0.1ab	75.6 ± 2.8a	167.3 ± 6.3a	1.95 ± 0.06fg
	180–315 µm	165.6 ± 1.5e	264.6 ± 1.5c	409.3 ± 9.8b	0.92 ± 0.03cd
	315–500 µm	312.0 ± 1.0g	463.3 ± 1.1de	671.6 ± 4.9c	0.78 ± 0.01bc
	≥500 µm	542.6 ± 5.6h	801.3 ± 25.0g	1243.3 ± 75.0e	0.90 ± 0.05bcd
	Unsieved powder	52.9 ± 0.5c	457.6 ± 5.1de	983.0 ± 7.8d	2.03 ± 0.01gh
18 000	≤180 µm	17.8 ± 0.2a	52.1 ± 0.6a	149.3 ± 1.5a	0.77 ± 0.01b
	180–315 µm	37.4 ± 0.1bc	181.0 ± 0.0b	332.6 ± 2.0b	0.48 ± 0.01a
	315–500 µm	290.0 ± 4.3f	442.6 ± 12.8d	644.6 ± 25.9c	0.80 ± 0.02bc
	≥500 µm	527.6 ± 4.7h	781.0 ± 14.1g	1186.6 ± 37.8e	0.84 ± 0.02bcd
	Unsieved powder	25.7 ± 0.3ab	228.6 ± 7.5bc	702.67 ± 55.7c	2.96 ± 0.16i

Means ± standard deviations followed by the same superscripted letters in the same column were not significantly different at P < 0.05 according to Tukey's HSD test.

Span of unsieved powders obtained at 6000, 12 000 and 18 000 rpm were equal to 2.11 ± 0.02, 2.03 ± 0.01 and 2.96 ± 0.16, respectively, showing that the width of the particle size distribution was larger at higher milling frequency. This could be explained by the higher proportion of fine particles when using a higher milling frequency, leading to increased heterogeneity of unsieved powders.

Total phenolic content

Figure 1 presents the total phenolic content of unsieved samples and granulometric classes of okra seed powders obtained at the three different milling frequencies. Higher total phenolic contents were observed for samples composed of smaller particles. Moreover, a greater milling frequency induced an increase in total phenolic content: for example, for unsieved powders, the total phenolic content equalled 2.34, 3.44 and 7.40 mg GAE g⁻¹ DM at 6000 rpm, 12 000 and 18 000 rpm milling frequencies, respectively. This can be explained by the fact that a more intense milling (here at higher milling frequency, but this could also have been obtained with longer milling duration) enhances the size reduction in plant material, which further facilitates the accessibility of solvents for the extraction of bioactive compounds owing to the higher specific surface area of smaller particles (Rosa et al., 2013). This observation was confirmed by the moderate negative correlation between total phenolic content and median particle size d50 evidenced by a Pearson’s correlation coefficient of r = −0.71 (P < 0.05). Indeed, the bioactive compounds are often associated with proteic and fatty fractions in plant extracts (Becker et al., 2016; Deli et al., 2019). As indicated in the introduction, a previous study focussed on the influence of granulometric characteristics of these powders on their proximate composition revealed that higher protein and fat contents were found in the smaller particle size fractions of okra powder seeds (Waiss et al., 2020). Thus, the discriminating power of the sieving process in matter of total phenolic content and proximate composition was enhanced at higher milling frequency.

Figure 1

Total phenolic content of granulometric classes (≤180 µm; 180–315 µm; 315–500 µm; ≥500 µm) and unsieved samples of okra seed powders. DM, dry matter; GAE, gallic acid equivalents. Bars topped by different letters were statistically different at P < 0.05 according to Tukey's HSD test.

Antioxidant activity

The DPPH^• and ABTS^•+ assays are spectrophotometric techniques based on quenching of stable coloured radicals which are widely used for the assessment of antioxidant activities of complex biological mixtures such as plant or food extracts.

DPPH^• radical scavenging activity assay

Results of antioxidant activity of okra seed powders obtained by DPPH^• test are presented in Fig. 2. All extracts were able to scavenge DPPH^• radicals. The percentage of DPPH^• inhibition was higher for ≤180 µm and ≥500 µm granulometric classes whatever the milling frequency. The ≤180 µm granulometric classes were richer in lipids and proteins and the ≥500 µm granulometric classes contained more carbohydrates, as shown in a previous study performed with the same powders as investigated in the present study (Waiss et al., 2020). It has been showed that bioactive molecules such as polyphenols are often associated with lipids, proteins, and polysaccharides in plant tissues (Zaiter et al., 2016). This observation partially meets the results obtained in the present study, as at 6000 and 12 000 rpm milling frequencies, DPPH^• scavenging activity decreased with median particle size for the ≤180, 180–315, and 315–500 µm granulometric classes; however, at a given milling frequency, the DPPH^• scavenging activity of the ≥500 µm granulometric classes was greater than for the other granulometric classes and unsieved powders, indicating that bioactive compounds may be mainly associated with carbohydrates, likely fibres, in okra seeds. The increasing trend of DPPH^• scavenging activity with median particle size at 18 000 rpm confirms this interpretation, as a more intense milling may have permitted to break fibres and decrease their size, increasing the DPPH^• scavenging activity of the granulometric classes of higher median particle size. Besides, micronisation of plant material generally leads to an increase in the antioxidant activity of plant extracts (Rosa et al., 2013), and this was confirmed by the present study where milling at 18 000 rpm induced a marked increase in the DPPH^• inhibition percentage. Milling is known to alter the physical structure of the plant matrix, increasing its specific surface area and leading to the exposition of antioxidant compounds of the plant material (Hemery et al., 2010). Last, the fact that the DPPH^• scavenging activity was systematically higher for granulometric classes than for unsieved powders (Fig. 2) showed that hydromethanolic extraction was improved after sieving, maybe by decreasing powder cohesion and then facilitating its dispersion and the exposure of bioactive molecules to the solvent (Salameh et al., 2016).

Figure 2

Inhibition of DPPH radical by hydromethanolic extracts of granulometric classes (≤180 µm; 180–315 µm; 315–500 µm; ≥500 µm) and unsieved samples of okra seed powders. Bars topped by different letters were statistically different at P < 0.05 according to Tukey's HSD test.

ABTS cation radical scavenging activity assay

The results of ABTS^•+ radical scavenging activity of granulometric classes and unsieved samples of okra seed powders obtained at 6000, 12 000 and 18 000 rpm milling frequencies are reported in Fig. 3. High antioxidant activity of all okra seed powders was observed during the ABTS assay. No great difference of ABTS^•+ inhibition was denoted between samples, in particular, no significant difference could be attributed to milling frequency, but a trend of increased antioxidant activity for powder samples having a smaller median particle size was apparent. This was confirmed by the Pearson’s correlation coefficient between median particle size d50 and ABTS^•+ inhibition percentage of r = −0.64 (P < 0.05), which denoted a moderate negative correlation between these parameters. Besides, the antioxidant activity obtained by the ABTS cation radical scavenging was significantly higher than for DPPH radical scavenging for all granulometric classes and unsieved powders of okra seeds (cf. Figs 2 and 3). Indeed, previous studies showed that the ABTS test leads to higher inhibition percentage than the DPPH test for fruits, vegetables and beverages (Floegel et al., 2011), owing to the better reactivity of high-pigmented and hydrophilic antioxidants with ABTS^•+. Obtained results are consistent with the fact that antioxidant activity of plant products is generally attributed to radical scavenging activity of phenolic compounds (Rahman and Moon, 2007; Becker et al., 2016; Deli et al., 2019). Indeed, the phenolic compounds such as flavonoids, polyphenols and tannins are the main bioactive molecules that contribute to antioxidant activity of plant extracts (Rosa et al., 2013; Brewer et al., 2014; Becker et al., 2017).

Figure 3

Inhibition of ABTS cation radical by hydromethanolic extracts of granulometric classes (≤180 µm; 180–315 µm; 315–500 µm; ≥500 µm) and unsieved samples of okra seed powders. Bars topped by different letters were statistically different at P < 0.05 according to Tukey's HSD test.

Moisture content and water activity

Moisture content is an important parameter that is indirectly related to powdered food ingredient stability during storage. Low levels of moisture content (up to 12% (w/w)) are known to limit microbial growth and extend food shelf-life (Aryee et al., 2006; Kaur et al., 2013). Results displayed in Table 2 show that the moisture content of granulometric classes and unsieved samples of okra seed powders ranged from 8.61 ± 0.08 to 12.07 ± 0.05 (w/w), presuming a good stability during storage. The moisture content was generally decreased at smaller median particle size and/or at higher milling frequency (cf. Table 2), as evidenced by the moderate positive correlation between moisture content and median particle size (Pearson’s correlation coefficient of r = 0.73, P < 0.05). Moisture content is not sufficient to characterise food stability during storage, as food structure also contributes to water availability for microorganisms and alteration reactions: the concept of water activity has been developed to account for these combined influences of moisture content and food structure (Tapia et al., 2007; Bhandari et al., 2013). Water activity ranged between 0.493 ± 0.010 and 0.583 ± 0.010 for investigated okra seed powders (cf. Table 2), which was below the threshold of 0.650 permitting the growth of microorganisms like xerophilic mould and bacteria (Tapia et al., 2007). Water activity was generally decreased at lower median particle size, similarly to water content. This was expected, as a more intense milling leads to higher stress, more friction, and thus higher heating of the plant material, leading to smaller particles with a lower moisture content (Becker et al., 2017).

Table 2

Moisture content, water activity, water absorption capacity, and water solubility index of granulometric classes (≤180 µm, 180–315 µm, 315–500 µm, ≥500 µm) and unsieved samples of okra seed powders obtained at 6000, 12 000 and 18 000 rpm milling frequencies

Milling frequency (rpm)	Powder samples	Moisture content (g/g) (Waiss et al., 2020)	Water activity (-)	Water absorption capacity, WAC (g/g)	Water solubility index, WSI (g/100 g)
6000	≤180 µm	9.59 ± 0.12cd	0.555 ± 0.001ef	4.16 ± 0.46abc	32.71 ± 0.53f
	180–315 µm	9.01 ± 0.67ab	0.556 ± 0.002ef	6.46 ± 0.20e	24.06 ± 0.10cd
	315–500 µm	9.21 ± 0.11bc	0.540 ± 0.001cd	5.58 ± 0.25cde	22.74 ± 0.74bc
	≥500 µm	12.07 ± 0.05i	0.524 ± 0.003b	4.76 ± 0.24abcd	15.98 ± 0.23a
	Unsieved powder	11.10 ± 0.07fg	0.583 ± 0.007h	5.07 ± 0.75cde	22.75 ± 0.37bc
12 000	≤180 µm	9.15 ± 0.10abc	0.565 ± 0.008efg	4.09 ± 0.28ab	37.75 ± 0.46g
	180–315 µm	8.61 ± 0.08a	0.524 ± 0.002b	5.77 ± 0.41de	32.53 ± 1.06f
	315–500 µm	9.22 ± 0.05bc	0.503 ± 0.004a	5.12 ± 0.30bcde	25.05 ± 0.57d
	≥500 µm	11.79 ± 0.06hi	0.531 ± 0.004bc	4.98 ± 0.23bcde	17.64 ± 0.16a
	Unsieved powder	11.27 ± 0.08fgh	0.550 ± 0.001de	5.19 ± 0.52bcd	22.00 ± 0.78b
18 000	≤180 µm	9.56 ± 0.08bc	0.572 ± 0.002gh	3.52 ± 0.12a	41.12 ± 1.16h
	180–315 µm	10.15 ± 0.02de	0.564 ± 0.005fgh	4.06 ± 0.87abc	36.77 ± 0.66g
	315–500 µm	11.32 ± 0.06gh	0.560 ± 0.002efg	5.60 ± 0.50cde	23.50 ± 0.07bc
	≥500 µm	13.20 ± 0.17j	0.550 ± 0.010ef	4.67 ± 0.12abc	17.00 ± 0.14a
	Unsieved powder	10.69 ± 0.09ef	0.493 ± 0.009a	5.55 ± 0.55 bcde	30.58 ± 0.80e

Means ± standard deviations followed by the same superscripted letters in the same column were not significantly different at P < 0.05 according to Tukey's HSD test.

Water interactions

The water absorption capacity (WAC) and water solubility index (WSI) results of all granulometric classes and unsieved samples of okra seed powders are presented in Table 2. WSI significantly differed according to the milling frequency. When the milling frequency was increased, the median particle size of powders was decreased, enhancing the water solubility index. This was confirmed by the high negative correlation between WSI and median particle size (Pearson’s correlation coefficient of r = −0.89, P < 0.05), indicating that interaction with water was improved for smaller particles (cf. Table 2). This can be explained by the fact that milling induces particle size reduction, leading to the improvement of material functionalities such as particle surface specific area and surface energy, enhancing the interactions of plant material with water (Ye et al., 2016). In addition, this may result from the exposure of hydrophilic groups (coming from cellulose and hemicellulose) at the surface of particles obtained by plant micronisation (Zhao et al., 2009; Ting et al., 2014). Zhao et al., (2009) also reported increased dispersibility and solubility for ginger powders produced with a more intense milling, owing to improved interactions with water evidenced by a higher WSI (Zhao et al., 2009). Water absorption capacity represents the ability of a product to absorb water. Table 2 shows that all powders had a high WAC whatever the granulometric class and the milling frequency. This can be explained by the high proportion of mucilage of okra seeds that highly contribute to water absorption (Aboubakar et al., 2008). No clear influence of median particle size on WAC was denoted, except for a small trend of higher WAC for intermediate granulometric classes. Thus, performed milling was sufficient to break cell walls and expose hydrophilic groups of all okra seed components. The results of the current study are consistent with the work of (Deli et al., 2019), who reported that WAC and WSI were enhanced for smaller particles of Boscia senegalensis seeds, Dichostachys glomerata fruits and Hibiscus sabdariffa calyxes. Hence, it can be hypothesised that okra seed powders obtained at higher milling frequency should interact more easily with water and be more soluble, which can facilitate their intestinal absorption during digestion.

Principal component analysis of physicochemical, water interaction and phytochemical properties

PCA is used to highlight similarities or oppositions between variables and identify the variables that are most correlated between them. All granulometric classes and unsieved samples of okra seed powders obtained at the three milling frequencies were compared using PCA on the basis of their physicochemical (moisture content, median particle size), water interaction (water activity, water solubility index and water absorption capacity) and phytochemical (total phenolic content, and antioxidant activity deduced from DPPH and ABTS tests) properties. The result of this PCA is presented in Fig. 4. Variables are presented in red, and principal components are represented on F1 and F2 axes which, respectively, explained 43.4% and 20.7% variability among powder properties, allowing the PCA to explain a total variability of 64.1% which was sufficient to outline the major links between physicochemical and phytochemical properties of okra seed powders. F1 axis was positively correlated with ABTS^•+ inhibition, WSI and total phenolic content, and negatively correlated with median particle size and moisture content. F2 axis was positively correlated with span and WAC, and negatively correlated with DPPH^• inhibition and water activity. It is worth noting that median particle size was negatively correlated to total phenolic content, WSI and ABTS^•+ inhibition percentage, and positively to moisture content, confirming the previous discussion of experimental results.

Figure 4

Representation of all granulometric classes (≤180; 180–315; 315–500; ≥500 µm) and unsieved samples of okra seed powders obtained at 6000 rpm (green), 12 000 rpm (black) and 18 000 rpm (blue) milling frequencies and their physicochemical and functional characteristics (red) on the principal components F1 and F2 deduced from PCA.

The representation of okra powder samples on the F1 × F2 biplot revealed a separation of powder samples according to median particle size (Fig. 4). Smaller granulometric classes (≤180 µm at 6000, 12 000 and 18 000 rpm milling frequencies and 180–315 µm at 18 000 rpm milling frequency) were located at the bottom right of the F1 × F2 biplot and characterised by their high total phenolic content and WSI, and obviously by their low median particle size. Higher granulometric classes (≥500 µm at 6000, 12 000 and 18 000 rpm milling frequency and 315–500 µm at 18 000 rpm milling frequency) were found at the left of the F1 × F2 biplot and characterised by high moisture content and median particle size, and low total phenolic content and WSI. The other granulometric classes of intermediate median particle size and unsieved powders were located in the top middle of F1 × F2 biplot and they were positively correlated with span and WAC, and negatively to water activity and DPPH^• inhibition.

Conclusion

The combination of milling and sieving processes was successful in generating granulometric classes with significantly different physicochemical, water interaction and phytochemical properties. The results evidenced a considerable link between median particle size of okra seeds powder and its total phenolic content, antioxidant activity deduced from ABTS and DPPH tests, moisture content, water activity and water solubility index. Powders with smaller median particle sizes (i.e. lower size granulometric classes and/or powders obtained at higher milling frequency) had lower moisture content, water activity and water absorption capacity, and higher total phenolic content, antioxidant activity and water solubility index. This can result from the higher level of disintegration of plant matrix structure induced by a more intense milling (i.e. undergone by smaller particles and/or at higher milling frequency) which makes its compounds more accessible to solvents, thus improving phenolic compounds extraction and powder solubility, and decreases the number of pores of the solid matrix, thus lowering water absorption capacity. For all samples, moisture content and water activity were sufficiently low to avoid food spoilage. Hence, okra seeds powder was biochemically and microbiologically stable, ensuring a good shelf-life during storage. Water interaction parameters showed high levels of water absorption and solubility, which seems to indicate that the reconstitutability of okra seeds powder could be improved by the successive milling and sieving process. Finally, all obtained results suggest that okra seed powders can be used in food manufacturing as functional ingredients enhancing antioxidant activity and water interactions.

Acknowledgements

The authors would like to thank Dr. Florentin MICHAUX, Dr. Jennifer BURGAIN, Pr. Stephane DESOBRY and Dr. Jordane JASNIEWSKI for their scientific advices. Special thanks should be addressed to Carole PERROUD-THOMASSIN and Carole JEANDEL for technical advices regarding powder production and analyses. The authors also acknowledge support of the LIBio by the ‘Impact Biomolecules’ project of the ‘Lorraine Université d'Excellence’ (Investissements d’avenir ANR).

Author contributions

Idriss Miganeh Waiss: Data curation (lead); Methodology (lead); Writing-original draft (equal). André Kimbonguilla: Conceptualization (equal); Funding acquisition (equal); Validation (equal). Fatouma Mohamed Abdoul-Latif: Resources (equal). Laurette Brigelia Nkeletela: Software (equal). Joël Scher: Conceptualization (equal); Project administration (equal); Supervision (equal). Jeremy Petit: Conceptualization (equal); Project administration (equal); Methodology (equal); Supervision (equal); Validation (equal); Writing-original draft (equal); Writing-review & editing (equal).

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

No experiment was performed on humans or animals: no ethics approval was not required for this research.

Peer review

The peer review history for this article is available at https://publons.com/publon/10.1111/ijfs.15099.

Data availability statement

Research data are not shared.