Protocellular CRISPR/Cas‐Based Diffusive Communication Using Transcriptional RNA Signaling

Abstract Protocells containing enzyme‐driven biomolecular circuits that can process and exchange information offer a promising approach for mimicking cellular features and developing molecular information platforms. Here, we employ synthetic transcriptional circuits together with CRISPR/Cas‐based DNA processing inside semipermeable protein‐polymer microcompartments. We first establish a transcriptional protocell that can be activated by external DNA strands and produce functional RNA aptamers. Subsequently, we engineer a transcriptional module to generate RNA strands functioning as diffusive signals that can be sensed by neighboring protocells and trigger the activation of internalized DNA probes or localization of Cas nucleases. Our results highlight the opportunities to combine CRISPR/Cas machinery and DNA nanotechnology for protocellular communication and provide a step towards the development of protocells capable of distributed molecular information processing.


Preparation of streptavidin-containing proteinosomes
In a typical experiment, BSA-NH2/PNIPAAm conjugates (final concentration of 8 mg ml -1 ), streptavidin (final concentrations of 4 or 10 μM) and 0.75 mg of PEG-bis(N-succinimidyl succinate) (Mw = 2,000, approximately 38 PEG units on average) were mixed in 15 μL of 50 mM sodium carbonate buffer (pH 9). A 300 μL volume of 2-ethyl-1-hexanol was immediately added and the mixture was shaken by hand for 20 s to produce a Pickering emulsion. After 2 h of sedimentation, the upper clear oil layer was discarded, 800 μL of 70% ethanol was added and the emulsion was gently shaken. The dispersion was then sequentially dialysed against 70% and 50% ethanol for 2 h each and finally against Milli-Q water for 24 h, resulting in polydisperse microcapsules (size range of 10-60 μm) containing a distribution of encapsulated streptavidin. The proteinosomes were then stored at 4°C for later use. To produce fluorescently labelled proteinosomes, a 1:7 mixture of labelled (FITC or DyLight 405)/unlabelled BSA-NH2/PNIPAAm conjugates was used.

DNA and RNA sequence design
DNA sequences were designed by hand with the help of a MATLAB script that generates random sequences with a desired fraction of each nucleotide. The sequences were screened with NUPACK [2] to detect any possible unintended interactions. Nicked T7 promotor regions of the transcriptional switches were based on previous publications. [3] The template strand of the genelet (denoted as T) were biotin functionalized to facilitate their binding to proteinosome-localized streptavidin. Transcriptional switches were activated using input strand that hybridize to the single-stranded region of the non-template strand (denoted as NT) and thereby complete the promoter region. Monitoring the activation of the transcriptional switches was achieved by either labeling the non-template strand with a fluorophore and the input strand with a corresponding quencher ( Figure 1) or by only labeling the input strands with a fluorophore ( Figure 3). All biotinylated gate complexes (denoted as FQ) in protocells were either labeled with a fluorophore and a corresponding quencher ( Figure 1) to observe the activation or only labeled with a fluorophore to detect the strand cleavage ( Figure 2). In addition, a nicked DNA target for (d)Cas9 with a removable quencher strand was designed based on earlier work [4] (Figure 3). Sequences of the Malachite Green (MG) genelet and the corresponding RNA aptamer are given in Supplementary Table S1. Sequences of the genelet, RNA signal and the fluorescent probe used in the 2-population signaling cascade are given in Supplementary Table S2. The sequences of double stranded DNA targets for (d)Cas9 and Cas12a with fluorophore and biotin modifications and also the corresponding guide RNA sequences are given in Supplementary Table S3. The sequences of the nicked dCas9 target, the corresponding genelet and the transcribed sgRNA are given in Supplementary Table S4. sgRNA synthesis sgRNA strands were synthesized with EnGenTM sgRNA Synthesis Kit, S. pyogenes (NEB), using manufacturer's suggested protocol. The sequences of the target-specific DNA strand and the resulting sgRNA strand are given in Supplementary  Table S3-4. Purification was performed with MonarchTM RNA Cleanup Kit (50 μg) (NEB) using manufacturer's suggested protocol. Yield was quantified using NanoDrop One (Thermo Scientific) spectrophotometer.

RNA synthesis
RNA strand were synthesized with T7 RNA Polymerase (ThermoFisher), using manufacturer's suggested protocol. The sequences of the DNA template (T1, NT1 and input1) and the resulting RNA strand (RNA1) are given in Supplementary  Table S2. Purification was performed with MonarchTM RNA Cleanup Kit (50 μg) (NEB) using manufacturer's suggested protocol. Yield was quantified using NanoDrop One (Thermo Scientific) spectrophotometer.

Labeling of T7 RNAP with Cy5
The Tris component in T7 RNAP was replaced with PBS by dialysis. T7 RNAP (4 μM) and Cy5 NHS (8 μM) were mixed in PBS to a final volume of 120 μL. The solution was incubated at room temperature for 45 minutes. Unreacted Cy5 was removed using Amicon Ultra 50kDa 0.5 mL centrifugal filter.

Labeling of dCas9 with Alexa546
Alex546-snap tag substrate was dissolved in fresh DMSO to a final concentration of 250 μM. EnGenTM Spy dCas9 (SNAP-tagTM) (4 μM) and Alexa546-SNAP-tag substrate (8 μM) were mixed in PBS with 1 μM DTT to a final volume of 30 μL. The solution was incubated in the dark at 37°C for 30 minutes. Unreacted substrate was removed using Amicon Ultra 50kDa 0.5 mL centrifugal filter.

Localization of genelets, FQ-probes, and DNA targets in streptavidin-containing proteinosomes
The buffer solution for all DNA localization experiments, unless otherwise specified, was 10 mM Tris (pH 8.0, Invitrogen) with 12 mM Mg 2+ (Invitrogen) and 0.1% vol/vol Tween 20 (Sigma). In a typical localization experiment, 10 μL of a dispersion of streptavidin-containing proteinosomes, 5 μL of 4x buffer and 2 μL of biotinylated DNA duplex (from a 10 μM stock solution) were gently mixed with a pipette in a 1.5 ml Eppendorf tube and incubated at room temperature for 1 h, followed by overnight incubation at 4°C. The excess unbound DNA strands were removed as follows: 10 μL of the supernatant was carefully removed from the top and discarded, 400 μL of buffer was added and the proteinosomes were resuspended by mixing with a pipette. The proteinosomes were allowed to sediment for 1-2 h or alternatively spun down using a microcentrifuge (4k RCF, for 5 min), then 400 μL of supernatant was removed from the top and discarded. This process was repeated, and the resulting suspension of streptavidin/DNA gate complex-containing proteinosomes was stored at 4°C.

Cy5-labeled T7 RNAP localization experiments
3 μL of proteinosomes with DNA genelet (Supplementary table S1) or 3 μL empty proteinosomes were suspended in NEB buffer 3.1 and then T7 RNAP-Cy5 (500 nM) was added to a final volume of 10 μL. The suspension was incubated in the dark at 37°C for 15 minutes and then imaged on glass slide using a confocal laser scanning microscope (CLSM, Leica SP8).

Alexa546-labeled dCas9 localization experiments
3 μL of proteinosomes with DNA target190 (Supplementary table S3) or 3 μL empty proteinosomes were suspended in NEB buffer 3.1 and then dCas9-Alexa546 (500 nM) and sgRNA (1 μM) were added to a final volume of 10 μL. The suspension was incubated in the dark at 37°C for 15 minutes and then imaged on glass slide using a confocal laser scanning microscope (CLSM, Leica SP8).

Design and fabrication of the microfluidic setup
We used a two-layer microfluidic chip to facilitate the physical trapping and in situ imaging of populations of streptavidin/DNA-loaded proteinosomes. The design of the chip ( Figure S16) is based on a previously shown device, [5] but was reduced in size to better facilitate its use in a temperature controlled setup. The chip consisted of a 1.5 mm X 2 mm or 1.5 mm X 1.5 mm localization chamber with PDMS pillars, a filtering chamber, inlet channels with pneumatically actuated Quake style push-up valves and an outlet channel. Master molds for the two layers were fabricated on separate silicon wafers (Silicon Materials) using standard photolithography techniques. [6] The molds for bottom and top layers were made by spin-coating SU8-3050 to a height of 50 μm, and spin-coating AZ 40xt to a height of 40 μm, respectively. After development the AZ 40xt mold was reflowed, resulting in rounded channels with a height of~60 μm at the center. The microfluidic chips were assembled from PDMS using standard multilayer soft lithography techniques [6] and plasma bonded to rectangular #1.5 glass coverslips.

Proteinosomes experiments in microfluidic devices
The microfluidic chip was inserted into a custom-made PID-controlled temperature control device ( Figure S16). The temperature regulator with the microfluidic chip was then mounted on the stage of a confocal laser scanning microscope (CLSM, Leica SP8). Control channels were filled with MilliQ water and actuated using a pneumatic valve array (FESTO), which was in turn actuated using a programmable logic controller (PLC, WAGO Kontakttechnik GmbH). The PLC was connected to the Ethernet port of a PC and controlled using a custom Matlab GUI. The pressure in the control channels was 2 bar. The pressure to the inlet channels was controlled using adjustable pressure regulators (Flow-EZ, Fluigent). In a typical experiment, buffer solution was connected to inlet port 1. First, air bubbles were pushed out of the flow channels by pressurizing the buffer channel at 1 bar and closing all other inlet and outlet valves, followed by thoroughly washing all the flow channels using the buffer solution. Next streptavidin/DNA-containing proteinosomes were loaded into the trap array from inlet port 2 at a pressure of 10 mbar. The inlet port 2 was then closed and the proteinosomes were gently washed (the exact pressure needed to achieve enough flow for effective washing, but without forcing the protocells out of the traps depended on the fluidic resistance of the microfluidic setup and was experimentally determined beforehand; 5 to 15 mbar was normally used) with buffer solution for 1 to 2 min to remove any unbound DNA. For multi-population experiments, in which it was critical that the unbound DNA was reduced to the minimum to avoid leakage reactions, the populations were loaded sequentially with intermittent washing periods for 1 to 2 min with a pressure of 5 to 15 mbar. To ensure that protocells of different populations were well mixed, additional mixing steps were performed when necessary. Mixing was achieved by first flowing the prococell suspension in reverse direction into one of the input channels and then back into the trapping chamber, this process randomizes the spatial distribution of protocells from different populations. The buffer solution for washing/filling the device and loading/washing the proteinosomes consists of 10 mM Tris (pH 8.0), 12 mM Mg 2+ and 0.1% v/v Tween 20. The temperature of the device was then set to 37°C. The confocal microscope was focused on the trapping chamber and time-lapse imaging was started. The initial steady-state signals were recorded as the baseline values.
The reaction buffer for all experiments in a microfluidic device, unless otherwise specified, was 1X NEB RNAPol reaction buffer supplemented with 15 mM MgCl2, 3 mM each NTP and 0.1 mg/mL BSA. Reagents (input DNA, T7 RNAP, Cas12a, Cas9, dCas9) were diluted (and mixed) to the desired concentration in the reaction buffer. The reactions were started by flowing the reagent solution into the trapping chamber using 10 mbar pressure for 20 s. The inlet and outlet valves were then closed and kept closed throughout the experiment.

Data acquisition and analysis
Fluorescence data were acquired using a confocal laser scanning microscope (CLSM, Leica SP8) equipped with solid-state lasers (405 nm for DyLight405, 488 nm for FITC, 552 nm for Cy3 and Alexa546, 638 nm for Cy5) and a hybrid detector. The time-lapse measurements were performed with a ×10/0.40 numerical aperture (NA) (1.55 × 1.55 mm2 field of view, 7 μm slice thickness) objective at a resolution of 512 × 512 pixels. The photon counting mode of the hybrid detector was used.
The RFU-to-concentration conversion factor was determined by measuring the average RFU (value across a horizontal line through the device) of specific DNA gate complex (0.25 μM, 0.5 μM, 1 μM and 2 μM) that was flown into the device. The conversion factor was then determined by plotting the RFU vs the concentration of DNA gate complex ( Figure S17).

Zeta potential measurement
All measurements ( Figure S1) were performed using a Malvern zetasizer nano-ZS instrument with a temperature controller. The samples (0.1 mg/ml) were measured in PBS buffer (pH 7.4, 0.1 mM) at room temperature.

Gel electrophoresis
Native Polyacrylamide gel electrophoresis (native-PAGE) was used to confirm the production of RNA from genelet circuit encoding aptamer sequence ( Figure S2). Gels was prepared at 12% monomer concentration of acrylamide in gel buffer (44.5 mM Tris, 44.5 mM Boric acid, 11.5mM MgCl2, pH 8.0). The gel was run in gel buffer for 1.5 h at 150 V at room temperature and post-stained with SYBR Gold. GeneRuler DNA ladder was included as a reference. Gels were imaged using an ImageQuant 400 Digital Imager (GE Healthcare).

Fluorescence measurements on a plate reader
Batch testing of the fluorescent malachite green aptamer transcription ( Figure S2), genelet-based activation of F1Q1 probe ( Figure S5), (d)Cas9 probe ( Figure S11) and genelet-based activation of dCas9 ( Figure S14) were performed at 37°C on a Biotek Synergy H1m plate reader using 384 well low-volume plates (Nunc) with a reaction volume of 12 μL. Excitation and emission wavelengths were 640 nm and 681 nm respectively. Fluorescence measurements were performed every 30 s.

Supplementary Figures
Supplementary Figure S1 In all experiments the concentrations of probe was 500 nM. Activation was only observed when (d)Cas9 (500 nM) and the corresponding sgRNA strand RNA4 (500 nM) were both present. A scrambled sgRNA (500 nM) together with dCas9 (500 nM) were used as control experiments. No (d)Cas9 and sgRNA were added in the blank. All experiments were performed in reaction buffer at 37°C (supplementary methods, plate reader). c) The RFU-to-concentration conversion factor was determined by fluorescence measurement over a range of concentrations (0.1, 0.2 and 0.4 μM) of active probes at 681 nm (excitation wavelength 640 nm). The data was fitted through a straight line and coefficient of determination (R 2 ) was calculated.