In Vitro Transcription–Translation in an Artificial Biomolecular Condensate

Biomolecular condensates are a promising platform for synthetic cell formation and constitute a potential missing link between the chemical and cellular stage of the origins of life. However, it has proven challenging to integrate complex reaction networks into biomolecular condensates, such as a cell-free in vitro transcription–translation (IVTT) system. Integrating IVTT into biomolecular condensates successfully is one precondition for condensation-based synthetic cell formation. Moreover, it would provide a proof of concept that biomolecular condensates are in principle compatible with the central dogma, one of the hallmarks of cellular life. Here, we have systemically investigated the compatibility of eight different (bio)molecular condensates with IVTT incorporation. Of these eight candidates, we have found that a green fluorescent protein-labeled, intrinsically disordered cationic protein (GFP-K72) and single-stranded DNA (ssDNA) can form biomolecular condensates that are compatible with up to μM fluorescent protein expression. This shows that biomolecular condensates can indeed integrate complex reaction networks, confirming their use as synthetic cell platforms and hinting at a possible role in the origin of life.

1. Detailed methods S2 2. Supporting tables and figures S8 S1: Overview of previous work on IVTT inside droplets S8 S2: Overview of systems used in this study S8 S3: Mixing order S9 S4: Bacterial cell lysate uptake in seven systems S10 S5: Final concentrations of charged species of various IVTT compositions S12 S6: IVTT uptake in seven systems S13 S7: deGFP expression as a function of IVTT composition S15 S8: Compatibility overview S15 S9: Turbidity measurements before and after droplet removal in four systems S16 S10: NPM1/rRNA droplet stability under reaction conditions S17 S11: Bleed-through from GFP-K 72 S18 S12: Phase-separation and fluorescence of GFP-K 72 -R97A mutant S18 S13: Partitioning into GFP-K 72 -R97A/ssDNA droplets S19 S14: Aggregation of lysate and feeding buffer at high concentrations S20 S15: Primers and important sequence information S21 S16: Spectrum of lysate-AF647 filter flow-through S22 NanoDrop One C using absorbance at 280 nm. The values were corrected using extinction coefficients based on protein sequence obtained from Protparam.
1.6 E. coli ribosomal RNA purification and labeling. rRNA was purified and labeled as described previously. 7 E. coli BL21 (DE3) cells were harvested from 1 liter LB at A 600 = 1.5 grown at 37 °C. Cells were pelleted and washed twice in lysis buffer (50 mM Tris-HCl pH 7.7, 60 mM potassium glutamate, 14 mM magnesium glutamate, 2 mM DTT). Cells were lysed at 1100 bar in a homogenizer (Stansted 'Pressure Cell' Homogenizer SPCH-EP, Homogenising Systems LTD) and cell debris was removed by centrifugation at 20,000 RCF for 25 minutes. Ribosomes were pelleted by ultracentrifugation for 3 hours at 50,000 RPM at 4 °C (Beckman Ti70.1 rotor). The ribosomes were resuspended in lysis buffer and rRNA was isolated using standard phenol-chloroform extraction protocols. E. coli rRNA concentration as determined using a NanoDrop One C and the 3'-end was labeled with AlexaFluor647-hydrazide (ThermoFisher) following Nelissen et al. 8 After labeling, rRNA was purified using isopropanol purification and ethanol purification or using an Amicon spin filter (Millipore). An agarose gel was used to check dye removal and sample concentrations were determined using the NanoDrop One C .
1.7 eGFP purification. E. coli BL21(DE3)/pLysS plus pET15b-His 6 -eGFP was grown overnight at 37 °C in 25 ml of LB medium containing 100 µg/ml ampicillin and 34 µg/ml chloramphenicol while shaking. The densegrown culture was used to inoculate 1 litre of the same medium and grown at 30 °C while shaking. At A 600 = 1.5, eGFP expression was induced by adding IPTG to 1 mM and cultivation was continued overnight. Cells were harvested by centrifugation at 5000 RPM in a Beckman JA-10 rotor at 4 °C for 10 minutes and the pellet was stored overnight at -80 °C. Next day, the pellet was resuspended in ice-cold buffer A (25 mM sodium phosphate buffer, 300 mM NaCl, 5 mM β mercaptoethanol, 0.5 mM EDTA, 20 mM imidazole, pH 8.0) containing 0.5 mg/ml lysozyme and 0. For labeling, relevant RNAP fractions were dialyzed against 100 volumes of labeling reaction buffer (100 mM NaCl, 10 mM sodium phosphate buffer pH 8, 0.1 mM EDTA, 0.1 mM DTT, 5 v/v% glycerol) in 3.5 kDa MWCO dialysis tubing overnight at 4 °C. Next morning, RNAP was dialyzed against fresh reaction buffer for 4 more hours and added directly to 1:5 RNAP:NHS-sulfoCy5 (Lumiprobe) molar ratio of dry NHS-sulofCy5 powder and a magnetic stirring bar. Conjugation took place at room temperature, for 4 hours, under continuous movement, and keeping the pH at 8 using sodium phosphate buffer. After 4 hours, 5 mM Tris-HCl pH 8 was added to react with remaining free NHS-sulfoCy5 for 1 hour at room temperature. The reaction solution was dialyzed in 3.5 kDa MWCO dialysis tubing against 1000 volumes of RNAP storage buffer (100 mM NaCl, 10 mM sodium phosphate buffer pH 7.5, 0.1 mM EDTA, 0.1 mM DTT, 50 v/v% glycerol) at 4 °C. Next day, the RNAP storage buffer was replenished and dialyzed for 4 more hours. RNAP-Cy5 was aliquoted and stored at -20 °C. Before use, RNAP-Cy5 was dialyzed for 5-6 hours in a 6-8 kDa MWCO D-Tube Dialyzer (Millipore) against 1000 volumes of RNAP working buffer (20 mM NaCl, 5 mM sodium phosphate buffer pH 7) at 4 °C and the RNAP-Cy5 concentration was determined using a Bradford assay.
1.9 E. coli ribosome purification and labeling. E. coli BL21 (DE3) cells at A 600 = 1.5 were harvested by centrifugation at 10,000 RCF for 10 minutes at 4 °C. Cells were lysed in monosome gradient buffer (20 mM Tris-HCl pH 8.0, 10 mM MgCl 2 , 140 mM KCl, 1 mM DTT, 0.1 mM EDTA, 20 U/ml Superase-In) using a Mini-Beadbeater (BioSpec). Cellular debris was removed by centrifugation at 33,000 RCF for 30 minutes at 4 °C and the supernatant was incubated at 37 °C for 80 minutes (ribosome run-off). Supernatant centrifugation was repeated and filtered using a 0.22 µm pore-size filter. Ribosomes in the supernatant were pelleted by ultracentrifugation for 3 hours at 211,000 RCF at 4 °C in a Beckman Ti70.1 rotor. Ribosome pellets were dissolved overnight at 4 °C in a small volume of monosome gradient buffer using a 10 RPM and 30° angle shaking program on a rotating mixer (no full rotations). Approximately 200 A 260 units of E.coli BL21 ribosomes were adjusted to 1 ml with monosome gradient buffer and loaded onto a 10-50% gradient of sucrose in monosome buffer in SW-28 tubes (Beckman) and centrifuged in a Beckman SW28-rotor for 20 hours at 58,000 RCF, at 4 °C. Gradients were harvested at 4 °C using a glass capillary connected to a peristaltic pump and a UV detector and the peak containing the 70S ribosomes 10 was fractionated in portions of ~1 ml. Ribosomes were subsequently pelleted by ultracentrifugation in a Ti70.1 rotor for 3 hours at 211,000 RCF, at 4°C. The pellet was dissolved in 500 µl of monosome buffer, the concentration determined, and flash frozen in liquid nitrogen and stored at -80 °C.
E. coli 70S ribosomes were labeled using a protocol optimized from the literature. 11,12 The labeling reaction was conducted in a buffer with the following composition: 50 mM Tris-HCl, pH 7.6, 15 mM MgCl 2 , 100 mM NH 4 Cl and 6 mM β-mercaptoethanol, containing 4.8 µM of ribosomes and 250 µM of DyLight650 NHS-ester (ThermoFisher). The reaction mixture was incubated at 37 °C for 30 minutes and precipitate was S4 removed by centrifugation for 1 minute at 10,000 RPM in a tabletop centrifuge. The supernatant containing ribosomes was concentrated in a centrifugal ultrafiltration device (Vivaspin 6, MWCO 30 kDa) and thoroughly washed with the labeling buffer to remove the excess of dye. The final concentration of labeled ribosomes was 6.3 µM and the DyLight650 concentration was 23.3 µM as determined using a Nanodrop 1000 spectrophotometer (Isogen). This corresponds to ~4 labels per ribosome. The ribosomes were flash frozen in liquid nitrogen and stored at -80 °C until further use.

E. coli lysate labeling and IVTT preparation.
The in vitro transcription-translation system used in the work has been previously described by Sun et al. 13 Some minor alterations were made. For lysate preparation, cell pellets were stored at -80 °C before lysis, cells were lysed using a cell homogenizer (Stansted 'Pressure Cell' Homogenizer SPCH-EP, Homogenising Systems LTD) at 1100 bar, and S30B buffer contained 14 mM magnesium glutamate and 150 mM potassium glutamate. Additionally, for the batch of lysate that was labeled with Alexa Fluor 647 NHS ester (ThermoFisher), a sodium phosphate buffer at pH 7.5 was used instead of Tris-HCl. Labeling was done following the instructions of the supplier, with an estimated molar dye concentration of <1:20 surface amines. Any remaining free dye was reacted with Tris-HCl and removed through overnight dialysis at 4 °C against 500 volumes of S30B buffer in 10 kDa MWCO dialysis cassette (Slide-A-Lyzer, ThermoFisher), followed by another 8 hours dialysis against 500 volumes of fresh S30B buffer. A small volume of labeled lysate was diluted 1:1 in S30B buffer and passed over a 10 kDa MWCO spin filter (Merck Millipore). A spectrum of the flow-through was taken (NanoDrop OneC) to check for remaining free dye (Fig. S16). Lysate-AF647 was aliquoted, flash frozen in liquid nitrogen, and stored at -80 °C.
Feeding buffer was either prepared in total as described in Sun et al., 13 but with the added amino acid mixture as described in Caschera and Noireaux, 14 or it was prepared in separate parts. For the partial preparation, the amino acids solution (AA), energy solution (ES) without 3-phosphoglyceric acid (3-PGA), and with 3-PGA were added separately to the IVTT reaction mixture. The typical lysate protein final concentration for an IVTT reaction was 10 mg/ml. For the feeding buffer components, the final concentrations can be found in Table S5. DNA was added as a linear fragment of p70a-deGPF or p70a-mmCherry.
1.12 Lysate mixing order and sequestration. Typically, coacervates (without lysate) were prepared by first mixing NaCl, Tris-HCl, MgCl 2 , MQ and the desired type of negatively charged species such as citrate, polyU, polyA, RNA, ssDNA, ATP or PAA in a microcentrifuge tube (0.5 ml, Eppendorf) at the required concentration, followed by the addition of positively charged (i.e. the relevant part) prot. sulf., spermine, GFP-K 72 , NPM1, pLys and PDADMAC from their respective stock solutions. The total volume was 20 µL. The S5 final concentration of NaCl is 0 or 100 mM and the final concentration of Tris-HCl and MgCl 2 are 50 mM and 5 mM, respectively. For NPM1/rRNA coacervates, the concentration of NaCl, Tris-HCl, and MgCl 2 are 0 to 150 mM, 10 mM, and 0 to 5 mM, respectively. Mixing was done by gentle pipetting. To test the uptake a minimal amount of lysate-AF647 (0.25 mg/ml final), coacervates were prepared as described above, but the negatively charged component, positively charged component, and lysate were added in a different order to the relevant mixture of MQ, salts, and buffer. The mixing order taken as the starting point for subsequent experiments can be found in Fig. S3. Droplets were observed in a passivized glass chamber using confocal microscopy.

IVTT sequestration.
The effect of an increasing ionic strength of an IVTT solution was explored similarly to lysate sequestration. Coacervate systems components were mixed first and a dilution of the full IVTT mixture (lysate plus feeding buffer) was added to the droplets. For the composition of the IVTT mixture, see Table S5. Here, it is important that the lysate and IVTT buffer are mixed in separately. Preparing a lysate/buffer mixture at increased concentrations (1.6x) leads to aggregation of the mixture (Fig. S13). The effect of the IVTT mixture on droplet morphology and lysate uptake was observed using confocal microscopy in a passivized glass chamber.

Coacervate compatibility with expression.
For the sequestering experiments, the various coacervate systems plus IVTT were prepared as described above. For each system, three samples were taken. For the control sample, a regular IVTT positive control mixture was prepared using the same batch of components used for the other samples. For the combined sample, the coacervate systems were mixed and combined with lysate and feeding buffer as described above. After 50 minutes incubation at room temperature, a small volume of droplets plus IVTT mixture was taken from the tube. For the depleted sample, the remaining volume was centrifuged for 5 minutes at room temperature and 5000 RCF in a tabletop centrifuge thereby pelleting the droplets (Fig. S9). A small volume of supernatant (dilute phase) was taken from the tube. For each sample, we tested the turbidity using absorbance measurements at 400 nm on a Tecan Spark plate reader. To determine the expression of each sample, 10 µl volumes of each sample was loaded onto a clear bottom 384-well plate and deGFP expression was followed using a Tecan Spark plate reader set to 30 °C for 16 hours. Each condition was tested in triplicate.

Stability in IVTT
mixture under reaction conditions. The stability over time was determined by incubating the droplets with a full IVTT system at 30 °C. Initially, GFP-K 72 /ssDNA, spermine/polyA, ATP/ pLys, and NPM1/rRNA were explored in this way. Here, coacervate droplets were formed first, after which the IVTT buffer was added, and lysate was added last. These were mixed by gentle pipetting. For the GFP-K 72 -R97A/ssDNA system, the lysate was added before the buffer (Fig. S12). To determine droplet stability, droplets were loaded into a passivized glass chamber and followed using an SP8x confocal microscope with a temperature control box set to 30 °C over a period of 16 hours. For the systems that did not prove stable in an IVTT reaction mixture, a reduced ionic strength IVTT mixture was used. For several systems, the effect of adding 1 U inorganic pyrophosphatase (IPP), 0.1 mM PMSF protease inhibitor, and/or 1 U Ribolock RNAse inhibitor (ThermoFisher) on stability was also explored.

Expression inside GFP-K 72 -R97A/ssDNA droplets.
Droplets were prepared from 24 µM GFP-K 72 -R97A, 0.05 mg/ml ssDNA, 5 mM Tris-HCl pH 7.5, 10 mg/ml lysate, and minimal ionic strength feeding buffer with the following changes: 1 mM amino acids, 30 mM 3-PGA, 6 mM magnesium glutamate, and no maltose. Feeding buffer either contained 10 nM p70a-deGFP linear fragment (DNA(+)) or no DNA ((DNA(-)). The total reaction volume was 20 µl. Droplets were incubated in 1.5 ml tubes (Eppendorf) at 30 °C in a thermoshaker for ~16 hours. After incubation, the tubes were spun briefly to concentrate all material in the bottom of the tube. The droplets were harvested with a pipette and put in a closed, passivized glass chamber and observed using confocal microscopy.

RNAP, ribosome, and eGFP partitioning.
Partitioning of E. coli RNAP-Cy5, E. coli ribosomes-DL650, and eGFP into GFP-K 72 -R97A/ssDNA droplets was determined using confocal microscopy. Coacervate droplets S6 were prepared in 1.5 ml tubes, RNAP-Cy5 (0.125 µM), ribosomes-DL650 (0.250 µM), and eGFP (7.5 µM) were added and incubated either in the tube or inside a closed chamber slide under various conditions. After incubation, the droplets were imaged using either the SP8x or SP8 liachroic confocal microscopes. The partitioning coefficient of each component into the droplets was determined by taking the average intensity of five droplets and dividing it by the average background intensity at five spots close to the droplets. All images were analyzed using ImageJ, including profile intensity measurements. Table S1. Overview of previous work on IVTT inside droplets. Also including the results from this work. Total expression = full systems of droplets and supernatant. Droplet phase expression = expression in droplet phase separated from dilute phase by centrifugation. In-droplet expression = expression levels of fluorescent protein inside actual droplets. Table S2. Overview of systems used in this work. References to various systems can be found in the main text. Figure S3. Mixing order. Mixing order-dependent effect of bacterial cell lysate on phase separation. Mixing order used as the basis for the lysate and IVTT experiments in Fig. 2, Fig. S4 Figure S3, except for spermine/polyA, where no labeled component was used. Additional salts and buffers are the same as in Figure S3. S11 5 mg/ml 10 mg/ml 1 mg/ml ATP/pLys 5 mg/ml 10 mg/ml 1 mg/ml  Fig. 2B and Fig. S6 at 0.5x. Reduced ionic strength: component concentrations used in Fig. 3A and beyond, unless specified otherwise. S12 S13 0.25x 0.5x 0.025x  Figure S3, except for spermine/polyA, where no labeled component was used. Additional salts and buffers are the same as in Figure S3, except for NPM1/rRNA, where 150 mM NaCl has been left out.  Figure S7. Expression of deGFP as a function of IVTT composition. Reducing concentrations of various feeding buffer components, and Mg/K-glut determined the minimal ionic strength condition that still gave appreciable expression. For selected concentrations see Table S5.  Table S8. Compatibility overview. Compatibility of each system with the corresponding step from Fig. 1. Green and vertically striped indicates broad compatibility. Red and diagonally striped indicates poor compatibility. Orange and horizontally striped indicates weak compatibility. Grey and clear indicates the systems has not been tested for the corresponding step. Definitions of when a system is considered compatible can be found in the main text. Briefly, mixing order (step 1) and lysate sequestering (step 2) required round, separated droplets without lysate aggregation, expression compatibility (step 3) required activity in the combined IVTT and droplet condition, and ideally no activity in the depleted condition, stability (step 4) required droplets to remain stable in the IVTT mixture under reaction conditions at 30 °C for 6 hours or more, and expression (step 5) required deGFP expression in the droplet sample. Some remarks: Spermine/polyU showed aggregation across all samples and was discontinued at the start. GFP-K72/tyRNA seemed less stable than the ssDNA variety. The ssDNA variety was preferred. IVTT mixture consisted of 10 mg/ml unlabeled lysate and the minimal ionic strenght feeding buffer as described in Table S5, but no linear fragment. Error bars represent standard error of N = 3.     (Table S5) is added before 10 mg/ml lysate. Condensate composition same as for S12A. Scale bars 50 µm. (D) GFP-K 72 -R97A/ssDNA droplet morphology when 10 mg/ml lysate is added before the feeding buffer mixture. Condensate composition same as for S12A. Compared to regular GFP-K 72 -R97A/ssDNA droplets in S12A, both S12C and S12D show clear lysate uptake in terms of droplet size. Scale bars 50 µm. (E) GFP-K 72 -R97A/ssDNA droplet stability in minimal ionic strenght IVTT (lysate, feeding buffer, no DNA) under expression conditions at 30 °C. Condensate composition same as for S12A. Scale bars 50 µm.  Table S15. Primers and important sequence information. Pertains particularly to site-directed mutagenisis for GFP-K 72 -R97A construction. Figure S16. Spectrum of lysate-AF647 filter flow-through. Flow-through was obtained from spinning diluted lysate-AF647 through a spin filter. Spectrum shows a peak for the nucleic acids passing the filter at 260 nm, and no peak around 647 nm, indicating little to no free dye remains in the sample.