Response Factor Correction for Quantitative Determination of Homooligomeric SsoSSB Binding to ssDNA by Native Mass Spectrometry

Abstract

Native mass spectrometry (nMS) has emerged as a complementary approach for elucidating molecular parameters of biological complexes relative to solution-phase experiments. Herein, we utilize nMS to determine the subunit binding affinities (K _d,i ) of the single-stranded DNA binding protein (SSB) from Saccharolobus solfataricus (Sso) to poly dT single-stranded DNA (ssDNA) compared with the apparent K _d ′ values obtained from solution-phase fluorescence anisotropy. This work resolves conflicting previous biochemical reports on the stoichiometry and affinities of SsoSSB while also highlighting the advantages and limitations of nMS quantification. Covalent concatemers of SsoSSB with increasing molecular weights were utilized as response factor (RF) standards to correct for physical and instrumental parameters that systematically underrepresent abundances from ionization of larger mass species. Furthermore, we show that regardless of the nMS quantification metric (peak area or intensity), meaningful comparative data can be extracted from multicomponent biochemical systems. Importantly, the binding affinities of the individual species (K _d,i ) determined by nMS approach the apparent K _d ′ from bulk solution-phase measurements but have the added benefit of separately quantifying individual binding steps within a multistep assembly process. Interestingly, the stoichiometries of SsoSSB binding to 15 or 30 nucleotides of ssDNA measured by nMS are subsaturating. The calculated binding affinities of the first and second SsoSSB molecules show some positive cooperativity, while the binding of the third is an order of magnitude weaker, suggesting that negative cooperativity is utilized to limit binding near the ends of the available length of the ssDNA.

Introduction

Single-stranded DNA (ssDNA) binding proteins (SSBs) are essential components of DNA-enzymatic assemblies that ensure appropriate replication and repair of genomes. Thus, SSBs are essential in all Domains of life to coat and protect any labile ssDNA that is present during genome maintenance processes. The SSB from the thermophilic archaeon, Saccharolobus solfataricus (SsoSSB), is integral for processes related to DNA replication, repair, and recombination, serving as a scaffold for recruiting essential enzymes involved in downstream processes. SsoSSB contains an archetypical oligonucleotide-binding fold (OB-fold), which binds ssDNA with a site size of ∼5–6 nucleotides (nts). Many SSBs organize into various oligomeric complexes to perform diverse regulatory functions, including human RAD51, the bacterial DnaA replication initiator, and several others, including the E. coli SSB (EcSSB) tetramer involved in genomic protection during replication. However, there is conflicting evidence on the active oligomeric state of SsoSSB, with monomer, dimer, and tetramer species having been previously observed in solution. ,− SsoSSB does not appear to wrap or condense ssDNA around any quaternary complexes like that known for EcSSB, where two binding modes have been shown that either bind two of the four subunits or wrap around all four subunits. − Instead, SsoSSB binding resembles that of bacteriophage T7 gp2.5 or T4 gp32, where a single OB-fold engages the ssDNA, allowing for a contiguous but one-dimensional mobile bound protective layer. , Yet, whether binding of a single SsoSSB nucleates a cooperative assembly of additional protein subunits along the ssDNA strand through interfacial binding energies has not been determined.

Traditional solution-phase biochemistry techniques have been used for more than 50 years to quantify the dissociation constants (K _d ) and cooperativity of protein–ligand binding. Although solution-phase approaches allow for convenient measurements, for systems like SSBs with multiple binding events to ssDNA, only ‘apparent’ dissociation constants (K _d ′) for the tightest binding step within the multisubunit assembly are determined as these measurements cannot distinguish individual binding steps in multicomponent systems (Figure A). More recently, native mass spectrometry (nMS) has been used for the characterization of biomolecular assemblies. In nMS, it is assumed that the equilibrium conditions in solution are maintained as the analytes are transferred to the gas phase. Therefore, the relative abundances of each species within the system can be directly measured, thereby allowing for the calculation of thermodynamic parameters that govern the behaviors of multisubunit protein–ligand complexes. − Furthermore, the fractional abundances of free and ligand-bound proteins from mass spectra can be used to determine individual dissociation constants (K _d,i ) of each binding event, even in a complex assembly process (Figure B). − By simultaneously quantifying the relative populations of species, we can not only determine individual K _d,i values for each oligomeric complex (i.e., monomer, dimer, trimer, etc.) but we can also compare them (K _d,1, K _d,2, K _d,3, etc.) to better understand whether independent or dependent assembly processes occur. nMS has provided numerous insights into binding stoichiometries, assembly pathways of multisubunit complexes, and the thermodynamic descriptors (K _d,i and cooperativity), which are critical for understanding cellular processes. −

1.

Schematic of oligomeric SsoSSB binding to ssDNA binding and quantifications. (A) Bulk solution-phase binding equilibria can be used to determine apparent dissociation constants (K _d ′) and Hill coefficients for cooperativity (n) in a multibinding system. (B) Alternatively, native mass spectrometry (nMS) can separate individual species by mass, which can be used to quantify fractional abundance. A quantitation window (shaded gray), where all species are present, can be used to calculate actual individual K _d,i values. (C) SsoSSB ssDNA binding schematic, species, and parameters.

nMS employs nanoelectrospray ionization (nESI) with low flow rates and voltages to maintain native protein folding and noncovalent interactions during transition from solution to the gas phase. It is generally assumed that the binding of a small molecule does not significantly alter the signal for protein complexes compared to the unbound (apo) protein. − However, a persistent challenge in nMS is that equimolar protein species with different masses, such as formation of protein–protein or protein-oligonucleotide complexes compared to unbound proteins, may not produce equivalent signals, which makes extracting quantitative measurements for calculating K _d,i values challenging. These variations in signal abundance can be due to experimental and/or instrumental parameters, including differences in sample ionization, ion transmission efficiency through the instrument, and detector sensitivity. −

Discrepancies in ionization efficiencies from different molecular weight (MW) species are difficult to predict a priori. Differences in the physical properties of the system in addition to mass, including solvent accessible surface area, polarity, and structural conformations, can cause variability in the abundance measurements that are required to calculate K _d,i . ,,, In general, small molecules binding to proteins can be measured directly using a titration approach without significant differences in ionization efficiencies for the unbound and bound complexes. ,, However, this problem is exacerbated within oligomerizing protein systems, as successively larger complexes may experience increasing signal suppression from additional factors, including enhanced retention of nonspecific adducts, competition for protons in ESI droplets, or other unknown factors. , Although implementation of nESI mitigates some of these factors by generating smaller ESI droplets, resulting in greater desolvation efficiency and less nonspecific adduction due to fewer ions per droplet, additional methods are needed to address the differences in homooligomerizing biomolecular systems with different MW and physical properties.

Response factors (RF), which quantify differences in ion signal as a function of MW can be used to normalize the signal abundance for proteins of different size. ,, Appropriate RF calibration methods enable the correction of systematic deviations in signal abundances for analytes with higher masses. Gross and co-workers determined the association constant, K _a , for the dimerization of gramicidin using a model that accounted for complex dissociation in the gas phase using a RF that accounted for significantly different signals of the complex compared to the free protein. Gabelica’s group used a titration approach with a constant internal standard to determine RF and K _a for intercalators binding to duplex DNA. Klassen’s group also found that the addition of a larger P particle altered the signal responses for monitoring the affinity of a octasaccharide ligand to an antibody fragment and that in-source dissociation of fatty acids for β-lactoglobulin can significantly affect K _a measurements. They were able to overcome these limitations by developing slow mixing mode (SLOMO) nMS which can simultaneously extract RF and K _d values from continuous measurements as a binding system approaches or departs from the steady-state equilibria state. Zenobi’s group determined the RF and K _a for the dimer-tetramer equilibrium of concanavalin A by correlating it with isothermal calorimetry (ITC) measurements. They also investigated RFs of different sized proteins by generating a series of maltose-binding protein (MBP) concatemers that varied only in MW and surface-area-to-volume ratios. Notably, when MBP concatemers of different MWs were analyzed, the linearity of signal abundance to concentration deviated at higher concentrations, which was attributed to differential ion suppression effects intrinsic to each species, thus providing an upper concentration limit on signal responses. Yet, constructing and purifying artificial concatemers for each protein system studied is challenging. Further, studies by the Zenobi lab demonstrated that larger protein complexes exhibit markedly lower response factors (RFs) per mole than smaller protein species, indicating nonlinear variations of signal abundance with MW. Empirical determination of RFs using appropriate reference proteins remains the most direct way to determine RFs for recombinant proteins and is currently necessary to achieve accurate quantitation of larger oligomeric assemblies from nMS. By applying biologically relevant concatemeric response factor protein standards, we can more accurately quantify and compare biological assembly processes, including multiple SSB proteins binding ssDNA molecules of various lengths.

nMS has been used previously to examine protein–DNA assemblies involved in DNA replication and repair, − but it has yet to be applied to understand the assembly of multiple SSB proteins along ssDNA to determine whether independent or dependent (i.e., cooperative) binding processes are utilized. Herein, we use nMS to directly and simultaneously examine individual species, oligomeric states, thermodynamic binding parameters, and any potential cooperativity parameters for the in vitro analysis of SsoSSB, allowing for a comprehensive analysis of this protein and resolving prior conflicting reports. We utilized artificial covalent dimers and trimers of SsoSSB and monomeric SsoSSB to determine RFs and correlate differences in signal abundance for noncovalent oligomeric species bound on ssDNA. By correcting nMS-derived abundances with RF standards, differences in RFs across homo-oligomeric species can be utilized for measuring the actual stepwise K _d,i values of individual binding events of SsoSSB with ssDNA. Our work bridges the gap between MS-based measurements and solution-phase binding equilibria by applying a well-defined, experimentally based correction strategy. This approach ensures that differences in RF and other experimental factors that may skew the signals of species with higher MW are addressed, leading to more reliable quantitative analyses.

Altogether, nMS has revealed that SsoSSB exists as a monomer in solution and initially binds ssDNA cooperatively for the first and second event which approach that of solution-based apparent K _d ′ values. However, we also find that the equilibria distributions of SsoSSB subunits on ssDNA that are measured by nMS are subsaturating compared to previous solution-phase stoichiometry measurements. This suggests SsoSSB imparts a degree of negative cooperativity for the binding of the last subunit that keeps the SsoSSB molecules one-dimensionally spaced along the ssDNA or away from the ends. These findings highlight the utility of RF calibrated nMS for decoding complex binding equilibria that advance our mechanistic understanding of biological assemblies including SsoSSB binding to ssDNA.

Experimental Procedures

Materials

HPLC purified oligonucleotides (Table S1) were purchased from IDT (Coralville, IA). Lyophilized ubiquitin from bovine erythrocytes was purchased from Sigma-Aldrich (St. Louis, MO). Optikinase and Superfi II were purchased from Fisher Scientific (Waltham, MA). Restriction enzymes were purchased from NEB (Ipswich, MA). All other materials were from commercial suppliers and were analytical grade or better.

Cloning and Protein Purification of WT Monomeric SsoSSB

SsoSSB (strain DSM 1617) was PCR amplified from genomic DNA (ATCC # 35092D-5) using specific primers (Table S1), cloned into the pENTR SD-TOPO Gateway cloning vector, and recombined into pDEST14 using LR Clonase II (Fisher Scientific) using manufacturer directions. The resulting plasmid was transformed into Rosetta2 Bl21 DE3 cells (MilliporeSigma, Burlington, MA) and expressed using an autoinduction protocol. The cells were pelleted, resuspended in Lysis buffer (see Table for a description of all buffers), lysed in a French press, and centrifuged for 30 min at 4 °C, 75,000 × g in a Avanti J-26 XP centrifuge (Beckman Coulter, Brea, CA) using a JA-25.50 rotor. The clarified supernatant was then heat-treated for 30 min and mixed every 5 min by inversion. The solution was then centrifuged again and syringe filtered using a 0.45 μm membrane filter (Adamas-Beta, Shanghai, China). The clarified solution was diluted by a factor of 4 and loaded onto a 5 mL HiTrap heparin HP column (Cytiva, Marlborough, MA) using an AKTA Pure FPLC (Cytiva) equilibrated with buffer A and eluted with a 2-h linear gradient of buffer B (Table ). Fractions (4 mL) were collected. SsoSSB-containing fractions were pooled, concentrated, and desalted with a 20 mL Pierce poly­(ether sulfone) 30K molecular-weight cut off (MWCO) Protein Concentrator (ThermoFisher, Waltham, MA) into storage buffer. Finally, protein samples were flash frozen in liquid nitrogen for storage at −80 °C. Analytical gel filtration was performed on a Superdex S 200 column in buffer A. The concentration was quantified using a NanoDrop 2000c Spectrophotometer (ThermoFisher) using (MW = 16,137.9 Da, extinction coefficient (ε) = 12,660 M^–1 cm^–1).

1. Buffer or Solution Compositions and Uses.

name	composition	use
lysis buffer	20 mM tris-base pH 7.0, 300 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, and 1 mM benzonase	cell lysis
buffer A	20 mM tris-base pH 7.0, 1 mM ethylenediaminetetraacetic acid (EDTA), and 1 mM 2-mercaptoethanol (BME)	purification
buffer B	20 mM tris-base pH 7.0, 2 M NaCl, 1 mM EDTA, and 1 mM BME	purification
storage buffer	20 mM tris-base pH 7.0, 365 mM NaCl, 1 mM EDTA, and 1 mM BME, 10% glycerol	–80 storage
buffer C	20 mM tris-base pH 7.0, 100 mM KCl, 20 mM imidazole, 1 mM EDTA, and 1 mM BME	purification
buffer D	20 mM tris-base pH 7.0, 800 mM KCl, 20 mM imidazole, 1 mM EDTA, and 1 mM BME	purification
buffer E	20 mM tris-base pH 7.0, 100 mM KCl, 1 M imidazole, 1 mM EDTA, and 1 mM BME	purification
FA buffer 1	20 mM tris-base pH 7.5, 100 mM NaCl, and 1 mM EDTA	FA
FA buffer 2	20 mM tris-base pH 7.5, 200 mM NaCl, and 1 mM EDTA	FA
NH4OAc	200 mM ammonium acetate (pH = 7.1 ± 0.1)	FA and nMS

Cloning and Protein Purification of SsoSSB Concatemers

A codon-optimized construct containing three SsoSSB subunits (Supplementary Figure S1) cloned in sequence (pUC-GW-AMP-SsoSSB-ABC) was purchased from Genewiz (South Plainfield, NJ) with a 6X His tag followed by a Tobacco Etch Virus (TEV) protease linker at the N-terminus. The first two SsoSSB sequences (A and B) included residues 1–124, removing the C-terminal linker, while the last C subunit contained the entire protein coding sequence including the C-terminal tail (residues 1–148). pET30a and pUC-GW-AMP-SsoSSB were digested with NdeI and XhoI, and the digested DNA bands were gel extracted, ligated with T4 DNA ligase (NEB, Ipswich, MA), and transformed into MachI cells. The pET30a-SsoSSB-ABC construct encoding the trimeric concatemer was digested with NcoI to generate pET30a-SsoSSB-BC by removing the A subunit, which was then religated and transformed into MachI cells. Positive transformants were confirmed by DNA sequencing (Plasmidsaurus, Eugene, OR).

The resulting plasmids were transformed into Rosetta2 Bl21 DE3 cells and expressed and purified similar to monomeric SsoSSB. The clarified supernatant was then loaded onto a His-trap Co²⁺ column with buffer C (Table ). Bound protein was washed with buffer D, again washed with buffer C, and eluted with 100% buffer E (Table ). SSB-containing fractions were pooled and desalted into buffer C using a 30 K MWCO concentrator. The purified protein was treated with TEV-protease for 16 h at 4 °C. The digested protein was again loaded on a His-trap Co²⁺ column and the flowthrough was collected, buffer exchanged, and concentrated in storage buffer (Table ) and flash frozen for storage at −80 °C. Concentrations were quantified using a NanoDrop 2000c Spectrophotometer (ThermoFisher) for monomeric SsoSSB-WT (MW = 16,137.9 Da, ε = 12,660 M^–1 cm^–1), dimeric SsoSSB-BC (MW = 29,437.9 Da, ε = 23,490 M^–1 cm^–1), and trimeric SsoSSB-ABC (MW = 42,906.1 Da, ε = 34,490 M^–1 cm^–1), respectively.

Fluorescence Anisotropy

Fluorescence anisotropy (FA) experiments were performed with 8 nM Cy5 tagged DNA129 (Supplemental Table S1) on a Tecan Spark (Männedorf, Switzerland) microplate reader in a 384 well black plate (Corning, Corning, NY). The fluorescence was excited at 630 nm (20 nm bandwidth), and the emission with various combinations of polarizers was monitored at 680 nm (20 nm bandwidth) for 30 flashes with an integration time of 40 μs each. The fluorescence anisotropy, r, was calculated automatically by the instrumental software using the equation:

$$r=\frac{I_{VV}-G{I}_{VH}}{I_{VV}+2G{I}_{VH}}$$ 1

where I is the polarized fluorescence intensity with subscripts V and H identifying vertically and horizontally polarized light, respectively. The G factor is a correction for the difference in sensitivities of detection for horizontally and vertically polarized light. Reactions were prepared to a volume of 80 μL in the dark using three solutions, FA buffer 1 and FA buffer 2 (with increasing concentrations of NaCl), or 200 mM ammonium acetate (NH₄OAc) (pH = 7.1 ± 0.1) (Table ). Samples were buffer exchanged from storage buffer via overnight dialysis at 4 °C. The reported data are averages with error bars representing ± one standard deviation of 10 technical replicates. The data was analyzed and fit using GraphPad Prism (San Diego, CA, v10.6) using a total site binding model that accounts for nonspecific binding as described below:

$$A=\frac{B_{\max }\times P}{K_d^{\prime }+P}+\mathrm{NS}\times P+A_0$$ 2

or a cooperativity model

$$\mathrm{A}=\frac{B_{\max }\times P^h}{{K_d^{\prime }}^h+P^h}+A_0$$ 3

where A is the change in anisotropy, P is the protein concentration, B _max is the saturation point for A, K _d ′ is the apparent dissociation constant, NS is the slope for any nonspecific binding between proteins in solution interacting with SsoSSB-ssDNA bound species at higher protein concentrations, A ₀ is the background signal of the unbound Cy5-ssDNA, and h is the Hill coefficient.

nESI-MS Acquisition Conditions

Protein samples were prepared in 200 mM NH₄OAc (pH = 7.1 ± 0.1) to a concentration of 5 μM following two rounds of buffer exchange with Bio-Spin P6 desalting columns (Bio-Rad, Hercules, CA). ssDNA ligands were resuspended and diluted in 200 mM NH₄OAc. Equimolar protein samples were combined with 200 mM NH₄OAc and equal volumes of ssDNA to make 10 μL samples and incubated on ice for 15 min prior to nESI-MS analysis. Capillaries were prepared from borosilicate glass tubes using a P-1000 Flaming/Brown Micropipette Puller System (Sutter Instrument, Novato, CA) to consistent nESI diameters (1.10 μm ± 0.05 μm) as measured by scanning electron microscopy (SEM, Hitachi TM3030 Plus, Tokyo, Japan) (Supplementary Figure S2A,B). For each nESI-MS experiment, 3 μL of sample was loaded using a microloader pipet tip (Eppendorf, Hamburg, Germany) and analyzed on a Synapt G2-S mass spectrometer (Waters MS-Technologies, Manchester, UK). The instrument was operated in resolution MS mode (V-Mode) at 0.5 scan/s. Mass spectra were acquired in positive-ion mode with a mass window of 50–12,000 m/z. Signals were averaged over the course of 1 min.

nESI direct infusion was performed using a custom nESI sample holder (Supplementary Figure S3). The capillary voltage was set to 1.1 kV. The instrumental parameters are as follows: source temperature, 65 °C; sample cone, 40 V; source offset, 0 V; transfer collision energy, 5 V; trap DC bias, 2 V; trap gas flow, 5 mL/min; helium cell gas flow, 110 mL/min, and the trap pressure, 1.82 ± 0.01 × 10^–2 mbar (Ar). External mass calibration was performed with 100 mg/mL sodium iodide (Sigma-Aldrich) dissolved in a 50:50, v/v, methanol/water solution.

MS Data Analysis and Processing

Data acquisition was performed using MassLynx software (v.4.2 Waters, Milford, MA) to measure species abundances as outlined in the Results. MS data extraction, text conversion, and charge deconvolution was performed using UniDec (v. 6.0.1).

A custom python-based script was written to integrate areas under overlapping peaks using CSV text files generated by UniDec. First, peaks were smoothed by boxcar averaging. Then, Gaussian functions were fitted to the smoothed intensity profile of the overlapping peaks using a nonlinear least-squares approach (Levenberg–Marquardt algorithm). For a specified number of species, the program simultaneously fit the Gaussian components using independent fitting parameters. For each fitted peak, the program returned the integrated peak area, which was calculated as

$$\mathrm{Area}=\mathrm{height}\times \sigma \times \sqrt{2\pi }$$ 4

where the height and σ are the height and standard deviation of the fitted peak, respectively. The resulting peak areas were either quantified or normalized to account for varying RFs (see below). Excel (Microsoft, Redmond, WA) and GraphPad Prism (v10.6) were used for data processing and visualization.

Direct nESI-MS Binding Measurements

To quantify the binding affinities of SsoSSB to ssDNA, a sequential binding model with independent K _d,i s was utilized. Binding of three SsoSSB proteins (P) stepwise to ssDNA (D) can be described by

$$D+P\rightleftharpoons \mathrm{DP}$$ 5

$$\mathrm{DP}+P\rightleftharpoons {\mathrm{DP}}_2$$ 6

$${\mathrm{DP}}_2+P\rightleftharpoons {\mathrm{DP}}_3$$ 7

For binding of SsoSSB (P) to a ssDNA sequence of 15 homopolymeric deoxythymidines (T15), only a dimer (DP ₂) was detected, utilizing eqs and , while binding to T30, a trimer (DP ₃) was detected, utilizing eqs –. The mass balance of all DNA and protein species are

$$[D{]}_{\mathrm{o}}=[D]+[\mathrm{DP}]+[{\mathrm{DP}}_2]+[{\mathrm{DP}}_3]$$ 8

$$[P{]}_{\mathrm{o}}=[P]+[\mathrm{DP}]+2[{\mathrm{DP}}_2]+3[{\mathrm{DP}}_3]$$ 9

where [D]_o and [P]_o are the initial concentrations of ssDNA and SsoSSB, respectively. The dissociation constants (K _d,i ) for sequential binding to ssDNA for each species (i) can be expressed by

$$K_{d,1}=\frac{[D][P]}{[DP]}=\frac{[P{]}_{\mathrm{o}}-nFs[D{]}_{\mathrm{o}}}{R_1}$$ 10

$$K_{d,2}=\frac{[DP][P]}{[D{P}_2]}=\frac{[P{]}_{\mathrm{o}}-nFs[D{]}_{\mathrm{o}}}{R_2/R_1}$$ 11

$$K_{d,3}=\frac{[D{P}_2][P]}{[D{P}_3]}=\frac{[P{]}_{\mathrm{o}}-nFs[D{]}_{\mathrm{o}}}{R_3/R_2}$$ 12

where n is the maximal number of bound species (i.e., dimer or 2 on the T15 substrate and trimer or 3 on the T30 substrate), R ₁, R ₂, R ₃ are the abundance ratios of free SsoSSB to the SsoSSB-bound ssDNA measured by

$$R_i=\frac{Ab{(P)}_i}{Ab{(PD)}_i}=\frac{[P_i]}{[{DP}_i]}$$ 13

and fractional abundances (FA _i ) for each species (i) are described by

$${\mathrm{FA}}_i=\frac{\sum i{R}_i}{3\left(1+\sum i{R}_i\right)}$$ 14

Any cooperativity relationships for equivalent binding of multiple SsoSSB molecules to ssDNA can be described by

$$2K_{d,1}=\frac{1}{2\alpha }{K}_{d,2}$$ 15

for two binding events or

$$3K_{d,1}=\frac{1}{\alpha }{K}_{d,2}=\frac{1}{3\beta }{K}_{d,3}$$ 16

for three binding events, where α and β are cooperativity parameters.

Determination of RF Coefficients and Abundance Corrections

RF coefficients were determined by analyzing equimolar mixtures (5 μM) of each SsoSSB concatemer and an internal reference protein, bovine ubiquitin (Ub) (MW = 8564.84 Da, ε = 1490 M^–1 cm^–1), or a mixture of the three SSB constructs. Ub was utilized as it remains stably soluble in 200 mM NH₄OAc and is observed as easily resolvable peaks in nonoverlapping m/z ranges compared to the experimental system. When each SsoSSB concatemer was analyzed separately, ubiquitin served as an internal standard. nESI-MS samples were prepared by mixing equimolar species and incubating on ice for 15 min prior to analysis. The summed signal abundances of each concatemer (all charge states) were normalized to the summed abundances of ubiquitin, plotted, and fit to an exponential decay curve using GraphPad Prism (v10.6)

$$\mathrm{RF}={\mathrm{RF}}_{\min }+({\mathrm{RF}}_0-{\mathrm{RF}}_{\min })\times e^{-k\times \mathrm{MW}}$$ 17

where RF_min is the plateau minima, RF₀ is the Y-intercept, and k is the exponential rate as a function of MW to calculate relative RF equation. Alternatively, when all three SsoSSB species were mixed directly, the summed signal abundances for each were normalized to the monomer and fit to a linear regression. The R of each species (i) from eq was then corrected to the RF of the free SsoSSB monomer (P) from the exact masses of i following

$$R_i^{\mathrm{RF}}=\frac{R_i\times [P]}{[P_i]}$$ 18

to give RF-corrected fractional abundances (R _i ).

Theoretical Population Distributions Based on Binding Site Size and Template Length

Using statistical partition functions, the expected populations of each ssDNA-bound species can be derived from

$${\mathrm{\Xi }}_{NN}^k=\sum _{k=0}^n{S}^k\times W_k(\alpha )$$ 19

where n is the total number of monomer units, S is a variable used to track complex size, k is the number of monomers in a complex from 0 to n, α is a system parameter to represent a statistical weight, and W _k (α) is the statistical weight for a k-mer. The expanded equations for expected populations of 3 or 6 SsoSSBs bound to ssDNA are described by eqs and , respectively.

$${\mathrm{\Xi }}_{NN}^3=1+\mathrm{S}(1)+\mathrm{S}2(1+\alpha )+\mathrm{S}3(1+3\alpha +3\alpha 2+\alpha 3)$$ 20

$${\mathrm{\Xi }}_{NN}^6=1+\mathrm{S}(1)+\mathrm{S}2(1+\alpha )+\mathrm{S}3(1+3\alpha +3\alpha 2+\alpha 3)+\mathrm{S}4(1+6\alpha +15\alpha 2+20\alpha 3+15\alpha 4+6\alpha 5+\alpha 6)+\mathrm{S}5(1+\alpha )10+\mathrm{S}6(1+\alpha )$$ 21

Results

SsoSSB Purifies and Elutes as a Monomer

Differences in the apparent K _d ′ of SsoSSB measured in solution-phase experiments have been reported based on the length of ssDNA, the type of experimental approach and conditions, and the fitting equations used, resulting in a broad reported range from 16 nM to 1 μM. − , Complicating this interpretation, there have been several reports of a range of oligomeric species of SsoSSB in solution that span from monomer to tetramer, with different binding affinities and cooperativities (Hill coefficients of 1–1.8). , It is possible that some of this confusion may be because SsoSSB (16.1 kDa) has a similar MW to EcSSB (18.8 kDa), which is inherently well expressed in the bacterial expression system often utilized to purify SsoSSB. We find that a careful elution gradient of NaCl from the heparin column can effectively separate SsoSSB from endogenous EcSSB (Supplementary Figure S4A,B). Positive SsoSSB fractions from the heparin column were applied to a size exclusion column to determine whether there is any indication of higher oligomeric species (Supplementary Figure S4C). SsoSSB eluted in a single peak just after the 17.6 kDa standard and there was no indication of any higher MW species consistent with dimer or tetramers.

Solution-Phase Binding of SsoSSB to ssDNA Is Equivalent in Tris and NH₄OAc

Solution-phase binding of protein to ssDNA can be easily and conveniently monitored with FA. The change in tumbling of a low concentration of a smaller fluorescently labeled ssDNA molecule upon binding of a protein can be measured directly with a fluorimeter equipped with vertical and horizontal polarizers. Ideally, this works well when there is a single binding event and the fit of the data is a hyperbola, where the 1/2Δr _max can provide the single-site dissociation constant, K _d . However, when there are multiple protein binding events that form a larger macromolecular complex, the fit of the data may not be ideal and requires additional parameters in the equation that may include a Hill coefficient (h) to indicate sigmodal cooperativity or further increases in slope at higher concentrations of protein in the titration that can indicate secondary or nonspecific (NS) binding. When multiple proteins bind along ssDNA in these solution experiments, the FA measurement is dominated by the lowest K _d value, even when multiple binding events occur, providing for an apparent K _d ′. For SSB binding, dimers, trimers, or even more SSB species can be present based on the length of ssDNA. When the binding affinity of one SSB does not depend on another or the position of binding along ssDNA, K _d values are equivalent.

To quantify the apparent K _d ′ of monomeric SsoSSB, we utilized FA with a Cy5-labeled ssDNA (18 mer with a random sequence) in three different solutions (20 mM Tris with either 100 mM or 200 mM NaCl or 200 mM NH₄OAc) to compare solution-phase binding in an MS-compatible solution (Figure ). The 18-mer DNA length can potentially accommodate three SsoSSB molecules based on the binding site size of 5–6 nts. , We employed both an equal sites model (solid line) with some nonspecific binding as well as cooperativity model (dashed line) to fit the data. For the equal sites model, the data was better fit when a nonspecific term was included (eq ) to account for deviations from the fit at higher protein concentrations, possibly from uncharacterized higher order or nonspecific interactions. For the cooperativity fit (eq ), the Hill coefficients (h) were modest but could be consistent with some positive cooperativity. The apparent K _d ′ values were not dissimilar between the fits but without other biophysical or analytical data there was no justification for one model over the other. Increasing NaCl concentrations in a Tris buffer (Table ) modestly increased the K _d ′ values (2.0–2.7 × 10^–7 M) as expected (Figure A,B) and provided a range to compare with binding in NH₄OAc. Therefore, SsoSSB would saturate the substrate based on the available protein, the individual molecular K _d , and the length of ssDNA. SsoSSB bound to the ssDNA substrate similarly in NH₄OAc with a fitted apparent K _d ′ of 2.2 × 10^–7 M (Figure C). Furthermore, SsoSSB-ssDNA binding in 200 mM NH₄OAc mimics that of 20 mM Tris with near physiological salt concentrations and can be utilized for comparison in nMS applications.

2.

Fluorescence anisotropy measurements of SsoSSB binding to ssDNA and determination of apparent K _d ′values. SsoSSB was titrated into a solution containing an 18mer Cy5-ssDNA substrate in (A) 20 mM Tris with 100 mM NaCl, (B) 20 mM Tris with 200 mM NaCl, or (C) 200 mM NH₄OAc. The reported anisotropy values were averaged from at least 8 technical replicates and plotted, where the error bars represent the standard deviation. The apparent K _d ′ values were calculated based on eqs or as indicated in each plot.

nMS Allows for Separation of Oligomeric Protein-ssDNA Species

nMS was utilized to quantify the individual populations of oligomeric SsoSSB complexes bound to different length homooligomeric T15 and T30 ssDNA substrates, as short poly dT DNA strands are devoid of any secondary structures. The concentration of SsoSSB was held constant at 5 μM, and the ssDNA was titrated from 0 to 5 μM to ensure detection of a broad distribution of protein species, both unbound and bound. Even though SsoSSB exists as a monomer (MW = 16,140 ± 10 Da), some minor populations of dimeric SsoSSB (MW = 32,280 ± 10 Da) were observed in the absence of ssDNA (Figure A, two red open circles), confirming the presence of low abundance, higher order species inferred from the nonspecific parameter (eq ) included in the FA experimental fits (Figure ). The most abundant charge states of the apo-proteins were +8 and +12 for the monomer and dimer, respectively.

3.

Representative nESI mass spectra of SsoSSB bound ssDNA. (A) SsoSSB was held constant at 5 μM, and the ssDNA concentration was increased from 0.5 to 5 μM for T15 (B–F) or T30 ssDNA (G–K). Open circles (○) indicate apo-species, while filled circles (●) denote SSB oligomers bound to ssDNA. Monomer, dimer, and trimer species are colored, blue, red, and green, respectively. Charge states are indicated. Insets in the upper right corner of some panels highlight lower-abundance species corresponding to higher-order protein–ssDNA complexes.

Upon addition of ssDNA substrates, T15 (Figure B–F) and T30 (Figure G–K), SsoSSB bound along the length of ssDNA to form higher order complexes in a concentration dependent manor. At lower concentrations of ssDNA (0.5–1.0 μM), there was a larger fractional abundance of higher order complexes (two red or three green solid circles) compared to higher concentrations of ssDNA (2.5–5 μM). As the concentration of either ssDNA substrate increased, complexes of monomeric SsoSSB bound to ssDNA became the most abundant species (Figure F,K). When ssDNA was less concentrated, higher order binding of dimers and trimers of SsoSSB bound to ssDNA were more abundant (Figure C–E,H–J). The quantities of the bound populations to ssDNA were normally distributed around 2 proteins per ligand, represented by the equilibria of the ssDNA concentration as expected, and did not appear to have strongly cooperative binding characteristics at lower DNA concentrations. However, the SsoSSB proteins did not appear to bind ssDNA to saturation, based on a site size of 5–6 nts. Theoretically, T15 could accommodate up to three SsoSSB molecules, however no trimer bound species were observed on T15. T30 could theoretically accommodate up to six SSBs, but only a small proportion of SsoSSB trimers bound to ssDNA were observed without any evidence of four or more SsoSSBs detected on T30 (Supplementary Figure S5). This limitation could be indicative of ESI artifacts such as in-source dissociation, or possibly intermolecular electrostatic repulsion between adjacent ssDNA-bound monomers. During ESI, it is assumed that intermolecular contacts are preserved during desolvation; however, the mobile binding nature of SsoSSB may permit noncanonical behavior, such as sliding off of the ends of the ssDNA substrate or repositioning of subunits along the length of the ssDNA. These phenomena could potentially limit the existence of higher order species observed by imparting negative cooperativity for formation of highest oligomeric species. The titration of SsoSSB onto low concentrations of ssDNA in the FA experiments would mask the presence of any negative cooperativity within the NS term. Instead, the nMS experiments titrate DNA with a constant SsoSSB concentration (5 μM) to better provide a direct measure of the equilibria distributions present, giving more insight into the higher order binding events.

The fractional abundances of each species were quantified from the summed peak intensities or integrated areas of the signal for all observed charge states. SsoSSB bound T15 or T30 in a concentration dependent manner with the monomer-bound species being the most abundant for T15, while T30 could accommodate dimer and some trimer species at the intermediate ssDNA concentrations (Figure A,B). Using peak intensities for fractional abundance calculations can quickly identify and quantify species, but it may also underrepresent larger m/z species that have more significant peak broadening occurring from less efficient desolvation and greater heterogeneity from adduct formation.

4.

Summed fractional abundance on T15 and T30. The summed fractional abundances of SsoSSB species bound to T15 or T30, respectively, were quantified separately using (A,B) intensities or (C,D) areas (gray). Each data point represents a distinct species either unbound (open symbols: ○ monomer or □ dimer) or bound (closed symbols: ● monomer, ■ dimer, or ▲ trimer) to ssDNA. Error bars indicate the standard deviation from two independent experiments each containing three technical replicates.

To more fully capture the abundance of species, especially for larger molecular weight species where peak broadening is more apparent, we also integrated the area of each peak using a custom python program (Supplementary Figure S6) and used these values to quantify the summed fractional abundances. Any overlapping m/z peaks of separate species were first smoothed using boxcar averaging to reduce the noise and enhance the signal for quantification separately and then quantified by fitting to separate Gaussian curves (Supplementary Figure S7). Afterward, the fractional abundances of the summed areas of multiple charge states from each species were plotted as a function of increasing ssDNA concentration (Figure C,D). The fractional distributions calculated from intensities are similar to the plots calculated from areas; however, the larger m/z dimer and trimer populations are in greater abundance using area and better represent their expected solution-phase populations. For example, the quantified bound dimer population at 1.25 μM T30 ssDNA increased from a fractional abundance of 0.38 ± 0.07–0.50 ± 0.05 (Figure B,D, red ■).

Covalent SsoSSB Dimers and Trimers Are Utilized to Correct for RF Differences

To test the hypothesis that larger MW species of SsoSSB dimers and trimers bound to ssDNA have different RFs compared to monomers, we engineered and purified covalent dimer (SsoSSB-BC) and trimer (SsoSSB-ABC) constructs of SsoSSB-WT (Supplementary Figure S8) as RF standards. Generally, RFs decrease with increasing MW, however, they are also affected by the concentration, solution conditions, instrumental parameters, and the intrinsic properties of the macromolecule making direct comparisons of signal intensities with abundance difficult. , The concatemeric SsoSSB proteins were covalently joined without the intervening negatively charged C-termini, except for the terminal subunit which retained its acidic C-terminal tail (Supplementary Figure S1) and was designed to mimic the noncovalently bound dimer and trimer species.

As one approach for RF correction, we included equimolar concentrations of ubiquitin as an internal standard with each of the SsoSSB C, BC, or ABC constructs and monitored the signals in separate nMS experiments (Supplementary Figure 9A–C). We tried other standards, including insulin, but could not maintain stable concentrations during the buffer exchange process due to precipitation of the protein. Consistent with the criteria for appropriate internal standard selection, ubiquitin was stable throughout the course of the experiment, is a globular protein with near neutral isoelectric point (pI = 6.8), and does not form any additional noncovalent complexes with other components in the reaction mixture. The ratios of the intensities or areas of all the charge states for each species were divided by the respective signals for ubiquitin and plotted (Supplementary Figure 9D,E). By establishing the RFs from a series of SsoSSB concatemers normalized to the standard ubiquitin, we attempted to standardize the signal response of different MW SsoSSB-ssDNA bound oligomers; however, unexplainably, the RF data fit better to exponential decay curves rather than simple linear regressions.

For a second RF correction approach, we instead mixed equimolar (5 μM) monomer, dimer, and trimeric SsoSSB species in a single nMS experiment and determined relative RFs to the monomeric SsoSSB species using both measured intensities and integrated areas (Figure ). With this analysis, the RF trend fits well to a linear decrease in signal with increasing mass for these similar species, better allowing for RF correction of abundances for experiments involving monomeric SsoSSB binding to ssDNA. Even so, these RF are not perfect as they do not contain ssDNA and can only provide approximate corrections to calculate actual abundances of noncovalently bound SsoSSB species to ssDNA. The data from Figure was corrected using the RF slopes from Figure to provide for a more accurate quantification of ssDNA-bound fractional abundances of SsoSSB (Figure ). Comparing the analyte abundances from intensities and areas for uncorrected and RF-corrected calculations shows that the larger MW species are underrepresented compared to the RF-corrected values (Supplementary Tables S2 and S3).

5.

ESI response factor testing with a mixture of SSB constructs. (A) Representative ESI native mass spectra of equimolar (5 μM) SsoSSB concatemers each of WT monomer C (blue ○), dimer BC (red □), or trimer ABC (green Δ) were used to determine relative RFs. Both (B) peak intensities and (C) peak areas (gray) were evaluated to estimate RFs. Individual points represent the sample means of three or more independent experiments with three or more technical replications for each capillary with error bars representing the standard error. The slopes and R ² values for the goodness of fit are indicated.

6.

Response factor-corrected summed fractional abundance of SsoSSB on T15 and T30. The RF-corrected and summed fractional abundances of SsoSSB species bound to T15 or T30, respectively, were quantified separately using (A,B) intensities or (C,D) areas (gray). Each data point represents a distinct species either unbound (open symbols: ○ monomer or □ dimer) or bound (closed symbols: ● monomer, ■ dimer, or ▲ trimer) to ssDNA. Error bars indicate the standard deviation from two independent experiments each containing three technical replicates.

Calculation of Dissociation Constants for Each Binding Event

Using the RF-corrected fractional abundances, the actual individual K _d,i s can now be determined using a sequential binding model for equivalent sites (Figure C). Theoretically, K _d,i values can be obtained from the distribution of species at any of the ssDNA concentrations; however, the broadest species distribution occurs at 1.25 μM ssDNA in these nMS experiments, which was chosen for the quantifications. The individual K _d,i values are compared for the uncorrected and RF-corrected abundances quantified from intensities and areas (Table ). Several important insights can be concluded from this analysis. First, K _d,i calculations from area values are generally lower (i.e., tighter binding), especially for K _d,2 and K _d,3 , than those calculated from intensities for both T15 and T30 based on a more complete integration of the signal for a more accurate representation of the equilibria distribution. Second, RF-corrected K _d,i values are also lower than uncorrected K _d,i values, again especially for K _d,2 and K _d,3 , calculated from either intensities or area indicating that an appropriate RF correction strategy is needed for more accurate quantification. Third, K _d,2 values are much larger than for K _d,1 on T15 with α-values greater than one, implying negative cooperativity for the binding of the second saturating SsoSSB. A similar trend is observed on T30, where K _d,3 values are significantly greater than K _d,1 or K _d,2 values with β-values greater than 1 consistent with negative cooperativity. Interestingly, K _d,1 or K _d,2 values on T30 are nearly equivalent, and importantly, these K _d,1 and K _d,2 values (2.7 or 3.2 × 10^–7 M, respectively) approach the apparent K _d ′ values measured from solution in NH₄OAc (2.2 × 10^–7 M) (Figure ), indicating the nMS can recapitulate binding affinities from solution but has the distinct advantage of identifying and quantifying all species with distinct mass from an equilibria population.

2. Calculated Binding Affinities for SsoSSB to 1 μM ssDNA from Nano-ESI MS Quantification.

	intensities					areas
T15	K d,1 (M)	K d,2 (M)	K d,3 (M)	α	β	K d,1 (M)	K d,2 (M)	K d,3 (M)	α	β
uncorrected	1.94 × 10–6	6.61 × 10–5	ND	8.5		1.24 × 10–6	1.56 × 10–5	ND	3.1
RF corrected	2.08 × 10–6	4.37 × 10–5	ND	5.3		1.32 × 10–6	1.07 × 10–5	ND	2.0
T30
uncorrected	1.97 × 10–7	6.00 × 10–7	1.21 × 10–5	0.3	6.8	2.53 × 10–7	4.75 × 10–7	1.77 × 10–6	0.6	0.8
RF corrected	2.09 × 10–7	3.91 × 10–7	7.35 × 10–6	0.6	3.9	2.66 × 10–7	3.20 × 10–7	3.30 × 10–6	0.4	1.4

^aND, not detected.

Based on statistical factors for multiple site binding, K _d,2 should be three times that of K _d,1 when no cooperativity is employed (eq ). The calculation of an α-factor of less than 1 on T30 indicates positive cooperativity is utilized for subsequent binding of the second SsoSSB. It is noteworthy that quantification by area and RF correction generally moderates and reduces the cooperativity factors because of a better quantification of the higher molecular weight species. Interestingly, while the binding of the second SsoSSB on T30 occurs with positive cooperativity (α = 0.4), the binding of the third SsoSSB shows some negative cooperativity (β = 1.4) that is similar to the binding of the second SsoSSB on T15 (α = 2.0). As the binding of SsoSSB molecules do not saturate based on a site size of 5–6 nucleotides and negative cooperativity is observed for binding of the ultimate SsoSSB molecule by nMS, we conclude that SsoSSB assembly occurs with some positive cooperativity when enough ssDNA length is present but limits binding through negative cooperativity at ssDNA termini.

Discussion

Compared to traditional bulk solution-phase binding assays, analogous measurements made using nMS provide more complete information on individual macromolecular abundances, including separation of oligomeric species. The ability to independently analyze every species present in a heterogeneous mixture at once is a distinct advantage for nMS. Herein, we demonstrate the advantages of utilizing nMS techniques to calculate thermodynamic properties of protein binding to ssDNA by directly comparing results with traditional solution-phase binding measurements. Both solution- and gas-phase experiments revealed that SsoSSB exists primarily as a monomer with very limited higher order oligomer formation and binds ssDNA with a modest apparent K _d ′ and no obvious positive cooperativity, settling previous conflicting reports. ,− SSB proteins are generally held at similar high concentrations used for nMS analysis with ∼5 μM tetrameric SSB reported in E. coli and ∼10 μM RPA reported in the human cell nucleus. Importantly for SsoSSB, FA and nMS assays yielded similar K _d values (K _d ′ or K _d,1 ), and these parameters correlate even more closely when a RF correction strategy is employed to account for decreasing signal response of larger MW species, validating that nMS can accurately quantify macromolecular binding parameters found in solution.

Interestingly, a very small fraction (<0.05) of nonspecific SsoSSB dimer at the micromolar concentrations used for nMS was detected, which may correlate with the nonspecific binding detected in the FA assays at higher concentrations. However, we find that this dimeric SsoSSB (off of ssDNA) is not stable in solution with a large, estimated K _d (>10^–5) that we were unable to accurately quantify because of the low abundance. This unbound dimer was not the product of nonspecific binding during ESI desolvation, as the emitter diameters and experimental conditions ensured the presence of <1 protein per droplet was analyzed prior to desolvation. Therefore, only species already present in solution should be observed, instead of these species being artificially generated during the ionization process. This dimer association may be from the interaction of the acidic C-terminal tail from one protein and the basic OB-fold domains from the other, but it is not expected to contribute significantly to regulating ssDNA binding and would exist in a different orientation on ssDNA. Other SSBs have been shown to have protein–protein interactions between analogous domains, including T7 gp2.5 and EcSSB, which do modulate their biochemical functions more significantly. ,

In calculating individual subunit binding K _d,i values from nMS, the use of ion intensity versus area quantifications were compared. While intensities are more readily measured and easily compared, they underrepresent the equilibrium, as successively larger complexes often present broader peak width distributions for lower charge states at greater m/z values, which reduces peak heights. Given that all species of the same charge state are likely similar in tertiary structure and composition, they can all be assumed to have equal contributions to ssDNA binding behaviors and species distributions and should be included in the quantification. However, using both methods of analysis, K _d,i values of similar magnitudes were obtained, differing by only 1–4-fold, suggesting that either of these quantification techniques are relatively robust.

Based on the observed decrease in RFs with increasing masses, RF standards consisting of artificial covalent dimer and trimer species of SsoSSB were utilized. These artificial SsoSSB dimers and trimers mimic the three-dimensional structure, polarity, and surface charge of the noncovalent bound oligomers and serve as useful calibrants for normalization of differential RFs. The SsoSSB concatemers are utilized as a biological mimic to calculate RFs for nearly equivalent noncovalent SsoSSB monomers binding along the length of DNA, which advances on the work from Zenobi who utilized MBP concatemers to investigate nESI signal response at different concentrations to show ion suppression at high concentrations. This study utilizes similar methodologies applied by Klassen who used nMS to examine the effect of Ca²⁺ on multiple molecules of glycan binding to a single lectin protein or from Foster who showed the distribution of ATP binding to the E. coli hexameric termination factor, Rho. Instead, we have expanded on these concepts to more complex systems by examining multiple SsoSSB proteins binding to a single ssDNA substrate to mimic in vivo nucleation and protection of labile ssDNA.

By utilizing RF corrections, generally the K _d,i values for the binding of the first SsoSSB increased while the binding for the subsequent larger molecules decreases, making the values further approach the apparent solution-phase K _d ′. Even though incremental improvements in K _d,i calculations from utilizing peak area and RF-correction made the values become even closer to and within error of the solution-phase measurements, the production of artificial concatemeric species as RF standards and the integration of peak areas are considerably more laborious than utilizing uncorrected peak intensities. Therefore, RF-corrected values should be utilized when precision is critical, rather than in comparative experiments. In some cases, intensity measurements can be superior based on the ease of data analysis for less critical applications, such as quality control, comparison between two different proteins, comparison of mutants or substrates, or in biopharmaceutical testing.

It is surprising and informative that in this work, T15 only accommodated a single SsoSSB monomer and T30 only accommodated two SsoSSB molecules with similar higher affinities. Based on the reported 5–6 nts site size, we expected that up to three or six molecules of SsoSSB could be bound tightly and simultaneously to T15 and T30, respectively; yet, we did not detect more than three SsoSSBs on the longer T30 substrate. It is possible that experimental ESI artifacts including charge repulsion or biophysical disruptions could impart a degree of negative cooperativity that change the one-dimensional distribution of SsoSSBs along ssDNA. Previously, it was shown using single molecule techniques that SsoSSB displays dynamic and independent monomer-to-monomer binding assembly on ssDNA, more similar to the analogous eukaryotic RPA, rather than EcSSB. Our data for binding on to T30 would support this model and add that positive cooperativity would account for this initial assembly so long as the ssDNA is of sufficient length. RPA also displays dynamic binding equilibria that alternates between bound states based on the five ssDNA binding domains within the RPA complex. − Furthermore, it was shown that there is a remarkable decrease in the RPA diffusion coefficient along ssDNA with increasing salt concentration measured on longer (<20 kb) ssDNA substrates. Based on the measured subsaturating distribution of SsoSSBs on DNA of two different lengths and that the initial assembly process utilizes positive cooperativity to saturate internal binding to the ssDNA while using negative cooperativity to limit binding to the ends, we hypothesize that SsoSSB binding is mobile and can be dynamically repositioned to effectively position SsoSSB subunits to maintain its genomic protective function while allowing other enzymes to act on the 3′ or 5′ ends. Whether contributions from the acidic C-terminal tail of SsoSSB or other molecular factors contribute to this repositioning remains to be determined.

Conclusions

Ultimately, the accuracy of measuring K _d,i values using nMS depends on the careful calibration of RFs and the rigorous control of experimental variables. Our work bridges the gap between raw nMS signal and true solution-phase binding equilibria by applying a well-defined, experimentally based correction strategy. The benefits of nMS to tease out individual K _d,i values for successive binding events from equilibria populations present a distinct advantage over bulk-phase techniques. This approach ensures that differences in RFs are adequately addressed, leading to more reliable and reproducible quantitative analyses in the study of protein–DNA interactions and discusses the advantages and limitations of RF correction based on the quantitative rigor required for biological interpretation. Using nMS, we have determined that SsoSSB initially binds to ssDNA with positive cooperativity and moderate affinities and that negative cooperativity limits binding near the ends to effectively coat unprotected regions of the genome.

Glossary

Abbreviations

apo

unbound

D

DNA

Ec

E. coli

EDTA

ethylenediaminetetraacetic acid

ESI

electrospray ionization

ε

extinction coefficient

FA

fluorescence anisotropy

k

exponential parameter

K _d ′

apparent dissociation constant

K _d,i

individual dissociation constant

MBP

maltose binding protein

MW

molecular weight

MWCO

molecular weight cutoff

NH4OAc

ammonium acetate

n

stoichiometries

h

Hill coefficient

nESI

nanoelectrospray ionization

nMS

native mass spectrometry

nts

nucleotides

OB

oligonucleotide binding

P

proteins

RF

response factor

ssDNA

single-stranded DNA

SSB

single-stranded DNA binding protein

Sso

Saccharolobus solfataricus

SsoSSB-ABC

covalent trimer

SsoSSB-BC

covalent dimer

TEV

Tobacco Etch Virus

All biochemical data presented in this study, including experimental values, gels, images, data sets, as well as any strains or plasmids, are available upon request to the corresponding author.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jasms.5c00446.

The authors declare no competing financial interest.