Advanced Hub Main Navigation Menu
Protocell Flow Reactors for Enzyme and Whole-Cell Mediated Biocatalysis
Abstract
The design and construction of continuous flow biochemical reactors comprising immobilized biocatalysts have generated great interest in the efficient synthesis of value-added chemicals. Living cells use compartmentalization and reaction-diffusion processes for spatiotemporal regulation of biocatalytic reactions, and implementing these strategies into continuous flow reactors can offer new opportunities in reactor design and application. Herein, the fabrication of protocell-based continuous flow reactors for enzyme and whole-cell mediated biocatalysis is demonstrated. Semipermeable membranized coacervate vesicles are employed as model protocells that spontaneously sequester enzymes or accumulate living bacteria to produce embodied microreactors capable of single- or multiple-step catalytic reactions. By packing millions of the enzyme/bacteria-containing coacervate vesicles in a glass column, a facile, cost-effective, and modular methodology capable of performing oxidoreductase, peroxidase and lipolytic reactions, enzyme-mediated L-DOPA synthesis, and whole-cell glycolysis under continuous flow conditions, is demonstrated. It is shown that the protocell-nested enzymes and bacterial cells exhibit enhanced activities and stability under deleterious operating conditions compared with their non-encapsulated counterparts. These results provide a step toward the engineering of continuous flow reactors based on cell-like microscale agents and offer opportunities in the development of green and sustainable industrial bioprocessing.
1 Introduction
The development of novel and cost-effective processes for a broad range of chemical reactions to reduce environmental impacts and economic stress is one of the major goals of sustainable chemistry.[1, 2] Biology offers an exquisite collection of enzymes displaying unique enantio- and regioselective transformations with the consequence that biocatalysis is rapidly emerging as a green technology for the modernization of organic synthesis.[3] In particular, there is an increasing emphasis on the confinement of biocatalytic cascade reactions in continuous flow reactors (CFRs) for the development of sustainable bioprocesses.[4, 5] Continuous flow biocatalysis offers several advantages such as flexibility over process control, efficient heat/mass transfer, reaction screening, and scale-up synthesis of value-added chemical and functional molecules for pharmaceuticals.[6, 7] The most common configurations of biocatalytic flow processes comprise fixed-bed reactors in which biocatalysts (enzymes or living cells) are entrapped in a caged network[8, 9] or directly anchored onto the packaging support material by simple physisorption[10, 11] or via chemical crosslinking on polymeric beads[6, 12] and inorganic materials.[13-15] Whilst these efforts have simplified heterogeneous catalysis in continuous flow systems, achieving the maximum loading of enzymes and whole cells on the support material without generating diffusion limitations and loss of catalytic activity continues to be a challenge.[16-18]
Compartmentalization is an essential feature of living cells that facilitates the simultaneous occurrence of a diverse range of metabolic processes. Recently, we and others have reported the design and construction of artificial cells (protocells) to mimic the basic behaviors of living cells.[19-22] In this regard, coacervate micro-droplets derived from liquid–liquid phase separation of oppositely charged polyelectrolytes have been extensively explored as a simple membrane-free cytomimetic model. Coacervate micro-droplets selectively compartmentalize key functional bio-components such as proteins,[23, 24] polynucleotides,[25] ribozymes,[26, 27] and chloroplasts,[28] and in living cells, coacervates have been found as P granules,[29] stress granules,[30] and Cajal bodies,[31] indicating their importance in intracellular processes and membraneless organelles. Coacervates as model protocells have been developed to realize a diverse range of cell-like behaviors including communication,[32] motility,[33] metabolism,[34] gene expression,[35, 36] phagocytosis,[37] artificial cytoskeleton reconstitution.[38] and membrane gating.[39] As most intracellular enzymatic reactions take place in molecularly crowded or confined environments,[40, 41] coacervate micro-droplets are a useful cytomimetic system as they possess chemically enriched viscoelastic interiors, biocompatibility, and large surface areas, and exhibit endogenous uptake of enzyme substrate molecules and dynamic diffusion properties. In addition, coacervates can spontaneously absorb and encapsulate bacterial cells[34] or form ultra-thin protective shells around living bacteria.[42] As such, coacervate-based model protocells should be outstanding biocompatible immobilization scaffolds for biocatalytic synthesis. However, little progress has been made in practical or macroscopic applications of coacervate micro-droplets as biocatalytic agents, mainly because they are fragile, prone to coalescence, and spontaneously disassemble under detrimental pH and temperature conditions.[43] Although the membranization of coacervate microdroplets using a fatty acid bilayer,[44] block copolymers,[45] erythrocyte membrane fragments,[46] nanoparticles,[39] or protein–polymer conjugates[47] has been reported recently, the resulting microscale constructs are often too fragile to be of practical value.
To advance the application of protocell research in sustainable bioprocessing, herein, we present the utilization of membrane-bound coacervate-based protocells as a host matrix for enzyme/whole cell-mediated biocatalytic transformations in continuous flow systems. To achieve this, functional coacervate micro-droplets prepared from polydiallydimethylammonium chloride (PDDA) and adenosine 5´-triphosphate (ATP) are reconfigured into membrane-bound coacervate vesicles by electrostatically-induced complexation of a polyanionic polyoxometalate (sodium phosphotungstate, PTA [PW11O39]7−) with PDDA macromolecules present at the surface of the coacervate droplets.[48] The resulting polyoxometalate coacervate vesicles (PCVs) comprise a robust semi-permeable PDDA/PTA/ATP shell that contains desired guest components and an internal aqueous lumen. By packing millions of enzyme-containing PCVs in a glass column, we show a facile and cost-effective methodology for the fabrication of protocell-based CFRs that operate as all-water catalytic platforms (Figure 1). We demonstrate that our methodology provides opportunities to build modular fixed-bed reactors either by co-encapsulation of biocatalytic cascades in a single population of PCVs or by segregation of biochemical networks in two different populations of the PCVs. The catalytic performance and stability of the PCV-CFRs are assessed using widely applied enzymes such as lipase, glucose oxidase (GOx)/horseradish peroxidase (HRP), and tyrosinase. We show that under constant flow conditions, the PCV-CFR biochemical transformations exhibit improved activity and storage stability. As a proof of concept, the application potential of our approach is demonstrated by using tyrosinase-containing PCV-CFRs for the synthesis of an active pharmaceutical ingredient, 3,4-dihydroxy-L-phenylalanine (Levodopa; L-DOPA), which is a non-protein amino acid widely used for the treatment of Parkinson's disease.[49] We also show that our strategy enables encapsulation and proliferation of the Gram-positive bacterium, Lactococcus lactis (L. lactis) within the PCVs to produce a hybrid protocell chassis capable of operating as living microbial cell factories for the generation of lactic acid under continuous flow conditions.
Taken together, our results highlight a facile cost-effective approach for the design and construction of programmable cytomimetic microcompartments that provide an effective milieu for the enhanced loading of biocatalysts, stabilization of biochemical processes, and diffusion-controlled spatiotemporal regulation of multi-enzymatic cascade reactions. We envisage that protocell-based chassis will provide opportunities to extend the catalytic lifetime of nested enzymes and microbial communities and facilitate the modulation of conventional fixed-bed reactor technologies for a range of biotechnological applications.
2 Results
2.1 Design and Construction of Protocell-Based Continuous Flow Reactors
We selected PCVs as a suitable protocell model for the fabrication of CFRs because of their stability across a wide range of pH and ionic strength values (Figure S1a, Supporting information), ability to encapsulate diverse functional guest molecules such as enzymes, nanoparticles and photocatalysts, facile synthesis protocol,[48] and scalability to gram-scale quantities. Notably, the semipermeable PCV membrane enabled the diffusion of low molecular weight substrates (molecular weight cut-off, ≈900 Da);[48] while, the transfer of macromolecules such as fluorescein isothiocyanate tagged-dextran (FITC-dextran, molecular weight 70 kDa) was limited (Figure S1b, Supporting information). Typically, aqueous solutions of PDDA (50 mm) and ATP (30 mm) were mixed at pH 8 to produce suspensions of coacervate microdroplets, followed by addition of PTA (100 mm) to trigger spontaneous restructuration of the droplets into membranized PCVs with a mean diameter of 27 µm. The PCVs comprised an external PTA/PDDA nanocomposite membrane that encased a secondary PDDA/ATP coacervate layer and water-filled lumen (Figure S2a–j, Supporting information). The production of biocatalytic PCVs was achieved by the spontaneous sequestration of enzymes, such as Dylight-405 tagged lipase (Dylight-405-lipase), into the coacervate microdroplets prior to the addition of the PTA solution. Reconfiguration into the PCVs resulted in localization of the enzymes within the inner compressed PDDA/ATP coacervate layer rather than in the water-filled vacuole (Figure S2b,f, Supporting information). Zeta potential measurements of the as-prepared PCVs showed a negative surface charge potential (−31.8 mV; Figure S2k, Supporting information). Scanning electron microscopy (SEM) images of the freeze-dried PCVs showed the presence of hollow microcompartments comprising a uniform organic/inorganic shell with thickness of ≈ 665 nm (Figure S2l, Supporting information). Energy dispersive X-ray (EDX) analysis of the PCVs showed a homogenous distribution of tungsten across the membrane (Figure S3, Supporting information).
We employed borosilicate glass columns (length 4 cm, inner diameter 0.5 cm) for the fabrication of customized and cost-effective continuous flow reactors. Typically, one end of the glass column was sealed with a commercial cotton plug, a suspension of freshly prepared biocatalytic PCVs was then injected into the column, and the open end of the glass tube was subsequently sealed with another cotton plug to produce a column reactor ≈1.2 cm in length and comprising a homogenously packed population of PCVs (volume number density ≈22 200 µL−1) (Figure 2a; Figure S4, Supporting information). Optical/fluorescence microscopy images (Figure 2b–d) and corresponding 3D reconstructions of the column bed reactor (Figure 2e) loaded with FITC-lipase-containing PCVs (FITC-lipase@PCVs) revealed that the enzyme-containing protocells were continuously and densely packed in random arrangements throughout the CFR. To test the potential of the FITC-lipase@PCV-loaded glass column as a biocatalytic flow reactor, resorufin butyrate (5 µm) solution was pumped through the column reactor at a pump rate of 20 µL min−1, and the time-dependent formation of the fluorescent product (resorufin) within the CFR was monitored by placing the column reactor under a confocal fluorescence microscope (Video S1, Supporting information). Corresponding time-lapse images showed that uptake of the enzyme substrate by the semi-permeable PCVs in the column gave rise to lipase-mediated hydrolysis and production of resorufin over a period of ≈30 min (Figure 2f). During the initial feeding stages, the fluorescent product was located both in the membrane and the lumen of the PCVs, indicating that the product was partially retained within the protocells. The biocatalytic reaction showed a progressive increase in fluorescence intensity with time (Figure 2h), suggesting that the catalytic activity of the encapsulated enzymes was retained under the flow conditions. Upon completion of the feeding cycle, the column was supplied with PBS buffer at a flow rate of 20 µL min−1 for 30 min to flush out the reaction product (Figure 2g). Significantly, the fluorescence intensity associated with the FITC-lipase@PCVs was maintained after the washing procedure, indicating that the enzyme was irreversibly trapped within the PCVs under the conditions employed.
To identify optimal operating conditions required to maintain the mechanical integrity of the PCVs, pressure drop measurements at different flow rates were performed. For this, a pressure sensor was connected to the inlet and outlet ends of a PCV-CFR (packing length, 1.2 cm; volume number density ≈22 200 µL−1). A continuous flow of water through the column showed a near-linear increase in pressure drop as the water flow rate increased from 0.5 to 8 mL min−1 (Figure 2i). An increase in flow rate to 9 mL min−1 caused a sudden rise in the pressure drop to 136 kPa and a decrease in the packing length of the PCV reactor bed from 1.2 to 1 cm, indicating a partial collapse or breakage of some of the PCVs due to the increased pressure. Optical microscopy images confirmed that most of the PCVs remained structurally intact and spherical in morphology under a pressure drop of up to 136 kPa (Figure S5, Supporting information), reflecting the operating mechanical strength of the PCVs. We then studied the flow type in the PCV-CFR by the step-response method.[50, 51] Given these observations, we determined the hold back value (H) and average residence time (τ) for the column; the H value was ≈0.046, and 98% of the substrate had a residence time within τ ± 0.3τ (Figure 2j; See Supporting Information). This data suggested that the flow in the column reactor was consistent with an ideal plug flow and that the flow properties of the reactor were not modulated by the cotton wool stoppers.
2.2 Single-Enzyme Biocatalysis in PCV Flow Reactors
The catalytic performance of the PCV-CFRs was examined with two widely used enzymes, GOx and lipase (Figure 3a). Typically, the encapsulation efficiencies of GOx and lipase in the PCVs were 66 ± 2% and 91 ± 2%, respectively, under the experimental conditions employed (Figure S6, Supporting information). The higher encapsulation efficiency of lipase was attributed to the increased hydrophobic character of the enzyme, which facilitated preferential accumulation within the coacervate phase with a lower polarity than the surrounding solution.[22] The activities of GOx and lipase within the PCVs under mild batch conditions (25 °C, MES buffer pH 6.0 [GOx], PBS buffer pH 7.4 [lipase]) were moderated after exposure to changes in pH and temperature (Figure S7, Supporting information). Both enzymes showed improved activity over the free enzymes in acidic (pH 3) and alkaline (pH 9) environments and enhanced thermal resistance when compared with their native counterparts.
Having established the robustness of the enzyme-containing PCVs, we prepared CFRs and tested their operational properties. For GOx@PCV-CFR with a continuous flow of β–D–glucose (substrate), the concentration of the product (H2O2) increased linearly with an increase in residence time of the substrate when the flow rates exceeded 200 µL min−1 and attained a steady state at lower flow rates (Figure 3b). Similar results were obtained for the lipase@PCV-CFR with a continuous feed of p-nitrophenyl acetate (PNPA), where the product (p-nitrophenol [PNP]) concentration increased linearly with the residence time of the PNPA for overflow rates ranging from 50 to 300 µL min−1 (Figure 3b). In both CFRs, the results approximated to the corresponding enzyme kinetics determined in homogenous solutions, where the reaction velocities remained constant in the initial stages and gradually decreased with increases in reaction time. Significantly, there was almost no loss in GOx or lipase activity after six reaction cycles (Figure 3c), indicating that the sequestered enzymes remained functional throughout repeated loading/washing cycles.
Given the recycling stability of the protocell-based CFRs, we employed the Michaelis–Menten equation to determine the reaction kinetic parameters in the PCV-CFRs.[52-54] Initial enzyme reaction velocities were obtained by changing the substrate concentration at a constant flow rate associated with linear residence time/product concentration conditions (GOx@PCV-CFR, 250 µL min−1; lipase@PCV-CFR, 200 µL min−1). Significantly, the initial velocity (Vo) values for GOx@PCV-CFR and lipase@PCV-CFR were higher than for the corresponding free enzymes. The GOx@PCV-CFR showed a 74% increase in Vo at 0.6 mm β–D–glucose concentration, whereas a nearly ninefold increase of Vo was observed for the lipase@PCV-CFR at 0.25 mm PNPA (Figure 3d,e). The initial velocity data were fitted with the Michaelis–Menten equation to obtain the maximum velocity (Vmax) and Michaelis constant (Km). The Vmax values for GOx@PCV-CFR and lipase@PCV-CFR were increased by 6% and 38%, respectively, when compared to free GOx and lipase in homogenous solutions (Figure 3f). We attributed the significant increase in Vmax values to the enrichment of the enzymes and their substrates within the two-layered PCV membrane and the rapid removal of reaction products under continuous flow conditions. Further, the Michaelis constant (Km) value for GOx@PCV-CFR (Km = 1.4±0.2 mm) was 1.6 times smaller than for free GOx (Km = 2.3 ± 0.5 mm), suggesting an undiminished substrate affinity of the GOx molecules nested in the PCVs and minimal diffusion restrictions in the PCV-CFRs.[50, 55, 56] Interestingly, the affinity of lipase for PNPA was significantly improved in the lipase@PCV-CFR, where the Km (8.9 ± 2.0 mm) was 5.1 times smaller than the free lipase (45.3 ± 6.8 mm) in solution. Lipase was known to work effectively at the water/oil interface,[57] suggesting that the enhanced affinity could be ascribed to the increase of interfacial area and contact frequency between lipase and PNPA when co-located within the model protocells.
Given the above observations, the potential shelf-life of the above PCV-CFRs was examined by storing the devices in appropriate buffers at 4 °C for 14 days. Significantly, the lipase@PCV-CFR retained 98% of initial activity after storage for 14 days; while, the activity of free lipase decreased by 29% following storage under the same conditions (Figure 3g). Similar results were observed for the GOx@PCV-CFR, indicating that the PCV microenvironment enhanced the stability of the enzymes during storage.
Based on the above results, the potential of protocell-based continuous flow reactors for the synthesis of value-added chemicals was explored by constructing a tyrosinase@PCV-CFR capable of generating L-DOPA. Tyrosinase-containing PCVs were prepared and packed into a glass tube to produce a reaction column of length 9 mm (176.4 µL). Plots of L-DOPA production at different pump rates indicated that protocell-mediated transformation of 1 mm of L-tyrosine gradually reached a steady state value at pump rates below 100 µL min−1, which decreased in concentration as the pump rate increased (Figure 3h). An operational compromise between conversion efficiency and reaction time (≈3.5 min.) was achieved by using a pump rate of 50 µL min−1, which gave a relatively high conversion rate (31%) and high production rate (1024 mg L−1 h−1) of L-DOPA. Significantly, the tyrosinase@PCV-CFR exhibited reusability, with 96% efficiency retained in the sixth reaction cycle (Figure 3i). The above L-DOPA production rates were higher than for reported biotechnological production studies,[58] suggesting that PCV-based CFRs could have application potential in the production of value-added chemicals.
2.3 Multi-Enzyme Cascades Using Modulated PCV-Based Continuous Flow Reactors
We exploited the high propensity of PDDA/ATP coacervate microdroplets to spontaneously sequester multiple enzyme molecules to generate individual PCVs or PCV binary populations comprising two-enzyme cascade reactions. In the former case, PCVs containing both GOx and HRP (GOx + HRP@PCVs) with encapsulation efficiencies of 50 ± 6% and 55 ± 4%, respectively, were prepared (Figure 4a,b; Figure S6, Supporting information); while, in the latter system, we prepared separate populations of GOx-containing PCVs (GOx@PCVs) and HRP-containing PCVs (HRP@PCVs). Two different biocatalytic CFR designs were then fabricated using either a single population of GOx + HRP@PCVs or by sequential loading of the GOx@PCVs upstream of the HRP@PCV populations (Figure 4c). The performance of the tandem biochemical reaction using the two different designs was investigated by simultaneously supplying the GOx substrate (β–D glucose, 0.05–6 mm) and HRP substrate (o-phenylenediamine [OPD, 2.5 mm]) at a pump rate of 300 µL min−1 and monitoring the production of fluorescent 2,3-diaminophenazine (2,3-DAP) (Figure S9, Supporting information). Interestingly, the reaction velocities determined for the GOx + HRP@PCV-CFR and GOx@PCV + HRP@PCV-CFR arrangements exhibited similar Vmax values (Vmax = 0.2 mm min−1) but were 2.5 times lower when compared with the free enzyme cascade (Vmax = 0.5 mm min−1, Figure 4d,e). Specifically, at lower substrate concentrations (β–D–glucose < 1 mm), the reaction velocities of the CFRs were higher than that of the free enzyme system, but the activity decreased with increase in substrate concentration. This was attributed to HRP fatigue due to the local increase in H2O2 concentration and concomitant saturation of the enzyme active sites. Significantly, the Km value of the GOx + HRP@PCV-CFR (376 µm) was lower than that of the GOx@PCV + HRP@PCV-CFR (407 µm) and was ≈4.6 times less than for the free enzyme cascade (1713 µm). The improved substrate affinity of the GOx + HRP@PCV-CFR was attributed to the homogeneous sequestration of the GOx/HRP cascade in the PCVs. As a result, the GOx/HRP co-assembly facilitated a substrate channelling effect, where the product H2O2 was directly transferred to HRP without equilibration with or diffusion through the surrounding bulk solutions.[59, 60]
2.4 Whole-Cell Biocatalysis in PCV-Based Continuous Flow Reactors
Living cells are considered as a cost-effective integrated catalytic system for a range of bio-mediated transformations.[61, 62] As whole-cell biocatalysis requires viable cells, appropriate immobilization methodologies are often required for sustainable catalytic efficiency. In this context, we sought to expand the feasibility of PCV-CFRs for use in whole-cell biocatalysis by demonstrating the capacity to support a viable bacterial population within the PCVs. Specifically, we employed cultures of L. lactis as a test system to implement whole-cell glycolysis within the PCVs under continuous flow conditions. Unlike the spontaneous internalized sequestration of enzymes, addition of L. lactis cells to an aqueous dispersion of PDDA/ATP coacervate microdroplets resulted in adsorption of the bacteria on the surface of the microdroplets (Figure 5a; Figure S10, Supporting information). Significantly, addition of PTA and reconfiguration into the PCVs did not displace the bacteria, which became attached primarily to the outer and inner surface of the newly formed PDDA/PTA/ATP hybrid membrane (L. lactis@PCV) (Figure 5b; Figure S11, Supporting information). Although the number of bacterial cells initially trapped in the PCV membrane was relatively low (mean number per PCV, 1200), incubation of the L. lactis@PCVs samples in M17 medium for 5 days supported bacterial growth, giving rise to two orders of magnitude increase in the number of viable bacterial cells (mean number per PCV, 408540) associated with the protocells (Figure 5c). Whilst the number of viable bacteria gradually decreased after 5 days, viability assays revealed that ≈79% and 60% of the encapsulated L. lactis population present at 5 days remained metabolically active within the PCVs after 20 and 31 days at 30 °C, respectively. To facilitate subsequent diffusion of nutrients into the PCV lumen, the L. lactis@PCVs were washed with PBS buffer to produce 1.2 ± 0.43 µm-sized macropores within the PCV membrane by a charged-induced (phosphate) etching process, which also released small numbers of the bacterial cells from the PCV membrane (Figure 5d,e). The number of pores formed within the membranes increased with an increase in exposure time to the buffer solution (Figure S12, Supporting information). Consequently, growth of the trapped bacterial population resulted in migration of the proliferating L. lactis cells from the PCV membrane into the central lumen to establish an encapsulated microbial community (Figure 5f,g; Figure S13, Supporting information). CLSM analysis and fluorescence intensity measurements on stained samples revealed that growth of the encapsulated L. lactis colony gave rise to a bacterial biofilm across the PCV lumen (Figure 5h). Corresponding SEM analysis of fractured L. lactis@PCVs at different cell density loadings showed the presence of a 3D interconnected network of bacterial cells (Figure 5i; Figure S14, Supporting information). Consequently, the encapsulated microbial population became self-immobilized and unable to escape from the PCVs (Video S2, Supporting information). Formation of the 3D biofilm network increased the viability of the encapsulated bacterial cells when incubated in acidic (pH 3), neutral (pH 7.4), or alkaline (pH 8) conditions, as well as after exposure to extreme temperatures (4 °C and 45 °C) and centrifugal forces (9610 × g) for 1 h. In each case, the percentage of dead bacteria was higher for free L. lactis cells compared to that of the L. lactis@PCVs (Figure 6a). This was consistent with the increased levels of reactive oxygen species (ROS) detected within the free bacterial cells, indicative of oxidative stress in response to adverse conditions (Figure 6b).
The biocatalytic performance of the L. lactis@PCVs was initially studied under batch conditions. First, we investigated the glycolysis activity of the L. lactis@PCVs by incubating the protocell-enclosed bacterial colonies in M17 medium supplemented with 1 wt% glucose and monitoring lactic acid production over 12 h (Figure 6c). Interestingly, the initial rate of lactic acid production in the L. lactis@PCVs was considerably higher than that for free L. lactis cells for the same bacterial numbers (1.0 × 1011 CFU) (Figure 6d). In both cases, the final lactate concentration after 12 h was 5 mg mL−1 under the conditions employed. The improved fermentation rate of the encapsulated L. lactis was attributed to the spontaneous uptake and sequestration of nutrients within the PCVs and the interconnected biofilm matrix, thereby enabling faster adaptation to the conditions and metabolic rates compared to the PCV-free counterparts. We then loaded the L. lactis@PCVs into glass column reactors and measured the residence times at different pump rates to establish conditions for the steady state production of lactic acid (Figure 6e) and the corresponding optimal rate for glycolysis (Figure 6f). Using these conditions (pump rate, 30 µL min−1; [glucose] = 1 wt%), the L. lactis@PCV-CFR was operated for 7 days. The reaction velocity for lactic acid production gradually increased in the first 24 h due to growth of the bacteria within the PCVs, and then, was maintained at a steady state for the subsequent 6 days (Figure 6g). During this period, the reaction rate for producing lactic acid was 3.13 mg mL−1 min−1 with a conversion rate of 82% on day 7. Significantly, most of the bacteria within the PCV-CFR were still alive after 7 days (Figure S15, Supporting information).
3 Conclusion
In this study, we present a proof-of-concept approach to the engineering of CFRs based on cell-like microscale agents. The facile polyoxometalate-mediated reconfiguration of molecularly crowded membraneless PDDA/ATP coacervate microdroplets into three-tiered semipermeable PCVs was exploited to generate distributed micro-chassis for single/multiple enzyme cascades and whole-cell biocatalysis under reaction flow conditions. The mechanical robustness of the protocell membrane facilitated retention and optimization of reaction performance, physical stability, and recyclability of the PCV-CFRs under adverse pH and temperature conditions. In general, the methodology was flexible, providing opportunities to fabricate a wide range of modulated CFRs by loading single PCV communities with co-encapsulated enzyme cascades or employing sequential loading of segregated PCV communities housing different biocatalysts. We demonstrated that high yields of an active pharmaceutical agent (L-DOPA) could be achieved using a tyrosinase@PCV-CFR, and that the accelerated production of lactic acid over 7 days was accomplished by packing a CFR with metabolically active L. lactis@PCVs as living bacterial cell factories.
Compared to conventional biocatalyst immobilization supports such as gel matrices or porous inorganic materials, the high biomolecule encapsulation efficiency of the coacervate-based protocells is determined by the molecularly crowded environment and associated non-covalent interactions within the hybrid organic–inorganic membrane.[24, 44, 48] The PCV membrane serves as a charge-dense buffer that mitigates potentially deleterious microenvironmental changes on biocatalyst performance[17, 63] and provides a cell-like semipermeable membrane that minimizes the extent of diffusive limitation on reactivity.[16] Interestingly, although macromolecular diffusion across the PCV membrane is limited, high levels of macromolecular binding to the semi-permeable PTA/PDDA/ATP nanocomposite shell would co-locate cognate enzymes and high molecular weight substrates such that the protocell micro-reactors should remain operational to substrates across a wide range of molecular weights. Moreover, the ability to etch the PCV membrane generates a pathway for the encapsulation and proliferation of metabolically active bacterial colonies with loadings of up to 40 cells µm−3 along with enhanced viability and improved fermentation rates. This facile procedure, which involves a combination of physical entrapment on the membrane, capture within the lumen, and self-immobilization by biofilms assembly, can provide a novel approach to whole-cell immobilization.[64] More generally, our approach opens opportunities to encapsulate communities of diverse living cells within discrete packages of PCVs that can be sequentially arranged to create modular CFRs for whole-cell flow biocatalysis. The construction of such integrated whole-cell CFRs by design is expected to offer advantages such as significantly reduced catalyst costs, higher stability, and minimal requirement for external cofactor addition.
Taken together, our work demonstrates a straightforward and cost-effective approach to the construction and operation of robust cytomimetic microcompartments that offer a molecularly crowded environment for the loading and stabilization of biocatalytic and metabolic processes, as well as for the spatiotemporal regulation of multi-enzymatic and metabolic reactions under directional flow.
We anticipate that the demonstrated methodology should provide a promising path to address the current challenges in biocatalyst immobilization through protocell research and enable the construction of protocell-based microreactors programmed with specific biochemical transformations as well as encapsulation of genetically-engineered whole cells for the conversion of desired feedstock into pharmaceutical agents and value-added chemical agents for the development of sustainable industrial bioprocessing, microbial fuel cells, and wastewater treatment.
Acknowledgements
H.M. and S.M. were funded by the CSC-UoB scholarship and ERC Advanced Grant schemes (EC-2016-674 ADG 740235), respectively. A.J.P acknowledges funding from the University of Bristol. The authors thank Dr R. Moreno Tortolero and Dr J.-C. Eloi for their assistance with pressure drop measurements and scanning electron microscopy imaging of the protocell samples, respectively.
Conflict of Interest
The authors declare no conflict of interest.
Author Contributions
A.J.P and S.M conceived and directed the project. All authors contributed to design and planning of the experiments. H.M performed the experiments. A.M. helped in the design and optimization of enzyme catalysis experiments. X.L. and A.H.N helped with bacterial cell work and expertise, respectively. All authors discussed the results and commented on the manuscript. H.M., A.J.P, and S.M. wrote the manuscript.
Open Research
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
