Cellular Repair of Synthetic Analogs of Oxidative DNA Damage Reveals a Key Structure-Activity Relationship of the Cancer-Associated MUTYH DNA Repair Glycosylase

Abstract

The base excision repair glycosylase MUTYH prevents mutations associated with the oxidatively damaged base, 8-oxo-7,8-dihydroguanine (OG), by removing undamaged misincorporated adenines from OG:A mispairs. Defects in OG:A repair in individuals with inherited MUTYH variants are correlated with the colorectal cancer predisposition syndrome known as MUTYH-associated polyposis (MAP). Herein, we reveal key structural features of OG required for efficient repair by human MUTYH using structure-activity relationships (SAR). We developed a GFP-based plasmid reporter assay to define SAR with synthetically generated OG analogs in human cell lines. Cellular repair results were compared with kinetic parameters measured by adenine glycosylase assays in vitro. Our results show substrates lacking the 2-amino group of OG, such as 8OI:A (8OI = 8-oxoinosine), are not repaired in cells, despite being excellent substrates in in vitro adenine glycosylase assays, new evidence that the search and detection steps are critical factors in cellular MUTYH repair functionality. Surprisingly, modification of the O8/N7H of OG, which is the distinguishing feature of OG relative to G, was tolerated in both MUTYH-mediated cellular repair and in vitro adenine glycosylase activity. The lack of sensitivity to alterations at the O8/N7H in the SAR of MUTYH substrates is distinct from previous work with bacterial MutY, indicating that the human enzyme is much less stringent in its lesion verification. Our results imply that the human protein relies almost exclusively on detection of the unique major groove position of the 2-amino group of OG within OG_syn:A_anti mispairs to select contextually incorrect adenines for excision and thereby thwart mutagenesis. These results predict that MUTYH variants that exhibit deficiencies in OG:A detection will be severely compromised in a cellular setting. Moreover, the reliance of MUTYH on the interaction with the OG 2-amino group suggests that disrupting this interaction with small molecules may provide a strategy to develop potent and selective MUTYH inhibitors.

Figure 1

Base excision repair pathway. Base excision repair (BER) glycosylases OGG1/MutM and MUTYH/MutY initiate the repair of OG:C and OG:A mispairs, respectively, that arise in DNA due to reactive oxygen species (ROS) and inaccurate replication. The baseless site is further processed by downstream BER enzymes to eventually restore a G:C base pair at the site of oxidative damage. Failure to capture the OG:A mismatch prior to replication leads to G:C to T:A transversion mutations.

Figure 2

MUTYH lesion-specific plasmid reporter. A) Plasmid map design of the OG:A-containing plasmid reporter which contains the dsRed gene as the transfection control followed by a P2A ribosome-skipping peptide. The OG:A mispair (position labeled as 8) is incorporated upstream of the GFP gene. The plasmid contains two nicking sites (blue) for removal in order to insert the OG- or OG-analog-containing oligonucleotide as well as a uniquely placed restriction enzyme site (orange). If no repair occurs, then transcription yields a stop codon in the mRNA that results in only dsRed expression during translation. If repair by MUTYH occurs to replace the A with C in the DNA template strand, then a glycine (Gly) codon in the mRNA is produced, which allows for translation read through and subsequent expression of GFP. B) Representative scheme generating the OG:A-containing or OG analog:A-containing GFP plasmid reporter. The plasmid is first nicked with the nickase Nb.Bpu10i. Ligation of the OG-containing oligonucleotide intentionally disrupts the AfeI restriction enzyme site, allowing for the parent plasmid to be digested and then degraded with T5 exonuclease. C) Representative gel of the plasmid products formed after each step to generate the OG:A-containing plasmid reporter. Note that the digested and OG:A plasmid lane refers to post-AfeI digestion. Subsequent T5 exonuclease treatment and purification provides the OG:A plasmid. Additional details and controls are shown in Figure S1.

Figure 3

Visualizing OG:A repair by MUTYH in human cells. A) Fluorescence microscopy imaging of OG:A-mediated repair in WT versus MUTYH^–/– HEK293FT cells at 10× magnification. B) Representative flow cytometry plots of compensated red (Y axis) versus compensated green (X axis) fluorescence in MUTYH^–/– HEK293FT compared to WT HEK293FT cell lines to quantify MUTYH-mediated OG:A repair versus the transfection control (pUC19, dsRed-/GFP-), negative control (pR/GFP OFF, dsRed+/GFP−), and positive control (pR/GFP ON, dsRed+/GFP+) plasmids. The percentage in each quadrant represents the percentage of cells within that population, where the lower left is untransfected, the upper left is dsRed + (transfected), the upper right is dsRed+GFP+ (transfected, repair positive), and the lower right would be cells that are only GFP+ (none detected, as expected).

Figure 4

Repair of A across from various OG analogs by MUTYH. A) Structure of OG:A mismatch. B) Structures of OG analogs with changes to the 2 position marked in orange, the N7 position marked in green, and the O8 position marked in gray. C) Normalized percent repair by human MUTYH as measured by flow cytometry, where repair in WT cells is marked in blue and MUTYH^–/– HEK293FT cells are marked in black. Percent repair is normalized by the pR/GFP ON (dsRed+/GFP+) positive control plasmid (eq 1). The error reported is the standard deviation from three trials. Data are reported in Tables S5 and S6. Note that pR/GFP OFF = a negative control plasmid containing T:A at the lesion site, and the small amount of green fluorescence observed is due to spectral overlap of the fluorophores, providing base levels for the detection of repair. OG = 8-oxo-7,8-dihydroguanine, 8SG = 8-thioguanine, 8OI = 8-oxoinosine, 7MOG = 7-methyl-8-oxo-7,8-dihydroguanine, G = guanine, and 8SI = 8-thioinosine. χ² test: †p = 0.0137 for 8OI:A; *p < 0.0001 for all other conditions.

Scheme 1. Minimal Kinetics Scheme for MUTYH

Figure 5

OG:A repair in mismatch repair-deficient cell lines. Percent repair by MUTYH of A across from various OG analogs in WT HEK293FT (blue) and mismatch repair-deficient HCT116 (gray) cells normalized by the pR/GFP ON (dsRed+/GFP+) plasmid. χ² test: *p < 0.0001 for all conditions.

Figure 6

OG-specific recognition by bacterial MutY and mouse Muyth. X-ray structure of OG recogntion site with Geobacillus stearothermophilus (Gs) MutY with A) OG DNA versus with B) G DNA (Gs MutY; PDB: 6U7T for OG and 6Q0C for G). C) Mouse Mutyh recognition sphere with OG (PDB: 7EF8) and D) structure of OG highlighting which features are essential for the initiation of human MUTYH repair.