Influence of Harvesting Stages on Phytonutrients and Antioxidant Properties of Leaves of Five Purple-Fleshed Sweet Potato (Ipomoea batatas) Genotypes

Sweet potatoes (Ipomoea batatas) are highly profitable, contribute to food security, and their leaves rich in phytonutrients. This study examined the optimal leaf harvesting stage by harvesting newly formed leaves (leaves 1 to 5) to achieve the highest concentration of carotenoids, phenolic compounds, antioxidant properties and mineral content. Leaves of five purple-fleshed sweet potato genotypes ‘2019-11-2’ and ‘2019-1-1’, ‘Purple-purple’, and from the USA ‘08-21P’ and ‘16-283P’ were harvested based on tuber life cycle [vegetative 8 weeks after planting (VS-8WAP), tuber initiation (TIS-12WAP), and tuber maturation phases (TMS-16WAP)]. At the 8WAP stage, leaves of genotype ‘2019-11-2’ had the highest concentrations of cyanidin-caffeoyl-sophoroside-glucoside (17.64 mg/kg), cyanidin-caffeoyl-feruloyl-sophoroside-glucoside (41.51 mg/kg), peonidin-caffeoyl-hydroxybenzoyl-sophoriside-glucoside (45.25 mg/kg), and peonidin caffeoyl-feruloyl-sophoriside-glucoside (24.47 mg/kg), as well as antioxidant scavenging activity. In contrast, ‘Purple-purple’ harvested at TIS-12WAP showed the highest concentration of caffeoylquinic acid derivatives. Zeaxanthin, lutein, all trans-β-carotene, and cis-β-carotene are the most abundant carotenoids in genotype ‘08-21P’ at VS-8WAP. As a result, local genotypes ‘2019-11-2’ harvested at 8WAP and ‘Purple-purple’ harvested at 12WAP are potential sources of anthocyanins and caffeoylquinic acid derivatives. Conversely, USA’s genotype ‘08-21P’ at the VS-8WAP stage is an excellent source of carotenoids. The leaves of USA’s ‘08-21P’ genotype and the local ‘2019-11-2’ genotype at TMS-16WAP exhibited the highest content of Fe and Mn, respectively. The study identified the optimal leaf stage for consumption of leaves and for use as a functional ingredient.


Introduction
Sweet potato (Ipomoea batatas), belonging to the family Convolvulaceae, could be a valuable food for coping with future changes in food supply and demand in developing countries [1].Apart from tubers, sweet potato leaves (SPLs) have the potential to serve as a natural dietary source with the possibility of being developed further as a sustainable crop for multiple uses in the food and medicinal industries [2].Leafy vegetable crops are an affordable source of bioactive compounds and antioxidants that can potentially combat hidden hunger in rural communities.These antioxidants prevent or inhibit other molecules during oxidation [3].Research investigations are focusing on identifying fruits and vegetables that are rich in essential phytochemicals and have nutritional benefits in preventing diseases and promoting good health by maintaining a healthy and balanced Foods 2024, 13, 1640 2 of 26 diet [4].One such vegetable that is gaining popularity in developing countries is SPLs.In Taiwan, purple SPLs are widely grown as a healthy vegetable [5].They are a leafy, edible vegetable that is nutritious and readily available throughout the season due to their dual-purpose cropping approach.Spinach, kale, mustard, lettuce, and traditional leafy vegetables like Amaranthus, spider plant, black nightshade, cucurbits, and cowpea leaves are among the most consumed worldwide [6].
Recently, there has been a growing interest in purple-fleshed sweet potatoes because they are rich in anthocyanins.Anthocyanins are a type of phenolic compound that offer many health benefits, and they can be found in the flesh, skin, and leaves of sweet potatoes [2,7].Researchers have identified 15 different types of anthocyanins in sweet potatoes, including peonidin 3-caffeoyl-p-coumaryl sophoroside-5-glucoside and a pelargonidin type, which were recently discovered [8,9].Interestingly, the number of anthocyanins in sweet potato cultivars varies, which is why some parts of the plant have a purple pigmentation.Although there is no official RDA for anthocyanins, China recommends 50 mg/day [10].SPLs are an excellent source of essential vitamins, minerals, and carotenoids, which are crucial for human health.According to studies conducted by Nguyen et al. [2] and Zhang et al. [11], SPLs have higher levels of these compounds compared to other leafy vegetables and even the roots of the plant.Additionally, SPLs can be harvested multiple times, until prior to the harvesting of tubers, making them a competitive option for leafy vegetable crops, as noted by Chirwa-Moonga et al. [12].Furthermore, SPL has a higher polyphenol content compared to spinach [13].The SPLs are available and ready to be harvested throughout the growing season, at multiple times within short intervals, as short as weekly harvesting [14].But harvesting too frequently may reduce the leaf's nutritional content and retard tuber growth when grown for dual purposes [15].Gakige et al. stated that the high total biomass could be obtained by harvesting SPL partially after 75 days [16].Therefore, the harvesting interval plays a role in the yield of harvestable leaves.Furthermore, SPL must be harvested along with young apical buds to supplement nutrients in addition to those of leaves [17].This is ideal because the apical buds showed the highest level of TPC and a more potent antioxidant activities (AA) than young tender leaves [18].In recent years, sweet potato breeding has focused on developing cultivars with nutritional benefits, high yield, long storability, and resistance to pests and diseases.While the leaves are identified as suitable for consumption.Purple-fleshed sweet potatoes are being introduced locally and might as well be promoted for dual-purpose use.Developing a niche market for the leaves of these genotypes is ideal as it will improve the cultivation of purple-fleshed sweet potato tubers, thus increasing the overall utilization of sweet potatoes.Currently, the Agriculture Research Council in South Africa has new breeding lines of purple genotypes as well as imported lines under trial for possible commercialisation, and there is little information available on the phytonutritional properties of these purple-fleshed genotypes.Phenolic contents differ with the harvesting stages but to fully understand this trend, the translocation of nutrients to which storage sink (organ) during the stage at which leaves were harvested must be taken into consideration.

Experimental Site
The trial was planted at the Agricultural Research Council for Vegetable, Industrial and Medicinal Plants known as ARC-VIMP located along Moloto/Kwa Mhlanga Road in Roodeplaat outside Pretoria in Gauteng (GPS coordinates: 25 ′ 56 • S; 28 ′ 35 • E), 1164 m above sea level [19].According to Köppen classification, the area is categorised as a humid subtropical (Cwa) region due to its warm temperate climate with a clay loam type of soil [20].Climatic data were collected by the WT1080 professional weather station throughout each growing season.External environmental factors such as temperature (minimum and maximum), rainfall, and humidity (%) from January to May in 2022 (Season 1, S1) and 2023 (Season 2, S2) were recorded during the growth period (Figure S2A,B of the Supplementary File).The PAR (µmol/m 2 s) on the day of harvest was also recorded using an ACCUPAR LP-80 (PAR/LAI) ceptometer (Decagon devices, Washington, DC, USA).

Cultivation
Five purple-fleshed genotypes were used as treatments, including 'Purple-purple' (local), '08-21P', and '16-283P' imported from the USA, as well as two ARC breeding lines, '2019-11-2' and '2019-1-1' (Figure S1 of Supplementary File).Cuttings that were approximately 30 cm long with 3 or 4 nodes were taken from mature vines and prepared before planting.Fertilisation and field preparation were conducted in accordance with the ARC guidelines for sweet potato production in South Africa [21].The field was prepared with ridges 1 m spacing apart and 0.3-0.4m high.Plants were spaced at 0.3 m between the plants on the ridge [19].The 2022 trial was planted at land no.41 N with a total of five blocks each with a population of 20 plants per treatment (5 genotypes) without replications.In 2023, after the generation advanced in the breeding scheme, the second trial was planted in land no. 30, which is adjacent to land no.41.Treatments (genotypes) were replicated in triplicate in a randomised complete block design with each block consisting of 20 plants per replicate.
Table S1 of Supplementary File shows the properties and nutritional status of the soil in plot no.41 and plot no.30.Based on the soil analysis and fertiliser recommendation report, 500 kg/ha of 1:0:1 (37%) plus 200 kg/ha of superphosphate (12%) of fertiliser was applied at planting.After the planting of the sweet potato cuttings, the top dressing was applied as 150 kg/ha of LAN (28).Twenty-one days after planting, a further 150 kg of LAN (28) plus 250 kg of potassium chloride per hectare were applied as top dressing.Irrigation was scheduled twice a week during the first six weeks after planting and then reduced to once a week after establishment.Irrigation was applied for 2 h with sprinkler irrigation supplying approximately 25-30 mm of water weekly, depending on the rainfall.To protect the plants from pests, Decis ® (Bayer CropScience, Maharashtra, India) was sprayed, and Cymoxazeb ® (Villa Crop Protection, Kempton Park, Gauteng, South Africa) was applied to protect the plants from fungal diseases.

Harvesting of Leaves and Preparation of Samples
The leaves were harvested three times over two growing seasons in 2022 and 2023 from 8 weeks after planting (8WAP) to 16 weeks after planting (16WAP) at four-week intervals.Week 8 occurred during the vegetative stage, week 12 at tuber initiation, and week 16 during the tuber maturation stage of sweet potato plant growth stages.SPLs were harvested by randomly hand-picking a maximum of five newly formed leaves.The leaves were from leaf developmental position 1 to 5 on the vine growth tips with each harvesting stage (Figure 1). Figure 1 also shows the colour development on the leaf blades of purple-fleshed genotypes within the first five harvestable leaves on vine tips.A set of 500 g per genotype were harvested and transported at 15 • C to the laboratory.Leaves were then sorted by removing petioles, discarding damaged leaves, and washed with running water to remove soil.After freezing at −80 • C, the leaf blades were freeze dried using a SP scientific virtis benchtop pro freeze dryer (BTP-9SGE0X, NY, USA) at −55 • C, at a total pressure of 2 × 10 + 3 mbar, for 7 days then ground into powder.The powders were used for all biochemical analysis.and Cymoxazeb ® (Villa Crop Protection, Kempton Park, Gauteng, South Africa plied to protect the plants from fungal diseases.

Harvesting of Leaves and Preparation of Samples
The leaves were harvested three times over two growing seasons in 2022 from 8 weeks after planting (8WAP) to 16 weeks after planting (16WAP) at four tervals.Week 8 occurred during the vegetative stage, week 12 at tuber initiation, 16 during the tuber maturation stage of sweet potato plant growth stages.SPLs vested by randomly hand-picking a maximum of five newly formed leaves.T were from leaf developmental position 1 to 5 on the vine growth tips with each h stage (Figure 1). Figure 1 also shows the colour development on the leaf blades fleshed genotypes within the first five harvestable leaves on vine tips.A set of genotype were harvested and transported at 15 °C to the laboratory.Leaves w sorted by removing petioles, discarding damaged leaves, and washed with runn to remove soil.After freezing at −80 °C, the leaf blades were freeze dried using a tific virtis benchtop pro freeze dryer (BTP-9SGE0X, NY, USA) at −55 °C, at a tota of 2 × 10 + 3 mbar, for 7 days then ground into powder.The powders were us biochemical analysis.

Morphological Analysis
The sweet potato's morphological characteristics were identified using inte descriptors as outlined by Huamán [22].Several phenotypic traits, such as le number of lobes, and abaxial vein pigmentation were recorded, along with th and immature leaf colours.

Analysis 2.5.1. Morphological Analysis
The sweet potato's morphological characteristics were identified using international descriptors as outlined by Huamán [22].Several phenotypic traits, such as leaf shape, number of lobes, and abaxial vein pigmentation were recorded, along with the mature and immature leaf colours.
Foods 2024, 13, 1640 5 of 26 2.5.2.Colour and Chlorophyll CR-400 chroma meter (Konica Minolta Sensing, Osaka, Japan) was used to measure the colour.A white plate was used to calibrate the Chroma meter before use [23].In each genotype, 10 readings were taken from different locations on the abaxial and adaxial leaf surfaces.Three sets of data were averaged and used as a representative sample.The colour value was recorded using the L*, a*, and b*.L* represents luminosity ranging from 0 to 100 (the highest value means the lighter the leaves); a* value ranges from −128 to 127 representing greenness to redness; and b* represents yellowness to blueness with values ranging from −128 to 127 [24].
A portable soil plant analysis development (SPAD) chlorophyll meter (SPAD-502 plus, Konica Minolta Sensing, Osaka, Japan) determined the chlorophyll content in SPLs non-destructively by recording the greenness from various leaf points.Before use, the SPAD meter was calibrated by clamping the measuring head without a sample.The measurements were replicated three times taking 10 repeated measures, and averages were used as a representative of three sets of data.The results were recorded in µg/cm 2 units.

Leaf Area Index (LAI)
The LAI was measured using the ACCUPAR LP-80 (PAR/LAI) ceptometer (Decagon devices, Washington, DC, USA).Ten repetitive measures were taken above and below the leaf surface with three replicates to obtain an LAI.

Mineral Analysis
The leaf samples were tested at ARC-Vegetable, Industrial and Medicinal Plants analytical facility in Rustenburg (ARC-VIMP).Micro Kjeldahl acid digestion process was used for N, P, and K 2 and dry-ashing and further digestion with Aqua Regia for Ca, Mg, and Micronutrient.

Extraction for Total Phenol Content (TPC) and Antioxidant Activities Assays
Ten mg of freeze-dried samples were weighed and mixed thoroughly with 2 mL of 80% methanol using a vortex mixer (BV1000, Benchmark Scientific Inc., New Jersey, USA) for 60 s, then sonicated for 10 min at 35 • C in a DC-150H MRC ultrasonic cleaner [23].After that, the sample was centrifuged for 20 min at 4427× g with a Hermle centrifuge (Z326k, Hermle Labortechnik GmbH, Wehingen, Germany).Extraction was repeated three times, and the resulting supernatants were pooled together into 1.Analysis of TPC, FRAP, ABTS, and DPPH was carried out using the collected supernatants.

TPC
The TPC was carried out following the method described by Phahlane et al. [25] using Folin-Cioclateu reagent with minor changes.A total of 0.5 mL of diluted extract (100 µg/mL) was mixed together with 2.5 mL of Folin-Ciocalteu reagent (diluted 10 times with water), and 2 mL of Na 2 CO 3 (75 g/L).The mixture was allowed to stand for 2 h at room temperature and absorbance was measured through a UV-Vis Spectrophotometer (SPECTROstar ® Nano BMG Labtech, Ortenberg, Germany) at 765 nm.The results were expressed as chlorogenic acid equivalent (CAE) mg/g Dw (Dry weight of leaf extract) based on the calibration curve (Y = 0.0034x + 0.0569, µg/mL, R 2 = 0.99, linear range: 0-100 µg/mL).Chlorogenic acid is the major phenolic acid in sweet potatoes.

Ferric Reducing Antioxidant Power (FRAP)
FRAP value was determined as described by Phahlane et al. [25] without any changes.The FRAP reagent was prepared by mixing acetic acid buffer pH 3.6, TPTZ solution (10 mmol/L), and FeCl 3 (20 mmol/L) solution according to the ratio (10:1:1).Then the mixture was incubated at 37 • C. Within 1-2 h, 240 µL of FRAP reagent was mixed with 10 µL of leaf extract (Section 2.5.5)into several wells on a microplate reader.The mixture was left a room temperature in darkness to react for 5 min before spectrophotometer Foods 2024, 13, 1640 6 of 26 (BMG LABTECH GmbH, Spectro Star Nano, Ortenberg, Germany) readings, measuring the absorbance at 593 nm.The results were calculated against the calibration curve Y = 0.0004x + 0.0825 (µg/mL, R 2 = 0.97, linear range: 0-100 µmol/mL) of the TROLOX standard and expressed as TEAC mM/g dw (dry weight), which was calculated as follows: where c is the sample concentration (µmol/mL) calculated against the TROLOX standard curve, V is the volume of the sample (mL), DF is the dilution factor, and m is the weight of the sample dry matter (g).
2.5.8.2,2-Azino-bis-3-ethylbenzothiazoline-6-sulfonic (ABTS) Scavenging Activity SPL scavenging activity with ABTS assay was carried out according to the procedure described by Phahlane et al. [25].Prior to use, 5 mL of ABTS stock solution was added to 45 mL of phosphate buffer.Serial dilutions from the sample extract (Section 2.5.5) were made and transferred into wells in a microplate reader and thereafter incubated for 30 min in the dark to allow a reaction with 200 µL of ABTS stock solution.The absorbance readings were obtained at 750 nm in a spectrophotometer and the results were recorded as IC 50 mg/mL Dw.IC 50 is the concentration of extracts required to scavenge 50% of the free radicals where a lower value shows higher radical scavenging efficiency [26].The formula used to calculate the % inhibition is = (A0 − A1) × 100/A0 whereby A0 and A1 represent the ABTS assay absorbance and samples absorbance, respectively.The % inhibitions (y) were plotted against the sample concentrations (x) at all points, and the respective regression line (y = ax + b) was drawn.The IC 50 value was calculated by substituting the value of Y with 50 in the regression equation.
2.5.9.2,2-Diphenyl-1-picrylhydrazyl (DPPH) Radical Scavenging Activity A DPPH assay was carried out using a DPPH solution according to Phahlane et al. [25].Briefly, 1.03 mg of DPPH was weighed and dissolved in 31.25 mL of absolute methanol to prepare the DPPH solution.An additional 15.625 mL of methanol diluted the DPPH solution to give the initial absorbance of 0.493 at 517 nm in a spectrophotometer.Thereafter, different concentrations (0-10 mg/mL) from the sample extract (Section 2.5.5) were made and 100 µL of each were transferred into a microplate reader wells (BMG LABTECH GmbH, Spectro Star Nano, Ortenberg, Germany) and incubated in darkness to allow a reaction with 200 µL of DPPH for 30 min before reading the absorbance.The formula used to calculate the % inhibition is = (A0 − A1) × 100/A0, whereby A0 and A1 represent the DPPH absorbance and sample absorbance, respectively [27].The % inhibitions (y) were plotted against the sample concentrations (x) at all points, and the respective regression line (y = ax + b) was drawn.The IC 50 value was calculated by substituting the value of Y with 50 in the regression equation.

Total Carotenoids
Total carotenoids were measured following the process performed by Phahlane et al. [25].For carotenoids extraction, 100 mg of freeze-dried sweet potato leaves were homogenised with 5 mL of 95% ethanol containing 0.1% (w/v) butylated hydroxytoluene (BHT) for 10 min.The mixture was saponified using KOH solution (20% in methanol, w/v) for 30 min under constant agitation.Carotenoids were then extracted three times with 3 mL of a mixture of (70:30, v/v) hexane/dichloromethane containing 0.1% BHT.A solution of 10% NaCl (w/v) was used for phase separation and then the mixture was centrifuged at 3900× g for 5 min.The upper layer was collected, combined, and the solvent was evaporated under nitrogen stream to dryness.A total of 2 mL of methanol/MTBE (50:50, v/v) prepared with 0.1% BHT was used to redissolve the carotenoids after drying under a nitrogen stream.A standard stock solution of β-carotene was prepared by dissolving 1 mg in 10 mL of methanol/MTBE.Standard solutions with concentrations of 0 to 0.1 mg/mL were prepared from serial dilution of stock solution.The absorbance of reconstituted extracts and standards were read at 460 nm using a spectrophotometer and the results were calculated against the β-carotene standard calibration curve Y = 0.0206x + 0.0318 (R 2 = 0.99, linear range: 0-100 µg/mL) on a dry weight basis in mg/g.2.5.11.Quantification of Phenolic and Anthocyanin Compounds Using UPLC-QTOF/MS A Waters Cyclic Quadrupole time-of-flight (qTOF) mass spectrometer (MS) connected to a Waters Acquity ultra-performance liquid chromatograph (UPLC) (Waters, Milford, MA, USA) was used for high-resolution UPLC-MS analysis.Column eluate first passed through a Photodiode Array (PDA) detector before going to the mass spectrometer, allowing simultaneous collection of UV and MS spectra.Electrospray ionization was used in positive mode with a cone voltage of 15 V, desolvation temperature of 275 • C, desolvation gas at 650 L/h, and the rest of the MS settings optimized for best resolution and sensitivity.Data were acquired by scanning from m/z 100 to 1500 m/z in resolution mode as well as in MSE mode.In MSE mode two channels of MS data were acquired, one at a low collision energy (4 V) and the second using a collision energy ramp (40-100 V) to obtain fragmentation data as well.Leucine enkaphalin was used as lock mass (reference mass) for accurate mass determination and the instrument was calibrated with sodium formate.Separation was achieved on a Waters HSS T3, 2.1 × 150 mm, 1.7 µm column.An injection volume of 0.5 µL was used and the mobile phase consisted of 0.1% formic acid (solvent A) and acetonitrile containing 0.1% formic acid as solvent B. The gradient started at 100% solvent A for 1 min and changed to 28% B over 22 min in a linear way.It then went to 40% B over 50 s and a wash step of 1.5 min at 100% B, followed by re-equilibration to initial conditions for 4 min.The flow rate was 0.3 mL/min, and the column temperature was maintained at 60 • C. Compounds were quantified in a relative manner against a calibration curve established by injecting a range of neochlorogenic acid and rutin standards from 0.2 to 5 mg/L rutin.After compression, centroiding and application of lock mass correction, the data were processed using MSDIAL and MSFINDER (RIKEN Center for Sustainable Resource Science: Metabolome Informatics Research Team, Kanagawa, Japan) [28,29].Databases searched by MSFinder: Functions 1 (unfragmented channel) and 2 (fragmented channel) of the Waters MSe data were processed by MSDial to produce MS1 and MS2 spectra as well as extracted ion chromatograms with associated peak height intensity data.Since calibration standards were not available for the majority of these compounds, the peak height intensity was converted to concentration in a semi-quantitative manner by interpolation off a calibration curve for neochlorogenic acid acquired under the same instrumental conditions.
Each deconvoluted feature (alignment in MSDial), together with its associated MS1 and MS spectra was exported from MSDial to MSFinder.Based on the accurate mass elemental compositions, possible compounds were identified from the listed databases and then subjected to in silico fragmentation.According to the spectral match between the in silico and measured spectra, a score (out of 10) is assigned to each of the possible compound matches with the highest score being accepted as the most likely (assuming a score of at least 4).The content of phenolic and anthocyanin compounds in the study is expressed as mg/kg.2.5.12.Quantification of Carotenoid Components Using HPLC-PDA After drying under a nitrogen stream, sample extracts were reconstituted in 0.6 mL of methanol/MTBE (50:50 v/v) prepared with 0.1% BHT and injected (10 µL) into an HPLC (Shimadzu Prominence-i HPLC System with DAD Detector, Shimadzu, Kyoto, Japan) system [25].The chromatographic separation was achieved on a YCM C30 column (3.6 × 250 mm, 3.6 µm) (Waters, Milford, MA, USA) kept at a consistent temperature (25 • C), with eluent A and B (solutions of 0.1% formic acid in methanol and 0.1% formic acid in methyl tert-butyl ether, respectively).The linear gradient was 0 min 0% B, 1 min 5% B, 37 min 50% B, and 38 min 0% B. Pure external standards (zeaxanthin, cis-β-carotene, lutein, all trans-β-carotene, and zeaxanthin) were used to identify and quantify carotenoids in the sample extract.The chromatograms were recorded at 460 nm.The calibration curves were obtained using different concentrations of the standards (0-0.1 mg/mL).Each carotenoid is presented as mg/100 g dry weight.

Statistical Analysis
The leaves were harvested three times in one season over two years.To see the significant difference between the harvesting stage and genotypes, a set of three data values was analysed statistically with two-way ANOVA using GenStat (64-bit Release 22.1, PC windows 10, VSN International Ltd., Hertfordshire, UK) and the results were presented as mean and standard deviation (n = 3).Tukey's HSD was adopted to differentiate the means between the treatments at a 5% significance level.

Morphological Characterisation
Table S2 of Supplementary File shows the variation in leaves among five purplefleshed sweet potato genotypes.The evaluation of leaf outline, leaf lobe type, leaf lobe number, and abaxial leaf vein was conducted to determine the key quality traits that explain the variation between sweet potato cultivars [19].Among the five genotypes, '08-21P' and '2019-11-2' have the same leaf outline but differ in leaf lobe type and number.The genotype '08-21P' can be distinguished from '2019-11-2' based on the number of young purple leaves (1 up to 3) in the vine tip (Figure 1).The occurrence of purple leaves in the vine tips is referred to as a red fading process that protects young leaves and prevents the limitation of photosynthetic capacity at a later stage [30].The genotypes 'Purple-purple', '16-283P', and '2019-1-1' had green leaves from leaves 1 to 5. Leaf developmental stages 1-3 showed slightly purple edges on genotype 'Purple-purple' and visible purple pigmentation on genotype '16-283P's' leaf blade (Figure 1).Genotype 'Purple-purple' had a slightly lobed leaf type and genotype '2019-1-1' had a very slightly lobed leaf type, whereas genotype '16-283P' differed completely from the other genotypes with a coordinate leaf outline.On the abaxial side, only 'Purple-purple' did not display purple veins.

Leaf Colour
Colour indicates the presence of primary pigmentation in plants, namely carotenoids (Yellow and orange), anthocyanins (Purple to blue), and chlorophyll (Green) [34].Table S3 of Supplementary File shows that for L* all main and interaction effects were highly significant (p < 0.001).L* represents the luminosity in leaves and the darker the leaves the higher the accumulation of dark pigments (anthocyanins).In Supplementary Table S5A (19.34), and '16-283P' x VS-8WAP (18.00).During the leaf harvesting stages, there was a noticeable trend (TMS-16WAP > TIS-12WAP > VS-8WAP) in the L* value on the adaxial side of the leaf.A similar trend can be seen in Figure 1 for the abaxial leaf surface as well as the adaxial leaf surface; '08-21P' x VS-12WAP (37.66), '08-21P' x TIS-16WAP (37.17), and 'Purple-purple' x TMS-16WAP (37.21) were lighter.The darker abaxial leaf L* colour was obtained in '16-283P' x VS-8WAP (25.6).

TPC
TPC was significantly influenced by genotype x harvesting stage x season interaction at p < 0.001 and all the main effects (Table S6 of Supplementary File).Table 1 shows '2019-11-2' x TIS-12WAP with the highest TPC (262.2 chlorogenic acid equivalent (CAE) mg/100 g dw), and the lowest TPC (85.0 CAE mg/g dw) was obtained in sweet potato leaves (SPLs) harvested at TMS-16WAP in genotype '2019-1-1'.The highest leaf TPC was obtained in a purple leaf genotype '2019-11-2' followed by '08-21P', (Table 1), these genotypes have hastate leaf outlines (Figure 1).According to the metabolic profiling by Tan et al. [38], high phenolics were correlated with lobed leaf types, this is consistent with the findings in this present study.Makori et al. [8] and Jia et al. [18] reported high TPC in green arial parts which contradicts the findings of this study.Tan et al. [38] showed a correlation between the leaf colour with flavonoids.
The trend was observed as follows: TIS-12WAP > VS-8WAP > TMS-16WA.SPLs reached their maximum growth peak during the vegetative and tuber initiation stage; thereafter, they ceased to grow and started shedding.The phenolics are secondary metabolites produced in the leaf blades, there is a probability that synthesis of these metabolites reaches the maximum point during these stages, and this could explain the variation in the accumulation of TPC during VS-8WAP and TIS-12WAP.TPC is also induced by temperature [2].Motsa et al. [4] indicated that favourable environmental conditions allow the plant to store phytochemicals in harvestable parts.On the other hand, elevated temperatures weaken the assimilation process which results in low polyphenol content [39].This is also evident from our study, the second-season (S2 2023) leaf samples showed lower TPC.On the contrary, Suárez et al. [40] obtained high TPC during the last harvest (H3).This could be explained by plant maturity stage, leaf developmental stage, and environmental conditions.The effect of the harvesting stage on the TPC was reported by Takács-Hájos and Vargas-Rubóczki [39], where beetroot leaves harvested 60 days after planting had higher polyphenols than leaves harvested after 85 days.Krochmal-Marczak et al. [41] reported different concentrations of phenolic compounds in leaves of sweet potato cultivars (Okinawa, Carmen Rubin, Radiosa, White Triumph, Molokai, Purple, Beauregard, Jewel, and Satsumo Imo) grown in Poland, harvested during three different growth stages.The results ranged between 85.0 to 262.2 CAE mg/g dw (Table 1).Sun et al. [42] evaluated 40 different sweet potato cultivars with TPC values ranging between 2730 and 12,460 (equivalent to 27.3-126.4)chlorogenic acid (CHA) mg/100 g dw.Similar findings were reported by Makori et al. [8] in SPLs of four cultivars harvested in Beijing China, the TPC was 7 130-12 080 (equivalent to 71.3-120.8mg/g dw) CAE mg/100 g dw.Our results are comparably higher, and this could be due to the genotype, harvesting stage, and environmental conditions [41].Data are means and standard deviation (n = 3).Small letters within the same column represent the significant differences between genotypes and harvesting stages.S1-season 1, S2-season 2, VS-8WAP-vegetative stage 8 weeks after planting, TIS-12WAP-tuber initiation stage 12 weeks after planting, TMS-16WAP-tuber maturation stage 16 weeks after planting.

UHPLC-QTOF-MS Identification and Characterisation of Phenolic Compounds
Table 1 shows that TPC in S1 > S2.This suggest that high temperatures and light intensity in S2 affected the accumulation of phenolic compounds compared to S1 (Supplementary Figure S2).Anthocyanins, phenolic acids, and flavonoids are phenolic compounds, hence S1 data were presented due to moderate temperatures.Moderate temperatures (20-25 • C) influenced the synthesis of anthocyanins composition [43].According to Sasaki et al. [44] the content of caffeic acid (CA) and caffeoylquinic acid (CQA) were higher in SPLs grown under 20-25 • C than in warm conditions (30 • C).
Across all samples, the peonidin/cyanidin ratio exceeded 3.5; however, cyanidincaffeoyl-feruloyl-sophoroside-glucoside, peonidin-caffeoyl-hydroxybenzoyl-sophorisideglucoside, and peonidin caffeoyl-feruloyl-sophoroside-glucoside were not detected in '16-283p' or 'Purple-purple' at any of the three different stages of harvesting (VS-8WAP, TIS-12WAP, and TMS-16WAP).Peonidin feruloyl-sophoroside-glucoside was present in all genotypes and at different leaf harvesting stages, including VS-8WAP, TIS-12WAP, and TMS-16WAP.Among all other genotypes and harvesting stages, 'Purple-purple' x TIS-12WAP exhibited the highest concentration of Peonidin feruloyl-sophoroside-glucoside.All five anthocyanin derivatives were detected in '2019-11-2' at leaf harvesting related to VS-8WAP and TIS-12WAP.Moreover, cyanidin-caffeoyl-sophoroside-glucoside, cyanidin-caffeoylferuloyl-sophoroside-glucoside, peonidin-caffeoyl-hydroxybenzoyl-sophoriside-glucoside, and peonidin-caffeoyl-feruloyl-sophoroside-glucoside had the highest concentrations in '2019-11-2' x VS-8WAP.Additionally, the levels of cyanidin-caffeoyl-sophoroside-glucoside detected in '2019-11-2' at TIS-12WAP were comparable to those found at VS-8WAP.As per Zhang et al. [47], there is a positive correlation between pigment sedimentation and anthocyanin biosynthesis genes in the leaves of purple sweet potato leaves.The location of IbMYB1 genes plays a crucial role in determining the colour variation in shoots [48].Furthermore, the expression of the IbMYB1 gene changes significantly during leaf development [47].It is likely that the expression of this gene is affected by the growth cycle of the tubers and the number of weeks after planting.However, additional research is needed to confirm this.Our study showed that the concentrations of various anthocyanin components in the leaves of purple-fleshed sweet potato are related to growth cycle of the tubers and number of weeks since planting.Mild temperature (20 • C) and prolonged exposure to the sun enhanced the synthesis of anthocyanins in SPLs [43].Moderate temperatures during TIS-12WAP induced the production of anthocyanins.Anthocyanins also provide photoprotection to leaf tissues which may have resulted in a high accumulation of cyanidin-caffeoyl-sophoroside-glucoside, cyanidincaffeoyl-feruloyl-sophoroside-glucoside, peonidin-caffeoyl-hydroxybenzoyl-sophorisideglucoside, and peonidin-caffeoyl-feruloyl-sophoroside-glucoside during VS-8WAP stage in genotype '2019-11-2'.Anthocyanins function as light filters during light stress [49].Data are means and standard deviation (n = 3).Means followed by the same letter within the row are not significantly different (p < 0.05), each of the samples was replicated three times, and the results are expressed as mean ± standard deviation.nd (No anthocyanin detected).VS-8WAP-vegetative stage 8 weeks after planting, TIS-12WAP-tuber initiation stage 12 weeks after planting, TMS-16WAP-tuber maturation stage 16 weeks after planting.
The purple sweet potato leaves from Poland contain 7 phenolic compounds, including 5CQA, 3CQA, 4CQA, 3,4-diCQA, 3,5-diCQA, quercetin-3-O-galactoside, and quercetin-3-O-glucoside.Among these, 3CQA was found to be the compound with the highest concentration at 13,720 mg/kg [41].On the other hand, the leaves of deep purple sweet potatoes from Korea had 4,5-diCQA and 3,5-diCQA as the compounds with the highest concentration (860 and 840 mg/kg, respectively) [50].Additionally, a purple-fleshed genotype cultivar from China had the highest levels of chlorogenic acid (9850 mg/kg) and 3,5-diCQA (22,670 mg/kg) compared to 18 other cultivars from China and one USA cultivar grown in China [51].According to our study, the purple sweet potatoes developed through Agriculture Research Council South Africa's breeding program '2019-1-1' harvested at the TIS-12WAP stage (1225.95mg/kg), '2019-11-2' harvested at the VS-8WAP stage (1337.75mg/kg), and the local 'Purple-purple' at the TIS-12WAP stage (3436.47mg/kg) showed lower 3CQA levels than those from Poland but higher than the Chinese and USA genotypes [51].It is worth noting that the phenolic compound profiles and concentrations in Purple sweet potato leaves vary depending on the cultivars and the region of cultivation [51].Data are means and standard deviation (n = 3).Small letters within the same row represent the significant differences between genotypes and harvesting stages.VS-8WAP-vegetative stage 8 weeks after planting, TIS-12WAP-tuber initiation stage 12 weeks after planting, TMS-16WAP-tuber maturation stage 16 weeks after planting.Means followed by the same letter within the row are not significantly different (p < 0.05), each of the samples was replicated three times, and the results are expressed as mean ± standard deviation.

Heat Map
A heat map was created based on the concentrations of phenolic compounds in all the samples.A colour block in each row represented each compound data point, with red boxes equivalent to higher levels and blue boxes indicating lower levels (Figure 2).Additionally, the heat map expressed the composition of phenolic compounds in the purple-fleshed sweet potato leaves relating to the tuber life cycle and number of days after planting of five genotypes.According to the heat map, the concentration of 1,3-diCQA and quercetin 3-sophoroside-7-rhamnoside was higher in '2019-1-1' x TIS-12WAP, quercetin 3-glucoside was higher in '16-283P' x TMS-16WAP.The highest concentrations of peonidin-caffeoylferuloyl-sophoriside-glucoside, peonidin-caffeoyl-hydroxybenzoyl-sophoriside-glucoside, cyanidin-caffeoyl-feruloyl-sophoroside-glucoside, and 3-O-caffeoyl-4-O-methylquinic acid were found in '2019-11-2' x VS-8WAP.

Heat Map
A heat map was created based on the concentrations of phenolic compoun the samples.A colour block in each row represented each compound data point, boxes equivalent to higher levels and blue boxes indicating lower levels (Figure tionally, the heat map expressed the composition of phenolic compounds in th fleshed sweet potato leaves relating to the tuber life cycle and number of days af ing of five genotypes.According to the heat map, the concentration of 1,3-diC quercetin 3-sophoroside-7-rhamnoside was higher in '2019-1-1' x TIS-12WAP, 3-glucoside was higher in '16-283P' x TMS-16WAP.The highest concentratio onidin-caffeoyl-feruloyl-sophoriside-glucoside, peonidin-caffeoyl-hydroxybe phoriside-glucoside, cyanidin-caffeoyl-feruloyl-sophoroside-glucoside, and 3-O 4-O-methylquinic acid were found in '2019-11-2' x VS-8WAP.

Total Carotenoids (TC) and Carotenoid Components
The study found that genotype '16-283P' x VS-8WAP had the highest lev

Total Carotenoids (TC) and Carotenoid Components
The study found that genotype '16-283P' x VS-8WAP had the highest level of TC, with a value of 7.69 mg/g dw, while genotype '2019-11-2' TMS-16WAP had the lowest value at 2.75 mg/g dw.Most of the genotypes showed a similar trend of VS-8WAP > TMS-16WAP > TIS-12WAP, except for 'Purple-purple', as shown in Table 1.According to Hossain et al. [52], carotenoid concentration in SPLs increased from 30 to 90 days after planting and decreased from 120 to 180 days after planting.In higher plants, carotenoids and Chl b, as well as Chl a-protein complexes, are known to be involved in light-harvesting structures.Chlorophyll pigments are also protected by carotenoids from sunlight damage [53].Furthermore, LAI was significantly reduced at TMS-16WAP, indicating that stress from shading or a large number of leaves could result in secondary metabolite synthesis.In stressed African eggplant leaves, however, carotenoids were significantly reduced, as shown by Mibei et al. [54].
The trend observed with individual carotenoids follows that of TC.The data presented are from S1 due to moderate temperature, although S2 had higher TC compared to S1. Elevated temperatures in S2 increased the carotenoids in plants to cope with climatic stress.Also, this suggest that VS-8WAP, which occurred in March with high temperature compared to April and May during TIS-12WAP and TMS-16WAP, respectively, in both seasons will favour the synthesis of carotenoids.Lutein, zeaxanthin, all-trans beta carotene, and cis beta carotene were identified and separated, as shown in Table 4. Ooko Abong et al. [55] also found the same compounds, while Drapal et al. [56] additionally identified violaxanthin.The levels of individual carotenoids were influenced by the genotype x harvesting stage and main effects (p < 0.001).The highest lutein content (128.35mg/100 g dw), all-trans β-carotene (60.20 mg/100 g dw), zeaxanthin (46.83 mg/100 g dw), and cis-β-carotene (9.20 mg/100g dw) were found in genotype '08-21P' x VS-8WAP, while the least were obtained in genotype '2019-1-1' x TMS-16WAP (52.43 mg/100 g dw), (9.11 mg/100 g dw), (9.11 mg/100 g dw) and (3.07 mg/100 g dw) of lutein, all trans-β-carotene, zeaxanthin and cis-β-carotene, respectively, as presented in Table 4.
The accumulation of carotenoid-derived metabolites can depend on the environment, affecting plant growth, development, and signalling processes [57].Therefore, a high accumulation was observed during VS-8WAP.In grapevine, β-carotene increased with an increase in temperature, while lutein showed the opposite trend.Similarly, kale and spinach showed high concentrations of lutein and β-carotene at temperatures under 30 • C and 10 • C, respectively.In this study, β-carotene was high during the vegetative stage and lower in the last harvest with lower temperatures [58].
Foods 2024, 13,1640 Table 4.The level of individual carotenoids isolated in sweet potato leaves harvested during three different stages.

Cis-β-carotene mg/100 g dw
Data are means and standard deviation (n = 3).Small letters within the same column represent the significant differences between genotypes and harvesting stages.*** represent significant levels at p < 0.001.VS-8WAPvegetative stage 8 weeks after planting, TIS-12WAP-tuber initiation stage 12 weeks after planting, TMS-16WAPtuber maturation stage 16 weeks after planting.

Antioxidant Activities
There was a significant effect observed when genotype, harvesting stage, and season were compared (p < 0.001, Table S6 of Supplementary File).The total phenolic content (TPC) and ferric-reducing antioxidant power (FRAP) showed a similar trend.Among all genotypes, the best antioxidant power (FRAP) value (53.92 TEAC mM/g dw) was obtained in TIS-12WAP of '2019-11-2', while the lowest value (15.17 TEAC mM/g dw) was observed in TMS-16WAP of '2019-1-1' (Table 5).The trend observed was TIS-12WAP > VS-8WAP > TMS-16WAP in all genotypes.However, in a study conducted by Suárez et al. [40], high FRAP was obtained in fresh leaves of cultivar Shangshu no.19 harvested at HP3 in China, which could be attributed to the environmental conditions and the stage at which the leaves were harvested.Data are means and standard deviation (n = 3).Small letters within the same column represent the significant differences between genotypes and harvesting stages.S1-season 1, S2-season 2, VS-8WAP-vegetative stage 8 weeks after planting, TIS-12WAP-tuber initiation stage 12 weeks after planting, TMS-16WAP-tuber maturation stage 16 weeks after planting.
Table 5 shows the ABTS scavenging activity of SPLs of five purple-fleshed genotypes.Genotype, harvesting stage, and season significantly influenced the antioxidant activity with ABTS assay in SPLs (p < 0.001) (Table S6 of Supplementary File).The genotype '2019-11-2' at VS-8WAP stage exhibited the highest ABTS scavenging activity (0.20 IC 50 mg/mL), which was not significantly different from that of '2019-11-2' x TIS-12WAP.According to Jia et al. [18], the strength of AA increases as the IC 50 value decreases.The genotype '2019-1-1' x TMS-16WAP exhibited a weaker ABTS with an IC 50 value of 1.31 mg/mL.Phahlane et al. [25] reported ABTS values ranging between 3.43 and 4.60 mg/mL in the leaves of orange flesh sweet potato local cultivars weaker compared to the present study.It is possible that the variation can be attributed to the harvesting stage, genotype, and environmental conditions.The IC 50 of ABTS was higher in SPLs harvested during TIS-12WAP and lower during the TMS-16WAP.Additionally, it was significantly higher in the genotype '2019-11-2' with purple-coloured leaves.
An important observation was made regarding the effect of genotype, harvesting stage, and season (p < 0.001), as well as the main effects, as shown in Table S6 of Supplementary File.According to Table 5, '2019-11-2' demonstrated a strong ability to scavenge DPPH radicals in SPLs harvested during VS-8WAP (0.66 IC 50 mg/mL) and TIS-12-WAP (0.58 IC 50 mg/mL) during both seasons, with no significant difference between the two.This value is comparable to the IC 50 reported by Zhang et al. [11], which was 0.758 mg/mL.In the current study, the combination of '08-21P' at TMS-16WAP harvesting stage showed the weakest DPPH with an IC 50 of 2.54 mg/mL.On the other hand, the DPPH values of the leaves of South African orange fleshed local cultivars ranged from 3.51 to 5.21 mg/mL, as reported by Phahlane et al. [25].The reason for this outcome may be the type of plant or the timing of when the leaves were harvested.The results showed that the leaves of 'Purple-purple', '08-21P', '2019-1-1', and '2019-11-2' genotypes at 16WAP demonstrated lower DPPH scavenging activity.Leaves with high anthocyanin content have been found to possess strong DPPH scavenging activity [59], hence a potent in genotype '2019-11-2'.Abiotic stress triggers the production of reactive oxygen species (ROS), which prompts plants to regulate their metabolic enzymes to produce antioxidants.When the production of free radicals exceeds that of free radical scavengers (antioxidants), the plant undergoes oxidative stress [60].The sweet potato leaf extracts of these genotypes indicate the presence of antioxidant properties with the ability to protect the plant against oxidative stress and the ability to provide physiological defence against oxidative and free-radical-mediated reactions promoting health benefits to consumers [2].Furthermore, SPLs could serve as a natural source of natural antioxidants in this growing interest over synthetic antioxidants because it will be safer and there are no health risk issues.In a recent study by Krochmal-Marczak et al. [41], the total antioxidant activity varied across different leaf samples harvested at three stages of growth from various cultivars as compared to our present study.The correlation between individual compounds and the antioxidant power (FRAP) are shown in Figure S3A of Supplementary File.Cyanidin-caffeoyl-sophoroside-glucoside (r = 0.60, p < 0.05) displayed the strongest correlation with the FRAP value, followed by cyanidin-caffeoyl-feruloylsophoroside-glucoside (r = 0.58, p < 0.05), peonidin-caffeoyl-hydroxybenzoyl-sophorisideglucoside (r = 0.56, p < 0.05), peonidin feruloyl-sophoroside-glucoside (r = 0.55, p < 0.05), peonidin-caffeoyl-feruloyl-sophoroside-glucoside (r = 0.54, p < 0.05), and chlorogenic acid (r = 0.42, p < 0.05).The highest correlation with the ABTS scavenging activity (Figure S3B of Supplementary File) were exhibited by peonidin-caffeoyl-feruloyl-sophoroside-glucoside (r = 0.59, p < 0.05), followed by cyanidin-caffeoyl-sophoroside-glucoside (r = 0.52, p < 0.05), 5CQA (r = 0.50, p < 0.05), and rutin (r = 0.40, p < 0.05), whereas for DPPH scavenging activity, the highest correlation (Figure S3C of Supplementary File) was with cyanidin-caffeoylsophoroside-glucoside (r = 0.67, p < 0.05), followed by peonidin-caffeoyl-hydroxybenzoylsophoriside-glucoside (r = 0.56, p < 0.05), peonidin feruloyl-sophoroside-glucoside (r = 0.54, p < 0.05), cyanidin-caffeoyl-feruloyl-sophoroside-glucoside (r = 0.47, p < 0.05), and rutin (r = 0.40, p < 0.05).The antioxidant effects of phenolic compounds had been associated with the number of hydroxy groups present in their structures [61].Moriyama et al. [62] reported that the superoxide anion-scavenging activity of acylated anthocyanins varied depending on the activity of each corresponding deacylated compound.Phenolic hydroxyl group in anthocyanins contribute substantially to scavenging reactive oxygen radicals [63].
SPLs in five purple-fleshed genotypes accumulated mineral contents differently during different harvesting stages.K was high in SPLs harvested during VS-8WAP, whereas contents of Fe were higher in SPLs harvested at TMS-16WAP except for genotype '2019-11-2' and '2019-1-1' in both seasons in all genotypes.Suárez et al. [40] determined mineral content during three harvesting periods.These mineral contents were higher at the first harvest in agreement with the findings in this present study.High accumulation of Fe at H1 contradict with the high content in 'Purple-purple', '08-21P' and '16-283P' during TMS-16WAP.K and Fe plays a role in plant growth and development, hence high accumulation during the vegetative stage [66,67].The accumulation of minerals can be induced by several factors.For instance, under biotic stress, plants become activated and generate cellular responses once they detect stress [68].Fe alleviates abiotic stresses through its ability to exist in two oxidation states as Fe +3 and Fe +2 whereas K offers reactive oxygen species (ROS) defence and hinders the accumulation of ROS [66,69,70].Therefore, under harsh environmental conditions sweet potatoes need to be supplemented with minerals to avoid plant stress and to maintain the synthesis of secondary metabolites.The mineral analysis obtained in our study is comparable to the results obtained by Suárez et al. [40] and Sun et al. [42].SPLs contain nutritional compounds found in leafy vegetables, such as spinach, kale, Chinese cabbage, black nightshade, pigweed, jaw's mallow, cowpea, pumpkin, and spider flower.Therefore, sweet potatoes are a good competitive leafy vegetable crop that is good for human consumption.

Conclusions
This study investigated the effects of tuber life cycles and days after planting on leaf anthocyanin and phenolic acids, predominantly chlorogenic acids derivatives and micronutrients.This helps in harvesting the correct leaf stage for consumption or for use as a functional ingredient.The use of sweet potato leaves as leafy vegetables is growing globally and is already very common in Asian countries; but, due to their functional properties, it is likely to become a niche market in the future.It would also be ideal to recommend purple-fleshed sweet potato genotypes that can be used as leafy vegetables in Southern Africa.The high anthocyanin and caffeoylquinic acid derivative content of purple-fleshed sweet potato genotypes '2019-11-2' and 'Purple-purple' warrants their commercialisation as leafy vegetables.In the future, it will be necessary to investigate the palatability and antinutritive components of the leaves in relation to the tuber life cycle and days after planting.

Figure 1 .
Figure 1.Colour development on the leaf blade of purple-fleshed genotypes within th harvestable leaves on vine tips.

Figure 1 .
Figure 1.Colour development on the leaf blade of purple-fleshed genotypes within the first five harvestable leaves on vine tips.

Foods
Year x harvesting stage x genotype significantly (p < 0.001) influenced the chlorophyll content (Table

Figure 2 .
Figure 2. Heat map.In the map, the various phenolic compounds found in different sw cultivars are coloured according to their concentration.The rows represent phenolic co and the columns represent the leaves of sweet potato genotype and harvesting stage relate life cycle (VS-8WAP-vegetative stage 8 weeks after planting, TIS-12WAP-tuber initiatio weeks after planting, TMS-16WAP-tuber maturation stage 16 weeks after planting).Th red and blue indicate high and low levels, respectively.

Figure 2 .
Figure 2. Heat map.In the map, the various phenolic compounds found in different sweet potato cultivars are coloured according to concentration.The rows represent phenolic compounds, and the columns represent the leaves of sweet potato genotype and harvesting stage related to tuber life cycle (VS-8WAP-vegetative stage 8 weeks after planting, TIS-12WAP-tuber initiation stage 12 weeks after planting, TMS-16WAP-tuber maturation stage 16 weeks after planting).The colours red and blue indicate high and low levels, respectively.

Table 1 .
The level of TPC and TC in leaves of five purple-fleshed genotypes harvested at different harvesting stages.

Table 2 .
Comparison of anthocyanin compounds detected in the leaves of different purple-fleshed sweet potato genotypes at different stages of harvesting.

Table 5 .
FRAP, ABTS, and DPPH antioxidant activity of sweet potato leaves harvested during three harvesting stages.