Figure 5. Cbf1 also binds and dissociates from nucleosomes significantly slower than from DNA.

(A) Ensemble PIFE measurement of a Cbf1 titration with a 94-bp DNA with and without the Cbf1-binding sites 1 bp from the 5′ end and Cy3 labeled. The normalized PIFE fits to a binding isotherm with an S_{1/2 Cbf1–DNA PIFE} = 1.3 ± 0.3 nM. Without the binding site, the Cy3 emission does not change, demonstrating that the observed PIFE is due to site-specific Cbf1 binding. (B) Cbf1 titration with Cy3-Cy5 labeled nucleosomes with the Cbf1 site at P8. The FRET fits to a binding isotherm with an S_{1/2 Cbf1–smNuc FRET} = 12.3 ± 1.6 nM. (C) Example time traces of single DNA molecules for two separate Cbf1 concentrations, where the black lines are the Cy3 fluorescence and the red lines are the two-state Hidden Markov Model fits. As the Cbf1 concentration increases, the immobilized DNA molecules shift to the high PIFE state. (D) The Cbf1–DNA primary binding and dissociation rates for increasing concentrations of Cbf1. These were determined from cumulative sums of Cbf1–DNA high PIFE and low PIFE dwell times that were fitted to double exponentials. The primary dissociation kinetics (blue) were constant with an average of k_{off Cbf1–DNA primary} = 0.30 ± 0.05 s⁻¹, while the primary binding kinetics (red) fit to a line with a slope that equals the overall binding rate of k_{on Cbf1–DNA primary} = 0.025 ± 0.006 s⁻¹ nM⁻¹. (E) Example time traces of single nucleosomes with two separate Cbf1 concentrations, where the black lines are the FRET efficiency data and the red lines are the two-state Hidden Markov Model fits. As the Cbf1 concentration increases, the immobilized nucleosome shift to the low FRET state. (F) The Cbf1–nucleosome binding and dissociation rates for increasing concentrations of Cbf1. These were determined from cumulative sums of Cbf1–nucleosome low FRET and high FRET dwell times that were fitted to single exponentials. The dissociation kinetics (blue) were constant with an average of k_{off Cbf1–Nuc} = 0.0111 ± 0.0007 s⁻¹, whereas the binding kinetics (red) fit to a line with a slope that equals the overall binding rate of k_{on Cbf1–Nuc} = 0.00021 ± 0.00002 s⁻¹ nM⁻¹.

Figure 5—figure supplement 1. Characterizing Cbf1 interactions with DNA and nucleosomes.

(A) Cy3 image of the EMSA of Cbf1 binding to a 94-bp DNA sequence with the Cbf1-binding site (S_{1/2 Cbf1–DNA EMSA} ≈ 1 nM). (B) Cy5 image of the EMSA of Cbf1 binding to nucleosomes with the Cbf1-binding site at P8 (S_{1/2 Cbf1–Nuc EMSA} ≈ 11.5 nM). (C) Cumulative sum of low PIFE (red) and high PIFE (blue) dwell times for DNA (left) and nucleosomes (right) for each Cbf1 concentration. The DNA cumulative sums were fitted with double exponentials whereas the nucleosome cumulative sums were fitted with single exponentials. (D) Table of the log-likelihood ratio used to test the assumption that a cumulative sum fits to a double exponential curve. If the test is more than 0.01 for three of the four Cbf1 concentrations, we rejected the assumption and fitted a single exponential instead. For Cbf1–DNA, both unbound and bound cumulative sums follow double exponentials. For Cbf1–nucleosomes, both unbound and bound cumulative sums follow single exponentials. (E) Plots of primary (left) and secondary (right) rates for Cbf1 low ⇒ high PIFE (red) and high ⇒ low PIFE (blue) transitions in the interaction with DNA at increasing Cbf1 concentrations. These are determined from the exponential fits of the cumulative sums in panel (C). (F) The fractions of the fast transitions for low ⇒ high PIFE (light grey) and high ⇒ low PIFE (dark grey) of Cbf1 on DNA fwith increasing Cbf1 concentrations. Fast transitions rates accounted for ~60% and ~70% of the low ⇒ high and high ⇒ low events, respectively. (G) Ratio of nucleosome-binding affinity to DNA-binding affinity for Cbf1 for both single-molecule (SM) and ensemble (Ens) measurements. The ratio obtained using the dominant rates from single-molecule measurements is within about a factor of 2 of the ratio from ensemble measurements.