Capsid-mediated control of adeno-associated viral transcription determines host range

SUMMARY

Adeno-associated virus (AAV) is a member of the genus Dependoparvovirus, which infects a wide range of vertebrate species. Here, we observe that, unlike most primate AAV isolates, avian AAV is transcriptionally silenced in human cells. By swapping the VP1 N terminus from primate AAVs (e.g., AAV8) onto non-mammalian isolates (e.g., avian AAV), we identify a minimal component of the AAV capsid that controls viral transcription and unlocks robust transduction in both human cells and mouse tissue. This effect is accompanied by increased AAV genome chromatin accessibility and altered histone methylation. Proximity ligation analysis reveals that host factors are selectively recruited by the VP1 N terminus of AAV8 but not avian AAV. Notably, these include AAV essential factors implicated in the nuclear factor κB pathway, chromatin condensation, and histone methylation. We postulate that the AAV capsid has evolved mechanisms to recruit host factors to its genome, allowing transcriptional activation in a species-specific manner.

In brief

Loeb et al. investigate the host range of avian AAV, which is limited in human cells due to transcriptional silencing. Swapping the avian AAV VP1 N terminus, however, with a primate AAV-derived domain unlocks transduction. The authors demonstrate that this domain recruits essential host factors and exerts control over viral transcription.

Graphical Abstract

INTRODUCTION

Adeno-associated viruses (AAVs) are single-stranded (ss) DNA viruses belonging to the genus Dependoparvovirus that rely on adenovirus or herpesvirus for helper-dependent replication.^1,2 Among the smallest known DNA viruses, the AAV capsid has a diameter of 25 nm and packages a genome approximately 4.7 kb in length. The AAV genome consists of four overlapping genes—rep, cap, aap, and maap—flanked between two palindromic inverted terminal repeats (ITRs).^1,2 Since the AAV ITRs are the only cis-acting elements required for genome packaging, the wild-type AAV genome can be replaced with any gene of interest to generate a recombinant vector. These attributes have been successfully exploited in biotechnology applications using AAV as a gene transfer vector.³

Interactions between the AAV capsid and host cells have been extensively mapped at different levels. Cell surface attachment of different AAV capsid serotypes is mediated by glycans such as heparan sulfate, sialic acid, or galactose conjugated to transmembrane proteins or phospholipids embedded in the outer cell membrane.⁴ Cellular uptake is mediated by KIAA0319A (AAVR), which is an essential entry factor for most AAV isolates.⁵ Other host factors determining cellular entry for AAV capsids, such as integrins, have been previously described.⁶ The post-entry interactions of AAV capsids with intracellular trafficking machinery, particularly in the Golgi and nucleus (e.g., AAVR, GPR108, nuclear importins) are well studied.^6–8 Further, interactions between the AAV genome and helper virus or host proteins involved in expression, replication, and packaging have also been documented. Intriguingly, several studies have described the capsid-dependent transcriptional activity of AAV genomes and the potential involvement of a network of host factors, including spliceosomal proteins (PHF5A, SF3B1),⁹ ubiquitin ligases (RNF121),¹⁰ and chromatin remodeling machinery (HUSH complex).¹¹ Additionally, mutational analysis of AAV capsids has revealed an impact on viral transcript levels^12,13 and histone modification markers.¹⁴ However, the underlying structural and mechanistic basis for how AAV capsids regulate the transcriptional activity of uncoated viral genomes has remained elusive.

While a majority of studies to date have focused on primate-derived AAV isolates (e.g., rhesus monkeys or humans), phylogenetic analysis reveals that AAVs infect a wide range of vertebrate hosts—including bats, birds, and reptile species. Investigation into the permissivity of human cells to non-mammalian AAV isolates offers an opportunity to unravel the mechanistic underpinnings behind capsid-specific AAV biology. Here, in order to understand the fundamental role of the AAV capsid in regulating vector transduction and host range, we examine the biology of avian AAV (isolate DA-1), a basal member of the genus Dependoparvovirus, which is transduction deficient in human cells. Our study reveals that the cell surface binding and uptake of avian AAV capsids is not rate limiting; rather, a lack of vector genome transcription appears to be the underlying cause. Through rational mutagenesis, we discover that the N-terminal domain of the VP1 capsid subunit is a structural determinant of species-specific transduction. By swapping the N-terminal domain from primate-derived AAVs (e.g., AAV8) onto non-mammalian AAVs (e.g., avian AAV), we unlock robust transduction in human cells, which is accompanied by markedly altered histone methylation on the AAV genome and high transcript levels. These results are recapitulated in the context of wild-type virus as well as recombinant vectors for in vivo gene transfer. Further investigation using proximity-ligation-based proteomics reveals that the VP1 N-terminal domain of primate-derived capsids selectively recruits host machinery, which controls AAV transduction. Notably, we demonstrate that the AAV8 N-terminal domain specifically recruits several AAV essential human host factors—interleukin-1 receptor-associated kinase 4 (IRAK4), EEF1A lysine methyltransferase 2 (EEF1AKMT2), and histone H1.0 (H1F0). These findings shed light on the structural determinants of capsid-mediated control over AAV transcription and on the histone modifications of AAV genomes, which may be critical for determining host range.

RESULTS

The AAV capsid VP1 N-terminal domain determines transduction and host range

Avian AAV DA-1 belongs to the Galliform AAV clade and is a basal member of the helper-virus-dependent branch of the genus Dependoparvovirus (Figure 1A). The capsid topology of avian AAV DA-1 is highly divergent, with only 60% VP3 amino acid identity to AAV8, a rhesus macaque isolate. Using AAV2 Rep and ITRs, a recombinant avian AAV vector was generated at titers similar to AAV8 (Figures S1A and S1B). Despite a divergent capsid surface, avian AAV appears to bind human cells at levels comparable to the primate-derived AAV8 (Figure 1B). However, recombinant avian AAV was unable to transduce human cells using either an ss-CBA-luciferase (ss-CBA-Luc) genome or a self-complementary CBh-GFP genome (sc-CBh-GFP) (Figures 1D and 1E; Figures S1C–S1E). Transduction was negligible across a wide variety of cell types, including primary human cells, regardless of different multiplicities of infection (Figures 1D and 1E; Figures S1C–S1E).

Figure 1. AAV capsid domain restricts host range.

(A) Maximum likelihood phylogenetic tree of the genus Dependoparvovirus, inferred from the Rep78 sequence of 294 Dependoparvovirus isolates.

(B) Capsid cell binding. U87 cells were dosed with the ss-CBA-Luc vector at an MOI of 2e4 vg/cell and incubated at 4°C for 1 h. Bound capsids were quantified by qPCR for luciferase. n = 4.

(C) Capsid structure, VP1 structure, and genome schematic. The avian-8.1 genome consists of the avian AAV genome (red) with the VP1/2u region of AAV8 up to the start of the aap gene (blue). Red motifs on the capsid surface indicate the AAVR binding footprint. Annotated domains on VP1 include the minimal N terminus (blue), PLA2 (green), VP1/2u linker (red), and VP3 (tan). Created with BioRender.com.

(D) Transduction of U87 cells with ss-CBA-Luc vector. Cells were dosed at an MOI of 1e5 vg/cell, and luciferase activity was assayed after 48 h. n = 3.

(E) Transduction of primary human cardiomyocytes with ss-CBA-Luc vector. Cells were dosed at an MOI of 2.5e5 vg/cell, and luciferase activity was assayed after 72 h. n = 5.

(F) Effect of capsid trans-supplementation. HEK293 cells were infected with 5e4 vg/cell AAV8 sc-CBh-GFP and avian AAV ss-CBA-Luc. Alternatively, HEK293 cells were transfected with pCMV-Nterm-AAV8 and infected with 5e4 vg/cell avian AAV ss-CBA-Luc. Luciferase activity was assayed after 48 h and compared to AAV8 and avian AAV. n = 5.

*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Figure error bars represent one standard error.

See also Figure S1.

Given our cell binding data, we hypothesized that the rate-limiting step that precludes avian AAV transduction in human cells is not related to cell entry but is rather related to post-entry steps. While the initial binding of AAV to cell surface glycans is mediated by surface-exposed residues on VP3, the intracellular stages of AAV transduction are mediated by internal peptides on the VP1/2 unique region (VP1/2u) that are exposed after capsid internalization (Figure 1C).¹⁵ Thus, we hypothesized that differences between the VP1/2u regions of avian AAV and AAV8 may play a role in transduction.^16,17 To test this hypothesis, we generated a chimeric capsid termed avian-8.1 by swapping the VP1/2u region of avian AAV with the corresponding region from AAV8 (Figure 1C). Strikingly, avian-8.1 displayed higher transduction efficiency compared to avian AAV across a wide range of cell lines and significantly outperformed AAV8 in both transformed cells (U87 and HEK293) as well as primary cells (human cardiomyocytes) (Figures 1D and 1E; Figures S1C–S1E). Avian AAV transduction remained undetectable under these conditions regardless of cell type or packaged transgene. To investigate whether the rescue of avian AAV transduction by VP1/2u swap is a cis or trans effect, cells were co-infected with avian AAV and AAV8. Additionally, cells over-expressing AAV8 VP1/2u were infected with avian AAV. Neither treatment, however, rescued avian AAV transduction, suggesting that rescue via VP1/2u swap occurs in cis—regulating transduction on a capsid-specific level (Figure 1F).

To further investigate the role of the VP1/2u region in controlling host range, we tested whether the VP1/2u region of a different primate-derived AAV isolate, AAV2, can rescue avian AAV transduction.¹⁸ Similar to our prior results, avian-2.1 transduced U87 cells with high efficiency, supporting the notion that the VP1/2 capsid region may control host range (Figure 2A). We then inquired whether other AAV capsids of non-mammalian origin can transduce human cells after an equivalent domain swap. To this end, we swapped the VP1/2u region of AAV8 onto bearded dragon AAV or snake AAV. Both chimeric AAV capsids, dragon-8.1 and snake-8.1, robustly transduced A549 cells, suggesting that VP1/2u swap enables successful vectorization of diverse non-mammalian AAVs for mammalian gene transfer (Figure 2B). Based on these results, we investigated the use of our chimeric capsids for gene transfer in vivo. To evaluate if non-mammalian AAVs can be applied in a therapeutic context, healthy C57BL/6J mice were injected intravenously with AAV vectors packaging acid alpha-glucosidase (GAA) under the control of a CBA promoter (ss-CBA-GAA) (Figure S2A). Autosomal recessive mutations in the acid alpha-glucosidase gene are known to cause Pompe disease, which affects 1:20,000 newborns and can potentially be rescued by supplementing correct copies of the GAA gene.^19,20 GAA enzyme levels were measured in mouse liver, heart, and muscle tissue after 6 weeks and normalized to GAA levels in mock-injected mice. We found that enzyme levels were above background for all of the tested capsids (Figure 2C). In heart tissue, dragon-8.1 and snake-8.1 GAA levels were comparable to AAV8, whereas avian-8.1 showed GAA levels significantly above AAV8 (Figure 2C). These results are concordant with our prior findings in human primary cardiomyocytes and suggest potential for improved cardiac gene transfer with avian-8.1 (Figures 1E and 2C).

Figure 2. Capsid domain swap rescues transduction in multiple non-mammalian AAVs.

(A) Avian-2.1 transduction. ss-CBA-Luc was added to U87 cells at an MOI of 1e5 vg/cell. Luciferase activity was assayed after 48 h. n = 2.

(B) Snake-8.1 and dragon-8.1 transduction. The ss-CBA-Luc vector was added to A549 cells at an MOI of 5e4 vg/cell. Luciferase activity was assayed after 48 h. n = 4.

(C) Chimeric AAV transduction in vivo. The ss-CBA-GAA vector was injected into 10-week-old female C57BL/6J mice at a dose of 1e11 vg per mouse via tail vein injection. Additionally, four mice were mock injected with PBS. After 6 weeks, mice were sacrificed, and tissue was collected. GAA enzymatic activity was assayed and normalized to mock-injected mice. n = 4.

*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Figure error bars represent one standard error.

See also Figure S2.

Next, we aimed to further narrow down the structural determinants within the AAV8 VP1/2u region responsible for rescuing avian AAV transduction. To this end, we generated a panel of domain swap capsids and evaluated transduction using the ss-CBA-Luc vector. Swapping the VP1/2u region (avian-8.1), the VP1u region (avian-8.2), or only the N-terminal domain (avian-8.3) rescued avian AAV transduction. Conversely, AAV8 capsids containing the VP1/2u region (AAV8-Av.1), the VP1u region (AAV8-Av.2), or the N-terminal domain of avian AAV (AAV8-Av.3) lost their ability to transduce human cells (Figure 3A). Swapping the phospholipase domain (PLA2) between AAV8 and avian AAV had no effect on the transduction of either chimeric capsid (Figure 3A).^17,21,22 Further swaps within the VP1 N-terminal domain did not reveal any additional fragments that rescue avian AAV transduction (Figure S3A). Thus, we discovered that a minimal 40 amino acid N-terminal domain derived from AAV8 was sufficient to rescue avian AAV transduction in mammalian cells (Figure S3C). Moreover, the equivalent minimal N-terminal domain from snake AAV (Avian-Sn.3) was unable to rescue avian AAV transduction, whereas robust transduction was observed using Avian-8.3 (Figure 3B). While the AAV capsid surface and host factors have generally been regarded as the sole determinants of AAV tropism and host range, our results strongly support the finding that the VP1 N-terminal domain plays a critical role in controlling both AAV transduction and host range.

Figure 3. Interrogation of AAV capsid trafficking and minimal essential N-terminal domains.

(A) Effect of swapping various regions from AAV8 capsid to avian AAV, and vice versa. Right schematic depicts the capsid sequence of each variant. Blue represents AAV8, while red represents avian AAV. U87 cells were dosed with ss-CBA-Luc at an MOI of 5e4, and luciferase activity was assayed after 48 h. n = 5.

(B) Avian-Sn.3 transduction. The ss-CBA-Luc vector was added to U87 cells at an MOI of 1e4 vg/cell. Luciferase activity was assayed after 48 h. n = 8.

(C) Wild-type AAV infectious titers. HEK293 cells were transfected with pXX80 and incubated for 24 h. Cells were then infected with wild-type AAV at an MOI of 1e3 vg/cell. Cells were harvested after 72 h and titrated using qPCR against the AAV ITRs. n = 4.

(D) AAV trafficking and processing efficiency. The ss-CBA-Luc vector was added to U87 cells at an MOI of 2e4 vg/cell and incubated at 37°C for 48 h. Nucleic and cytoplasmic fractions were isolated and subject to qPCR for luciferase to calculate the total internalized and nuclear-localized AAV. Samples were treated with DNase I to calculate encapsidated AAV genomes, nuclease P1 to calculate double-stranded AAV genomes, and exonuclease V to calculate episomal AAV genomes. The efficiency of each step in AAV trafficking and processing was then quantified as a percentage. n = 4.

(E) AAV transcription. The ss-CBA-Luc vector was added to U87 cells at an MOI of 2e4 vg/cell and incubated at 37°C for 48 h. RNA was isolated and reverse transcribed. cDNA was quantified by qPCR for luciferase and normalized to human beta actin. n = 5.

*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Figure error bars represent one standard error.

See also Figure S3.

These observations are also recapitulated with corresponding wild-type AAV viral strains. The cap gene of AAV8, avian AAV, or our minimal N-terminal domain swap capsids were cloned into a wild-type AAV plasmid containing rep from AAV2 and flanked by AAV2 ITRs. All four viruses were effectively produced in HEK293 cells (Figure S3C). To test infectivity, HEK293 cells were dosed with each wild-type AAV strain after transfection with an adenovirus helper plasmid to enable replication. We only observed propagation of AAV8 and avian-8.3 viruses, with avian-8.3 infectivity significantly exceeding that of AAV8 (Figure 3C; Figure S3B). Only background levels were detected for avian AAV and AAV8-Av.3. Thus, the N terminus of avian AAV VP1 appears to preclude infectious cycling in a wild-type context, while the N terminus of AAV8 appears to enable infection in human cells.

The VP1 N-terminal domain controls AAV genome transcription

To understand the post-entry steps influenced by the VP1 N-terminal domain, we investigated each step of the AAV infectious pathway, i.e., cell binding, internalization, nuclear localization, genome release, second-strand synthesis, and episome formation (Figures 3D; Figures S3F and S3G). First, we observed that AAV8 and avian AAV were internalized at equivalent levels, with a slight defect in uptake noted for our chimeric capsid. Similar to AAV8, both avian-8.1 and avian-8.3 were highly dependent on AAVR, a canonical uptake factor for most AAV isolates (Figures S3D and S3E).⁵ Second, we observed that avian AAV nuclear entry was significantly reduced compared to AAV8 (Figure 3D; Figure S3F and S3G). The chimeric capsid, however, undergoes nuclear entry at levels equivalent to avian AAV. Thus, VP1/2u swap did not mediate improved transduction for the chimeric capsid via increased cellular uptake or nuclear entry. Next, we observed that avian-8.1 had a modest yet statistically significant increase in genome release compared to avian AAV. However, there was no statistically significant difference between AAV8 and avian AAV genome release. Thus, we examined post-release genome processing. No significant improvements in second-strand synthesis leading to double-stranded genomes or subsequent episome formation were observed for the avian-8.1 chimera. Finally, we measured the transcript levels for AAV8, avian-8.1, and avian AAV. While genomes delivered by AAV8 and avian-8.1 were transcribed at similar levels, avian AAV transcription was completely abrogated and was below background levels (Figure 3E). Thus, we concluded that the rate-limiting step for avian AAV transduction in human cells is vector genome transcription, which is rescued by the AAV8 VP1 N-terminal domain.

The VP1 N-terminal domain determines chromatin accessibility and histone modification of AAV genomes

Inorder to determine why avian AAV transcription is silenced in human cells, we conducted a chromatin accessibility assay. This experiment measures the sensitivity of vector genomes to nuclease treatment. Higher enrichment ratios indicate greater chromatin accessibility to nuclease treatment and an open chromatin state. Notably, avian AAV had an enrichment ratio equivalent to the heterochromatin control, suggesting a closed, heterochromatic state (Figure 4A). In contrast, genomes delivered by avian-8.1 had enrichment ratios greater than or equal to a euchromatin control locus and similar to AAV8, suggesting an open, euchromatic state (Figure 4A).

Figure 4. AAV capsid VP1 N terminus influences chromatin status and histone modifications of the AAV genome.

(A) Chromatin accessibility assay. U87 cells were dosed with the ss-CBA-Luc vector at an MOI of 2.5e4 vg/cell and incubated at 37°C for 5 days. Chromatin was extracted, and cells were treated with nuclease P1 to digest any remaining single-stranded AAV genomes. Chromatin accessibility of AAV genomes was then assayed using the EpiQuick Chromatin Accessibility Kit. qPCR primers targeted luciferase or euchromatin and heterochromatin control loci. n = 4.

(B–D) AAV CUT&RUN profiling. U87 cells were infected with ss-CBA-Luc vectors at an MOI of 5e3 vector vg/cell and collected after 5 days. Histone modification was then assayed with the Cell Signaling Technology CUT&RUN kit. Antibodies targeted H3K4me3 (C) and H3K9me3 (D). Reads were normalized to an immunoglobulin G (IgG) isotype control antibody. Total histone modification is reported as mean reads per base, normalized to an IgG isotype control antibody (B). n = 1.

*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Figure error bars represent one standard error.

See also Figure S4.

Based on these results, we hypothesized that the N terminus of VP1 may modulate histone acetylation, a key regulator of chromatin condensation. Thus, we tested avian AAV transduction after cells were treated with a histone deacetylase inhibitor (Trichostatin A) or a histone acetylase activator (CTPB) (Figure S4A).^23,24 Neither compound rescued avian AAV transduction, indicating that cell-wide modulation of histone acetylation is not sufficient to rescue avian AAV. To investigate whether avian AAV capsids fail to recruit histone acetylases to the vector genome, we targeted histone acetylase P300 to the AAV genome using dCas9.²⁵ While this approach significantly increased AAV8 transduction, no effect on avian AAV was observed (Figure S4B).

We then carried out a CUT&RUN assay to profile histone methylation across ss-CBA-Luc genomes delivered by avian AAV and by our minimal chimeric capsid, avian-8.3 (Figures 4B–4D). Specifically, we targeted H3K4me3, a marker for active transcription, and H3K9me3, a marker for transcriptional repression.²⁶ Overall, H3K4me3 and H3K9me3 modifications were found on genomes delivered by both avian AAV and avian-8.3. No correlation, however, was observed between H3K4me3 or H3K9me3 marks on the two genomes, suggesting that histone modification profiles were significantly altered between the two capsids (Figures 4C and 4D; Figures S4C–S4D). Strikingly, fewer H3K4 and H3K9 modifications (~2-fold) were observed across the avian AAV vector genome than avian-8.3 (Figure 4B). The reduced prevalence of both activating and repressive markers across the avian AAV vector genome is consistent with our prior observation of a closed chromatin state for avian AAV. Local chromatin accessibility plays a crucial role in the function of histone methyltransferases, and increased chromatin compaction has been shown to limit both H3K4 and H3K9 trimethylation.^27,28 Our CUT&RUN results suggest that changes to the VP1 N terminus can exert a profound impact on the chromatin modeling of AAV genomes. More broadly, these findings corroborate earlier reports that AAV capsid proteins can influence genome transcription by altering association with histone complexes.^29,30

Rescue of avian AAV transduction by N-terminal swap is GPR108 independent

Previous studies have identified G-protein-coupled receptor 108 (GPR108) as an essential factor for certain AAV isolates.⁸ Specifically, GPR108 dependence has been linked to the N-terminal domain of VP1. To assess whether GPR108 is responsible for the rescue of avian AAV transduction upon N-terminal swap, we evaluated vector transduction in GPR108 knockout (KO) cells. As expected, our minimal chimeric capsid, avian-8.3, was unable to transduce GPR108 KO cells since AAV8 is a GPR108-dependent serotype (Figures S4E–S4F). Next, we generated a different chimeric capsid by swapping the N-terminal domain of a GPR108-independent capsid, AAV5, onto avian AAV.⁸ When avian-5.3 was tested on GPR108 KO cells, robust transduction was observed at levels far surpassing avian AAV. Thus, the rescue of avian AAV transduction by N-terminal swap is not solely dependent on GPR108.

AAV8 VP1 N-terminal domain recruits essential host factors

Since the rescue of avian AAV transduction by N-terminal domain swap occurred only in cis (Figure 1F), we rationalized that the N-terminal domain of primate-derived AAV isolates may recruit host factors, which activate vector genome expression in human cells. Thus, we carried out a BioID proximity ligation experiment to identify host factors that specifically interact with the AAV8 N-terminal domain. Briefly, we fused the VP1u region of avian-8.3 or avian AAV to BioID2 biotin ligase with a 6× GGGS linker (Figure S5A).³¹ A BioID2-only construct acted as a control to capture non-specific interactions. Mass spectrometry was then utilized to identify enriched interacting partners for the AAV8 N-terminal domain in U87 cells (Figure S5A). Overall, 4,412 proteins were identified, and 134 were enriched at levels significantly above the BioID2-only control (fold change > 2; p < 0.05) (Figure 5A; Figure S5B). Of these 134 proteins, 66 interacted with only the AAV8 N terminus. Notably, the top AAV8 N-terminal-domain-interacting proteins included RNA-binding proteins, transcription factors, and histone proteins. Gene Ontology analysis reveals that AAV8 N terminus interactors are significantly enriched for terms related to transcriptional regulation and RNA processing (Figure 5B; Figure S5C). These results corroborate the notion that the VP1 N-terminal domain of AAV8, but not of avian AAV, recruits host machinery, even in the absence of AAV genomes.

Figure 5. AAV capsid VP1 N terminus recruits host transcription machinery.

(A) Proximity ligation interactors. Proteins with at least 2-fold enrichment (p < 0.05) from the BioID2-only control were considered significantly enriched. n = 3. Created with BioRender.com.

(B) Gene Ontology analysis. Significant (p < 0.01) biological process terms are shown for AAV8 N terminus-specific interactors (blue) and for proteins enriched for both constructs (gray).

(C) AAV co-immunoprecipitation (coIP)-qPCR. U87 cells were transfected with pEL-CMV-IRAK4-HA, pEL-CMV-EEF1AKMT2-HA, or pEL-CMV-H1F0-HA and incubated for 48 h. The ss-CBA-Luc vector was added at an MOI of 1e4 vg/cell. After 2 h, cells were harvested and cross-linked. HA-tagged proteins were immunoprecipitated, and bound capsid was quantified by qPCR for the luciferase gene. Schematic was created with BioRender.com. n = 3.

(D) AAV8 transduction after IRAK4, EEF1AKMT2, and H1F0 knockdown. HEK293 cells were transfected with lentiCRISPRv2, with four guides per gene. After selection with puromycin, cells were incubated for 6 days to allow for protein turnover. Cells were then dosed with AAV8 packaging sc-CBh-GFP at an MOI of 2e4 vg/cell. Mean fluorescence intensity (MFI) was measured by flow cytometry after 48 h and quantified as MFI relative to scramble knockout.

(E) Proposed model. Both avian AAV and AAV8 are successfully internalized and release their genomes that undergo episome formation. However, avian AAV is not transcribed. We propose that the N terminus of AAV8 recruits essential host factors to the AAV genome. This promotes downstream signaling, which leads to chromatin remodeling and activation of AAV transcription. The N terminus of avian AAV does not interact with these factors, leading to repression of AAV transcription.

*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Figure error bars represent one standard error.

See also Figure S5.

To validate these findings, we performed co-immunoprecipitation qPCR (Figure 5C). Three top candidates were selected from our BioID dataset for further evaluation—IRAK4, EEF1AKMT2, and H1F0. Based on Gene Ontology analysis, these genes have been implicated in the nuclear factor κB (NF-κB) pathway, in chromatin condensation, and in histone methyltransferase activity. Each gene was over-expressed, and the corresponding protein was immunoprecipitated after cells were dosed with either AAV8 or avian AAV. Encapsidated vector genomes were then quantified by qPCR. For all three genes, bound AAV8 levels were significantly greater than bound avian AAV, supporting our mass spectrometry data (Figure 5C). Next, we knocked out each of the genes in HEK293 cells and evaluated the transduction of AAV8 and avian AAV. KO of all three genes significantly reduced AAV8 transduction when using either the ss-CBA-Luc or sc-CBh-GFP vector, indicating that these factors are essential for AAV transduction (Figure 5D; Figures S5D–S5F). While the exact role of these genes in regulating AAV chromatin modification remains to be elucidated, our data suggest that the AAV8 capsid, specifically its N-terminal domain, likely recruits human host factors essential for vector transduction, whereas avian AAV fails to recruit these factors. Swapping the N-terminal domain of AAV8 onto avian AAV enables recruitment of these factors and rescues transduction of the vector.

DISCUSSION

Much of our current understanding of the infectious biology of AAV stems from studies based on primate-derived isolates. Within this framework, AAV tropism and host range have primarily been attributed to interactions of the capsid surface with cell surface carbohydrates and receptors.⁶ Paradoxically, non-mammalian isolates such as avian or serpentine AAV, which can engage human glycans,³² are unable to transduce mammalian cells.³³ Here, while evaluating AAVs of non-mammalian origin, we discover that capsid proteins can exert control over viral genome transcription in a host-specific manner.

A putative role for the AAV capsid in transcriptional regulation has been reported earlier by multiple groups.^9–12 Further, mutations at the 2-fold axis of symmetry in the AAV capsid have been shown to impact transcription,¹³ as well as the histone modification of AAV genomes and, in turn, host range.¹⁴ However, the structural and mechanistic underpinnings of how the AAV capsid may exert control over transcription from vector genomes have remained elusive. Here, we discover that a short span of amino acid residues in the N-terminal domain of VP1 determine the ability of AAV capsids to influence vector genome transcription in a host-specific manner. Remarkably, avian AAV successfully binds to the surface of human cells, is effectively internalized, traffics to the nucleus, undergoes uncoating, and forms episomes—as observed with other AAVs of primate origin. However, a significant lack of transcriptional activity and an overall decrease in histone markings were observed with genomes delivered by avian AAV capsids. Notably, genomes delivered by avian AAV appear largely heterochromatic, corroborating the lack of transcriptional activity. These deficits were rescued by swapping the N-terminal domain of avian AAV with the corresponding domain from primate-derived capsids. Remarkably, no rescue was observed by swapping the corresponding domain from a reptile-derived AAV onto avian AAV, corroborating the importance of selective interactions between mammalian capsids and mammalian host proteins. Genomes delivered by our chimeric capsid are euchromatic and support robust vector genome transcription. Thus, the N-terminal domain of VP1, which remains buried within the AAV capsid lumen prior to cellular uptake,^34,35 appears to play a significant role in genome transcription.

In addition to avian AAV, robust transduction efficiency was also observed by swapping the N-terminal capsid domain of AAV8 onto other non-mammalian AAV capsids, such as snake AAV and bearded dragon AAV. These non-mammalian chimeric capsids are functional both in vitro and in vivo. We demonstrate the increased expression of GAA, the Pompe disease gene, in mouse tissue following the administration of our chimeric capsids. These results provide an initial proof of concept for chimeric non-mammalian AAV vectors as a potential recombinant vector platform for application in human gene therapy. Future evaluation in large-animal models will help determine the potential for these AAV capsids in therapeutic applications.

Limitations of the study

Several limitations with regard to our mechanistic understanding of the findings reported in this study are worth noting. First, while we observed that avian AAV is unable to recruit essential mammalian host factors, a direct link with transcriptional activity remains to be established. Further, exactly how these host factors influence chromatin remodeling and histone methylation/acetylation patterns of the AAV genome remains unclear. Our proximity ligation analysis of the primate-derived AAV8 VP1 N-terminal domain yielded hits such as IRAK4, EEF1AKMT2, and H1F0. Gene Ontology annotation implicates these proteins in the regulation of the NF-κB pathway, chromatin condensation, and histone methyltransferase activity. However, exactly how these factors influence the chromatin modification of AAV vector genomes and impact transcription remains a topic of continued investigation.

Nevertheless, based on the results obtained thus far, we postulate that the N-terminal domain of AAV capsids has likely evolved to control viral genome transcription in a host-specific manner. Overall, our study not only provides important insight into virus-host interactions but also provides a roadmap for utilizing chimeric AAV capsids to further our understanding of these properties while ultimately promising potentially improved recombinant vectors for gene therapy.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources, reagents, and code should be directed to and will be fulfilled by the lead contact, Aravind Asokan (aravind.asokan@duke.edu).

Materials availability

All materials generated in this paper are available upon request to the lead contact upon completion of materials transfer agreement, if relevant.

Data and code availability

• Accession numbers utilized in this study are Avian AAV DA-1 (GenBank: AY629583), AAV8 (GenBank: NC_006261), AAV2 (GenBank: NC_001401), Bearded dragon parvovirus 2014 (GenBank: KP733794), Snake AAV (GenBank: AY349010).

• This paper does not contain original code.

• Data and new materials reported in this paper will be shared by the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines and primary cells

The sex of utilized cell lines and primary cells are as follows: HeLa (female), A549 (male), U-87 MG (male), HEK293 (female), HuH7 (male), C2C12 (female), TH1 (unspecified), primary human cardiomyocytes (male). All cells were cultured at 37 C° with 5% CO_2. Cell lines were maintained using Dulbecco’s Modified Eagle Medium (4.5 g/L D-glucose, 110 mg/L sodium pyruvate), supplemented with 10% fetal bovine serum, 100 μg/mL penicillin, and μg/mL 100 streptomycin. Primary cells were maintained on Matrigel-coated dishes using RPMI-1640 with 1X B27 Supplement.

Mouse experimentation

All animal procedures were performed in accordance with protocols approved by the Institutional Care and Use Committee (IACUC) at Duke University, accredited by the Association for Assessment and Accreditation of Laboratory Animal Care-International (AAALAC). Female C57BL/6J mice (4-weeks old) were purchased from Jackson Laboratories (#000664) and maintained at the Duke University School of Medicine with the assistance of Duke’s Division of Laboratory Animal Resources (DLAR). Mice were housed in a temperature-controlled (18–23 C, 40–60% humidity) and enriched environment, with a 12h light/hour dark cycle. Mice were acclimated to the habitation environment for 6 weeks prior to experimentation. Standard food/water were available ad libitum.

METHOD DETAILS

Phylogenetic analysis

Genomes for all available Dependoparvovirus isolates were accessed from NCBI GenBank. Rep78 sequences were extracted, translated, and aligned using MUSCLE.³⁶ A maximum likelihood phylogenetic tree was then inferred using IQ-TREE.³⁷ Sequences were clustered into classes based on sequence identity.

Structure prediction

VP1 structures for avian AAV and AAV8 were predicted with ColabFold,³⁸ an AlphaFold2 implementation using MMseq2, and annotated using ChimeraX.³⁹

AAV vector production

AAV vector was produced by transfecting HEK293 cells, grown to 80% confluence with 12 μg adenovirus helper plasmid (pXX680), 10 μg RepCap plasmid, and 6 μg ITR transfer plasmid. Transfection was accomplished using polyethyleneimine (PEI) (PolyScience) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated in a 37°C CO2 incubator and media was collected on days 3 and 7 post-transfection. Media was centrifuged at 2,400 g for 5 min to remove cell debris. PEG-8000 was added to the media to a final concentration of 8%. PEG-media was incubated at 4°C overnight. Next, PEG-media was centrifuged at 4,000 g for 1 h to pellet the flocculate containing viral vector. The pellet was resuspended in 2 μL DNase buffer (10 mM tris-HCL, 10 mM MgCl₂, 2 mM CaCl₂) and treated with 20 μL DNase I (10 mg/mL). Next, an iodixanol gradient was prepared by layering 3 mL 17% iodixanol, 3 mL 25% iodixanol, 4 mL 40% iodixanol, and 3 mL 60% iodixanol in a clear ultracentrifuge tube (Beckman-Coulter). Vector solution was overlaid on top of the gradient and centrifuged at 30,000 rpm for 15 h. The top two layers were discarded and 1 mL Iodixanol fractions were then collected. Fractions were tittered with qPCR to determine AAV genome concentration. Peak fractions were then pooled and purified by diafiltration, using a 100k protein concentrator (ThermoFisher). Final vector preps were eluted in PBS containing 0.001% Pluronic F68 and 1 mM MgCl₂. Final vector concentrations were determined by qPCR (see below).

AAV vector quantification

AAV samples were treated with DNase I, along with standards and negative controls, to remove non-encapsidated DNA. 10 μL of each sample was combined with 89 μL DNase buffer (10 mM tris-HCL, 10 mM MgCl₂, 2 mM CaCl₂) and 1 μL DNase I (10 mg/mL) (Sigma Aldrich) and incubated at 37°C for 1 h. Samples were then treated with 100 μL EDTA (20 mM) to inactivate DNase I. Next, samples were diluted 1:100 with 0.05% Tween 20 to inhibit non-specific genome-capsid interactions. Standards were then subject to 7 5-fold dilutions in H₂O. Finally, 2 μL of each sample was combined with 5 μL SYBR Green qPCR Master Mix (Roche), 2.5 μL H₂O, and 0.5 μL primer mix (10 μM each forward and reverse primer). Concentrations were then determined by qPCR with primers amplifying inverted terminal repeat regions.

Transduction assays

Cells were seeded in an opaque 96-well plate at a density of 1e5–2e5 cells/mL and grown in a 37°C CO2 incubator overnight. AAV vector was diluted in PBS (0.001% Pluronic F68, 1 mM MgCl₂) to achieve the desired MOI. 20 μL diluted vector then added to each well and allowed to incubate at 37°C for 48–72 h. For luciferase assays, media was aspirated from each well and cells were then treated with 20 μL 1X passive lysis buffer (Promega) for 30 min. 20 μL firefly luciferase assay reagent (Promega) was added to each well. Luminescence was then measured for a duration of 1 s by the VICTOR X Plate Reader (PerkinElmer). For analysis of GFP expression by flow cytometry, cells were dissociated in TripLE (Gibco) and washed with PBS. Cells were then resuspended in 10% bovine serum albumin and passed through a 100-μm cell strainer. Fluorescent intensity was measured using a Sony SH800 cell sorter following excitation with a 488-nm laser and analyzed using FlowJo.

Generation of primary cardiomyocytes

A human induced pluripotent stem cell (hiPSC) line (DM-03) was maintained under feeder-free culture on Matrigel (Corning) in mTeSR medium (Stemcell Technologies). DM-03 hiPSCs were colony-passaged as small clusters every 5 days. DM-03 hiPSCs were then differentiated into cardiomyocytes (CMs) via small molecule-based modulation of Wnt signaling.⁴³ Briefly, DM-03 hiPSCs were plated into 10-cm Matrigel-coated dishes at a cellular density of 2.8e6 cells/dish with 5uM Y27632 (Stemcell Technologies). To induce cardiac differentiation, cells were incubated in 8 μM CHIR9902 (Stemcell Technologies) in RPMI-1640 (ThermoFisher) with B27(−) insulin (ThermoFisher). Exactly 24 h later, CHIR9902 was removed, and the media replaced with 5 μM IWR-1 (Stemcell Technologies) and 50 μg/ml ascorbic acid in RPMI-1640 with B27(−) insulin. This was repeated again on day 3. On day 4, media was replaced with 5 μM IWR-1 in RPMI-1640 with B27(−) insulin. From day 6 onwards, the cells were cultured in RPMI-1640 with B27(+) insulin (ThermoFisher). Spontaneous contraction of CMs began on days 5–7 of differentiation. At the end of the differentiation period, cultures were dissociated into single cells using Accutase (Stemcell Technologies) and re-plated onto fresh Matrigel-coated dishes to remove dead cells and debris. Cells were then maintained in media consisting of RPMI-1640 with B27(+) insulin. CMs were seeded in a Matrigel-coated 96-well plate at a density of 4e4 cells per well 24 h prior to experimentation.

In vivo study

AAV vector was generated, as described above, packaging acid alpha-glucosidase (GAA) under the control of a CBA promoter. Four capsids packaging CBA-GAA were evaluated: AAV8, Avian-8.1, Snake-8.1, and Dragon-8.1. 10-week-old female C57BL/6J mice were administered ss-CBA-GAA vector at a dose of 1e11 vg per mouse via tail vein injection. 4 mice were dosed per capsid, in addition to 4 PBS mock-injected mice. After 6 weeks, mice were sacrificed, and tissue was collected. GAA enzymatic activity was then assayed in liver, heart, and skeletal muscle tissue.¹⁹ Briefly, tissue was homogenized and clarified by centrifugation. Samples were then evaluated in duplicate in a black 96-well plate. 20 μL of each sample was incubated at 37°C with 4-methylumbelliferyl α-D-glucopyranoside substrate, pH 4.3 (Sigma-Aldrich). The reaction was stopped after 1 h by the addition of sodium carbonate buffer (pH 10.7). Fluorescence was then measured using the Varioskan LUX (ThermoFisher) with 360 nm excitation and 460 nm emission. GAA activity was then calculated using a standard curve made with 4 mM methylumbelliferyl (Sigma-Aldrich). GAA activity was normalized to total protein content, measured by DC Protein Assay (Bio-Rad). GAA activity from AAV treated mice was then normalized to background GAA levels from mock-injected mice.

AAV wildtype production

HEK293 cells were seeded in a 6-well plate at a density of 6e5 cells/mL and grown in a 37°C CO2 incubator overnight. Each well was transfected with 1.2 μg adenovirus helper plasmid (pXX680) and 0.6 μg wildtype AAV plasmid (pTR-RepCap). Transfection was accomplished using PEI (PolySciences) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated in a 37°C CO2 incubator. 72 h after transfection, media and cells were collected. Cells were vortexed vigorously with lysis buffer (10 mM tris-HCl, 10 mM MgCl₂, 2 mM CaCl₂, 0.5% Triton X-100) and then treated with RNase A (ThermoFisher), and Halt Protease Inhibitor Cocktail (ThermoFisher) for 1 h. AAV titers were then determined by qPCR for each sample, as described above. Cell and media yields were summed to determine total packaged vector yield.

AAV wildtype infection

HEK293 cells were seeded in a 6-well plate at a density of 6e5 cells/mL and grown in a 37°C CO2 incubator overnight. Each well was transfected with 1.2 μg adenovirus helper plasmid (pXX680). Transfection was accomplished using PEI (PolySciences) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated in a 37°C CO2 incubator. 24 h after transfection, virus was added to cells at an MOI of 1e3. 72 h later, cells were collected and vortexed vigorously with lysis buffer (10 mM tris-HCl, 10 mM MgCl₂, 2 mM CaCl₂, 0.5% Triton X-100) and then treated with RNase A (ThermoFisher), and Halt Protease Inhibitor Cocktail (ThermoFisher) for 1 h. AAV titers were then determined by qPCR for each sample, as described above.

Generation of cell lines

For generation of U87 GPR108 knockout and HEK239 AAVR knockout, recombinant lentivirus was generated via triple transfection of lentiCRISPRv2, containing guides against the target gene, psPax2, and pVSVG. Lentivirus containing media was harvested at 48 h post transfection. Cells were seeded at 1e5–3e5 cells/well in a 6-well plate and incubated with media containing recombinant lentivirus. Cells were treated with 8 μg/mL polybrene for 10 min, followed by spinoculation for 30 min at 400 g. Cells were incubated for 2–4 h and media was then replaced with media containing lentivirus without polybrene. 48 h post spinoculation, cells were selected with puromycin for 7 days. Clonal lines were then generated by serial dilution of cells.⁴⁴

Binding assay

U87 cells were seeded in a 6-well plate at a density of 3e5 cells/mL and grown in a 37°C CO2 incubator for 48 h. Cells were then preincubated at 4°C for 30 min. AAV vector packaging ss-CBA-Luc was diluted in PBS to achieve an MOI of 2e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 4°C for 1 h. Media was aspirated and cells were collected. Cells were then washed three times with PBS to remove unbound vector.⁴⁵ Cells were vortexed vigorously with lysis buffer (10 mM tris-HCl, 10 mM MgCl₂, 2 mM CaCl₂, 0.5% Triton X-100) and then treated with RNase A (ThermoFisher), and Halt Protease Inhibitor Cocktail (ThermoFisher) for 1 h. AAV titers were then determined by qPCR for each sample, as described above.

AAV packaging and egress assay

HEK293 cells were seeded in a 96-well plate at a density of 1e5 cells/mL and grown in a 37°C CO2 incubator overnight. Each well was transfected with 0.12 μg adenovirus helper plasmid (pXX680), 0.10 μg RepCap plasmid, and 0.06 μg ITR transfer plasmid (pTR-sc-CBh-eGFP). Transfection was accomplished using PEI (PolySciences) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated in a 37°C CO2 incubator. 24 h after transfection, media was discarded and replaced to remove untransfected DNA. At days 2, 4, and 6 post-transfection, the media and cells were collected from three wells per vector and frozen at −80°C. Cells were then treated with 1X passive lysis buffer (Promega), DNase I (Sigma-Aldrich), RNase I (Sigma-Aldrich), and Halt Protease Inhibitor Cocktail (ThermoFisher) for 30 min. Packaged vector yields for each sample was determined by quantitative PCR, as described above. Cell and media yields were summed to determine total packaged vector yield. Cell yields were divided by the total packaged vector yields to determine % vector egress.

Trafficking and genome processing assay

U87 cells were seeded in a 6-well plate at a density of 3e5 cells/mL and grown in a 37°C CO2 incubator for 48 h. AAV vector packaging ss-CBA-Luc was diluted in PBS to achieve an MOI of 2e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 37°C for 48 h. Cells were collected and washed three times with PBS to remove unbound vector. Nuclear and cytoplasmic fractions were then isolated using the Thermo Scientific NE-PER kit. For cell uptake and nuclear localization, samples were diluted 1:2000 and quantified by qPCR. For capsid uncoating, samples were treated with DNase I (Sigma-Aldrich) to degrade non-encapsidated genomes, diluted 1:2000 and quantified by qPCR. For second strand synthesis, samples were heated at 95C to denature the capsids and treated with nuclease P1 (NEB) to degrade single stranded genomes. Samples were then diluted 1:2000 and quantified by qPCR. For episome formation, samples were heated at 95°C to denature the capsids and treated with exonuclease V (NEB) to degrade non-circular genomes. Samples were then diluted 1:2000 and quantified by qPCR.

Transcription assay

U87 were seeded in a 6-well plate at a density of 3e5 cells/mL and grown in a 37°C CO2 incubator for 48 h. AAV vector packaging ss-CBA-Luc was diluted in PBS to achieve an MOI of 2e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 37°C for 48 h. RNA was isolated using the IBI Tri-Isolate RNA Pure Kit. Residual DNA was degraded using the Invitrogen TURBO DNA-Free kit. RNA was then reverse transcribed using the Applied Biosystems High-Capacity RNA-to-cDNA kit. cDNA was then diluted 1:10 in H₂O and quantified by qPCR using primers against luciferase and β-actin. Fold enrichment was calculated as delta-delta Ct.

Chromatin accessibility

U87 were seeded in a 6-well plate at a density of 3e5 cells/mL and grown in a 37°C CO2 incubator for 48 h. AAV vector packaging ss-CBA-Luc was diluted in PBS to achieve an MOI of 2.5e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 37°C for five days, replacing media every 48 h. Cells were collected and processed using the EpiQuik Chromatin Accessibility Assay Kit (EpigenTek). The protocol was followed as described by the manufacturer; however, nuclease P1 (NEB) was added to all samples during chromatin digestion to degrade single stranded AAV genomes. qPCR was then performed using primers against luciferase and control loci. Fold enrichment was calculated as 2^(Ct_{MNase treated} - Ct_{No Mnase}).

Histone acetylation modulation

U87 cells were seeded in an opaque 96-well plate at a density of 1e5 cells/mL and grown in a 37°C CO2 incubator for 48 h. Media was supplemented with 1 μM Trichostatin A (Sigma-Aldrich) or 250 μM CTPB (Active Motif).^46,47 AAV vector packaging ss-CBA-Luc was then diluted in PBS to achieve an MOI of 4e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 37°C for 48 h. Luciferase activity was then assayed, as described above.

P300-dCas9 assay

HEK293 cells were seeded in an opaque 96-well plate at a density of 1e5 cells/mL and grown in a 37°C CO2 incubator for 48 h. Cells were transfected with 150 ng pLV-dCas9-p300-P2A-Puro and 50 ng pSRJ7,²⁵ containing guides against the AAV ITRs. Transfection was accomplished using PEI (PolyScience) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated at 37°C for 24 h. AAV vector packaging ss-CBA-Luc was then diluted in PBS to achieve an MOI of 2e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 37°C for 48 h. Luciferase activity was then assayed, as described above.

CUT&RUN

U87 cells were seeded in a 12-well plate at a density of 1e5 cells/mL and grown in a 37°C CO2 incubator for 24 h. AAV vector packaging ss-CBA-Luc was then diluted in PBS to achieve an MOI of 5e3 vg/cell. ss-CBA-Luc vector was then added to each well and allowed to incubate at 37°C for five days, replacing media every 48 h. Cells were harvested and processed using the Cell Signaling CUT&RUN Assay kit.⁴⁸ Antibodies used were Tri-Methyl-Histone H3 Lys9 Rabbit mAb (Cell Signaling; D4W1U), Tri-Methyl-Histone H3 Lys4 Rabbit mAb (Cell Signaling; C42D8), and IgG XP Isotype Control Rabbit mAb (Cell Signaling; DA1E). DNA was prepared for sequencing using NEBNext Ultra II DNA Library Prep with Sample Purification Beads (NEB) and NEBNext Multiplex Oligos for Illumina (NEB). Samples were then sequenced with Illumina sequencing, yielding approximately 70 million reads per sample. Reads were trimmed with cutadapt and mapped to the ss-CBA-Luc genome using bowtie2.^40,41 Mapped reads were then converted into coverage per base. Coverage was normalized as counts per million reads (CPM) according to sequencing depth for each sample. H3K4 or H3K9 CPM was then normalized for non-specific antibody binding by subtracted IgG control CPM at each base. Total modification was calculated by averaging IgG normalized CPM across the AAV genome.

BioID experiment

The VP1u sequence of avian AAV or avian-8.3 was cloned into a plasmid containing BioID2–6x-GGGGS. A BioID2 only plasmid was also used as a negative control. Three 15 cm plates of U87 cells were grown in a 37°C CO2 incubator until 80% confluent. Each plate of cells was transfected with 6000 ng of BioID2 construct plasmid.³¹ Transfection was accomplished using PEI (PolyScience) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated at 37°C for 48 h. Media was then replaced with 20 mL of DMEM-10% FBS containing 50 μM biotin. Cells were incubated for an additional 24 h and then lysed with 600 μL Urea Buffer (8M Urea, 50 mM tris-HCl, 1 mM DTT, 1X Halt Protease Inhibitor Cocktail [ThermoFisher]. pH 7.4). Cells were further lysed with 150 μL of 20% Triton X-100 and sonication. Samples were centrifuged and biotinylated proteins were pulled down using High-Capacity Streptavidin Agarose (ThermoFisher). Samples were washed with buffer containing 8M urea and 50 mM tris-HCl (pH 7.5), and eluted in buffer containing 25 mM Tris, 50 mM NaCl, 10 mM DTT, 2% SDS, and 3M biotin. Samples were then subject to mass spectrometry, as described below.

Mass spectrometry sample preparation

Samples were spiked with undigested bovine casein at a total of either 1 or 2 pmol as an internal quality control standard. Next, samples were supplemented with 5.9 μL of 20% SDS, reduced with 10 mM dithiothreitol for 30 min at 80°C, alkylated with 20 mM iodoacetamide for 30 min at room temperature, then supplemented with a final concentration of 1.2% phosphoric acid and 874 μL of S-Trap (Protifi) binding buffer (90% MeOH/100mM TEAB). Proteins were trapped on the S-Trap micro cartridge (Protifi), digested using 20 ng/μL sequencing grade trypsin (Promega) for 1 h at 47°C, and eluted using 50 mM TEAB, followed by 0.2% FA, and lastly using 50% ACN/0.2% FA. All samples were then lyophilized to dryness. Samples were resolubilized using 12 μL of 1% TFA/2% ACN with 12.5 fmol/μL yeast ADH.

LC/MS/MS analysis

Quantitative LC/MS/MS was performed on 3 μL input using an MClass UPLC system (Waters Corp) coupled to a Thermo Orbitrap Fusion Lumos high resolution accurate mass tandem mass spectrometer (Thermo) equipped with a FAIMSPro device via a nanoelectrospray ionization source. Briefly, the sample was first trapped on a Symmetry C18 20 mm × 180 μm trapping column (5 μL/min at 99.9/0.1 v/v water/acetonitrile), after which the analytical separation was performed using a 1.8 μm Acuity HSS T3 C18 75 μm × 250 mm column (Waters Corp.) with a 90-min linear gradient of 5–30% acetonitrile with 0.1% formic acid at a flow rate of 400 nL/min with a column temperature of 55°C. Data collection on the Fusion Lumos mass spectrometer was performed for three difference compensation voltages (−40v, −60v, −80v). Within each CV, a data-dependent acquisition (DDA) mode of acquisition with an r = 120,000 (@ m/z 200) full MS scan from m/z 375–1500 with a target AGC value of 4e5 ions was performed. MS/MS scans were acquired in the ion trap in Rapid mode with a target AGC value of 1e4 and max fill time of 35 ms. The total cycle time for each CV was 0.66 s, with total cycle times of 2 s between like full MS scans. A 20 s dynamic exclusion was employed to increase depth of coverage. The total analysis cycle time for each injection was approximately 2 h.

Quantitative mass spectrometry data analysis

Following LC/MS/MS analyses, data were imported into Proteome Discoverer 2.5 (Thermo Scientific). In addition to quantitative signal extraction, the MS/MS data was searched against the SwissProt H. sapiens database (downloaded in Nov 2019), BioID sequences, a common contaminant/spiked protein database (bovine albumin, bovine casein, yeast ADH, human keratin, etc.), and an equal number of reversed-sequence “decoys” for false discovery rate determination. Sequest with Infernys enabled (v 2.5, Thermo PD) was utilized to produce fragment ion spectra and to perform the database searches. Database search parameters included fixed modification on Cys (carbamidomethyl) and variable modification on Met (oxidation). Search tolerances were 2 ppm precursor and 0.8 Da product ion with full trypsin enzyme rules. Peptide Validator and Protein FDR Validator nodes in Proteome Discoverer were used to annotate the data at a maximum 1% protein false discovery rate based on q-value calculations. Note that peptide homology was addressed by only using unique peptides for quantitation. Protein homology was addressed by grouping proteins that had the same set of peptides to account for their identification. A master protein within a group was assigned based on % coverage. Prior to imputation, a filter was applied such that a peptide was removed if it was not measured in at least 2 unique samples (50% of a single group). After that filter, any missing data missing values were imputed using the following rules; 1) if only one single signal was missing within the group of three, an average of the other two values was used or 2) if two out of three signals were missing within the group of three, a randomized intensity within the bottom 2% of the detectable signals was used. To summarize the protein level, all peptides belonging to the same protein were summed into a single intensity.

Gene ontology analysis

Significant GO terms for each gene set were identified using the Comparative Toxicogenomics Database Gene Set Analyzer.⁴² Enriched biological process terms were queried using a corrected p value cutoff of 0.01. Terms were then filtered to include only GO level 5 and above.

Co-IP/qPCR

U87 cells were seeded in a 6-well plate at a density of 3e5 cells/mL and grown in a 37°C CO₂ incubator for 48 h. Cells were transfected with 1250 ng pEL-CMV-IRAK4-HA, pEL-CMV-EEF1AKMT2-HA, or pEL-CMV-H1F0-HA. Transfection was accomplished using PEI (PolyScience) at a ratio of 3.5 μg PEI per 1 μg DNA. Cells were incubated at 37°C for 48 h. AAV vector packaging ss-CBA-Luc was then diluted in PBS to achieve an MOI of 2e4 vg/cell. AAV vector was then added to each well and allowed to incubate at 37°C for 2 h. Cells were collected, cross-linked with 1% formalin, and quenched with 1.25 M glycine. Cells were then treated with 1X RIPA buffer (Sigma-Aldrich), DNase I (Sigma-Aldrich), RNase A (Sigma-Aldrich), and Halt Protease Inhibitor Cocktail (ThermoFisher). Lysate was added to anti-HA agarose beads (Pierce) and allowed to incubate at 4°C for 24 h. Agarose beads were washed three times were TBST (20 mM Tris, 150 mM NaCl, 0.1% Tween 20), eluted three times with 40 μL 0.1 M glycine (pH 2.0), and neutralized with 4 μL tris (pH 9.5). AAV genomes were then quantified by qPCR, as described above.

BioID candidate knockout/transduction assay

HEK293 cells were seeded in an opaque 96-well plate at a density of 1e5 cells/mL and grown in a 37°C CO₂ incubator for 48 h. Cells were transfected with lentiCRISPRv2, with four guides per gene (250 ng per plasmid). Transfection was accomplished using PEI (PolyScience) at a ratio of 3.5 μg PEI per 1 μg DNA. After selection with puromycin, cells were incubated for 6 days to allow for protein turnover. AAV vector (ss-CBA-Luc or sc-CBh-GFP) was then added to cells at an MOI of 1e4 vg/cell. Luciferase or GFP activity was assayed after 24 h.

Statistical analysis

Statistical analysis was performed using GraphPad Prism software. Statistical calculations comparing two or more groups were made were made using one or two-way ANOVA. Multiple comparisons were then performed, and statistical significance was determined using a Holm-Sidak posttest with an alpha value of 0.05. The number of biological replicates is specified in figure legends. Data is visualized as box and whisker plots, where whiskers represent minimum and maximum values. Correlation was calculated using simple linear regression and an R squared goodness of fit test. All reported replicates were taken from distinct samples. ns = p > 0.05, * = p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Figure error bars represent one standard error.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Tri-Methyl-Histone H3 Lys9 Rabbit mAb	Cell Signaling	D4W1U
Tri-Methyl-Histone H3 Lys4 Rabbit mAb	Cell Signaling	C42D8
IgG XP Isotype Control Rabbit mAb	Cell Signaling	DA1E
Pierce anti-HA agarose	Thermo Fisher	26181
Bacterial and virus strains
AAV8, CBA-Luciferase	This manuscript	N/A
AAV8, CBh-GFP	This manuscript	N/A
AAV8, CBA-GAA	This manuscript	N/A
AAV8, Wildtype	This manuscript	N/A
Avian AAV, CBA-Luciferase	This manuscript	N/A
Avian AAV, CBh-GFP	This manuscript	N/A
Avian AAV, Wildtype	This manuscript	N/A
Avian-8.1, CBA-Luciferase	This manuscript	N/A
Avian-8.1, CBh-GFP	This manuscript	N/A
Avian-8.1, CBA-GAA	This manuscript	N/A
Avian-8.2, CBA-Luciferase	This manuscript	N/A
Avian-8.3, CBA-Luciferase	This manuscript	N/A
Avian-8.3, Wildtype	This manuscript	N/A
Avian-8.3.A, CBA-Luciferase	This manuscript	N/A
Avian-8.3.B, CBA-Luciferase	This manuscript	N/A
Avian-8.3.C, CBA-Luciferase	This manuscript	N/A
Avian-8.4, CBA-Luciferase	This manuscript	N/A
AAV8-Av.1, CBA-Luciferase	This manuscript	N/A
AAV8-Av.2, CBA-Luciferase	This manuscript	N/A
AAV8-Av.3, CBA-Luciferase	This manuscript	N/A
AAV8-Av.3, Wildtype	This manuscript	N/A
AAV8-Av.4, CBA-Luciferase	This manuscript	N/A
Avian-2.1, CBA-Luciferase	This manuscript	N/A
Dragon-8.1, CBA-Luciferase	This manuscript	N/A
Dragon-8.1, CBA-GAA	This manuscript	N/A
Snake-8.1, CBA-Luciferase	This manuscript	N/A
Snake-8.1, CBA-GAA	This manuscript	N/A
AAV5, CBA-Luciferase	This manuscript	N/A
Avian-5.3, CBA-Luciferase	This manuscript	N/A
Sealion AAV, CBA-Luciferase	This manuscript	N/A
Avian-Sn.3, CBA-Luciferase	This manuscript	N/A
Chemicals, peptides, and recombinant proteins
Trichostatin A	Sigma-Aldrich	T8552
CTPB	Active Motif	14064
4-Methylumbelliferyl α-D-glucopyranoside	Sigma-Aldrich	69591
Critical commercial assays
CUT&RUN Assay Kit	Cell Signaling	86652
Chromatin Accessibility Kit	EpigenTek	P-1047–48
Luciferase Assay Reagent	Promega	E1483
NE-PER Nuclear and Cytoplasmic Extraction Kit	Themo Fisher	78833
Tri Isolate RNA Pure Kit	IBI Scientific	IB47630
High-Capacity cDNA Reverse Transcription Kit	Thermo Fisher	4368814
NEBNext Ultra II DNA Library Prep Kit	New England BioLabs	E7103S
Deposited data
AAV8 Cap	GenBank	AY629583
Avian AAV DA-1 Cap	GenBank	NC_006261
AAV2 Cap	GenBank	NC_001401
Bearded dragon parvovirus 2014 Cap	GenBank	KP733794
Snake AAV Cap	GenBank	AY349010
Experimental models: Cell lines
HeLa	ATCC	CRM-CCL-2
A549	ATCC	CRM-CCL-185
U-87 MG	ATCC	HTB-14
HEK293	ATCC	CRL-1573
HuH7	CLS	300156
C2C12	ATCC	CRL-1772
TH1	Kerafast	ECH001
U87 GPR108 KO	This manuscript	N/A
HEK293 AAVR KO	This manuscript	N/A
Experimental models: Organisms/strains
C57BL/6J mice	Jackson Laboratories	000664
Oligonucleotides
ITR qPCR FWD: AACATGCTACGCAGAGAGGGAGTGG	This manuscript	N/A
ITR qPCR REV: CATGAGACAAGGAACCCCTAGTGATGGAG	This manuscript	N/A
Luciferase qPCR FWD: CCTTCGCTTCAAAAAATGGAA	This manuscript	N/A
Luciferase qPCR REV: AAAAGCACTCTGATTGACAAATAC	This manuscript	N/A
Recombinant DNA
pXX680	This manuscript	N/A
pLV-dCas9-P300-PuroR	Hilton et al.25	https://doi.org/10.1038/nbt.3199
pSJR7-ITR-guide-3	This manuscript	N/A
pSJR7-ITR-guide-5	This manuscript	N/A
pSJR7-ITR-guide-scr	This manuscript	N/A
pEL-DA1-VP1u-6x-Linker-BioID2	This manuscript	N/A
pEL-DA1-VP1u-Nterm-AAV8–6x-Linker-BioID2	This manuscript	N/A
LentiCRISPRv2	Addgene	52961
pEL-CMV-IRAK4-HA	This manuscript	N/A
pEL-CMV-EEF1AKMT2-HA	This manuscript	N/A
pEL-CMV-H1F0-HA	This manuscript	N/A
Software and algorithms
GraphPad Prism software, version 10.0.2	GraphPad Software	https://www.graphpad.com/
Unipro UGENE software, version 41.0	Unipro	http://ugene.net/
MUSCLE	Edgar et al.36	https://doi.org/10.1186/1471-2105-5-113
IQ-TREE 2	Minh et al.37	https://doi.org/10.1093/molbev/msaa015
ColabFold	Mirdita et al.38	https://doi.org/10.1038/s41592-022-01488-1
UCSF ChimeraX	Pettersen et al.39	https://doi.org/10.1002/2Fpro.3943
Cutadadapt	Martin et al.40	https://doi.org/10.14806/ej.17.1.200
Bowtie 2	Langmead et al.41	https://doi.org/10.1038/nmeth.1923
Comparative Toxicogenomics Database	Davis et al.42	https://doi.org/10.1093/nar/gkac833

Highlights.

• Avian AAV cannot infect human cells due to transcriptional silencing

• Swapping VP1 N-terminal domain from AAV8 onto avian AAV rescues infection

• N-terminal swap increases transcription, chromatin accessibility, and histone methylation

• VP1 N-terminal domain recruits essential host factors to the AAV genome