Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Aug 18;112(33):E4571-80.
doi: 10.1073/pnas.1507709112. Epub 2015 Aug 4.

Intrinsic Mutagenic Properties of 5-chlorocytosine: A Mechanistic Connection Between Chronic Inflammation and Cancer

Affiliations
Free PMC article

Intrinsic Mutagenic Properties of 5-chlorocytosine: A Mechanistic Connection Between Chronic Inflammation and Cancer

Bogdan I Fedeles et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

During chronic inflammation, neutrophil-secreted hypochlorous acid can damage nearby cells inducing the genomic accumulation of 5-chlorocytosine (5ClC), a known inflammation biomarker. Although 5ClC has been shown to promote epigenetic changes, it has been unknown heretofore if 5ClC directly perpetrates a mutagenic outcome within the cell. The present work shows that 5ClC is intrinsically mutagenic, both in vitro and, at a level of a single molecule per cell, in vivo. Using biochemical and genetic approaches, we have quantified the mutagenic and toxic properties of 5ClC, showing that this lesion caused C→T transitions at frequencies ranging from 3-9% depending on the polymerase traversing the lesion. X-ray crystallographic studies provided a molecular basis for the mutagenicity of 5ClC; a snapshot of human polymerase β replicating across a primed 5ClC-containing template uncovered 5ClC engaged in a nascent base pair with an incoming dATP analog. Accommodation of the chlorine substituent in the template major groove enabled a unique interaction between 5ClC and the incoming dATP, which would facilitate mutagenic lesion bypass. The type of mutation induced by 5ClC, the C→T transition, has been previously shown to occur in substantial amounts both in tissues under inflammatory stress and in the genomes of many inflammation-associated cancers. In fact, many sequence-specific mutational signatures uncovered in sequenced cancer genomes feature C→T mutations. Therefore, the mutagenic ability of 5ClC documented in the present study may constitute a direct functional link between chronic inflammation and the genetic changes that enable and promote malignant transformation.

Keywords: 5-chloro-deoxycytidine; carcinogenesis; hypochlorite; inflammatory bowel disease; myeloperoxidase.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
A model of the 5ClC biomarker as a functional link between chronic inflammation and the genetic changes observed in tumors. During inflammation, activated neutrophil granulocytes produce the enzyme myeloperoxidase (MPO); other inflammatory processes produce hydrogen peroxide, among other reactive oxygen species. Using hydrogen peroxide and chloride, MPO catalyzes the formation of HOCl. Exposure of the surrounding cells to HOCl can generate 5ClC in their genomic DNA, a type of DNA damage occurring either by direct chlorination of cytosines or by the incorporation of damaged nucleotides from the pool. The present study showcases the mutagenic ability of 5ClC, which can directly induce C→T mutations. Additionally, 5ClC is known to induce epigenetic changes that could lead to the inactivation of tumor-suppressor genes such as MLH1 by hypermethylation, in turn possibly leading to additional mutagenic outcomes such as small indels and microsatellite instability. Together, the biological effects of 5ClC generate a characteristic mutational signature (the relative distribution of all possible point mutations and indels over all possible sequence contexts), which could account for certain mutational signatures found in malignancies associated with chronic inflammation. The mutational signature shown is a cartoon representation of signature 6 from Alexandrov et al. (56), highlighting the preponderance of C→T transitions.
Fig. 2.
Fig. 2.
The biological consequences of 5ClC in vivo. (A) General strategy for characterizing the in vivo properties of 5ClC. Oligodeoxynucleotides containing 5ClC, the negative controls cytosine (C) or 5-methylcytosine (m5C), or the positive control 3-methylcytosine (m3C) were ligated into M13 genomes, which were then introduced into E. coli cells of different genetic backgrounds and allowed to replicate. DNA from the resulting viral progeny was analyzed to determine the replication efficiency (a measure of how easily the lesion is bypassed) and replication fidelity (the ability of the lesion to induce mutations). A detailed graphical overview of the methods is shown in Fig. S1. (B) In vivo bypass efficiency of 5ClC and controls. The colors correspond to the three E. coli strains used, as indicated. Each graphed value corresponds to the mean of three independent biological replicates; error bars indicate one SD. *P < 0.05. Data are also tabulated in Table S1. (C) The mutagenicity of 5ClC in vivo. Each graph shows the base composition at the lesion site in the progeny phage coming from the constructed genomes containing the indicated input base at the lesion site. Each graphed value corresponds to the mean of three independent biological replicates; error bars indicate one SD. Data are also tabulated in Table S2.
Fig. S1.
Fig. S1.
Outline of the in vivo experimental strategy. 16mer oligonucleotides containing 5ClC or controls at a defined site were used to construct M13 ssDNA viral genomes. The lesion-containing genomes were mixed with a “+3” competitor genome in a fixed ratio and were transformed into several strains of E. coli cells by electroporation. The resulting phage progeny were collected and reamplified. DNA was extracted from the phage progeny, and a region of interest was PCR amplified with assay-specific primers. Restriction endonucleases and 32P end-labeling were used to radiolabel an 18mer piece containing the lesion site at its 5′ end. The CRAB assay evaluated the ratio between the 18mer band (coming from the lesion-containing genome) and the 21mer band (given by the “+3” competitor genome) and normalized it to the equivalent ratio of the unmodified dC control to obtain the relative bypass percentage. The lower the bypass percentage, the stronger is the replication-blocking ability of the lesion and therefore its toxic effect on the cell. In the REAP assay, the 18mer oligonucleotide was isolated using PAGE and was digested with P1 nuclease to release the radioactive monophosphate nucleotide from its 5′ end. TLC separation and phosphorimagery then provided the identity and amount of the base at the lesion site.
Fig. S2.
Fig. S2.
Oligonucleotide purity analysis. ESI Q-TOF mass spectrometry traces of the 16mer oligodeoxynucleotides of the sequence 5′-GAAGACCTXGGCGTCC-3′ (in which X = 5ClC) after one round (1×) and two rounds (2×) of HPLC purification described in Materials and Methods. Insets show the parent peaks, highlighting their characteristic isotope distribution envelopes.
Fig. S3.
Fig. S3.
HPLC separation of 16mer oligonucleotides that differ by only one mass unit. Oligodeoxynucleotides of the sequence 5′-GAAGACCTXGGCGTCC-3′, in which X is either deoxyuridine (dU) or deoxycytidine (dC), were separated on an Acclaim Polar Advantage (Thermo Scientific) column using 10 mM ammonium acetate and acetonitrile as solvents. The gradient used was 2–12% acetonitrile over 100 min. The peaks then were analyzed on an ESI Q-TOF mass spectrometer (Agilent). The isotope distribution envelope and the theoretical mass (for the −4-charged species) of each oligonucleotide are shown.
Fig. S4.
Fig. S4.
Purity analysis of the 5ClC-containing genome. (A) Lesion integrity assay. Scaffolds containing G or A opposite the lesion site and restriction enzymes (with the restriction sites indicated) were used to isolate a 32P-radiolabeled monophosphate of the base X (base at the lesion site, i.e., 5ClC) and evaluate its identity and purity on TLC (Cellulose-PEI). (B) The result of the lesion integrity assay applied to an M13 genome containing 5ClC. Two mixtures of deoxynucleotide monophosphate standards, S1 (5ClU, 5ClC, G) and S2 (C, U, T, G, A) were loaded on the TLC plate alongside three independent replicates for each scaffold. The area encircled by the red dotted line indicates the location where any putative 5ClU contaminant would appear. Quantification revealed that the constructed genomes contained >99% 5ClC at the lesion site.
Fig. 3.
Fig. 3.
The mutagenesis of 5ClC in vitro. (A) Strategy for analyzing the mutagenesis of 5ClC in vitro. The modified M13 genomes, used as templates, were incubated with the REAP PCR primers (blue and orange) and PfuTurbo polymerase. The first incubation extended only the blue primer, generating a double-stranded product, which was then PCR amplified with the aid of the orange primer. The identity and amount of base N is determined by the REAP assay. (B) In vitro mutagenesis of 5ClC and controls. The graphed values are averages of three replicates, with error bars representing one SD. Data are also tabulated in Table S3. (C) Schematic of the iREAP strategy. The M13 genome containing a lesion (5ClC) or a control base (cytosine or m5C) at a defined site was used as a template for a primer extension reaction. The primer (iPrimer) contained a unique 5′ overhang that did not anneal to the M13 genome. The primer extension was carried out with a chosen test polymerase at 37 °C. Subsequently, the extended product was purified to remove the polymerase and unextended iPrimer, and a PCR was performed so that only the newly synthesized strand was amplified. The PCR product was then analyzed using the REAP assay to identify the type and amount of base (denoted N) present at the lesion site. The results of this assay with eight different test polymerases are summarized in Table 1. (D) A control experiment demonstrating the specificity of the primers used in the iREAP assay. The PCR products obtained from control iREAP experiments 1–5 with a C genome have been separated on a 2% agarose gel and stained with ethidium bromide. In each experiment, the iPrimer, test polymerase (Klenow fragment), the iREAP PCR primers (or REAP PCR primers) were added (+) or left out (−), as indicated.
Fig. 4.
Fig. 4.
Structure of pol β binary and nonmutagenic ternary complex with 5ClC-containing oligonucleotide. Fo-Fc omit maps (green) are contoured at 3σ. Helix N of pol β, primer terminus (O3′), 5Cl of 5ClC, and key active-site residues are indicated. (A) The binary pol β complex shown in the open conformation. (B) A different viewing angle of the structure highlights the anti-geometry of the 5ClC base and the distances (in Ångstroms) between the 5Cl atom and the oxygen atoms of the phosphate backbone. (C) The pol β ternary complex with an incoming dGMP(CH2)PP base-pairing with 5ClC. Mg2+ ions are shown as red spheres in the catalytic and nucleotide-binding sites. The distance (in Ångstroms) between O3′ and Pα is shown. (D) The 5ClC and dGMP(CH2)PP are shown in a stick representation (purple), and the protein is shown in surface representation (gray). The distances correspond to the Watson–Crick hydrogen bonding between 5ClC and the dGTP analog. The DNA major groove edge is indicated. For reference the dC:dGTP (PDB ID code 4UB4) ternary complex is superimposed (light blue).
Fig. 5.
Fig. 5.
Structures of the pol β closed ternary insertion complex during mutagenic 5ClC lesion bypass. Fo-Fc omit maps (green) are contoured at 3σ. Helix N, primer terminus (O3′), 5Cl of 5ClC, and key active-site residues are indicated. (A) The pol β mismatch ternary complex with an incoming dAMP(CH2)PP opposite 5ClC with putative hydrogen-bonding interaction distances (in Ångstroms) shown. Mn2+ ions are shown as purple spheres in the catalytic and nucleotide-binding sites. Water molecules are shown in blue. (B) The view in A with the protein (gray) and DNA (yellow) shown in surface representation. Incoming dAMP(CH2)PP is shown in stick representation. (C) The nonmutagenic ternary complex with an incoming dGMP(CH2)PP base-pairing with 5ClC is shown with the protein (gray) and DNA (purple) in surface representation. The Mg2+ ions are shown in red, and the incoming dGMP(CH2)PP is shown in stick representation. (D) The anti-5ClC and dAMP(CH2)PP are shown in stick representation (yellow), and the protein is shown in surface representation (gray). The DNA major groove edge is indicated. For reference, the mismatch syn-dC:dATP (PDB ID code 3C2L) ternary complex is superimposed (green). The C5 position of the nondamaged cytosine is indicated.

Similar articles

See all similar articles

Cited by 19 articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback