Skip to main content
NCBI home page
JACS Au logoLink to JACS Au
. 2025 Jun 16;5(7):3533–3544. doi: 10.1021/jacsau.5c00568

Controlled Formation of DNA Condensates as Model Nuclei in Monodisperse Giant Vesicles

Ryotaro Yoneyama 1, Naoya Morikawa 1, Ryota Ushiyama 1, Tomoya Maruyama 2,4, Reiko Sato 1, Mamiko Tsugane 1, Masahiro Takinoue 2,3,4, Hiroaki Suzuki 1,*
PMCID: PMC12308404  PMID: 40747041

Abstract

Several studies have attempted to replicate the complex hierarchy of eukaryotic cells for the bottom-up construction of artificial cells. Specifically, reconstruction of liquid–liquid phase separation systems as membrane-less organelles is one of the key focuses of this research field, with DNA condensates acting as versatile building blocks whose associative interactions can be precisely controlled via sequence design. However, such control is only possible at the nanoscale as control over the size and morphology of the lipid vesicles and liquid–liquid phase separation systems at the meso-to-microscale is determined by the kinetic aspects of their formation processes. Microfluidics is well-suited for controlling dynamic molecular assemblies at the cellular scale. In this study, we report the controlled condensation of DNA nanostars in mass-produced monodisperse giant vesicles (GVs) generated using a microfluidic device by manipulating the concentrations of DNA and salt associated with the GV volume changes. Our approach facilitates the precise control of the dynamics of DNA condensate formation, final size of condensates, formation of multiple condensates, and reversible formation/dissociation of condensates in GVs serving as a chassis for an artificial cell. Furthermore, our approach eliminates the need for thermal annealing prior to DNA condensation, supporting the coexistence of enzyme-containing biochemical reaction systems, such as gene expression systems.

Keywords: artificial cell, artificial organelle, giant vesicle (GV), microfluidics, DNA condensate

graphic file with name au5c00568_0010.jpg

graphic file with name au5c00568_0008.jpg

1. Introduction

Giant vesicles (GVs) are widely used as containers for artificial cells in bottom-up biology. − Many studies have attempted to encapsulate reaction systems with various components, such as nucleic acid replication, amplification, ,, and transcription/translation systems, , into GVs. Based on previous studies on artificial cell reconstruction, extensive efforts have been made to construct hierarchical internal structures to recapitulate complex functions. − Eukaryotic cells possess multiple membrane-bound organelles, such as the nucleus, mitochondria, endoplasmic reticulum, Golgi apparatus, and lysosomes, which locally change reaction systems and environments different from the cytosol and segregate macromolecules and membrane proteins. Because of these hierarchical structures, eukaryotic cells perform more specialized molecular processing functions than prokaryotic cells.

Recently, an extensive line of studies highlighted the importance of droplet-like substructures, such as stress granules and ribonucleoprotein particles, in living cells. − These droplets reversibly formed via liquid–liquid phase separation (LLPS) regulate the localization of specific molecules and reaction systems. Increasing interest in artificial LLPS systems has prompted attempts to reconstruct these systems in membrane-bound artificial cells. For example, Keating et al. showed that aqueous droplets generated via segregative interactions between two immiscible polymers, wrapped by a layer of small unilamellar vesicles, control the partition of molecules into distinct phases. The LLPS system formed via electrostatic associative interactions of the polycation and polyanion, termed a complex coacervate, is also a possible component of artificial cells. Previously, controlled formation of complex coacervate droplets in GVs was demonstrated using pH changes, optical stimulation, and membrane permeation of anionic small molecules. − LLPS in GVs exhibits adhesion via wetting, endo/exocytosis-like encapsulation, and passage across membranes by interacting with the membranes. , Compared to organelles consisting of lipid membranes, these membrane-less organelle models facilitate the easy regulation of molecular transport/storage. Lipid membranes control the selective transport and permeation of molecules via membrane proteins and remodeling. In LLPS systems, these functions can be controlled by modulating the polymer concentration and affinity. Therefore, control over formation and dissolution of LLPS systems in GVs would facilitate the construction of artificial cells whose hierarchical structures and functional partitions can be dynamically controlled.

Among various LLPS systems, DNA condensates (DNA droplets and hydrogels) are attracting considerable attention as artificially programmable nanomaterials whose physical properties and behaviors can be controlled via sequence design. − Particularly, condensates of branched DNA nanostructures (DNA nanostars) formed by oligomers have been extensively studied. By adding sticky ends (SEs) <10 bp to the tips of multiarm structures (e.g., Y- and X-motifs), microscopic condensates are formed via complementary bonds between SEs, creating fluid droplets or hydrogels. Physical properties and molecular recognition ability of these DNA condensates can be precisely modulated via sequence design, molecular modification, and environmental factors, such as temperature and salt concentrations. Additionally, these condensates can be used for the bottom-up reconstruction of cell-mimetic structures and various functions, such as target molecule detection, − protein synthesis scaffolds, , and cytoskeleton. ,

Several studies have attempted to form DNA condensates in microcompartments simulating artificial cells. Tran et al. encapsulated Y-motif DNA hydrogels into water-in-oil (W/O) droplets in fluorinated oil and polydisperse liposomes and demonstrated the segregation of orthogonal DNA droplets by RNase or light stimulation. Zhao et al. showed that homogeneous DNA hydrogel droplets are formed by encapsulating the Y-motif DNA in monodisperse W/O droplets in fluorinated oil. They further demonstrated that the partitioning of specific molecules can be controlled via photoswitching using azobenzene-tethered DNA. These pioneering studies have demonstrated DNA condensates as artificial and controllable intracellular organelles. However, they used thermal annealing to condense the DNA nanostructures in compartments, similar to conventional bulk preparation methods. In this approach, further artificial cell applications are difficult as heat shock damages the lipid membranes and encapsulated proteins. Furthermore, although it is desirable for artificial cells to be homogeneous, no study has demonstrated DNA condensate formation in uniform lipid-based GVs to date.

In this study, we investigated the formation dynamics of DNA condensates without thermal annealing in uniform GVs produced using a microfluidic system we described previously. , DNA condensation was triggered in GVs by gradually increasing the concentrations of encapsulated ingredients (DNA nanostars and salt) at room temperature. This operation was performed via osmotic shrinkage of GVs. , With this technique, differences in the formation behaviors of DNA condensates based on SE length and component concentrations were studied. Orthogonal DNA condensates formation, reversible condensation/decondensation, and protein expression of DNA condensate-bound genes showcase potential functions for dynamic and hierarchical artificial cells.

2. Experimental Section

2.1. Microfluidic Device Fabrication

A microfluidic channel to produce monodisperse lipid-stabilized water-in-oil-in-water (W/O/W) droplets as GV templates was fabricated using a standard one-step replica molding process, as previously described. , A mold was formed on a 2-in. silicon wafer (Matsuzaki Seisakusho Co., Ltd., Japan) with a negative photoresist (SU-8 3025; Kayaku Advanced Materials, Inc., MA, USA) at a height of 37 μm via photolithography. This mold was used to transfer the micropattern to polydimethylsiloxane (PDMS; Sylgard-184; Dow Corning) with a thickness of approximately 5 mm. After opening three inlet and two outlet holes using a biopsy punch (φ 0.75 mm), O2 plasma was applied (O2 flow rate: 20 sccm; power: 75 W; 5 s) using a reactive ion etching apparatus (FA-1; Samco Inc., Japan) to bond the PDMS channel to a glass slide. The device was used as-is after inserting Teflon tubing (inner diameter: 0.30 mm; outer diameter: 0.76 mm).

Next, a perfusion device for osmotic manipulation of GVs was fabricated to verify the reversibility of DNA condensate formation. Using the same procedure as described above, a PDMS device with microchambers of 100 μm diameter and approximately 50 μm depth was fabricated (Figure S1). After opening inlet (φ = 1.5 mm) and outlet (φ = 0.75 mm) holes using a biopsy punch, the microchamber device was made hydrophilic by applying O2 plasma (O2 flow rate: 20 sccm; power: 75 W; 5 s). Two pieces of double-sided tape (Nitto Denko No. 5302A; 85 μm thickness) were immediately used to bond the coverslip to the PDMS, forming a flow channel. This device was used within 30 min before the loss of hydrophilicity.

2.2. Preparation of the DNA Nanostar-Containing Inner Aqueous Phase

Based on our previous work, we prepared and used two types of Y-motif DNA nanostars with palindromic SEs of 4 and 6 nt, referred to as YA-4nt and YA-6nt, respectively. Additionally, we prepared YB-6nt with a 6 nt SE with a sequence different from that of YA-6nt (i.e., orthogonal to the SE of YA-6nt). The sequences of all DNA oligomers are listed in Table S1. Oligonucleotides were purchased from Eurofins Genomics (Tokyo, Japan). Stock solutions of DNA oligomers (100 μM each), followed by buffer and NaCl solutions, were mixed to prepare the inner aqueous solution (Table ) at a final volume of 100 μL. To label the Y-motifs, one of the three DNA oligomers making up each Y-motif was replaced with a DNA oligomer possessing the same stem sequence but modified with a fluorescent dye at a 10% molar ratio. YA-4nt and YA-6nt were labeled with a fluorescein-tagged oligomer (FAM; ex/em = 493/517 nm). For the experiment simultaneously generating two orthogonal DNA condensates, YB-6nt labeled with a cyanine 5 (Cy5; ex/em = 633/647 nm)-tagged oligomer was used in addition to YA-6nt to distinguish two orthogonal condensates in GV (Table S2). A mixed solution containing two types of orthogonal Y-motifs was prepared by separately mixing the three oligomers for each Y-motif (YA-6nt and YB-6nt), waiting for 10 min, and combining them before adding the buffer and NaCl. This preparation was performed immediately before microfluidic vesicle formation, and encapsulation of the inner solution was completed within 30 min of preparation.

1. Basic Compositions of Three Solutions Used for Giant Vesicle (GV) Production.

Inner aqueous phase
YA1-4nt or YA1-6nt 1 μM
YA2-4nt or YA2-6nt 0.9 μM
YA2-FAM 0.1 μM
YA3-4nt or YA3-6nt 1 μM
fluorescent dextran 40 μg/mL
sucrose 100 mM
NaCl 0–35 mM
Tris-HCl (pH 8.5) 4 mM
     
Outer aqueous phase
glucose 100 mM
NaCl 0–35 mM
Tris-HCl (pH 8.5) 4 mM
pluronic F-68 1 %v/v
PVA 0.5 %w/v
     
Oil phase
squalene:1-octanol 8:2 v/v
egg PC 6 mg/mL
rhodamine-PE 2 μg/mL
Open in a new tab

2.3. Preparation of the Oil and Outer Aqueous Phases

Compositions of the oil phase and the outer aqueous phase are also listed in Table . Oil phase, which forms the shell of the W/O/W droplets, was prepared by adding a 100-mg/mL stock solution of EggPC dissolved in chloroform to the oil phase consisting of squalene (Wako, Japan) and 1-octanol (Wako, Japan) at a volume ratio of 8:2 to achieve a final EggPC concentration of 6 mg/mL, as previously described. To visualize the oil phase and lipid membrane, rhodamine-DHPE (L1392; Thermo Fisher Scientific, Inc., MA, USA) was added to the oil phase at a final concentration of 2 μg/mL. The outer solution primarily consisting of sugars and salts was prepared as shown in Table . Among the listed components, amphiphilic copolymer pluronic F68 (Thermo Fisher Scientific, Inc.) was used to lower the interfacial tension and facilitate the transfer of droplets across the interface. Poly­(vinyl alcohol) (PVA; 87% hydrolyzed; MW 13–23 k; Sigma-Aldrich, MO, USA) was added to stabilize the W/O/W droplets and GVs after transfer. Inclusion of sucrose in the inner solution and glucose in the outer solution at equimolar concentrations ensured that the initial osmotic pressure was balanced.

2.4. Device Operation and Imaging

Uniform thin-shelled W/O/W droplets were produced by operating the microfluidic device as described in our previous studies. , Briefly, the microchannel device was placed on an indium-tin-oxide glass heater (SS-100; Blast Inc., Japan) mounted on a microscope stage, and two aqueous solutions (inner and outer water phases) and an oil phase containing lipids were injected using a three-channel pressure pump (On-Chip Droplet Generator; On-chip Biotechnologies Co., Ltd., Japan). The flow rate of each solution was adjusted by tuning the pressure applied to the reservoir. Initially, all solutions were subjected to high pressure (30 kPa) to purge any air from the tubing and microchannels. Once the laminar flow was stabilized, the pressure of all inlets was reduced to approximately 10 kPa. Finally, pressure values on the inner water and oil phases were slightly adjusted to control the size of W/O droplets and width of the oil phase flow, respectively. The flow inside the microchannels was monitored using an inverted microscope (Ti; Nikon, Japan) and recorded using the CMOS camera (GS3-U3-23S6M-C; Teledyne FLIR, OR, USA). W/O/W droplets produced in the device floated in water due to the buoyancy of the oil layer. Subsequently, W/O/W droplets collected in the tube were sampled from the top layer using a micropipette.

The collected droplets were observed by placing them between two pieces of a cover glass separated by a spacer. Practically, cover glasses (25 mm × 25 mm; thickness 0.17 mm; Matsunami Glass, Japan) were passivated by applying 2.5 w/v% PVA and dried at 120 °C for 20 min. Immediately before use, the cover glass was hydrated for over 5 min to extend the polymer brushes. This process effectively prevented the spread of the oil cap attached on GVs (Figure S2). After applying approximately 10 μL of W/O/W droplet solution on an untreated coverslip with a double-sided tape (Nitto No. 5000NS; thickness 160 μm) having a φ13 mm opening, a PVA-coated cover glass was placed on top. Next, a confocal laser scanning microscope (LSM-700; Carl Zeiss, Germany) equipped with a water-immersion 40× lens (C-Apochromat 40×/1.2 w CorrECS M27) was used to acquire the z-stack images of W/O/W droplets or GVs at 1 μm intervals in the upper layer of the sample space. Then, maximum cross-sectional areas of GVs and DNA condensates were measured from binalized fluorescent images using the particle analysis plugin in the ImageJ image analysis software (NIH).

2.5. Osmotic Shrinkage of GVs

A solution containing high glucose and salt concentrations was prepared to induce the osmotic shrinkage of GVs to the desired volume (Table S3). This solution was mixed with the GV solution collected from the microfluidic device on a glass slide at a 1:1 volume ratio (2 μL each). Then, GVs were shrunk to achieve final volume ratios (V/V 0) of 1.0, 0.8, 0.6, 0.4, and 0.2 based on the concentration difference between the encapsulated inner aqueous phase and added outer aqueous phase. Finally, GVs were observed under a confocal microscope, and images were analyzed using the ImageJ software. In the time-lapse observations, the start of the observation was set as time zero. Although there was some variation in the time between droplet collection and observation (approximately 3–5 min), this variation is sufficiently small compared to the characteristic time of the condensate formation dynamics. Additionally, we measured osmolarity of the inner solution and the final outer solution used to make V/V 0 = 0.2 using the a freezing-point depression osmometer (OM815, Vogel MedTec GmbH, Germany). Because the major contributors to the osmolarity in the aqueous solution compositions in the present experimental design are sugars (sucrose/glucose) and NaCl, ratio of osmolarity was nearly the same with the ideal (designed) values.

2.6. Reversibility Test of DNA Condensate Formation

Direct shrinkage method was used in this experiment. Briefly, five-times high-concentration glucose/salt solution (500 mM glucose, 175 mM NaCl, 1%v/v pluronic F-68, 0.5% w/v PVA, and 20 mM Tris-HCl) was used as the outer aqueous phase for the microfluidic production of GVs, directly creating a concentration difference between the inner and outer aqueous phases upon generation. A solution containing W/O/W droplets discharged from the microfluidic channel was directly introduced into the perfusion device (Figure S1), in which the droplets were trapped inside the microchambers. After trapping, a low-concentration glucose/salt solution (100 mM glucose, 35 mM NaCl, 1%v/v pluronic F-68, 0.5% w/v PVA, and 4 mM Tris-HCl) and high-concentration outer solution were alternately perfused using a syringe pump (KDS100; K d Scientific Inc., MA, USA) at a flow rate of 50 μL/h to repeat dilusion and concentration operation within GVs.

2.7. In Vesicle Expression of a Gene Embedded in the DNA Condensate

Green fluorescent protein (GFP)-coding DNA fragment with SE was prepared by cutting the GFP-coding plasmid (pETg5tag) with the restriction enzyme, PluTI (New England Biolabs, MA, USA), creating 5′-GCGC-3′ SEs. Additionally, GFP-coding DNA fragments with blunt ends were prepared by cutting with PvuII (Figure S3). These DNA fragments (both approximately 5 kbp) were purified using the QIAquick PCR purification kit (Qiagen, Germany), according to the manufacturer’s protocol. Covalent rhodamine staining of these DNA fragments was performed using the Label IT CX-Rhodamine nucleic acid labeling kit (Mirus Bio LLC, WI, USA), following the manufacturer’s protocol. A direct shrinkage approach was also used to shrink the uniform GVs initially containing halved concentrations of YA-4nt (2.5 μM) and PURE system solution (PUREfrex 1.0; Gene Frontier Inc., Japan) to achieve V/V 0 ∼ 0.5. GVs collected from the microfluidic device were placed on a cover glass slide, incubated at 37 °C on the microscope stage heater, and observed using the fluorescent confocal laser scanning microscope. Subsequently, we evaluated localization of the GFP-coding gene in the PURE system solution, DNA condensate formation in GVs, and GFP expression in GVs separately because covalent staining of the GFP-coding gene inhibited its expression. The detailed solution compositions are provided in Tables S4 and S5.

3. Results and Discussion

3.1. Generation of GVs Using Microfluidics

A schematic and an actual snapshot of W/O/W droplet formation in the microfluidic channel are shown in Figure a,b, respectively. First, hydrodynamic flow focusing was used to generate W/O droplets stabilized by lipids, in which DNA nanostars and salts were encapsulated at lower concentrations. Next, downstream, an outer aqueous phase was introduced from the side of the main channel, forming a water–oil (W–O) laminar flow where lipids were assembled at the interface. As the width of the microchannel narrowed from 200 to 100 μm along the W–O laminar flow, W/O droplets were pressed against the W–O interface. In the curved section of the channel, W/O droplets passed through the W–O interface, forming W/O/W droplets with thin oil shells. At the branching section at the end of the microchannel, oil and outer aqueous streams were separated, and an aqueous solution containing W/O/W droplets was collected into a test tube. Observation of droplets directly placed on a coverslip revealed that excess oil in the shell of the droplets underwent dewetting to form an oil cap, and the remaining area became a lipid bilayer. Dewetting (right panel in Figure a) began approximately 10 min after collection and was completed within 30 min for all droplets. GVs floated on the water surface of the tube owing to the buoyancy effect of the attached oil cap. In the subsequent experiments involving changes in vesicle volume, the oil cap functions as a lipid reservoir.

1.

1

Open in a new tab

(a) Schematic of monodisperse giant vesicle (GV) generation in the microfluidic channel by transferring W/O droplets across the W–O interface, followed by dewetting of excess oil in W/O/W droplets. (b) Microscopic image of the actual microfluidic device operation.

3.2. Generation of DNA Condensates Induced via Osmotic Shrinkage of GVs

We investigated whether DNA condensates can be formed by continuously increasing the concentration of DNA Y-motif without thermal annealing. A schematic of the experimental procedure is shown in Figure a. Upon the microfluidic formation of GVs, 1 μM of each DNA oligomer and 35 mM NaCl were encapsulated. In this initial state, Y-motif nanostructures were formed via complementary binding of the 34-nt stem sequences due to the high binding free energy, regardless of thermal annealing, as confirmed by electrophoresis (Figure S4). However, due to the low binding free energy of the short (4-nt or 6-nt) SEs, Y-motifs remain dispersed in the solution, and no condensates were formed (in previous study, DNA condensation experiments were designed with DNA oligomer and NaCl concentrations of 5–15 μM and 350 mM, respectively). By adding a solution containing high concentrations of sugar and salt outside of GVs, water was expelled through the lipid bilayer, causing the GVs to shrink. Accordingly, concentrations of the encapsulated Y-motifs and salt increased. Complementary binding of SEs located at the ends of the arms was dependent on the concentrations of SEs and salt. With increasing concentrations, SE binding increased, leading to the formation of DNA condensates.

2.

2

Open in a new tab

(a) Schematic of DNA condensate formation via concentration elevation of Y-motif and salt induced by the osmotic shrinkage of GVs. (b) Bright-field and confocal fluorescent images of GVs 2 h after osmotic shrinkage. (c) Three-dimensional (3D) constructed image of a GV 2 h after osmotic shrinkage generated from 47 z-sliced images with 1-μm spacing.

Figure b,c show a wide field view and a 3D confocal reconstruction image of the W/O/W droplets 2 h after osmotic shrinkage. The W/O/W droplets right after generation were surrounded by a relatively thick oil layer of approximately 1 μm and had a relatively large internal volume (typically 40–50 μm diameter). After adding a high-concentration glucose/salt solution, dewetting occurred after 10–30 min. Excess oil formed an oil cap on the upper side of the GV, as shown in Figure c, as a large red blob over the blue internal volume of the GV. Concurrently, the internal volume was significantly reduced (typically approximately 20 μm in diameter), and a single DNA condensate (marked green) was observed within the internal aqueous lumen of the GV. Both GV and DNA condensates were quite uniform, with diameters of 19.6 ± 0.4 and 6.8 ± 0.5 μm (N = 37, the number following ± represents the standard deviation), respectively, in the typical example displayed in Figure b. Additionally, we performed the fluorescent recovery after photobreaching experiment in the bulk condition (Figure S5). The Y-motif condensate with 6nt SE maintained fluidity even at a room temperature.

GV generation and shrinkage operation were repeatedly performed by varying the initial concentrations of DNA/NaCl and extent of shrinkage, which was adjusted by the concentration of the hypertonic solution added later (Table S3). The final volume of the GV and the size of the resulting DNA condensates were evaluated from their maximum diameters in z-stack images acquired using a confocal microscope. As the GVs shrank, the fluorescence intensity of the internal marker (fluorescently tagged dextran, shown in blue) encapsulated in the solution increased (Figure a). This confirms that the impermeable molecules were concentrated as the GV shrunk. Assuming that the internal aqueous phase of the GV was spherical, the GV volume relative to the initial volume, V/V 0, was calculated using the maximum diameter of the GV in the z-stack images. The designed V/V 0 value based on the osmolarity difference approximately matched the observed V/V 0 value (Figure b). This result indicates that, by adjusting the osmolarity differences, it is possible to precisely control the GV size to concentrate the encapsulated molecules.

3.

3

Open in a new tab

Shrinkage assay to concentrate DNA Y-motif (YA-6nt) to form the DNA condensate. (a) Fluorescent confocal microscopic images of the internal volume (blue) and YA-6nt DNA condensates (green) 2 h after the exchange of the outer medium at various salt and DNA concentrations. Superposition of >30 z-stacks is shown to visualize the maximum diameters of the DNA condensate and GV in the 2D image. (b) Actual (measured) relative volume V/V 0 of GVs vs designed V/V 0 predicted from the osmolarity difference. Red dotted line represent the theoretical curve. (c) Dependence of the DNA condensate volume on designed V/V 0. In (b) and (c), averages and standard deviations of n = 10 GV samples under each condition are shown.

The final size of the DNA condensates depended on the initial DNA and NaCl concentrations. First, the three conditions in the upper part of Figure a highlight that the final size of the DNA condensate depends on the salt concentration. When the initial DNA concentration was 1 μM but NaCl was absent, no DNA condensates formed even when the GV was shrunk to V/V 0 = 0.2 (Figure a, first row). When the initial NaCl concentration was 10 mM, no condensates were observed at V/V 0 = 0.8 to 0.4, but visible DNA condensates appeared at V/V 0 = 0.2 (Figure a, second row). Furthermore, when the initial NaCl concentration was 35 mM, small condensates appeared at V/V 0 = 0.8, and their sizes increased with the extent of shrinkage (Figure a, third row). These results indicate that the complementary binding of SEs is dependent on salt concentration, with higher NaCl concentrations facilitating the formation of DNA condensates.

The two conditions in the lower rows of Figure a compare the differences in the final sizes of the DNA condensates based on the initial DNA concentration. Under an initial NaCl concentration of 35 mM, DNA condensates began to form at a designed V/V 0 = 0.2, and the diameter remained constant at approximately 5 μm at V/V 0 = 0.4 and beyond, with an initial DNA concentration of 1 μM. On the other hand, with an initial DNA concentration of 5 μM, DNA condensates were observed even without shrinkage, and the diameter reached approximately 9 μm at V/V 0 = 0.2. Comparing conditions with the same designed V/V 0 value, it is clear that a higher initial DNA concentration results in a larger DNA condensate at the final stage.

Figure c shows the volume of the DNA condensates calculated by assuming a spherical shape based on the maximum cross-sectional area of the DNA condensates observed in the z-stack images. Under the initial concentrations of DNA and NaCl of 1 μM and 35 mM (orange line), the size of the DNA condensates increased with the extent of shrinkage (decreasing V/V 0), reaching a constant value (approximately 60 μm3) at V/V 0 = 0.4–0.2. At the same DNA concentration but with an initial salt concentration of 10 mM (blue line), the volume of the DNA condensates formed at a designed V/V 0 = 0.2 showed a similar value. On the other hand, with an initial DNA concentration of 5 μM and a NaCl concentration of 35 mM (yellow line), the size of the DNA condensates continued to increase from V/V 0 = 0.8, and there was no indication that the size would reach a plateau. At V/V 0 = 0.8, the volume of the condensates was approximately 400 μm3. Comparing the condensate sizes based on the initial DNA concentration, when the DNA concentration was increased 5-fold, the condensate volume also increased by approximately 6.6 times (Figure c, orange and yellow lines). Considering the estimation error based on the assumption of a sphere, this result is in reasonable agreement with the theoretical prediction based on the premise that the size of the DNA condensate is determined by the number of DNA nanostars enclosed within the compartment. From this result, it was possible to estimate the DNA concentration in the condensed phase. Assuming that 1 μM DNA in a spherical vesicle with 40 μm initial diameter was condensed to become 60 μm3 in volume, the DNA concentration in the condensed phase is calculated to be approximately 560 μM.

Notably, a single DNA condensate was formed in each individual GV under all conditions, and as the GVs were uniform, the resulting DNA condensates were also uniform in size. When DNA condensates prepared in bulk are encapsulated directly into uniform compartments, the initial sizes of the condensates are uneven, resulting in the presence of multiple nonuniform DNA condensates inside the uniform GVs. Therefore, in previous studies using uniform compartments (W/O droplets and solid microchambers), heat annealing was employed to obtain uniform condensates. , In this study, however, we demonstrated that with a gradual change in concentration, a single uniform condensate could be formed within each compartment, as seen in the coefficient of variation of 8–22% in volume for n = 10 GVs at each condition presented in Figure c. These results indicate that this system can serve as a reliable platform to produce uniform DNA condensates within uniform GVs, and that their volumes can be adjusted by the initial DNA concentration, NaCl concentration, and extent of shrinkage of the GVs.

3.3. Dependence on SE Length

Next, we investigated the effect of the SE length of the Y motif on the condensate formation dynamics. Uniform GVs were produced encapsulating Y-motif DNA and NaCl at 1 μM and 35 mM, respectively, and subjected to an osmotic shrinkage to achieve V/V 0 = 0.2. Differences in the formation behavior between SE with lengths of 4 and 6 nt were observed (Figure a). In the case of YA-4nt, multiple nonspherical DNA condensates of varying sizes appeared immediately after shrinkage. These small condensates coalesced over time and merged into a single condensate after 30 min, and it remained spherical thereafter. In contrast, with YA-6nt, multiple nonspherical condensates were formed within approximately 15 min. As time elapsed, the small condensates adhered together. However, unlike YA-4nt, they did not readily adopt a spherical shape due to the smaller fluidity arising from the greater binding energy. Even after 60 min, the shape gradually approached a spherical form; however, in some GVs, even after 2 h, the condensates did not fully merge into a single condensate, and multiple condensates remained.

4.

4

Open in a new tab

Comparison of DNA condensate formation with different sticky end (SE) lengths. (a) Time-lapse images of DNA condensate formation with different SE lengths within GV shrunk to V/V 0 = 0.2. Superposition of >30 z-stacks is shown to visualize the maximum size of the DNA condensate and GV in the 2D image. Scale bars = 10 μm. (b) Relationship between designed V/V 0 and total projected area of the DNA condensate. Averages and standard deviations of n = 10 GV samples under each condition are shown.

We evaluated the amount of DNA condensate in confocal z-stack images after 2 h at various shrinkage ratios (Figure b). As the shape and number of condensates in each GV varied, especially with YA-6nt, precise evaluation of volume was difficult. Therefore, we evaluated the sum of the largest areas of the individual condensates in the stacked images. For YA-4nt, DNA condensate formation started at designed V/V 0 = 0.8, whereas for YA-6nt, it started at V/V 0 = 0.6. In both cases, the condensate volume became almost constant at V/V 0 ≤ 0.4. Furthermore, comparison of condensate sizes for different V/V 0 values indicated that a longer SE resulted in a smaller condensate. However, as shrinkage progressed, the size difference diminished, and the condensates reached a similar volume of 15–20 μm3 at V/V 0 = 0.2. This result indicates that the DNA concentrations in condensates are similar regardless of the SE length.

3.4. Reversibility of DNA Condensation

Next, we examined the reversibility of DNA condensate formation in the GV system. Specifically, we tested whether the DNA condensate formed by concentrating YA-4nt via GV shrinkage dissociates when the GV is re-expanded. To streamline the experimental procedure, instead of creating GVs in an isotonic outer solution and adding a high-osmolarity solution, we used a high-osmolarity solution as the outer aqueous phase in the microfluidic channel upon GV formation. This ensured that GV shrinkage began immediately after W/O/W droplet generation. We confirmed that even with such rapid shrinkage, shrinkage ratio was close to the designed value. A solution containing droplets discharged from the microfluidic channel was introduced directly into the perfusion device (Figure S1). The perfusion device was equipped with a microchamber array on its ceiling of the channel, where the GVs floating due to the buoyancy of the oil cap, were trapped. GVs were formed with initial DNA and NaCl concentrations of 1 μM and 35 mM, respectively. GV shrinkage occurred immediately after dewetting to V/V 0 = approximately 0.2, and a single DNA condensate was formed within 2 h (Figure a).

5.

5

Open in a new tab

Reversibility of DNA condensate generation. Confocal microscopy images of GV containing YA-4nt DNA trapped in the microfluidic perfusion device. A GV in the identical microchamber (indicated by a dotted line) is shown. (a) 2 h after GV production. GVs shrunk due to the osmolarity difference between the inner and outer aqueous phases. (b) 3 h after medium exchange (hypotonic solution) to expand the GV volume. (c) 3h after medium exchange to hypertonic solution. (d) 1.5 h after medium exchange to hypotonic solution. Blue, internal volume marker of GV; green, DNA condensate; red, lipid membrane. White and yellow arrows indicate the GV membrane and DNA condensate, respectively.

Next, we exchanged the solution surrounding trapped GVs with a low-concentration sugar and salt medium using a syringe pump by perfusing at a flow rate of 50 μL/h, causing the GV to expand. Because the GVs were trapped in the microchambers, they remained in place, allowing for continuous observation of the same GVs during perfusion. After 3 h of perfusion, the GVs expanded to a volume similar to that in the initial state (Figure b). The fluorescence of the internal marker was diminished, indicating that the inner contents were diluted. Concomitantly, the DNA condensate present before perfusion disappeared. We continued another round of shrinkage and expansion, and succeeded in repeated generation and dissolution of the condensate (Figure c,d). This result demonstrates that the formation of DNA condensates can be reversibly controlled by concentration changes that accompany GV volume modulation. Scale bars = 100 μm.

3.5. Controlled Formation of Immiscible DNA Condensates

We then conducted an experiment to encapsulate two types of Y-motifs with immiscible (orthogonal) SEs, that is, SE sequences that do not form complementary bonds with each other, to generate two distinct DNA condensates within the same GV. By shrinking the GV, the concentration of the coencapsulated orthogonal Y-motifs (YA-6nt and YB-6nt) increased simultaneously to form condensates (Figure a). Each of the two Y-motifs contained a fluorescently tagged oligomer at a 10% molar ratio, which allowed us to distinguish them with their fluorescent colors.

6.

6

Open in a new tab

Generation of mutually orthogonal DNA Y-motifs in uniform GVs. (a) Schematic of the experimental system. (b) Time-lapse confocal microscopic images of GV and two DNA condensates. Blue, internal volume marker of GV; green, YA-6nt; yellow, YB-6nt. (c) Fluorescent images and calculated volume of two condensates with various initial YB-6nt concentrations (X) and initial YA-6nt concentration fixed to 1 μM. Scale bars = 10 μm. In the graph, averages and standard deviations of n = 10 GV samples under each condition are shown.

In this experiment, we again used the direct shrinkage approach, inducing shrinkage immediately after W/O/W droplet generation. Figure b shows the time-dependent formation of DNA condensates after shrinkage. Within 15 min, polydisperse nonspherical condensates were formed. Over time, these condensates gradually coalesced into similar condensates, and only two distinct condensates with nonspherical shapes remained after 2 h. After 6 h, two spherical DNA condensates of nearly equal sizes were observed.

By varying the encapsulated concentration of each Y-motif, we could individually control the sizes of the two types of DNA condensates. Here, initial concentration of YA-6nt (marked in green) was fixed at 1 μM, whereas that of YB-6nt (marked in yellow) was varied. Figure c shows the microscopic images of the two different DNA condensates with controlled sizes and a plot of their volumes evaluated from the images. The size of the YA-6nt DNA condensate (green) remained constant, whereas that of the YB-6nt DNA condensates (yellow) increased as the initial concentration increased. Assuming a spherical shape, the condensate volumes were calculated from the maximum diameters of the condensates (Figure c). The condensate volumes were approximately proportional to the initial concentrations. These results suggest that different DNA condensates can coexist in the same GV vai sequence design and that their volumes can be controlled via the loading amounts.

3.6. Expression of a Gene Embedded in the DNA Condensate

Finally, we induced in vesicle expression of a gene embedded in the DNA condensate. First, we prepared a GFP-coding DNA fragment with an SE identical to that of YA-4nt at both 5′ and 3′ ends by cutting the plasmid (pETg5tag5) using the restriction enzyme, PluTI. As a preliminary test, we examined whether Y-motif DNA condensates are formed in the buffer of the PURE system, which was used for gene expression, and whether DNA fragments with SEs interact with Y-motif condensates under bulk conditions. After mixing 5 μM YA-4nt and 1.2 ng/μL (0.38 nM) GFP gene, with the latter covalently stained with rhodamine, Y-motif DNA condensate was successfully formed without the addition of NaCl. Ionic components in PURE buffer possibly facilitated complementary bonds between SEs. Furthermore, GFP gene fragment having the SE was attached to the surface of the YA-4nt condensate (Figure S6a). In contrast, the same GFP gene fragment with blunt ends did not interact with the Y-motif DNA condensate (Figure S6b). These observations confirmed that the GFP-coding gene with SEs interacted with the SEs of the Y motif, although they did not completely dissolve into the condensates. This is because relatively long (approximately 5 kb) double-stranded DNA is too rigid to enter the flexible network of the Y-motif.

We further tested whether the DNA condensate is formed in GVs with the buffer used for in vitro gene expression , via GV shrinkage. After encapsulating YA-4nt at 5 μM together with the halved PURE system buffer (detailed compositions of inner and outer solutions are listed in Table S4), GV was subjected to osmotic shrinkage to V/V 0 = approximately 0.5. We observed the development of a single DNA condensate in GV (Figure a), similar to experiments conducted under fundamental solution conditions (Table ). This result confirmed that DNA condensate formation in GVs was compatible with the PURE buffer conditions.

7.

7

Open in a new tab

Microscopy images of GVs showing the in vesicle green fluorescent protein (GFP) expression of the gene embedded in the Y-motif DNA condensate using the PURE system. (a) Y-motif condensation in the solution for gene expression. Blue and green fluorescent colors show the internal volume marker of GVs and YA-4nt condensate, respectively. (b, c) GFP expression in the (b) absence and (c) presence of YA-4nt. Left and right panels show images before and after 2-h incubation at 37 °C. Red and green fluorescent colors show the membrane (rhodamine-DHPE) and synthesized GFP, respectively. Scale bars = 10 μm.

Next, we enclosed 0 or 5 μM YA-4nt and PURE system into GVs together with 1.2 ng/μL (0.38 nM) of the GFP-coding DNA fragment with 4nt SE (detailed compositions of inner and outer solutions are listed in Table S5) and incubated them at 37 °C for 4 h. The fluorescently tagged oligomer (YA2-FAM) was omitted to avoid overlapping fluorescent colors (also see Figure S7 for the simultaneous observation of DNA condensation and GFP expression). In the presence of 2.5 μM YA-4nt (∼5 μM after shrinkage to V/V 0 = 0.5, Figure c), a similar amount of GFP was produced compared to the case without YA-4nt DNA (Figure b). GFP on the order of 0.1 mg/mg was synthesized in GV in our previous work, so we suppose a similar amount was synthesized in the present experiments. These results confirm that our DNA condensate control system, which does not require thermal annealing, is compatible with essential biochemical reactions for the artificial cell research.

4. Conclusions

This study demonstrated a method for the precise control of the dynamics of DNA condensate formation, final size, reversible formation/dissolution and multiple condensate formation via sequence design and GV volume modulation induced by osmotic action. This micofluidic technique facilitates the mass-production of uniform-sized DNA condensates in uniform microscale lipid vesicles. Notably, formation of a single DNA condensate in a GV indicated that this structure resembles a cell nucleus, in a sense that a locally condensed phase of DNA is present in the cytoplasm bounded by a lipid membrane. Additionally, as the formation of DNA condensates did not require temperature-annealing-based condensation, our method can be extended to more complex cell-mimetic systems incorporating proteins, such as enzymes, without any risk of heat denaturation, thus showing high scalability. The final demonstration of the coexistence of DNA condensation and protein expression provides a promising pathway for establishing a production hub of macromoleculessimilar to a real cell nucleuswithin artificial cells in the future. To date, enhanced gene expression in a molecularly crowded environment has been reported by several research groups. ,− We are currently investigating the effect of DNA condensates on the protein synthesis reaction in detail. We foresee that this hierarchical structure can be used for various purposes, such as molecular recognition and gene expression in cell-mimetic environments.

Supplementary Material

au5c00568_si_001.pdf (2.8MB, pdf)

Acknowledgments

We thank Prof. Yusuke Sato (Kyushu Institute of Technology) for fruitful discussion. We also thank Prof. Tomoaki Matsuura (Institute of Science Tokyo) for kindly providing the GFP-conding plasmid. This study was supported by JSPS Grants 19H02576, 19H00901, 20H05935, 24K01320, 24K08214, and 24H01155 and the Institute of Science and Engineering at Chuo University.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c00568.

  • Figure S1: Overview, structure, and design of the microchamber device for GV trapping and perfusion; Figure S2: Comparison of the spreading of the oil cap of GV after dewetting; Figure S3: Construction of GFP-coding gene; Figure S4: Electrophoresis of DNA oligomers for Y-motifs A having different SEs; Figure S5: Fluorescent recovery after photobreaching experiment; Figure S6: Fluorescent confocal images of YA-4nt DNA condensate incubated together with the rhodamine-stained GFP-coding DNA fragment in bulk condition; Figure S7: Confocal microscope images of GVs after shrinkage and incubation showing fluorescence signals from GFP and DNA condensate; Table S1: Sequences for Y-motifs; Table S2: Composition of the inner aqueous phase for the experiment producing orthogonal Y motifs; Table S3: Composition of outer solutions to induce osmotic shrinkage; Table S4: Compositions for three solutions used for GV production to test the generation of DNA condensates within GVs in PUREsystem buffer; Table S5: Compositions for three solutions used for GV production to test the GPF expression within GVs coencapsulating Y-motif DNA (PDF)

R.Yoneyama established the experimental system and conducted most of the experiments. N.Morikawa conducted the gene expression experiment. R.Ushiyama established the microfluidic formation of monodisperse GV to encapsulate DNA Y-motifs. R.Sato and M.Tsugane prepared the SE-modified gene and helped with molecular biology experiments. T.Maruyama and M.Takinoue supervised the experimental design of DNA condensates. H.Suzuki directed and supervised the entire project, and R.Yoneyama and H.Suzuki wrote the manuscript.

The authors declare no competing financial interest.

References

  1. Walde P., Cosentino K., Engel H., Stano P.. Giant vesicles: preparations and applications. ChemBioChem. 2010;11(7):848–865. doi: 10.1002/cbic.201000010. [DOI] [PubMed] [Google Scholar]
  2. Buddingh B. C., van Hest J. C. M.. Artificial Cells: Synthetic Compartments with Life-like Functionality and Adaptivity. Acc. Chem. Res. 2017;50(4):769–777. doi: 10.1021/acs.accounts.6b00512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Salehi-Reyhani A., Ces O., Elani Y.. Artificial cell mimics as simplified models for the study of cell biology. Exp. Biol. Med. 2017;242(13):1309–1317. doi: 10.1177/1535370217711441. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Lu Y., Allegri G., Huskens J.. Vesicle-based artificial cells: materials, construction methods and applications. Mater. Horiz. 2022;9(3):892–907. doi: 10.1039/D1MH01431E. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Sunami T., Kita H., Hosoda K., Matsuura T., Suzuki H., Yomo T.. Detection and Analysis of Protein Synthesis and Rna Replication in Giant Liposomes. Methods Enzymol. 2009;464:19–30. doi: 10.1016/S0076-6879(09)64002-7. [DOI] [PubMed] [Google Scholar]
  6. Kurihara K., Tamura M., Shohda K., Toyota T., Suzuki K., Sugawara T.. Self-reproduction of supramolecular giant vesicles combined with the amplification of encapsulated DNA. Nat. Chem. 2011;3(10):775–781. doi: 10.1038/nchem.1127. [DOI] [PubMed] [Google Scholar]
  7. Tsugane M., Suzuki H.. Reverse Transcription Polymerase Chain Reaction in Giant Unilamellar Vesicles. Sci. Rep. 2018;8:9214. doi: 10.1038/s41598-018-27547-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Nishimura K., Tsuru S., Suzuki H., Yomo T.. Stochasticity in Gene Expression in a Cell-Sized Compartment. ACS Synth. Biol. 2015;4(5):566–576. doi: 10.1021/sb500249g. [DOI] [PubMed] [Google Scholar]
  9. Noireaux V., Bar-Ziv R., Godefroy J., Salman H., Libchaber A.. Toward an artificial cell based on gene expression in vesicles. Phys. Biol. 2005;2(3):1–8. doi: 10.1088/1478-3975/2/3/P01. [DOI] [PubMed] [Google Scholar]
  10. Liu Z., Zhou W., Qi C., Kong T.. Interface Engineering in Multiphase Systems toward Synthetic Cells and Organelles: From Soft Matter Fundamentals to Biomedical Applications. Adv. Mater. 2020;32(43):2002932. doi: 10.1002/adma.202002932. [DOI] [PubMed] [Google Scholar]
  11. Wang X., Liu X., Huang X.. Bioinspired Protein-Based Assembling: Toward Advanced Life-Like Behaviors. Adv. Mater. 2020;32(25):2001436. doi: 10.1002/adma.202001436. [DOI] [PubMed] [Google Scholar]
  12. Yu X., Zhou L., Wang G., Wang L., Dou H.. Hierarchical Structures in Macromolecule-Assembled Synthetic Cells. Macromol. Rapid Commun. 2022;43(14):2100926. doi: 10.1002/marc.202100926. [DOI] [PubMed] [Google Scholar]
  13. Wang X., Qiao X., Chen H., Wang L., Liu X., Huang X.. Synthetic-Cell-Based Multi-Compartmentalized Hierarchical Systems. Small Methods. 2023;7(12):2201712. doi: 10.1002/smtd.202201712. [DOI] [PubMed] [Google Scholar]
  14. Shin Y., Brangwynne C. P.. Liquid phase condensation in cell physiology and disease. Science. 2017;357(6357):eaaf4382. doi: 10.1126/science.aaf4382. [DOI] [PubMed] [Google Scholar]
  15. Bracha D., Walls M. T., Brangwynne C. P.. Probing and engineering liquid-phase organelles. Nat. Biotechnol. 2019;37(12):1435–1445. doi: 10.1038/s41587-019-0341-6. [DOI] [PubMed] [Google Scholar]
  16. Visser B. S., Lipinski W. P., Spruijt E.. The role of biomolecular condensates in protein aggregation. Nat. Rev. Chem. 2024;8(9):686–700. doi: 10.1038/s41570-024-00635-w. [DOI] [PubMed] [Google Scholar]
  17. Dewey D. C., Strulson C. A., Cacace D. N., Bevilacqua P. C., Keating C. D.. Bioreactor droplets from liposome-stabilized all-aqueous emulsions. Nat. Commun. 2014;5:4670. doi: 10.1038/ncomms5670. [DOI] [PubMed] [Google Scholar]
  18. Last M. G. F., Deshpande S., Dekker C.. pH-Controlled Coacervate-Membrane Interactions within Liposomes. ACS Nano. 2020;14(4):4487–4498. doi: 10.1021/acsnano.9b10167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Love C., Steinkühler J., Gonzales D. T., Yandrapalli N., Robinson T., Dimova R., Tang T. Y. D.. Reversible pH-Responsive Coacervate Formation in Lipid Vesicles Activates Dormant Enzymatic Reactions. Angew. Chem., Int. Ed. 2020;59(15):5950–5957. doi: 10.1002/anie.201914893. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Deshpande S., Brandenburg F., Lau A., Last M. G. F., Spoelstra W. K., Reese L., Wunnava S., Dogterom M., Dekker C.. Spatiotemporal control of coacervate formation within liposomes. Nat. Commun. 2019;10:1800. doi: 10.1038/s41467-019-09855-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Mu W., Ji Z., Zhou M., Wu J., Lin Y., Qiao Y.. Membrane-confined liquid-liquid phase separation toward artificial organelles. Sci. Adv. 2021;7(22):eabf9000. doi: 10.1126/sciadv.abf9000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lu T. M., Liese S., Schoenmakers L., Weber C. A., Suzuki H., Huck W. T. S., Spruijt E.. Endocytosis of Coacervates into Liposomes. J. Am. Chem. Soc. 2022;144(30):13451–13455. doi: 10.1021/jacs.2c04096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lu T., Hu X., van Haren M. H. I., Spruijt E., Huck W. T. S.. Structure-Property Relationships Governing Membrane-Penetrating Behaviour of Complex Coacervates. Small. 2023;19(38):2303138. doi: 10.1002/smll.202303138. [DOI] [PubMed] [Google Scholar]
  24. Lattuada E., Leo M., Caprara D., Salvatori L., Stoppacciaro A., Sciortino F., Filetici P.. DNA-GEL, Novel Nanomaterial for Biomedical Applications and Delivery of Bioactive Molecules. Front. Pharmacol. 2020;11:01345. doi: 10.3389/fphar.2020.01345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Dong Y. H., Yao C., Zhu Y., Yang L., Luo D., Yang D. Y.. DNA Functional Materials Assembled from Branched DNA: Design, Synthesis, and Applications. Chem. Rev. 2020;120(17):9420–9481. doi: 10.1021/acs.chemrev.0c00294. [DOI] [PubMed] [Google Scholar]
  26. Wang D., Liu P., Luo D.. Putting DNA to Work as Generic Polymeric Materials. Angew. Chem., Int. Ed. 2022;61(4):e202110666. doi: 10.1002/anie.202110666. [DOI] [PubMed] [Google Scholar]
  27. Udono H., Gong J., Sato Y., Takinoue M.. DNA Droplets: Intelligent, Dynamic Fluid. Adv. Biol. 2023;7(3):2200180. doi: 10.1002/adbi.202200180. [DOI] [PubMed] [Google Scholar]
  28. Takinoue M.. DNA droplets for intelligent and dynamical artificial cells: from the viewpoint of computation and non-equilibrium systems. Interface Focus. 2023;13(5):20230021. doi: 10.1098/rsfs.2023.0021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Gong J., Tsumura N., Sato Y., Takinoue M.. Computational DNA Droplets Recognizing miRNA Sequence Inputs Based on Liquid-Liquid Phase Separation. Adv. Funct. Mater. 2022;32(37):2202322. doi: 10.1002/adfm.202202322. [DOI] [Google Scholar]
  30. Udono H., Fan M., Saito Y., Ohno H., Nomura S. I. M., Shimizu Y., Saito H., Takinoue M.. Programmable Computational RNA Droplets Assembled via Kissing-Loop Interaction. ACS Nano. 2024;18(24):15477–15486. doi: 10.1021/acsnano.3c12161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Maruyama T., Gong J., Takinoue M.. Temporally controlled multistep division of DNA droplets for dynamic artificial cells. Nat. Commun. 2024;15(1):7397. doi: 10.1038/s41467-024-51299-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Park N., Kahn J. S., Rice E. J., Hartman M. R., Funabashi H., Xu J. F., Um S. H., Luo D.. High-yield cell-free protein production from P-gel. Nat. Protoc. 2009;4(12):1759–1770. doi: 10.1038/nprot.2009.174. [DOI] [PubMed] [Google Scholar]
  33. Park N., Um S. H., Funabashi H., Xu J. F., Luo D.. A cell-free protein-producing gel. Nat. Mater. 2009;8(5):432–437. doi: 10.1038/nmat2419. [DOI] [PubMed] [Google Scholar]
  34. Kurokawa C., Fujiwara K., Morita M., Kawamata I., Kawagishi Y., Sakai A., Murayama Y., Nomura S. I. M., Murata S., Takinoue M.. et al. DNA cytoskeleton for stabilizing artificial cells. Proc. Natl. Acad. Sci. U.S.A. 2017;114(28):7228–7233. doi: 10.1073/pnas.1702208114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sato Y., Takinoue M.. Capsule-like DNA Hydrogels with Patterns Formed by Lateral Phase Separation of DNA Nanostructures. JACS Au. 2022;2(1):159–168. doi: 10.1021/jacsau.1c00450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Tran M. P., Chatterjee R., Dreher Y., Fichtler J., Jahnke K., Hilbert L., Zaburdaev V., Göpfrich K.. A DNA Segregation Module for Synthetic Cells. Small. 2023;19(13):2202711. doi: 10.1002/smll.202202711. [DOI] [PubMed] [Google Scholar]
  37. Zhao Q., Cao F., Luo Z., Huck W. T. S., Deng N.. Photoswitchable Molecular Communication between Programmable DNA-Based Artificial Membraneless Organelles. Angew. Chem., Int. Ed. 2022;61(14):e202117500. doi: 10.1002/anie.202117500. [DOI] [PubMed] [Google Scholar]
  38. Ushiyama R., Koiwai K., Suzuki H.. Plug-and-play microfluidic production of monodisperse giant unilamellar vesicles using droplet transfer across Water-Oil interface. Sens. Act. B Chem. 2022;355:131281. doi: 10.1016/j.snb.2021.131281. [DOI] [Google Scholar]
  39. Ushiyama R., Nanjo S., Tsugane M., Sato R., Matsuura T., Suzuki H.. Identifying Conditions for Protein Synthesis Inside Giant Vesicles Using Microfluidics toward Standardized Artificial Cell Production. ACS Synth. Biol. 2024;13(1):68–76. doi: 10.1021/acssynbio.3c00629. [DOI] [PubMed] [Google Scholar]
  40. Deng N. N., Vibhute M. A., Zheng L. F., Zhao H., Yelleswarapu M., Huck W. T. S.. Macromolecularly Crowded Protocells from Reversibly Shrinking Monodisperse Liposomes. J. Am. Chem. Soc. 2018;140(24):7399–7402. doi: 10.1021/jacs.8b03123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Vibhute M. A., Schaap M. H., Maas R. J. M., Nelissen F. H. T., Spruijt E., Heus H. A., Hansen M. M. K., Huck W. T. S.. Transcription and Translation in Cytomimetic Protocells Perform Most Efficiently at Distinct Macromolecular Crowding Conditions. ACS Synth. Biol. 2020;9(10):2797–2807. doi: 10.1021/acssynbio.0c00330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sato Y., Sakamoto T., Takinoue M.. Sequence-based engineering of dynamic functions of micrometer-sized DNA droplets. Sci. Adv. 2020;6(23):eaba3471. doi: 10.1126/sciadv.aba3471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sokolova E., Spruijt E., Hansen M. M. K., Dubuc E., Groen J., Chokkalingam V., Piruska A., Heus H. A., Huck W. T. S.. Enhanced transcription rates in membrane-free protocells formed by coacervation of cell lysate. Proc. Natl. Acad. Sci. U.S.A. 2013;110(29):11692–11697. doi: 10.1073/pnas.1222321110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Nishio T., Yoshikawa Y., Yoshikawa K., Sato S. I.. Longer DNA exhibits greater potential for cell-free gene expression. Sci. Rep. 2021;11(1):11739. doi: 10.1038/s41598-021-91243-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tomohara K., Minagawa Y., Noji H.. Artificial cells with all-aqueous droplet-in-droplet structures for spatially separated transcription and translation. Nat. Commun. 2025;16(1):627. doi: 10.1038/s41467-024-55366-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

au5c00568_si_001.pdf (2.8MB, pdf)

Articles from JACS Au are provided here courtesy of American Chemical Society

RESOURCES