Essential Oils from The Dropped Flowers of Camellia japonica: Extraction Optimization, Chemical Profile and Antibacterial Property

These authors contributed equally Abstract: Camellia japonica is one of the most economically valuable species that can be taxonomically placed in the Theaceae family. Many horticultural varieties of C. japonica have been cultivated as ornamental trees worldwide, owing to their charming flowers. However, young C. japonica flowers usually drop in a “decapitated” form during the peak flowering stage, causing a resource waste. In this study, Essential Oils from the Dropped C. japonica Flowers (EODCJF) were extracted with hydrodistillation and process was optimized by response surface methodology. The chemical profile and antibacterial activity were tested by gas chromatography-mass spectrometry and by disc diffusion, respectively. The optimal extraction parameters were: Immersion time of 9 h, extraction time of 8 h and liquid-to-solid ratio of 39: 1 (mL/g), under these conditions, the yield was 3.20%. The actual yield was in compliance with the predicted result with an error margin of 3.61%. Thirty-three volatile compounds accounting for 94.46% were authenticated in EODCJF with alkanes of 53.96% and alcohols of 28.59%, respectively. EODCJF exerted certain antibacterial capacities against Staphylococcus aureus, Escherichia coli, Bacillus subtilis and Bacillus pumilus. Their antibacterial activities could be contributed by the long-chain alcohols in EODCJF and alkanes may potentiate the dispersion of these liposoluble alcohols in EODCJF.


Introduction
Essential Oils (EOs) are a kind of secondary metabolites in plants. They usually comprise diverse components that bear smaller molecular weights and aromatic odors (Zhang et al., 2018a). According to the chemical profile difference, EOs can be classified into aromatics, aliphatics and terpenoids, most of which are featured by a higher volatility and a lower boiling point (Tian et al., 2019). EOs exert a series of bioactivities, such as antioxidant (Imane et al., 2020), antitumor (Andrade et al., 2018), antimicrobial (Zhang et al., 2018b), anti-inflammatory (Sen et al., 2019) and immunomodulatory (Valdivieso-Ugarte et al., 2019) properties. Thus the search for novel bioactive EOs remains ongoing concern.
The genus Camellia is native to Southeast Asia and China. Up to now, approximately 280 species of Camellia have been found to be placed in the Theaceae family taxonomically, such as Camellia sinensis var. sinensis, Camellia sinensis var. assamica, Camellia japonica, Camellia oleifera, Camellia chrysantha and Camellia reticulate (Li et al., 2019). Among them, C. japonica is one of the most economically valuable species. Many horticultural varieties of C. japonica have been developed as ornamental trees worldwide, due to the beautiful and charming flowers (Kim et al., 2019;Páscoa et al., 2019). In addition to horticultural application, C. japonica and its principles elicit various beneficially physiological and pharmacological effects. The ethanol extract from flowers of C. japonica exerts antioxidant capacity via scavenging excessive free radicals and reactive oxygen species and via elevating activities of antioxidases (Piao et al., 2011). The quercetin glycosides, gallic acid and its glucoside conjugates isolated from the hot water extract of C. japonica flowers show higher 1, 1-Diphenyl-2-Picrylhydrazyl (DPPH) radical-scavenging capacity than those of other compounds, indicating that the antioxidant activity of C. japonica flowers could be chiefly contributed by these bioactive phenolic acids and flavonoid glycosides (Lee et al., 2011). The Oleanane triterpenes isolated from C. japonica flowers exhibit inhibitory effects against the replication of porcine epidemic diarrhea virus (Yang et al., 2015) and the methanol extracts from the leaves and flowers of C. japonica display certain inhibitory effects against some food-borne pathogen and food spoilage bacteria (Lee et al., 2005). However, to the best of our knowledge, there is little information on the research about the essential oils from C. japonica flowers.
Unlike other ornamental flowers, C. japonica flowers usually drop in a young "decapitated" form during the peak flowering stage, which causes a tremendous waste of this resource. In present work, Essential Oils from the Dropped C. japonica Flowers (EODCJF) were extracted for The first time and the extraction process was optimized. After characterization, the antibacterial potential was appraised. It is hoped that this study will highlight the comprehensive development and utilization of dropped C. japonica flowers.

Materials
The dropped C. japonica flowers were collected from April to June 2020 in the campus of Changshu Institute of Technology (Changshu, Jiangsu, China) and authenticated by Dr. Zhumei Cui, School of Biology and Food Engineering, Changshu Institute of Technology. After being dried at 50-60℃, the dried C. japonica flowers were deposited at the School of Biology and Food Engineering, Changshu Institute of Technology with a voucher number of SW-2020-068. Before the experiment began, dried C. japonica flowers were sheared and sieved to 40-mesh.
All the reagents and solvents involved were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd (Shanghai, China) and were of analytical grade.

Essential Oils Extraction
Essential Oils from the Dropped C. japonica Flowers (EODCJF) were extracted with hydrodistillation method in a Clevenger-type apparatus (Memarzadeh et al., 2015). Twenty grams of dried powders of C. japonica flowers were immersed in a certain volume of distilled water for several hours. After a certain time of extraction starting from water boiling, EODCJF were obtained and then stored at 0℃ in a sealed vial. The extraction yield (w/w, %) was calculated according to the following formula: = The weight of obtained essential oils (g) W2 = The weight of dried powders of C. japonica flowers (g)

Experimental Design for Extraction Optimization
First, single-factor experiment was performed to assess the effects of three independent variables, including immersion time (2-12 h), liquid-to-solid ratio (10-60: 1 mL/g) and extraction time (2-12 h) on EODCJF yield. Then, the extraction process was optimized with Response Surface Methodology (RSM) (Cui et al., 2018). A three-factor-three-level Box-Behnken Design (BBD) encompassing seventeen runs was applied at the center point (Table 1).
Regression analysis was conducted to analyze data and fitted to the second-order polynomial model as below: The predicted extraction yield of EODCJF β0 = An intercept βi = The linear coefficient βii = The quadratic coefficient βij = The interactive term Xi and Xj = The coded independent variables

The Chemical Profile of Essential Oils
The chemical composition of EODCJF was determined by a gas chromatography-mass spectrometry (GCMS-QP2010, Shimadzu, Kyoto, Japan) equipped with a Rxi-5Sil MS column (30 m ×0.25 mm ID, 0.25 μm of film thickness) and interfaced with an ion trap detector. Analysis was operated under the following conditions: (1) Helium was adopted as the carrier gas at a flow rate of 1.0 mL/min. (2) Temperatures of transfer line and ion trap were set as 250 and 230℃, respectively. (3) Column temperature was initially raised from 40 to 60℃ at 5℃/min and kept for 2 min, then elevated to 110℃ at 15℃/min and held for 4 min, from 110 to 180℃ at 5℃/min and maintained for 2 min, finally increased up to 280℃ at 25℃/min and kept for 5 min. One microliter of EODCJF diethyl ether solution at concentration of 10 mg/mL was injected into the analyzer at a split ratio of 1: 30 under a scan range of 35-500 amu and an electron ionization energy of 70 eV. Compound identification was assigned by referring the retention index of nalkanes (C8-C40) standards and by comparing the mass spectrum with NIST 05 database or with reported literatures (Tian et al., 2019).

The Antibacterial Activity of Essential Oils
The paper disks with diameter of 6 mm were soaked by EODCJF in Dimethyl Sulfoxide (DMSO) and placed on the surface of solid medium that had been inoculated with the test bacteria (S. aureus, E. coli, B. subtilis and B. pumilus) at 6.0 log CFU/mL. Normal saline in Dimethyl Sulfoxide (DMSO) and chloramphenicol in DMSO were served as negative and positive controls, respectively. All the solid medium plates were cultured at 37℃ for 24 h. Diameter of inhibition zones was recorded (Zhang et al., 2018b).

Statistical Analysis
Extraction optimization was designed and analyzed with a Design-Expert software 8.0.6.1 (Stat-Ease, Minneapolis, MN, USA). Experimental data was represented as means or means ± SD (standard deviation). SPSS19.0 Software (SPSS Inc., Chicago, USA) was used to appraise the statistical difference. The significance of distances between means was evaluated with t-test and one-way Analysis Of Variance (ANOVA).

Effects of Immersion Time on Yield
As shown in Fig. 1, when liquid-to-solid ratio and extraction time were fixed as 40: 1 (mL/g) and 8 h, respectively, EODCJF yield increased with the prolongation of immersion time ranging from 2 to 8 h, but when immersion time exceeded 8 h, yield decreased. The optimal immersion time was 8 h.

Effects of Extraction Time on Yield
Fix liquid-to-solid ratio of 40: 1 (mL/g) and immersion time of 8 h to explore the effects of extraction time on yield. As shown in Fig. 2, in the range of 2-8 h, EODCJF yield was raised with the increase of extraction time, maximized to 2.83±0.12% at 8 h and then declined with lengthening extraction time.

Effect of Liquid-to-Solid Ratio on Yield
As shown in Fig. 3, when immersion time and extraction time were all set as 8 h, EODCJF yield was heightened within the liquid-to-solid ratio of 10:1 (mL/g)-40: 1 (mL/g), maximized to 3.3±0.1% at the ratio of 40: 1 (mL/g) and then dropped with the elevation of liquid-to-solid ratio.

The Extraction Optimization of EODCJF with RSM
Based on principles of BBD, according to the results of single-factor experiments, three independent variables, including immersion time (h, X1), extraction time (h, X2) and liquid-to-solid ratio (mL/g, X3) were selected and applied to RSM optimization with EODCJF extraction yield as response variable. Seventeen possible orderings and response values were achieved and represented in Table 2. It can be observed that the EODCJF yields were within a range of 1.88-3.30%, which can be described by the following second-order polynomial equation: The one-way Analysis Of Variance (ANOVA) for fitted equation was shown in Table 3. For the source of model, the high F-value (148.11) and low P-value (<0.0001) demonstrated that the obtained model is highly significant and the regression equation is desirable (Cheng et al., 2017). The determination coefficient of model (R 2 ) and adjusted coefficient (Adj R 2 ) were 0.9948 and 0.9881, respectively. These values implied that the accuracy of model is sufficient to interpret 98.81% of the total variability of model response. For lack of fit, the F-value and P-value were all greater than 0.05, indicating a non-significance, which further validated the model. The linear terms X1 and X2 were highly significant (P<0.01) and X3 was not significant (P>0.05). The quadratic parameters X1 2 , X2 2 and X3 2 were all highly significant (P<0.01). The interaction parameters X1X2 were highly significant (P<0.01), but X1X3 and X2X3 were not significant (P>0.05).
To better understand the interaction effects between two independent variables, the contour plots (Figs. 4 to 6) were achieved by RSM. From Fig. 4, it can be clearly seen that the increase in immersion time or extraction time potentiated the improvement of EODCJF yield. In Fig. 5, EODCJF yield increased with the increase of immersion time or liquid-to-solid ratio in the initial stage, but experienced a slow elevation with the further increase of any parameters. From Fig. 6, we can observe that EODCJF yield was raised with increasing extraction time or liquidto-solid ratio at first, but obviously descended afterwards. These results and trends were consistent with those of single-factor experiments.
The optimized extraction parameters obtained from Design-Expert 8.0.6.1 were as follows: Immersion time of 8.69 h, extraction time of 8.24 h, liquid-to-solid ratio of 39.26: 1 (mL/g) and the predicted yield was 3.32%. In order to facilitate practical operation, the predicted optimal conditions were further modified as: Immersion time of 9 h, extraction time of 8 h and liquid-to-solid ratio of 39: 1 (mL/g), under which the actual yield was 3.20%. The actual results were in line with the predicted data with an error margin of 3.61%.

The Chemical Profile of EODCJF
As shown in Table 4, thirty-three volatile compounds accounting for 94.46% were authenticated in EODCJF with GC-MS. It can be noted that EODCJF are abundant in alkanes and alcohols, corresponding to 53.96 and 28.59% of the total amount, respectively. N-tetracosane was the volatile compound with highest content identified in EODCJF (27.7%), followed by 1-tetracosanol (14.3%), N-nonacosane (9.03%) and phytol (7.92%).

The In Vitro Antibacterial Activity of EODCJF
As shown in Table 5 and Fig. 7, EODCJF elicited certain antibacterial capacities against E. coli, B. pumilus, B. subtilis and S. aureus. The inhibition zone diameters of EODCJF against these test bacteria were larger than those of normal control, but smaller than those of positive control, indicating that the antibacterial activities of EODCJF against these bacteria are inferior to those of chloramphenicol, a well-known chemically obtained antibacterial agent.   1.20 RI: Retention Index from NIST 05 database; RI * : Retention Index reckoned by referring to that of C8-C40 n-alkanes standard. EODCJF: Essential Oils from the Dropped C. japonica Flowers

Discussion
At the beginning, single-factor experiments for EODCJF extraction were conducted for obtaining the independent variables of RSM optimization. From Figs. 1 to 3, we can observe that prolonged pre-immersion and extraction time as well as overlarge liquid-tosolid ratio lead to a sharp decline in EODCJF yield. These unexpected phenomena could be explained by the following reasons: (1) pre-immersion of plant materials is beneficial for the dispersion of essential oils in water, thereby increasing extraction yield, however, overtime immersion before extraction would cause the loss of volatile compounds, leading to a decrease in extraction yield (Hui et al., 2015); (2) long-term hydrodistillation at water boiling temperature would alter the chemical compositions of essential oils and make certain substances that are vulnerable to heating degrade, thus causing a decline in extraction yield (Zheljazkov et al., 2014); (3) larger amount of water could result in an excessive swelling of raw materials to absorb some substances in EODCJF, leading to a decrease in extraction yield (Xiao et al., 2008).
Then, a three-factor-three-level RSM was achieved using BBD, followed by giving the predicted optimal parameters for EODCJF extraction. Under the predicted optimal extraction conditions, the actual yield of EODCJF was almostly equivalent to the predicted yield, and the error margin was smaller than ±5%, indicating that RSM coupled with BBD could be a feasible way to optimize the extraction parameters of EODCJF.
GC-MS was performed to disclose the chemical profile of EODCJF (Table 4). The major volatile ingredients in the Essential Oils from C. japonica Seeds (EOCJS) are quite different from those of EODCJF. α-Terpineol, α-terpenyl acetate, αterpinolene, α-pinene and β-pinene were observed to be present more in EOCJS (Ramachandran et al., 2020). The significant differences on the chemical profiles of the essential oils derived from different parts of C. japonica would be due to the fact that biological components in plant are apt to assemble in a manner of tissue-specificity and different metabolic enzymes that are responsible for the modulation of secondary metabolites biosynthesis are expressed in different plant organs (Ji et al., 2016). Except C. japonica, variations on the essential oils from different parts for a same species were also found in Blumea balsamifera, Ruta chalepensis, Artemisia persica and so on (Tian et al., 2019).
The antibacterial capacities of EODCJF against E. coli, B. pumilus, B. subtilis, and S. aureus could be partially contributed by their long-chain alcohols (Kubo et al., 1995). As summarized in Table 4, the total amount of alcohols in EODCJF was found to be 28.59% and overwhelming majority of them can be classified as long-chain alcohols. Meanwhile, according to similarity and intermiscibility principle, the abundant alkanes (Table 4) may ameliorate the dispersity of these liposoluble alcohols in EODCJF as benign solvents, thereby enhancing the interactions with test bacteria (Zheng et al., 2019). Whether or not EODCJF exerts inhibitory effects on other bacteria, such as fungus deserves further investigation. Besides, the determination of minimal inhibitory concentration and the underlying antibacterial mechanisms are equally needed to be explored.

Conclusion
In order to utilize the young dropped flowers of C. japonica, a commonly cultivated ornamental tree. In present investigation, the extraction optimization, chemical profile and antibacterial potential of EODCJF were explored for the first time. After being optimized by RSM and BBD, the yield of EODCJF increased up to 3.20%. EODCJF elicited certain inhibitory effects on E. coli, B. subtilis, B. pumilus and S. aureus. The antibacterial capacity could be contributed by the long-chain alcohols present in EODCJF. Further studies are still pending: (1) Antimicrobial screenings with wider bacteria spectrum; (2) The underlying antibacterial mechanisms; (3) Fractionate EODCJF to disclose the structure-activity relationship; (4) Comparison on the chemical constituents with the essential oils from fresh C. japonica flowers, etc. It is hoped that this work would highlight the comprehensive utilization and development of wasted C. japonica flowers.