Subcritical dimethyl ether extraction as a simple method to extract nutraceuticals from byproducts from rice bran oil manufacture

The byproducts of rice bran oil processes are a good source of fat-soluble nutraceuticals, including γ-oryzanol, phytosterol, and policosanols. This study aimed to investigate the effects of green technology with low pressure as the subcritical fluid extraction with dimethyl ether (SUBFDME) on the amount of γ-oryzanol, phytosterol, and policosanol extracted from the byproducts and to increase the purity of policosanols. The SUBFDME extraction apparatus was operated under pressures below 1 MPa. Compared to the chemical extraction method, SUBFDME gave the highest content of γ-oryzanol at 924.51 mg/100 g from defatted rice bran, followed by 829.88 mg/100 g from the filter cake, while the highest phytosterol content was 367.54 mg/100 g. Transesterification gave the highest extraction yield of 43.71% with the highest policosanol content (30,787 mg/100 g), and the SUBFDME method increased the policosanol level from transesterified rice bran wax to 84,913.14 mg/100 g. The results indicate that the SUBFDME method is a promising tool to extract γ-oryzanol and phytosterol and a simple and effective technique to increase the purity of policosanol. The study presented a novel technique for the potential use of SUBSFDME as an alternative low-pressure and low-temperature technique to extract γ-oryzanol and phytosterol. The combination of transesterification and the SUBFDME technique is a potential simple two-step method to extract and purify policosanol, which is beneficial for the manufacture of dietary supplements, functional foods and pharmaceutical products.


Scientific Reports
| (2020) 10:21007 | https://doi.org/10.1038/s41598-020-78011-z www.nature.com/scientificreports/ the chemical extraction method of transesterification (TE) was used because it could not be directly used by the SUBFE in the DME method. Releasing DME into the environment decreases the temperature of DME to lower than − 11 °C (data from our experiment), which makes RBW freeze (the wax will solidify at temperatures lower than 25 °C). FC was used as a sample to compare the effect of SUBFDME method with the TE chemical reaction.
Nutraceutical extraction of byproducts by SUBFDME. The lab-scale apparatus of SUBFDME with a capacity of 10 g was used to extract nutraceuticals, including γ-oryzanol, phytosterol, and PCs, and a schematic illustration is shown in Fig. 1. The sample was put into a cellulose thimble (Whatman 30 mm × 100 mm), and a known amount of DME was released into the apparatus. The temperature of the reactor, pressure (< 1 MPa), extraction time, and stirring speed rate were set at one condition, and the extraction time was 30 min. The extracted sample was filtered through a metal filter (7 µM) and stored at − 20 °C for further application. The obtained extract from the machine was processed to analyze the bioactive composition.
Nutraceutical extraction of byproducts by TE. The TE was prepared using a modified method of Aryusuk 22 by dissolving samples in a solution of NaOH in EtOH (2%) and stirring at 80 °C for 15 min. The solution was allowed to react and mixed with a warm solution of isooctane and EtOH. The isooctane layer was separated and kept in a refrigerator (4 °C) overnight. The crystallized wax that formed was filtered on a Buchner funnel and washed twice with EtOH. The precipitate was kept and dried in a hot air oven at 60 °C for 5 h. The dried extract was ground and stored at − 20 °C for other applications and chemical analysis.
Increasing the policosanol content from transesterified RBW and FC. RBW and FC were used as samples to increase the PC content (the purity of PCs). Since the results from the first part indicate that TE increased the PC content of the samples, the sample was pretreated by chemical reaction through TE. Then, the SUBFDME extraction was processed for the transesterified samples, and the solvent extraction method (SE) was used to compare the results. The pretreatment conditions of the samples by TE were identical to those mentioned above.
Increasing purity of policosanols by SUBFDME and by SE. To increase the PC content, the transesterified samples were processed in a SUBFE machine. The condition of the SUBFDME is identical to that for the aforementioned nutraceutical extraction. The nutraceutical contents of the obtained PCs powder from the machine were determined. The SE by toluene was used to compare the result with SUBFE. For SE, transesterified samples were extracted according to the method described by Wongwaiwech et al. 1 with some modification. The transesterified sample was dissolved in toluene. The mixture was shaken for 30 min and subsequently centrifuged at 4000 rpm at 10 °C for 15 min. The supernatant was collected into a round flask and evaporated until dry. The residues were gathered and stored at − 20 °C for further analysis.   weighed in a 20-mL headspace vial, and the vial was closed with an aluminum crimp cap equipped with a Tefloncoated butyl rubber septum. The samples were incubated at 70 °C for 30 min and subsequently injected using the  automated headspace sampler. The headspace sampler and transfer line were set at 70 °C and 150 °C. The loop  equilibrium, loop fill, and GC cycle times were 0.05, 0.10, and 50 min, respectively. The instrument assembly consisted of an Agilent Technologies 6890 coupled with an Agilent 7694 Headspace-Sampler. The capillary column was an Agilent DB-5 ms (30 m × 0.25 mm × 0.25 µm, USA). Helium was used as a carrier gas at a flow rate of 1.0 mL/min. The pressure exerted by a constant column of 6.75 psi, whereas the sample inlet was held at 250 °C. Oven temperatures originally set at 35 °C (5 min) were subsequently raised to 300 °C (1 min) at a rate of 10 °C/min. The injector, MS quad temperatures, transfer line, and MS source were 250, 150, 280, and 230 °C, respectively. The residual solvent was identified and quantified by the SIM (single ion monitoring) mode according to their retention times and MS spectra. Isooctane, DME, ethanol, toluene, hexane, acetone and acetaldehyde were used as external standards.
Nutraceutical analyses. Analysis of γ -oryzanol contents. γ-Oryzanol was extracted and determined according to a previous report by Wongwaiwech et al. 1 using LC-MS. Briefly, samples were extracted with a mixture of chloroform and methanol. A solution of 500 µL of supernatant was mixed with a solution of 500 µL of acetonitrile, methanol, and isopropanol, which was subsequently injected into the LC-MS.
γ-Oryzanol was separated on an Agilent Technologies 1100 with a diode array detector (DAD) chromatographic system equipped with an ultraviolet (UV) detector set at 298 and 325 nm. The sample was separated on an Agilent Zorbax Eclipse XDB-C18 column (4.6 m × 150 mm × 5 µm, U.S. A), and the column temperature was set at 40 °C. The mass spectrometer was an Agilent Technologies LC/MSD SL equipped with an electrospray ion source (ESI). The ESI-MS spectra were acquired in the positive ionization mode with a capillary voltage of 4000 V, nebulizer pressure of 50 psi, gas temperature of 350 °C, drying gas of 13.01 L/min and recorded on a mass range of m/z 200-800. A standard mixture of γ-oryzanol was used as an external standard to identify the peaks by Agilent Mass Hunter software based on their retention times.
Analysis of phytosterol contents. The phytosterol composition of the samples was analyzed using GC-MS according to a previous method 1 . In summary, the sample was extracted by 60% KOH, 95% ethanol, and 10% NaCl under nitrogen gas (N 2 ). The saponified solution was extracted twice with a mixture of hexane and ethyl acetate (9:1, v/v). The unsaponifiable residue was collected and evaporated to dryness. The derivatization was performed by N,O-bis (trimethylsilyl)-trifluoroacetamide, and the quantification and identification of phytosterol were performed by an Agilent Technologies 7683 on DB-5 ms (30 m × 0.25 mm × 0.25 µm, USA). Trimethylsilyl-phytosterols were identified and quantified in the SIM mode according to their mass spectra and retention times.
Analysis of PCs contents. The extraction of PCs was performed according to a previous report by Wongwaiwech et al. 1 . All crude extracts were extracted by the saponification reaction. In brief, a sample was extracted with 0.2 M NaOH (10 mL) in methanol solution and subsequently extracted again with toluene. The upper layer was collected and filtered through a 0.45-µm filter. Identification and quantification of the PCs were determined using an Agilent Technologies 6890 fitted with an Agilent DB-5 ms fused silica capillary column (30 m × 0.25 mm × 0.25 µm, USA). The SIM mode was set for identified and quantified PCs; docosanol (C 22 ), tetracosanol (C 24 ), hexacosanol (C 26 ), octacosanol (C 28 ) and triacontanol (C 30 ) were identified and quantified according to their molecular target ion and retention times.
Statistical analysis. Analysis of variance (ANOVA) was performed to analyze the data with Duncan's tests using the SPSS 19 software (SPSS Inc., Chicago, IL, USA). The significant difference level was set at 0.05. Each reported value is expressed as the mean ± standard deviation (SD) based on the dry weight from three replications.

Results and discussion
Extraction yield and nutraceutical contents by SUBFDME and TE. The contents of γ-oryzanol, phytosterol, and PCs from the crude extract were investigated by the SUBFDME and TE methods. The samples extracted by SUBFDME were a yellowish brown oily substance, while those extracted by the chemical reaction of TE were in a solid form, and their color was similar to their origin samples (Fig. 2a). The extraction yield and conditions for SUBFDME and TE are shown in Table 1, and the nutraceutical contents, including γ-oryzanol, phytosterol and PCs, by SUBFDME and TE are shown in Table 2. For SUBFE, DFRB-C gave a higher extraction yield (9.71%) than DFRB-S (3.60%). From Table 2, our previous study 1 reported that the DFRB-C byproduct contained more γ -oryzanol, phytosterol, and PCs (280.74 mg/100 g) than DFRB-S (77.93 mg/100 g). However, the total amount of nutraceutical compounds found from DFRB-C by SUBFE was slightly lower (1194.96 mg/100 g) than that found from DFRB-S (1233.51 mg/100 g) using the same extraction technique. Since we focused on three bioactive compounds, the higher extraction yield from DFRB-C might contribute to the oil contents and other bioactive components, such as tocopherol, tocotrienal, squalene, phytic acid, lecitin, inositol and wax 3,24 . There are reports of higher amounts of tocopherol tocotrienal, phenolic, phytic acid, inositol and γ-oryzanol in defatted rice bran 24,25 . Compared to the report by Wongwaiwech et al. 1  www.nature.com/scientificreports/ higher amounts of γ-oryzanol and phytosterol were found from crude extracts from SUBFDME. The results suggest a high ability of SUBFDME to release γ-oryzanol (924.51 and 737.46 mg/100 g) and phytosterol (257.12 and 367.54 mg/100 g) from these byproducts ( Table 2). For both SUBFDME and TE extraction methods, the most abundant phytosterol contents in the crude extract from all byproducts were stigmasterol and β-sitosterol, and the data were consistent with our previous report 1 . These data were also confirmed by Derakhshan-Honarparvar et al. 26 and Sawadikiat and Hongsprabhas 27 , which shows that the predominant forms of phytosterol components in rice bran were stigmasterol and β-sitosterol. Regarding FC, SUBFE gave a higher extraction yield (52.14%) and bioactive compounds, including γ-oryzanol (829.88 mg/100 g) and phytosterol (312.34 mg/100 g), than TE (18.82%). There was a slightly lower oryzanol content (829.88 mg/100 g) of crude extract from the SUBFDME than the report of oryzanol content in FC (1058.28 mg/100 g) by Wongwaiwech et al. 1 . This is due to the analysis of γ-oryzanol by Wongwaiwech et al. 1 , who used the solvent extraction method, which is well known for its high ability to extract bioactive compounds 28 . Compared to SUBFDME, the TE method for FC gave dramatically higher PC contents (6100.12 mg/100 g). There was no significant difference in PC contents of the crude extract from all byproducts by SUBFDME; therefore, SUBFDME had no effect on PC contents regardless of the type of byproduct. Wongwaiwech et al. 1 reported a high amount of PCs in RBW (332.79 mg/100 g). Many 1 reports indicate that RBW is a good source of PC compound 29 . Interestingly, the TE method released very high amounts of PCs from RBW at 30,787.89 mg/100 g (Fig. 2a), which is nearly 93 times the amount previously reported by Wongwaiwech et al. 1 . The chemical reaction of TE is very efficient as an extraction method for PCs from RBW 30,31 . Wang et al. 32 and Ning-ning et al. 33 used the transesterification method to extract policosanol from rice bran wax and found that the yield of policosanol was approximately 21%. Here, we found a lower yield of PC (13.46%) by TE extraction compared to Wang et al. (a) Byproducts of rice bran oil refining processes and the resulting crude wax extracted by SUBFDME and TE; (b) PCs extract by SUBFDME and SE of transesterified RBW and FC. DFRB-S defatted rice bran from the solvent extraction process, DFRB-C defatted rice bran from the cold pressed extraction process, FC filtered cake, RBW rice bran wax, TE extracted by transesterification, SUBFDME extracted by the subcritical dimethyl ether extraction, SE extracted by the solvent extraction. Table 1. Comparison of the extraction yield of nutraceutical from byproducts using the SUBFDME and TE processes. DFRB-S defatted rice bran from the solvent extraction system, DFRB-C defatted rice bran from the cold-pressed extraction system, FC filtered cake from the cold-pressed extraction system, RBW rice barn wax from the solvent extraction system, SUBFDME extracted by the subcritical dimethyl ether extraction, TE tranesterification. The results indicate that a powerful extraction technique for γ-oryzanol and phytosterol is the SUBFDME method, whereas the TE method is effective in PC extraction. Since DME can dissolve a wide range of nonpolar substances 34 , it can increase mass transfer by developing hydrogen bonds with extractable substances 35 . The principle of subcritical fluid extraction promotes the DME temperature to rise above the boiling point and applies sufficient pressure to help maintain the DME in a liquid state. Under such conditions, DME has features that help promote the extraction process, such as high diffusion coefficients, low viscosity and high solvent strength. Furthermore, an increased temperature produces high solubility and high diffusion rates of the solutes in the solvent, while pressure helps to force DME into the sample matrix and allows it to be filled faster 36 . The data also suggest that the SUBFDME technique can liberate fat soluble bioactive compounds, including γ-oryzanol and phytosterol, and it liberates the same amount of PCs regardless of the types of sample tested.

Increasing PC recovery of transesterified rice bran wax and filter cake by SUBFDME. Both SE
and SUBFDME methods can be used to increase the PC content of transesterified FC and RBW; therefore, they were compared in terms of PC content and purity. From the first part of the results, the data suggest that the TE extraction technique released a large amount of PC compounds from both types of byproducts, FC and RBW. FC and RBW were promising sources of PCs and selected as samples to process to increase the PC content. The extraction yields of crude PCs extract from transesterified FC and RBW are shown in Table 3. The extraction yields of crude PCs from transesterified FC and RBW by the SE method (10.84-18.24%) were higher than those from SUBFDME (1.32-2.49%). The color of crude PCs extracted by the SUBFDME and SE methods changed from a yellowish-brown powder to pure white and yellowish-white powder, respectively (Fig. 2b). The color of Table 2. Comparison of γ-oryzanol, phytosterol and policosanol contents extracted from byproducts using various extraction methods. Each value represents the mean ± S.D. Values with different superscript letters in the same row are significantly different (P < 0.05). Values in the table are expressed on a dry basis. The γ-oryzanol, phytosterol and policosanol contents of the original samples were extracted by the solvent extraction method (data from our previous study 1 . TE extracted by transesterification, SUBFDME extracted by the subcritical dimethyl ether extraction, DFRB-S defatted rice bran from the solvent extraction system, DFRB-C defatted rice bran from the cold pressed extraction system, RBW rice bran wax from the solvent extraction system, FC filtered cake from the cold pressed extraction system.  Table 3. Comparison of extraction yields from transesterified RBW and FC using the SUBFDME and SE methods. TE-samples tranesterified samples, RBW rice barn wax from the solvent extraction system, FC filtered cake from the cold pressed extraction, SE extracted by the solvent extraction (toluene) , SUBFDME extracted by the subcritical fluid dimethyl ether extraction.  37 reported that crude PC extracted from RBW appeared dark brown, but after the purification step, the color of the obtained PCs changed to white powder. A comparison of γ-oryzanol, phytosterol and PC contents of extracts from transesterified FC and RBW using the SUBFDME and SE methods is shown in Table 4. The results show that a high amount of γ-oryzanol was detected in the extract from transesterified FC (258.02 mg/100 g), followed by transesterified RBW (114.37 mg/100 g) by the SE method, whereas it was not detected by SUBFDME. These data suggest that SE, not SUBFDME, increases the γ-oryzanol content of the extract from transesterified FC and RBW samples by 3 and 2.6 times, respectively ( Table 2). The phytosterol contents of transesterified FC and RBW were 93.28 and 59.71 mg/100 g, respectively ( Table 2). The SE method increased the phytosterol content of the extract of transesterified FC and RBW by 6-and 7.7-fold, respectively. Although SUBFDME increased the phytosterol contents of transesterified samples, the increase in phytosterol by SE was more effective than that by SUBFDME. However, the results from Table 2 indicate that direct extraction of byproducts by SUBFDME was the most effective method for both oryzanol and phytosterol, since there was a higher yield with a shorter extraction time than chemical extraction. Chotimarkorn et al. 38 , Lilitchan et al. 39 and Lai et al. 40 applied a solvent (methanol) extraction method to extract gamma-oryzanol from Thai rice bran and Japonica rice bran. They reported that the yields of oryzanol were 56.0-108.0 mg/100 g, 343-367 mg/100 g and 160-180 mg/100 g, respectively. Bhatnagar et al. 41 used Soxhlet extraction to extract phytosterols from broken rice and rice germ, and the yields of phytosterols were 12.87 and 76.96 mg/100 g, respectively. Compared with solvent extraction [38][39][40][41] , this study confirmed that SUFDME was significantly more effective for gamma-oryzanol and phytosterol extraction from byproducts (DFRB-S (924.51 and 257.12 mg/100 g, respectively), DFRB-C (737.46 and 367.54 mg/100, respectively), and FC (829.88 and 312.34 mg/100 g, respectively). The PC content was significantly different from the γ-oryzanol and phytosterol contents. The SE method increased the PC content of transesterified FC and RBW to 72,318.21 and 62,717.72 mg/100 g, respectively, while SUBFDME was more effective in increasing the PC content of transesterified samples (84,398.86 and 84,913.14 mg/100 g, respectively). The distinctly increasing amounts were increased by 14 and 3 times. This result suggests that FC and RBW are good sources of PCs, and the combination methods of TE and SUBFDME are promising methods to extract and purify PCs from FC and RBW, which are coproducts from the rice bran oil process. Many studies focused on extraction and purification methods for policosanol, such as solvent extraction, supercritical fluid extraction, ultrasonic-assisted extraction, and other extraction methods. Chen et al., 2007 42 reported that the extraction of rice bran wax by transesterification and purification by molecular distillation made the yield of policosanol 53.8%. Lorenz et al. 43 used the supercritical carbon dioxide extraction method to isolate policosanol from quine wax and achieved a PC content of 36,410 mg/100 g. Ishaka and colleagues 29 used solvent ultrasonic-assisted extraction to extract PCs from rice bran wax and rice bran oil, and the yields of policosanol were 10,820 and 9810 mg/100 g, respectively.
This study is the first report using SUBFDME to increase the purity (approximately 84%) of PC compounds. The data also indicate that the remaining (16%) PC compounds may be dotriacontanol (C 32 H 66 O), tetratriacontanol (C 34 H 70 O), phytosterol and vit E (data not shown). Table 4. Comparative contents of gamma-oryzanol, phytosterol and policosanol from tranesterified RBW and FC using the SUBFDME and SE techniques. Each value represents the mean ± SD. Values with different superscript letters in the same row are significantly different (P < 0.05). Values in the table are expressed on a dry basis. ND amount detected below the LOD; the LOD of gamma-oryzanol is 0.05 ppm, TE-samples tranesterified samples, SUBFDME extracted by the subcritical fluid dimethyl ether extraction, SE extracted by solvent extraction (toluene) , RBW rice bran wax from the solvent extraction system, FC filtered cake from the cold-pressed extraction system.