Successive grinding and sieving as a new tool to fractionate polyphenols and antioxidants of plants powders: Application to Boscia senegalensis seeds, Dichrostachys glomerata fruits, and Hibiscus sabdariffa calyx powders

Abstract The present investigation aimed at evaluating the effect of powder fractionation based on particle size on the chemical composition in macronutrients, polyphenol contents, and antioxidant properties of powders of Boscia senegalensis seeds, Dichrostachys glomerata fruits, and Hibiscus sabdariffa calyces. Significant differences (p < 0.05) among granulometric classes of each plant were observed for the chemical composition in macronutrients. A decrease in particle size of plant powders was associated with an increase in ash, protein, and fat contents, while carbohydrate content was lowered. The following Granulometric classes, [0–180 µm] for Boscia senegalensis, [180–212 µm] for Dichrostachys glomerata, and [212–315 µm] for Hibiscus sabdariffa, respectively, were found to maximize total phenolic content and antioxidant activity. These results confirm that the grinding and controlled differential screening technology is an approach may serve as a useful guide to obtain optimum polyphenol extraction and enhance antioxidant activity of plant products.

currently brought up many criticisms (Palmade-Le Dantec & Picot, 2010;Baudelaire, 2013): Worries have been reported in relation to the dangerousness of many solvents used for extraction (e.g., dichloromethane, acetyl acetate, and toluene), their impact on atmosphere, environment, and human health, the costs associated with the treatment of generated toxic wastes, and the impact on extracts quality and safety (Palmade-Le Dantec & Picot, 2010).
Recently, alternation of drying and grinding process (ADG) and CDS extraction, that is, grinding and controlled differential screening, consisting in combinations of drying, grinding, and controlled sieving processes, has received increasing attention due to the raising desire to develop ecological extraction technologies of natural and active ingredients (Baudelaire, 2013;Becker et al., 2017Becker et al., ,2016Brewer, Kubola, Siriamornpun, Herald, & Shi, 2014;Karam, Petit, Zimmer, Djantou, & Scher, 2016;Li et al., 2015;Lucas-González, Viuda-Martos, Pérez-Álvarez, & Fernández-López, 2017;Zaiter, Becker, Baudelaire, & Dicko, 2018;Zaiter, Becker, Petit, et al., 2016). Indeed, the competitive advantage of plant powders in comparison with conventional bioactive molecules obtained from plants by solvent extraction mainly resides in the preservation of bioactive ingredients, and more especially of their bioactivity, for human interest. Becker et al. (2017) reported that the sieving process separates plant powders by granulometric differentiation through sieves of decreasing mesh, leading to selective distribution of bioactive molecules in the different granulometric fractions. Moreover, these authors explained that various parts of the same plant, more or less hard, more or less fibrous, are thus more or less difficult to crush, and thus, different plant parts may likely lead to particle fractions presenting different chemical compositions and/or structures. The sieving process leads to the separation of particles according to their sizes, thus leading to different physicochemical properties of resulting particle size classes (Guerrero-Beltrán, Jiménez-Munguía, Welti-Chanes, & Barbosa-Cánovas, 2009;Toth et al., 2005;Wang & Flores, 2000). The fine particles issued from the combination of drying, grinding, and sieving processes enable a better release of bioactive substances owing to their high specific surface area (Rosa, Barron, Gaiani, Dufour, & Micard, 2013;Zhao et al., 2018). Several recent studies have reported the link between particle size range of plant powders, bioactive ingredient contents, physicochemical properties, and functionalities (Becker et al., 2017;Sharma, Kadam, Chadha, Wilson, & Gupta, 2013;. With the increasing quantity and variety of powdered ingredients used in industry (Sharma et al., 2013), production of plant powders constitutes an important step for their valorization in various industrial sectors. Plant processing into powders not only allows the production of functionally adequate products, but also ensures their preservation during an extended shelf life, while supplying bioactive molecules under an adapted form for market of food supplements.
Africa is home to some of the most important species-rich biodiversity regions in the world (Linder, 2014). In several African countries, natural products constitute an important part of human diet and are also an excellent source of bioactive molecules. First plant of medicinal interest of the current study, Hibiscus sabdariffa, is an edible plant, for which previous studies on alcoholic and aqueous extracts from its calyx reported anti-inflammatory, antioxidant, hypolipidemic (Chung, Kong, Choi, & Kong, 2017;Medina-Carrillo et al., 2015), and anti-hypertensive effects (Abubakar, Ukwuani, & Mande, 2015), owing to its wealth of bioactive compounds (polyphenols, flavonoids, saponins, tannins, alkaloids, etc.). Then, it has also been showed that the aqueous, alcoholic, and hydroalcoholic extracts of the fruits of Dichrostachys glomerata, which is the second plant of medicinal interest of the study, exhibit antioxidant (Kuate, Etoundi, Soukontoua, Ngondi, & Oben, 2010), anti-hypertensive, hypoglycemic (Fankam, Kuete, Voukeng, Kuiate, & Pages, 2011), anti-inflammatory, and anti-hyperlipidemic activities (Kuate et al., 2013); these numerous biological activities were attributed to their contents in a broad range of bioactive molecules such as alkaloids, saponins, tannins, mucilage, glucocapparins, and sterols. Likewise, the seed extracts of the third plant of medicinal interest of the current study, Boscia senegalensis, are rich in saponins, tannins, anthraquinone, alkaloids, and flavonoids; it has been shown that they have anti-inflammatory, anti-hyperglycemic, and antioxidant properties (Dongmo, Dogmo, & Njintang, 2017).
This work aims at evaluating the effect of grinding and sieving on chemical composition in macronutrients, phytochemical contents, and antioxidant activity of powders of Boscia senegalensis seeds, Dichrostachys glomerata fruits, and Hibiscus sabdariffa calyces. It intended to find the granulometric fractions presenting the highest antioxidant activity for their use in nutraceutical formulations.

| Plant material
Samples were collected in May 2015 from different localities: Sundried red calyces of Hibiscus sabdariffa were purchased from local markets in the Adamawa region of Cameroon, sun-dried fruits of Dichrostachys glomerata were bought in a market located in Yaounde (Central Region, Cameroon), and dry fruits of Boscia senegalensis were purchased from rural farmers in a local market of Oum-Madjer (Batha state, Chad). The fruits of Dichrostachys glomerata and Hibiscus sabdariffa calyces were manually separated from inorganic materials, dirt, and dust particles before grinding. On other hand, the fruits of Boscia senegalensis were manually decorticated and the obtained seeds were taken into grinding.

| Plant grinding
Approximately 1 kg of B. senegalensis, 1 kg of D. glomerata, and 1.2 kg of H. sabdariffa were ground by 50 g batches at 6,200 g rotor speed and ambient temperature of about 20°C. This rotor speed was chosen as a compromise between grinding efficiency and local temperature increase in plant parts during grinding, as the latter is known to be enhanced at high rotor speed and lead to bioactive compounds alteration Zaiter, Becker, Petit, et al., 2016).

| Plant powder sieving
The sieving process is based on the separation of particles from a granular material by making them pass through several sieves of decreasing mesh size (315, 212, and 180 µm in the current study).
Basically, 100 g ground plant sample was sieved in permanent vibratory mode at 0.5 mm amplitude for 10 min. The fraction of the powder retained on each sieve was recovered and weighed for the calculation of the mass fraction of each granulometric class. A sample of unsieved plant powder was kept for comparison purposes.
Resulting plant powders were then put in sealed polyethylene plastic bags and stored at 10°C until analyses.

| Particle size distribution
Particle size distribution of each the powder classes and unsieved plant powders was determined by laser diffraction (Mastersizer 3000, Malvern Instruments France, Orsay, France) at ambient temperature, supplied with the Aero S dry dispersion unit that uses high-pressure air to disperse particles. The obscuration level was set lower than 2% to avoid multiple scattering by adjusting dispersion conditions at 1.5 bar air pressure: 30% air pressure, 30% feed rate, 2.5 mm hopper length for powder samples of Dichrostachys glomerata and Hibiscus sabdariffa; 70% air pressure, 70% feed rate, 2.5 mm hopper length for Boscia senegalensis powder samples.
Particle sizes were expressed in terms of equivalent spherical diameters in volume. Characteristic sizes of the particle size distribution, D10, D50, and D90, were measured, where DX means that X % of the volume of particles has a diameter inferior to DX. Also, the span, a common parameter related to the width of particle size distribution, was evaluated as follows:

| Macronutrient composition
Moisture content was determined by drying 5 g plant powder in an oven at 103°C during 24 hr until reaching constant weight, according to AOAC method 925.10 (AOAC, 1990). Total ash content was determined by incinerating from 3 to 5 g plant powder sample in a furnace at 550°C for 6 hr, then weighing the residue after cooling to room temperature in a desiccator, following the AOAC method 920.87 (AOAC, 1990). The crude nitrogen content was obtained using the Kjeldahl method by the AOAC method 991.20 (AOAC, 1990), and the protein content was deduced from it with a conversion factor of 6.25 (AACC, 1999). In the procedure of fat content determination [34], 8 g of powder samples was added to 30 ml of chloroform/methanol (2/1 [v/v]) and mixed for 20 min.
The mixture was filtered under dinitrogen, and the residue was re-extracted in 20 ml of the same solvent and filtered. The extracts were then mixed and allowed to separate after the addition of 0.02 ml of NaCl solution at 0.7 g/100 ml. The fat was recovered by rotary evaporation at 40°C under liquid nitrogen (Rotavapor R-144 Büchi, Flawil, Switzerland), and the fat contents were calculated by weight difference. Powder samples were hydrolyzed in 1.5 N sulfuric acid, and the available sugars were quantified by the phenol method (Dubois, Gilles, Hamilton, Ribers, & Smith, 1956).
The mixture was subjected to maceration by stirring at 300 rpm for 24 hr at room temperature (18 ± 2°C) and then filtered through with a Whatman filter paper (GE Healthcare companies, China) of 2-3 µm pore size. Thereafter, the supernatant was brought to 15 ml by addition of extraction solvent and stored at 4°C until analysis. This choice of extraction procedure (maceration under agitation) allowed shortening the process of extraction, thus minimizing the contact time of plant sample with solvent and better preserving the bioactivity of extracted molecules. Besides, the extraction at ambient temperature was also a compromise between extraction efficiency and limitation of thermal alteration of extracted biomolecules (Ćujíc et al., 2015).

Determination of total phenolic content
Total phenolic content of the hydromethanolic extract of plant powders was determined using the Folin-Ciocalteu method (Wafa, Amadou, Larbi, & Héla, 2014). Briefly, 20 µl of filtered extracts was diluted with 2,980 µl distilled water. Then, 500 µl of 10% (v/v) Folin-Ciocalteu reagent was added and the mixture was mixed. After 3 min, 400 µl of saturated solution of sodium carbonate Na 2 CO 3 (20% [w/v]) was added. After stirring, the tubes were placed at room temperature for 60 min and absorbance was measured at 760 nm using a spectrophotometer (Shimadzu UV-VIS 1605, Tokyo, Japan). A calibration curve (R 2 = 0.99) was prepared using standard solutions of gallic acid (40,80,120,160,200,240, and 280 g/L). Thus, total phenolic content was expressed as milligram gallic acid equivalents per gram dry weight (mg GAE/g DW) of plant powder.

Determination of flavonoid content
The total content in flavonoid compounds of the different samples was measured following the method of Dewanto, Wu, Adom, and Liu (2002). In brief, 0.1 ml diluted and filtered hydromethanolic extract was added to 2.4 ml of distilled water and 0.15 ml of 5% Na 2 NO 2 (w/v). After 6 min, 0.3 ml of 10% aluminum chloride (AlCl 3 ·6H 2 O) (w/v) was added. The mixture was kept at room temperature for 5 min, and 1 ml NaOH (1 M) was added. Absorbance was then measured at 510 nm by UV/visible spectrophotometry against the extraction solvent as blank. A calibration curve (R 2 = 0.99) was prepared using 20, 40, 80, 100, 120, and 140 g/L rutin as standards. The results were expressed in milligrams rutin equivalents per gram of dry weight (mg RE/g DW).

Determination of condensed tannins content
The content in condensed tannins of plant powders was evaluated using concentrated sulfuric acid to depolymerize tannins, which allowed them to react with vanillin and produce red anthocyanidins that were detected at 510 nm by UV/visible spectrophotometry (Sun, Ricardo-Da-Silva, & Spranger, 2008). 0.05 µl of diluted and filtered hydromethanolic extract was mixed with 3 ml of 4% vanillin (w/v), and 1.5 ml concentrated sulfuric acid was added. The mixtures were stirred and then kept at room temperature for 30 min.

| Determination of antioxidant activity
DPPH radical scavenging activity assay Antioxidant activity was first evaluated by the DPPH method (Zhang & Yasumori, 2004), in which the electron donating capacity of the extracts was measured by whitening of the purple-colored solution of 1,1-diphenyl-2-picrylhydrazyl (DPPH) cation radical.
This assay is based on the ability of antioxidants to scavenge the DPPH cation radical. Briefly, 2 ml of 0.1 mM DPPH methanolic solution was added to 0.5 ml hydromethanolic extract of plant sample at different concentrations (0.025, 0.05, 0.1, 0.5, 1, 5, 10, 100 mg/ ml). The mixture was thoroughly stirred and incubated in the dark for 1 hr at room temperature. After that, absorbance of the mixture was measured at 517 nm by UV/visible spectrophotometry. Lower absorbance of the reaction mixture indicated higher free radical scavenging activity. The scavenging activity was estimated based on the percentage of scavenged DPPH radical using the following: The antioxidant activity was expressed as the concentration required to cause 50% DPPH scavenging, referred as IC 50 (µg/ml). Ascorbic acid which was used as reference standard at the same concentrations as plant extracts showed an IC 50 value of 14.33 ± 0.58 µg/ml.

ABTS cation radical scavenging activity assay
The ABTS radical cation scavenging activity was measured according to the method described by Re et al. (1999) with slight modifications. In brief, ABTS solution was generated as follows: 6.62 mg of potassium persulfate and 38.4 mg of ABTS reagent were weighed in a glass beaker, 10 ml distilled water was added, and then the mixture was perfectly mixed. This solution was kept away from light and let stand for 16 hr at room temperature to yield a blue-green-colored solution containing the ABTS cation radical. Afterward, the ABTS +• solution was diluted with absolute ethanol until reaching an absorbance of 0.70 ± 0.22 at 734 nm. Free radical scavenging activity was assessed by mixing 150 µl of each test sample at 0.025, 0.05, 0.1, 0.5, 1, and 5, 10, 100 mg/ml with 2 ml ABTS +• radical solution. The decrease in absorbance was measured 1 min after mixing the solution. The scavenging activity was calculated according to Equation 2. The antioxidant activity was expressed as the concentration required to cause 50% ABTS inhibition, noted IC 50 (µg/ml). Ascorbic acid standard showed an IC 50 value of 12.33 ± 0.58 µg/ml.

| Statistical analysis
Results of chemical composition, phytochemical contents, and antioxidant activity were each subjected to analysis of variance (ANOVA) to determine the occurrence of statistically significant differences among them (p < 0.05). Duncan multiple range test was used to determine the degree of significance of the difference be-

| Particle size distribution
Particle size distributions of granulometric fractions and unsieved powders of Boscia senegalensis seeds, Dichrostachys glomerata fruits, and Hibiscus sabdariffa calyces are displayed in Figure 1. Unsieved powders of the three plants were bimodal, with the major population around 200 µm and the minor one around 20 µm for Dichrostachys glomerata and Hibiscus sabdariffa, and inversely for Boscia senegalensis. The lower particle size of Boscia senegalensis unsieved powder was also confirmed by its lower D50 of 28.60 ± 3.45 µm, compared to the D50 of Dichrostachys glomerata (166.00 ± 1.21 µm) and Hibiscus sabdariffa (126.00 ± 1.71 µm) unsieved powders.
In addition, the results of sieve fractionation and particle size characteristics of powder fractions (Table 1) showed that the grinding/sieving procedure was effective in separating Dichrostachys glomerata and Hibiscus sabdariffa powders into sufficiently different size classes. Indeed, D50 of granulometric fractions was comprised in or close to the sieve mesh range and span values were rather low (generally under 2), which is the sign of a relatively narrow particle size distribution and a quite homogeneous powder (Zhang, Xu, & Li, 2009). However, it is important to emphasize that during sieving analysis Boscia senegalensis was extremely sticky and cohesive, so it cannot be said that sieving analysis provided reliable results for Boscia senegalensis particle size. Sieving did not allow particles of Boscia senegalensis powder to be perfectly separated. Indeed, the high span values (over 4) of Boscia senegalensis fractions classes showed that they each were constituted of several particle size populations. Particle size distribution results of Boscia senegalensis powder samples were consistent with sieving issues encountered when processing this plant: Boscia senegalensis powder adhered to sieve walls and agglomerated upon vibrations, limiting its passage through sieve meshes, hence resulting in poorly efficient sieving difficulties.

| Biochemical composition
The proximate composition of fractions and unsieved powders is shown in Table 2. The moisture content of plants was ranging from 4.66% for Boscia senegalensis to 8.22% for Hibiscus sabdariffa, which makes them more stable during storage and packaging. Indeed, higher moisture content (generally over 10%) induces the development of microorganisms and product deterioration (Kaur, Kaushal, & Sandhu, 2011). No clear trend about the influence of particle size can be drawn from moisture results. A similar observation was reported by Becker et al. (2016): According to these authors, the heating effect of grinding, more pronounced for smaller particles and leading to moisture content decrease by heat-induced evaporation, could be compensated by the higher specific surface of small particles, facilitating the absorption of surrounding air humidity.
The results also showed that the chemical composition depends on particle size. A significant difference (p < 0.05) was observed between the protein content of the different fractions of Boscia senegalensis: Proteins were more concentrated in finer particles (<180 µm).
On the contrary, protein contents were similar or at least very close for the different powder fractions of the two other plants.
It is often observed that the smallest particles are richer in minerals, because fibrous plant parts, containing fewer minerals, are harder to grind, resulting in larger particles (Becker et al., 2017. Similar observations were made on Eucalyptus grandis powders. In this respect, fractions with smaller particle size were found to possess higher ash content and smaller fiber (hemicellulose and cellulose) content (Flávia, Edwil, & Fernando, 2016 Note. For each plant, means ± SD followed by the same superscripted letter were not significantly different (p < 0.05). It can be noted that the content in phenolic components considerably varied between unsieved plant powders and powder fractions of the same plant. The <180 µm granulometric fraction of Boscia senegalensis presented a higher total phenolic content. This can be related to the highest total protein and lipid contents of this particle size fraction. Indeed, large particles, richer in carbohydrates and probably in fibers (Oghbaei & Prakash, 2016), are expected to contain less bioactive compounds . of Hibiscus sabdariffa (Hakimeh, Shahryar, & Majid, 2016) and some other plants such as Halimium halimifolium (Rebaya et al., 2014) and

| DPPH radical scavenging activity
The results of DPPH radical scavenging activity presented in Table 4 were expressed in terms of concentration required to cause 50% DPPH inhibition (IC 50  respectively. This suggests that phenolic compounds significantly contributed to the antioxidant activity of investigated plant powders. This is consistent with the fact that antioxidant activity of plant products is generally attributed to radical scavenging activity of phenolic compounds such as flavonoids, polyphenols, and tannins (Rahman & Moon, 2007). The antioxidant activity of phenolic compounds is mainly due to their redox properties, which can play an important role in adsorbing and neutralizing free radicals, quenching singlet and triplet oxygen, or decomposing peroxides (Hasan et al., 2008). The high contents in bioactive compounds of studied plant powders are likely responsible for their significant antioxidant activity.  between the IC50 scavenging activities and the total phenol content.

| Principal Component Analysis of physicochemical and phytochemical properties of studied plant powders
In addition, the D50 was significantly correlated negatively with protein (r = 0.61; p < 0.01) and fat (r = −0.76; p < 0.001) meaning that as the particle size increased, the protein and fat contents decreased.
We can also observe in Figure 2 the opposition between the total carbohydrate and the phenol contents. In this respect, antioxidant activities of the powders were increased by the total phenol, flavonoid, TA B L E 3 Contents in total phenols, flavonoids, and condensed tannins in hydromethanolic extracts of the different granulometric fractions and unsieved powders of Boscia senegalensis seeds, Dichrostachys glomerata fruits, and Hibiscus sabdariffa calyces and tannin contents while the protein and fat contents tended to alter them. These observations were in concordance with previous findings (Marhuenda et al., 2016;Singh, Singh, Ashish, & Salim, 2013).  between. This representation revealed the general tendency of increasing phenols content (more pronounced in Dichrostachys glomerata sample) as the particle size of the powder fraction decreased.

| CON CLUS ION
The aim of this study was to determine the granulometric class of

ACK N OWLED G M ENTS
This study is part of the project "Equipe de recherche" financed by the Agence Universitaire de la Francophonie (AUF, Francophone University Association) we wished to thank. The project awarded the first author DELI Markusse, a scholarship facilitating its studies at the LIBio (Biomolecular Engineering Laboratory, University of Lorraine, France) as well as at the LABBAN (Laboratory of Biophysics, Food Biochemistry and Nutrition of Ngaoundéré University, Cameroon).
The Extrapole consortium, funded by the former Lorraine region (France), and more especially its project leader Elie DJANTOU BAUDELAIRE are also thanked for the initiative of the study presented in this manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N
NYN, EBN, and JS have made substantial contributions to conception and design, while MD, TJNM, and JP contributed to acquisition, analysis, and interpretation of data, and drafting the manuscript. All the authors critically revised and approved the final submitted version of the manuscript. Prior to submitting the article, all authors agreed on the order in which their names are listed in the manuscript.

E TH I C A L S TATEM ENT
This study does not involve any human nor animal testing.