Preferential Interactions of a Crowder Protein with the Specific Binding Site of a Native Protein Complex

Abstract

Nonspecific binding of crowder proteins with functional proteins is likely prevalent in vivo, yet direct quantitative evidence, let alone residue-specific information, is scarce. Here we present nuclear magnetic resonance (NMR) characterization showing that bovine serum albumin weakly but preferentially interacts with the histidine carrier protein (HPr). Notably, the binding interface overlaps with that for HPr's specific partner protein, EIN, leading to competition. The crowder protein thus decreases the EIN-HPr binding affinity and accelerates the dissociation of the native complex. In contrast, Ficoll-70 stabilizes the native complex and slows its dissociation, as one would expect from excluded-volume and microviscosity effects. Our atomistic modeling of macromolecular crowding rationalizes the experimental data and provides quantitative insights into the energetics of protein-crowder interactions. The integrated NMR and modeling study yields benchmarks for the effects of crowded cellular environments on protein-protein specific interactions, with implications for evolution regarding how nonspecific binding can be minimized or exploited.

Figure 1.

Global fitting of HPr CSPs as a function of EIN concentration. CSPs of 100 μM ¹⁵N-labeled HPr were obtained in (A) the buffer (20 mM Tris•HCl at pH 7.4 with 150 mM NaCl), (B) 10% w/v Ficoll-70, or (C) 10% w/v BSA. The residues included in the global fits and the resulting K_D values are shown in the legend.

Figure 2.

Off-resonance R_1ρ rates measured for the Q51 amide protons of 300 μM HPr in complex with equimolar EIN at increasing spin-lock field strengths. The experiments were performed in (A) the buffer, (B) 10% w/v Ficoll-70, or (C) 10% w/v BSA, at two magnetic fields (red, 950 MHz; black, 600 MHz). The k_ex value from a global fit of the two relaxation dispersion curves is shown in the legend. The error bar represents the standard error from the R_1ρ exponential fitting at each spin-lock field strength.

Figure 3.

Overlap between the specific binding site of EIN and the preferential interaction site of BSA on HPr. (A) The native complex between EIN and HPr (PDB entry 3EZA). Left: EIN is represented by its molecular surface colored according to electrostatic potential (red: positive; blue: negative), and HPr is shown in cartoon representation in cyan. Right: HPr, rotated 180° to expose the EIN-binding site, is rendered by its electrostatic potential. (B) The EIN-binding site on HPr, with a hydrophobic core (side chains shown as sticks with carbon atoms in green) and a positively charged rim (side chains shown as sticks with carbon atoms in cyan). Note that F48 was assigned a new rotamer such that this side chain retracts back to EIN. (C) A representative complex between BSA and HPr with significant nonspecific interactions. This complex was selected by applying the FMAP method (cf. Figure 5A), whereby an HPr molecule was inserted into many positions in a BSA solution and the interaction energy of the HPr molecule at each position with the BSA molecules was calculated. The selected complex has the most favorable interaction between the HPr molecule and a single BSA molecule, and is displayed in a view where the HPr molecule has an orientation similar to that in (A). HPr is presented as a cartoon on the left and an electrostatic potential (with 180° rotation) on the right. BSA is presented as electrostatic potential. (D) Residues in free HPr that make the largest contributions to the interaction energy with BSA (c.f. Figure 5B).

Figure 4.

Intermolecular PRE values for backbone ¹H^N protons of 300 μM [²H, ¹⁵N]-labeled EIN, mixed with equimolar unlabeled HPr conjugated with EDTA-Mn²⁺ at E5C (top), E25C (middle), or E66C (bottom). The PRE data were collected in (A) the buffer, or (B) 10% w/v Ficoll-70, or (C) 10% w/v BSA. The error bar of the PRE value corresponds to the standard deviation from the experimental measurements of ¹H^N Γ₂ value using three incremental delays.

Figure 5.

Interactions between the protein crowder BSA and the partner proteins EIN and HPr. (A) Scheme illustrating how the crowding effect of BSA on the binding free energy of EIN and HPr is accounted for. The binding free energies in the dilute and crowded conditions are ΔG_b0 and ΔG_b, respectively. Instead of calculating ΔG_b0 and ΔG_b (horizontal paths) to find the difference $\Delta \Delta G_{\text{c}}^{\text{b}}=\Delta G_{\text{b}}-\Delta G_{\text{b}0}$, in the FMAP method, one calculates the transfer free energies of the two proteins in the free and bound states, denoted as Δμ_A + Δμ_B and Δμ_C, respectively, from the dilute to the crowded solution (vertical paths). Because the horizontal and vertical paths close a thermodynamic cycle, one has $\Delta \Delta G_{\text{c}}^{\text{b}}=\Delta {\mu }_{\text{C}}-(\Delta {\mu }_{\text{A}}+\Delta {\mu }_{\text{B}})$. The right panel displays the full view of the EIN-HPr native complex in a box of BSA crowders. (B) Residue-level decomposition of the interaction energies of HPr and EIN with BSA, in either the free or bound state.

Figure 6.

NMR probe of the soft interactions between the protein crowder BSA and free HPr. (A) Longitudinal relaxation rates R₁ (1^st row), transverse relaxation rates R₂ (2^nd row), the ratio between R₂ and R₁ (3^rd row), and the product of R₁ and R₂ rates (4^th row), measured for the amide nitrogens of the ¹⁵N-labeled free HPr in three solution conditions. The rigid limit of R₁ × R₂, with ϖ_N × τ_c ≫ 1 and S² = 0.85, is shown as a gray line. Error bars represent the standard errors of the fits of the peak intensities at different delay times (1^st and 2^nd rows) and propagated errors (3^rd and 4^th rows). (B) Overlay of the NMR HSQC spectra of [²H,¹⁵N]-labeled HPr in the buffer (black), or in the presence of increasing concentrations of BSA. (C) The CSPs obtained at 30% w/v BSA, mapped to the structure of HPr in the native complex with EIN (PDB code 3EZA, colored red), with key interacting residues denoted. (D) The CSPs for six HPr residues are globally fit to a single-site binding isotherm, yielding the BSA-HPr dissociation constant shown in the legend.