Based upon the trend of higher yields from both longer sonication burst periods and longer total sonication time (Table 1 and Supplementary Figure S3), we increased the sonication burst period to 10 or 20 min with a 2-min 4°C cooling period. For a direct comparison to the traditional extract preparation protocol with a high-pressure homogenizer, cell aliquots from the same shake flask fermentations (with and without MOPS) reported in Supplementary Figure S1 were used with sonication. The protein production yields obtained with CFPS extracts prepared using the 10- or 20-min sonication burst and a 2-min cooling period are shown in Figure 2, A and B. Similar results were obtained from cells harvested from either fermentation (with or without MOPS). The extracts prepared by sonication were also suitable for CFPS using linear DNA templates (see Supplementary Figure S5). Also, continuing the sonication burst-cooling cycle beyond the first sonication burst-cooling cycle was not necessary to obtain a CFPS extract as productive as that obtained from high-pressure homogenization (Figure 2, A, B, and D). In addition, the CFPS extract productivity remained fairly constant over 100 min total sonication time (10 cycles for the 10-min sonication burst, and 5 cycles for the 20-min sonication burst) (Figure 2, A and B).
One concern of lysis by sonication is sample heating, and Kigawa et al. (4) postulated that sample heating is a possible reason why in their tests sonication was not suitable for preparing CFPS extract. The temperature of cell samples was measured over 120 min with continuous cycling between the 10- or 20-min sonication burst and 2 min cooling. The temperature of three independent experiments was measured by a K-type mini thermocouple, and the average remained below 15°C throughout the 120-min run (Figure 3, triangle and square data). In comparison, the average temperature of the cell samples processed by high-pressure homogenization was much higher with the maximum approaching 34°C. Also, sonicated CFPS extract productivity did not significantly decrease over 100 min of sonication, suggesting the extract was not damaged by heat.
The efficiency of lysis by sonication was also assessed relative to a French press-style high-pressure homogenizer by plating dilutions of lysate on LB broth agar Petri dishes. Lysis efficiency increased with total sonication time, with 10 cycles of 10 s sonication resulting in 98.898% lysis, 40 cycles of 10 s sonication resulting in 99.789% lysis, and 10 min of continuous sonication resulting in 99.988% lysis. The high-pressure homogenizer had the highest efficiency at 99.9996%. By comparing lysis efficiency alone, it appears that very high lysis efficiency (>99.98%) is needed for the consistent preparation of productive E. coli extract for CPFS. Although, other factors also play a role, such as the formation of soluble inverted membrane vesicles to facilitate oxidative phosphorylation CFPS, as reported with extracts prepared using a high-pressure homogenizer (34, 47).Performance of cell extract prepared using bead vortex mixing
With simplifying the extract preparation as our ultimate goal, we also sought to simplify the bead milling process for extract preparation by using a table top vortex mixer with 0.1-mm diameter glass beads. This method is significantly more economical than the commercial bead mill method (approximately $10,000–$40,000 for a commercial bead mill; Supplementary Table S1) compared with about $350 for the commonly available table top vortex mixer (Supplementary Table S1). Commercial bead milling equipment has been regularly used for protein purification (48, 49), DNA extraction (50), cell-free extract preparation (4, 51), and lipid extraction (52).
Initial attempts with bead vortex mixing were performed with 10%, 20%, 50%, and 80% (w/v) bead. Of the different combinations, only the bead concentration recommended by the manufacturer [80% (w/v) bead-to-cell buffer ratio] produced a significant amount of protein (results not shown) and was therefore chosen for further experiments. Also, cell lysates obtained by bead vortex mixing were observed to have a higher viscosity, and two centrifugation steps of 30 min each were required to adequately clarify the lysate. The protein yield obtained with extract prepared by cycling between bead vortex mixing and cooling in ice water resulted in lower protein production yields and higher extract-to-extract variability (Figure 2C and Supplementary Figure S4) as compared with protein production yields from extracts prepared by high-pressure homogenization and 10+ min of sonication. Lysis efficiency following two and five cycles of 10-min bead vortex mixing was 99.378% and 99.479%, respectively which is lower than that observed with the high-pressure homogenization and 10 min of sonication. The bead vortex mixing lysis efficiency observed is similar to that observed at lower sonication times, which also resulted in CFPS extract with lower protein yields and higher variability between replicate extract preparations. Increasing the vortex mixing time beyond 80 min seemed unreasonable for a streamlined extract preparation method. Also of note is the large temperature swings observed by cycling between bead vortex mixing and cooling, although temperatures higher than that obtained with the high-pressure homogenizer were not observed (Figure 3). Although there are challenges associated with this method, yields up to 600 µg/mL sfGFP production were obtained from CFPS extract prepared with bead vortex mixing.