Artificial Cell Fermentation as a Platform for Highly Efficient Cascade

Dec 20, 2017 - *E-mail: [email protected]. ... Because of its high specificity and stereoselectivity, cascade reactions using enzymes have been ...
0 downloads 24 Views 2MB Size
Letter Cite This: ACS Synth. Biol. XXXX, XXX, XXX−XXX

pubs.acs.org/synthbio

Artificial Cell Fermentation as a Platform for Highly Efficient Cascade Conversion Kei Fujiwara,* Takuma Adachi, and Nobuhide Doi Department of Biosciences and Informatics, Faculty of Science and Technology, Keio University, Yokohama 223−8522, Japan S Supporting Information *

ABSTRACT: Because of its high specificity and stereoselectivity, cascade reactions using enzymes have been attracting attention as a platform for chemical synthesis. However, the sensitivity of enzymes outside their optimum conditions and their rapid decrease of activity upon dilution are drawbacks of the system. In this study, we developed a system for cascade enzymatic conversion in bacteria-shaped liposomes formed by hypertonic treatment, and demonstrated that the system can overcome the drawbacks of the enzymatic cascade reactions in bulk. This system produced final products at a level equivalent to the maximum concentration of the bulk system (0.10 M, e.g., 4.6 g/L), and worked even under conditions where enzymes normally lose their function. Under diluted conditions, the conversion rate of the artificial cell system was remarkably higher than that in the bulk system. Our results indicate that artificial cells can behave as a platform to perform fermentative production like microorganisms. KEYWORDS: liposomes, bioconversion, cascade reactions, synthetic biology, artificial cells As an intermediate field between in vivo and in vitro synthetic biology, artificial cell systems have recently gained attention.23−25 Particularly, cell-sized unilamellar liposomes, also called artificial cells (ARTCELLs), can contain enzymes inside a lipid bilayer, and have attracted attention as a life-mimicking system.23,26−28 We predicted that enzymatic conversion in ARTCELLs could overcome the drawbacks of in vitro synthetic biology because of the characteristic similarities of these systems to microorganisms. However, to our knowledge, the maximum concentration produced by the system was less than 10 mM of final products so far. Hence, a highly efficient cascade conversion to levels such as those by produced through in vivo and in vitro synthetic approaches is desirable in the ARTCELL system. However, effective cascade conversion to levels such as 100 mM have not been achieved using this system, and thus, its feasibility is limited. To demonstrate that artificial cells can be used in highly efficient substance conversion systems, we constructed an artificial ethanol synthesis pathway consisting of a three-step reaction (Figure 1A). In this system, lactate dehydrogenase (LDH) initiated the conversion of lactate to pyruvate in a coenzyme NAD+-dependent manner. Subsequently, pyruvate decarboxylase (PDC) converted pyruvate to acetaldehyde, and finally, alcohol dehydrogenase (ADH) produced ethanol from acetaldehyde and NADH. In this study, LDH was derived from Bacillus subtilis var. natto, which is a microorganism used in a Japanese traditional fermented food, natto. PDC was derived from Zymomonas mobilis, a microorganism used for tequila

A

characteristic feature of live cells is that membranes work as a boundary between inside and outside the cell, maintaining intracellular homeostasis through the uptake and emission of molecules and energy. The process of fermentative production using microorganisms utilizes this feature; inexpensive substances taken into cells are converted to beneficial and/or precious chemicals like foods, drugs, and biofuels. To expand the capabilities of fermentation, metabolic engineering using genetic engineering, also called in vivo synthetic biology, has been established.1,2 This strategy is based on introducing artificial metabolic pathways into commonly used microorganisms such as Escherichia coli and fermentative yeasts.3−7 Although in vivo synthetic biology is effective, it requires various steps, such as inactivation of inhibitory metabolic pathways and consideration of side reactions and toxicity. As an alternative to fermentation using live cells, in vitro enzymatic conversion, also called in vitro synthetic biology, has been recently utilized.8−10 In this method, enzymes are used as catalysts in an aqueous system. Because of its high specificity and enzyme stereoselectivity, this system is easily applied to one-pot synthesis using cascade reactions.9,11,12 For simple reactions, this system can convert substrates to products at concentrations of near 15 g/L (∼100 mM).13 In recent years, studies have shown that this system can be applied for complicated reactions such as converting sugars to hydrogen, fatty acids, and other compounds.14−18 Although the reuse and recycle of biocatalysts is possible under certain factors,19,20 enzymes are very sensitive to environmental conditions such as pH and salt concentrations, and in the case of complicated reactions, their conversion efficiency drops sharply when diluted.21,22 © XXXX American Chemical Society

Received: October 12, 2017

A

DOI: 10.1021/acssynbio.7b00365 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology

Figure 1. Artificial lactate-ethanol conversion pathway used in this study. (A) A schematic representation of the artificial cell conversion system in this study. (B) Ethanol production under various enzyme concentrations after a 12-h reaction. All enzymes were used at the indicated concentrations. Concentration of the cofactor, NAD+, was fixed at 1.5 mM. Error bars indicate standard deviations (n = 3).

Figure 2. Lactate-ethanol conversion by enzymes contained in ARTCELLs. (A) A schematic representation of the conversion system using ARTCELLs in this study. (B) Fluorescent images of the ARTCELLs used in this study. (C) Digestion of enzymes after a 1-h Proteinase K (ProK) treatment. (D) Dependence of ethanol production on enzyme concentrations inside ARTCELLs. Error bars indicate standard deviation (n = 3). Notably, ethanol synthesis by the ARTCELLs was lower than that by the bulk system (Figure, 1B) because of the low concentration of total enzymes in the ARTCELL system (approximately 4 or 12 nM in the reaction tubes).

the outside and release the final product to the outside (Figure 2A). We prepared ARTCELLs containing LDH, PDC, ADH, and the coenzyme NAD+ by the droplet transfer method.32 In this method, enzymes are first encapsulated by water-in-oil droplets covered with a lipid monolayer, and then the droplets are transferred to cell-sized liposomes by passing through the oil−water interface covered with another lipid monolayer (Supplementary Figure S2A). We selected POPC as a lipid because this lipid yields high rates of liposome formation. To visualize the encapsulation of small molecules and macromolecules inside the ARTCELLs, a small green fluorescence molecule (calcein) and a red fluorescent protein (mCherry) were also added as markers (Figure 2B). Under our experimental conditions, ARTCELLs formation slightly changed the enzyme concentration from that of the initial bulk solution (Supplementary Figure S2B). However, 96.9 ± 1.2% of enzymes in the initial solution failed to encapsulate inside ARTCELLs owing to a low conversion rate from droplet to liposomes, and therefore, they remained in the

fermentation, and ADH was derived from Saccharomyces cerevisiae, a microorganism used for fermenting beer, wine, and bread. Both lactate and ethanol in microorganisms are final products in the fermentation process to acquire energy anaerobically, and conversion from lactate to ethanol supplies neither metabolites nor energy. Thus, this system is not active in natural cells unless LDH, PDC, and ADH are overexpressed. The concentrations of ethanol synthesized were quantified by an enzymatic method using alcohol oxidase and peroxidase29 which specifically detects ethanol in this artificial pathway (Supplementary Figure S1). In bulk (test tubes), this system synthesized about 80 mM ethanol when 150 mM lactate was added as the initial substrate (Figure 1B). Next, this artificial pathway was encapsulated in ARTCELLs. Since lipid membranes are semipermeable and efficiently pass small molecules less than 100 Da,30 lactate (90 Da) and ethanol (44 Da) can permeate through the membrane by diffusion.31 This means that ARTCELLs can uptake initial substrates from B

DOI: 10.1021/acssynbio.7b00365 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology

Figure 3. Conversion using the ARTCELL system proceeds in a solution detrimental to enzymes. (A) NAD+ dependence of ethanol production in bulk and using ARTCELL. Ethanol production levels were normalized to the values obtained under NAD+-supplied conditions in each case (bulk and ARTCELL). (B,C) Ethanol production using lactate at various pH values. Ethanol production levels were normalized to the values obtained from lactate pH 6.7 in the case of bulk reaction (B) or in the case of conversion using ARTCELL (C). Substrate pH indicates the pH of lactate outside artificial cells. Error bars indicate standard deviations (n = 3). RU: relative unit.

utilize a substrate in a solution that is detrimental to the enzyme. Another possible merit of the ARTCELL system compared with enzymes in bulk is resistance to dilution of the system concentration. “System concentration” is defined as the concentration of all enzymes and cofactors in the system where their concentration ratios are kept constant. For example, a system concentration of 1/2 means that concentrations of all enzymes and NAD+ were halved. In complex reaction networks using multiple enzymes, production rates of final products do not necessarily show a linear relationship with the system concentration.22 Especially, cooperative relationships, such as that between dehydrogenases and NAD+, can behave as a second-order reaction in the case of system concentrations lower than the critical concentration (Supplementary Discussion). Therefore, the probability of their cooperative reaction decreases in the case of a test tube under diluted conditions. In contrast, inside ARTCELLs, the system concentrations are not changed by dilution, and therefore, their cooperative reactions still proceed efficiently. This led to the hypothesis that ARTCELL systems can efficiently progress cascade reactions under diluted conditions compared with bulk systems when the reaction rates show a nonlinear response to the system concentration. This hypothesis was verified by diluting the bulk and ARTCELL systems (Figure 4A). ARTCELLs or their catalytic elements (LDH, PDC, ADH, and NAD+) in bulk were simultaneously diluted, and then their resulting ethanol synthesis was quantified. As expected, ethanol synthesis levels decreased nonlinearly when catalytic elements were diluted in the bulk system (Figure 4A, black line). In contrast, diluting the amount of ARTCELLs linearly decreased ethanol synthesis (Figure 4A, gray line). The results were recapitulated by simulating the ethanol synthesis pathway using Cell Designer, a general-purpose biochemical reaction simulator35 (Figure 4B, see details in Methods). The simulation suggested that the nonlinear decrease of ethanol synthesis upon dilution in the bulk system was derived from cooperative catalysis between enzymes and NAD+. Next, to improve conversion efficiency, ARTCELLs were deformed to mimic bacterial morphology. Here, we adopted enforced deformation using osmotic pressure.36,37 Because lipid bilayers act as semipermeable membranes, hypertonic treatments promote removal of water from the inner media. Consequently, liposome volume decreases without changing the surface area, increasing the surface area-to-volume ratio (S/ V) and concentrations of impermeable molecules (Figure

external solution. To remove the enzymes from the external solution, buffers were replaced twice, and Proteinase K (ProK) was added to degrade proteins (Figure 2A, see Methods). ProK is a strong protease with broad specificity and is used to remove external proteins.33 Actually, ProK digested input enzymes in a 1-h reaction (Figure 2C). Due to washing twice and the ProK treatment, a control experiment using liposomes without entrapped enzymes showed that ethanol was not produced by the enzymes outside liposomes at detectable levels. Thus, it was expected that ethanol produced by the system originated from conversion in the ARTCELLs. After a 40-h reaction using 150 mM lactate as a substrate, the ARTCELL system synthesized a significant amount of ethanol. The levels depended on the amount of enzyme encapsulated inside ARTCELLs (Figure 2D). The long reaction time for the ARTCELL system raised a concern of bacterial contamination in the external solution. However, no contamination was observed after the 40-h reaction (Supplementary Figure S3) because antibiotics were added in the external solution. The artificial ethanol synthesis pathway was NAD+-dependent, and did not function in the absence of NAD+ in bulk (Figure 3A, black bars). On the other hand, since NAD+ was contained in ARTCELLs and is not permeable through lipid membranes, the presence or absence of NAD+ outside was expected not to affect ethanol synthesis. Actually, although NAD+ was not added to the outer solution, the ARTCELL system successfully synthesized ethanol in the experiments described above. Supplementation of NAD+ in the outside solution brought no significant improvement of ethanol synthesis using the ARTCELL system, confirming that external NAD+ was not necessary (Figure 3A, gray bars). These results strongly suggested that conversion proceeded inside the ARTCELLs. Analyzing the effect of pH shift on ethanol synthesis further supported that the reaction proceeded within the ARTCELLs. Since the usage of lactate with low pH reduces enzyme function in bulk systems, the amount of synthesized ethanol drastically decreased in low-pH solutions (Figure 3B). This decrease was not observed in the case of the ARTCELL system (Figure 3C). We confirmed that addition of ARTCELLs did not change the pH of the solutions (see Methods). Although H+ passes through lipid membranes, its rate of transport is low because of its charge.34 Furthermore, consumption of lactate increases pH which may cancel the effect of H+ influx. This suggests that the pH gradient between the outside and inside of artificial cells was maintained. Considering these results and that ProK was present in the external solution, the ARTCELL system can C

DOI: 10.1021/acssynbio.7b00365 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology

became more apparent when the levels were compared with those of bulk enzyme conditions. In the case of bulk enzymes using the same levels of components as used in the ARTCELL system, ethanol synthesis after 40 h of reaction was 1.8 mM. On the other hand, the ARTCELL system synthesized 6.6 mM ethanol under isotonic conditions and 21.3 mM under 3-fold osmotic pressure conditions. Taken together, the ethanol synthesis rate by ARTCELL was 12-fold higher than that of the bulk system (Figure 5B). Finally, we demonstrated that the ARTCELL system could realize a product concentration of 100 mM upon cascade conversion. In the bulk system, the lactate-ethanol conversion system used in this study did not convert more than 100 mM. This difference was assumed to result from the depletion of lactate as a substrate and the increased pH during reactions as acidic lactate was consumed. In addition, low enzyme stability also contributed (Supplementary Figure S5). Therefore, we tested a long-term reaction using ARTCELL at 4 °C to stabilize proteins, and a 1/50 volume of 6 M lactate was added after 1, 2, and 5 days to supply substrates and maintain optimal pH. These treatments increased the final amount of ethanol produced. Quantification of ethanol concentration after 1, 2, 5, and 10 days showed that ethanol levels increased over time, with the levels exceeding 100 mM on average after 10 days (Figure 5C). These results resembled those of fermentation by microorganisms, proceeding slowly but showing efficient bioconversion. In this study, we demonstrated that ARTCELLs can function as a platform for cascade reactions to give a product at concentrations up to 100 mM (4.6 g/L). An artificial ethanol synthesis pathway entrapped in ARTCELLs converted lactate to ethanol, even under conditions where enzymes lose their function in bulk. By mimicking bacterial morphology with osmotic pressure, reaction rates were increased by approximately 3-fold, and synthesis efficiency was 12-fold higher than that of a bulk enzyme solution with the same levels of components. Furthermore, a low-temperature reaction with periodical substrate addition resulted in final product levels exceeding 100 mM. Moreover, the ethanol production pathway we used in this study can be reconstituted in a cell-free protein expression system (Supplementary Figure S6). Thus, rather than enzymes, genes using a cell-free protein expression system can be used in this system. These results showed that by using ARTCELLs, highly efficient conversion can be achieved even with a small amount of materials, similar to in living cells. Thus, we propose that the ARTCELL conversion system can be referred to as artificial cell fermentation. Artificial cell fermentation is beneficial for producing substances that are difficult to obtain from live cells because of their high toxicity, combinatorial synthesis in various ARTCELLs containing different enzymes, and hybrid reaction systems using a mixture of enzymes and chemical catalysts. The current bottlenecks of these application are (i) the limited permeability of lipid membranes (only passes