An AI-Designed Lysosomal Degradation System Targeting HIV-1 Reverse Transcriptase: A New Strategy for Achieving Functional Cure

Abstract

The primary obstacle to achieving a functional cure for HIV-1 is the persistent presence of the latent viral reservoir, which cannot be eradicated by current antiretroviral therapies. This paper proposes a simplified protein-based therapeutic strategy that combines AI-driven de novo protein design with lysosome-targeted degradation. This strategy utilizes computationally designed high-affinity microproteins that recognize conserved regions of the HIV-1 reverse transcriptase and Gag protein. These recognition modules are fused to a lysosome-targeting signal peptide, thereby directing viral proteins to lysosomes for degradation. To deliver the therapeutic proteins to helper T cells, PEGylated liposomes are used as carriers, surface-modified with AI-designed gp120 receptor-binding domain mimetics for targeted recognition and a CD47 mimetic peptide for immune evasion. A PEST sequence is also incorporated into the system to enable the therapeutic protein to self-destruct at a predetermined time after completing its task. This approach requires only two core components—"recognition domain + sorting signal"—and does not rely on viral protease activation or the ubiquitin-proteasome system, thereby providing a minimalist antiviral strategy with a well-defined mechanism. This paper discusses the design logic, rationale, feasibility, and potential validation pathways of this approach.