Tropism is not a fixed property of the capsid — it is the output of surface chemistry
AAV capsid tropism is often discussed as if it were an intrinsic binary property: AAV2 transduces retinal cells, AAV9 crosses the blood-brain barrier, AAV8 targets hepatocytes. This framing is useful shorthand but misleading for engineering purposes. Tropism is the emergent result of the capsid surface's interaction with cell-surface receptors and glycans — interactions that are determined by the three-dimensional arrangement of amino acid residues on the surface-exposed loops of the VP1, VP2, and VP3 viral proteins.
Because tropism arises from surface chemistry, it is modifiable. The surface-exposed regions of the AAV capsid — particularly the variable regions (VRs) I–IX located on the protruding protrusions of the icosahedral surface — can be mutated, inserted, or chimeric-swapped to redirect receptor binding specificity. This is tropism engineering: the rational or directed-evolution-guided manipulation of capsid surface residues to achieve a desired biodistribution outcome.
This article covers the molecular basis of AAV tropism, the rational mutation strategies for surface loops, the functional consequences of those mutations, and the tradeoffs that make tropism engineering considerably more complicated than swapping receptor-binding residues one at a time.
The structural basis of AAV surface interactions
The AAV capsid is a T=1 icosahedral particle approximately 25 nm in diameter, assembled from 60 copies of VP1, VP2, and VP3 (in a typical 1:1:10 ratio) arranged with fivefold, threefold, and twofold symmetry axes. The primary receptor-binding interactions are mediated by amino acid residues located at the protrusions formed at the threefold symmetry axes — the "three-fold protrusions" visible in cryo-EM reconstructions.
The variable regions (VRs) are defined as the surface-exposed loops that distinguish AAV serotypes from one another. While the core beta-barrel structure of the VP monomers is highly conserved across AAV serotypes (reflecting shared capsid assembly requirements), the VRs diverge significantly in sequence and structure. These VR differences underlie the serotype-specific receptor binding and tropism profiles:
- HSPG (heparan sulfate proteoglycan) binding in AAV2: Positively charged residues R484, R487, R585, R588, and K532 form the HSPG binding footprint on the threefold protrusion. These arginines and lysines interact electrostatically with the negatively charged sulfate groups of heparan sulfate. AAV2's high HSPG affinity (Kd approximately 1–10 nM) drives its efficient transduction of HSPG-expressing cells including retinal cells, neurons, and hepatocytes.
- Galactose binding in AAV9: W503 (tryptophan at position 503 in the VP1 numbering) forms the primary galactose contact on the AAV9 threefold protrusion, explaining AAV9's glycan receptor preference and contributing to its ability to bind endothelial glycoproteins and cross the blood-brain barrier after IV administration.
- LamR (37/67 kDa laminin receptor) binding in AAV2/AAV3: A separate receptor interaction contributing to broad transduction. LamR-expressing cells include neurons, hepatocytes, and epithelial cells.
Rational mutation strategies
Three broad categories of rational capsid engineering have been explored for tropism redirection and enhancement.
Surface loop substitution. The simplest conceptual approach is to replace the receptor-binding residues of one AAV serotype with those of another to transfer tropism. In practice, VR grafting — taking the VR sequence from a serotype with desired tropism and inserting it into the backbone of a serotype with favorable manufacturing characteristics — has been pursued for several applications. Success is partial: transferring the binding footprint often transfers receptor affinity but does not always transfer in vivo biodistribution, because tropism in vivo depends on more than primary receptor binding (post-entry trafficking, intracellular transport, nuclear import all differ between serotypes).
Peptide insertion into surface loops. Rather than swapping loop sequences, peptide ligands (receptor-targeting peptides, integrin-binding RGD motifs, tumor-targeting peptides, CNS-targeting peptides) can be inserted into exposed surface positions without replacing the native loop. The most commonly used insertion site in engineering studies is the position between VP1 residues 587/588 in AAV2 (within VR VIII, the dominant receptor-binding protrusion). Insertions at this site have been used to add integrin-binding motifs (RGD for αv integrin targeting), single-chain antibody fragments, DARPins, and targeting peptides derived from phage display selections.
The key constraint for peptide insertion is steric: the inserted sequence must fold in a way that presents the targeting moiety to solvent while not disrupting threefold protrusion assembly or VR VIII geometry to the point where capsid assembly fails. Insertions of 7–15 residues are generally tolerated; larger insertions frequently abolish or severely reduce capsid titer. Flexible linkers (Gly-Ser repeats) flanking the insert improve tolerability.
Directed evolution (SELEX / capsid shuffling). Rather than rational design, directed evolution approaches generate libraries of capsid variants (by error-prone PCR, DNA shuffling of multiple parental serotypes, or structured VR randomization) and then select for variants that transduce target cells efficiently in tissue culture or in vivo. Several AAV variants isolated via directed evolution — including AAV-PHP.B and AAV-PHP.eB (CNS-targeting after IV in rodents) and Anc80L65 (synthetic ancestral capsid with broad tropism) — have emerged from directed evolution approaches rather than rational design. Directed evolution is currently more productive than purely rational mutation for achieving quantitative improvements in transduction of specific cell types.
Consequences of surface loop mutations: what changes beyond receptor binding
Capsid surface mutations do not affect only receptor binding. This is the principal complexity of tropism engineering — the capsid surface participates in multiple functions simultaneously, and mutations intended to change receptor specificity often have collateral effects.
Antibody recognition epitopes. The surface-exposed VRs that determine receptor binding are the same regions that are recognized by neutralizing antibodies in human serum (anti-AAV NAbs). Mutations within VRs that change receptor binding often simultaneously change the capsid's immunological profile — the antibody epitopes change, which alters NAb recognition. This is a double-edged effect: mutations can reduce NAb seroprevalence (because the mutant capsid is less recognized by circulating anti-serotype antibodies) but can also generate new immunological characteristics that are difficult to predict without empirical NAb profiling.
Capsid stability and manufacturing yield. The protrusions formed at the threefold axis are the sites of VP:VP interactions between neighboring pentamers in the icosahedral assembly. Mutations within the threefold protrusions that introduce steric clashes or remove stabilizing hydrogen bonds can reduce capsid assembly efficiency, lowering manufacturing yield. Mutations that reduce thermal stability of the assembled capsid affect storage characteristics. For clinical programs, manufacturing yield (expressed as genome-containing particles / mL at harvested titer) and full:empty capsid ratio (ratio of genome-containing to genome-empty capsids) are as important as in vivo function — a highly active capsid variant that yields 10-fold less titer than the parental serotype may not be manufacturable at the dose required for clinical use.
Endosomal trafficking. AAV enters cells via receptor-mediated endocytosis and must escape from the early/late endosome to reach the nucleus for transduction. The VP2 unique region contains a phospholipase A2 (PLA2) domain (in the VP1 unique region, accessible via conformational change at low pH) that is responsible for endosomal membrane penetration. Surface mutations far from this domain can nonetheless affect endosomal escape efficiency through changes in the conformational dynamics of the capsid under the acidic endosomal environment. VP1/VP2 externalization — the process by which the N-terminal PLA2 domain becomes surface-exposed after endosomal acidification — can be altered by mutations in the VRs that affect capsid breathing dynamics.
AAVR (AAV receptor) interaction. AAVR (also known as KIAA0319L) is a broadly required co-receptor for AAV entry identified relatively recently. Most characterized AAV serotypes require AAVR for efficient transduction, interacting through the AAVR polycystic kidney disease (PKD) domain 1 repeat. The AAVR binding footprint on the capsid partially overlaps with the HSPG and other receptor footprints in some serotypes. Mutations engineered for receptor retargeting may unintentionally disrupt AAVR interactions, reducing overall transduction independently of the intended receptor change.
The tyrosine surface residue modification story
One of the most extensively studied rational capsid modification strategies targets surface-exposed tyrosine residues. Tyrosines on the capsid surface are substrates for ubiquitin-mediated proteasomal degradation: tyrosine phosphorylation by EGFR kinase recruits ubiquitin ligases that target the capsid for proteasomal processing before nuclear entry. This degradation pathway reduces transduction efficiency.
Systematic mutation of surface-exposed tyrosines to phenylalanine (Y→F mutations) — which preserves the aromatic ring but removes the hydroxyl group that is the phosphorylation substrate — has been shown to significantly increase transduction efficiency for several serotypes (AAV2 Y444F, Y500F, Y730F; AAV8 Y447F, Y733F; AAV9 Y446F, Y731F) in both in vitro and in vivo studies. These mutations are now routinely incorporated into capsid engineering strategies as a baseline improvement to transduction efficiency independently of tropism targeting.
The Y→F set of mutations illustrates a principle applicable to tropism engineering generally: the same mutation can have multiple simultaneous effects (reducing proteasomal degradation, changing surface hydrophobicity, potentially altering interaction with serum proteins), and the net in vivo result is the product of all effects, not just the intended one.
What this means for vector selection in clinical programs
The complexity of surface loop mutations has practical implications for programs deciding between naturally occurring AAV serotypes, engineered capsid variants, and non-viral delivery alternatives.
For clinical programs targeting the liver via IV delivery, AAV8 and AAV-LK03 (a primate clade F variant with favorable human hepatocyte tropism) provide well-characterized hepatotropism without requiring capsid engineering. The in vivo performance data for these serotypes from clinical trials in hemophilia and other metabolic liver diseases provides a regulatory precedent. Engineered capsid variants, however scientifically promising, carry a higher regulatory characterization burden than natural serotypes.
For programs targeting tissues beyond the liver's naturally high AAV tropism, the choice between rational capsid engineering, directed evolution, and switching to LNP-based delivery must weigh: (1) the maturity of the targeting engineering (is it in clinical stage, preclinical, or still in discovery?); (2) the immune evasion requirements (does the patient population have high NAb prevalence for the parental serotype?); and (3) the payload size constraint (does the therapeutic cargo fit within the AAV limit given the insert requirements of the engineering strategy?).
LNP delivery lacks the elegant natural receptor-targeting tropism that makes liver-tropic AAV serotypes so efficient for hepatic programs — but LNP has no packaging limit, no pre-existing immunity, and can be re-dosed. The optimal vector selection depends on which constraint is binding for a given program, not on ideology about viral versus non-viral delivery platforms.
Gendelivr focuses on LNP formulation optimization and does not engineer AAV capsids. This article provides context for understanding AAV tropism engineering because many of our customers are making vector choice decisions where a thorough understanding of AAV surface biology is necessary for an informed comparison.