Variations in compositions and antioxidant activities of essential oils from leaves of Luodian Blumea balsamifera from different harvest times in China

Xanthoxylin was the main compound (content 44.92% of total volatiles) in the leaves of Luodian B. balsamifera, which might be the key cause of failure in collecting essential oil (EO) of the leaves using general hydrodistillation in Clevenger apparatus. A modified hydrodistillation equipped with Clevenger apparatus was designed for isolating EO from the leaves. Six EOs of Luodian B. balsamifera harvested once a month from September to next February were collected successfully. The main components of EOs were δ-elemene, α-cubenene, caryophyllene, caryophyllene epoxide, γ-eudesmol, xanthoxylin, and α-eudesmol. The EOs of Luodian B. balsamifera collected from October to December had higher antioxidant activities (ACs). Combining the principal component analysis of chemical components with the results of ACs and the yields of six EOs, the leaves of Luodian B. balsamifera were suitable to be harvested in November and December to obtain EO with high quality.


Extraction of total volatiles
Fifty grams dried leaves of B. balsamifera were distilled to collect 5 L of distillate (distillation time: 10 h), and the distillate was collected for extracting total volatiles (TV) of the leaves using same volume anhydrous diethyl ether under room temperature. The extract solution was dehydrated with anhydrous sodium sulfate and concentrated to obtain total volatiles. The TV was stored at -20˚C in dark glass bottles until required. The process was executed in ten replicates.

Hydrodistillation
The general hydrodistillation was executed in Clevenger apparatus. Fifty grams of leaves samples and 1 L distilled water were added into 2 L flask and distilled for 10 h in Clevenger apparatus (Jingbo, Jintan, China) ( Fig 1A). In the end, there was no EO to be collected. However, there were white volatile on the surface of condenser inner wall and an aqueous phase (aromatic water) in Clevenger apparatus. The aqueous phase was extracted by diethyl ether anhydrous (100 mL), and the extraction solution was concentrated under vacuum. White volatiles (WV) and the extract of aqueous phase (EAP) were weighed and stored at 4˚C in dark glass bottles until analysis, respectively. The experiments were executed with six replicates.

Modified hydrodistillation procedure
Fifty grams of leaves samples and 1 L distilled water were added into 2 L flask. At first, the mixture was heated to boiling temperature, and 500 mL distillate was collected and then extracted using diethyl ether anhydrous (500 mL), the extraction solution was concentrated to obtain the extract which was named "Aifen" (AF). The residual mixture was added with 500 mL of deionized water and then connected to the Clevenger apparatus for being distilled continuously. In the end, two phases were observed in Clevenger apparatus, an EO and an aqueous phase ( Fig 1B). The EO was collected and dried with anhydrous Na 2 SO 4 . The yield of EO was used as a reference to optimize the distillation time (from 2 h to 10 h). AF and EO were stored at 4˚C in dark glass bottles until analysis. The processes were executed with ten replicates.

Gas chromatography (GC)
The oil samples for GC analysis were prepared in 99% n-hexane. The final concentration was 0.5 mg/mL. GC analysis was executed on a Shimadzu GC-2010 gas chromatograph operated with a split/splitless injector and a Shimadzu AOC-20i autoinjector (Shimadzu, Kyoto, Japan The internal standard method was used in the quantitative determination of l-borneol and xanthoxylin and was performed on GC-FID system mentioned above. Naphthalene was used as the internal standard (IS). The samples and IS were prepared in ethyl acetate, and the final concentration was 0.2 mg/mL. The validation of quantitative determination as follows: the linear range was 0.6-1875 μg/mL for l-borneol and xanthoxylin; the accuracy was verified by eight replicate analyses with standard solution, the calculated RSD values for l-borneol and xanthoxylin were 0.16% and 0.21%, respectively; moreover, the recoveries of two compounds were 97% and 96%.

Identification of constituents
Individual compound was identified by the following two methods, (1) mass spectrum was compared with those reported in NIST and WILEY libraries, Adam's records and those of literature data; (2) Linear RIs relative to (C 8 -C 30 ) n-alkanes were used in the identification of chemical components and matched with those of our own authentic compounds, Adam's records and literature data.

Quantification of constituents
The quantitation of EO components was based on the peak area normalization with response factors (RFs). The constituents of EO were complex, and these compound standards were lack, so it was unrealistic to determine the RFs of all compounds. Therefore, we used a solution introduced by Costa et al. [24], based on EO constituents classified by their functional groups and chemical classes. RFs are the average values of the response factors measured by individual standard compounds within the same chemical class and are recorded in Table 1. The quantification was applied in the concentration determination (%) of all oil components.

Antioxidant activity
Antioxidant activities of EOs were evaluated by three tests (DPPH radical scavenging test, βcarotene bleaching (BCB) test, and thiobarbituric acid reactive species (TBARS) assay), and three tests referred to our previous experimental process [23].

Statistical analysis
All experiments were performed in five replicates, and the data were expressed as mean values ± standard deviation. The analysis of variance (ANOVA) was used in the statistical analysis of the data and was computed in SPSS Statistics 18.0 (SPSS Inc., Shanghai, China) software, and a probability value of P < 0.05 was considered to represent a statistically significant difference among mean values. Principal component analysis was calculated in Origin Pro 2016 (OriginLab Corporation, USA) software.

Hydrodistillation
In the general hydrodistillation process, EO should float on the surface of aqueous phase and be collected in Clevenger apparatus ( Fig 1A) [19]. Unfortunately, EO was not collected on the surface of aqueous phase, while a white volatile matter (WV) and an aqueous phase (AP) were observed in the condenser and collecting tube. In the previous study, there was a large number of high purity l-borneol on the inner wall of the condenser in hydrodistillation [22]. As seen from Table 2, the mean yields of WV and the extract of aqueous phase (EAP) were 0.58% and 0.80% (w/w), respectively. The sum of two yields was 1.38% (w/w) and was less than the yield of total volatiles (TV, 2.42%, w/w). Thus, the volatiles of Luodian B. balsamifera leaves did not entirely exist in the collecting tube of Clevenger apparatus. Qualitative and quantitative analyses of WV and EAP are shown in Table 3, l-borneol (content 90.81%) was the main component of WV. It meant that l-borneol was mainly present on the surface of the condenser inner wall of the glassware (Clevenger apparatus), and it had little effect on the extraction of EO. Xanthoxylin (content 34.96%) and l-borneol (content 17.82%) were the key components of EAP, and xanthoxylin was the important component dispersed in AP during the process of general hydrodistillation. In other words, xanthoxylin could affect the collection of EO. Besides, the contents of camphor, α-cubenene, and caryophyllene in WV were more than 1%, and the contents of caryophyllene, caryophyllene epoxide, γ-eudesmol were above 5% in EAP. The contents of l-borneol and xanthoxylin in TV, WV, and EAP were calculated and compared. The sum of l-borneol contents in WV (53.54%) and EAP (14.14%) accounted for 67.68% of total l-borneol content in leaves. The content of xanthoxylin was rare in WV, and 25.69% of total xanthoxylin was present in EAP. Other parts of total l-borneol and xanthoxylin were still kept in the leaves.
For the failure of extracting EO by the general hydrodistillation, we assumed that one or some volatile components had adverse effects on the extraction process of EO. As seen from Table 3, the melting points of several compounds extracted from TV and AP were lower than 80˚C, as a result, these compounds in AP (measured temperature: 85˚C) were liquid state in the collecting tube, such as caryophyllene, dehydro-aromadendrane, caryophyllene epoxide, Table 2. Yields of l-borneol and xanthoxylin in the different volatiles from the leaves of Luodian Blumea balsamifera.     RIs reported from NIST standard library mass spectral data and Adam's records [25]; guaiol, 10-epi-γ-eudesmol, γ-eudesmol, β-eudesmol, α-eudesmol, xanthoxylin, etc. A large number of components in volatiles possessed lower density than water and would float on the surface of aqueous phase. However, the density of xanthoxylin (1.172 g/mL) was higher than water, and xanthoxylin would sink to the bottom of the water. Besides, xanthoxylin was the main component of TV (content 44.92%) and EAP. Moreover, the concentrate of xanthoxylin was at a high level (28 mg/mL) in AP in the collecting tube. According to the physicochemical properties of xanthoxylin, we speculated that xanthoxylin dispersed into AP and made the property of AP similar to organic solution. Low-density components would also follow xanthoxylin disperse into AP rather than float on the surface of AP. The total concentration of volatile components in AP was at a high level (80 mg/mL). Thus, the volatile components and water circulated simultaneously in Clevenger apparatus. The coaction of these volatile components made AP play the role of organic solvent to dissolve EO so that EO could not float on the surface of AP to be collected. In a word, the high content of xanthoxylin was the primary cause of the failure to obtain EO of Luodian B. balsamifera leaves.

Isolation of essential oil by the modified hydrodistillation
As inferred from the above deduction, if a part of xanthoxylin is removed from the leaves, we designed the modified hydrodistillation procedure to isolate EO of Luodian B. balsamifera leaves ( Fig 1B). Meanwhile, the design also referenced the improved hydrodistillation in the previous study [22,23]. However, the objective of the previous method was the efficient extraction of l-borneol and was different from this study. In the beginning, the first 500 mL distillate was distilled to remove partial volatiles (namely "Aifen") from the leaves, and then the residual leaves were further distilled for collecting EO using Clevenger apparatus. Fortunately, we successfully isolated EO of Luodian B. balsamifera leaves, and the modified procedure showed that the above deduction was credible. Next, the distillation time was optimized, as seen in Fig  2. The yield of EO did not significantly increase after 8 h. Therefore, eight hours was considered to be the optimal distillation time. As shown in Table 2, the yield of "Aifen" (AF) was 1.16% (w/w), accounted for 47.93% of total volatiles (2.42%, w/w) in leaves (Fig 3B). It is important to know whether the modified procedure negatively affects the chemical composition of EO or not. Firstly, two key components (l-borneol and xanthoxylin) were investigated. l-Borneol (content 68.12%) and xanthoxylin (content 22.03%) were two main compositions in AF, and the contents of other volatiles were less (content < 2%). Moreover, lborneol existed in AF was 79.80% of total l-borneol in leaves. According to the result of general hydrodistillation, the internal surface of condenser was stained with white volatile matter (WV), and l-borneol was the main constituent (content 90.81%) of WV. Because the sublimation property of l-borneol, even if l-borneol was not distilled and removed in the first 500 mL distillate, l-borneol was not in the EO and existed on the surface of condenser inner wall of the glassware.
Besides, the content of xanthoxylin was monitored in the hydrodistillation process. As shown in Fig 3A, the yield of xanthoxylin decreased gradually with the increase of distillation volume, and 23.85% of total xanthoxylin was distilled and removed in the first 500 mL distillate. The above hydrodistillation rule had been confirmed in simultaneous distillation and extraction in the previous study [16]. Therefore, we believed that the removal of partial l-borneol and xanthoxylin did not affect the composition of EO. Likewise, the effects of other low content volatile components were little and could be ignored in the collection of EO. Thus, the modified hydrodistillation procedure was feasible to collect EO of Luodian B. balsamifera leaves.
To have an overview of all results and to analyse the difference among six EOs harvested at different times. A principal component analysis (PCA) was applied to the data set of six EOs. Fig 5 shows the score-plot for PC1 and PC2 from PCA of different EO samples. PC1 and PC2 explained a great part of the total variance (70.80%). The score-plot was better to understand the relationships between harvest times of the leaves and chemical compounds of EOs. The regions occupied by the six EO samples were distributed in three parts (Fig 5). PCA shows that the chemical components of EO2-EO4 belonged to one category, and were different from that of EO1, EO5, and EO6. It indicated that there were significant differences in chemical components of B. balsamifera at different growth stages. However, the chemical components of EOs

Antioxidant activity
Antioxidant activities (ACs) of six EOs were evaluated by DPPH radical-scavenging test, β-carotene bleaching (BCB) test, and thiobarbituric acid reactive species (TBARS) assay. The results of AC tests are shown in Fig 6. Six EOs exhibited ACs in three tests, which were in agreement with our previous study [23]. Several studies have reported that the ACs of volatile oils and EOs of B. balsamifera were significantly different in several cultivation areas [4,6,15,17,18,23]. However, to our knowledge, there is no study on the ACs of EOs of Luodian B. balsamifera in different growth stages. DPPH test, BCB test, and TBRAS test exhibited that the ACs of EO sample No. 2-4 were significantly higher (P <0.05). It indicated that the ACs of EOs from the leaves of B. balsamifera growing from October to December were at a high level. This observation is consistent with previous studies [13]. The joint interpretation of PCA of chemical components and AC tests presented that the AC of EO from B. balsamifera was highly correlated with the chemical composition of EO. Besides, the yields and the ACs of EOs from B. balsamifera harvested in November and December were high, and the leaves of B. balsamifera was suitable for harvesting and extracting EO in these two months.

Conclusion
The high content of xanthoxylin in the leaves of Luodian B. balsamifera was considered to be the primary cause of the unsuccessful collection of EO using Clevenger apparatus in the general hydrodistillation. A modified hydrodistillation was designed to isolate EO from Luodian B. balsamifera leaves. Six EOs of Luodian B. balsamifera leaves harvested from different growth periods were successfully obtained by the modified hydrodistillation. The EOs contained a considerable amount of caryophyllene and xanthoxylin, which was significantly different from B. balsamifera cultivated in other regions. The chemical components of EOs from B. balsamifera leaves harvested from October to December were comparatively consistent, and these EOs had good ACs. Considering the yields and the ACs of EOs, for extracting EO with high quality, Luodian B. balsamifera leaves should be harvested in November and December. The EO may be used extensively in pharmaceutical, food, and cosmetic industries.
Supporting information S1 Data. Data set-raw data of tables and figs. (XLSX)