SASO was prepared in our laboratory and obtained by solvent extraction method with an acid value of 5.5 ± 0.3 mgKOH/g. All the commercial lipases (lipase A from Candida sp. (CALA), Eversa^® Transform 2.0, and lipase B from Candida sp. (CALB) were acquired from Novozymes A/S (Denmark). The ethanol was supplied from Kemiou Chemical Reagent Co., Ltd (Tianjin, China). Molecular sieves 4A (4A-MS) was purchased from Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). Blue silica gel was acquired from Komeo Chemical Reagent Co., Ltd. (Tianjin, China). Acrylic acid super absorbent polymer (SAP) was bought from Wanhua Chemical Group Co., Ltd. (Yantai, China). Dialysis Bag MD 44 was obtained from Sobalai Technology Co., Ltd. (Beijing, China).

Under the conditions of enzyme addition of 4% liquid lipase (w/w, SASO), 0.8% enzyme dry powder (w/w, SASO), 7:1 (mol/mol) substrate ratio of alcohol to oil, 0.6 g/g of dehydrating agent, 35°C, the reaction was carried out in a 25 mL round bottom flask and then placed in a constant temperature water bath oscillator. Samples were taken out at regular intervals for centrifugation in a centrifuge tube, and a certain amount of the upper sample was used for GC and acid value analysis.

The purification of the synthesized FAEE product of SASO was carried out using molecular distillation under the following conditions: heating temperature of 120°C, scraping speed of 50 rpm, injection volume of 2 mL/min, and pressure of 50 Pa. The light phase product was obtained as the FAEE product.

The FAEE product of SASO were tested for acid value, moisture content, iodine value, oxidative stability, cetane number, and other indexes in accordance with ASTM standards.

All experiments were performed at least in triplicate. Results were expressed as average ±S.E.M. SPSS 20.0 was employed to analyze the data. Statistical significance was considered at p < 0.05.

Lipase contains an amphiphilic amino acid chain, the lid, which is responsible for the activation of the enzyme due to its mobility and protection of the enzyme’s active site (Khan et al., 2017). Differences in the lid structure, active center structure, and the amount of water needed to activate lipase resulted in different specificities and catalytic activities (Secundo et al., 2006; Tongboriboon et al., 2010).

What’s a side note, the catalytic activity of the liquid lipase was greater than that of the corresponding enzyme dry powder, probably because the right amount of water facilitates the opening of the lid structure covered by the active site of lipase, and maintains the stability of the lipase structure and provides a useable oil-water interface (Huang Z et al., 2015).

From Supplementary Figure S3, it can be seen that FAEE product has the highest yield when catalyzed by the liquid lipase Eversa^® Transform 2.0. This was due to the fact that the lipase Eversa^® Transform 2.0 has a smaller lid structure, which allows for easy opening of the lid at the interfaces of oil and water so that shows a higher catalytic activity. However, the lipase CALB has a lid structure of 5 amino acids and CALA has 9–10 amino acids. (Uppenberg et al., 1994; Ericsson et al., 2008). In addition, the “channel-like” stucture of the lipase Eversa^® Transform 2.0 made binding of SASO to ethanol and release of the product easier, which resulted in a faster reaction rate than the lipases CALA and CALB. It also suggests that liquid lipase has better tolerance to ethanol (Li et al., 2013).

For maximum FAEE yield, the lipase should be able to convert all glycerides (including mono-, di-, and triglycerides) as well as free fatty acids into FAEE. The result showed that the lowest acid value of FAEE (0.2 ± 0.12 mgKOH/g) was found in the final system of the reaction using dry enzyme powder of CALB as a catalyst. This may be due to the absence of interfacial activation of lipase CALB (Tongboriboon et al., 2010) or the low water requirement to activate lipase CALB (Martinelle et al., 1995). The main reason is attributed to the higher catalytic esterification capacity of the dry enzyme powder of CALB.

As a consequence, the reaction requires both a favorable liquid enzyme with the ability to catalyze esterification to maintain a high yield, and a certain amount of enzyme dry powder capable of esterification with free fatty acids to reduce the acid value to further maximize the yield. (Christopher et al., 2014). Therefore, a combination of enzymes (liquid lipase Eversa^® Transform 2.0 and enzyme dry powder of CALB) was used for the next studies.

As can be seen from the experimental results (Supplementary Figure S4), preparation of liquid lipase Eversa^® Transform 2.0 catalyzed under different dehydrating agent conditions, the FAEE yield was: 4A-MS (86.3% ± 1.2%) > blue silica (83.1% ± 1.5%) > SAP (76.2% ± 1.0%), and the acid value of FAEE product was: SAP (4.5 ± 0.2 mgKOH/g) > blue silica gel (4.4 ± 0.1 mgKOH/g) > 4A-MS (3.2 ± 0.1 mgKOH/g). Three water removals have similar FAEE yield, but 4A-MS is able to achieve a lower acid value while maintaining a high FAEE yield.

During the enzymatic synthesis of FAEE, a relatively small amount of water was required in organic solvent to maintain enzyme activity (Atadashi et al., 2012). It was demonstrated that for every 1% of fatty acids in the raw material, 650 mg/kg of water was produced by the esterification reaction (Liu et al., 2013). Due to the high content of free fatty acids in raw material used to synthesize FAEE, a large amount of water could be generated during the esterification reaction with ethanol. However, excess water can facilitate hydrolysis of the oil. Furthermore, water in the reaction system may inhibit the positive progression of the lipase-catalyzed reaction and result in the high acid value of the product (Hájek et al., 2012; Nguyen et al., 2018). During the reaction, the surface of the lipase Eversa^® Transform 2.0 requires a layer of essential water to maintain a certain active conformation for its catalytic activity (Molina-Gutiérrez et al., 2021). Therefore, the dehydration capacity of the water removal agent should be appropriate, so that not only the acid value of the FAEE product is desired, but also an essential layer of water is not taken away from the surface of the enzyme.

Difference ability of dehydrating agents in water removal can be explained based on the principle of removing water and the different specific surface areas of the materials. Firstly, SAP is a white powder with small particle size (0.2–0.9 mm), large specific surface area, and many grooves on the surface, which provides a large space for SAP to store water (Gu et al., 2015). In addition, SAP has a strong water removal capacity and it can efficiently take away the essential water from the surface of the lipase Eversa^® Transform 2.0 molecule, thus putting the lipase in a rigid inactivation state and reducing the catalytic activity. These resulted in the lowest FAEE yield and the highest acid value when the reaction was catalyzed with the lipase Eversa^® Transform 2.0. Secondly, compared with other water removing agents (4A-MS, SAP), blue silica gel has a smaller specific surface area and weaker adsorption surface polarity. In addition, the pore size of blue silica gel is larger than the molecular size of ethanol, and the adsorption of ethanol by silica gel may negatively affect the reaction and thus getting a poorer result (Wang et al., 2006). Finally, as the specific surface area of 4A-MS is similar to that of blue silica gel, but the polar effect of the adsorption surface is stronger and its water removal effect was better than blue silica gel. Therefore, 4A-MS was the best choice of water removal agent for this work. In contrast to other dehydrating agents, the lipase Eversa^® Transform 2.0 in the 4A-MS system had higher catalytic activity and lower acid value of the FAEE product, indicating that 4A-MS did not take away essential water from the enzyme surface. For consequence, 4A-MS was used for the next research.

In the screening of enzyme types, we chose Eversa^® Transform 2.0, a low-cost liquid enzyme with good catalytic effeciency, to be compounded with the dry powder enzyme CALB. The function of CALB is to catalyze the esterification of free fatty acids with ethanol to obtain more product that reduces the acid value. Also, Eversa®Transform 2.0 is effective in increasing yield due to its non-specificity (Christopher et al., 2014) and better tolerance to ethanol and phospholipids (Li et al., 2013). As shown in Supplementary Figure S5, at a certain enzyme dry powder CALB (0.8%) addition, with the increasing addition load of liquid lipase Eversa^® Transform 2.0 from 1% to 8%, there was an increase in FAEE yield (from 91.2% ± 1.2% to 97.6% ± 1.5%). However, a substantial rise in the acid value of the product was observed (from 2.7 ± 0.1 mgKOH/g to 6.2 ± 0.1 mgKOH/g). The reason is that the volume of enzyme added not only affected the reaction rate, but also imported more water into the reaction system, which in turn affected the acid value of FAEE product.

The relationship between the acid value of FAEE and the moisture content of the system is displayed in Supplementary Figure S6. In fact, the proportion of the complex enzyme not only affects the rate of the reaction, at the same time, since the liquid lipase itself contains water, its addition increases the moisture content of the system which in turn affects the acid value of the FAEE product. As can be deduced from the linear relationship between moisture and acid value in the system that the acid value of biodiesel catalyzed by lipase could meet the requirement (AV≤0.5 mgKOH/g) only when the moisture content in the system was no more than 500 mg/kg. In order to achieve the required acid value of the FAEE product (≤0.5 mgKOH/g), a combined complex enzyme with 1% liquid lipase Eversa^® Transform 2.0% and 0.8% enzyme dry powder CALB were the best choice.

With the increase of 4A-MS addition load, the acid value of FAEE product was gradually decreased. But the yield of FAEE was also slowly declined (Supplementary Figure S7). This may be due to the fact that the removal of water from the reaction system is enhanced by the increasing addition load of 4A-MS, which in turn promotes the positive direction of the esterification. However, the increase in 4A-MS addition load limited enzyme diffusion, increased resistance to migration of the reaction, impeded access of substrates (SASO and ethanol) to the complex enzymes active sites, and retarded the response velocity (Nguyen et al., 2018). In addition, the increasing addition load of 4A-MS would take away the essential water from the lipase surface layer, which leaves the enzyme molecule in a rigid deactivated state and in turn weakens the catalytic activity of the enzyme and slows down the reaction rate.

Accordingly, to reduce the acid value of the FAEE product below the required standard (≤0.5 mgKOH/g) and hold high FAEE yield, 0.8 g/g 4A-MS was added to carry out the following experiments.

When the alcohol-oil substrate is added from 3:1 (mol/mol) to 9:1 (mol/mol), the FAEE yield was sharply increased from 33.3% ± 1.0% to 90.8% ± 1.5%, while the acid value of product was dramatically decreased from 3.8 ± 0.1 mgKOH/g to 0.36 ± 0.06 mgKOH/g (Supplementary Figure S8). This is owing to the fact that an increase in the alcohol-oil substrate ratio can promote a shift in the reaction equilibrium toward more FAEE production of SASO. With a continued increase in the substrate ratio (>9:1), the yield of FAEE was decreased, while the acid value of the product was also increased. This may be due to the reason that an increase in ethanol concentration in the reaction system affects the folding structure of lipase protein (Eversa^® Transform 2.0 and CALB), which leads to a decrease in the catalysing ability of lipase (Liu et al., 2013). Alternatively, there is also a reason that high alcohol concentration weakens the activity of the enzyme by stripping the essential water molecules and unfavorably unfolding it into a more helical state (Lotti et al., 2015). At the same time, the increase of ethanol addition also increased the water content in the reaction system, which inhibited the positive transesterification reaction and resulted in an increase of acid value. In addition, high alcohol-oil ratio can affect the separation of products. In the meantime, the additional ethanol can make the mass concentration of SASO in the system smaller, resulting in a slow formation rate of enzyme-substrate complex.

This demonstrated that an alcohol-oil ratio of 9:1 (mol/mol) was the appropriate substrate ratio for the complex enzyme catalyzed the preparation of FAEE product of SASO, which was greater than the alcohol-oil molar ratio for the immobilized lipase (from Pseudomonas cepacia) catalyzed ethanolysis of Jatropha curcas hair oil (4:1) (Aransiola, 2013). The results also showed that the combination of the complex enzymes had high alcohol resistance and could not easily be inactivated in high ethanol environments.

Reaction temperature is an important parameter affecting the catalytic activity of lipase and reaction rate. The yield of FAEE was increased to a maximum value of 90.83% ± 1.5% when the temperature was varied up from 25°C to 35°C, and the acid value of FAEE product was decreased to 0.36 ± 0.06 mgKOH/g (Supplementary Figure S9). This may be due to the fact that higher temperatures foster the collision of lipase and substrate as well as the formulation of enzyme-substrate composite, thus increasing the efficiency of the reaction. In addition, elevating temperature also reduces the viscosity of reaction system, thereby alleviating mass transfer limitations and facilitating the effective diffusion of the substrates (Guo et al., 2020). It was observed, however, that there was a significant decrease in FAEE yield when the temperature was increased from 35°C to 65 °C. This may be due to the influence of higher temperature (>35°C) on the structure of lipase, resulting in the inactivation of lipase.

Therefore, the reaction temperature of 35°C is the optimum temperature for the preparation of FAEE catalyzed by the complex enzymes. Compared with the preparation of FAEE by ethanolysis of murumuru oil using the complex enzymes, Lipase AK (from Pseudomonas fluorescens) and Lipase PS (from P. cepacia) (the optimum temperature of 45°C), the reaction temperature (35°C) in our work was lower and less energy consuming (Tongboriboon et al., 2010).

SASO mainly consists of TG and FFA, and two parallel reactions simultaneously exist in this reaction system: i) ester exchange reaction (TG is reacted with ethanol to form MG, DG and FAEE), and ii) esterification reaction (FFA is reacted with ethanol to form FAEE). It can be observed from Supplementary Figure S10 that the yield of FAEE was rapidly increased to 65.2% ± 1.3% during the first 6 h of the reaction. This was followed by a gradual increase to 89.2% ± 1.8%. When reaction time was further increased to longer than 24 h, the yield of FAEE was remained almost constant (∼90%). The equilibrium time of this reaction (24 h) was similar with that of the lipase Rhizomucor miehei catalyzed preparation of FAEE from microalgae oil (24 h) (Huang J et al., 2015).

In the first 6 h of the reaction, the content of TG was rapidly decreased and the content of MG was gradually increased. At the same time, the content of DG was gradually increased during the first 3 h of the reaction. However, when reaction time was >6 h, the contents of TG, DG and MG were all gradually reduced. This indicated that the transesterification reaction of TG with ethanol was the main reaction in the first 6 h. However, when the reaction time was longer than 6 h, the main transesterification reactions of DG and MG with ethanol were observed. At the same time, the FFA content was gradually dropped throughout the reaction, and to less than 0.25% at 12 h. These suggested that the esterification reaction also existed simultaneously in this reaction system, but was not the main reaction.

The physical and chemical indexes (except oxidative stability) of the SASO FAEE product were in compliance with the requirements of ASTM D6751 (as shown in Supplementary Table S1). To address the issue of oxidative stability of FAEE product of SASO, antioxidants such as butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) were added in the subsequent study. The addition of antioxidants both improves the oxidative stabilities of biodiesel product and decreases the emission of nitrogen oxides (Ryu, 2010). Biodiesel prepared from SASO has a lower viscosity (4.54 mm²/s) than that from rapeseed oil (4.83 mm²/s) and palm oil (5.70 mm²/s), which was more conducive to the flow of biodiesel in engines (Ryu, 2010). The density of the FAEE product prepared with SASO was 0.870 g/cm³, which is lower than the 0.880 g/cm³ prepared from sunflower oil, as well as 0.882 g/cm³ for rapeseed oil and 0.899 g/cm³ for castor oil in other studies (Karmakar et al., 2010). The SASO FAEE product has a flash point of 174°C, which is qualified by ASTM D6751. These suggested that the biodiesel prepared by SASO had relatively excellent injection and atomization properties, which can be used as a potential renewable oil resource for biodiesel preparation.