Soil samples were collected from the 0-20 cm horizon using an auger, following the method described by Saba etal. [16], from soils in the Oubri region (SO) (Latitude: N 12˚39'20.67'', Longitude: W 1˚14'50.226'') and the Guiriko region (SG) (Latitude: N 11˚21'00.00'', Longitude: W 4˚41'90.00''). The Guiriko site corresponds to a cultivated experimental soil, whereas the Oubri site corresponds to a degraded and previously abandoned soil with low agricultural productivity. At each site, five experimental plots established for this study were selected and five subsamples were collected from each plot and combined to form one composite soil sample. Sampling was performed along two diagonals with a central point. The 0 - 20 cm depth was chosen because it corresponds to the main root zone of Hibiscussabdariffa and represents the soil layer most influenced by agricultural practices and nutrient availability. The collected samples were air-dried at room temperature for 7 days, gently crushed, and sieved before being packed in plastic bags and transported to the Laboratory for Soil Fertility Studies and Research for chemical analyses. Soil analyses were carried out over two consecutive years. In the first year, soils were characterized before sowing to determine their physicochemical properties and chemical fertility classes. Hibiscus sabdariffa was then cultivated, and the harvested calyces were analyzed for their phenolic and nutritional composition. In the second year, only soil analysis was performed after the rainy season to assess the evolution of soil physicochemical characteristics and chemical fertility.

The R72-1 variety of Hibiscussabdariffa was selected for this study. The seeds were obtained from the Institute of Environmental and Agricultural Research (INERA) in Farako-Bâ (Bobo-Dioulasso). Manual sowing was carried out from July 14 to 16, 2024. Four to five seeds were placed per hole, with spacing of 0.40 m between holes and 0.80 m between rows [13].

The harvesting of calyces was carried out in successive stages from November 15 to December 20, 2024. After harvest, the calyces were air-dried on racks for seven days, protected from direct sunlight and humidity. The dried calyces were then crushed, packed in zip-lock bags, and stored for subsequent analyses.

Soil pH was determined by direct reading using an electronic pH meter in a 1:2.5 soil-to-distilled water suspension for pH-H₂O and in a 1 M potassium chloride (KCl) solution for pH-KCl. Organic carbon (C-org) and total nitrogen (N-total) were measured by dry combustion after decarbonation using the Dumas method [17]. Total phosphorus (P-total) was quantified by automated colorimetry of the mineralized product after reaction with ammonium molybdate and ascorbic acid [18]. Available phosphorus (P-avail) was determined at 720 nm using a spectrophotometer after extraction with a mixed solution of ammonium chloride and hydrochloric acid followed by reaction with ammonium molybdate [19]. Total potassium (K-total) was measured using a flame photometer on the filtrate obtained after mineralization [20]. Available potassium (K-avail) and exchangeable bases (K⁺, Ca²⁺, Mg²⁺, and Na⁺) were extracted with 1 M ammonium acetate at pH 7. Calcium (Ca²⁺) and magnesium (Mg²⁺) were quantified by atomic absorption spectrophotometry, while K⁺, Na⁺, and K-avail were determined by flame photometry. Cation exchange capacity (CEC) was measured by percolating soil columns with 1 M ammonium acetate buffered at pH 7.

The assessment of soil fertility was based on the complete soil fertility scale described by Jongschaap, 1995 [21]. The classification was performed according to predefined threshold ranges for each parameter, allowing the soils to be categorized into different fertility classes.

Three types of extracts were prepared: a total phenolic-rich extract, an anthocyanin-rich extract, and an aqueous extract.

The total phenolic-rich extract (TPE) was prepared by macerating 100 g of plant material in 300 mL of an acetone/water/acetic acid mixture (70:29.5:0.5, v/v/v) for 72 h at 4˚C [22]. The resulting extract was then subjected to liquid-liquid partitioning with two 50 mL portions of chloroform to remove non-phenolic compounds [23].

For the anthocyanin-rich extract (ARE) and the aqueous extract (AE), 100 g of crushed dried calyx were macerated respectively in 300 mL of acidified methanol (1% HCl) and 300 mL of distilled water for 72 h at 4˚C, following one hour of ultrasonic sonication [24].

All procedures were performed in triplicate. The filtrates and fractions (TPE, ARE, and AE) were pooled and concentrated to dryness under reduced pressure (40 mbar) using a rotary evaporator and stored at 4˚C. The aqueous extract was subsequently lyophilized.

The different extraction yields were calculated using the following formula:

The anthocyanin extract was purified following the method described by Giusti etal. [25]. The crude extract was dissolved in acidified water (pH = 1), filtered, and loaded onto an Amberlite XAD-7 column (24 cm × 3 cm) packed with 200 g of resin. The column was first washed with 1000 mL of distilled water to remove sugars and other polar non-anthocyanin constituents. Elution was then carried out with 600 mL of methanol acidified with 1% HCl, and fractions of 200 mL were collected. The anthocyanin-enriched fractions were pooled, concentrated to a minimal volume, precipitated with ethyl acetate, and dried under vacuum in a desiccator.

The total phenolic content (TPC) of the extracts was determined using the Folin-Ciocalteu reagent method as described by Boizot etal. [26]. For each measurement, 60 µL of Folin-Ciocalteu reagent was added to 60 µL of diluted extract, followed by the addition of 120 µL of 7.5% sodium carbonate after 8 minutes. The mixture was then incubated at 37˚C for 30 minutes, and absorbance was measured at 760 nm. Total phenolic contents were calculated using a gallic acid calibration curve and expressed as mg GAE/g of plant material.

Total flavonoid content was assessed using the aluminum trichloride method described by Lebreton etal. [27]. A total of 50 µL of diluted extract was mixed with 150 µL of distilled water and 15 µL of 5% NaNO₂, followed by incubation for 5 min at room temperature. Then, 15 µL of 10% aluminum chloride was added and the mixture was incubated for 6 min, followed by the addition of 50 µL of 1 N NaOH. Absorbance was measured at 510 nm, and flavonoid contents were calculated using a quercetin calibration curve and expressed as mg QE/g of plant material.

Total anthocyanin content was measured using the pH-differential method according to Wrolstad etal. [24]. Two buffer systems (KCl, pH 1.0 and sodium acetate, pH 4.5) were used, and 0.5 mL of extract was mixed with 9.5 mL of the corresponding buffer and incubated for 15 min in the dark. Absorbance was measured at 510 nm and 700 nm, and the absorbance difference (A) was calculated between the two pH conditions. Monomeric anthocyanin content was determined using the following standard formula and expressed as mg/g:

where A is the absorbance difference, M W the molecular weight, D F the dilution factor, V the volume, ε the molar extinction coefficient, l the path length (cm), and m the sample mass (g).

Hydrolyzable tannins were quantified using the method of Mole and Waterman [28]. For each analysis, 40 µL of diluted extract was mixed with 200 µL of 2.5% KIO₃ solution and incubated for 2 min (extracts) or 4 min (tannic acid standard). Absorbance was then measured at 550 nm against a reagent blank. Hydrolyzable tannin content was determined using a tannic acid calibration curve and expressed as mg TAE/g of dry material.

Condensed tannins were determined according to the method proposed by Swain and Hillis [28]. For each analysis, 23 µL of diluted extract was mixed with 171 µL of 4% vanillin solution and 86 µL of concentrated HCl, then incubated for 20 min. Absorbance was measured at 500 nm, and results were calculated using a catechin calibration curve and expressed as mg CE/g of dry material.

Total protein content was measured by the Kjeldahl method as described by Saez-Plaza etal. [29], based on nitrogen quantification and subsequent conversion into protein content. The distillation step was carried out using a UDK 149 distillation unit.

Total sugars were determined using the phenol-sulfuric acid method described by R’Zina etal. [30]. To 50 µL of extract were added 50 µL of 5% phenol solution and 250 µL of concentrated sulfuric acid. After mixing, the reaction mixture was incubated for 20 min at room temperature, and the absorbance was measured at 490 nm. Results were expressed as mg of glucose equivalents per gram, based on a calibration curve established using glucose as the standard.

All absorbance readings were carried out using a Multiskan SkyHigh microplate spectrophotometer (Thermo Scientific, USA). All calibration curves used for quantification are provided in the Supplementary Materials (Figure S1toS5).

High-Performance Thin-Layer Chromatography (HPTLC) was performed using a CAMAG system. Solutions of the dry extracts were prepared at a concentration of 10 mg/mL and filtered through 0.22 µm syringe-compatible filters. An aliquot of 10 µL of each solution was automatically applied onto glass HPTLC plates (20 × 5 cm) coated with silica gel 60 F254 (MERCK KGaA, Germany) using a CAMAG applicator. Chromatographic development was carried out using a solvent system composed of ethyl acetate/formic acid/glacial acetic acid/water (100:11:11:26, v/v). Plates were visualized under visible light. No reference standards were used, as the objective of this analysis was not the formal identification of individual compounds but rather the assessment of the chromatographic profile. The number of detected bands and their corresponding Rf values were used to estimate the diversity of compounds present in the extracts [31].

Antioxidant activity of the extracts was evaluated using the DPPH, ABTS, and FRAP methods, as described by Lamien-Meda etal. [32], Miller etal. [33] and Pulido etal. [34], respectively. In these assays, Trolox was used as a reference antioxidant to establish calibration curves and express the antioxidant activity of the samples. Standard solutions of Trolox were prepared at concentrations ranging from 0 to 0.1 mg·mL⁻¹.

In addition, the radical-scavenging activity of the extracts was further assessed using the DPPH method described by Dosseh etal. [35]. For this purpose, Trolox and ascorbic acid were used as reference antioxidants, and their standard solutions were prepared over a concentration range of 0 - 25 µg·mL⁻¹. Similarly, extract solutions were prepared over concentration ranges of 0 - 50, 0 - 75, 0 - 100, or 0 - 150 µg·mL⁻¹, depending on the extract. These ranges were used to establish regression curves of absorbance versus concentration for both reference antioxidants and sample extracts. The reaction mixtures were incubated for 15 min at room temperature in the dark before measuring the absorbance.

IC₅₀ values were determined from the dose-response curves as the concentrations required to achieve 50% inhibition of the radical signal. The antioxidant activity index (AAI) was subsequently calculated to compare the radical-scavenging efficiency of the extracts. All measurements were performed in triplicate, and results were expressed as mean ± standard deviation. Statistical analyses were conducted to assess significant differences between samples (p < 0.05).

Based on the AAI values, classification was performed according to the scale proposed by Scherer etal. [36]:

AAI < 0.5 : low antioxidant activity 0.5 < AAI < 1 : moderate antioxidant activity 1 < AAI < 2 : strong antioxidant activity AAI > 2 : very strong antioxidant activity

The calibration curves of Trolox and ascorbic acid for the DPPH, ABTS, and FRAP assays, as well as the curves used for the determination of IC₅₀ values of Trolox, ascorbic acid, and the different extracts in the DPPH assay, are presented in the Supplementary Materials (Figure S6toS18).

Statistical analyses were performed using Statistica software version 12.3. Results are expressed as mean ± standard deviation (SD). Data were analyzed using one-way analysis of variance (ANOVA), followed by pairwise comparisons between groups. For soil parameters, comparisons were performed between soil types (Guiriko vs Oubri) within the same year, and between years for each soil type (year 1 vs year 2). For extract analyses, comparisons were systematically performed between samples from Guiriko and Oubri for each type of extract. Differences were considered statistically significant at p < 0.05. p-values were reported in the tables to facilitate interpretation of the observed differences.

Soil physicochemical parameters, as well as nutrient-related chemical properties, measured before sowing (year 1) and after the rainy season (year 2), are presented in Table 1 and Table 2.

Table 1 highlights the differences between the two soil types (Guiriko and Oubri) for each year. Overall, regardless of the year considered, Guiriko soils exhibited significantly higher nutrient contents than Oubri soils.

Table 2, on the other hand, illustrates the variations in physicochemical properties within the same soil type across the two years. The results indicate that, depending on the parameter considered, some variables showed significant differences between year 1 and year 2, whereas others did not vary significantly.

Table 1. Comparison of soil physicochemical parameters and nutrient-related properties between Guiriko and Oubri soils before sowing (year 1) and after the rainy season (year 2).

Rat: Rating; SG: soil from Guiriko region;SO: soil from Oubri region, OM: organic matter (1.724 × organic carbon), *p < 0.05: significant.

Table 2. Changes in soil physicochemical parameters and nutrient-related properties within each soil type (Guiriko and Oubri) between year 1 (before sowing) and year 2 (after the rainy season).

Rat: Rating; SG: soil from Guiriko region;SO: soil from Oubri region, OM: organic matter (1.724 × organic carbon), *p < 0.05: significant.

The extraction yields of the different extracts are presented in Figure 1. For all extracts, calyces harvested from Guiriko soils generally exhibited higher extraction yields than those obtained from Oubri soils.

Figure 1. Extraction yields of calyx extracts. (CG: calyces extract from the Guiriko region; CO: calyces extract from the Oubri region; AE: aqueous extract; TPE: total phenolic-rich extract; ARE: anthocyanin-rich extract; PA: purified anthocyanins. Horizontal lines indicate comparisons between groups, and asterisks (*) denote statistically significant differences between CG and CO for each extract (p < 0.05)).

The contents of phenolic compounds and nutrients in calyx extracts from Guiriko and Oubri soils are presented in Table 2. Guiriko samples had significantly higher levels of total phenolics, flavonoids, anthocyanins, proteins, and sugars than Oubri samples (p < 0.05). In contrast, condensed tannins and hydrolyzable tannins were significantly higher in Oubri samples.

Table 3. Phenolic compounds groups and nutrients contents of calyx extracts from Guiriko and Oubri soils.

CG: calyces extract from the Guiriko region; CO: calyces extract from the Oubri region; TPC(mgGAE/g): Total phenolics content (mg of gallic acid equivalents per g); TFC(mgQE/g): Total flavonoids content (mg of quercetin equivalents per g); TAC(mgECy/g): Total anthocyanins content (mg of cyanidin equivalents per g); CT(mgCAT/g): Condensed tannins (mg of catechin equivalents per g); HT(mgGAE/g): Hydrolyzable tannins (mg of gallic acid equivalents per g); TP: Total proteins (%); TS(mgGE/g): Total sugars (mg of glucose equivalents per g); Values are mean ± SD. * indicates significant differences between CG and CO (p < 0.05).

The chromatographic profiles obtained after purification of anthocyanin extracts from Hibiscussabdariffa calyces grown on Guiriko and Oubri soils are illustrated in Figure 2.

Figure 2. Chromatographic profiles of anthocyanin extracts from Hibiscussabdariffa calyces according to the soil. CG: calyces extract from the Guiriko region; CO: calyces extract from the Oubri region; Rf: retention factor.

Six similar chromatographic bands were observed in the purified anthocyanin extracts from Hibiscussabdariffa calyces. These bands appeared in samples from both Guiriko and Oubri soils, although the bands from the Oubri soil showed lower intensities.

The total antioxidant capacity of the calyx extracts, obtained from Guiriko and Oubri soils and evaluated using the DPPH, ABTS, and FRAP methods, is presented in Table 3.

Regardless of the method used, extracts from Guiriko soils consistently exhibited significantly higher antioxidant contents than those obtained from Oubri soils (p < 0.05).

Table 4. Total antioxidant capacity of calyx extracts from Guiriko and Oubri soils, determined by DPPH, ABTS, and FRAP methods.

CG: calyces extract from the Guiriko region; CO: calyces extract from the Oubri region; AE: aqueous extract; TPE: total phenolic-rich extract; ARE: anthocyanin-rich extract; PA: purified anthocyanins, TE: trolox equivalent; * indicates statistically significant differences between CG and CO samples (p < 0.05).

Figure 3 shows the antioxidant activities of calyx extracts from Guiriko and Oubri soils, along with standard antioxidants, expressed both as IC₅₀ values and as antioxidant activity indices (AAI).

A statistically significant difference was observed between the IC₅₀ and AAI values of CG and CO samples (p < 0.05), with CG samples exhibiting lower IC₅₀ values and consequently higher AAI values.

Figure 3. IC₅₀ and AAI of Hibiscussabdariffa calyx extracts and standards. CG: calyces extract from the Guiriko region; CO: calyces extract from the Oubri region; AE: aqueous extract; TPE: total phenolic-rich extract; ARE: anthocyanin-rich extract; PA: purified anthocyanins. Horizontal lines indicate comparisons between groups, and asterisks (*) denote statistically significant differences between CG and CO for each extract (p < 0.05).