Rapid and automated design of two-component protein nanomaterials using ProteinMPNN.

The design of protein-protein interfaces using physics-based design methods such as Rosetta requires substantial computational resources and manual refinement by expert structural biologists. Deep learning methods promise to simplify protein-protein interface design and enable its application to a wide variety of problems by researchers from various scientific disciplines. Here, we test the ability of a deep learning method for protein sequence design, ProteinMPNN, to design two-component tetrahedral protein nanomaterials and benchmark its performance against Rosetta. ProteinMPNN had a similar success rate to Rosetta, yielding 13 new experimentally confirmed assemblies, but required orders of magnitude less computation and no manual refinement. The interfaces designed by ProteinMPNN were substantially more polar than those designed by Rosetta, which facilitated in vitro assembly of the designed nanomaterials from independently purified components. Crystal structures of several of the assemblies confirmed the accuracy of the design method at high resolution. Our results showcase the potential of deep learning-based methods to unlock the widespread application of designed protein-protein interfaces and self-assembling protein nanomaterials in biotechnology.

Accurate computational design of three-dimensional protein crystals.

Protein crystallization plays a central role in structural biology. Despite this, the process of crystallization remains poorly understood and highly empirical, with crystal contacts, lattice packing arrangements and space group preferences being largely unpredictable. Programming protein crystallization through precisely engineered side-chain-side-chain interactions across protein-protein interfaces is an outstanding challenge. Here we develop a general computational approach for designing three-dimensional protein crystals with prespecified lattice architectures at atomic accuracy that hierarchically constrains the overall number of degrees of freedom of the system. We design three pairs of oligomers that can be individually purified, and upon mixing, spontaneously self-assemble into >100 µm three-dimensional crystals. The structures of these crystals are nearly identical to the computational design models, closely corresponding in both overall architecture and the specific protein-protein interactions. The dimensions of the crystal unit cell can be systematically redesigned while retaining the space group symmetry and overall architecture, and the crystals are extremely porous and highly stable. Our approach enables the computational design of protein crystals with high accuracy, and the designed protein crystals, which have both structural and assembly information encoded in their primary sequences, provide a powerful platform for biological materials engineering.

Rapid and automated design of two-component protein nanomaterials using ProteinMPNN.

The design of novel protein-protein interfaces using physics-based design methods such as Rosetta requires substantial computational resources and manual refinement by expert structural biologists. A new generation of deep learning methods promises to simplify protein-protein interface design and enable its application to a wide variety of problems by researchers from various scientific disciplines. Here we test the ability of a deep learning method for protein sequence design, ProteinMPNN, to design two-component tetrahedral protein nanomaterials and benchmark its performance against Rosetta. ProteinMPNN had a similar success rate to Rosetta, yielding 13 new experimentally confirmed assemblies, but required orders of magnitude less computation and no manual refinement. The interfaces designed by ProteinMPNN were substantially more polar than those designed by Rosetta, which facilitated in vitro assembly of the designed nanomaterials from independently purified components. Crystal structures of several of the assemblies confirmed the accuracy of the design method at high resolution. Our results showcase the potential of deep learning-based methods to unlock the widespread application of designed protein-protein interfaces and self-assembling protein nanomaterials in biotechnology.

Hierarchical design of pseudosymmetric protein nanoparticles.

Discrete protein assemblies ranging from hundreds of kilodaltons to hundreds of megadaltons in size are a ubiquitous feature of biological systems and perform highly specialized functions 1-3. Despite remarkable recent progress in accurately designing new self-assembling proteins, the size and complexity of these assemblies has been limited by a reliance on strict symmetry 4,5. Inspired by the pseudosymmetry observed in bacterial microcompartments and viral capsids, we developed a hierarchical computational method for designing large pseudosymmetric self-assembling protein nanomaterials. We computationally designed pseudosymmetric heterooligomeric components and used them to create discrete, cage-like protein assemblies with icosahedral symmetry containing 240, 540, and 960 subunits. At 49, 71, and 96 nm diameter, these nanoparticles are the largest bounded computationally designed protein assemblies generated to date. More broadly, by moving beyond strict symmetry, our work represents an important step towards the accurate design of arbitrary self-assembling nanoscale protein objects.

Hierarchical design of pseudosymmetric protein nanoparticles.

Discrete protein assemblies ranging from hundreds of kilodaltons to hundreds of megadaltons in size are a ubiquitous feature of biological systems and perform highly specialized functions1-3. Despite remarkable recent progress in accurately designing new self-assembling proteins, the size and complexity of these assemblies has been limited by a reliance on strict symmetry4,5. Inspired by the pseudosymmetry observed in bacterial microcompartments and viral capsids, we developed a hierarchical computational method for designing large pseudosymmetric self-assembling protein nanomaterials. We computationally designed pseudosymmetric heterooligomeric components and used them to create discrete, cage-like protein assemblies with icosahedral symmetry containing 240, 540, and 960 subunits. At 49, 71, and 96 nm diameter, these nanoparticles are the largest bounded computationally designed protein assemblies generated to date. More broadly, by moving beyond strict symmetry, our work represents an important step towards the accurate design of arbitrary self-assembling nanoscale protein objects.

Fast and versatile sequence-independent protein docking for nanomaterials design using RPXDock.

Computationally designed multi-subunit assemblies have shown considerable promise for a variety of applications, including a new generation of potent vaccines. One of the major routes to such materials is rigid body sequence-independent docking of cyclic oligomers into architectures with point group or lattice symmetries. Current methods for docking and designing such assemblies are tailored to specific classes of symmetry and are difficult to modify for novel applications. Here we describe RPXDock, a fast, flexible, and modular software package for sequence-independent rigid-body protein docking across a wide range of symmetric architectures that is easily customizable for further development. RPXDock uses an efficient hierarchical search and a residue-pair transform (RPX) scoring method to rapidly search through multidimensional docking space. We describe the structure of the software, provide practical guidelines for its use, and describe the available functionalities including a variety of score functions and filtering tools that can be used to guide and refine docking results towards desired configurations.

Computational design of constitutively active cGAS.

Cyclic GMP-AMP synthase (cGAS) is a pattern recognition receptor critical for the innate immune response to intracellular pathogens, DNA damage, tumorigenesis and senescence. Binding to double-stranded DNA (dsDNA) induces conformational changes in cGAS that activate the enzyme to produce 2'-3' cyclic GMP-AMP (cGAMP), a second messenger that initiates a potent interferon (IFN) response through its receptor, STING. Here, we combined two-state computational design with informatics-guided design to create constitutively active, dsDNA ligand-independent cGAS (CA-cGAS). We identified CA-cGAS mutants with IFN-stimulating activity approaching that of dsDNA-stimulated wild-type cGAS. DNA-independent adoption of the active conformation was directly confirmed by X-ray crystallography. In vivo expression of CA-cGAS in tumor cells resulted in STING-dependent tumor regression, demonstrating that the designed proteins have therapeutically relevant biological activity. Our work provides a general framework for stabilizing active conformations of enzymes and provides CA-cGAS variants that could be useful as genetically encoded adjuvants and tools for understanding inflammatory diseases.

De novo design of self-assembling helical protein filaments.

We describe a general computational approach to designing self-assembling helical filaments from monomeric proteins and use this approach to design proteins that assemble into micrometer-scale filaments with a wide range of geometries in vivo and in vitro. Cryo-electron microscopy structures of six designs are close to the computational design models. The filament building blocks are idealized repeat proteins, and thus the diameter of the filaments can be systematically tuned by varying the number of repeat units. The assembly and disassembly of the filaments can be controlled by engineered anchor and capping units built from monomers lacking one of the interaction surfaces. The ability to generate dynamic, highly ordered structures that span micrometers from protein monomers opens up possibilities for the fabrication of new multiscale metamaterials.

Development of a thermostable nanoemulsion adjuvanted vaccine against tuberculosis using a design-of-experiments approach.

BACKGROUND: Adjuvants have the potential to increase the efficacy of protein-based vaccines but need to be maintained within specific temperature and storage conditions. Lyophilization can be used to increase the thermostability of protein pharmaceuticals; however, no marketed vaccine that contains an adjuvant is currently lyophilized, and lyophilization of oil-in-water nanoemulsion adjuvants presents a specific challenge. We have previously demonstrated the feasibility of lyophilizing a candidate adjuvanted protein vaccine against Mycobacterium tuberculosis (Mtb), ID93 + GLA-SE, and the subsequent improvement of thermostability; however, further development is required to prevent physicochemical changes and degradation of the TLR4 agonist glucopyranosyl lipid adjuvant formulated in an oil-in-water nanoemulsion (SE). MATERIALS AND METHODS: In this study, we took a systematic approach to the development of a thermostable product by first identifying compatible solution conditions and stabilizing excipients for both antigen and adjuvant. Next, we applied a design-of-experiments approach to identify stable lyophilized drug product formulations. RESULTS: We identified specific formulations that contain disaccharide or a combination of disaccharide and mannitol that can achieve substantially improved thermostability and maintain immunogenicity in a mouse model when tested in accelerated and real-time stability studies. CONCLUSION: These efforts will aid in the development of a platform formulation for use with other similar vaccines.

Lyophilization of an Adjuvanted Mycobacterium tuberculosis Vaccine in a Single-Chamber Pharmaceutical Cartridge.

Although substantial effort has been made in the development of next-generation recombinant vaccine systems, maintenance of a cold chain is still typically required and remains a critical challenge in effective vaccine distribution. The ability to engineer alternative containment systems that improve distribution and administration represents potentially significant enhancements to vaccination strategies. In this work, we evaluate the ability to successfully lyophilize a previously demonstrated thermostable tuberculosis vaccine formulation (ID93 + GLA-SE) in a cartridge format compared to a traditional vial container format. Due to differences in the shape of the container formats, a novel apparatus was developed to facilitate lyophilization in a cartridge. Following lyophilization, the lyophilizate was assessed visually, by determining residual moisture content, and by collecting melting profiles. Reconstituted formulations were assayed for particle size, protein presence, and GLA content. Based on assessment of the lyophilizate, the multicomponent vaccine was successfully lyophilized in both formats. Also, the physicochemical properties of the major components in the formulation, including antigen and adjuvant, were retained after lyophilization in either format. Ultimately, this study demonstrates that complex formulations can be lyophilized in alternative container formats to the standard pharmaceutical glass vial, potentially helping to increase the distribution of vaccines.

Particle Sizing of Nanoparticle Adjuvant Formulations by Dynamic Light Scattering (DLS) and Nanoparticle Tracking Analysis (NTA).

Dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA) are two orthogonal and complementary methods of measuring size of particles in a sample. These technologies use the theory of Brownian motion by analyzing the random changes of light intensity scattered by particles in solution. Both techniques can be used to characterize particle size distribution of proteins and formulations in the nanometer to low micron range.Each method has benefits over the other. DLS is a quick and simple measurement that is ideal for monodisperse particles and can also analyze a distribution of particles over a wide range of sizes. NTA provides a size distribution that is less susceptible to the influence of a few large particles, and has the added benefit of being able to measure particle concentration. Here we describe methods for measuring the particle size and concentration of an oil-in-water nanoemulsion.

Quantification of Multiple Components of Complex Aluminum-Based Adjuvant Mixtures by Using Fourier Transform Infrared Spectroscopy and Partial Least Squares Modeling.

Fourier transform infrared (FTIR) spectroscopy is widely used in the pharmaceutical industry for process monitoring, compositional quantification, and characterization of critical quality attributes in complex mixtures. Advantages over other spectroscopic measurements include ease of sample preparation, quantification of multiple components from a single measurement, and the ability to quantify optically opaque samples. This method describes the use of a multivariate model for quantifying a TLR4 agonist (GLA) adsorbed onto aluminum oxyhydroxide (Alhydrogel®) using FTIR spectroscopy that may be adapted to quantify other complex aluminum based adjuvant mixtures.

Quantitative measurement of Toll-like receptor 4 agonists adsorbed to Alhydrogel(®) by Fourier transform infrared-attenuated total reflectance spectroscopy.

Aluminum salts have a long history as safe and effective vaccine adjuvants. In addition, aluminum salts have high adsorptive capacities for vaccine antigens and adjuvant molecules, for example, Toll-like receptor 4 (TLR4) agonists. However, the physicochemical properties of aluminum salts make direct quantitation of adsorbed molecules challenging. Typical methods for quantifying adsorbed molecules require advanced instrumentation, extreme sample processing, often destroy the sample, or rely on an indirect measurement. A simple, direct, and quantitative method for analysis of adsorbed adjuvant molecules is needed. This report presents a method utilizing Fourier transform infrared spectroscopy with a ZnSe-attenuated total reflectance attachment to directly measure low levels (<30 µg/mL) of TLR4 agonists adsorbed on aluminum salts with minimal sample preparation.

Cryogenic transmission electron microscopy of recombinant tuberculosis vaccine antigen with anionic liposomes reveals formation of flattened liposomes.

Development of lipid-based adjuvant formulations to enhance the immunogenicity of recombinant vaccine antigens is a focus of modern vaccine research. Characterizing interactions between vaccine antigens and formulation excipients is important for establishing compatibility between the different components and optimizing vaccine stability and potency. Cryogenic transmission electron microscopy (TEM) is a highly informative analytical technique that may elucidate various aspects of protein- and lipid-based structures, including morphology, size, shape, and phase structure, while avoiding artifacts associated with staining-based TEM. In this work, cryogenic TEM is employed to characterize a recombinant tuberculosis vaccine antigen, an anionic liposome formulation, and antigen-liposome interactions. By performing three-dimensional tomographic reconstruction analysis, the formation of a population of protein-containing flattened liposomes, not present in the control samples, was detected. It is shown that cryogenic TEM provides unique information regarding antigen-liposome interactions not detectable by light-scattering-based methods. Employing a suite of complementary analytical techniques is important to fully characterize interactions between vaccine components.

Elimination of the cold-chain dependence of a nanoemulsion adjuvanted vaccine against tuberculosis by lyophilization.

Next-generation rationally-designed vaccine adjuvants represent a significant breakthrough to enable development of vaccines against challenging diseases including tuberculosis, HIV, and malaria. New vaccine candidates often require maintenance of a cold-chain process to ensure long-term stability and separate vials to enable bedside mixing of antigen and adjuvant. This presents a significant financial and technological barrier to worldwide implementation of such vaccines. Herein we describe the development and characterization of a tuberculosis vaccine comprised of both antigen and adjuvant components that are stable in a single vial at sustained elevated temperatures. Further this vaccine retains the ability to elicit both antibody and TH1 responses against the vaccine antigen and protect against experimental challenge with Mycobacterium tuberculosis. These results represent a significant breakthrough in the development of vaccine candidates that can be implemented throughout the world without being hampered by the necessity of a continuous cold chain or separate adjuvant and antigen vials.

Modulating potency: Physicochemical characteristics are a determining factor of TLR4-agonist nanosuspension activity.

Activity of adjuvanted vaccines is difficult to predict in vitro and in vivo. The wide compositional and conformational range of formulated adjuvants, from aluminum salts to oil-in-water emulsions, makes comparisons between physicochemical and immunological properties difficult. Even within a formulated adjuvant class, excipient selection and concentration can alter potency and physicochemical properties of the mixture. Complete characterization of physicochemical properties of adjuvanted vaccine formulations and relationship to biological response is necessary to move beyond a guess-and-check paradigm toward directed development. Here we present a careful physicochemical characterization of a two-component nanosuspension containing synthetic TLR-4 agonist glucopyranosyl lipid adjuvant (GLA) and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) at various molar ratios. Physicochemical properties were compared with potency, as measured by stimulation of cytokine production in human whole blood. We found a surprising, nonlinear relationship between physicochemical properties and GLA-DPPC ratios that corresponded well with changes in biological activity. We discuss these data in light of the current understanding of TLR4 activation and the conformation-potency relationship in development of adjuvanted vaccines.

In vitro evaluation of TLR4 agonist activity: formulation effects.

Effective in vitro evaluation of vaccine adjuvants would allow higher throughput screening compared to in vivo studies. However, vaccine adjuvants comprise a wide range of structures and formulations ranging from soluble TLR agonists to complex lipid-based formulations. The effects of formulation parameters on in vitro bioactivity assays and the correlations with in vivo adjuvant activity is not well understood. In the present work, we employ the Limulus amebocyte lysate assay and a human macrophage cellular cytokine production assay to demonstrate the differences in in vitro bioactivity of four distinct formulations of the synthetic TLR4 agonist GLA: an aqueous nanosuspension (GLA-AF), an oil-in-water emulsion (GLA-SE), a liposome (GLA-LS), and an alum-adsorbed formulation (GLA-Alum). Furthermore, we demonstrate the importance of the localization of GLA on in vitro potency. By comparing to previous published reports on the in vivo bioactivity of these GLA-containing formulations, we conclude that the most potent activators of the in vitro systems may not be the most potent in vivo adjuvant formulations. Furthermore, we discuss the formulation considerations which should be taken into account when interpreting data from in vitro adjuvant activity assays.

Working together: interactions between vaccine antigens and adjuvants.

The development of vaccines containing adjuvants has the potential to enhance antibody and cellular immune responses, broaden protective immunity against heterogeneous pathogen strains, enable antigen dose sparing, and facilitate efficacy in immunocompromised populations. Nevertheless, the structural interplay between antigen and adjuvant components is often not taken into account in the published literature. Interactions between antigen and adjuvant formulations should be well characterized to enable optimum vaccine stability and efficacy. This review focuses on the importance of characterizing antigen-adjuvant interactions by summarizing findings involving widely used adjuvant formulation platforms, such as aluminum salts, emulsions, lipid vesicles, and polymer-based particles. Emphasis is placed on the physicochemical basis of antigen-adjuvant associations and the appropriate analytical tools for their characterization, as well as discussing the effects of these interactions on vaccine potency.