The concentration of the E. coli cells used for extract preparation was determined using the wet-weight as measured following cell harvest (1 trillion E. coli cells per gram wet weight). Following lysis and before clarification by centrifugation, the lysate was diluted 50, 2500, 120,000, and 24,000,000 fold. Twenty microliters from the dilutions were plated on sterile Luria-Bertani agar Petri dishes (without antibiotics) and incubated overnight at 37°C. The lysis efficiency was determined by comparing the numbers of colonies on each plate to the total number of cells that would be present in the corresponding same volume and dilution if the cells were not lysed.Cell-free protein synthesis reaction
PANOx-SP energized cell-free reactions were performed on a flat bottomed 96-well black microtiter plate (Fisher Scientific, Waltham, MA, USA) with reaction volume of 15 µL at 37°C for 3 h with the following modifications: (i) radiolabeled leucine was not added to the reaction mixture, (ii) cell extract was used at 24%–25% (w/v) concentration, and (iii) BL21 Star DE3 E. coli cell strain was used for extract preparation, eliminating the need for addition of purified T7 RNA polymerase (13). Phosphoenolpyruvate and E. coli tRNA mixture were obtained from Roche Molecular Biochemicals (Indianapolis, IN, USA). All other reagents were obtained from Sigma-Aldrich. The super-folder green fluorescent protein (sfGFP) expression plasmid used in this work has been reported previously (18). The protein yield was determined by diluting the reaction volume to 60 µL with MilliQ water (Millipore, Billerica, MA, USA) in flat-bottomed 96-well black microtiter plates and measuring the fluorescence with a SYNERGY MX microplate reader (BioTek Instruments, Winooski, VT, USA) at a sensitivity setting of 50 and excitation/emission wavelengths of 485 and 510 nm, respectively. A calibration curve was used to determine sfGFP concentration (see Supplementary Figure S2).Results and discussion Shake flask cell culture
The use of shake flask fermentation and a common commercial E. coli strain, BL21 Star (DE3) for extract preparation simplifies the fermentation and eliminates the need to add independently purified RNA polymerase to the reaction or the acquisition of a specialized cell strain. While effective at producing E. coli cells viable for producing CFPS extract (7), the nutrient concentrations and pH are not directly controlled during shake flask fermentation. To assess the effect of the pH change during fermentation on extract viability, E. coli cells were fermented with and without the presence of 100 mM MOPS. Although, over the course of the fermentation, the pH exhibited a smaller change with MOPS (Supplementary Figure S1), the average sfGFP production yields from the extract prepared from cells grown in MOPS was within a standard deviation of the extract prepared from cells grown without MOPS (see Figure 2D). Thus the inclusion of MOPS in the fermentation did not significantly affect extract performance.
Performance of extract prepared with sonication
The purpose of using sonication cell lysis as opposed to the previously reported lysis by bead mill and high-pressure homogenizers is to reduce the capital cost and enable researchers without access to these specialized equipment the opportunity to assess and use CFPS for their desired application. The capital cost of the bead mill or high-pressure homogenizers used in this work and reported by others to produce CPFS extract is approximately $10,000 to $40,000 (Supplementary Table S1) compared with the commonly available sonicator, which costs about $4000 (Supplementary Table S1).
For cell lysis, the sonication burst periods and 4°C cooling periods were initially selected based upon commonly reported protocols for cell lysis, which use 10- to 60-s sonication burst periods and 3–10 cycles as shown in Table 1 (43-46). However, following these sonication protocols for cell lysis, the resulting CFPS extract produced protein at yields <25% of that achieved using a high-pressure homogenizer (Supplementary Figure S3). Also, different from the simplified Kim et al. (6) protocol using a French press-style high-pressure homogenizer (Figure 1), centrifugation for 30 min was needed to clarify the extract to appear similar to that obtained by high-pressure homogenization and 10 min of centrifugation. Although the sonication-produced extract was less productive than that produced by high-pressure homogenization, continued sonication burst-cooling cycles, and thus longer total sonication times, did in most cases result in a more productive CFPS extract (Supplementary Figure S3). However, a significant limitation was the variation in extract productivity observed among replicates (Table 1 and Supplementary Figure S3). This result was not surprising given that Kigawa et al. (4) claimed from their tests that sonication was not suitable for E. coli-based CFPS extract preparation. Although Kigawa et al. did not report the method used or the actual data from testing sonication; it is likely they also used common sonication protocols for E. coli cell lysis.