Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Artificial photosynthetic cell producing energy for protein synthesis

Artificial photosynthetic cell producing energy for protein synthesis ARTICLE https://doi.org/10.1038/s41467-019-09147-4 OPEN Artificial photosynthetic cell producing energy for protein synthesis 1 1 2,3 Samuel Berhanu , Takuya Ueda & Yutetsu Kuruma Attempts to construct an artificial cell have widened our understanding of living organisms. Many intracellular systems have been reconstructed by assembling molecules, however the mechanism to synthesize its own constituents by self-sufficient energy has to the best of our knowledge not been developed. Here, we combine a cell-free protein synthesis system and small proteoliposomes, which consist of purified ATP synthase and bacteriorhodopsin, inside a giant unilamellar vesicle to synthesize protein by the production of ATP by light. The photo-synthesized ATP is consumed as a substrate for transcription and as an energy for translation, eventually driving the synthesis of bacteriorhodopsin or constituent proteins of ATP synthase, the original essential components of the proteoliposome. The de novo photosynthesized bacteriorhodopsin and the parts of ATP synthase integrate into the artificial photosynthetic organelle and enhance its ATP photosynthetic activity through the positive feedback of the products. Our artificial photosynthetic cell system paves the way to construct an energetically independent artificial cell. Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, Bldg. FSB-401, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8562, Japan. Earth-Life Science Institute, Tokyo Institute of Technology, 2-12-1-IE-1, Ookayama, Meguro-ku, Tokyo 152- 8550, Japan. JST, PRESTO, Saitama 332-0012, Japan. Correspondence and requests for materials should be addressed to T.U. (email: ueda@edu.k.u-tokyo. ac.jp) or to Y.K. (email: kuruma@elsi.jp) NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ecent advances in synthetic biology allow us to challenge found that only 25% bR were maintaining the proper membrane whole reconstruction of cell from simple non-living orientation (Supplementary Fig. 3C). To improve this ratio, we 1–4 Rmolecules and redesigned minimal genome . Such did some modifications in the preparation method by changing attempts for the construction of artificial cell would lead not only the timing of bR addition (Supplementary Fig. 3A), i.e., empty to determining the necessary requirements for life phenomena liposomes were first roughly preformed and, then, the purified but also to developing as a biodevice toward industrial applica- bR was combined before completely removing the detergent. By tion . A cell-mimicking artificial cell is constructed by encapsu- this method, 70% bR was properly reconstructed in the PLs lating a cell-free protein synthesis system inside giant vesicle. (Supplementary Fig. 3C). The improvement of the membrane Cell-free system has been widely applied to researches in the field orientation faithfully reflected into the proton-pump activity of synthetic biology, and especially a reconstructed cell-free (Supplementary Fig. 3D). Since the efficiency of proton gradient system (PURE system) has been used as a basic technology for generation directly affects the F F activity, we employed this o 1 the artificial cell construction because all constituent enzymes are optimized method for all of the following experiments. known. This would be rather important when we try to recon- During the light illumination, we observed a decrease of proton struct self-reproducing artificial cells that have to synthesize all concentration at the outside of bR-PLs in proportion to bR their own components. Although several cellular functions or concentration (Fig. 1c), suggesting that the protons were phenomenon have been reconstructed so far in the artificial cell transported from the outside to inside of the bR-PL lumen 7–12 system , an energy self-supplying system for the internal (Supplementary Fig. 1A). In addition to the proton-pump protein synthesis has not been achieved. To develop the artificial activity, we also observed a rapid return of the proton cell into the energetically independent system, it is necessary to concentration when the illumination ceased. This indicates set up a circulating energy-consumption and production system proton leakage from the inside to outside of the bR-PL lumen. driven by an unlimited external physical or chemical energy The proton leak was accelerated when the lateral fluidity of the source. For this purpose, a biomimetic artificial organelle pro- bR-PL membranes was increased by temperature rise (Supple- ducing adenosine triphosphate (ATP) by collaborating ATP mentary Fig. 4). For the sake of inhibiting the leak through the synthase and bacteriorhodopsin is applicable as a rational energy membrane, we added 30% cholesterol into the lipid composition 13–18 18 20 generating system for artificial cells . Recently, Lee et al. of bR-PLs , which resulted in 30% reduction of the proton leak performed ATP synthesis using similar photosynthetic artificial (Supplementary Fig. 5). Thus, we kept this condition throughout organelle, where they demonstrated carbon fixation (in vitro) the study. and actin polymerization within giant unilamellar vesicle (GUV). Next, we estimated the membrane orientation of the This result evokes us to apply the artificial organelle into the reconstituted bR by evaluating the binding sensitivity of a artificial cell system, i.e., protein synthesis based on the photo- histidine-tag, which elongated at the C-terminus of recombinant synthesized ATP inside GUV. In this study, we performed ATP bR, to the Ni-NTA-conjugated magnet beads (Supplementary synthesis by light-driven artificial organelle inside GUV. Through Fig. 6). If the reconstructed bR was keeping the working optimization for the preparation method of proteoliposomes orientation, the C-terminus histidine-tag can bind to the magnet containing bacteriorhodopsin and ATP synthase, we succeeded to beads and be eluted in the elution fraction. The ratio of bR produce millimolar level ATP inside GUVs, wherein 4.6 µmol obtained in the elution fraction was normalized with the ratio of ATP per mg ATP synthase was produced after 6 h of illumina- control experiment in which bR was monodispersed by dissolving tion. By combining the artificial organelle and PURE system, we the PLs with detergent (Triton). In the control experiment, design and construct an artificial photosynthetic cell that pro- 91% bR was collected in the elution fraction, although that duces ATP for the internal protein synthesis. The produced ATP should be 100% theoretically (Supplementary Fig. 6). Considering was consumed as a substrate of messenger RNA (mRNA), or as this result, we calculated that 86% bR was reconstructed in an energy for aminoacylation of transfer RNA (tRNA) and for the working (outward C-terminus) orientation within the PL −1 phosphorylation of guanosine diphosphate (GDP) (Fig. 1a and membrane; i.e., Elu. Elu. 100%. It should be noted −Triton +Triton Supplementary Fig. 1). Additionally, we also demonstrated pho- that the opposite orienting bRs (inward C-terminus) pump tosynthesis of bacteriorhodopsin or a membrane portion of ATP protons from the inside to outside of the PLs. Thus, the net- synthase, which is the original component of the artificial orga- working ratio of the reconstituted bR is calculated as 72% nelle. Our artificial cell system enables the self-constitution of its (Supplementary Table 1). Taking account of the bR membrane own parts within a structure of positive feedback loop. orientation, the initial reaction rate of bR was calculated as −2.87 ± −1 −1 −1 −1 0.53 ΔpH min nmol or −0.11 ± 0.02 ΔpH min mg , mean ± S.D. (Fig. 1c and Supplementary Table 1). On the other hand, Results the net-working ratio of the reconstituted F F was 65.1% o 1 Construction of light-driven artificial organelle. Light-driven after the normalization as with bR (Supplementary Fig. 7 and artificial organelle was composed of two kinds of membrane Supplementary Table 1), and the initial reaction rate was 128 ± 3.2 −1 −1 −1 −1 proteins, bacteriorhodopsin (bR) and F-type ATP synthase ATP nmol min nmol or 223 ± 6.1 ATP nmol min mg (F F ). bR was isolated from a purple membrane of Halobacter- (Fig. 1d and Supplementary Table 1). The reverse function of o 1 ium salinarum by ultra-centrifugation with sucrose density F F , ATP-dependent proton-pump activity, was also detected o 1 gradient (Fig. 1b and Supplementary Fig. 2). F F of Bacillus PS3 (Supplementary Fig. 8), suggesting the full functionality of the o 1 was purified as recombinant protein from Escherichia coli cells reconstituted F F -PLs. o 1 (Fig. 1b). The isolated bR were reconstructed as bR-embedding To construct artificial organelle, we assembled purified bR and proteoliposomes (bR-PLs) for the measurement of light- F F to form bRF F -PLs. We prepared PLs in various proportion o 1 o 1 dependent proton-pump activity. The size of bR-PLs were of bR against F F and illuminated with visible light passing a o 1 mostly 100–200 nm as diameter. We used phosphatidylcholine 500 nm long-pass filter. The amount of produced ATP was extract from soybean to form PLs which are stable in the reaction measured by means of luciferin and luciferase. The highest ATP mixture of PURE system and also maintain the F F activity . photosynthesis was obtained in the case of 176 µM bR and 1 µM o 1 The formation of bR-PLs was carried out by reducing the F F . This means that approximately 0.6 × 10 ATP was produced o 1 detergent concentration in the mixture of lipids and purified by a single bRF F -PL within 4 h of illumination (Fig. 1e). The o 1 protein according to the previous report ; however, we have maximum turnover number for ATP synthesis in the initial 5 min 2 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 1 F F -PLs o 1 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE a bc bR F F o 1 kDa 250 75 150 α H β 20 μM bR+FCCP H DNA ADP ATP H 0 GDP ATP hv 37 γ synthase –0.5 ATP + ADP GTP H + mRNA 25 + H + + a –1.0 H H 20 25 H b 15 –1.5 + ε Protein Bacteriorhodopsin –2.0 10 10 0 120 240 360 480 600 Time (s) d ef ×10 ×10 In GUVs 600 1.0 5E+4 1.6 176:1 Ratio of bR:F F (μM) 30 nM F F o 1 o 1 4E+4 *** 1.4 used for PLs preparation 3E+4 140:1 1.2 0.8 70:1 2E+4 1.0 400 1E+4 35:1 *** 20 nM F F 176:1 o 1 0.8 0.6 012 345 Time (min) 0.6 140:1 0.4 70:1 0.4 0.2 35:1 10 nM F F o 1 30 nM F F +FCCP GUVs + ++ – 0.2 o 1 –– 0 +FCCP PKin + – 0 20 40 60 80 100 120 Dark PKout ++ + Time (s) 0 Light ++ – + Time (h) Fig. 1 Light-driven adenosine triphosphate (ATP) synthesis by artificial organelle. a Schematics of the artificial photosynthetic cell encapsulating artificial organelle, which consists of bacteriorhodopsin (bR) and F F -ATP synthase (F F ). Synthesized ATP are consumed as substrates for messenger RNA o 1 o 1 (mRNA) (➀), as energy for phosphorylation of guanosine diphosphate (GDP) (➁) or as energy for aminoacylation of transfer RNA (tRNA) (➂). b Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of purified bR and F F . The positions of molecular markers and F F component o 1 o 1 proteins are indicated beside the gels. c Light-driven proton-pump activity of bR reconstituted in a proteoliposome (PL). Proton-pump activity of bR was measured by monitoring the proton concentration at the outside of bR-PLs where fluorescent proton-sensor ACMA (9-amino-6-chloro-2-methoxy acridine) was added. We defined as ΔpH = pH (original, outside) − pH (after illumination, outside). The ΔpH caused by bR activity was measured with the various bR concentrations as indicated. White and gray areas indicate light ON and OFF, respectively. An uncoupler, FCCP (carbonyl cyanide 4- (trifluoromethoxy) phenylhydrazone), was used as a control experiment. d ATP synthesis activity of F F reconstituted as F F -PLs. ATP synthesis reactions o 1 o 1 were initiated by adding F F -PLs at 30 s with various F F concentrations, as indicated. The synthesized ATP was measured by means of luciferin and o 1 o 1 luciferase (see Methods section for the experiment details). FCCP was used for control. e Light-driven ATP synthesis. The amount of the photosynthesized ATP by bRF F -PLs, which was constituted in various proportions of bR against F F , were measured by luciferin and luciferase. FCCP and dark conditions o 1 o 1 were also performed as controls. The inset indicates initial rate of the each PL. f Light-driven ATP synthesis inside giant unilamellar vesicle (GUV). bRF F - o 1 PLs were illuminated inside GUVs in the presence or absence of proteinase K (PK) that degrades the F F . The in vitro experiment was also performed for o 1 comparison. ***p < 0.001. P values were from two-sided t-test. All experiments were repeated at least three times, and their mean values and standard deviations (S.D.) are shown. Source data are provided as a Source Data file −1 was 8.3 ± 0.3 s in the case of 176 µM bR and 1 µM F F . This synthase. The efficiency of ATP production in GUVs was o 1 was almost double compared to the previous report . Here, in a roughly one-third that of the in vitro system, perhaps caused single PL, 3560 of the working bRs drive 18 F F (Supplementary by lower light intensity inside a GUV. Since our artificial o 1 Table 1). In all cases, we used 10 mM NaN to inhibit the reverse organelle can produce ATP inside GUV at the comparable (ATPase) activity of F F . We found that the ATP production concentration as a real living cell, we proceeded to design and o 1 plateaued when the illumination was higher than 10 mW per cm construct the photosynthetic artificial cell system that synthesize (Supplementary Fig. 9). protein by light. The same reaction was also performed inside GUVs in which about 1.1 × 10 bRF F -PLs are contained in a 10 µm diameter o 1 GUV. After 6 h of illumination, we observed photosynthesized Light-driven protein synthesis inside the artificial cell.We ATP from the inside of the GUVs (Fig. 1f), where 1.8 mM performed green fluorescent protein (GFP) synthesis inside ATP was produced in a single GUV (Supplementary Table 1). GUVs by means of the photosynthesized ATP to demonstrate This represents that 4.6 µmol ATP was produced per mg ATP that the constructed artificial organelle works in an artificial cell NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 3 Synthesized ATP (nM) Produced ATP/bRF F -PL o 1 ΔpH at the bR-PLs exterior Produced ATP/bRF F -PL o 1 In vitro ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 system. For this purpose, we combined bRF F -PLs with the confirmed in vitro (without GUVs) (Fig. 2b and Supplementary o 1 PURE system which is a cell-free protein synthesis system. The Fig. 13) and synchronized with timing of the ATP photosynthesis PURE system is reconstituted from purified translation factors , (Supplementary Fig. 14). These results indicate that GFP synth- and therefore we can customize the component factors suited esis inside the GUVs was driven by the photosynthesized for the designed artificial system. The PURE system was modified ATP (Supplementary Fig. 1B). We found that 50–60% of the as shown in Supplementary Table 2 to allow the photosynthesized GUVs emitted more fluorescence than the nonilluminated con- ATP be specifically used for the aminoacylation of tRNA (Sup- trol GUVs (Fig. 2c) by flow-cytometry analysis. We also found plementary Fig. 10), and supplied with a mRNA encoding a certain percentage of GUVs were not showing fluorescent GFP together with bRF F -PLs and NaN . NaN did not inhibit even when illuminated. Although the definitive cause is not o 1 3 3 protein synthesis at concentrations below 50 mM (Supplementary unclear, it has been reported that the encapsulation efficiency Fig. 11). The prepared reaction mixture was encapsulated of PLs is affected by the size of PLs; i.e., less than 35% GUVs inside GUVs, and illuminated to induce protein synthesis. A large can encapsulate the PLs when their size are over 200 nm majority of the GUV population appeared in a range of 10–20 µm (diameter) . Additionally, we cannot deny the possibility that as diameter (n = 200) (Supplementary Fig. 12). After 6 h, we inactivity of the internal artificial organelle by the fusion of observed the fluorescence of internally synthesized GFP by bRF F -PLs and GUV membranes is limiting the successful o 1 confocal microscopy (Fig. 2a). This GFP synthesis was also artificial cell formation. ab c Light Dark 200K bRF F -PL + + o 1 100K bR-PL – – – – F F -PL o 1 ADP + + ATP – – Light + – 10K GFP 300 300 25 kDa 4K 10 100 1K 10K 0 0 0 180 0 160 Pixel Pixel GFP fluorescent (A.U.) d e h 200K bRF F -PL + +++ + + o 1 100K NDK – – ++ + – ++–– – – GTP 300 GDP – –– ++ + 0 200 Pixel ++ + + – + ADP Light ++ + + + + 10K 12 3 4 5 6 GFP 4K 25 kDa 0 10 100 1K 10K 0 200 Pixel GFP fluorescent (A.U.) fg Light Dark bRF F -PL ++ + – + o 1 GFP-DNA ++ + + – T7RNAP ++ – + + ATP – – – – – 0 200 Pixel ++ – + + Light GFP 25 kDa 0 250 Pixel Fig. 2 Protein synthesis inside giant unilamellar vesicle (GUV) driven by light. Green fluorescent protein (GFP) was synthesized from its messenger RNA (mRNA) (a–e) or DNA (f–h) inside light illuminated GUV (a, c, e–h) or in vitro (b, d). GFP was synthesized inside GUV (a) or in vitro (the PURE system) (b) in which the photosynthesized adenosine triphosphate (ATP) was consumed for the aminoacylation of transfer RNA (tRNA). The insets in a, e and g indicate plot profile of green and red colors on the thin yellow line. c Flow-cytometric analysis of the GUVs of a. The illuminated GUVs are shown as green, whereas the GUVs incubated in the dark are shown as black. The X- and Y-axes represent the fluorescent intensity and the area of forward scattering, respectively. d GFP synthesis coupled with guanosine 5’-triphosphate (GTP) generation. GFP was synthesized in the PURE system with or without nucleoside-diphosphate kinase (NDK), GTP, guanosine diphosphate (GDP) and adenosine 5’-diphosphate (ADP). e The same reactions as in lanes 4 and 6 of d were performed inside GUVs as indicated as NDK+ and NDK−, respectively. f GFP synthesis from its DNA. A gene of whole GFP was introduced in the PURE system with or without bRF F -PLs, T7 RNA polymerase (T7RNAP), ATP and light. g A small part of GFP (GFP11: 15 amino acids) o 1 was synthesized from its encoding DNA inside GUVs containing T7RNAP, another large part of GFP (GFP1-10) purified form E. coli cells, and the PURE system lacking NDK. h The same GUVs of g were analyzed by flow-cytometer as in d. The synthesized GFP in b, d, and f were labeled with [ S] methionine. Scale bar: 10 µm. Source data are provided as a Source Data file 4 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications Int. Int. NDK– NDK+ Int. Int. Int. FSC-A FSC-A Int. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE Next, we omitted GTP from the reaction mixture but the GUVs were not encapsulating liposomes, many larger-size introduced GDP and nucleoside-diphosphate kinases (NDKs), puncta appeared, suggesting aggregation of the synthesized bR- which allows the photosynthesized ATP to be consumed for GFP (Supplementary Fig. 17). These results imply that the synthesis of GTP that is a direct energy source of translation bR-GFP synthesized in the GUVs (Fig. 3b) localized onto the (Supplementary Fig. 10). The results showed that synthesized GFP internal PL membrane avoiding protein aggregation. The mem- was clearly detected by the sodium dodecyl sulfate–polyacrylamide brane localization of bR was further confirmed in vitro by flota- gel electrophoresis (SDS-PAGE) analysis when adenosine 5’- tion assay. When bR was synthesized in a standard PURE system diphosphate (ADP), GDP and NDK were added (Fig. 2d, lane 4). in the presence of liposomes, we found that the synthesized We performed the same reaction inside GUVs and observed bR appeared in the liposome fractions (Fig. 3c) after ultra- the fluorescence emission from the GUV lumen (Fig. 2e), centrifugation with a sucrose cushion, whereas almost all bR suggesting that the photosynthesized energy was consumed not appeared in the pellet fraction when liposomes were omitted. only for aminoacylation of tRNAs but also directly for translation This result directly shows the membrane localization of the de inside GUVs. novo bR onto the PL membrane. Additionally, the membrane In real cells, ATP is consumed not only as energy but also as a localized bR showed the proton-pump activity in response to substrate of transcription. To build up this, we performed a the duration of protein synthesis reaction (Fig. 3d). Here, 11, 61, transcription-and-translation coupled reaction in the artificial 124 or 233 bRs per one liposome were synthesized at the time photosynthetic cell system. When T7 RNA polymerase and of 10, 30, 60 or 180 min reaction, respectively (Supplementary template DNA coding-GFP were introduced into the PURE Fig. 18). These results lead us to conclude that the de novo system, photosynthesis of GFP was clearly detected by SDS- photosynthesized bRs spontaneously localized onto the internal PAGE analysis (Fig. 2f). However, we could not detect significant PL membrane and may have increased the proton-pump fluorescence by microscopy observation and flow-cytometer activity there. analysis when we performed inside GUVs. This is because the If the de novo photosynthesized bRs are functionable on the PL synthesized GFP level was lower than the detection limit. To membrane, the ATP production rate of PL should be enhanced overcome this problem, we applied the split-GFP method according to the increase of the number of bR per PL. To confirm developed by Cabantous et al. , i.e., GFP is split into two parts: this, we measured ATP concentration in the PURE system a small peptide (GFP11) and another large partner protein reaction mixture during the photosynthesis of de novo bR (GFP1-10). The fluorescence of GFP1-10 was restored by (bR ). In this experiment, we used the PLs consisting of a low wt incorporating GFP11 (Supplementary Fig. 15). Although the concentration bR (i.e. 5 µM bR) to emphasize the effects of intensity was rather weak, we observed the emission of GFP the de novo photosynthesized bR. The effect of the de novo fluorescence from the GUVs when GFP11 was photosynthesized photosynthesized bR was determined by comparing to the control from the template DNA (Fig. 2g). In this reaction, we experiment synthesizing a mutant bR (bR ) which does not mut encapsulated the PURE system modified for transcription-and- have any proton-pomp activity (Supplementary Fig. 19), and translation reaction (Supplementary Table 2), and the purified therefore the ATP production rate of the PLs containing bR is mut GFP1-10. The successful photosynthesis of GFP11 was also constant throughout the bR photosynthesis. In the case of bR wt confirmed in an in vitro reaction (Supplementary Fig. 16). By photosynthesis, the ATP concentration was higher than that in flow-cytometry analysis, we found that about 15% of the total the case of bR photosynthesizing in all three independent mut GUVs emitted significant fluorescence as a consequence of measurements (Supplementary Fig. 20), especially after 10 min reaction. This is consistent with the result of proton-pump transcription and translation inside (Fig. 2h). These results show that the photosynthesized ATP was consumed both as the activity of the bR synthesized in PURE system (Fig. 3d). substrates for mRNA transcription and as the energy for protein Here, the difference in the ATP concentration between bR wt translation, just as in real cells. and bR photosynthesizing reactions represents the effect of mut de novo photosynthesized bR . The ATP concentration in the wt bR -photosynthesizing reaction was approximately 1.5-fold wt Self-production of the artificial organelle components. The two higher than that of bR (Supplementary Fig. 20). It should be mut kinds of component proteins of the artificial organelle produced noted that the obtained ATP concentration indicates the net of ATP, and the resulting ATP drove protein synthesis. To test the photosynthesized ATP minus the consumed ATP for the whether our artificial photosynthetic cell system can synthesize protein synthesis. The synthesis rate of the bR and bR was wt mut the component proteins of its own artificial organelle, we tried adjusted to be the same by regulating the amount of template to photosynthesize bR, as well as F F . In this reaction, we used mRNA (Supplementary Fig. 21), and thus the ATP consumption o 1 the translation-only PURE system (see For mRNA start in Sup- rates were equal in both cases. To directly compare all three plementary Table 2). We expected that the newly photo- measurements, we normalized each obtained result with the synthesized de novo bRs localize onto the bRF F -PL membrane ATP per PL value at the endpoint time (120 min) of the de novo o 1 and increase ATP photosynthesis activity of the artificial orga- bR -photosynthesizing reaction. The ratio of the increased mut nelle as a consequence of activity enhancement in the proton artificial organelle activity is shown in Fig. 3e. We also confirmed gradient generation of the bRF F -PL (Fig. 3a and Supplementary the photosynthesized bR by SDS-PAGE analysis (Fig. 3f) in which o 1 −1 −1 Fig. 1E and F). The fluorescence of the synthesized bR, which the photosynthesis rate was 2.5 nmol ml min . After 2 h fused with GFP (bR-GFP), was mostly homogeneously observed of reaction, the number of working bRs per PL increased from inside the GUV lumen but not on the GUV membrane (Fig. 3b), 100 working bRs per PL (original) to 110 bRs per PL (after indicating that the synthesized bR-GFP localized onto the internal photosynthesis). These series of results indicate that the ATP PL membrane. This directed membrane localization is controlled production rate was enhanced during the photosynthesis of de by means of cholesterol which inhibits spontaneous membrane novo bR because the ability of proton gradient generation was wt integration of protein . We added 40% (mol%) cholesterol in the improved by increasing the number of functional bRs on a PL. lipid composition of GUV membrane but not in the internal Finally, we challenged to photosynthesize de novo F F in vitro o 1 PL membrane. When bR-GFP was synthesized inside GUVs and to observe the enhancement of ATP production activity containing liposomes, the same homogeneous fluorescent dis- of the resulting PLs. Unlike bR, F F consists of eight kinds of o 1 tribution was observed within the GUV lumen. In contrast, when subunit proteins. Thus, we first try to synthesize these eight kinds NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 a bc Light Dark Liposome Top F1 F2 F3 Bot. aa tRNA 25 kDa ATP ADP %Frac. 42 28 19 10 1 ARS 25 kDa aa-tRNA %Frac. 00 1 1 98 Min 0 30 60 120 Protein mRNA Rbs De novo 25 kDa bR d eh g 0.2 2.0 ×10 ** *** 4.0 1.5 F synth. with a o wt F synth. with a o mut 3.0 1.5 –0.2 1.0 2.0 –0.4 1.0 1.0 bR synth. –0.6 time 0.5 0 min 030 60 90 120 0.5 10 min –0.8 30 min Time (min) 60 min 180 min 0 i –1.0 Min 0 60 120 180 0 120 240 360 480 600 De novo F -a Time (s) o 25 kDa De novo F -b Fig. 3 Self-constituting protein synthesis in artificial photosynthetic cells. a Schematics of self-constituting protein synthesis. The numbers indicate the order of reactions; ➀: adenosine triphosphate (ATP) synthesis, ➁: aminoacylation of transfer RNA (tRNA) by aminoacyl-tRNA synthetase (ARS), ➂: translation by ribosomes (Rbs), ➃: de novo bacteriorhodopsin (bR) synthesis, and ➄: de novo F synthesis. b Light-induced bR-GFP synthesis in giant unilamellar vesicles (GUVs). Bar: 10 µm. c Membrane localization of bR. The bRs synthesized in the PURE system with or without liposomes were fractionated by ultra-centrifugation with sucrose cushion. The percentages of bR in each fraction (%Frac.) are indicated at the bottom of the gels. d Proton- pump activity of bR synthesized in the PURE system. The measurement was performed as in Fig. 1c. The reaction times of protein synthesis are indicated by different colors. The white and gray areas represent light ON and OFF, respectively. e Enhanced artificial organelle by de novo bR. Wild-type (bR ) wt or mutant (bR ) bRs were photosynthesized in the PURE system containing bRF F -PLs. The ATP concentrations at the time 2 h was measured and mut o 1 converted into ATP per proteoliposome (PL). The value of the de novo bR -containing PL was normalized to that of the de novo bR -containing PL. wt mut ***P < 0.001. f Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the de novo photosynthesized bR. g Light-driven ATP synthesis by PLs consist of cell-free synthesized F . Wild-type (a ) or mutant (a ) a-subunit protein was synthesized together with b- and c-subunits in o wt mut the PURE system containing purified F and bR-PLs. The measured ATP concentrations were converted into the produced ATP per PL. h Enhanced artificial organelle by de novo F . a or a was photosynthesized together with b- and c-subunits in the presence of purified F and bRF F -PLs. The ATP o wt mut 1 o 1 concentrations at the time 3 h was measured and converted into ATP per PL. The value of the de novo a -containing PL was normalized by that of the wt de novo a -containing PL. **P < 0.01. i SDS-PAGE analysis of the de novo photosynthesized F . P values were from two-side t-test. All experiments mut o were performed at least three times and their means and S.D. are shown. Source data are provided as a Source Data file of proteins by adding their corresponding template DNAs alanine , was synthesized instead of the wild-type a. This further into a standard PURE system supplemented with liposomes. supports that the cell-free synthesized F formed functional F F o o 1 However, unfortunately, we could not detect a significant activity onto the PL membrane and synthesized ATP, and thus we next of the F F due to low yields. We next synthesized only three tried to photosynthesize F and observed the enhancement of o 1 o component proteins of F , a-, b- and c-subunits, in the presence ATP photosynthesis activity in the resulting bRF F -PLs. The o o 1 of purified F and bR-PLs. After the reaction, the resulting photosynthesis reaction of F was performed in the translation 1 o bRF F -PLs were isolated from the reaction mixture and only system. The a-, b- and c-subunit proteins form the complex o 1 illuminated with supplying ADP. The result shows that ATP structure of F in the stoichiometry of 1, 2 and 10, respectively. photosynthesis of the PLs was detected in proportion to In order to find the best proportion of these three templates for illumination time, when wild-type a-subunit (a ) protein was obtaining the highest F F activity, we tested various proportions wt o 1 synthesized (Fig. 3g) with other b- and c-subunit proteins. This of the template DNA mix, at first. The multi-protein synthesis indicates the cell-free synthesized F localized onto the bR-PL for F was performed in the presence of liposomes and purified o o membrane and photosynthesized ATP by co-working with F . We detected the highest F F activity when 4, 2 and 10 nM 1 o 1 bR. Contrary, we could not detect any activity when a mutant template DNA of a-, b- and c-subunit, respectively, were added a-subunit (a ), which has an amino acid substitution at R169 to (Supplementary Fig. 22). Following this, we photosynthesized de mut 6 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications m = 260, R = 0.99 wt ΔpH Relative activity bR mut bR wt Relative activity Fo-a mut Fo-a wt Produced ATP/bRF F -PL o 1 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE novo F component proteins in the presence of bRF F -PLs minimal number of enzymes and molecules since we used a o o 1 (0.3 µM F F and 140 µM bR) and purified F in the PURE system reconstructed artificial organelle and cell-free protein synthesis o 1 1 suited for mRNA start (Supplementary Table 2 and Supplemen- system . The functional significance in our artificial cell would −1 −1 tary Fig. 1F). The F photosynthesis rate was 50 fmol ml min accelerate the researches of artificial cell (or synthetic cell) in and reached to the plateau level at 7 h (Supplementary Fig. 23). the field of synthetic biology, as well as the development of a After the F photosynthesis reaction, PLs were isolated from biodevice sensing light and promoting protein and RNA synth- the reaction mixture and illuminated in the presence of ADP. In esis. For example, our artificial cell technique would be applicable order to distinguish the effect of de novo photosynthesized F ,we into the study of drug delivery that can control spatiotemporal also synthesized a instead of a and compared them, same as production of aptamer or single chain Fv within a vesicle capsule. mut wt in the case of de novo bR photosynthesis. The PL-containing More promising application of the artificial organelle is to use de novo F -a (PLs-a ) showed higher ATP photosynthetic as the phosphate recycling system in cell-free system. The o wt wt activity than the PL-containing de novo F -a (PLs-a ) in all current cell-free system is using creatine phosphate as a primary o mut mut three independent measurements (Supplementary Fig. 24A). The energy source; however, since this is unidirectional reaction, enhancement of ATP photosynthesis rate per PL was 1.38-fold. free phosphates accumulate in the system as the reaction goes Since we recovered the PLs-a and PLs-a from the reaction on. Our artificial organelle can avoid this problem by recharge wt mut mixtures, the amount of PLs analyzed was same in both samples the free phosphate onto ADP after the ATP consumption. (Supplementary Fig. 24B), and therefore the difference in Artificial cells have been employed as a model of protocell activities of PLs-a and PLs-a is thought to be reflecting the or primordial cell, which are thought to have existed before wt mut 2,18,25–28 enhanced activity by the de novo photosynthesized F . We also modern cells, in the study of origin of life . Especially, confirmed the same amount of F component proteins were how the primordial cell gained the ability to produce an energy photosynthesized in both samples (Supplementary Fig. 24C). to drive primitive metabolism is a big argument . The genes of Based on these results, we analyzed the enhanced ATP ATP synthase are highly conserved beyond the species and have photosynthesis rate in the resulting PLs. After 7 h of photosynth- been thought to exist from early stage of life . However, what esis, the net concentration of de novo photosynthesized F mechanism generated a proton gradient to drive ATP synthase within the reaction mixture was 20 nM (Supplementary Fig. 23). before the completion of the complicated electron transfer system Since 18 nM PLs were contained in the reaction mixture, is still unknown. Our work demonstrated that a simple bio-sys- statistically one de novo F was assigned to one PL. This reflected tem, which consists of two kinds of membrane proteins, is able into the enhanced ATP photosynthesis rate of the PL as 101.4 to supply sufficient energy for operating gene expression inside −1 −1 ATP PL min (Supplementary Fig. 24A), which represents a microcompartment. Thus, we think that primordial cells using −1 −1 −1 101.4 ATP F F min (turnover number: 1.7 s ). The sunlight as a primal energy source could have existed in the o 1 calculated result consistent to the specific activity of F F early stage of life before evolving into an autotrophic modern cell o 1 −1 −1 reconstructed into F F -PLs, 118±3.2 nmol ATP min nmol system. We believe the attempts to construct living artificial o 1 (Supplementary Table 1). On the other hand, when PLs-a was cell will reveal the boundary state of the transition from non- mut photosynthesized, the ATP photosynthesis rate showed 268 ATP living to living matters that actually happened in the early Earth −1 −1 PL min . Since five working-F F were contained in one PL, it environment. o 1 −1 −1 can be converted as about 50 ATP F F min that is lower o 1 than the de novo F activity which we cannot explain well. Methods Overall, the obtained result of the F photosynthesis seems o Materials. All reagents utilized in experiments were of the highest purity and reasonable. As we showed above, recursive production of F grade. These include POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine), cholesterol, PEG2000PE (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N- portion of F F was definitely observed, though it did not enhance o 1 [methoxy (polyethylene glycol)-2000] (ammonium salt)), soybean phosphati- exponentially. Although the photosynthesis level is still low, dylcholine (SoyPC) extract; liquid paraffin (Wako); pH gradient sensitive fluor- we engineered a self-constituting protein synthesis positive ophore ACMA (9-amino-6-chloro-2-methoxy acridine), protonophore FCCP feedback loop in the artificial photosynthetic cells. (carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone), ADP (adenosine 1 5 5’-diphosphate monopotassium salt), P ,P -di(adenosine-5′) pentaphosphate pentasodium salt (AP5A), potassium-specific ionophore valinomycin, Triton X-100, all-trans retinal (ATR) (Sigma); ATP (adenosine 5’-triphosphate), GTP Discussion (guanosine 5’-triphosphate), CTP (cytidine 5’-triphosphate), UTP (uridine We show that our artificial cell system containing the artificial 5’-triphosphate) (geneACT inc); OG (octyl β-D-glucopyranoside); DDM (n-dode- cyl-β-D-maltoside) (Dojindo); Tween-20 (Calbiochem); sodium cholate (Wako); organelle was able to first transduce light energy into an elec- and ATP bioluminescence assay kit CLS II (Roche). trochemical potential, and then convert into the chemical energy of ATP inside GUV. The produced ATP was consumed for the Isolation of purple membrane. Purple membrane patches containing firmly reaction of aminoacylation of tRNA as well as for the generation packed two-dimensional crystals of bR were isolated from Halobacterium sali- of GTP which was eventually consumed for translation. We also narum R1 following the previous protocol with slight modification. In brief, the showed that the produced ATP was converted into a mRNA that H. salinarum colonies were cultured in 6 L tryptone media containing 4 M NaCl subsequently translated into a part of GFP. The biochemical under high oxygen tension and continuous illumination with 200 W LED lamp for almost 192 h at 39 °C. The cells were then collected and resuspended in basal salt reactions performed in our artificial cell system mimic that is −1 in the presence of 17 µg ml of DNaseI (Sigma-Aldrich). This was followed by occurring in real living cells. Finally, we performed the photo- overnight dialysis against 100 mM NaCl at 4 °C. The translucent red lysate was synthesis of bR and F . The photosynthesized de novo bR loca- centrifuged at 45,000 × g for 60 min, and the precipitate was washed 6 times with lized onto the membrane of internal artificial organelle and 100 mM NaCl solution and distilled water. Eventually, the purple membrane enhanced the activity of ATP production, indicating the func- precipitate was overlaid on a 30–50% linear sucrose gradient and centrifuged at 100,000 × g for 17 h using SW28 rotor. At the end of the centrifugation, the purple tional engagement of protein synthesis and energy production band was collected. The sucrose solution was then removed by centrifuging the reactions. Because bR is the original compound of the artificial purple suspension at 45,000 × g for 1 h. The purple membrane sediment was organelle, we demonstrated that the artificial cell synthesized its resuspended and later stored in 50 mM Tris-HCl (pH 7.6), 150 mM NaCl and own part in a positive feedback loop. Furthermore, another 10% glycerol. Stock concentration of the purified bR was 450 µM. membrane-embedding component, F , was photosynthesized and its functional contribution in ATP photosynthesis was detected. It Overexpression and purification of bR from E. coli. C43(DE3) E. coli strain was should be noted that all these reactions were reconstructed with a transformed with pET21c-bR construct (see the DNA sequence in Supplementary NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 Table 5). The resulting recombinant bR was bearing hexa-his-tag at its C-terminus. Tris-HCl (pH 8.0), 150 mM NaCl and 10% glycerol. Stock concentration of the The bacterial culture was set in 6 L 2× YT media at 37 °C with shaking. The purified GFP1-10 was 100 µM. incubation continued until OD 0.6–0.7 and the culture media were supplemented On the other hand, the template DNA for in vitro or in vesicle synthesis of the with 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and 10 µM ATR. The smaller partner (GFP11) was prepared by PCR using P18 and P19 primers, where incubation was continued for 1 h at 30 °C and then for 3 h at 37 °C. The cells were P17 oligo was used as the template. The amplified DNA was cloned into pET29a collected and washed. The collected cells were disrupted by 5 passes of French press vector using NdeI and EcoRI. The resulting construct was used as a template homogenizer at 550 bar in a lysis buffer (50 mM MES (pH 6.0), 1 mM EDTA, DNA for the second PCR using T7 promoter and terminator primers (P22 and 300 mM NaCl, proteinase inhibitor cocktail). The membrane fractions were P23) to prepare the template DNA for GFP11 synthesis in the PURE system collected by centrifuging the pure lysate at 234,788 × g for 1 h. The collected (Supplementary Fig. 15). The further PCR for making the template DNA for membrane fraction was solubilized overnight at 4 °C in a buffer containing 50 mM GFP11 photosynthesis in vesicle (Fig. 2g, h) was carried out using P24 and P20 MES (pH 6.0), 300 mM NaCl, 5 mM imidazole and 1.5% DDM. The solubilized primers, and the resulting template DNA omits the non-translational sequence bR was purified in Ni-affinity chromatography (pre-equilibrated in a buffer con- after the stoop codon. taining 50 mM MES (pH 6.0), 300 mM NaCl, 0.2% DDM and 40 mM imidazole). The protein was eluted in a linear imidazole gradient of 40–300 mM. Further purification was done using Mono-Q column in the presence of 0.2% DDM. Reconstruction of PLs. The reconstitution of PLs with either bR or F F or the o 1 Finally, bR was eluted with linear NaCl gradient of 10–300 mM. Stock con- co-reconstitution of bRF F -PLs has been performed based on complete detergent o 1 10,19,34 centration of the purified recombinant bR was 1 µM. solubilization of liposomes following previous literature by the incorporation of the necessary modifications. Buffer PA6-5 was used as a reconstitution buffer unless otherwise indicated. First, lipid powder was suspended in buffer PA6-5 (pH 7.3), composed of 10 mM HEPES, 3 mM MgCl , 10 mM NaH PO and Expression and purification of recombinant F F and F . Thermophilic 2 2 4 o 1 1 200 mM sucrose, at a concentration of 41.3 mM. Later, the lipid suspension was Bacillus PS3 F F -ATP synthase (F F ) was overexpressed in DK8 E. coli strain o 1 o 1 (unc minus) carrying pTR19ASDSεΔc construct . The culture was made in 3 L completely solubilized by 6% octyl β-D-glucopyranoside (W/V) for about 1 h at room temperature. Then, the detergent was removed by first-round addition of 2× YT media for 21 h at 37 °C. The purification of F F was undertaken in o 1 accordance with previous work with modification. The cells were disrupted by a of 200 mg of pre-equilibrated SM2 Bio-Beads (Bio-Rad) and incubated at room temperature for 30 min. This was followed by the addition of bR (as a purple tip-sonication in a buffer containing 10 mM HEPES-KOH (pH 7.5), 5 mM MgCl , 10% Glycerol and 28 mM β-mercaptoethanol. The cell debris was removed by membrane), F F or both bR and F F at a given final concentration. The o 1 o 1 incubation was continued for additional 30 min, before the second-round centrifuging the lysate at 5500 × g for 30 min at 4 °C accompanied by collecting the membrane faction containing F F complex by ultra-centrifugation at addition of 300 mg of Bio-Beads by rotation mixing. After adding the second- o 1 round Bio-Beads, the proteoliposome was left mixing at room temperature for 225,000 × g for 1 h. The membrane fraction was homogenized in buffer I (pH 7.5) (10 mM HEPES-KOH, 5 mM MgCl , 2% Triton X-100, 0.5% cholate, 10% 90–120 min. Then, the turbid upper fraction of the proteoliposome was collected and stored at −80 °C until use. glycerol and protease inhibitor cocktail) and incubated at 30 °C for 30 min with mild shaking. The sample was then centrifuged at 311,000 × g for 20 min at 30 °C and supernatant was collected. To the supernatant, 70 ml of buffer II (pH 7.5) (20 mM potassium phosphate, 100 mM KCl, 24 mM imidazole and protease Light-dependent proton-pump activity of bR-PLs. For assaying proton pumping activity of bR, 800 µl of 1 R buffer (10 mM HEPES-KOH (pH 7.3) and 3 mM inhibitor cocktail) was added. This was later applied to Ni-NTA (Qiagen) pre- −1 equilibrated with buffer A (pH 7.5) (20 mM potassium phosphate, 100 mM KCl, MgCl ), pre-warmed at 37 °C, was supplemented with 30 µl bR-PL and 0.5 µg ml ACMA. Thereafter, the fluorescence trace of ACMA was measured every 0.2 s 20 mM imidazole and 0.15% DDM). Finally, the F F was eluted with buffer A o 1 (pH 7.5) containing 200 mM imidazole. The homogeneity and purity of the protein at the excitation and emission wavelength of 410 nm and 480 nm, respectively, was judged by 10–20% gradient SDS-PAGE. Finally, pure and homogenous using JASCO FP6500 spectrofluorometer that is customized by attaching a light fractions were collected and buffer was exchanged with a buffer containing 20 mM source (Tokina, Techno light KTS-150RSV, equipped with 15 V 150 W lamp KPi (pH 7.6), 100 mM NaCl and 0.15% DDM. This was concentrated eventually and 520 nm long-pass filter) at the lid. For the bR synthesized in the PURE system, by 50 kDa Amicon. Stock concentration of the purified F F was 99 µM. an assay buffer composed of 10 mM HEPES-KOH (pH 7.5), 5 mM MgCl and o 1 2 The F complex was purified following the previous literature by Suzuki 100 mM KCl was used. The proton-pump activity of bR was initiated by illumi- et al. . Briefly, E. coli cells expressing F F were disrupted and the cytosol fraction nating the sample with the light source. o 1 was obtained by centrifugation. The supernatant was incubated at 67 °C for 15 min, and the aggregated E. coli proteins were removed by centrifugation. The resulting yellow supernatant was subjected to a Ni-NTA column. The eluted F was Ultrapurification of ADP. The purchased ADP-monopotassium salt powder was incubated for 60 min at 25 °C, then ammonium sulfate was added to the solution. first dissolved with 10 mM HEPES-KOH (pH 7.3) at the concentration of 400 The resulting solution was applied to a phenyl-Toyopearl column. After washing, µmol, then applied to pre-equilibrated mono-Q column and eluted with linear salt the column was eluted with a linear reverse gradient of ammonium sulfate (1–0 M), gradient of 0–300 mM NaCl. The peak fraction for ADP was collected and then fractions containing F were collected and precipitated with ammonium lyophilized (using Taitec VD-800F freeze-dryer), and suspended with 10 mM sulfate. The F was further purified with a Superdex 200HR column. The purified HEPES-KOH (pH 7.3) at the stock concentration of 80 mM. protein was stored at −80 °C. The stock concentration of F was 5.7 µM. Light-driven ATP synthesis activity of bRF F -PL. For the ATP synthesis activity o 1 assay, 20 µl PL was mixed with 7.3 mM ultrapurified ADP and 10 mM NaN , and Preparation of split-GFP. The split-GFP was prepared as previously described . 3 The split-GFP was prepared firstly by introducing 7 point mutations into the gene the total volume to 40 µl with buffer PA6-5 was filled up. After illuminating the PLs for a given time length and light intensity with a halogen lamp, the ATP synthesis of superfolder-GFP (sfGFP), then by dividing into two portions. We used P1–P14 primer sets (Supplementary Table 3) to introduce point mutations N39I, T105K, was terminated by breaking the PLs with 2.5% trichloroacetic acid (TCA). This was accompanied by neutralizing the mixture with equal volume of buffer N (250 mM E111V, I128T, K166T, I167V and S205T, by quick change PCR. The DNA strand for a larger part (GFP1-10) of the split-GFP was amplified by PCR using P15 and Tris-HCl (pH 9.5) and 4 mM EDTA). Then, the mixture was injected into 800 µl of P16 primes and inserted into pET29a vector between the restriction enzyme sites of buffer R (pH 8.3) (20 mM tricine, 20 mM succinic acid, 80 mM NaCl and 0.6 mM KOH) which was pre-supplemented with 100 µl of luciferase/luciferin mix (CLSII). NdeI and EcoRI using infusion cloning technique. Thereby, a hexa-histidine-tag was introduced at the N-terminus of the open reading frame. The luminescence was measured (using AB-2270 ATTO luminometer) every 1 s for a total of 150 s. The trace of luminescence signal was used to calculate the The resulting construct, pET29aGFP1-10 (see the DNA sequence in Supplementary Table 5), was introduced into BL21(DE3) strain for overexpression synthesized amount of ATP based on the standard curve where the luminescence intensity is plotted against known concentrations of ATP. and purification of GFP1-10. The E. coli strain B21(DE3) harboring pET29aGFP1-10 was cultured in LB −1 media containing 50 µg ml Kanamycin. The culture was incubated under shaking at 37 °C until OD 0.6. Then, 1 mM IPTG was added for induction and the ΔpH-dependent ATP synthesis activity of F F -PL. The ΔpH-dependent ATP o 1 incubation continued for additional 6 h at 30 °C. The cells were then collected and synthesis activity of F F -PL was measured by acid-base transition o 1 10,24 washed one time. The collected cells were sonicated being suspended in a buffer assay . SoyPC extract was reconstituted with 1 µM F F to form F F -PL at the o 1 o 1 −1 containing 20 mM Tris-HCl (pH 8.0), 200 mM NaCl, 0.005 % Triton X-100 and concentration of 16 mg ml . This was used for acid-base transition assay. The protease inhibitor cocktail. After removing debris by centrifugation, the lysate was lumen of 30 µl F F -PL was acidified in acidification buffer (20 mM succinic acid, o 1 injected to His-Trap Ni-column which was pre-equilibrated with buffer A (pH 8.0) 0.6 mM KOH, 2.5 mM MgCl , 10 mM NaH PO (pH 4.5)) containing 286 nM 2 2 4 (20 mM Tris-HCl, 20 mM NaCl and 10 mM imidazole). After the washing, the valinomycin, 0.8 mM ultrapure ADP and 0.07 mM AP5A for 5 min at room protein was eluted with linear gradient of 10–300 mM of imidazole. Further temperature. The ΔpH-dependent ATP synthesis was assayed by injecting the purification was carried out using anion exchange chromatography (mono-Q acidified PL into a basic buffer (20 mM Tricine, 130 mM KOH, 2.5 mM MgCl , column) after exchanging the buffer with buffer C (pH 8.0) (20 mM Tris-HCl, 10 mM NaH PO , 7.3 mM ultrapure ADP (pH 8.8)) containing luciferin and 2 4 10 mM NaCl). The protein was eluted with linear gradient of NaCl from 10 to luciferase reagents. The generated ATP level was estimated by injecting 0.2 nmol 500 mM. The purified protein was stored at −80 C in a buffer containing 10 mM of ATP three times as a standard. 8 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE 35,36 Preparation of GUVs. GUVs were prepared as described previously , with two-step PCR (i.e., where P30 was used as a reverse PCR). The DNA templates for slight modification. GUVs were prepared from a fresh lipid-paraffin mix always. the cell-free expression of F complex (uncB, uncF and uncE) were prepared using Lipids dissolved in chloroform at a desired stock concentration were mixed toge- a primer set of P32 and P35, P36 and P37, or P38 and P39, respectively. Point ther considering the lipid-paraffin mix volume of 500 µl. This mixture was mixed mutation in uncB (R169A) was incorporated by overlap PCR using P33 and P34. well and flushed with a flow of N gas. To enhance the lipid mixing and completely Finally, T7 promoter, ribosome binding site and 3’ extra 14 nucleotides were added remove the remaining chloroform, the lipid-paraffin mix was heated at 80 °C for to it by P32 and P35. 20 min and vigorously vortexed. This was further flushed with a flow of N gas and the vials were sealed tightly. The vials containing the lipid mix were then sonicated Co-flotation assay. Spontaneous membrane integration of cell-free synthesized in warm water bath (55 °C) for 30–60 min. bR was performed as described previously with modification. bR was synthesized The lipid mix was let to cool at room temperature. Subsequently, 300 µl of the in the PURE system from 5 nM of bR template DNA (with upstream T7 promoter lipid-paraffin solution was mixed with 30 µl inner solution supplemented with 35 −1 and ribosome binding site) in the presence of [ S]methionine and 8 mg ml 200 mM of sucrose. Water-in-oil emulsion was created by gently pipetting up and SoyPC extract liposomes, which was filtered by 200 nm pore size membrane using a down for few 10 s. The emulsion was slowly laid over the outer solution (composed micro-extruder (Avanti Polar). The synthesis reaction was terminated by adding of equal volume of 400 mM glucose and PURE buffer, which is the same −1 RNaseA at the concentration of 20 ng µl followed by incubation at 37 °C for composition as GUV inside but without tRNAs) and kept on ice for 10 min. Next, 30 min. Then, 10 µl of synthesized bR (i.e., bR-PL) was overlaid on the top of the sample was centrifuged at 10,000 × g for 30 min followed by collecting the 0, 25 and 30% sucrose layers, where each layer was with a volume of 300 µl in precipitate by piercing the bottom of the flat-bottomed Eppendorf tube by 18-gauge sucrose flotation buffer (50 mM HEPES-KOH (pH 7.6), 100 mM KCl, 10 mM needle. Finally, the collected GUV suspension was centrifuged again at 9100 × g for MgCl ). This was centrifuged at 157,800 × g for 30 min using Beckman Coulter 10 min and precipitated GUV was resuspended in 50 µl of fresh outer solution. Optima Max XP Ultracentrifuge using TLS 55 rotor. Fractions of 200 µl were collected from the top in a total of four fractions. To the collected fractions, ice- Light-driven ATP synthesis inside GUV. The GUVs were prepared from lipid chilled TCA was added at the concentration of 10% and centrifuged at 20,400 × g mixture of POPC, cholesterol and PEG2000PE (5.75:4:0.25 molar ratio). The for 30 min at 4 °C. The supernatant was removed and the pellet was resuspended in bRF F -PLs were constituted from 176 µM bR and 1 µM F F with buffer-1 (pH o 1 o 1 200 µl of acetone and sonicated for at least 3 min. The suspension was centrifuged 7.3), and then encapsulated inside GUVs together with 7.3 mM ultrapurified ADP again at 20,400 × g for 30 min. The pellet was left at room temperature for 5 min, and 10 mM NaN . After the formation of GUVs, the outer buffer was replaced by a 3 mixed with 10–20 µl of loading dye and water-bath sonicated for 1 min. The fresh buffer-1 containing 200 mM glucose. Proteinase K (PK) was supplemented sample was finally analyzed by 15% SDS-PAGE. either to the outer or both the inner and outer solution at the concentration of 1 µM. After PK addition, the samples were incubated at 37 °C for 2 h prior Membrane orientation assay. Membrane orientation of the reconstructed bR- (to degrade unencapsulated bR and F F ) to the main light or dark incubation. o 1 PLs or F F -PLs was assessed by observing the binding affinity of each protein o 1 These GUV samples were incubated at 37 °C under light or dark. As a control, component to Ni-NTA Magnetic Beads, and 41.3 mM of lipid mixture (SoyPC an in vitro reaction mixture was also prepared in the same condition as the extract/cholesterol with the molar ratio of 70:30) was reconstituted with either F F o 1 encapsulated reaction mixture and illuminated. or bR at the protein concentration of 0.2 µM. The β-subunit of the recombinant F F and C-terminus of bR were bearing a His-tag. Magnetic beads conjugated o 1 Light-driven protein synthesis in GUV or in vitro. For the light-driven translation with Ni were used to trap His-tagged terminus of the proteins facing outside of reaction, bRF F -PL (176 µM bR and 1 µM F F ) was resuspended in buffer-1 o 1 o 1 liposome (cytosol side). As a control, PLs were solubilized with 0.5% Triton X-100 (Supplementary Table 4) supplemented with 500 nM of mRNA encoding sfGFP or and had to undergo the same procedure as the experimental samples. Finally, the bR::sfGFP (see the DNA sequence in Supplementary Table 5), 7.3 mM ADP, 10 mM collected fractions were run in 10–20% gradient gel and bands were visualized by NaN , 40 A per ml tRNA and 200 mM sucrose. This was mixed with the corre- western blotting with anti-His-tag antibody. sponding enzyme mix (Supplementary Table 2) and then encapsulated inside GUVs. The vesicle suspension was incubated for 6 h at 37 °C under the light illumination Photosynthesis of bR or F . A mutant bR (bR ) containing substitutions passing 500 nm long-pass filter. For the light-driven transcription-and-translation o mut of D85N and K216N was used as a negative control of photosynthesis of bR. reaction, the reaction mixture was prepared in same way, but using buffer-2 (Sup- The point mutations were introduced into pEXP-5-ct-bRwt construct, sequentially, plementary Table 4) supplemented with 1 nM DNA encoding GFP11 (see the DNA by quick change PCR using primer sets of P32-and-P33 and P34-and-P35 sequence in Supplementary Table 5) and 5 µM of the purified GFP1-10 protein. (see the DNA sequence in Supplementary Table 5). Next, linear DNA template The reaction was carried out for 7 h at 37 °C under the light illumination. was prepared using P22 and P23 to be used as a template for in vitro transcription. The vesicles synthesizing sfGFP were observed by confocal laser scanning For light-dependent expression of either bR or functional bR , the reaction microscopy (CLSM) (Zeiss LSM 550). For the population analysis, the vesicle mut wt mixture of PURE system was supplemented with bRF F -PL (where the PL was suspension was first diluted 10 times with buffer PA6-5 and then the fluorescence o 1 reconstituted with 5 µM bR and 1 µM F F ), 100 µM of ATR and either 0.4 µM of intensity of 100,000 vesicles was analyzed by fluorescence-activated cell sorter o 1 bR mRNA or 0.8 µM bR mRNA. The reaction mix was incubated at 37 °C (FACS Aria III). wt mut In vitro, the light-driven transcription-translation reaction was performed using under light and 2 µl sample was aliquoted at each given time for subsequent ATP level quantification. the same reaction mixture as mentioned above in the presence of [ S]methionine. The synthesized protein was subjected to 15% SDS-PAGE and visualized by On the other hand, for functional in vitro assembly of F F , the F complex o 1 o was expressed from the uncB, uncF and uncE-mRNAs (i.e., at the concentration autoradiography (Fujifilm). For the light-driven transcription-and-translation of 40 nM, 20 nM and 100 nM, respectively) coding for their respective subunits reaction, the sample was reacted for 13 h and subjected to column filtration a, b and c. The expression of the F complex was supplemented with 300 nM of (RNase-Free Micro Bio-Spin Columns with Bio-Gel P-30). Radioactivity of the o purified F complex for co-translational assembly of the two partners. The reaction resulting sample was counted by Liquid Scintillation Counter (ALOKA LSC-6100). mixture of PURE system (buffer-2 containing 200 mM of sucrose) was initially Concerning light-driven translation reaction using transient light illumination, seeded with 18 nM of bRF F -PL that was composed of 140 µM bR and 0.29 µM 500 μM of ADP, 500 nM mRNA coding GFP and [ S]methionine were o 1 F F . After incubating the reaction mixture for 7 h under light at 37 °C, 5 times supplemented in the translation reaction mixture. At a given time, aliquots were o 1 diluted PL was isolated by centrifuging at 230,000 × g for 30 min. The precipitated taken for 15% SDS-PAGE and ATP amount analysis. PL was resuspended in assay buffer (pH 7.3) (composed of 10 mM HEPES, 7.3 mM ADP, 10 mM NaH PO , 3 mM MgCl , 10 mM NaN and 200 mM sucrose), 2 4 2 3 bR::sfGFP synthesis in GUV. GUVs were prepared with 6 mM POPC and 4 mM making the final concentration of 44 nM. Later, the light-driven ATP synthesis cholesterol with encapsulating a cell-free reaction mix (PUREfrex Ver.2) together was assayed as described before. For negative control, the mutant uncB (R169A) with 5 nM linear DNA of bR::sfGFP construct (see the DNA sequence in Sup- −1 mRNA template was used. plementary Table 5), 150 µM ATR and 8 mg ml of SoyPC liposome. The reaction was performed in the presence or absence of liposomes for 4 h at 37 °C without light illumination. The fluorescence pattern of the synthesized bR::GFP fusion Reporting Summary. Further information on experimental design is available in protein in GUV lumen was analyzed by CLSM. the Nature Research Reporting Summary linked to this article. Preparation of bR or F DNA for the cell-free expression. The bR gene (with the Data availability codon optimized for expression in the PURE system) was amplified by primer P26 The authors declare that all the relevant data supporting the findings of the study are and P31, in the first step of PCR, whereas P21 and P31 primers were used in final available in this article and its Supplementary Information file, or from the step of PCR. Furthermore, for bR::sfGFP fusion construct, 3’ and 5’ overlap corresponding author (Y.K.) upon reasonable request. sequences were added to the bR and sfGFP gene by separate PCR using a primer set of P25 and P27 or P28 and P29, respectively. The resulting PCR products were used for overlap PCR in the presence of P25 and P29 as a forward and reverse Received: 16 August 2018 Accepted: 26 February 2019 primer, respectively. Eventually, ribosome binding site and T7 promoter site were added to the fusion construct by the primer P26 and P21, respectively, in the final NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 References 28. Kurihara, K. et al. Self-reproduction of supramolecular giant vesicles 1. Rasmussen, S. et al. Evolution. Transitions from nonliving to living matter. combined with the amplification of encapsulated DNA. Nat. Chem. 3, Science 303, 963–965 (2004). 775–781 (2011). 2. Luisi, P. L. The Emergence of Life: From Chemical Origins to Synthetic Biology 29. Deamer, D. & Weber, A. L. Bioenergetics and life’s origins. Cold Spring Harb. (Cambridge University Press, Cambridge, 2006). Perspect. Biol. 2, a004929 (2010). 3. Mann, S. Systems of creation: the emergence of life from nonliving matter. 30. Yoshida, M., Muneyuki, E. & Hisabori, T. ATP synthase–a marvellous rotary ACC Chem. Res. 45, 2131–2141 (2012). engine of the cell. Nat. Rev. Mol. Cell Biol. 2, 669–677 (2001). 4. Hutchison, C. A. 3rd et al. Design and synthesis of a minimal bacterial 31. Shiu, P. J., Ju, Y.-H., Chen, H.-M. & Lee, C.-K. Facile isolation of purple genome. Science 351, aad6253 (2016). membrane from Halobacterium salinarum via aqueous-two-phase 5. Xu, C., Hu, S. & Chen, X. Artificial cells: from basic science to applications. system. Protein Expr. Purif. 89, 219–224 (2013). Mater. Today (Kidlington, Engl.) 19, 516–532 (2016). 32. Suzuki, T., Ueno, H., Mitome, N., Suzuki, J. & Yoshida, M. F of ATP synthase 6. Shimizu, Y. et al. Cell-free translation reconstituted with purified components. is a rotary proton channel. J. Biol. Chem. 277, 13281–13285 (2002). Nat. Biotechnol. 19, 751–755 (2001). 33. Suzuki, T. et al. F F -ATPase/synthase is geared to the synthesis mode by 0 1 7. Kuruma, Y., Stano, P., Ueda, T. & Luisi, P. L. A synthetic biology approach to conformational rearrangement of epsilon subunit in response to proton the construction of membrane proteins in semi-synthetic minimal cells. motive force and ADP/ATP balance. J. Biol. Chem. 278, 46840–46846 (2003). Biochim. Biophys. Acta 1788, 567–574 (2009). 34. Rigaud, J. L., Paternostre, M. T. & Bluzat, A. Mechanisms of membrane 8. Ichihashi, N. et al. Darwinian evolution in a translation-coupled RNA protein insertion into liposomes during reconstitution procedures involving replication system within a cell-like compartment. Nat. Commun. 4, 2494 the use of detergents. 2. Incorporation of the light-driven proton pump (2013). bacteriorhodopsin. Biochemistry 27, 2677–2688 (1988). 9. Matsubayashi, H., Kuruma, Y. & Ueda, T. In vitro synthesis of the E. coli 35. Fujii, S., Matsuura, T., Sunami, T., Kazuta, Y. & Yomo, T. In vitro evolution Sec translocon from DNA. Angew. Chem. (Int. Ed. Engl.) 53, 7535–7538 of alpha-hemolysin using a liposome display. Proc. Natl Acad. Sci. USA 110, (2014). 16796–16801 (2013). 10. Kuruma, Y. & Ueda, T. The PURE system for the cell-free synthesis of 36. Fujii, S. et al. Liposome display for in vitro selection and evolution of membrane proteins. Nat. Protoc. 10, 1328–1344 (2015). membrane proteins. Nat. Protoc. 9, 1578–1591 (2014). 11. Furusato, T. et al. De novo synthesis of basal bacterial cell division proteins 37. Pykäläinen, A. et al. Pinkbar is an epithelial-specific BAR domain protein FtsZ, FtsA, and ZipA inside giant vesicles. ACS Synth. Biol. 7, 953–961 (2018). that generates planar membrane structures. Nat. Struct. Mol. Biol. 18, 902–907 12. Rampioni, G. et al. Synthetic cells produce a quorum sensing chemical signal (2011). perceived by Pseudomonas aeruginosa. Chem. Commun. (Camb.) 54, 2090–2093 (2018). Acknowledgements 13. Deisinger, B. et al. Purification of ATP synthase from beef heart mitochondria We thank Professor D. Oesterhelt and Dr. S.V. Gronau for the Halobacterium salinarum (F0F1) and co-reconstitution with monomeric bacteriorhodopsin into strain R1, Dr. Kazuhito Tabata for discussion on ATP synthase, Dr. Satoshi Fujii for liposomes capable of light-driven ATP synthesis. Eur. J. Biochem. 218, operation of FACS, Dr. Toshiharu Suzuki for the constructs of F F and Dr. T.Z. Jia for 377–383 (1993). o 1 advice on manuscript preparation. This work was supported by JSPS KAKENHI (Grant 14. Freisleben, H. J. et al. Reconstitution of bacteriorhodopsin and ATP synthase Numbers 16H06156, 16KK0161, 16H00797, J26106003 to Y.K. and 16H02089 to T.U.), from Micrococcus luteus into liposomes of the purified main tetraether the Astrobiology Center Project of the National Institutes of Natural Sciences (NINS) lipid from Thermoplasma acidophilum: proton conductance and light-driven (Grant Number AB291017 to Y.K.), and JST, PRESTO (Grant Number JPMJPR18K5 ATP synthesis. Chem. Phys. Lipids 78, 137–147 (1995). to Y.K.). 15. Matuschka, S., Zwicker, K., Nawroth, T. & Zimmer, G. ATP synthesis by purified ATP-synthase from beef heart mitochondria after coreconstitution with bacteriorhodopsin. Arch. Biochem. Biophys. 322, 135–142 (1995). Author contributions 16. Richard, P., Pitard, B. & Rigaud, J. L. ATP synthesis by the F0F1-ATPase from S.B. performed all experiments. Y.K. conceived the idea, designed most of the the thermophilic Bacillus PS3 co-reconstituted with bacteriorhodopsin into experiments, analyzed the data, prepared the figures and wrote the manuscript. T.U. liposomes. Evidence for stimulation of ATP synthesis by ATP bound to a gave advice for manuscript preparation. noncatalytic binding site. J. Biol. Chem. 270, 21571–21578 (1995). 17. Choi, H. J. & Montemagno, C. D. Artificial organelle: ATP synthesis from cellular mimetic polymersomes. Nano Lett. 5, 2538–2542 (2005). Additional information 18. Lee, K. Y. et al. Photosynthetic artificial organelles sustain and control ATP- Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- dependent reactions in a protocellular system. Nat. Biotechnol. 36, 530–535 019-09147-4. (2018). 19. Pitard, B., Richard, P., Dunach, M., Girault, G. & Rigaud, J. L. ATP synthesis Competing interests: The authors declare no competing interests. by the F0F1 ATP synthase from thermophilic Bacillus PS3 reconstituted into liposomes with bacteriorhodopsin. 1. Factors defining the optimal Reprints and permission information is available online at http://npg.nature.com/ reconstitution of ATP synthases with bacteriorhodopsin. Eur. J. Biochem. 235, reprintsandpermissions/ 769–778 (1996). 20. Deng, D., Jiang, N., Hao, S. J., Sun, H. & Zhang, G. J. Loss of membrane Journal peer review information: Nature Communications thanks the anonymous cholesterol influences lysosomal permeability to potassium ions and protons. reviewers for their contribution to the peer review of this work. Peer reviewer reports are Biochim. Biophys. Acta 1788, 470–476 (2009). available. 21. Bald, D. et al. ATP synthesis by F0F1-ATP synthase independent of noncatalytic nucleotide binding sites and insensitive to azide inhibition. J. Biol. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in Chem. 273, 865–870 (1998). published maps and institutional affiliations. 22. Cabantous, S., Terwilliger, T. C. & Waldo, G. S. Protein tagging and detection with engineered self-assembling fragments of green fluorescent protein. Nat. Biotechnol. 23, 102–107 (2005). Open Access This article is licensed under a Creative Commons 23. Nakamura, S., Suzuki, S., Saito, H. & Nishiyama, K. I. Cholesterol blocks spontaneous insertion of membrane proteins into liposomes of Attribution 4.0 International License, which permits use, sharing, phosphatidylcholine. J. Biochem. 163, 313–319 (2018). adaptation, distribution and reproduction in any medium or format, as long as you give 24. Kuruma, Y., Suzuki, T., Ono, S., Yoshida, M. & Ueda, T. Functional analysis of appropriate credit to the original author(s) and the source, provide a link to the Creative membranous Fo-a subunit of F1Fo-ATP synthase by in vitro protein Commons license, and indicate if changes were made. The images or other third party synthesis. Biochem. J. 442, 631–638 (2012). material in this article are included in the article’s Creative Commons license, unless 25. Adamala, K. & Szostak, J. W. Nonenzymatic template-directed RNA synthesis indicated otherwise in a credit line to the material. If material is not included in the inside model protocells. Science 342, 1098–1100 (2013). article’s Creative Commons license and your intended use is not permitted by statutory 26. Soga, H. et al. In vitro membrane protein synthesis inside cell-sized vesicles regulation or exceeds the permitted use, you will need to obtain permission directly from reveals the dependence of membrane protein integration on vesicle volume. the copyright holder. To view a copy of this license, visit http://creativecommons.org/ ACS Synth. Biol. 3, 372–379 (2014). licenses/by/4.0/. 27. Buddingh, B. C. & van Hest, J. C. M. Artificial cells: synthetic compartments with life-like functionality and adaptivity. ACC Chem. Res. 50, 769–777 © The Author(s) 2019 (2017). 10 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Artificial photosynthetic cell producing energy for protein synthesis

Loading next page...
 
/lp/springer-journals/artificial-photosynthetic-cell-producing-energy-for-protein-synthesis-B5wAr8nx6o

References (41)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/s41467-019-09147-4
Publisher site
See Article on Publisher Site

Abstract

ARTICLE https://doi.org/10.1038/s41467-019-09147-4 OPEN Artificial photosynthetic cell producing energy for protein synthesis 1 1 2,3 Samuel Berhanu , Takuya Ueda & Yutetsu Kuruma Attempts to construct an artificial cell have widened our understanding of living organisms. Many intracellular systems have been reconstructed by assembling molecules, however the mechanism to synthesize its own constituents by self-sufficient energy has to the best of our knowledge not been developed. Here, we combine a cell-free protein synthesis system and small proteoliposomes, which consist of purified ATP synthase and bacteriorhodopsin, inside a giant unilamellar vesicle to synthesize protein by the production of ATP by light. The photo-synthesized ATP is consumed as a substrate for transcription and as an energy for translation, eventually driving the synthesis of bacteriorhodopsin or constituent proteins of ATP synthase, the original essential components of the proteoliposome. The de novo photosynthesized bacteriorhodopsin and the parts of ATP synthase integrate into the artificial photosynthetic organelle and enhance its ATP photosynthetic activity through the positive feedback of the products. Our artificial photosynthetic cell system paves the way to construct an energetically independent artificial cell. Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, Bldg. FSB-401, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8562, Japan. Earth-Life Science Institute, Tokyo Institute of Technology, 2-12-1-IE-1, Ookayama, Meguro-ku, Tokyo 152- 8550, Japan. JST, PRESTO, Saitama 332-0012, Japan. Correspondence and requests for materials should be addressed to T.U. (email: ueda@edu.k.u-tokyo. ac.jp) or to Y.K. (email: kuruma@elsi.jp) NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ecent advances in synthetic biology allow us to challenge found that only 25% bR were maintaining the proper membrane whole reconstruction of cell from simple non-living orientation (Supplementary Fig. 3C). To improve this ratio, we 1–4 Rmolecules and redesigned minimal genome . Such did some modifications in the preparation method by changing attempts for the construction of artificial cell would lead not only the timing of bR addition (Supplementary Fig. 3A), i.e., empty to determining the necessary requirements for life phenomena liposomes were first roughly preformed and, then, the purified but also to developing as a biodevice toward industrial applica- bR was combined before completely removing the detergent. By tion . A cell-mimicking artificial cell is constructed by encapsu- this method, 70% bR was properly reconstructed in the PLs lating a cell-free protein synthesis system inside giant vesicle. (Supplementary Fig. 3C). The improvement of the membrane Cell-free system has been widely applied to researches in the field orientation faithfully reflected into the proton-pump activity of synthetic biology, and especially a reconstructed cell-free (Supplementary Fig. 3D). Since the efficiency of proton gradient system (PURE system) has been used as a basic technology for generation directly affects the F F activity, we employed this o 1 the artificial cell construction because all constituent enzymes are optimized method for all of the following experiments. known. This would be rather important when we try to recon- During the light illumination, we observed a decrease of proton struct self-reproducing artificial cells that have to synthesize all concentration at the outside of bR-PLs in proportion to bR their own components. Although several cellular functions or concentration (Fig. 1c), suggesting that the protons were phenomenon have been reconstructed so far in the artificial cell transported from the outside to inside of the bR-PL lumen 7–12 system , an energy self-supplying system for the internal (Supplementary Fig. 1A). In addition to the proton-pump protein synthesis has not been achieved. To develop the artificial activity, we also observed a rapid return of the proton cell into the energetically independent system, it is necessary to concentration when the illumination ceased. This indicates set up a circulating energy-consumption and production system proton leakage from the inside to outside of the bR-PL lumen. driven by an unlimited external physical or chemical energy The proton leak was accelerated when the lateral fluidity of the source. For this purpose, a biomimetic artificial organelle pro- bR-PL membranes was increased by temperature rise (Supple- ducing adenosine triphosphate (ATP) by collaborating ATP mentary Fig. 4). For the sake of inhibiting the leak through the synthase and bacteriorhodopsin is applicable as a rational energy membrane, we added 30% cholesterol into the lipid composition 13–18 18 20 generating system for artificial cells . Recently, Lee et al. of bR-PLs , which resulted in 30% reduction of the proton leak performed ATP synthesis using similar photosynthetic artificial (Supplementary Fig. 5). Thus, we kept this condition throughout organelle, where they demonstrated carbon fixation (in vitro) the study. and actin polymerization within giant unilamellar vesicle (GUV). Next, we estimated the membrane orientation of the This result evokes us to apply the artificial organelle into the reconstituted bR by evaluating the binding sensitivity of a artificial cell system, i.e., protein synthesis based on the photo- histidine-tag, which elongated at the C-terminus of recombinant synthesized ATP inside GUV. In this study, we performed ATP bR, to the Ni-NTA-conjugated magnet beads (Supplementary synthesis by light-driven artificial organelle inside GUV. Through Fig. 6). If the reconstructed bR was keeping the working optimization for the preparation method of proteoliposomes orientation, the C-terminus histidine-tag can bind to the magnet containing bacteriorhodopsin and ATP synthase, we succeeded to beads and be eluted in the elution fraction. The ratio of bR produce millimolar level ATP inside GUVs, wherein 4.6 µmol obtained in the elution fraction was normalized with the ratio of ATP per mg ATP synthase was produced after 6 h of illumina- control experiment in which bR was monodispersed by dissolving tion. By combining the artificial organelle and PURE system, we the PLs with detergent (Triton). In the control experiment, design and construct an artificial photosynthetic cell that pro- 91% bR was collected in the elution fraction, although that duces ATP for the internal protein synthesis. The produced ATP should be 100% theoretically (Supplementary Fig. 6). Considering was consumed as a substrate of messenger RNA (mRNA), or as this result, we calculated that 86% bR was reconstructed in an energy for aminoacylation of transfer RNA (tRNA) and for the working (outward C-terminus) orientation within the PL −1 phosphorylation of guanosine diphosphate (GDP) (Fig. 1a and membrane; i.e., Elu. Elu. 100%. It should be noted −Triton +Triton Supplementary Fig. 1). Additionally, we also demonstrated pho- that the opposite orienting bRs (inward C-terminus) pump tosynthesis of bacteriorhodopsin or a membrane portion of ATP protons from the inside to outside of the PLs. Thus, the net- synthase, which is the original component of the artificial orga- working ratio of the reconstituted bR is calculated as 72% nelle. Our artificial cell system enables the self-constitution of its (Supplementary Table 1). Taking account of the bR membrane own parts within a structure of positive feedback loop. orientation, the initial reaction rate of bR was calculated as −2.87 ± −1 −1 −1 −1 0.53 ΔpH min nmol or −0.11 ± 0.02 ΔpH min mg , mean ± S.D. (Fig. 1c and Supplementary Table 1). On the other hand, Results the net-working ratio of the reconstituted F F was 65.1% o 1 Construction of light-driven artificial organelle. Light-driven after the normalization as with bR (Supplementary Fig. 7 and artificial organelle was composed of two kinds of membrane Supplementary Table 1), and the initial reaction rate was 128 ± 3.2 −1 −1 −1 −1 proteins, bacteriorhodopsin (bR) and F-type ATP synthase ATP nmol min nmol or 223 ± 6.1 ATP nmol min mg (F F ). bR was isolated from a purple membrane of Halobacter- (Fig. 1d and Supplementary Table 1). The reverse function of o 1 ium salinarum by ultra-centrifugation with sucrose density F F , ATP-dependent proton-pump activity, was also detected o 1 gradient (Fig. 1b and Supplementary Fig. 2). F F of Bacillus PS3 (Supplementary Fig. 8), suggesting the full functionality of the o 1 was purified as recombinant protein from Escherichia coli cells reconstituted F F -PLs. o 1 (Fig. 1b). The isolated bR were reconstructed as bR-embedding To construct artificial organelle, we assembled purified bR and proteoliposomes (bR-PLs) for the measurement of light- F F to form bRF F -PLs. We prepared PLs in various proportion o 1 o 1 dependent proton-pump activity. The size of bR-PLs were of bR against F F and illuminated with visible light passing a o 1 mostly 100–200 nm as diameter. We used phosphatidylcholine 500 nm long-pass filter. The amount of produced ATP was extract from soybean to form PLs which are stable in the reaction measured by means of luciferin and luciferase. The highest ATP mixture of PURE system and also maintain the F F activity . photosynthesis was obtained in the case of 176 µM bR and 1 µM o 1 The formation of bR-PLs was carried out by reducing the F F . This means that approximately 0.6 × 10 ATP was produced o 1 detergent concentration in the mixture of lipids and purified by a single bRF F -PL within 4 h of illumination (Fig. 1e). The o 1 protein according to the previous report ; however, we have maximum turnover number for ATP synthesis in the initial 5 min 2 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 1 F F -PLs o 1 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE a bc bR F F o 1 kDa 250 75 150 α H β 20 μM bR+FCCP H DNA ADP ATP H 0 GDP ATP hv 37 γ synthase –0.5 ATP + ADP GTP H + mRNA 25 + H + + a –1.0 H H 20 25 H b 15 –1.5 + ε Protein Bacteriorhodopsin –2.0 10 10 0 120 240 360 480 600 Time (s) d ef ×10 ×10 In GUVs 600 1.0 5E+4 1.6 176:1 Ratio of bR:F F (μM) 30 nM F F o 1 o 1 4E+4 *** 1.4 used for PLs preparation 3E+4 140:1 1.2 0.8 70:1 2E+4 1.0 400 1E+4 35:1 *** 20 nM F F 176:1 o 1 0.8 0.6 012 345 Time (min) 0.6 140:1 0.4 70:1 0.4 0.2 35:1 10 nM F F o 1 30 nM F F +FCCP GUVs + ++ – 0.2 o 1 –– 0 +FCCP PKin + – 0 20 40 60 80 100 120 Dark PKout ++ + Time (s) 0 Light ++ – + Time (h) Fig. 1 Light-driven adenosine triphosphate (ATP) synthesis by artificial organelle. a Schematics of the artificial photosynthetic cell encapsulating artificial organelle, which consists of bacteriorhodopsin (bR) and F F -ATP synthase (F F ). Synthesized ATP are consumed as substrates for messenger RNA o 1 o 1 (mRNA) (➀), as energy for phosphorylation of guanosine diphosphate (GDP) (➁) or as energy for aminoacylation of transfer RNA (tRNA) (➂). b Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of purified bR and F F . The positions of molecular markers and F F component o 1 o 1 proteins are indicated beside the gels. c Light-driven proton-pump activity of bR reconstituted in a proteoliposome (PL). Proton-pump activity of bR was measured by monitoring the proton concentration at the outside of bR-PLs where fluorescent proton-sensor ACMA (9-amino-6-chloro-2-methoxy acridine) was added. We defined as ΔpH = pH (original, outside) − pH (after illumination, outside). The ΔpH caused by bR activity was measured with the various bR concentrations as indicated. White and gray areas indicate light ON and OFF, respectively. An uncoupler, FCCP (carbonyl cyanide 4- (trifluoromethoxy) phenylhydrazone), was used as a control experiment. d ATP synthesis activity of F F reconstituted as F F -PLs. ATP synthesis reactions o 1 o 1 were initiated by adding F F -PLs at 30 s with various F F concentrations, as indicated. The synthesized ATP was measured by means of luciferin and o 1 o 1 luciferase (see Methods section for the experiment details). FCCP was used for control. e Light-driven ATP synthesis. The amount of the photosynthesized ATP by bRF F -PLs, which was constituted in various proportions of bR against F F , were measured by luciferin and luciferase. FCCP and dark conditions o 1 o 1 were also performed as controls. The inset indicates initial rate of the each PL. f Light-driven ATP synthesis inside giant unilamellar vesicle (GUV). bRF F - o 1 PLs were illuminated inside GUVs in the presence or absence of proteinase K (PK) that degrades the F F . The in vitro experiment was also performed for o 1 comparison. ***p < 0.001. P values were from two-sided t-test. All experiments were repeated at least three times, and their mean values and standard deviations (S.D.) are shown. Source data are provided as a Source Data file −1 was 8.3 ± 0.3 s in the case of 176 µM bR and 1 µM F F . This synthase. The efficiency of ATP production in GUVs was o 1 was almost double compared to the previous report . Here, in a roughly one-third that of the in vitro system, perhaps caused single PL, 3560 of the working bRs drive 18 F F (Supplementary by lower light intensity inside a GUV. Since our artificial o 1 Table 1). In all cases, we used 10 mM NaN to inhibit the reverse organelle can produce ATP inside GUV at the comparable (ATPase) activity of F F . We found that the ATP production concentration as a real living cell, we proceeded to design and o 1 plateaued when the illumination was higher than 10 mW per cm construct the photosynthetic artificial cell system that synthesize (Supplementary Fig. 9). protein by light. The same reaction was also performed inside GUVs in which about 1.1 × 10 bRF F -PLs are contained in a 10 µm diameter o 1 GUV. After 6 h of illumination, we observed photosynthesized Light-driven protein synthesis inside the artificial cell.We ATP from the inside of the GUVs (Fig. 1f), where 1.8 mM performed green fluorescent protein (GFP) synthesis inside ATP was produced in a single GUV (Supplementary Table 1). GUVs by means of the photosynthesized ATP to demonstrate This represents that 4.6 µmol ATP was produced per mg ATP that the constructed artificial organelle works in an artificial cell NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 3 Synthesized ATP (nM) Produced ATP/bRF F -PL o 1 ΔpH at the bR-PLs exterior Produced ATP/bRF F -PL o 1 In vitro ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 system. For this purpose, we combined bRF F -PLs with the confirmed in vitro (without GUVs) (Fig. 2b and Supplementary o 1 PURE system which is a cell-free protein synthesis system. The Fig. 13) and synchronized with timing of the ATP photosynthesis PURE system is reconstituted from purified translation factors , (Supplementary Fig. 14). These results indicate that GFP synth- and therefore we can customize the component factors suited esis inside the GUVs was driven by the photosynthesized for the designed artificial system. The PURE system was modified ATP (Supplementary Fig. 1B). We found that 50–60% of the as shown in Supplementary Table 2 to allow the photosynthesized GUVs emitted more fluorescence than the nonilluminated con- ATP be specifically used for the aminoacylation of tRNA (Sup- trol GUVs (Fig. 2c) by flow-cytometry analysis. We also found plementary Fig. 10), and supplied with a mRNA encoding a certain percentage of GUVs were not showing fluorescent GFP together with bRF F -PLs and NaN . NaN did not inhibit even when illuminated. Although the definitive cause is not o 1 3 3 protein synthesis at concentrations below 50 mM (Supplementary unclear, it has been reported that the encapsulation efficiency Fig. 11). The prepared reaction mixture was encapsulated of PLs is affected by the size of PLs; i.e., less than 35% GUVs inside GUVs, and illuminated to induce protein synthesis. A large can encapsulate the PLs when their size are over 200 nm majority of the GUV population appeared in a range of 10–20 µm (diameter) . Additionally, we cannot deny the possibility that as diameter (n = 200) (Supplementary Fig. 12). After 6 h, we inactivity of the internal artificial organelle by the fusion of observed the fluorescence of internally synthesized GFP by bRF F -PLs and GUV membranes is limiting the successful o 1 confocal microscopy (Fig. 2a). This GFP synthesis was also artificial cell formation. ab c Light Dark 200K bRF F -PL + + o 1 100K bR-PL – – – – F F -PL o 1 ADP + + ATP – – Light + – 10K GFP 300 300 25 kDa 4K 10 100 1K 10K 0 0 0 180 0 160 Pixel Pixel GFP fluorescent (A.U.) d e h 200K bRF F -PL + +++ + + o 1 100K NDK – – ++ + – ++–– – – GTP 300 GDP – –– ++ + 0 200 Pixel ++ + + – + ADP Light ++ + + + + 10K 12 3 4 5 6 GFP 4K 25 kDa 0 10 100 1K 10K 0 200 Pixel GFP fluorescent (A.U.) fg Light Dark bRF F -PL ++ + – + o 1 GFP-DNA ++ + + – T7RNAP ++ – + + ATP – – – – – 0 200 Pixel ++ – + + Light GFP 25 kDa 0 250 Pixel Fig. 2 Protein synthesis inside giant unilamellar vesicle (GUV) driven by light. Green fluorescent protein (GFP) was synthesized from its messenger RNA (mRNA) (a–e) or DNA (f–h) inside light illuminated GUV (a, c, e–h) or in vitro (b, d). GFP was synthesized inside GUV (a) or in vitro (the PURE system) (b) in which the photosynthesized adenosine triphosphate (ATP) was consumed for the aminoacylation of transfer RNA (tRNA). The insets in a, e and g indicate plot profile of green and red colors on the thin yellow line. c Flow-cytometric analysis of the GUVs of a. The illuminated GUVs are shown as green, whereas the GUVs incubated in the dark are shown as black. The X- and Y-axes represent the fluorescent intensity and the area of forward scattering, respectively. d GFP synthesis coupled with guanosine 5’-triphosphate (GTP) generation. GFP was synthesized in the PURE system with or without nucleoside-diphosphate kinase (NDK), GTP, guanosine diphosphate (GDP) and adenosine 5’-diphosphate (ADP). e The same reactions as in lanes 4 and 6 of d were performed inside GUVs as indicated as NDK+ and NDK−, respectively. f GFP synthesis from its DNA. A gene of whole GFP was introduced in the PURE system with or without bRF F -PLs, T7 RNA polymerase (T7RNAP), ATP and light. g A small part of GFP (GFP11: 15 amino acids) o 1 was synthesized from its encoding DNA inside GUVs containing T7RNAP, another large part of GFP (GFP1-10) purified form E. coli cells, and the PURE system lacking NDK. h The same GUVs of g were analyzed by flow-cytometer as in d. The synthesized GFP in b, d, and f were labeled with [ S] methionine. Scale bar: 10 µm. Source data are provided as a Source Data file 4 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications Int. Int. NDK– NDK+ Int. Int. Int. FSC-A FSC-A Int. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE Next, we omitted GTP from the reaction mixture but the GUVs were not encapsulating liposomes, many larger-size introduced GDP and nucleoside-diphosphate kinases (NDKs), puncta appeared, suggesting aggregation of the synthesized bR- which allows the photosynthesized ATP to be consumed for GFP (Supplementary Fig. 17). These results imply that the synthesis of GTP that is a direct energy source of translation bR-GFP synthesized in the GUVs (Fig. 3b) localized onto the (Supplementary Fig. 10). The results showed that synthesized GFP internal PL membrane avoiding protein aggregation. The mem- was clearly detected by the sodium dodecyl sulfate–polyacrylamide brane localization of bR was further confirmed in vitro by flota- gel electrophoresis (SDS-PAGE) analysis when adenosine 5’- tion assay. When bR was synthesized in a standard PURE system diphosphate (ADP), GDP and NDK were added (Fig. 2d, lane 4). in the presence of liposomes, we found that the synthesized We performed the same reaction inside GUVs and observed bR appeared in the liposome fractions (Fig. 3c) after ultra- the fluorescence emission from the GUV lumen (Fig. 2e), centrifugation with a sucrose cushion, whereas almost all bR suggesting that the photosynthesized energy was consumed not appeared in the pellet fraction when liposomes were omitted. only for aminoacylation of tRNAs but also directly for translation This result directly shows the membrane localization of the de inside GUVs. novo bR onto the PL membrane. Additionally, the membrane In real cells, ATP is consumed not only as energy but also as a localized bR showed the proton-pump activity in response to substrate of transcription. To build up this, we performed a the duration of protein synthesis reaction (Fig. 3d). Here, 11, 61, transcription-and-translation coupled reaction in the artificial 124 or 233 bRs per one liposome were synthesized at the time photosynthetic cell system. When T7 RNA polymerase and of 10, 30, 60 or 180 min reaction, respectively (Supplementary template DNA coding-GFP were introduced into the PURE Fig. 18). These results lead us to conclude that the de novo system, photosynthesis of GFP was clearly detected by SDS- photosynthesized bRs spontaneously localized onto the internal PAGE analysis (Fig. 2f). However, we could not detect significant PL membrane and may have increased the proton-pump fluorescence by microscopy observation and flow-cytometer activity there. analysis when we performed inside GUVs. This is because the If the de novo photosynthesized bRs are functionable on the PL synthesized GFP level was lower than the detection limit. To membrane, the ATP production rate of PL should be enhanced overcome this problem, we applied the split-GFP method according to the increase of the number of bR per PL. To confirm developed by Cabantous et al. , i.e., GFP is split into two parts: this, we measured ATP concentration in the PURE system a small peptide (GFP11) and another large partner protein reaction mixture during the photosynthesis of de novo bR (GFP1-10). The fluorescence of GFP1-10 was restored by (bR ). In this experiment, we used the PLs consisting of a low wt incorporating GFP11 (Supplementary Fig. 15). Although the concentration bR (i.e. 5 µM bR) to emphasize the effects of intensity was rather weak, we observed the emission of GFP the de novo photosynthesized bR. The effect of the de novo fluorescence from the GUVs when GFP11 was photosynthesized photosynthesized bR was determined by comparing to the control from the template DNA (Fig. 2g). In this reaction, we experiment synthesizing a mutant bR (bR ) which does not mut encapsulated the PURE system modified for transcription-and- have any proton-pomp activity (Supplementary Fig. 19), and translation reaction (Supplementary Table 2), and the purified therefore the ATP production rate of the PLs containing bR is mut GFP1-10. The successful photosynthesis of GFP11 was also constant throughout the bR photosynthesis. In the case of bR wt confirmed in an in vitro reaction (Supplementary Fig. 16). By photosynthesis, the ATP concentration was higher than that in flow-cytometry analysis, we found that about 15% of the total the case of bR photosynthesizing in all three independent mut GUVs emitted significant fluorescence as a consequence of measurements (Supplementary Fig. 20), especially after 10 min reaction. This is consistent with the result of proton-pump transcription and translation inside (Fig. 2h). These results show that the photosynthesized ATP was consumed both as the activity of the bR synthesized in PURE system (Fig. 3d). substrates for mRNA transcription and as the energy for protein Here, the difference in the ATP concentration between bR wt translation, just as in real cells. and bR photosynthesizing reactions represents the effect of mut de novo photosynthesized bR . The ATP concentration in the wt bR -photosynthesizing reaction was approximately 1.5-fold wt Self-production of the artificial organelle components. The two higher than that of bR (Supplementary Fig. 20). It should be mut kinds of component proteins of the artificial organelle produced noted that the obtained ATP concentration indicates the net of ATP, and the resulting ATP drove protein synthesis. To test the photosynthesized ATP minus the consumed ATP for the whether our artificial photosynthetic cell system can synthesize protein synthesis. The synthesis rate of the bR and bR was wt mut the component proteins of its own artificial organelle, we tried adjusted to be the same by regulating the amount of template to photosynthesize bR, as well as F F . In this reaction, we used mRNA (Supplementary Fig. 21), and thus the ATP consumption o 1 the translation-only PURE system (see For mRNA start in Sup- rates were equal in both cases. To directly compare all three plementary Table 2). We expected that the newly photo- measurements, we normalized each obtained result with the synthesized de novo bRs localize onto the bRF F -PL membrane ATP per PL value at the endpoint time (120 min) of the de novo o 1 and increase ATP photosynthesis activity of the artificial orga- bR -photosynthesizing reaction. The ratio of the increased mut nelle as a consequence of activity enhancement in the proton artificial organelle activity is shown in Fig. 3e. We also confirmed gradient generation of the bRF F -PL (Fig. 3a and Supplementary the photosynthesized bR by SDS-PAGE analysis (Fig. 3f) in which o 1 −1 −1 Fig. 1E and F). The fluorescence of the synthesized bR, which the photosynthesis rate was 2.5 nmol ml min . After 2 h fused with GFP (bR-GFP), was mostly homogeneously observed of reaction, the number of working bRs per PL increased from inside the GUV lumen but not on the GUV membrane (Fig. 3b), 100 working bRs per PL (original) to 110 bRs per PL (after indicating that the synthesized bR-GFP localized onto the internal photosynthesis). These series of results indicate that the ATP PL membrane. This directed membrane localization is controlled production rate was enhanced during the photosynthesis of de by means of cholesterol which inhibits spontaneous membrane novo bR because the ability of proton gradient generation was wt integration of protein . We added 40% (mol%) cholesterol in the improved by increasing the number of functional bRs on a PL. lipid composition of GUV membrane but not in the internal Finally, we challenged to photosynthesize de novo F F in vitro o 1 PL membrane. When bR-GFP was synthesized inside GUVs and to observe the enhancement of ATP production activity containing liposomes, the same homogeneous fluorescent dis- of the resulting PLs. Unlike bR, F F consists of eight kinds of o 1 tribution was observed within the GUV lumen. In contrast, when subunit proteins. Thus, we first try to synthesize these eight kinds NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 a bc Light Dark Liposome Top F1 F2 F3 Bot. aa tRNA 25 kDa ATP ADP %Frac. 42 28 19 10 1 ARS 25 kDa aa-tRNA %Frac. 00 1 1 98 Min 0 30 60 120 Protein mRNA Rbs De novo 25 kDa bR d eh g 0.2 2.0 ×10 ** *** 4.0 1.5 F synth. with a o wt F synth. with a o mut 3.0 1.5 –0.2 1.0 2.0 –0.4 1.0 1.0 bR synth. –0.6 time 0.5 0 min 030 60 90 120 0.5 10 min –0.8 30 min Time (min) 60 min 180 min 0 i –1.0 Min 0 60 120 180 0 120 240 360 480 600 De novo F -a Time (s) o 25 kDa De novo F -b Fig. 3 Self-constituting protein synthesis in artificial photosynthetic cells. a Schematics of self-constituting protein synthesis. The numbers indicate the order of reactions; ➀: adenosine triphosphate (ATP) synthesis, ➁: aminoacylation of transfer RNA (tRNA) by aminoacyl-tRNA synthetase (ARS), ➂: translation by ribosomes (Rbs), ➃: de novo bacteriorhodopsin (bR) synthesis, and ➄: de novo F synthesis. b Light-induced bR-GFP synthesis in giant unilamellar vesicles (GUVs). Bar: 10 µm. c Membrane localization of bR. The bRs synthesized in the PURE system with or without liposomes were fractionated by ultra-centrifugation with sucrose cushion. The percentages of bR in each fraction (%Frac.) are indicated at the bottom of the gels. d Proton- pump activity of bR synthesized in the PURE system. The measurement was performed as in Fig. 1c. The reaction times of protein synthesis are indicated by different colors. The white and gray areas represent light ON and OFF, respectively. e Enhanced artificial organelle by de novo bR. Wild-type (bR ) wt or mutant (bR ) bRs were photosynthesized in the PURE system containing bRF F -PLs. The ATP concentrations at the time 2 h was measured and mut o 1 converted into ATP per proteoliposome (PL). The value of the de novo bR -containing PL was normalized to that of the de novo bR -containing PL. wt mut ***P < 0.001. f Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the de novo photosynthesized bR. g Light-driven ATP synthesis by PLs consist of cell-free synthesized F . Wild-type (a ) or mutant (a ) a-subunit protein was synthesized together with b- and c-subunits in o wt mut the PURE system containing purified F and bR-PLs. The measured ATP concentrations were converted into the produced ATP per PL. h Enhanced artificial organelle by de novo F . a or a was photosynthesized together with b- and c-subunits in the presence of purified F and bRF F -PLs. The ATP o wt mut 1 o 1 concentrations at the time 3 h was measured and converted into ATP per PL. The value of the de novo a -containing PL was normalized by that of the wt de novo a -containing PL. **P < 0.01. i SDS-PAGE analysis of the de novo photosynthesized F . P values were from two-side t-test. All experiments mut o were performed at least three times and their means and S.D. are shown. Source data are provided as a Source Data file of proteins by adding their corresponding template DNAs alanine , was synthesized instead of the wild-type a. This further into a standard PURE system supplemented with liposomes. supports that the cell-free synthesized F formed functional F F o o 1 However, unfortunately, we could not detect a significant activity onto the PL membrane and synthesized ATP, and thus we next of the F F due to low yields. We next synthesized only three tried to photosynthesize F and observed the enhancement of o 1 o component proteins of F , a-, b- and c-subunits, in the presence ATP photosynthesis activity in the resulting bRF F -PLs. The o o 1 of purified F and bR-PLs. After the reaction, the resulting photosynthesis reaction of F was performed in the translation 1 o bRF F -PLs were isolated from the reaction mixture and only system. The a-, b- and c-subunit proteins form the complex o 1 illuminated with supplying ADP. The result shows that ATP structure of F in the stoichiometry of 1, 2 and 10, respectively. photosynthesis of the PLs was detected in proportion to In order to find the best proportion of these three templates for illumination time, when wild-type a-subunit (a ) protein was obtaining the highest F F activity, we tested various proportions wt o 1 synthesized (Fig. 3g) with other b- and c-subunit proteins. This of the template DNA mix, at first. The multi-protein synthesis indicates the cell-free synthesized F localized onto the bR-PL for F was performed in the presence of liposomes and purified o o membrane and photosynthesized ATP by co-working with F . We detected the highest F F activity when 4, 2 and 10 nM 1 o 1 bR. Contrary, we could not detect any activity when a mutant template DNA of a-, b- and c-subunit, respectively, were added a-subunit (a ), which has an amino acid substitution at R169 to (Supplementary Fig. 22). Following this, we photosynthesized de mut 6 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications m = 260, R = 0.99 wt ΔpH Relative activity bR mut bR wt Relative activity Fo-a mut Fo-a wt Produced ATP/bRF F -PL o 1 NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE novo F component proteins in the presence of bRF F -PLs minimal number of enzymes and molecules since we used a o o 1 (0.3 µM F F and 140 µM bR) and purified F in the PURE system reconstructed artificial organelle and cell-free protein synthesis o 1 1 suited for mRNA start (Supplementary Table 2 and Supplemen- system . The functional significance in our artificial cell would −1 −1 tary Fig. 1F). The F photosynthesis rate was 50 fmol ml min accelerate the researches of artificial cell (or synthetic cell) in and reached to the plateau level at 7 h (Supplementary Fig. 23). the field of synthetic biology, as well as the development of a After the F photosynthesis reaction, PLs were isolated from biodevice sensing light and promoting protein and RNA synth- the reaction mixture and illuminated in the presence of ADP. In esis. For example, our artificial cell technique would be applicable order to distinguish the effect of de novo photosynthesized F ,we into the study of drug delivery that can control spatiotemporal also synthesized a instead of a and compared them, same as production of aptamer or single chain Fv within a vesicle capsule. mut wt in the case of de novo bR photosynthesis. The PL-containing More promising application of the artificial organelle is to use de novo F -a (PLs-a ) showed higher ATP photosynthetic as the phosphate recycling system in cell-free system. The o wt wt activity than the PL-containing de novo F -a (PLs-a ) in all current cell-free system is using creatine phosphate as a primary o mut mut three independent measurements (Supplementary Fig. 24A). The energy source; however, since this is unidirectional reaction, enhancement of ATP photosynthesis rate per PL was 1.38-fold. free phosphates accumulate in the system as the reaction goes Since we recovered the PLs-a and PLs-a from the reaction on. Our artificial organelle can avoid this problem by recharge wt mut mixtures, the amount of PLs analyzed was same in both samples the free phosphate onto ADP after the ATP consumption. (Supplementary Fig. 24B), and therefore the difference in Artificial cells have been employed as a model of protocell activities of PLs-a and PLs-a is thought to be reflecting the or primordial cell, which are thought to have existed before wt mut 2,18,25–28 enhanced activity by the de novo photosynthesized F . We also modern cells, in the study of origin of life . Especially, confirmed the same amount of F component proteins were how the primordial cell gained the ability to produce an energy photosynthesized in both samples (Supplementary Fig. 24C). to drive primitive metabolism is a big argument . The genes of Based on these results, we analyzed the enhanced ATP ATP synthase are highly conserved beyond the species and have photosynthesis rate in the resulting PLs. After 7 h of photosynth- been thought to exist from early stage of life . However, what esis, the net concentration of de novo photosynthesized F mechanism generated a proton gradient to drive ATP synthase within the reaction mixture was 20 nM (Supplementary Fig. 23). before the completion of the complicated electron transfer system Since 18 nM PLs were contained in the reaction mixture, is still unknown. Our work demonstrated that a simple bio-sys- statistically one de novo F was assigned to one PL. This reflected tem, which consists of two kinds of membrane proteins, is able into the enhanced ATP photosynthesis rate of the PL as 101.4 to supply sufficient energy for operating gene expression inside −1 −1 ATP PL min (Supplementary Fig. 24A), which represents a microcompartment. Thus, we think that primordial cells using −1 −1 −1 101.4 ATP F F min (turnover number: 1.7 s ). The sunlight as a primal energy source could have existed in the o 1 calculated result consistent to the specific activity of F F early stage of life before evolving into an autotrophic modern cell o 1 −1 −1 reconstructed into F F -PLs, 118±3.2 nmol ATP min nmol system. We believe the attempts to construct living artificial o 1 (Supplementary Table 1). On the other hand, when PLs-a was cell will reveal the boundary state of the transition from non- mut photosynthesized, the ATP photosynthesis rate showed 268 ATP living to living matters that actually happened in the early Earth −1 −1 PL min . Since five working-F F were contained in one PL, it environment. o 1 −1 −1 can be converted as about 50 ATP F F min that is lower o 1 than the de novo F activity which we cannot explain well. Methods Overall, the obtained result of the F photosynthesis seems o Materials. All reagents utilized in experiments were of the highest purity and reasonable. As we showed above, recursive production of F grade. These include POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine), cholesterol, PEG2000PE (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N- portion of F F was definitely observed, though it did not enhance o 1 [methoxy (polyethylene glycol)-2000] (ammonium salt)), soybean phosphati- exponentially. Although the photosynthesis level is still low, dylcholine (SoyPC) extract; liquid paraffin (Wako); pH gradient sensitive fluor- we engineered a self-constituting protein synthesis positive ophore ACMA (9-amino-6-chloro-2-methoxy acridine), protonophore FCCP feedback loop in the artificial photosynthetic cells. (carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone), ADP (adenosine 1 5 5’-diphosphate monopotassium salt), P ,P -di(adenosine-5′) pentaphosphate pentasodium salt (AP5A), potassium-specific ionophore valinomycin, Triton X-100, all-trans retinal (ATR) (Sigma); ATP (adenosine 5’-triphosphate), GTP Discussion (guanosine 5’-triphosphate), CTP (cytidine 5’-triphosphate), UTP (uridine We show that our artificial cell system containing the artificial 5’-triphosphate) (geneACT inc); OG (octyl β-D-glucopyranoside); DDM (n-dode- cyl-β-D-maltoside) (Dojindo); Tween-20 (Calbiochem); sodium cholate (Wako); organelle was able to first transduce light energy into an elec- and ATP bioluminescence assay kit CLS II (Roche). trochemical potential, and then convert into the chemical energy of ATP inside GUV. The produced ATP was consumed for the Isolation of purple membrane. Purple membrane patches containing firmly reaction of aminoacylation of tRNA as well as for the generation packed two-dimensional crystals of bR were isolated from Halobacterium sali- of GTP which was eventually consumed for translation. We also narum R1 following the previous protocol with slight modification. In brief, the showed that the produced ATP was converted into a mRNA that H. salinarum colonies were cultured in 6 L tryptone media containing 4 M NaCl subsequently translated into a part of GFP. The biochemical under high oxygen tension and continuous illumination with 200 W LED lamp for almost 192 h at 39 °C. The cells were then collected and resuspended in basal salt reactions performed in our artificial cell system mimic that is −1 in the presence of 17 µg ml of DNaseI (Sigma-Aldrich). This was followed by occurring in real living cells. Finally, we performed the photo- overnight dialysis against 100 mM NaCl at 4 °C. The translucent red lysate was synthesis of bR and F . The photosynthesized de novo bR loca- centrifuged at 45,000 × g for 60 min, and the precipitate was washed 6 times with lized onto the membrane of internal artificial organelle and 100 mM NaCl solution and distilled water. Eventually, the purple membrane enhanced the activity of ATP production, indicating the func- precipitate was overlaid on a 30–50% linear sucrose gradient and centrifuged at 100,000 × g for 17 h using SW28 rotor. At the end of the centrifugation, the purple tional engagement of protein synthesis and energy production band was collected. The sucrose solution was then removed by centrifuging the reactions. Because bR is the original compound of the artificial purple suspension at 45,000 × g for 1 h. The purple membrane sediment was organelle, we demonstrated that the artificial cell synthesized its resuspended and later stored in 50 mM Tris-HCl (pH 7.6), 150 mM NaCl and own part in a positive feedback loop. Furthermore, another 10% glycerol. Stock concentration of the purified bR was 450 µM. membrane-embedding component, F , was photosynthesized and its functional contribution in ATP photosynthesis was detected. It Overexpression and purification of bR from E. coli. C43(DE3) E. coli strain was should be noted that all these reactions were reconstructed with a transformed with pET21c-bR construct (see the DNA sequence in Supplementary NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 Table 5). The resulting recombinant bR was bearing hexa-his-tag at its C-terminus. Tris-HCl (pH 8.0), 150 mM NaCl and 10% glycerol. Stock concentration of the The bacterial culture was set in 6 L 2× YT media at 37 °C with shaking. The purified GFP1-10 was 100 µM. incubation continued until OD 0.6–0.7 and the culture media were supplemented On the other hand, the template DNA for in vitro or in vesicle synthesis of the with 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and 10 µM ATR. The smaller partner (GFP11) was prepared by PCR using P18 and P19 primers, where incubation was continued for 1 h at 30 °C and then for 3 h at 37 °C. The cells were P17 oligo was used as the template. The amplified DNA was cloned into pET29a collected and washed. The collected cells were disrupted by 5 passes of French press vector using NdeI and EcoRI. The resulting construct was used as a template homogenizer at 550 bar in a lysis buffer (50 mM MES (pH 6.0), 1 mM EDTA, DNA for the second PCR using T7 promoter and terminator primers (P22 and 300 mM NaCl, proteinase inhibitor cocktail). The membrane fractions were P23) to prepare the template DNA for GFP11 synthesis in the PURE system collected by centrifuging the pure lysate at 234,788 × g for 1 h. The collected (Supplementary Fig. 15). The further PCR for making the template DNA for membrane fraction was solubilized overnight at 4 °C in a buffer containing 50 mM GFP11 photosynthesis in vesicle (Fig. 2g, h) was carried out using P24 and P20 MES (pH 6.0), 300 mM NaCl, 5 mM imidazole and 1.5% DDM. The solubilized primers, and the resulting template DNA omits the non-translational sequence bR was purified in Ni-affinity chromatography (pre-equilibrated in a buffer con- after the stoop codon. taining 50 mM MES (pH 6.0), 300 mM NaCl, 0.2% DDM and 40 mM imidazole). The protein was eluted in a linear imidazole gradient of 40–300 mM. Further purification was done using Mono-Q column in the presence of 0.2% DDM. Reconstruction of PLs. The reconstitution of PLs with either bR or F F or the o 1 Finally, bR was eluted with linear NaCl gradient of 10–300 mM. Stock con- co-reconstitution of bRF F -PLs has been performed based on complete detergent o 1 10,19,34 centration of the purified recombinant bR was 1 µM. solubilization of liposomes following previous literature by the incorporation of the necessary modifications. Buffer PA6-5 was used as a reconstitution buffer unless otherwise indicated. First, lipid powder was suspended in buffer PA6-5 (pH 7.3), composed of 10 mM HEPES, 3 mM MgCl , 10 mM NaH PO and Expression and purification of recombinant F F and F . Thermophilic 2 2 4 o 1 1 200 mM sucrose, at a concentration of 41.3 mM. Later, the lipid suspension was Bacillus PS3 F F -ATP synthase (F F ) was overexpressed in DK8 E. coli strain o 1 o 1 (unc minus) carrying pTR19ASDSεΔc construct . The culture was made in 3 L completely solubilized by 6% octyl β-D-glucopyranoside (W/V) for about 1 h at room temperature. Then, the detergent was removed by first-round addition of 2× YT media for 21 h at 37 °C. The purification of F F was undertaken in o 1 accordance with previous work with modification. The cells were disrupted by a of 200 mg of pre-equilibrated SM2 Bio-Beads (Bio-Rad) and incubated at room temperature for 30 min. This was followed by the addition of bR (as a purple tip-sonication in a buffer containing 10 mM HEPES-KOH (pH 7.5), 5 mM MgCl , 10% Glycerol and 28 mM β-mercaptoethanol. The cell debris was removed by membrane), F F or both bR and F F at a given final concentration. The o 1 o 1 incubation was continued for additional 30 min, before the second-round centrifuging the lysate at 5500 × g for 30 min at 4 °C accompanied by collecting the membrane faction containing F F complex by ultra-centrifugation at addition of 300 mg of Bio-Beads by rotation mixing. After adding the second- o 1 round Bio-Beads, the proteoliposome was left mixing at room temperature for 225,000 × g for 1 h. The membrane fraction was homogenized in buffer I (pH 7.5) (10 mM HEPES-KOH, 5 mM MgCl , 2% Triton X-100, 0.5% cholate, 10% 90–120 min. Then, the turbid upper fraction of the proteoliposome was collected and stored at −80 °C until use. glycerol and protease inhibitor cocktail) and incubated at 30 °C for 30 min with mild shaking. The sample was then centrifuged at 311,000 × g for 20 min at 30 °C and supernatant was collected. To the supernatant, 70 ml of buffer II (pH 7.5) (20 mM potassium phosphate, 100 mM KCl, 24 mM imidazole and protease Light-dependent proton-pump activity of bR-PLs. For assaying proton pumping activity of bR, 800 µl of 1 R buffer (10 mM HEPES-KOH (pH 7.3) and 3 mM inhibitor cocktail) was added. This was later applied to Ni-NTA (Qiagen) pre- −1 equilibrated with buffer A (pH 7.5) (20 mM potassium phosphate, 100 mM KCl, MgCl ), pre-warmed at 37 °C, was supplemented with 30 µl bR-PL and 0.5 µg ml ACMA. Thereafter, the fluorescence trace of ACMA was measured every 0.2 s 20 mM imidazole and 0.15% DDM). Finally, the F F was eluted with buffer A o 1 (pH 7.5) containing 200 mM imidazole. The homogeneity and purity of the protein at the excitation and emission wavelength of 410 nm and 480 nm, respectively, was judged by 10–20% gradient SDS-PAGE. Finally, pure and homogenous using JASCO FP6500 spectrofluorometer that is customized by attaching a light fractions were collected and buffer was exchanged with a buffer containing 20 mM source (Tokina, Techno light KTS-150RSV, equipped with 15 V 150 W lamp KPi (pH 7.6), 100 mM NaCl and 0.15% DDM. This was concentrated eventually and 520 nm long-pass filter) at the lid. For the bR synthesized in the PURE system, by 50 kDa Amicon. Stock concentration of the purified F F was 99 µM. an assay buffer composed of 10 mM HEPES-KOH (pH 7.5), 5 mM MgCl and o 1 2 The F complex was purified following the previous literature by Suzuki 100 mM KCl was used. The proton-pump activity of bR was initiated by illumi- et al. . Briefly, E. coli cells expressing F F were disrupted and the cytosol fraction nating the sample with the light source. o 1 was obtained by centrifugation. The supernatant was incubated at 67 °C for 15 min, and the aggregated E. coli proteins were removed by centrifugation. The resulting yellow supernatant was subjected to a Ni-NTA column. The eluted F was Ultrapurification of ADP. The purchased ADP-monopotassium salt powder was incubated for 60 min at 25 °C, then ammonium sulfate was added to the solution. first dissolved with 10 mM HEPES-KOH (pH 7.3) at the concentration of 400 The resulting solution was applied to a phenyl-Toyopearl column. After washing, µmol, then applied to pre-equilibrated mono-Q column and eluted with linear salt the column was eluted with a linear reverse gradient of ammonium sulfate (1–0 M), gradient of 0–300 mM NaCl. The peak fraction for ADP was collected and then fractions containing F were collected and precipitated with ammonium lyophilized (using Taitec VD-800F freeze-dryer), and suspended with 10 mM sulfate. The F was further purified with a Superdex 200HR column. The purified HEPES-KOH (pH 7.3) at the stock concentration of 80 mM. protein was stored at −80 °C. The stock concentration of F was 5.7 µM. Light-driven ATP synthesis activity of bRF F -PL. For the ATP synthesis activity o 1 assay, 20 µl PL was mixed with 7.3 mM ultrapurified ADP and 10 mM NaN , and Preparation of split-GFP. The split-GFP was prepared as previously described . 3 The split-GFP was prepared firstly by introducing 7 point mutations into the gene the total volume to 40 µl with buffer PA6-5 was filled up. After illuminating the PLs for a given time length and light intensity with a halogen lamp, the ATP synthesis of superfolder-GFP (sfGFP), then by dividing into two portions. We used P1–P14 primer sets (Supplementary Table 3) to introduce point mutations N39I, T105K, was terminated by breaking the PLs with 2.5% trichloroacetic acid (TCA). This was accompanied by neutralizing the mixture with equal volume of buffer N (250 mM E111V, I128T, K166T, I167V and S205T, by quick change PCR. The DNA strand for a larger part (GFP1-10) of the split-GFP was amplified by PCR using P15 and Tris-HCl (pH 9.5) and 4 mM EDTA). Then, the mixture was injected into 800 µl of P16 primes and inserted into pET29a vector between the restriction enzyme sites of buffer R (pH 8.3) (20 mM tricine, 20 mM succinic acid, 80 mM NaCl and 0.6 mM KOH) which was pre-supplemented with 100 µl of luciferase/luciferin mix (CLSII). NdeI and EcoRI using infusion cloning technique. Thereby, a hexa-histidine-tag was introduced at the N-terminus of the open reading frame. The luminescence was measured (using AB-2270 ATTO luminometer) every 1 s for a total of 150 s. The trace of luminescence signal was used to calculate the The resulting construct, pET29aGFP1-10 (see the DNA sequence in Supplementary Table 5), was introduced into BL21(DE3) strain for overexpression synthesized amount of ATP based on the standard curve where the luminescence intensity is plotted against known concentrations of ATP. and purification of GFP1-10. The E. coli strain B21(DE3) harboring pET29aGFP1-10 was cultured in LB −1 media containing 50 µg ml Kanamycin. The culture was incubated under shaking at 37 °C until OD 0.6. Then, 1 mM IPTG was added for induction and the ΔpH-dependent ATP synthesis activity of F F -PL. The ΔpH-dependent ATP o 1 incubation continued for additional 6 h at 30 °C. The cells were then collected and synthesis activity of F F -PL was measured by acid-base transition o 1 10,24 washed one time. The collected cells were sonicated being suspended in a buffer assay . SoyPC extract was reconstituted with 1 µM F F to form F F -PL at the o 1 o 1 −1 containing 20 mM Tris-HCl (pH 8.0), 200 mM NaCl, 0.005 % Triton X-100 and concentration of 16 mg ml . This was used for acid-base transition assay. The protease inhibitor cocktail. After removing debris by centrifugation, the lysate was lumen of 30 µl F F -PL was acidified in acidification buffer (20 mM succinic acid, o 1 injected to His-Trap Ni-column which was pre-equilibrated with buffer A (pH 8.0) 0.6 mM KOH, 2.5 mM MgCl , 10 mM NaH PO (pH 4.5)) containing 286 nM 2 2 4 (20 mM Tris-HCl, 20 mM NaCl and 10 mM imidazole). After the washing, the valinomycin, 0.8 mM ultrapure ADP and 0.07 mM AP5A for 5 min at room protein was eluted with linear gradient of 10–300 mM of imidazole. Further temperature. The ΔpH-dependent ATP synthesis was assayed by injecting the purification was carried out using anion exchange chromatography (mono-Q acidified PL into a basic buffer (20 mM Tricine, 130 mM KOH, 2.5 mM MgCl , column) after exchanging the buffer with buffer C (pH 8.0) (20 mM Tris-HCl, 10 mM NaH PO , 7.3 mM ultrapure ADP (pH 8.8)) containing luciferin and 2 4 10 mM NaCl). The protein was eluted with linear gradient of NaCl from 10 to luciferase reagents. The generated ATP level was estimated by injecting 0.2 nmol 500 mM. The purified protein was stored at −80 C in a buffer containing 10 mM of ATP three times as a standard. 8 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 ARTICLE 35,36 Preparation of GUVs. GUVs were prepared as described previously , with two-step PCR (i.e., where P30 was used as a reverse PCR). The DNA templates for slight modification. GUVs were prepared from a fresh lipid-paraffin mix always. the cell-free expression of F complex (uncB, uncF and uncE) were prepared using Lipids dissolved in chloroform at a desired stock concentration were mixed toge- a primer set of P32 and P35, P36 and P37, or P38 and P39, respectively. Point ther considering the lipid-paraffin mix volume of 500 µl. This mixture was mixed mutation in uncB (R169A) was incorporated by overlap PCR using P33 and P34. well and flushed with a flow of N gas. To enhance the lipid mixing and completely Finally, T7 promoter, ribosome binding site and 3’ extra 14 nucleotides were added remove the remaining chloroform, the lipid-paraffin mix was heated at 80 °C for to it by P32 and P35. 20 min and vigorously vortexed. This was further flushed with a flow of N gas and the vials were sealed tightly. The vials containing the lipid mix were then sonicated Co-flotation assay. Spontaneous membrane integration of cell-free synthesized in warm water bath (55 °C) for 30–60 min. bR was performed as described previously with modification. bR was synthesized The lipid mix was let to cool at room temperature. Subsequently, 300 µl of the in the PURE system from 5 nM of bR template DNA (with upstream T7 promoter lipid-paraffin solution was mixed with 30 µl inner solution supplemented with 35 −1 and ribosome binding site) in the presence of [ S]methionine and 8 mg ml 200 mM of sucrose. Water-in-oil emulsion was created by gently pipetting up and SoyPC extract liposomes, which was filtered by 200 nm pore size membrane using a down for few 10 s. The emulsion was slowly laid over the outer solution (composed micro-extruder (Avanti Polar). The synthesis reaction was terminated by adding of equal volume of 400 mM glucose and PURE buffer, which is the same −1 RNaseA at the concentration of 20 ng µl followed by incubation at 37 °C for composition as GUV inside but without tRNAs) and kept on ice for 10 min. Next, 30 min. Then, 10 µl of synthesized bR (i.e., bR-PL) was overlaid on the top of the sample was centrifuged at 10,000 × g for 30 min followed by collecting the 0, 25 and 30% sucrose layers, where each layer was with a volume of 300 µl in precipitate by piercing the bottom of the flat-bottomed Eppendorf tube by 18-gauge sucrose flotation buffer (50 mM HEPES-KOH (pH 7.6), 100 mM KCl, 10 mM needle. Finally, the collected GUV suspension was centrifuged again at 9100 × g for MgCl ). This was centrifuged at 157,800 × g for 30 min using Beckman Coulter 10 min and precipitated GUV was resuspended in 50 µl of fresh outer solution. Optima Max XP Ultracentrifuge using TLS 55 rotor. Fractions of 200 µl were collected from the top in a total of four fractions. To the collected fractions, ice- Light-driven ATP synthesis inside GUV. The GUVs were prepared from lipid chilled TCA was added at the concentration of 10% and centrifuged at 20,400 × g mixture of POPC, cholesterol and PEG2000PE (5.75:4:0.25 molar ratio). The for 30 min at 4 °C. The supernatant was removed and the pellet was resuspended in bRF F -PLs were constituted from 176 µM bR and 1 µM F F with buffer-1 (pH o 1 o 1 200 µl of acetone and sonicated for at least 3 min. The suspension was centrifuged 7.3), and then encapsulated inside GUVs together with 7.3 mM ultrapurified ADP again at 20,400 × g for 30 min. The pellet was left at room temperature for 5 min, and 10 mM NaN . After the formation of GUVs, the outer buffer was replaced by a 3 mixed with 10–20 µl of loading dye and water-bath sonicated for 1 min. The fresh buffer-1 containing 200 mM glucose. Proteinase K (PK) was supplemented sample was finally analyzed by 15% SDS-PAGE. either to the outer or both the inner and outer solution at the concentration of 1 µM. After PK addition, the samples were incubated at 37 °C for 2 h prior Membrane orientation assay. Membrane orientation of the reconstructed bR- (to degrade unencapsulated bR and F F ) to the main light or dark incubation. o 1 PLs or F F -PLs was assessed by observing the binding affinity of each protein o 1 These GUV samples were incubated at 37 °C under light or dark. As a control, component to Ni-NTA Magnetic Beads, and 41.3 mM of lipid mixture (SoyPC an in vitro reaction mixture was also prepared in the same condition as the extract/cholesterol with the molar ratio of 70:30) was reconstituted with either F F o 1 encapsulated reaction mixture and illuminated. or bR at the protein concentration of 0.2 µM. The β-subunit of the recombinant F F and C-terminus of bR were bearing a His-tag. Magnetic beads conjugated o 1 Light-driven protein synthesis in GUV or in vitro. For the light-driven translation with Ni were used to trap His-tagged terminus of the proteins facing outside of reaction, bRF F -PL (176 µM bR and 1 µM F F ) was resuspended in buffer-1 o 1 o 1 liposome (cytosol side). As a control, PLs were solubilized with 0.5% Triton X-100 (Supplementary Table 4) supplemented with 500 nM of mRNA encoding sfGFP or and had to undergo the same procedure as the experimental samples. Finally, the bR::sfGFP (see the DNA sequence in Supplementary Table 5), 7.3 mM ADP, 10 mM collected fractions were run in 10–20% gradient gel and bands were visualized by NaN , 40 A per ml tRNA and 200 mM sucrose. This was mixed with the corre- western blotting with anti-His-tag antibody. sponding enzyme mix (Supplementary Table 2) and then encapsulated inside GUVs. The vesicle suspension was incubated for 6 h at 37 °C under the light illumination Photosynthesis of bR or F . A mutant bR (bR ) containing substitutions passing 500 nm long-pass filter. For the light-driven transcription-and-translation o mut of D85N and K216N was used as a negative control of photosynthesis of bR. reaction, the reaction mixture was prepared in same way, but using buffer-2 (Sup- The point mutations were introduced into pEXP-5-ct-bRwt construct, sequentially, plementary Table 4) supplemented with 1 nM DNA encoding GFP11 (see the DNA by quick change PCR using primer sets of P32-and-P33 and P34-and-P35 sequence in Supplementary Table 5) and 5 µM of the purified GFP1-10 protein. (see the DNA sequence in Supplementary Table 5). Next, linear DNA template The reaction was carried out for 7 h at 37 °C under the light illumination. was prepared using P22 and P23 to be used as a template for in vitro transcription. The vesicles synthesizing sfGFP were observed by confocal laser scanning For light-dependent expression of either bR or functional bR , the reaction microscopy (CLSM) (Zeiss LSM 550). For the population analysis, the vesicle mut wt mixture of PURE system was supplemented with bRF F -PL (where the PL was suspension was first diluted 10 times with buffer PA6-5 and then the fluorescence o 1 reconstituted with 5 µM bR and 1 µM F F ), 100 µM of ATR and either 0.4 µM of intensity of 100,000 vesicles was analyzed by fluorescence-activated cell sorter o 1 bR mRNA or 0.8 µM bR mRNA. The reaction mix was incubated at 37 °C (FACS Aria III). wt mut In vitro, the light-driven transcription-translation reaction was performed using under light and 2 µl sample was aliquoted at each given time for subsequent ATP level quantification. the same reaction mixture as mentioned above in the presence of [ S]methionine. The synthesized protein was subjected to 15% SDS-PAGE and visualized by On the other hand, for functional in vitro assembly of F F , the F complex o 1 o was expressed from the uncB, uncF and uncE-mRNAs (i.e., at the concentration autoradiography (Fujifilm). For the light-driven transcription-and-translation of 40 nM, 20 nM and 100 nM, respectively) coding for their respective subunits reaction, the sample was reacted for 13 h and subjected to column filtration a, b and c. The expression of the F complex was supplemented with 300 nM of (RNase-Free Micro Bio-Spin Columns with Bio-Gel P-30). Radioactivity of the o purified F complex for co-translational assembly of the two partners. The reaction resulting sample was counted by Liquid Scintillation Counter (ALOKA LSC-6100). mixture of PURE system (buffer-2 containing 200 mM of sucrose) was initially Concerning light-driven translation reaction using transient light illumination, seeded with 18 nM of bRF F -PL that was composed of 140 µM bR and 0.29 µM 500 μM of ADP, 500 nM mRNA coding GFP and [ S]methionine were o 1 F F . After incubating the reaction mixture for 7 h under light at 37 °C, 5 times supplemented in the translation reaction mixture. At a given time, aliquots were o 1 diluted PL was isolated by centrifuging at 230,000 × g for 30 min. The precipitated taken for 15% SDS-PAGE and ATP amount analysis. PL was resuspended in assay buffer (pH 7.3) (composed of 10 mM HEPES, 7.3 mM ADP, 10 mM NaH PO , 3 mM MgCl , 10 mM NaN and 200 mM sucrose), 2 4 2 3 bR::sfGFP synthesis in GUV. GUVs were prepared with 6 mM POPC and 4 mM making the final concentration of 44 nM. Later, the light-driven ATP synthesis cholesterol with encapsulating a cell-free reaction mix (PUREfrex Ver.2) together was assayed as described before. For negative control, the mutant uncB (R169A) with 5 nM linear DNA of bR::sfGFP construct (see the DNA sequence in Sup- −1 mRNA template was used. plementary Table 5), 150 µM ATR and 8 mg ml of SoyPC liposome. The reaction was performed in the presence or absence of liposomes for 4 h at 37 °C without light illumination. The fluorescence pattern of the synthesized bR::GFP fusion Reporting Summary. Further information on experimental design is available in protein in GUV lumen was analyzed by CLSM. the Nature Research Reporting Summary linked to this article. Preparation of bR or F DNA for the cell-free expression. The bR gene (with the Data availability codon optimized for expression in the PURE system) was amplified by primer P26 The authors declare that all the relevant data supporting the findings of the study are and P31, in the first step of PCR, whereas P21 and P31 primers were used in final available in this article and its Supplementary Information file, or from the step of PCR. Furthermore, for bR::sfGFP fusion construct, 3’ and 5’ overlap corresponding author (Y.K.) upon reasonable request. sequences were added to the bR and sfGFP gene by separate PCR using a primer set of P25 and P27 or P28 and P29, respectively. The resulting PCR products were used for overlap PCR in the presence of P25 and P29 as a forward and reverse Received: 16 August 2018 Accepted: 26 February 2019 primer, respectively. Eventually, ribosome binding site and T7 promoter site were added to the fusion construct by the primer P26 and P21, respectively, in the final NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09147-4 References 28. Kurihara, K. et al. Self-reproduction of supramolecular giant vesicles 1. Rasmussen, S. et al. Evolution. Transitions from nonliving to living matter. combined with the amplification of encapsulated DNA. Nat. Chem. 3, Science 303, 963–965 (2004). 775–781 (2011). 2. Luisi, P. L. The Emergence of Life: From Chemical Origins to Synthetic Biology 29. Deamer, D. & Weber, A. L. Bioenergetics and life’s origins. Cold Spring Harb. (Cambridge University Press, Cambridge, 2006). Perspect. Biol. 2, a004929 (2010). 3. Mann, S. Systems of creation: the emergence of life from nonliving matter. 30. Yoshida, M., Muneyuki, E. & Hisabori, T. ATP synthase–a marvellous rotary ACC Chem. Res. 45, 2131–2141 (2012). engine of the cell. Nat. Rev. Mol. Cell Biol. 2, 669–677 (2001). 4. Hutchison, C. A. 3rd et al. Design and synthesis of a minimal bacterial 31. Shiu, P. J., Ju, Y.-H., Chen, H.-M. & Lee, C.-K. Facile isolation of purple genome. Science 351, aad6253 (2016). membrane from Halobacterium salinarum via aqueous-two-phase 5. Xu, C., Hu, S. & Chen, X. Artificial cells: from basic science to applications. system. Protein Expr. Purif. 89, 219–224 (2013). Mater. Today (Kidlington, Engl.) 19, 516–532 (2016). 32. Suzuki, T., Ueno, H., Mitome, N., Suzuki, J. & Yoshida, M. F of ATP synthase 6. Shimizu, Y. et al. Cell-free translation reconstituted with purified components. is a rotary proton channel. J. Biol. Chem. 277, 13281–13285 (2002). Nat. Biotechnol. 19, 751–755 (2001). 33. Suzuki, T. et al. F F -ATPase/synthase is geared to the synthesis mode by 0 1 7. Kuruma, Y., Stano, P., Ueda, T. & Luisi, P. L. A synthetic biology approach to conformational rearrangement of epsilon subunit in response to proton the construction of membrane proteins in semi-synthetic minimal cells. motive force and ADP/ATP balance. J. Biol. Chem. 278, 46840–46846 (2003). Biochim. Biophys. Acta 1788, 567–574 (2009). 34. Rigaud, J. L., Paternostre, M. T. & Bluzat, A. Mechanisms of membrane 8. Ichihashi, N. et al. Darwinian evolution in a translation-coupled RNA protein insertion into liposomes during reconstitution procedures involving replication system within a cell-like compartment. Nat. Commun. 4, 2494 the use of detergents. 2. Incorporation of the light-driven proton pump (2013). bacteriorhodopsin. Biochemistry 27, 2677–2688 (1988). 9. Matsubayashi, H., Kuruma, Y. & Ueda, T. In vitro synthesis of the E. coli 35. Fujii, S., Matsuura, T., Sunami, T., Kazuta, Y. & Yomo, T. In vitro evolution Sec translocon from DNA. Angew. Chem. (Int. Ed. Engl.) 53, 7535–7538 of alpha-hemolysin using a liposome display. Proc. Natl Acad. Sci. USA 110, (2014). 16796–16801 (2013). 10. Kuruma, Y. & Ueda, T. The PURE system for the cell-free synthesis of 36. Fujii, S. et al. Liposome display for in vitro selection and evolution of membrane proteins. Nat. Protoc. 10, 1328–1344 (2015). membrane proteins. Nat. Protoc. 9, 1578–1591 (2014). 11. Furusato, T. et al. De novo synthesis of basal bacterial cell division proteins 37. Pykäläinen, A. et al. Pinkbar is an epithelial-specific BAR domain protein FtsZ, FtsA, and ZipA inside giant vesicles. ACS Synth. Biol. 7, 953–961 (2018). that generates planar membrane structures. Nat. Struct. Mol. Biol. 18, 902–907 12. Rampioni, G. et al. Synthetic cells produce a quorum sensing chemical signal (2011). perceived by Pseudomonas aeruginosa. Chem. Commun. (Camb.) 54, 2090–2093 (2018). Acknowledgements 13. Deisinger, B. et al. Purification of ATP synthase from beef heart mitochondria We thank Professor D. Oesterhelt and Dr. S.V. Gronau for the Halobacterium salinarum (F0F1) and co-reconstitution with monomeric bacteriorhodopsin into strain R1, Dr. Kazuhito Tabata for discussion on ATP synthase, Dr. Satoshi Fujii for liposomes capable of light-driven ATP synthesis. Eur. J. Biochem. 218, operation of FACS, Dr. Toshiharu Suzuki for the constructs of F F and Dr. T.Z. Jia for 377–383 (1993). o 1 advice on manuscript preparation. This work was supported by JSPS KAKENHI (Grant 14. Freisleben, H. J. et al. Reconstitution of bacteriorhodopsin and ATP synthase Numbers 16H06156, 16KK0161, 16H00797, J26106003 to Y.K. and 16H02089 to T.U.), from Micrococcus luteus into liposomes of the purified main tetraether the Astrobiology Center Project of the National Institutes of Natural Sciences (NINS) lipid from Thermoplasma acidophilum: proton conductance and light-driven (Grant Number AB291017 to Y.K.), and JST, PRESTO (Grant Number JPMJPR18K5 ATP synthesis. Chem. Phys. Lipids 78, 137–147 (1995). to Y.K.). 15. Matuschka, S., Zwicker, K., Nawroth, T. & Zimmer, G. ATP synthesis by purified ATP-synthase from beef heart mitochondria after coreconstitution with bacteriorhodopsin. Arch. Biochem. Biophys. 322, 135–142 (1995). Author contributions 16. Richard, P., Pitard, B. & Rigaud, J. L. ATP synthesis by the F0F1-ATPase from S.B. performed all experiments. Y.K. conceived the idea, designed most of the the thermophilic Bacillus PS3 co-reconstituted with bacteriorhodopsin into experiments, analyzed the data, prepared the figures and wrote the manuscript. T.U. liposomes. Evidence for stimulation of ATP synthesis by ATP bound to a gave advice for manuscript preparation. noncatalytic binding site. J. Biol. Chem. 270, 21571–21578 (1995). 17. Choi, H. J. & Montemagno, C. D. Artificial organelle: ATP synthesis from cellular mimetic polymersomes. Nano Lett. 5, 2538–2542 (2005). Additional information 18. Lee, K. Y. et al. Photosynthetic artificial organelles sustain and control ATP- Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- dependent reactions in a protocellular system. Nat. Biotechnol. 36, 530–535 019-09147-4. (2018). 19. Pitard, B., Richard, P., Dunach, M., Girault, G. & Rigaud, J. L. ATP synthesis Competing interests: The authors declare no competing interests. by the F0F1 ATP synthase from thermophilic Bacillus PS3 reconstituted into liposomes with bacteriorhodopsin. 1. Factors defining the optimal Reprints and permission information is available online at http://npg.nature.com/ reconstitution of ATP synthases with bacteriorhodopsin. Eur. J. Biochem. 235, reprintsandpermissions/ 769–778 (1996). 20. Deng, D., Jiang, N., Hao, S. J., Sun, H. & Zhang, G. J. Loss of membrane Journal peer review information: Nature Communications thanks the anonymous cholesterol influences lysosomal permeability to potassium ions and protons. reviewers for their contribution to the peer review of this work. Peer reviewer reports are Biochim. Biophys. Acta 1788, 470–476 (2009). available. 21. Bald, D. et al. ATP synthesis by F0F1-ATP synthase independent of noncatalytic nucleotide binding sites and insensitive to azide inhibition. J. Biol. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in Chem. 273, 865–870 (1998). published maps and institutional affiliations. 22. Cabantous, S., Terwilliger, T. C. & Waldo, G. S. Protein tagging and detection with engineered self-assembling fragments of green fluorescent protein. Nat. Biotechnol. 23, 102–107 (2005). Open Access This article is licensed under a Creative Commons 23. Nakamura, S., Suzuki, S., Saito, H. & Nishiyama, K. I. Cholesterol blocks spontaneous insertion of membrane proteins into liposomes of Attribution 4.0 International License, which permits use, sharing, phosphatidylcholine. J. Biochem. 163, 313–319 (2018). adaptation, distribution and reproduction in any medium or format, as long as you give 24. Kuruma, Y., Suzuki, T., Ono, S., Yoshida, M. & Ueda, T. Functional analysis of appropriate credit to the original author(s) and the source, provide a link to the Creative membranous Fo-a subunit of F1Fo-ATP synthase by in vitro protein Commons license, and indicate if changes were made. The images or other third party synthesis. Biochem. J. 442, 631–638 (2012). material in this article are included in the article’s Creative Commons license, unless 25. Adamala, K. & Szostak, J. W. Nonenzymatic template-directed RNA synthesis indicated otherwise in a credit line to the material. If material is not included in the inside model protocells. Science 342, 1098–1100 (2013). article’s Creative Commons license and your intended use is not permitted by statutory 26. Soga, H. et al. In vitro membrane protein synthesis inside cell-sized vesicles regulation or exceeds the permitted use, you will need to obtain permission directly from reveals the dependence of membrane protein integration on vesicle volume. the copyright holder. To view a copy of this license, visit http://creativecommons.org/ ACS Synth. Biol. 3, 372–379 (2014). licenses/by/4.0/. 27. Buddingh, B. C. & van Hest, J. C. M. Artificial cells: synthetic compartments with life-like functionality and adaptivity. ACC Chem. Res. 50, 769–777 © The Author(s) 2019 (2017). 10 NATURE COMMUNICATIONS | (2019) 10:1325 | https://doi.org/10.1038/s41467-019-09147-4 | www.nature.com/naturecommunications

Journal

Nature CommunicationsSpringer Journals

Published: Mar 22, 2019

There are no references for this article.