GRK2 Mediates β-Arrestin Interactions with 5-HT₂ Receptors for JC Polyomavirus Endocytosis

As intracellular parasites, viruses require a host cell to replicate and cause disease. Therefore, virus-host interactions contribute to viral pathogenesis.

ABSTRACT

JC polyomavirus (JCPyV) infects the majority of the population, establishing a lifelong, asymptomatic infection in the kidney of healthy individuals. People who become severely immunocompromised may experience JCPyV reactivation, which can cause progressive multifocal leukoencephalopathy (PML), a neurodegenerative disease. Due to a lack of therapeutic options, PML results in fatality or significant debilitation among affected individuals. Cellular internalization of JCPyV is mediated by serotonin 5-hydroxytryptamine subfamily 2 receptors (5-HT₂Rs) via clathrin-mediated endocytosis. The JCPyV entry process requires the clathrin-scaffolding proteins β-arrestin, adaptor protein 2 (AP2), and dynamin. Furthermore, a β-arrestin-interacting domain, the Ala-Ser-Lys (ASK) motif, within the C terminus of 5-HT_2AR is important for JCPyV internalization and infection. Interestingly, 5-HT₂R subtypes A, B, and C equally support JCPyV entry and infection, and all subtypes contain an ASK motif, suggesting a conserved mechanism for viral entry. However, the role of the 5-HT₂R ASK motifs and the activation of β-arrestin-associated proteins during internalization have not been fully elucidated. Through mutagenesis, the ASK motifs within 5-HT_2BR and 5-HT_2CR were identified as being critical for JCPyV internalization and infectivity. Furthermore, by using biochemical pulldown techniques, mutagenesis of the ASK motifs in 5-HT_2BR and 5-HT_2CR resulted in reduced β-arrestin binding. When small-molecule chemical inhibitors and RNA interference were used, G protein receptor kinase 2 (GRK2) was determined to be required for JCPyV internalization and infection by mediating interactions between β-arrestin and the ASK motif of 5-HT₂Rs. These findings demonstrate that GRK2 and β-arrestin interactions with 5-HT₂Rs are critical for JCPyV entry by clathrin-mediated endocytosis and the resultant infection.

IMPORTANCE As intracellular parasites, viruses require a host cell to replicate and cause disease. Therefore, virus-host interactions contribute to viral pathogenesis. JC polyomavirus (JCPyV) infects most of the population, establishing a lifelong asymptomatic infection within the kidney. Under conditions of severe immunosuppression, JCPyV may spread to the central nervous system, causing the fatal demyelinating disease progressive multifocal leukoencephalopathy (PML). Individuals living with HIV or undergoing immunomodulatory therapies are at risk for developing PML. The mechanisms of how JCPyV uses specific receptors on the surface of host cells to initiate internalization and infection are poorly understood processes. We have further identified cellular proteins involved in JCPyV internalization and infection and elucidated their specific interactions that are responsible for the activation of receptors. Collectively, these findings illuminate how viruses usurp cellular receptors during infection, contributing to current development efforts for therapeutic options for the treatment or prevention of PML.

INTRODUCTION

JC polyomavirus is the causative agent of progressive multifocal leukoencephalopathy (PML), a debilitating neurological disease for which there are no effective therapeutics (1, 2). Seroepidemiology studies suggest that upwards of 75% of the human population harbors JCPyV within the kidney, as a lifelong, asymptomatic infection (2, 3). Although the occurrence of PML is relatively rare, individuals at heightened risk for developing PML are those infected with HIV (4, 5) and individuals undergoing prolonged immunomodulatory therapies, including natalizumab treatment for multiple sclerosis (MS) (6, 7). Currently, the first line of treatment following a PML diagnosis is treatment of the underlying immunosuppression through implementation of highly active antiretroviral therapy (HAART) for those infected with HIV-1 or through the cessation of immunomodulatory therapy for those suffering from immune-related disorders, often resulting in exacerbation of the underlying disease and fatality (4, 5, 8–12). Although removal of immunomodulatory therapies can prolong life, a cure for PML is not yet available, and if left untreated, PML can result in fatality within months (13). A deeper understanding of the JCPyV infectious life cycle within the host may inform the development of improved antiviral therapeutics.

As a member of the Polyomaviridae family, JCPyV is a nonenveloped double-stranded DNA (dsDNA) virus (14). Polyomavirus capsids are comprised of three viral proteins, namely, viral protein 1 (VP1), VP2, and VP3 (15, 16). This virus family also includes other polyomaviruses, including simian virus 40 (SV40) and BK polyomavirus (BKPyV), of close relation to JCPyV (17). Expressed on the exterior of the capsid, VP1 serves as the point of attachment between JCPyV and host cell surface receptors (18) through direct interactions with α2,6-sialic acid containing lactoseries tetrasaccharide c (LSTc) or nonsialylated glycosaminoglycans (GAGs) (18–21). However, recent findings demonstrate that polyomaviruses, including JCPyV, may also be packaged into extracellular vesicles as a means of establishing infection in cells independent of attachment factor expression (22–25). Following attachment, JCPyV entry is facilitated by the 5-hydroxytryptamine (5-HT) serotonin subtype 2 family receptors (5-HT_2AR, 5-HT_2BR, and 5-HT_2CR) (26–29) by clathrin-mediated endocytosis (CME), usurping the endocytic protein β-arrestin (27, 29, 30). While JCPyV utilizes CME for uptake within cells, other polyomaviruses studied, including SV40, utilize either caveolin- or nonclathrin/noncaveolin-mediated endocytosis (31–35). Moreover, proteins critical for CME of JCPyV are not required for SV40 infection, suggesting that proteins involved in the activation of CME are dispensable for SV40 infection (29). Utilization of the CME entry pathway is unique to JCPyV among polyomaviruses; however, following CME, virions traffic to the endoplasmic reticulum (ER) prior to nuclear translocation, similar to other polyomaviruses studied (36–41).

5-HT₂Rs are G protein-coupled receptors (GPCRs) that can be activated by G protein-dependent or β-arrestin-mediated signaling pathways, resulting in differing signaling outcomes (42, 43). 5-HT₂Rs can be internalized by CME in an agonist- and cell-type-specific fashion (44–47) through the recruitment of scaffolding proteins, including clathrin, β-arrestin, and adaptor protein 2 (AP2) (44, 45, 47). The activation of these proteins ultimately dictates the signaling outcomes of the receptor and associated cargo (42–44, 48), facilitating the delivery of 5-HT₂Rs to endocytic vesicles resulting in recycling, trafficking, degradation of the receptor, or activation of specific signaling cascades (44). We have previously determined that JCPyV usurps the CME proteins β-arrestin, AP2, and clathrin to facilitate a productive infection and yet had no impact on SV40 infection (29). While these proteins were demonstrated to be important for viral infection, β-arrestin and clathrin were specifically identified to facilitate viral entry (29). Expression of 5-HT receptors 2A, 2B, or 2C can confer infection in poorly permissive cell types, suggesting a reliance on conserved signaling inherent to 5-HT₂Rs (27). Furthermore, HEK293A cells stably expressing 5-HT_2AR, 5-HT_2BR, or 5-HT_2CR can promote infection; however, when β-arrestin is knocked down, these cells become resistant to JCPyV infection (28, 29). This knowledge highlights the importance of β-arrestin in the viral usurpation of 5-HT₂R signaling capabilities, likely occurring in a conserved manner.

The C terminus of 5-HT_2AR contains a conserved, tripeptide motif, the Ala-Ser-Lys (ASK) motif, which is a β-arrestin binding domain; this motif has been identified to be critical for endocytosis and trafficking of 5-HT_2AR within cells (47). Mutagenesis of this motif alters the signaling capability of 5-HT_2AR and negates receptor reliance on β-arrestin for intracellular trafficking and recycling (47). Upon β-arrestin interactions with the receptor, further endocytic scaffolding machinery, including AP2 and additional β-arrestin, are recruited to the site of the activated receptor and can directly interact with clathrin, assisting in the formation of the clathrin-coated pit (43, 45, 49). Through these direct interactions β-arrestin can promote downstream signaling in the cell, including the activation of the mitogen-activated protein kinase (MAPK) cascade (50–52), which has been demonstrated to drive JCPyV infection (53–56). Mutagenesis of the ASK motif in 5-HT_2AR drastically reduces JCPyV internalization and infection (29), suggesting a dependence on this domain for β-arrestin-induced signaling events critical for viral infection. Interestingly, in addition to 5-HT_2AR, the ASK motif is also conserved in related 5-HT₂R subtypes 2B and 2C (NCBI accession numbers 3356, 3357, and 3358, respectively). However, the importance of the conserved ASK motif for the canonical function of the 5-HT_2BR and 5-HT_2CR or a reliance on β-arrestin for their intracellular trafficking is not well characterized.

β-Arrestin-mediated signaling is facilitated by G protein receptor kinases (GRKs), which promote uncoupling of G proteins through specific phosphorylation events (46, 47). GRKs, a family of seven kinases, are recruited to GPCRs and recognize specific amino acid sequences, or “bar codes,” on the receptor, resulting in phosphorylation of mainly Ser or Thr residues within the signal sequence (57). This phosphorylation event prevents the activation of G proteins and promotes β-arrestin recruitment and binding to the receptor, resulting in rapid desensitization, or internalization, of 5-HT₂Rs by CME. Furthermore, the β-arrestin binding domain of 5-HT_2AR, the ASK motif, is phosphorylated by GRK2 (47). This phosphorylation promotes the recruitment of β-arrestin which engages the phosphate residue-tagged Ser within the ASK motif (47). The direct interaction between β-arrestin and 5-HT_2AR then promotes internalization of the receptor by CME. Trafficking of the receptor within the cell is reliant on β-arrestin engagement of the ASK motif as mutagenesis of Ser within the motif results in β-arrestin-independent internalization of the receptor with drastically different signaling capabilities (47). Interestingly, mutagenesis of the Ser residue within the ASK motif of 5-HT_2AR significantly reduces JCPyV entry within cells (29). Therefore, the presence of the ASK motif within 5-HT₂Rs is critical for canonical signaling functionality of the receptor, and as such, β-arrestin is a regulator of receptor trafficking and signaling outcomes (29, 44, 47).

The focus of this study was to determine the signaling networks responsible for the induction of JCPyV internalization through 5-HT₂Rs and β-arrestin. We have identified a cellular kinase and β-arrestin-interacting domains within the 5-HT₂Rs responsible for promoting JCPyV internalization and infection. Collectively, the results presented herein further define JCPyV entry strategies, as well as provide improved insight into JCPyV utilization of 5-HT₂Rs as functional entry receptors for viral infection and pathogenesis.

RESULTS

The ASK motifs of 5-HT_2BR and 5-HT_2CR are required for JCPyV infection.

β-Arrestin and the Ser residue of the β-arrestin-binding ASK motif within 5-HT_2AR are necessary for JCPyV entry and infection (29). 5-HT_2BR and 5-HT_2CR also support JCPyV entry and infection, and the β-arrestin-binding ASK motif is conserved in 5-HT₂Rs (29). Interactions between β-arrestin and the Ser residue within the motif of 5-HT_2AR ultimately regulate receptor internalization and subsequent trafficking (47). Thus, investigation of the ASK motif in the 5-HT_2BR and 5-HT_2CR is necessary to define the mechanism of 5-HT₂R-mediated entry of JCPyV. To determine whether the ASK motif is required for JCPyV utilization of 5-HT_2BR and 5-HT_2CR for infection, site-directed mutagenesis was employed, engineering single or double amino acid mutations within the ASK motif of either receptor in fusion with yellow fluorescent protein (YFP). Through mutagenesis, the Ser and Lys residues of 5-HT_2BR and 5-HT_2CR were replaced with Ala at residue positions 323 and 324 (5-HT_2BR) and positions 310 and 311 (5-HT_2CR) (depicted in Fig. 1A). Wild-type (WT) and mutated receptors were expressed in HEK293A cells, which were subsequently infected with JCPyV, and viral infectivity was determined through the quantitation of nuclear T antigen (T-Ag) expression by indirect immunofluorescence (Fig. 1B and C). 5-HT_2BR-S323A (AAK) and 5-HT_2BR-SK323-24A (AAA) demonstrated approximately 80% reductions in JCPyV infectivity compared with 5-HT_2BR-K324A (ASA), which demonstrated infectivity levels comparable to 5-HT_2BR (Fig. 1B). Interestingly, 5-HT_2CR-S310A (AAK) and 5-HT_2CR-SK310-11AA (AAA) also demonstrated significant reductions of nearly 80% in infection compared with 5-HT_2CR-K311A (ASA) and 5-HT_2CR (Fig. 1C).

FIG 1.

β-Arrestin-binding Ala-Ser-Lys (ASK) motif within 5-HT_2BR and 5-HT_2CR is required for JCPyV infection. (A) Schematic demonstrating the conserved β-arrestin-binding ASK motif positions within 5-HT₂Rs. HEK293A cells were transfected with plasmids containing wild-type 5-HT_2BR-YFP or 5-HT_2CR-YFP (WT), 5-HT_2BR-YFP or 5-HT_2CR-YFP with amino acid point mutations within the ASK motif, or empty vector YFP. (B, C) Transiently transfected cells were infected with JCPyV (MOI of 1 FFU/cell) and fixed at 48 hours postinfection (hpi). Cells were stained using a JCPyV-specific T-Ag antibody, and infectivity was measured by indirect immunofluorescence. Data represent the percentage of infected cells/visual field normalized to the total number of DAPI+ cells/visual field for five 10× fields of view for triplicate samples and were independently repeated three times. Samples were normalized to the WT receptor within each receptor subtype (100%). A two-sample Student’s t test was used to determine statistical significance. Scale bars = 20 μm. NS, no significance. Error bars are SD. *, P < 0.05.

To determine the expression of the mutated receptors at the cell surface compared with the wild type (WT), cell surface expression was measured by confocal microscopy. The surface of the cell was defined by staining with an antibody (Ab) that detects the cell surface proteins cadherins (29). Individual fields of view were measured, and the percentage of cell surface expression was determined for individual cells expressing mutated 5-HT_2BR and 5-HT_2CR compared with their respective WT, unmutated receptors (Fig. 2A and B). Specifically, the colocalization between the expressed receptor and the cell surface marker was determined for individual cells using ImageJ software (Fig. 2B). Moreover, the cellular distribution of the expressed receptors was determined by quantifying the corrected total cell fluorescence (CTCF; total plasma membrane and intracellular fluorescence), as well as the corrected intracellular fluorescence (CICF; only intracellular fluorescence) for individual cells using ImageJ software (Fig. 2C). Mutagenesis of the Ser or Lys residues in either receptor did not significantly alter cell surface expression compared with the WT receptors (Fig. 2B), nor did mutagenesis significantly modify the total cellular receptor expression or subcellular expression, as determined by receptor fluorescence intensity (Fig. 2C). Collectively, these results suggest that the β-arrestin-binding ASK motif, conserved in all three subtypes of 5-HT₂Rs, is necessary for JCPyV infectivity and that, specifically, the Ser residue within the ASK motif is important for infection across the 5-HT₂R subtypes.

FIG 2.

Mutagenesis of β-arrestin-binding ASK motif within 5-HT_2BR and 5-HT_2CR does not alter cellular receptor expression. (A, B) Receptor cell surface expression of WT or mutated receptors was measured by confocal microscopy at 60× magnification (top). Cell surface expression was determined using ImageJ software for colocalization between expressed receptor in YFP (green) and a cell surface marker (pan-cadherin-647, pseudocolored orange) using Mander’s coefficient (n = 30 cells/sample/replicate) for 3 independent experiments. (C, D) Total cellular and intracellular receptor fluorescence were determined using ImageJ software to measure total cell fluorescence and intracellular fluorescence. Corrected fluorescence intensity was determined by ImageJ software; integrated density of receptor fluorescence was measured and corrected for ROI area and background for individual cells (n = 30 cells/sample/replicate) for three independent experiments. CTCF, corrected total cell fluorescence; CICF, corrected intracellular fluorescence. Data are represented as a box-and-whisker plot (bottom) from one experiment that is representative of three independent experiments. Outliers are indicated in grey. Upper and lower whiskers represent 1.5 times the interquartile range. Statistical significance was determined using a pairwise Wilcoxon signed-rank test. Scale bars = 20 μm.

The serine residue of the 5-HT_2BR and 5-HT_2CR ASK motif is critical for JCPyV internalization.

JCPyV requires the Ser residue within the ASK motif of 5-HT_2AR for internalization and productive infection (29), and the ASK motifs of 5-HT_2BR and 5-HT_2CR are critical for JCPyV infection (Fig. 1). Thus, the impact of mutagenesis of the ASK motifs within 5-HT_2BR and 5-HT_2CR during JCPyV attachment and entry was explored. HEK293A cells were transfected with WT and mutated 5-HT_2BR and 5-HT_2CRs, each in fusion with YFP. The transiently transfected cells were incubated with JCPyV labeled with Alexa Fluor 647 (JCPyV-647) on ice for viral attachment. The cells were fixed, and viral attachment was measured by flow cytometry. The mean fluorescence intensity was comparable in cells expressing the mutated ASK motifs compared with the unmutated 5-HT_2BR or 5-HT_2CR, suggesting that mutagenesis of the Ser or Lys residues did not impact viral attachment (Fig. 3A and B). To measure viral entry, HEK293A cells expressing WT and mutated 5-HT₂Rs were incubated with JCPyV-647 on ice for synchronized viral attachment, and then cells were shifted to a 37°C incubator for 2 h for viral internalization prior to fixation. Viral entry was measured by confocal microscopy and single-cell quantitative analysis (Fig. 3C and D) (29). The relative fluorescence intensity for internalized JCPyV-647 was measured by drawing individual regions of interest excluding the plasma membrane, defined by differential inference contrast (DIC) overlay, utilizing Olympus Fluoview software (29). Mutagenesis of either 5-HT_2BR or 5-HT_2CR with an altered Ser residue (AAK or AAA) resulted in reductions in JCPyV internalization by approximately 75%, while mutation of the Lys residue alone (ASA) was not significantly different compared with WT receptors (Fig. 3C and D). Together, these findings highlight the importance of the conserved 5-HT₂R ASK motif and specifically the Ser residue within the motif for internalization of JCPyV, as disruption of this residue significantly diminishes viral entry.

FIG 3.

Mutagenesis of serine residue within ASK motif of either 5-HT_2BR or 5-HT_2CR reduces JCPyV entry. (A, B) HEK293A cells were transfected with plasmids containing wild-type 5-HT_2BR-YFP or 5-HT_2CR-YFP (WT), 5-HT_2BR-YFP or 5-HT_2CR-YFP with amino acid point mutations within the ASK motif, or empty vector YFP. Cells were removed from plates, prechilled on ice, and subsequently incubated with JCPyV-647 (MOI of 5 FFU/cell) on ice for 1 h for viral attachment. Cells were analyzed by flow cytometry. Histograms represent the mean fluorescence intensities for cells without treatment (grey), empty YFP vector (no virus, black), or cells treated with JCPyV-647. Data are representative of 20,000 events for 3 independent experiments and were performed in triplicate. (C, D) To measure JCPyV entry, cells were prechilled, and then incubated with JCPyV-647 (MOI of 5 FFU/cell, pseudocolored magenta) on ice for 1 h for viral attachment. Cells were then shifted to 37°C for 2 h for viral entry prior to fixation. Cell nuclei were stained with DAPI (blue). Viral entry was analyzed by confocal microscopy at 60× magnification (top). Fluoview software single-measurement analysis was used to define regions of interest (ROI) using z-sectioning excluding the plasma membrane to measure internalized virus per cell. JCPyV internalization was defined as the relative fluorescence per cell for at least 30 cells in triplicate per sample for WT, 5-HT_2BR-YFP, or 5-HT_2CR-YFP with amino acid point mutations within the ASK motif (green) using Olympus Fluoview 10-ASW software. Data are depicted as a raincloud plot where the height and width of each raincloud represent the distribution of each sample (bottom). Individual rain points (black) denote the percentage of virus internalized for each cell measured. Samples were normalized to the average of WT (100%). Microscopy experiments were performed independently 3 times, and at least 30 cells in triplicate samples per replicate were analyzed; graphs are comprised of 3 independent replicates (n = 90). Statistical significance was determined using a pairwise Wilcoxon signed-rank test. Scale bars = 20 μm. NS, no significance. *, P < 0.05.

Mutagenesis of the ASK motif in 5-HT₂Rs reduces β-arrestin-receptor interactions.

The Ser residue of the conserved ASK motifs of 5-HT₂Rs are required for JCPyV infection and entry (Fig. 1 and 3) (29). Furthermore, the Ser residue of the ASK motif of 5-HT_2AR has been implicated as a direct β-arrestin-binding domain when 5-HT induces canonical signaling events or pathways (47). To determine if JCPyV induces β-arrestin interactions with 5-HT₂Rs during entry and whether mutagenesis of the ASK motif reduces this interaction, biochemical pulldown assays were employed. HEK293A cells were transfected with WT or mutated receptors, namely, 5-HT_2AR-YFP, 5-HT_2BR-YFP, or 5-HT_2CR-YFP, in which the ASK motif was mutated to AAK or ASA, as these mutations did not alter JCPyV entry or infection (Fig. 1 and 3). Cells were incubated with either 5-HT or JCPyV or were not treated at 4°C for 1 h for attachment and then shifted to 37°C for 15 min for entry, and then they were subsequently harvested and lysed. Magnetic agarose beads preconjugated to a GFP antibody for detection of YFP were incubated with the cellular lysates for immunoprecipitation of 5-HT₂R-YFPs and direct-interacting proteins. Samples containing the bound fraction of proteins were resolved by SDS-PAGE and analyzed through Western blotting using anti-GFP and anti-β-arrestin antibodies. Quantitation of the percentage of bound β-arrestin relative to 5-HT₂R-YFP was determined through Li-Cor ImageStudio analysis. Briefly, the borders of each protein band were defined by region of interest (ROI), resulting in determination of the intensity of the band; β-arrestin band intensity was then internally normalized to the 5-HT₂R band intensity to normalize for protein loaded/well. (Fig. 4A). Incubation of WT-expressing cells with either 5-HT or JCPyV resulted in increased β-arrestin binding, compared with cells that did not receive agonist treatment (Fig. 4B). Furthermore, AAK-expressing HEK293A cells treated with either 5-HT or JCPyV demonstrated reduced β-arrestin binding at 15 min posttreatment compared with WT controls, and β-arrestin binding in ASA-expressing cells was similar to that of the WT (Fig. 4B). Mutagenesis of the Ser residue in the ASK motif of the 5-HT₂Rs (AAK) reduced β-arrestin binding for JCPyV-treated samples by 60% (5-HT_2AR), 60% (5-HT_2BR), and 70% (5-HT_2CR) compared with that of the WT receptors (Fig. 4B). Together, these results suggest that upon stimulation of 5-HT₂Rs by either 5-HT or JCPyV, β-arrestin localizes with the receptor, resulting in direct interactions between β-arrestin and the receptor. Furthermore, these findings also demonstrate that the Ser residue of the ASK motif is critical for conferring β-arrestin interactions with 5-HT₂Rs when receptors are activated by either 5-HT or JCPyV.

FIG 4.

Mutagenesis of ASK motif of 5-HT₂R reduces β-arrestin-receptor interactions during JCPyV internalization. HEK293A cells were transfected with unmutated 5-HT_2AR-YFP, 5-HT_2BR-YFP, 5-HT_2CR-YFP (WT), or receptors containing a point mutation within the ASK motif (AAK, ASA). (A) Cells were incubated with JCPyV (MOI of 12 FFU/cell), serotonin (5-HT, 200 μM/well), or received no treatment (NT) at 4°C for 1 h (attachment) and then were shifted to 37°C for 15 min prior to lysis (whole-cell lysate [WCL]). Lysed cells were incubated with magnetic agarose beads conjugated to a GFP antibody for immunoprecipitation (IP) of YFP-conjugated 5-HT₂Rs. Bound samples for WT, AAK, and ASA receptors were processed by SDS-PAGE and immunoblotted (IB) for β-arrestin and GFP (corresponding 5-HT₂R). WCL samples were stained for total protein. (B) The percentage of β-arrestin bound to 5-HT₂R was determined using Li-Cor ImageStudio software. Data represent the quantitation of the relative amounts of β-arrestin that are bound to 5-HT₂Rs at 15 min postinternalization, normalized to the 5-HT WT (100%), and contain data from individual replicates performed in three independent experiments. Error bars = SDs. A two-sample Student’s t test was used to determine statistical significance. *, P < 0.05.

GRKs are required for JCPyV infection.

Prior to recruitment of β-arrestin for endocytosis, 5-HT_2AR is phosphorylated by GRK2 (46, 47). Furthermore, 5-HT_2AR internalization by β-arrestin has been reported to be both GRK2 dependent and GRK2/GRK5 independent (45, 47, 58). The involvement of either GRK2 or GRK5 is dependent on cell type and agonist (45) and ultimately determines receptor internalization and downstream signaling cascades (46, 47). To determine whether GRK2 and/or GRK5 are required for JCPyV infection of glial cells, SVG-A cells were transfected with small interfering RNAs (siRNAs) targeting an irrelevant control (CTL) (29), GRK2, or GRK5. Cells were then infected with either JCPyV or SV40 and scored for infectivity by indirect immunofluorescence of newly synthesized VP1 or processed for Western blot analysis to determine the efficiency of the protein knockdown (Fig. 5A). SV40 infection of SVG-A cells occurs independently of cellular proteins involved in clathrin-mediated endocytosis, including β-arrestin (29, 33). Furthermore, mutation of the ASK motif in the 5-HT_2AR had no impact on SV40 infection (29); however, the necessity of GRKs has not been studied for this closely related polyomavirus. Interestingly, treatment of SVG-A cells with siRNA targeting GRK2 or GRK5 reduced JCPyV infection by approximately 70% and 40%, respectively. (Fig. 5A). In contrast, infection of SVG-A cells by SV40 was not hindered by knockdown of GRK2, while siRNA targeting GRK5 resulted in reduced SV40 infection by approximately 45% (Fig. 5A). Combinatory siRNA knockdown of both GRK2 and GRK5 further reduced JCPyV infectivity beyond that seen for GRK2 alone by approximately 7%, while SV40 infection was not further reduced (Fig. 5A).

FIG 5.

G protein receptor kinases 2 and 5 are necessary for JCPyV infection. SVG-A (A) or HEK293A (B) cells stably expressing the 5-HT₂R subtypes were transfected with siRNA targeting an irrelevant control (CTL), GRK2, or GRK5, and were either processed for protein knockdown by Western blot analysis (below) or infected with either JCPyV (MOI of 1 FFU/cell) or SV40 (MOI of 0.001 FFU/cell) and fixed at either 72 (SVG-A) or 48 (HEK293A-5-HT₂R) hpi. Noninternalized virus was neutralized using JCPyV-neutralizing antiserum (1:10,000). Cells were stained with an antibody specific for VP1 (SVG-A) or T-Ag (HEK293A-5-HT₂R), and infectivity was measured by indirect immunofluorescence. Data represent the percentage of infected cells/visual field normalized to the number of DAPI+ cells/visual field for five 10× fields of view for triplicate samples and were independently repeated three times. Samples were normalized to the control siRNA for each cell type (100%). Error bars = SD. *, P < 0.05. (C) SVG-A cells were transfected with plasmids for GFP (CTL), wild-type GRK2 and GFP (WT), or a kinase-deficient GRK2 containing a K220R point mutation and GFP (K220R). Cells were infected with either JCPyV (MOI of 1 FFU/cell) or SV40 (MOI of 0.001 FFU/cell) and fixed at 72 hpi. Cells were stained with an antibody specific for VP1, and infectivity was measured by indirect immunofluorescence. Data represent the percentage of infected cells/visual field normalized to the number of DAPI+ cells/visual field for five 10× fields of view for triplicate samples and were independently repeated three times. Samples were normalized to the CTL sample (100%). A two-sample Student’s t test was used to determine statistical significance. NS, no significance. Error bars = SD. *, P < 0.05.

To determine whether the reliance on GRKs is required for infection of kidney cells and attributed to specific 5-HT₂Rs, HEK293A cells stably expressing 5-HT_2AR, 5-HT_2BR, or 5-HT_2CR were similarly treated with siRNA targeting an irrelevant control (CTL), GRK2, or GRK5 prior to JCPyV challenge. JCPyV infectivity was quantified by T-Ag expression at 48 h postinfection by indirect immunofluorescence. JCPyV infection of HEK293A cells expressing 5-HT₂Rs, regardless of subtype, exhibited similar reductions (∼70% to 80%) in JCPyV infection when GRK2 was knocked down. Infection following GRK5 knockdown was reduced by approximately 40% (Fig. 5B). In addition, to determine if a functioning GRK2 kinase domain is important for either JCPyV or SV40 infection, SVG-A cells were transiently transfected with GFP (CTL), wild-type GRK2 and GFP (WT), or kinase-deficient GRK2 (K220R) and GFP (47) prior to viral challenge. Expression of kinase-deficient K220R significantly hindered JCPyV infection compared with control GFP and the overexpressed WT GRK2 by nearly 50% (Fig. 4C), while SV40 infection was not reduced (Fig. 5C). Interestingly, overexpression of WT GRK2 did not enhance JCPyV infection beyond that of the GFP control (Fig. 5C). Collectively, these results suggest that JCPyV requires the kinase activity of GRK2 for infection of either glial or kidney cell types, regardless of the 5-HT₂R subtype used. Furthermore, these results suggest that both JCPyV and SV40 infection of glial cells require GRK5, and due to the disparate mechanisms in JCPyV and SV40 entry, this is perhaps through a conserved postentry polyomavirus effect.

GRK2, but not GRK5, is required for JCPyV internalization.

Internalization of 5-HT_2AR involves the recruitment of GRKs resulting in phosphorylation of 5-HT_2AR and leads to the recruitment of β-arrestin and subsequent internalization of the receptor (47). Inhibition of GRK2 and GRK5 reduced JCPyV infectivity (Fig. 5); thus, the necessity of these kinases in promoting either JCPyV attachment or internalization warranted further characterization. To determine whether a reduction of GRK2 or GRK5 impacted JCPyV attachment or entry, SVG-A cells were transfected with siRNA targeting an irrelevant control (CTL), GRK2, or GRK5; then cells were incubated with JCPyV-647 on ice and fixed; and viral attachment was measured by flow cytometry. The mean fluorescence intensity of virus attached to cells treated with GRK2 or GRK5 siRNA was equivalent to that of the CTL-siRNA samples (Fig. 6A). To determine whether GRK2 or GRK5 are required for JCPyV internalization, SVG-A cells treated with siRNAs were incubated with JCPyV-647 for viral synchronized attachment on ice and then shifted to a 37°C incubator for 2 h for internalization. Cells were fixed and viral entry was measured by confocal microscopy and single-cell quantitative analysis. The relative fluorescence intensity for internalized JCPyV-647 was measured by drawing individual regions of interest excluding the plasma membrane, defined by a pan-cadherin stain, and analyzed using Olympus Fluoview software (Fig. 6B). Compared with irrelevant control siRNA-treated samples, JCPyV internalization in cells treated with GRK2 siRNA was reduced by approximately 80% (Fig. 6B). Interestingly, GRK5 knockdown in SVG-A cells did not impact JCPyV entry (Fig. 6B). Collectively, these results suggest a requirement for GRK2 in facilitating JCPyV internalization, for which GRK5 is dispensable.

FIG 6.

GRK2 is critical for JCPyV internalization. (A and B) SVG-A cells were transfected with siRNA targeting either an irrelevant control (CTL), GRK2, or GRK5. (A) Cells were stripped from plates, prechilled on ice, incubated with JCPyV-647 (MOI of 5 FFU/cell) on ice for 1 h, and then analyzed by flow cytometry for viral attachment. Histograms represent the mean fluorescence intensities for cells transfected with either CTL siRNA (blue), GRK2 or GRK5 siRNA (green), or cells alone (no siRNA or virus, shaded grey). Data are representative of 20,000 events for at least 3 independent experiments. (B) To measure JCPyV entry, cells were prechilled, and then incubated with JCPyV-647 (MOI of 5 FFU/cell; pseudocolored magenta) on ice for 1 h for viral attachment. Cells were then shifted to 37°C for 2 h for viral entry prior to fixation. Cell nuclei were stained with DAPI (blue), and the plasma membrane was stained with an anti-pan-cadherin antibody (pseudocolored green). Viral entry was analyzed by confocal microscopy at 60× magnification (top). Fluoview software single-measurement analysis was used to define regions of interest (ROI) using z-sectioning excluding the plasma membrane to measure internalized virus per cell. JCPyV internalization was defined as the relative fluorescence per cell for at least 30 cells per sample for CTL and GRK2 and GRK5 siRNA-treated samples using Olympus Fluoview 10-ASW software. Data are depicted as a raincloud plot where the height and width of each raincloud represent the distribution of each sample (bottom). Individual rain points (black) denote the percentage of virus internalized for each cell measured. Samples were normalized to the average of CTL-treated samples (100%). Microscopy experiments were performed independently 3 times in triplicate, containing at least 30 cells per sample per replicate; graphs are comprised of 3 independent replicates (n = 90). Statistical significance was determined using a pairwise Wilcoxon signed-rank test. Scale bars = 20 μm. NS, no significance. *, P < 0.05.

GRK5 is required for JCPyV or SV40 infection in a posttrafficking step.

While GRK2 is required for JCPyV infectivity (Fig. 5A and B) and internalization (Fig. 6B), GRK5 knockdown reduced infectivity of both JCPyV and SV40 (Fig. 5A); yet, siRNA knockdown of GRK5 did not hinder JCPyV internalization (Fig. 6B). These data suggest that GRK5 may function to facilitate a step in polyomavirus infection following entry. Thus, the impact of GRKs for the promotion of either JCPyV or SV40 infection was further explored. To determine if either GRK2 or GRK5 are required for infection postinternalization, SVG-A cells were transfected with siRNA targeting the irrelevant control (CTL), GRK2, or GRK5, and were incubated for 72 h. Cells were then transfected with the infectious clone of either JCPyV or SV40, bypassing viral attachment, entry, and trafficking. Cells were fixed at either 4 or 7 days postinfection, and viral infectivity was determined by indirect immunofluorescence of newly synthesized VP1 (54). siRNA knockdown of GRK2 did not impact JCPyV infection, while knockdown of GRK5 reduced infection by approximately 40% and 55% at days 4 and 7 postinfection, respectively (Fig. 7A). Meanwhile, SV40 infection was also not impacted following GRK2 siRNA, although knockdown of GRK5 resulted in reduced SV40 infection at both day 4 and day 7, each by approximately 50% (Fig. 7B). Collectively, these results suggest that the requirement of GRK2 for JCPyV infection occurs prior to viral transcription, confirming the effects on viral entry (Fig. 5), while GRK5 plays a role in both JCPyV and SV40 infection following localization of these viruses to the nucleus.

FIG 7.

Knockdown of GRK5 prevents polyomavirus infection post entry. (A and B) SVG-A cells were transfected with siRNA targeting either an irrelevant control (CTL), GRK2, or GRK5. Cells were then transfected with an infectious clone of either JCPyV (A) or SV40 (B). Cells were fixed at either 4 or 7 days posttransfection and stained with an antibody specific for VP1, and infectivity was measured by indirect immunofluorescence. Data are representative of the number of infected cells/10× visual field for 5 fields of view per sample in triplicate. Data are representative of results from three independent experiments. Error bars indicate SDs. A two-sample Student’s t test was used to determine statistical significance. NS, no significance. *, P < 0.05.

GRK2 mediates β-arrestin-5-HT₂R interactions for viral internalization.

GRK2 has been identified to phosphorylate the Ser residue within 5-HT_2AR, and this phosphorylation is critical for β-arrestin engagement of the receptor and subsequent internalization and signaling (47). Furthermore, JCPyV infection of cells expressing 5-HT₂Rs relies on β-arrestin and GRK2 (Fig. 5) (29), and the conserved ASK motifs of 5-HT_2AR, 5-HT_2BR, and 5-HT_2CR are critical for JCPyV entry and infection of glial and kidney cell types (Fig. 1 and 3) (29). Additionally, mutagenesis of the ASK motif results in reduced β-arrestin engagement with 5-HT₂Rs (Fig. 4). To determine if either GRK2 or GRK5 plays a role in recruiting β-arrestin to 5-HT₂Rs during JCPyV infection, HEK293A cells stably expressing WT 5-HT₂Rs were transfected with siRNA targeting either an irrelevant control (CTL), GRK2, or GRK5 for 72 h. Serum starved, prechilled cells were incubated with either 5-HT or JCPyV or were not treated at 4°C for 1 h for synchronized attachment. Cells were then shifted to 37°C for 15 min for entry and subsequently harvested and lysed. Cellular lysates were incubated with magnetic agarose beads preconjugated to a GFP antibody for detection of YFP for immunoprecipitation of 5-HT₂R-YFPs and direct-interacting proteins. Bound protein fractions were resolved by SDS-PAGE, and quantitation of the percentage of bound β-arrestin was employed using anti-GFP and anti-β-arrestin antibody detection utilizing Li-Cor ImageStudio software (Fig. 8). Briefly, the borders of each protein band were defined by ROI, resulting in determination of the intensity of the band; the β-arrestin band intensity was then internally normalized to the 5-HT₂R band intensity to normalize for protein loaded/well. siRNA knockdown of GRK2 resulted in reduced β-arrestin interactions with 5-HT₂Rs when either 5-HT (approximately 60% to 65%) or JCPyV (approximately 65% to 70%) served as the agonist compared with CTL-siRNA samples (Fig. 8A, B, and C). However, siRNA knockdown of GRK5 did not significantly alter β-arrestin binding to 5-HT₂Rs, regardless of agonist (Fig. 8A, B, and C). Together, these results demonstrate the necessity of GRK2 in facilitating β-arrestin-receptor interactions, when either 5-HT or JCPyV is the cellular stimulus, regardless of 5-HT₂R subtype, and reinforce the necessity of GRK5 following viral entry.

FIG 8.

Knockdown of GRK2 reduces β-arrestin-receptor interactions during JCPyV entry. HEK293A cells stably expressing 5-HT_2AR-, 5-HT_2BR-, or 5-HT_2CR-YFP were transfected with siRNA targeting an irrelevant control (CTL), GRK2, or GRK5. (A) 5-HT_2AR-YFP, (B) 5-HT_2BR-YFP, or (C) 5-HT_2CR-YFP cells were incubated with either JCPyV (MOI of 12 FFU/cell) or serotonin (5-HT; 200 μM/well) or received no treatment (NT) at 4°C for 1 h for attachment and then were incubated at 37°C for 15 min. Cells were then removed from plates, lysed (whole-cell lysate [WCL]), and incubated with magnetic agarose beads conjugated to a GFP antibody for immunoprecipitation (IP) of 5-HT₂Rs in fusion with YFP. WCL samples were stained for total protein. Bound samples for CTL, GRK2 siRNA, and GRK5 siRNA samples were processed by SDS-PAGE followed by immunoblotting (IB) for detection of β-arrestin and GFP (corresponding 5-HT₂R). The percentage of bound β-arrestin was determined using Li-Cor ImageStudio software. Data represent the quantitation of the relative amounts of β-arrestin that are bound to 5-HT₂Rs at 15 min postinternalization in the presence or absence of either GRK2 or GRK5 protein, normalized to 5-HT CTL samples (100%), and contain data from individual replicates performed in three independent experiments. Error bars indicate SDs. A two-sample Student’s t test was used to determine statistical significance. *, P < 0.05.

DISCUSSION

Viral utilization of a cellular receptor for functional internalization is a tightly controlled and yet complex process. In this study, we have further defined the mechanism by which JCPyV usurps human 5-HT₂Rs for internalization. A β-arrestin-binding domain within 5-HT₂Rs, the ASK motif, is critical for viral entry and infection (Fig. 1 and 3). Furthermore, G protein-coupled receptor kinases GRK2 and GRK5 are important for JCPyV infection (Fig. 5), with a specific role for GRK2 in facilitating JCPyV internalization (Fig. 6), while GRK5 promotes JCPyV and SV40 infection at a postinternalization step (Fig. 7). Additionally we have defined here that upon JCPyV or 5-HT activation of 5-HT₂Rs, β-arrestin directly engages the receptor. Furthermore, mutagenesis of the Ser residue within the ASK motif of 5-HT₂Rs reduces these β-arrestin-receptor interactions (Fig. 4), which are facilitated by GRK2 (Fig. 8). These results demonstrate a mechanism by which JCPyV activates β-arrestin and usurps its functionality to promote viral internalization, further defining and characterizing key host cell proteins involved in the utilization of 5-HT₂Rs for entry of JCPyV (Fig. 9).

FIG 9.

Model of JCPyV infectious cycle. JCPyV infection is initiated by binding to LSTc or glycosaminoglycans (GAGs) through capsid protein VP1 (1). Entry then requires 5-HT₂Rs, leading to recruitment of GRK2 (2) which, through GTPase activity, likely phosphorylates the ASK motif in the intracellular domains of the 5-HT₂ receptors. Once phosphorylated, activated β-arrestin can be recruited to the 5-HT₂Rs (3). This interaction leads to the further activation of the receptors and recruitment of intracellular proteins, clathrin, and AP2 (4), which are involved in receptor-mediated internalization. Through the formation of a clathrin-coated pit, the receptor and associated viral cargo could be internalized into an early endosome (5) (*, internalization of 5-HT₂Rs results in delivery of receptors into an early endosome but has not been demonstrated for JCPyV) and transferred to a caveolin-1+ vesicle (6) (demonstrated for 5-HT_2AR), prior to delivery of the virus to the endoplasmic reticulum (ER) (7). Within the ER, partial viral uncoating occurs allowing for nuclear delivery of the virion for transcription and replication (8), followed by eventual release of viral progeny (9).

By hijacking a cellular receptor, viruses select for specific signaling networks to transverse the plasma membrane and ensure arrival at the appropriate compartment for replication (48). Interestingly, polyomaviruses utilize different cellular attachment and entry factors (37, 59–63). The majority of polyomaviruses studied to date, including SV40, rely on ganglioside receptors for internalization by caveolin- or nonclathrin/noncaveolin-mediated endocytosis (37, 64). Internalization of polyomaviruses results in delivery to the same destination (59–62), the endoplasmic reticulum (ER), for uncoating prior to nuclear translocation (36–41). In contrast, while JCPyV also arrives at the ER for partial uncoating, after a low-affinity binding event to sialic acid-containing LSTc receptors or GAGs, it usurps the proteinaceous 5-HT₂R family for functional internalization by CME (21, 27, 63). Although this dichotomy in internalization mechanisms among polyomaviruses exists, it suggests that crosstalk occurs between endocytic pathways; this crosstalk is important for ensuring polyomavirus delivery to the appropriate intracellular compartment and replicative success.

Agonist-mediated activation of GPCRs recruits specific subsets of proteins, ultimately resulting in a cascade of signaling within the cell (46, 47). This receptor activation results in the following two signaling outcomes: G protein-dependent or β-arrestin-dependent/G protein-independent signaling pathways. GRKs are involved in the direct identification and phosphorylation of GPCRs, determining the signaling pathway activated by the receptor (65, 66). Furthermore, due to the requirement of β-arrestin for internalization (29), JCPyV likely utilizes a pathway similar to the canonical 5-HT-activated signaling cascade (65, 66). In phosphorylating GPCRs and mediating β-arrestin binding, GRKs sterically block the binding of G proteins from the receptor (46). While seven GRKs have been identified to date (57), the implication of either GRK2 or GRK5 in facilitating the internalization of 5-HT₂Rs have yielded mixed results (45, 47, 58). Through GRK2 phosphorylation of the Ser residue of 5-HT_2AR, β-arrestin then binds to the phosphate tag on the receptor initiating endocytosis (47). Phosphorylation by GRKs and subsequent binding by β-arrestin commit the receptor to β-arrestin-dependent internalization (51, 67–69). Interestingly, the rat-origin 5-HT_2AR contains an Asn-Cys-Thr (NCT) motif in place of the ASK motif, and internalization of rat-origin 5-HT_2AR is independent of β-arrestin and GRK2 (47, 58). However, expression of the human ASK motif in place of the NCT in the rat 5-HT_2AR confers a reliance on β-arrestin and GRK2 for internalization (47). Collectively, these findings demonstrate the critical nature of the human-derived ASK motif in promoting 5-HT_2AR-β-arrestin-dependent internalization.

Following GRK phosphorylation of 5-HT₂Rs initiating receptor activation, β-arrestin is recruited to the site of the activated receptor (42, 45, 67, 70) where direct interactions between β-arrestin and the phosphate residues on the receptor occur (Fig. 4 and 8) (47). In addition to the ASK motif, there are several other identified binding sites for β-arrestin on the third intracellular loop and in the C terminus, including the Asn-Pro-X-X-Tyr (NPXXY) and the Asp-Arg-Try (DRY) (28, 29, 71, 72). Interestingly, altering the ASK motif of 5-HT_2AR changes not only the β-arrestin bias of the receptor but also the recycling rate of the receptor, with the mutated receptor recycling more rapidly than the unmutated receptor, likely due to the contributions of β-arrestin (47). While receptor expression levels are not consistent across receptor subtypes, mutated 5-HT_2B or 5-HT_2C receptors do not exhibit differential intracellular or cell surface expression compared with their wild-type counterparts (Fig. 2). JCPyV utilization of 5-HT₂Rs for internalization requires an intact ASK motif, and disruption of this interaction either through receptor mutagenesis or siRNA of β-arrestin significantly reduces JCPyV internalization and infection (Fig. 1 and 3) (28, 29), although the surface expression of these receptors following mutagenesis is not significantly altered, suggesting that accessibility of the receptor by the virus is not hindered (Fig. 2) (29). However, mutagenesis of the ASK motif does not completely ablate β-arrestin binding (Fig. 4), suggesting that multiple receptor interactions with β-arrestin may be necessary to support JCPyV internalization and infection. While mutagenesis of the ASK motif correlates with a reduction in JCPyV entry, infection, and binding of β-arrestin with 5-HT₂Rs (Fig. 1, 3, and 4), the location of the ASK motif differs between the different receptor subtypes. This result suggests that the conservation of the ASK motif within the receptor, rather than the location of the motif, may be the critical driver for JCPyV utilization of this receptor subfamily. Moreover, while GRK2 had been previously identified to facilitate β-arrestin binding with the ASK motif of 5-HT_2AR for canonical 5-HT-stimulated signaling, we demonstrate here that a similar mechanism occurs for the 5-HT_2BR and 5-HT_2CR, regardless of whether 5-HT or JCPyV serves as the receptor stimulus (Fig. 4 and 8). Collectively these results suggest that JCPyV may be hijacking these cellular receptors and activating the typical 5-HT driven pathway, in a conserved manner, to ensure delivery of the virus within the cell. In addition, as β-arrestin binding does not completely correlate with the reduction in JCPyV infection following GRK2 siRNA (Fig. 4 and 8), there may be additional roles for GRK2 in JCPyV entry aside from the identification of the ASK motif β-arrestin binding site in 5-HT₂Rs. Alternatively, a recent study by Assetta et al. suggested that the Pro residue adjacent to the DRY motif is necessary for successful JCPyV infection and may contribute to additional β-arrestin receptor binding (28). However, the Pro residue is thought to contribute to phosphorylation-independent β-arrestin-mediated endocytosis (73). Therefore, GRK2 may identify an additional β-arrestin binding site for internalization. Although GRK2 and GRK5 are both G protein receptor kinases, they belong to different GRK subfamilies, with GRK2 belonging to the βARK subfamily (GRK2 and GRK3) and GRK5 to the GRK4-like subfamily (GRK4, GRK5, and GRK6) (46). Compensation between GRKs has been reported to occur within GRK subfamilies, with a compensation effect observed between GRK5 and GRK6 (74). However, a GRK5/6 knockout mouse results in lethality, suggesting the inability of compensation for the loss of these proteins by GRKs from other subfamilies (74). Furthermore, knockout of GRK2 in mice also results in a lethal phenotype, again highlighting the importance of distinct individual roles of these proteins within cells (75). While both GRK2 and GRK5 siRNAs individually reduce JCPyV infection (Fig. 5), JCPyV infection is further reduced upon a combinatory siRNA of both GRK2 and GRK5, suggesting independent roles during infection (Fig. 5).

Although knockdown of GRK2 or GRK5 reduces JCPyV infection (Fig. 5), only GRK2 reduces JCPyV internalization (Fig. 6), suggesting the involvement of GRK5 in a postinternalization step in the infection process. Moreover, transfection of either JCPyV or SV40 infectious clones, bypassing viral attachment, entry, and trafficking, results in reduced infection following GRK5 siRNA (Fig. 7). Interestingly, previous studies have demonstrated a role for GRKs independent of GPCR kinase function. GRK5 has been identified to play roles in receptor tyrosine kinase activity (76), and reduced GRK5 levels correlate with diminished ERK signaling (76), of which both are necessary for productive JCPyV infection (53–56). GRKs, including GRK2 and GRK5, have also been implicated with the regulation of cytoskeletal dynamics and microtubule rearrangement through modulation of proteins in the ezrin-radixin-moesin (ERM) family (77–80). Furthermore, GRK5 has been associated with the regulation of nuclear factors, including activation of NF-κB signaling (81, 82), a DNA damage response pathway that not only promotes JCPyV replication (83) but also is activated by SV40 during infection (84). Moreover, GRK5 also plays a role in facilitating activation of nuclear factor of activated T cells (NFAT) (85), a transcription factor downstream of the MAPK cascade. Importantly, NFAT has been associated with the promotion of SV40 and JCPyV transcription and replication (86, 87). While SV40 does not utilize the MAPK cascade for replication (54, 88), GRK5 may play a role in facilitating activation of NFAT outside the MAPK cascade during infection. Collectively, these findings may suggest potential roles for GRK5 in JCPyV and SV40 infection following attachment, internalization, and trafficking.

The importance of 5-HT₂Rs, GRKs, and β-arrestin introduces an enticing target for the development of GPCR-directed antiviral therapies (89). Gaining a deeper understanding of the mechanism by which JCPyV utilizes 5-HT₂Rs to mediate entry will further elucidate targets for GPCR therapies. Moreover, 5-HT receptor antagonists prevent the internalization of other viruses, including chikungunya virus, coronavirus mouse hepatitis virus, and Ebola virus (90, 91). The common effect of 5-HT receptor antagonists suggests that multiple viruses may target overlapping GPCR-mediated signaling pathways. Thus, development of GPCR-mediated antiviral therapies could potentially be applied broadly for the treatment of multiple significant viral pathogens. Developing therapies to prevent GPCR activation through β-arrestin- or 5-HT₂R-specific therapies represents an area for the development of new antivirals (92, 93). β-Arrestin-based therapeutics, including carvedilol and TRV120027, have demonstrated promising results when used as therapies for heart failure (94–97). Furthermore, β-arrestin has been shown to be upregulated in individuals with MS, a patient group that is at risk for JCPyV-induced PML, while taking immunosuppressive therapies like natalizumab (98), warranting exploration of a potential role for β-arrestin in PML development.

Within this study, we have identified that 5-HT₂Rs, capable of supporting JCPyV infection, bind to β-arrestin, and that this interaction is critical for JCPyV internalization. Furthermore, we have identified that a key cellular kinase, GRK2, is required to mediate JCPyV internalization and infection. These results further define JCPyV-activated 5-HT₂R signaling pathways that are critical for JCPyV infection, contributing to our understanding of viral usurped internalization pathways. Collectively, conclusions from this work identify new potential targets for the design of therapeutic agents to treat or prevent JCPyV infection and the development or progression of PML.

MATERIALS AND METHODS

Cells, viruses, antibodies, plasmids, reagents, and siRNAs.

SVG-A cells (99) were cultured in complete minimum essential medium (MEM) (Corning) supplemented with 10% fetal bovine serum (FBS), 1% penicillin-streptomycin (P/S) (Mediatech, Inc.), and 0.2% Plasmocin prophylactic (Invivogen). HEK293A cells were maintained using Dulbecco’s modified Eagle medium (DMEM) (Corning) containing 10% FBS, 1% P/S, and 0.2% Plasmocin prophylactic. HEK293A cells stably expressing 5-HT₂R subtypes 2A, 2B, and 2C (27) (NCBI accession numbers 3356, 3357, and 3358, respectively) in fusion with YFP were cultured in complete DMEM additionally containing 1% G418 (Corning). All cell types were grown in a humidified incubator at 37°C with 5% CO₂. Cell lines used for this study were generously provided by the Atwood Laboratory (Brown University), of which all were verified by ATCC.

JCPyV strain Mad1-SVEΔ was generated and produced (provided by the Atwood Laboratory, Brown University) as described previously (100, 101). Labeling of JCPyV with Alexa Fluor 647 was previously described (63), and determination of JCPyV titer was performed in SVG-A cells by fluorescent focus unit (FFU) assay for infectivity. Noninternalized virus was neutralized using JCPyV-neutralizing antiserum (1:10,000; provided by Atwood Laboratory). Antibodies used to detect infectivity by FFU assay include PAB597 (1:10), an anti-JCPyV VP1 monoclonal antibody (mAb) grown from a hybridoma (generously provided by Ed Harlow); PAB962 (1:3), a mAb grown from a hybridoma for specific detection of JCPyV large T antigen (T-Ag; generously provided by the Tevethia Laboratory, Penn State University) (102); and secondary polyclonal goat anti-mouse and anti-rabbit Alexa Fluor 594 or 647 antibodies (1:1,000; Thermo Fisher). Antibodies used for Western blot analysis include rabbit GRK5 polyclonal (1:500; catalog number 64943; Abcam), mouse GRK2 mAb (1:1,000; MA5-15840; ThermoFisher), β-arrestin 1/2 rabbit mAb (1:500; 4674S; Cell Signaling Technology [CST]), GFP rabbit mAb (1:1,000; for detection of YFP; 2956; CST), and mouse and rabbit GAPDH (1:2,000; 9484, and 9485, respectively; Abcam). Plasmids used in this study include WT GRK2 (a gift from Robert Lefkowitz, Addgene; RRID Addgene_14691) (103) and kinase-defective GRK2-K220R (a gift from Robert Lefkowitz, Addgene; RRID Addgene_35403) (104). Plasmids for mutated 5-HT_2B and 5-HT_2CRs, as well as accession numbers, are included in Table 1. 5-HT_2AR plasmids used in this study have been previously described (29). Plasmids used for transfection of the infectious clones include JCPyV-Puc19, JCPyV strain JC12 DNA, a subclone of Mad1-SVEΔ, subcloned into Puc19 at a BamHI site (105); and SV40-Puc19, SV40 strain 776 DNA subcloned into Puc19 at a EcoRI site (generously provided by the Atwood Laboratory, Brown University). siRNAs used in this study targeted EGFR (irrelevant control; 6482; Cell Signaling Technology), GRK2 (AM51331; ThermoFisher), and GRK5 (AM16708; ThermoFisher). Concentrations of siRNAs used in this study are as described in the “siRNA treatment” Materials and Methods section. siRNAs were screened for toxicity through a propidium iodide flow cytometry assay as in reference 29.

TABLE 1.

Conditions for site-directed mutagenesis used in this study

Mutation	Template length (ng)	Primer sequences (5′-3′) for mutagenesis of 5-HT2AR-YFP		Primer conc (ng) per reaction	Phusion HF polymerase (units)	Annealing temp (°C)	GenBank accession no.
		Forward primer	Reverse primer
5-HT2B-S323A-R	50	ATCCCTAGGACCTTTGCGGCTCTCTGTTCGTTG	CAACGAACAGAGAGCCGCAAAGGTCCTAGGGAT	175	2	64	MT647189
5-HT2B-SK323-24AA-R	50	AAAACACAATCCTAGGACCGCTGCGGCTCTCTGTTCGTTGGAAA	TTTCCAACGAACAGAGAGCCGCAGCGGTCCTAGGGATTGTGTTTT	175	2	63	MT647190
5-HT2B-K324A-R	50	CATTTCCAACGAACAGAGAGCCTCAGCGGTCCTAGGGATT	AATCCCTAGGACCGCTGAGGCTCTCTGTTCGTTGGAAATG	175	2	63	MT647191
5-HT2C-S310A-R	50	CCCAAGGACTTTCGCAGCTTTTCTTTCATTGTTGATAGC	GCTATCAACAATGAAAGAAAAGCTGCGAAAGTCCTTGGG	175	2	64	MT647192
5-HT2C-SK310-11A-R	50	ACAATCCCAAGGACTGCCGCAGCTTTTCTTTCATTGTTGATAGCCTGC	GCAGGCTATCAACAATGAAAGAAAAGCTGCGGCAGTCCTTGGGATTGT	175	2	65	MT647193
5-HT2C-K311A-R	50	CACAAAGAAAACAATCCCAAGGACTGCCGAAGCTTTTCTTTCATTGTTGATA	TATCAACAATGAAAGAAAAGCTTCGGCAGTCCTTGGGATTGTTTTCTTTGTG	175	2	63	MT647194

Site-directed mutagenesis of 5-HT₂ receptors.

Generation of 5-HT_{2A, 2B, 2C}R-YFP fusion construct plasmids was previously described (27, 102). Individual amino acids within the ASK motif of 5-HT_2BR and 5-HT_2CR were altered by site-directed mutagenesis. 5-HT_2AR ASK motif mutagenesis was previously described (29). Primers for each desired mutation were designed using Agilent QuickChange primer design software and were high-performance liquid chromatography (HPLC) purified by Integrated DNA Technologies (IDT) (Table 1). Individual mutagenesis reactions were designed containing wild-type DNA template, forward and reverse primers, Phusion high-fidelity (HF) DNA polymerase (New England BioLabs), and 5× Phusion HF buffer (New England BioLabs), as described (Table 1). All sample mutagenesis reactions were conducted in a BioRad thermocycler (c1000) as follows: samples were denatured at 95°C for 30 sec, followed by 30 cycles of 95°C for 30 sec, annealing for 1 min (various temperatures listed in Table 1), 72°C for 5 min, and a final extension at 72°C for 7 min. Following mutagenesis, template DNA was digested by incubation with 10 U of Dpn1 enzyme (New England BioLabs) in a 37°C water bath for 1 h. Plasmids were then transformed in DH5α-competent cells (Invitrogen), and sequences were verified by The University of Maine Sequencing Facility, analyzed by MacVector (version 15.5), and purified by using a plasmid DNA Maxiprep kit (Qiagen). GenBank accession numbers are provided in Table 1.

Transfection of plasmids in HEK293A and SVG-A cells.

For transfection of control GFP, GRK2 wild-type, or GRK2 K220R mutant plasmids, SVG-A cells were plated to 80% confluence in a 24-well plate in 500 μl in MEM only supplemented with 10% FBS (lacking antibiotics). Transfection complexes containing 1 μg DNA:1 μl of Lipofectamine 2000 transfection reagent (Invitrogen)/well for were incubated at room temperature (RT) for 20 min prior to addition to cells. Cells transfected with wild-type GRK2 or GRK2 K220R plasmids were cotransfected with GFP to allow for determination of transfection efficiency (0.5 μg of wild-type GRK2 or GRK2 K220R DNA and 0.5 μg of GFP DNA/well). Cells treated with transfection complexes were incubated at 37°C for 4 h. The transfection medium was removed, and cells were fed with complete media and incubated at 37°C for 20 h. Transfection efficiency was determined through GFP expression by fluorescence microscopy and determined to be at least 70%, and cells were infected as indicated. For expression of 5-HT_{2A, 2B, 2C}R-YFP wild-type and mutated plasmids, HEK293A cells were seeded to 80% confluence in 500 μl/well (24-well plate) or in 100 μl/well (96-well number 1.5 glass-bottom plates; CellVis). All cells were plated in DMEM supplemented with 10% FBS (lacking antibiotics). Transfection complexes were performed as described above with a final concentration of 1 μg DNA/well (24-well plate) and 0.16 μg DNA/well (96-well plate). Transfection efficiency was determined by fluorescence microscopy for expression of YFP, which was determined to be at least 70%, and cells were fixed or infected as indicated.

Cell surface and total cellular fluorescence expression of 5-HT_2B and 5-HT_2C wild-type and mutated receptors.

Cell surface expression was performed as previously described (29). Briefly, following transfection of wild-type and mutated 5-HT_2BR- and 5-HT_2CR-YFP plasmids in HEK293A cells in number 1.5 96-well glass bottom plates, cells were incubated for 24 h and fixed in 4% paraformaldehyde (PFA) at RT for 10 min. After fixation, cells were washed with 1× phosphate-buffered saline (PBS) 3 times and incubated in blocking reagent (1× PBS containing 2% goat serum, 0.2% Triton X-100, and 0.1% bovine serum albumin [BSA]) at RT for 1 h. Cells were then stained with an anti-pan-cadherin antibody (1:200; ab6528; Abcam) in blocking reagent at 4°C rocking overnight (O/N). Cells were washed with 1× PBS and subsequently stained with secondary Alexa Fluor 647 antibody (1:1,000) in blocking reagent at RT for 1 h. Cells were washed with 1× PBS, and nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; 1:1,000) at RT for 10 min prior to imaging by confocal microscopy. Samples were imaged utilizing an Olympus laser scanning confocal microscope (model IX81) at 60× magnification (oil immersion) and Fluoview software (version 04.01.01.05). Fields of view were defined using DAPI staining and then viewed via z sectioning for background-corrected samples within an applied threshold, as previously described (29). Images were acquired using 405/635- and 543/488-nm argon laser lines; at least 10 fields of view were acquired for each sample. The cell surface expression of mutant receptors was quantified using ImageJ by defining the percentage of overlap between the pan-cadherin cell surface marker and YFP expression of transfected receptors. Colocalization analysis in ImageJ was used to determine the Mander’s overlap coefficient (n = 30 cells/sample/replicate) for 3 independent experiments. Total cellular and intracellular receptor fluorescence were determined using ImageJ software by drawing regions of interest outside individual cells, excluding the plasma membrane (total cell fluorescence) and a second region of interest inside the plasma membrane (intracellular fluorescence), as defined by pan-cadherin cell surface marker. Then, corrected fluorescence intensity was determined by ImageJ software by measuring the integrated density of receptor fluorescence, corrected for size and background of the ROI, using the equation corrected cell fluorescence = integrated density of the ROI – (ROI area × image background intensity) for each cell measured (n = 30 cells/sample/replicate) for 3 independent experiments (106). Data are depicted as box and whisker plots performed using R (version 3.6.1) equipped with R Studio (version 1.2.1335; 2019; Integrated development of RStudio, Inc.) for three independent experiments.

JCPyV and SV40 infection.

Following siRNA treatment, chemical inhibitor treatment, or transfections, SVG-A or HEK293A-5-HT₂R cells were infected with JCPyV or SV40 at the multiplicity of infections (MOIs) indicated in the figure legends. Virus inoculum was prepared using either MEM (SVG-A) or DMEM (HEK293A), containing 2% FBS and 1% P/S, and 200 μl (24-well plates) or 300 μl (12-well plates) was added per well. Cells were incubated in a 37°C humidified incubator containing 5% CO₂ for 1 h, complete medium was added to each well, and cells were further incubated for 48 h (T-Ag; HEK293A-5-HT₂R cells) or 72 h (VP1, SVG-A cells). Cells were washed with 1× PBS and fixed with 4% paraformaldehyde (PFA) at RT for 10 min, followed by staining for indirect immunofluorescence. For infectivity staining, cells were washed with 1× PBS and permeabilized with 1% Triton X-100-PBS (ThermoFisher) at RT for 15 min. Cells were incubated with mAb PAB962 (T-Ag) or PAB597 (VP1) to detect T-Ag or VP1 protein in SVG-A or HEK293A-5-HT₂R cells at 48 or 72 hpi, respectively (107), at 37°C for 1 h. Following incubation, cells were washed with 1× PBS three times, followed by incubation with polyclonal secondary antibody goat anti-mouse Alexa Fluor 488 or 594 antibodies. Infection was determined through quantifying the number of VP1-positive (VP1+) or T-Ag-positive (T-Ag+) cells per 10× visual field divided by number of DAPI-positive cells within the field (percent infection). The number of DAPI-positive cells per field was determined using a binary code as previously described (29, 54), using Nikon NIS-Elements Basic Research software (version 4.50.00; 64 bit). The average percent infection was then normalized to that of the control as indicated.

Viral attachment by flow cytometry.

Following siRNA treatment or transfection of wild-type or mutated receptors, SVG-A or HEK293A cells were removed from plates using the CellStripper reagent (Corning) and centrifuged at 414 × g at 4°C for 5 min. Cells were then washed with 1× PBS, centrifuged, and resuspended in cell type-specific phenol-free MEM or DMEM (Corning) containing 10% FBS. Cells were prechilled on ice for 45 min. Chilled cells were pelleted by centrifugation, resuspended in 500 μl of either complete phenol-free MEM or DMEM containing Alexa Fluor 647-labeled JCPyV (JCPyV-647; MOI indicated in figure legend), and incubated on ice for 1 h for viral attachment. Cells were centrifuged, washed with 1× PBS, and fixed in 4% PFA for 10 min prior to resuspension in 300 μl of 1× PBS. Cells were then analyzed by flow cytometry for viral attachment using a FACSCanto system (BD Biosciences), equipped with 488 and 633 AP-C laser lines (Becton, Dickinson, and Company) for at least 20,000 events, followed by analysis with BD FACSDiva (Becton, Dickinson, and Company) and FlowJo (TreeStar, Inc.) software. Gating was performed to exclude complex and dead cells using FlowJo software.

Viral entry by confocal microscopy.

Cells were plated to 50% confluence in number 1.5 96-well glass bottom plates (CellVis) prior to siRNA treatment or transfection with wild-type or mutated plasmids. After transfection, cells were prechilled on ice for 45 min and then incubated with JCPyV-647 (MOIs indicated in figure legends) on ice for 1 h for viral attachment; plates were then shifted to a 37°C humidified incubator with 5% CO₂ for 2 h for viral entry. Following viral entry, cells were washed with 1× PBS and fixed in 4% PFA for 10 min at RT. Cells were washed with 1× PBS and then subsequently incubated in blocking reagent (1× PBS containing 2% goat serum, 0.2% Triton X-100, and 0.1% BSA) at RT for 1 h while rocking. Cells were then incubated with a primary anti-pan-cadherin antibody (1:200) in blocking reagent, at 4°C overnight. Cells were washed with 1× PBS and incubated with secondary Alexa Fluor 488 or 647 antibody (1:1,000) in blocking reagent at RT for 1 h. Cells were washed with 1× PBS and then incubated with DAPI (1:1,000) for 10 min in 1× PBS for nuclear visualization. Samples were visualized by confocal microscopy utilizing an Olympus laser scanning confocal microscope (model IX81) at 60× magnification (oil immersion) and Fluoview software (version 04.01.01.05). Fields of view were defined using DAPI staining, and Fluoview software single-measurement analysis was used to define regions of interest (ROI) using z-sectioning and excluding the plasma membrane (using either pan-cadherin or DIC overlay) to measure relative internalized virus by relative fluorescence units per cell for background-corrected samples within an applied threshold, as previously described (29). Images were acquired using 405/635- and 543/488-nm argon laser lines. Cross sections of individual cells were analyzed (at least 30 cells per sample) for siRNA-treated samples (SVG-A cells) or cells expressing wild-type or mutated receptors (HEK293A cells). Microscopy experiments were performed independently 3 times in triplicate, containing at least 30 cells per sample per replicate; graphs are comprised of 3 independent replicates (n = 90).

siRNA treatment.

SVG-A or 5-HT₂R stably expressing HEK293A cells were seeded to 50% confluence in 12-well plates. siRNA transfections targeting either EGFR (irrelevant control), GRK2, or GRK5 were performed using RNAiMax reagent (ThermoFisher) per the manufacturer’s instructions (7.5 pmol/well). Following siRNA transfection, cells were incubated in a humidified incubator at 37°C with 5% CO₂ for 72 h. Efficiency of the siRNA transfection was determined using Block-iT Alexa Fluor red oligonucleotide control (Life Technologies) at 48 h following transfection. Following incubation with siRNAs, cells were either processed for Western blot analysis, infected with JCPyV or SV40, (MOIs indicated in figure legends), or were transfected with the infectious clones of JCPyV or SV40.

siRNA Western blot analysis.

To determine the efficiency of knockdown by siRNA, Western blot analysis was performed as described in reference 29. In brief, SVG-A or HEK293A-5-HT₂R cells were washed with 1× PBS and subsequently removed from wells with a cell scraper. Cells were pelleted by centrifugation at 414 × g at 4°C for 5 min; pellets were resuspended in Tris-HCl lysis buffer (1 mM EDTA, 50 mM Tris-HCl, and 120 mM NaCl) containing protease inhibitors (1:10; Sigma-Aldrich) and phosphatase inhibitors (1:100; Sigma-Aldrich) and incubated on ice for 10 min. Cellular pellets were then sonicated at 35% amplitude, on ice, for 10 s, which was repeated three times. Insoluble cellular material was then pelleted at 21,130 × g at 4°C for 10 min. Cellular protein-containing lysates were mixed 1:1 with Laemmli sample buffer (Bio-Rad), boiled at 95°C for 5 min, and processed by SDS-PAGE with a 10% TGX minigel (Bio-Rad). Resolved proteins were transferred to a nitrocellulose membrane (Bio-Rad) using a semidry transblot apparatus (Bio-Rad) at 2.5 amps (25 V) for 3 min. Membranes were equilibrated in 1× Tris-buffered saline (TBS) for 5 min, blocked with Odyssey blocking buffer (Li-Cor) at RT for 1.5 h, and then washed with 1× TBS-T for 5 min three times. Membranes were incubated with primary antibodies targeting either GRK2 (1:1,000) or GRK5 (1:500) and GAPDH (housekeeping; 1:2,000) in Odyssey blocking buffer at 4°C overnight with rocking. Following extensive washing with 1× TBS with Tween 20 (TBST), membranes were incubated with secondary antibodies (Li-Cor) in Odyssey blocking buffer at RT for 1 h while rocking and then were washed. Imaging was performed using a Li-Cor Odyssey CLx system. siRNA-induced knockdown of protein expression was determined by defining the relative fluorescence for each housekeeping control protein band and target protein band using Li-Cor ImageStudio software (version 5.2). Target protein bands were normalized to the housekeeping control, and the reduction in protein expression was determined in comparison with the control siRNA-treated samples. Knockdown by siRNA was determined to be at least 70% for all experiments reported or shown in Table 2.

TABLE 2.

Average knockdown of cellular proteins following siRNA

Cell type	Target protein	Avg knockdown (%)
SVG-A	GRK2	71.5
SVG-A	GRK5	75.74
SVG-A	GRK2	70.22
	GRK5	72.04
HEK293A-5-HT2AR	GRK2	85.11
HEK293A-5-HT2AR	GRK5	79.05
HEK293A-5-HT2BR	GRK2	72.51
HEK293A-5-HT2BR	GRK5	82.85
HEK293A-5-HT2CR	GRK2	78.78
HEK293A-5-HT2CR	GRK5	90.29

Transfection of JCPyV and SV40 infectious clones.

SVG-A cells were plated in 24-well plates to 50% confluence followed by transfection of either EGFR (CTL), GRK2, or GRK5 siRNA using the RNAiMax reagent (ThermoFisher) per the manufacturer’s instructions (3.75 pmol/well). Following siRNA transfection, cells were incubated in a humidified incubator at 37°C with 5% CO₂ for 72 h. Linearization of viral plasmids was accomplished through digestion with BamHI (JCPyV-Puc19; Promega) or EcoRI (SV40-Puc19; Promega) at 37°C for 2 h, and the successful digestion of plasmids was determined by agarose gel electrophoresis. Cells were then transfected with 1 μg of DNA containing the linearized plasmids of either JCPyV-Puc19 or SV40-Puc19 DNA using the Fugene reagent (Roche) at a ratio of 1.5 μl Fugene:1 μg of DNA prior to incubation at 37°C. At 24 h posttransfection, cells were fed with 500 μl of complete medium containing antibiotics and additionally supplemented with 1% amphotericin B (Corning). Cells were fixed and stained for newly synthesized VP1 at either 4 or 7 days posttransfection.

5-HT₂R pulldown in HEK293A and HEK293A-5-HT₂R cells.

5-HT₂R pulldown assays were adapted from reference 47. Briefly, HEK293A or HEK293A-5-HT₂R cells were plated to 90% confluence in 6-well plates in DMEM containing 10% FBS (no antibiotics). Cells were then either transfected with siRNAs for 72 h (HEK239A-5-HT₂R) or WT or mutated 5-HT_2AR, 5-HT_2BR, or 5-HT_2CR plasmids (HEK293A; 10 μg DNA/well/sample) at 37°C for 4 h. At 4 h posttransfection, the medium was replaced with incomplete DMEM (no FBS or antibiotics) for serum starvation O/N (transfected samples) or for 4 h (siRNA-treated samples). Cells were then prechilled at 4°C for 45 min, followed by incubation with either 5-HT (200 μM/well) or JCPyV (MOI indicated in figure legend) at 4°C for 1 h for attachment. Cells were spiked with warm DMEM (10% FBS) and shifted to a 37°C incubator for 15 min prior to removal from plates by scraping. Cells were pelleted by centrifugation at 414 × g at 4°C for 5 min and resuspended in 200 μl lysis buffer (10 mM Tris HCl, 150 mM NaCl, 0.5 mM EDTA, and 0.5% Triton X-100 [pH 7.5]) additionally supplemented with phosphatase (1:100; Sigma-Aldrich) and protease (1:10; Sigma-Aldrich) inhibitors for 30 min with vigorous pipetting every 10 min. Samples were pelleted by centrifugation at 21,130 × g at 4°C for 10 min, and 50 μl of supernatant was mixed 1:1 with Laemmli sample buffer (Bio-Rad) and boiled at 95°C for 5 min (input). GFP-Trap magnetic agarose beads (25 μl/sample) were washed 3 times with wash buffer (10 mM Tris HCl, 150 mM NaCl, and 0.5 mM EDTA [pH 7.5]) and separated using a magnetic tube rack until clear per manufacturer instructions (CST). The remaining supernatant was mixed with beads and incubated at RT for 2 h while rotating. Beads were pelleted using a magnetic tube rack, and 75 μl of supernatant was mixed 1:1 with Laemmli sample buffer (Bio-Rad) and boiled at 95°C for 5 min (nonbound). Beads were washed three times with wash buffer, transferred to a new sterile microcentrifuge tube in 50 μl of wash buffer, mixed 1:1 with Laemmli sample buffer (Bio-Rad), and boiled at 95°C for 10 min. After magnetic separation, the supernatant was transferred to a new microcentrifuge tube (bound). Negative-control beads (magnetic agarose beads lacking a preconjugated antibody) were also used to detect nonspecific bead interactions (not shown). Half of the bound samples were processed by SDS-PAGE using a 10% TGX mini gel (Bio-Rad). Resolved proteins were transferred to a nitrocellulose membrane (Bio-Rad) using a semidry transblot apparatus (Bio-Rad) at 2.5 amps (25 V) for 3 min. Membranes were then equilibrated in 1× TBS at RT for 5 min and blocked with Odyssey blocking buffer (Li-Cor) at RT for 1.5 h. Membranes were washed with 1× TBS-T three times for 5 min, incubated with a primary monoclonal antibody targeting β-arrestin 1/2 (1:500; 4674; CST), and then diluted in blocking buffer at 4°C O/N while rocking. Membranes were washed and imaged on a Li-Cor Odyssey CLx system, stripped using Restore Western blot stripping buffer according to manufacturer’s instructions (Bio-Rad), and reprobed with a primary monoclonal anti-GFP antibody (1:1,000; 2956; CST) for 5-HT₂R-YFP detection. The percentage of bound β-arrestin was determined by defining the relative fluorescence for each β-arrestin protein band and corresponding receptor protein band using Li-Cor ImageStudio software (version 5.2). β-Arrestin bands were normalized to the receptor expression within each sample, and the percentage of bound β-arrestin was determined compared with the WT- or CTL-siRNA-transfected samples treated with 5-HT.

Statistical analysis.

All experiments were performed in triplicate with three independent replicates. Microscopy experiments were performed independently 3 times, containing at least 30 cells for triplicate samples per replicate; graphs are comprised of 3 independent replicates (n = 90). A two-sample Student’s t test was used to determine statistical significance, assuming unequal variance, by comparing the mean values of triplicate samples collected and was performed in Microsoft Excel. Additionally, the normality of distribution of the data was determined using the Shapiro-Wilk’s normality test and a quantile-quantile (Q-Q) plot in RStudio (version 1.2.1335, 2019; Integrated development of RStudio, Inc.). A pairwise Wilcoxon signed-rank test was used for comparing data populations that were not normally distributed (microscopy experiments containing samples of least 90 cells). For these data, the statistical analysis was performed in RStudio, and plots were generated using ggplot2 by plotting the raw values that had been imported from Excel into RStudio using library XLConnect. Significance was determined by a P value of <0.05.

Data availability.

Sequences of mutated 5-HT₂Rs were deposited in GenBank with the following accession numbers: MT647189 (5-HT_2B-S323A-R), MT647190 (5-HT_2B-SK323-24AA-R), MT647191 (5-HT_2B-K324A-R), MT647192 (5-HT_2C-S310A-R), MT647193 (5-HT_2C-SK310-11A-R), and MT647194 (5-HT_2C-K311A-R) (also displayed in Table 1).