High efficient extracellular production of recombinant Streptomyces PMF phospholipase D in Escherichia coli

Background Currently, Streptomyces is widely used in the preparation of phospholipase D (PLD) with high transphosphatidylation activity. However, the yield of PLD from Streptomyces was low and the culture period was long. Therefore, an efficient and cost-effective method is needed urgently. Results Firstly, PLDs from Streptomyces PMF and Streptomyces racemochromogenes were separately over-expressed in E. coli to compare their transphosphatidylation activity based on the synthesis of phosphatidylserine (PS), and PLD PMF was determined to have higher activity. To further improve PLD PMF synthesis, a secretory expression system suitable for PLD PMF was constructed and optimized with diﬀerent signal peptides. The highest secretory efficiency was observed when the PLD PMF gene was expressed together with its native signal peptide (Nat) and the signal peptide PelB from E. coli . For the application of recombinant PLD to PS synthesis, the PLD properties were characterized and 30.2 g/L of PS was produced after 24 h of bioconversion when 50 g/L phosphatidylcholine (PC) was added. Conclusions We succeeded in over-expressing PLD from Streptomyces PMF in E. coli with high transphosphatidylation activity and enhanced the yield by secretory expression. The secreted PLD was successfully used in the production of PS. Our work makes the large-scale production of PLD and PS feasible.

advantages of high unit activity and low cost.
PLD has been characterized in many microorganisms and is most commonly found in Streptomyces strains, such as Streptomyces PMF [7], S. lividans [8], and Streptomyces sp. YU100 [9]. For microbial production of PLD, Streptomyces strains are most widely used due to the high transphosphatidylation activity of native PLD and natural PLD secretion activity during fermentation. For example, Saovanee et al. isolated Streptomyces sp. SC734 from soil-contaminated palm oil, and the PLD it produced exhibited high activity with a conversion rate of phosphatidylcholine (PC) to PS of up to 94.7% in 100 min [10]. Ogino et al. constructed an overexpression system for secretory production of PLD in S.
lividans and the amount of PLD secreted reached a maximum level of 118 mg/L [11]. However, the genetic-transfer systems for Streptomyces remain largely inefficient, which limits efficient production of PLD. Thus, the heterologous expression of Streptomyces PLDs in other model microorganisms, such as yeast or E. coli is highly desired.
Using Pichia pastoris as the host, Liu et al. developed a yeast cell surface display system to express PLD from S. chromofuscus, and the displayed PLD converted 67.5% of PC to PS within 10 h [12]. PLDs from different sources have also been successfully expressed with E. coli used as the host. For example, Carlo et al. expressed the PLD from Streptomyces PMF in E. coli BL21(DE3)pLysS, and 5 mg/L PLD was finally obtained with an enzyme activity of 15 mU/mL [13]. For the high-level and stable production of PLD, several engineering strategies were carried out in E. coli, including optimizing and tightly regulating promoter strength, optimizing codon usage and amino acid supplementation, and maintaining the best cellular state by supplementing nutrition. Finally, a large amount of PLD (81.5 mg/L) was obtained in batch culture [14]. Although there has been considerable progress in heterologous production of PLD, it is still not enough for industrial applications of PLD.
Developing an efficient expression system for PLD production is urgently needed.
One of the biggest obstacles to efficient PLD production is that overexpressed PLD is toxic to the host, which may cause plasmid instability, cell lysis and PLD leakage [14]. Secretory production of heterologous proteins has great advantages compared with conventional cytosolic protein production, especially when the heterologous proteins are toxic. In addition, the secretory production of heterologous proteins could simplify the purification processes and reduce cost since cell disruption is not required. Many reports have proposed strategies for improving the secretory production of heterologous proteins, such as optimizing the environmental conditions [15], constructing leaky strains [16] or co-expressing the secretory pathway [17]. The production of PLD in the secretory form seems to be a promising approach to address this issue.
In this study, PLDs from Streptomyces PMF and Streptomyces racemochromogenes were separately overexpressed in E. coli to compare their transphosphatidylation activity based on synthesis of PS.
Recombinant PLD PMF exhibited higher activity. To further improve the synthesis of PLD PMF , a secretory expression system suitable for PLD PMF was constructed and optimized with different signal peptides.
The highest secretory efficiency was observed when the PLD PMF gene was expressed together with its native signal peptides (Nat) and PelB. After optimizing induction conditions including induction temperature, induction pH, IPTG concentration, induction time and addition of metal ions, 10.5 U/ml PLD was detected in the fermentation medium. For the application of recombinant PLD to PS synthesis, the PLD properties were characterized and 30.2 g/L of PS was produced after bioconversion for 24 h when 50 g/L PC was added.

Materials And Methods Microorganisms and media
The strains used and constructed in this paper are listed in Table 1. The E. coli strains were cultured in Luria-Bertani medium (tryptone 10 g/L, NaCl 5 g/L and yeast extract 5 g/L) containing appropriate antibiotics at the following concentrations: 50 mg/L kanamycin (kana) and 100 mg/L ampicillin (Amp).
Plasmid pET28a was used as the original plasmid.  PhoA-Nat-PLD* pET22b derivative; P T7−lacO -PhoA-Nat-PLD* 6 respectively. The signal peptide genes OmpC, OmpF, OmpT, LamB, PhoA, MalE and pelB were synthesized by Sprin GenBioTech Co. LTD (Nanjing, China). The codon optimization procedure for the two PLD genes were conducted by Sprin GenBioTech Co. LTD (Nanjing, China). We used the primers P1 and P2 to amplify the PLD PMF gene and inserted it into the plasmid pET28a between the NcoI and EcoRI sites, yielding the recombinant plasmid pET28a-PLD PMF . The PLD SR fragment was amplified using the primer P3 with a NcoI restriction site and primer P4 with an EcoRI restriction site, and was ligated into pET-28a vector, yielding the plasmid pET28a-PLD SR . Table 2 Primers used in this study General reverse primer CTCGAGCGGAGCGTTGCAGATACC AC For secretory expression of PLD, PLD PMF was cloned into plasmid pET22b together with different signal peptides. Primers P5/P6 were used for PCR amplification of the PLD PMF gene containing the native signal peptide (Nat), while the primers P7/P8 were used to obtain the fragment OmpA-Nat-PLD*, and primers P9/P10 were used to obtain the fragment OmpA-PLD*. These three fragments were inserted into NcoI/XhoI sites of plasmid pET22b to yield the plasmids pET22b-Nat-PLD*, pET22b-OmpA-Nat-PLD* and pET22b-OmpA-PLD* respectively. To optimize the secretory efficiency, seven other signal peptides OmpC, OmpF, OmpT, LamB, PhoA, MalE and PelB were amplified using appropriate primers listed in Table 2 to replace OmpA of plasmid pET22b-OmpA-Nat-PLD*.

Protein expression and cell fractionation
The engineered E. coli was cultivated in 100 mL LB medium with 0.1 mM of appropriate antibiotics at 37 °C on a rotatory shaker (200 rpm). When the cell-culture density at 600 nm (OD 600 ) reached 0.6, 0.5 mM β-d-1-thiogalactopyranoside (IPTG) was added into the culture. Then the cells were incubated at 28 °C for 12 h. Cells were harvested by centrifugation at 8000 rpm for 10 min. The supernatant was concentrated eight times using a rotary evaporator (20 °C, 15 mbar) and was then used as the extracellular fraction. The collected cells were resuspended in water to an OD 600 of 20 and the crude extracts were prepared on ice by ultrasonication: 20 min pulsing (0.3 ms, 0.2 ms pause) at 40% amplitude.

Enzyme assay
The transphosphatidylation activity of PLD was measured according to the production of PS. One unit (U) was defined as 1 µmol PS produced per 1 min. To determine the PLD activity, a catalytic reaction was carried out in a two-phase system containing an aqueous phase and organic phase, and the volume ratio was 1:1. The aqueous phase contained 1 mL enzyme solution and 1 mL 0.2 M sodium acetate buffer (pH 5.5) containing 1.4 g/L serine. The organic phase was 2 mL methylene chloride containing 10 g/L PC. The molar ratio of substrate PC to serine was 1:60. The reaction mixture was incubated in a 200 rpm shaker at 30 ºC for 4 h. The reaction mixture was boiled for 5 min to completely arrest the reaction and the mixture was subsequently centrifuged at 6000 rpm for 5 min.
After that, a 100 µL organic sample was taken from the mixture and diluted 10 times with a mixture containing chloroform and methanol with a volume ratio of 1:

Intracellular expression of PLD in E. coli
The host strain E. coli BL21(DE3) is an efficient expression system for various recombinant proteins.
Here we attempted to use it for the production of PLD. Two PLDs in the plasmids pET28a-PLD PMF and pET28a-PLD SR were separately introduced into BL21(DE3). Enzyme production was induced by the addition of IPTG and the activity of crude PLD extracts was compared. As shown in Fig. 1, both PLDs were functionally expressed in E. coli, and crude extracts of the strain BL21(DE3)/pET28a-PLD PMF exhibited higher transphosphatidylation activity. Using the crude extracts of the strain BL21(DE3)/pET28a-PLD PMF for the bioconversion of PC to PS, PS reached 0.37 g/L after 8 h, which is 1.4-fold higher than that of BL21(DE3)/pET28a-PLD SR . Therefore, PLD PMF was applied to further optimize expression.

Secretory expression of PLD in E. coli by optimizing signal peptides
To investigate the secretory expression of PLD, the signal peptides Nat (Native signal peptide from PLD PMF ), OmpA and the fused signal peptide OmpA-Nat were co-expressed with the PLD* (PLD PMF without Nat sequence) gene (Fig. 2a). First, the effect of Nat and OmpA on the PLD secretory efficiency was compared. No PS was detected using the extracellular fraction of the strain BL21(DE3)/pET22b-Nat-PLD* for the conversion of PC to PS (Fig. 2b). In contrast, the PS yield reached 40.68% after bioconversion for 24 h using the extracellular fraction of the strain BL21(DE3)/pET22b-OmpA-PLD*, indicating that Nat is not functional for directing the secretion of heterologous PLD in E.
coli. Subsequently, to identify whether the cleavage of Nat sequence in the N-terminus of PLD PMF affected PLD activity, PLD* and PLD PMF were separately expressed after being fused with OmpA in E.
coli. We found that the PLD from the strain BL21(DE3)/pET22b-OmpA-Nat-PLD* exhibited higher transphosphatidylation activity with PS yield of 45.72% after bioconversion for 24 h. This result indicated that the Nat signal peptide is important for maintaining the transphosphatidylation activity of recombinant PLD expressed in E. coli. Thus, the expression of PLD PMF directed by E. coli homologous signal peptides was best for the secretory expression of PLD in E. coli.
To further determine the optimum signal peptide for directing secretion expression of PLD in E. coli, seven different signal peptides were employed to replace OmpA (Fig. 3a). As shown in Fig. 3b, different signal peptides directed export of PLD with varying efficiencies. Compared with OmpA, the signal peptides OmpF, OmpT, LamB and MalE are less efficient and lower PS yield was observed. In contrast, more efficient PLD secretion was obtained with the signal peptides OmpC, PhoA and PelB resulting in higher PS yield. Among them, the highest level of extracellular PLD was found after expressing the plasmid PelB-Nat-PLD* in E. coli, where PS yield was increased by 86.51% compared to that from OmpA. Thus, the recombinant strain BL21(DE3)/pET22b-PelB-Nat-PLD* with the highest PLD secretory expression activity was selected for the following experiment.

Effect of fermentation conditions on the secretory expression of PLD
To further improve PLD synthesis, fermentation conditions, including induction temperature, induction pH, cell density at induction and IPTG concentration were optimized. For the control group, the induction temperature was 28 °C, IPTG concentration was 0.5 mM, cultivation pH was 7.0 and the induction OD 600nm was 0.6. Induction temperature is an important factor influencing heterologous protein expression in E. coli (Fig. 4a). The recombinant strain was incubated at a temperature ranging from 16 °C to 36 °C, and the maximum PLD activity was obtained at 20 °C with an increase of 19.4% compared to the control group (Fig. 4a). The effect of the concentration of IPTG was evaluated by varying the concentration from 0.4 mM to 0.8 mM. The highest PLD activity was achieved when 0.7 mM IPTG was added, which resulted in a 68.2% increase in PS yield (Fig. 4b). The optimal induction OD 600 and time were also determined at the induction OD 600 of 1.4 after induction for 12 h ( Fig. 4c and 4d). Varying pH of the growth environment change bacterial metabolic pathways, which might negatively affect the expression of heterologous protein in E. coli. In addition, pH also affects the charge state on the cell surface, and thus the permeability of the cell membrane, which has an important impact on the exchange of substances and the secretion of recombinant proteins. When the engineered BL21(DE3)/pET22b-PelB-Nat-PLD* was cultivated at pH ranging from 5.0 to 8.0, the PLD activity reached the highest level at pH of 6.5 (Fig. 4e).
To further improve the secretory expression of recombinant PLD, the effects of surfactant addition were evaluated. As shown in Fig. 4f, seven different surfactants were separately added into the fermentation medium. Compared to the control group, all surfactants benefitted the secretory expression of PLD, among which, the group with 3 g/L Span60 added exhibited the best PLD activity.

Characterization of the recombinant PLD activity
To characterize the recombinant PLD activity, the effects of reaction temperature, pH and metal ion additives were evaluated. To determine the optimal reaction temperature, bioconversion was carried out at 20, 25, 30, 35, or 40 °C (Fig. 5c). From 20 °C and 30 °C, the PLD activity clearly increased with increasing temperature, and reached the highest level at 30 °C. When the temperature was higher than 30° C, the PLD activity sharply decreased, indicating the temperature sensitivity of recombinant PLD. The PLD activity increased with a rise in reaction pH (from 4.0 to 8.0) and reached a maximum at pH 6.5 (Fig. 5b).
Several metal ions have been reported to play an important role in maintaining the activity of PLD [18]. To evaluate the effect of metal ions on recombinant PLD activity, metal ions including Co 2+ , Ni 2+ , Zn 2+ , Cu 2+ , Ca 2+ , K + , Fe 2+ , Mg 2+ and Mn 2+ were added. The reaction performed without any metal ions served as the control group. The addition of Co 2+ , Ca 2+ and Mg 2+ showed positive effects on the PLD activity, and the addition of Ca 2+ gave the highest level of PLD activity (Fig. 5c).
The application of recombinant PLD PMF for the bioconversion of PC to PS The optimum expression system, fermentation conditions and PLD traits were determined based on the above results. With the extracellular recombinant PLD produced by engineered E. coli, the capacity of producing PS from PC by PLD was tested under the optimal reaction conditions. The reaction was performed with PC substrate concentrations of 10 g/L, 30 g/L and 50 g/L. Samples were taken at reaction times of 4 h, 8 h, 12 h and 24 h to detect the amount of product PS and substrate PC (Fig. 6). In the case of 10 g/L PC, after bioconversion of 24 h, the production of PS reached 9.2 g/L with a molar yield of 88.05%. When the concentration of PC was increased to 30 g/L, the final PS titer of 18.2 g/L was obtained with a molar yield of 58.02% after bioconversion of 24 h. Further increasing PC concentration to 50 g/L, increased PS titer to 30.2 g/L with a molar yield of 57.81%.

Discussion
Phospholipase has achieved significant attention in recent years for its applications in the production of various high value rare phospholipids [19]. To achieve high production of PLD, overexpression of native PLD or heterologous expression of various PLDs in model microorganisms, including E. coli [20], yeast [21], and Bacillus subtilis [22] have been performed. Among them, E. coli is the most frequently used host strain for the expression of heterologous proteins due to its well-characterized genetics, high protein expression levels and rapid growth rate [23]. However, the toxicity from overexpressing PLD has limited its production in E. coli. In this work, after screening PLD sources, a secretory PLD expression system was developed and optimized by investigating the effects of different signal peptides for efficient PLD production.
After determining a suitable source of PLD, OmpA, a signal peptide that has been reported to guide the secretory expression of heterologous proteins with high efficiency in E. coli [24], was first employed to direct the secretory production of recombinant PLD. The effect of native signal peptide (Nat) in PLD sequence was also evaluated. Fortunately, the PLD production efficiency was largely enhanced with the secretory expression system. Moreover, the presence of Nat signal peptides resulted in a higher extracellular PLD activity, suggesting that the Nat sequence might contribute to the correct fold of recombinant PLD in E. coli. It is well known that the N-terminal signal sequence can guide the protein to the Sec-translocon through the post-translational SecB-targeting pathway or the co-translational signal recognition particle (SRP)-targeting pathway and then fold correctly [25].
However, there is currently no general rule in selecting a proper signal sequence for a given recombinant protein [26]. To identify a more efficient signal peptide to direct PLD secretion in E. coli, seven other signal peptides including OmpF, OmpT, OmpC, LamB, MalE, PhoA, and PelB were compared. The highest secreted PLD activity occurred with PelB (Fig. 3b), indicating the important role of signal peptide for the efficient production of recombinant PLD.
For the secretory production of recombinant proteins, membrane permeability might be a limiting factor since the cellular membrane often retards the entry of substrate into the cellular systems and prevents the product from being released from the cellular system for an easy recovery [27]. For applying PLD to the synthesis of high value added phospholipid, the properties of recombinant PLD were also characterized. As shown in Fig. 5a, recombinant PLD is sensitive to temperature changes. The best pH for PLD activity was observed under pH 5.5, which coincides with other reports in which PLD exerted high transphosphatidylation activity in a weak acid environment [28]. For the metal ion additives, the highest PLD activity was observed with the addition of Ca 2+ [29]. Ca 2+ binding to PLD has been reported to cause a conformational change in the PLD that enhances binding of protein to zwitterionic interfaces [10]. Ca 2+ is also an activator when other soluble substrates are used [30]. Ca 2+ possibly coordinates with enzymes, improving their stability. Finally, the recombinant PLD was applied for the bioconversion of PC to PS.  Table 3, and the highest PS concentration so far was obtained in our study.

Availability Of Data And Materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
Not applicable.

Consent for publication
The present manuscript has been seen and approved by all authors. They agree with its content and submission to Microbial Cell Factories.