High resolution biosensor to test the capping level and integrity of mRNAs

Abstract 5′ Cap structures are ubiquitous on eukaryotic mRNAs, essential for post-transcriptional processing, translation initiation and stability. Here we describe a biosensor designed to detect the presence of cap structures on mRNAs that is also sensitive to mRNA degradation, so uncapped or degraded mRNAs can be detected in a single step. The biosensor is based on a chimeric protein that combines the recognition and transduction roles in a single molecule. The main feature of this sensor is its simplicity, enabling semi-quantitative analyses of capping levels with minimal instrumentation. The biosensor was demonstrated to detect the capping level on several in vitro transcribed mRNAs. Its sensitivity and dynamic range remained constant with RNAs ranging in size from 250 nt to approximately 2700 nt and the biosensor was able to detect variations in the capping level in increments of at least 20%, with a limit of detection of 2.4 pmol. Remarkably, it also can be applied to more complex analytes, such mRNA vaccines and mRNAs transcribed in vivo. This biosensor is an innovative example of a technology able to detect analytically challenging structures such as mRNA caps. It could find application in a variety of scenarios, from quality analysis of mRNA-based products such as vaccines to optimization of in vitro capping reactions.


Plasmid construction for protein expression and purification
Plasmid pET28a+ (Novagen, UK) was used for protein expression. The synthetic gene encoding B4E was PCR amplified with primers 1 and 2 (Supplementary Table 1), the pET28a+ backbone was amplified with the primers 3 and 4 and both parts assembled via Gibson DNA assembly (40), to generate the plasmid pET28a-B4E-vi. A second version of the expression plasmid, pET28a-B4E-v.ii, was prepared by deleting some additional residues from N-and C-termini of B4E-v-i. Here primers 5 and 6 were used to amplify B4E from the original gene synthesis vector and primers 7 and 8 to amplify the pET28a+ plasmid backbone. Both fragments were joined using Gibson assembly. For the construction of the plasmid pET28a-βLac the ampR gene was amplified with primers 9 and 10 from the pET-15b plasmid and the pET28a+ backbone was amplified with primers 11 and 12. The fragments were joined using Gibson DNA assembly.

SDS-PAGE analysis of B4E
B4E was expressed and purified as described in the main manuscript. Figure S1 shows, the SDS-PAGE analysis of the various fractions obtained from his-tag purification. Elutions 2, 3 and 4 (lanes 8-10) were combined and buffer exchanged using a centrifugal concentrator with a 30 kDa MWCO (Vivaspin). 200 µL of purified B4E at 663 µg/mL were obtained (as measured by absorbance at 280 nm). insoluble fraction; lanes 4: flow through (fraction not bound to the resin); lanes 5 and 6: washes; lanes 7 to 11: elutions 1 to 5. The expected size of the B4E protein is 56.2 kDa.

Total RNA extraction from CHO cells for biosensor control experiments
The total RNA fromCHO cells was extracted from 10 6 CHO cells collected on day 4 of culture produced as indicated in the materials and methods section. The RNA was purified from 0.5 mL of cell lysate with a RiboPure TM RNA purification kit (Thermo), with the addition of the Phenol:Chloroform:IAA step (omitting the use of zirconia beads) and continuing with the purification as indicated by the manufacturer.

Capping efficiency of IVT mRNAs
To estimate the percentage of capped RNAs incorporated either co-transcriptionally for cRNA1, cRNA2 and cRNA3 or post-transcriptionally for mRNAv, 5 µg of RNA was first digested with an RNA 5'  Figure S3). The percentage of capped RNA was calculated by densitometry by comparison with a control reaction lacking the ribonuclease using the image analysis software Fiji 2.

Biosensor optimization
Several alternative assay conditions were studied to decrease non-specific binding between B4E and RNA. The pH was increased in order to decrease positive charges on the surface of the protein.
However, increasing pH was not effective, since even though the non-specific binding was decreased, the B4E binding to capped RNA was also drastically reduced ( Figure S8). The optimal pH for the interaction between eIF4e and m 7 GTP is approximately 7.25, and is influenced by zwitterionic nature of the m 7 GTP group, which binds in its cationic form to eIF4e (1).
Alternatively, altering the reducing conditions in the buffer was a more effective way to reduce nonspecific binding. Figure S9 shows the response of the biosensor with and without addition of DTT to buffer B. Adding DTT considerably reduces non-specific binding, likely by preventing intermolecular B4E oligomerisation via disulfide bond formation (2) and electrostatic interaction between the positively charged region of the oxidised cysteines of the B4E and the negatively charged mRNA, pdT25 oligo and beads. This observation is also in agreement with the fact that eIF4E is naturally present in the cell nucleus and cytoplasm, which are reducing environments. Finally, further reduction of non-specific binding to an almost complete reduction of the background signal was achieved by the removal of nonfunctional amino acids on the C and N termini of the B4E ( Figure S10).

Biosensor m 7 GTP inhibition analysis
A biosensor assay using a buffer with a low level of stringency to permit non-specific binding further confirmed the specificity of the interaction between B4E and the mRNA cap during the actual biosensor assay. The presence of the m 7 GTP is expected only to affect the B4E-cap 0 interaction, since some of the eIF4e peptides will bind the methylated nucleotide instead of capped RNA, while the controls should remain unaffected. This was confirmed experimentally, as shown in Figure S11. The presence of m 7 GTP leads to a decrease in the signal of the biosensor when cRNA1 is used, while the signal due to non-specific binding observed for RNA1 or the pdT25 oligo-functionalised beads was unchanged. This further confirms the existence of a specific interaction between the cap binding region of B4E and the cap 0 of mRNA while immobilised on the beads used for the biosensor assay. In addition, this experiment suggests that the background signal originates from non-specific interactions with other regions of the protein.    Table S4. Amounts of RNAs bound to the beads during the first incubation step of the biosensor assay. RNA concentrations of the samples before and after binding to the beads were measured on a BioDrop DUO+ microspectrophotometer. The percentage of the bound mRNA that is then bound to a E4B molecule was estimated using a calibration curve of the activity of the free protein. The calibration curve is shown in Figure S4.      Figure S8. Biosensor assay for RNA1 analysing the effect of the pH. The response of the sensor using buffer B at pH 7.4 (solid line) and pH 8 (dashed line). The background signal was almost eliminated at pH 8; However, the increase in pH also heavily reduces the specific interaction between B4E and cRNA1. Figure S9. Biosensor assay analysing the effect of the addition of the reducing agent DTT to buffer B. DTT decreases the non-specific binding between B4E and uncapped RNA and beads. The signal for capped RNA was also reduced, likely due to screening out the non-specific interaction in this case as well.