Chemoenzymatic conversion of glycerol to lactic acid and glycolic acid

Catalytic valorization of raw glycerol derived from biodiesel into high-value chemicals has attracted great attention. Here, we report chemoenzymatic cascade reactions that convert glycerol to lactic acid and glycolic acid. In the enzymatic step, a coenzyme recycling system was developed to convert glycerol into 1,3-dihydroxyacetone (DHA) with a yield of 92.3% in potassium phosphate buffer (300 mM, pH 7.1) containing 100 mM glycerol, 2 mM NAD+, 242 U/mL glycerol dehydrogenase-GldA and NADH oxidase-SpNoxK184R at 30 °C. Subsequently, NaOH or NaClO2 catalyzes the formation of lactic acid and glycolic acid from DHA. The high yield of lactic acid (72.3%) and glycolic acid (78.2%) verify the benefit of the chemoenzymatic approaches. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s40643-022-00561-z.


Introduction
With the urgency of reducing the dependence on fossil fuels, (Kosamia et al. 2020) the fuel energy and chemical commodities is increasingly preferred towards the use of clean liquid biofuels, ethanol and biodiesel. However, large amount of crude glycerol was produced as a byproduct accounting for 10% of the produced biodiesel (Veluturla et al. 2018), with an estimated amount of 4 × 10 6 tons by 2024 (Singh and Ahuja 2018). Moreover, since glycerol is a by-product with low economic value and cannot be safely disposed in the environment, the generated excess glycerol may become an environmental threat. Therefore, it is highly profitable to transform lowvalue but versatile glycerol to high value-added chemicals (Walgode et al. 2020). Glycerol is an important C3 platform compounds ) that can be upgraded into various chemical building blocks through chemical oxidation, reduction (Xu and Ni 2015), dehydration, deoxydehydration reaction or a combination of the aforementioned methods, e.g., to allyl alcohol (Tshibalonza and Monbaliu 2017), acrylic acid (Sun et al. 2017), divinyl glycol (Sotto et al. 2016), DHA (Lari et al. 2015), etc.
The conversion of glycerol to DHA was performed by glycerol dehydrogenase (EC 1.1.1.6, EC 1.1.1.72, EC 1.1.99.22). Glycerol dehydrogenase-GldA (from Escherichia coli, EC 1.1.1.6) is an enzyme present in bacterial cytoplasm, converts glycerol to 1,3-dihydroxyacetone accompanied by NAD + to NADH. It should be noted that it catalyzes the reaction of glycerol for a reversible reaction, glycerol oxidation in alkaline conditions and DHA reduction in acidic conditions. Therefore, it is necessary to pay attention to the change of reaction pH when using it for experiments. For the cofactor regeneration, often NADH oxidase which catalyzes the oxidation of NADH by reducing O 2 to H 2 O through four-electron reduction was used.
Lactic acid and glycolic acid are important building blocks of biodegradable polymer, poly-lactic acid (PLA), poly-glycolic acid (PGA) and poly lactic-co-glycolic acid (PLGA). Lactic acid (López-Gómez et al. 2019) is a three carbon alpha hydroxycarboxylic acid widely used in food, medicine, cosmetics, detergents and dairy industries (Marianou et al. 2019) as well as the synthesis of polylactic acid (Huang et al. 2021), lactate ester (Hasegawa et al. 2008), pyruvate (Liu et al. 2010), acrylic acid (Wang et al. 2008), 1,2-propanediol (Oude Elferink et al. 2001), 2,3-pentanedione (Niu et al. 2019), etc. The production of lactic acid by microbial fermentation is the main production method (Wang et al. 2020), because of the wide source of raw materials, low production cost, high optical purity and high safety. Currently, more than 90% of the global lactic acid production is made by the microbial fermentation (Zhang et al. 2019). Arabinose, pentose, sucrose and lignocellulosic hydrolysates have been used as substrates for microbial fermentation of lactic acid. Glycolic acid is a two-carbon α-hydroxy acid, which is widely used in pharmaceutical, textile and food. It is often used in emulsion polymers, solvents and additives in inks and paints to improve flow performance and gloss (Lasprilla et al. 2012). Furthermore, glycolic acid can copolymerize with lactic acid to form poly PLGA, which is an important biodegradable material with unique properties such as non-toxic, biocompatible and degradable in vivo (Shi et al. 2006). PLGA is widely used in the fields of medical polymers such as suture and tonic materials, surgical sutures, drug sustained-release carriers, tissue engineering, etc. (Huang et al. 2020). Glycolic acid is mainly produced by chemical processes, e.g., carbonylation of formaldehyde with synthesis gas (Shi et al. 2020b) or treatment of carbon monoxide with water (Shi et al. 2020a), electrolytic reduction of oxalic acid and hydrolysis of glycolonitrile . Microbial fermentation with engineered metabolic pathways have been developed for the production of glycolic acid. Using the lack of a functional carbon-concentrating mechanism (CCM), glycolic acid was accumulated through an inefficient and incomplete photorespiratory pathway (Yun Eun et al. 2021). Using glucose as substrate, glycolysis pathway was integrated with glycolic acid pathway, and finally the yield of glycolic acid with the potency of 41 g/L/h was achieved (Zhu et al. 2021). The titer of glycolic acid in 5-L bioreactor reached 40.9 g/L after multiple rounds of strain engineering using biosensor-based high-throughput method . Chemical method is commonly used in industrial production of glycolic acid (Shi et al. 2019). Chemoenzymatic cascade reactions are synergistic tandem reactions by combining a variety of chemical catalysts with biocatalysts in vitro, which can effectively reduce by-products and wastes and make the products easier to be purified (Luo et al. 2021).
Inspired by a previous report on lactic acid generation through chemical catalysis , such as KOH, Ba(OH) 2 , Ca(OH) 2 (Lux and Siebenhofer 2013), DHA was firstly dehydrated to pyruvic aldehyde and then isomerized to lactic acid via the intramolecular Cannizzaro reaction under alkaline conditions (Evans and Cornthwaite 2002). And the C-C bond of DHA was selectively cleaved and then oxidized to glycolic acid and formate in the presence of H 2 O 2 and NH 4 OH (Pullanikat et al. 2010). We decided to use NaOH as catalysts for the conversion of DHA into lactic acid, and NaClO 2 as oxidant for the oxidation of DHA to glycolic acid.
In this study, two-step chemoenzymatic cascade reactions were developed to convert glycerol to lactic acid and glycolic acid, respectively. Reaction conditions were optimized to achieve desirable yield. This is a novel pathway for converting glycerol to lactic acid and glycolic acid.

Design of the cascade reactions
As shown in Scheme 1, the cascade reaction starts with the glycerol dehydrogenase (GldA) catalyzed oxidation of glycerol into DHA, which is subsequently transformed to lactic acid and glycolic acid in different chemical reaction systems. For the generation of lactic acid, NaOH was used to promote the rearrangement of DHA into lactic acid where DHA was firstly dehydrated to pyruvic aldehyde and then isomerized to lactate via the intramolecular Cannizzaro reaction under alkaline conditions. For the generation of glycolic acid, DHA was oxidized to glycolate acid by NaClO 2 .

Enzymatic conversion of glycerol to DHA
For the initial step of this cascade, in vitro biotransformation of glycerol to DHA was achieved with glycerol dehydrogenase-GldA (Gene ID: CP034658.1) (from Escherichia coli) (Piattoni et al. 2013) using NAD + as a cofactor and the NADH oxidase-SpNox K184R (from Streptococcus pyogenes) (Oscar;and Jan-E 2015) for NAD + regeneration, which requires only O 2 as the cosubstrate. The GldA gene was cloned into T7 promoterbased plasmid pET21b to construct pET21b-GldA. The SpNox gene with an amino acid substitution K184R was synthesized and cloned into the pRSFDuet-1. pET21b-GldA and pRSFDuet-SpNox K184R were expressed in E. coli/BL21-Gold (DE3), respectively. After the strain was cultured and expressed, the protein was collected from the broken strain and then for the following purification.
Firstly, three kinds of buffer solution (potassium phosphate buffer (100 mM K 2 HPO 4 and KH 2 PO 4 , pH7.1), sodium phosphoric buffer (100 mM NaH 2 PO 4 and Na 2 HPO 4 , pH7.1), Tris-HCl buffer (100 mM Tris-HCl, pH7.1)) were investigated for the oxidation of glycerol catalyzed by dehydrogenase-GldA and NADH oxidase-SpNox K184R (Xu et al. 2017). Enzyme activity of GldA and SpNox K184R was determined by the slop of decrease and increase of NADH observed at 340 nm, 30 °C. As shown in Fig. 1, glycerol dehydrogenase-GldA exhibited acceptable activity in potassium phosphate buffer, but nearly inactive in Tris-HCl buffer. Meanwhile, SpNox K184R showed acceptable activity in Tris-HCl buffer and potassium phosphate buffer. Considering the compatibility of these two enzymes in the same reaction system, potassium phosphate buffer was selected for the GldA-SpNox K184R coupled enzyme system for further investigation.
After the selection of the buffer solution, the optimum pH value of GldA and SpNox K184R were investigated with the potassium phosphate buffer, ranging from pH 6.1-7.1 (Fig. 2). The activity of GldA and SpNox K184R was gradually increased with the elevated pH value. Since GldA and SpNox K184R can reach highest activity at pH 7.1, the following experiments were carried out at pH7.1.
The concentration of the potassium phosphate buffer was optimized for the optimum activity of GldA and SpNox K184R , ranging from 20 to 500 mM. Figure 3 shows that 300 mM is the optimal concentration of potassium phosphate buffer for GldA and 50 mM was the optimal concentration for SpNox K184R . On overall consideration, as a key enzyme, GldA's activity affects the catalytic efficiency of the whole system, so 300 mM potassium phosphate buffer is selected for subsequent experiments.
The enzymatic activity of GldA and SpNox K184R were determined to be 24.2 U mg −1 and 34.62 U mg −1 , respectively. One unit of the enzyme activity was defined as catalyzing the conversion of 1 μmol NADH to NAD + per minute under standard conditions for GldA or vice versa for SpNox K184R .
And then we conducted an experiment on the thermal stability of the enzyme, and the results are shown Scheme 1 A diagram of the cascade reaction in this experiment The activity of GldA (orange circle) and SpNox K184R (blue square) at different pH, ranging from 6.1 to 7.1 (100 mM K 2 HPO 4 and KH 2 PO 4 ) in Fig. 4. In potassium phosphate buffer (300 mM, pH 7.1), GldA and SpNox K184R still maintained more than 50% activity after 25.5 h at 30 °C.
Finally, the optimal concentration of GldA-SpNox K184R coupled system and substrate glycerol was investigated to achieve higher yield of DHA. In the presence of 2 mM NAD + , the yield of DHA at the experimental setup of GldA (24.2-242 U mL −1 ), SpNox K184R (24.2-242 U mL −1 ) and glycerol (100-1000 mM) was quantified. As shown in Fig. 5, at pH 7.1, 300 mM potassium phosphate buffer, 30 °C, the highest yield of DHA, 92.3%, was achieved with 100 mM glycerol and 242 U mL −1 GldA and 242 U mL −1 SpNox K184R . To the best of our knowledge, this is the highest yield for enzymatic conversion of glycerol to DHA. After reaction, the enzyme will appear in a state of aggregation, the aggregated enzyme will lose activity. This may be caused by undesired enzyme interactions (Sanchez et al. 2016).

Chemoenzymatic conversion of glycerol to lactic acid
The conversion of DHA to lactic acid was performed in NaOH aqueous solution based on our previous study . Firstly, the enzymatic production of DHA with 100 mM glycerol as substrate was carried out and subsequently enzymes was removed from the reaction solution. Then, the reaction solution was added into 5 M NaOH solution with a volume ratio of 1:1. The reaction solution was incubated for 8 h (30 °C, 1000 rpm). Finally, the yield of lactic acid was detected by HPLC. DHA in NaOH solution was completely consumed and the major product was lactic acid. The yield of lactic acid from glycerol in the two-step chemoenzymatic system with GldA, SpNox K184R and NaOH as catalysts was 72.3% (Table 1). Using the coupled GldA-SpNox K184R and sodium hydroxide, the chemoenzymatic synthesis of lactic acid from glycerol was successfully established.

Chemoenzymatic conversion of glycerol to glycolic acid
The conversion of DHA to glycolic acid was carried out using NaClO 2 as oxidant. Firstly, the enzymatic conversion of glycerol to DHA with 100 mM glycerol as   substrate was carried out. Subsequently, enzymes were removed from the reaction solution by ultrafiltration. Then, NaClO 2 solution (four times of the concentration of DHA) was supplemented to the reaction system. The reaction mixture was incubated for 24 h (50 °C, 1000 rpm) and the yield of glycolic acid was detected by HPLC. DHA in NaClO 2 solution was completely consumed and the major product was glycolic acid. The yield of glycolic acid in the chemoenzymatic system with GldA, SpNox K184R and NaClO 2 as catalysts was 78.2% (Table 1).

Conclusions
In summary, two-step chemoenzymatic approaches for the conversion of glycerol into lactic acid and glycolic acid were established and optimized using coupled enzyme system and NaOH/NaClO 2 with high overall yield. Glycerol-dehydrogenase-GldA from Escherichia coli and NADH oxidase-SpNox K184R from Streptococcus Pyogenes were used to act together to achieve the dynamic balance of NAD + and NADH transformation, and convert glycerol to DHA. In the enzymatic step, a coenzyme recycling system was developed to convert glycerol into dihydroxyacetone (DHA) with a yield of 92.3% in potassium phosphate buffer (300 mM, pH 7.1) containing 100 mM glycerol, 2 mM NAD + , GldA and SpNox K184R concentration of 242 U/mL at 30 °C. The conversion of glycerol to DHA by the coenzyme cycle pathway reduced the generation of by-products and accelerated the catalytic efficiency of the enzymes. After the removal of the enzymes, NaOH or NaClO 2 catalyzes the formation of lactic acid and glycolic acid from DHA. The high yield of lactic acid (72.3%) and glycolic acid (78.2%) verify the benefit of the chemoenzymatic approaches. The chemoenzymatic production of lactic acid and glycolic acid from low-cost glycerol in this study opens a new route for the utilization of low-cost renewable compound.

Preparation of standard curve of NADH
A standard solution of NADH was prepared, diluted with ddH 2 O, and the absorbance was measured at 340 nm. The linear range was observed to be 0-1 mM. The standard curve of NADH is shown in Additional File 1: Figure S2.

Enzyme activity detection
The activity of GldA was assayed using purified enzyme (20 μg mL −1 ) at 30 °C in potassium phosphate buffer (300 mM K 2 HPO 4 and KH 2 PO 4 , pH 7.1) containing 2 mM NAD + and 100 mM glycerol. SpNox K184R was assayed using purified enzyme (25 μg mL −1 ) at 30 °C in potassium phosphate buffer (300 mM K 2 HPO 4 and KH 2 PO 4 , pH 7.1) containing 2 mM NADH. The increase and decrease of absorbance at 340 nm was detected for 20 min by a microtiter plate reader. One unit of enzyme activity was defined as the amount of enzyme required to produce 1 μmol of NADH per min (GldA). One unit of enzyme activity was defined as the amount of enzyme required to oxidize 1 μmol of NADH per min (SpNox K184R ). The data are average of two independent experiments (triplicates in each experiment). Enzyme activity calculation formula: where D: dilution rate; ∆A 1 : change in NADH absorbance; ∆A 2 : change in absorbance value of blank control; V t : the total reaction system, mL; e: molar absorbance, 6.22 × 10 -3 mol L −1 cm −2 ; V s : enzyme product, mL; d: optical path distance, cm; and c: protein concentration, mg mL −1 . Transforming the formula according to the actual requirements: where ΔT is the reaction time.
According to the standard curve of NADH, the generation of NADH (GldA) or the consumption of NADH (SpNox K184R ) were calculated and substituted into the formula to finally calculate the enzyme activity. (1)

Optimization of reaction buffer
Three kinds of buffer solutions (pH 7.1) were investigated as following: potassium phosphate buffer (100 mM K 2 HPO 4 and KH 2 PO 4 , pH 7.1), sodium phosphoric buffer (100 mM NaH 2 PO 4 and Na 2 HPO 4 , pH 7.1), Tris-HCl buffer (100 mM Tris-HCl, pH 7.1). GldA activity was assayed using purified enzyme (20 μg mL −1 ) at 30 °C in different buffer containing 2 mM NAD + and 100 mM glycerol. SpNox K184R was assayed using purified enzyme (25 μg mL −1 ) at 30 °C in different buffer containing 2 mM NADH. The increase and decrease of absorbance at 340 nm was detected for 20 min by a microtiter plate reader. One unit of enzyme activity was defined as the amount of enzyme required to produce 1 μmol of NADH per min (GldA). One unit of enzyme activity was defined as the amount of enzyme required to oxidize 1 μmol of NADH per min (SpNox K184R ). The data are average of three independent experiments (triplicates in each experiment). The optimum buffer was selected according to the activity of the two enzymes.

Optimization of the pH value and ionic strength of the buffer
According to the characteristics of the enzyme, pH 6.1-7.1 was selected as the enzyme activities were detected in different pH buffers. GldA activity was assayed using purified enzyme (20 μg mL −1 ) at 30 °C in different buffer containing 2 mM NAD + and 100 mM glycerol. SpNox K184R was assayed using purified enzyme (25 μg mL −1 ) at 30 °C in different buffer containing 2 mM NADH. The increase and decrease of absorbance at 340 nm was detected for 20 min by a microtiter plate reader. One unit of enzyme activity was defined as the amount of enzyme required to produce 1 μmol of NADH per min (GldA). One unit of enzyme activity was defined as the amount of enzyme required to oxidize 1 μmol of NADH per min (SpNox K184R ). The data are average of three independent experiments (triplicates in each experiment). The optimum buffer was selected according to the activity of the two enzymes. After selecting pH and buffer type, the buffer concentration was optimized. 20-500 mM potassium phosphate buffers (K 2 HPO 4 and KH 2 PO 4 ) were prepared at pH 7.1. GldA activity was assayed using purified enzyme (20 μg mL −1 ) at 30 °C in different buffer containing 2 mM NAD + and 100 mM glycerol. SpNox K184R was assayed using purified enzyme (25 μg mL −1 ) at 30 °C in different buffer containing 2 mM NADH. The increase and decrease of absorbance at 340 nm was detected for 20 min by a microtiter plate reader. One unit of enzyme activity was defined as the amount of enzyme required to produce 1 μmol of NADH per min (GldA). One unit of enzyme activity was defined as the amount of enzyme required to oxidize 1 μmol of NADH per min (SpNox K184R ). The data are average of three independent experiments (triplicates in each experiment). The optimum buffer was selected according to the activity of the two enzymes.

Enzymatic stability of GldA and SpNox K184R
In potassium phosphate buffer (300 mM, pH 7.1), GldA and SpNox K184R were stored at 30 °C for 25.5 h. Then the enzyme activity was detected and compared with the original enzyme activity to determine the stability of the enzyme.

Biooxidation of glycerol to DHA by GldA coupled with SpNox K184R
The ratio of GldA and SpNox K184R was adjusted according to the enzyme specific activity, with activity ratio of 1:1. The 500 μL reaction mixture in 30 °C, potassium phosphate buffer (300 mM K 2 HPO 4 and KH 2 PO 4 , pH7.1) containing 100 mM-1 M glycerol, 2 mM NAD + , 24.2-242 U mL −1 GldA, SpNox K184R was the same enzyme activity as GldA (the enzyme activity ratio of GldA and SpNox K184R is 1:1). The reactions were carried out in 4.5 mL tubes at 30 °C with shaking (1000 rpm) for 14 h.

The high-performance liquid chromatography (HPLC) analysis for DHA
The HPLC method was used for the detection of DHA.
After removing the enzymes using the concurrent filter unit (Millipore UFC500 3 kDa Filter Devices, The United States) through ultrafiltration at 8000 rpm for 20 min, 30 μL sample solution was added in 1.5 mL microcentrifuge tube containing 140 μL 20 mg mL −1 PFBHA solution (PFBHA was dissolved into citric acid buffer (0.1 M, citric acid and NaOH, pH 4) and adjusted to pH 4.0 with 1 M NaOH), 30 μL 50 mM veratryl alcohol (internal standard). After reaction at 1000 rpm, 30 °C for 30 min, 200 μL acetonitrile and 100 μL water were added in the tube. The samples were further filtered through 0.22 μm organic filtration membrane and was used for HPLC detection. Detection conditions after derivatization by HPLC (Thermo, UltiMate 3000): Ultimate XB-C18 column, 4.6 × 250 mm, 5 μm; mobile phase: water, acetonitrile; UV absorption wavelength: 263 nm; the flow rate: 1.2 mL min −1 ; column temperature: 30 °C; sample injection volume: 20 μL. The standard curve of DHA is shown in Additional File 1: Figure S3.

The synthesis of lactic acid from glycerol
Firstly, glycerol was converted to DHA using the coenzyme circulation system (500 μL) including 100 mM glycerol, 2 mM NAD + , 242 U mL −1 GldA and 242 U mL −1 SpNoX K184R and 300 mM potassium phosphate buffer (pH7.1). At the end of the reaction, the enzymes were removed and injected slowly into 500 μL NaOH (5 M) through the channel injection pump with a syringe at a rate of 100 μL h −1 . The reaction was conducted for 8 h (1000 rpm, 30 °C). After 8 h, reaction mixture was acidified with dilute sulfuric acid to pH 1-2 for HPLC analysis. The HPLC system was carried out on an Agilent 1260 system equipped with a UV detector (210 nm), and fitted with Sugar 10H column (DEVOTE): mobile phase: 5 mM H 2 SO 4 , flow rate: 0.5 mL min −1 , sample volume: 20 μL, column temperature: 35 °C. Triplicates in each experiment. The standard curve of lactic acid is shown in Additional File 1: Figure S4.

The synthesis of glycolic acid from glycerol
Firstly, glycerol was converted to DHA using the coenzyme circulation system (500 μL) including 100 mM glycerol, 2 mM NAD + , 242 U mL −1 GldA and 242 U mL −1 SpNoX K184R and 300 mM potassium phosphate buffer (pH7.1). The 500 μL reaction solution was added into 1.5 mL microcentrifuge tube containing 500 μL NaClO 2 solution (the concentration of NaClO 2 solution is about 4 times concentration of DHA). The reaction was conducted for 24 h (1000 rpm, 50 °C). After, the glycolic acid production was detected by HPLC directly. The HPLC system was carried out on an Agilent 1260 system equipped with a UV detector (210 nm), and fitted with Sugar 10H column (DEVOTE): mobile phase: 5 mM H 2 SO 4 , flow rate: 0.5 mL min −1 , sample volume: 20 μL, column temperature: 35 °C. Triplicates in each experiment. The standard curve of glycolic acid is shown in Additional File 1: Figure S5.

Yield calculation
The yields of all products were calculated with the following equations: (3) Yield (DHA from glycerol) = produced DHA glycerol × 100%, Yield (lactic acid from DHA) = produced lactic acid generated DHA × 100%, Yield (lactic acid from glycerol) = produced lactic acid glycerol × 100%, Yield (glycolic acid from DHA) = produced glycolic acid generated DHA × 100%, Yield (glycolic acid from glycerol) = produced glycolic acid glycerol × 100% HPLC analysis of chemoenzymatic conversion of glycerol to lactic acid and glycolic acid is shown in Additional File 1: Figure S6.