Engineering of a Microbial Cell Factory for the Extracellular Production of Catalytically Active Phospholipase A₂ of Streptomyces violaceoruber

Abstract

Phospholipase A₂ (PLA₂) from Streptomyces violaceoruber is a lipolytic enzyme used in a wide range of industrial applications including production of lysolecithins and enzymatic degumming of edible oils. We have therefore investigated expression and secretion of PLA₂ in two workhorse microbes, Pichia pastoris and Escherichia coli. The PLA₂ was produced to an activity of 0.517 ± 0.012 U/ml in the culture broth of the recombinant P. pastoris. On the other hand, recombinant E. coli BL21 star (DE3), overexpressing the authentic PLA₂ (P-PLA₂), showed activity of 17.0 ± 1.3 U/ml in the intracellular fraction and 21.7 ± 0.7 U/ml in the culture broth. The extracellular PLA₂ activity obtained with the recombinant E. coli system was 3.2-fold higher than the corresponding value reached in a previous study, which employed recombinant E. coli BL21 (DE3) overexpressing codon-optimized PLA₂. Finally, we observed that the extracellular PLA₂ from the recombinant E. coli P-PLA₂ culture was able to hydrolyze 31.1 g/l of crude soybean lecithin, an industrial substrate, to a conversion yield of approximately 95%. The newly developed E. coli-based PLA₂ expression system led to extracellular production of PLA₂ to a productivity of 678 U/l•h, corresponding to 157-fold higher than that obtained with the P. pastoris-based system. This study will contribute to the extracellular production of a catalytically active PLA₂.

Introduction

Phospholipase A₂ (PLA₂, EC 3.1.1.4) hydrolyzes the ester bond in the sn-2 position of phospholipids, producing free fatty acids and the corresponding lysophospholipids. In comparison with native lecithins, lysolecithins prepared by PLA₂ not only exhibit enhanced O/W emulsifying properties but also form stable emulsions under various process conditions [1]. Thus, lysolecithins are used in a wide range of industrial applications such as food, cosmetics, and pharmaceuticals [2]. In particular, PLA₂ can be used extensively for enzymatic degumming, a key process in the refining of vegetable and other edible oils [3].

Previous studies have been mainly focused on expression and characterization of eukaryotic secretory PLA₂s. While eukaryotic PLA₂s have been successfully expressed in yeast [4-7] and fungus [8], inclusion bodies were formed when expressed in Escherichia coli due to the presence of five to eight disulfide bonds [9, 10]. Nevertheless, it would be desirable to establish an E. coli-based PLA₂ expression system because eukaryotic systems are generally considered as time consuming and uneconomic in comparison to prokaryotic systems. For this reason, several research groups developed expression systems for soluble expression of the eukaryotic PLA₂ in E. coli using maltose-binding protein (MBP) [11], thioredoxin [12], and protein disulfide bond isomerase (DsbC) [13] as fusion partners. These expression systems, however, require additional steps to eliminate fusion partners and hence are not economic for practical use [14].

It has been relatively easy to express prokaryotic PLA₂ in E. coli because it has only two or zero disulfide bonds [15]. The first PLA identified in prokaryotes was from Streptomyces violaceoruber A-2688, a soil bacterium [16]. It is a small protein with molecular weight of 14 kDa containing two disulfide bonds and requires Ca²⁺ for catalytic activity. The PLA₂ from S. violaceoruber was successfully produced extracellularly by P. pastoris [17] and E. coli [18]. The PLA₂ expressed in P. pastoris, however, contains a part of its signal sequence at the N-terminal end of mature PLA₂ protein, which might alter properties of the authentic PLA₂.

In this study, expression and secretion levels of the authentic PLA₂ in P. pastoris and E. coli were compared. Since the amount of extracellular PLA₂ produced in E. coli was 8.4 times higher than that in P. pastoris, we sought to develop an efficient PLA₂ expression system in E. coli. To do so, effects of the following factors on extracellular production of PLA₂ were systematically investigated: (1) codon optimization, (2) various host strains, and (3) attachment of aspartate tags.

Materials and Methods

Strains and Plasmids

E. coli TOP10 strain was used for genetic manipulation, and P. pastoris X-33, E. coli BL21 star (DE3), Origami 2 (DE3), BL21 (DE3), BL21 RIL (DE3), C41 (DE3), and C43 (DE3) strains were used for PLA₂ production. For expression of PLA₂ in P. pastoris, codon-optimized PLA₂ gene was cloned behind the AOX1 promoter in plasmid pPICZαA and their transcription was induced by adding methanol. Codon optimization was carried out by using the program (https://zendto.bioneer.co.kr/codon/index.py) provided by Bioneer (Korea). For expression of PLA₂ in E. coli, the natural and codon-optimized PLA₂ genes were located behind the T7 promoter in plasmid pET- 26b(+) and their transcription was induced by adding isopropyl-β-D-thiogalactopyranoside (IPTG). The natural and codon-optimized PLA₂ genes were synthesized by Bioneer. Strains and plasmids used in this study are listed in Table 1.

Table 1.

Strains and plasmids used in this study.

Name	Description	Reference
E. coli
E. coli TOP10	F- mcrA Δ(mrr-hsdRMS-mcrBC) Φ80lacZΔM15 ΔlacX74 recA1 araD139 Δ(ara-leu)7697 galU galK rpsL(StrR) endA1 nupG	Invitrogen
(Carlsbad, CA, USA)
E. coli BL21 (DE3)	F- ompT hsdS (rB- mB-) dcm gal (DE3 [lacI lacUV5-T7 gene 1 ind1 sam7 nin5])	Novagen
(Darmstadt, Germany)
E. coli BL21 star (DE3)	BL21 rne131 (DE3)	Invitrogen
E. coli BL21 CodonPlus-RIL (DE3)	BL21 (DE3) dcm+ Tetr endA Hte [argU ileY leuW Cam′]	Agilent technologies
(Santa Clara, CA, USA)
E. coli C41 (DE3)	BL21 (DE3 [lacI lac-T7 gene 1 ind1 sam7 nin5])	Lucigen
(Middleton, WI, USA)
E. coli C43 (DE3)	C41 (DE3) derivative	Lucigen
E. coli Origami (DE3)	∆ara–leu7697 ∆lacX74 ∆phoAPvuII phoR araD139 ahpC galE galK rpsL F'[lac+(lacIq)pro] gor522 ::Tn10 (TetR) trxB::kan (DE3)	Novagen
SK33	BL21 star (DE3) containing pET-26b(+)	This study
SK31	BL21 star (DE3) containing pHLK01	This study
SK32	BL21 star (DE3) containing pHLK02	This study
SK35	Origami (DE3) containing pHLK01	This study
SK36	BL21 (DE3) containing pHLK01	This study
SK37	BL21 RIL (DE3) containing pHLK01	This study
SK38	C41 (DE3) containing pHLK01	This study
SK39	C43 (DE3) containing pHLK01	This study
SK54	BL21 star (DE3) containing pHLK03	This study
SK55	BL21 star (DE3) containing pHLK04	This study
SK56	BL21 star(DE3) containing pHLK05	This study
SK57	BL21 star(DE3) containing pHLK06	This study
SK87	BL21 star(DE3) containing pSHK01	This study
P. pastoris
P. pastoris X-33	Wild type	Invitrogen
PX	X-33 containing pPICZαA	This study
PP	X-33 containing pMFα-PLA2	This study
Plasmids
pET-26b(+)	pBR322 origin, T7 promoter, PelB signal sequence, His-tag, KanR	Novagen
pHLK01	Expression vector containing P-PLA2, KanR	This study
pHLK02	Expression vector containing P-Opt. PLA2, KanR	This study
pHLK03	Expression vector containing P-D3-PLA2, KanR	This study
pHLK04	Expression vector containing P-D5-PLA2, KanR	This study
pHLK05	Expression vector containing P-D7-PLA2, KanR	This study
pHLK06	Expression vector containing P-D9-PLA2, KanR	This study
pSHK01	Expression vector containing PLA2, KanR	This study
pPICZαA	pUC origin, AOX1 promoter, MFα signal sequence, His-tag, ZeocinR	Invitrogen
pMFα-PLA2	Expression vector containing M-Opt. PLA2, ZeocinR	This study

Genetic Manipulation

The natural PLA₂ or codon-optimized PLA₂ genes without native signal sequence (Fig. S1) were PCR amplified with primers of HL01 (with MscI site) and HL02 (with XhoI site) or HL03 (with MscI site) and HL04 (with XhoI site). After the gene amplification, PCR products were cut with MscI and XhoI and then ligated with plasmid pET- 26b(+) digested with the same enzymes to construct pHLK01 and pHLK02 (Table 1). To attach various lengths of aspartate residues at the N-terminal end of PLA₂, plasmid pHLK01 was amplified with the primer sets and then ligated after MscI treatment. The primer sets used for amplification of the DNA fragments are as follows: HL07 (with MscI site) and HL11 (with MscI site) for pHLK03; HL08 (with MscI site) and HL11 (with MscI site) for pHLK04; HL09 (with MscI site) and HL11 (with MscI site) for pHLK05; HL10 (with MscI site) and HL11 (with MscI site) for pHLK06. Plasmid pSHK01 is identical to pHLK01 except that it does not contain the PelB signal sequence. To make this change, a DNA fragment without PelB signal sequence was amplified with primers SH05 (with NdeI site) and SH06 (with NdeI site) using pHLK01 as template. This linear DNA was digested with NdeI and ligated to construct pSHK01.

The codon-optimized PLA₂ gene for expression in P. pastoris was PCR-amplified with primers F_PLA₂ (with NheI site) and R_PLA₂ (with SpeI site). The plasmid pPICZαA was amplified with primers F_pPICZαA (with NheI site) and R_pPICZαA (with SpeI site). These two linear DNA fragments were digested with NheI and SpeI, and ligated to construct pMFα-PLA₂. Transformation of the cassette for overexpressing PLA₂ was performed using the Pichia EasyComp Kit (Invitrogen, USA). Plasmid pMFα-PLA₂ was cut with MssI and transformed. Transformants were selected on YPDS medium (10 g/l yeast extract, 20 g/l peptone, 20 g/l glucose, and 182 g/l sorbitol) containing 100 μg/ml Zeocin. PCR amplification was done with primers (F_ch_AOX1p and R_ch_pPICZα) to verify positive transformants. All plasmids and the check PCR products were sequenced by automatic sequencing (Cosmogenetech, Korea). Names of recombinant PLA₂ gene products and schematic structures are shown in Fig. 1, and primers used for plasmid constructions and confirmations are listed in Table S1.

Fig. 1. Schematic diagrams of the structures of recombinant PLA₂ expression cassettes.

Symbols: AOX1 promoter (AOX1p), T7 promoter (T7p), translational stop codon (stop), and the genes coding for the signal sequence of S. cerevisiae mating factor α (MFα), the signal sequence of pectate lyase B from Erwinia carotovora (PelB), S. violaceoruber phospholipase A₂ (PLA₂), codon-optimized PLA₂ (Opt. PLA₂), 3 aspartates (D3), 5 aspartates (D5), 7 aspartates (D7), and 9 aspartates (D9).

Media and Culture Conditions

P. pastoris was pre-cultured in 100 ml BMGY medium (10 g/l yeast extract, 20 g/l peptone, 13.4 g/l yeast nitrogen base, 3 × 10⁻⁴ g/l biotin, 10 g/l glycerol, and 100 mM potassium phosphate (pH 6.0)) at 30oC and 200 rpm for 24 h. Pre-cultured cells were then inoculated into 100 ml BMMY medium containing 10 g/l yeast extract, 20 g/ l peptone, 13.4 g/l yeast nitrogen base, 3 × 10⁻⁴ g/l biotin, 5 or 10 g/l methanol, and 100 mM potassium phosphate (pH 6.0). Expression of PLA₂ was induced by adding methanol every 24 h at a final concentration of 5 or 10 g/l.

E. coli cells were pre-cultured in LB medium (5 g/l yeast extract and 10 g/l bacto-trypton) at 37oC and 230 rpm for 12 h. After harvesting the cells, the cell pellets were used for inoculation. Batch fermentations were carried out in a 500 ml baffled flask containing 100 ml of Riesenberg medium [13.5 g/l KH2PO4, 4.0 g/l (NH₄)₂HPO₄, 1.7 g/l citric acid, 1.4 g/l MgSO₄•7H₂O, 10 ml/l trace element solution (10 g/l Fe(III) citrate, 2.25 g/l ZnSO₄•7H₂O, 1.0 g/l CuSO₄•5H₂O, 0.35 g/l MnSO₄•H₂O, 0.23 g/l Na₂B₄O₇•10H2O, 0.11 g/l (NH₄)₆Mo₇O₂₄, 2.0 g/l CaCl₂•2H₂O), pH 6.8] with 20 g/l glucose. Agitation speed was maintained at 200 rpm. When OD₆₀₀ reached 0.8-1.2, 0.2 mM IPTG was added to the culture broth. After induction, cultivation was continued at 25oC for an additional 24 h.

Preparation of Protein

After IPTG induction of 24 h, the culture broth was centrifuged at 15,000 ×g for 10 min to collect the medium fraction. The remaining pellet was resuspended in B-PERTM reagent (Thermo Fisher Scientific, USA) and lysed as specified by the manufacturer. The total, soluble, and insoluble fractions of intracellular proteins were prepared as described in the previous report [19].

Protein Purification

A 20 ml-scale column containing 750 μl of Ni-NTA agarose (QIAGEN, Germany) was washed with 20 ml of the His-tag binding buffer (pH 7.4) containing 20 mM NaH₂PO₄, 20 mM Na₂HPO₄, 0.5 M NaCl, and 40 mM imidazole. After 100 ml of the medium fraction prepared as described above was mixed with 300 ml of the His-tag binding buffer, the prepared mixture was loaded into the column. The proteins eluted from the column were collected, and protein concentration was determined using the Bio-Rad protein assay kit with bovine serum albumin (BSA) as the standard.

Analysis of Protein Expression and Enzyme Assay

To visualize recombinant PLA₂s, the protein samples were electrophoresed in 12% sodium dodecyl sulfate– polyacrylamide gel, and were either stained using Coomassie blue or were transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; EMD Millipore, USA). The membrane was then probed with anti- 6X His tag antibody (abcam, UK). After incubating the membrane with goat F(ab')2 Anti-mouse IgG (abcam) conjugated to alkaline phosphatase, the blot was developed using the BCIP/NBT chromogenic substrate solution (SurModics, Eden Prairie, USA) as specified by the manufacturer. The quantification of band intensity was carried out using the densitometry software (Total Lab 1.01; Nonlinear Dynamics Ltd.).

PLA₂ activity was measured using the sPLA₂ Assay Kit (Cayman Chemical, USA) according to the manufacturer’s instructions. The absorbance change at 37°C and 414 nm of wavelength was monitored by a spectrophotometer (OPTIZEN POP, Mecasys, Korea) after addition of enzyme solution. One unit (U) of PLA₂ activity was defined as the amount of PLA₂ able to hydrolyze 1 μmol of diheptanoyl thio-phosphatidylcholine in one minute.

Biotransformation of Soybean Lecithin

The 10 ml of reaction mixture was formulated with 0.5 mM Tris-HCl, 6 mM CaCl₂, 31.1 g/l crude soybean lecithin (Sigma-Aldrich, catalog number P3644) (pH 8.0), and 10% (v/v) of enzyme solution. The reaction conditions of 37°C and 400 rpm were maintained using a stirring heating mantle (LKLAB KOREA, Korea). According to the manufacturer’s information, soybean lecithin consists of an average of 55% (42–63%) L-α- phosphatidylcholine and 20% (10-32%) phosphatidylethanolamine. Concentrations of fatty acids were determined by gas chromatography/mass spectrophotometry (GC/MS), as previously reported [20-22]. Fatty acids present in 500 μl of samples were extracted by mixing with 2 ml of isopropyl alcohol, 500 μl of heptane, and 50 μl of sulfuric acid. For the derivatization, 25 μl of N-methyl-N-(trimethylsilyl)trifluoroacetamide (TMS) (TCI Chemicals, Tokyo, Japan) dissolved in 75 μl of pyridine was added to 60 μl of sample solution containing lauric acid (TCI Chemicals) as the internal standard. Concentrations of TMS derivatives were determined using GC/MS (Agilent Technologies, USA) equipped with a flame ionization detector, split injection system, and nonpolar capillary column (30 m length, 0.25 mm film thickness, HP-5MS, Agilent Technologies). Column temperature was controlled by the following gradient program: 235°C for 3 min; increase at a rate of 25°C/min; 270°C for 10 min; increase at a rate of 5°C/min; 300°C for 1 min. Mass spectra and scan spectra were obtained by electron impact ionization at 70 eV and within the range of 100–600 m/z, respectively. Selected ion monitoring was used for the detection and fragmentation analysis of the reaction products.

Results and Discussion

Production of PLA₂ in P. pastoris X-33

To construct expression plasmid for the PLA₂ from S. violaceoruber in P. pastoris X-33, the codon-optimized PLA₂ gene without its native signal sequence was cloned into pMFα-PLA₂ vector under the transcriptional control of the alcohol oxidase 1 (AOX1) gene promoter [23] containing a signal sequence of S. cerevisiae mating factor α (MFα) [24]. A schematic diagram for the PLA₂ expression cassette in pPICZαA plasmid and its name are displayed in Fig. 1. We noted that recombinant PLA₂ expressed previously in P. pastoris (PLA₂-Pp) was designed to contain a part of its native signal sequences (Ala-Pro-Pro-Gln-Ala) [17] whereas these five amino acids are not present in the authentic mature PLA₂ and recombinant PLA₂ produced in other previous studies (PLA₂-Ec), which employed E. coli as a host strain (Fig. S2) [16, 18, 25]. In addition to the presence of native signal sequences, amino acid sequences of the PLA₂-Pp were not identical (Fig. S2) to those of the PLA₂-Ec because these two PLA₂s were originated from different S. violaceoruber sources: the PLA₂-Pp was from S. violaceoruber 2917 whereas the PLA₂- Ec was from S. violaceoruber A-2688. The presence of additional five amino acids at the N-terminal end of PLA₂- Pp resulted in a lower optimum pH of 6.0 [17] compared to the PLA₂-Ec, which has optimum pH of 7.3–8.3. In addition to optimum pH, this factor might alter the expression level and some properties of the enzyme as reported for lipase B from Candida antarctica (CalB) [26]. Therefore, the PLA₂-Ec, which does not have its native signal peptide, was used in this study for accurate comparison of PLA₂ production in P. pastoris and E. coli.

As expected, growth of the control strain containing the empty plasmid (pPICZαA) and the P. pastoris X-33 harboring pMFα-PLA₂ was virtually identical regardless of methanol concentrations (Fig. 2A), indicating that expression of PLA₂ in P. pastoris had no obvious detrimental effect on growth in general. A batch fermentation of the P. pastoris X-33 harboring pMFα-PLA₂ with intermittent addition of 1.0% methanol led to an extracellular production of PLA₂-Ec to an activity of 0.517 ± 0.012 U/ml in 120 h (Fig. 2B) (see the Materials and Methods for the activity assay). This value is much lower than the corresponding value (34.7 U/ml) obtained by a batch fermentation of P. pastoris overexpressing the PLA₂-Pp [17]. This is likely due to the difference of PLA₂ sequences and activity assay methods. While the extracellular PLA₂ activity with addition of 0.5% methanol was similar to that with 1.0% methanol, addition of 1.0% methanol shortened overall fermentation time from 144 h to 120 h. Although PLA₂ was difficult to identify using Coomassie blue staining, it was clearly detected by western hybridization analysis using monoclonal anti-His antibodies (Fig. 2C). A band corresponding to the 14 kDa predicted molecular mass of PLA₂ was visible in both 0.5% and 1.0% methanol induction conditions. A protein band of approximate molecular mass of 16 kDa was also detected (Fig. 2B), and we speculated that this protein band corresponds to glycosylated PLA₂. This result is consistent with a previous study showing that a part of PLA₂ expressed in P. pastoris was glycosylated as it has three putative glycosylation sites [17].

Fig. 2. Batch production of PLA₂ in recombinant P. pastoris.

(A and B) growth curves (A) and extracellular PLA₂ activities (B) of P. pastoris X-33 harboring pPICZαA (Control) and pMFα-PLA₂. Batch production of PLA₂ was induced in duplicate by adding 0.5% (v/w) or 1.0% methanol every 24 h. The activities of crude PLA₂s in the extracellular fraction were measured in triplicate using diheptanoyl thio-phosphatidylcholine as a substrate. (C) Western blotting for His-tagged PLA₂ from the extracellular fraction of the recombinant P. pastoris strains. Lanes: M, prestained SDS-PAGE standards; 1, the control strain; 2, P. pastoris X-33 harboring pMFα-PLA₂ induced with 0.5% methanol; 2, P. pastoris X-33 harboring pMFα-PLA₂ induced with 1.0% methanol. The thin and thick arrows point to the protein bands of PLA₂ with and without glycosylation, respectively.

Production of PLA₂ in Various E. coli Strains

The authentic PLA₂ gene contains several rare codons for E. coli including Leu (CTC). This codon bias problem could be solved by codon optimization of the gene or by supplying rare-codon tRNAs. Here, we investigated effects of codon optimization of PLA₂ gene on its expression and secretion. The PLA₂ gene expression system was constructed with and without the PelB signal sequence (Fig. 1), which is involved in targeting the proteins to the periplasmic space [27, 28]. As expected, the PLA₂ without the signal sequence showed a basal level of lipase activity in both intracellular and extracellular fractions (Fig. 3A). On the other hand, the lipase activities increased up to 9.4 ± 1.5 U/ml in the intracellular fraction and 16.1 ± 0.9 U/ml in the culture broth of the recombinant E. coli BL21 star (DE3) overexpressing the authentic PLA₂ gene (P-PLA₂). This is 6.3- and 4.8-times higher than the corresponding values of the case of codon-optimized PLA₂ (P-Opt. PLA₂) (Fig. 3A). In addition to the enzyme activity assay, SDS–PAGE analysis showed the high secretion of P-PLA₂ in the culture medium (Fig. S3). This study and earlier studies [29, 30] suggest that a faster expression from the optimized gene could lead to higher concentration of target protein, which in turn results in degradation and/or misfolding of the protein.

Fig. 3. Effects of codon optimization (A) and E. coli host strains (B) on activities of recombinant PLA₂ in intracellular and extracellular fractions.

The activities of crude PLA₂s in the soluble and extracellular fractions (see the Materials and Methods for details) collected 24 h after IPTG induction were measured in triplicate using diheptanoyl thio-phosphatidylcholine as a substrate.

Protein expression in E. coli BL21 (DE3), BL21 RIL (DE3), BL21 star (DE3), Origami 2 (DE3), C41 (DE3), and C43 (DE3) were analyzed by SDS-PAGE to select a host for PLA₂ production. The expression level of PLA₂ was the highest in E. coli BL21 star (DE3) which has a mutation in the gene encoding RNaseE (rne131 mutation), indicating that protection of mRNAs from RNases plays an important role in PLA₂ expression (Fig. S4). Therefore, the highest activities in both intracellular and extracellular fractions were obtained for recombinant E. coli BL21 star (DE3) overexpressing P-PLA₂ (Fig. 3B). PLA₂ activities in culture broth of E. coli BL21 (DE3) and BL21 RIL (DE3) exhibited 63.6 and 14.3% of extracellular activity in E. coli BL21 star (DE3). Thus, E. coli BL21 star (DE3) strain was chosen as the host of PLA₂ production.

We concluded from these data that alleviating codon bias by codon optimizing the PLA₂ gene or by supplying rare-codon tRNAs has negative effects on correct folding of PLA₂. This conclusion supports the hypothesis that the translation is a bottleneck in functional expression of PLA₂, and hence an overall delay in PLA₂ expression gives protein machineries more time to fold PLA₂ correctly.

Hydrolysis of Soybean Lecithin by Extracellular PLA₂s from the Recombinant E. coli and P. pastoris

The extracellular PLA₂ activities of the recombinant E. coli P-PLA₂ and P. pastoris X-33 M-Opt. PLA₂ were examined by using an industrial substrate (i.e., crude soybean lecithin). When the extracellular fraction of P. pastoris X-33 M-Opt. PLA₂ culture (shown in Fig. 2B) was added into the reaction medium containing 31.1 g/l of crude soybean lecithin (see the Materials and Methods for details), linoleic acid, which was the major fatty acid constituent of soybean lecithin, was produced to 20.1 mM at t = 240 min (Fig. 4). This indicated that approximately 70% of soybean lecithin was hydrolyzed into lysolecithin and linoleic acid. The extracellular fraction of E. coli P-PLA₂ culture displayed a biotransformation profile similar to that of P. pastoris X-33 M-Opt. PLA₂ culture (Fig. 4). Remarkably, linoleic acid was produced to 36.3 mM at t = 240 min (Fig. 4). This indicated that approximately 95% of soybean lecithin was converted into lysolecithin and linoleic acid. Moreover, the initial conversion rate was 2.6-fold greater than that of P. pastoris X-33 M-Opt. PLA₂ culture. Besides, the cultivation time of E. coli P-PLA₂ was significantly shorter than that of P. pastoris X-33 M-Opt. PLA₂ (32 h vs. 120 h). It was thereby assumed that the E. coli-based PLA₂ expression system would be superior to the P. pastoris system in terms of extracellular PLA₂ productivity.

Fig. 4. Biotransformation of crude soybean lecithin into linoleic acid and lysolecithin by recombinant PLA₂ collected from recombinant P. pastoris X-33 and E. coli BL21 star (DE3) overexpressing PLA₂.

Biotransformation was initiated by adding 10-fold concentrated extracellular crude enzyme solutions from the recombinant P. pastoris X-33 and E. coli BL21 star (DE3) to the reaction mixture consisting of 0.5 mM Tris-HCl, 6 mM CaCl₂, and 31.1 g/l crude soybean lecithin (pH 8.0). Results are the mean of triplicate experiments and error bars indicate standard deviations.

Effects of N-terminal Repeat of Aspartate Residues on Specific Activity and Expression of PLA₂

We previously reported that fusion tag systems composed of the PelB signal sequence and repeated aspartate tags improved both expression and secretion of CalB and asparaginase isozyme II (AnsB) from E. coli [31, 32]. To investigate whether or not repeated aspartate residues would improve the secretion and activity of PLA₂, various lengths of aspartate residues were introduced into the N-terminal end of PLA₂ gene to construct the cassettes P- D3-PLA₂, P-D5-PLA₂, P-D7-PLA₂, and P-D9-PLA₂ as shown in Fig. 1. Crude PLA₂ enzymes present in the intracellular and extracellular fractions were subjected to SDS-PAGE (Fig. S5) and activity (Fig. 5A) analyses. Among a series of repeated amino acids consisting of 3, 5, 7, or 9 aspartates, the three aspartates facilitated the secretion of PLA₂, and hence comparison of the band intensities from the extracellular fractions showed that the band corresponding to the P-D3-PLA₂ was 64% greater than P-PLA₂ (Fig. S5). However, the intracellular and extracellular lipase activities obtained for recombinant E. coli BL21 star (DE3) overexpressing P-D3-PLA₂ were instead 42.6% and 32.3% lower than the corresponding values obtained in the case of P-PLA₂ (Fig. 5A). These results suggested that the presence of three aspartate residues at the N-terminal end of PLA₂ might alter the specific activity of PLA₂. To confirm the hypothesis, PLA₂ and P-D3-PLA₂ were His-tag purified and subjected to activity assay. As expected, specific activity of P-D3-PLA₂ was 6.1 times lower than that of PLA₂ (Fig. 5B). This result is consistent with previous studies reporting that the attachment of repeated aspartates altered catalytic efficiency of CalB and α-1,2-fucosyltransferase (FucT2) from Helicobacter pylori [31, 33]. More research is in progress to find PLA₂ from other bacteria with increased stability, of which specific activity is not affected significantly by the attachment of repeated aspartates.

Fig. 5. Activity assays of recombinant PLA₂s with various lengths of aspartate tags (A) and His-tag purified P-PLA₂ and P-D3-PLA₂ (B) to investigate the effects of aspartate tags on expression in E. coli and specific activity of PLA₂.

Results are the mean of triplicate experiments and error bars indicate standard deviations.

In conclusion, this study demonstrated that an E. coli-based PLA₂ production system could be more efficient in terms of PLA₂ productivity, as compared to the P. pastoris-based system. Among the E. coli host strains harboring the authentic PLA₂ gene (P-PLA₂) or codon- optimized PLA₂ (P-Opt. PLA₂), the recombinant E. coli BL21 star (DE3) P-PLA₂ has exhibited the highest activities of 21.7 ± 0.7 U/ml in the culture broth and 17.0 ± 1.3 U/ml in the intracellular fraction. Moreover, the extracellular PLA₂s from the recombinant E. coli P-PLA₂ culture was able to hydrolyze 31.1 g/l of crude soybean lecithin to linoleic acid and lysolecithin at a conversion yield of at least 95%. Therefore, it was concluded that the recombinant E. coli P-PLA₂ system could be used as a microbial cell factory to produce a catalytically active PLA₂ for hydrolysis of the selective sn-2 position of plant lecithins.

Supplemental Materials

Supplementary data for this paper are available on-line only at http://jmb.or.kr.