Overexpression of the Plant Medium-chain Acyl-carrier Protein Thioesterase BTE for the Overproduction of the nC14-surfactin Isoform with Promising MEOR Applications


 BackgroundSurfactin, a representative biosurfactant of popeptide mainly produced by Bacillus subtilis, consists of a cyclic heptapeptide linked to a β-hydroxy fatty acid chain. The functional activity of surfactin is closely related to the length and isomerism of the fatty acid chain. ResultsIn this study, the plant medium-chain acyl-carrier protein (ACP) thioesterase (BTE) from Umbellularia californica was overexpressed in a recombinant surfactin production strain based on B. subtilis 168. As a result, the surfactin yield after 24 h of cultivation improved by 23%, and the production rate increased from 0.112 to 0.177 g/L/h. The isoforms identified by RP-HPLC and GC-MS showed that the proportion of nC14-surfactin increased 6.4 times compared to the control strain. A comparison of further properties revealed that the product with more nC14-surfactin had higher surface activity and better performance in oil-washing. Finally, the product with more nC14-surfactin isoform had a higher hydrocarbon-emulsification index, and it increased the water-wettability of the oil-saturated silicate surface. Conclusion﻿The obtained results provide an original approach to modify the fatty acid chain of surfactin and further demonstrate the importance of the length and isomerism of the β-hydroxy fatty acid chain for the MEOR application of surfactin.


Background
Microbial enhanced oil recovery (MEOR) has been widely used in the eld of crude oil extraction as an environmentally-friendly and biodegradable alternative to chemical surfactants [1,2]. Strategies for the application of MEOR include in-situ and ex-situ approaches. The former often relies on injecting nutrients to activate the microbes that generate bioproducts inside the reservoir [3], while the latter is based on direct injection of highly concentrated fermented functional products into oil reservoirs to improve oil recovery [4]. During the exsitu MEOR process, functional products such as biosurfactants can be used directly as oil-displacing agents without considering the microorganism's adaptability to the extreme conditions inside the reservoir [5]. Among the reported biosurfactants, the cyclic lipopeptide surfactin has demonstrated the greatest potential for oil eld applications.
Surfactin is a secondary metabolite produced by Bacillus strains that can reduce the surface tension of water from 72 to 27 mN/m at a concentration of 1×10 -5 mol/L (10 mg/L) [6]. Moreover, it also possesses advantages of high temperature stability up to 121 o C [7], highly salt tolerance [8], wide pH adaptability [9] and high interfacial activity [10]. Based on these characteristics, a number of studies have shown that surfactin-based extraction has good prospects in MEOR [11]. The industrial application of surfactin is not limited by yield, since the surfactin production of engineered strains reached 10-20 g/L [12]. At present, the point of concern is the heterogeneity of surfactin structure, which closely affects the functional activity of surfactin preparations.
Surfactin is composed of a β-hydroxy fatty acid chain with a length of 13-15 carbon atoms connected to a peptide ring of 7 amino acids (L-Glu-L-Leu-D-Leu-L-Val-L-Asp-D-Leu-L-Leu) [13]. Due to the preference of B. subtilis for branched-chain fatty acids, the β-hydroxy fatty acid chain of surfactin is also mainly branched, accounting for about 78% of the total [14,15]. The isoforms of the β-hydroxy fatty acid moiety mainly include iso, anteiso and straight (n) chains, mainly with chain lengths of C 14 and C 15 [16,17]. Youssef et al. analyzed the relationship between the surface activity and the fatty acyl structure of 8 surfactin isomers, and the results showed that the iso-odd fatty acyl isomer has higher oil displacement activity than the n-even fatty acyl isomers [14]. Coutte et al. proved that C 14 surfactin has higher foaming capacity than C 13 and C 15 surfactin [18].
Our own research also showed that a higher C 15 -surfactin content results in better oil-washing and oildisplacement e ciency [11]. Therefore, the composition of surfactin isoforms in the reparation merits more attention than the total output of surfactin.
At present, strategies proposed to modify the structure of surfactin are all based on the metabolism of branchedchain amino acids, which are closely related to the fatty acyl moiety of surfactin [12]. For example, the proportion of surfactin with even β-hydroxy fatty acid components C 14 and C 16 increased with the addition of Arg, Gln or Val, whereas the addition of Cys, His, Ile, Leu, Met, and Ser enhanced the proportion of odd β-hydroxy fatty acids in B. subtilis TD7 [19]. Coutte et al. further revealed that knocking out the lpdV gene responsible for the nal degradation of branched-chain amino acids can increase the proportion of nC 14 surfactin 2.5 times [16].
Here, we propose a rational strategy for modifying the fatty acyl moiety of surfactin. We increased the proportion of nC 14 -surfactin by overexpressing the medium-chain acyl carrier protein (ACP) thioesterase BTE derived from the plant Umbellularia californica. The biosurfactant production, surfactin isoforms composition, and enhanced oil recovery properties including surface activity, oil-washing e ciency, emulsifying activity and wettability alteration were analyzed, revealing promising application potential of the production strain developed in this study.

Reagents and strain construction
Yeast extract and peptone were purchased from Oxoid (Hampshire, England). BSTFA (N,O-bis(trimethylsilyl)tri uoroacetamide) was purchased from Macklin (Shanghai, China). Other chemicals were purchased from the China National Pharmaceutical Group Corporation (Shanghai, China).
All strains and plasmids used in this study are listed in Table 1. The marker-free knockout and knock-in approach has been described in detail before [20,21]. Brie y, the left anking region (LF) (∼ 800 bp), target genes (IG), direct repeat (DR) sequence (∼ 500 bp), PC cassette (1900 bp), and right anking (RF) region (∼ 800 bp) fragments were rst ampli ed using appropriate primers (Table 2), and then fused using overlap-extension PCR in the order LF, DR, PC cassette, and RF or LF, IG, DR, PC cassette, and RF. The resulting puri ed PCR products LF-DR-PC-RF/ LF-IG-DR-PC-RF were used to transform corresponding competent cells and further selected on chloramphenicolcontaining agar plates and MGY-Cl medium. Oligonucleotide primers were synthesized by GenScript (Nanjing, China). The transformation of Bacillus was carried out according to published protocols [22,23]. B. subtilis BSFX022 was used as the parental strain to construct the strains BSFX023, BSFX024 and BSFX025. The surfactin concentration was determined using high-performance liquid chromatography (HPLC) on a U-3000 instrument (Thermo Fisher Scienti c, USA) equipped with an Amethyst C18-P column (4.6 × 250 mm, 5 μm). The mobile phase was composed of 90% (v/v) methanol and 10% (v/v) water, with 0.05% tri uoroacetic acid at a ow rate of 0.8 mL/min. Authentic surfactin (98%) was purchased from Sigma Aldrich (USA). The cell growth was monitored by measuring the optical density at 600 nm (OD 600 ).

Whole cell uorescence measurements
The coding sequence of green uorescent protein GFP fused with P veg promoter or the P 43 promoter, was inserted at the amyE locus of the genome of B. subtilis 168 using the primers shown in Table 2. The expression of GFP from the different promoters was monitored by measuring whole-cell uorescence using a Spectra Max M3 multimode microplate reader (Shanghai Huanxi Medical Equipment Co., Ltd, China). After culturing for 12 hours, 1 ml of fermentation broth was centrifuged at 10,956 × g for 10 minutes, the supernatant was discarded, and the cells were washed 3 times. The washed cells were re-suspended in deionized water to the same optical density (OD 600 ). The excitation wavelength and emission wavelengths were 485 and 525 nm respectively. B. subtilis 168 without the chromosomal gfp expression cassette was used as the negative control. Standard deviations are based on a minimum of three statistically independent experiments.

Isolation, puri cation and isoform analysis of surfactin
Surfactin was extracted and puri ed using the acid precipitation method [24]. After fermentation, the cells were removed by centrifugation at 10,956 × g for 10 min. Then, 6 mol/L HCl was added to the supernatant to achieve a pH of 2.0 for acid precipitation, and allowed to settle at 4 °C overnight. The acid precipitate was collected by centrifugation at10,956 × g for 10 min. The nal pH was adjusted to 7.0 with 5 mol/L NaOH, and the neutralized precipitate was lyophilized. The dried surfactin components were further extracted with methanol and dried on a rotary evaporator under vacuum.
Surfactin components were analyzed by reverse-phase UPLC-MS (UPLC, Agilent, 1290) coupled with a single quadrupole MS (Q-TOF, Agilent, 6550) on an extend C18 column (2.1×50mm 1.7µm; Agilent) using a method based on a acetonitrile/water (acidi ed with 0.1% formic acid) gradient that allowed the simultaneous detection of all three lipopeptide families. Elution was started at 10% acetonitrile at a ow rate of 0.50 mL/min. After 7 min, the percentage of acetonitrile was increased to 95% and held until 5 min. Then, the column was re-equilibrated with 10% acetonitrile for 1 min. The compounds produced by BSFX024 and BSFX025 were compared. Ionization and source conditions were set as follows: source temperature, 150˚C; desolvation temperature, 350˚C; nitrogen ow, 15 L/min; voltage, 4000V.
Fatty acid side-chain analysis The fatty acid side-chains were analyzed according to the method reported by Mu et al. [25]. Critical micelle concentration (CMC) of the biosurfactant As the concentration increases, the decreasing rate of surface tension will suddenly change at CMC [24]. The surfactin samples obtained from recombinant strains BSFX024 and BSFX025 were dissolved in distilled water at different concentrations (0-100 mg/L). Then, the surface tension was measured using the platinum plate method on an automated tensiometer (BZY-3B; Shanghai Automation Instrumentation Sales Center, China) at 25°C. The CMC was determined based on the in ection point of surface tension versus concentration.

Measurement of Emulsi cation Activity
For the measurement of emulsi cation activity, 2 ml of a solution containing 200 mg/L surfactin obtained from recombinant strain BSFX024 or BSFX025 and 2 ml of different hydrocarbons (dodecane, tetradecane, hexadecane, octadecane, p-xylene and liquid para n) were mixed in cylindrical glass vials, respectively. The mixtures were vortexed at maximum speed (QT-2; Qite Corp., Shanghai, China), and then incubated at 25°C for 24 h. The emulsi cation activity was calculated using the emulsi cation index (EI 24 ) formula [4]: Oil washing e ciency The standard oil sand for measuring the oil-washing e ciency was prepared according to a reported method [11,26]. Brie y, 170 g of quartz sand, 4 g of arti cial crude oil (Shengli Oil eld, China) and 10 ml of petroleum ether were mixed. The mixtures were heated at 80 ℃ for 1 h to remove the petroleum ether and then aged at 60 ℃ for 7 days. Subsequently, 2 g of the aged oil sand was placed into a ask with 20 ml of different surfactin

Results
In uence of the plant thioesterase BTE on the production of surfactin Expression of BTE in Escherichia coli was reported to promote the conversion of long-chain acyl ACP to C 12 and C 14 free fatty acids, which regulate the early steps in fatty acid biosynthesis [27]. Fatty acids are important precursors for the synthesis of surfactin and the core of the precursor supply module in the synthesis pathway of surfactin [28]. Here, to reduce competition for direct precursors, the nonribosomal peptide plipastatin synthetase gene pps and polyketide synthase gene pks were knocked out, as shown in Fig.1A. The obtained mutants were named BSFX023 (Δpps) and BSFX024 (Δpps, Δpks). Subsequently, the bte coding sequence under the control of the P veg promoter was integrated into the ackA locus (acetate kinase gene) to obtain the mutant BSFX025(Δpps, Δpks, bte). As shown in Fig. 1B, the surfactin titer did not improve signi cantly when the NRPS/PKS pathways were deleted, reaching 3.32 ± 0.08 g/L and 3.51 ± 0.083 g/L, respectively. The surfactin yield of BSFX025 reached 4.02 ± 0.085 g/L, representing a 34% improvement compared to BSFX022. To further analyze the in uence of BTE, the surfactin accumulation and strain growth was compared between BSFX024 and BSFX025, as depicted in Fig 1C. Although the expression of BTE had a weak inhibitory effect on the growth of the strain, the production rate of surfactin increased from 0.112 to 0.177 g/L/h at 24 h. These results indicated that the bte gene is an effective regulator of surfactin synthesis.
In our previous research, integrating the bte coding sequence under the control of the strong constitutive promoter P 43 into the genome signi cantly improved the production of surfactin by 23% [29]. Here, the bte coding sequence was placed under the control of the strong promoter P veg , in which the -35 and -10 regions were optimized for high transcriptional e ciency. The transcription intensity of P veg and P 43 was identi ed by measuring the GFP activity [30]. Fig. 1D illustrated the strength of the P veg and P 43 promoters according to the GFP uorescence measured using a microplate reader. P veg was more than 8-fold stronger than the P 43 promoter. This may be the reason why the output of surfactin was further improved in the corresponding strain.

HPLC-MS analysis of the isomeric composition of surfactin
HPLC analysis indicated that the overexpression of BTE also caused changes in the isomeric composition of surfactin. As shown in Fig. 2, a total of nine components were detected and there were obvious differences in the surfactin peaks of BSFX025 and BSFX024, especially the peaks at positions 1, 2, and 6. To further con rm the isomeric components of surfactin at each position, ESI-MS was performed. The molecular weight information of each component was obtained by mass spectrometry analysis (Figs. 2B and C). According to the results of mass spectrometry, we deduced the likely structure represented by each peak, as summarized in Table 3. The study of the peptide structure showed that Glu, Leu, Asp and Leu at positions 1, 3, 5 and 6 are usually conserved, and the amino acids at positions 2, 4 and 7 can be replaced by Leu, Val or Ile [31]. Due to the speci city of the C-domain of non-ribosomal peptide synthase (NRPS) for substrate recognition, the fatty acid chain length of surfactin is usually 13-15 [32].
According to the structural characteristics of surfactin and the m/z of each component, the likely structure represented by each peak was deduced as summarized in Table 3. The molecular mass of peak 1 was 979, with the major [M+H] + peak at m/z 980.63 the same as the reported of Mu et. al [33]. Therefore, its structure may be 1036 have been proved to contain C 13 -surfactin, C 14 -surfactin and C 15 -surfactin in our previous work [34]. It can be seen that the overexpression of BTE obviously enhanced the production of C 14 -surfactin and may even induce the production of C 12 -surfactin. However, to further con rm the detailed isoform of the fatty acid side-chain, GC-MS analysis was needed.  respectively, on the basis of the MS Agilent NIST 05 Chemical Structure Library (Fig, 3C). The proportion of each fatty acid component in the two surfactin samples was further summarized in Table 4, which clearly showed that the proportion of nC 14 -surfactin in the extract of strain BSFX025 was 6.4 times higher than that of BSFX024.  15 ) accounts for the highest proportion of the β-hydroxy fatty acid chains of surfactin [35]. In this study, the straight-chain fatty acid con guration became the main component in the β-hydroxy fatty acid side-chain of surfactin produced by strain BSFX025, which means that the properties and functional activity of surfactin may change accordingly. In order to further understand the in uence of fatty acid chain length and con guration on surfactin activity, the critical micelle concentration (CMC), emulsi cation activity, oil-washing e ciency and wettability of two surfactin samples were compared.
Comparison of the critical micelle concentration. CMC is a crucial parameter in the assessment of biosurfactants. Oil-washing e ciency of the surfactin produced by the two strains. The oil-washing e ciency can indirectly re ect the potential of surfactin for applications in MEOR. The performance of the surfactin produced by strains BSFX024 and BSFX025 in oil-sand washing at different concentrations is presented in Fig. 5. The difference of oil-washing e ciency become more signi cant as the concentration decreased. Fig. 5A shows photographs of dried oil sand after washing with different concentrations of surfactin. The de nite oil-washing e ciency value of the different surfactin solutions was further quanti ed as shown in Fig. 5B. The oil-washing e ciency of the two surfactin preparations was above 85% in both cases, and the color of the oil sands after washing were similar at concentrations of 0.2 and 0.15 g/L. However, at concentrations of 0.1 and 0.05 g/L, the color of the oil sands washed with the surfactin from strain BSFX025 was slightly lighter than that of the oil sand washed with the surfactin form strain BSFX024. Accordingly, the oil-washing e ciency was also relatively higher by 9.2% and 14.5%. Because the oil-washing e ciency in reservoirs is in uenced by two successive processes, stripping and emulsi cation of crude oil, it is necessary to compare the emulsi cation and wettability alteration capability of the two surfactin preparations.
Emulsi cation activity of surfactin preparations produced by the two strains. The amphipathic structure of surfactin gives it emulsifying properties. This feature can be used to reduce the viscosity of heavy oil, which is rich in asphaltenes and gums, and its high viscosity and poor uidity cause great di culties in oil exploitation [36]. Surfactin can emulsify strongly hydrophobic substrates such as pentadecane, diesel, hexane and kerosene [37][38][39][40]. A comparison of the emulsi cation activity of the two surfactin preparations at 200 mg/L is shown in Fig. 5.

Discussion
In this work, the overexpression of the plant thioesterase BTE changed the proportion of surfactin isoforms, resulting in surfactin with the nC 14 isoform as the predominant component. As reported, two kinds of thioesterases are useful in fatty acid production-the TesA thioesterase from E. coli and medium chain thioesterases from plants. Wu et al. overexpressed the TesA thioesterase from E. coli with a strong preference for long-chain acyl-ACP substrates and deleted the rst gene of the fatty acid degradation pathway in the engineered B. subtilis 168, which increased the output of surfactin by nearly 55% [28,41]. Regretfully, they did not report if there were any changes in the isomeric composition of surfactin. In fact, the proportion of nC 14 component is very small among the reported surfactin production strains (Table 5). For example, among the β-hydroxyl fatty acids of surfactin produced by the natural strain B. subtilis HSO121, the nC 14 component only accounts for 8.95% of the total [25], which is consistent with the results of our control strain BSF024. The proportion of these components can be altered to a certain extent by changing the medium composition, adding branched chain amino acids and speci c genetic modi cations. However, most of these changes are re ected in the isoand anteiso-components, while an improvement of the n-component is rarely seen. The only approach to speci cally overproduce surfactin with a C 14 FA chain was reported by Coutte et al., who increased the proportion of the surfactin C 14 isoform 2.5 times through genetic engineering of the branched-chain amino acid degradation pathway [16]. To our best knowledge, this is the rst study to show that the overexpression of thioesterase BTE can increased the proportion of nC 14 -surfactin by 6.4 times. The surface activity was enhanced due to the increase in the percentage of nC 14 -surfactin. In addition, it was observed that the product with more nC 14 -surfactin was more effective in washing oil sand, emulsifying hydrocarbons and altering the wettability of oil-wet solid. Youssef et al. showed that the optimal hydrophiliclipophilic balance required for the highest surface activity is exhibited by C 14 -surfactin [14]. Similarly, Bacon et al.
proved that C 14 -surfactin possesses higher surface activity [44]. This is consistent with our research results, indicating that C 14 -surfactin has the best surface activity among all the surfactin components reported so far.
Similarly, Yakimov et al. found that an increase in the percentage of branched-chain fatty acids in lichenysin A decreased the surface activity, whereas an increase in the percentage of straight-chain 3-hydroxytetradecanoate (n-3OH-C 14 ) enhanced the surface activity [45]. Liu et al. reported that a higher C 15 -surfactin content leads to better oil-washing e ciency at low concentrations. We found that the product with more nC 14 -surfactin also had a better performance in oil-washing. This phenomenon signi es that the isomerism of the fatty acid side-chain may have a greater in uence on the activity of surfactin than the length. The emulsi cation activity also depends on the structure of the biosurfactant. According to Tao et al., the hydrophilic group of surfactin stretches towards the water phase and the lipophilic group inserts itself into the oil phase, forming a dense lm on the interface [46].
The hydrophobic fatty acid chain of the surfactin produced by strain BSFX025 has a higher percentage of straight-chain 3-hydroxytetradecanoate. As a consequence, the emulsifying activity also changed signi cantly.
The emulsi cation capacity observed for the surfactin produced by strain BSFX025 suggests that it can potentially be used in the oil industry for cleaning the sludge in storage tanks, oil mobilization and MEOR.
Wettability alteration towards a strongly hydrophilic state is a favorable modi cation for obtaining higher oil recovery rates, because capillary forces will change from negative towards positive values due to this wettability alteration [47]. As a result, the changed surface activity is more conducive for water passing through porous rock, thereby further enhancing the displacement of trapped oil. The mechanisms of wettability alteration by SDBS and C12TAB indicates that the extent of wettability alteration is dictated by the structure of the applied surfactant. SDBS cannot alter the wettability towards a strongly hydrophilic state because of the larger numbers of hydrophobic side chain groups [48]. By contrast, C12TAB can induce better water-wettability of the surface due to the formation of ion pairs between the cationic heads of C12TAB and the acidic components of crude oil absorbed onto the rock surface [49]. Biosurfactants have been proved to alter the wettability of surfaces to the same extent as chemical surfactants, whereby the alteration of wettability is also related to the isomeric structure [50]. Here, the change in the hydrophobic fatty acid side-chain of surfactin caused the difference in wettability alteration of the two surfactin preparations. The increase in the percentage of straight-chain 3hydroxytetradecanoate led to stronger surface changes and better water-wettability. Further experiments should be carried out for improving the production of nC 14