Non-enzymatic oligonucleotide ligation in coacervate protocells sustains compartment-content coupling

Abstract

Modern cells are complex chemical compartments tightly regulated by an underlying DNA-encoded program. Achieving a form of coupling between molecular content, chemical reactions, and chassis in synthetic compartments represents a key step to the assembly of evolvable protocells but remains challenging. Here, we design coacervate droplets that promote non-enzymatic oligonucleotide polymerization and that restructure as a result of the reaction dynamics. More specifically, we rationally exploit complexation between end-reactive oligonucleotides able to stack into long physical polymers and a cationic azobenzene photoswitch to produce three different phases-soft solids, liquid crystalline or isotropic coacervates droplets-each of them having a different impact on the reaction efficiency. Dynamical modulation of coacervate assembly and dissolution via trans-cis azobenzene photo-isomerization is used to demonstrate cycles of light-actuated oligonucleotide ligation. Remarkably, changes in the population of polynucleotides during polymerization induce phase transitions due to length-based DNA self-sorting to produce multiphase coacervates. Overall, by combining a tight reaction-structure coupling and environmental responsiveness, our reactive coacervates provide a general route to the non-enzymatic synthesis of polynucleotides and pave the way to the emergence of a primitive compartment-content coupling in membrane-free protocells.

Fig. 1. Salt- and light-dependent phase behavior of end-interacting oligonucleotides and azobenzene cations.

a Sequence and hybridization of the self-complementary DNA Dickerson dodecamer (DD) used in the study (left), chemical structure of trans-azoTAB and reversible photo-isomerization to cis-azoTAB (right), and schematic representation of interactions involved in DD/trans-azoTAB complexation (middle). b Bright-field optical microscopy images of charge-balanced mixtures of DD (5 mM nucleobase concentration) and trans-azoTAB (5 mM) prepared at different NaCl concentrations, as indicated, showing the formation of irregular soft solids (100 mM NaCl), liquid crystal (LC) coacervate droplets (200 mM NaCl), isotropic (ISO) coacervate droplets (350 mM NaCl), or absence of phase separation (550 mM NaCl). Scale bars, 20 μm. Insets show zoomed images of the white squared areas under 90° crossed polarizers. Scale bars, 10 μm. c Schematic views of the self-assembled complexes produced at the different salt concentrations shown in (b). d Equilibrium phase diagram of charge-balanced mixtures of DD or DD₂ (5 mM nucleobase concentration) and azoTAB (5 mM total concentration) at varying trans:cis fraction (for DD) and at 100% trans-azoTAB (for DD₂) and at varying NaCl concentration. The complexes produced for each condition were characterized by optical microscopy (as shown in Supplementary Fig. 7) and classified as solid-like aggregates (dark gray), soft solids (light gray), LC coacervates (cyan), ISO coacervates (green), and no phase separation (orange). The fractions of cis- and trans-azoTAB produced under UV or blue light are also indicated (dashed vertical lines). e Optical microscopy images of DD/trans-azoTAB ISO coacervate droplets prepared at 350 mM NaCl under blue light or UV light, as indicated. Scale bars, 20 μm.

Fig. 2. Oligonucleotide ligation depends on the nature of the DDp/trans-azoTAB phase.

a Cropped gel images of a 15% PAGE run of the ligation product obtained after 24 h DDp/trans-azoTAB solutions (equimolar charge mixtures, 10 mM total charge concentration) in different phases, as labeled (soft solids: 80 mM EDC; LC droplets: 100 mM NaCl + 80 mM EDC; ISO droplets: 200 mM NaCl + 80 mM EDC; single phase: 400 mM NaCl + 80 mM EDC; supernatant obtained from centrifuged LC droplets). Ladder contains mixed 12-, 24-, 48- and 96-mer DNA oligomers derived from the repetition of the 12-mer DD sequence. b Fluorescence intensity (i_f, colored areas), extracted from gel images shown in (a) and cumulative weight distribution (C(n), colored dots) plotted as a function of the degree of polymerization, n, derived from rescaling (see Supplementary Note 3). Solid lines show fits of C(n) data points using Flory’s theory for linear polymer condensation (see Supplementary Note 2). c Time-dependent evolution of polymerization yield, p, for the different solutions prepared as in (a) (gray: soft solids, cyan: LC coacervates, green: ISO coacervates, orange: single phase, black: supernatant). Yields were extracted by analyzing samples from the same experiment (n = 1) run in the same PAGE gel (except for single phase, whose gel was run in parallel) shown in Supplementary Fig. 13 using two fitting procedures, as detailed in Supplementary Note 3, and are reported as mean ± SD of the two values obtained from fits. Error bars thus represent errors associated with uncertainties in the fits used to extract the reaction yields. Solid lines are fits using a catalyzed step-growth polymerization model to estimate the characteristic reaction time, τ (see Supplementary Note 4). d Weight fraction distributions, P(n), after 24 h of ligation reaction in the different solutions prepared as in (a) (black: supernatant, gray: soft solids, cyan: LC coacervates, green: ISO coacervates, orange: single phase). Solid lines show fits of P(n) data points using Flory’s theory for linear polymer condensation (see Supplementary Note 2). Source data are provided as Source Data files.

Fig. 3. Oligonucleotide ligation is reversibly controlled with light.

a Molar fraction of DD strands sequestered in ISO and LC coacervate droplets in the dark, under UV light, and under blue light (after exposure to UV light). The average value of three independent experiments (shown as black dots) is reported. Error bars represent the associated standard deviation. b, c Polymerization yield, p, calculated from fitted product distributions (see Supplementary Notes 2 and 3) for reactive LC (b, 100 mM NaCl + 80 mM EDC) and ISO (c, 200 mM NaCl + 80 mM EDC) coacervate droplets kept in the dark (black dots) or under constant UV light (purple dots), or exposed to cycles of UV and blue light (cyan or green dots, solid line; purple regions show when UV light was on). Normalized data for UV/blue cycles (open dots, dotted line) was obtained by subtracting the contribution of ligation occurring during UV exposure measured in dilute samples constantly exposed to UV light. Yields were extracted by analyzing samples from the same experiment (n = 1) run in the same PAGE gel (except for constant UV exposure, whose gel was run in parallel) shown in Supplementary Fig. 19 using two fitting procedures, as detailed in Supplementary Note 3, and are reported as mean ± SD of the two values obtained from fits. Error bars thus represent errors associated with uncertainties in the fits used to extract the reaction yields. Solid lines for dark and UV conditions are fits using a catalyzed step-growth polymerization model; other solid lines are guides to the eye. Source data are provided as Source Data file.

Fig. 4. Structural transformation of coacervates during DD oligonucleotide ligation.

a Bright-field optical microscopy images of charged-balanced reactive mixtures of DDp (5 mM nucleobase concentration) and trans-azoTAB (5 mM) prepared at different total ionic strength (soft solids: 80 mM EDC (top); LC droplets: 100 mM NaCl + 80 mM EDC (middle); ISO droplets: 200 mM NaCl + 80 mM EDC (bottom)) at t = 0 and after 24 h, as labeled. Scale bars, 20 μm. Insets show the white squared areas under 90° crossed polarizers. Scale bars, 10 μm. b Confocal fluorescence microscopy images of reactive ISO coacervate droplets (5 mM DDp nucleobase concentration, 5 mM trans-azoTAB, 200 mM NaCl + 80 mM EDC) doped with 1× SYBR Gold at t = 0 and after 24 h of reaction. Scale bars, 20 μm. Insets show zoomed images of the white squared areas. Scale bars, 10 μm. c Schematic representation of the transformation of ISO coacervate droplets into multiphase LC-in-ISO coacervate droplets during ligation. d Optical microscopy images of single DDp/trans-azoTAB ISO coacervate droplets prepared under reactive conditions (200 mM NaCl + 80 mM EDC) and exposed to cycles of UV and blue light (same light cycles as used in Fig. 3c, i.e., 0–4 h: dark, 4–8 h: UV, 8–24h: blue, 24–28 h: UV, 28–32 h: blue, 32–48 h: UV). Black arrows point to LC subdomains. Scale bars, 5 µm. e, f Optical microscopy images of single droplets before and after UV (e, t = 24 h) or blue (f, t = 28 h) light irradiation during the light cycles shown in (d). The dotted line identifies the LC (blue) and ISO (green) phases. Blue arrows point to the LC subdomains. Scale bars, 5 µm. g Schematic representation of the selective dissolution of the outer ISO coacervate phase with UV light in multiphase LC-in-ISO coacervate droplets. The remaining LC subphase containing longer polynucleotides acts as a seed for blue light-induced regrowth of the outer ISO coacervate phase.