Recognition and binding of mismatch repair proteins at an oncogenic hot spot

Abstract

Background: The current investigation was undertaken to determine key steps differentiating G:T and G:A repair at the H-ras oncogenic hot spot within the nuclear environment because of the large difference in repair efficiency of these two mismatches.

Results: Electrophoretic mobility shift (gel shift) experiments demonstrate that DNA containing mismatched bases are recognized and bound equally efficiently by hMutSalpha in both MMR proficient and MMR deficient (hMLH1-/-) nuclear extracts. Competition experiments demonstrate that while hMutSalpha predictably binds the G:T mismatch to a much greater extent than G:A, hMutSalpha demonstrates a surprisingly equal ratio of competitive inhibition for both G:T and G:A mismatch binding reactions at the H-ras hot spot of mutation. Further, mismatch repair assays reveal almost 2-fold higher efficiency of overall G:A repair (5'-nick directed correct MMR to G:C and incorrect repair to T:A), as compared to G:T overall repair. Conversely, correct MMR of G:T --> G:C is significantly higher (96%) than that of G:A --> G:C (60%).

Conclusion: Combined, these results suggest that initiation of correct MMR requires the contribution of two separate steps; initial recognition by hMutSalpha followed by subsequent binding. The 'avidity' of the binding step determines the extent of MMR pathway activation, or the activation of a different cellular pathway. Thus, initial recognition by hMutSalpha in combination with subsequent decreased binding to the G:A mismatch (as compared to G:T) may contribute to the observed increased frequency of incorrect repair of G:A, resulting in the predominant GGC --> GTC (Gly --> Val) ras-activating mutation found in a high percentage of human tumors.

Figure 1

Specific interaction of hMSH6 and hMSH2 with site-specific mismatched DNA at H-ras codon 12. Protein-DNA binding reactions and gel shifts were performed using nuclear extracts from HCT 116 + Ch. 3 and equal cpm of [³²P]-G:T-oligo (69-mer) in the presence of 100X molar excess of cold (unlabeled) homoduplex (G:C). BSA, goat-, or rabbit-nonspecific IgG (lanes 1, 2, 4 respectively), goat anti-hMSH6 (lane 3), and rabbit anti-hMSH2 (lane 5) were also included in the binding reactions as indicated. The lower gel shift band in each lane is due to biotin end-labeling of the probe.

Figure 2

Comparison of binding avidity of G:T-containing oligomer by MMR proficient (HCT 116 + Ch. 3) and MMR deficient (HCT 116; hMLH1^-/-) nuclear extracts. Equal aliquots of nuclear extracts from HCT 116 + Ch. 3 (lanes 1–3) or HCT 116 (lanes 4–5) were incubated with equal cpm of [³²P]-G:T-oligo (69-mer) and 50X molar excess cold (unlabeled) oligo as indicated.

Figure 3

Binding avidity of hMutSα to G:T versus G:A at H-ras codon 12 (a) Equal cpm of [³²P]-G:T-oligo (69 mer) were incubated with equal concentrations of nuclear extract (HCT116 + Ch. 3) in the presence of increasing concentrations of unlabeled (cold) G:T- or G:A-containing oligo, and 100X cold G:C oligo. (b) Equal aliquots of nuclear extract were incubated for 2.5 hours (versus 30 minutes) with equal cpm of each [³²P]-G:T- or [³²P]-G:A-oligo alone (lanes 1 and 3), or also with respective 100X molar excess cold G:T or G:A oligo (lanes 2 and 4). Bar graphs are densitometry results of corresponding radioactive band intensities.

Figure 4

Repair model of hMutSα MMR. Model of hMutSα initial recognition (equal) of G:T and G:A, followed by repair (differential) of DNA containing a G:T or a G:A mismatch at the H-ras codon 12 location. hMutSα recognizes and forms an initial recognition complex (IRC) with both mismatches equally, and then binds more strongly to G:T, perhaps by undergoing an additional conformational step to the ultimate recognition complex (URC), which does not occur with G:A [40]. This results in more accurate repair of G:T, but more frequent total repair of G:A. See text for further discussion.