Interactions among DNA ligase I, the flap endonuclease and proliferating cell nuclear antigen in the expansion and contraction of CAG repeat tracts in yeast

Abstract

Among replication mutations that destabilize CAG repeat tracts, mutations of RAD27, encoding the flap endonuclease, and CDC9, encoding DNA ligase I, increase the incidence of repeat tract expansions to the greatest extent. Both enzymes bind to proliferating cell nuclear antigen (PCNA). To understand whether weakening their interactions leads to CAG repeat tract expansions, we have employed alleles named rad27-p and cdc9-p that have orthologous alterations in their respective PCNA interaction peptide (PIP) box. Also, we employed the PCNA allele pol30-90, which has changes within its hydrophobic pocket that interact with the PIP box. All three alleles destabilize a long CAG repeat tract and yield more tract contractions than expansions. Combining rad27-p with cdc9-p increases the expansion frequency above the sum of the numbers recorded in the individual mutants. A similar additive increase in tract expansions occurs in the rad27-p pol30-90 double mutant but not in the cdc9-p pol30-90 double mutant. The frequency of contractions rises in all three double mutants to nearly the same extent. These results suggest that PCNA mediates the entry of the flap endonuclease and DNA ligase I into the process of Okazaki fragment joining, and this ordered entry is necessary to prevent CAG repeat tract expansions.

Figure 1.

MMS sensitivity of mutants. Stationary yeast cells were exposed to MMS for the times indicated, washed, and plated to determine survival. These determinations were repeated such that each strain was examined independently in three or more experiments. The points shown in the graphs are the mean values. No variation in relative sensitivity among strains occurred during the multiple trials. (A) MMS sensitivity of DNA ligase I and flap endonuclease mutants. (B) MMS sensitivity of DNA ligase I and flap endonuclease mutations combined with rad51Δ. (C) MMS sensitivity of DNA ligase I and flap endonuclease mutations combined with the PCNA mutation pol30-90.

Figure 2.

Co-immunoprecipitation of PCNA and DNA ligase I. Extracts were prepared from wild-type and cdc9-p mutant cells in which the CDC9 copies had been appended at their C terminus with three HA epitopes. (A) HA-tagged DNA ligase I was precipitated with an antibody bound to agarose beads and the precipitate was subjected to electrophoresis. The gel was blotted and probed for PCNA with an anti-PCNA antibody. Controls included a blot probed with anti-HA antibody to control for successful precipitation of DNA ligase I and input controls to assess the amounts of DNA ligase I and PCNA in the extracts. Inputs are 0.1% of the precipitated extract in the case of the HA-tagged DNA ligase I and 0.5% in the case of PCNA. The relative amounts of PCNA precipitated from the two extracts was measured with a Fuji FLA-5000 imager using the amount of precipitated HA-tagged DNA ligase I for comparison. We repeated the precipitations four times from independent extracts and found reductions of fivefold or greater each time (11, 15, 18, and 19%). (B) PCNA was precipitated with an antibody and the precipitate was subjected to electrophoresis. The gel was blotted and probed with an anti-HA antibody for DNA ligase I. Input and recovery controls were as described above except that inputs were 10× as great. These precipitations were repeated five times independently and found to yield a twofold reduction or greater in each trial (29, 32, 33, 38, and 39%). (C) The PCNA precipitation was done on extracts that had been treated with DNaseI. The amounts of precipitate loaded and the control precipitates are the same as those described for B. The lanes labeled 5× represent the lanes directly above them in which five times more precipitate was subjected to electrophoresis.

Figure 3.

A model that shows communication between DNA ligase I and the flap endonuclease bound to the front face of PCNA at the site of flap cleavage and ligation. In this model, both flap endonuclease and DNA ligase I are bound to the same molecule of PCNA. Rotation of the flap endonuclease is dependent on the association of DNA ligase I with PCNA. The rotation occurs either in a pocket between DNA ligase I and PCNA or by movement of DNA ligase I as shown. In wild-type cells (A) the reactions are coupled whereas in the cdc9-p mutant (B) the reduction in binding of DNA ligase I prevents a signal from passing through PCNA from DNA ligase I to the flap endonuclease. This is the most parsimonious model based on the crystal structures that show that both enzymes bind to the front face of PCNA (Chapados et al. 2004; Pascal et al. 2004; Sakurai et al. 2005). Additional models in which the flap endonuclease and DNA ligase I bind to separate molecules of PCNA are also possible.