mRNA-based therapeutics offer significant potential, providing unparalleled speed and flexibility in the development of vaccines and cancer immunotherapies. However, despite their success, traditional non-viral delivery methods have been limited by an implicit rule: they have mostly relied on electrostatic adsorption. To compact and protect the negatively charged genetic material, researchers have almost always used positively charged carriers such as synthetic cationic polymers, ionizable lipids, and cationic peptides.
Electrostatic interactions are undeniably strong and simple, but they also lack directionality. In the complex biological environment inside the body, this indiscriminate positive charge can have negative effects. In our routine laboratory work, we frequently observed the hidden drawbacks of this method: cationic condensing agents act like tiny detergents, nonspecifically inserting into cell membranes and disrupting their structure. This leads to unintended membrane damage, cell toxicity, and unwanted immune reactions. This persistent membrane disruption prompted us to ask a seemingly simple yet radical question: Are positive charges truly necessary for mRNA delivery? Is it possible to fully condense intact mRNA without them, relying solely on neutral molecules?
Our inspiration came from the fundamental structure of nucleic acids. Unlike double-stranded nucleic acids, where bases are tightly packed inside the helix, single-stranded nucleic acids expose many unpaired bases, offering a strong opportunity for molecular recognition. We hypothesized that by combining specific hydrogen-bond recognition with the cooperative effects of supramolecular self-assembly, we could replace the need for strong electrostatic attraction. With this in mind, we designed Electrically Neutral Self-Assembling Peptides (ENSPs), composed of three parts: an inner segment for recognizing nucleobases, a neutral α-helical coiled-coil region in the center, and hydrophilic neutral sugar layer on the outside. Our aim was to promote the natural condensation of single-stranded mRNA by means of complementary hydrogen bonding and hydrophobic interactions between helices, thereby entirely removing the requirement for cationic elements.
However, progress was challenging. When we tested our first-generation prototype, which included only a single adenine base, the binding to single strand DNA was very weak. We required an extremely high base-to-base ratio of 12,800:1 just to observe any interaction. With such low efficiency, condensing complex, full-length mRNA was practically impossible. This significant deficiency made it clear that a design relying on a single hydrogen bond was too weak to effectively bind the mRNA.
This challenge prompted us to reconsider the peptide design in a systematic and rational way, resulting in our second-generation system. First, we incorporated tandem double nucleobases to enable “multivalent recognition.” Next, to enhance the driving force behind self-assembly, we introduced a non-natural hydrophobic amino acid, 2-aminoisobutyric acid (Aib), into the hydrophobic interface of the coiled coil. The inclusion of Aib proved to be a significant breakthrough: it helped preserve a stable helical structure while substantially improving hydrophobic and cooperative interactions among the peptide molecules.
After resolving the affinity issue, we encountered an even greater challenge: how to deliver sequence-diverse, full-length mRNAs without needing to create dozens of different custom peptides for each new nucleic acid. By conducting a thorough statistical analysis of dinucleotide tandem repeat frequencies across mRNA sequences, we devised a combinatorial strategy using only four universal building blocks: (thy)₂-U₈, (gua)₂-U₈, (cyt)₂-U₈, and (ade)₂-U₈. When these were combined and annealed at carefully optimized stoichiometric ratios, the outcome was remarkable. These four neutral peptides effectively condensed full-length EGFP and luciferase mRNA, reducing the optimal base ratio to 3:1. This represented a 4,300-fold increase in condensation efficiency compared to our initial prototype.
The most thrilling aspect was undoubtedly observing the structure directly through Cryo-Transmission Electron Microscopy (Cryo-TEM). Instead of forming typical precipitates, the neutral peptides cooperatively wrapped around the mRNA strand and spontaneously assembled into highly ordered, hollow nanotubes with 11 nm outer diameter and 3 nm core. It is closely resembling the natural Tobacco Mosaic Virus (TMV) capsid.
Ultimately, this breakthrough was validated in our biological experiments. These neutral nanotubes not only facilitated highly efficient cellular uptake via glucose transporters and generated strong in vivo protein expression, but they also demonstrated excellent biosafety. In precise in vitro and in vivo hemotoxicity tests, commercial lipid nanoparticles (LNPs) and cationic carriers caused significant red blood cell deformation and hemolysis. In contrast, red blood cells treated with ENSPs maintained their normal, healthy biconcave disc shape perfectly.
Looking back, this study reflects our effort to explore alternative principles in nanocarrier design. Rather than relying on conventional electrostatic interactions, it suggests that well-balanced, cooperative supramolecular forces can achieve effective assembly on their own. We hope these findings provide a useful perspective for developing safer and more biocompatible next-generation mRNA delivery systems.