13 Apr 2021

Structure-guided multivalent nanobodies block SARS-CoV-2 infection and suppress mutational escape (Science)

The pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) continues to spread, with devastating consequences. For passive immunization efforts, nanobodies have size and cost advantages over conventional antibodies. In this study, Wu, N. et al. generated four neutralizing nanobodies that target the receptor binding domain of the SARS-CoV-2 spike protein. They used x-ray crystallography and cryo-electron microscopy to define two distinct binding epitopes. On the basis of these structures, they engineered multivalent nanobodies with more than 100 times the neutralizing activity of monovalent nanobodies. Biparatopic nanobody fusions suppressed the emergence of escape mutants. Several nanobody constructs neutralized through receptor binding competition, whereas other monovalent and biparatopic nanobodies triggered aberrant activation of the spike fusion machinery. These premature conformational changes in the spike protein forestalled productive fusion and rendered the virions noninfectious.

13 Apr 2021

Bat and pangolin coronavirus spike glycoprotein structures provide insights into SARS-CoV-2 evolution (Nature Communications)

In recognizing the host cellular receptor and mediating fusion of virus and cell membranes, the spike (S) glycoprotein of coronaviruses is the most critical viral protein for cross-species transmission and infection. Here we determined the cryo-EM structures of the spikes from bat (RaTG13) and pangolin (PCoV_GX) coronaviruses, which are closely related to SARS-CoV-2. All three receptor-binding domains (RBDs) of these two spike trimers are in the "down" conformation, indicating they are more prone to adopt the receptor-binding inactive state. However, Wang, XQ et al. found that the PCoV_GX, but not the RaTG13, spike is comparable to the SARS-CoV-2 spike in binding the human ACE2 receptor and supporting pseudovirus cell entry. They further identified critical residues in the RBD underlying different activities of the RaTG13 and PCoV_GX/SARS-CoV-2 spikes. These results collectively indicate that tight RBD-ACE2 binding and efficient RBD conformational sampling are required for the evolution of SARS-CoV-2 to gain highly efficient infection.

29 Mar 2021

Heterogeneity of Glycan Processing on Trimeric SARS-CoV-2 Spike Protein Revealed by Charge Detection Mass Spectrometry (J. Amer. Chem. Soc.)

The heterogeneity associated with glycosylation of the 66 N-glycan sites on the protein trimer making up the spike (S) region of the SARS-CoV-2 virus has been assessed by charge detection mass spectrometry (CDMS). CDMS allows simultaneous measurement of the mass-to-charge ratio and charge of individual ions, so that mass distributions can be determined for highly heterogeneous proteins such as the heavily glycosylated S protein trimer. The CDMS results are compared to recent glycoproteomics studies of the structure and abundance of glycans at specific sites. Interestingly, average glycan masses determined by “top-down” CDMS measurements are 35–47% larger than those obtained from the “bottom-up” glycoproteomics studies, suggesting that the glycoproteomic measurements underestimated the abundances of larger, more-complex glycans. Moreover, the distribution of glycan masses determined by CDMS is much broader than the distribution expected from the glycoproteomics studies, assuming that glycan processing on each trimer is not correlated. The breadth of the glycan mass distribution therefore indicates heterogeneity in the extent of glycan processing of the S protein trimers, with some trimers being much more heavily processed than others. This heterogeneity may have evolved as a way of further confounding the host’s immune system.

20 Aug 2021

Highly accurate protein structure prediction for the human proteome (Nature)

Protein structures can provide invaluable information, both for reasoning about biological processes and for enabling interventions such as structure-based drug development or targeted mutagenesis. After decades of effort, 17% of the total residues in human protein sequences are covered by an experimentally determined structure1. Here, Kathryn Tunyasuvunakool, Jonas Adler, Demis Hassabis et. al. markedly expand the structural coverage of the proteome by applying the state-of-the-art machine learning method, AlphaFold2, at a scale that covers almost the entire human proteome (98.5% of human proteins). The resulting dataset covers 58% of residues with a confident prediction, of which a subset (36% of all residues) has very high confidence. They introduce several metrics developed by building on the AlphaFold model and use them to interpret the dataset, identifying strong multi-domain predictions as well as regions that are likely to be disordered. Finally, they provide some case studies to illustrate how high-quality predictions could be used to generate biological hypotheses. They are making their predictions freely available to the community and anticipate that routine large-scale and high-accuracy structure prediction will become an important tool that will allow new questions to be addressed from a structural perspective.

20 Aug 2021

Highly accurate protein structure prediction with AlphaFold (Nature)

Proteins are essential to life, and understanding their structure can facilitate a mechanistic understanding of their function. Through an enormous experimental effort, the structures of around 100,000 unique proteins have been determined, but this represents a small fraction of the billions of known protein sequences. Structural coverage is bottlenecked by the months to years of painstaking effort required to determine a single protein structure. Accurate computational approaches are needed to address this gap and to enable large-scale structural bioinformatics. Predicting the three-dimensional structure that a protein will adopt based solely on its amino acid sequence—the structure prediction component of the ‘protein folding problem’—has been an important open research problem for more than 50 years. Despite recent progress, existing methods fall far short of atomic accuracy, especially when no homologous structure is available. Here, John Jumper, Richard Evans, Demis Hassabis et. al.provide the first computational method that can regularly predict protein structures with atomic accuracy even in cases in which no similar structure is known. They validated an entirely redesigned version of our neural network-based model, AlphaFold, in the challenging 14th Critical Assessment of protein Structure Prediction (CASP14), demonstrating accuracy competitive with experimental structures in a majority of cases and greatly outperforming other methods. Underpinning the latest version of AlphaFold is a novel machine learning approach that incorporates physical and biological knowledge about protein structure, leveraging multi-sequence alignments, into the design of the deep learning algorithm.

20 Aug 2021

Artificial intelligence in structural biology is here to stay (Nature – Editorial)

Machine learning will transform our understanding of protein folding. And it’s essential that all data be open. “I didn’t think we would get to this point in my lifetime.” That’s how one research leader in structural biology responded to last week’s publication of research in which artificial intelligence (AI) was used to predict the structure of more than 20,000 human proteins, as well as that of nearly all the known proteins produced by 20 model organisms such as Escherichia coli, fruit flies and yeast, but also soya bean and Asian rice. That is a combined total of around 365,000 predictions.

20 Aug 2021

Huge protein structure database could transform biology (Science - In Depth)

Earlier this month, two groups unveiled the culmination of years of work by computer scientists, biologists, and physicists: advanced modeling programs that can predict the precise 3D atomic structures of proteins. Last week, the biggest payoff of that work arrived. One team used its newly minted artificial intelligence (AI) programs to solve the structures of 350,000 proteins from humans and 20 model organisms, such as Escherichia coli bacteria, yeast, and fruit flies, all mainstays of biological research. In the coming months, the group says it plans to expand its efforts to all catalogued proteins—some 100 million molecules.

20 Aug 2021

AlphaFold2 and the future of structural biology (Nature - Correspondence)

AlphaFold2 is a machine-learning algorithm for protein structure prediction that has now been used to obtain hundreds of thousands of protein models. The resulting resource is marvellous and will serve the community in many ways. Here Patrick Cramer (MPI for Biophysical Chemistry Göttingen) discusses the implications of this breakthrough achievement, which changes the way we do structural biology.

16 Aug 2021

MyD88 TIR domain higher-order assembly interactions revealed by microcrystal electron diffraction and serial femtosecond crystallography (Nature Communications)

MyD88 and MAL are Toll-like receptor (TLR) adaptors that signal to induce pro-inflammatory cytokine production. The authors previously observed that the TIR domain of MAL (MALTIR) forms filaments in vitro and induces formation of crystalline higher-order assemblies of the MyD88 TIR domain (MyD88TIR). These crystals are too small for conventional X-ray crystallography, but are ideally suited to structure determination by microcrystal electron diffraction (MicroED) and serial femtosecond crystallography (SFX). Here, they present MicroED and SFX structures of the MyD88TIR assembly, which reveal a two-stranded higher-order assembly arrangement of TIR domains analogous to that seen previously for MALTIR. They demonstrate via mutagenesis that the MyD88TIR assembly interfaces are critical for TLR4 signalling in vivo, and they show that MAL promotes unidirectional assembly of MyD88TIR. Collectively, their studies provide structural and mechanistic insight into TLR signal transduction and allow a direct comparison of the MicroED and SFX techniques.

16 Aug 2021

Cryo-EM structure of Helicobacter pylori urease with an inhibitor in the active site at 2.0 Å resolution (Nature Communications)

Infection of the human stomach by Helicobacter pylori remains a worldwide problem and greatly contributes to peptic ulcer disease and gastric cancer. Without active intervention approximately 50% of the world population will continue to be infected with this gastric pathogen. Current eradication, called triple therapy, entails a proton-pump inhibitor and two broadband antibiotics, however resistance to either clarithromycin or metronidazole is greater than 25% and rising. Therefore, there is an urgent need for a targeted, high-specificity eradication drug. Gastric infection by H. pylori depends on the expression of a nickel-dependent urease in the cytoplasm of the bacteria. Here, H. Luecke et. al. report the 2.0 Å resolution structure of the 1.1 MDa urease in complex with an inhibitor by cryo-electron microscopy and compare it to a β-mercaptoethanol-inhibited structure at 2.5 Å resolution. The structural information is of sufficient detail to aid in the development of inhibitors with high specificity and affinity.

16 Aug 2021

Cryo-EM structure of the human MT1–Gi signaling complex (Nature Structural and Molecular Biology)

Melatonin receptors (MT1 and MT2) transduce inhibitory signaling by melatonin (N-acetyl-5-methoxytryptamine), which is associated with sleep induction and circadian rhythm modulation. Although recently reported crystal structures of ligand-bound MT1 and MT2elucidated the basis of ligand entry and recognition, the ligand-induced MT1 rearrangement leading to Gi-coupling remains unclear. Here T. Nishizawa, O. Nurek et. al. report a cryo-EM structure of the human MT1–Gi signaling complex at 3.3 Å resolution, revealing melatonin-induced conformational changes propagated to the G-protein-coupling interface during activation. In contrast to other Gi-coupled receptors, MT1 exhibits a large outward movement of TM6, which is considered a specific feature of Gs-coupled receptors. Structural comparison of Gi and Gscomplexes demonstrated conformational diversity of the C-terminal entry of the Gi protein, suggesting loose and variable interactions at the end of the α5 helix of Gi protein. These notions, together with their biochemical and computational analyses, highlight variable binding modes of Gαi and provide the basis for the selectivity of G-protein signaling.