Adeno-associated Virus (AAV) Serotypes Have Distinctive Interactions with Domains of the Cellular AAV Receptor

ABSTRACT Adeno-associated virus (AAV) entry is determined by its interactions with specific surface glycans and a proteinaceous receptor(s). Adeno-associated virus receptor (AAVR) (also named KIAA0319L) is an essential cellular receptor required for the transduction of vectors derived from multiple AAV serotypes, including the evolutionarily distant serotypes AAV2 and AAV5. Here, we further biochemically characterize the AAV-AAVR interaction and define the domains within the ectodomain of AAVR that facilitate this interaction. By using a virus overlay assay, it was previously shown that the major AAV2 binding protein in membrane preparations of human cells corresponds to a glycoprotein with a molecular mass of 150 kDa. By establishing a purification procedure, performing further protein separation by two-dimensional electrophoresis, and utilizing mass spectrometry, we now show that this glycoprotein is identical to AAVR. While we find that AAVR is an N-linked glycosylated protein, this glycosylation is not a strict requirement for AAV2 binding or functional transduction. Using a combination of genetic complementation with deletion constructs and virus overlay assays with individual domains, we find that AAV2 functionally interacts predominantly with the second Ig-like polycystic kidney disease (PKD) repeat domain (PKD2) present in the ectodomain of AAVR. In contrast, AAV5 interacts primarily through the first, most membrane-distal, PKD domain (PKD1) of AAVR to promote transduction. Furthermore, other AAV serotypes, including AAV1 and -8, require a combination of PKD1 and PKD2 for optimal transduction. These results suggest that despite their shared dependence on AAVR as a critical entry receptor, different AAV serotypes have evolved distinctive interactions with the same receptor. IMPORTANCE Over the past decade, AAV vectors have emerged as leading gene delivery tools for therapeutic applications and biomedical research. However, fundamental aspects of the AAV life cycle, including how AAV interacts with host cellular factors to facilitate infection, are only partly understood. In particular, AAV receptors contribute significantly to AAV vector transduction efficiency and tropism. The recently identified AAV receptor (AAVR) is a key host receptor for multiple serotypes, including the most studied serotype, AAV2. AAVR binds directly to AAV2 particles and is rate limiting for viral transduction. Defining the AAV-AAVR interface in more detail is important to understand how AAV engages with its cellular receptor and how the receptor facilitates the entry process. Here, we further define AAV-AAVR interactions, genetically and biochemically, and show that different AAV serotypes have discrete interactions with the Ig-like PKD domains of AAVR. These findings reveal an unexpected divergence of AAVR engagement within these parvoviruses.

immortalized cell lines derived from a variety of human tissues were isolated and probed in this assay (Fig. 1A). The 150-kDa AAV-BP was detected strongly in membrane fractions of AAV2-permissive cells (HeLa S3, K562, HeLa, HEK293, KB, and Hep-2 cells) but was weakly detected or undetectable in cell lines known to be less permissive to AAV2 transduction (HL60 and UT7/Epo cells) (17,21,24). We went on to purify AAV-BP from the cell membrane fraction of HeLa cells, optimizing the purification process by using Pisum sativum agglutinin (PSA) lectin-coated beads and then jacalin-conjugated beads, which bound AAV-BP well. Upon mass spectrometry (MS) analysis of the protein excised from the gel in a region corresponding to where AAV binding activity was detected, we identified peptide sequences from a number of proteins, including the low-density lipoprotein receptor precursor (LDLR), apolipoprotein E receptor 2 (ApoER2), AAVR (KIAA0319L), oxygen-regulated protein 150 (ORP150), and integrin ␣5 (Fig. 1B). By using specific antibodies against each of the proteins identified, only those against AAVR precisely overlapped the region that displayed AAV2 binding activity in the virus overlay assay ( Fig. 1C and D). These results indicate that the protein with the strongest binding activity in a virus overlay assay (AAV-BP) is identical to AAVR, the multiserotype receptor identified in an unbiased genetic screen for AAV2 transduction (21). This was further validated by an AAV2 virus overlay assay using isogenic knockout cells lines (created by using CRISPR-mediated genome editing), where the 150-kDa Membrane proteins were extracted from various types of cells, and 100 g of membrane proteins was used to perform a virus overlay assay with purified wild-type AAV2 particles (56). The arrow indicates a strong binding band at 150 kDa, designated AAV-BP. (B) Summary of the top five genes that correspond to the peptide sequences from the mass spectrometry analysis of the AAV-BP band. AAVR is also denoted KIAA0319L. (C) One hundred micrograms of PSA-purified HeLa S3 membrane proteins was separated on a 2-D gel and transferred onto a PVDF membrane for a virus overlay assay with rAAV2, followed by reprobing with anti-LDLR, anti-ApoER2, anti-ORP150, or anti-AAVR antibody. (D) One hundred micrograms of N-and O-deglycosylated crude HeLa S3 cell membrane proteins was separated by 2-D electrophoresis and underwent a virus overlay assay, followed by reprobing with a rabbit polyclonal antibody to integrin ␣5. Squares indicate an identical area of the membrane.
band of AAV-BP was detectable in wild-type (WT) but not in AAVR knockout (AAVR KO ) cells ( Fig. 2A and B). AAVR glycosylation is not essential for AAV2-AAVR interactions or AAV2 transduction. AAVR has a predicted protein molecular mass of 108 kDa, but it is detectable at 150 kDa after immunoblotting (21), likely due to the effect of N-and O-linked glycosylation on protein migration during gel electrophoresis. To assess the importance of AAVR glycosylation on its interaction with AAV2, we treated cellular membrane preparations with deglycosylation enzymes (Fig. 3A and B). N-Glycosidase treatment led to a shift in migration, implying that AAVR is an N-linked glycosylated protein. Concomitant treatment with N-and O-linked glycosidases and neuraminidase (both of the latter two enzymes are required to cleave O-linked disaccharides) led to a further shift, suggesting that AAVR is also an O-linked glycosylated protein. None of the deglycosylation treatments reduced the binding of AAV2 in a virus overlay assay (Fig.  3A), demonstrating that glycosylation is not required for binding. To test the effect of N-linked glycosylation in a functional transduction assay, we used the stable expression of AAVR mutants in HeLa AAVR KO cells. We have previously shown that the expression of an AAVR minimal mutant (miniAAVR), consisting of a signal peptide, PKD domains 1 to 3 of the ectodomain, the transmembrane domain, and the C-terminal cytoplasmic region, is able to efficiently rescue AAV2 transduction in AAVR KO cells (21). This indicates that these PKD domains are sufficient for binding AAV2 and mediating entry. Within the ectodomain of miniAAVR, there are 5 putative asparagine glycosylation sites (indicated in Fig. 3C), which were mutated to alanine (individually or in combination) to prevent N-linked glycosylation. Mutation N525A and the quintuple mutant mildly affected AAV2 transduction, whereas the other mutations did not appreciably affect the transduction efficiency compared to that of the wild-type construct (Fig. 3D). Thus, by this cellular assay, N-linked glycosylation of AAVR facilitates optimal AAV2 transduction but is not strictly required.
AAVR PKD2 is critical for the interaction of AAV2 with AAVR. We previously identified AAVR PKD domains 1 to 3 to include the binding domain for AAV2 (21). To further characterize the AAVR-AAV2-interacting region, we created glutathione S-transferase (GST)-tagged AAVR ectodomain mutants consisting of individual PKD domains or combinations of sequential domains (Fig. 4A). We expressed the respective mutants in Escherichia coli and performed a virus overlay assay, where we observed binding of AAV2 to PKD domains 1 to 5, PKD domains 1 and 2, and PKD domain 2 alone ( Fig. 4B and C). In addition, AAV2 transduction of AAVR KO cells stably expressing AAVR deletion mutants consisting of individual domain deletions ( Fig. 4D and F) demonstrated that only the deletion of PKD2 completely abrogated AAV2 transduction (Fig. 4E) in transduction efficiency. These data indicate that PKD2 is the critical region for AAV2 binding and transduction, with PKD1 playing a possible accessory role.
To further define the region of PKD2 necessary for AAV2 transduction, we constructed domain swap mutants in the AAVR homologue, KIAA0319, which shares approximately 60% amino acid sequence similarity with AAVR. AAVR is known to be KIAA0319-like because of its structural likeness to KIAA0319 (Fig. 5A). The two proteins both comprise a MANEC domain along with 5 PKD domains in their ectodomain and are both glycosylated, type I transmembrane proteins (25). Moreover, PKD1 and PKD2 of KIAA0319 share 72% and 80% similarities, respectively, with the corresponding PKD domains in AAVR (Fig. 5B). Importantly, despite its similarity to AAVR, KIAA0319 cannot be utilized by AAV2 to successfully transduce cells (Fig. 5C). Thus, we utilized KIAA0319 as a backbone in which regions of KIAA0319 were replaced with AAVR ( Fig. 5A, D, and E). We then stably expressed these mutants in AAVR KO cells and assessed AAV2 transduction. A swap-in of AAVR PKD domains 1 to 3 into KIAA0319 (mutant a) and PKD2 alone (mutant b) fully rescued AAV2 transduction, whereas a swap-in of PKD1 alone (mutant g) did not rescue transduction (Fig. 5C). This is consistent with the binding and deletion mutant data and reinforces the major role of AAVR PKD2 for AAV2. Additionally, a swap-in of only an N-terminal region of AAVR PKD2 (ϳ30 amino acids [aa]) partially rescued transduction, indicating that this region is particularly important for AAV2 transduction (mutant c). Reiterating this point, swapping of the same region of KIAA0319 into AAVR (mutant f) eliminated the capacity of AAVR to rescue AAV2 transduction. Swapping of the C-terminal region of AAVR PKD2 into KIAA0319 (mutants d and e) did not rescue transduction. Using a collection of deletion constructs of PKD2, we observed that none of the mutants, even the ones containing small N-and C-terminal PKD2 deletions ( Fig. 6A and B), were able to bind to AAV2 in a virus overlay assay ( Fig. 6C and D). Thus, N-terminal of AAVR PKD2 is able to partially rescue AAV2 transduction in the domain swaps yet is not able to bind AAV2 particles in the in vitro binding assay. This suggests that full PKD2 of AAVR is required for optimal binding and transduction. AAV5 requires AAVR PKD1 for binding and functional transduction. Our previous study demonstrated that AAVR is an essential receptor for multiple AAV serotypes (21), including AAV5, which is evolutionarily distant from AAV2 and shows only 57% identity in the capsid amino acid sequence (26)(27)(28). To explore the hypothesis that the interaction interfaces of different AAV serotypes with AAVR are distinct, we used the collection of domain swap mutants and probed AAV5 transduction. In stark contrast to the results obtained with AAV2, neither full-length PKD2 (mutant b) nor the N-terminal   mutant (mutant c) rescued AAV5 transduction when swapped into KIAA0319 (Fig. 7A). Additionally, the AAVR mutant with a partial PKD2 swap of KIAA0319 (mutant f), which completely lost its ability to facilitate AAV2 transduction, was fully competent in mediating AAV5 transduction. Interestingly, when the most membrane-distal PKD domain, PKD1, was swapped into KIAA0319 (mutant g), we observed a rescue of transduction to the same levels as those with full-length AAVR, suggesting that AAV5 evolved to require PKD1 rather than PKD2 for functional transduction (Fig. 7A). In agreement with this, an AAV5 virus overlay assay using individual AAVR PKD ectodomain mutants demonstrated that AAV5 bound specifically to AAVR PKD1 and not PKD2 (Fig. 7B). Consistent with data from the virus overlay assay and swap experiments, AAV5 transduction of AAVR KO cells stably expressing AAVR single deletion mutants confirmed the strict requirement of AAVR PKD1 for AAV5 transduction (Fig. 7C). To test the differential requirements for PKD domain usage further, we used a virus transduction assay, where increasing amounts of soluble variants of AAVR domains were incubated with cells during AAV2 or AAV5 transduction. The transduction of both AAV2 and AAV5 was inhibited by increasing amounts of PKD1-5, while the addition of PKD3 had no    (Fig. 7D). AAV2 transduction was potently blocked by PKD2 and, to a lesser degree, by PKD1, in concurrence with the deletion mutant data (Fig. 4E). Conversely, AAV5 transduction was potently blocked by PKD1 and unaffected by the presence of PKD2 at low concentrations, in agreement with the transduction and virus overlay data (Fig. 7D). However, at a higher concentration of soluble PKD2, we observed a significant decrease in AAV5 transduction (Fig. 7E), suggesting that PKD2 may play a minor role in the AAV5-AAVR interaction. The functional relevance of this is unclear because the deletion of PKD2 does not reduce AAV5 transduction (Fig. 7C). Together, these results imply complexity in the interaction of AAVR with different serotypes, where AAV2 utilizes PKD2 and, to a lesser degree, PKD1 to enable vector transduction, whereas AAV5 primarily requires PKD1. Serotype-specific utilization of AAVR PKD domains 1 and 2. In view of the disparity in AAVR domain dependence between AAV2 and AAV5, we evaluated the transduction of AAV1 (Fig. 8A) and AAV8 (Fig. 8B) in AAVR KO cells stably expressing   AAVR single deletion mutants. Notably, AAV1 and AAV8 were similar to AAV2 in their critical requirement for PKD2, although they appeared to depend more strongly on PKD1 than AAV2 did for transduction. Both AAV1 and AAV8 bound to only PKD2 in a virus overlay assay ( Fig. 8C and D), suggesting a higher affinity for PKD2 in this assay.

DISCUSSION
A thorough understanding of how AAV interacts with its multiserotype receptor AAVR has the potential to lead to improvements in future vector design and modulations in tissue-targeting strategies. In this study, we establish that AAV-BP, the component of cell membrane extracts reported 2 decades ago to have the greatest AAV2 binding activity, is identical to the multiserotype receptor, AAVR. We further uncover that several different AAV serotypes have evolved distinct interactions for engaging the same receptor, AAVR, upon which they all have a strict dependence.
It has long been established that AAV2 attaches to cells via interactions with heparan sulfate proteoglycan (10). However, it was unclear if a proteinaceous receptor was a critical requirement for AAV2 infection. We recently identified AAVR in an unbiased genetic screening approach as being an essential receptor required for infection by AAV2 as well as by several other AAV serotypes (21). To further characterize AAVR, we revisited the first reported AAV receptor study (17) to determine if characteristics of the previously described AAV2 binding protein (AAV-BP) matched those of AAVR. AAV-BP, a highly glycosylated protein with a molecular mass of 150 kDa, was found to have dominant binding to AAV2 in cell membrane extracts from several human cell types. We show here, using mass spectrometry, modified virus overlay assays, and deglycosylation treatments, that AAVR is identical to the previously described AAV-BP. Moreover, using isogenic AAVR knockout cell lines, we unambiguously demonstrate that AAV-BP is present only in cells that express AAVR. This study thus provides an independent yet complementary approach to the identification of AAVR as an essential AAV2 receptor, demonstrating that the traditional virus overlay assay combined with a membrane protein purification procedure can still be a powerful approach to identify viral receptors. Notably, several other proteinaceous coreceptors for AAV2 have been described (18)(19)(20), but none scored in the genetic screens or are implicated in the current proteomic analysis of virus overlays.
Although the role of AAVR in cellular physiology is unclear, the structure indicates that it is a glycoprotein, containing Ig-like PKD domains in its ectodomain (21,29). Many glycoproteins that contain repeats of Ig-like domains have been identified as viral receptors, including the human rhinovirus (HRV) major receptor, ICAM-1, and the coxsackievirus/adenovirus receptor, CAR (23). Here, we observed that the enzymatic removal of both O-and N-glycosylation decreased AAVR to a molecular mass of around 100 kDa, approximately the predicted molecular mass without modifications, which suggests that AAVR itself is not a heparan sulfate proteoglycan. We show that different AAV serotypes specifically interact with the Ig-like PKD domains of AAVR but demonstrate that glycosylation is not a prerequisite for AAV2 binding and transduction. This is similar to what has been noted for other nonenveloped viruses such as HRV and coxsackievirus, where the interactions between the virus and Ig-like domains in their receptors occur independently of N-linked glycosylation (30,31). Interestingly, our glycosylation studies showed a slight reduction in functional transduction upon mutation of the Asp 525 residue located in AAVR PKD3 as well as in the mutant carrying all 5 putative glycosylation sites. Because PKD3 is not critical for AAV2 transduction per se (Fig. 4E), we hypothesize that this observed reduction could be due to a disturbed folding or trafficking of AAVR, which potentially might require N-linked glycosylation at this site.
AAVR shares striking homology with KIAA0319, a candidate gene that may contribute to dyslexia (25,32). Specifically, the ectodomain of KIAA0319 also contains 5 Ig-like PKD domains, which have an implied role in cell-to-cell adhesion (22,33), and KIAA0319 is a highly glycosylated, transmembrane protein. Despite these similarities, AAV cannot utilize KIAA0319 instead of AAVR to gain entry into the cell. Thus, to determine which domain of AAVR is critical for transduction, we took advantage of the homology between AAVR and KIAA0319 and swapped regions of AAVR into KIAA0319, and vice versa, evaluating which mutants would rescue AAV transduction. Notably, we found that although AAVR is required for multiple serotypes, its interaction with AAV differs between serotypes. We show that AAV has evolved to interact specifically with the most membrane-distal domains (PKD1 and PKD2). In particular, AAV2 functionally interacts mainly with PKD2, with some contribution by PKD1, while AAV5 predominantly interacts with PKD1. It is interesting to note that, similarly to AAV, other viruses that engage receptors with Ig-like domains also utilize the most membrane-distal domain (34)(35)(36)(37)(38)(39) as their contact site. For example, ICAM-1 contains 5 Ig-like domains (D1 to D5), of which the most membrane-distal domain (D1) directly interacts with HRV14, D2 contributes somewhat to binding, and D3 to D5 provide spacing to allow accessibility to the binding sites (38,39).
The difference in AAVR interactions among the various serotypes is intriguing, as we have tested serotypes from different clades. AAV1 (clade A) and AAV2 (clade B) functionally interact mainly with PKD2, with some contribution by PKD1, while AAV5 (no designated clade) predominantly interacts with PKD1, and AAV8 (clade E) requires both PKD1 and -2 to transduce cells. These striking differences in AAV-AAVR functional interactions could be related to the origin from which the serotype was initially isolated, as AAV5 was isolated from human tissues (40,41), while many other AAVs, including AAV1 and AAV2, were first found as contaminants in laboratory adenovirus stocks (42)(43)(44)(45)(46), and AAV8 was isolated from nonhuman primate samples (47). Additionally, AAV5 speciated early and is quite distinct from the evolutionary lineage to which the other AAV serotypes belong (28). As sequence diversity between AAV serotypes has allowed AAV to evolve footprints that preferentially interact with different glycan receptors, such as heparan sulfate and sialic acid (9), our results suggest that AAV serotypes also evolved distinct interactions with AAVR. This may or may not be related to their glycan dependence, something to be investigated in future studies for a more in-depth understanding of virus-receptor engagement. Moving forward, structural studies that complement the data presented here will further define the AAV-AAVR interaction domain in greater molecular detail and will be necessary to establish the exact binding sites, both in AAVR and on the viral capsids.

MATERIALS AND METHODS
Cell lines and primary cells. All cell lines were grown in medium supplemented with 10% fetal calf serum (FCS) (Sigma, St. Louis, MO), 100 IU/ml penicillin-streptomycin (Sigma, St. Louis, MO), and 2 mM L-glutamine (Sigma, St. Louis, MO) and grown in a humidified incubator at 37°C with 5% CO 2 . KB, HeLa, HeLa S3, 293, Hep-2, K562, and HL60 cells were purchased from the American Type Culture Collection (ATCC) (Manassas, VA). HeLa S3 cells were cultured as a suspension in F-12 medium, and other cells were maintained as a monolayer in Dulbecco's modified Eagle's minimal medium (DMEM). UT7/Epo cells were provided by N. Komatsu (Jichi Medical School, Japan) and cultured in DMEM with the same supplements as those described above plus erythropoietin (EPO) at 1 U/ml. The isogenic AAVR knockout cell lines used in this study were generated by using CRISPR-Cas9 technology and were described previously (21). They were maintained in the same way as parental cell lines. The AAVR KO cells referred to in the genetic experiments shown in Fig. 3 to 8 are HeLa AAVR KO cells.
Viruses and recombinant vectors. WT AAV2 was propagated and purified by CsCl ultracentrifugation, performed as previously reported (48). The protein concentration of the virus preparation was measured by using the bicinchoninic acid (BCA) protein assay reagent (Pierce Co., Rockford, IL).
Enzymatic The fractions with AAV2 binding activity were combined and concentrated by ultrafiltration. The above-described concentrated protein preparation was further separated by isoelectric focusing (IEF) at pH 4 to 6, followed by a second sodium dodecyl sulfate (SDS)-6% polyacrylamide gel electrophoresis (PAGE) step. AAV-BP was located by performing a virus overlay assay. The protein spot with binding activity (at pI ϳ4.5) was cut off and subjected to protein identification. Protein identification was performed at the Harvard Microchemistry Facility (Cambridge, MA) by microcapillary reverse high-performance liquid chromatography (HPLC)-nanoscale electrospray tandem mass spectrometry (LC-MS/MS) on a Finnigan LCQ quadrupole ion trap mass spectrometer.
Two-dimensional gel electrophoresis of cell membrane proteins. Each aliquot of the sample (100 l) was dialyzed overnight against 5 mM Tris-HCl (pH 6.8) using a 6,000-to 8,000-molecular-weight (MW)-cutoff (MWCO) membrane at 5°C. Each aliquot was lyophilized and redissolved in 50 l of a 1:1 dilution of SDS boiling buffer-urea sample buffer before loading. Two-dimensional (2-D) electrophoresis was performed according to the carrier ampholyte method of isoelectric focusing (50,51) by Kendrick Labs, Inc. (Madison, WI), as follows. Isoelectric focusing was carried out in a glass tube with an inner diameter of 2.3 mm using 2% pH 3 to 10 or pH 4 to 6 Servalytes (Serva, Heidelberg, Germany) for 9,600 V · h. One microgram of an IEF internal standard, tropomyosin, was added to the sample. This protein migrates as a doublet with a lower polypeptide spot at a MW of 33,000 and pI 5.2. The enclosed tube gel pH gradient plot for this set of Servalytes was determined with a surface pH electrode. After equilibration for 10 min in buffer "O" (10% glycerol, 50 mM dithiothreitol, 2.3% SDS, and 0.0625 M Tris [pH 6.8]), each tube gel was sealed to the top of a stacking gel, which was on top of an 8% acrylamide slab gel (0.75 mm thick) poured with lower-gel buffer at pH 8.0. SDS slab gel electrophoresis was carried out for about 10 h at 15 mA/gel. After slab gel electrophoresis, the gel was stained Coomassie brilliant blue R-250 or silver stained. The gels for blotting were placed into transfer buffer (10 mM N-cyclohexyl-3-aminopropanesulfonic acid [CAPS] [pH 11], 10% methanol [MeOH]) and transblotted onto a polyvinylidene difluoride (PVDF) membrane overnight at 200 mA and approximately 100 V per two gels.
To generate the PKD deletion constructs of AAVR, two PCR products were generated for Gibson cloning. The extreme 5=-end primer (5=-GACTCTAGTCCAGTGTGGTG-3=) and 3=-end primer (5=-ATCCAG AGGTTGATTGTCGAG-3=) were similar for the constructs. Primers used to amplify the fragments for the pLenti-CMV-Puro-DEST vector and used as described previously (52). Lentivirus was produced by using HEK293 cells and was utilized to transduce the respective cell lines overnight. Cells stably expressing the gene of interest were selected by treatment with 1 to 3 g/ml puromycin over 2 days (InvivoGen).
Immunoblot analysis. Proteins were separated by SDS-PAGE or 2-D electrophoresis and electrotransferred onto a nitrocellulose or PVDF membrane. Membranes were blocked by incubation with 1ϫ phosphate-buffered saline (PBS) containing 5% NFT for 1 h at room temperature (RT). Membranes were subsequently incubated overnight at 4°C with primary antibodies in blocking buffer. Membranes were then washed 3 times for 5 min by using wash buffer (1ϫ PBS with 0.1% Tween 20) and further incubated with HRP-conjugated secondary antibodies (anti-mouse and anti-rabbit at 1:5,000 in blocking buffer) (GeneTex) for 1 h at RT. After another set of three washes, antibody-bound proteins were visualized on film by using West Pico and Extended Duration chemiluminescence peroxide solutions (Thermo-Scientific, USA). rAAV transduction assay. Cells were seeded at 10,000 cells/well (96-well plate) overnight. They were then infected with scAAV2-CMV-RFP or scAAV5-CMV-GFP at a multiplicity of infection (MOI) of 20,000 viral genomes (vg)/cell, in complete DMEM. Virus infectivity was determined at 48 h postransduction by measuring transgene expression (red fluorescent protein [RFP] or green fluorescent protein [GFP]) by using flow cytometry. All transductions were performed in triplicate. To measure virus transgene expression (RFP or GFP) in all other experiments, cells were trypsinized 48 h after infection, and a BD LSR Fortessa flow cytometer (BD, Franklin Lakes, NJ, USA) was used to detect fluorescent cells. Data were analyzed and assembled by using FlowJo software (TreeStar, Inc., Ashland, OR, USA).
Virus neutralization assays. (i) Luciferase assay (Fig. 7D). HeLa cells were seeded into a 96-well plate at 1 ϫ 10 4 cells per well 1 day prior to transduction. Purified PKD or the GST protein control was mixed with rAAV(Luc/mCherry) in a volume of 100 l of PBS (pH 7.4) at various concentrations for 30 min at room temperature. At the time of infection, the medium was removed, and 100 l of the rAAV(Luc/ mCherry)-protein mixture was added to the wells of cells (at an MOI of 7,500 vg per cell). At 3.5 h postinfection at 37°C with 5% CO 2 , the inoculum was removed, the cells were washed with DMEM with 10% FCS briefly, and 100 l of medium was added to each well. The cells were incubated for 48 h, followed by quantification of luciferase expression. Luciferase activity was quantified by using a luciferase assay system kit (catalogue number E1500; Promega, Madison, WI) according to the manufacturer's instructions, with measurements being taken on a Synergy H1 microplate reader (BioTek, Winooski, VT). The construction and purification of the His-tagged PKD1, PKD2, and PKD3 proteins used in this experiment are described above. The maltose binding protein-PKD1-5 fusion protein used as a control was described previously (21).
(ii) Fluorescence assay (Fig. 7E). HeLa cells were seeded into 96-well plates at 10,000 cells per well overnight. Purified, soluble His-tagged AAVR PKD domain constructs were introduced to the medium at the specified concentrations. Cells were then transduced with scAAV-CMV-GFP or scAAV5-CMV-GFP at an MOI of 12,000 vg per cell and incubated for 24 h at 37°C. This was followed by trypsinization and measurement of transgene expression by flow cytometry.
(Z.Y.); the Weston Havens Foundation (J.E.C., S.P., and A.S.P.); and the Stanford SPARK program (S.P. and A.S.P.). J.E.C. is a David and Lucile Packard Foundation fellow. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. S.P., A.S.P., N.L.M., O.D., M.S.C., and J.E.C. are inventors on a patent describing the utilization of AAVR to modulate AAV transduction.