5.2 Summary of results
5.2.2 Modulation of Mcm2-7 activity by Cdt1
In budding yeast, Cdt1 is required for nuclear import and loading of Mcm2-7 onto chromatin in early G1 phase of the cell cycle. To investigate the interaction between Mcm2-7 and Cdt1 in isolation, individual components purified separately from bacterial- expression systems were combined and the complexes purified by size exclusion chromatography (Figure 3.1). Cdt1 co-eluted with each of the six Mcm subunits with the peak of elution equivalent to ~670 kDa and with equal stoichiometry (Figure 3.1A).
We next addressed whether the Mcm2-7•Cdt1 complex behaved differently from Mcm2-7 alone. Initially, the ability of the two complexes to hydrolyze ATP was compared. During both the titration of protein (Figure 3.2A) and time course experiments (Figure 3.2B), we consistently observed a lower ATPase activity with the Mcm2-7•Cdt1 complex compared to the Mcm2-7 complex. Overall, there was a two-fold reduction in the hydrolysis of ATP by Mcm2-7•Cdt1 relative to that of Mcm2-7 alone.
The two-fold reduction, rather than a complete inhibition, in ATPase activity by Mcm2-7•Cdt1 suggested that this reconstitution comprised a mixture of Mcm2-7 complexes with and without Cdt1. To this end, we asked whether the hydrolysis of ATP by the Mcm2-7•Cdt1 complex would be further inhibited if recombinant Cdt1 were added in solution (Figure 3.2C). While the rate of ATP hydrolysis by Mcm2-7 and Mcm2-7•Cdt1 was comparable after adding two-fold excess Cdt1, a maximum reduction in Mcm2-7•Cdt1 ATPase activity was reached with an equimolar amount of Cdt1 added. These data indicated that most, if not all, of the Cdt1 was associated with Mcm2-7 in the Mcm2-7•Cdt1 reconstitution.
We further compared the activity of Mcm2-7 and Mcm2-7•Cdt1 complexes in an established helicase assay 30. In this regard, complexes were incubated with radiolabelled synthetic dsDNA fork substrate containing a biotin-streptavidin moiety and the ssDNA generated by unwinding were separated from duplex DNA by native polyacrylamide gel electrophoresis. Similar to ATPase activity, the DNA unwinding activity by Mcm2- 7•Cdt1 was approximately two-fold lower than by Mcm2-7 alone (Figure 3.3). Unfortunately, the presence of nuclease contamination in the Cdt1 preparation nullified our ability to assess the effect on DNA unwinding by adding Cdt1 in solution to the Mcm2-7•Cdt1 complex.
We next addressed whether the DNA binding activity of Mcm2-7•Cdt1 was different from that of Mcm2-7. For this, radiolabelled Mcm2-7 or Mcm2-7•Cdt1 was incubated with either single-stranded M13 DNA (Figure 3.4) or double-stranded plasmid DNA (Figure 3.5). Whereas ssDNA binding increased in a DNA-dependent fashion, little to no difference in the ability of the two complexes to bind ssDNA was found. For dsDNA binding, Mcm2-7•Cdt1 demonstrated slightly lower binding relative to Mcm2-7 at all except the lowest concentration of dsDNA tested. Consistent with the poor ability of Mcm2-7 to bind dsDNA 31, 32, the overall amount of binding to dsDNA by either complex was much lower than that to ssDNA. Taken together, these results indicate that Cdt1 has little to no effect on the DNA binding activity of Mcm2-7. In addition and
consequently, the ability for Cdt1 to inhibit the DNA unwinding activity of Mcm2-7 was not due to a defect in DNA binding by this helicase complex.
To further investigate the individual roles of Mcm2-7 and Cdt1, we set out to assemble the pre-RC on replication origins from pure proteins and proteins purified from G1-arrested yeast extracts. To this end, Mcm2-7 was added to biotinylated ARS1 origin DNA coupled to streptavidin-coated magnetic beads in the presence of ORC, Cdc6 and ATP. Beads were washed with either low salt or high salt buffer to differentiate between Mcm2-7 complexes that were associated with versus loaded onto origin DNA, respectively. Consistent with published results 8, 10, 33, 34, we found that origin loading of Mcm2-7 was dependent on ORC, Cdc6 and Cdt1 (Figures 3.6 and 3.7). We also observed a greater extent of loading using Mcm2-7 and ORC purified from G1-arrested yeast extracts compared to pure proteins. To determine the cause for this discrepancy, we examined these complexes in more detail. Analysis of complexes run on SDS-PAGE gels (Figures 3.8A and B) revealed that there was approximately equal stoichiometry of subunits in the respective complexes. Since the hydrolysis of ATP by ORC and Cdc6 is required for the stable association of Mcm2-7 with replication origins, we determined the ATPase activity of the pre-RC factors in isolation or in combination with each other (Figure 3.8C). While low, the hydrolysis of ATP by bacterial-expressed ORC was two- fold lower than that of ORC purified from G1-arrested yeast extracts. In contrast, and regardless of the source of ORC, the ATPase activity of ORC and Cdc6 combined was similar.
Intriguingly, we found that in the presence of Cdt1, the Mcm2-7 complex was comprised of subunits at or near equal stoichiometry. In contrast, in the absence of Cdt1, subcomplexes of Mcm subunits were more pronounced (Figure 3.1; M. O’Donnell and L. Langston, unpublished). In light of this, we predicted that Cdt1 plays a role in stabilizing Mcm2-7. To this end, we evaluated by gel filtration the formation of Mcm2-7 subcomplexes in the presence or absence of Cdt1 using purified proteins (Figure 3.9). Similar to that observed for Mcm2-7, a subcomplex containing Mcm2, 4 and 6 neared stoichiometry in the presence as opposed to the absence of Cdt1 (Figures 3.9A and B). Whereas the addition of Cdt1 had no discernible effect on the stoichiometry of a Mcm357
(Mcm odds) complex, the formation of both trimers and hexamers of Mcm357 was observed in the absence of Cdt1 (Figures 3.9C and D). Furthermore, we observed that when all six Mcm subunits were expressed in yeast, and when ATP was omitted from affinity chromatography purification, only the Mcm odds were recovered (Figure 3.10). Encouraged by the above results and the recent finding that, in the presence of the poorly hydrolysable ATP analog ATPγS, only Mcm3, 5 and 7 were associated with origin DNA upon mixing Mcm2-7 in a pre-RC loading reaction lacking Cdt1 33, we performed a loading assay in the absence of Mcm2, 4, 6 and Cdt1 (Figure 3.11). Consequently, after treating loading reactions with low salt buffer, we observed that the Mcm357 subcomplex could associate with ARS1 in the absence of Mcm2, 4, 6 and Cdt1. While the association of the Mcm odds complex was dependent on nucleotide, there was a greater amount of origin recruitment in the presence of ATPγS relative to ATP. Additionally and consistent with published data 10, 33, none of the pre-RC factors bound to origin DNA in the absence of ORC and Cdc6. Finally, we also observed non-specific binding of ORC and Cdc6 in the absence of ATP.