Formation of dsRNA by-products during in vitro transcription can be reduced by using low steady-state levels of UTP

Abstract

Introduction: Exogeneous messenger ribonucleic acid (mRNA) can be used as therapeutic and preventive medication. However, during the enzymatic production process, commonly called in vitro transcription, by-products occur which can reduce the therapeutic efficacy of mRNA. One such by-product is double-stranded RNA (dsRNA). We therefore sought to limit the generation of dsRNA by-products during in vitro transcription. Materials and methods: In vitro transcription was performed with a DNA template including a poly(A)-tail-encoding region, dinucleotide or trinucleotide cap analogs for cotranscriptional capping, and relevant nucleoside triphosphates. Concentrations of UTP or modified UTP (m1ΨTP) and GTP were reduced and fed over the course of the reaction. mRNA was analyzed for dsRNA contamination, yield of the reaction, RNA integrity, and capping efficiency before translational activity was assessed. Results: Limiting the steady-state level of UTP or m1ΨTP during the enzymatic reaction reduced dsRNA formation, while not affecting mRNA yield or RNA integrity. Capping efficiency was optimized with the use of a combined GTP and UTP or m1ΨTP feed, while still reducing dsRNA formation. Lower dsRNA levels led to higher protein expression from the corresponding mRNAs. Discussion: Low steady-state concentrations of UTP and GTP, fed in combination over the course of the in vitro transcription reaction, produce mRNA with high capping and low levels of dsRNA formation, resulting in high levels of protein expression. This novel approach may render laborious purification steps to remove dsRNA unnecessary.

Keywords: RNA capping; double-stranded RNA; in vitro transcription; mRNA translatability; mRNA-based therapeutics.

FIGURE 1

Increasing cap analog concentration leads to higher capping efficiency (A) and a reduction in dsRNA formation (B): RNAs produced by in vitro transcription with increasing concentrations of either a dinucleotide (Di) or a trinucleotide (Tri) cap analog using a GTP fed-batch reaction were analyzed for capping efficiency (A) and dsRNA levels (B). Representative example of several experiments with similar outcomes.

FIGURE 2

In vitro transcription reactions with a UTP feed resulted in production of mRNA with the lowest level of dsRNA. (A) J2 dotblot of in vitro transcription reactions with a dinucleotide cap analog and a DNA template encoding a poly(A)-tail at the 3′ end, in which different nucleotide triphosphates as indicated were fed over time: (B) J2 dotblot of dsRNA levels from in vitro transcription RNA from DNA templates encoding different 3′ ends either with no nucleotide triphosphate feed or with a UTP feed: Quantification of dsRNA levels as shown in (C, D). Representative example of several experiments with similar outcomes.

FIGURE 3

Using a combined GTP and UTP feed keeps capping efficiency high while still reducing dsRNA formation. Analysis of in vitro transcribed RNA with a dinucleotide cap analog fed with the indicated nucleotides. Capping efficiency is reduced using a UTP feed, but can be rescued to GTP feed levels with a combined GTP/UTP feed (A): dsRNA is reduced using UTP feed or GPT/UTP dual feed (B): Representative example of several experiments with similar outcomes.

FIGURE 4

The effects of nucleotide feeds on dsRNA and capping are independent of the type of cap analog and whether modified or unmodified UTP is used. RNAs were in vitro transcribed either in the presence of a dinucleotide or a trinucleotide cap analog and either using UTP or m1ΨTP with feeding of the indicated nucleotide triphosphates. The resulting RNAs were then analyzed for dsRNA levels (A) and capping (B). Please note the different scales for the dsRNA levels in (A).

FIGURE 5

Maximal protein expression is observed with mRNAs transcribed using a combined feed of GTP and UTP or m1ΨTP. The mRNAs as also used for the experiments in Figure 4, all coding for luciferase, were transfected into human immature dendritic cells. Upon transfection, luciferase activity was measured 6, 24, 48, and 72 h after transfection. The luciferase activity is depicted for mRNAs capped with a dinucleotide cap analog and using UTP in (A), capped with a trinucleotide cap analog and using UTP in (B), capped with a dinucleotide cap analog and using m1ΨTP in (C), and capped with a trinucleotide cap analog and using m1ΨTP in (D). In each panel, the corresponding feeds are indicated at the bottom. Representative example of several experiments with similar outcomes.