Cells rely on shape. The contours of a membrane, the arrangement of proteins, and the geometry of molecular scaffolds all determine how cellular components recognize one another and respond. But biological shapes are often fleeting and difficult to reprogram. Now, researchers have built a synthetic cell system that couples nanoscale reshaping of engineered DNA structures to microscale reshaping of artificial cell membranes, creating programmable channels that open and close on demand.

The work, published in Nature Materials, demonstrates a versatile platform in which signal responsive DNA nanorafts attached to giant unilamellar vesicles can trigger reversible membrane deformations, perforate lipid bilayers, and regulate the transport of large cargo molecules across membranes. Assisted by bacterial pores, the system enables the formation of synthetic channels far larger than those found in nature, offering new possibilities for synthetic biology and cellular engineering.

Building with DNA

DNA has become an ideal construction material for bottom up synthetic biology. Its sequence specific interactions allow researchers to design structures with nanometer precision. DNA origami, in which a long single strand of DNA is folded into predetermined shapes using short staple strands, has been used to mimic membrane sculpting proteins, create nanopores, and build molecular motors.

In this study, the team designed DNA origami based nanorafts that undergo reversible shape changes driven by toehold mediated strand displacement reactions. These nanorafts can switch between a nearly square configuration, measuring roughly 71 nanometers by 55 nanometers with an aspect ratio of 1.3, and an elongated rectangular form, stretching 190 nanometers by 20 nanometers with an aspect ratio of 9.5. The transformation is triggered by adding unlocking DNA strands, and the process reverses when locking strands are introduced.

Crucially, the nanorafts are anchored to lipid membranes via cholesterol tags. Twelve cholesterol anchors extend from the bottom surface of the origami structure, embedding into the membrane and ensuring stable attachment. The pattern formed by these cholesterol anchors also adapts as the DNA raft changes shape, creating a dynamic interface between the nanostructure and the lipid bilayer.

From Nanoscale to Microscale

The researchers first characterized the behavior of the DNA rafts on supported lipid bilayers using atomic force microscopy. In the square state, the rafts displayed a disordered arrangement on the membrane. After adding unlocking strands, nearly all the square rafts transformed into elongated rectangles. These high aspect ratio structures tend to align side by side, creating short range local order on the membrane. This ordering can be explained theoretically as a two dimensional system of Brownian hard needle-like objects, where the alignment of the elongated rafts inhibits their translational and rotational diffusion perpendicular to their long axes.

When locking strands were added, the elongated rafts reverted to the square configuration and disorder returned. Importantly, when elongated rafts were assembled directly rather than through dynamic reconfiguration, no prominent local order emerged. This indicates that the reshaping process itself, which drives rearrangement of the rafts on the membrane, is essential for creating the locally ordered state.

The team then tested whether this nanoscale reconfiguration could influence the shape of microscopic vesicles. Giant unilamellar vesicles, or GUVs, are synthetic cell models composed of lipid bilayers that enclose an aqueous interior. After incubating square DNA rafts with GUVs and allowing them to bind, the vesicles initially remained spherical. Upon adding unlocking strands to trigger the transformation to elongated rafts, notable membrane deformations appeared within 30 minutes and became more pronounced over time.

Control experiments confirmed that the deformation was tied to the dynamic reshaping process. Adding non-specific DNA strands did not deform the vesicles, and GUVs bound with pre-assembled elongated rafts also remained spherical. The rearrangement of the elongated rafts into locally ordered domains generates steric pressure that bends the membrane, collectively converting local deformations into large scale remodeling. This effect is reminiscent of protein crowding on biological membranes and submembrane scaffolds.

When locking strands were added, the elongated rafts transformed back to the square state, disorder returned, and the deformed GUVs recovered their spherical shape. Fluorescence recovery after photobleaching experiments showed that the square rafts and their reverted counterparts were highly mobile on the membrane, whereas the interacting elongated rafts displayed very low mobility and effectively sculpted the vesicle morphology.

Engineering the Interaction

The researchers systematically investigated key parameters that influence the interaction between DNA rafts and GUVs. Surface density played a critical role. At low densities, around 36 rafts per square micrometer, neither the conformation state of the rafts nor osmotic conditions notably affected vesicle shape. At an intermediate density of approximately 60 rafts per square micrometer, substantial differences among the three states became apparent. At higher densities, surface crowding began to dominate, and deformations occurred even in the square state.

The number and pattern of cholesterol anchors also proved important. Free, non-patterned cholesterol DNA strands failed to induce deformations. Using the programmability of DNA origami, the team designed variants with 4, 8, 12, and 16 cholesterol attachment sites arranged in different patterns. With only 4 or 8 anchors, the membrane affinity was too weak to offset the energy required for remodeling, and no significant deformations occurred. With 16 anchors, strong interactions caused deformations even in the square state.

Among the 12 anchor variants, the pattern that produced the most distinct cholesterol arrangement between square and elongated states yielded the highest degree of vesicle reshaping. This pattern featured tightly spaced cholesterol sites arranged in parallel lines that expanded widely from the center to the ends, reinforcing the elongated shape and anchoring the raft firmly to the membrane.

Forming Synthetic Channels

The most striking discovery came when the researchers integrated biogenic pores into the system. They reconstituted OmpF, a bacterial outer membrane protein with a pore size of about 1.1 nanometers, onto the vesicle membranes. OmpF allows the passage of small molecules under 600 daltons but excludes larger proteins like green fluorescent protein, which weighs roughly 27 kilodaltons.

The team set up a four stage experiment. In stage one, square raft bound GUVs were introduced into a solution containing GFP. The vesicles remained spherical and no GFP entered. In stage two, unlocking strands triggered the transformation to elongated rafts, and the expected deformations appeared, but the interior of the vesicles remained dark. In stage three, OmpF was added to the membrane. Gradually, the vesicles recovered to a spherical shape, yet still no GFP influx occurred.

This recovery process was driven by the exchange of small solutes such as sucrose, sodium, magnesium, and chloride ions through the OmpF pores, which restored osmotic balance across the membrane. As the concentrations of these small molecules equalized, the free energy required to restore the spherical shape was supplied by the entropy gained from molecular diffusion. During this dynamic recovery, strong fluctuations in membrane curvature mediated membrane perforation by the locally ordered elongated raft architectures, likely from their ends given the high aspect ratio.

Remarkably, in stage four, after full shape recovery, GFP began to enter the vesicles. Within roughly 30 minutes, a homogeneous distribution of GFP appeared inside. Because OmpF excludes GFP by size, other membrane channels with dimensions far larger than the OmpF pore must have formed. These synthetic channels were stable, as the influx process persisted for over 40 minutes.

Control experiments confirmed the necessity of both dynamic reconfiguration and dynamic shape recovery. When pre-assembled elongated rafts and OmpF were bound to vesicles, no GFP influx occurred. When square rafts at high density, which directly induce deformations, were combined with OmpF, shape recovery occurred but no channels formed. On supported lipid bilayers, where no genuine three dimensional membrane effects exist, no synthetic channels were detectable by atomic force microscopy or electrophysiology.

The data suggest that the elongated rafts, which align their hydrophilic surfaces side by side in local order, partially insert their hydrophobic surfaces into the membrane. During the shape recovery driven by small solute exchange, membrane curvature fluctuations facilitate perforation, and the aligned rafts likely shield lipid tails from hydration, stabilizing transmembrane channels.

Controlling Channel Size and Gating

To estimate channel dimensions, the researchers tested the transport of fluorescently labeled dextran molecules of various molecular weights. Small molecules like Cy3 maleimide at 650 daltons and 20 kilodalton dextran passed through almost completely. Larger molecules showed broader distributions, with 70 kilodalton dextran exhibiting roughly 50 percent permeability. Molecules of 150 and 500 kilodaltons were mostly excluded, and 2,000 kilodalton dextran was nearly impermeable. Based on the hydrodynamic diameter of 70 kilodalton dextran, the estimated channel size is around 15 nanometers, far exceeding the dimensions of most protein pores.

The synthetic channels could also be reversibly sealed. Adding locking strands at stage four triggered transformation of the elongated rafts back to the square state, lifting membrane perforation. Fluorescence recovery after photobleaching showed that GFP within the vesicles could no longer be replenished from outside, confirming channel closure. Conversely, without locking strands, fluorescence rapidly recovered, indicating sustained influx through open channels.

To demonstrate temporal control, the team added locking strands at different intervals after the initial deformation. Adding them after 30 minutes resulted in very low GFP influx, as the transformation to square rafts disrupted local order and inhibited perforation. Adding them at 60 or 80 minutes allowed progressively more transport, while addition at 120 minutes, when equilibrium was nearly reached, had minimal effect on the already completed influx.

Enzyme Reactions in Synthetic Cells

The distinct sizes and functions of the synthetic channels and OmpF enabled step by step transport of different reactants into vesicles, offering high spatiotemporal control over enzyme cascades. The researchers encapsulated glucose oxidase, a 160 kilodalton enzyme, inside GUVs during their formation. At stage four, when synthetic channels were open, they added the substrate Amplex Red, which diffuses passively across membranes, and then introduced Cy5 labeled myoglobin, a 17 kilodalton protein, to the exterior.

The synthetic channels allowed myoglobin to enter the vesicles, as evidenced by homogeneous distribution in the vesicle lumen. Locking strands then sealed the channels, encapsulating myoglobin and preventing its diffusion during a washing step to remove external residuals. Subsequently, glucose, a small 180 dalton molecule, diffused into the vesicles via OmpF, immediately triggering the glucose oxidase and myoglobin enzyme cascade and producing fluorescent resorufin inside the vesicles.

The researchers also demonstrated that the phenomenon is not specific to OmpF. They replaced it with the protein translocase of the outer membrane, or TOM, which has a pore size of approximately 2.5 nanometers and a molecular weight cutoff around 6 kilodaltons. Using mitochondrial preproteins of roughly 40 kilodaltons as cargo, they observed the same sequence of membrane deformations, vesicle shape recovery, synthetic channel formation, and cargo transport. Importantly, preproteins were not effectively translocated in vesicles reconstituted with TOM alone, demonstrating that the synthetic channels bypass the need for threading cargo through nanoscale pores or engaging substrate specific translocases.

Why It Matters

The work outlines a versatile platform for interfacing reconfigurable DNA nanostructures with synthetic cells. By coupling programmable nanoscale shape changes to microscale membrane remodeling, the system translates molecular engineering into cellular behavior. The locally ordered DNA raft architectures can perforate membranes in cooperation with biogenic pores, creating synthetic channels that enable transport of large cargo, up to about 70 kilodaltons, and that can be sealed on demand.

These synthetic channels overcome limitations of protein pores, which are often fragile and restricted to a narrow size range up to around 5 nanometers. Going beyond this size limit opens possibilities for directly shuttling or sensing large protein complexes and enzymes across membranes without intricate biological machinery. The modularity of DNA origami allows the rafts to be functionalized with diverse moieties, potentially enabling them to recognize and programmably puncture diseased cells or control the release of therapeutics from synthetic cells.

The system also provides a testbed for exploring how membrane proteins and cells interact. By systematically varying DNA raft density, cholesterol anchor patterns, and osmotic conditions, researchers can probe the interplay between molecular shape and cellular morphology. Such insights may shed light on how proteins have evolved their biological functions.

The reshaping of DNA rafts and vesicles reveals an intimate mutual influence. The DNA rafts impose deformation on the vesicle, and in turn, the dynamic shape recovery of the vesicle assists the rafts in forming locally ordered architectures that perforate the membrane. This feedback loop, absent in simpler systems, highlights the importance of genuine three dimensional membrane effects in synthetic cell engineering.

Rather than attempting to completely replicate the complexity of living cells, the work exemplifies a strategy of designing fully artificial DNA based platforms with engineered features that do not necessarily have biological equivalents but can function in biological environments. The result is a new way of thinking about synthetic biology, one that emphasizes programmability, reversibility, and the creative use of DNA nanotechnology to expand the functional repertoire of artificial cells.

Credit & Disclaimer: This article is a popular science summary written to make peer-reviewed research accessible to a broad audience. All scientific facts, findings, and conclusions presented here are drawn directly and accurately from the original research paper. Readers are strongly encouraged to consult the full research article for complete data, methodologies, and scientific detail. The article can be accessed through https://doi.org/10.1038/s41563-024-02075-9