Decorating outer membrane vesicles with organophosphorus hydrolase and cellulose binding domain for organophosphate pesticide degradation

Chemical Engineering Journal 308 (2017) 1–7

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Decorating outer membrane vesicles with organophosphorus hydrolase and cellulose binding domain for organophosphate pesticide degradation Fu-Hsiang Su, Ian Dominic Flormata Tabañag, Chih-Yun Wu, Shen-Long Tsai ⇑ Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei 10607, Taiwan

h i g h l i g h t s

g r a p h i c a l a b s t r a c t


An OPH and a CBD were appended to

OMVs via genetic fusion.
The engineered OMVs exhibited an

enhanced activity toward paraoxon degradation.
Recovery of the engineered OMVs could be achieved easily via cellulose pull-down.
OPHs tethered on OMVs revealed improved temperature, pH and longterm stabilities.

a r t i c l e

i n f o

Article history: Received 19 July 2016 Received in revised form 8 September 2016 Accepted 9 September 2016 Available online 10 September 2016 Keywords: Outer membrane vesicles Organophosphorus hydrolase Cellulose binding module Paraoxon Biodegradation

a b s t r a c t Outer membrane vesicles (OMVs) are nanoscale spheres naturally released from Gram-negative bacteria. They contain a diverse array of proteins and lipopolysaccharide but do not replicate, which increases their safety profile and renders them attractive for environmental applications. Herein, an efficient and reusable biocatalyst for enhanced degradation of organophosphate pesticides was developed. Organophosphorus hydrolase (OPH) was tethered onto OMVs via a genetically fused ice nucleation protein (INP) to form OMV-based biocatalysts. To accomplish quick purification and easy recovery of the engineered OMV using cellulose, a cellulose binding module (CBM) was collaterally tethered on the OMV. The OPH-decorated OMVs exhibited an enhanced degradation rate when assayed with paraoxon as a substrate. In addition, the thermal stability and pH tolerance were also enhanced remarkably. Furthermore, the resulting biocatalysts could still retain more than 80% activity even after 15 cycles of recovery and reuse, demonstrating their potential use in bio-catalytic decontamination of organophosphate compounds. Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction Due to intensive industrialization and manufacturing activities, leaks and inappropriate disposal of organic compounds have resulted in the contamination of environment. Bioremediation, which refers to the use of biological systems such as bacteria, fungi and enzymes to degrade pollutants, has been widely recognized as ⇑ Corresponding author. E-mail address: [email protected] (S.-L. Tsai). http://dx.doi.org/10.1016/j.cej.2016.09.045 1385-8947/Ó 2016 Elsevier B.V. All rights reserved.

a powerful tool for the treatment of environmental contaminants [1]. In most cases, bioremediation usually exploits mixed cultures of microorganisms in either natural or engineered environments, wherein the specific enzymes produced by the microorganisms and their mechanistic actions involved in pollutant biodegradation are not yet fully established in some systems [2]. With recent advances in molecular biology, mechanisms of pollutant degradation can be determined with such ease and accuracy that enzymes and microbial systems can be characterized and utilized more efficiently [3]. Compared to the utilization of microbial

2

F.-H. Su et al. / Chemical Engineering Journal 308 (2017) 1–7

systems in pollutant degradation, enzymatic systems are more efficient and attractive due to the fact that it can degrade the pollutant of interest with an increased tolerance, higher rate and specificity, resulting in less by-products and biomass formation [4]. For example, enzymes that can hydrolyze organophosphorus compounds have been identified and characterized from different microbial species thus, exploiting these enzymes for effective detection and biodegradation of organophosphate pesticide seems to be an emerging approach [5,4,6]. However, major limitations for enzymatic bioremediation are the high cost, and tedious work required for enzyme extraction and purification [7,8]. Enzyme immobilization is a process that makes the enzymes more resistant to harsh environments and enables the enzymes to be recovered and recycled by attaching the enzyme to a solid support (e.g. polymers, metals, nanomaterials) [9,10]. Furthermore, enzyme activities after immobilization are sometimes reduced due to enzyme conformational change, mass transfer limitation, or detachment of enzymes from supports [11]. For example, an elegant study on enzymatic bioremediation of organophosphorus insecticides by recombinant organophosphorus hydrolase showed significant reduction in enzyme activity after immobilization [12]. Therefore, an efficient enzyme immobilization technology that can provide strong interaction between supports and enzymes without compromising the activity of enzymes is highly desired. Outer membrane vesicles (OMVs) are protein, lipid, and polysaccharide-based vesicles derived from the outer membrane of many gram-negative and gram-positive bacteria as part of their natural growth cycle [13]. They are produced when small portions of the outer membrane bulge away from the cell, pinch off, and then release. Although it is still poorly understood why certain membrane proteins are preferentially enriched in OMVs while others are excluded, it has been demonstrated that Escherichia coli (and other gram-negative bacteria) can be engineered to incorporate heterologously expressed proteins into either the membrane or the lumen fraction of OMVs [14,15]. Engineered OMVs are particularly attractive for enzyme immobilization applications owing to their nanoscale size, ease and low cost of production via fermentation and their extreme stability at ambient conditions [16]. However, the recovery of OMVs requires the utilization of an ultracentrifugation step, which is cost, energy and labor intensive, especially for environmental applications. Therefore, a simple and cost-effective method for OMV recovery is considered a necessity. Cellulose binding domains (CBDs) are proteins that can bind rapidly, tightly, and specifically to cellulose in a wide range of pH and temperatures [17]. Therefore, they are commonly used in biotechnology as a fusion partner for immobilizing proteins on cellulose [18]. Cellulose is a naturally abundant, inexpensive, and chemically inert material with inherently low binding characteristics [19]. The unique properties of CBDs, in addition to the low cost of cellulose, may not only eliminate the aforementioned problems in the recovery of OMVs, but also enable the commercial application of OMV-based enzymatic bioremediation. Herein, we investigated the possibility of targeting an organophosphorus hydrolase (OPH) and a CBD to the surface of OMVs using two complementary surface anchors for quick and simple recovery of the engineered OMVs using cellulose. In addition, we also demonstrated the potential use of the OPH-and-CBD-decorated OMVs in decontaminating an organophosphate pesticide, paraoxon.

2. Materials and methods 2.1. Strains and plasmids E. coli strains JM109 and JC8031 were used in this study for plasmid construction and OMV production, respectively. To display

the OPH on the surface of OMVs, plasmid pVLT33-INPOPH6 encoding a fusion protein that contains an ice nucleation protein (INP) and the OPH was constructed by PCR amplification of the corresponding DNA fragment from pINCOP [20] by using the forward primer FEcorI3INP: 50 -gggggaattcaggaaacgatgaatatcg-30 and the reverse primer ROPHHis6HindIII: 50 -ccccaagctttcagtggtggtggtggtg gtgtacgcccaaggtcg-30 . The PCR product was then digested with EcoRI and HindIIII and inserted into linearized pVLT33 [21]. Plasmid pUCBD [22] that has been described elsewhere was used for surface display of the CBD on OMVs. A control E. coli strain, JC8031 harboring pVLT33 and pUC18 plasmids, was developed for comparison.

2.2. Protein expression and OMVs collection The recombinant E. coli was pre-cultured in 3 mL of LB medium supplemented with appropriate antibiotics at 37 °C for 12 h. Then, the cells were transferred into 100 mL of LB medium containing 10 mmol/L of CoCl2 and appropriate antibiotics, and then incubated at 37 °C with continuous shaking at 250 rpm in the dark before the optical density (OD600nm) reaches 0.8. The cells were then cooled to 30 °C, and isopropyl-b-D-thiogalactopyranoside (IPTG) was added to a final concentration of 100 lmol/L. After 4 h of induction, cells were harvested by centrifugation at 5000 g for 10 min at 4 °C. The cell-free supernatant was collected and passed through a 0.45 lm pore-size vacuum filter (Millipore). Vesicles were collected by centrifugation at 50,000 g for 2 h and resuspended in 50 mmol/L of Tris-HCl buffer (pH 8) supplemented with 10 mmol/L of CoCl2. 2.3. Protein characterization Proteins on OMVs were characterized via SDS-PAGE using Laemmli method (also known as glycine-based buffer system). SDS-PAGE gels were blotted onto a polyvinylidene difluoride membrane. OPHs on OMVs were probed by a primary anti-His tag mouse monoclonal antibody (Genetex) and labeled with a secondary AP-conjugated goat anti-mouse IgG antibody (Jackson Immuno). The expression of CBM on OMVs was characterized by assaying the binding ability of OMVs on microcrystalline cellulose Avicel (Sigma-Aldrich). Briefly, 5 lg of OMVs were mixed with Avicel in 50 mmol/L of Tris-HCl (pH 8) containing 10 mmol/L of CaCl2 for 10 min. After centrifugation at 5000 rpm for 3 min, the OMVbound Avicel was precipitated and washed three times with icecold Tris-HCl (pH 8). The pellets were collected and then tested for their activities on paraoxon (Sigma-Aldrich) degradation. The amount of OPH on Avicel was quantitatively determined via densitometry after Western blot analysis.

2.4. Immunofluorescence microscopy The display of OPH enzyme on the E. coli surface was verified via immunofluorescence microscopy. Briefly, E. coli cells displaying OPH were harvested by centrifugation. Collected cells were resuspended in 250 lL of PBS containing 1 g/L bovine serum albumin and 0.5 lg of anti-His6 immunoglobulin G (IgG; Genetex) for 4 h with occasional mixing. Then, the probed cells were pelleted, washed with PBS, and resuspended in PBS plus 1 g/L bovine serum albumin and 0.5 lg anti-mouse IgG conjugated with Alexa 488 (Thermo scientific) for labeling. After incubation for 2 h, the labeled cells were pelleted, washed twice with PBS, and resuspended in PBS to an OD600nm of 1. Five microliter of the labeled suspensions were spotted onto glass slides and were further analyzed under a fluorescence microscope (Olympus IX73, USA).

F.-H. Su et al. / Chemical Engineering Journal 308 (2017) 1–7

3

2.5. Scanning electron microscopy

3. Results and discussion

Samples were resuspended in 100 lL of 8% glutaraldehyde and 4% paraformaldehyde in 0.1 mol/L of sodium cacodylate buffer. Two microliter of diluted samples were spotted on a glass slide and dehydrated in an oven at 45 °C after being washed twice with 0.1 mol/L of sodium cacodylate buffer. The dehydrated samples were coated with gold palladium and micrographs were obtained by SEM (Hitachi, Japan) at different resolutions.

OPH exhibits a remarkable enzymatic activity towards a wide range of organophosphates and has been successfully expressed in many microorganisms, making it a promising candidate for enzymatic bioremediation [23]. As the OPH always exists as a free-floating ‘soluble’ enzyme, enzyme immobilization has become a critical strategy for improving its stability and recovery. In this study, a CBD was fused to a surface-anchoring Lpp-OmpA protein [24], and an OPH was fused to another surface-anchoring protein INP to create an efficient biocatalyst with high activity and easy recovery capability towards paraoxon degradation (Fig. 1). By fusing CBD and OPH to the Lpp-OmpA and INP fusion vehicles, respectively, the two proteins were displayed together on the E. coli surface which later drove the decoration of the two proteins on released OMVs. By taking advantage of the system of OMVs, enzymes could be continuously produced by cells and released into the medium without the need to disrupt the cells. The released OMVs could be readily captured via affinity interaction between the displayed CBD and cellulose before being recovered via gravity precipitation.

2.6. OPH activity measurement Enzyme activities were determined in 50 mmol/L of Tris-HCl buffer (pH = 9.00) at 30 °C. Rates of p-nitrophenol production were determined from the measured absorbance readings at 412 nm using a spectrophotometer (JASCO V-630, Japan). The calibration plot of p-nitrophenol is provided in Fig. S1 (Supplementary Content). Paraoxon was immediately added to the samples prior to the analyses. To evaluate the performance of OPH in different configurations (e.g. free OPH, OPH-OMV, OPH-OMV-Cellulose), the Michaelis–Menten model, as described in Eq. (1), was used to relate reaction velocity to paraoxon concentration for each configuration.

v¼

V max ½S K M þ ½S

ð1Þ

where v is the initial degradation rate. Vmax represents the maximum velocity, KM is the substrate concentration at which the reaction velocity is half of the Vmax. [S] is the concentration of the substrate S. Furthermore, kcat, a parameter referred to as the turnover number, represents the number of substrate molecule each enzyme site converts to product per unit time and operationally defined as kcat = Vmax/[E] where [E] is the enzyme concentration. Taking the reciprocal of Eq. (1) gives

1

v

¼

KM 1 1  þ V max ½S V max

ð2Þ

which transforms Michaelis–Menten model into one that gives a straight-line plot. The plot of 1/v versus 1/[S], called LineweaverBurk plot, yields an intercept of 1/Vmax and a slope of KM/Vmax. Enzyme kinetics experiments were done at paraoxon concentrations ranging from 10 lmol/L to 1 mmol/L. Initial rates were determined by the slope of the first 5 min of reaction with the paraoxon substrate. A Lineweaver-Burk plot was utilized to determine the enzyme kinetic parameters (Vmax and Km) of Michaelis– Menten model. All data were obtained from three independent experiments. 2.7. Temperature and pH effects and reusability The thermal stabilities of both free and immobilized OPH were evaluated as described elsewhere with some modifications [23]. Briefly, enzyme samples were heated for 30 min at different temperatures ranging from 20 °C to 80 °C. The heated samples were cooled down to room temperature prior to the determination of enzyme activity as described above. The effect of pH on OPH activity was monitored under different pH values ranging from 5 to 11 at 30 °C. To test the reusability of the engineered OMVs, samples were added into a microcentrifuge tube containing paraoxon and shook at 30 °C. Samples were monitored periodically and the enzymatic activity for the reaction was determined using the method described above. Subsequently, the used OMVs were collected and then reused in another reaction.

3.1. Protein expression and functionality The E. coli strain harboring the two plasmids pVLT33-INPOPH6 and pUCBD was induced with 100 lmol/L of IPTG to express the OPH and CBD on the cell surface. To check whether the OPH was successfully expressed and correctly translocated onto the cell surface, immunofluorescent labeling of cells was carried out with anti-His tag serum and Alexa 488-conjugated goat anti-mouse IgG. As shown in Fig. 2A, cells expressing OPH enzymes were brightly fluorescent. In contrast, no fluorescence was observed in the control E. coli cells. Since monoclonal antibodies are not able to penetrate the cell wall [25], the fluorescence images confirmed that the OPH enzymes are displayed on the cell surface. To investigate if the OPH and CBD proteins displayed on the cell surface could be properly incorporated into the released OMVs, the OMVs collected through ultracentrifugation were systematically characterized. The presence of OPH on the OMVs was analyzed via Western blotting. OMVs were suspended in SDS-PAGE loading dye and boiled. After being separated by SDS-PAGE, proteins on OMVs were blotted onto polyvinylidene fluoride membranes and detected with antibodies specific for the His-tag (Fig. 2B). The single band observed at ca. 83 kDa, which corresponds to the molecular size of INP and OPH fusion protein as predicted by its nucleotide sequence confirmed the expression and incorporation of OPH onto OMVs. The functionality of CBD on OMVs to bind onto cellulose was visualized by SEM images (Fig. 2C). OMVs released from the control strain did not possess CBDs. Therefore, they could not attach to the cellulose surface. In contrast, OMVs released from the designed strain were able to attach onto cellulose via the specific interaction between CBD and cellulose, demonstrating the existence and functionality of CBDs on OMVs. To examine the coexistence and functionality of both proteins on released OMVs, a pull-down activity assay was performed. After exposing the OMVs to celluloses, the celluloses were pulled down via gravity centrifugation. Ideally, the OMVs should be captured by the cellulose if they contained CBDs. In addition, the captured OMVs should be able to hydrolyze paraoxon and produce pnitrophenol if they also contained OPHs. In contrast, there should not be any activity from the cellulose incubated with the OMVs that only possess OPH. As shown in Fig. 2D, the appreciable activity difference between the control OMVs and the OMVs possessing both OPH and CBD indicated that the functionalities of both proteins on the OMVs were preserved. More importantly, although

4

F.-H. Su et al. / Chemical Engineering Journal 308 (2017) 1–7

Fig. 1. Schematic illustration of the design, production and recovery of engineered OMVs. Genetically engineered E. coli was employed to produce OPH and CBD decorated OMVs, which could be used for paraoxon degradation and recycled via cellulose precipitation.

Fig. 2. (A) Immunofluorescent microscope images of the control E. coli cells and the E. coli cells displaying OPH. The scale bar represents 200 nm. (B) Western blot analysis of OPH extracted from the engineered OMV. (C) SEM photograph of the cellulose surface after exposed to control OMVs and engineered OMVs. (D) Pull-down enzyme activity assay of the cellulose after exposed to control OMVs and engineered OMVs.

F.-H. Su et al. / Chemical Engineering Journal 308 (2017) 1–7

Lpp-OmpA and INP fusion vehicles have individually been demonstrated to be engineered to incorporate heterologously expressed proteins into the membrane fraction of OMVs [14], their coexistence in a single OMV has not yet been demonstrated. Though it is still poorly understood why certain membrane proteins are preferentially enriched in vesicles while others are excluded [26], to our best knowledge, this is the first study which demonstrated that both Lpp-OmpA and INP can anchor to the same OMV without eliminating each other.

3.2. Paraoxon degradation and kinetics In a significant number of examples, enzyme activities were reduced after immobilization as the native configuration of the protein is perturbed and the substrate availability is reduced [27–29]. However, there are also cases suggesting that immobilization increases the apparent activity of enzymes [23,30–32]. Different from chemical conjugation and physical adsorption, the immobilization method used in this study offers several advantages. First, the OPH is attached on the surface of OMVs via polypeptide bond, which could not only provide higher affinity than physical adsorption bur also reduce the mass transfer barrier. Second, enzyme immobilization is occurred in vivo, which can prevent the denaturation and activity reduction of OPHs due to chemical conjugation. Third, the interaction between OMVs and cellulose is through the high affinity binding of CBD towards cellulose, preserving the independence and integrity of OPHs. Based on these characteristics, it is expected that this strategy will not diminish the activity of enzymes. To investigate the impact of OMV and OMV-CBD-Cellulose configurations on the activity of OPH, the time profile of paraoxon degradation of different configurations was monitored. As shown in Fig. 3A, in comparison to the freely diffusing OPH (control) in solution, paraoxon degradation rates of both the OPHs on OMVs and the OPHs on cellulose-bound OMVs were notably enhanced. To demonstrate that OPH of the OMV and OMV-CBD-Cellulose configurations can similarly improve the catalytic activity even towards other organophosphate pesticides, parathion and methyl-parathion were tested (Fig. S2, Supplementary Content). Again, faster degradation rates were observed for the OPHs on OMVs and the OPHs on cellulose-bound OMVs. With the goal of creating highly robust and active OPH composites for decontamination and detoxification of pesticides, the result was very encour-

5

aging as it suggested that the activity of OPH was not disturbed, but appreciably increased after immobilizing on OMVs. In order to understand the potential mechanism causing such an elevated enzyme performance, we then thought to compare the enzyme kinetic characteristics for the free and the immobilized OPHs. Different OPH configurations (e.g. free OPH, OPH-OMV, OPHOMV-Cellulose) were exposed to various concentrations of paraoxon substrate. OPH catalyzes the hydrolysis of paraoxon to generate an equimolar amount of p-nitrophenol, which has a yellowish color with an absorption maximum at 412 nm. This property facilitates a quick and easy spectrophotometric measurement of paraoxon biodegradation. Therefore, initial rates of substrate consumption were determined from p-nitrophenol production. As seen in Fig. 3B, for both the OPH on OMV and the OPH on cellulosebound OMV complexes, their initial rates were constantly and significantly higher compared to the initial rates of equal amounts of free OPH. This implies that OPH immobilization on either OMV or cellulose-bound OMV complexes confers an increase in activity. Among the three configurations, the OPH on OMV configuration exhibited the highest degradation rate which was around 2-fold higher than that of the free OPHs. This increase in activity for the OPH-OMV configuration further implies that OPH is displayed on the OMV surface at a higher localized density such that the substrate can be efficiently converted into the product with insignificant product and substrate diffusion effects. The degradation rate of the OPHs on cellulose-bound OMVs was slightly lower than that of the OPHs on OMVs, presumably due to the steric hindrance effect (which impose some diffusion limitations of either product, substrate, or both product and substrate on the OPH active sites) caused by cellulose or the spatial rearrangement of OPHs after the OMVs attached onto cellulose. Nevertheless, such a small loss in enzyme activity will not affect its practical use as this result clearly demonstrated that the activity of OPHs could be greatly enhanced by OMV surface display technology. The kinetic parameters of the free OPHs, the OPHs on OMVs and the OPHs on cellulose-bound OMVs were calculated and fitted into a Michaelis–Menten kinetics model. The Michaelis–Menten constant KM, the catalytic rate kcat, and specificity constant kcat/KM were estimated utilizing a Lineweaver-Burk analysis and the data are listed in Table 1. The KM for the free and OMVs displayed OPHs were comparable, confirming that this immobilization approach did not cause any significant diffusion limitation between OPHs and paraoxon. However, when compared to the free OPHs, the immobilized OPHs (both the OPH-OMV and OPH-OMV-Cellulose

Fig. 3. (A) Performance of paraoxon degradation by different configuration of OPHs. (B) Degradation rates of different configuration of OPHs as a function of paraoxon concentration. OPH (free): freely suspended OPH; OPH (OMV): OPH tethered on freely suspended OMV; OPH (OMV/C): OPH tethered on cellulose-bound OMV. Experiments were conducted in triplicate and with same amounts of immobilized or free enzymes.

6

F.-H. Su et al. / Chemical Engineering Journal 308 (2017) 1–7

Table 1 Enzyme kinetics of OPH from different configuration. Format

KM (lM)

kcat (s1)

kcat/KM (lM1 s1)

OPH (free) OPH (OMV) OPH (OMV/C)

47.95 ± 9.36 42.14 ± 5.22 51.27 ± 8.14

3513 ± 216 5716 ± 379 5579 ± 336

73.26 ± 19.28 135.64 ± 63.86 108.82 ± 18.48

complexes) exhibited a roughly 1.7-fold higher kcat, confirming the role of OMV structure in enhancing the overall catalytic effect. It has been reported that OMVs carry some metal acquisition systems that are able to bind several metals in the environment, providing the bacteria with access to these essential compounds [33]. Since OPH is a metalloenzyme that requires binuclear metals to activate, interaction between the OMV-bound Co2+ and the OMVdisplayed OPH may stimulate the activity [34]. Such an activation effect of enzyme carriers has also been reported from our lab and other labs as well [23,35,36]. As OMVs also contain other surface proteins, another possible factor for the enhanced catalytic activity was the increased capture or localization of substrate in vicinity of OPH surface by interaction with OMV surface proteins. Improved local enzyme density and proximity effects are also speculated to be the potential reasons. On the other hand, for the OPHs on cellulose-bound OMVs, the kcat were compatible to that of the OPHs on OMVs. However, the KM was slightly increased, indicating that cellulose attachment had a modest impact on the substrate diffusion (as discussed in the previous paragraph) towards the OPHs on the cellulose-bound OMVs.

Richins et al. used a direct fusion of OPH with CBD for paraoxon degradation and the results showed a remarkable increase in KM and a noticeable decrease in kcat after cellulose attachment [37]. In comparison from the said results, although KM was slightly increased, the kcat was not significantly changed in our study. This result clearly demonstrated the benefit of using OMV display technology to preserve the independence and integrity of OPHs after cellulose attachment. 3.3. Characteristics and reusability of engineered OMVs Though immobilization does not necessarily lead to the stabilizing of enzyme, immobilization techniques have also been used to improve the operational efficiency of enzymes in environments that possess pH, temperature, and substrate concentration which ranges beyond those typically experienced in native enzymatic habitats. To assess the effects of OMV structure on the thermal and pH stability of OPHs, activities of different formats of OPHs were evaluated under various pH and temperatures. Fig. 4A shows the thermal stability of the OPHs after being heated for 1 h at several different temperatures ranging from 30 to 70 °C. Interestingly, both OMVs displayed OPHs and cellulose-bound OMV/OPHs offered a noticeable increase in thermal stability. In addition, the OPHs that are displayed on the cellulose bound OMVs exhibited an even better improvement in thermal stability than those displayed on OMVs only. Similarly, OPHs displayed on OMVs and cellulose-bound OMVs also improved the operational pH range to some extent (Fig. 4B), suggesting that the steric constraint imposed

Fig. 4. (A) Relative enzyme activity of different configuration of OPHs after being treated at different temperature. The activity measured at 30 °C was defined as 100%. (B) Relative enzyme activity of different configuration of OPHs after exposed to different pH. The activity measured at pH = 9 was defined as 100%. (C) Performance of the OPH on cellulose-bound OMVs towards paraoxon degradation after different number of repeated cycles. The activity measured before recycling was defined as 100%. (D) Long-term stability of different configuration of OPHs. OPH (free): freely suspended OPH; OPH (OMV): OPH tethered on freely suspended OMV; OPH (OMV/C): OPH tethered on cellulose-bound OMV. Experiments were conducted in triplicate and with same amounts of immobilized or free enzymes.

F.-H. Su et al. / Chemical Engineering Journal 308 (2017) 1–7

by OMVs and cellulose had a favorable effect on maintaining the OPH structure and making denaturation more difficult. When developing a biocatalyzed process, the enzyme reusability has become one of the key parameters to consider for the reduction of industrial costs. To investigate the reusability of the OPHs on cellulose-bound OMVs, the biocatalyst was recycled via gravity precipitation and then subjected to another run of enzymatic reaction; the activity obtained after each run was compared. Encouragingly, only around 15% reduction in OPH activity was observed throughout the fifteen paraoxon degradation and reuse cycles, demonstrating an excellent reusability of the developed biocatalyst (Fig. 4C). Apart from the reusability, the long term stability of enzymes is also important as enzymes are notoriously unstable and are susceptible to reduced activity over a period of time. To evaluate the stability of enzymes after days of storage at 4 °C the retained activity of each different configurations of OPHs was monitored periodically. As shown in Fig. 4D, while the free OPH was completely deactivated within 30 days, the OPHs on OMVs and cellulose-bound OMVs still retained 20% and 30% activity, respectively, even after 40 days of storage. Clearly, displaying OPHs on cellulose-bound OMVs can help the stabilization of OPHs and make the OPHs less susceptible to inactivation, making this system a powerful and robust tool for the degradation of organophosphorus pesticides. 4. Conclusion Via genetic engineering, OPH was successfully incorporated and displayed on the OMV surface. With the addition of CBD on the engineered OMV, the OPH could be readily recovered through the affinity interaction between CBD and cellulose. The cellulosebound OMV biocatalyst exhibited an enhanced activity towards paraoxon degradation. Besides, the robust structure of the cellulose and OMV was able to improve the operational efficiency of the engineered OPH in pH and temperature settings beyond those typically experienced in habitats of the free OPH. The excellent reusability and the remarkable durability also suggested its potential applications in biocatalytic degradation of organophosphorus pesticides. Acknowledgments This research was funded by the Ministry of Science and Technology, Taiwan (Grant Number: MOST 104-2221-E-011-009). The authors are grateful to Prof. Wilfred Chen of the University of Delaware for the plasmid pUCBD. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2016.09.045. References [1] S. Gupta, B. Pathak, M.H. Fulekar, Molecular approaches for biodegradation of polycyclic aromatic hydrocarbon compounds: a review, Rev. Environ. Sci. Biotechnol. 14 (2015) 241–269. [2] L.W. Perelo, Review: in situ and bioremediation of organic pollutants in aquatic sediments, J. Hazard. Mater. 177 (2010) 81–89. [3] P.J. Lien, Z.H. Yang, Y.M. Chang, Y.T. Tu, C.M. Kao, Enhanced bioremediation of TCE-contaminated groundwater with coexistence of fuel oil: effectiveness and mechanism study, Chem. Eng. J. 289 (2016) 525–536. [4] M. Alcalde, M. Ferrer, F.J. Plou, A. Ballesteros, Environmental biocatalysis: from remediation with enzymes to novel green processes, Trends Biotechnol. 24 (2006) 281–287. [5] B.K. Singh, A. Walker, Microbial degradation of organophosphorus compounds, FEMS Microbiol. Rev. 30 (2006) 428–471. [6] K.E. Le Jeune, J.R. Wild, A.J. Russell, Nerve agents degraded by enzymatic foams, Nature 395 (1998) 27–28.

7

[7] M.D. Aitken, Waste treatment applications of enzymes: opportunities and obstacles, Chem. Eng. J. 52 (1993) B49–B58. [8] M.A. Rao, R. Scelza, F. Acevedo, M.C. Diez, L. Gianfreda, Enzymes as useful tools for environmental purposes, Chemosphere 107 (2014) 145–162. [9] C. Hou, Y. Wang, H. Zhu, H. Wei, Construction of enzyme immobilization system through metal-polyphenol assisted Fe3O4/chitosan hybrid microcapsules, Chem. Eng. J. 283 (2016) 397–403. [10] E. Woo, H. Kwon, C. Lee, Preparation of nano-magnetite impregnated mesocellular foam composite with a Cu ligand for His-tagged enzyme immobilization, Chem. Eng. J. 274 (2015) 1–8. [11] R.C. Rodrigues, C. Ortiz, A. Berenguer-Murcia, R. Torres, R. Fernandez-Lafuente, Modifying enzyme activity and selectivity by immobilization, Chem. Soc. Rev. 42 (2013) 6290–6307. [12] M. Kapoor, R. Rajagopal, Enzymatic bioremediation of organophosphorus insecticides by recombinant organophosphorous hydrolase, Int. Biodeterior. Biodegrad. 65 (2011) 896–901. [13] J.L. Baker, L.X. Chen, J.A. Rosentha, D. Putnam, M.P. DeLisa, Microbial biosynthesis of designer outer membrane vesicles, Curr. Opin. Biotechnol. 29 (2014) 76–84. [14] N.C. Kesty, M.J. Kuehn, Incorporation of heterologous outer membrane and periplasmic proteins into Escherichia coli outer membrane vesicles, J. Biol. Chem. 279 (2004) 2069–2076. [15] J.Y. Kim, A.M. Doody, D.J. Chen, G.H. Cremona, M.L. Shuler, D. Putnam, M.P. DeLisa, Engineered bacterial outer membrane vesicles with enhanced functionality, J. Mol. Biol. 380 (2008) 51–66. [16] B.S. Collins, Gram-negative outer membrane vesicles in vaccine development, Discov. Med. 62 (2011) 7–15. [17] M. Linder, T.T. Teeri, The roles and function of cellulose-binding domains, J. Biotechnol. 57 (1997) 15–28. [18] I. Levy, O. Shoseyov, Cellulose-binding domains biotechnological applications, Biotechnol. Adv. 20 (2002) 191–213. [19] M. Jonoobi, R. Oladi, Y. Davoudpour, K. Oksman, A. Dufresne, Y. Hamzeh, R. Davoodi, Different preparation methods and properties of nanostructured cellulose from various natural resources and residues: a review, Cellulose 22 (2015) 935–969. [20] M. Shimazu, A. Mulchandani, W. Chen, Cell surface display of organophosphorus hydrolase using ice nucleation protein, Biotechnol. Progr. 17 (2001) 76–80. [21] V. Delorenzo, L. Eltis, B. Kessler, K.N. Timmis, Analysis of Pseudomonas GeneProducts Using Laciq Ptrp-Lac Plasmids and Transposons That Confer Conditional Phenotypes, Gene 123 (1993) 17–24. [22] A.A. Wang, A. Mulchandani, W. Chen, Whole-cell immobilization using cell surface-exposed cellulose-binding domain, Biotechnol. Progr. 17 (2001) 407– 411. [23] S.H. Leng, C.E. Yang, S.L. Tsai, Designer oleosomes as efficient biocatalysts for enhanced degradation of organophosphate nerve agents, Chem. Eng. J. 287 (2016) 568–574. [24] C. Stathopoulos, G. Georgiou, C.F. Earhart, Characterization of Escherichia coli expressing an Lpp’OmpA(46–159)-PhoA fusion protein localized in the outer membrane, Appl. Microbiol. Biotechnol. 45 (1996) 112–119. [25] M.L. Rodrigues, L.R. Travassos, K.R. Miranda, A.J. Franzen, S. Rozental, W. de Souza, C.S. Alviano, E. Barreto-Bergter, Human antibodies against a purified glucosylceramide from Cryptococcus neoformans inhibit cell budding and fungal growth, Infect. Immun. 68 (2000) 7049–7060. [26] E.Y. Lee, J.Y. Bang, G.W. Park, D.S. Choi, J.S. Kang, H.J. Kim, K.S. Park, J.O. Lee, Y.K. Kim, K.H. Kwon, K.P. Kim, Y.S. Gho, Global proteomic profiling of native outer membrane vesicles derived from Escherichia coli, Proteomics 7 (2007) 3143– 3153. [27] C. Forsyth, T.W.S. Yip, S.V. Patwardhan, CO2 sequestration by enzyme immobilized onto bioinspired silica, Chem. Commun. 49 (2013) 3191–3193. [28] G. Kaur, S. Saha, M. Tomar, V. Gupta, Influence of immobilization strategies on biosensing response characteristics: a comparative study, Enzyme Microb. Technol. 82 (2016) 144–150. [29] S. Hudson, J. Cooney, E. Magner, Proteins in mesoporous silicates, Angew. Chem. Int. Ed. 47 (2008) 8582–8594. [30] J.C. Breger, S.A. Walper, E. Oh, K. Susumu, M.H. Stewart, J.R. Deschamps, I.L. Medintz, Quantum dot display enhances activity of a phosphotriesterase trimer, Chem. Commun. 51 (2015) 6403–6406. [31] Y.F. Zhang, J. Ge, Z. Liu, Enhanced activity of immobilized or chemically modified enzymes, ACS Catal. 5 (2015) 4503–4513. [32] S.L. Tsai, M. Park, W. Chen, Size-modulated synergy of cellulase clustering for enhanced cellulose hydrolysis, Biotechnol. J. 8 (2013) 257–261. [33] C. Schwechheimer, M.J. Kuehn, Outer-membrane vesicles from Gram-negative bacteria: biogenesis and functions, Nat. Rev. Microbiol. 13 (2015) 605–619. [34] J.K. Grimsley, B. Calamini, J.R. Wild, A.D. Mesecar, Structural and mutational studies of organophosphorus hydrolase reveal a cryptic and functional allosteric-binding site, Arch. Biochem. Biophys. 442 (2005) 169–179. [35] J. Ge, J.D. Lei, R.N. Zare, Protein-inorganic hybrid nanoflowers, Nat. Nanotechnol. 7 (2012) 428–432. [36] L.B. Wang, Y.C. Wang, R. He, A. Zhuang, X.P. Wang, J. Zeng, J.G. Hou, A new nanobiocatalytic system based on allosteric effect with dramatically enhanced enzymatic performance, J. Am. Chem. Soc. 135 (2013) 1272–1275. [37] R.D. Richins, A. Mulchandani, W. Chen, Expression, immobilization, and enzymatic characterization of cellulose-binding domain-organophosphorus hydrolase fusion enzymes, Biotechnol. Bioeng. 69 (2000) 591–596.