Systematic computer-aided disulfide design as a general strategy to stabilize prefusion class I fusion proteins.

Numerous enveloped viruses, such as coronaviruses, influenza, and respiratory syncytial virus (RSV), utilize class I fusion proteins for cell entry. During this process, the proteins transition from a prefusion to a postfusion state, undergoing substantial and irreversible conformational changes. The prefusion conformation has repeatedly shown significant potential in vaccine development. However, the instability of this state poses challenges for its practical application in vaccines. While non-native disulfides have been effective in maintaining the prefusion structure, identifying stabilizing disulfide bonds remains an intricate task. Here, we present a general computational approach to systematically identify prefusion-stabilizing disulfides. Our method assesses the geometric constraints of disulfide bonds and introduces a ranking system to estimate their potential in stabilizing the prefusion conformation. We hypothesized that disulfides restricting the initial stages of the conformational switch could offer higher stability to the prefusion state than those preventing unfolding at a later stage. The implementation of our algorithm on the RSV F protein led to the discovery of prefusion-stabilizing disulfides that supported our hypothesis. Furthermore, the evaluation of our top design as a vaccine candidate in a cotton rat model demonstrated robust protection against RSV infection, highlighting the potential of our approach for vaccine development.

Systematic computer-aided disulfide design as a general strategy to stabilize prefusion class I fusion proteins.

Numerous enveloped viruses, such as coronaviruses, influenza, and respiratory syncytial virus (RSV), utilize class I fusion proteins for cell entry. During this process, the proteins transition from a prefusion to a postfusion state, undergoing substantial and irreversible conformational changes. The prefusion conformation has repeatedly shown significant potential in vaccine development. However, the instability of this state poses challenges for its practical application in vaccines. While non-native disulfides have been effective in maintaining the prefusion structure, identifying stabilizing disulfide bonds remains an intricated task. Here, we present a general computational approach to systematically identify prefusion-stabilizing disulfides. Our method assesses the geometric constraints of disulfide bonds and introduces a ranking system to estimate their potential in stabilizing the prefusion conformation. We found, for the RSV F protein, that disulfides restricting the initial stages of the conformational switch can offer higher stability to the prefusion state than those preventing unfolding at a later stage. The implementation of our algorithm on the RSV F protein led to the discovery of prefusion-stabilizing disulfides, providing evidence that supports our hypothesis. Furthermore, the evaluation of our top design as a vaccine candidate in a cotton rat model demonstrated robust protection against RSV infection, highlighting the potential of our approach for vaccine development.

A general computational design strategy for stabilizing viral class I fusion proteins.

Many pathogenic viruses rely on class I fusion proteins to fuse their viral membrane with the host cell membrane. To drive the fusion process, class I fusion proteins undergo an irreversible conformational change from a metastable prefusion state to an energetically more stable postfusion state. Mounting evidence underscores that antibodies targeting the prefusion conformation are the most potent, making it a compelling vaccine candidate. Here, we establish a computational design protocol that stabilizes the prefusion state while destabilizing the postfusion conformation. With this protocol, we stabilize the fusion proteins of the RSV, hMPV, and SARS-CoV-2 viruses, testing fewer than a handful of designs. The solved structures of these designed proteins from all three viruses evidence the atomic accuracy of our approach. Furthermore, the humoral response of the redesigned RSV F protein compares to that of the recently approved vaccine in a mouse model. While the parallel design of two conformations allows the identification of energetically sub-optimal positions for one conformation, our protocol also reveals diverse molecular strategies for stabilization. Given the clinical significance of viruses using class I fusion proteins, our algorithm can substantially contribute to vaccine development by reducing the time and resources needed to optimize these immunogens.

A general computational design strategy for stabilizing viral class I fusion proteins.

Many pathogenic viruses, including influenza virus, Ebola virus, coronaviruses, and Pneumoviruses, rely on class I fusion proteins to fuse viral and cellular membranes. To drive the fusion process, class I fusion proteins undergo an irreversible conformational change from a metastable prefusion state to an energetically more favorable and stable postfusion state. An increasing amount of evidence exists highlighting that antibodies targeting the prefusion conformation are the most potent. However, many mutations have to be evaluated before identifying prefusion-stabilizing substitutions. We therefore established a computational design protocol that stabilizes the prefusion state while destabilizing the postfusion conformation. As a proof of concept, we applied this principle to the fusion protein of the RSV, hMPV, and SARS-CoV-2 viruses. For each protein, we tested less than a handful of designs to identify stable versions. Solved structures of designed proteins from the three different viruses evidenced the atomic accuracy of our approach. Furthermore, the immunological response of the RSV F design compared to a current clinical candidate in a mouse model. While the parallel design of two conformations allows identifying and selectively modifying energetically less optimized positions for one conformation, our protocol also reveals diverse molecular strategies for stabilization. We recaptured many approaches previously introduced manually for the stabilization of viral surface proteins, such as cavity-filling, optimization of polar interactions, as well as postfusion-disruptive strategies. Using our approach, it is possible to focus on the most impacting mutations and potentially preserve the immunogen as closely as possible to its native version. The latter is important as sequence re-design can cause perturbations to B and T cell epitopes. Given the clinical significance of viruses using class I fusion proteins, our algorithm can substantially contribute to vaccine development by reducing the time and resources needed to optimize these immunogens.

Epitopes in the Glycosylphosphatidylinositol Attachment Signal Peptide of Trypanosoma cruzi Mucin Proteins Generate Robust but Delayed and Nonprotective CD8+ T Cell Responses.

Infection with the protozoan parasite Trypanosoma cruzi elicits substantial CD8+ T cell responses that disproportionately target epitopes encoded in the large trans-sialidase (TS) gene family. Within the C57BL/6 infection model, a significant proportion (30-40%) of the T. cruzi-specific CD8+ T cell response targets two immunodominant TS epitopes, TSKb18 and TSKb20. However, both TS-specific CD8+ T cell responses are dispensable for immune control, and TS-based vaccines have no demonstrable impact on parasite persistence, a determinant of disease. Besides TS, the specificity and protective capacity of CD8+ T cells that mediate immune control of T. cruzi infection are unknown. With the goal of identifying alternative CD8+ T cell targets, we designed and screened a representative set of genome-wide, in silico-predicted epitopes. Our screen identified a previously uncharacterized, to our knowledge, T cell epitope MUCKb25, found within mucin family proteins, the third most expanded large gene family in T. cruzi. The MUCKb25-specific response was characterized by delayed kinetics, relative to TS-specific responses, and extensive cross-reactivity with a large number of endogenous epitope variants. Similar to TS-specific responses, the MUCKb25 response was dispensable for control of the infection, and vaccination to generate MUCK-specific CD8+ T cells failed to confer protection. The lack of protection by MUCK vaccination was partly attributed to the fact that MUCKb25-specific T cells exhibit limited recognition of T. cruzi-infected host cells. Overall, these results indicate that the CD8+ T cell compartment in many T. cruzi-infected mice is occupied by cells with minimal apparent effector potential.

Inhibition of a malaria host-pathogen interaction by a computationally designed inhibitor.

Malaria is a substantial global health burden with 229 million cases in 2019 and 450,000 deaths annually. Plasmodium vivax is the most widespread malaria-causing parasite putting 2.5 billion people at risk of infection. P. vivax has a dormant liver stage and therefore can exist for long periods undetected. Its blood-stage can cause severe reactions and hospitalization. Few treatment and detection options are available for this pathogen. A unique characteristic of P. vivax is that it depends on the Duffy antigen/receptor for chemokines (DARC) on the surface of host red blood cells for invasion. P. vivax employs the Duffy binding protein (DBP) to bind to DARC. We first de novo designed a three helical bundle scaffolding database which was screened via protease digestions for stability. Protease-resistant scaffolds highlighted thresholds for stability, which we utilized for selecting DARC mimetics that we subsequentially designed through grafting and redesign of these scaffolds. The optimized design small helical protein disrupts the DBP:DARC interaction. The inhibitor blocks the receptor binding site on DBP and thus forms a strong foundation for a therapeutic that will inhibit reticulocyte infection and prevent the pathogenesis of P. vivax malaria.

Deep learning of protein sequence design of protein-protein interactions.

MOTIVATION: As more data of experimentally determined protein structures are becoming available, data-driven models to describe protein sequence-structure relationships become more feasible. Within this space, the amino acid sequence design of protein-protein interactions is still a rather challenging subproblem with very low success rates-yet, it is central to most biological processes. RESULTS: We developed an attention-based deep learning model inspired by algorithms used for image-caption assignments to design peptides or protein fragment sequences. Our trained model can be applied for the redesign of natural protein interfaces or the designed protein interaction fragments. Here, we validate the potential by recapitulating naturally occurring protein-protein interactions including antibody-antigen complexes. The designed interfaces accurately capture essential native interactions and have comparable native-like binding affinities in silico. Furthermore, our model does not need a precise backbone location, making it an attractive tool for working with de novo design of protein-protein interactions. AVAILABILITY AND IMPLEMENTATION: The source code of the method is available at https://github.com/strauchlab/iNNterfaceDesign. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Sampling of structure and sequence space of small protein folds.

Nature only samples a small fraction of the sequence space that can fold into stable proteins. Furthermore, small structural variations in a single fold, sometimes only a few amino acids, can define a protein's molecular function. Hence, to design proteins with novel functionalities, such as molecular recognition, methods to control and sample shape diversity are necessary. To explore this space, we developed and experimentally validated a computational platform that can design a wide variety of small protein folds while sampling shape diversity. We designed and evaluated stability of about 30,000 de novo protein designs of eight different folds. Among these designs, about 6,200 stable proteins were identified, including some predicted to have a first-of-its-kind minimalized thioredoxin fold. Obtained data revealed protein folding rules for structural features such as helix-connecting loops. Beyond serving as a resource for protein engineering, this massive and diverse dataset also provides training data for machine learning. We developed an accurate classifier to predict the stability of our designed proteins. The methods and the wide range of protein shapes provide a basis for designing new protein functions without compromising stability.

Large-scale design and refinement of stable proteins using sequence-only models.

Engineered proteins generally must possess a stable structure in order to achieve their designed function. Stable designs, however, are astronomically rare within the space of all possible amino acid sequences. As a consequence, many designs must be tested computationally and experimentally in order to find stable ones, which is expensive in terms of time and resources. Here we use a high-throughput, low-fidelity assay to experimentally evaluate the stability of approximately 200,000 novel proteins. These include a wide range of sequence perturbations, providing a baseline for future work in the field. We build a neural network model that predicts protein stability given only sequences of amino acids, and compare its performance to the assayed values. We also report another network model that is able to generate the amino acid sequences of novel stable proteins given requested secondary sequences. Finally, we show that the predictive model-despite weaknesses including a noisy data set-can be used to substantially increase the stability of both expert-designed and model-generated proteins.

Design of protein-binding proteins from the target structure alone.

The design of proteins that bind to a specific site on the surface of a target protein using no information other than the three-dimensional structure of the target remains a challenge1-5. Here we describe a general solution to this problem that starts with a broad exploration of the vast space of possible binding modes to a selected region of a protein surface, and then intensifies the search in the vicinity of the most promising binding modes. We demonstrate the broad applicability of this approach through the de novo design of binding proteins to 12 diverse protein targets with different shapes and surface properties. Biophysical characterization shows that the binders, which are all smaller than 65 amino acids, are hyperstable and, following experimental optimization, bind their targets with nanomolar to picomolar affinities. We succeeded in solving crystal structures of five of the binder-target complexes, and all five closely match the corresponding computational design models. Experimental data on nearly half a million computational designs and hundreds of thousands of point mutants provide detailed feedback on the strengths and limitations of the method and of our current understanding of protein-protein interactions, and should guide improvements of both. Our approach enables the targeted design of binders to sites of interest on a wide variety of proteins for therapeutic and diagnostic applications.

Targeting Viral Surface Proteins through Structure-Based Design.

The emergence of novel viral infections of zoonotic origin and mutations of existing human pathogenic viruses represent a serious concern for public health. It warrants the establishment of better interventions and protective therapies to combat the virus and prevent its spread. Surface glycoproteins catalyzing the fusion of viral particles and host cells have proven to be an excellent target for antivirals as well as vaccines. This review focuses on recent advances for computational structure-based design of antivirals and vaccines targeting viral fusion machinery to control seasonal and emerging respiratory viruses.

De novo design of picomolar SARS-CoV-2 miniprotein inhibitors.

Targeting the interaction between the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike protein and the human angiotensin-converting enzyme 2 (ACE2) receptor is a promising therapeutic strategy. We designed inhibitors using two de novo design approaches. Computer-generated scaffolds were either built around an ACE2 helix that interacts with the spike receptor binding domain (RBD) or docked against the RBD to identify new binding modes, and their amino acid sequences were designed to optimize target binding, folding, and stability. Ten designs bound the RBD, with affinities ranging from 100 picomolar to 10 nanomolar, and blocked SARS-CoV-2 infection of Vero E6 cells with median inhibitory concentration (IC50) values between 24 picomolar and 35 nanomolar. The most potent, with new binding modes, are 56- and 64-residue proteins (IC50 ~ 0.16 nanograms per milliliter). Cryo-electron microscopy structures of these minibinders in complex with the SARS-CoV-2 spike ectodomain trimer with all three RBDs bound are nearly identical to the computational models. These hyperstable minibinders provide starting points for SARS-CoV-2 therapeutics.

Skp1 Dimerization Conceals Its F-Box Protein Binding Site.

Skp1 is an adapter that links F-box proteins to cullin-1 in the Skp1/cullin-1/F-box (SCF) protein family of E3 ubiquitin ligases that targets specific proteins for polyubiquitination and subsequent protein degradation. Skp1 from the amoebozoan Dictyostelium forms a stable homodimer in vitro with a Kd of 2.5 µM as determined by sedimentation velocity studies yet is monomeric in crystal complexes with F-box proteins. To investigate the molecular basis for the difference, we determined the solution NMR structure of a doubly truncated Skp1 homodimer (Skp1ΔΔ). The solution structure of the Skp1ΔΔ dimer reveals a 2-fold symmetry with an interface that buries ∼750 Å2 of predominantly hydrophobic surface. The dimer interface overlaps with subsite 1 of the F-box interaction area, explaining why only the Skp1 monomer binds F-box proteins (FBPs). To confirm the model, Rosetta was used to predict amino acid substitutions that might disrupt the dimer interface, and the F97E substitution was chosen to potentially minimize interference with F-box interactions. A nearly full-length version of Skp1 with this substitution (Skp1ΔF97E) behaved as a stable monomer at concentrations of ≤500 µM and actively bound a model FBP, mammalian Fbs1, which suggests that the dimeric state is not required for Skp1 to carry out a basic biochemical function. Finally, Skp1ΔF97E is expected to serve as a monomer model for high-resolution NMR studies previously hindered by dimerization.

Parallelized identification of on- and off-target protein interactions.

Genetic selection combined with next-generation sequencing enables the simultaneous interrogation of the functionality and stability of large numbers of naturally occurring, engineered, or computationally designed protein variants in parallel. We display hundreds of engineered proteins on the surface of yeast cells, select for binding to a set of target molecules by flow cytometry, and sequence the starting pool as well as selected pools to obtain enrichment values for each displayed protein with each target. We show that this high-throughput workflow of multiplex genetic selections followed by large-scale sequencing and comparative analysis allows not only the determination of relative affinities, but also the monitoring of specificity profiles for hundreds to thousands of protein-protein and protein-small molecule interactions in parallel. The approach not only identifies new interactions of designed proteins, but also detects unintended and undesirable off-target interactions. This provides a general framework for screening of engineered protein binders, which often have no negative selection or design step as part of their development pipelines. Hence, this method will be generally useful in the development of protein-based therapeutics.

An integrated device for the rapid and sensitive detection of the influenza hemagglutinin.

Influenza is a viral respiratory tract infection responsible for up to 5 million cases of severe infection and nearly 600 000 deaths worldwide each year. While treatments for influenza exist, diagnostics for the virus at the point of care are limited in their sensitivity and ability to differentiate between subtypes. We have developed an integrated two-dimensional paper network (2DPN) for the detection of the influenza virus by the surface glycoprotein, hemagglutinin. The hemagglutinin assay was developed using proteins computationally designed to bind with high affinity to the highly-conserved sialic acid binding site. The integrated 2DPN uses a novel geometry that allows automated introduction of an enzymatic amplification reagent directly to the detection zone. This assay was integrated into a prototype device and demonstrated successful detection of clinically relevant virus concentrations spiked into 70 µL of virus-free pediatric nasal swab samples. Using this novel geometry, we found improved assay performance on the device (compared to a manually-operated dipstick method), with a sensitivity of 4.45 × 102 TCID50 per mL on device.

Computational design of trimeric influenza-neutralizing proteins targeting the hemagglutinin receptor binding site.

Many viral surface glycoproteins and cell surface receptors are homo-oligomers, and thus can potentially be targeted by geometrically matched homo-oligomers that engage all subunits simultaneously to attain high avidity and/or lock subunits together. The adaptive immune system cannot generally employ this strategy since the individual antibody binding sites are not arranged with appropriate geometry to simultaneously engage multiple sites in a single target homo-oligomer. We describe a general strategy for the computational design of homo-oligomeric protein assemblies with binding functionality precisely matched to homo-oligomeric target sites. In the first step, a small protein is designed that binds a single site on the target. In the second step, the designed protein is assembled into a homo-oligomer such that the designed binding sites are aligned with the target sites. We use this approach to design high-avidity trimeric proteins that bind influenza A hemagglutinin (HA) at its conserved receptor binding site. The designed trimers can both capture and detect HA in a paper-based diagnostic format, neutralizes influenza in cell culture, and completely protects mice when given as a single dose 24 h before or after challenge with influenza.

Rapid Diagnostic Assay for Intact Influenza Virus Using a High Affinity Hemagglutinin Binding Protein.

Influenza is a ubiquitous and recurring infection that results in approximately 500 000 deaths globally each year. Commercially available rapid diagnostic tests are based upon detection of the influenza nucleoprotein, which are limited in that they are unable to differentiate by species and require an additional viral lysis step. Sample preprocessing can be minimized or eliminated by targeting the intact influenza virus, thereby reducing assay complexity and leveraging the large number of hemagglutinin proteins on the surface of each virus. Here, we report the development of a paper-based influenza assay that targets the hemagglutinin protein; the assay employs a combination of antibodies and novel computationally designed, recombinant affinity proteins as the capture and detection agents. This system leverages the customizability of recombinant protein design to target the conserved receptor-binding pocket of the hemagglutinin protein and to match the trimeric nature of hemagglutinin for improved avidity. Using this assay, we demonstrate the first instance of intact influenza virus detection using a combination of antibody and affinity proteins within a porous network. The recombinant head region binder based assays yield superior analytical sensitivity as compared to the antibody based assay, with lower limits of detection of 3.54 × 107 and 1.34 × 107 CEID50/mL for the mixed and all binder stacks, respectively. Not only does this work describe the development of a novel influenza assay, it also demonstrates the power of recombinant affinity proteins for use in rapid diagnostic assays.

Multiplex pairwise assembly of array-derived DNA oligonucleotides.

While the cost of DNA sequencing has dropped by five orders of magnitude in the past decade, DNA synthesis remains expensive for many applications. Although DNA microarrays have decreased the cost of oligonucleotide synthesis, the use of array-synthesized oligos in practice is limited by short synthesis lengths, high synthesis error rates, low yield and the challenges of assembling long constructs from complex pools. Toward addressing these issues, we developed a protocol for multiplex pairwise assembly of oligos from array-synthesized oligonucleotide pools. To evaluate the method, we attempted to assemble up to 2271 targets ranging in length from 192-252 bases using pairs of array-synthesized oligos. Within sets of complexity ranging from 131-250 targets, we observed error-free assemblies for 90.5% of all targets. When all 2271 targets were assembled in one reaction, we observed error-free constructs for 70.6%. While the assembly method intrinsically increased accuracy to a small degree, we further increased accuracy by using a high throughput 'Dial-Out PCR' protocol, which combines Illumina sequencing with an in-house set of unique PCR tags to selectively amplify perfect assemblies from complex synthetic pools. This approach has broad applicability to DNA assembly and high-throughput functional screens.

Computational design of a pH-sensitive IgG binding protein.

Computational design provides the opportunity to program protein-protein interactions for desired applications. We used de novo protein interface design to generate a pH-dependent Fc domain binding protein that buries immunoglobulin G (IgG) His-433. Using next-generation sequencing of naïve and selected pools of a library of design variants, we generated a molecular footprint of the designed binding surface, confirming the binding mode and guiding further optimization of the balance between affinity and pH sensitivity. In biolayer interferometry experiments, the optimized design binds IgG with a Kd of ∼ 4 nM at pH 8.2, and approximately 500-fold more weakly at pH 5.5. The protein is extremely stable, heat-resistant and highly expressed in bacteria, and allows pH-based control of binding for IgG affinity purification and diagnostic devices.

Community-wide assessment of protein-interface modeling suggests improvements to design methodology.

The CAPRI (Critical Assessment of Predicted Interactions) and CASP (Critical Assessment of protein Structure Prediction) experiments have demonstrated the power of community-wide tests of methodology in assessing the current state of the art and spurring progress in the very challenging areas of protein docking and structure prediction. We sought to bring the power of community-wide experiments to bear on a very challenging protein design problem that provides a complementary but equally fundamental test of current understanding of protein-binding thermodynamics. We have generated a number of designed protein-protein interfaces with very favorable computed binding energies but which do not appear to be formed in experiments, suggesting that there may be important physical chemistry missing in the energy calculations. A total of 28 research groups took up the challenge of determining what is missing: we provided structures of 87 designed complexes and 120 naturally occurring complexes and asked participants to identify energetic contributions and/or structural features that distinguish between the two sets. The community found that electrostatics and solvation terms partially distinguish the designs from the natural complexes, largely due to the nonpolar character of the designed interactions. Beyond this polarity difference, the community found that the designed binding surfaces were, on average, structurally less embedded in the designed monomers, suggesting that backbone conformational rigidity at the designed surface is important for realization of the designed function. These results can be used to improve computational design strategies, but there is still much to be learned; for example, one designed complex, which does form in experiments, was classified by all metrics as a nonbinder.