Adeno-associated virus (AAV) has emerged as a leading platform for gene therapy, enabling the delivery of therapeutic DNA to target cells. However, the potential of AAV to deliver protein payloads has been unexplored. In this study, we engineered a protein carrier AAV (pcAAV) to package and deliver proteins by inserting binding domains on the interior capsid surface. These binding domains mediate the packaging of specific target proteins through interaction with cognate peptides or protein tags during the capsid assembly process. We demonstrate the packaging of multiple proteins, including green fluorescent protein, Streptococcus pyogenes Cas9, Cre recombinase, and the engineered peroxidase APEX2. Packaging efficiency is modulated by the binding domain insertion site, the viral protein isoform containing the binding domain, and the subcellular localization of the target protein. We show that pcAAV can enter cells and deliver the protein payload and that enzymes retain their activity after packaging. Importantly, this protein packaging capability can be translated to multiple AAV serotypes. Our work establishes AAV as a protein delivery vehicle, significantly expanding the utility of this viral vector for biomedical applications.

AAV is widely used for efficient delivery of DNA payloads. The extent to which the AAV capsid can be used to deliver a protein payload is unexplored. Here, we report engineered AAV capsids that directly package proteins - Protein Carrier AAV (pcAAV). Nanobodies inserted into the interior of the capsid mediate packaging of a cognate protein, including Green Fluorescent Protein (GFP), Streptococcus pyogenes Cas9, Cre recombinase, and the engineered peroxidase APEX2. We show that protein packaging efficiency is affected by the nanobody insertion position, the capsid protein isoform into which the nanobody is inserted, and the subcellular localization of the packaged protein during recombinant AAV capsid production; each of these factors can be rationally engineered to optimize protein packaging efficiency. We demonstrate that proteins packaged within pcAAV retain their enzymatic activity and that pcAAV can bind and enter the cell to deliver the protein payload. Establishing pcAAV as a protein delivery platform may expand the utility of AAV as a therapeutic and research tool.

Adeno-associated virus (AAV) has been remarkably successful in the clinic, but its broad tropism is a practical limitation of precision gene therapy. A promising path to engineer AAV tropism is the addition of binding domains to the AAV capsid that recognize cell surface markers present on a targeted cell type. We have recently identified two previously unexplored capsid regions near the 2/5-fold wall and 5-fold pore of the AAV capsid that are amenable to insertion of larger protein domains, including nanobodies. Here, we demonstrate that these hotspots facilitate AAV tropism switching through simple nanobody replacement without extensive optimization in both VP1 and VP2. Our data suggest that engineering VP2 is the preferred path for maintaining both virus production yield and infectivity. We demonstrate highly specific targeting of human cancer cells expressing fibroblast activating protein (FAP). Furthermore, we found that the combination of FAP nanobody insertion plus ablation of the heparin binding domain can reduce off-target infection to a minimum, while maintaining a strong infection of FAP receptor-positive cells. Taken together, our study shows that nanobody swapping at multiple capsid locations is a viable strategy for nanobody-directed cell-specific AAV targeting.

An important quality attribute of a recombinant adeno-associated virus (rAAV) as a therapeutic vector is its infectivity. Current assays to quantify infectious rAAV rely on coinfection with a helper virus such as adenovirus (Ad), which requires helper virus preparation and introduces additional variability. A simple method that has high sensitivity and removes the need for helper virus would improve assay consistency and facilitate high-throughput applications such as rAAV producer cell line development. In this study, we describe a stable assay cell line that was generated by integrating the coding sequences for AAV Rep68 and Ad E4orf6 and DNA binding protein under the control of inducible promoters. The Rep68 protein expression was further modulated by a ligand-responsive destabilization domain. In several benchmarks, the cell line gave comparable titers with those obtained using a classical Ad coinfection method. The cell line was also used to titer vectors of multiple rAAV serotypes. This cell line has the potential to serve as an effective and robust tool for quantifying infectious rAAV titers to advance gene therapy vector biomanufacturing.

Several evolved properties of adeno-associated virus (AAV), such as broad tropism and immunogenicity in humans, are barriers to AAV-based gene therapy. Most efforts to re-engineer these properties have focused on variable regions near AAV's 3-fold protrusions and capsid protein termini. To comprehensively survey AAV capsids for engineerable hotspots, we determined multiple AAV fitness phenotypes upon insertion of six structured protein domains into the entire AAV-DJ capsid protein VP1. This is the largest and most comprehensive AAV domain insertion dataset to date. Our data revealed a surprising robustness of AAV capsids to accommodate large domain insertions. Insertion permissibility depended strongly on insertion position, domain type, and measured fitness phenotype, which clustered into contiguous structural units that we could link to distinct roles in AAV assembly, stability, and infectivity. We also identified engineerable hotspots of AAV that facilitate the covalent attachment of binding scaffolds, which may represent an alternative approach to re-direct AAV tropism.

Synthetic biology and protein engineering have become major components in developing 21st century medicines for a host of diseases. However, due to a relative paucity of biophysical models of protein domain compatibility, rational design of synthetic multi-domain proteins for biomedical applications remains a significant challenge.

A longstanding goal in protein science and clinical genetics is to develop quantitative models of sequence, structure, and function relationships and delineate the mechanisms by which mutations cause disease. Deep Mutational Scanning (DMS) is a promising strategy to map how amino acids contribute to protein structure and function and to advance clinical variant interpretation. Here, we introduce 7,429 single residue missense mutation into the Inward Rectifier K+ channel Kir2.1 and determine how this affects folding, assembly, and trafficking, as well as regulation by allosteric ligands and ion conduction. Our data provides high-resolution information on a cotranslationallyfolded biogenic unit, trafficking and quality control signals, and segregated roles of different structural elements in fold-stability and function. We show that Kir2.1 trafficking mutants are underrepresented in variant effect databases, which has implications for clinical practice. By comparing fitness scores with expert-reviewed variant effects, we can predict the pathogenicity of ‘variants of unknown significance’ and disease mechanisms of know pathogenic mutations. Our study in Kir2.1 provides a blueprint for how multiparametric DMS can help us understand the mechanistic basis of genetic disorders and the structure-function relationships of proteins.

A long-standing goal in protein science and clinical genetics is to develop quantitative models of sequence, structure, and function relationships to understand how mutations cause disease. Deep mutational scanning (DMS) is a promising strategy to map how amino acids contribute to protein structure and function and to advance clinical variant interpretation. Here, we introduce 7429 single-residue missense mutations into the inward rectifier K+ channel Kir2.1 and determine how this affects folding, assembly, and trafficking, as well as regulation by allosteric ligands and ion conduction. Our data provide high-resolution information on a cotranslationally folded biogenic unit, trafficking and quality control signals, and segregated roles of different structural elements in fold stability and function. We show that Kir2.1 surface trafficking mutants are underrepresented in variant effect databases, which has implications for clinical practice. By comparing fitness scores with expert-reviewed variant effects, we can predict the pathogenicity of 'variants of unknown significance' and disease mechanisms of known pathogenic mutations. Our study in Kir2.1 provides a blueprint for how multiparametric DMS can help us understand the mechanistic basis of genetic disorders and the structure-function relationships of proteins.

As with many other proteins, ion channels are tremendously complex. In normal physiology for an ion channel to function it must fold, traffic to the surface, respond to stimuli, and conduct ions. In disease, mutations breaks an ion channel by altering one of these stages. However, traditionally, we only study several mutations at a time which limits our ability to mechanistically understand the molecular basis of disease or how a protein normally functions in biology. In the inwardly rectifying potassium channel, Kir2.1, mutations give rise to developmental and cardiac disorder through altering PIP2 dependent activation, folding, and functions.

Inward Rectifier K+ (KIR) channels play key roles in the operation of cells in neuromuscular and other tissue. Pathogenic variants are linked to numerous neurological, cardiovascular, and metabolic disorders. Although some variants cause gating defects in KIR by altering ligand regulation or ion permeation, there is growing evidence that many -perhaps most- variants cause defects in folding and trafficking of KIR. Despite the central role for folding and trafficking in the disease etiology, there have been to date no comprehensive large-scale studies that determine sequence and structural determinants of folding and trafficking robustness in KIR. We recently developed Saturated Programmable Insertion Engineering (SPINE) and differential Domain Insertion Profiling through Sequencing (dDIP-Seq) that allowed us determine the folding and trafficking phenotype of hundreds of thousands variants of Kir2.1, Kir3.1/Kir3.2, and Kir6.2/SUR1 in a single experiment. We will present the SPINE/dDIP-Seq methodology as a generalizable method to study ion channel folding and trafficking, which we have applied (in addition to KIR)to voltage-dependent K+ channels (Kv1.3), acid sensing channels (ASIC1a), and purinoreceptors (P2X3). Focusing on inward rectifiers, we will present our findings that most determinants of folding and trafficking are conserved in KIR and that they correlate with dynamic protein features, such as stiffness. Applying machine learning on the genotype/phenotype datasets plus sequence and structural properties of KIR allowed us to distill the fundamental protein features that determine variant effect. We will also enumerate paralog-specific folding and trafficking determinants and discuss these in the context of adaptive changes in the KIR family that enabled each paralog to respond to different allosteric ligands (e.g., ATP, G proteins, PIP2).

Protein domains are the basic units of protein structure and function. Comparative analysis of genomes and proteomes showed that domain recombination is a main driver of multidomain protein functional diversification and some of the constraining genomic mechanisms are known. Much less is known about biophysical mechanisms that determine whether protein domains can be combined into viable protein folds. Here, we use massively parallel insertional mutagenesis to determine compatibility of over 300,000 domain recombination variants of the Inward Rectifier K+ channel Kir2.1 with channel surface expression. Our data suggest that genomic and biophysical mechanisms acted in concert to favor gain of large, structured domain at protein termini during ion channel evolution. We use machine learning to build a quantitative biophysical model of domain compatibility in Kir2.1 that allows us to derive rudimentary rules for designing domain insertion variants that fold and traffic to the cell surface. Positional Kir2.1 responses to motif insertion clusters into distinct groups that correspond to contiguous structural regions of the channel with distinct biophysical properties tuned towards providing either folding stability or gating transitions. This suggests that insertional profiling is a high-throughput method to annotate function of ion channel structural regions.

Deep mutational scanning (DMS) facilitates data-driven models of protein structure and function. Here, we adapted Saturated Programmable Insertion Engineering (SPINE) as a programmable DMS technique. We validate SPINE with a reference single mutant dataset in the PSD95 PDZ3 domain and then characterize most pairwise double mutants to study epistasis. We observe wide-spread proximal negative epistasis, which we attribute to mutations affecting thermodynamic stability, and strong long-range positive epistasis, which is enriched in an evolutionarily conserved and function-defining network of "sector" and clade-specifying residues. Conditional neutrality of mutations in clade-specifying residues compensates for deleterious mutations in sector positions. This suggests that epistatic interactions between these position pairs facilitated the evolutionary expansion and specialization of PDZ domains. We propose that SPINE provides easy experimental access to reveal epistasis signatures in proteins that will improve our understanding of the structural basis for protein function and adaptation.

Prime editing brings immense promise to correct a large number of human pathogenic mutations and enact diverse edit types without introducing widespread undesired editing events. Delivery of prime editors in vivo would enable such edits to be introduced in a clinical setting. The coding sequence for prime editor, however, is too large to fit within the size-constrained adeno-associated virus (AAV) genome. Herein, we describe a split Staphylococcus aureus prime editor capable of being delivered by dual AAVs. We characterize the editing ability of plasmid-based versions of an S. aureus prime editor in vitro at a variety of loci with diverse edit types. We investigate various split prime editor architectures and alternative dimerization domains. Finally, we demonstrate the capacity of prime editor to be co-delivered by dual AAVs in vitro. While editing rates are lower than desired, this approach presents an important step to translate prime editing for in vivo delivery.

Understanding the biophysical mechanisms that govern the combination of protein domains into viable proteins is essential for advancing synthetic biology and biomedical engineering. Here, we use massively parallel genotype/phenotype assays to determine cell surface expression of over 300,000 variants of the inward rectifier K + channel Kir2.1 recombined with hundreds of protein motifs. We use machine learning to derive a quantitative biophysical model and practical rules for domain recombination. Insertional fitness depends on nonlinear interactions between the biophysical properties of inserted motifs and the recipient protein, which adds a new dimension to the rational design of fusion proteins. Insertion maps reveal a generalizable hierarchical organization of Kir2.1 and several other ion channels that balances stability needed for folding and dynamics required for function. Summary Massively parallel assays reveal interactions between donor domains and recipient proteins govern domain compatibility