No abstract

Single-strand-selective monofunctional uracil-DNA glycosylase 1 (SMUG1) is a base excision repair enzyme that removes uracil and oxidised pyrimidines from DNA. We show that SMUG1 interacts with the pseudouridine synthase Dyskerin (DKC1) and colocalizes with DKC1 in nucleoli and Cajal bodies. As DKC1 functions in RNA processing, we tested whether SMUG1 excised modified bases in RNA and demonstrated that SMUG1 has activity on single-stranded RNA containing 5-hydroxymethyldeoxyuridine, but not pseudouridine, the nucleoside resulting from isomerization of uridine by DKC1. Moreover, SMUG1 associates with the 47S rRNA precursor processed by DKC1, and depletion of SMUG1 leads to a reduction in the levels of mature rRNA accompanied by an increase in polyadenylated rRNA. Depletion of SMUG1, and, in particular, the combined loss of SMUG1 and DKC1, leads to accumulation of 5-hydroxymethyluridine in rRNA. In conclusion, SMUG1 is a DKC1 interaction partner that contributes to rRNA quality control, partly by regulating 5-hydroxymethyluridine levels.

Both a DNA lesion and an intermediate for antibody maturation, uracil is primarily processed by base excision repair (BER), either initiated by uracil-DNA glycosylase (UNG) or by single-strand selective monofunctional uracil DNA glycosylase (SMUG1). The relative in vivo contributions of each glycosylase remain elusive. To assess the impact of SMUG1 deficiency, we measured uracil and 5-hydroxymethyluracil, another SMUG1 substrate, in Smug1 −/− mice. We found that 5-hydroxymethyluracil accumulated in Smug1 −/− tissues and correlated with 5-hydroxymethylcytosine levels. The highest increase was found in brain, which contained about 26-fold higher genomic 5-hydroxymethyluracil levels than the wild type. Smug1 −/− mice did not accumulate uracil in their genome and Ung −/− mice showed slightly elevated uracil levels. Contrastingly, Ung −/− Smug1 −/− mice showed a synergistic increase in uracil levels with up to 25-fold higher uracil levels than wild type. Whole genome sequencing of UNG/SMUG1-deficient tumours revealed that combined UNG and SMUG1 deficiency leads to the accumulation of mutations, primarily C to T transitions within CpG sequences. This unexpected sequence bias suggests that CpG dinucleotides are intrinsically more mutation prone. In conclusion, we showed that SMUG1 efficiently prevent genomic uracil accumulation, even in the presence of UNG, and identified mutational signatures associated with combined UNG and SMUG1 deficiency.