The Importance of the Fifth Nucleotide in DNA: Uracil

Abstract

Uracil is a ribonucleotide found in both DNA and RNA, with the main difference between the two being the presence of thymine in DNA and uracil in RNA. Although thymine and uracil are similar in function and can form the same base pairs with adenine, the presence of uracil in DNA can affect DNA stability and modulate cell-specific functions. Without repair mechanisms to remove uracil from DNA, cytosine deamination can occur, resulting in gene drift that is not tolerable in organisms. While the deamination of cytosine in DNA signals damage, a corresponding deamination in RNA would yield normal RNA constituents. To correct this, uracil DNA glycosylases detect and remove uracil bases from uracil-containing DNA, but not natural thymine-containing DNA. The mechanisms of uracil incorporation into DNA, its roles in DNA, cellular mechanisms to detect and remove uracil, and the clinical utility of uracil in DNA will be discussed in this chapter.

1. Introduction

Conservation of DNA integrity is important during replication to ensure that daughter cells have accurately replicated DNA to promote genetic continuity. The accumulation of aberrations within the DNA sequence, if left unrepaired, can lead to genetic drift and subsequent detrimental effects following subsequent rounds of replication. Given that the rate of DNA replication occurs at a frequency of 500 nucleotides/minute/replication fork, only a small number of errors (~1 nucleotide per 1 × 10⁹ nucleotides) arise during the replication process [1]. Evolution has enabled the cell to develop proof-reading mechanisms that minimise the potential disruption and preservation of its genetic code [2, 3]; however, errors remain and with time contribute to increased genomic instability and an altered metabolic landscape as required for a sufficient supply of macromolecules, including nucleotides, to drive proliferation [4, 5, 6].

Uracil is one of the most frequently occurring error bases in DNA, occurring through mutagen hydrolytic deamination of cytosine to uracil or through substantial uracil DNA misincorporation, and the cell has, therefore, evolved different strategies to target and repair this type of DNA damage. In the absence of uracil-DNA repair, relatively fast cytosine deamination and the toxicity of the resulting uracil will result in a gene drift which is likely not tolerated by an organism.

2. Evolution of the genome: why thymine was key

The genome of all organisms on earth today is coded by 4 nucleobases, which are the two purines, adenine (A) and guanine (G), and the 2 pyrimidines, cytosine (C) and thymine (T). The nucleobases are commonly conjugated to (deoxy)ribose, which are termed nucleosides, and they can then be further conjugated to phosphate groups, giving them the name nucleotides. As deoxynucleotide triphosphates (dNTPs), the nucleobases A, T, G and C are commonly incorporated into DNA, held in a sequence via covalent attachments to the DNA backbone (made up of covalently attached deoxyribose and phosphate in a chain), making a DNA strand. A DNA strand is paired with another DNA strand (forming the classic double-helix structure of DNA) which are complementary to one another via hydrogen bonds between nucleobases within adjacent strands of the DNA helix. In ‘normal’ DNA, A:T and G:C always pair with one another. Uracil (U) is another nucleobase that is mostly found in RNA, which is synthesised via nucleotide triphosphates (NTPs). RNA can form similar structures to DNA, except that A:U pair together instead of A:T and that the RNA backbone incorporates ribose instead of deoxyribose. RNA is also able to encode genetic information (mRNA) but is more diverse and carries out functions similar to proteins (tRNA and ribozymes). One might ask what the need is for these two similar systems for carrying genetic information and why are they different? To answer these questions, it helps to explore the evolution of the genome.

Life on earth is thought to have originated ~3.8 billion years ago from what is termed the ‘primordial soup’ (or prebiotic soup). In this prebiotic world the ‘RNA world’ hypothesis states that the first complex organic molecules to form were RNA based. To support this hypothesis, analysis of carbonaceous meteorites that have fallen to earth have found to contain a range of carbon-molecules, including U, A and G (but not C), which is thought to represent the composition of a very young earth (reviewed here [7]). Additionally, a formamide-based scenario has purposed that formamide (available in the prebiotic earth [8]) could be the starting point for generating all the RNA-components, under conditions that are thought to be present in the prebiotic earth, with U being generated in good yield [9, 10, 11, 12]. This leads to the belief that the hereditary genetic information might have been RNA. It is thought that, eventually, there was transition to a DNA-based hereditary system since the deoxyribose-containing DNA backbone is much more stable than ribose-containing RNA backbone [13, 14, 15] and RNA replication seems to be far more error prone than DNA replication [16]. In turn, this allowed the evolution of larger, more complex genomes and, therefore, complex multicellular organisms to form.

Potentially, at some point RNA and other molecules were concentrated in a membrane-like structure (making certain catalytic reactions feasible), forming the first RNA cells with metabolism. RNA can catalyse reactions (ribozymes), encode genetic information, transport amino acids (tRNA) and catalyse peptide-bond formation (ribosomes). The idea that RNA was first to carry out these critical functions of the cell is based on ribosomal RNA being extensively involved in peptide-bond formation, suggesting that proteins potentially became essential later in evolution [17]. To allow for the transition from an RNA to DNA cell, the evolution of a mechanism to convert NTPs (containing ribose) to dNTPs (containing deoxyribose) must have occurred. Ribonucleotide reductases (RNRs) frequently catalyses the conversion of nucleotide diphosphates (NDP)/NTP in eukaryotic/prokaryotic cells into dNDP/dNTP, respectively, and are thought to have a common ancestor [18, 19]. Interestingly, it is thought that the first DNA cell would have incorporated U (instead of T) into its genome. This is backed up by the fact that RNR can directly convert ATP, UTP, GTP and CTP into its corresponding dNTPs; however, dTTP needs extra steps involving deoxyuridine monophosphate (dUMP) conversion to dTMP, via thymidylate synthase (TS or TYMS), and two phosphorylation steps by kinases to produce dTTP [20]. Due to this convoluted route to produce dTTP, yet dUTP is synthesised in a simpler manner, it would make sense in an evolutionary context that the initial DNA cell first incorporated U into its DNA, which was then later replaced by T.

One might ask why the need of T-based DNA when U-based DNA performs the same task and is energetically easier to make? C deamination produces U, which happens at a relatively fast rate. A deaminated C will produced a G:U mismatch and led to mutated DNA during replication. In U-based DNA, a U produced via C deamination and U that is normally incorporated in DNA are chemically identical and, therefore, a G:U mismatch would be difficult to identify as damaged DNA, in the context of primitive cells without sophisticated DNA repair mechanisms. To overcome this issue, cells evolved to incorporate T instead of U into their DNA, making U produced from C deamination completely foreign instead. This meant that T-based DNA produced more stable genomes, which is especially important for evolving and maintaining large complex genomes found in, for example, multicellular organisms.

For T-based DNA to be viable, the cell evolved three key enzymes: dUTP nucleotidohydrolase (dUTPase), thymidylate synthase (TS) and uracil-DNA glycosylases (UDGs). After dUTP synthesis via RNR, dUTP is potently dephosphorylated to dUMP by dUTPase. In a two-fold mechanism, this reduces dUTP levels, reducing U misincorporation into the DNA, and produces the TS substrate (dUMP, as discussed in the previous paragraph), leading to an increase in dTTP. The relative levels of the dUTP:dTTP pools determines the rate of U misincorporation into DNA, due to DNA polymerase having difficulty distinguishing between dTTP and dUTP, which are identical molecules except for a single methyl group. In fact, DNA polymerases readily incorporate dUTP as well as dTTP based on their representative concentration and nucleotide availability [21]. If U is misincorporated or C deamination occurs, a UDG removes U, creating an abasic site where no nucleobase is present in a strand of DNA, and a primitive form of DNA damage repair could have corrected it. With the evolution of these 3 proteins, cells were able to detect C deamination (which is always mutagenic) rapidly, repair it and reduce U misincorporation as well.

3. Biosynthesis of NTPs

Nucleotide synthesis begins with U(ridine monophosphate) and I(nosine monophosphate) [22, 23, 24]. The success of DNA replication is dependent upon a number of factors including the availability of balanced nucleotide pools and the cellular ability to recognise misincorporated bases in DNA and initiate their removal and subsequent repair. In the quiescent state, the genomic content of the cell is at its’ minimum. Following stimulation to enter the cell cycle and initiate S-phase, DNA biosynthesis is up-regulated and therefore access to the required deoxynucleotide triphosphates (dNTPs) is vital for errorless replication of the genetic material [25]. The enzyme ribonucleotide reductase (RNR) is the initiating factor that induces the reduction of ribonucleoside diphosphates into their respective deoxyribonucleoside diphosphates (dNDPs) [26, 27]. Additional phosphorylation of the nucleosides occurs through the action of nucleoside diphosphate kinase (NDPK), converting the dNDP’s to dNTP. However, the synthesis of dTTP requires additional steps within the biosynthesis pathway (Figure 1; for an in-depth review, see ref. [28]).

Figure 1.

Uracil biosynthesis and incorporation into genomic RNA and DNA. A simplified schematic showing how uracil (U, purple) and thymidine (T, red) are incorporated into DNA. Normally, uracil is metabolised to UDP then UTP, allowing it to be incorporated into RNA. UDP can also be processed by ribonucleotide reductase (RNR, composed of RRM1/2/2B subunits) to dUDP, which is further metabolised to dUTP. Cells have evolved a system where dUTP levels are kept low by dUTPase converting it into dUMP, which is a substrate for dTTP production. This maintains a high cellular dTTP:dUTP ratio, ensuring minimal uracil misincorporation into DNA since DNA polymerase has similar affinities for dTTP and dUTP. Arrows indicate the direction of a metabolic reaction by an enzyme, double ended arrows indicate a single enzyme can perform a metabolic reaction in both directions. Abbreviations: U = uracil, rU = uridine, dU = deoxyuridine, (d)UMP = (deoxy)uridine monophosphate, (d)UDP = (deoxy)uridine diphosphate, (d)UTP = (deoxy)uridine triphosphate, dT = deoxythymidine, (d)TMP = (deoxy)thymidine monophosphate, (d)TDP = (deoxy)thymidine diphosphate, (d)TTP = (deoxy)thymidine triphosphate, CH2-FH4 = 5,10-methylenetetrahydrofolate, FH2 = dihydrofolate, DNA Polβ = DNA polymerase β, RNR = ribonucleotide reductase, and dUTPase = dUTP nucleotidohydrolase.

4. Uracil

Uracil is a base analogue of thymine and is found in both DNA and RNA (Figure 1). As a pyrimidine nucleotide it has an amino acid group at the 2′ position and a carboxy group at the 5′ position. U and T are equivalent in their information storage, transmission and their base pairing with adenine.

In DNA, U is present in pyrimidine dimers, tri- and tetradiphosphate conjugates, DNA repair intermediates and DNA damage-based U incorporation. It is classified as an antimetabolite, which means it can block enzymes that participate in cellular metabolism, where its incorporation into DNA has several biological and pharmacological effects.

Uracil’s presence in DNA has been identified to be an important key intermediate in both adaptive and innate immunity (see Section 4.2). In adaptive immunity, U is introduced into DNA by the enzyme activation-induced deaminase (AID), where in B lymphocytes its presence modulates the antibody binding site or initiation of Ig isotype switching [29, 30]. In innate immunity, U in DNA serves as an intermediate in the restriction of viral pathogens with the assistance of the family of APOBEC polynucleotide deaminases, where there is a C:U hypermutation of the viral genome leading to its degradation [31].

Beyond these positive immunity roles of U in DNA, the DNA incorporation of U can negatively impact the cellular functionality of the genetic information storage system, destabilize the Watson-Crick DNA helical structure, inhibit DNA replication, cell growth and development, highlighting the importance of cellular mechanisms to detect and remove inappropriate uracil from DNA.

4.1 Modifications to uracil

Of all the bases, U is the most modified in RNA [32, 33], while in DNA, U has been reported to undergo modifications including methylation [32], hydroxymethylation [34] and oxidation [35]. Methylation, the addition of a methyl group to the nitrogen atom in the ring structure of uracil, can occur at several sites on the base, including the N1 and N3 positions [32]. Hydroxymethylation, the addition of a hydroxymethyl group to the N3 position, is a common modification [32, 35]. Oxidation can also lead to the formation of several different uracil derivates, such as 5-hydroxymethyluracil and 5-formyluracil [35]. Other less prevalent modifications may include glutathionylation, nitration or phosphorylation [36]. These modifications can affect the stability and function of the DNA molecule and may also play a role in regulating gene expression and the development of diseases such as cancer [37, 38].

4.2 Maintenance of uracil-free DNA uracil

Hydrolytic depurination produces about 10,000 abasic sites per cell per day, while hydrolytic deamination of cytosines produces 70–200 uracil bases per day in DNA [13]. The inappropriate U is recognized by family of human UDGs, which cleaves the N-glycosidic bond and thereby generates an abasic sites in the DNA, which are themselves cytotoxic and potentially mutagenic [39]. As mentioned in Section 2, U can arise in DNA either by C deamination or by U mis-incorporation. Deamination of C to U is always mutagenic, if not corrected, as U:G mismatches always leads to C→T or G→A transitions during DNA synthesis and is, interestingly, the most frequent spontaneous mutation of cells, and often found in human tumours. On the other hand, misincorporation of U should not lead to a mutation during DNA synthesis, since U:A pairs would lead to T:A pairs; however, DNA repair mechanisms can be error-prone and, therefore, it is best for the cell to avoid U DNA misincorporation from the onset, as discussed in Section 2.

In almost all the organisms, the nucleotide pools are essential for the correct DNA replication and U is one of the most frequently occurring error bases in DNA; different strategies for “keeping free” the organism of unbalanced levels of nucleotides exists [22]. In this direction, four superfamilies of NTP (nucleoside triphosphate) pyrophosphatases including the nudix hydrolases, trimeric dUTPase, inosine triphosphate pyrophosphatases (ITPases) and all α NTP pyrophosphatases function to hydrolyze the α-β phosphodiester bond of (d)NTPs to monophosphate and pyrophosphoric acid (PPi) focussing on the non-canonical NTPs [22, 23]. The deoxycytidine triphosphatase (dCTPase) and dUTPase are the two main nucleotide hydrolases involved in the elimination of non-canonical nucleotides [22, 24].

The most important mechanisms to maintain U-free-DNA are dUTPase and UDGs [40, 41, 42].

4.2.1 dUTPase

The function of dUTPase is to hydrolyze dUTP to dUMP and pyrophosphate, providing a dUMP precursor for the dTMP synthesis, maintaining the balanced dUTP/dTTP ratio and ultimately DNA integrity [43]; this reaction facilitates the cells avoidance of dUTP DNA misincorporation by DNA polymerases during replication (Figure 1) [44].

Two distinct protein isoforms of dUTPase, one nuclear and the other mitochondrial, in human cells have been reported [45]. Nuclear dUTPase (DUT-N) and mitochondrial dUTPase (DUT-M) are encoded by two distinct mRNA species of 1.1 and 1.4 kilobases respectively, nonetheless, the dUTPase gene (DUT) encode both nuclear and mitochondrial isoforms and arise to mature form by splicing process using different exon patterns [46].

In normal cells, the expression of dUTPase varies, where fluctuations in expression are dependent upon the current state of the cell cycle. Stimulation from mitogenic signals to initiate mitosis triggers the cell to progress from the resting G₀-phase into G₁/S-phase [47]. It is during S-phase of the cell cycle when the DUT-N is increased to provide the dUMP substrate required for dTMP synthesis and subsequent deoxythymidine triphosphate (dTTP) for incorporation into newly synthesized DNA [48]. Overall, the specific activity is over 16,000 nmol of dUMP hydrolyzed per min/mg of dUTPase [49].

4.2.2 Base excision repair

There are four UDGs present in mammalian cells: UNG, SMUG1, TDG and MBD4 [50] that function to recognize and remove U from DNA. This family of enzymes provides redundancy which may be required for specific circumstances and highlights the importance of this repair process. UNG remains central to the repair of U:A misincorporated uracil, whereas all family members are involved in the U:G repair [29].

The process of uracil repair produces an abasic sites, which forms part of Base Excision Repair (BER) pathway in humans, a process that repairs small, non-helix distorting base lesions in the genome [51]. Briefly, a UDG detects U within DNA and ‘flips’ it out of the double-helix and cleaves the U leaving an abasic site. An AP-endonuclease (APE) then cleaves the DNA-backbone 5′ of the abasic site, creating a single-strand break [52]. During short-patch BER, DNA polymerase β (POLβ) inserts the correct nucleotide into the abasic site and has lyase activity that removes the deoxyribosephosphate (dRP) left over from the abasic site [53]. The open 3′ end (that is left over after DNA polymerase activity) can then be sealed by DNA ligase III (LIG3) and its co-factor XRCC1, removing what is a single-strand break [54, 55]. Alternatively, during long-patch BER; POLδ, POLε and PCNA inserts multiple nucleotides from the abasic site, displacing the downstream DNA (which also contains dRP at its 5′ end), creating a flap [53]. The flap endonuclease 1 (FEN1) then removes this 5′ flap of DNA [56] and DNA ligase 1 (LIG1) is able to seal the single-strand break in the DNA backbone left after this process [57]. See Figure 2 for BER schematic.

Figure 2.

Detection and removal of uracil in DNA by base excision repair pathway. A simplified schematic showing the two arms of the base excision repair (BER) pathway in the repair of uracil that is either misincorporated (dU:dA) or mismatched (dU:dG) in DNA. A uracil DNA glycosylase, UDG/SMUG1/TDG/MBD4 in humans, detects and ‘flips’ out the uracil base from the DNA and cleaves it, leaving an abasic site. AP endonuclease (APE) then nicks the DNA backbone 5′ to the abasic site, creating a single-strand DNA break. From there, BER pathway will either proceed down the short patch BER or Long patch BER depending on the type of damage, stage of the cell cycle, and cell differentiation state. In short patch BER, DNA Polβ fills in the gap of the abasic site with the correct base and is also cleaves the deoxyribosephosphate (5′ dRP) left over from the abasic site. After DNA polymerisation, a single-strand DNA break is present and is ligated by LIG3:XRCC1. In long patch BER, DNA Polδ:Polε:PCNA inserts multiple bases from the abasic site, creating a ‘flap’ of single-stranded DNA. FEN1 is able to cleave this flap and LIG1 seals the single-strand DNA break left over from the process. Abbreviations: 5′ dRP = deoxyribosephosphate, dA = deoxyadenine, dG = deoxyguanine, dU = deoxyuridine, dT = deoxythymidine, APE = Apurinc endonuclease, BER = base excision repair, DNA Polβ = DNA polymerase, FEN1 = flap endonuclease 1, LIG = DNA ligase, MBD4 = methyl-CpG-binding domain protein 4, SMUG1 = single-strand selective monofunctional uracil DNA glycosylase, TDG = thymine DNA glycosylase, UDG = uracil DNA glycosylase, and XRCC1 = X-ray repair cross-complementing protein 1.

4.3 Uracil-DNA glycosylases

4.3.1 UNG

In humans, two splice variants of Uracil-N-Glycosylase (UNG) are expressed from its gene (UNG), with both isoforms containing an identical sequence except for their N-terminal which is unique to each protein. UDG/UNG are interchangeably used to refer to this protein; however, in this text we will refer to the protein as UNG and the superfamily encompassing all the uracil-targeting DNA glycosylases will continue to be referred to as UDG. UNG1 is expressed constitutively in the mitochondria, first as a 35 kDa precursor which is then processed at the N-terminal to a 29 kDa protein, and UNG2 is a 36 kDa serine/threonine phosphoprotein located in the nucleus, which maintains its N-terminal sequence. U detection and removal is predominantly carried out by UNG, in human cells, since it has the highest activity with single-stranded (ss)DNA and is very active with double-stranded (ds)DNA compared to the other UDGs present in human cells (SMUG1, TDG and MBD4) [50, 58, 59]. Additionally, UNG has at least 10^1–3 times higher turnover than the other three human UDGs [50, 59, 60], with UNG2 being the only DNA glycosylase present in the nucleus that is able to remove U:A pairs close to passing replication forks [61, 62], where UNG is mostly located in replication foci during S-phase [59, 61, 63]. Due to the efficiency of UNG compared to the other UDGs in humans, it is thought that UNG is predominantly responsible for removing U:G mismatches [59, 60]; however, this has not been directly reported except of U:G mismatches produced by activation-induced cytosine deaminase (AID), which are critical in the adaptive immune system.

When an infection occurs in humans, our adaptive immune system will try to generate antibodies specifically for that antigen, which will allow the infection to be efficiently cleared out. B lymphocytes are responsible for generating antibodies, but to generate specific antibodies they need a mechanism to induce heterogeneity, which is achieved via somatic hypermutation (SHM) and class switch recombination (CSR) [29]. The Ig loci produces the heavy chain of an antibody and codes for the Ig variable region (produces part of heavy chain that directly binds to antigen) and the Ig constant region (determines class of antibody, which could be IgM, IgG, IgA or IgE). AID deaminates Cs in specific regions of this loci [31]. During CSR, UNG2 targets U and, with APE, generates abasic sites with single-strand breaks. Since AID generates clustered regions of U-containing DNA, this can produce double-strand breaks once UNG2 and AP endonuclease have processed enough of them. This then triggers non-homologous end joining and connects the Ig variable region to a new constant region and determines the class of the antibody. During SHM, AID introduces U into Ig variable region and UNG2 removes these producing abasic sites and error prone polymerases then introduce mutations and alter the DNA sequence. Overall, this means various B lymphocytes produce unique antibodies coded from their Ig loci. The unique antibody is exposed on the surface of the B lympocyte and if that antibody has affinity to the antigen presented to it, it will survive and produce Plasma cells, which produce the cloned antibody and then allow the adaptive immune system to target the antigen [64].

In addition to UNG’s role in the adaptive immune system, UNG is also involved in the innate immune system. Virally infected cells are exposed to proviral DNA (viral DNA that is yet to become active), which will allow the virus to propagate further by hijacking cellular functions. To counteract this, cells express APOBEC3 enzymes (another DNA cytosine deaminase) that can associate with proviral DNA, before it integrates into the cell’s genome, and proceeds to deaminate C→U [65]. This then either leads to degradation of the proviral DNA (with the help of AP endonuclease) or if the proviral DNA does integrate within the genome of the host it is hypermutated (G:C→A:T) and, therefore, non-functional. Other functions of UNG may also include removal of some oxidative products of C including alloxan, isodialuric acid and 5-hydroxyuracil; but it is unknown if this is true in vivo [66]. Additionally, UNG2 might be involved in TET-mediated demethylation of cytosine, which would reverse the epigenetic silencing of certain genes [67].

4.3.2 SMUG1

Single-strand-selective Monofunctional Uracil-DNA Glycosylase (SMUG1) was named so because it was originally thought to prefer ssDNA to dsDNA as a substrate [68]; however, it was later found to be specific for dsDNA [60]. SMUG1 is expressed as a 30 kDa protein and is evenly distributed in the nucleus, accumulates in nucleoli and is also found in the cytosol [59]; additionally, unlike the UNG gene, the SMUG1 gene is not regulated by the cell cycle [69]. Like UNG, SMUG1’s substrate specificity is greater for U:G mismatches than U:A pairs; however, the catalytic activity of SMUG1 is slower than UNG’s [59]. While SMUG1 has been thought to act as a backup to UNG in SHM and CSR in mice, it is worth noting that mice express higher levels of SMUG1 (relative to UNG) than humans do and that UNG/SMUG1 have different roles in initiating BER in mice vs. humans [60, 70]. In addition to its role as a UDG, SMUG1 is the major DNA glycosylase in removing hydroxymethyluracil (5-hmU, an epigenetic modification [71]) from DNA [59, 72, 73, 74], with UNG seemingly not having any significant involvement [75].

4.3.3 TDG

Thymine DNA Glycosylase (TDG) is a 46 kDa protein that is located in the nucleus and plasma membrane of the cell. Despite UNG, SMUG1 and TDG sharing less than 10% amino acid similarity they all have similar structures and are located on the same arm of chromosome 12; additionally, they all may have evolved from the same ancestral gene [76]. In contrast to UNG, TDG is highly expressed during G1 and G2-M phases of the cell cycle and not in S-phase, while UNG is highly expressed only in S-phase [77]. TDG is known to remove U:G and T:G mismatches (derived from deaminated 5-methylcytosine (5-mC)) from DNA, with better efficiency with U:G mismatches [78]; however, seems to overall have a very slow turnover rate. Additionally, TDG has higher affinity for U:G mismatches, but also has a very high affinity for abasic sites opposite Gs [79] and it has been reported that TDG SUMOylation helps dissociate it from abasic sites [80]. TDG is also able to remove 5-hmU, thymine glycol halogenated pyrimidines and εC (caused by lipid peroxidation) when they are paired with G. Unlike UNG and SMUG1, when TDG is knocked out of mice it is embryonically lethal [81]. Rather than associating the lethality with TDG’s U-DNA glycosylase activity, it is thought this is due to disruption of TDG’s association with promoters, transcription factors, transcriptional coactivators and DNA methyltransferase which impairs the epigenetic regulation of developmental genes [81, 82]. Furthermore, TDG has been reported to directly induce DNA demethylation by removing 5-formylC (5-fC) and 5-carboxyC (5-aC) (both derived from TET-mediated oxidation of 5-mC), which leads to BER and repair of the site a non-modified C [83, 84].

4.3.4 MBD4

Methyl-CpG Binding Domain 4 DNA glycosylase (MBD4) is a 66 kDa protein that is predominantly found in the nucleus of the cell. While UNG, SMUG1 and TDG have a similar structure to one another, MBD4 has a N-terminal methyl-binding domain (MBD) which is connected to the C-terminal glycosylase domain that, additionally, has a different structural fold to the other 3 UDGs [85]. However, similar to TDG, MBD4 is able to remove U, T, 5-hmU thymine glycol halogenated pyrimidines and εC when they are paired with G, and MBD4 is not regulated by the cell cycle, the same as SMUG1 [69]. SMUG1’s DNA glycosylase activity might predominantly occur near CpG sites since mice with MBD4 knocked out had an increase in C→T transition mutations at CpG sites (DNA methylation sites) [86, 87]. MBD4 has also been reported to have additional roles in apoptosis, transcriptional regulation and active demethylation [82].

4.4 UNG vs. SMUG1 vs. mismatch repair

While UNG seems to be the predominant UDG in removing genomic U, it does not appear to be essential in the overall process. SMUG1 knockout mice’s organs had no increase in genomic U, while UNG knockout mice’s organs had a 1.9–2.2-fold increase [75]. However, a UNG/SMUG1 double knockout a much greater increase, especially in the liver (25-fold increase), suggesting that while SMUG1 cannot fully compensate for UNG loss it is able to act as a backup [75]. Surprisingly, mice and humans deficient in UNG induce issues related to immunity like problems with CSR and inducing lymphoid hyperplasia [88, 89, 90, 91]; however, UNG/SMUG1 single and double knockout mice do not have reduced one-year survival rates [72]. This would suggest that cells can tolerate U, to a certain level at least, and that the major issues that arise from UNG loss seem to be immune-related, likely due to its importance in the adaptive immune system (as discussed in Section 4.2). Paradoxically, cell death induced by too high a levels of genomic U might be induced by the cell’s own DNA repair mechanism that induces ssDNA breaks that could lead to dsDNA breaks that are cytotoxic (similar to the mechanism of CSR), in a cancer setting at least [92]. If this is the case, then that would indicate a significant role for UNG/SMUG1 in inducing the cell death in this setting, rather than the high levels of genomic U.

While UNG/SMUG1 double knockout does not reduce one-year survival rates, a triple knockout including MutS homolog 2 (MSH2), a critical enzyme in the DNA repair mechanism termed Mismatch Repair (MMR) severely reduced survival rates when compared to mice with a single MSH2 knockout [72]. This could suggest that while UNG and SMUG1 are complementary to each other’s function in removing genomic U, MMR might act as a last resort when the two proteins are lost; however, it is worth mentioning that MSH2 loss alone significantly decreased one-year survival by itself [72]. Furthermore, UNG/SMUG1/MSH2 triple knockout mice also had an increased chance of cancer development (compared to other knockout combinations), which was mostly lymphoma (maybe due to high C deaminase activity in immune cells via AID), suggesting that these three proteins are key to maintaining genomic stability, potentially via the removal of genomic U [72].

Overall, one could hypothesis that high levels of genomic U, in the acute setting, could lead to cell death induced by the DNA repair machinery not being able repair the damage it inflicts initially for repair; however, if high levels of genomic U occur when UNG/SMUG1/MMR are not present then the cells would not initially die but over time would acquire a high amount of mutations, leading to either death or cancerous phenotypes. Though, a lot more work is needed to validate this hypothesis but visualising cell death in a UNG/SMUG1/MMR deficient cell with high levels of genomic U might reveal some answers.

5. Therapeutic-induced thymine-less death

As mentioned, U:A pairs in DNA arise from dUTP misincorporation by DNA polymerase since they use dTTP and dUTP with similar efficiencies. In fact, several studies have confirmed that DNA-cytosine deaminase or U:G mismatches constitute a major mutational DNA lesion that contributes to various diseases, including cancer development and progression [93]. To counteract uracil in DNA, the cell has evolved the two proteins TS and dUTPase (see Section 4.1), which lowers the dUTP:dTTP ratio and thereby reduces dUTP use by DNA polymerases by having more dTTP available. The first step is by dUTPase, which converts dUTP into dUMP; thus, dUTP levels are decreased and TS’s substrate (dUMP) is increased simultaneously. TS then converts dUMP into dTMP by attaching a methyl group to the C5 position of U’s aromatic ring, the methyl group being donated by 5,10-methylenetetrahydrofolate (5,10-CH₂THF, a folate derivative). Two extra kinase steps are then needed to convert dTMP into dTTP, which can be readily processed for DNA synthesis.

Originally, it was rationalized that targeting the pyrimidine biosynthesis pathway would be an effective method in treating cancer, based on several findings including the observation that U was specifically incorporated into the nucleic acid fractions of rat tumours [94]. This eventually led to the development of the fluoropyrimidine 5-Fluoruracil (5-FU) in 1957 [95], one of the best know chemotherapies for cancer treatment in modern times. 5-FU’s mechanism of action as an anti-cancer therapy seem to be complicated but what is known is that metabolites produced from it can are incorporated into DNA and RNA, which induces cellular stress and then subsequent cell death in cancer cells. Additionally, one of 5-FU’s metabolites (FdUMP) is able to inhibit and it is through this method a range of fluoropyrimidines and antifolates (which indirectly inhibit TS) have been designed and used for treatment of large range or cancer types, in which their uses still seem to be growing today [96].

Building on 5-FU’s clinical success, several additional cancer therapies such Pemetrexed, Capecitabine, Methotrexate and Raltitrexed were developed and are currently used in cancer chemotherapy regimens to inhibit TS and modify the cancer cell’s viability as a consequence of the depletion of dTTP pools, called thymine less-death. Thymine less-death results from an imbalance in dUTP/TTP levels, where dTTP is depleted and there is an increase in dUTP, misincorporation of uracil into DNA and following attempted repair results in DNA double-strand breaks (DSB) [97, 98]. To a global understanding of involved mechanism it is important to highlight that the dUMP pools needed for TMP biosynthesis depends on dCMP deamination and UDP reduction by deoxycytidylate deaminase (DCD) and RNR, respectively [39].

6. Conclusions

Uracil, a non-canonical DNA base, has been identified as one of the major bases misincorporated into DNA either through the process of cytosine deamination or through the introduction by DNA polymerase. The presence of uracil in DNA, while important for specific adaptive and innate immune functions, in other contexts threatens genetic stability and continuity, where dysregulated genomic uracil levels has been linked to various disease, including cancer. Several key enzymes have been identified to collaborate in maintaining uracil-free DNA through modulation of uracil levels within the cell by dUTPase, as well as the uracil-DNA glycosylase family (UNG, SMUG1, TDG, MBD4), who function to recognize and excise uracil from DNA, where UNG1 and UNG2 are the most competent and widely used for uracil misincorporation in DNA.

Acknowledgments

We would like to thank the staff, past and present, of the Nucleotide Metabolism Research Group at Queen’s University Belfast for their dedication to the research on uracil metabolism. In addition, we would like to thank Dr Peter Wilson and Dr Robert Ladner for their valuable insights into the biology and many valuable conversations that have helped shape the research.

Conflict of interest

The authors declare no conflict of interest.

Nomenclature

5-aC	5-carboxyC
5-fC	5-formylC
5-FU	5-Fluorouracil
5-hmU	Hydroxymethyluracil
A	Adeninie
AID	Activation-induced deaminase
APE	AP-endonuclease
APOBEC	Apolipoprotein B MRNA editing enzyme catalytic subunit
ATP	Adenosine triphosphate
BER	Base exicision repair
C	Cytosine
CSR	Class switch recombination
CTP	Cytosine triphosphate
dCTPase	Deoxycytidine triphosphatase
DNA	Deoxyribonucleic Acid
dNTPs	Deoxynucleotide triphosphosphates
dRP	Deoxyribosephosphate
DSB	Double-strand breaks
dTMP	Deoxythymidine monophosphate
dTTP	Deoxythymidine triphosphate
dUMP	Deoxyuridine monophosphate
dUTP	Deoxyuridine triphosphate
dUTPase	dUTP nucleotidehydrolase
FEN1	Flap endonuclease 1
G	Guanine
GTP	Guanosine triphosphate
ITPase	Inosine triphosphate pyrophosphatases
LIG3	DNA ligase III
MBD4	Methyl-CpG binding domain 4 DNA glycosylase
MMR	Mismatch repair
mRNA	Messenger RNA
MSH2	MutS homolog 2
NDPK	Nucleoside diphosphate kinase
NTPs	Nucleotide triphosphosphates
PCNA	Proliferating cell nuclear antigen
POLβ	DNA polymerase β
RNA	Ribonucleic Acid
RNR	Ribonucleotide reductase
SHM	Somatic hypermutation
SMUG1	Single-strand-selective monofunctional uracil-DNA glycosylase
T	Thymine
TDG	Thymine DNA glycosylase
tRNA	Transfer RNA
TS	Thymidylate synthase
U	Uracil
UDG	Uracil-DNA glycosylase
UNG	Uracil-N-Glycosylase
UTP	Uracil triphosphate