A Simple, Robust, and Low-Cost Method To Produce the PURE Cell-Free System

: We demonstrate a simple, robust, and low-cost method for producing the PURE cell-free transcription − translation system. Our OnePot PURE system achieved a protein synthesis yield of 156 μ g/mL at a cost of 0.09 USD/ μ L, leading to a 14-fold improvement in cost normalized protein synthesis yield over existing PURE systems. The one-pot method makes the PURE system easy to generate and allows it to be readily optimized and modi ﬁ ed.

It has been thought that precise control over the PURE system composition is required to achieve reproducible, and high protein expression yields and it has been suggested that a simple one-pot method would not be a viable option for robustly generating the PURE system [17]. However, we observed that variations in overnight culture densities (Supplementary Fig. S4) did not lead to substantial differences in OnePot PURE protein content ( Fig. 1d, Supplementary Fig. S2). We observed high protein expression robustness across four biological replicates, especially for the 38% and 47% EF-Tu formulations, with coefficients of variation (CV) of 8% and 12%, respectively (Fig. 1e, f). In comparison, the CV for a technical replicate of PURExpress and HomeMade PURE were 5% and 12%, respectively.
To avoid significant total protein concentration differences across replicates, we adjusted the concentration of the protein mixture to 1.6 mg/mL in the final reaction. This optimal concentration was chosen based on titrations of OnePot PURE (47% EF-Tu) replicate A ( Supplementary Fig. S5).
We compared the protein composition of our OnePot PURE system to the commercially available PURExpress (NEB) and our HomeMade PURE system prepared based on the Shimizu protocol with minor adjustments [7]. From gels we determined that the overall composition of the PURExpress and HomeMade PURE systems were quite similar to one another as expected (Fig. 2a). Both PURExpress and HomeMade PURE had a higher relative percentage of EF-Tu and a lower total protein concentration (1 mg/mL for HomeMade PURE) than OnePot PURE. The relative intensities of individual proteins in the OnePot PURE deviated from the PURExpress and HomeMade PURE standards although the protein expression yield of the OnePot PURE system (47% EF-Tu) was similar to PURExpress, 1.6 times higher than our HomeMade PURE and 5 times higher than TraMOS (Fig. 2b).
One of the main factors limiting the use of the PURE system is its high cost. We performed a detailed cost analysis of different PURE systems: two systems prepared from individually purified protein components (PURExpress and HomeMade PURE), as well as two systems prepared from batch cultures and pooled purifications (OnePot and TraMOS) (Fig. 2c,Supplementary  We demonstrated that it is possible to robustly produce a highly functional PURE system at low cost using a practical single batch culture and purification approach. The OnePot PURE system described here achieved a protein synthesis yield of 156 µg/mL at a cost of 0.09 USD/µL. At 1.7 µg/USD the cost normalized protein synthesis yield is over a magnitude higher than the commercial PURE system and substantially higher than TraMOS. We also showed that it is possible to adjust and optimize the OnePot PURE system by varying the inoculation fraction of an individual strain. This simple, low-cost, and robust protocol for producing the PURE system should broaden access to the technology and enable new applications which hitherto were not feasible due to the high cost and complexity of producing the PURE system.

OnePot protein preparation
Lysogeny broth (LB) used for OnePot protein component preparation was supplemented with 100 µg/mL ampicillin and all cultures were grown at 37 • C, 260 RPM. To allow for fast and easy inoculation, the different strains were stored as a glycerol stock in a single 96 well microplate. All overnight cultures were inoculated by a 96-well replicator (VP 408FS2AS, V & P Scientific), except for the EF-Tu strain, and grown in 0.3 mL of LB in a deep-well microplate (96 wells, void volume 1.5 mL). The strain expressing EF-Tu was grown in 3 mL of LB in a standard 14 mL culture tube. Overnight cultures (in total 3.6 mL) were used to inoculate 500 mL of LB media in a 1 L baffled flask. The exact composition of the inoculation cultures for different OnePot systems are given in Supplementary Table S4. Cells were grown 2 h before induction with 0.1 mM IPTG for 3 h, then harvested by centrifugation (4,000 RPM, 10 min, 4 • C) and stored at −80 • C overnight. Cells were resuspended in 7.5 mL buffer A and lysed by sonication on ice (Vibra cell 75186 and probe tip diameter: 6 mm, 4 × 20s:20s pulse, 70% amplitude). Cell debris was removed by centrifugation (15,000 RPM, 20 min, 4 • C). The supernatant was mixed with 2 mL of equilibrated resin, prepared 9 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018. ; https://doi.org/10.1101/420570 doi: bioRxiv preprint as described below, and incubated for 3 h, at 4 • C. After the incubation, unbound lysate was allowed to flow through the column. The column was washed with 25 mL of a wash buffer (95% buffer A, 5% buffer B) and eluted with 5 mL of elution buffer (10% buffer A, 90% buffer B). Instead of dialysis, buffer exchange was done using a 15 mL Amicon Ultra filter unit with a 3 kDa molecular weight cutoff (Merck). All centrifugation steps were performed at 4,000 RPM and 4 • C. The elution fraction was diluted with 25 mL of HT buffer and concentrated to 1 mL (2 × 60 min). The concentrated sample was then diluted with 10 mL of HT buffer, concentrated to 1.5 mL (60 min), and mixed with 1.5 mL of stock buffer B. The protein solution was then concentrated (14,000 RPM, 30 min, 4 • C) using a 0.5 mL Amicon Ultra filter unit with a 3 kDa molecular weight cutoff (Merck) and stored at −80 • C. Total protein concentration in the OnePot protein mixture was determined using a microplate Bradford protein assay with bovine gamma-globulin as a standard (Bio-Rad). Samples were diluted 1:25 and 5 µL of the diluted sample was mixed with 250 µL of Bradford reagent. Absorbance at 595 nm was measured using a SynergyMX platereader (BioTek). The OnePot protein mixture was then adjusted to a concentration of 12.25 mg/mL.

HomeMade PURE protein preparation
Proteins were prepared by Ni-NTA gravity-flow chromatography. The LB medium used was supplemented with 100 µg/mL of ampicillin and/or 50 µg/mL of kanamycin (Supplementary Table S5), and all cultures were grown at 37 • C, 250 RPM. Overnight cultures were grown in 3 mL of LB. Each strain was then individually inoculated in a flask with 2 L of LB. Cells were grown 2 h before induction with 0.1 mM of IPTG for 3 h, then harvested by centrifugation and stored at −80 • C overnight. The cells were resuspended in 30 mL of buffer A and lysed by sonication on ice (Vibra cell 75186 and probe tip diameter: 6 mm, 8 × 20s:20s pulse, 70% amplitude). Cell debris was removed by centrifugation (25,000 RCF, 20 min, 4 • C). The supernatant was mixed with 2-3 mL of equilibrated resin (described below), and incubated for 1-2 h, at 4 • C. After the incubation, unbound lysate was allowed to flow through the column. The column was washed with 30 mL of a wash buffer (95% buffer A, 5% buffer B) and eluted with 15 mL of an elution buffer (10% buffer A, 90% buffer B). The elution fraction was dialysed against HT buffer (2×) and stock buffer and stored at −80 • C.
Protein concentrations were estimated by absorbance at 280 nm and calculated protein ex-10 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018.

Cost calculations
To estimate the cost of PURE systems, we analyzed in detail the costs of the different subsets: protein components, ribosomes, and energy solution. The calculation for protein subset costs varies with the type of the system. For the TraMOS system, the reported cost of 0.052 USD/µL was used [17]. For our OnePot system, the cost was estimated based on the calculations given in Supplementary Table S1, with the assumptions that some of the materials can be reused and that four purifications can be done simultaneously in one working day. In the case of the HomeMade PURE system, our estimate was based on the price charged by the EPFL protein expression core facility: 300 USD per 2 L expression culture, which corresponds to our calculation for OnePot PURE of 83 USD per 0.5 L culture (332 USD for 2 L, Supplementary Table S1). Although the total price of this PURE system is high, the total amount of proteins purified is higher as well which can generate at least 40 mL of PURE HomeMade system (based on the volume of the protein limiting the preparation, in our case EF-Tu). Therefore, the price per µL of HomeMade protein components is 0.27 USD.
Two different possibilities were taken into account in the case of the ribosome subset. In 13 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018. ; https://doi.org/10.1101/420570 doi: bioRxiv preprint the first system, commercial ribosomes (Supplementary Table S6) were used for the PURE reactions (TraMOS). In the second system, purified ribosomes were used (HomeMade and OnePot PURE). The cost calculations for purified ribosomes are given in Supplementary   Table S2, with the assumptions that some of the materials can be reused and that hands-on time for one purification is a single working day.
The cost calculation for the OnePot energy solution is described in Supplementary Table   S3, with the assumption that half a day is necessary for the preparation of 20 mL of energy solution. For the TraMOS energy solution and the additional protein components, the costs were recalculated based on the component's price that would apply for the preparation of the given solutions (Supplementary Table S6). For some of the additional protein components, we were not able to determine the exact protein which was purchased and its amount used, mostly due to a difference in the type of units reported in the paper as compared to the units specified by the supplier. However, we arrived at a very similar cost estimate as given in the original calculation. Furthermore, we assumed that the work required for the solution preparation is taken into account in the purification cost calculation, so we did not consider it.
In the case of PURExpress, the total cost was based on the commercial price. The values used in the cost calculation were derived from experience with the actual experiments while preparing the different subsets. All costs for the different components were based on the prices given in our internal EPFL system when performing the calculation; no delivery costs were taken into account.

ACKNOWLEDGEMENTS
We thank Yoshihiro Shimizu (RIKEN) for kindly supplying plasmids for the PURE system, members of the EPFL Protein Crystallography and Protein Expression Core Facilities, especially Jean Philippe Gaudry for providing us with purified PURE protein components and help with ribosome purification and Nadanai Laohakunakorn for helpful discussions.
We also thank Corinna Tuckey and John DeMartino from NEB for providing us with custom PURExpress kits. This work was supported by theÉcole Polytechnique Fédérale de Lausanne.
14 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018.

DECLARATION OF INTERESTS
The authors declare no competing interests.

15
. CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available

16
. CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available  17 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available  18 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available  19 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018.  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Figure S4: OD600 measurement of overnight cultures used for inoculation of the mixed culture.

20
. CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available

21
. CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018. ; https://doi.org/10.1101/420570 doi: bioRxiv preprint 23 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available 25 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available The copyright holder for this preprint (which was this version posted September 18, 2018. ;https://doi.org/10.1101/420570 doi: bioRxiv preprint   27 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available     30 . CC-BY 4.0 International license under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available