There has been considerable recent progress in designing new proteins using deep-learning methods1–9. Despite this progress, a general deep-learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher-order symmetric architectures, has yet to be described. Diffusion models10,11 have had considerable success in image and language generative modelling but limited success when applied to protein modelling, probably due to the complexity of protein backbone geometry and sequence–structure relationships. Here we show that by fine-tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of designed symmetric assemblies, metal-binding proteins and protein binders. The accuracy of RFdiffusion is confirmed by the cryogenic electron microscopy structure of a designed binder in complex with influenza haemagglutinin that is nearly identical to the design model. In a manner analogous to networks that produce images from user-specified inputs, RFdiffusion enables the design of diverse functional proteins from simple molecular specifications.

Heavy alcohol consumption appears to be the primary driver of changes in the circulating microbiome associated with a shift in its inferred metabolic functions. (Hepatology 2018;67:1284-1302).

There has been considerable recent progress in designing new proteins using deep learning methods. Despite this progress, a general deep learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher order symmetric architectures, has yet to be described. Diffusion models have had considerable success in image and language generative modeling but limited success when applied to protein modeling, likely due to the complexity of protein backbone geometry and sequence-structure relationships. Here we show that by fine tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding, and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold Diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of new designs. In a manner analogous to networks which produce images from user-specified inputs, RFdiffusion enables the design of diverse, complex, functional proteins from simple molecular specifications.

SUMMARY Broadly HIV-1 neutralizing VRC01 class antibodies target the CD4-binding site of Env. They are derived from VH1–2*02 antibody heavy chains paired with rare light chains expressing 5-amino acid-long CDRL3s. They have been isolated from infected subjects but have not yet been elicited by immunization. Env-derived immunogens capable of binding the germline forms of VRC01 B cell receptors on naive B cells have been designed and evaluated in knockin mice. However, the elicited antibodies cannot bypass glycans present on the conserved position N276 of Env, which restricts access to the CD4-binding site. Efforts to guide the appropriate maturation of these antibodies by sequential immunization have not yet been successful. Here, we report on a two-step immunization scheme that leads to the maturation of VRC01-like antibodies capable of accommodating the N276 glycan and displaying autologous tier 2 neutralizing activities. Our results are relevant to clinical trials aiming to elicit VRC01 antibodies.

Summary Recent advances in single-particle cryo-electron microscopy (cryoEM) have resulted in determination of an increasing number of protein structures with resolved glycans. However, existing protocols for the refinement of glycoproteins at low resolution have failed to keep up with these advances. As a result, numerous deposited structures contain glycan stereochemical errors. Here, we describe a Rosetta-based approach for both cryoEM and X-ray crystallography refinement of glycoproteins which is capable of correcting conformational and configurational errors in carbohydrates. Building upon a previous Rosetta framework, we introduced additional features and score terms enabling automatic detection, setup and refinement of glycan-containing structures. We benchmarked this approach using twelve crystal structures and showed that glycan geometries can be automatically improved while maintaining good fit to the crystallographic data. Finally, we used this method to refine carbohydrates of the human coronavirus NL63 spike glycoprotein and of an HIV envelope glycoprotein, demonstrating its usefulness for cryoEM refinement.

Small heat shock proteins (sHSPs) are essential ‘holdase’ chaperones that form large assemblies and respond dynamically to pH and temperature stresses to protect client proteins from aggregation. While the alpha-crystallin domain (ACD) dimer of sHSPs is the universal building block, how the ACD transmits structural changes in response to stress to promote holdase activity is unknown. We found that the dimer interface of HSPB5 is destabilized over physiological pHs and a conserved histidine (His-104) controls interface stability and oligomer structure in response to acidosis. Destabilization by pH or His-104 mutation shifts the ACD from dimer to monomer but also results in a large expansion of HSPB5 oligomer states. Remarkably, His-104 mutant-destabilized oligomers are efficient holdases that reorganize into structurally distinct client–bound complexes. Our data support a model for sHSP function wherein cell stress triggers small perturbations that alter the ACD building blocks to unleash a cryptic mode of chaperone action.DOI: http://dx.doi.org/10.7554/eLife.07304.001

Cyclic nucleotide-gated (CNG) and hyperpolarization-activated cyclic nucleotide-regulated (HCN) ion channels play crucial physiological roles in phototransduction, olfaction, and cardiac pace making. These channels are characterized by the presence of a carboxylterminal cyclic nucleotide-binding domain (CNBD) that connects to the channel pore via a C-linker domain. Although cyclic nucleotide binding has been shown to promote CNG and HCN channel opening, the precise mechanism underlying gating remains poorly understood. Here we used cryoEM to determine the structure of the intact LliK CNG channel isolated from Leptospira licerasiae-which shares sequence similarity to eukaryotic CNG and HCN channels-in the presence of a saturating concentration of cAMP. A short S4-S5 linker connects nearby voltage-sensing and pore domains to produce a non-domain-swapped transmembrane architecture, which appears to be a hallmark of this channel family. We also observe major conformational changes of the LliK C-linkers and CNBDs relative to the crystal structures of isolated C-linker/CNBD fragments and the cryoEM structures of related CNG, HCN, and KCNH channels. The conformation of our LliK structure may represent a functional state of this channel family not captured in previous studies.yclic nucleotide-gated (CNG) and hyperpolarization-activated cyclic nucleotide-regulated (HCN) channels are cationpermeable ion channels regulated by the direct binding of cyclic nucleotides (cAMP or cGMP) (1). CNG channels are present in retinal photoreceptors and olfactory sensory neurons, where they perform chemoelectrical energy conversion in response to light or odor stimuli, respectively. Mutations in CNG channels have been associated with numerous inherited retinal degenerative disorders, achromatopsia, and anosmia (2). HCN channels are found in the cardiac sinoatrial node and throughout the nervous system, where they open in response to membrane hyperpolarization and generate a depolarizing current responsible for rhythmic firing (1). HCN channel mutations and mistrafficking have been associated with several disorders, including sinus bradycardia, epilepsy, and autism (3, 4).CNG and HCN channels possess a cyclic nucleotide-binding domain (CNBD) in their carboxyl-terminal region, and binding of cyclic nucleotide produces a large increase in the open probability of the channel pore. Cyclic nucleotide binding to HCN channels also shifts the voltage dependence of activation to more depolarized potentials, increasing the rate and extent of channel opening (5). CNG and HCN channels are part of a family that includes KCNH potassium channels (Fig. 1A and Fig. S1). KCNH channels, however, are distinct in that they possess a cyclic nucleotide-binding homology domain (CNBHD) occupied by an "intrinsic ligand" and are not directly regulated by cyclic nucleotides (6-8).CNG and HCN channels also harbor a C-linker domain situated between the pore and the CNBD. Based on its position, along with mutagenesis and cross-linking studies, this domain is thought t...

VRC01 broadly neutralizing antibodies (bnAbs) target the CD4-binding site (CD4BS) of the human immunodeficiency virus-1 (HIV-1) envelope glycoprotein (Env). Unlike mature antibodies, corresponding VRC01 germline precursors poorly bind to Env. Immunogen design has mostly relied on glycan removal from trimeric Env constructs and has had limited success in eliciting mature VRC01 bnAbs. To better understand elicitation of such bnAbs, we characterized the inferred germline precursor of VRC01 in complex with a modified trimeric 426c Env by cryo-electron microscopy and a 426c gp120 core by X-ray crystallography, biolayer interferometry, immunoprecipitation, and glycoproteomics. Our results show VRC01 germline antibodies interacted with a wild-type 426c core lacking variable loops 1–3 in the presence and absence of a glycan at position Asn276, with the latter form binding with higher affinity than the former. Interactions in the presence of an Asn276 oligosaccharide could be enhanced upon carbohydrate shortening, which should be considered for immunogen design.