Partitioning behavior of recombinant lipase in Escherichia coli by ionic liquid-based aqueous two-phase systems

Abstract

There has been significant interest in ionic liquid aqueous two-phase systems (ILATPSs) with properties such as rapid phase segregation which can lead to a reduction in time taken for protein recovery. Evaluations of ILATPSs were performed with various types of ionic liquid ((Emim)BF₄, (Emim)Br, (Bmim)BF₄, and (Bmim)Br) and salts (potassium-, sodium-, and magnesium-based) as phase components to figure out their competencies in the recovery of lipase from a fermentation broth of E. coli using banana waste as a substrate. The results of this study revealed that an ILATPS comprising (Emim)Br/potassium phosphate significantly enhanced lipase recovery upon partitioning lipase. Optimization of the composition of the ILATPS using response surface methodology (RSM) significantly improved the purification factor (P_F) and partition coefficient (K_e). The influences of crude loading (C_L), pH changes, and the presence of NaCl on the recovery performance were also studied. Recovery of E. coli BL21 lipase was accomplished in an (Emim)Br/potassium phosphate ILATPS using 26.5% (w/w) (Emim)Br, 19% (w/w) potassium phosphate at pH 7.6, 3% NaCl and a crude loading of 7% (w/w). Using this ILATPS, lipase was successfully recovered in a single purification step, which gave a yield, P_F and K_e of 93.75%, 3.394 and 1.352, respectively. A high P_F value indicates that (Emim)Br/potassium phosphate is capable of attaining an excellent degree of lipase purity, suggesting that the proposed ILATPS is suitable for implementation in a large scale process for lipase purification.

---

Introduction

Lipases (triacylglycerile hydrolases; EC 3.1.1.3) are industrially important enzymes due to their regio-, stereo- and chemo-selective reactions and kinetic resolution of racemic mixtures.¹ Of late, lipases are garnering widespread interest because of their great potential in various industries mainly due to their availability and stability.² Lipases are used in the lipid technology industry, and have ex situ multifaceted applications in the food, detergent and pharmaceutical industries. Lipase is also used in bioenergy and biodiesel production. The unique features of lipases have led to intensive studies about the suitable and excellent techniques of recovery and purification which provide better degrees of specificity and yield.¹

In general, about 80% of the manufacturing cost of protein is contributed to by the complexity of the recovery and purification of the target protein from the crude extract.³ Numerous purification strategies have been reported for lipase recovery with varying degrees of success. The common purification methods which include precipitation along with chromatographic techniques such as gel filtration, ion exchange and affinity chromatography are often associated with low yields and are time intensive, expensive and require sophisticated operations. Biomanufacturing industries prefer purification processes that are less capital intensive, high yielding, rapid and robust with ease of up-scaling the procedure for maximal profit.⁴

In recent years, there has been a growing interest in ionic liquid-based ATPSs (ILATPS) due to several of their advantages exhibited as a novel separation strategy. ILATPSs, a new type of green system, have shown many advantages in separation and purification such as being easy to scaled up, rapid mass transfer and balance, high extraction efficiency, low viscosity, negligible emulsion formation, not requiring the use of volatile organic solvents, and a gentle biocompatible environment. Therefore, these novel ATPSs have been successfully used to separate proteins, amino acids, alkaloids, and saccharides among others.⁵

In view of the fact that the use of an ionic liquid based aqueous two-phase system (ILATPS) is an ideal purification method for the separation, extraction and concentration of biomolecules, this study evaluated the partitioning efficiency of a recombinant lipase from E. coli based on an aqueous two phase system which was composed of an ionic liquid and a salt as a recovery strategy for microbial lipase. Response surface methodology (RSM) was used for optimization of the IL’s components and maximizing the response of interest (e.g. purification factor and yield). RSM is a great analytical tool for the determination of the optimum composition of the IL and salt which could be used to minimize the operating cost of ILATPSs for lipase purification with an improvement in the purification factor (P_F) and yield.

The purpose of this study was to construct an easy to operate and sustainable green downstream method based on an aqueous two phase system which was composed of an ionic liquid and salt as a recovery strategy for microbial lipase and also to optimize the IL’s components. In this context, several extraction parameters of the ILATPS, namely the influence of the phase components, phase composition of the ILATPS, crude loading, pH and the presence of salt were evaluated for their improvement in the recovery of lipase from E. coli BL21.

Materials and methods

Experimental

Materials. 1-Ethyl-3-methylimidazolium tetrafluoroborate (EmimBF₄, ≥97.0%), 1-ethyl-3-methylimidazodium bromide (EmimBr, ≥97.0%), 1-butyl-3-methylimidazolium tetrafluoroborate (BmimBF₄, ≥97.0%), and 1-butyl-3-methylimidazolium bromide (BmimBr, ≥97.0%) were obtained from Sigma Aldrich, USA. The protein assay kit and albumin standard were supplied by Bio-Rad, USA and Thermo Scientific Pierce, respectively. Luria Bertani (LB) broth, glucose and sodium chloride (NaCl) were sourced from Merck (Darmstadt, Germany). Sodium citrate (C₆H₅Na₃O₇·2H₂O) and potassium phosphate were purchased from SAFC (St. Louis, MO, USA). Isopropylthio-β-galactoside (IPTG) was acquired from Calbiochem (Billerica, MA, USA). All chemicals used in this study were of analytical grade except the ionic liquids which were of chemical grade.

Microorganism and fermentation. Recombinant E. coli BL21(DE3) pLysS transformed with pET 51b-lipase was used in this study. A stock culture of E. coli BL21 was cultured at 37 °C and kept in 10% (v/v) glycerol aseptically at −20 °C. Each inoculum was prepared by inoculating 2% (v/v) of stock culture into 50 mL of the LB broth supplemented with 50 mg L⁻¹ ampicillin (Calbiochem, Darmstadt, Germany) and 35 mg L⁻¹ chloramphenicol (A.G. Scientific, San Diego, California) in a 250 mL shake flask. The ampicillin acted as an agent to prevent the overgrowth of cells that have lost their plasmid and the chloramphenicol was added to avoid growth of foreign cells in the culture. The culture was grown at 37 °C in an incubator shaker (Infors HT, Switzerland) with constant shaking at 250 rpm for 16 h, prior to use as a standard inoculum for all fermentations.

Lipase production by recombinant E. coli. All fermentations were conducted under aerobic conditions in a 250 mL shake flask with 50 mL of the production medium consisting of (g L⁻¹): banana pseudostem juice, 200; NaCl, 3; tryptone, 10; and yeast extract, 25. The medium was inoculated with 4% (v/v) inoculum and incubated at 30 °C in an incubator shaker, agitated at 200 rpm. After 2 h of fermentation, IPTG was added to a final concentration of 0.5 mM, to initiate the expression of lipase and the culture was further incubated for 4 h prior to harvesting. The cultures were centrifuged at 4000 × g for 30 min at 4 °C (rotor model 1189, Universal 22R centrifuge, Hettich AG, Switzerland). The cell pellet was collected and resuspended in 0.5 mM potassium phosphate buffer at pH 7 and disrupted by glass beads (1.5 g per mL) at 250 rpm for 30 min in an incubator shaker at 20 °C. The disrupted cell suspension was centrifuged at 4000 × g for 30 min to get rid the cell debris and the supernatant that contained the crude lipase enzyme, which was then collected and kept at −20 °C prior to use in experiments.

ILATPSs

Preparation of phase diagrams for IL/Salt based aqueous two-phase systems. A binodal curve is the borderline between one and two-phases in an ATPS.⁶ This binodal curve was plotted based on a turbidimetric titration method.⁷ Salt solutions at 40 wt% and aqueous solutions of the different hydrophilic ILs at 100 wt% were prepared and used for the phase diagrams’ determination. Known concentrations of ionic liquids and salts (18 points) were added to form several ILATPSs, which corresponded to various total compositions. The phase system had a turbid appearance because of the immiscibility between the two phase components. Using the turbidimetric titration method, distilled water was added drop-wise into the turbid phase system until the appearance became clear. The diminished turbidity showed that the titration method had reached a critical point of the two phase components. The weight of distilled water added to reach the critical point was taken and the binodal nodes were plotted based on it. The final concentrations of the ILs and salts were calculated and binodal curves were constructed at different concentrations of ionic liquid and salt.

Partition experiments of the lipase enzyme. Aqueous two-phase extraction experiments were implemented by mixing appropriate amounts of ionic liquids ((Emim)BF₄, (Emim)Br, (Bmim)BF₄ and (Bmim)Br), salts, and 7% (w/w) crude lipase and distilled water. The centrifuge tube was then vortexed vigorously to ensure a homogeneous state, followed by centrifugation at 1000 × g for 1 min at 25 °C for faster separation. The top and bottom phases were separated for the determination of total protein content and lipase activity. Measurement of the volume ratios (V_r) of each ILATPS was performed. All of the ILATPs’ phase system experiments were conducted in triplicate.

Effect of IL and salt composition on ILATPS partitioning using response surface methodology. Response surface methodology (RSM) was employed to optimize the purification of lipase and to investigate the relative and interactive effects of the phase compositions of the ILs and salts. A five-level, full-factorial central composite design (CCD) with three independent variables, that is, the concentration of the ionic liquid (IL), salt and type of IL was used in this study. A total of 52 sets of experimental runs consisting of 16 factorial (cubic points), 16 axial (star points), and 20 replicates of center points was designed. The concentrations of the IL and salt were numerical variables while the type of IL was a categorical variable. The pH and crude loading were fixed at 7.0 and 3.75% (w/w) as constant variables, respectively. To comprehend the purification performance of the ILATPSs, two responses comprising the purification factor and recovery yield were initially calculated. The observed responses from the CCD design were then fitted to the following polynomial eqn (1):where Y is the predicted response; i and j are the index numbers for the pattern; β is the offset term; β_i, β_ii, and β_ij are the coefficients for the linear, quadratic, and interaction effects, respectively; x_i and x_j are the coded variables; and ε is the error. The regression equation was optimized by an iterative method to achieve the optimum values. Design Expert (version 7.1.6, Stat-Ease Inc., Minneapolis, MN, USA) was used for regression modeling and data interpretation. The optimum phase composition of IL (g) and salt (g) would be used for further parameter analysis such as crude loading, pH and NaCl.

Determination of partition coefficient, selectivity, purification fold and yield. Fractionation of lipase and total protein in the top and bottom phases was calculated by the partition coefficient of lipase (K_e) (eqn (2)) and the partition coefficient of protein (K_p) (eqn (3)):where A_T and A_B refer to lipase activity (U mL⁻¹) in the top and bottom phases and P_T and P_B are the total protein concentrations (mg mL⁻¹) in the top and bottom phases, respectively.

The specific activity (SA) of lipase, defined as the ratio between lipase activity (AU) and total protein concentration (mg), was calculated according to eqn (4):

The purification factor (P_F) of lipase, referred to as the ratio of lipase specific activity in the ionic liquid of the top phase (SA_T) to the specific activity of lipase in the crude extract (SA_C), was determined using eqn (5):

The yield (Y_T) of lipase was calculated using eqn (6):

Lipase activity assay. Lipase activity was determined according to method described by Gupta et al.,⁸ with some modifications. The reaction mixture contained 25 μL of the crude enzyme extract, 200 μL 0.05 M phosphate buffer (pH 6.5), 5 μL of Triton-X and 25 μL of 0.02 M p-nitrophenyl-laurate (p-NPL) in ethanol. The assay mixture was incubated at 37 °C for 15 min and the absorbance was measured at 405 nm using a microplate reader (Dynamica Halo MPR-96, Switzerland). The amount of p-nitrophenol (p-NP) released from p-NPL was measured and calculated using the standard curve of p-NP. One unit of lipase activity was defined as the amount of lipase liberated in 1 μmol of p-NP per min.

Bradford assay. The total concentration of protein was determined by the Bradford method using bovine serum albumin (BSA) as a standard.⁹ A total of 20 μL of sample was added to 200 μL of diluted reagent and incubated at room temperature for 15 min. The absorbance was measured at 595 nm.

Sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE). To determine the molecular weight of the lipase molecule, SDS-PAGE was performed according to the method of Abbasiliasi et al.¹⁰ with minor modifications. The SDS-PAGE analysis was carried out using polyacrylamide gel made up of a 12% (v/v) resolving gel and a stacking gel of 5% (v/v) in a Mini-Vertical SE250 electrophoresis unit (Amersham Biosciences, USA) at 80 V constant voltage for 10 min and was proceeded with 120 V for 60 minutes. Then, the gel was stained with a staining solution which was 0.05% (v/v) Coomassie Brilliant Blue G-250 overnight and destained with a destaining solution for about 15 min. After destaining, the images were subsequently viewed using Gel Image (model Bio-Rad Molecular Image Gel Doc™ XR⁺, Universal Hood, USA).

Results and discussion

Effect of the phase components on lipase activity

The effect of the salt phase components on the relative activity of lipase was evaluated in order to determine the type of salt that was suitable to serve as a phase for the construction of the ILATPSs (Table 1). The pH of all salts was fixed at pH 7. Inorganic salts of potassium and sodium based on charged anions such as phosphate, acetate and citrate showed a high relative activity of lipase at lower concentrations [10–20% (w/w)]. On the other hand, sodium nitrate and magnesium sulphate gave the lowest relative activity compared to the other salts. These salts were excluded from further experiments for use as phase components. The highest relative activity of lipase (262.10 U mL⁻¹) was recorded at 20% w/w of potassium phosphate, out of all the other potassium-based salts. For the sodium-based salts, the highest relative activity (247.91 U mL⁻¹) was obtained by sodium citrate at 20% (w/w). High concentrations of the salts (30% (w/w)), either potassium- or sodium-based decreased lipase’s relative activity significantly. Thus, low concentrations of salt were preferred for the recovery of lipase to reduce loss of lipase activity and protein precipitation. Inorganic salts usually have a higher affinity towards water molecules in ATPSs.^11,12 When the salt concentration is too high [i.e. 30% (w/w)], it will have a strong water affinity that will lead to over-migration and less interaction of the water molecules with lipase, which subsequently contributes to the protein precipitation phenomenon, which is known as salting-out.¹³

Table 1 Effect of salt phase and IL components on the activity of E. coli BL21 lipase. For each component, a constant percentage of crude enzyme at 20% (w/w) was used. The separation mixture was incubated for 30 min at room temperature prior to the determination of lipase activity

Salt phase components	Concentration, % (w/w)	Lipase activity (U mL^−1)
Potassium phosphate	10	259.66
	20	262.10
	30	227.42
Potassium citrate	10	199.52
	20	201.93
	30	144.66
Sodium citrate	10	246.10
	20	247.91
	30	241.76
Sodium acetate	10	164.57
	20	134.46
	30	117.74
Sodium nitrate	10	31.73
	20	35.98
	30	36.84
Magnesium sulphate	10	15.24
	20	14.60
	30	16.16

---

Ionic liquid phase components	Concentration, % (w/w)	Lipase activity (U mL^−1)
(Emim)BF4	20	261.79
	40	238.25
	60	222.22
(Bmim)BF4	20	266.44
	40	243.48
	60	216.30
(Emim)Br	20	288.22
	40	286.47
	60	286.07
(Bmim)Br	20	286.87
	40	282.19
	60	261.91

---

The type of IL chosen as the top phase component in the ATPS is able to influence the stabilization of lipase during the recovery process.¹⁴ In order to determine the influence of each IL on the E. coli lipase relative activity, crude lipase was loaded into IL solutions at various concentrations [i.e. 20–40% (w/w)]. The results are shown in Table 1. Lipase has greater stability and higher activity at lower concentrations of the ILs [20% (w/w) and 40% (w/w)]. However its relative activity was reduced at high concentrations of the ILs [60% (w/w)]. The lipase relative activity was slightly decreased at 60% (w/w) (Emim)Br, in which lipase stability was still maintained even at high concentrations of (Emim)Br. In the other ILs, lipase activity was decreased abruptly at high IL concentrations. The stability of lipase activity is greatly dependent on several factors, including hydrophobic interactions, electrostatic capacity, and the salting out factor due to differences in the concentration and type of IL.¹³ Higher concentrations of the IL leads to a high salting out effect which suppresses the relative activity of lipase. Thus, this can be curtailed by selecting ILs at a lower concentration range [20–40% (w/w)]. Research on ILATPSs for the recovery of biomaterials put forth that the IL was able to enhance protein stabilization and to promote high relative activity, but a greater concentration of the ILs suppressed protein stabilization due to the salting out factor that led to protein precipitation.^5,13,15–17

As a result, the concentrations of the ILs [20–40% (w/w)] and potassium phosphate [10–20% (w/w)] were selected for the low and high limits for determining the effect of the composition of the ILs and salt via response surface methodology (RSM) in the subsequent experiments.

Construction of phase diagrams of ILATPSs based on various ILs and salts

Studies on the precipitation of different ILs when reacted with different salts. This preliminary study is crucial in order to determine the phase formation when different ILs reacted with different salts prior to the construction of the phase diagrams for the ILATPSs. Surprisingly, these findings revealed that potassium-based salts such as potassium phosphate mixed with (Emim)BF₄ and (Bmim)BF₄ became solidified and precipitated throughout the determination of the phase diagram (Table 2). Thus, the construction of the phase diagram for these two types of ILs could only be done with sodium citrate as the bottom phase which has the second highest relative lipase activity (247.91 U mL⁻¹). The reason for the non-compatibility of the ILs [(Emim)BF₄ and (Bmim)BF₄] with the potassium phosphate lies in the Hofmeister series of inorganic salts or kosmotropic salts. Some kosmotropic salts are not able to form a biphasic layer with anion- or cation-based ILs according to the Hofmeister series. The presence of ions on the macromolecular configuration is able to dominate the bulk water molecules through fracturing and then structuring.¹⁸ In the Hofmeister series, potassium-based salts are weakly hydrated compared to sodium-based salts which draw away the water molecules from the anion-based IL phase. Hence, as a result the common [BF₄⁻] anion based ILs will become solid and precipitate.¹⁸ Instead, cation-based ILs [(Emim)Br and (Bmim)Br] were found to be compatible and able to form a biphasic phase with the potassium-based anion salt.¹⁸ As a result, sodium citrate was used as the bottom phase for the [BF₄] based ILs ((Emim)BF₄ and (Bmim)BF₄), while potassium phosphate was implemented for the [Br] based ILs ((Emim)Br and (Bmim)Br) during phase diagram construction for the ILATPSs.

Table 2 Effect of various ILs with different salts on system composition^a

System composition	Appearance
a [BF4]-based ILs: (Emim)BF4 and (Bmim)BF4; [Br]-based ILs: (Emim)Br and (Bmim)Br.
[BF4]-based ILs/sodium citrate	Phase formation
[Br]-based ILs/potassium phosphate	Phase formation
[Br]-based ILs/sodium citrate	Phase formation
[BF4]-based ILs/potassium phosphate	Precipitation occurs

---

Phase diagrams. The ILs are considered to be chaotropic salts (soluble in water) whereas inorganic salts are considered to be kosmotropic salts, in which the layer between the two phases will be formed when they are joined together due to the salting out effect.¹⁵ The extent of exclusion is directly proportional to the concentration of the ILs or salt. Eventually, phase formation in the ILATPSs would become favorable. It was observed that ILs formed the top phase while the salt formed the bottom phase. From the phase diagrams, some important information could be extracted including the concentration of the phase-forming components that is essential to create an ILATPS in equilibrium, V_r of the ILATPS and the concentration of the phase components in the system.¹¹

The binodal curve of (Bmim)BF₄/sodium citrate was shifted downwards and closer to its origin point, indicating greater phase segregation compared to (Emim)BF₄/sodium citrate (Fig. 1A). (Bmim)BF₄/sodium citrate was easily segregated into biphasic layers with a lower concentration of the IL and salts. Conversely, (Emim)BF₄/sodium citrate required a higher concentration of the IL and salt in order to reach two phase layers at the same point as the (Bmim)BF₄/sodium citrate.¹³

Fig. 1 Phase diagrams of the ILATPSs. Phase diagram of four ILATPSs were constructed at 37 °C and salt pH 7. The (A) sodium citrate based binodal curves of (Emim)BF₄ (♦) and (Bmim)BF₄ (■) ILATPSs and (B) potassium phosphate based binodal curves of (Emim)Br/potassium phosphate (♦) and (Bmim)Br/potassium phosphate (■) were tabulated.

The binodal curve of (Bmim)Br/potassium phosphate was slightly segregated compared to the binodal curve of (Emim)Br/potassium phosphate, indicating that these two types of ILs have an almost equivalent phase segregation ability to become biphasic layers (Fig. 1B). As the binodal curve of (Emim)Br/potassium was shifted slightly upwards compared to the binodal curve of (Bmim)Br/potassium phosphate. Thus a slight increase in the concentration of the ILs and salts was required in order to form a two phase system at the same point on the binodal curve as (Bmim)Br/potassium phosphate.¹¹

Optimization of IL and salt composition in ILATPS partitioning using the response surface methodology (RSM)

Table 3 displays the CCD design matrix of the phase composition, together with the actual and predicted responses, that is, the purification factor and partition coefficient. The maximum P_F (3.152) was found in (Emim)Br/potassium phosphate under experimental conditions of IL = 25% and salt = 17.5%. Results of this study also showed that (Bmim)BF₄/sodium citrate gave the lowest P_F compared to that obtained in (Emim)Br/potassium phosphate under the same experimental conditions which was considered to not be acceptable. Results from this study also revealed that the differences between the experimental and predicted values obtained for (Emim)Br/potassium phosphate were very small compared to the differences observed for the others IL. The experimental response from (Emim)Br/phosphate was used for the interpretation of the response surface model.

Table 3 (Bmim)BF₄ and (Emim)BF₄/sodium citrate and (Bmim)Br and (Bmim)Br/potassium phosphate ATPS composition performance based on RSM^a

No.		Phase composition		Partition coefficient, Ke		Purification factor, PF
		IL, % (w/w)	Salt, % (w/w)	Experimental	Predicted	Experimental	Predicted
a n.p. is no phase formation.
1	(Emim)BF4/sodium citrate	17.9	17.5	0.780	0.71	1.467	1.47
2		25.0	17.5	1.016	0.61	2.938	2.51
3		25.0	17.5	0.686	0.61	3.159	2.51
4		25.0	17.5	0.768	0.61	3.102	2.51
5		25.0	17.5	1.194	0.61	2.995	2.51
6		25.0	17.5	0.974	1.25	2.538	2.35
7		20	10	n.p.	n.p	n.p.	n.p.
8		25	28.1	0.823	0.90	0.770	1.26
9		20	25	0.160	0.61	0.338	2.51
10		30	25	0.446	0.38	1.088	0.37
11		32.1	17.5	0.096	0.50	0.295	1.05
12		30	10	0.927	0.69	1.934	1.36
13		25	6.9	n.p.	0.00	n.p.	n.p.
14	(Bmim)BF4/sodium citrate	25	17.5	0.167	0.16	0.422	0.45
15		25	17.5	0.080	0.16	0.643	0.45
16		25	17.5	0.060	0.16	0.586	0.45
17		25	17.5	0.345	0.16	0.477	0.45
18		25	17.5	0.125	0.16	0.120	0.45
19		30	10	2.640	1.97	2.663	2.00
20		30	25	0.660	0.21	1.814	1.14
21		20	10	1.300	1.31	2.589	2.85
22		25	6.9	1.738	2.11	2.612	2.81
23		17.9	17.5	0.889	0.63	2.240	1.80
24		25	28.1	0.189	0.26	0.558	0.77
25		20	25	0.218	0.44	0.569	0.82
26		32.1	17.5	0.231	0.93	0.575	1.43
27	(Emim)Br/potassium phosphate	17.9	17.5	n.p.	n.p	n.p.	n.p.
28		25.0	17.5	1.380	1.02	2.988	3.02
29		25.0	17.5	1.050	1.02	3.209	3.02
30		25.0	17.5	1.132	1.02	3.152	3.02
31		25.0	17.5	1.558	1.02	3.145	3.02
32		25.0	17.5	1.338	1.02	2.588	3.02
33		20	10	n.p.	n.p.	n.p.	n.p.
34		25	28.1	1.232	1.61	2.967	3.11
35		20	25	1.863	1.67	2.844	2.49
36		30	25	1.035	1.19	2.931	2.89
37		32.1	17.5	1.324	1.43	2.814	2.93
38		30	10	1.301	1.42	2.67	2.35
39		25	6.9	n.p.	n.p.	n.p.	n.p.
40	(Bmim)Br/potassium phosphate	25	17.5	1.718	1.69	2.673	2.68
41		25	17.5	1.685	1.69	2.894	2.68
42		25	17.5	1.470	1.69	2.837	2.68
43		25	17.5	1.896	0.87	2.730	2.68
44		25	17.5	1.676	1.69	2.273	2.68
45		30	10	1.196	0.97	2.828	2.46
46		30	25	1.185	1.07	2.428	2.25
47		20	10	n.p.	n.p.	n.p.	n.p.
48		25	6.9	n.p.	n.p.	n.p.	n.p.
49		17.9	17.5	n.p.	n.p.	n.p.	n.p.
50		25	28.1	1.019	1.11	1.559	1.77
51		20	25	1.683	1.69	1.842	1.53
52		32.1	17.5	0.983	1.15	2.802	3.05

---

Regression analysis

A Fisher F-test for P_F [F = 17.73] with a very low probability value (p model > F = 0.0008) demonstrated high significance for the regression model as shown in ANOVA (Table 4). Meanwhile for K_e, the Fisher F-test [F = 7.16] with a very low probability value (p model > F = 0.0112) also showed a high significance for the model. The value of the determination coefficient for P_F (R² = 0.93) was calculated and only 7% of the total variations are not explained by the model. The value of the adjusted determination coefficient (Adj. R² = 0.8745) was also high, supporting the high significance of the model. The R² value must always lie from 0 to 1. The model becomes stronger and better in its prediction of response based on the experimental data if the value of R² is closer to 1.0.¹⁹ The coded quadratic model for the purification factor (P_F) and partition coefficient (K_e) is given in eqn (7) and (8), respectively.

Purification factor = +3.02 (7)

Partition coefficient = +1.29 − 0.21 × A² − 0.24 × B² (8)

Table 4 ANOVA for the quadratic model: estimated regression model of the relationship between IL% and potassium phosphate% on (A) purification factor (P_F) and (B) partition coefficient (K_e)

(A)
Source	Sum of squares	df	Mean square	F value	p-Value Prob > F
Model	18.73	5	3.75	17.73	0.0008	Significant
A – IL%	5.67	1	5.67	26.85	0.0013
B – phosphate	6.66	1	6.66	31.53	0.0008
AB	1.67	1	1.67	7.89	0.0262
A^2	2.83	1	2.83	13.41	0.0081
B^2	2.50	1	2.50	11.85	0.0108
Residual	1.48	7	0.21
Lack of fit	1.22	3	0.41	6.36	0.0530	Not significant
Pure error	0.26	4	0.064
Cor total	20.21	12
R-Squared	0.9268
Adj R-squared	0.8745
Pred R-squared	0.5499

---

(B)
Source	Sum of squares	df	Mean square	F value	p-Value Prob > F
Model	3.83	5	0.77	7.16	0.0112	Significant
A – IL%	0.69	1	0.69	6.42	0.0390
B – phosphate	1.39	1	1.39	13.02	0.0087
AB	1.13	1	1.13	10.58	0.0140
A^2	0.31	1	0.31	2.92	0.1311
B^2	0.38	1	0.38	3.59	0.0999
Residual	0.75	7	0.11
Lack of fit	0.58	3	0.19	4.73	0.0837	Not significant
Pure error	0.16	4	0.041
Cor total	4.58	12
R-Squared	0.8364
Adj R-squared	0.7195

---

In order to make sure that the functional relationship between the experimental factors and the response variable was adequately described, the lack of fit (p-value > 0.05) was calculated. The lack of fit (p) values for P_F and K_e were 0.0530 and 0.0837 respectively, reflecting that the experimental data was suited well to the model, and the model is capable of predicting the P_F and K_e. The values of the determination coefficient (R²) for P_F and K_e in the regression model were 0.9268 and 0.8364 while the values of the adjusted determination coefficient (R_adj²) for P_F and K_e were 0.8745 and 0.7195 respectively, expressing a high level of agreement between the experimental and the predicted values.

Interpretation of the response surface model

Three-dimensional (3-D) plots for the purification factor (P_F) and partition coefficient (K_e) as a function of the IL% and salt% are given in Fig. 2. The P_F ranged from −0.7 to 3.5 whereas the K_e ranged from −0.4 to 1.9. The lowest P_F was observed at the lower concentration of the IL [20% w/w] and salt% [10% w/w] in the (Emim)Br/phosphate system (Fig. 2A). It can be seen that an increase in the percentage of IL and salt would increase the values of P_F and K_e. However, the values of P_F and K_e were slightly decreased at high concentrations for both experimental conditions due to the salting out phenomenon.¹³ It is interesting to note that the stability and affinity of lipase towards the IL phase were still maintained even at a high concentrations of IL% ((Emim)Br) and salt% (potassium phosphate).

Fig. 2 Three-dimensional (3D) plots of the (A) purification factor (P_F) and (B) partition coefficient (K_e) as a function of IL% and salt% for the (Emim)Br/potassium phosphate ATPS.

Verification of the predictive model

An optimization study was conducted to determine the optimal operating conditions for the recovery of lipase with a high purification factor. One combination solution should be selected to maximize the response (i.e. P_F and K_e) and the criteria should include the efficiency, energy conserved and its feasible of operation.¹⁹ A verification experiment was conducted to validate the predictive model for the desired P_F and K_e that correspond to IL of 26.50% and salt of 19% with crude loading of 3.75% (Table 5). These optimum conditions gave a P_F value of 3.24. The experimental data (P_F of 3.24) validated that there was only a small difference with the predicted P_F value of 3.33 which gave a yield of 78.89%. As a result, the predictive model can be used to determine the preferred composition of IL (26.5%) and salt (19%) for the partitioning of lipase on ILATPS for the parameters in the subsequent experiments.

Table 5 Predictive model and experimental conditions with optimum P_F and K_e

Optimum solutions
IL%	Salt%	Purification factor, PF		Partition coefficient, Ke
		Predicted	Validated	Predicted	Validated
26.50	19.0	3.33	3.24	1.40	1.32

---

Effect of crude loading (C_L) on partitioning lipase in the (Emim)Br/potassium phosphate ILATPS

To attain the highest yield of lipase upon partitioning in the ILATPS, experiments on the amount of crude lipase that was loaded in the ILATPS were performed at various concentrations of crude loading [i.e. 3.75–19% (w/w)]. The yield trend showed an increase proportionally with increasing C_L% up to 13% (w/w) of C_L. The yield (Y_T) was reduced at higher C_L% (15%, 17% and 19%) (Fig. 3). A reduction in Y_T% at higher C_L [15–19% (w/w)] was commonly caused by the loss of lipase at the interface of the system’s phases and precipitated due to volume exclusion effect. An increase in the C_L%, usually accommodates higher amounts of impurities or protein contaminants which caused a reduction in the P_F and K_e values though it provides the highest Y_T% in the recovery process.⁵ The presence of impurities in the crude loading led to modification of the electrostatic forces and combining of the phases, which in turn, decreased lipase’s affinity towards the IL phase (top phase). The intensity or affinity of lipase to fractionate in the IL top phase of the ILATPS is explained by the value of K_e where the value is reduced with the presence of high impurities.¹¹ In order to determine the optimum C_L% with the highest yield%, the system must also provide suitable and acceptable P_F and K_e values.¹¹ The highest yield (91.47%) was obtained at 13% (w/w) of C_L in the ILATPS, with lower values of P_F (2.482) and K_e (1.045) indicating that the recovered protein still contained some protein impurities. Previously, the validated prediction model of RSM corresponded to a composition of IL of 26.50% and salt of 19% with a crude loading of 3.75%, in which, the highest values of P_F (3.24) and K_e of (1.32) were obtained. Since this IL gave very low yield of recovery (78.89%), it can be considered to not be favorable. The second highest P_F (2.837), K_e (1.045) and Y_T (89.99%) were obtained at 7% (w/w) crude, where an acceptable yield recovery of the protein was obtained and is much better than that obtained in the system with 3.75% (w/w) crude loading. Thus, 7% (w/w) crude was chosen as the optimal crude loading.

Fig. 3 Influence of crude loading% on the recovery of lipase in the (Emim)Br/potassium phosphate ILATPS. An (Emim)Br/potassium phosphate ILATPS comprising 26.5% (w/w) (Emim)Br and 19% (w/w) potassium phosphate was applied.

Influence of the pH of the (Emim)Br/potassium phosphate ILATPS on lipase partitioning with a constant crude loading of 7% (w/w)

The unique features of target proteins such as surface charges and isoelectric point (pI) have provided a powerful tool for understanding lipase partitioning in ILATPSs by manipulating the pH of the salt phase. The experiments were performed at pH values ranging from 6 to 8 which corresponded to 26.5% (w/w) of IL, 19% (w/w) of salt and 7% (w/w) of crude loading. The highest recovery of lipase was observed at pH 7.6 which gave P_F, K_e and Y_T values of 3.137, 1.140 and 97.12% respectively. The Y_T value obtained at pH 8 (97.85%) was comparable to that obtained at pH 7.6. Since the system at pH 8 gave very low values of P_F (3.081) and K_e (1.071), this system was not suitable for selection. The lowest values of P_F (2.874) and K_e (1.111) were obtained at pH 6.6 (Fig. 4). Since lipase usually exhibits an isoelectric point (pI) at 6.3,²⁰ the salt system at pH 6.6 will give the protein a small net charge that leads to an enormous hydrophobicity effect on lipase, which will make fractionation into the IL phase more difficult. Results of this study also revealed that the salt at pH 7.6 had a significant influence on the high electrostatic force and high net charge of lipase. This condition promotes excellent fractioning of lipase into the IL phase. Thus, a pH of 7.6 for the salt phase was chosen as a parameter for subsequent experiments.

Fig. 4 Effect of pH changes on the recovery of lipase in the (Emim)Br/potassium phosphate ILATPS. The (Emim)Br/potassium phosphate ILATPS comprised 26.5% (w/w) (Emim)Br, 19% (w/w) potassium phosphate and 7% (w/w) crude loading.

Effect of salt (NaCl) on lipase partitioning in (Emim)Br/potassium phosphate ILATPS

The electrostatic capacity and hydrophobic factor of target proteins are important criteria that influence the fractioning of lipase in ATPSs, which can be achieved by the addition of NaCl at various concentrations [1–5% (w/w)].^21,22 The most favorable recovery of lipase was obtained in the ILATPS with the presence of 3% (w/w) NaCl, in which P_F (3.394) and K_e (1.352) were significantly increased. However, the effect of addition of high concentrations of NaCl [4 and 5% (w/w)] on lipase partitioning in the ILATPS greatly reduced the P_F values to 2.956 and 2.813 respectively (Fig. 5). NaCl usually promotes high electrostatic capacity in ILATPS that encourages the fractionation of lipase in the IL in the top phase.²¹ Conversely, lipase activity and stability were reduced at high concentrations of NaCl [4–5% (w/w)] due to the salting out phenomenon on ILATPSs.¹³ In conclusion, the suggested compositions of (Emim)Br/potassium phosphate for lipase recovery are as follows: 26.5% (w/w) (Emim)Br, 19% (w/w) potassium phosphate, 7% (w/w) of crude loading and 3% (w/w) NaCl as an enhancer at pH 7.6.

Fig. 5 Effect of NaCl on the recovery of lipase in the (Emim)Br/potassium phosphate ILATPS. The (Emim)Br/potassium phosphate ILATPS comprised 26.5% (w/w) (Emim)Br, 19% (w/w) potassium phosphate and 7% (w/w) crude loading at pH 7.6.

Lipase recovery from ILATPS (SDS-PAGE)

The lipase that had been purified in the ILATPS which was composed of 26.5% (w/w) (Emim)Br, 19% (w/w) potassium phosphate, 7% (w/w) crude loading with the addition of 3% (w/w) NaCl at pH 7.6 was assayed by 12% SDS-PAGE analysis. SDS-PAGE profiles of a standard protein marker (lane M), crude lipase extract (lanes 1 & 2) and recovered lipase in the ILATPS (lanes 4 & 5) are shown in Fig. 6. Results from this study demonstrated that lipase from the crude extract has been successfully purified using the ILATPS, which was proven from the results of the SDS-PAGE assay (lanes 4 & 5), though a small quantity of contaminants were still present. From the protein ladder, the molecular weight of the purified lipase was about 40 kDa. The exact molecular weight of intracellular E. coli BL21 lipase is 43 kDa.^23,24 The crude extract (lanes 2 & 3) had many protein bands in its profile, as it accommodated higher protein impurities resulting from the cell disruption process (Fig. 6.). In conclusion, the (Emim)Br/potassium phosphate ILATPS can be implemented as a single step process for the recovery and purification of lipase from the crude extract.

Fig. 6 SDS-PAGE profiles for E. coli BL21 lipase that had been purified in the ILATPS using 12% SDS-PAGE analysis. Note: lane M = a standard marker protein; lanes 1 and 2 = crude extract with duplicate loading; lanes 4 and 5 = recovered lipase from the top phase of the ILATPS.

Conclusion

(Emim)Br was used as the top phase whereas potassium phosphate was used as the bottom phase to successfully form a biphasic system, which provided a high degree of purity for the recovery of lipase using the ILATPS. The preferred composition of the phases for the ILATPS employed 26.5% (w/w) (Emim)Br and 19% (w/w) potassium phosphate to attain the highest P_F and K_e via a predictive model which used RSM. High efficiency recovery of E. coli BL21 lipase was accomplished using the (Emim)Br/potassium phosphate ILATPS with the composition as follows: 26.5% (w/w) (Emim)Br, 19% (w/w) potassium phosphate at pH 7.6, the presence of 3% NaCl and 7% (w/w) crude enzyme. The crude enzyme extract was loaded into the ILATPS with one step procedure which gave the highest purification performance [K_e (1.352), P_F (3.394), and Y_T (93.75%)]. ILs in ILATPSs posses properties such as low viscosity, the ability to offer recyclability and rapid phase separation, as well as being easy and simple to scale up. Use of ILATPSs can be considered to be a neat and sustainable process that has great potential to be applied in large scale industrial processes for enzyme recovery and purification.

Competing interests

The authors declare that they have no competing interests.

Acknowledgements

This study was funded by the short-term grant of Universiti Sains Malaysia (Project code: P4544-304/PTENKIND/6313233) and Fundamental Research Grant Scheme (F1047-2015/0564, 203/PTENKIND/6711499).

References

1. R. Aravindan, P. Anbumathi and T. Viruthagiri, Indian J. Biotechnol., 2007, 6, 141.
2. R. Sharma, Y. Chisti and U. C. Banerjee, Biotechnol. Adv., 2001, 19, 627–662.
3. N. Pietruszka, I. Y. Galaev, A. Kumar, Z. Brzozowski and B. Mattiasson, Biotechnol. Prog., 2000, 16, 408–415.
4. R. Gupta, N. Gupta and P. Rathi, Appl. Microbiol. Biotechnol., 2004, 64, 763–781.
5. H. S. Ng, C. W. Ooi, P. L. Show, C. P. Tan, A. Ariff, M. N. Moktar, E.-P. Ng and T. C. Ling, Sep. Purif. Technol., 2014, 138, 28–33.
6. P.-Å. Albertsson, Partition of cell particles and macromolecules: separation and purification of biomolecules, cell organelles, membranes, and cells in aqueous polymer two-phase systems and their use in biochemical analysis and biotechnology, Wiley, New York etc., 1986.
7. C. W. Ooi, B. T. Tey, S. L. Hii, A. Ariff, H. S. Wu, J. C. W. Lan, R. S. Juang, S. M. M. Kamal and T. C. Ling, Biotechnol. Bioprocess Eng., 2010, 14, 811–818.
8. N. Gupta, P. Rathi and R. Gupta, Anal. Biochem., 2002, 311, 98–99.
9. M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
10. S. Abbasiliasi, J. S. Tan, T. A. Ibrahim, S. Kadkhodaei, H. S. Ng, F. Vakhshiteh, Z. Ajdari, S. Mustafa, T. C. Ling, R. A. Rahim and A. B. Ariff, Food Chem., 2014, 151, 93–100.
11. R. Hatti-Kaul, Mol. Biotechnol., 2001, 19, 269–277.
12. M. G. Freire, C. M. Neves, I. M. Marrucho, J. N. C. Lopes, L. P. N. Rebelo and J. A. Coutinho, Green Chem., 2010, 12, 1715–1718.
13. Y. Pei, J. Wang, K. Wu, X. Xuan and X. Lu, Sep. Purif. Technol., 2009, 64, 288–295.
14. I.-H. Pan and Y.-K. Li, J. Chromatogr. B: Biomed. Sci. Appl., 2001, 754, 179–184.
15. S. P. M. Ventura, S. G. Sousa, M. G. Freire, L. S. Serafim, Á. S. Lima and J. A. P. Coutinho, J. Chromatogr. B, 2011, 879, 2679–2687.
16. G. A. Baker and W. T. Heller, Chem. Eng. J., 2009, 147, 6–12.
17. Z. Li, Y. Pei, H. Wang, J. Fan and J. Wang, Trends Anal. Chem., 2010, 29, 1336–1346.
18. J. Flieger and A. Czajkowska-Żelazko, Ionic Liquids in Separation Techniques, INTECH Open Access Publisher, 2011.
19. A. C. Atkinson, A. N. Donev and R. D. Tobias, Optimum experimental designs, with SAS, Oxford University Press Oxford, 2007.
20. G. Tasnádi, E. Forró and F. Fülöp, Tetrahedron: Asymmetry, 2009, 20, 1771–1777.
21. C. W. Ooi, B. T. Tey, S. L. Hii, S. M. M. Kamal, J. C. W. Lan, A. Ariff and T. C. Ling, Process Biochem., 2009, 44, 1083–1087.
22. R. Hatti-Kaul, Aqueous two-phase systems: methods and protocols, Humana Press, 2000.
23. T. H. T. A. Hamid, M. A. Eltaweel, R. N. Z. R. A. Rahman, M. Basri and A. B. Salleh, Ann. Microbiol., 2009, 59, 111–118.
24. J. S. Tan, S. Abbasiliasi, Y. K. Lin, M. S. Mohamed, M. R. Kapri, S. Kadkhodaei, Y. J. Tam, R. N. Z. R. A. Rahman and A. B. Ariff, Sep. Purif. Technol., 2014, 133, 328–334.

---