Engineering Synthetic Protein Assemblies to Create Bottom-Up RNA Delivery Vehicles

Molecular cloning and plasmid construction

All plasmids were constructed using established molecular cloning methods. DNA fragments were generated by PCR with Platinum SuperFi II Master Mix (Thermo Fisher Scientific) and appropriate oligonucleotides (IDT DNA), by digestion of plasmids with standard restriction enzymes (NEB), or by commercial gene synthesis (Twist Bioscience or IDT DNA). Fragment assemblies were performed with NEBuilder HiFi DNA Assembly Mix (NEB) or Instant Sticky-end Ligase Master Mix (NEB). The assembled products were transformed into self-prepared chemically competent Escherichia coli DH5α cells. Candidate clones were screened by plasmid extraction using the Monarch Plasmid Miniprep Kit (NEB) and verified by Sanger sequencing (Azenta) or rolling-circle amplification directly from bacterial cells (Microsynth). Confirmed plasmids were purified with a Plasmid Maxiprep Kit (QIAGEN) and used for cell transfection. All sequences generated in this study are provided in Supplementary Table 1.

Plasmid transfection protocol

Cells were seeded one day before transfection at 3.0 × 10⁴ cells per well in 96-well plates, 2.2 × 10⁵ cells per well in 24-well plates, 7.5 × 10⁵ cells per well in 6-well plates, or 4.0 × 10⁶ cells per 10-cm dish. JetOptimus DNA transfection reagent (Polyplus transfection) was used according to the manufacturer’s instructions with 75 ng DNA per well in 96-well plates, 300 ng DNA per well in 24-well plates, 1 µg DNA per well in 6-well plates, and 5 μg DNA per 10-cm dish.

Cell culture conditions and engineered cell lines

HEK293T cells, provided by the Institute of Developmental Genetics, Helmholtz Munich, were cultured at 37 °C in a humidified atmosphere containing 5% CO₂. Cells were maintained in Dulbecco’s modified Eagle medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco) and 1% penicillin–streptomycin (Gibco). The HEK293T split-Luc reporter cell line was generated by Cas9-mediated cleavage at the AAVS1 locus followed by homology-directed integration of a donor construct encoding LgBiT, the carboxy-terminal fragment of Fluc separated by a P2A sequence, and a puromycin-resistance gene. Three days after transfection, cells were selected for 2 weeks in medium containing 2 µg ml⁻¹ puromycin (Thermo Fisher Scientific).
HEK293T cells stably expressing EGFR or IL-7Rα were generated by amplifying the EGFR sequence from Addgene plasmid 23935, provided by W. Hahn and D. Root, or by synthesizing the IL-7Rα sequence (Twist Bioscience). Each coding sequence was cloned into an AAVS1 knock-in donor plasmid, cotransfected with AAVS1-targeting Cas9, and selected with 2 µg ml⁻¹ puromycin. Dual-positive EGFR/IL-7Rα receptor cells were produced by cloning IL-7Rα into an AAVS1 donor plasmid containing a blasticidin-resistance gene. Transfected cells were selected in medium containing 10 µg ml⁻¹ blasticidin (Thermo Fisher Scientific).

Quantification of STV-mediated target RNA release

Supernatants from STV-producing cells were collected 48 h after production and passed through 0.45-µm polyvinylidene fluoride (PVDF) filters (Merck Millipore). RNA was extracted using the Monarch Total RNA Miniprep Kit (NEB). The purified RNA was analysed by one-step reverse-transcription quantitative PCR (RT–qPCR) with the Luna Universal One-Step RT-qPCR Kit (NEB) and a custom primer/FAM-probe set specific to EGFP mRNA (Metabion). Fluorescence was measured with a QuantStudio 7 Flex real-time PCR system (Thermo Fisher Scientific).

Incorporation of diffusion-designed symmetric oligomers into STV constructs

Previously designed RFdiffusion symmetric oligomers were screened for successful assembly using size-exclusion data³. All D2-symmetric oligomers were excluded. The remaining 39 sequences were synthesized as eBlocks (IDT DNA) and cloned as C-terminal fusions to the additional STV components PHPLC, SynL and tdPCP.

Integration of structure-mined membrane-binding domains into STV design

The pleckstrin homology domain (PDB: 1MAI) was used as the query structure for a structure-based homology search with FoldSeek^18,40. Ten sequences from three groups—other species, human proteins and metagenomic proteins—were selected according to their highest structural homology to the input domain. Each sequence was synthesized by IDT DNA and fused to the amino terminus of the previously identified ideal STV construct containing SynL, tdPCP and HE0690.

Sequence and structural alignment of membrane-binding domains

The amino acid sequences of the structure-mined membrane-binding domains were aligned with the corresponding sequence from R. norvegicus PHPLCδ using the residues resolved in the X-ray structure (PDB: 1MAI). Alignments were generated with the MAFFT v.7 add tool⁴¹ using default settings: the auto strategy, BLOSUM62 scoring matrix, a gap-opening penalty of 1.53 and an offset value of 0.0.
For structural comparisons, the membrane-binding domain regions were extracted from AlphaFold2 predictions⁴² for the human and other-species proteins or from ESMFold predictions¹⁹ for the metagenomic proteins. The extracted structures were aligned, and root mean square deviation relative to the X-ray structure (PDB: 1MAI) was calculated using the PyMOL super alignment tool.

Screening symmetric oligomers and membrane-binding proteins for RNA transfer

Cells were seeded in 96-well plates and transfected with individual oligomeric or membrane-binding STV constructs together with plasmids encoding VSV-G and N-split-Luc-PP7 at a 2:1:7 ratio. At 24 h post-transfection, 5 µl of supernatant was collected, mixed with 45 µl phosphate-buffered saline (PBS), and analysed using the Nano-Glo HiBiT Lytic Detection System (Promega) on a Centro LB960 luminometer (Berthold Laboratories) with a 0.5-s integration time.
At 48 h post-transfection, 120 µl of supernatant was collected and clarified by centrifugation through a 0.45-µm PVDF 96-well filter plate (Sigma-Aldrich) at 1,500g for 20 min at 4 °C. The cleared supernatant was transferred to a seeded 96-well plate containing C-split-Luc reporter cells. After a further 24 h, the supernatant was completely removed and cells were analysed using the Nano-Glo Dual-Luciferase Reporter Assay (Promega). STV uptake was quantified from NanoLuc substrate luminescence, whereas uptake and expression of N-split-Luc-PP7 mRNA were determined from Fluc substrate luminescence. Total STV protein transfer and Fluc protein expression in recipient cells were additionally quantified by comparing screen-derived luminescence with a Fluc reference sample (Abcam) or a HiBiT control protein (Promega).

Flow-cytometry validation of STV-mediated EGFP mRNA delivery

HEK293T producer cells were transfected in 24-well plates with plasmids encoding STV constructs, VSV-G and EGFP-PP7 at a 2:1:7 ratio. STV-containing supernatants were collected on 2 consecutive days, filtered through 0.45-μm PVDF membrane filters and concentrated five- to tenfold with Lenti-X Concentrator (Takara Bio) in fresh DMEM. Resuspended STVs, 10–20 µl per well, were added to HEK293T cells in 96-well plates. After 24 h, cells were detached with StemPro Accutase (Thermo Fisher Scientific), resuspended in FACS buffer containing EDTA and bovine serum albumin (BSA), and passed through tubes fitted with cell strainers. Living single cells were gated, and EGFP mRNA uptake and expression were measured by flow cytometry using a BD FACSaria III instrument (BD Biosciences). Data were processed with BD FACSDiva v.6.1.3 and FlowJo v.10.

Design of additional C8-symmetric oligomers

Additional C8-symmetric oligomers were designed with the open-source version of RFdiffusion and the script provided for symmetric oligomer generation³. Calculations were performed on a single A100 GPU.

Subcellular localization of STV constructs

HEK293T cells were transfected with STV constructs containing different membrane-binding domains. After 24 h, cells were fixed with 10% formalin (Sigma-Aldrich) and permeabilized with 1% BSA and 0.5% Triton X-100 diluted in PBS. Cells were incubated overnight at 4 °C with primary anti-HA antibody (Sigma-Aldrich, H3663), washed and stained overnight at 4 °C with Alexa 488-conjugated donkey anti-mouse antibody (Thermo Fisher Scientific, A21202). Samples were mounted with ProLong Diamond reagent (Thermo Fisher Scientific) and imaged using an Axio Imager M2 fluorescence microscope (Carl Zeiss).

Flow-cytometry analysis of STV packaging capacity

EGFP-PP7-STVs were produced in 24-well plates as described above. Producer cells were also transfected with mRuby3-PP7 constructs containing random untranslated-region (UTR) sequences of different lengths. Concentrated STVs were transferred to HEK293T target cells, and EGFP and mRuby3 expression were quantified by flow cytometry 24 h later.

Ultracentrifugation-based STV concentration

Producer cells were seeded in poly-l-lysine-coated 10-cm dishes (Sigma-Aldrich) and transfected with plasmids encoding the STV-C8 components required for each experiment. Unless stated otherwise, supernatants were collected for 3 consecutive days and stored at 4 °C until collection was complete. Pooled supernatants were centrifuged for 5 min at 1,000g and filtered through a 0.45-μm PVDF membrane. The filtrate was layered over a 20% (w/v) sucrose cushion prepared in PBS (Sigma-Aldrich). Samples were ultracentrifuged at 26,000 rpm for 2 h at 4 °C in an SW28 rotor using an Optima L-60 ultracentrifuge (Beckman Coulter). The supernatant and sucrose cushion were removed, and the pellet was resuspended in 50 µl ice-cold 1× PBS (Thermo Fisher Scientific) on an orbital shaker at 150 rpm for 45 min at 4 °C. The suspension was clarified at 1,000g for 5 min at 4 °C to remove debris and stored at −80 °C. This procedure produced an approximately 300-fold concentration.

Assessment of STV purity

STV-C8 preparations concentrated by ultracentrifugation were assessed by silver staining. Samples were heated for 10 min at 98 °C in 2× Laemmli buffer (Sigma-Aldrich) and separated on 4–15% gradient TGX gels (Bio-Rad) in 1× Tris/glycine/SDS running buffer (Bio-Rad) for 60 min at 130 V. Gels were silver-stained according to the manufacturer’s instructions (Serva). An identical gel was run in parallel and transferred to a nitrocellulose membrane for 60 min at 100 V and 4 °C in Tris/glycine transfer buffer (Bio-Rad). STV-C8 protein position was identified by imaging the membrane with the Nano-Glo HiBiT Blotting System (Promega) on a Fusion SL Vilber instrument (Peqlab). The HiBiT signal was used to locate STV proteins on the corresponding silver-stained gel.

Cryo-electron tomography sample and grid preparation

Producer cells were seeded in poly-l-lysine-coated 10-cm dishes and transfected with plasmids encoding STV-C8 and EGFP-PP7. Twenty-four hours later, cells were washed with PBS and cultured in serum-free DMEM. After an additional 24 h, the supernatant was collected and concentrated by ultracentrifugation as described above. Purified STV-C8 vesicles were diluted to 10⁹ particles per microlitre in PBS. Samples were applied to 200-mesh copper EM grids containing Quantifoil R 3.5/1 holey carbon films (Quantifoil) covered with a homemade 3-nm continuous carbon film prepared by flotation. Grids were glow-discharged at 4 mA for 10 s, blotted and plunge-frozen in liquid ethane with a Vitrobot IV (Thermo Fisher Scientific) at 95% humidity and 10 °C. Sixteen grids were prepared across two experiments.
For additional size-distribution analysis, STV-C8 vesicles were separated on an iodixanol gradient containing 15%, 25%, 40% and 60% iodixanol in PBS-MgCl₂/KCl/NaCl buffer. FBS-containing STV-C8 supernatant was prepared as described above, loaded onto the gradient and ultracentrifuged at 26,000g for 4.30 h at 4 °C. Fractions were collected by puncturing the tube wall with a 27-G needle. Samples were applied to 200-mesh copper EM grids with Quantifoil R 3.5/1 holey carbon films and a homemade 3-nm continuous carbon layer. After glow discharge at 4 mA for 10 s, grids were blotted and plunge-frozen in a liquid ethane/propane mixture using a Vitrobot IV at 95% humidity and 4 °C. Eighteen grids were prepared in one experiment, with two grids for each triplicate across the 15%, 25% and 40–60% iodixanol conditions.

Cryo-ET data acquisition, reconstruction and vesicle-size analysis

Tilt series were acquired with Tomo5 software on a Titan Krios G4 transmission electron microscope (Thermo Fisher Scientific) equipped with a cold field-emission gun operated at 300 kV, a Falcon IVi camera and a Selectris X energy filter. Images were collected at ×81,000 magnification, corresponding to a pixel size of 1.63 Å, across a −60° to 60° angular range with 2° increments using a dose-symmetric tilt scheme. The total dose was 122 e⁻ per Ų, and the target defocus ranged from −2.5 to −4 µm. Data were recorded in EER format. Vesicle-size distribution statistics were obtained from search maps collected at ×11,500 magnification using Tomo5. At least 500 search-map micrographs were collected per grid in the first experiment with purified STV-C8 vesicles in PBS. In the iodixanol-gradient experiment, 1,125 search-map micrographs were collected per grid.
Tilt-series alignment and tomogram reconstruction were performed with AreTomo3⁴³ using fourfold binning, resulting in a final pixel size of 6.52 Å per pixel. Frame alignment and contrast-transfer-function estimation used the MotionCor3⁴³ and GCtfFind⁴⁴ implementations in AreTomo3, respectively. Aligned tilt series were inspected manually, and problematic tilts were removed before reconstruction. Membranes were segmented with MemBrain-seg, while particles were manually segmented in Amira⁴⁵ (Thermo Fisher Scientific).
Assembly-size homogeneity was determined by morphometric analysis of cryo-electron tomograms. Diameters were measured directly from tomographic slices for 500 individual assemblies using calibrated pixel distances. Values were converted to physical units using the tomogram pixel size and compiled into an assembly-size frequency distribution.

Characterization of STV-C8 particles from iodixanol gradients

STV-C8(EGFP) vesicles were purified using iodixanol gradients as described above. RNA was extracted from three gradient fractions, and the relative abundance of EGFP mRNA was measured by RT–qPCR. STV protein content was assessed by HiBiT measurement. Delivery activity was evaluated by adding each fraction to target cells and measuring EGFP expression by flow cytometry 24 h later.

AlphaFold3 prediction of the STV-C8 multimer structure

The STV-C8 sequence, UaPHPLC-SynL-tdPCP-HE0690, was submitted to the AlphaFold3 server using octameric settings⁴⁶. The predicted structure was coloured according to pLDDT (predicted local distance difference test) scores and captured in two orientations.

Live-cell measurement of STV-C8 uptake kinetics

VSV-G-enveloped STV-C8 vesicles were produced and added to split-Luc reporter cells seeded the previous day in black-walled 96-well plates. Nano-Glo Endurazine live-cell substrate (Promega) was added to the treated cells, and plates were placed in a Cytation 3 plate reader (Agilent). Luminescence from reconstituted HiBiT/LgBiT NanoLuciferase was recorded every 15 min for 3 days.

RNA-sequencing analysis of STV-C8 cargo

STV-C8 particles were produced and purified as described above. RNA was isolated from the particles and matched producer cells with the Monarch Total RNA Miniprep Kit (NEB). Illumina RNA-sequencing libraries were prepared and sequenced on a NovaSeq instrument, generating 20 million paired-end reads per sample. Reads were mapped to the human reference transcriptome with STAR, and differential expression was assessed using DESeq2. Library preparation, sequencing and bioinformatic analysis were performed by Azenta (Leipzig).

Mass-spectrometry analysis of STV-C8 protein cargo

STV-C8 particles were generated and purified as described above. Total protein was extracted with PreOmics lysis buffer supplemented with cOmplete Protease Inhibitor (Roche) and quantified by BCA assay (Thermo Fisher Scientific). A 10-µg aliquot from each sample was processed by filter-aided sample preparation⁴⁷ and analysed using a Q Exactive HFx mass spectrometer coupled online to an Ultimate 3000 RSLC system (Thermo Fisher Scientific). Label-free quantification was performed with MaxQuant 2.4.9.0 (MPI)⁴⁸ using a combined database containing SwissProt human proteins and sequences of exogenously expressed proteins. Statistical analyses were conducted in Perseus (MPI)⁴⁹.

Gene-set enrichment analysis of STV RNA and protein cargo

Significantly enriched or depleted genes were defined using an adjusted P < 0.005 and log₂[fold change] thresholds of +3 or −3. Proteins were considered significant using −log q < 0.05 and log₂[fold change] thresholds of +3 or −3. Gene-set enrichment analysis was performed with gProfiler2 using default settings (e111_eg58_p18_30541362).

Native PAGE analysis of purified STV-C8 vesicles

HEK293T cells were transfected with STV-C8 and an unrelated control plasmid in poly-l-lysine-coated T175 flasks. Twenty-four hours after transfection, cells were washed with PBS and supplied with serum-free DMEM. After a further 24 h, supernatants were collected and concentrated by ultracentrifugation. Concentrated samples were lysed with M-PER (Thermo Fisher Scientific) for 10 min at room temperature, mixed 1:4 with native PAGE sample buffer (Invitrogen), and loaded onto 12% Tris-glycine gels alongside NativeMark Unstained Protein Standards (Invitrogen). Electrophoresis was performed for 2 h at 150 V in Tris/glycine buffer (Bio-Rad). Gels were stained with Coomassie solution (Thermo Fisher Scientific) for 30 min and imaged using a Fusion SL Vilber system (Peqlab).

Comparison of STV-mediated EGFP mRNA delivery with SEND, EPN and VLP systems

SEND/MmPeg10 was obtained from Addgene (174858; provided by F. Zhang). EPN-MCP, VLP-MCP, EGFP-MS2 and SEND-EGFP constructs were synthesized by Twist Bioscience and cloned into CAG-promoter expression backbones. For each platform, capsid-scaffold and EGFP cargo-RNA plasmids were cotransfected with VSV-G in 24-well plates. Supernatants were produced for 48 h and concentrated as described above. Concentrated delivery vehicles were added to HEK293T, Vero E6, N2a and HepG2 cells. The other cell lines were provided by the Institute of Virology, Helmholtz Munich. EGFP expression was quantified by flow cytometry 24 h later.
A reporter plasmid containing a lox-stop-lox cassette upstream of the EGFP coding sequence was also constructed. VLP, EPN and SEND coding plasmids (Addgene 205525, 205555 and 174858) and Cre cargo plasmids (Addgene 174862 and 205559) were obtained from Addgene and compared directly with unmodified STV-C8 packaging of Cre mRNA. Each platform was cotransfected with its respective Cre cargo plasmid and VSV-G in 10-cm dishes. Supernatants were produced for 72 h and concentrated by ultracentrifugation. HEK293T cells were transfected with the reporter plasmid and, 24 h later, treated with two concentrations of concentrated supernatant. EGFP expression was measured by flow cytometry after an additional 24 h.

In vitro and in vivo comparison of STV-C8 and lipid nanoparticles

A plasmid encoding EGFP under a T7 promoter was generated and linearized downstream of the stop codon, leaving a 3′ UTR approximately equivalent in length to that of the STV cargo plasmid. The linearized product was purified with the Monarch DNA Cleanup Kit (NEB) and used for in vitro transcription with the HiScribe T7 Quick High Yield RNA Synthesis Kit (NEB). RNA was purified with the Monarch RNA Cleanup Kit (NEB), capped with the Vaccinia Capping System (NEB), purified again and polyadenylated with E. coli poly(A) polymerase (NEB). Following a final purification, EGFP mRNA was diluted to 150 ng µl⁻¹ in 20 mM citrate buffer, pH 4.0.
Lipid nanoparticles (LNPs) contained ALC-0315 (Cayman Chemical, 34337), DOPE (Avanti Polar Lipids, 850725), cholesterol (ChemCruz, sc-202539) and DMG-PEG 2000 (Avanti Polar Lipids, 880151) at a ratio of 50:10:38.5:1.5. Lipid and RNA solutions were rapidly mixed at a 1:3 volume ratio to produce a final weight ratio of 40:1. For particle-size analysis, 1 µl of the formulation was diluted in 3 ml PBS in a Sarstedt cuvette and measured by dynamic light scattering using a Zetasizer Pro (Malvern Panalytical). STV-C8 particles containing EGFP mRNA were prepared by ultracentrifugation. Absolute STV-C8 protein levels were calculated by comparison with a HiBiT Control Protein standard curve (Promega). EGFP mRNA content was determined by absolute RT–qPCR using an in vitro-transcribed EGFP mRNA standard, and STV-C8 particle numbers were measured by dynamic light scattering.
STV-C8 and LNP delivery efficiency were compared in two ways. First, the mRNA dose required to produce EGFP expression in the same proportion of target cells was determined by titrating each vehicle and analysing EGFP expression by flow cytometry. Second, expression levels were compared after titrating STV-C8 or LNPs to produce EGFP expression in 50% of cells. EGFP fluorescence at the corresponding mRNA concentrations was used to calculate the theoretical mRNA dose required to produce one mean fluorescence intensity unit.
For in vivo analysis, 8-week-old female C57BL/6 mice received intravenous injections of 150 µl LNPs or concentrated STV-C8 formulated with Akaluc mRNA. Both formulations were prepared as described above. Twenty-four hours later, mice received 150 µl of 33 mM TokeOni substrate by intraperitoneal injection. Bioluminescence was recorded with an IVIS Lumina S5 imaging system (PerkinElmer) using a 20-s exposure. Animal procedures followed institutional guidelines at the Helmholtz Munich Center German Mouse Clinic and were approved by the Ethical Review Board of the Government of Upper Bavaria.

Evaluation of STV-C8-induced interferon signalling

A549-IFN-GFP reporter cells, provided by R. Bartenschlager, were used to monitor interferon activation through GFP expression. Cells were transfected with luciferase plasmid DNA as a positive control or treated with STV-C8 particles containing luciferase mRNA. GFP expression was assessed 24 h later using an EVOS imaging system (Thermo Fisher Scientific).
In vitro-transcribed mTagBFP2 mRNA was prepared as described above using either N¹-methylpseudouridine (Jena BioScience) or unmodified uridine. A549-IFN-GFP cells were treated with STV-C8 containing mTagBFP2 mRNA or with LNPs containing modified or unmodified mTagBFP2 mRNA. Vehicle amounts were adjusted to produce similar mTagBFP2 expression levels. Interferon–GFP activation was subsequently quantified by flow cytometry.

Comparison of LNP- and STV-C8-associated cytotoxicity

HEK293T cells were seeded in 96-well plates and treated with 50 ng EGFP mRNA-containing LNPs or purified STV-C8(EGFP) particles. Both delivery systems were applied at concentrations expected to produce EGFP expression in approximately 50% of cells. After 24 h, cells were detached with 0.05% trypsin (Thermo Fisher Scientific), resuspended in Annexin V binding buffer containing 10 mM HEPES, 140 mM NaCl and 2.5 mM CaCl₂, pH 7.4, and stained 1:100 with Annexin V-iFluor 680 (Abcam). Annexin V fluorescence was measured by flow cytometry.

Engineering cell-type-specific STV-C8 using peptide binders

Previously designed EGFRn, EGFRc and IL-7Rα minibinders²⁸, as well as a CD19 scFv, were displayed on the STV-C8 surface. Each binder was expressed as a fusion containing a signal peptide, minibinder sequence and transmembrane domain, together with STV-C8 components and an LDLR-binding-deficient VSV-G mutant (K63Q, R370Q²⁷). Constructs were transfected into 6-well plates with EGFP mRNA cargo. Supernatants were collected for 48 h and concentrated using Lenti-X Concentrator (Takara Bio). Thirty microlitres of concentrated supernatant was added to wild-type HEK293T cells or HEK293T cells stably expressing EGFR or IL-7Rα. EGFP expression was quantified by flow cytometry 24 h later.

STV-C8 delivery of EGFP mRNA to human retinal pigment epithelium spheroids

Human retinal spheroids were differentiated from the healthy-donor hiPS cell line F49B7, which had been validated for pluripotency markers and germ-layer differentiation potential. hiPS cells were cultured on Matrigel-coated six-well plates in mTeSR Plus medium (STEMCELL Technologies), with medium replacement every 2 days. At approximately 70% confluence, cells were passaged as small clumps using 0.5 mM EDTA (0.5 M, pH 8.6; Thermo Fisher Scientific). On day 0, hiPS cells were dissociated into small aggregates with 0.5 mM EDTA, suspended in cold growth-factor-reduced Matrigel (Corning), and incubated at 37 °C for 20 min to allow gel formation. Aggregates were gently dispersed in neural induction medium consisting of DMEM/F12 with GlutaMAX, 1% B27 with vitamin A, 0.5% N2, 0.1 mM 2-mercaptoethanol, 2 mM GlutaMAX and 1% penicillin–streptomycin. Cultures were maintained in ultra-low-attachment six-well plates. On day 5, floating cysts were transferred to Matrigel-coated six-well plates. On day 15, cysts were detached with dispase (0.5 mg ml⁻¹ in DMEM/F12; STEMCELL Technologies) for 3–4 min at 37 °C, washed with DMEM/F12 and cultured in retinal differentiation medium containing DMEM/F12 with GlutaMAX, 2% B27 without vitamin A, 1% NEAA and 1% penicillin–streptomycin. On day 25, immature retinal spheroids were transferred to retinal maturation medium containing DMEM/F12 with GlutaMAX, 8% FBS, 2% B27 without vitamin A, 1% NEAA, 1% antibiotic–antimycotic and 1% 100 mM taurine. Half the medium was replaced every 2–3 days. Spheroids were maintained at 37 °C and 5% CO₂ until analysis.
Retinal pigment epithelium developed as patches attached to the retinal spheroids. On day 200, RPE spheroids were dissected and transferred to 96-well U-bottom ultra-low-attachment plates, with 3–4 spheroids per well. Spheroids were treated with 10 µl STV-C8(EGFP)/VSV-G or STV-C8(EGFP), fixed 2 days later, and gradually dehydrated in 10% sucrose and 30% sucrose at room temperature followed by 50% sucrose overnight at 4 °C. Samples were embedded in Tissue-Tek O.C.T. compound (Sakura), frozen at −80 °C and sectioned at 10 µm with a Leica CM3050 S cryostat. Sections were rehydrated, blocked in 5% chemo-blocker solution (Merck) for 30 min and permeabilized with 0.3% Triton X-100 for 30 min. Sections were incubated overnight at 4 °C with anti-RPE65 (Proteintech, 17939-1-AP) and anti-GFP (Santa Cruz, sc-101536) antibodies diluted in 5% chemo-blocker solution. After three PBS washes, sections were incubated for 1 h at room temperature with goat anti-rat Alexa Fluor 488 and donkey anti-rabbit Alexa Fluor 555 secondary antibodies (Thermo Fisher Scientific). Sections were washed, mounted with Fluoroshield containing DAPI (Sigma-Aldrich) and imaged with a Leica TCS SP8 spectral confocal microscope.

STV-mediated EGFP delivery to human monocytes

Primary human monocytes (ATCC, CRL-3622) were seeded in 96-well plates and treated with 5 µl concentrated STVs containing EGFP mRNA. EGFP expression was quantified by flow cytometry 24 h after treatment.

Isolation of primary mouse astroglia and delivery of Ascl1 mRNA

Primary astrocytes were isolated from the cerebral cortex of postnatal day 5 C57BL/6N mice. Cortices were dissected, cut into small pieces and mechanically dissociated by vigorous pipetting. The cell suspension was centrifuged for 7 min at 1,300 rpm, and the pellet was plated in a T25 flask. Cells were cultured for 7–13 days in DMEM/F12 with GlutaMAX supplemented with 10% FBS, 10% penicillin–streptomycin, 5% horse serum, 4.5 g l⁻¹ d-(+)-glucose, 2% B27, 10 ng ml⁻¹ bFGF and 10 ng ml⁻¹ EGF. At 90% confluence, cells were passaged with 0.05% Trypsin/EDTA and approximately 75,000 cells were seeded onto poly-d-lysine-coated glass coverslips. After 24 h, 15 µl concentrated STV-C8 containing EGFP or Ascl1-P2A-EGFP was added. After 48 h, cells were fixed with 10% formalin and incubated overnight at 4 °C with anti-GFP (Abcam, ab13970) or anti-Mash1 (Abcam, ab211327) antibodies in PBS containing 1% BSA and 0.3% Triton X-100. Cells were washed and stained for 1–2 h in the dark at room temperature with Alexa 488-conjugated donkey anti-chicken or Alexa 594-conjugated donkey anti-rabbit secondary antibodies. Samples were counterstained with DAPI, mounted in Aqua Poly/Mount (Polyscience) and imaged with an Axio Imager M2 fluorescence microscope.

Exon 51 deletion in primary porcine fibroblasts

Two Cas9 sgRNA plasmids containing PP7 motifs in the sgRNA stem–loop and spacers targeting porcine dystrophin were constructed. The sgRNAs targeted intron 50 (AGAGTTCCTAAGGTAGAGAG) and intron 51 (ATAAAGATAAGAGCTGGCAG) to delete exon 51 (ref. ¹³). A plasmid encoding Cas9 fused to nuclear localization and export signals, together with a 3′ UTR PP7 motif, was also generated. HEK293T producer cells were seeded in poly-l-lysine-coated 10-cm dishes and cotransfected with Cas9 mRNA, both sgRNA plasmids, STV-C8 plasmids and VSV-G in a 1:1:1 ratio for the Cas9 and sgRNA components. STV-C8 particles were collected and concentrated by ultracentrifugation. Primary porcine fibroblasts were seeded in collagen-coated 48-well plates in DMEM⁵⁰ supplemented with 1% NEAA, 10 mM HEPES, 15% FBS and 2-mercaptoethanol. Cells were treated with 20 µl STVs for 72 h. Genomic DNA was extracted with the Monarch Genomic DNA Purification Kit (NEB), and a 2-kb region spanning the targeted deletion was amplified using Platinum SuperFi II Master Mix with primers CCCATGACATTTACCCTATTATTATCCC and GCTAATGTTCATTTTAAAAAGGAATCTGTC. Products were separated on a 1.5% agarose gel and imaged.

STV delivery of Cas13d-NCS to SARS-CoV-2-infected human lung cells

For lung differentiation, hiPS cells (ISFi001-A, RRID: CVCL_YT30) were cultured in StemMACS medium (Miltenyi Biotec) on Geltrex Reduced Growth Factor-coated plates. At 70% confluence, colonies were dissociated into single cells with Accutase for 5 min at 37 °C, neutralized with StemMACS medium and centrifuged for 3 min at 200g at room temperature. Cells, 1.0–1.2 × 10⁶, were seeded in non-adherent six-well plates in StemMACS medium containing 10 μM Y-27632. Differentiation basal medium consisted of DMEM/F12 1:1 with GlutaMAX, 1× NEAA, 0.1% Albumax and 1× B27. Embryoid-body formation was induced by switching to a 1:1 mixture of StemMACS medium and differentiation basal medium containing 20 ng ml⁻¹ activin A. The medium was then replaced with differentiation basal medium containing 20 ng ml⁻¹ activin A for 48 h.
Definitive endoderm was induced from days 0–5 by plating embryoid bodies at 7 bodies per cm² on Geltrex-coated plates in differentiation basal medium containing 150 ng ml⁻¹ activin A and 25 ng ml⁻¹ BMP4, with daily medium replacement. From days 6–10, anteriorization was induced with differentiation basal medium containing 50 ng ml⁻¹ EGF, 50 ng ml⁻¹ bFGF, 3 μM SB431542 and 10 ng ml⁻¹ Noggin. Lung progenitors capable of generating type II alveolar epithelial cells were produced from days 10–17 in medium containing 50 ng ml⁻¹ BMP2, 50 ng ml⁻¹ FGF10, 50 ng ml⁻¹ BMP4, 50 ng ml⁻¹ bFGF and 50 ng ml⁻¹ WNT3A. Differentiation was confirmed by measuring ACE2 and SLC34A2 expression by RT–qPCR.
NLS- and NES-containing Cas13d-NCS³² was cloned into a PP7-containing backbone with the PP7 motif in the 3′ UTR. A PP7 motif was also attached 3′ to a crRNA targeting the SARS-CoV-2 3′ UTR sequence GUCAUCCAAUUUGAUGGCACCUG. Lung progenitors were seeded at 2 × 10⁴ cells per well in Geltrex-coated 96-well plates and differentiated for 7 days. Differentiated lung cells were treated with 40 μl concentrated STV-C8 carrying Cas13d-NCS and either SARS-CoV-2-targeting or non-targeting crRNA. After 24 h, cells were infected with SARS-CoV-2-GFP at a multiplicity of infection of 10. Viral replication was monitored for 72 h using an Incucyte S3 live-cell imaging system.

Assessment of STV-C8 inactivation by human blood components

Peripheral blood mononuclear cells (PBMCs) were isolated by diluting blood two- to fourfold with PBS. Thirty-five millilitres of diluted blood were carefully layered onto 15 ml Ficoll with a density of 1.077 g ml⁻¹ in a Falcon tube and centrifuged without braking at 400g for 30 min at 20 °C. The PBMC layer at the interface was transferred to a fresh tube, filled with PBS and centrifuged at 300g for 10 min at 20 °C. Pellets were resuspended in PBS and counted by Trypan blue exclusion. For long-term storage, PBMCs were frozen at 1 × 10⁷ cells ml⁻¹ in FBS containing 20% dimethyl sulfoxide.
Blood for serum collection was obtained in EDTA-free tubes, gently inverted and allowed to clot at 4 °C for 3–4 h. Samples were centrifuged at 2,500g for 10 min at room temperature, and the clear serum layer was transferred to sterile tubes and stored at −80 °C. STV-C8(N-split-Luc) particles were produced in 24-well plates for 48 h and concentrated with Lenti-X (Takara Bio). Thirty microlitres of concentrated STV-C8 were mixed with 30 µl of serum diluted 1:10 and 30 µl resuspended PBMCs, approximately 3.0 × 10⁵ cells, or PBS. Mixtures were incubated at 37 °C for 60 min. Fifty microlitres of each mixture were then transferred to split-Luc reporter cells in 96-well plates. N-split-Luc RNA expression was measured the following day with the ONE-GloEX Luciferase Assay (Promega).

STV-C8 storage-stability testing

STV-C8(N-split-Luc) particles were produced in 6-well plates for 2 days, concentrated with Lenti-X (Takara Bio) and stored for 7 days at either 4 °C or −80 °C. Fifty microlitres of stored material were added to split-Luc reporter cells. N-split-Luc RNA expression was measured the following day using the ONE-GloEX Luciferase Assay (Promega).

STV-C8 delivery of OpenCRISPR-1

The OpenCRISPR-1 coding sequence was synthesized by Twist Bioscience and cloned into a CAG-promoter expression plasmid. The sequence was fused to two NLS signals and one NES signal, and a PP7 aptamer was inserted into the 3′ UTR. An sgRNA containing a PP7 aptamer in the stem–loop and a spacer targeting the stop codon in eTLR cells was also cloned (GCUCCCACAACGAAGACUGAC; cells provided by the Institute of Synthetic Biomedicine, Helmholtz Munich)³⁰. STV-C8 particles containing OpenCRISPR-1 or Cas9 and sgRNA were produced in 6-well plates for 3 days and concentrated using Lenti-X. Twenty microlitres of concentrated particles were added to eTLR cells in 96-well plates. Cells were imaged 3 days later with an EVOS imaging system.

Whole-body mouse biodistribution of STV-C8-mediated EGFP expression

STV-C8(EGFP) and empty STV-C8 vesicles were produced in coated 10-cm dishes for 3 days and concentrated by ultracentrifugation. Fifty microlitres of concentrated STV-C8(EGFP) were injected intravenously into 4-week-old female BALB/c wild-type mice (Charles River Laboratories) under approved institutional animal-care procedures. Mice were euthanized 24 or 72 h after injection and perfused intracardially with heparinized PBS containing 10 U ml⁻¹ heparin followed by 4% paraformaldehyde. Skin was removed, and bodies were fixed overnight in 4% paraformaldehyde at 4 °C.
Whole-body vDISCO staining and clearing were performed as described previously³³. Briefly, samples underwent decolourization with 25% CUBIC reagent in PBS, decalcification with 10% (w/v) EDTA in PBS, signal enhancement with anti-GFP nanobodies (Chromotek, anti-GFP-AF647), dehydration with tetrahydrofuran, delipidation with dichloromethane and refractive-index matching in benzyl alcohol/benzyl benzoate. Light-sheet imaging was performed with a Blaze system (LaVision BioTec) providing 4-µm axial resolution. Whole-body scans used a ×4 objective (Olympus XFLUOR ×4, 0.28 numerical aperture, 10-mm working distance). High-magnification tile scans used 22% overlap and an 80% reduced light-sheet width. The z-step was 6 µm, with exposure times of 40 ms for the 488-nm background channel and 60 ms for the 640-nm signal channel. Fiji was used to stitch raw TIFF files into complete planes. Planes were merged into three-dimensional files with Imaris Converter and visualized in Imaris³³.

Immune response and liver-toxicity analysis after systemic STV-C8 administration

Female C57BL/6J mice aged 11 weeks received intravenous PBS as an untreated control or 50 µl STV-C8. Mice were euthanized by carbon dioxide inhalation on days 1 and 3 after injection according to institutional animal-care procedures approved by the Ethical Review Board of the Government of Upper Bavaria (ROB-2532.Vet_02-23-143). Blood was collected by cardiac puncture, and liver tissue was excised immediately. Serum alanine aminotransferase was measured with an ALAT (GPT) FS (IFCC mod.) assay using a respons910 random-access clinical chemistry analyser (DiaSys Diagnostic Systems). Liver tissue was homogenized, and total RNA was extracted from the aqueous phase with a NucleoSpin RNA Mini Kit (Macherey-Nagel). Four hundred nanograms of RNA were reverse transcribed with a PrimeScript RT Reagent Kit (TaKaRa). RT–qPCR was performed using PowerUp SYBR Green Master Mix (Applied Biosystems) on a QuantStudio 3 Real-Time PCR System.

EGFP mRNA delivery to differentiated mouse myotubes

C2C12 mouse myoblasts, provided by the Institute of Developmental Genetics, Helmholtz Munich, were differentiated into myotubes in DMEM supplemented with 2% horse serum. Differentiation was confirmed by the formation of multinucleated cells. Mature myotubes were treated with STV-C8(EGFP) for 2 days, and EGFP expression was measured by flow cytometry.

In vivo exon 51 deletion in porcine muscle

All large-animal procedures were approved and monitored by the Bavarian local authority (ROB-55.2-2532.Vet_02-19-39). STV-C8(Cas9/sgRNA) particles were produced in coated 10-cm dishes and concentrated by ultracentrifugation. For proof-of-concept testing, a 3-month-old, 25-kg wild-type German Landrace pig was sedated by intramuscular ketamine and azaperone. Fentanyl was administered intravenously through a 20-G ear-vein catheter for analgesia. The right hind limb was shaved and disinfected, and 1 ml concentrated STV-C8 was injected 1.75 cm deep into the right M. biceps femoris using a 1-ml Luer-lock syringe fitted with a 22-G safety needle. The animal was monitored until recovery and assessed twice daily for 3 days for infection, inflammation or other adverse reactions. After 3 days, the pig was sedated according to the approved protocol and euthanized by intravenous pentobarbital. Tissue samples were collected from areas surrounding the injection site, the uninjected contralateral leg and the M. latissimus dorsi.
Genomic DNA was extracted with the Monarch Genomic DNA Purification Kit (NEB). PCR was performed with Platinum SuperFi II Master Mix (Thermo Fisher Scientific) using primers CCCATGACATTTACCCTATTATTATCCC and GCTAATGTTCATTTTAAAAAGGAATCTGTC. Deletion efficiency was assessed by agarose-gel band-intensity comparison. Bands corresponding to the 2-kb wild-type product and 1-kb deletion product were excised and verified by Sanger sequencing (Microsynth). PCR products were also analysed by Oxford Nanopore sequencing (Eurofins Genomics), and deletion frequency was quantified with Geneious Prime 2025.1.2 (Dotmatics).

Immune-response analysis after local STV-C8 injection in pig muscle

RNA was extracted from muscle samples by phenol–chloroform extraction. Expression of inflammation-associated genes after STV-C8 treatment was measured by RT–qPCR using TBP as the endogenous control. Samples from STV-C8-injected muscle represented the treatment group. Non-injected muscle from the same animal and muscle from untreated animals served as negative controls, while samples from rejected tissue were used as positive indicators of a strong inflammatory response.

hiPS cell differentiation into myotubes and treatment of patient-derived muscle cells

Control and DMDΔ52 hiPS cells were differentiated into skeletal muscle using SKM-KIT (Amsbio). For STV-C8 treatment, wild-type and patient-derived myoblasts were seeded at 40,000 cells cm⁻² on collagen I-coated plates (5 µg cm⁻²; Sigma-Aldrich, 122-20) in skeletal muscle myoblast medium (SKM02). At confluence, cells were switched to skeletal muscle myotube medium (SKM03). After 4 days, differentiated myotubes were treated with STV-C8 or left untreated. Live-cell images were acquired at 24 and 72 h after transduction. At 72 h, cells were fixed and stained with DAPI. For STV-C8(Cas9) experiments, RNA was collected 96 h after transduction. RNA was reverse transcribed with the Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific), and the DMD transcript was amplified with primers binding exons 49 and 54. RPS18 was amplified as an RNA-quality control. Treated and untreated cDNA products were analysed by agarose-gel electrophoresis, and splicing-event integrity was confirmed by nanopore sequencing of the PCR products.

Statistical analysis and experimental reproducibility

Numerical data were statistically analysed and plotted with GraphPad Prism. Unless otherwise noted, immunofluorescence and gel-electrophoresis images are representative of at least three independent experiments. Sample sizes were selected based on prior experience and anticipated experimental variability, with at least biological duplicates unless otherwise specified. Animals were randomly assigned to experimental groups after enrolment. Blinding was not used for most in vitro and molecular experiments because outcomes were based on objective quantitative measurements, predefined criteria or standardized analysis pipelines. Samples from all experimental groups were processed in parallel under identical conditions. Automated image analysis and computational processing were used where applicable to reduce potential bias.

Research reporting summary

Additional information about the experimental design and reporting requirements is available in the Nature Portfolio Reporting Summary linked to this article.