The rapid expansion of genomic medicine has brought a fundamental physical bottleneck to the forefront: the geometric volume of viral capsids. Viral vectors, particularly Adeno-associated viruses (AAVs), are the gold standard for delivering therapeutic genetic material into human cells due to their high transduction efficiency and low toxicity. However, as detailed by Huifang Peng et al., a wild-type AAV capsid has a rigid, unyielding physical capacity limit of approximately 4.7 to 4.9 kilobases of single-stranded DNA.
This strict cargo constraint creates an immediate engineering conflict with the scale of human genetics and modern biotechnology. Many critical therapeutic targets—such as the DMD gene required to treat Duchenne muscular dystrophy or the F8 gene for hemophilia A—are far too large to fit inside a standard viral vector. Furthermore, next-generation genome editing tools, including CRISPR-Cas9 base editors, prime editors, and epitranscriptomic modifiers, require fusing large enzymatic domains to Cas proteins, pushing their total genetic size well beyond the packaging threshold of common delivery vehicles.
To bypass these physical boundaries, researchers have developed a suite of creative strategies designed to shrink the molecular machinery, optimize the vector's internal real estate, or distribute the genetic cargo across multiple vehicles.
Shrinking the Payload: Protein and Promoter Engineering
The most direct way to fit a large genetic sequence into a small capsid is to eliminate non-essential components. Many large human proteins are modular, containing structural domains or repetitive spacer regions that are not strictly required for their primary therapeutic function.
In the case of muscular dystrophy, the full-length dystrophin protein is massive. To make it deliverable, Jackie L. McCourt et al. evaluated internally truncated "mini-dystrophin" variants that preserve only the essential anchoring domains. While these condensed proteins fit within viral limits, manual truncation can inadvertently compromise structural stability. To systematize this process, Ethan Baron et al. introduced a machine learning model called SCISOR, which uses discrete diffusion to learn the evolutionary rules of protein length. By simulating the removal of non-essential amino acids, this approach allows researchers to algorithmically "shrink" therapeutic proteins while maintaining their core biological activity.
Standard Expression Cassette:
[--- Extensive Promoter (1.5 kb) ---][====== Large Therapeutic Gene (4.0 kb) ======]
Total: 5.5 kb (Exceeds AAV Capacity)
Optimized Expression Cassette:
[- MiniPromoter (0.4 kb) -][=== Algorithmically Shrunk/Truncated Gene (4.0 kb) ===]
Total: 4.4 kb (Fits AAV Capacity)
Beyond the protein-coding sequence itself, an expression cassette requires regulatory elements like promoters to drive transcription. Traditional tissue-specific promoters can occupy a significant fraction of the vector's capacity. To reclaim this space, A. Gomes et al. engineered highly compact "MiniPromoters." By isolating the minimal functional DNA-binding motifs required for precise cellular targeting, these engineered promoters minimize the regulatory footprint, freeing up critical headroom inside the capsid for larger therapeutic genes.
Splitting the Cargo Across Multiple Vectors
When a genetic sequence simply cannot be shrunk enough to fit into a single capsid, researchers treat the payload like a piece of modular furniture: they take it apart, pack it into separate vectors, and rely on the target cell to reassemble it.
For example, the F8 gene for hemophilia A cannot fit into a single AAV alongside its regulatory elements. To overcome this, Jianping Zhang et al. utilized a multi-vector strategy where the necessary genome-editing nucleases, guide RNAs, and donor templates are distributed across separate viral vehicles. When these vectors are co-administered, they co-transduce the same target cells. Once inside the nucleus, the separate components are expressed simultaneously and interact to perform the desired genomic modification.
As noted in a comprehensive review on retinal interventions, Gene therapy and genome surgery in the retina, multi-vector systems are highly effective for delivering complex CRISPR setups. However, this approach introduces a statistical bottleneck. As Emre Bektik et al. observed, relying on cells to take up multiple distinct viral particles simultaneously can lower overall efficiency and increase cellular toxicity compared to single-vector delivery, making precise dosing a critical factor.
Optimizing the Vector and Finding Smaller Machinery
An alternative to modifying the cargo is to modify the delivery vehicle itself or search nature for more compact molecular tools.
For larger viral platforms like adenoviruses, the packaging capacity can be expanded by deleting non-essential viral genes. Michael A. Thomas et al. demonstrated that systematically deleting genes within the viral Early Region 4 (E4) frees up roughly 1,000 base pairs of payload space. This expanded capacity allows the virus to accommodate complex, multi-component antigen cassettes without compromising its structural integrity or replication performance.
In the realm of CRISPR-based genome editing, the widely used Streptococcus pyogenes Cas9 (SpCas9) enzyme is notoriously bulky, leaving little room for guide RNAs or regulatory sequences within an AAV vector. To address this, Josh Tycko et al. characterized the smaller Staphylococcus aureus Cas9 (SaCas9) ortholog. Because SaCas9 is significantly shorter than SpCas9, it fits comfortably inside standard viral delivery volumes while maintaining high editing efficiency, providing a natural solution to the cargo problem.
Bypassing the Capsid Entirely
When the physical constraints of viral capsids prove entirely incompatible with a therapeutic design, researchers bypass genetic delivery altogether by using physical or chemical transduction methods.
Viral Delivery: [ Genetic Blueprint ] ---> Packed into Capsid (Strict Size Limit)
Molecular Bypass: [ Functional Protein ] + [ Cell-Penetrating Peptide ] ---> Direct Cell Entry
One approach is direct protein delivery. Kevin J. Sonnemann et al. bypassed the need to deliver the massive utrophin gene by chemically fusing the functional utrophin protein directly to a cell-penetrating TAT peptide. This modification allowed the full-length, fully formed protein to cross cell membranes and enter muscle tissues systemically, completely avoiding the packaging limitations of a viral genome.
For ex vivo applications like engineering immune cells, physical membrane disruption offers another alternative. Justin A. Jarrell demonstrated that microfluidic vortex shedding can gently generate transient pores in cell membranes, allowing large macromolecular payloads like mRNA to diffuse directly into primary human T cells. By eliminating the need for a viral middleman, these non-viral platforms remove the geometric ceiling on cargo size, paving the way for highly complex, multi-gene cellular therapies.
Key Foundational Work
Effects of the Deletion of Early Region 4 (E4) Open Reading Frame 1 (orf1), orf1-2, orf1-3 and orf1-4 on Virus-Host Cell Interaction, Transgene Expression, and Immunogenicity of Replicating Adenovirus HIV Vaccine Vectors (2013) — contributed 3 concept elements to this direction.
Functional Substitution by TAT-Utrophin in Dystrophin-Deficient Mice (2009) — contributed 1 concept elements to this direction.