Protein Design Research Articles

Modeling membrane geometries implicitly in Rosetta.

Interactions between membrane proteins (MPs) and lipid bilayers are critical for many cellular functions. In the Rosetta molecular modeling suite, the implicit membrane energy function is based on a "slab" model, which represent the membrane as a flat bilayer. However, in nature membranes often have a curvature that is important for function and/or stability. Even more prevalent, in structural biology research MPs are reconstituted in model membrane systems such as micelles, bicelles, nanodiscs, or liposomes. Thus, we have modified the existing membrane energy potentials within the RosettaMP framework to allow users to model MPs in different membrane geometries. We show that these modifications can be utilized in core applications within Rosetta such as structure refinement, protein-protein docking, and protein design. For MP structures found in curved membranes, refining these structures in curved, implicit membranes produces higher quality models with structures closer to experimentally determined structures. For MP systems embedded in multiple membranes, representing both membranes results in more favorable scores compared to only representing one of the membranes. Modeling MPs in geometries mimicking the membrane model system used in structure determination can improve model quality and model discrimination.

Using quantum annealing to design lattice proteins

Quantum annealing has shown promise for finding solutions to difficult optimization problems, including protein folding. Recently, we used the D-Wave Advantage quantum annealer to explore the folding problem in a coarse-grained lattice model, the HP model, in which amino acids are classified into two broad groups: hydrophobic (H) and polar (P). Using a set of 22 HP sequences with up to 64 amino acids, we demonstrated the fast and consistent identification of the correct HP model ground states using the D-Wave hybrid quantum-classical solver. An equally relevant biophysical challenge, called the protein design problem, is the inverse of the above, where the task is to predict protein sequences that fold to a given structure. Here, we approach the design problem by a two-step procedure implemented and executed on a D-Wave machine. In the first step, we perform a pure sequence-space search by varying the type of amino acid at each sequence position, and seek sequences which minimize the HP-model energy of the target structure. After mapping this task onto an Ising spin-glass representation, we employ a hybrid quantum-classical solver to deliver energy-optimal sequences for structures with 30–64 amino acids, with a 100% success rate. In the second step, we filter the optimized sequences from the first step according to their ability to fold to the intended structure. In addition, we try solving the sequence optimization problem using only the quantum processing unit (QPU), which confines us to sizes ≤20, due to exponentially decreasing success rates. To shed light on the pure QPU results, we investigate the effects of control errors caused by an imperfect implementation of the intended Hamiltonian on the QPU, by numerically analyzing the Schrödinger equation. We find that the simulated success rates in the presence of control noise semiquantitatively reproduce the modest pure QPU results for larger chains. Published by the American Physical Society 2024

Computational design of α-amylase from Bacillus licheniformis to increase its activity and stability at high temperatures

The thermostable α-amylase derived from Bacillus licheniformis (BLA) has multiple advantages, including enhancing the mass transfer rate and by reducing microbial contamination in starch hydrolysis. Nonetheless, the application of BLA is constrained by the accessibility and stability of enzymes capable of achieving high conversion rates at elevated temperatures. Moreover, the thermotolerance of BLA requires further enhancement. Here, we developed a computational strategy for constructing small and smart mutant libraries to identify variants with enhanced thermostability. Initially, molecular dynamics (MD) simulations were employed to identify the regions with high flexibility. Subsequently, FoldX, a computational design predictor, was used to design mutants by rigidifying highly flexible residues, whereas the simultaneous decrease in folding free energy assisted in improving thermostability. Through the utilization of MD and FoldX, residues K251, T277, N278, K319, and E336, situated at a distance of 5 Å from the catalytic triad, were chosen for mutation. Seventeen mutants were identified and characterized by evaluating enzymatic characteristics and kinetic parameters. The catalytic efficiency of the E271L/N278K mutant reached 184.1 g L−1 s−1, which is 1.88-fold larger than the corresponding value determined for the WT. Furthermore, the most thermostable mutant, E336S, exhibited a 1.43-fold improvement in half-life at 95 ℃, compared with that of the WT. This study, by combining computational simulation with experimental verification, establishes that potential sites can be computationally predicted to increase the activity and stability of BLA and thus provide a possible strategy by which to guide protein design.

ForceGen: End-to-end de novo protein generation based on nonlinear mechanical unfolding responses using a language diffusion model.

Through evolution, nature has presented a set of remarkable protein materials, including elastins, silks, keratins and collagens with superior mechanical performances that play crucial roles in mechanobiology. However, going beyond natural designs to discover proteins that meet specified mechanical properties remains challenging. Here, we report a generative model that predicts protein designs to meet complex nonlinear mechanical property-design objectives. Our model leverages deep knowledge on protein sequences from a pretrained protein language model and maps mechanical unfolding responses to create proteins. Via full-atom molecular simulations for direct validation, we demonstrate that the designed proteins are de novo, and fulfill the targeted mechanical properties, including unfolding energy and mechanical strength, as well as the detailed unfolding force-separation curves. Our model offers rapid pathways to explore the enormous mechanobiological protein sequence space unconstrained by biological synthesis, using mechanical features as the target to enable the discovery of protein materials with superior mechanical properties.

DIProT: A deep learning based interactive toolkit for efficient and effective Protein design

The protein inverse folding problem, designing amino acid sequences that fold into desired protein structures, is a critical challenge in biological sciences. Despite numerous data-driven and knowledge-driven methods, there remains a need for a user-friendly toolkit that effectively integrates these approaches for in-silico protein design. In this paper, we present DIProT, an interactive protein design toolkit. DIProT leverages a non-autoregressive deep generative model to solve the inverse folding problem, combined with a protein structure prediction model. This integration allows users to incorporate prior knowledge into the design process, evaluate designs in silico, and form a virtual design loop with human feedback. Our inverse folding model demonstrates competitive performance in terms of effectiveness and efficiency on TS50 and CATH4.2 datasets, with promising sequence recovery and inference time. Case studies further illustrate how DIProT can facilitate user-guided protein design.

De novo design of cavity-containing proteins with a backbone-centered neural network energy function

The design of small-molecule-binding proteins requires protein backbones that contain cavities. Previous design efforts were based on naturally occurring cavity-containing backbone architectures. Here, we designed diverse cavity-containing backbones without predefined architectures by introducing tailored restraints into the backbone sampling driven by SCUBA (Side Chain-Unknown Backbone Arrangement), a neural network statistical energy function. For 521 out of 5816 designs, the root-mean-square deviations (RMSDs) of the Cα atoms for the AlphaFold2-predicted structures and our designed structures are within 2.0Å. We experimentally tested 10 designed proteins and determined the crystal structures of two of them. One closely agrees with the designed model, while the other forms a domain-swapped dimer, where the partial structures are in agreement with the designed structures. Our results indicate that data-driven methods such as SCUBA hold great potential for designing de novo proteins with tailored small-molecule-binding function.

Next-generation direct reprogramming.

Tissue repair is significantly compromised in the aging human body resulting in critical disease conditions (such as myocardial infarction or Alzheimer's disease) and imposing a tremendous burden on global health. Reprogramming approaches (partial or direct reprogramming) are considered fruitful in addressing this unmet medical need. However, the efficacy, cellular maturity and specific targeting are still major challenges of direct reprogramming. Here we describe novel approaches in direct reprogramming that address these challenges. Extracellular signaling pathways (Receptor tyrosine kinases, RTK and Receptor Serine/Theronine Kinase, RSTK) and epigenetic marks remain central in rewiring the cellular program to determine the cell fate. We propose that modern protein design technologies (AI-designed minibinders regulating RTKs/RSTK, epigenetic enzymes, or pioneer factors) have potential to solve the aforementioned challenges. An efficient transdifferentiation/direct reprogramming may in the future provide molecular strategies to collectively reduce aging, fibrosis, and degenerative diseases.

Sparks of function by de novo protein design.

Information in proteins flows from sequence to structure to function, with each step causally driven by the preceding one. Protein design is founded on inverting this process: specify a desired function, design a structure executing this function, and find a sequence that folds into this structure. This 'central dogma' underlies nearly all de novo protein-design efforts. Our ability to accomplish these tasks depends on our understanding of protein folding and function and our ability to capture this understanding in computational methods. In recent years, deep learning-derived approaches for efficient and accurate structure modeling and enrichment of successful designs have enabled progression beyond the design of protein structures and towards the design of functional proteins. We examine these advances in the broader context of classical de novo protein design and consider implications for future challenges to come, including fundamental capabilities such as sequence and structure co-design and conformational control considering flexibility, and functional objectives such as antibody and enzyme design.

Machine learning for functional protein design.

Recent breakthroughs in AI coupled with the rapid accumulation of protein sequence and structure data have radically transformed computational protein design. New methods promise to escape the constraints of natural and laboratory evolution, accelerating the generation of proteins for applications in biotechnology and medicine. To make sense of the exploding diversity of machine learning approaches, we introduce a unifying framework that classifies models on the basis of their use of three core data modalities: sequences, structures and functional labels. We discuss the new capabilities and outstanding challenges for the practical design of enzymes, antibodies, vaccines, nanomachines and more. We then highlight trends shaping the future of this field, from large-scale assays to more robust benchmarks, multimodal foundation models, enhanced sampling strategies and laboratory automation.

Pycofitness-Evaluating the fitness landscape of RNA and protein sequences.

The accurate prediction of how mutations change biophysical properties of proteins or RNA is a major goal in computational biology with tremendous impacts on protein design and genetic variant interpretation. Evolutionary approaches such as coevolution can help solving this issue. We present pycofitness, a standalone Python-based software package for the in silico mutagenesis of protein and RNA sequences. It is based on coevolution and, more specifically, on a popular inverse statistical approach, namely direct coupling analysis by pseudo-likelihood maximization. Its efficient implementation and user-friendly command line interface make it an easy-to-use tool even for researchers with no bioinformatics background. To illustrate its strengths, we present three applications in which pycofitness efficiently predicts the deleteriousness of genetic variants and the effect of mutations on protein fitness and thermodynamic stability. https://github.com/KIT-MBS/pycofitness.

Protein design: the experts speak.

Protein design: the experts speak.

Learning the shape of protein microenvironments with a holographic convolutional neural network

Proteins play a central role in biology from immune recognition to brain activity. While major advances in machine learning have improved our ability to predict protein structure from sequence, determining protein function from its sequence or structure remains a major challenge. Here, we introduce holographic convolutional neural network (H-CNN) for proteins, which is a physically motivated machine learning approach to model amino acid preferences in protein structures. H-CNN reflects physical interactions in a protein structure and recapitulates the functional information stored in evolutionary data. H-CNN accurately predicts the impact of mutations on protein stability and binding of protein complexes. Our interpretable computational model for protein structure-function maps could guide design of novel proteins with desired function.

Protein engineering and design.

Protein engineering and design.

Geometry-complete perceptron networks for 3D molecular graphs.

The field of geometric deep learning has recently had a profound impact on several scientific domains such as protein structure prediction and design, leading to methodological advancements within and outside of the realm of traditional machine learning. Within this spirit, in this work, we introduce GCPNet, a new chirality-aware SE(3)-equivariant graph neural network designed for representation learning of 3D biomolecular graphs. We show that GCPNet, unlike previous representation learning methods for 3D biomolecules, is widely applicable to a variety of invariant or equivariant node-level, edge-level, and graph-level tasks on biomolecular structures while being able to (1) learn important chiral properties of 3D molecules and (2) detect external force fields. Across four distinct molecular-geometric tasks, we demonstrate that GCPNet's predictions (1) for protein-ligand binding affinity achieve a statistically significant correlation of 0.608, more than 5%, greater than current state-of-the-art methods; (2) for protein structure ranking achieve statistically significant target-local and dataset-global correlations of 0.616 and 0.871, respectively; (3) for Newtownian many-body systems modeling achieve a task-averaged mean squared error less than 0.01, more than 15% better than current methods; and (4) for molecular chirality recognition achieve a state-of-the-art prediction accuracy of 98.7%, better than any other machine learning method to date. The source code, data, and instructions to train new models or reproduce our results are freely available at https://github.com/BioinfoMachineLearning/GCPNet.

De novo protein design targeting voltage-gated sodium channel 1.7 using RFdiffusion and AlphaFold2

De novo protein design targeting voltage-gated sodium channel 1.7 using RFdiffusion and AlphaFold2

Transfer learning to leverage larger datasets for improved prediction of protein stability changes

Amino acid mutations that lower a protein's thermodynamic stability are implicated in numerous diseases, and engineered proteins with enhanced stability can be important in research and medicine. Computational methods for predicting how mutations perturb protein stability are, therefore, of great interest. Despite recent advancements in protein design using deep learning, in silico prediction of stability changes has remained challenging, in part due to a lack of large, high-quality training datasets for model development. Here, we describe ThermoMPNN, a deep neural network trained to predict stability changes for protein point mutations given an initial structure. In doing so, we demonstrate the utility of a recently released megascale stability dataset for training a robust stability model. We also employ transfer learning to leverage a second, larger dataset by using learned features extracted from ProteinMPNN, a deep neural network trained to predict a protein's amino acid sequence given its three-dimensional structure. We show that our method achieves state-of-the-art performance on established benchmark datasets using a lightweight model architecture that allows for rapid, scalable predictions. Finally, we make ThermoMPNN readily available as a tool for stability prediction and design.

Design of functional intrinsically disordered proteins.

Many proteins do not fold into a fixed three-dimensional structure, but rather function in a highly disordered state. These intrinsically disordered proteins pose a unique challenge to protein engineering and design: How can proteins be designed de novo if not by tailoring their structure? Here, we will review the nascent field of design of intrinsically disordered proteins with focus on applications in biotechnology and medicine. The design goals should not necessarily be the same as for de novo design of folded proteins as disordered proteins have unique functional strengths and limitations. We focus on functions where intrinsically disordered proteins are uniquely suited including disordered linkers, desiccation chaperones, sensors of the chemical environment, delivery of pharmaceuticals, and constituents of biomolecular condensates. Design of functional intrinsically disordered proteins relies on a combination of computational tools and heuristics gleaned from sequence-function studies. There are few cases where intrinsically disordered proteins have made it into industrial applications. However, we argue that disordered proteins can perform many roles currently performed by organic polymers, and that these proteins might be more designable due to their modularity.

DexDesign: an OSPREY-based algorithm for designing de novo D-peptide inhibitors.

With over 270 unique occurrences in the human genome, peptide-recognizing PDZ domains play a central role in modulating polarization, signaling, and trafficking pathways. Mutations in PDZ domains lead to diseases such as cancer and cystic fibrosis, making PDZ domains attractive targets for therapeutic intervention. D-peptide inhibitors offer unique advantages as therapeutics, including increased metabolic stability and low immunogenicity. Here, we introduce DexDesign, a novel OSPREY-based algorithm for computationally designing de novo D-peptide inhibitors. DexDesign leverages three novel techniques that are broadly applicable to computational protein design: the Minimum Flexible Set, K*-based Mutational Scan, and Inverse Alanine Scan. We apply these techniques and DexDesign to generate novel D-peptide inhibitors of two biomedically important PDZ domain targets: CAL and MAST2. We introduce a framework for analyzing de novo peptides-evaluation along a replication/restitution axis-and apply it to the DexDesign-generated D-peptides. Notably, the peptides we generated are predicted to bind their targets tighter than their targets' endogenous ligands, validating the peptides' potential as lead inhibitors. We also provide an implementation of DexDesign in the free and open source computational protein design software OSPREY.

Multi-indicator comparative evaluation for deep learning-based protein sequence design methods.

Proteins found in nature represent only a fraction of the vast space of possible proteins. Protein design presents an opportunity to explore and expand this protein landscape. Within protein design, protein sequence design plays a crucial role, and numerous successful methods have been developed. Notably, deep learning-based protein sequence design methods have experienced significant advancements in recent years. However, a comprehensive and systematic comparison and evaluation of these methods have been lacking, with indicators provided by different methods often inconsistent or lacking effectiveness. To address this gap, we have designed a diverse set of indicators that cover several important aspects, including sequence recovery, diversity, root-mean-square deviation of protein structure, secondary structure, and the distribution of polar and nonpolar amino acids. In our evaluation, we have employed an improved weighted inferiority-superiority distance method to comprehensively assess the performance of eight widely used deep learning-based protein sequence design methods. Our evaluation not only provides rankings of these methods but also offers optimization suggestions by analyzing the strengths and weaknesses of each method. Furthermore, we have developed a method to select the best temperature parameter and proposed solutions for the common issue of designing sequences with consecutive repetitive amino acids, which is often encountered in protein design methods. These findings can greatly assist users in selecting suitable protein sequence design methods. Overall, our work contributes to the field of protein sequence design by providing a comprehensive evaluation system and optimization suggestions for different methods.

Energy, water, and protein folding: A molecular dynamics-based quantitative inventory of molecular interactions and forces that make proteins stable.

Protein folding energetics can be determined experimentally on a case-by-case basis but it is not understood in sufficient detail to provide deep control in protein design. The fundamentals of protein stability have been outlined by calorimetry, protein engineering, and biophysical modeling, but these approaches still face great difficulty in elucidating the specific contributions of the intervening molecules and physical interactions. Recently, we have shown that the enthalpy and heat capacity changes associated to the protein folding reaction can be calculated within experimental error using molecular dynamics simulations of native protein structures and their corresponding unfolded ensembles. Analyzing in depth molecular dynamics simulations of four model proteins (CI2, barnase, SNase, and apoflavodoxin), we dissect here the energy contributions to ΔH (a key component of protein stability) made by the molecular players (polypeptide and solvent molecules) and physical interactions (electrostatic, van der Waals, and bonded) involved. Although the proteins analyzed differ in length, isoelectric point and fold class, their folding energetics is governed by the same quantitative pattern. Relative to the unfolded ensemble, the native conformations are enthalpically stabilized by comparable contributions from protein-protein and solvent-solvent interactions, and almost equally destabilized by interactions between protein and solvent molecules. The native protein surface seems to interact better with water than the unfolded one, but this is outweighed by the unfolded surface being larger. From the perspective of physical interactions, the native conformations are stabilized by van de Waals and Coulomb interactions and destabilized by conformational strain arising from bonded interactions. Also common to the four proteins, the sign of the heat capacity change is set by interactions between protein and solvent molecules or, from the alternative perspective, by Coulomb interactions.