High-Yield and Sustainable Production of Phosphatidylserine in

Nov 24, 2017 - Theoretically, the production cost of PS can be decreased significantly through the phospholipase D (PLD)-mediated transphosphatidylati...
1 downloads 6 Views 3MB Size
Download PDF
Article Cite This: J. Agric. Food Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/JAFC

High-Yield and Sustainable Production of Phosphatidylserine in Purely Aqueous Solutions via Adsorption of Phosphatidylcholine on Triton-X-100-Modified Silica Xiaoli Zhang,*,‡ Binglin Li,‡ Jiao Wang, Huanyu Li, and Binxia Zhao Department of Chemical Engineering, Northwest University, 229 North Taibai Road, Xi’an, 710000 Shaanxi, China S Supporting Information *

ABSTRACT: Triton X-100 was covalently bound to a surface of silica and acted as an anchor molecule to facilitate the adsorption of phosphatidylcholine (PC) in a purely aqueous solution. The silica-adsorbed PC obtained was successfully used for phospholipase D (PLD)-mediated transphosphatidylation in the production of phosphatidylserine (PS). Organic solvents were completely avoided in the whole production process. The PC loading and PS yield reached 98.9 and 99.0%, respectively. Two adsorption models were studied, and the relevant parameters were calculated to help us understand the adsorption and reaction processes deeply. In addition, the silica-adsorbed PC provides a promising way to continuously biosynthesize PS. A packed-bed reactor was employed to demonstrate the process flow of the continuous production of PS. The recyclability and stability of the Triton-X-100-modified silica were excellent, as demonstrated by its use 30 times during continuous operation without any loss of the productivity. KEYWORDS: transphosphatidylation, phosphatidylserine, adsorption, aqueous-solid system, continuous production

■

work.15 The surface of silica was employed as an “artificial interphase” between PC and PLD. The nature of the surface of the silica was changed with PC attachment, allowing the creation of a hydrophobic microenvironment for minimal hydrolysis and maximal reaction efficiencies. However, the preparation of the carrier-attached PC was complicated and used many volatile organic compounds (VOCs). Meanwhile, a portion (∼10%) of the PC was wasted in the preparation of the silica-attached PC, which is unacceptable in industrial production. Therefore, an improvement for this system seems to be necessary, and the ideal system is one in which a high yield is achieved while the waste of PC and the use of VOCs are avoided. In this paper, a novel method for highly efficient biosynthesis of PS in purely aqueous solutions without environmental pollution has been demonstrated. Triton X-100 was covalently bound to the surface of silica and acted as an anchor molecule to facilitate the adsorption of phosphatidylcholine (PC) in purely aqueous solutions. We systematically studied the effect of different contents of Triton X-100 on modified silica. Two adsorption models were studied, and the relevant parameters were calculated to help us understand the adsorption and reaction processes deeply. In addition, a packed-bed reactor was employed to demonstrate the process flow of the continuous production of PS.

INTRODUCTION Phosphatidylserine (PS), a kind of naturally rare phospholipid (PL), has many applications in functional food and pharmaceutical industries because of its physicochemical characteristics.1,2 Clinical studies have indicated that PS can revitalize brain cell membranes to improve memory performance in old people with memory impairments and children with cognitive and mood problems.2−4 So far, commercial sources of PS come from plant sources. Its low availability and the sophisticated extraction technology required result in its high price. Theoretically, the production cost of PS can be decreased significantly through the phospholipase D (PLD)-mediated transphosphatidylation of phosphatidylcholine (PC) with L-serine.5,6 The widely used system for transphosphatidylation is the liquid−liquid system, consisting of a water-immiscible organic phase and an aqueous phase. 7−9 The serious drawbacks of this system are unsatisfactory yields, serious hydrolysis, and complicated separation operations. Moreover, the use of toxic organic solvents not only results in enzyme denaturation but also entails the risk of environmental pollution. With regard to PS as a food or medicine for human use, the use of toxic organic solvents should be avoided. Next, the biosynthesis of PS in less toxic solvents (e.g., ionic liquids, γ-valerolactone, and deep eutectic solvents) has appeared.10−12 The complicated process of separation and the expensive prices of these solvents are limiting factors in their industrial applications. Transphosphatidylation has also been carried out in aqueous solutions using PC aggregates.13,14 The “effective reaction interface” is very low, thus decreasing the reaction rate and increasing the consumption of enzymes. In this field, a highly efficient and environmentally friendly aqueous−solid system has been introduced in our recent © XXXX American Chemical Society

Received: Revised: Accepted: Published: A

October 12, 2017 November 22, 2017 November 24, 2017 November 24, 2017 DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Figure 1. Route for the preparation of the Triton-X-100-modified silica.

■

Determination of a Model for the Adsorption of PC on the Triton-X-100-Modified Silica. Triton-X-100-modified silica (5.34 × 10−9 molTriton/gsilica) was used to determine the adsorption model. The adsorption process was investigated by two models (the one-step and two-step models).16 In the one-step model, the surfactant directly interacts with the active site to form a solloid (hemi-micelle) as below:

MATERIALS AND METHODS

Chemicals and Enzymes. Phospholipase D (CAS Registry No. 9001-87-0, from a Streptomyces sp.), phosphatidylcholine (PC), phosphatidylserine (PS), and phosphatidic acid (PA) were purchased from Sigma-Aldrich Company (St. Louis, MO). PLD was diluted and stored in 0.2 M acetate buffer (3.30 × 10−4 gprotein/mL, pH 5.5) at 4 °C. Silica gel G (HG/T2354) was obtained from Qingdao Haiyang Chemical Company, Ltd. (China). More details with respect to the silica used can be found in the Supporting Information (Table S1 and Figure S1). Adsorption of PC on Surfaces of Triton-X-100-Modified Silica. The route for the preparation of Triton-X-100-modified silica is shown in Figure 1. The experimental procedure can be found in the Supporting Information. The contents of Triton X-100 on the five kinds of Triton-X-100-modified silica were 5.34 × 10−11, 5.34 × 10−10, 5.34 × 10−9, 5.10 × 10−8, and 4.99 × 10−7 molTriton/gsilica. One milligram of PC was simply dispersed in 2 mL of 0.2 M acetate buffer (pH 5.5) in an ultrasonic bath for 5 min, and the solution was then stirred (500 rpm) for 4 h. Two milligrams of silica (either the purchased silica or the Triton-X-100-modified silica) was added to the solution. The mixture was incubated for 70 min at 30 °C while being stirred at 500 rpm. Fifty microliters of the reaction mixture was taken at specified time intervals and separated by centrifugation (69g, 1 min, 30 °C). Samples (20 μL) taken from the upper liquid were mixed with 20 μL of 0.08 M sodium hydroxide to react the acetic acid, then the mixture was dried under a nitrogen stream and dissolved in a 100 μL mixture of diethyl ether/distilled water (2:1, v/v). Samples (10 μL) were taken from the upper liquid to analyze the concentration of PC by high-performance liquid chromatography (HPLC), as described previously.15 Due to its poor solubility in water, PC aggregates may mix with silica after centrifugation, resulting in an error in the calculation of the amount of PC adsorbed. The process of the pretreatment of PC is done to reduce the sizes of the PC aggregates. The centrifugal force during the process of separation is another critical factor. A preliminary experiment was carried out to study the effect of centrifugal force on the efficiency of separation (data not shown). Finally, the conditions of centrifugation were 69g of centrifugal force, 1 min of operational time, and 30 °C. General Procedure for the Production of PS. Silica-adsorbed PC was prepared as described above using 1 mg of PC and 2 mg of Triton-X-100-modified silica (5.34 × 10−10, 5.34 × 10−9, or 5.10 × 10−8 molTriton/gsilica), incubated at 30 °C for 70 min with stirring at 500 rpm. Then, 200 mg of L-serine were dissolved in the above mixture. Fifty microliters of the PLD solution were added and allowed to react for 60 min at 30 °C with stirring at 500 rpm. Samples (20 μL) were taken from the reaction mixture at specified time intervals and intermittently mixed with 10 μL of 1 N HCl to inactivate the enzyme. The yield of PS was analyzed as described previously.15

site + (n)monomer = hemi‐micelle The general isotherm equation of the one-step model is

Γ = kC n Γ∞ − Γ

(1)

where Γ is the adsorption density, Γ∞ is the maximum adsorption density, k is the equilibrium constant, C is the equilibrium concentration, and n is the aggregation number. We take the logarithm of eq 1 and obtain a linear equation:

log

Γ = n log C + log k Γ∞ − Γ

(2)

In the two-step model, the surfactant monomer first adsorbs on the solid surface at concentrations below the hemi-micelle-aggregation concentration, and then the adsorbed surfactant monomers act as anchors for the formation of hemi-micelles:

site + monomer = adsorbed monomer (n − 1)monomers + adsorbed monomer = hemi‐micelle The general isotherm equation of the two-step model is Γ=

( 1n + k 2Cn−1)

Γ∞k1C

1 + k1C(1 + k 2C n − 1)

(3)

where k1 and k2 are the equilibrium constants for the first and second reactions, respectively. Then, the standard free energies for surface aggregation (ΔG0hm) in two models are calculated using the equations 0 ΔG hm = − (1/n)RT ln k

(4)

0 ΔG hm = − (1/n)RT ln k 2

(5)

The standard free energy for surface aggregation can be compared with that for micellization: 0 ΔGmic = − RT ln(cmc)

(6)

where cmc is the critical micelle concentration. The hemi-micelle concentration (hmc) is defined as the concentration above which adsorption increases dramatically as B

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry hemi-micelles form on the solid/liquid interface. For S-type isotherms, the results can be obtained from17 hmc =

⎛ n − 1 ⎞(n + 1)/ n −1/ n ⎜ ⎟ ·K ⎝ n + 1⎠

⎧ 3(n − 1) − [3(n2 − 1)]1/2 1 ⎫1/ n hmc = ⎨ · ⎬ ⎩ 3(n − 1) + [3(n2 − 1)]1/2 K ⎭

(7)

(8)

where K = k or K = k1 · k2. Operational Stability of the Free-PLD Solution. The operational stability of the free-PLD solution was studied. The reaction time was 20 min. After each batch reaction, the free-PLD solution was collected and used for the next batch. In each batch reaction, 1 mg of PC and 2 mg of Triton-X-100-modified silica (5.24 × 10−9 molTriton/ gsilica) were used. Recycling of the Triton-X-100-Modified Silica. Recycling of the Triton-X-100-modified silica (5.24 × 10−9 molTriton/gsilica) was investigated. After each batch reaction, Triton-X-100-modified silica was collected and washed with distilled water and the eluent. The renewable Triton-X-100-modified silica was used for the next batch. Continuous Synthesis of PS. The process flow for the continuous production of PS in a packed-bed reactor was shown in Figure 2. The packed-bed reactor was a glass tube (25 mm inside

Figure 3. Effect of carriers on the adsorption of PC: purchased silica (black), Triton-X-100-modified silica with 5.34 × 10−11 (red), 5.34 × 10−10 (blue), 5.34 × 10−9 (violet), 5.10 × 10−8 (green), and 4.99 × 10−7 (yellow) molTriton/gsilica.

Figure 4. Comparison of the adsorption capacities of PC on the purchased and Triton-X-100-modified silica.

Figure 5. Scheme of the adsorption of PC on the silica surface with three surface-content levels of Triton X-100.

Figure 2. Scheme of the process flow for the continuous production of PS in a packed-bed reactor.

3. Although PC showed only trace adsorption on the purchased silica, adsorption was facilitated dramatically by the use of the Triton-X-100-modified silica. It was worth noting that the maximum value of the adsorption density (Γ) was controlled at 2.65 × 10−6 mol/m2 (0.5 gPC/gsilica), which was confirmed as the optimal value for this system in our previous study.15 According to the Somasundaran−Fuerstenau theory, the adsorption process can be characterized by four regions.18−20 In region I, the driving force for adsorption is controlled by electrostatic interactions. In region II, the driving force transforms from electrostatic interactions into lateral interactions between hydrocarbon chains, which is accompanied by a sharp increase in adsorption density. In region III, the rate of adsorption begins to reduce because of the increase in steric interferences and the disappearance of electrostatic attractions.

diameter) containing sand core G3. Triton-X-100-modified silica (0.1 g, 5.24 × 10−9 molTriton/gsilica) and 1.0 g of glass beads (1 μm diameter) were packed in the reactor using the wet method. Fifty milliliters of a 1 mg/mL PC aqueous solution (0.2 M acetate buffer, pH 5.5) was pumped into the reactor and cycled 10 times. Then, 20 mL of the 1.15 × 10−4 gprotein/mL PLD solution (0.2 M acetate buffer, pH 5.5) containing 4 g of L-serine was cycled 10 times at room temperature. After the reaction, the packed-bed reactor was washed, and the eluent was pumped to elute PS.

■

RESULTS AND DISCUSSION Adsorption of PC on Surfaces of Triton-X-100Modified Silica. Six kinds of silica were compared to examine their performances in the adsorption of PC, as shown in Figure C

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Figure 6. Adsorption of PC on silica: left, PC aqueous solution without adsorption; middle, purchased silica; right, Triton-X-100modified silica (5.34 × 10−9 molTriton/gsilica).

Figure 8. (a) Adsorption isotherms of PC on the Triton-X-100modified silica (5.34 × 10−9 molTriton/gsilica) at 30 °C, plotted according to eq 2. (b) Adsorption isotherms of PC on the Triton-X100-modified silica. The black solid line represents the fitting of the one-step model to the experimental data. The blue solid line represents the value of the derivative of the adsorption density (Γ) versus the equilibrium concentration (C). The dashed line representing a Γ of 2.65 × 10−6 mol/m2 is provided to guide the eye only.

Figure 7. Transphosphatidylation using Triton-X-100-modified-silicaadsorbed PC.

In region IV, the adsorption density reaches the saturation adsorption density. The electrostatic interaction between PC and the purchased silica (silanol groups) is too weak to have the desired adsorption density of PC, which is expected to begin in region II.20 The adsorption of PC on the Triton-X-100-modified silica can be also characterized by four regions. Compared with the four regions of the Somasundaran−Fuerstenau isotherm described above, the essential difference is in region I. As shown in Figure 4, the covalently bound Triton X-100 acted as an anchor molecule for PC in the synergistic adsorption, and the driving force was due to lateral interactions between hydrocarbon chains. Compared with the electrostatic attraction, the lateral association made the adsorption process rapidly reach a specified adsorption density in region I and smoothly shift to region II. At the onset of region II, PC began to form surface aggregations, and a sharp increase in adsorption density was obtained. The decrease in the electrostatic repulsion between the polar headgroups of PC by the nonionic surfactant (Triton X-100) was another factor in the synergistic adsorption.19,21 A similar phenomenon can be found in the adsorption of surfactant mixtures. One surfactant has little effect on the force and adsorbs very little by itself, whereas a

trace of another surfactant coadsorbed on the carrier surface will promote adsorption.21−24 On the other hand, it was observed that for contents of Triton X-100 below 5.34 × 10−9 molTriton/gsilica, the adsorption capacity of the Triton-X-100-modified silica increased with the content of Triton X-100, whereas for contents above this value, decreases in the adsorption capacity were observed. The covalently bound Triton X-100 molecule, as an initiator, played a crucial role in the adsorption process. The lateral hydrophobic interaction between PC molecules in region II was triggered by PC molecules that were adsorbed early by the Triton X-100 molecules. Therefore, the adsorption capability of the Triton-X100-modified silica is strongly dependent on its content of Triton X-100. Figure 5 is a scheme of the adsorption of PC on the surface of Triton-X-100-modified silica with three content levels of Triton X-100. The “effective adsorption surface” was lower when the surface content of Triton X-100 was lower. Therefore, there were fewer active sites on the surfaces of the Triton-X-100-modified silica, which can improve the affinity D

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Figure 11. Investigation of the reuse of the free-PLD solution: 20 min (gray) and 3 h (blue) of reaction time in each run.

Figure 9. Adsorption isotherms of PC on the Triton-X-100-modified silica (5.34 × 10−9 molTriton/gsilica) at 30 °C. The black solid line represents the fitting of the two-step model to the experimental data. The blue solid line represents the value of the derivative of the adsorption density (Γ) versus the equilibrium concentration (C). The dashed line representing a Γ of 2.65 × 10−6 mol/m2 is provided to guide the eye only.

Table 1. Values for the Adsorption of PC on Triton-X-100Modified Silica from an Aqueous Solution at 30 °C adsorbate PC temperature (K) R2 a cmc (10−4 mol/L) hmc (10−6 mol/L) experimental eq 7 eq 8 n k k1 k2 ΔG0mic (kJ/mol) ΔG0hm (kJ/mol)

one-step model

two-step model

303 0.994

0.994 4.36 2.76

2.96 3.49 3.06 5.25 × 1015  

3.64 4.28 3.65  1.88 × 104 2.78 × 1014

Figure 12. Recycling of the Triton-X-100-modified silica.

19.5 −29.8

−23.0

a

The value of the cmc was determined in a preliminary experiment (Figure S3 in the Supporting Information).

Figure 10. Local concentration of PC on the surface of purchased and Triton-X-100-modified silica. Figure 13. Effect of the number of cycles on PC loading. After each run, the residual concentration was measured to calculate the PC loading.

between PC and silica in region I. At high content levels of Triton X-100, the effect of steric hindrance was not negligible. Although a higher effective adsorption surface was provided, the unoccupied surface of silica was reduced. This highly crowded environment did not permit compact packing of PC molecules, and the possibility of multilayer adsorption on the local surface increased with consequent decreases in the

adsorption rate and the adsorption density.19 In addition, a very high content of Triton X-100 would change the polarity of the surface of the silica from hydrophilicity to hydrophobicity; E

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

high density of the silica used (3.7 g/cm3), it also provides an alternative way to collect these silica-adsorbed phospholipids. After adsorption, the mixture was left standing for a few minutes to separate the silica from the aqueous solution. As shown in Figure 6, the aqueous solution (right) that used Triton-X-100-modified silica (5.24 × 10−9 molTriton/gsilica) as the adsorbent was transparent, and the edges of the magnetic rotor were very clear. The concentration of PC in this aqueous solution was analyzed, indicating that most PC molecules had been adsorbed. On the other hand, the aqueous solution (middle) that used purchased silica as the adsorbent was turbid. The PC concentration was the same as that of the initial aqueous solution of PC (left). Therefore, the turbidity of the aqueous solution can be taken as a standard of naked-eye detection to roughly estimate the end of adsorption. In addition, the adsorption of PC was studied without the pretreatment of the PC. One milligram of PC was directly mixed with 2 mg of Triton-X-100-modified silica (5.24 × 10−9 molTriton/gsilica) and incubated under stirring (500 rpm) for 12 h. A PC loading of 99.0% was obtained, indicating that the pretreatment can be simplified in the practical application. The chemical stability of the Triton-X-100-modified silica was also checked, and the results demonstrated that no “leakage” of Triton X-100 into the solution had taken place. This work is also the first description of the efficient adsorption of PC on silica in an aqueous solution. Transphosphatidylation Using Triton-X-100-Modified-Silica-Adsorbed PC. Figure 7 shows the time courses of the conversion of PC into PS. The results indicated that high yields of PS were obtained in all cases. Although our previous study found that the properties of the carrier surface had an effect on the yield of PS,15 the properties of the surface of the Triton-X-100-modified silica were controlled by the properties of PC because of the relatively higher content of PC (6.54 × 10−4 molPC/gsilica) than of Triton X-100 (5.34 × 10−10, 5.34 × 10−9, or 5.10 × 10−8 molTriton/gsilica) after the surface was occupied by PC molecules. It was also observed that yields of PA were always less than 1%. A reasonable explanation was that after being occupied by PC molecules, the surface of Triton-X100-modified silica, which acted as an “artificial interface”, allowed the creation of a hydrophobic microenvironment between PC and PLD and thus minimized hydrolysis. This also explained why the production of PS was accompanied by serious hydrolysis in aqueous-suspension systems.13,15 Many PC aggregates were suspended in the water, and the surface of calcium sulfate did not act as a reaction interface. Interestingly, there is a significant difference in the reaction rates of transphosphatidylations that use different carriers. This may be explained by the content of Triton X-100 on the surface of the silica being correlated with the structures of the PC surface aggregations.20,24,25 When the content of Triton X-100 was at relatively low levels, the effective adsorption surface was lower; thus, the local concentration of PC around each Triton X-100 molecule was higher at the same amount of PC adsorbed, increasing the possibility of multilayer adsorption on the local surface. At high contents, Triton X-100 occupied a large portion of the surface of the silica and increased the steric hindrance in the compact packing of PC.20,26,27 Determination of a Model for the Adsorption of PC on the Triton-X-100-Modified Silica. After Triton-X-100modified silica with a content of 5.34 × 10−9 molTriton/gsilica was confirmed to be the most suitable candidate for transphosphatidylation, the models of PC adsorption were system-

Figure 14. Operational stability of Triton-X-100-modified silica used 30 times during the continuous production of PS.

Figure 15. Investigation of the reuse of the free-PLD solution in the continuous production of PS.

Table 2. Comparison between Two Methods of PS Production

PC loading yield of PS use of VOCs recycling of the adsorption system recycling of the freePLD solution continuous biosynthsis of PS a

reported aqueous−solid systema

aqueous−solid system using Triton-X-100-modified silica

90.5 99.2 yes no

98.9 99.0 no yes

yes

yes

no

yes

Data can be found in ref 15.

it might reduce the dispersity of the silica in the aqueous solution and increase the repulsion between the polar headgroups of PC and the surface of silica. Although the process of centrifugation can be easily operated in the laboratory, industrial centrifuges are very expensive and require the consumption of energy. Sedimentation is the most popular method for separation of the product because of its low cost, lack of pollution, and simple operation. Because of the F

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Operational Stability of the Free-PLD Solution. The operational stability of the free-PLD solution in this system has been investigated, and the results are shown in Figure 11. After being used for seven batches, the yield of PS remained 30.3%. Although part of the enzymes was adsorbed on the carriers, the amount of that after each run was minimal, ranging from 2 to 6%, and enzyme contamination of the product could be avoided through thorough washing. The main reason for the reduction of activity was the loss of the PLD protein due to the small amount of the reaction mixture. Fortunately, a remedy was provided to prevent the reduction of the yield through prolonging the reaction time (3 h). As shown in Figure 11 (blue bar), the yield of PS reached 85.0% in the seventh batch, highlighting the presumable cost-effectiveness of the enzyme. The accumulation of the byproduct choline, which caused an inhibition of transphosphatidylation, was another reason for the decline of enzyme activity in repeated reuses.28 Recycling of the Triton-X-100-Modified Silica. The recyclability of the Triton-X-100-modified silica was examined. As shown in Figure 12, the recyclability of the Triton-X-100modified silica was excellent, as demonstrated in the 20 reaction cycles performed during the recycling experiment, which did not result in any loss of productivity. The recycling and reuse of the Triton-X-100-modified silica can reduce pollution and the cost of raw materials. Continuous Operation for PS Production. From an industrial view, continuous production is a more efficient strategy to maximize throughput. A process flow for the continuous production of PS was systematically investigated. As shown in Figure 13, the PC loading reached 90.8% and slightly increased after 7 cycles. In this packed-bed reactor, the “effective surface” of the Triton-X-100-modified silica decreased slightly because of the tight packing. Meanwhile, mass-transfer resistance and dead space were other limiting factors for future improvement of PC loading. Fortunately, the waste of PC can be avoided by reusing the residual PC solution for the next batch or by increasing the amount of Triton-X-100-modified silica used. Next, the result of the transphosphatidylation indicated that a PS yield of 99.3% was achieved. In the reported continuous operation for phospholipid synthesis, the activity of immobilized PLD decreased significantly after it was used for four batches.29 Therefore, the reactor needed to be frequently opened to replace the enzymes. Compared with the immobilized enzymes, the stability of the Triton-X-100modified silica is excellent. No loss of productivity was observed after recycling it 30 times, as shown in Figure 14. Moreover, the operational stability of the free-PLD solution was also investigated (Figure 15), and the results indicated that 86.2% of the yield of PS was obtained after the solution was used for ten batches. These excellent results make this system a more promising candidate for the industrial production of PS. In conclusion, we have developed an environmentally friendly and highly efficient method for the production of PS. Compared with our recent work,15 the waste of PC and the use of VOCs can be completely avoided in the whole process of transphosphatidylation. Moreover, this method provides a promising way to continuously biosynthesize PS. A detailed comparison between two methods can be found in Table 2.

atically investigated to help us understand the adsorption and reaction processes deeply. In the one-step model, the result of a linear adjustment of eq 2 to the data of the experimental adsorption isotherm was given in Figure 8a. The relevant parameters were calculated, and the simulation of the adsorption isotherm was described by the curve (black, solid) in Figure 8b. In the two-step model, a nonlinear adjustment of the eq 3 to the data was shown in Figure 9, and the relevant parameters were shown in Table 1. It is worth pointing out that the determination coefficient, R2, has the same value of 0.994 for both models, indicating the agreement of these two models with the experimental results. The driving forces for PC adsorption can be characterized by two parts: (1) the synergistic adsorption by the hydrophobic interaction between the immobilized Triton X-100 molecules and the free PC molecules and (2) the hydrophobic interaction between the adsorbed PC molecules and the free PC molecules. From this viewpoint, the two-step model may be the better candidate for our system. However, the value of k1 (1.88 × 104) implied that the adsorption rate in region I was very fast. Unlike monosite adsorption (head/head Coulombic interactions, i.e., each adsorbed molecule occupies one site) provided by the silanol groups, each immobilized Triton X-100 molecule can adsorb more than one PC molecule in three-dimensional space through lateral interactions between tail/tail hydrophobic chains in region I. The theoretical mechanisms behind this observation are the same as hydrophobic interactions between the adsorbed and free PC molecules in region II, which results in an agreement of the one-step model with the experimental data. As shown in Figures 8b and 9, the blue solid line represents the value of the derivative of the adsorption density (Γ) versus the equilibrium concentration (C). When the value of the adsorption density was above 2.65 × 10−6 mol/m2 (dashed line), an obvious decrease in the growth rate of the adsorption density was observed, implying the value of the monolayer adsorption density. Interestingly, the maximum value of the adsorption density (5.85 × 10−6 mol/m2) was approximately two times that of the monolayer adsorption (2.65 × 10−6 mol/ m2). If a bilayer structure formed, the repulsion forces derived from the head/head Coulombic interactions would hinder further adsorption (Figure S2 in the Supporting Information). Then the values of the standard free energy for surface aggregation (ΔG0hm) based on two models were calculated and compared with that for micellization. The results showed that ΔG0hm was a negative value and always less than ΔG0mic; thus, the surface aggregation was spontaneously formed and energetically favored.20 A sharp decrease was observed in the value of the hmc, which was ∼150 times less than that of the cmc. Even if PC adsorbed on the purchased silica, PC would form the monosite adsorption in region I, as shown in Figure 10 (left). Until these available monosites were completely occupied, the PC concentration around each adsorbed PC would begin to increase, and there would be the possibility of the formation of surface aggregations. For Triton-X-100-modified silica, each Triton X-100 molecule can adsorb more than one PC molecule in three-dimensional space through lateral interactions in region I, as shown in Figure 10 (right). The content of Triton X-100 was less than that of silanol groups on the carrier surfaces, and thus, the “local concentration” of PC around each Triton X-100 molecule was significantly higher than that around each silanol group, leading to a very low hmc value.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jafc.7b04744. G

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

■

(11) Yang, S.; Duan, Z. Insight into enzymatic synthesis of phosphatidylserine in deep eutectic solvents. Catal. Commun. 2016, 82, 16−19. (12) Bi, Y.; Duan, Z.; Li, X.; Wang, Z.; Zhao, X. Introducing biobased ionic liquids as the nonaqueous media for enzymatic synthesis of phosphatidylserine. J. Agric. Food Chem. 2015, 63, 1558−1561. (13) Iwasaki, Y.; Mizumoto, Y.; Okada, T.; Yamamoto, T.; Tsutsumi, K.; Yamane, T. An aqueous suspension system for phospholipase Dmediated synthesis of PS without toxic organic solvent. J. Am. Oil Chem. Soc. 2003, 80, 653−657. (14) Dittrich, N.; Ulbrich-Hofmann, R. Transphosphatidylation by immobilized phospholipase D in aqueous media. Biotechnol. Appl. Biochem. 2001, 34, 189−194. (15) Li, B.; Wang, J.; Zhang, X.; Zhao, B.; Niu, L. Aqueous−Solid System for Highly Efficient and Environmentally Friendly Transphosphatidylation Catalyzed by Phospholipase D To Produce Phosphatidylserine. J. Agric. Food Chem. 2016, 64, 7555−7560. (16) Gu, T.; Zhu, B. Y. The S-type isotherm equation for adsorption of nonionic surfactants at the silica gel-water interface. Colloids Surf. 1990, 44, 81−87. (17) Zhu, B.-Y.; Gu, T. Surfactant adsorption at solid-liquid interfaces. Adv. Colloid Interface Sci. 1991, 37, 1−32. (18) Somasundaran, P.; Healy, T. W.; Fuerstenau, D. W. The aggregation of colloidal alumina dispersions by adsorbed surfactant ions. J. Colloid Interface Sci. 1966, 22, 599−605. (19) Koopal, L. K.; Lee, E. M.; Böhmer, M. R. Adsorption of Cationic and Anionic Surfactants on Charged Metal Oxide Surfaces. J. Colloid Interface Sci. 1995, 170, 85−97. (20) Zhang, R.; Somasundaran, P. Advances in adsorption of surfactants and their mixtures at solid/solution interfaces. Adv. Colloid Interface Sci. 2006, 123−126, 213−229. (21) Somasundaran, P.; Fu, E.; Xu, Q. Coadsorption of anionic and nonionic surfactant mixtures at the alumina-water interface. Langmuir 1992, 8, 1065−1069. (22) Zhang, L.; Somasundaran, P. Adsorption of mixtures of nonionic sugar-based surfactants with other surfactants at solid/liquid interfaces. J. Colloid Interface Sci. 2006, 302, 20−24. (23) Lokar, W. J.; Ducker, W. A. Forces between glass surfaces in mixed cationic-zwitterionic surfactant systems. Langmuir 2004, 20, 4553−4558. (24) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. Theory of selfassembly of hydrocarbon amphiphiles into micelles and bilayers. J. Chem. Soc., Faraday Trans. 2 1976, 72, 1525−1568. (25) Zhang, R.; Somasundaran, P. Aggregate formation of binary nonionic surfactant mixtures on hydrophilic surfaces. Langmuir 2005, 21, 4868−4873. (26) Matsson, M. K.; Kronberg, B.; Claesson, P. M. Adsorption of alkyl polyglucosides on the solid/water interface: Equilibrium effects of alkyl chain length and head group polymerization. Langmuir 2004, 20, 4051−4058. (27) Alami, E.-O.; Holmberg, K. Hetergemini surfactants. Adv. Colloid Interface Sci. 2003, 100-102, 13−46. (28) Juneja, L. R.; Taniguchi, E.; Shimizu, S.; Yamane, T. Increasing productivity by removing choline in conversion of phosphatidylcholine to phosphatidylserine by phospholipase D. J. Ferment. Bioeng. 1992, 73, 357−361. (29) Juneja, L. R.; Hibi, N.; Yamane, T.; Shimizu, S. Repeated batch and continuous operations for phosphatidylglycerol synthesis from phosphatidylcholine with immobilized phospholipase D. Appl. Microbiol. Biotechnol. 1987, 27, 146−151.

Information on the silica gel used: the SEM micrograph, the specific surface area, and the average pore diameter; scheme of the adsorption of PC on the Triton-X-100modified-silica surface with the saturation adsorption density; determination of the value of the cmc of PC using the surface tension measurement; preparation of Triton-X-100-modified silica. (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Xiaoli Zhang: 0000-0003-3702-3700 Author Contributions ‡

X.Z. and B.L. contributed equally to this work.

Funding

This work was supported by Northwest University Doctorate Dissertation of Excellence Funds (Program No. YYB17014). Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Sakai, M.; Yamatoya, H.; Kudo, S. Pharmacological effects of phosphatidylserine enzymatically synthesized from soybean lecithin on brain functions in rodents. J. Nutr. Sci. Vitaminol. 1996, 42, 47−54. (2) Kato-Kataoka, A.; Sakai, M.; Ebina, R.; Nonaka, C.; Asano, T.; Miyamori, T. Soybean-derived phosphatidylserine improves memory function of the elderly Japanese subjects with memory complaints. J. Clin. Biochem. Nutr. 2010, 47, 246−255. (3) Claro, F. T.; Patti, C. L.; Abílio, V. C.; Frussa-Filho, R.; Silva, R. H. Bovine brain phosphatidylserine attenuates scopolamine induced amnesia in mice. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 2006, 30, 881−886. (4) Hirayama, S.; Terasawa, K.; Rabeler, R.; Hirayama, T.; Inoue, T.; Tatsumi, Y.; Purpura, M.; Jäger, R. The effect of phosphatidylserine administration on memory and symptoms of attention-deficit hyperactivity disorder: a randomised, double-blind, placebo-controlled clinical trial. J. Hum. Nutr. Diet. 2014, 27, 284−291. (5) Ogino, C.; Negi, Y.; Matsumiya, T.; Nakaoka, K.; Kondo, A.; Kuroda, S.; Tokuyama, S.; Kikkawa, U.; Yamane, T.; Fukuda, H. Purification, characterization, and sequence determination of phospholipase D secreted by Streptoverticillium cinnamoneum. J. Biochem. 1999, 125, 263−269. (6) Hatanaka, T.; Negishi, T.; Kubota-Akizawa, M.; Hagishita, T. Purification, characterization, cloning and sequencing of phospholipase D from Streptomyces septatus TH-2. Enzyme Microb. Technol. 2002, 31, 233−241. (7) Zhang, Y.-N.; Lu, F.-P.; Chen, G.-Q.; Li, Y.; Wang, J.-L. Expression, purification, and characterization of phosphatidylserine synthase from Escherichia coli K12 in Bacillus subtilis. J. Agric. Food Chem. 2009, 57, 122−126. (8) Hosokawa, M.; Shimatani, T.; Kanada, T.; Inoue, Y.; Takahashi, K. Conversion to docosahexaenoic acid-containing phosphatidylserine from squid skin lecithin by phospholipase D-mediated transphosphatidylation. J. Agric. Food Chem. 2000, 48, 4550−4554. (9) Mao, X.; Liu, Q.; Qiu, Y.; Fan, X.; Han, Q.; Liu, Y.; Zhang, L.; Xue, C. Identification of a novel phospholipase D with high transphosphatidylation activity and its application in synthesis of phosphatidylserine and DHA-phosphatidylserine. J. Biotechnol. 2017, 249, 51−58. (10) Duan, Z.-Q.; Hu, F. Highly efficient synthesis of phosphatidylserine in the eco-friendly solvent γ-valerolactone. Green Chem. 2012, 14, 1581. H

DOI: 10.1021/acs.jafc.7b04744 J. Agric. Food Chem. XXXX, XXX, XXX−XXX