Optimization of tea‐leaf saponins water extraction and relationships between their contents and tea (Camellia sinensis) tree varieties

Abstract Resulting from the year‐on‐year increase in tea plantations and the saturated consumption of tea leaves, the relative overcapacity in China's tea‐leaf production appears. Discovering the new utilization of tea leaves is helpful to alleviate this phenomenon. The feasibility of extracting saponins from aged tea leaves was investigated and confirmed; three major variables in water extraction were optimized by Box‐Behnken designs. The significant variable found in Box‐Behnken designs, liquid–solid ratio, was went through single‐variable experiments for a more accurate optimization. Seventy‐five ml/g, 1 hr, and 80°C were optimal values and tea‐leaf saponins yield of tea tree variety Longjing 43 reached 12.19% ± 0.0030% after optimizations, higher than the yield of tea‐seed saponins from Camellia oleifera seed meals using the same extraction method (water extraction based on optimizations). According to correlation analyses, tea tree's leaf type and germination stage affected tea‐leaf saponins contents positively, indicating tea trees with larger leaves and later germination stage would have a higher content of tea‐leaf saponins with a higher yield of tea‐leaf saponins under the same extraction method.

However, it is feasible and necessary to investigate the extraction yield of saponins from tea leaves and factors influencing tea-leaf saponins contents, which benefits to deepen the understanding of tea leaves and contributes to the comprehensive utilization of excessive tea leaves.
This research aimed at investigating the feasibility of extracting tea-leaf saponins from tea (C. sinensis) trees with the assistance of water extraction and at the same time, optimizing three major variables (liquid-solid ratio, extraction time, and extraction temperature) in water extraction to acquire a higher yield of tea-leaf saponins.
Factors impacting the content of tea-leaf saponins between different tea tree varieties were also analyzed and discussed.

| Materials and chemicals
Six varieties of tea trees, namely Longjing 43, Anji white tea, Zisun, Huangjinya, Jiukengzao, and Zhenong 117 with details given in All leaves were wiped to remove dust and then dried in an oven at 60 °C until they were in constant weights and sieved through a No. 60 mesh variety-by-variety.

Chemicals and reagents of analytical grade bought from
Sinopharm Chemical Reagent Co., Ltd, Shanghai, China were used in this research. Deionized water (≥18.2 MΩ) was utilized for the preparation of aqueous solutions.

| Methods
The well-established vanillin-sulfuric acid method (Makkar & Becker, 2007) was employed to establish the standard curve and to determine the content of saponins in extraction solutions with the aid of a UV-visible spectrophotometer (Cary 60 UV-Vis; Agilent Technologies, USA).
Three-level Box-Behnken designs were designed to optimize three major variables in tea-leaf saponins water extraction, which were liquid-solid ratio (ml/g), extraction time (hr), and extraction temperature (°C) with ranges designed consulting existing literatures (Qi & Zhang, 2014;Zhang et al., 2003). The variety, Longjing 43, was selected for optimization experiments due to its popularity and universality in Zhejiang Province, and 0.5 g was set as the amount for each measurement. Because the real content of tea-leaf saponins could not be determined for the moment, making the extraction rate, which used the real content of tea-leaf saponins as the denominator, not applicable for comparing the performances of TA B L E 1 Information on six varieties of tea trees

Variety
Leaf type

Germination stage Adaptability
Longjing 43 Medium (1) Extremely early (1) Green tea (1) Large (2) Early (2) Green tea (1) Zhenong 117 Medium (1) Early (2) Green tea (1) Note. Information acquired from Tea Tree Cultivation (Luo 2013). optimizations. Thus, the yield (%), which used the mass of tea-leaf powder as the denominator, was applied as the dependent variable in Box-Behnken designs. The yield (%) of tea-leaf saponins was calculated as equation (1): where c saponin was the concentration of each extraction solution (g/ ml) and calculated from the standard curve and the dilution factor, V extraction solution was the volume of each extraction solution (ml) obtained by the measuring cylinder after a 0.45μm microporous filtering film and m tea powder was the mass of tea-leaf powder utilized for each extraction (g) and determined by an electronic balance accurately (BSA224S; Sartorius scientific instruments (Beijing) Co., Ltd.).
The significant level was 0.05, and data acquired from Box-Behnken designs were analyzed by Design-Expert 11 (Stat-Ease, Inc., USA).

For significant variable without significant interactions in Box-
Behnken designs, single-variable experiments were employed as an auxiliary method to define the effect of this variable more accurately on tea-leaf saponins yields.
Tea-leaf saponins yields of other five tea tree varieties were obtained by the optimized water extraction method. The extraction and measurement were performed three times for each variety.

| Statistical analysis
Independent sample t test of tea-leaf saponins yields from six varieties of tea trees as well as the correlation between yields and leaf type, germination stage along with adaptability was carried out and presented by SAS University Edition (SAS Institute Inc., USA).

TA B L E 3 Analysis of Variance (ANOVA) for the linear model selected from Box-Behnken designs
Tea-leaf saponins yield was 9.61% ± 0.0034% under the extraction condition of 40 ml/g, 1 hr, and 80°C, only having a 0.52% relative standard deviation (R.S.D) with the value predicted by Box-Behnken designs, which proved the predictive ability of the reduced linear model.
Water extraction was not a common method in extracting tea-seed saponins from C. oleifera seed meals, although analysis of range's results (Chen et al., 2017) showed that the three influencing variables on tea-seed saponins yields were in the following sequence: solid-liquid ratio > extraction temperature > extraction time, being consistent with the results of this study. Nonetheless, results of Lin-jian Li, Wu, et al. (2016) were not the same: the sequence of these three variables was extraction temperature > extraction time > liquid-solid ratio. Three close liquid-solid ratio points set for the experiment with values of 5:1, 6:1 and 7:1 ml/g might be the cause of the inconsistency. Different extraction objects were also one of the reasons for the different influencing sequence of optimized variables.

| Single-variable experiments for liquid-solid ratio
A broader range of liquid-solid ratio was designed for the additional experiments with 1 hr and 80°C selected as extraction time and extraction temperature, considering the experimental operability and energy consumption. the tendency of liquid-solid ratio had an ascent at first followed with a decline afterward and the curve came to the peak at 75 ml/g.
The optimal liquid-solid ratio of 75 ml/g was not similar as that in tea-seed saponins extraction from C. oleifera seed meals, suggesting tea leaves required more water to obtain a higher yield of saponins than tea seeds, while too much water reduced the yield because the content of tea-leaf saponins was a constant and the excessive water diluted the concentration of tea-leaf saponins.
The tea-leaf saponins yield of tea tree variety Longjing 43 The 12.19% yield of tea-leaf saponins was higher than the yield of tea-seed saponins with the same method, water extraction based on optimizations, whose values were around 8.5% (Chen et al., 2017; F I G U R E 1 Effects of optimized variables on tea-leaf saponins yields by Box-Behnken designs Li, Wu, et al., 2016;Li, Yu, et al., 2016), which presented that it was feasible to extract saponins from aged tea leaves, in favor of the utilization of excessive tea leaves.

| Influential factors on tea-leaf saponins yields and contents
Under the same extraction conditions, tea-leaf saponins yields of different tea tree varieties could be approximately equivalent to tea-leaf saponins contents of different tea tree varieties; hence, as Table 5 presents, significant differences of tea-leaf saponins contents existed between various tea tree varieties. Table 6 provides the correlation coefficients between tea-leaf saponins yields and three potential influential factors of tea trees both from Pearson's and Spearman's methods. These three factors are tea trees' inherent properties, changing as tea tree varieties changed. Leaf type reflects the size of tea leaves, germination stage is the time tea trees grow new shoots and adaptability refers to the tea type, such as green tea, black tea, and oolong tea, that leaves picked from this tea tree variety are more suitable to produce.
Because of the tea type Zhejiang Province mainly produced is green tea, adaptability of tea trees in this study lacked extensiveness and differences, might causing the inaccuracy of the correlation coefficient between tea-leaf saponins yields and adaptability, nonetheless, it could be concluded that influential factors, leaf type, and germination stage had positive correlations with tea-leaf saponin's contents and leaf type affected stronger, which indicated the tea tree with larger leaves as well as later germination stage enjoyed a higher content of tea-leaf saponins and a higher yield under the same extraction conditions.
Saponin is a secondary metabolite of tea trees, and it is possible that its content increases as tea tree's leaf size increases and germination stage postpones. Nevertheless, the exact correlation coefficient still requires more experiments with a wider scale of tea tree varieties to analyze and come to a more accurate conclusion. All the same, it was feasible to extracting saponins from tea leaves, which benefitted the comprehensive utilization of tea leaves and contributed to solve the relative surplus of tea-leaf production.

| CON CLUS ION
This study investigated and confirmed the feasibility of extracting saponins from tea (Camellia sinensis) leaves and optimized three major variables in water extraction, which were liquid-solid ratio, extraction time, and extraction temperature. After optimization by Box-Behnken designs, the significant variable liquid-solid ratio was conducting an additional experiment to acquire a more accurate optimization using single-variable method. Seventy-five ml/g, 1 hr, and 80°C were optimal values and tea-leaf saponins yield of tea tree variety Longjing 43 reached 12.19% ± 0.0030% after optimizations, higher than the yield of tea-seed saponins from Camellia oleifera seed meals with the same extraction method (water extraction based on optimizations).
Factors affecting the yield of tea-leaf saponins between different tea tree varieties were also analyzed and discussed. Correlation analyses between tea-leaf saponins yields and three inherent properties of tea trees, leaf type, germination stage along with adaptability were carried out, and results showed that the variety of tea tree with larger leaves as well as later germination stage would have a higher content of tea-leaf saponins as well as a higher yield under the same extraction conditions.
Confirming the feasibility of extracting saponins from tea leaves not only benefits to the in-depth understanding of tea leaves, but also contributes to the comprehensive utilization of tea leaves, especially excessive aged tea leaves in autumn and winter, which was helpful to alleviate the relative surplus of tea-leaf production in China.

ACK N OWLED G M ENTS
This work was supported by National Key R&D program of China, (grant number: 2016YFD0700304).

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing.

O RCI D
Xiao-Lan Yu http://orcid.org/0000-0002-1780-7850 TA B L E 6 Correlations between tea-leaf saponins yields and leaf type, germination stage along with adaptability of tea trees