In a groundbreaking advance that promises to revolutionize genetic engineering in postmitotic cells, researchers have unveiled a method that dramatically enhances precise genome editing efficiency in adult mouse brains. By harnessing the natural cellular repair machinery with an innovative design of microhomology (µH) tandem repeat repair arms, this new approach unlocks the combined power of two DNA repair pathways previously thought to be challenging to co-opt: nonhomologous end joining (NHEJ) and microhomology-mediated end joining (MMEJ). This dual activation strategy results in unprecedented rates of in-frame gene tagging, offering a powerful tool for neuroscience and countless other fields that depend on precise in vivo genetic modifications.
Traditional strategies for targeted gene tagging often struggle in nondividing cells such as neurons, where the standard homology-directed repair (HDR) pathways become virtually inactive. As a result, most gene editing efforts in such cells rely heavily on NHEJ-dependent methods like homology-independent targeted integration (HITI), which, while functional, typically lead to inefficient or out-of-frame integrations. This limitation has stalled progress in tagging endogenous proteins with fluorescent markers in adult brain tissue, curbing our ability to study protein localization and function in intact neural circuits.
The new study confronted this longstanding problem by ingeniously incorporating frame-retentive µH tandem repeat sequences into the repair template. These sequences are designed to promote engagement of both NHEJ and MMEJ pathways simultaneously. MMEJ, a repair mechanism that had been underutilized in this context, leverages short regions of sequence homology to drive precise and predictable DNA integration, offering a promising avenue to overcome the inefficiencies of traditional NHEJ-only approaches.
To test their strategy, the authors targeted Tubb2a, a neuronal-specific tubulin gene known for its critical role in microtubule dynamics within axons and somas. Employing a dual-viral system, they delivered adeno-associated viruses (AAVs) into adult mouse brains: one encoding the Cas9 nuclease and the other carrying a guide RNA (gRNA) targeting the 3′ end of Tubb2a, alongside a promoterless eGFP sequence designed for in-frame fusion, as well as a constitutively expressed mCherry marker for verifying successful viral transduction.
Three weeks post-administration, meticulous histological examination revealed eGFP fluorescence precisely localized in neurons from virus-infected brain regions, demonstrating endogenous expression from the modified Tubb2a locus. More strikingly, whole-brain volumetric imaging via mesoSPIM, after optimized optical clearing with the wildDISCO method, visualized eGFP distribution along neuronal projections in cortical and hippocampal areas, confirming effective integration and expression in intact neural networks.
To validate the molecular identity of the targeted fusion protein, immunoprecipitation followed by Western blotting was employed. The results revealed a clear band matching the combined molecular weight of Tubb2a and eGFP exclusively in brains receiving the dual AAV system, ruling out nonspecific expression or artifacts and underscoring the precision of the genome-editing outcome.
Delving deeper into the DNA repair mechanisms at play, the team performed deep sequencing of the expected Tubb2a–eGFP junction site from two separate hemispheres of the treated brains without selection bias for eGFP-expressing cells. This comprehensive profiling illuminated the complex landscape of endogenous DNA repair at the target locus. Notably, while NHEJ-mediated integration was present, it constituted a mere 1.8% of editing events, highlighting the limitations of relying solely on this pathway in adult neurons.
In stark contrast, the inclusion of µH tandem repeats activated the MMEJ pathway robustly, accounting for approximately 8.6% of editing outcomes with tight reproducibility across samples. The dominant MMEJ repair outcome was a highly predictable six-nucleotide deletion connecting the donor and target sequences, a result harmonizing with prior computational predictions via the inDelphi platform. This microhomology-driven mechanism ensured that the majority of integration events retained the proper reading frame and avoided insertion/deletion scars, which are often problematic in functional protein tagging.
Through this dual-pathway approach, the frequency of in-frame mutations increased nearly fivefold, and scar-free, precise gene tagging doubled compared to conventional NHEJ or HITI methods alone. These quantitative improvements are transformative, especially in nonproliferating tissue contexts where the efficiency and fidelity of genome engineering have traditionally been limiting factors.
The significance of this research extends far beyond fluorescent protein tagging. By enabling predictable and in-frame targeted gene modifications in adult neurons, the µH tandem repeat-mediated integration platform opens new horizons for investigations into neuronal biology, neurodegenerative disease models, and potential therapeutic genome editing applications. Precise manipulation of neural proteins in vivo could facilitate mapping of protein interactions, monitoring of dynamic cellular processes, and even correction of pathogenic mutations with unprecedented fidelity.
Moreover, the underlying design principle—exploiting microhomologies to co-opt MMEJ in tandem with NHEJ—presents a versatile template engineering paradigm likely applicable across diverse cell types and organisms. This could revolutionize the field of precise genome editing, expanding its reach into tissues and developmental stages traditionally refractory to genetic modification.
The integration of computational prediction tools, exemplified by inDelphi, in template design was a critical factor in the method’s success. This deep-learning-assisted approach enabled anticipation of the predominant repair outcomes, guiding the creation of µH microhomology arms that preserved the protein reading frame. Such synergy between predictive modeling and molecular engineering exemplifies the cutting-edge convergence of bioinformatics and experimental biology.
Visual evidence from whole-brain imaging substantiates not only the molecular precision of the editing but also its biological relevance, as the eGFP-tagged Tubb2a protein was observed localized appropriately within neuronal compartments. This confirms that the approach yields functional fusion proteins faithfully recapitulating endogenous expression patterns.
Beyond neuroscience, the concept of frame-retentive µH tandem repeat repair arms engaging dual DNA repair pathways has broader implications for gene therapy, synthetic biology, and the development of genetically programmable cellular systems. Improving the efficiency and accuracy of targeted integrations in postmitotic or slowly dividing cells has long been a critical hurdle for translational genome engineering, and this method offers a promising path forward.
In practical terms, the use of AAVs for delivery, a proven clinical vector system, underscores the translational potential of this technology. By circumventing the necessity for HDR-dependent repair, the method may enhance the applicability of in vivo genome editing strategies in mature tissues, including human brains.
The experimental confirmation that µH-mediated integration operates effectively in adult mammalian neurons challenges previous paradigms suggesting limited repair flexibility in these cells. This reinforces a new understanding of endogenous DNA repair repertoire accessible for therapeutic exploitation.
Overall, this study exemplifies how precise molecular design, informed by computational prediction and elegant biological experiments, can substantially enhance the efficiency and predictability of gene editing. It sets a new benchmark for endogenous gene tagging and paves the way for a future where genetic interventions in the brain and other postmitotic tissues are routine, safe, and precise.
By engaging both the NHEJ and MMEJ pathways through intelligent template design, the authors successfully transcended the traditional boundaries of genome editing efficiency in neurons. This could catalyze a wide range of applications from fundamental neuroscience to clinical gene therapies, illustrating the power of integrating synthetic biology with computational foresight.
The capacity to generate scar-free, predictable, and functionally relevant in-frame integrations in differentiated cells may herald a new era in genetic engineering, accelerating discoveries and treatments that were previously constrained by technical bottlenecks.
In conclusion, this innovative approach of µH tandem repeat-mediated integration enhances the precision, efficiency, and functional outcome of genome editing in the adult mouse brain. It provides a robust platform for endogenous protein tagging and gene modification that overcomes prior limitations in postmitotic cells. As genome editing rapidly evolves, this development represents a leap forward toward unlocking the full potential of genetic manipulations in complex tissues.
Subject of Research: Precise genome editing strategies to improve in-frame gene tagging in nonproliferating cells, specifically adult neurons.
Article Title: Precise, predictable genome integrations by deep-learning-assisted design of microhomology-based templates.
Article References:
Naert, T., Yamamoto, T., Han, S. et al. Precise, predictable genome integrations by deep-learning-assisted design of microhomology-based templates. Nat Biotechnol (2025). https://doi.org/10.1038/s41587-025-02771-0
Image Credits: AI Generated
