Materials and Methods

3.1. Materials

CleanCap eGFP mRNA (996 nt) was purchased from Trilink Biotechnologies (CA, USA), the mRNA includes a 120 nt poly adenosine (poly(A)) tail. This transcript was chosen because of its use in previous stability studies (Barros et al., 2025).

The lipids used for formulation of LNPs came from MedChemExpress (NJ, USA) and included (4 hydroxybutyl)(azanediyl)bis(hexane-6,1-diyl)bis(2-hexyldecanoate) (ALC-0315), cholesterol, 1,2 distearoyl-sn-glycero-3-phosphocholine (DSPC), and 2-[(poly(ethylene glycol))-2000]-N,N ditetradecylacetamide (ALC-0159).

Nuclease-Free Water from Invitrogen (MA, USA) and tris(hydroxymethyl)aminomethane buffer from Thermo Fisher Scientific (MA, USA) were used as solvents for the mRNA thermal stability studies. Sodium citrate, citric acid, and ethanol were purchased from Thermo Fisher Scientific (MA, USA). Induro® Reverse Transcriptase, 5x Induro® RT Reaction Buffer, NEBNext® Quick Ligation Reaction Buffer, T4 DNA Ligase 2M U/ml, and Murine RNase Inhibitor were purchased from New England Biolabs (MA, USA).

Agencourt RNAClean XP beads purchased from Beckman Coulter Life Sciences (CA, USA). All of these reagents are part of the additional materials required for RNA sequencing, as per the protocol provided by Oxford Nanopore Technologies (SQK-004). 

3.2. Instrumentation 

For Direct RNA sequencing, the GridION Mk1 Sequencer with FLO-MIN004RA flow cells and the Direct RNA Sequencing Kit, SQK-RNA004, from Oxford Nanopore Technologies (Oxford, UK) were used.

Automated capillary electrophoresis was performed on the Agilent Tapestation 4150 with the RNA ScreenTape, RNA ScreenTape Ladder, and RNA ScreenTape Sample Buffer from Agilent (CA, USA).

Lipid nanoparticles were prepared using microfluidic mixing with the Ignite™ NanoAssemblr™ NxGen™ nanoparticle formulation system from Precision Nanosystems (Vancouver, Canada). The size and polydispersity index of the LNPs were measured using dynamic light scattering with a Malvern Zetasizer Ultra from Malvern Panalytical Ltd. (Malvern, U.K.). Ribogreen assay fluorescence was measured on the BioTek Synergy H1 fluorometer (Agilent, CA, USA) with an excitation wavelength of 485 nm and emission wavelength of 528 nm. 

3.3. Lipid Nanoparticle preparation and initial characterisation 

The Precision Nanosystems Ignite system was used for manufacturing lipid nanoparticles, according to the protocol described by McKenzie et al. (2023). Briefly, each lipid was dissolved in ethanol to a concentration of 10 mg/mL, eGFP mRNA was diluted to a final concentration of 136 μg/mL in 50 mM citrate buffer, pH 4.1. The four lipids were then mixed to produce a final molar ratio percentage of 50:1.5:10:38.5 (ALC-0315: ALC-0159: DSPC: cholesterol). 

The N/P ratio used was 6, aqueous: organic ratio was set to 3 and the total flow rate (TFR) 20 mL/min. The formulation was buffer exchanged to 10 mM Tris pH 7.5 using the Pierce Protein Concentrators with MWCO of 10 kDa. Size of the lipid nanoparticles was measured using the Malvern Red Ultra system via Dynamic Light Scattering (DLS) equipped with a polystyrene cuvette at 25°C in triplicate. Polydispersity index was also measured during the same sampling.

The encapsulation efficiency measurement was determined using the Ribogreen assay. Briefly, RNA standard (100 μg/mL) was diluted in either TE buffer, or in TE buffer/2% Triton to produce a seven point calibration curve from 1 μg/mL to 20 ng/mL. LNP samples were diluted in 1:20 ratio in either TE buffer (control samples) or TE/2% Triton (lysed samples) and incubated at 37°C for ten minutes. The Ribogreen solution (100 μL) was then added to both samples and controls and incubated for 10 minutes at room temperature. Fluorescence was measured on the BioTek Synergy H1 fluorometer (Agilent, CA, USA) with an excitation wavelength of 485 nm and emission wavelength of 528 nm. The percent encapsulation efficiency was measured by comparing the fluorescence of the samples in TE buffer and in TE/Triton X-100 buffer using the equation below (Equation 1) from Barros et al. (2025). 

where FD is the fluorescence of an LNP sample exposed to 2% (v/v) Triton X-100, which fully disrupts the LNPs, and F0 is the fluorescence of an LNP sample in TE buffer only. 

3.4. mRNA extraction and integrity testing 

mRNA from LNP samples was extracted via ammonium acetate precipitation, using the protocol by Malburet et al. (2024). Briefly, each LNP sample was mixed with 900 ul 60 mM ammonium acetate in isopropanol. The sample was mixed and incubated at -70°C for 30 minutes and centrifuged for 10 minutes at 14,000g. This was followed by a wash step with pure ice-cold isopropanol and another centrifugation step. The pellet was then dried in vacuo and reconstituted in 16 μL nuclease-free water. 

3.5. Experimental design

Thermal stability studies were carried out to simulate accelerated degradation conditions. eGFP mRNA was diluted in nuclease free water to a final concentration of 40 μg/mL. Stability studies were performed at 25°C, 35°C and 50°C in order to investigate the effects of accelerated degradation conditions on mRNA stability. For each temperature, there were seven timepoints T0 (fresh) to T6, which varied depending on the temperature (Table 1). At every timepoint, three vials (triplicate sample, n=3) were removed from incubators and were stored at -40°C until analysis.

LNP-encapsulated eGFP mRNA samples were diluted to a final concentration of 36.6 μg/mL in Tris buffer pH 7.5. Each aliquot contained 4 μg of eGFP mRNA. The LNPs were incubated at 35°C for 33 days, collecting samples at 5 timepoints throughout the incubation period. These timepoints and their corresponding time in weeks are listed in Table 2.

3.6. Polyadenylation of mRNA

Polyadenylation was performed using the NEB PolyA polymerisation kit, according to instructions by Oxford Nanopore Technologies. Briefly, naked mRNA was dried in vacuo and resuspended in 15 μL of NFW, extracted mRNA was not altered further. The mRNA was mixed with 1 μL of PolyA polymerase, 1 μL of ATP and 2 μL of reaction buffer were added. The mixture was briefly vortexed and spun, followed by 90-second-incubation at 37°C. The reaction was quenched by the addition of 5 μL 50 mM RNAse free EDTA. The mRNA was then purified using Agencourt RNAClean XP beads. 45μL of beads were added, incubated on a rotary shaker for 5 minutes, washed twice with 70% ethanol and quickly dried. The RNA was eluted in 12 μL NFW. The final concentration was assessed on UV Nanophotometer from Implen (Munich, Germany). 

3.7. Analytical Methods 

3.7.1. Capillary Electrophoresis

mRNA integrity was assessed using an Agilent TapeStation 4150 with the standard RNA ScreenTape kit from Agilent was used. Briefly, 1 μL of sample was mixed with 5 μL of sample buffer. The mixture was vortexed, spun and heated to induce denaturation. The denatured mixture was then cooled down on ice and loaded onto the instrument. For experiments with naked mRNA 1 μL of sample was added from the corresponding vial, while for extracted mRNA, 1 μL of the 16 μL was used for CE analysis and the rest was saved for polyadenylation. 

3.7.2. Direct RNA Sequencing with Oxford Nanopore

The libraries for Direct Sequencing were prepared as per the protocol provided by the Oxford Nanopore Technologies (SQK-RNA-004). The main deviation from the standard protocol is RNA was not quantified before analysis. Instead, 8 μL of sample were processed according to the protocol, which involves a reverse transcription reaction and an enzymatic digestion. The prepared libraries were loaded onto primed flow cells, and each sample was run for 24 hours with no real time basecalling.

Due to the limited number of available ONS flow cells only certain samples were chosen for Oxford Nanopore Sequencing. The specific replicate selected for analysis was selected based on the measured concentration from the Tapestation readings. These samples are listed below. 

3.7.3. Sequence Data Analysis 

Following Direct RNA sequencing using the GridION sequencer, the raw sequencing data is stored in the MinKNOW software in .pod5 (compressed FAST5) file format. The raw data is a measurement of the changes in ionic current as the RNA passes through the pore. Post-sequencing basecalling was done within the MinKNOW desktop application, converting the raw data into sequence data in .fastq file format.

The MinKNOW High accuracy (HAC) basecalling model, which specifies a Phred Q-score cutoff of >9 for reads to be put into the pass folder, was chosen for basecalling. This basecaller is based on Dorado, ONT’s proprietary basecalling model. By specifying a Q-score of 9, the probability of an incorrect basecall occurring is 12.6% (Ewing and Green, 1998).

3.7.4. Data Handling 

Following basecalling in the MinKNOW software, further data manipulation was performed using command line tools on a Windows 11 laptop using the Windows Subsystem for Linux (WSL2). The environment was based on Ubuntu 24.04.1 LTS running the Linux kernel 6.6.87.2 (x86_64).

Using bash commands and seqtk v1.5 (Li, 2014; https://github.com/lh3/seqtk), the .fastq files from each basecalled sequencing run were converted and concatenated into a single .fasta file for each sample.

The alignment software Minimap2 (Li, 2018; https://github.com/lh3/minimap2) was used for aligning the reads contained in the concatenated .fasta files to the reference gene eGFP.fasta. This generated aligned .sam files for each sample which were subsequently converted to .bam files using samtools (Danecek et al., 2021). Samtools were also used to sort and index the .bam files. A diagram showing the file types used is shown below (Figure 9). 

The mRNA size distribution graphs were generated from the concatenated .fasta files in R using the following R packages: biostrings, ggplot2, dplyr, tidyr, purrr, glue, and plotly.

Depth coverage .tsv files were generated from sorted .bam files using the bedtools suite (Quinlan and Hall, 2010). Depth coverage graphs were generated in R from the .tsv files using the following packages: dplyr, purrr, readr, ggplot2, and stringr.

RNAfold Web Server from ViennaRNA was used to generate a model of the secondary structure of the full-length mRNA strand (Lorenz et al., 2011).