Chemical composition, antioxidant, and antiproliferative activities of nine Chinese proso millet varieties

ABSTRACT Interest in proso millet as a functional food is growing. However, studies on the chemical composition, antioxidant, and antiproliferative activities’ differences of proso millet cultivars in China are limited. Nine Chinese proso millet varieties (five waxy varieties – S, D, N2, N4, N8 and four non-waxy varieties – J1, J3, J7, J9) were investigated. Results showed that the non-waxy varieties were significantly higher than waxy varieties in the content of amylase and resistant starch (22.95%, 0.23%, 2.02%, and 0.78%, respectively; P < .05). Lys was the first limit amino acid (AAS: 16.08%). Linolenic acid (61.74%) and oleic acid (22.16%) were the dominant fatty acids. Overall, the antioxidant content and antioxidant activities of J1 and J3 were higher than other cultivars. And, J7, J3, and J1 possessed better antiproliferative effects than other proso millet varieties. These results are anticipated to provide useful information on the development of proso millet-based functional food.


Introduction
Proso millet (Panicum miliaceum L.) is a kind of the oldest cultivated millet crop and consumed as a staple food among the majority of people who live in arid and semiarid tropics of the world, such as Asia, Africa, and parts of Europe (Lu et al., 2009). Several reports have shown that millet is superior to other major cereals in nutritional value (Pathak, Srivastava, & Grover, 2000). Starch, which is the most abundant nutrient in proso millet, is nearly 70%. However, proso millet-based products own a lower glycemic index (GI) than corn, which means proso millet appears to be a good ingredient for producing low-GI products (McSweeney, Seetharaman, Ramdath, & Duizer, 2017). Studies have also shown that protein accounts for 12% and the essential amino acid index (EAAI) of proso millet was higher (51%) compared to wheat (Kalinova & Moudry, 2006). Besides, proso millet also contains fat, crude fibre, minerals, vitamins, and other phytochemicals, such as phenolic, flavonoid, proanthocyanidin, and phytic acid (Shahidi & Chandrasekara, 2013).
Recently, proso millet has attracted attention for its considerable health benefits and functional components, particularly with respect to the antioxidant, anti-diabetic, anti-cancer, anti-liver injury, and antiproliferative effects. A large number of studies have shown that the aqueous/ethanol/methanol extracts of proso millet possess antioxidant properties (Chandrasekara & Shahidi, 2011a, 2011bChandrasekara & Shahidi, 2012;Ragaee, Abdel-Aal, & Noaman, 2006). Park, Ito, Nagasawa, Choi, and Nishizawa (2008) suggest that proso millet protein concentrate may have potential for therapeutic intervention in type 2 diabetes. Nishizawa et al. (2002) suggested that proso millet is considered to be another preventive food for liver injury by examining the effects of dietary protein from proso millet on liver injury induced by D-galactosamine in rats.
In spite of good characteristics, proso millets are still underutilized, and there is little information on the differences of Chinese proso millet varieties. Therefore, the present study was carried out to (1) compare the chemical compositions, particularly with respect to amino acid and fatty acid components, (2) evaluate the antioxidant activities and antiproliferative effect on MDA-MB-231-Breast Cell Lines of extracts from nine Chinese proso millet varieties and (3) investigate correlations between the antioxidant activities and antiproliferative effect.

Proso millet materials
Nine proso millet varieties, including five waxy proso millet varieties (S, D, N2, N4, and N8) and four non-waxy proso millet varieties (J1, J3, J7, and J9), were collected. S and D were procured from Millet Research Institute, Shanxi Academy of Agricultural Sciences, China. The others were procured from Gansu Academy of Agricultural Science, China. All samples were dried at 40°C, dehusked in a centrifugal sheller, ground in a laboratory mill (Landert-Motoren AG, Buelach, Switzerland), and passed through a Ф 0.178 mm screen sieve successively to obtain the whole proso millet flour.

Nutritional composition analyses
Crude starch, amylose, and resistant starch were measured using the Megazyme assay kits. Protein and fat contents were determined according to the methods of the Association of Official Analytical Chemists (2005) 979.09, 996.11. Amino acids were determined by an amino acid analyzer (L-8800; Hitachi, Japan) after sample pretreatment (Qin et al., 2014). Besides, the tryptophan was analysed alone after alkaline hydrolysis (Guo et al., 2017). EAAI was calculated as a geometric average of amino acid score (AAS) of all essential amino acids. AAS = 100× the amino acid content in tested protein/content of the amino acid in standard protein according to FAO/WHO. Fatty acid analyses were performed according to the method of Miao et al. (2010), using gas chromatography (Agilent Agilent Technologies, Palo Alto, CA) equipped with a flame ionisation detector.

Preparation of proso millet extracts
Extracts for the determination of antioxidant compounds, antioxidant activities, and antiproliferation activity were prepared according to the method of Shi, Yao, Zhu, and Ren (2017) and some modifications were made. In brief, two grams of proso millet flours were mixed with 20 mL of 70% methanol in a 50-mL centrifuge tube. Then, the mixture was placed in an ultrasonic bath (Powe sonic 420, Hwashin co., Korea) and sonicated at 50°C for 45 min. Subsequently, the mixture was cooled. Then, the mixture was centrifuged at 5000 rpm/min for 10 min. The supernatants were transferred into culture tubes and stored at −4°C for the determination of antioxidant content and antioxidant activities, and the remaining supernatants were dried by a stream of nitrogen for the measurement of cytotoxicity and antiproliferative experiments.

Determination of antioxidant compounds
The measurement of the total saponin content (TSC) was done following the method of Wu, Lin, and Chau (2001) using oleanolic acid as a standard. The results were expressed as mg/g of extracts. Total phenolic content (TPC) measurement was based on the Folin-Ciocalteu method of Dewanto, Wu, Adom, and Liu (2002) using gallic acid as a standard. The results were expressed as mg/g of extracts. Total flavonoid content (TFC) detection was based on aluminium chloride calorimetric assay of Kim, Jeong, and Lee (2003) using rutin as a standard. The results were expressed as mg/g of extracts.

Determination of antioxidant activity
The DPPH radical scavenging activity was estimated according to the method of Cheung, Cheung, and Ooi (2003) and calculated based on a calibration curve using Trolox and expressed as μM/g of extracts.
The scavenging capacity of 2,2'-azino-di-[3-ethylbenzthiazoline sulphonate] (ABTS •+ radical cations) was measured using a total antioxidant capacity assay kit (the ABTS method). Briefly, 200-μL ABTS + solution and 10 μL of the supernate (distilled water or solution of Trolox standard) were added to a 96-well microplate. The mixture was mixed and left for 2-6 min at room temperature under dim light, and then the absorbance was measured at 734 nm using a microplate reader. The results were reflected in μM/g of extracts.
The FRAP assay was determined using a total antioxidant capacity assay kit (the FRAP method). Briefly, 180-μL FRAP solution and 5 μL of the supernate (distilled water or solution of FeSO 4 standard) were added to a 96-well microplate. The mixture was mixed and left for 3-5 min at room temperature under dim light, and then the absorbance was measured at 593 nm using a microplate reader. The results were expressed as the concentration of extracts (μM/g of extracts) having a ferric reducing ability equivalent to that of 1 mmol/L FeSO4.

Measurement of cytotoxicity by methylene blue assay
The cytotoxicity was determined using the method reported by Wang et al. (2014) and Yoon and Liu (2008). Briefly, MDA-MB-231 cells were seeded into a 96-well microplate (4.0 × 10 4 cells/well) in 100 μL of growth medium/well and were incubated for 24 h. Then, the cells were treated with 100 μL of medium with different concentrations (31.25, 62.5, 125, 250, 500 μg/mL) of proso millet extracts; the wells that received medium without proso millet extract served as the control. After 24 h of incubation, the cells were stained using 50-μL methylene blue coloured liquid (98% HBSS, 0.7% glutaraldehyde and 0.6% methylene blue). After 1 h of incubation, the cells were washed six times in deionised water, and 100 μL of elution buffer (49% PBS, 50% ethanol, and 1% acetic acid) was added. The absorbance was measured at 570 nm in a microplate reader (Bio Rad, IMAX, Hercules, USA) after shaking for 20 min on a tablet oscillator. The cytotoxicity was determined as percentage compared to the control. If a certain concentration of proso millet extract reduced cell viability compared to the control, then that concentration was considered to be cytotoxic (Felice, Sun, & Liu, 2009).

Measurement of antiproliferative activity by methylene blue assay
The antiproliferative activities of proso millet extracts were assessed using the method of Wang et al. (2014) and Yoon and Liu (2008). Briefly, MDA-MB-231 cells were seeded into the central 96-well microplate (2.5 × 10 4 cells/well), and cell-free medium was added to the peripheral wells of the 96-well microplate at 100 μL of growth medium/well. After 8 h of incubation, the medium was removed and 100 μL of medium with different concentrations (31.25, 62.5, 125, 250, 500 μg/mL) of proso millet extracts was added; the wells that received medium without proso millet extract served as the control. Following 96 h of incubation, the staining solution was removed and the 96-well microplates were washed six times in deionised water. Then, 100 μL of elution buffer (49% PBS, 50% ethanol, 1% acetic acid) was added to each well. The 96-well microplates were transferred to a tablet oscillator for 20 min. Absorbance was measured at 570 nm using a microplate reader (Bio Rad, IMAX, Hercules, USA). The antiproliferative activities were evaluated by the EC 50 (Effective concentration of 50% cell proliferation)values, which were expressed as μg of proso millet extracts/mL.

Statistical analysis
All determinations were made in triplicate. The experimental data were expressed as the mean ± SD. One-way analysis of variance and Dunnett's multiple range tests (SPSS 17.0) were used to determine the significant differences between a group means at P < .05.Correlations between DPPH, ABTS + , FRAP, and EC 50 values of antiproliferative activity were identified using Spearman's correlation (SPSS 17.0). Correlations were considered highly significant at P < .01.

Total starch, amylose, and resistant starch
The total starch content of these 9 proso millet varieties ranged from 65.82% to 78.59% (Table 1). The average amylase content of these 4 non-waxy proso millets (22.95%) were significantly higher (P < .05) than that of these 5 waxy proso millets (0.23%), which were in agreement with the earlier reports of Ragaee et al.(2006). The RS content of these 9 proso millet varieties ranged from 0.64% (S) to 2.21% (J7), and non-waxy proso millet varieties were significantly richer (P < .05) than waxy proso millet varieties, which were in agreement with the report of Ragaee et al. (2006). RS is a fraction of starch that resists digestion within the small intestine, reaching the large intestine intact. RS-rich foods may be particularly useful for managing diabetes and improving gut microbiota composition (Matsumoto et al., 2016;Nielsen, Theil, Purup, NØrskov, & Knudsen, 2015). Thus, the non-waxy proso millets, especially the variety of J7, may be a good ingredientfor the prevention of diabetes and improvement of intestinal flora.

Total protein and amino acid profile
The contents of protein and 18 kinds of amino acids in 9 Chinese proso millet varieties are shown in Table 2. The average protein content of waxy proso millet varieties was higher than non-waxy proso millet varieties (13.38% and 12.05%, respectively). All varieties of proso millet were particularly rich in Glu, Leu, and Ala (averaging 32.55 mg/g, 16.19 mg/g, and 12.25 mg/g, respectively). Besides, the AAS, which is based on the theory of balance of essential amino acids, was used to judge the limit amino acid (Table 3). It can be found that Lys was the first limit amino acid for all these nine proso millet cultivars and sulphurous amino acid (Met and Cys) was the second limit amino acid (AAS: 16.08% and 62.00%, respectively), which were in agreement with the earlier reports of Kalinova and Moudry (2006). The average EAAI of waxy proso millet varieties (80.64%) was higher than non-waxy proso millet varieties (77.57%), and the highest value of protein (EAAI) was found in J7(85.22%). The results were higher than that of wheat and the report of Kalinova and Moudry (2006), and this is possibly due to the difference of varieties and the conditions for growth.

Total fat and fatty acid profile
The fat and fatty acid compositions of nine Chinese proso millet varieties are presented in Table 4. The fat contents ranged from 1.54% (J3) to 3.77% (D), which were lower than that in brown grains but in agreement with the range of 1.2% to 3.8% in polished and whole grains reported by Devisetti, Yadahally, and Bhattacharya (2014). There were clear differences in the fatty acid contents of the 9 Chinese proso millet cultivars. Linolenic acid and oleic acid were the two dominant fatty acids in the nine proso millet varieties, and their average contents were 61.74% and 22.16%, respectively. These results were consistent with the report (65.5% and 20.0%, respectively) of Annor, Marcone, Corredig, Bertoft, and Seetharaman (2015). D and N8 had the highest linolenic acid (64.23%) and linolenic acid (2.51%) among all proso millet cultivars, respectively. Linolenic acid and linolenic acids are essential fatty acid and must be supplied via the diet (Kim, Nam, Kim, Hayes, & Lee, 2014;Malcicka, Visser, & Ellers, 2017). It is reported that linolenic acid has many benefits for human health, such as cardiovascular-protective, anti-cancer, neuroprotective, and anti-inflammatory effects (Kim et al., 2014). Therefore, proso millet may be beneficial as a healthy food ingredient.

The phytochemical compounds and antioxidant activities of methanol extracts
It has been reported that methanol is an effective solvent in extracting phenolics and other polar substances in grains (Suma & Urooj, 2012). In this study, 70% methanol Note: Data are expressed as mean ± standard deviation of triplicate samples; mean values with different superscripts in the same row are significantly different (P < .05). *Means there are significantly different (P < .05) between the waxy group and non-waxy group within the same row.

FOOD AND AGRICULTURAL IMMUNOLOGY
extracts from proso millets were used for the determination of phytochemical compounds, antioxidant, and antiproliferative activities. The TSC, TPC, and TFC ranged from 100.93 to 162.11 mg oleanolic acid equivalents, 9.28-19.05 mg gallic acid equivalents, and 3.51-7.33 mg rutin equivalents per gram of extracts, respectively (Table 5). Non-waxy proso millets were significantly higher than waxy proso millet in TSC (136.12 mg/g and 108.51 mg/g, respectively; P < .05). J1 had the highest TSC and TPC, while S possessed the lowest. The results were in agreement with the studies of Choi, Jeong, and Lee (2007) and higher than wheat, barley, and rice (Kim, Hyun, & Kim, 2010). Table 3. Amino acid content (AA, mg/g protein), amino acid score (AAS, %), and essential amino acid index (EAAI, %) of nine Chinese proso millet cultivars. The antioxidant properties of proso millets were evaluated on the basis of measuring scavenging for DPPH radicals, ABTS radical cations, and FRAP by proso millet methanol extracts. The average proso millet extracts' values were 35.60 μmol TE/g, 75.76 μmol TE/g, and 86.36 μmol FeSO 4 /g as determined by the DPPH, ABTS, and FRAP assays, respectively (Table 5). As a whole, non-waxy proso millet possessed a higher antioxidant activity than waxy proso millet. Among all these nine Chinese proso millet varieties, N2 displayed the highest DPPH radical scavenging activity (41.28 μM/g), J3 had the highest ABTS radical scavenging activity and FRAP value (104.16 μM/g and 117.54 μM/g, respectively), whereas S showed the lowest antioxidant activity. The differences between DPPH, ABTS + radical scavenging activity, and FRAP among the varieties may be due to the content and composition of phenolics and flavonoids of the extracts (Chandrasekara and Shahidi, 2011).

Cytotoxicity and antiproliferative effects on MDA-MB-231-Breast Cell Lines
The cytotoxic effects and inhibition of MDA-MB-231-Breast Cell Lines' proliferation by the proso millet extracts are presented in Figure 1. As shown in Figure 1(a), all of the extracts had no cytotoxic effect (<10% reduction in the absorbance reading at 570 nm compared to the control) on MDA-MB-231-Breast Cell Lines in a low dosage range (0-125 μg/mL). Figure 1(b) shows that all extracts inhibited MDA-MB-231-Breast Cell Lines' proliferation at doses of 31.25, 62.5, 125, 250, 500 μg/mL, and the inhibition was dose-dependent. For the different cultivars of proso millet, the EC 50 values of antiproliferative activities are shown in Figure 2. The EC 50 of J7 (88.26 μg/mL), J3 (93.98 μg/mL), and J1 (94.98 μg/mL) were significantly lower than all other cultivars (P < .05). Therefore, the methanol extracts of J7, J3, and J1 possessed better antiproliferative effects than other proso millet varieties. The result suggested that the inhibitory effect was not attributed to a cytotoxic effect but to the anticancer effects of the extracts in a certain extract range. Zhang, Liu, and Wei (2014)  also reported that both free and bound phenolic extracts of edible proso millet showed antiproliferative activities towards MDA cells. However, further researches are needed to elucidate which function compounds are responsible for the antiproliferative activity of proso millet.

Correlations
The correlations between antioxidant activity (DPPH, ABTS, and FRAP value) and antiproliferative activity (EC 50 value) are shown in Table 6. The results showed that antiproliferative activity significantly negatively correlates with ABTS (P < .01) and FRAP (P < .05). The negative correlation indicates that the higher ABTS values resulted in a higher inhibition of cancer cell proliferation. However, there was no significant linear

Conclusion
Although all the nine kinds of Chinese proso millet had good nutritional characteristics of antioxidant and antiproliferative activities, there were still significant differences among the evaluated proso varieties. The non-waxy varieties were significantly higher than waxy varieties in the content of amylase and resistant starch (P < .01). All of the proso millets showed high content of protein and suitable amino acid constitution (EAAI = 79.27%). Oleic and linolenic acids were the dominant fatty acids. In general, non-waxy proso millet, especially J3 and J7, possessed a higher content of phytochemical compounds, antioxidant, and antiproliferative activity than waxy proso millet. Besides, significant negative correlations (P < .05) of antiproliferative activity with ABTS and FRAP were observed. The nutritional and biological activities data suggest that the proso millets, particularly non-waxy proso millets, hold promise as healthy food ingredients. These results are anticipated to provide useful information on the selection of daily diets or the development of proso millet-based functional food.

Disclosure statement
No potential conflict of interest was reported by the authors.