Search

This Perspective reviews tools developed over the past five years in the macromolecular modeling, docking and design software Rosetta.

A method for de novo design of peptide macrocyles called RFpeptides has been developed. RFpeptides is an extension of RoseTTAFold2 and RFdiffusion and combines structure prediction and protein backbone generation for rapid and custom design of macrocyclic peptide binders.

The EMDataResource Ligand Model Challenge aimed at assessing the reliability and reproducibility of modeling ligands bound to protein and protein–nucleic acid complexes in cryo-EM maps determined at near-atomic resolution. This analysis presents the results and recommends best practices for assessing cryo-EM structures of liganded macromolecules.

AfCycDesign: Cyclic offset to the relative positional encoding in AlphaFold2 enables accurate structure prediction, sequence redesign, and de novo hallucination of cyclic peptide monomers and binders.

New detector technology has improved the resolution of cryo-electron microscopy (cryo-EM), but tools for structure determination from high-resolution maps have lagged behind. DiMaio et al. report structure determination from high-resolution cryo-EM maps using a homologous structure as a starting model. Also in this issue, Wang et al. describe a de novo approach for structure determination that does not require a starting model.

New detector technology has improved the resolution of cryo-electron microscopy (cryo-EM), but tools for structure determination from high-resolution maps have lagged behind. Wang et al. describe a de novo approach for structure determination from high-resolution cryo-EM maps. Also in this issue, DiMaio et al. report structure determination using a homologous structure as a starting model.

Computational methods are becoming an increasingly important part of biological research. Using the Rosetta framework as an example, the authors demonstrate how community-driven development of computational methods can be done in a reproducible and reliable fashion.

The authors in this work introduce RosettaVS, an AI-accelerated open-source drug discovery platform. They apply this tool to multi-billion compound libraries, where it was able to identify compounds that bind important targets KLHDC2 and Na_V1.7.

An approach combining the sampling methodology and energy function of Rosetta with the X-ray refinement methodology of Phenix enables improved low-resolution crystallographic refinement.

A multi-laboratory study in the form of a community challenge assesses the quality of models that can be produced from cryo-EM maps using different software tools, the reproducibility of models generated by different users and the performance of metrics used for model validation.

RoseTTAFoldNA extends the RoseTTAFold2 platform to predict the structures of protein–DNA and protein–RNA complexes.

As cryo-EM becomes commonplace in drug discovery, tools for automating small molecule structure determination are needed. Here, authors show a map-guided ligand modeling approach to building ligand structures at resolutions common in cryo-EM.

Engineering the tunability of protein assembly in response to pH changes within a narrow range is challenging. Here the authors report the de novo computational design of pH-responsive protein filaments that exhibit rapid, precise, tunable and reversible assembly and disassembly triggered by small pH changes.

Streptomyces are discovered to produce antibacterial protein complexes that selectively inhibit the hyphal growth of related species, a function distinct from that of the small-molecule antibiotics they are known for.

Fine-tuning the RoseTTAFold structure prediction network on protein structure denoising tasks yields a generative model for protein design that achieves outstanding performance on a wide range of protein structure and function design challenges.

Recently, a pipeline for the design of protein-binding proteins using only the structure of the target protein was reported. Here, the authors report that the incorporation of deep learning methods into the original pipeline increases experimental success rate by ten-fold.

The results of the Fifth RNA-Puzzles contest highlights advances in RNA three-dimensional structure prediction and uncovers new insights into RNA folding and structure.

The online game Foldit invites players to solve problems involving protein structure prediction. Now Foldit players have been recruited to work on a modeling problem and ultimately solve the crystal structure of a retroviral protease that had resisted previous determination.

XMRV is a retrovirus that has been linked to prostate cancer and chronic fatigue syndrome. The crystal structure of the XMRV protease differs from the structures of other retropepsins and instead resembles those of pepsin-like enzymes, suggesting that this protease may represent a distinct evolutionary branch of the aspartic proteases.

Flexible filamentous plant viruses, which cause substantial crop damage worldwide, have eluded structural characterization so far. The cryo-EM structure of BaMV now reveals the virus architecture and the structural basis of its flexibility.

The trRosetta neural network was used to iteratively optimise model proteins from random 100-amino-acid sequences, resulting in ‘hallucinated’ proteins, which when expressed in bacteria closely resembled the model structures.

Roy et al. describe a generalized method for computationally designing miniproteins selective for a single integrin heterodimer and conformational state. The designed αvβ6 inhibitor remains monomeric and maintains biological activity following aerosolization and shows excellent efficacy in bleomycin induced lung fibrosis.

An approach for the design of protein pores is demonstrated by the computational design and subsequent experimental expression of both an ion-selective and a large transmembrane pore.

Sea urchin hyperpolarization-activated cyclic nucleotide-gated (spHCN) ion channels channels are activated by membrane hyperpolarization instead of depolarization and undergo inactivation with hyperpolarization. Here authors apply transition metal ion FRET, patch-clamp fluorometry and Rosetta modeling to measure differences in the structural rearrangements between activation and inactivation of spHCN channels.

Arlt et al. report a structure of yeast seipin, a lipid droplet formation protein, and suggest that it forms a flexible, oligomeric cage in the endoplasmic reticulum membrane, enabling triacylglycerol phase separation, lipid droplet growth and budding toward the cytoplasm.

FliD forms a cap complex at the tip of bacterial flagella and is essential for flagellum filament assembly. Here, the authors present the cryo-EM structure of the Campylobacter jejuni cap complex, revealing a pentameric assembly of FliD and further show that the C. jejuni flagellum filament is 11-stranded.

Structural and functional analyses of BRCA1/BARD1 E3 ligase interactions with the H2A C-terminal tail reveal both the basis of its substrate specificity and the consequences of cancer-associated BRCA1/BARD1 mutations.

Type VI secretion systems (T6SSs) translocate effector proteins into eukaryotic and bacterial recipient cells and are present in many Gram-negative bacteria. Here the authors present the 3.7 Å cryoEM structure of the E.coli T6SS baseplate wedge comprising TssK–TssF–TssG and propose a model for the T6SS baseplate and needle complex.

Bacterial ClpB is a disaggregase that solubilizes protein aggregates. Here the authors present the 2.9 Å cryo-EM structure of a hyperactive variant of ClpB bound to the substrate casein in active translocation states and discuss its polypeptide translocation mechanism.

Regulation of centriole duplication is essential for normal chromosome segregation and the maintenance of genomic stability, and Polo-like kinase 4 (Plk4) is a known regulator with unclear mechanisms. Here the authors show that Plk4 undergoes phase separation into an assembling body through autophosphorylation which drives centriole biogenesis.

Transition metal FRET and Rosetta modeling reveal that the S4 helix in the voltage-sensing ___domain of the HCN channel moves downward and its carboxy-terminal portion tilts during hyperpolarization activation.

The electron cryo-microscopy structure of Sulfolobus islandicus pili enabled the identification of SiL_2606 as the main pilin in these filaments and revealed that the pili are glycosylated, which probably explains how these structures remain soluble and stable even when cells grow at pH 3 and 80 °C.

Cryo-EM structures of MERS-CoV S glycoprotein trimer in complex with different sialosides reveal how the virus engages with sialylated receptors, providing insight into receptor specificity and selectivity.

Computational design incorporating modular buried hydrogen networks produces highly orthogonal protein heterodimers.

Structural insight into TIR-___domain interactions, which are essential for the recruitment of signaling adapters to Toll-like receptors during innate immune responses, demonstrates a conserved interaction mode involved in both TLR and IL-1R signaling.

Cryo-EM and mass spectrometry analyses of the spike glycoprotein trimer from coronavirus HcoV-NL63 reveal an extensive glycan shield that covers the protein surface, including an epitope targeted by neutralizing antibodies against several coronaviruses.

RosettaES, an algorithm that uses a fragment-based sampling strategy, improves macromolecular structure modeling from cryo-EM data at 3–5-Å resolution.

The cryo-electron microscopy structure of the bacteriophage ɸ6 dsRNA genome shows that the genome is packaged in a spooled manner that is more similar to dsDNA viruses than to other dsRNA viruses.

Proteins designed de novo by players of the online protein-folding game Foldit can be expressed in Escherichia coli and adopt the designed structure in solution.

The high-resolution cryo-electron microscopy structure of a pre-fusion coronavirus spike trimer from mouse hepatitis virus is presented; the structure reveals architectural similarities to paramyxovirus F proteins, suggesting that these fusion proteins may have evolved from a distant common ancestor.

Cryo-electron microscopy analysis of yeast Rad26 bound to RNA polymerase II provides insight into the initiation of the transcription-coupled DNA repair mechanism in eukaryotes.

StarMap is a software package that increases the accuracy of macromolecular structures by refining models using state-of-the-art Rosetta algorithms. StarMap’s graphical user interface is fully integrated into UCSF ChimeraX.

The fit of atomic models of protein structures to high-resolution cryo-electron microscopy maps can be assessed with a validation tool, EMRinger.

The X-ray crystal structures of trypanosome and mammalian quiescin sulphydryl oxidase are determined; these structures and follow-up biochemical studies show that large conformational changes occur as the enzyme relays disulphide bonds through its redox-active sites.

The structure of yeast Hrd1 in complex with Hrd3 shows that Hrd1 forms an aqueous cavity with a lateral seal within the endoplasmic reticulum membrane, shedding light on how misfolded proteins are transported out of the endoplasmic reticulum.

Daniel Mann et al. describe a higher-resolution structure of the maintenance of lipid asymmetry inner membrane complex (MlaBDEF) in the Gram-negative pathogen, Acinetobacter baumannii. With this improved structural map, the authors clarify the secondary structure elements of MlaE helices and report on potential lipid dynamics by the MLA system that could inform the development of future therapeutics against A. baumannii infection.