Antioxidant activity, nutritional and physicochemical characteristics, and toxicity of minimally refined brown sugar and other sugars

Abstract Minimally refined brown sugar (MRBS) is a brown sugar derived from sugarcane that has a low glycemic index. This study aimed to determine and compare the antioxidant contents and nutritional and physicochemical properties of MRBS, refined sugar (RS), and brown sugar (BS). In addition, the toxicity of these sugars was evaluated via in vitro cytotoxicity method and by using a zebrafish model. Results showed that MRBS was better than the two other sugars because it has a lower moisture content and higher ash content. The contents of potassium and manganese of MRBS were higher than those of the two other sugars. Surprisingly, MRBS also contained selenium, which was not detected in RS and BS. The major phenolics in MRBS are 4‐hydroxybenzoic acid, chlorogenic acid, protocatechuic acid, trans‐Ferulic acid, and apigenin. All sugar solutions and their antioxidant‐containing extracts were not cytotoxic to 3T3‐L1 adipocytes.

MRBS derived from sugarcane is not genetically modified (Jaffé, 2015). The fewer processes involved in producing MRBS retain some naturally occurring trace minerals (such as calcium, magnesium, and potassium), vitamins, amino acids, antioxidants, and phytochemicals (Lee et al., 2018). Given that health professionals and officials recommend increasing the intake of foods rich in antioxidants, substituting RS with MRBS may offer a potential extra source of antioxidants.
In general, all types of cane BS are nontoxic sweetener widely consumed worldwide. Interest in polyphenols, including flavonoids and phenolic acids, has considerably increased because several studies have demonstrated that the toxic effects of phenolic compounds depend on phenolic concentration (Guimarães et al., 2007;Payet, Sing, & A., Smadja, J, 2006;Valli et al., 2012). Numerous studies have evaluated the potential toxicity of natural sweeteners, such as honey, molasses, and syrups (Guimarães et al., 2007;Payet et al., 2006;Seguí, Calabuig-Jiménez, Betoret, & Fito, 2015;Valli et al., 2012). MRBS has been recently commercialized in Malaysia because of its antioxidant content. However, the potential health effects of cane MRBS as a substitute for RS on food have not been assessed. Moreover, its toxicity has not been evaluated yet. In this study, we investigated the antioxidant activity, nutritional and physicochemical characteristics, vitamin and mineral contents, and toxic effects of MRBS via in vitro chemical and biological assays. We evaluated the cytotoxic effects of these sugars via in vitro cytotoxicity method and a zebrafish model (van Eyk, 2015;Valli et al., 2012).
Finally, we compared MRBS with commercial BS and sugars in terms of the said parameters. RS, commercial BS, and MRBS were obtained in triplicate from the Central Sugar Refinery Sdn Bhd, Malaysia. All sugar samples (5.0 kg each) were stored at room temperature (25°C) until use. All chemicals, standards, and solvents of analytical grade were purchased from Sigma-Aldrich Company (Germany).

| Extraction of polyphenols
Polyphenols were extracted following the method of Chen, Zhao, and Yu (2015) with some modifications to obtain extracts of good quality. Solid samples were ground into powder and thoroughly homogenized with liquid nitrogen. Polyphenols were extracted by mixing 1.0 g of sugar samples with 10 ml of acidified ethanolic solution (1.6 M HCl in 60% ethanol,v/v). The mixtures were vigorously stirred for 1 min and then ultrasonicated using a Powersonic 405 ultrasonic homogenizer for 20 min (Hwashin Technology Co). The solvent was removed by a rotary vacuum evaporator at 45°C, and the aqueous extracts obtained were used to determine antioxidants via in vitro assays and to evaluate the cytotoxic effects of the sugar extracts.

| Determination of fats, proteins, and carbohydrates
Proximate compositions of the sugar samples were determined following AOAC methods (AOAC, 2012). In accordance with AOAC method 991.36, fats were extracted using petroleum ether as the extraction solvent in a Soxhlet extractor. The sugar samples were hydrolyzed according to the Weibull-Stoldt method to extract fats.
The sugar samples (100 g) were hydrolyzed using 100 ml of 3 M HCl for 1 hr under reflux.
Protein contents of the sugar samples were determined following the AOAC method 981.10 and the Kjeldahl method. Total nitrogen contents were determined using a Kjeltec automatic analyzer (Protein-Gerhardt VAP50 Vapodest, Konigswinter, Germany). The recovery rate of the Kjeldahl method was 80%-120%.
Carbohydrate contents were determined as total sugar content on the basis of the AOAC method 982.14 with some modifications (AOAC, 2012). Carbohydrate content was calculated as.
where R and R′ are the peak heights and peak area of sample or standard, respectively; V is the volume of sample; W is the weight of sample; and C is the concentration of sugar standard (g/ml). Each experiment was performed in triplicate.

| Determination of total dietary fibers, moisture contents, ashes, and minerals
Moisture contents were determined by placing a 2 g sugar sample in a drying oven at 105 ± 2°C until a constant weight was achieved following the AOAC standard method no. 991.43 (dietary fiber) (AOAC, 2012). Ash contents were determined by placing a 2 g %component = (R∕R�) × (C∕W) × V × 100, sugar sample in a crucible and heating it in a muffle furnace to 550°C for 6 hr according to the AOAC method no. 923.03 (AOAC, 2012).
Mineral contents 2 g sugar samples were determined according to the AOAC method nos. 968.08 and 965.09 (AOAC, 2012).
All mineral standards had a recovery of 73%-111%. Each experiment was performed in triplicate.

| Determination of B vitamins
B vitamins were determined on the basis of methods described by Ekinci and Kadakal (2005)  The total run time of HPLC analysis was set to 20 min.
The standards used for B vitamins were thiamin, riboflavin, niacin, pantothenic acid, pyridoxine, and folic acid. Different concentrations (5, 10, 25, 50, and 100 ppm) of the standards were prepared for standard calibration.

| Estimation of total phenolic, total flavonoid, and total anthocyanin contents
Total phenolic contents (TPC) were estimated using the Folin-Ciocalteu method described by Wu et al. (2006) with slight modifications. An aliquot of 0.125 ml of the sugar solution (25 mg/ml) was added with 0.5 ml of Folin-Ciocalteu reagent in a test tube. All tubes were vortexed for 15 s and then incubated at room temperature (25°C). After 6 min, 1.25 ml of 7.5% sodium carbonate solution was added to the mixture and topped up to 10 ml with distilled water.
Absorbance was measured at 760 nm by using a Hewlett Packard UV-vis spectrophotometer (Palo Alto, CA). TPC was calculated on the basis of a standard curve of gallic acid solutions at five concentrations ranging from 0.05 to 2.0 mg/ml. Results were expressed as gallic acid equivalents (mg GAE/100 g).
Total flavonoid contents (TFC) were determined via the aluminum chloride colorimetric method (Seguí et al., 2015). In brief, 1.5 ml of the extracts was mixed with 1.5 ml of aluminum chloride solution (2% w/v in methanol). The mixture was vigorously shaken and allowed to react for 10 min. Absorbance of the mixture at 368 nm was measured, and flavonoid content was expressed in milligram of quercetin equivalents per gram of fresh weight.
Total anthocyanin contents were estimated via the spectrophotometric pH differential method (Wu et al., 2006). The solutions (25 mg/ml) were thoroughly mixed with 0.025 M potassium chloride at pH 1.0. Absorbance of the mixture at 510 and 700 nm was measured using distilled water. The mixtures were then combined with 0.4 M sodium acetate buffer at pH 4.5, and their absorbance was measured at the same wavelength. Absorbance was measured as follows: Total anthocyanin contents were calculated as the total monomeric anthocyanins: where A is the absorbance of diluted sugar solutions, MW is the molecular weight of cyanidin-3-galactoside (484.84), DF is dilution factor, and ε is equal to 34,300 M/cm. Each experiment was performed in triplicate.

| Determination of phenolic compounds
Phenolic compounds were separated via the method described by Zakaria et al. (2019) with some modifications by using a Hypersil GOLD column (1.9 µm, 2.1 mm × 100 mm) (Thermo Scientific). The stationary phase was attached to a Dionex Ultimate 3000 Series UHPLC (Thermo Scientific). Phenolic compounds were separated on the basis of a gradient elution of formic acid in water (A) and formic acid in acetonitrile (B) at a flow rate of 0.4 ml/min and injection volume of 2.0 µl. Gradient run was set as follows: 5%-90% B for 5.0 min, 90%-90% B for 2 min, 90%-5% for 2 min, and 5%-5% for 3 min. Total run time of LCMS analysis was 12 min with a 3 min postrun time.
Heated electrospray ionization conditions were set with capillary temperature of 320°C, spray voltage of 3,700 V, gas flow rate of 18 L/min, and nebulizer pressure of 45 psi. Mass spectrometry data were documented in negative ion mode. Mass spectrometry data were acquired within 50-750 m/z for both positive and negative ion modes. Data were analyzed using Xcalibur 2.2 software (Thermo Fisher Scientific, Waltham, MA). Phenolic compounds were quantified on the basis of standard calibration curves of phenolic acids (benzoic acid, 4-hydroxybenzoic acid, caffeic acid, chlorogenic acid, p-coumaric acid, trans-Ferulic acid, protocatechuic acid, syringic acid, and vanillic acid) and flavonoids (apigenin, luteolin, tricin, and vanillin). These phenolic acid standards were obtained from Sigma-Aldrich (M) Sdn Bhd (Selangor, Malaysia). The coefficient correlation values (R 2 ) of all standard calibration curves were higher than 0.9.

| DPPH radical scavenging capacity assay
DPPH radical scavenging capacity assay was performed following the method of Wu et al. (2006) with slight modifications. In brief, an aliquot of 0.1 ml of 10 mg/ml sugar solutions was mixed with 2.0 ml of 0.6 mM DPPH radical solution in 80% ethanol. The tubes were vortexed for 15 s and allowed to stand at room temperature for 30 min in the dark. Absorbance of the mixture at 517 nm was measured using a UV-vis spectrophotometer (SECOMAM SA, Alѐs, France). A blank of 80% ethanol was used for this assay. The control used was a DPPH radical solution without the test samples.
Inhibition activity was calculated using the following equation:

| Ferric ion reducing antioxidant power assay
Antioxidant activity was estimated on the basis of ferric reducing antioxidant power (FRAP) assay described by Wu, Lin, Lin, Ken, and Wen (2007) and Pulido, Bravo, and Saura-Calixto (2000) with some modifications. FRAP reagent was freshly prepared by mixing 2.5 ml of 10% trichloroacetic acid, 2.5 ml of 1% potassium ferricyanide, and 2.5 ml of 0.2 M phosphate buffer at pH 6.6. In brief, 2.5 ml of the FRAP reagent was added with 2.5 ml of distilled water, 0.5 ml of 1% iron (III) chloride, and 1.0 ml of 10 mg/ml sugar solutions and then incubated at 50°C for 30 min. Reagent blank was prepared by mixing the test samples without reagent. Absorbance of the reac-

| Cytotoxicity of 3T3-L1 cell lines
3T3-L1, a preadipocyte cell line (ATCC), was obtained from the Institute of Biosciences, Universiti Putra Malaysia. The adipocytes were cultured and maintained in a complete growth medium containing high-glucose DMEM supplemented with 10% fetal bovine serum and 100 U/mL penicillin-streptomycin in a 5% (CO 2 ) incubator at 37°C (Castillo, González, & Moore-Carrasco, 2019). Cytotoxicity was measured using the MTT assay (Bahuguna, Khan, Bajpai, & Kang, 2017). In brief, 3T3-L1 cells were seeded at 2 × 10 3 cells per well in a 96-well plate containing 100 µl of complete growth medium per well. The plate was incubated in the CO 2 incubator until the cells reached 70%-80% confluence. The medium was then removed and replaced with a freshly prepared medium containing 0.05-200 mg/ml of the sugar samples. The plates were incubated for 24, 48, and 72 hr before measuring sugar cytotoxicity via the MTT assay. Afterward, the medium (containing media and samples) was removed, and the plate was washed with phosphate buffer saline (PBS, pH 7.4) to remove dead cells and debris. Subsequently, a medium mixture containing 20 µl of MTT reagent (5 mg/ml in PBS) was prepared, and then, 100 µl of the medium was added to each well of the plate and wrapped with aluminum foil. The plate was then incubated at 37°C in the CO 2 incubator for 44 hr until the purple formazan was visible under the microscope. The medium was then discarded, and the crystalized formazan was dissolved with 100 µl of dimethyl sulfoxide and swirled on a microplate shaker for 15 min.
The dissolved formazan was measured using a microplate reader at 540 nm (Molecular Devices Corp., VERSAmax, and Sunnyvale, CA, USA). All experiments were run in three replicates on three different days.

| Staging of fish embryos
The procedures for zebrafish acute toxicity experiments were adopted from Ding and Chen (2012) and Sata, Bakar, Ramlan, and Ibrahim (2016) with slight modifications. The health of zebrafish embryos was examined before performing the acute toxicity tests. In the tests, coagulated eggs were replaced with new healthy embryos ( Figure 1).

| Exposure of embryos to sugar solutions
All embryos at 24 hr postfertilization (hpf) were cultivated in 96well cell culture plates by using a Pasteur pipette (Zhang, Lu, Gelinas, Ciruna, & Sun, 2011). The hatching, survival, and heart rates of all embryos were measured at each check-point. Before treatment, the embryo media were discarded from the wells to ensure that the sugar solution was treated at the desired concentration. The solution was transferred into each well by using a pipette, and each well contained five embryos. Embryonic development was observed and recorded at 24 hr intervals. The hatching rate, scoliosis, and heartbeat (beats/min) of the embryos were recorded and analyzed at the end of the treatment (96 hr). Images of the morphology of zebrafish embryos were obtained using an inverted microscope with a built-in camera (Hoage, Ding, & Xu, 2013;Ismail et al., 2017).

| Statistical analysis
All analyses were conducted in triplicate (n = 3) for each sugar sample. Minitab version 18 was used for statistical analysis. Significant differences between the mean values of the samples were statistically analyzed using one-way ANOVA coupled with LSD multiple comparisons.

| Proximate composition analysis
The proximate compositions of all sugar samples are presented in Table 1. Results showed that BS samples had the highest protein, dietary fiber, and moisture contents, but the samples had the lowest carbohydrate content at p < .05. MRBS samples had the highest ash content (0.09%). These findings were somewhat consistent with those of Lee et al. (2018), who reported that unrefined sugars have higher ash contents (0.02%-1.54%) than refined sugars. Previous studies suggested that the ash content of unrefined sugars should not exceed 2.2% because a high ash content is attributed to a high potassium content, which may impart an unpleasant taste and hinder sugar crystallization (Lee et al., 2018;Lopes & Borges, 2004).
In the present study, the ash contents of the sugars (0.01%-0.09%) were below the level proposed by Lopes and Borges (2004). The moisture content of the tested sugars ranged from 0.1% to 0.15%.
The moisture content of BS (0.15%) was significantly higher than that of RS (0.1%) and MRBS (0.11%). Variations in moisture content among sugars are due to differences in processing conditions (Jaffé, 2015). However, a high moisture content promotes crystal dissolution, biochemical degradation reactions, microbial deterioration, and cobblestone formation, all of which shorten the shelf life of sugars (Guerra & Mujica, 2010). However, the difference in energy content among the samples was not significant (at p > .05). Further assays revealed that RS had the lowest amounts of many nutrients, such as protein, ash, dietary fiber, and moisture. Aside from these nutrients, dietary fiber was not detected in RS samples. By contrast, RS samples had the highest carbohydrate content because white cane RS contains low to trace amounts of antioxidants (Phillips, Carlsen, & Blomhoff, 2009) and other nutrients besides sucrose. HPLC revealed that the sugar samples contained 97% sucrose as the main carbohydrate. However, glucose, fructose, and maltose were not detected. As shown in Table 1, RS contained the lowest amount of sucrose (96.57%), whereas MRBS and BS had slightly higher sucrose content (97.3%). The same amount of sucrose in MRBS and BS was probably due to the fact that both sugars are regular BS. The only difference between these sugars is their processing procedures. MRBS undergoes fewer refining processes than BS. Raw BS has about 92% of total sugar, of which 84.5% is sucrose (Jaffé, 2015). In the present study, the sucrose content of both MRBS and BS samples was higher than that of raw BS reported in the literature. This high amount of disaccharides in BS correlates well with the low GI compared. The influence of different raw materials and various manufacturing processes may explain the variations observed herein. In addition, mass accumulation of sucrose in sugar cane is hypothesized to be caused by the activities of certain enzymes in cane tissues (Rohwer & Botha, 2001). For instance, sucrose phosphorylase is known to catalyze sucrose synthesis from uridine diphosphate glucose (Cardini, Leloir, & Chiriboga, 1955). Therefore, BS and MRBS are supposed to retain diphosphate glucose because they undergo less treatment than RS (Valli et al., 2012).

| Physicochemical characteristics
Physicochemical characteristics, such as acidity, Brix, and color, were determined. Results showed that the physicochemical characteristics were significantly different among all sugar samples (Table 1) (Table 1).
A sugar sample with high acidity indicates that it has undergone fewer processing and refining steps. MRBS had the highest acidity, thus confirming that it was the least processed cane sugar. The higher acidity of MRBS than the two other sugar samples may confer it with some advantages for use as a natural sweetener. High acidity helps regulate the acidity of a juice or drink added with MRBS. MRBS had the highest Brix value; thus, it had the highest amount of glucose and sweeter than the two other sugars. Similar findings in refined and unrefined sugars have been reported (Lee et al., 2018). Given that RS undergoes numerous processing steps, most of its nutrients, including minerals, are removed or lost because of heavy processing, polishing, and refining (Hansen, Harholt, Oikawa, & Scheller, 2012).

| Determination of vitamins and minerals
The presence of selenium in MRBS makes it an ideal candidate for low GI sweetener with antioxidants. Other cane sugars may also exhibit contain high selenium content. Given that selenium is an insulin mimic (Zhu et al., 2013), MRBS is a good choice of sweetener for patients with diabetes. MRBS is also rich in potassium. Potassium has antihypertensive effects. A recent population-based cross-sectional study revealed that a low dietary intake of sodium increases the risk of diabetes (Sun et al., 2013).

| Identification and quantification of phenolic acids
The phenolic compounds of the sugar samples were determined via UHPLC-ESI-MS. HPLC chromatograms and TICs of the sugar samples are shown in Figure 2. Except for 4-hydroxybenzoic acid, protocatechuic acid, and chlorogenic acid, the phenolic acids among the sugar samples were significantly different. As shown in Abbreviations: C3GE: cyanidin-3-galactoside equivalent; FRAP: ferric ion reducing antioxidant power; GAE: gallic acid equivalent; MRBS: minimally refined brown sugar; ND: not detected; TAC: total anthocyanin content; TFC: total flavonoid content; TPC: total phenolic content.
Although MRBS had trace amounts of chlorogenic acid and 4-hydroxybenzoic acid, MRBS had the highest concentrations of these phenolic acids (data not shown).
As shown in Figure 2, compound A had a monoisotopic mass of m/z 391. Therefore, the compound (2,3,5,6-tetrachlorophenol) with a mass of m/z 229 was the basic unit of compound A. Similarly, the mass of fragment ions (A1-A5) were m/z 367, 359, 337, 323, 229, and 213, respectively. The major phenolic compound A and its fragments ions with molecular structures are presented in Figure 3. Further investigation of this phenolic compound (A) revealed that it was a type of phenolic derivative that was assumed to have formed from the hydrolysis of sugar solutions with a strong acid (HCl) during the extraction of polyphenols (Barrera, Betoret, & Seguí, 2020).
All flavonoids were detected in trace amounts. These flavonoids were represented exclusively by flavones (apigenin, luteolin, and tricin; Table 4). Although the HPLC peaks of these flavonoids were detectable, their concentrations were too low for quantification. Consequently, the concentrations of apigenin, luteolin, and tricin were detected but below the quantification limits. The results were in consistent with those of previous studies, which reported that flavones (apigenin, luteolin, and tricin) are found in sugarcane in low amounts (Colombo, Yariwake, Queiroz, Hdjoko, & Hostettmann, 2005;Duarte-Almeida et al., 2011). MRBS contained the lowest concentrations of luteolin (but the results were not significant), but it had the highest amounts of apigenin. By contrast, BS contained the highest tricin content. Barrera et al. (2020) reported that tricin and apigenin are the most abundant phenolics in raw sugars. No significant differences in the concentrations of flavonoids were found among all samples. Other phenolic compounds were not detected in the sugar samples because they were probably lost during the extraction, especially during thermal degradation (Payet et al., 2006). Finally, the concentrations of inverted sugars in MRBS also affected the evaluation of phenolic and flavonoid contents (Seguí et al., 2015).

| Antioxidant activities of sugar samples
The antioxidant activities of the sugar samples were determined via DPPH radical scavenging effect and FRAP assays. As shown in Note: Data are shown as mean ± standard error (µg/g sample).
a The values of flavonoids in the sugar samples are not quantifiable due to concentrations are below the quantification limit. Trace (<1.0 µg/g samplebelow the quantification limit).

F I G U R E 3
Major compounds identified using LCMS and their molecular profile      activity of branded unrefined sugars are available (Lee et al., 2018).

However, the DPPH inhibition activities of light and dark BS
were not significantly different. A previous study reported that granulated RS (0.013) had FRAP values (mmol/130 g) lower than those of BS (0.516-0.986) and raw cane sugar (0.186) (Phillips et al., 2009). Raw cane sugar has a lower DPPH inhibition activity (22.1%) than all the samples studied (>88%); the inhibition activity of commercial BS ranges from 14.5% to 26.9% (Payet et al., 2005).
Another study reported that raw cane sugar has a DPPH scavenging activity of 0.4-0.9 mmol Trolox equivalent/100 g sample and a FRAP value of 0.204 mmol/100 g (Jaffé, 2015). Thus, the antioxidant activity of the sugars is attributed to their polyphenolic constituents (mainly flavonoids, polyphenol, and phenolic acids) (Yao et al., 2004). The results for antioxidant activity were in good agreement with those of previous studies in terms of total phenols, flavonoids, and color values. For instance, tricin and apigenin are the most abundant phenolics in cane sugars. These phenolics are considered as an important bioactive constituent of foods and are postulated as nutraceuticals with antiproliferative and chemopreventive agents (Barrera et al., 2020). Several studies have proved that the correlation between total polyphenols and antioxidant activity are generally good but depends on the nature of the sample and the effects of other compounds on antioxidant activity (Guimaraes et al., 2007;Kadam et al., 2008;Moure et al., 2001). Barrera et al. (2020) found that the physicochemical and antioxidant properties of raw sugars are strongly related to the degree of refinement of each product. In the present study, total polyphenols and antioxidant activity were not significantly correlated. Pearson correlation coefficient analysis showed that the TPC of the sugar samples was not significantly correlated with DPPH inhibition. Aside from DPPH, the TPC of the sugar samples was strongly correlated with FRAP values (r = 0.981, p < .001).
The TAC of the sugar samples was also strongly correlated with FRAP values (r = 0.826, p < .01). Payet et al. (2005) also demonstrated that the percentage of DPPH inhibition is not significantly correlated with TPC (r = 0.462, p > .05). By contrast, Feng, Luo, Zhang, Zhong, and Lu (2014) and Seguí et al. (2015) found that the percentage of DPPH inhibition is strongly and positively correlated with TPC (r = 0.881-0.996, p < .01). Thus, the present study demonstrated that antioxidant content had no correlation with DPPH inhibition activity. However, an increase in phenolic concentration would increase antioxidant activity.

| Cytotoxic effects of sugar samples
Few data on the cytotoxic effects of natural sugars in vivo or ex vivo systems are available. Thus, this study aimed to determine the cytotoxicity of RS, MRBS, and BS. All sugar samples were not cytotoxic to 3T3-L1 adipocytes (

| Effects of sugar solutions on hatching and survival rates
As shown in Figure  Compared with the control, most concentrations delayed hatching.
As shown in Figure 5b,  The survival rate of zebrafish embryos treated with MRBS solution was better than that of embryos treated with BS ( Figure 5c). The survival rate of the embryos treated with MRBS at the solution concentration of 10 mg/ml for 96 hr was significantly higher (80%) than that of the embryos treated with BS (70%). However, the embryos treated with both sugar solutions had a high survival rate, demonstrating that both types of BS did not affect the survival rate of (or were not toxic to) zebrafish embryos.

| Effects of sugar solutions on heart rate
The heart rate of zebrafish embryos ranged from 120 beats/min to 180 beats/min ( Figure 6). Heart rates were measured by direct visual examination of ventricles as they beat. Embryos of the negative control group had heart rates of 120-130 beats/min (Chan, Lin, & Cheng, 2009).
Results showed that the heart rate of zebrafish embryos decreased F I G U R E 6 Heart rates of zebrafish embryos treated with sugar solutions of (A) refined sugar, (B) brown sugar and (C) MRBS. *C: control (paracetamol) (a) (b) (c) due to treatment with RS solution at concentrations between 2.5 and 10.0 mg/ml (Figure 6a). However, the embryos treated with paracetamol had slightly lower heartbeat (beats/min), indicating that the RS solution slightly increased the heart rate of zebrafish embryos.
As shown in Figure 6, treatment with MRBS and BS solutions also slightly increased the heart rate of the embryos. The heart rate of the embryos increased as the concentration of BS solution increased. At the same concentration of 10 mg/ml, the embryos treated with paracetamol had significantly a lower heart rate (beats/min) than the embryos treated with BS solution. These results demonstrated that the BS solution increased the heart rate of the zebrafish embryos because BS had high antioxidant properties.
By contrast, treatment with MRBS only slightly reduced heart rate as solution concentration increased. Although the heart rate of embryos increased as the concentration of MRBS increased from 0.31 to 2.5 mg/ml, the heart rate significantly decreased at the sugar concentration of 5 mg/ml and only slightly increased at the sugar concentration of 10 mg/ml. Kim et al. (2015) reported that phenolic compounds may affect the heartbeat of aquatic fish. For instance, they have found that phlorotannin isolated from a brown alga considerably decreased the rate heart of zebrafish embryos treated with 150 mM glucose. In the present study, the phytochemicals in MRBS, such as 4-hydroxybenzoic acid, chlorogenic acid, and protocatechuic acid, and trans-Ferulic acid, within the concentration range of 5-10 mg/ml, decreased the heart rate of the zebrafish embryos compared with BS.

| CON CLUS ION
BS, especially MRBS, had fair amounts of antioxidants. The minimal refining process of MRBS retained some of its phytochemicals, vitamins, and minerals. RS had lower amounts of nutrients than BS, except for sucrose. Heavy sugar refinement or processing causes a considerable decrease in antioxidants (polyphenols) in sugarcane.
Minimal processing of BS retains some of its phytochemicals, such as phenolic acids and flavonoids. This study suggested that small amounts of phenolics in BS pose no threat to humans. This result was further confirmed by in vitro and in vivo assays. Total polyphenols and antioxidant activity were not significantly correlated. Overall, MRBS had a better nutritional quality in terms of physicochemical characteristics than the two other sugar samples. Moreover, MRBS also contained selenium, which is not found in many types of sweeteners. These antioxidants and nutrients found in BS, especially in MRBS, may provide extra benefits compared with RS. However, the amounts of these antioxidants and nutrients were lower in MRBS than those in other sugar alternatives. Therefore, future studies should investigate the antidiabetic effects of MRBS and assess the efficacy of polyphenols in preventing metabolic-related diseases.

ACK N OWLED G M ENTS
We would like to acknowledge Central Sugar Refinery Sdn Bhd, Malaysia for providing us with the research funding and the sugar samples.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.