SARS-CoV-2 proteins structural studies using synchrotron radiation

Abstract

In the process of the development of structural biology, both the size and the complexity of the determined macromolecular structures have grown significantly. As a result, the range of application areas for the results of structural studies of biological macromolecules has expanded. Significant progress in the development of structural biology methods has been largely achieved through the use of synchrotron radiation. Modern sources of synchrotron radiation allow to conduct high-performance structural studies with high temporal and spatial resolution. Thus, modern techniques make it possible to obtain not only static structures, but also to study dynamic processes, which play a key role in understanding biological mechanisms. One of the key directions in the development of structural research is the drug design based on the structures of biomolecules. Synchrotron radiation offers insights into the three-dimensional time-resolved structure of individual viral proteins and their complexes at atomic resolution. The rapid and accurate determination of protein structures is crucial for understanding viral pathogenicity and designing targeted therapeutics. Through the application of experimental techniques, including X-ray crystallography and small-angle X-ray scattering (SAXS), it is possible to elucidate the structural details of SARS-CoV-2 virion containing 4 structural, 16 nonstructural proteins (nsp), and several accessory proteins. The most studied potential targets for vaccines and drugs are the structural spike (S) protein, which is responsible for entering the host cell, as well as nonstructural proteins essential for replication and transcription, such as main protease (M_pro), papain-like protease (PL^pro), and RNA-dependent RNA polymerase (RdRp). This article provides a brief overview of structural analysis techniques, with focus on synchrotron radiation-based methods applied to the analysis of SARS-CoV-2 proteins.

Introduction

Research in the field of structural biology provides results of high significance and is directly related to human health problems. One of the key directions in the development of structural research is the design of drugs based on the structures of biomolecules. The development of new drugs such as human immunodeficiency virus (HIV) protease inhibitors (Weber et al. 2021) and influenza neuraminidase inhibitors (Mahal et al. 2021) is a direct consequence of structural biology research. This trend is expected to continue and become increasingly important in the fight against many re-emerging viral and microbial pathogens that threaten human population.

The structural information obtained from various techniques helps researchers understand the function of different proteins within virus and can provide insights into potential drug targets. SARS-CoV-2 virions are predominantly spherical or ellipsoidal in shape with an average shell diameter of 108 ± 8 nm (Hardenbrook and Zhang 2022). SARS-CoV-2 consists of four main structural proteins, spike glycoprotein (S), envelope glycoprotein (E), membrane glycoprotein (M), and nucleocapsid protein (N), and 16 nonstructural and 9 accessory proteins (Gorkhali et al. 2021) (Fig. 1).

Fig. 1.

SARS-CoV-2 virion structure and genome organization. Abbreviations: ORF is open reading frame; bp is base pair. a A schematic representation of the virion structure: spike, membrane, and envelope proteins incorporated into viral envelope and nucleoprotein-encapsulated RNA. b The genome organization. c Structure of PL.^pro (nsp3) with inhibitor GRL0617 (PDB ID 7CMD (Gao et al. 2021)). d Structure of non-liganded M_pro (PDB ID 6Y2E (Zhang et al. 2020)). e Structure of RdRp complex composed of nsp12, nsp7, and nsp8 (PDB ID 6YYT (Hillen et al. 2020)). f Structure of prefusion S protein with one receptor-binding domain (RBD) up (PDB ID 6VSB (Wrapp et al. 2020)). g N protein monomer structure (PDB ID 8FD5 (Casasanta et al. 2023)). Modified under CC BY 4.0 from Jamison et al. (2022)

Some of these proteins are being actively explored as potential targets for the development of therapeutic agents and vaccines (Table 1).

Table 1.

Most promising targets for vaccine and drug development

Protein name	Function	PDB ID
S protein	Mediates the virus entry into host cells by recognition and binding to the ACE2 receptor (Martínez-Flores et al. (2021); Almehdi et al. (2021))	7ODL, 7R1B, 7V7D, 7UB0, 7TOU, 7V7A, 7LYN, 7KDJ
N protein	Responsible for organizing the viral genome by encapsulating the viral RNA (Bai et al. (2021))	6VYO, 6WKP, 6YUN, 7C22, 7XWX, 7SUE
Mpro (nsp5)	Responsible for the cleavage and release of most nonstructural proteins, accelerates their maturation (Hu et al. (2022))	6Y2E, 6Y2F, 6M2N, 7L0D, 6M2Q, 7JKV, 7JPY, 7JPZ
PLpro (nsp3)	Involved in replication and transcription, maturation of other viral proteins, modulation of host immune response (Osipiuk et al. (2021))	6YWL, 6WEY, 7CZ4, 7CJD, 7LLZ, 7LOS, 7CMD, 7JIW
RdRp (nsp12)	In combination with cofactors nsp7 and nsp8, essential for viral RNA synthesis (Peng et al. (2020))	7BW4, 6M71, 6M71, 7BTF, 7BV1, 7BW4

Knowledge of this proteins’ structures makes it possible to localize functionally significant sites, in relation to which the search for substances capable of blocking the function of the protein is subsequently carried out. There are several common approaches for determining the structures of proteins:

• Cryo-electron microscopy (Cryo-EM). Cryo-EM is an experimental technique that captures high-resolution 3D structures of proteins. This method allows obtaining non-crystalline structures of macromolecules and their complexes with high resolution (Ye et al. 2022; Shi et al. 2023).

• Nuclear magnetic resonance (NMR) spectroscopy. NMR spectroscopy is another method for obtaining protein structures both in solid state and in solution, as well as conformational and interactional dynamics (Hu et al. 2021). There are NMR approaches for intracellular research (Sugiki et al. 2020; Theillet and Luchinat 2022).

• X-ray crystallography. X-ray crystallography is used to determine the structures of molecules in the crystalline state with the highest resolution achieved up to 0.48 Å to date (Schmidt et al. 2011). Modern X-ray sources and methods of macromolecular crystallography make it possible to conduct time-resolved studies at physiologically significant temperatures (Pearson and Mehrabi 2020).

• Small-angle X-ray scattering (SAXS). SAXS is a widely used method for structural studies of proteins, which provides information on macromolecular forms, conformation, packing, aggregation, state, and dynamics of macromolecules in solution (Kursula 2021).

Along with experimental techniques for protein structure determination, computational prediction methods are widely used, such as:

• Homology modeling. This approach relies on known protein structures as templates to predict the structure of a target protein with a similar sequence. It works well when there is a close evolutionary relationship between the target and template proteins (Hameduh et al. 2020). Several tools like SWISS-MODEL and MODELLER can be used for homology modeling (Dong et al. 2020; Wierbowski et al. 2021).

• Ab initio (de novo) modeling. In this method, the protein structure is predicted without relying on known templates. It uses physical principles and energy minimization algorithms (Yousef et al. 2019). Examples of ab initio modeling tools include Rosetta and I-TASSER (Rodríguez et al. 2009; Bhowmick et al. 2022).

• Machine learning-based modeling. This technique is based on the use of machine learning algorithms, primarily deep learning, and known protein structures to predict the structures of target proteins (Jumper et al. 2021; Baek and Baker 2022). Examples of such systems are AlphaFold2 and RoseTTAFold. The emergence of this approach, especially AlphaFold2, which is able to predict the structure of proteins with atomic precision, has a significant impact on research in structural biology and bioinformatics (Pereira et al. 2021; Gomes et al. 2022; Kilim et al. 2023; Yang et al. 2023).

Despite recent advances in the prediction of protein structures, the role of experimental data is still significant due to several factors. Computational methods primarily predict a single structure for a single sequence, while the structural distribution corresponding to a variety of possible conformations is of considerable interest (Yuan et al. 2015; Kim and Porter 2021; Lane 2023). There are also a number of questions related to the structure of disordered regions, macromolecular complexes, post-translational modifications of proteins, etc. (Muhammed and Aki-Yalcin 2019; Buel and Walters 2022; Gomes et al. 2022; Akdel et al. 2022). In addition, modern high-precision predictive methods currently cannot provide a deeper understanding of the protein folding process (Outeiral et al. 2022). Computational and experimental techniques complement each other, which has led to the development of hybrid approaches to the determination of protein structures (Webb et al. 2018; Robertson et al. 2019; Seffernick and Lindert 2020; Terwilliger et al. 2022).

The most widely used among the listed experimental methods is X-ray crystallography. More than 85% known structures in the protein data bank (PDB) were obtained by this method, over 80% of them using synchrotron radiation (Berman 2000). With the use of frozen crystals, crystallographic studies of large multicomponent complexes have become almost routine. The development of X-ray crystallography and SAXS is closely related to synchrotron radiation (SR).

SR is electromagnetic radiation emitted by accelerating electrons moving at speeds close to the speed of light. SR has a high brightness, a wide continuous spectrum, a narrow directivity, and a high degree of coherence and polarization. The use of undulators and wigglers makes it possible to increase the spectral power of the radiation by several orders of magnitude and make it more energetic, to obtain monochromatic radiation, and also to change the polarization characteristics. SR also makes it possible to obtain ultrashort high-power pulses.

The development of synchrotrons began in the 1950s with the appearance of first-generation facilities for high-energy physics research, with synchrotron radiation occurred only as an accompanying phenomenon. The first facilities dedicated to generating synchrotron radiation appeared in the early 1980s (Baumgärtel 1981; Munro 1997). The long way of development has resulted in more than 50 synchrotron radiation sources operating and under construction, including several fourth-generation facilities such as ESRF-EBS in Grenoble, France, and Sirius in São Paulo, Brazil. Another one 4 + generation synchrotron radiation facility “Siberian Circular Photon Source” (SRF SCPS) is currently under construction in Russia (Bukhtiyarov et al. 2022). Synchrotron sources have made it possible to implement projects that represent complex biophysical research (Sanchez-Cano et al. 2021).

X-ray crystallography

The method is based on the interaction of X-ray radiation with the electrons of sample atoms. Due to the fact that X-ray photons weakly interact with atoms, radiation penetrates deep into the sample under study in the absence of significant multiple scattering (Als-Nielsen and McMorrow 2011). X-ray radiation, which interact with the electrons of atoms in a crystal, is scattered into discrete diffraction spots named reflections. Diffraction spots are the result of interference of beams of incoherent SR incident on the crystal, diffracted on different planes of the crystal. In the case of coherent SR, which is the case in free-electron lasers, there is no need to work with crystals (Miyashita and Joti 2017; Bielecki et al. 2020). In the process of obtaining diffraction data, only the positions of the reflections and the intensity corresponding to the squared amplitude of the scattered waves are directly measured. The positions of the reflections in the diffraction pattern correspond to discrete scattering angles that satisfy the Laue conditions and the Bragg law and reflect the structure of the crystal lattice. The intensities and phases of reflections are related to the structural factor F(hkl), where h, k, and l are the indices of the reflecting plane. The structural factor is the Fourier transform of electron density. Therefore, the electron density can be defined as a summation over all positions (hkl) in reciprocal space:

$$\rho \left(x,y,z\right)=\frac{1}{V}\sum _h\sum _k\sum _l|F\left(h,k,l\right)|e^{-2\pi i\left(h,x,+,k,y,+,l,z\right)+i\alpha (hkl)}$$ 1

where ρ(xyz) is the electron density at position (xyz) in the unit cell, V is the crystal unit cell volume, |F(hkl)| is the structure factor amplitude, and α(hkl) in the exponent determines the phase of the diffracting wave. On the other hand, the total intensity I(hkl) measured by the detector, on condition that the crystal rotates with the angular velocity ω in the reflection position, has the form

$$I\left(h,k,l\right)=\frac{{\lambda }^3}{\omega V^2}\times (\frac{e^2}{m{c}^2})_{}^2\times V_{\mathit{cr}}\times I_0\times L\times P\times T_r\times {|F\left(h,k,l\right)|}^2$$ 2

In addition to |F(hkl)|², the complete intensity equation contains terms associated with the scattering by the electrons themselves, with the wavelength of incident radiation λ, with the total crystal volume V_cr, and with the volume V of the crystal unit cell. I₀ is the intensity of incident radiation; I is the integral intensity measured by detector; e and m are the charge and mass of the electron, respectively; c is the speed of light; L is the Lorentz coefficient, depending on the data acquisition technique; the polarization factor P is related to the features of scattering by electrons, which leads to a change in polarization; and Tr is the transmittance coefficient, which is related to the absorption of radiation passing through the sample.

In this case, to calculate the electron density using Eqs. (1) and (2), it is necessary to solve the phase problem: due to the fact that the phase cannot be directly measured experimentally, it must be restored from the available data. If there is a homologous structure in the protein data bank, the most common molecular replacement (MR) method can be used (McCoy 2007; Dodson 2021) including the developing MR approach using in silico models (McCoy et al. 2022; Medina et al. 2022). If no homologous structures are available, additional experimental methods are used. The methods could include single isomorphous replacement (SIR) (Foos et al. 2022) or multiple isomorphous replacement (MIR) (Lawrence et al. 2020), in which diffraction data is collected from a natural crystal as well as a number of derivatives containing heavy atoms. Single-wavelength anomalous dispersion (SAD) (Koelmel et al. 2021; Mou et al. 2022) or multiple-wavelength anomalous dispersion (MAD) (Hendrickson and Ogata 1997; Truong et al. 2021) methods are also widely used, based on wavelength-dependent differences in the contribution of anomalous scattering of certain crystal atoms. In the absence of both homologous structures and additional experimental data, ab initio methods (Morgan et al. 2019; Arnal and Millane 2022; Kingston and Millane 2022) based on the use of natural mathematical conditions can be used. Based on the obtained diffraction pattern and the reconstructed phase, the electron density map of an individual crystal lattice cell can be calculated. The electron density map is used to fit atoms and build a model of the studied molecule. To assess the resulting structure, resolution, the factors R, Rfree, real space R-value (RSR), RSR Z-score (RSRZ), clashscore, and some other values are used (Wang 2015; Gore et al. 2017; Shabalin et al. 2018). R factor shows the model’s ability to explain the experimental data that was used to build the model, and Rfree factor attributes the data that were not used. RSR is a measure of the similarity of the electron density map calculated from the crystallographic model with the electron density map calculated from experimental data. RSRZ allows to determine the proportion of residues that are incompatible with the local electron density. The clashscore is the number of pairs of model atoms that are located too close to each other, normalized per 1000 pairs.

Many proteins retain functional activity in their crystalline state (Schlichting and Goody 1997), and macromolecular crystallography offers opportunity to study them in a variety of media and conditions. At modern synchrotron stations, data acquisition times does not exceed several seconds at crystal rotation speeds greater than 90°s⁻¹, and the frame rate of the diffraction image reaches more than 100 Hz (Cohen 2021). Collection, control, and processing of data occur automatically, which allows to conduct research in a streaming mode. Moreover, today structural research goes beyond the solution of individual static structures. The series of structural images are used to obtain detailed information about the positions and motion of atoms, which determine not only the static positions of atoms, but also the dynamics of studied processes.

Modern X-ray sources and methods of macromolecular crystallography are also used to study structures at physiologically significant temperatures. A significant step in this direction has been the development of multicrystalline strategies (Pearce et al. 2017; Nam 2020), which represent an alternative approach to cryopreservation to overcome the limitations of radiation damage. This approach, originally applied using X-ray free-electron lasers, is based on the fact that a short X-ray pulse creates a stationary diffraction image before significant radiation-induced changes occur inside the crystal (de la Mora et al. 2020; Zhu et al. 2020; Barends et al. 2022). Due to the fact that in this case the crystal is destroyed by irradiation, a number of samples must be exposed in a process called serial crystallography to obtain a complete dataset (Nam 2022). This technology has made it possible to enhance the structural information available from small radiation-sensitive crystals and enable rapid measurements with high temporal resolution at room temperature. Later, this approach also found application in the field of synchrotron macromolecular research. By distributing the X-ray dose over many crystals, the use of serial crystallography techniques at the synchrotron allows structural studies of smaller crystals, including studies of the dynamics of irreversible reactions (Martiel et al. 2019). Data analysis uses approaches originally developed for X-ray free-electron laser experiments (Kirian et al. 2011; Mehrabi et al. 2021). Time-resolved synchrotron macromolecular crystallography makes it possible to obtain structures with atomic resolution with microsecond time resolution using monochromatic X-rays or subnanosecond resolution using polychromatic PinkBeam (Martin-Garcia 2021). PinkBeam is operating mode with a relative spread of photon energy ΔE/E ≈ 5 × 10^–2, with a flux about 100 times greater than that of a monochromatic beam (Meents et al. 2017; Fischer 2021; Kim and Nam 2022). Time-resolved crystallographic studies of macromolecules at room temperature make it possible to visualize functionally significant details and alternative conformations changed by cryopreservation. Maintaining the conformational flexibility at room temperature allows greater access to the ligand binding pockets, providing additional information for studying interactions (Russi et al. 2017; Yun et al. 2018).

X-ray crystallography combined with biophysical, virological, computational, and other methods has found wide application both in large screening studies of the search for potential inhibitors of viral infections and in studies of their mechanisms of action. Zhang et al. (2020) were the first to obtain the X-ray structure of non-liganded SARS-CoV-2 M_pro and its complex with the α-ketoamide inhibitor, which showed promising pharmacokinetic results, and it was proposed as the basis for the development of pyridone-containing inhibitors of anticoronavirus drugs. Another of the early large-scale studies on M_pro was the work of Jin et al. (2020), in which the crystal structure of M_pro was obtained in combination with the Michael acceptor inhibitor (Ekici et al. 2004) N3, which previously showed activity against coronaviruses. A comparative analysis with the M_pros of other known coronaviruses showed a high conservation of the substrate-binding pocket, which led to the conclusion that inhibitors targeting this pocket have a wide spectrum of action against coronaviruses. Subsequently, new putative M_pro inhibitors were developed based on this structure using quantum mechanics/molecular mechanics (QM/MM) simulations (Arafet et al. 2021). Compounds based on boceprevir, telaprevir, and α-acyloxymethylketone and others that can inhibit M_pro with nanomolar IC50 have been developed (Fu et al. 2020; Mody et al. 2021; Baker et al. 2021; Bai et al. 2022). Gold compounds, which not only inhibit SARS-CoV-2 replication by binding to the M_pro catalytic domain, but also suppress the expression of inflammatory cytokines, have also been shown to be a promising therapeutic agents (He et al. 2022). At the same time, detailed studies of the structure of M_pro and the study of the mechanisms that can affect the protein suggest the possibility of mutations that can potentially lead to drug resistance (Lee et al. 2022a). It has also been shown that M_pro structures obtained at room temperature may be physiologically more significant than those obtained by cryocrystallography (Kneller et al. 2020). Along with M_pro, the SARS-CoV-2 S protein is of great interest, because it is responsible for the entrance of the virus into the host cell and plays a key role in transmission from natural hosts to humans. In a study by Shang et al. (2020), researchers determined the crystal structure of the RBD S1 subunit of the S protein in complex with human ACE2, a comparative analysis of which with SARS-CoV RBD showed the structural features of the RBD of the new coronavirus responsible for increasing its affinity for ACE2. In addition, it has been shown that ACE2 may also be a receptor for the bat coronavirus RaTG13 protein, suggesting the possibility of transmission of the virus directly from bats, without an intermediate host, the pangolin, as previously suggested (Lam et al. 2020). Because the S protein is a key target for vaccines and therapeutic antibodies, their effectiveness has decreased against newly emerging variants, especially Omicron subvariants, which contain an unusually large number of mutations in the S protein (Ou et al. 2022; Chakraborty et al. 2022). As a result, there is a need to study the mechanisms of resistance to existing antibodies and the search for conservative protective epitopes of the S protein. One of these epitopes was 3-2A2-4, identified by analyzing the structure of the RBD complex with immunized alpaca nanobodies (Li et al. 2022). Also, based on the crystal structure of the RBD complex of the triple mutant B.1.351 with the P2C-1F11 antibody, the basis of the mechanism of antibody neutralization and escape was proposed (Wang et al. 2021). Many similar studies investigated all key coronavirus proteins involved in the recognition, entry, replication, and exit of the virus (Ionescu 2020; Aatif et al. 2021; Newman et al. 2021; Lee et al. 2022b; Sanders et al. 2022; Ma et al. 2022; Guo et al. 2022; Avelar et al. 2023). In total, more than 3000 structures of SARS-CoV-2 proteins and their associated complexes available in the PDB have been released, with more than 60% of them obtained by X-Ray diffraction (Berman 2000).

Small-angle X-ray scattering

Another widely used method for structural studies of proteins, often in combination with X-ray crystallography, is small-angle x-ray scattering (SAXS), which provides information on macromolecular forms, conformation, packing, aggregation, state, and dynamics of macromolecules in solution, which in turn largely determine the functional properties of proteins. Although SAXS has a much lower resolution (Putnam et al. 2007; Da Vela and Svergun 2020) than X-ray crystallography, it is a powerful tool for the analysis of intrinsically disordered proteins (Gräwert and Svergun 2020). These two methods combined provide enough information to obtain accurate structures for the studying of the mechanisms of the functioning of macromolecules. SAXS is a technique based on the detection of X-rays scattered by electrons of the studied sample. The range of scattering angles, which is the most informative in biological studies, ranges from 0.03° to 5° at a distance of about 1.5 m from the sample (Brosey and Tainer 2019). Azimuthal integration of the recorded intensity at each value of momentum transfer, correction for scattering in the buffer, and extrapolation to infinity dilution gives a one-dimensional X-ray scattering profile averaged over the ensemble. Further, application of using the mathematical apparatus based to the nature of the obtained scattering curve allows to extract the information about the structure and dynamics, sample quality, conformational flexibility, particle size, and density. Examples of such information are the degree of structural order, macromolecule packing data, etc. Additional information can be extracted using special transformations such as Fourier, and Porod-Debye transform (Da Vela and Svergun 2020). Modern approaches such as time-resolved SAXS (Cho et al. 2021) and super-resolution SAXS (Rambo and Tainer 2013) in combination with the molecular dynamics (MD) are tools with great potential in the field of structural and dynamic analysis of conformationally complex systems such as proteins (Boldon et al. 2015; Byer et al. 2023). In the development of the structure- and interaction-based drug design (SIBDD) procedure (Mironov et al. 2022) based on structure-based drug design (SBDD) procedure (Śledź and Caflisch 2018), SAXS is used to assess the stability of complexes and determine the ratio of binding of the developed aptamer and target. SIBDD is the approach proposed for the accelerated design of targeting molecules for newly emerging variants of pathogenic agents. Using the developed procedure, the authors proposed several aptamers that can effectively bind the RBD domain of the SARS-CoV-2 S protein. The data obtained using SAXS enabled to retrieve information about the conformation of the full-length nucleocapsid protein, which cannot be acquired by other methods (Różycki and Boura 2022).

In addition to those presented, other methods for studying protein structures using synchrotron radiation are also used. For example, the Zn-induced interaction of the SARS-CoV-2 accessory open reading frame 7a (orf7a) protein with human bone marrow stromal cell antigen 2 (BST2)/tetherin was studied using X-ray absorption spectroscopy (XAS). This study resulted in the proposal of a mechanism for disrupting the antiviral activity of BST2 by the orf7a protein (Petrosino et al. 2021). However, the application of such methods in protein studies is quite specific and has not yet become frequent practice.

Conclusion

In this review, we briefly discussed the role of synchrotron radiation in structural biology using the example of studies of SARS-CoV-2. Modern sources of synchrotron radiation make it possible to conduct high-performance structural studies with high temporal and spatial resolution, which plays a key role in understanding biological mechanisms. At the same time, X-ray crystallography is the most powerful method of structural biology, which, in combination with SAXS and computational methods, has great potential, both in the study of fundamental biological mechanisms and in the effective development of new drugs.

Author contribution

All authors contributed to the study conception and design. The first draft of the manuscript was written by M.K., and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. A.R. led the work and reviewed the manuscript.

Funding

This study was supported by the Ministry of Science and Higher Education of the Russian Federation (agreement №075-15-2021-1355).

Data availability

Not applicable.

Code availability

Not applicable.

Declarations

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Conflict of interest

The authors declare no competing interests.