A solid-state NMR tool box for the investigation of ATP-fueled protein engines

Highlights

• Solid-state NMR approaches to assign and investigate large proteins and complexes thereof.

• ATP-mimics to study ATP hydrolysis.

• Conformational and dynamic changes during enzymatic reaction cycles monitored by NMR.

• Protein-nucleotide contacts probed by solid-state NMR spectroscopy.

• Fast Magic-Angle spinning and proton-detection to identify protons involved in noncovalent interactions.

Abstract

Motor proteins are involved in a variety of cellular processes. Their main purpose is to convert the chemical energy released during adenosine triphosphate (ATP) hydrolysis into mechanical work. In this review, solid-state Nuclear Magnetic Resonance (NMR) approaches are discussed allowing studies of structures, conformational events and dynamic features of motor proteins during a variety of enzymatic reactions. Solid-state NMR benefits from straightforward sample preparation based on sedimentation of the proteins directly into the Magic-Angle Spinning (MAS) rotor. Protein resonance assignment is the crucial and often time-limiting step in interpreting the wealth of information encoded in the NMR spectra. Herein, potentials, challenges and limitations in resonance assignment for large motor proteins are presented, focussing on both biochemical and spectroscopic approaches. This work highlights NMR tools available to study the action of the motor domain and its coupling to functional processes, as well as to identify protein-nucleotide interactions during events such as DNA replication. Arrested protein states of reaction coordinates such as ATP hydrolysis can be trapped for NMR studies by using stable, non-hydrolysable ATP analogues that mimic the physiological relevant states as accurately as possible. Recent advances in solid-state NMR techniques ranging from Dynamic Nuclear Polarization (DNP), ³¹P-based heteronuclear correlation experiments, ¹H-detected spectra at fast MAS frequencies >100 kHz to paramagnetic NMR are summarized and their applications to the bacterial DnaB helicase from Helicobacter pylori are discussed.

Introduction

ATP-fuelled proteins, among which motor proteins are very important representatives, are involved in a variety of cellular processes in all kingdoms of life, such as cellular transport of vesicles and organelles, genetic encoding, synthesis of proteins in a cell, cell motility and chemotaxis [1]. They convert chemical energy from the hydrolysis of the fuel, a nucleotide triphosphate, to the corresponding diphosphate into mechanical work, e.g. ATP is hydrolysed to ADP and inorganic phosphate [2]. ATP hydrolysis takes place in the motor domains of such protein engines which are composed of structural motifs that are conserved among various motor proteins [3], [4], [5]. The action in the motor protein is then typically coupled to functional processes such as movement of a protein along a substrate [1], [6], [7].

Motor proteins are grouped into cytoskeleton filament motor proteins, nucleic-acid motor proteins and rotary motor proteins [1]. Cytoskeleton motor proteins such as dyneins and kinesines [8] are able to generate forces and movement on microtubules and therefore play a key role in cellular transport processes or cell division [9]. Myosins drive muscle contraction and move along the cytoskeleton protein actin filaments in all eukaryotic cells [10], whereas dyneins and kinesines are responsible for both the transport of cargos along microtubules and dynamic processes in mitosis and meiosis [11]. Kinesines move towards the plus end of a microtubule, whereas dyneins towards the minus end (the minus and plus ends of a microtubule are composed of α-tubulin and β-tubulin, respectively, and are thus structurally and functionally distinct) [8]. Nucleic-acid motor proteins comprise particularly polymerases [12], helicases [13], clamp loaders [14] and topoisomerases [15], which are all involved in DNA replication [16]. The third class of motor proteins contains rotary motor proteins, such as the F₀F₁ ATPase [17], [18], [19], [20] working by a rotary movement of one group of subunits relative to the rest. Another class of ATP-fuelled protein engines contains ATP-binding cassette (ABC) transporters that are involved in catalysing transport reactions across a cellular membrane [21], [22], [23]. Note that ABC transporters are not formally classified as motor proteins, since they do not move along a substrate, but nonetheless they contain a motor domain and also utilize the chemical energy from ATP hydrolysis for functional processes. This review article focuses mainly on DNA helicases as a characteristic example for which solid-state NMR strategies have been described in the last couple of years allowing studies of structure and functioning of such engines.

DNA helicases are classified into six different superfamilies (SF1-6) based on the sequence identity among the conserved helicase motifs [24], [25]: helicases of SF1 and SF2 act as monomers and dimers, while helicases of SF3-SF6 form ring-shaped oligomeric assemblies [24], [25], [26]. DnaB helicases will be taken as examples in this review, in particular that from Helicobacter pylori (Hp) which belongs to the SF4 superfamily and forms (double-) hexameric assemblies [25]. DnaB helicases are essential in DNA replication due to their capability of unwinding double-stranded DNA by coupling ATP-hydrolysis to the movement of the protein along DNA. In case of ring-shaped helicases, strand separation is achieved by encircling one strand of the DNA in the central pore while the second strand is excluded from the helicase during DNA unwinding [13], [27]. DNA helicases are classified as P-loop (phosphate-binding loop) NTPases in which the motor domain adopts a fold called ASCE (Additional Strand Catalytic E) and which are subdivided in the ATPases Associated protein with diverse cellular Activities (AAA+) [28], [29], [30], [31] or RecA superfamilies’ [32] (for a schematic drawing see Fig. 1).

The Walker A and B motifs belong to the conserved part (in terms of their amino acid sequences) of such motor domains [33]. The Walker A motif (phosphate-binding loop, P-loop) is characterized by an amino-acid sequence of the form GXXXXGK[S/T], in which X stands for any amino acid. Lysine and serine/threonine residues typically coordinate to the β- and γ- phosphate groups of ATP as well as a divalent metal ion [5], [28], [33]. The Walker A motif is located between the first strand of the central half-β barrel and the following α-helix. In contrast, the Walker B motif, typically containing the primary sequence X₄DE (X represents a hydrophobic residue), coordinates the positively charged metal ion and is also involved in hydrolysis [28]. It is located next to the third strand of the β-core. Further characteristic motifs of ASCE proteins include a conserved glutamate that activates a water molecule for ATP hydrolysis, a polar residue probably acting as a sensor of the nucleotide state and a conserved arginine (R-finger) that is connected with inter-subunit interactions and believed to act as an activating moiety [5], [13], [34], [35]. Such structural elements are highlighted in Fig. 2 on the low-resolution X-ray structure of the (double-) hexameric bacterial DnaB helicase from Helicobacter pylori [36]. Several models for the ATP hydrolysis mechanism in oligomeric helicases have been proposed spanning the range from a fully stochastic process (each subunit acts independently) to a concerted mechanism (cooperative interactions among the subunits) [13], [25], [37], [38]. The action of the motor domain investigated by solid-state NMR is presented in Section 7.

Typically, the mechanism of ATP-hydrolysis and DNA binding to such protein engines has been studied by X-ray crystallography [39], [40], [41], which can determine conformations of the protein along the enzymatic reaction pathway. X-ray crystallography studies have led to structural insights into helicases for nearly all superfamilies, although with “blind-spots” for SF4 helicases relevant in bacteria (only two single-crystal structures for an SF4 helicase complexed with DNA and an ATP-analogue have been reported [40], [42]) and SF6 helicases [13]. This lack of structural information is partially caused by difficulties in the crystallization of such large protein complexes. This is especially true when the proteins are complexed with DNA, since the flexibility of DNA might hinder the crystallization process. However, in the last years, advances in cryo-Electron Microscopy (cryo-EM) allowed determination of high-resolution structures of oligomeric enzymes, such as proteins of the replisome [43], [44], [45]. As an example, two subnanometer structures of the eukaryotic Cdc45-MCM-GINS helicase bound to a DNA fork have been determined, suggesting an interplay between the motor domain and the N-terminal domain during DNA translocation [43]. A strength of cryo-EM is the possibility of studying large protein complexes of the replisome. In that vein, the sub-4-Å-resolution structures of the E. coli helicase-loader complex (composed of the helicase DnaB and the loader DnaC) in pre- and post-DNA engagement states have been described leading to new insights into the loading of helicases onto DNA [46]. Structures obtained at such high resolution allow identification of protein-DNA contacts and following of their modulations during functional processes [46].

Nuclear Magnetic Resonance (NMR) is a further important player in studying structures and dynamics of proteins. The inherent size of motor proteins limits the use of solution-state NMR, since NMR resonances are strongly broadened due to reduced molecular tumbling for large proteins (life-time broadening effects) [47] and approaches such as methyl-TROSY experiments [48] on selective ¹³CH₃-labeled methyl groups in highly deuterated proteins have to be applied [49], [50], [51], which do not allow the entire protein structure to be studied directly. Solid-state NMR is not affected by this size limitation and does thus not require a priori specific isotope labeling schemes. The technique is based on a rotation of the protein sample, typically filled in a ZrO₂ rotor, around the magic angle (approximately 54.7° with respect to the external magnetic field direction) at rotation frequencies between roughly 10 and 110 kHz [52], [53], [54]), leading to high-resolution Magic-Angle Spinning (MAS) spectra.

Solid-state NMR is well-suited for the investigation of difficult-to-crystallize large biomolecular assemblies, because the proteins can be studied in their sedimented state [55], [56]. Details of the sample preparation protocol for solid-state NMR will be discussed in Section 2. Additionally, solid-state NMR allows proteins to be studied directly in their lipid environment. This has been particularly employed for a subunit of the E. coli F₀F₁ ATPase synthase motor protein [57], [58], [59], as well as for ABC transporters [60], [61], [62], [63]. A further strength of NMR is the possibility of studying dynamic processes which are very often responsible for signal transduction between different proteins and are the basis for many enzymatic reactions [64]. In that vein, NMR approaches not only allow for identification of flexible parts of the protein and molecular motions in particular as a consequence of protein–ligand recognition, but also offer a way to determine the different timescales of the involved dynamic processes. NMR allows characterization of correlation times ranging from ~1 ps to ~1 s [65]. In contrast, the static picture obtained by X-ray crystallography and cryo-EM might yield an incomplete structural description not addressing dynamical aspects involved in enzymatic reactions. Also, NMR has the possibility to perform the experiments at room temperature and in cellular-like environments (e.g. the protein is still hydrated to roughly 50–60% in the sedimented state [56]), whereas for EM cryogenic conditions are required under which the motions involved are no longer present. For example, solid-state NMR allows studies of the structures and dynamic properties of different arrested states along the DNA-binding and ATP-hydrolysis reactions by probing a variety of different ATP-analogues [66]. However, despite the difficulties in sample preparation for cryo-EM (e.g. to orient the proteins properly on the EM grid [67]), hybrid approaches combining cryo-EM with NMR are an extremely powerful tool in protein structure determination [68], [69], [70]. In such approaches, NMR offers a high sensitivity towards structural changes, a simple sample preparation based on sedimentation and the possibility to detect dynamic processes and flexible parts of the protein.

This review aims to present recent progress in solid-state NMR spectroscopy, allowing for the functional characterization of large, ATP-fuelled protein engines. An NMR-spectroscopic handling of such large proteins requires not only advanced and sensitive pulse schemes, but also demands biochemical techniques that enable protein resonance assignment, which is crucial for exploring the entire wealth of information encoded in the NMR data. Section 2 describes sample preparation for solid-state NMR on ATP-fuelled protein engines based on sedimentation in an external ultracentrifuge. This simple sample preparation protocol allows characterization of functionalized protein complexes, in particular those involving nucleotides. A particular challenge of studying large protein engines is posed by NMR resonance assignment, since only assigned residues can be used in the analysis of site-specific conformational and dynamic changes. Several approaches such as “divide-and-conquer”, segmental labeling or selective labeling (or unlabeling) schemes will be discussed in Section 2 and applications to a bacterial DnaB helicase from Helicobacter pylori will be reviewed. From a spectroscopic point-of-view, the benefit of fast MAS and ¹H-detected experiments in the assignment process of large motor proteins is discussed. A combination of NMR spectroscopic and biochemical approaches allowed to assign >50% of the 488 residues of the bacterial DnaB helicase. This in particular allows for tracking conformational and dynamic changes during enzymatic reaction pathways. Chemical-shift perturbations, as a convenient tool to follow such changes, are discussed in Section 3. Section 4 introduces paramagnetic NMR to identify residues within or in the vicinity of nucleotide-binding domains by exploiting paramagnetic relaxation enhancements or pseudo-contact shifts. Section 5 summarizes solid-state NMR techniques to study protein–protein complexes, in which the interaction interface is often of particular interest. Solid-state NMR strategies to determine such interactions in the case of weakly interacting proteins, such as a helicase-primase complex, are described.

The conformational events occurring upon ATP hydrolysis and thus the action of the motor domain in motor proteins are most conveniently studied by an ATP hydrolysis equivalence scheme in which the real states of the ATP hydrolysis are mimicked by ATP-analogues. The procedure is described in Section 6 together with ³¹P MAS NMR techniques, which are a powerful tool to monitor nucleotide and DNA binding to motor proteins. Solid-state NMR benefits from the possibility of detecting such small molecules that are difficult to localize precisely at low resolution by X-ray crystallography or cryo-EM. An ATP hydrolysis equivalence scheme was used for studying conformational and dynamic changes of a bacterial DnaB helicase during ATP hydrolysis and the biologically relevant results are discussed in Section 7.

Solid-state NMR strategies to identify protein-DNA contacts in motor proteins are described in detail in Section 8. Such experiments turn out to be challenging in solids, due to the low efficiencies of NMR polarization-transfer. Herein, heteronuclear correlation experiments in the solid state exploiting the ³¹P nuclei of ATP/DNA are presented. Dynamic Nuclear Polarization (DNP) is one elegant option to overcome sensitivity limitations, and its application to motor proteins will be discussed herein. This section will also focus on the DNA translocation process. The unwinding of DNA by DnaB helicases proceeds by a strand exclusion model in which one strand runs through the central channel of the helicase while the other is excluded from it [71], [72], [73]. However, the detailed mechanism of DNA unwinding is still a topic of discussion, and mechanisms like the Brownian ratchet [7], [74] and power stroke (stepping mechanism) mechanism [75] are described. Molecular dynamic simulations revealed the importance of lysine sidechains in pulling the DNA through the central pore of such helicases [6]. Solid-state NMR strategies to monitor the DNA translocation reaction coordinate will be described in Section 8. NMR chemical-shift values contain direct information about the chemical binding state, e.g. they are extremely sensitive towards ¹H-bonds which are one of the key binding elements in protein-DNA interactions [40], [76], [77]. In that vein, ¹H-detected experiments to probe protein-DNA interactions are presented. Altogether, this review shows how state-of-the art approaches in biomolecular NMR are combined with novel techniques to characterize large and functionalized proteins, discussed using the example of motor proteins.