Allostery without Conformational Change: A Native Mass Spectrometry Perspective

Abstract

Native electrospray ionization-mass spectrometry (nESI-MS) enables studies of intact proteins, protein complexes, and protein-ligand complexes. Variable temperature (vT)-nESI-MS, where the temperature of the solution contained in the ESI emitter can be varied from 2 to 100 °C, adds new capabilities for dissecting the thermodynamics for protein-ligand binding. Here, vT-nESI-MS and ion mobility spectrometry (IMS) are used to compare the effects of temperature and nESI buffers on nucleotide (ADP) binding for the GroEL single ring mutant (SR1). Temperature-dependent shifts for average charge states (Z_avg) and rotationally averaged collision cross sections (CCS) for both apo- and nucleotide-bound SR1 complexes (SR1-ADP_n, n = 1-7) indicate that nESI buffers alter structure, stabilities, and dynamics. These studies report nucleotide (ADP) binding affinities (K_a) and insight into cooperativity and enthalpy-entropy compensation (EEC). Specifically, we focus on three commonly used native ESI buffers: ammonium acetate (AmAc), triethylammonium acetate (TEAA), and ethylenediammonium acetate (EDDA). In AmAc solutions, ADP binding is highly cooperative at low temperatures (2-21 °C) but is significantly diminished at higher temperatures (21-31 °C). While cooperative ADP binding is only observed at low temperatures (4 °C) for TEAA solutions, it is absent in EDDA solutions. Collectively, these findings illustrate very different influences of ammonium and alkyl ammonium ions on the SR1 conformation and dynamics as manifested by changes in Z_avg (change of solvent-accessible surface area) and thermodynamics for nucleotide binding. Moreover, temperature-dependent changes in Z_avg and ligand binding provide additional experimental data that support prior work on the effects of hydration on cold protein folding. These results also align with recent computational work for the effects of hydration water on protein binding sites as well as membrane protein complex-lipid binding. The observed temperature-dependent changes in Z_avg, buffer-dependent nucleotide binding, EEC, and changes in heat capacity strongly suggest that ADP influences the conformational states of the SR1 complex. Note, however, that large-scale structural changes in the SR1 complex are not observed in the IMS CCS experiments. Collectively, these results suggest that ADP binding alters key structural and/or dynamic properties of SR1, changes that are not observed in the overall, macroscopic structure of the complex. We suggest that SR1-ADP binding is an archetypal example of "allostery without (measurable) conformational change".

1

(A–E) Effects of solution temperature contained in the ESI emitter on the Z _avg for SR1 and SR1-ADP _n complexes in 200 mM AmAc, EDDA, and TEAA buffers containing 1 mM MgAc₂, 1 μM SR1, and various concentrations of ADP. 20 μM ADP was present in the AmAc solution (A). Owing to low abundances for SR1-ADP complexes at low ADP concentrations, the Z _avg changes for SR1 in TEAA were obtained at (B) 10 μM and (C) 100 μM ADP concentrations; similarly, data for EDDA were obtained at 20 μM (D) and 300 μM (E) ADP concentrations. The averaged data were generated from triplicate measurements. Collision cross sections for SR1 and SR1-ADP _n are shown in (F) with the error bar showing the peak widths.

2

(A–C) Deconvoluted mass spectra of ADP binding products of 1 μM SR1 in 200 mM AmAc, TEAA, and EDDA buffer containing 1 mM MgAc₂ and 20 μM ADP at cold (4 °C), medium (21 °C), and high (31 °C) temperatures. Note that each set of spectra compares the result of heating (4–31 °C) and cooling (31–4 °C) the solution contained in the ESI emitter. (D–F) Mole fraction plots for SR1-ADP _n (n = 0–9) complexes in AmAc, TEAA, and EDDA buffer at low (4 °C), medium (21 °C), and high (31 °C) temperatures at varying ADP concentrations. Note that the x-axis has a larger range for EDDA. Nonspecific ADP binding products SR1-ADP_8–9 are observed at high ADP concentrations. (G–I) The bar charts show intrinsic binding constants (K _a) for individual SR1-ADP binding steps at 4 °C (blue), 21 °C (red), and 31 °C (yellow) in 200 mM solutions of AmAc, TEAA, and EDDA, respectively. All binding constants are generated from triplicated data sets, and error bars are standard deviations of the three replicates.

3

Van’t Hoff fitting plots are shown in (A) 200 mM AmAc, (B) 200 mM TEAA, and (C) 200 mM EDDA. The corresponding entropy, enthalpy, and free-energy values for individual ADP binding steps at 25 °C are shown in bar charts in (D–F), respectively. Plots comparing EECs in (G) 200 mM AmAc, (H) 200 mM TEAA, and (I) 200 mM EDDA. The slope of the fitted line is close to unity; a perfect EEC should have a slope of 1. All values are generated from triplicated data sets and error bars are standard deviations of the three replicates.

4

Effects of EDDA and TEAA on SR1-ADP binding in AmAc buffer with increasing concentrations of EDDA or TEAA at (A) the original pH of individual buffers (AmAc at pH 6.8, EDDA at pH 6.3, and TEAA at pH 7) and (B) pH 7 with 1 mM MgAc₂ and 25 μM ADP at 25 °C. The total buffer concentration is kept at 200 mM. 25 μM ADP was added to SR1, which was prepared in the buffer containing 1 mM MgAc₂ and buffer molecules with the amounts indicated in each panel.