Example 3
We tested the ability of mouse embryonic stem cells to tolerate s4U metabolic RNA-labeling after 12 h or 24 h at varying s4U concentrations (FIG. 6A). As reported previously, high concentrations of s4U compromised cell viability with an EC50 of 3.1 mM or 380 μM after 12 h or 24 h labeling, respectively (FIG. 6A). Hence, we employed labeling conditions of 100 μM s4U, which did not severely affect cell viability. Under these conditions, we detected a steady increase in s4U-incorporation in total RNA preparation 3 h, 6 h, 12 h, and 24 h post labeling, as well as a steady decrease 3 h, 6 h, 12 h, and 24 h after uridine chase (FIG. 6B). As expected, the incorporation follows a single exponential kinetics, with a maximum average incorporation of 1.78% s4U, corresponding to one s4U incorporation in every 56 uridines in total RNA (FIG. 6C). These experiments establish s4U-labeling conditions in mES cells, which can be employed to measure RNA biogenesis and turnover rates under unperturbed conditions.
To test the ability of the method to uncover s4U incorporation events in high throughput sequencing datasets we generated mRNA 3′ end libraries (employing Lexogen's QuantSeq, 3′ mRNA-sequencing library preparation kit) using total RNA prepared from cultured cells following s4U-metabolic RNA labeling for 24 h (FIG. 7) (Moll et al., supra). Quant-seq 3′ mRNA-Seq Library Prep Kit generates Illumina-compatible libraries of the sequences close to the 3′end of the polyadenylated RNA, as exemplified for the gene Trim28 (FIG. 8A). In contrast to other mRNA-sequencing protocols, only one fragment per transcript is generated and therefore no normalization of reads to gene length is needed. This results in accurate gene expression values with high strand-specificity.
Furthermore, sequencing-ready libraries can be generated within only 4.5 h, with ˜2 h hands-on time. When combined with the invention, Quant-seq facilitates the accurate determination of mutation rates across transcript-specific regions because libraries exhibit a low degree of sequence-heterogeneity. Indeed, upon generating libraries of U-modified RNA through the Quant-seq protocol from total RNA of mES cells 24 h after s4U metabolic labeling we observed a strong accumulation of T>C conversions when compared to libraries prepared from total RNA of unlabeled mES cells (FIG. 8B). In order to confirm this observation transcriptome-wide, we aligned reads to annotated 3′ UTRs and inspected the occurrence of any given mutation per UTR (FIG. 9). In the absence of s4U metabolic labeling, we observed a median mutation rate of 0.1% or less for any given mutation, a rate that is consistent with Illumina-reported sequencing error rates. After 24 h of s4U metabolic labeling, we observed a statistically significant (p<10−4, Mann-Whitney test), 25-fold increase in T>C mutation rates, while all other mutations rates remained below expected sequencing error rates (FIG. 9). More specifically, we measured a median s4U-incorporation of 2.56% after 24 h labeling, corresponding to one s4U incorporation in every 39 uridines. (Note, that median incorporation frequency for mRNA are higher than estimated by HPLC in total RNA [FIG. 6C], most certainly because stable non-coding RNA species, such as rRNA, are strongly overrepresented in total RNA.) These analyses confirm that the new method uncovers s4U-incorporation events in mRNA following s4U-metabolic RNA labeling in cultured cells.
We expect the same incorporation results of other modified nucleotides, such as s6G or 5-ethynyluridine, as reported previously (Eidinoff et al., Science. 129, 1550-1551 (1959); Jao et al. PNAS 105, 15779-15784 (2008); Melvin et al. Eur. J. Biochem. 92, 373-379 (1978); Woodford et al. Anal. Biochem. 171, 166-172 (1988)).
