Studying Conformational Changes of Biomolecules with Spatiotemporal Resolution and in cell using Pulsed Dipolar EPR Spectroscopy

Abstract

The function of biomolecules is closely linked with their structure and dynamics. Furthermore, the interconversion of different conformations of a biomolecule determines its activity and is vital for biomolecular processes. Thus, for a thorough understanding of biomolecular function, it is crucial to monitor conformational changes over space and time. Here, the ligand-induced helix movement in a cyclic nucleotide-binding domain (CNBD) was studied with spatiotemporal resolution using pulsed electron-electron double resonance (PELDOR) spectroscopy and microsecond freeze-hyperquenching (MHQ). The PELDOR-derived distance distributions distinguished between the ligand-free <em>apo</em> state and the <em>holo</em> state, which was obtained upon binding of cyclic adenosine monophosphate (cAMP). By freezing a mixture of the CNBD and cAMP in the MHQ device at ageing times between 82 µs and 668 µs, the <em>apo</em>-to-<em>holo</em> transition could be monitored with Angstrom and microsecond spatiotemporal resolution. The PELDOR data revealed a gradual depletion of the <em>apo</em> state population and a simultaneous buildup of the <em>holo</em> state population, but no intermediates between the two states could be detected. This observation suggested that the helix movement occurred on a sub-microsecond time scale and thus could not be monitored by MHQ/PELDOR. Molecular dynamics (MD) simulation confirmed this notion by showing that the helix movement proceeds within a few nanoseconds. Saturation experiments revealed that cAMP binding was accomplished within the dead time of the MHQ device at cAMP-to-CNBD ratios above 67; therefore, ligand binding could be excluded as a potential cause of the population shift, as MHQ/PELDOR experiments were performed with 100-fold excess of cAMP. A mechanism was proposed to interpret these experimental and theoretical results based on dwell times: In a mechanistic picture, the <em>apo</em>-to-<em>holo</em> transition involves two free-energy barriers, ligand binding and the conformational change, which are both crossed within nanoseconds. Upon ligand binding, an <em>apo</em>-ligand-complex is formed in which the protein is structurally in the <em>apo</em> state but with the ligand settled in the binding pocket. The two barriers are separated by a dwell time, in which the <em>apo</em>-ligand-complex acquires thermal energy to cross the second barrier and to transit to the <em>holo</em> state. This dwell time is in the microsecond range and is thus the rate-limiting step on the trajectory from <em>apo</em> to <em>holo</em>. Since the dwell time is individual for each protein molecule and as PELDOR monitors a protein ensemble, MHQ/PELDOR could resolve the dwell-time distribution of the CNBD upon cAMP binding. <br /> The structure of a biomolecule may depend on its environment, i.e. the physiologically active conformation <em>in cell</em> may differ from the one observed <em>in vitro</em>. While X-ray crystallography, nuclear magnetic resonance spectroscopy, and electron microscopy provide a high-resolution structure <em>in vitro</em>, electron paramagnetic resonance (EPR) pulsed dipolar spectroscopy (PDS) allows studying biomolecules <em>in cell</em>. Performing a PDS experiment <em>in cell</em> requires spin labels that are stable in the reductive cellular environment. While the commonly-used nitroxide labels such as MTSL are reduced quickly, (tris)-tetrathiatriarylmethyl (trityl)-based spin labels (TSLs) are stable under cellular conditions. Moreover, they have a narrow EPR spectrum, which allows for highly sensitive single-frequency PDS experiments at nanomolar concentrations. The performance of PDS in combination with TSLs was assessed on a construct of the <em>Yersinia</em> outer protein O (YopO) labelled with Mal-TSL, a label obtained by esterification of the so-called Finland trityl radical. It was found that the double-quantum coherence (DQC) experiment outperforms PELDOR and the single-frequency technique for refocusing dipolar couplings (SIFTER) in terms of the modulation depth and the signal-to-noise ratio. This was attributed to the more efficient phase cycle in DQC, which extracts the dipolar signal by double-quantum filtering. The distance distribution obtained with Mal-TSL was broader than the one obtained with MTSL, which is related to the long and flexible linker in Mal-TSL.

To narrow the distribution and to increase the distance resolution, the ester moiety in Mal-TSL was replaced by a single methylene group, giving rise to the short-linked maleimide trityl label SLIM. Shortening the linker from five rotatable bonds in Mal-TSL to two bonds in SLIM reduced the flexibility and thus the conformational freedom of the label. As a result, SLIM lead to a narrower distance distribution than Mal-TSL. This experimental finding was confirmed by <em>in silico</em> spin labelling, showing that the volume sampled by SLIM is more than twofold lower compared with Mal-TSL. The sensitivity of PDS with trityls was illustrated by a DQC experiment on SLIM-labelled YopO at 90 nM protein concentration. Exploiting the high sensitivity and the stability of SLIM, DQC could be performed upon injecting YopO into oocytes of the African clawed frog (<em>Xenopus laevis</em>). Of note, a conformational change could be observed upon translocating YopO into cells, which illustrates the need to study biomolecular structures in the native environment. The sensitivity was enhanced further by replacing the methyl groups in SLIM with hydroxyethyl groups, giving rise to a trityl label called Ox-SLIM. The absence of methyl substituents in the vicinity of the electron spin increased the phase-memory time and thus allowed for a DQC experiment at only 45 nM protein concentration. Additionally, the hydroxyethyl groups increased the hydrophilicity of the label and thus reduced aggregation and unspecific interactions with the biomolecule. <br /> Beyond distance measurements on biomolecules, two trityl-based model compounds were used to study the electron-spin exchange interaction by continuous wave EPR, DQC, and density functional theory (DFT). Strong antiferromagnetic coupling was observed in a biphenyl-linked trityl biradical; the exchange-coupling constant was determined from the temperature dependence of the half-field signal and the interspin distance distribution was obtained by DQC. Taking spin-density delocalization into account, the experimental distance distribution could be confirmed by MD simulation. Furthermore, DFT was used to study the exchange coupling in a trityl radical connected by a phenyl bridge with copper(II) tetraphenyl porphyrin (TPP). The bridge dynamics were shown to modulate the exchange interaction: In the energy minimum with the phenyl ring almost perpendicular to the planes of TPP and trityl, weak ferromagnetic exchange was observed. Upon rotating the phenyl ring about the connection axis, strong antiferromagnetic coupling was observed if the phenyl ring was in-plane with TPP and the trityl, thus demonstrating that the exchange interaction sensitively depends on the orientation of the phenyl bridge.