Artificial morphogen-mediated differentiation in synthetic protocells

The design and assembly of artificial protocell consortia displaying dynamical behaviours and systems-based properties are emerging challenges in bottom-up synthetic biology. Cellular processes such as morphogenesis and differentiation rely in part on reaction-diffusion gradients, and the ability to mimic rudimentary aspects of these non-equilibrium processes in communities of artificial cells could provide a step to life-like systems capable of complex spatiotemporal transformations. Here we expose acoustically formed arrays of initially identical coacervate micro-droplets to uni-directional or counter-directional reaction-diffusion gradients of artificial morphogens to induce morphological differentiation and spatial patterning in single populations of model protocells. Dynamic reconfiguration of the droplets in the morphogen gradients produces a diversity of membrane-bounded vesicles that are spontaneously segregated into multimodal populations with differentiated enzyme activities. Our results highlight the opportunities for constructing protocell arrays with graded structure and functionality and provide a step towards the development of artificial cell platforms capable of multiple operations.


Supplementary Figures
Supplementary Figure 1. (a) Optical image showing diffusion of SDS (left side) and/or POM clusters (sodium phosphotungstate; right side) into a 2D droplet array. Experiments involving unidirectional diffusion of POM or SDS, or counter-diffusion of POM and SDS were undertaken. The custom-made acoustic trapping device consists of a square acoustic trapping chamber (20 x 20 mm) with four rectangular water chambers to provide cooling. SDS or POM was injected specifically from one edge of the trapping chamber to generate a unidirectional gradient of a single morphogen within the pre-organized array of PDDA/ATP coacervate micro-droplets. Alternatively, SDS and POM were injected simultaneously at the left and right edges of the square acoustic trapping chamber, respectively, to generate opposing reaction-diffusion gradients of the morphogens. (a-f) Time-dependent optical microscopy images recorded from an acoustically formed array of PDDA/ATP coacervate micro-droplets at 0, 3, 6, 9, 12 and 15 min after addition of a stirred solution of polyanionic POM clusters (sodium phosphotungstate; final concentration = 20 mM) into the sample chamber. Images show 5 droplets in a single row of the array viewed in the central observation window. The membrane-free coacervate droplets (a) immediately collapse on addition of the POM clusters (b-c) due to changes in osmotic pressure associated with the high POM concentration. As a consequence, transition of the droplets to spherical POM/coacervate vesicles (PCV) is kinetically inhibited. After 6 min, multiple internalized water micro-droplets (MCV) form in the collapsed droplets and induce re-swelling as the POM/PDDA membrane is established (d-e). The MCV intermediates subsequently transform over a period of ca. 3 min into the PCV morphological type (f). Scale bar = 100 µm. , 20 µm. The sample was prepared by adding 500 µL of SDS (40 mM) to a PDDA/ATP coacervate suspension (500 µL, 10 mM). After 30 min, the sample was centrifuged three times at 5000 rpm for 5 min to remove the supernatant and re-dispersed in Milli-Q water. The suspension was then lyophilized, and the obtained powder coated with a 30 nm-thick coating of carbon for SEM imaging and EDX analysis. Lyophilization resulted in extensive damage to the soft surfactant/polymer shell as shown in (a). Corresponding energy dispersive X ray (EDX) maps for tungsten (b), sulphur (c) and phosphorus (d) and analysis profile (d) showing co-location of POM (phosphotungstate) and SDS in the membrane. The low P count is consistent with the removal of ATP. Scale bars for b and c, 10 µm. The sample was prepared by adding a mixture of SDS and POM solution (500 µL, SDS/POM, 20/4 mM) to a PDDA/ATP coacervate suspension (500 µL, 10 mM). After 30 min, the sample was centrifuged three times at 5000 rpm for 5 min to remove the supernatant and re-dispersed in Milli-Q water. The suspension was then lyophilized, and the obtained powder was coated with 30 nm-thick coating of carbon for SEM imaging and EDX analysis. Images are recorded at 0 (a), 40 (b), 80 (c), 120 (d), 160 (e) and 200 s (f) after injection of a mixture of Amplex red and H2O2 followed by vigorous stirring to achieve homogenous concentrations (final concentrations in the chamber, 2.5 and 10 µM, respectively) of the substrates across the 2D array. Formation of the product, resorufin, gives rise to red fluorescence specifically in the PCV and PCB populations but not in the domains containing the PSWV and PSCV protocells, which remain dark. The differentiated protocells were prepared using opposing gradients of SDS and POM (SDS : POM = 2.3 (70/30 µL; 50 mM). Scale bar 500 µm.
Confocal microscopy imaging was performed by mounting the custom-made acoustic device on a Leica SP5-II laser scanning microscope attached to a Leica DMI 6000 inverted epifluorescence microscope and equipped with a ×10 or ×20 objective (0.4 NA and 0.7 NA, respectively). High contrast images against the background solution were obtained after ca. 45 minutes of in situ pattern formation. 3D reconstructions were processed with Icy software, and all images were consistent with the images shown in the main text. All glass slides used for imaging were functionalized with PEG-TMS.
Labelling of horseradish peroxidase RITC-labelled HRP was prepared by dissolving the enzyme (10 mL, 4 mg/mL) in sodium carbonate buffered solutions (100 mM, pH 8), followed by addition of a dimethyl sulfoxide (DMSO) solution of RITC (200 μL, 2 mg/mL). The reaction mixture was kept at 4 °C for 12 h, and then dialysed (molecular weight cut 15 kDa) against Milli-Q water over three days with regular changes in water. The fluorescently tagged enzyme was lyophilized and stored in the dark before use. UV/vis spectroscopy (ε(559 nm) = 6.21 x 10 4 M -1 cm -1 for RITC) typically gave a RITC : HRP molar ratio of 1 : 40.

Mapping of the morphological landscape
To determine the range of possible morphological types produced by the interaction of POM or SDS, or mixtures of the two additives on preformed membrane-free PDDA/ATP coacervate droplets, a matrix of 30 samples was prepared and investigated. In each case, the coacervate micro-droplets were prepared by adding ATP (1 mL, 50 mM) to a solution of PDDA (10 mL, 5 mM, monomer, 100 -200 kDa), followed by centrifugation of the suspensions at 2000 rpm for 5 min, and removal of the supernatant. The coacervate phase was then re-dispersed in 5 mL of Milli-Q water and solutions of the morphogens (total volume, 500 μL) added to 500 μL of the re-dispersed micro-droplets to give the following

Determination of CMC of SDS at different POM concentration
The CMC values of SDS at different concentrations of POM were determined by using pyrene as a fluorescent probe. The measurements were performed on a Fluoromax 4 fluorescence spectrometer with excitation wavelength of 334 nm and the emission spectrum was recorded from 350 to 450 nm; the excitation/emission slits were set as 4/2 mm. For each measurement, 1 μL of pyrene/ethanol solution (0.66 mM) was added to 1 mL of SDS and SPT solution. The ratio (I3/I1) between the intensities of the first (I1=372 nm) peak and the third (I3=383 nm) peak in the fluorescence emission spectra was used to determine the CMC of SDS.

Simulation methods
The simulated area consisted of the entire chamber (20 x 20 mm). Results are shown corresponding to the 5x5 mm square in the centre of the device (observation window; see Supplementary Figures 14 and 15). The simulations assumed that the chemical morphogens (POM or SDS) were injected from the left-hand side of the device and diffused along the x direction. In general, diffusion of SDS and POM was restricted to a 2D plane to simulate addition of the morphogens at the base of the acoustic trapping device (20 x 20 x 2 mm). Restricting the diffusion plane to a height (z) of 500 μm gave simulated induction times (ca. 3 min) commensurate with the experimental data for known injected morphogen concentrations. In contrast, larger values of z (2 mm) gave simulated induction times of around 15-30 min.
The simulated concentration gradient across the viewing area along the diffusion direction was defined as; ∆C = Cin -Cout. The 2D diffusion profiles of POM were directly determined using Fick's equations: where the diffusion coefficient of POM (D(POM)) is 2.48 x 10 -10 m 2 s -1 (Ref 1), t signifies time, and x and y are dimensions along and perpendicular to the diffusion direction, respectively (see Figure 14a). Simulations included: (i) plots of ∆C against time across the observation window and along the diffusion direction, and (ii) 2D plots of the spatial and temporal distributions of the POM concentration in a row of protocells aligned perpendicular or parallel to the diffusion direction in the centre of the viewing window (plots showing changes in concentration with time as a function of relative position in the row). The simulations were approximations as they did not consider binding of the POM clusters at the coacervate surface or morphogen depletion during the reaction-diffusion process.
Simulation of the SDS gradients was complicated by the continuous interchange of SDS between molecular dispersed monomers and self-assembled micelles (critical micelle concentration, C CMC = 8.2 mM; 62 molecules per micelle) (Ref.
2). The diffusion coefficients of the SDS molecule (C SDS Mol ) and SDS micelle (C SDS Mic ) were 5.3 x 10 -10 m 2 s -1 (Ref. 2), and 9.2 x 10 -11 m 2 s -1 (Ref. 3), respectively. Diffusion of both the molecular and micellular components were modelled using a finite-difference solution to the 2D Fick's diffusion equation. We assumed instant inter-conversion of SDS molecules and micelles during the diffusion process such that the interchange was determined by assessing the total surfactant concentration relative to the CMC. Hence, at each time step (typically 3s), the total SDS concentration at each location in the observation window was evaluated relative to the CMC and the concentration partitioned into molecular and micellular species according to: and < ℎ = (3) After partition, simulations of the diffusion gradients of the molecular and micellular components were undertaken by separate application of the finite-difference method for a discrete time step, after which the evaluation process was repeated. To ensure the stability of the simulation, the stability factor (Fo) was set to less than 0.25, where L was 100 µm (representing one step of the lattice) and ∆t was set as 3 s. Using a finite explicit approach, dynamic adjustments of the CMC of SDS in the POM concentration gradient were calculated in the same time interval (3 s) and the adjusted values inserted into the chemical concentration gradients and then iterated into the next round of calculations.