Parvoviruses are highly attractive templates for the engineering of safe, efficient, and specific gene therapy vectors, as best exemplified by adeno-associated virus (AAV). Another candidate that currently garners increasing attention is human bocavirus 1 (HBoV1). Notably, HBoV1 capsids can cross-package recombinant (r)AAV2 genomes, yielding rAAV2/HBoV1 chimeras that specifically transduce polarized human airway epithelia (pHAEs). Here, we largely expanded the repertoire of rAAV/BoV chimeras, by assembling packaging plasmids encoding the capsid genes of four additional primate bocaviruses, HBoV2–4 and GBoV (Gorilla BoV). Capsid protein expression and efficient rAAV cross-packaging were validated by immunoblotting and qPCR, respectively. Interestingly, not only HBoV1 but also HBoV4 and GBoV transduced pHAEs as well as primary human lung organoids. Flow cytometry analysis of pHAEs revealed distinct cellular specificities between the BoV isolates, with HBoV1 targeting ciliated, club, and KRT5+ basal cells, whereas HBoV4 showed a preference for KRT5+ basal cells. Surprisingly, primary human hepatocytes, skeletal muscle cells, and T cells were also highly amenable to rAAV/BoV transduction. Finally, we adapted our pipeline for AAV capsid gene shuffling to all five BoV isolates. Collectively, our chimeric rAAV/BoV vectors and bocaviral capsid library represent valuable new resources to dissect BoV biology and to breed unique gene therapy vectors.

No abstract

The virtual screening of large numbers of compounds against target protein binding sites has become an integral component of drug discovery workflows. This screening is often done by computationally docking ligands into a protein binding site of interest, but this has the drawback of a large number of poses that must be evaluated to obtain accurate estimates of protein-ligand binding affinity. We here introduce a fast pre-filtering method for ligand prioritization that is based on a set of machine learning models and uses simple pose-invariant physicochemical descriptors of the ligands and the protein binding pocket. Our method, Rapid Screening with Physicochemical Descriptors + machine learning (RASPD+), is trained on PDBbind data and achieves a regression performance that is better than that of the original RASPD method and traditional scoring functions on a range of different test sets without the need for generating ligand poses. Additionally, we use RASPD+ to identify molecular features important for binding affinity and assess the ability of RASPD+ to enrich active molecules from decoys.