Novel Adeno-associated Viral Vectors:Synergy Of Evolution And Engineering

Adeno-associated virus(AAV)vectors have become the preferred viral vectors for in vivo gene delivery due to their excellent properties,being widely used in basic biomedical research as well as gene therapies.As application widens,the requirements for AAV vector performance keep increasing.Expanding the application scope of AAV vectors in central nervous system gene delivery faces two main challenges:(1)For precise in situ labeling in a small area,it is necessary to develop AAV vectors with higher targeting specificity and confined diffusion;and(2)for long-range transduction over a large area,novel AAV vectors need to be designed to cover specific brain regions or cell populations that cannot be transduced by current vectors.To address these challenges,it is essential to fully exploit the natural evolutionary resources of AAV and iteratively upgrade the protein engineering technology of artificially modified AAV vectors.Firstly,we investigated AAV13,one of the 13 naturally evolved serotypes and has not yet been used for gene delivery purposes,found that AAV 13 could infect mouse brain tissue with a highly confined diffusion range.The diffusion range of AAV13 is smaller than that of AAV2,which is commonly used for in situ transduction,making it more suitable for precisely targeting neurons in the brain.Therefore,we developed single-virus and dual-virus sparse labeling strategies for AAV 13 to sparsely label local neurons in the brains of C57BL/6 or Cre-transgenic mice.Additionally,through injection in the ventral tegmental area and the following neurobehavioral test,we demonstrated that AAV13 carrying Cre recombinase to drive the expression of Credependent calcium-sensitive indicator can be used for functional monitoring.In conclusion,the AAV 13-based vector toolbox provides new means for precise local gene delivery,which can be applied to in situ targeting of small nuclei,sparse labeling,functional monitoring,etc.Secondly,our investigation of another natural serotype,AAV11,which has not yet been widely used as a gene delivery vector,revealed that AAV11 possesses a more efficient and rigorous ability to retrogradely label projection neurons in the mouse brain.Retrogradely transduced AAV vectors that can be absorbed by neuronal axon terminals have been used to trace input networks of neurons in a given brain region in neuroscience research and for network drug delivery in CNS gene therapy.We developed an efficient and convenient system for producing high-titer AAV11.AAV 11 outperforms or complements the commonly used AAV2-retro in retrograde tracing across various circuit connections.In combination with fiber photometry,AAV 11 can be used to monitor neuronal activities in the functional network by retrograde delivering calcium-sensitive indicator under the control of a neuron-specific promoter or the Crelox system.Furthermore,we showed that GfaABC1D promoter embedding AAV 11 was superior to AAV8 and AAV5 in astrocytic tropism in vivo.Combined with anterograde axoastrocytic transfer technology,AAV11 can be used to study neuron-astrocyte connection.Finally,we showed that AAV11 allows for analyzing neural connectivity difference in the brains of the Alzheimer’s disease and control mice.These properties make AAV11 a promising tool for mapping and manipulating neural circuits,and for gene therapy of proper neurological and neurodegenerative disorders.Finally,targeting amino acids 442-469 in AAV2’s capsid variable region IV,we devised a computer-aided directed evolution approach.Initial screening of AAV2 mutant libraries revealed two novel AAV vectors with more than tenfold higher transduction efficiency in the central nervous system compared to AAV2.Given the passive and low-throughput nature of using naturally evolved AAV resources for novel vector development,it is crucial to improve the technical approach for artificial AAV modification.Although many studies have been carried out on the engineering of the capsid VR-VIII region,few attempts have been made in the VR-IV region.Here,we established a novel engineering paradigm of computer-aided directed evolution,based on training samples from previous datasets,to obtain a viral vector library containing nearly 105 mutants.As proof of concept,we performed a preliminary library screening and identified two AAV2 mutants(AAV2.A1 and AAV2.A2)with 10-15 times higher CNS transduction efficiency than AAV2.The established diverse mutant library is poised for screening high-performance AAV2 variants,providing novel tools for gene delivery and neuroscience.This study integrates bioinformatics,AI and synthetic biology to revolutionize vector engineering.In summary,this study meticulously examined the transduction properties of two naturally evolved AAVs in the central nervous system,leading to the development of the AAV13 vector for precise in situ labeling within limited range,and the AAV11 vector for efficient retrograde transduction and superior astrocyte transduction.Furthermore,by utilizing AAV2 as the subject of study,we enhanced the technical paradigm for artificial evolution by creating a low-redundancy,high-diversity AAV2 mutant library through machine learning-aided directed evolution.We also identified two novel AAV vectors with superior transduction efficiency compared to AAV2.Collectively,these research outcomes provide innovative vector tools for neuroscience research,advanced technical methods for optimizing AAV vector performance,and novel possibilities for future gene therapy vector selection.