Phoenix tree (Firmiana simplex) seeds were supplied by the Xintai Nursery Filed (Jiangsu, China). Lipozyme TLIM (from Thermomyces lanuginose, TLL) was obtained from Novozymes A/S (Bagsvaerd, Denmark). Potassium hydroxide and sodium hydroxide were purchased from Tianjin Kemiou Chemical Reagent Co. Ltd (Tianjin, China). Other chemicals were of analytical grade.

Phoenix tree seeds were dried and impurities such as pebbles and leaves were removed. Then, the seeds were crushed using a pulverizer and passed through a 40 mesh sieve to obtain seed powder. Next, Phoenix tree seed powder was extracted using n-hexane in a 1000 mL beaker which was placed in an ultrasonic bath. The supernatant was then centrifuged and the solvent was removed using a rotary evaporator. The oil obtained was used for the hydrolysis reaction.

Hydrolysis reactions were conducted using 250 mL three-necked flasks. The mixture of Phoenix tree seed oil and water (ratio of water to oil 10:1, w/w) was prepared using a magnetic stirrer (200 rpm) at different temperatures (30–55 ^oC) in oil bath. The reaction was started by the addition of Lipozyme TLIM (10%, relative to the weight of all substrates). The samples were taken out at specified time intervals, dissolved and centrifuged.

High temperature GC (Agilent 7980B) equipped with a DB-1H capillary column (30 m×0.25 mm, 0.1 μm of film thickness) and flame ionization detector (FID) was also used to evaluate the hydrolysis product formation during the hydrolysis. The detector and injector temperatures were set at 350 ^oC and 400 ^oC, respectively. Initial column temperature was 100 ^oC, programmed to 360 ^oC at the rate of 20 ^oC/min. Helium was used as carrier gas at a flow rate of 1 mL/min. Acid values were analyzed according to the AOCS Official Method Cd 3d-63 (Firestone et al., 1998).

DH (%) was calculated from the following equation:

Where AV₀ is the initial acid value of Phoenix tree seed oil, and AV_t is the acid value of Phoenix tree seed oil at time t. SV is the saponification value of Phoenix tree seed oil.

A 3-level-3-factor Box–Behnken experimental design was applied to evaluate the interaction effect of reaction variables on the hydrolysis. The factors and levels used in experimental design were as follows: reaction time (3, 14 and 25 h), reaction temperature (30, 45 and 60 °C), enzyme load (1%, 8% and 15%; relative to the weight of total substrates) (Table 1).

Design-Expert 8.0 was used to analyze the experimental data, which can generalize second-order polynomial model as follows:

Where Y is the predicted response (DH); X_i and X_j represent the independent variables. Regression coefficients β₀, β_i, β_ii, and β_ij are the intercept, linear, quadratic, and interaction terms, respectively.

All experiments were performed at least in triplicate. Results were expressed as average ± SEM. A two-way analysis of variance (ANOVA) was used. Statistical significance was considered at p < 0.05.

Figure 1 shows that during the hydrolysis process, the triacylglycerol (TAG) of Phoenix tree seed oil was firstly converted to diacylglycerol (DAG), and then DAG was hydrolyzed to monoacylglycerol (MAG), and finally fatty acids can be produced by the hydrolysis of MAG, DAG and TAG. The low content of MAG (~3%) during the hydrolysis progress indicated that MAG can be quickly converted to free fatty acids when MAG was formed. And the relative high DAG content (~20% at 2 h) indicated that the formation rate of DAG is higher than that of the hydrolysis of DAG. These results showed that the limitation step for the hydrolysis of the oil was attributed to the first step (the formation of DAG by the hydrolysis of TAG). Figure 1 also shows that the content of fatty acids in the product sharply increased with the increase in hydrolysis time from 0 to 12 h, and at 24 h, and the hydrolysis reaction reached equilibrium at 24 h. Similar results can also be found in the chemical (Fe-Zn double-metal cyanide)-catalyzed hydrolysis of oil (Satyarthi et al., 2011).

Reaction temperature played an important role in the hydrolysis of Phoenix tree seed oil, which influenced the mass transfer of the reaction system and catalytic activity of Lipozyme TLIM. To evaluate the effect of the reaction temperature on the hydrolysis of Phoenix tree seed oil, experiments were performed at different temperatures from 30 ^oC to 55 ^oC and the results are shown in the Figure 2.

Figure 2A shows that the DH of Phoenix tree seed oil increased with the increase in reaction temperature from 30 ^oC to 45 ^oC. However, when the temperature was above 45 ^oC, the DH of Phoenix tree seed oil decreased, which may be ascribed to the deactivation of Lipozyme TLIM due to the adsorption of glycerol on the immobilized lipase surface and higher temperature (>45 ^oC) in the excess water system. Therefore, the optimum temperature for maximum DH was 45 ^oC, which was lower than that of the Lipozyme TLIM-catalyzed hydrolysis of oil in a supercritical carbon dioxide medium (60 ^oC) (Prado and Saldaña, 2013), which was attributed to the difference in hydrolysis media.

The initial hydrolysis rate (V₀, mol/(L∙min)) of Phoenix tree seed oil, defined as the initial hydrolysis degree per unit time, can be obtained from six experimental points of the hydrolysis degree-time profile corresponding to low DH (15% or less) at different reaction temperatures. Activation energy can be calculated by the Arrhenius law described as follows:

Where A represents the Arrhenius constant, and Ea represents the activation energy. R and T are gas constant and absolute temperature (K), respectively.

A good linearity between LnV₀ and 1/T was found (Figure 2B). The value of Ea calculated by equation (3) was 39.25 KJ/mol, which was higher than the Ea value (0.97-34.5 KJ/mol) of general lipase-catalyzed reactions (Phuah et al., 2012). The differences can be ascribed to the greater steric hindrance of sterculic acid. Therefore, the Arrhenius equation for the hydrolysis of Phoenix tree seed oil can be re-written as LnV₀ = 9.12 –4721/T.

To explore the effect of mass ratio of water to oil on the hydrolysis reaction, experiments with different mass ratios were carried out (Figure 3). With the increase in mass ratio of water to oil from 1:1 to 10:1 (w/w), the DH of Phoenix tree seed oil increased significantly to its maximum (~96%) after the reaction reached equilibrium. When the mass ratio of water to oil increased to 15:1, the maximum DH of Phoenix tree seed oil was the same as that of 10:1. However, the initial hydrolysis rate (0.0041 mol/(L∙min)) of 15:1 is lower than that (0.0038 mol/(L∙ min)) of 10:1. These results showed that enough water is necessary for the hydrolysis of Phoenix tree seed oil. Similar effect of water on the lipase-catalyzed hydrolysis of oil can also be found in previous reports (Chen et al., 2014; Prado and Saldaña, 2013).

Enzyme load will affect mass transfer and initial reaction rate. In order to evaluate the effect of enzyme load on the hydrolysis of Phoenix tree seed oil, different enzyme loads from 1% (relative on the weight of all substrates) to 15% were added to the reaction. The DH of Phoenix tree seed oil increased with the increase in enzyme load from 1% to 10% (Figure 4). However, when the enzyme load exceeded 10%, the initial DH of Phoenix tree seed oil decreased. When the hydrolysis reached equilibrium with a higher enzyme load (≥5%), the maximum DH of Phoenix tree seed oil were similar (~96.5%). These results can be explained by the fact that sufficient enzyme is necessary for the hydrolysis. However, excessive enzyme also leads to aggregate and great mass limitation, which results in the low hydrolysis rate at the highest enzyme load (15%). A similar effect of high enzyme load on the reaction can also be found in other reactions (Albasi et al., 1999; Al-Zuhair et al., 2003; Sun and Chen, 2015).

The linear relationship between the reaction rate and enzyme loading (Figure 5A) showed that the effect of external mass transfer limitation on the hydrolysis reaction can be neglected, and the hydrolysis was controlled by kinetics. In order to obtain the kinetic parameter of the hydrolysis reaction, several experiments were performed by varing oil concentration (0.02 mol/L-0.3 mol/L). The initial reaction rate firstly increased to the maximum (V₀ = 0.120 mol/(L∙min)) (Figure 5B) and then decreased with the increase in oil concentration. According to Ping-pong Bi-Bi, the initial reaction rate can be described as follows:

Due to the presence of significantly excess water, the contention of H₂O can be regarded as a constant. Hence, the initial reaction rate can be described as:

Where V₀ is the initial hydrolysis rate; Vmax is the maximum reaction rate; [A] and [B] are Michanelis-Meton constants of oil and water, respectively; K_IA is the inhibition contant of oil, and K^/m is the apparent Michaelis-Meton.

A good Lineweaver-Burk plot was obtained between the reciprocal of the initial rate versus the reciprocal concentration of oil (Figure 5C). The value for Vmax and K^/m were calculated as 0.232 mol/(L*min) and 0.084 mol/L, respectively.

RSM is a comprehensiveness of statistical and numerical technology which is used to optimize experimental conditions (Syam et al., 2016). The experimental values obtained from the model are shown in Table 1. The data were regressed by employing a quadratic polynomial equation to calculate the relationship between the DH of Phoenix tree seed oil and hydrolysis variables.

The analysis of variance (ANOVA) for the model is shown in Table 2. The low P-value (P < 0.001) and high coefficient (R² = 0.9636) indicated that the model was adequate to explain the relationship between reaction variables and the DH of Phoenix tree seed oil. The order of hydrolysis variables influencing the DH of Phoenix tree seed oil was reaction time > enzyme load > hydrolysis temperature. The regression equation explaining the DH of Phoenix tree seed oil was given as follow:

Figure 6A shows the effect of reaction time, temperature, and their mutual interaction on the hydrolysis of Phoenix tree seed oil with an 8% enzyme load. High DH (>96.4%) of Phoenix tree seed oil can be obtained at long reaction times (14–25 h) and moderate temperatures (36–54 ^oC). Figure 6B shows the 3D surface plot of DH drawn for the interaction of enzyme load and reaction time, which were similar to those of the interaction of reaction time and reaction temperature. Proper enzyme load and longer reaction time were beneficial for the hydrolysis. Low DH at high enzyme load was attributed to the aggregation of excessive enzyme. Figure 6C shows the effect of enzyme load, temperature and their mutual interaction on the hydrolysis reaction. The maximum DH (97.2%) of Phoenix tree seed oil could be achieved at 40–45 ^oC and 8%–13% enzyme load. The deactivation of Lipozyme TLIM at a higher temperature (>54 ^oC) can also be found in Figure 6C.

The maximum DH (98.2±0.8%) of Phoenix tree seed oil can be achieved at the following optimized conditions: 45 ^oC, 18h, 10% enzyme load and 10:1 mass ratio of water to oil. Good agreement between the observed and predicted values indicates the validation of the model. In the hydrolysis product, the content of sterculic acid was similar to that (~23%) of Phoenix tree seed oil.