Metabolic labeling and recovery of nascent RNA to accurately quantify mRNA stability

Abstract

Changes in the rate of mRNA decay are closely coordinated with transcriptional changes and together these events have profound effects on gene expression during development and disease. Traditional approaches to assess mRNA decay have relied on inhibition of transcription, which can alter mRNA decay rates and confound interpretation. More recently, metabolic labeling combined with chemical modification and fractionation of labeled RNAs has allowed the isolation of nascent transcripts and the subsequent calculation of mRNA decay rates. This approach has been widely adopted for measuring mRNA half-lives on a global scale, but has proven challenging to use for analysis of single genes. We present a series of normalization and quality assurance steps to be used in combination with 4-thiouridine pulse labeling of cultured eukaryotic cells. Importantly, we demonstrate how the relative amount of 4sU-labeled nascent RNA influences accurate quantification. The approach described facilitates reproducible measurement of individual mRNA half-lives using 4-thiouridine and could be adapted for use with other nucleoside analogs.

Keywords: 4-Thiouridine labeling; Digital PCR; RNA biotinylation; mRNA decay; mRNA stability.

Figure 1. Typical work-flow for the assessment of individual mRNA half-lives using 4sU metabolic labeling

Sub-confluent cell cultures are incubated with 4sU for a pre-determined time. Following incubation, total RNA is extracted and DNAse treated. A positive control (4sU labeled) RNA is spiked into the total RNA sample for downstream normalization and overall quality control. The RNA is conjugated to biotin followed by clean up and precipitation. Fractionation of total RNA to separate nascent and pre-existing transcripts is performed using magnetic streptavidin beads and nascent RNA is recovered by elution with DTT. The nascent and total RNA fractions are analyzed by reverse transcription and digital PCR. Abundance of the mRNA of interest and the positive control mRNA are determined in the total and nascent fraction and used in the equation shown to calculate half-life.

Figure 2. The experimental system must be at steady state in order to accurately estimate mRNA half-life

A key assumption of this approach is that the experimental system is at steady state. In other words, the mRNA abundance, transcription rate, and rate of decay do not vary during the labeling period. The bar charts show the amounts of pre-existing and newly synthesized mRNA during a 2 hour labeling period for hypothetical transcripts with different expression profiles. The times above each bar indicate the half-life that would be calculated based on the ratio of nascent to total RNA. Panel A shows a theoretical transcript with a 60 minute half-life, while B depicts a transcript with a much longer half-life of around 13 hr. The total amount of RNA does not change during the course of the experiment in either case and the same half-life would be calculated whether the RNA was harvested at 60 min or 120 min. However, the half-life calculation would be significantly more reliable for A than B because in B the nascent:total ratio is ≥0.9 and therefore outside of the acceptable range (see Figure 3). Panels C and D depict the same short-lived theoretical transcript as in A but the half-life (C) or transcription rate (D) change between 60 and 120 min. This leads to changes in the overall abundance of the transcript and a different assessment of the mRNA half-life at 120 versus 60 min.

Figure 3. 4sU labeling periods should be adjusted depending on half-life of the RNA of interest

Panel A depicts the dependence of the calculated half-life on nascent/total ratios determined for a 2 hour labeling period. When the N/T ratio falls within the red shaded areas (below 0.15 or above 0.9), background binding, pipetting inaccuracies, partial loss or decay of RNA samples, and other errors have a disproportionate impact on the half-life calculation. When the ratio falls in the center of the curve (green shaded areas) normal experimental variation will have less impact on the half-life. The chart shown in Panel B depicts the range of half-lives that can be reliably determined with labeling periods ranging from 0.5 to 8 hours. It is recommended that the 4sU labeling period be adjusted so that the expected half-life of the mRNA of interest falls close to the center of the range shown.

Figure 4. Tapestation assessment of the quality of the RNA fractions used in half-life calculations

Equal volumes of total (T), nascent (N) and pre-existing (P) RNA from iPSCs treated with siRNAs targeting GFP or PCBP2 were interrogated by Tapestation. RNA integrity number equivalents (RIN_e) are shown. A RIN value ≥8 is acceptable for downstream abundance measurements [39]. Concentration of each fraction was determined using a Qubit with the Qubit RNA Broad Range Assay Kit.

Figure 5. Knockdown of PCBP2 results in the stabilization of the GPR56 transcript

iPSCs were transfected with siRNAs targeting GFP or PCBP2 as described in Supplementary Data. 4-thiouridine metabolic labeling was used to determine the half-life of GPR56 mRNA. Error bars are SEM derived from 4 independent replicates.