Binding dynamics of a monomeric SSB protein to DNA: a single-molecule multi-process approach

Single-stranded DNA binding proteins (SSBs) are ubiquitous across all organisms and are characterized by the presence of an OB (oligonucleotide/oligosaccharide/oligopeptide) binding motif to recognize single-stranded DNA (ssDNA). Despite their critical role in genome maintenance, our knowledge about SSB function is limited to proteins containing multiple OB-domains and little is known about single OB-folds interacting with ssDNA. Sulfolobus solfataricus SSB (SsoSSB) contains a single OB-fold and being the simplest representative of the SSB-family may serve as a model to understand fundamental aspects of SSB:DNA interactions. Here, we introduce a novel approach based on the competition between Förster resonance energy transfer (FRET), protein-induced fluorescence enhancement (PIFE) and quenching to dissect SsoSSB binding dynamics at single-monomer resolution. We demonstrate that SsoSSB follows a monomer-by-monomer binding mechanism that involves a positive-cooperativity component between adjacent monomers. We found that SsoSSB dynamic behaviour is closer to that of Replication Protein A than to Escherichia coli SSB; a feature that might be inherited from the structural analogies of their DNA-binding domains. We hypothesize that SsoSSB has developed a balance between high-density binding and a highly dynamic interaction with ssDNA to ensure efficient protection of the genome but still allow access to ssDNA during vital cellular processes.


Supplementary Methods
Ensemble fluorescence assays PIFE experiments were performed by exciting the Cy3 dye at 545 nm and calculating the emission area from 560 to 650 nm. The concentration of 12-mer ssDNA was 10 nM and the KCl concentration was 10 mM. PIFE measurements were performed at room temperature (20 °C) and 65 °C. The calculated Cy3 emission area at each concentration of SsoSSB was normalized with respect to the emission area in the absence of protein. Data were fitted using non-linear squares to a Hill binding model described by Equation S1: where Bmax is the maximum specific binding, KD is the concentration required for half-maximum binding, and  is the Hill coefficient.
Quenching experiments were performed by exciting Alexa647-SsoSSB at 645 nm and recording the fluorescence spectra from 650 to 800 nm. The percentage of Alexa647 quenching at each concentration of 12-mer DNA added was obtained by comparing the emission area at each DNA concentration with respect to that in the absence of DNA. The concentration of SsoSSB was 10 nM and the KCl concentration was 10 mM. All quenching experiments were performed at room temperature. The binding isotherm was fitted to a Hill binding model using Equation S1 (1).
Inter-molecular FRET experiments were performed exciting the Cy3 donor (exc 545 nm) as above and the energy transfer efficiency was calculated using the donor quenching method as described by Intra-molecular FRET measurements were performed under magic angle conditions to avoid anisotropy artifacts and analysed by exciting the donor dye, Cy3 at 545 nm and recording the emission spectrum from 560 to 800 nm. The acceptor dye (Alexa647) emission spectrum was also recorded using an excitation wavelength of 645 nm and emission monitored from 650 to 800 nm. The efficiency of energy transfer was calculated using the RatioA method because using this method variations in the donor and acceptor quantum yield due to protein binding do not interfere with the estimation of the FRET efficiency (2).

PELDOR measurements
All PELDOR data were recorded as previously described (3). With the exception of the following timings all settings and optimisation procedures were used as described previously (4). The pump pulse was set to 20 ns at X-band and 12 -14 ns at Q-band, 1 to 380 ns, 2 to 3 -4 s at X-band and 6 s at Qband, and the shot repetition time to 2.5 -3 ms, averaging the data for approximately 12 h at X-band or < 3.5 h at Q-band. Raw PELDOR data were subjected to background correction assuming a monoexponential decay, followed by Tikhonov regularisation in DeerAnalysis2013 (5). Resulting distance distributions were validated with the validation tool within DeerAnalysis2013, using a noise level of 1.50 (noise increased by 50% over the experimental noise) and 5 trials, and varying the background start time from 5% to 95% of the total data acquisition time window in 6 trials, resulting in 30 trials in total. A prune level of 1.15 was applied, i.e. retaining only data sets within 1.15 times of the best root mean square deviation. If less than 50% of the trials remained upon pruning (SSB at X-band and SSB + 9A at Q-band) data were cut by 10% and subsequently by 5% until more than 50% of the trials remained. This procedure eliminated validations dominated by distortions of the raw data at long dipolar evolution times. Shown in the validation figures are the ± 2 ×  confidence intervals as coloured shaded areas.

Molecular modelling of the SSB:DNA complex
The crystallographic structure 1JMC of RPA70 with single stranded DNA (1JMC) was used for generating the model. A new strand of 9C was generated using the 8C DNA in 1JMC as template, preserving the structure of the 3 first and 3 last nucleotides. Later, two SsoSSB structures (1O7I) were added aligning them with corresponding position of the RPA domains respect to the DNA. CCP4 and PyMol were used in this process. The structure was protonated and immersed in a rectangular water box using sodium atoms as counter-ions. The energy of whole system was minimized, and finally it was equilibrated up to 298 K. NAMD with Amber force-field parameters was used for these later steps. Figure S1. SsoSSB binding to a 12-mer Cy3-labelled single-strand DNA at 65° C monitored by protein induced fluorescence enhancement (PIFE). The solid line represents the fit to a Hill model (Equation S1

Supplementary Figures
). We obtained values of 15  1 nM and 1.4  0.2 for the dissociation constant and the Hill coefficient, respectively. Figure S2. Representative single-molecule PIFE trajectories obtained at 1 nM (a) and 10 nM (b) concentration of SsoSSB obtained with 50 ms integration time. SsoSSB association and dissociation can be observed as PIFE events in the single molecule trace. The Cy3 intensity in the absence of SssoSSB has been normalized to unity and this has been taken as the signal reference for the singlemolecule trajectories to quantify the relative increase in Cy3 intensity due to SsoSSB binding             Figure S16. Representative single-molecule donor (green) and acceptor (dark red) intensity trajectories (upper panels), FRET trace (middle panels, grey) and total intensity (bottom panels, black) showing stepwise disassembly events of the (SsoSSB)2:DNA complex marked with an asterisk. Upper row: disassembly events taking place as transitions from PIFE to FRET states corresponding to dissociation of a single monomer and the subsequent disruption of acceptor quenching. Lower row: disassembly events taking place as a stepwise decrease in Cy3 intensity corresponding to loss of PIFE due to single monomer dissociation. Both types of disassembly events are very rare and represent less than 15% of the total number of dissociation events. Figure S17. Single-molecule dwell-time histograms obtained for the dissociation of (SsoSSB)2:DNA complexes ( 3→ 1 ) formed by monomer-to-monomer incorporation leading to free ssDNA at the indicated concentrations of SsoSSB. Solid lines represent the results from monoexponential fitting of the dwell-time histogram to extract the dissociation rate at the indicated concentrations of SsoSSB. Figure S18. Single-molecule dwell-time histograms obtained for the association (a) of both monomers to a surface-immobilized 12-mer dC single-strand DNA in a single step (S1S4) at the indicated concentrations of SsoSSB. The association rate ( 1→ 4 ) was determined from the dwelltime of unbound states (S1) directly leading to PIFE events (see main text for details). Solid lines represent the result from fitting the dwell-time histograms to a mono-exponential decay function to extract each the association rate at each concentration. (b) Single-molecule dwell-time histograms obtained for the dissociation of (SsoSSB)2:DNA complexes ( 4→ 1 ). Only those complexes formed following a single S1S4 step at the indicated concentrations of SsoSSB were taken for this analysis. Solid lines represent the results from monoexponential fitting of the dwell-time histogram to extract the dissociation rate at the indicated concentrations of SsoSSB.