Human Rap1 modulates TRF2 attraction to telomeric DNA

More than two decades of genetic research have identified and assigned main biological functions of shelterin proteins that safeguard telomeres. However, a molecular mechanism of how each protein subunit contributes to the protecting function of the whole shelterin complex remains elusive. Human Repressor activator protein 1 (Rap1) forms a multifunctional complex with Telomeric Repeat binding Factor 2 (TRF2). Rap1–TRF2 complex is a critical part of shelterin as it suppresses homology-directed repair in Ku 70/80 heterodimer absence. To understand how Rap1 affects key functions of TRF2, we investigated full-length Rap1 binding to TRF2 and Rap1–TRF2 complex interactions with double-stranded DNA by quantitative biochemical approaches. We observed that Rap1 reduces the overall DNA duplex binding affinity of TRF2 but increases the selectivity of TRF2 to telomeric DNA. Additionally, we observed that Rap1 induces a partial release of TRF2 from DNA duplex. The improved TRF2 selectivity to telomeric DNA is caused by less pronounced electrostatic attractions between TRF2 and DNA in Rap1 presence. Thus, Rap1 prompts more accurate and selective TRF2 recognition of telomeric DNA and TRF2 localization on single/double-strand DNA junctions. These quantitative functional studies contribute to the understanding of the selective recognition of telomeric DNA by the whole shelterin complex.


Electrostatic contribution to binding affinity
The number of ion pairs formed upon protein-DNA binding and corresponding electrostatic contribution to overall binding affinity (Ka) could be derived from the dependence of the binding constant on salt concentration according to the equation: log Ka = log Ka nel  Z•φ•log [NaCl] (1) where Ka is association constant representing the overall affinity, Ka nel is non-electrostatic component of association constant, Z is the number of DNA phosphates that interact with the protein, φ is the number of Na + cations released from the interaction with the phosphate group after binding of protein to DNA. For double-stranded B-DNA molecule with a length of 24 bp and shorter, the value for φ is approximately 0.64 (26). The association constant (Ka) were calculated as the reciprocal value of the dissociation constant. From the association constant is possible to calculate the value of the Gibbs free energy according to the equation: The binding energy could be divided into electrostatic and non-electrostatic terms: The electrostatic component disappears when the salt concentration approaches 1 M and the overall energy of binding is given only by the non-electrostatic component:

Surface plasmon resonance analysis
Protein-protein interaction between TRF2 and Rap1 was analyzed on the ProteOn GLC sensor chip (Bio-Rad).
Protein TRF2 in concentrations from 10 to 2.5 μg/ml was immobilized on the chip in a PBST buffer onto five of the six ligand channels using the amine coupling reagents 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride (EDAC, 133 mM) and N-hydroxysulfosuccinimide (sulfo-NHS, 33 mM). The sixth channel was not modified and served as a reference channel. Remaining unmodified surface groups were deactivated with 1 M ethanolamine HCl, pH 8.5. A sample of running buffer was used for ligand injection in the reference channel. For analyte binding, five solutions of Rap1 were prepared at concentrations of 80, 40, 20, 10 and 5 nM by serial dilution in PBST buffer. The sixth channel was used as reference channel using only the running buffer. The samples were injected for 240 s at a flow rate of 100 μl/min; dissociation time was 1200 sec. The obtained sensorgrams were processed for baseline alignment and reference channels subtraction ( Figure S4 A). Equilibrium response levels were plotted against the analyte concentration. The equilibrium response plot was fitted using a simple bimolecular equilibrium model (Figure S4 B), from which the equilibrium dissociation Kd was determined.

DNA oligonucleotides used in SPR studies
Three different double-stranded oligonucleotides were used for SPR analyses. DNA duplex SR2 containing two human telomeric repeats 5'-TTAGGG-3' was prepared by annealing the 5' end biotinylated oligonucleotide (oligo) 5'-CTAACCCTAACCCTAAGTTAG-3' to the oligo 5'-CTAACTTAGGGTTAGGGTTAG-3'. DNA duplex Overhang_SR2 which consisted of a 21-bp stem with a 6 nucleotide overhang was prepared by annealing the 5' end Rap1-TRF2 was prepared in 1:1 molar ratio. The sixth channel was used as a reference channel using only the running buffer. The samples were injected for 350 s at a flow rate of 70 μl/min; the dissociation time was 900 s. The obtained sensorgrams were processed for baseline alignment and reference channels subtraction ( Figure S78). Kinetic analyses were performed for the five analyte concentrations. Values of rate dissociation constant kd (k-off) was obtained from concentrations of analyte, which gave the best response signal (Table S4). Table S1. Primers used to prepare plasmid constructs for expression of Rap1 and TRF2.   Values of koff were obtained by fitting of dissociation part of sensorgrams in Figure S8 from the starting value Rmax of response unit (RU). Ch 2 is Chi-Square goodness of fit.          Figure S11. The release of telomeric DNA pre-bound to TRF2 full length or TRF2 lacking N-terminal basic domain (TRF2ΔB) after Rap1 addition. Either TRF2 or TRF2ΔB was allowed to bind with telomeric DNA R2 (7.5 nM) labelled by Alexa Fluor 488. The formation of TRF2-DNA or TRF2ΔB-DNA complex was demonstrated by an increase of anisotropy value. Next, instead of TRF2 or TRF2ΔB, Rap1 was added to the solution, which led to an immediate drop in the anisotropy. More DNA is released when TRF2ΔB is used, which is in accordance with our previous findings. Figure S10. Rap1 improves selectivity of TRF2 binding to telomeric DNA. A: The quantification of binding association constant of TRF2 with or without of Rap1 to the telomeric DNA duplex R2 or nontelomeric DNA duplex N2 based on fluorescence anisotropy measurements (B). The comparison of binding affinity ratios for binding to telomeric and nontelomeric DNA (selectivity) revealed that Rap1 increases selective binding of TRF2 to telomeric DNA. Oligonucleotides in the cuvette (7.5 nM) were labelled by Alexa Fluor® 488. Measurement conditions were the same as described in materials and methods above. Note that binding saturations occur at significantly higher protein concentrations in case of protein binding to the nontelomeric DNA duplex N2 when compared to binding to the telomeric duplex R2.
Figure S12. Rap1 does not bind DNA duplex R2 or Overhang_SR2. The interaction of Rap1 to the fully hybridized DNA duplex R2 (A) or Overhang_SR2 (B) was monitored by fluorescence anisotropy and compared with TRF2 binding to DNA duplex R2. Oligonucleotides in the cuvette (7.5 nM) were labelled by Alexa Fluor® 488. Assay conditions were the same as described in materials and methods above. Note that binding was not observed even at significantly higher Rap1 concentrations compared to TRF2 concentrations. Figure S13. Rap1 induces a partial release of TRF2 from a preformed TRF2-DNA complex also at DNA concentrations well below Kd. Reaction mixtures contained the same amount of DNA (0.1 pmol), constant amount of TRF2 (1.2 pmol) and increasing amount of Rap1 (0.2-0.4 pmol). The numbers above electrophoretic lanes represent molar ratios of proteins and DNA in individual wells. Note that the saturation molar ratio of TRF2-DNA complex had to be changed due to the DNA concentration being significantly lower than Kd of TRF2 binding to DNA. The binding saturation occurs at TRF2:DNA ratio 12:1. DNA duplex R2 was labelled with Alexa Fluor 488. The reaction mixtures (15 μl) were resolved on horizontal 4% non-denaturing polyacrylamide gel. The reaction mixtures were resolved on horizontal 4% nondenaturing polyacrylamide gels. Note that after addition of Rap1 to TRF2 pre-bound to DNA Ov with single-stranded overhang (B) no release of free DNA is observed.