An Essential N-Terminal Serine-Rich Motif in the AAV VP1 and VP2 Subunits That May Play a Role in Viral Transcription

Abstract

Adeno-associated virus (AAV) is one of the most researched, clinically utilized gene therapy vectors. Though clinical success has been achieved, transgene delivery and expression may be hindered by cellular and tissue barriers. Understanding the role of receptor binding, entry, endosomal escape, cytoplasmic and nuclear trafficking, capsid uncoating, and viral transcription in therapeutic efficacy is paramount. Previous studies have shown that N-terminal regions of the AAV capsid proteins are responsible for endosomal escape and nuclear trafficking, however the mechanisms remain unknown. We identified a highly-conserved three-residue serine/threonine (S/T) motif in the capsid N-terminus, previously uncharacterized in its role in intracellular trafficking and transduction. Using alanine scanning mutagenesis, we found S155 and the flanking residues, D154 and G158, are essential for AAV2 transduction efficiency. Remarkably, specific capsid mutants show a 5 to 9-fold decrease in viral mRNA transcripts, highlighting a potential role of the S/T motif in transcription of the viral genome.

Introduction

Although the safety and efficacy of AAV vectors have been demonstrated in numerous phase I, II, and III clinical trials involving gene therapy for multiple human diseases,¹ the full therapeutic potential of AAV vectors has yet to be realized. Capsid engineering is one such way to improve cell transduction efficiency, which may provide improved clinical success.² Furthermore, since efficient gene delivery is hindered by major barriers at the cellular level including receptor attachment and entry, post entry trafficking, and nuclear import and viral transcription,³ the AAV intracellular trafficking pathway may reveal ways to develop improved gene therapy vectors.

The process of AAV intracellular trafficking involves initial entry into the cell, endosomal escape, entry into the nucleus, uncoating of the capsid, and release of the transgene.³ Within the AAV capsid, the highly conserved functional domains PLA2 and BR3 – both found in the N-terminus of VP subunits – are essential for endosomal escape and nuclear localization, respectively.^4,5 Other amino acid residues described as determinants of AAV trafficking include four interacting amino acids (E563, H526, R389, and Y704), noted as the pH quartet in AAV8.⁶ While most mutations in these residues show normal titers, significant effects on transduction were observed. Specifically, Y704A resulted in a defect greater than 7 log, despite normal cell entry, nuclear trafficking, and viral uncoating phenotypes. In addition, mRNA levels of Y704A were reduced 450-fold, which suggests failure to facilitate transgene transcription for efficient transduction. These data suggest that AAV capsids may play a role in modulating second-strand synthesis from the single-stranded viral genome and/or transcription from the double-stranded DNA molecule post second-strand synthesis. Furthermore, impaired viral transduction correlated to transcription defects in mutants with normal cellular entry, subcellular fractionation, and uncoating has also been demonstrated in AAV2 mutants D529A, K692A, D528A, and D564A.⁷ Transcription occurred at rates 300-fold lower (D529A and K692A) or 5 logs lower (D528A and D564A) when compared to wild type (wt). Taken together, these results indicate that, in certain capsid mutants, diminished transcription is responsible for a significant decrease in transduction.

Though progress has been made to understanding aspects of AAV intracellular trafficking, several mechanisms remain to be elucidated: What regulates the major processes such as viral trafficking through the cytoskeletal network? How exactly does uptake occur via the nuclear pore complex? How does capsid uncoating release the genome? How does the capsid play a role in viral transcription? Given that post-translational modifications, such as phosphorylation, are often-used biological strategies to alter the intracellular trafficking of proteins, we explored whether commonly modified amino acids (such as serine residues¹) are found in the VP1/2 N-terminus of the AAV2 capsid – the capsid domain known to house other elements involved in viral intracellular trafficking. We focused on one particular stretch of serine/threonine residues, and using alanine scanning mutagenesis, we demonstrate the motif is likely to play a role in viral transcription.

Materials and Methods

Mutant Capsid Gene Cloning

All AAV2 capsid mutants were created using the QuikChange Site-Directed Mutagenesis Kit (Agilent) to modify pXX2, a plasmid containing AAV serotype 2 (AAV2) rep and cap genes. Specifically, alanine scanning mutagenesis was performed on AAV2 capsid amino acid residues 148–161. All mutant capsid genes were verified by Sanger sequencing (Genewiz).

Cell Culture

HeLa and HEK 293T cells were maintained in Dulbecco’s Modified Eagle Medium (Life Technologies) supplemented with 10% fetal bovine serum (FBS, Atlanta-Biologicals) and 1% penicillin and streptomycin (Life Technologies), and cultured at 37°C in 5% CO₂.

Virus Production

Viruses were generated through a triple plasmid co-transfection of HEK 293T cells with pscGFP (self-complementary GFP transgene flanked with AAV2 ITRs), pXX6–80 (adenoviral helper genes), and a plasmid encoding AAV2 rep with either wt AAV2 cap or a mutated AAV2 cap gene. The HEK293T producer cells were seeded in 15cm dishes that had been pre-coated with poly-L-lysine 24h before transfection, which occurred when the cells were 70–80% confluent. Cells were harvested 48h post-transfection, resuspended in media, transferred to a 50ml conical tube, and centrifuged at 1000xg for 15min at 4°C. After aspirating the supernatant, each pellet was then re-suspended in 13ml 1x gradient buffer (GB) containing 10 mM MgCl₂, 150 mM NaCl, 10 mM Tris, at pH

7.6. Cell lysate was stored at −80°C until virus separation.

Frozen cell lysate was thawed in a 37°C water bath then snap frozen in liquid nitrogen twice for a total of three thaws. The thawed solution received 2.5μL benzonase nuclease (250 units/uL; Sigma) and was incubated for 40min in a 37°C water bath. The solution was centrifuged at 3000xg for 20min at 4°C, and the supernatant was subsequently collected for iodixanol density gradient separation (Optiprep, Beckman Beckman #344326, Quick-Seal Ultra Clear 25 × 89 mm centrifuge tubes). Samples were then sealed and centrifuged at 48,000rpm for 1h 45min at 18°C. Viruses were extracted from the 40% iodixanol layer and stored at 4°C.

Quantification of Virus Titers

Genomic viral titers were determined using qPCR. Samples were denatured by adding 10μl 2M NaOH to 10μl of virus, incubated at 75˚C for 30min, then neutralized with 10μl 2M HCl. 10μl of the resulting mixture was then combined with 490μl filler DNA [10ng/μl sheared salmon sperm DNA (Thermo Fisher Scientific) in UltraPure H₂O]. Viral titers were quantified with SYBR green (Applied Biosystems) using a C1000 thermal cycler (Bio-Rad) and primers against the CMV promoter (forward: 5′-TCACGGGGATTTCCAAGTCTC-3′ and reverse: 5′-AATGGGGCGGAGTTGTTACGA-3′).

Western Blot

Viruses were run on a 7% Tris-Acetate gel (Life Technologies) and transferred to nitrocellulose (GE Healthcare) at 30V for 90 min. Blocking was performed in 5% skim milk in phosphate-buffer saline (PBS) with 0.1% Tween-20 (PBS-T) for 1h while rocking. Blots were then rinsed three times and rocked for 20min in PBS-T. Primary antibody B1 (monoclonal mouse anti-VPs, American Research Products, diluted 1:50) was applied to blots overnight at 4°C in PBS with 3% BSA (3% BSA-PBS). After washing with PBS-T, goat anti-mouse peroxidase-conjugated secondary antibody (Jackson ImmunoResearch) was applied at a 1:2000 dilution in 5% skim milk in PBS-T for 1h. Blots were then washed three times for 15min with PBS-T while rocking. Imaging was performed on a Fujifilm LAS 4000 with Lumi-Light Western blotting substrate (Roche).

Cellular Internalization Assay

HEK 293T cells were seeded in 48-well plates. At 90% confluency, cells were transduced with virus at a MOI of 1000 for 2h at 37˚C. After 2h, cells were trypsinized and collected for DNA isolation and quantification. Total cell DNA was isolated using E.Z.N.A Tissue DNA kit (Omega Bio-tek), and DNA concentration was quantified with a NanoDrop spectrophotometer. The number of viral genomes per μg of total DNA was quantified by qPCR.

Subcellular Fractionation Assay

HeLa cells were seeded in 6-well plates and pulse-transduced at 90% confluency at a MOI (multiplicity of infection) of 1000 viral genomes per cell: cells were transduced with AAV2 vectors encoding GFP at 4˚C for 30min and then washed gently with cold PBS and incubated at 37˚C. At 24h post-transduction, cells were harvested by trypsinization. Cytoplasmic and nuclear fractions were separated using a NE-PER Nuclear and Cytoplasmic Extraction Reagent kit (Thermo Scientific). Both fractions were subsequently incubated in PB buffer (Qiagen) containing 3M NaOAc for 10min at room temperature. Samples were purified using a QiaQuick Purification kit (Qiagen). The number of viral genomes in each sample was quantified using qPCR and total DNA was measured with a NanoDrop spectrophotometer.

Nuclear Uncoating Assay

HeLa cells were seeded in 6-well plates and pulse-transduced at 90% confluency at a MOI of 1000. Cells were transduced with AAV2 viruses encoding GFP at 4˚C for 30min. Cells were washed gently with cold PBS and incubated at 37˚C. 48h post-transduction, cells were harvested by trypsinization. Cells were then processed for DNA isolation. DNA was eluted in 40μl in water (Qiagen DNeasy and Blood Tissue kit). To digest viral genomes that did not successfully get converted into circular episomes post-capsid uncoating, we used the T5 exonuclease assay.⁸ 2μg of DNA were sham- and T5 exonuclease-treated (New England Biolabs, M0363S) at 37˚C for 1h and heat-activated at 70˚C for 10m. Digested samples were diluted to 5ng/ml in salmon sperm DNA (Thermo Scientific). Circular viral episomes were quantified with qPCR using GFP primers as described below.

Cellular Transduction Assay

HEK 293T and HeLa cells were seeded in 12- or 48-well plates. After 24h, cells were transduced with AAV2 encoding GFP at a MOI of 1000. For a standard transduction assay, cells were incubated in virus-containing, serum-free media at 37˚C for 12h. At 12 h post-transduction, serum-containing media was added and cells were incubated for an additional 36 h. At 48h post-transduction, cells were harvested for flow cytometry (FACSCanto II) to measure GFP expression. Flow cytometry data were analyzed using FlowJo software, and transduction efficiency was measured by calculating the normalized transduction index (nTI).

Nuclease Protection Assay

Five microliters of virus sample was added to 45μl of Endo Buffer (1.5 mM MgCl₂, 0.5 mg/ml BSA, 50 mM Tris, pH 8.0). After mixing by inversion, 20μl of sample was aliquoted into two PCR tubes, and one was treated with 0.5μl of vehicle control while the other received 0.5μl of benzonase nuclease (250 units/μl; Sigma). The tubes were mixed well by inversion, collected by brief spin and incubated at 37˚C for 30min. To neutralize the enzyme, 0.5μl of 0.5M EDTA was added to each tube. Viral genomes were quantified using qPCR (BioRad, C1000) with GFP primers as described below.

RT-PCR

HeLa cells were seeded in 6-well plates 24h prior to transduction. Cells were transduced with AAV2 virus encoding GFP at a MOI of 1000. At 48h post-transduction, cells were harvested, RNA was extracted with a Qiagen RNeasy Mini kit, samples were treated with DNase, and total RNA was measured by a NanoDrop spectrophotometer (Qiagen). RNA samples were transcribed to cDNA using a Verso cDNA kit (Thermo Scientific). Real-time PCR (Bio-Rad) was used to quantify GFP transcripts with primers against the GFP transgene (forward primer: TGA TGC CAC ATA CGG AAA GC; and reverse primer AAA AGC ACT GCA CGC CAT AG). The housekeeping gene, 18S, was used to normalize the gene expression. GFP expression was compared to wt for absolute quantification.

RESULTS

Conservation of DSSSG N-terminal S/T motif in AAV VP1

We aligned 140 sequences of AAV VP1 available on GenBank (accession numbers in Supplemental File S1) and found a highly conserved stretch of amino acids containing serine/threonine residues in the N-terminus of the VP1 subunit: D154, S155, S156, S157, and G158. G158 appears to be the most conserved across AAV variants with a prevalence of 96.4%. D154, S155, and S156 have frequencies above 90% and S157 displays the lowest conservation at 52.9%. At 40.7%, T is the second most prevalent amino acid at position 157 (Figure 1). Overall, the DSSSG motif, called the N-terminal S/T motif here out, appears highly conserved among AAV variants.

Figure 1. Multiple sequence alignment of 140 AAV VP1 sequences.

(a) A sequence logo illustrates the conservation of D154, S155, S156, S157, and G158 residues in AAV. The bit height profile measures the certainty and frequency levels for conservation of each residue. (b) Prevalence of D154, S155, S156, S157, and G158 residues of AAV.

Impact of alanine substitutions in the N-terminal S/T motif

We generated AAV2 capsid variants with amino acid residues within the N-terminal S/T motif mutated to alanine, and the resulting vectors were characterized for changes in capsid formation and genome protection ability (Figure 2). No significant changes were observed in mutant vector titers compared to wt (Figure 2a), suggesting that alanine mutagenesis of this N-terminal S/T motif does not disrupt viral assembly or genome packaging. Viral capsid subunit incorporation was assessed by western blot, revealing all mutants displayed the expected VP1, VP2, and VP3 band patterns and exhibited capsid compositions akin to that of wt (Figure 2b). We also assessed the ability of the mutants to protect their genomes from external nuclease digestion, and when compared to wt, all mutants displayed similar abilities to protect their encapsidated genomes (Figure 2c). Our results demonstrate that alanine substitutions in the N-terminal S/T motif do not impact capsid formation, genome packaging, or genome protection.

Figure 2. Impact of alanine substitutions in the N-terminal S/T motif.

(a) Virus titers from two independent virus preparations quantified by qPCR. (b) Each well was loaded with 1.5 vg of iodixanol-purified virus. Capsid subunits were probed with the B1 antibody, which detects a C-terminal epitope shared among the VPs. (c) AAV2 alanine mutants were incubated with benzonase nuclease to quantify the protected viral genomes (vgs) using qPCR. Wt AAV2 capsid vector and naked plasmid DNA were included as controls. Error bars represent SEM of two independent experiments performed in duplicate.

Differential transduction efficiency of AAV2 N-terminal S/T motif mutants

Next, we determined the transduction efficiencies of the alanine mutant panel (Figure 3). When compared to wt, D154A, S155A, G158A, and S155–7A exhibit significantly decreased transduction efficiencies, while S157A shows behavior on par with wt in both HEK 293T and HeLa cells. D154A and G158A mutants display the most dramatic reductions in transduction – greater than a 10-fold decrease compared to wt. Of the serine mutants, S155A and S155–7A exhibit 9- and 15-fold decreases in HeLa cells, respectively, compared to that of wt. The S156A mutant shows approximately a 2-fold decrease in HeLa cells but not in HEK293T cells, implying that results may be cell type-dependent. The results thus far demonstrate that residues D154 and G158 are critical for viral transduction, and S155 may be the most essential residue of the serine triplet.

Figure 3. Differential transduction efficiency of AAV2 N-terminal S/T motif mutants.

(a) HEK293T and (b) HeLa cells were treated with vectors at a multiplicity of infection (MOI) of 500. Flow cytometry was conducted 48 h post-transduction. The transduction Index (TI) was calculated based on %GFP-positive cells multiplied by geometric mean fluorescence intensity. The TI was normalized to wt (nTI). Error bars represent SEM of two independent experiments performed in duplicate. *P < 0.05 by one-way ANOVA analysis.

Mutation to N-terminal S/T motifs does not affect cellular internalization or nuclear uptake

To test for cellular uptake of the alanine mutant panel, we quantified viral genome internalization in HeLa cells (Figure 4). After a two hour incubation with the viruses, the number of viral genomes in the cells was determined. Interestingly, all mutants internalized into cells at a similar rate as the wt capsid. These results suggest that the N-terminal S/T motif does not play a major role in cellular uptake of the vectors.

Figure 4. Mutation to N-terminal S/T motifs does not affect cellular internalization.

Vectors at an MOI of 1000 were added to HeLa cells. Cells were harvested 2 h post-transduction and qPCR was used to quantify the number of internalized vgs normalized to total DNA content. Error bars represent SEM of two independent experiments performed in duplicate.

To assess nuclear uptake of the alanine mutants, we performed a subcellular fractionation and quantified the number of viral genomes in the nucleus versus the cytoplasm of cells at 24 hours post-transduction (Figure 5). A comparison of absolute numbers of viral genomes reveals that all mutants are able to enter the nucleus as effectively as wt (Figure 5a). All viruses display similar levels of compartmentalization between the nucleus and cytoplasm, with slightly more than half of the viral genomes located in the cytoplasm (Figure 5b). Taken together, the mutants described here do not display significant differences in cellular internalization or subcellular fractionation when compared to wt.

Figure 5. Mutation to N-terminal S/T motifs does not affect nuclear uptake.

Vectors at an MOI of 1000 were added to HeLa cells, and cells were harvested 24h post-transduction. After harvest, cytoplasmic and nuclear extracts were isolated for quantification of DNA. (a) The number of vgs in the cytoplasm versus the nucleus was quantified by qPCR and normalized to total DNA. (b) The fraction of intracellular viral genomes is shown for the cytoplasm and nucleus. Error bars represent SEM of two independent experiments performed in duplicate.

N-terminal S/T motif mutation has no impact on nuclear uncoating but inhibits transcription

We next determined if any of the alanine mutations affected nuclear uncoating. We used the T5 exonuclease assay⁸ to detect viral genomes that are episomal, indicating successful capsid uncoating, second strand synthesis, and circularization of the double-stranded DNA. Episomes would be resistant to T5 exonuclease digestion, therefore T5-resistant genomes reflect the population of viral genomes that have successfully uncoated. After 48h post-transduction, cells were assessed to quantify the number of T5-resistant viral genomes (Figure 6a). All mutants demonstrated similar uncoating compared to wt. These data suggest that capsid uncoating is an unlikely cause for the defective transduction efficiencies observed in the mutants. Furthermore, similar levels of genome circularization suggest the ITRs, which have previously been identified as playing an important role in circularization,⁹ are intact in the mutants.

Figure 6. N-terminal S/T motif mutation has no impact on nuclear uncoating but inhibits transcription.

Vectors at an MOI of 1000 were added to HeLa cells, and cells were harvested 48h post-transduction. After cell harvesting, (a) DNA was isolated for sham- and T5 exonuclease-treatment. qPCR was then used to quantify the number of vgs remaining. Error bars represent SEM of two independent experiments performed in duplicate. Circular plasmid DNA encoding GFP (GFP only) was used as positive control. (b) RNA was extracted and relative GFP mRNA expression of the mutants relative to wt AAV2 capsid vector was determined by qPCR. Error bars represent SEM of four independent experiments done in duplicate. *P < 0.05 by one-way ANOVA analysis.

Following the nuclear uncoating assay, we measured GFP mRNA expression levels, which revealed defective GFP mRNA expression in all mutants (Figure 6b). Of the mutant panel, S157A retained the highest level of GFP mRNA expression, which was the closest to wt levels. For the remaining serine mutants, S155A, S156A, and S155–7A, GFP mRNA expression decreased nearly 5-fold, while the most robust decrease, about 9-fold, occurred in the flanking residue mutants D154A and G158A when compared to wt. It is of note that when compared to S156A, the S155A mutant has substantially decreased GFP protein expression but similar GFP mRNA expression. The explanation for this discrepancy is currently unclear. Overall, these results suggest that defects in viral transcription leads to diminished transduction efficiency of the mutants.

DISCUSSION

Exploration of the AAV2 VP1/2 N-terminus by alanine scanning mutagenesis of a triple serine motif and its flanking residues has uncovered essential amino acids that are necessary for transduction. In particular, serine residues of a triplet motif located in positions 155–157, including the flanking residues, D154 and G158 are important for efficient transduction. With the exception of S157A, most of the mutants (D154A, S155A, S156A, S155–7A, and G158A) displayed reduced transduction efficiency, warranting the study of viral intracellular trafficking to elucidate the mechanism of impaired efficiency.

In this study, major defects in viral genome transcription may explain the diminished transduction efficiency. Salganik et al., in an exploration of mutant phenotypes in the pH quartet of AAV8,⁶ also found another mutant (Y704A) that displayed reduced transduction efficiency as the result of impaired viral transcription. It is of further interest to understand whether these capsid mutations prevent critical host cellular proteins from binding to the capsid, thus inhibiting important intracellular mechanisms, or conversely if the mutations recruit undesired host factors that then hinder viral transcription. Currently, the biological mechanism explaining the transcriptional defects of the N-terminal S/T mutants or the Y704A mutant, and whether the same mechanism is affected in both mutational contexts, is unknown.

Although the mechanistic processes involving serine/threonine/tyrosine capsid residues remain to be elucidated, some studies have been conducted to understand the role of these residues in AAV intracellular transport. One such study suggests that AAV nucleolar trafficking modulates transgene expression,¹⁰ with potential effects of AAV nuclear location on viral infectivity. Other groups have reported the importance of cellular proteins in AAV transduction, including FKBP52,¹¹ a phosphatase that mediates dephosphorylation of serine/threonine residues. Zhao et al. found that both cytosolic and nuclear FKBP52 attaches to the D-sequence of AAV inverted terminal repeats (ITRs) to participate in AAV intracellular trafficking.¹¹ Cytosolic FKBP52 facilitates AAV cellular transport and nuclear FKBP52 inhibits second-strand synthesis. Interaction of host factors with the capsid may result in increased susceptibility of AAV to degradation by ubiquitination or proteasomal machinery, thereby negatively impacting transduction efficiency.^12–14 Modification of exposed tyrosine residues on the viral capsid to phenylalanine dramatically affects transduction and potentially cell type tropism.^15,16 These types of capsid changes can also lead to reductions in hepatotoxicity associated with major histocompatibility complex class I presentation.¹⁷ In addition to phosphorylation, putative sites for N- and O-linked glycosylation within the VP1/2 region of the AAV capsid have been identified, including SGXG and SGLG motifs at positions 157 and 195, respectively.¹⁸ We do not currently have any evidence demonstrating that the 154–158 motif investigated in this study is being post-translationally modified, and further studies will be required to answer this important question.

An important consideration in the design of any AAV vector is determining whether a capsid motif that plays a role in one context also plays a similar role in other contexts. For instance, at this point it is unclear if the N-terminal S/T motif plays an important role in in vivo transduction by AAV, and if state of the host – healthy versus disease – impacts the biological mechanism at play. It is also unclear if the observed effects are cell/tissue-type dependent or AAV capsid serotype-dependent. Future studies aimed at uncovering these context-specific questions would help inform the design of improved AAV capsids.

CONCLUSIONS

In conclusion, we have investigated an S/T domain in the N-terminus of the VP1/2 capsid subunit and found it is essential for AAV transduction. Further mechanistic studies are needed to uncover the biological mechanism behind the observations and to gain a better understanding of AAV intracellular events. Phosphorylation of the AAV capsid appear to impact the fate of the virus intracellularly,^13,14,19 but it is unclear at this point if the S/T motif investigated here is being phosphorylated. Continued research on potential post-translational modification of the AAV capsid and its impact on transduction efficiency in vitro and in vivo will broaden our understanding of AAV biology – knowledge which can help inform new ways to develop improved AAV vectors for therapeutic use.