SP-R210 isoforms of Myosin18A modulate endosomal sorting and recognition of influenza A virus infection in macrophages

Abstract

Influenza A virus (IAV) infection causes acute and often lethal inflammation in the lung. The role of macrophages in this adverse inflammation is partially understood. The surfactant protein A receptor 210 (SP-R210) consists of two isoforms, a long (L) SP-R210_L and a short (S) SP-R210_S isoform encoded by alternative splicing of the myosin 18A gene. We reported that disruption of SP-R210_L enhances cytosolic and endosomal antiviral response pathways. Here, we report that SP-R210_L antagonizes type I interferon β (IFNβ), as depletion of SP-R210_L potentiates IFNβ secretion. SP-R210 antibodies enhance and attenuate IFNβ secretion in SP-R210_L replete and deficient macrophages, respectively, indicating that SP-R210 isoform stoichiometry alters macrophage function intrinsically. This reciprocal response is coupled to unopposed and restricted expression of viral genes in control and SP-R210_L-deficient macrophages, respectively. Human monocytic cells with sub-stoichiometric expression of SP-R210_L resist IAV infection, whereas alveolar macrophages with increased abundance of SP-R210_L permit viral gene expression similar to murine macrophages. Uptake and membrane binding studies show that lack of SP-R210 isoforms does not impair IAV binding and internalization. Lack of SP-R210_L, however, results in macropinocytic retention of the virus that depends on both SP-R210_S and interferon-inducible transmembrane protein-3 (IFITM3). Mass spectrometry and Western blot analyses indicate that SP-R210 isoforms modulate differential recruitment of the Rho-family GTPase RAC1 and guanine nucleotide exchange factors. Our study suggests that SP-R210 isoforms modulate RAC-dependent macropinosomal sorting of IAV to discrete endosomal and lysosomal compartments that either permit or prevent endolysosomal escape and inflammatory sensing of viral genomes in macrophages.

Introduction

The ability of influenza A virus to evade innate immune recognition depends on its capacity to gain entry and rapidly release its viral genome through fusion of endosomal and viral membrane in acidified endosomes. This process enables transport and then replication of viral genomes in the nucleus [1]. IAV infection can still be productive or abortive despite infecting cells. Productive infection results in packaging and release of viral progeny as occurs in epithelial cells across the respiratory tract. Infection of macrophages is largely abortive wherein the virus retains the ability to replicate viral genes and express viral proteins while packaging and release of viral progeny is attenuated [2-5]. Both productive and abortive infection can suppress activation of Retinoic acid-inducible gene-I and nuclear factor kappaB (NFκB) signaling pathways, mainly through the actions of non-structural protein 1 (NS1). NS1 interferes with mRNA processing to attenuate host protein expression in favor of viral gene expression. As a consequence, NS1 ensures survival of the virus and facilitates innate immune escape during the incubation period of the virus in vivo [1, 6, 7].

The biological consequences and mechanisms of restrictive infection, however, are less clear. Studies suggest that endo-lysosomal containment of the virus enhances survival, activation of innate immune signaling, and sorting of viral components for antigen presentation [8-11]. IAV is internalized through both clathrin- and actin-dependent endocytic and macropinocytic mechanisms [12-15]. Macrophages and dendritic cells utilize both constitutive and stimulated mechanisms to capture and process antigen critical to development of cell-mediated immunity [16]. Known activators of macrophage macropinocytosis include lipopolysaccharide (LPS), macrophage colony stimulating factor 1 (M-CSF), and the chemokine C-X-C motif chemokine 12 (CXCL12) [17, 18]. Constitutive and induced expression IFITM3 in different cell types suppresses IAV infection through a number of endosomal mechanisms that may depend on cell type. IFITM3 prevents endosomal escape of IAV genomes by preventing fusion of IAV hemagglutinin with endosomal membrane and by promoting trafficking of IAV for degradation in lysosomes [9, 19-21]. IFITM3 functions depend on cholesterol, palmitoylation, and other posttranslational modifications. IFITM3 does not affect the late stage blocks that result in abortive infection of macrophages [10, 11, 19-26]. On the other hand, IFITM3 suppresses inflammatory cytokine production in response to viral entry in endosomes of myeloid cells [8].

Here, we studied the role of the SP-A receptor SP-R210 in internalization and endocytic fate of IAV. We reported that dominant negative disruption of SP-R210_L alters phenotype, activation, trafficking of innate immune receptors, and enhances antiviral responses to IAV infection in Raw264.7 macrophages [27, 28]. Macrophages express a long SP-R210_L and short SP-R210_S isoform encoded from alternatively spliced MYO18A mRNA variants. These encode MYO18Aα and MYO18Aβ protein isoforms. MYO18Aα isoforms are distinguished by an amino-terminal extension that contains a KE actin-binding motif and a PDZ domain. Splicing of small exons results in MYO18Aα and MYO18Aβ isoform variation in different cell types [29, 30]. Both SP-R210 isoforms are expressed by terminally differentiated alveolar and peritoneal macrophages, whereas circulating monocytes and immature monocytic cells exhibit predominant expression of the short isoform [31, 32]. SP-R210 isoforms modulate macrophage activation and phagocytosis of SP-A-coated bacteria [33-36]. Ectopic expression of SP-R210_S in COS-1 cells, which lack endogenous expression of SP-R210_S and non-muscle myosin IIA, confers attachment but not internalization of SP-A-opsonized bacteria [29, 37]. MYO18Aα variants have been localized to the cell-surface, cortical membrane, Golgi, and endoplasmic reticulum modulating phagocytosis, intracellular transport, exocytosis, differentiation, and cell migration in different cell types [38-41]. MYO18Aα localizes to integrin adhesion complexes in membrane ruffles of K562 lymphoblasts [42]. IFNγ enhances localization of MYO18Aα to macrophage phagosomes [43].` Furthermore, MYO18Aα modulates B lymphocyte differentiation and development of humoral immunity [41]. The CD245α and CD245β isoforms of MYO18A enhance cytotoxic activities in natural killer cells [40]. MYO18A translocates from the plasma membrane to the viral assembly complex of the human cytomegalovirus in fibroblasts [44]. In hepatocytes, MYO18Aα mediates budding of hepatitis C virus [45]. Liver cells, however, express MYO18A mRNA but not the macrophage SP-R210 isoforms [29], indicating expression of additional tissue specific MYO18A variants in liver. Our previous studies using LPS showed that SP-R210 isoforms coordinate phlogistic and non-phlogistic macropinocytosis with intracellular retention and trafficking of toll-like receptor 4 (TLR4) to the cell membrane, respectively [28]. Here, we show that SP-R210 isoforms play discrete roles in macropinosomal sorting, escape, and restriction of IAV in the endo-lysosomal compartment.

Methods

Cell isolation and culture

Murine SP-R210_L dominant negative (DN) RAW 264.7 macrophages (SP-R210_L(DN)) were generated and characterized as described previously by stable transfection of a pTriex-2 vector expressing the carboxy-terminal domain of SP-R210 [27, 31, 35]. Control cells were stably transfected with empty vector. SP-R210(KO) with depletion of both SP-R210 isoforms was achieved using CRISPR/Cas9 as describe previously [28]. The KG-1 human macrophage cell line was obtained from American Tissue Culture Collection (ATCC Cat # CCL-246). Human alveolar macrophages (hAMs) were isolated post-mortem by bronchoalveolar lavage from donated human lungs rejected for transplant as described previously [46, 47]. Lungs were procured through the Gift of Life Donors Program and all procedures were approved by Penn State College of Medicine’s Institutional Review Board. All cell lines were cultured in DMEM/10% heat-inactivated Fetal Bovine Serum (FBS).

Plasmid and siRNA transfection

IFITM2 and IFITM3 siRNA was transfected using Lipofectamine 3000 (ThermoFisher Cat # L3000015) according to manufacturer’s recommendations. Briefly, 1.5x10⁵ cells were seeded into 24-well plates and allowed to adhere for at least 6 hours to at most 12 hours. For plasmid transfection, 3 μl/well of Lipofectamine 3000 in 125 μl/well of Opti-MEM media (Gibco, Cat # 31985070) was mixed with 5 μg/well of plasmid and 2 μl/well of P3000 reagent in 125 μl/well of Opti-MEM, while for siRNA transfection, transfection 3 μl/well of Lipofectamine LTX in 125 μl/well of Opti-MEM media was mixed with 0.1 ng/well of siRNA in 125 μl/well of Opti-MEM. Mixture was incubated at room temperature for 20 minutes. Cells were washed once with PBS. After incubation, 250 μl of the mixture was added to the cells and incubated at 37°C for 24 hours. siRNA used are IFITM2 (ThermoFisher Cat # 4390771), IFITM3 (ThermoFisher Cat # AM16708), and negative control siRNA (ThermoFisher Cat # 4390843).

Virus Preparation

The mouse adapted influenza virus strain A/Puerto Rico/8/34 (PR8), recombinant PR8-GFP H1N1 strain [48, 49], A/Aichi/202/2009/H3N2 (Aichi), and A/Philippines/1982/(Phil82) [50] were propagated in the allantoic fluid of embryonated chicken eggs as described previously [51]. Viral titer was determined by fluorescent plaque assays using Madin-Darby Canine Kidney (MDCK) epithelial cells (ATCC Cat # CRL-2936). Cells were incubated at 37°C with serially diluted purified virus for 2 hours, at which point virus-containing media were replaced with virus-free media and cells were incubated for another 6 hours. Cells were then fixed and stained with an IAV nuclear protein (NP) antibody (Sigma Aldrich Cat # MAB8251, 1:100 in PBS) for 30 minutes at 4°C and subsequently labeled with Rhodamine conjugated anti-mouse IgG (Jackson ImmunoResearch, Cat # 115-026-062, 1:100 in PBS). Fluorescently labeled nuclei were counted using a Nikon Eclipse TE2000-U at 20x magnification.

Influenza infection

Macrophages were infected as described recently in detail [48]. Cells were seeded at a density of 150-200,000 cells per 24-well plate for 12 hours prior to infection unless otherwise indicated in figure legends. Influenza strains A/Puerto Rico/8/1934/H1N1 (PR8) [48], recombinant PR8-GFP H1N1 strain [48, 49], A/Aichi/202/2009/H3N2 (Aichi), and A/Philippines/1982/(Phil82) [50] were used in these studies. Cells were washed twice with phosphate buffered saline (PBS) and virus was added at the designated MOI in infection media consisting of 1:1 ratio Dulbecco’s modified Eagles medium (DMEM) w/o serum to PBS. Cells were then incubated with virus in infection media at 37°C for two hours, unless otherwise indicated in figure legends. Infection media were removed and replaced with DMEM/10% FBS. Infection was allowed to proceed at 37°C for the desired incubation time. Infection was assessed via viral protein production by Western blot, flow cytometry, and microscopy as described previously [48] and below.

Treatment of cells with TLR agonists

Control, SP-R210_L(DN), and SP-R210(KO) RAW 264.7 were plated at a density of 300,000 cells/well in 24-well plates and cultured for 24-hours in DMEM/10% fetal bovine serum (FBS) in a humidified incubator set at 5% CO₂. The media were then removed, cells washed twice with 1 mL PBS, and media replaced with 0.5 mL of 1:1 ratio DMEM w/o serum to PBS (infection medium above) in the presence or absence of 20 μg/mL Poly(I:C) (InvivoGen cat # tlr-pic-5) or 5 μg/mL Imiquimod (R837) (InvivoGen cat # tlrl-imqs). After 2-hours of incubation, media were collected for the 2-hour time point and then replaced with 0.5 mL DMEM/10% FBS and cultured for 2, 6, and 22-hours (for the 4, 8, and 24-hour time points, respectively). Media harvested at the end of each time point were stored frozen at −80°C until cytokine measurements by ELISA.

Flow cytometry

Control, SP-R210_L(DN), and SP-R210(KO) RAW 264.7, and human alveolar macrophages (hAMs) were detached using 0.25% Trypsin-EDTA (Corning Cat # 25-050-Cl). Non-adherent KG-1 cells were collected by centrifugation. Cells were washed by centrifugation in PBS containing 2% fetal bovine serum (FBS) and then fixed using 100 μL IC Fixation Buffer (eBioscience Cat # 00-8222-49) for 15 minutes. Fixed cells were washed and then blocked with Fc block (BD Biosciences Cat # 553142) in PBS at a concentration of 12.5 μg/mL and 2% FBS for 10 minutes at room temperature. For intracellular staining, after blocking, cells were permeabilized using 1x permeabilization buffer (eBioscience Cat # 00-8333-56). Permeabilized cells were stained with fluorescein isothiocyanate (FITC)-NP (EMD Millipore Cat # MAB8257F, 1:40) or Biotin-NP (EMD Millipore Cat # MAB8257B, 1:40) in 1x permeabilization buffer for 30 minutes at 4°C or pH-sensitive anti-hemagglutinin antibody (pHHA) antibody [52, 53]. Cells stained with Biotin-NP were washed once, then stained with Streptavidin PE-Cy5 (BioLegend Cat # 405205, 1:40) in 1x permeabilization buffer for 30 minutes at 4°C. For extracellular staining, blocked cells were incubated with FITC-conjugated MAA (Fischer Scientific Cat # 50-199-151; 1:40) or FITC-conjugated SNA (ThermoFisher Ca t# L32479, 1:40) in PBS with 2% FBS for 30 minutes at 4°C. Cells were washed once more and resuspended in Hanks Balanced Salt Solution (HBSS) with 2% FBS and 0.02% sodium azide until assessment via BD LSRII flow cytometer. A minimum of 30,000 events were collected and data were analyzed via FlowJo 9.8.8 using a previously described gating strategy [48].

Western blot analysis

Cells were either harvested using 0.25% Trypsin-EDTA (Corning Cat # 25-053-Cl) and centrifuged at 15,000xg for 5 minutes at 4°C (Eppendorf 5430R) or washed in plate and lysed in 60 μL SDS Lysis buffer (1% SDS, 50mM Tris-HCL pH 8.1, 10 mM EDTA pH 8.0) with 1x Protease/Phosphatase Inhibitor Cocktail (Cell Signaling Cat # 5872S) and placed at −20°C. Samples were thawed and sonicated at 50% amplitude for 10 seconds in 2 second ON/OFF intervals, centrifuged at 15,000xg for 5 minutes at 22°C, and supernatants mixed with 4x LDS sample buffer (Invitrogen Cat # NP007) to a final concentration of 1x.

Protein was separated on 4-12% Bis-Tris gels (Invitrogen, Cat # NP0321BOX, NP0349BOX) and transferred using the eBlot L1 Transfer system (GenScript). Blots were blocked in Tris-buffered saline containing 0.1% Tween 20 (TBST) and supplemented with 5% Bovine Serum Albumin (BSA). Blots were probed at 4°C overnight with the relevant primary antibody in TBST containing 1% BSA. Antibodies used are as follows: MYO18A (Proteintech Cat # 14611-1-AP, 1:1000) [27, 31], IFITM3 (ThermoFisher Cat # PA5-11274) and anti-β-actin (Sigma Aldrich Cat # A2228-100 μL, 1:20000). Blots were washed and then incubated with IRDye 680RD anti-Mouse IgG (LI-COR Cat # 926-68070, 1:15,000) and IRDye 800 CW anti-Rabbit IgG (LI-COR Cat # 926-32211, 1:10,000) in TBST, 1% BSA for 2 hours at room temperature. Blots were imaged using LI-COR Odyssey CLx and Image Studio 4.0. Band densitometry was acquired and images were adjusted using Image Studio Lite 5.2.5.

Immunofluorescence

PR8 infected cells were stained after the indicated incubation period, fixed using 3.7% paraformaldehyde for 25 minutes, and then permeabilized in 0.3% Triton X-100 in PBS for 15 minutes. Fixed and permeabilized cells were blocked with 10% donkey serum and 3% BSA in PBS. Blocked cells were probed using ras-related protein in brain 5 (Rab5) (Cell Signaling Technology Cat # 3547S, 1:200, Cell signaling Technology Cat#46449S, 1:400) or 7 (Rab7)(Cell Signaling Technology Cat#95746S, 1:400) with IFITM3 (Proteintech Cat # 11714-1-AP, 1:400), lysosomal associated membrane protein 1 (LAMP1) (Cell Signaling Taechnology Cat # E5N8Z, 1:500), or IAV NP (BioRad Cat # MCA400, 1:1000; EMD Millipore Cat # MAB8257, 1:500) overnight at 4°C, and labeled using AF488 (Jackson ImmunoResearch Cat#715-545-150, 1:500), FITC (ThermoFisher Cat # PI31541) or tetramethylrhodamine (TRITC)-conjugated anti-mouse IgG (Jackson ImmunoResearch Cat#115-025-062, 1:500;) and AF594 (Jackson ImmunoResearch Cat#711-585-152, 1:500) or TRITC-conjugated anti-rabbit IgG (Jackson ImmunoResearch Cat#111-025-144, 1:500 or ThermoFisher Cat # T2769) for 2 hours at room temperature. Slides were washed with PBS and mounted with Prolong Diamond Antifade with 4,6-diamidino-2-phenylindole (DAPI) (Invitrogen Cat # P36962). Images were acquired using either a LEICA AOBS SP8 confocal microscope (Leica microsystems, Heidelberg, Germany) and Imaris software or Nikon C2 confocal microscope (Nikon Instruments, Melville, NY). Images were captured using either IMARIS (Bitplane, Switzerland) or Nikon Elements 4.30.01 software. For 3D projections and surface rendering, optical projections were obtained every 0.33 μm over the entire cell volume.

Colocalization analysis

Colocalization analysis was performed on images captured using Nikon Elements 4.30.01 and imported into FIJI/ImageJ v2.3.0. Regions of Interest (ROI) were drawn around single cells. Specific ROI were analyzed for colocalization between AF488 channel and AF594 channels using Coloc2 for the entire captured stack. Colocalization was reported out via Pearson’s Correlation Coefficient (PCC) and Intensity Correlation Quotient (ICQ). PCC measures colocalization and signal intensity at a given pixel where the fluorescence intensity of one marker is likely associated with the intensity of the second [54, 55]. ICQ measures pixel ratios where fluorescence of both channels is both above or below the average to account for variability of Rab and IFITM3 staining intensity throughout the cell [54, 55].

Cytokine analysis

Cytokines were assessed from clarified supernatant using R&D Quantikine enzyme linked immunosorbent assay (ELISA) kits for tumor necrosis factor alpha(TNFα) (R&D Systems, Cat#MTA00B), DuoSet ELISA kits for interferon beta (IFNβ) (R&D Systems, Cat#DY8234-05). Cell media were spun at 15,000 x g for 10 minutes to remove debris. Clarified supernatants were aliquoted and frozen before utilizing in ELISA experiments. ELISAs were performed following manufacture protocol without diluting supernatant samples.

Virus labeling

PR8 virus was labeled according to the manufacturers’ protocol for AF488 (ThermoFisher, Cat#A20000) and pHrodo Red (ThermoFisher Cat # P36011).

Virus binding to isolated cell membranes

Binding of IAV PR8 to isolated cell membranes from control, SP-R210_L(DN), and SP-R210(KO) cells was assessed in ELISA assays as described in detail recently [48].

Endocytic assay

For experiments utilizing pH-sensitive Rodo (pHRodo)-labeled probes, cells were washed and incubated with 125 μL cargo-containing 1:1 PBS/DMEM media for 15 minutes at room temperature (for Dextran and Transferrin) or on ice (for PR8). pHrodo Green Dextran MW10000 (ThermoFisher, Cat#P35368) and pHrodo Red Transferrin (ThermoFisher, Cat#P35376) were used at 30 μg/mL, while pHrodo PR8 was incubated at 15 MOI using a cell density of 2x10⁵. After incubation, media was removed, cells were washed, and 200 μL of 1:1 DMEM/PBS was added. Cells were then incubated at 37°C for the specified time (10, 20, 30, or 45 minutes). Cells were washed and 300 μL of PBS with 2%FBS and 0.02% sodium azide was added. Cells were removed via cell scraper and fluorescence data was collected using LSRII or Fortessa flow cytometer and analyzed using FlowJo 9.8.8. For experiments utilizing FITC-Dextran (ThermoFisher, Cat # D1820), AF488-Transferrin (ThermoFisher, Cat # T13342) or AF488 labeled IAV PR8, cells were washed and incubated with 125 μL 1:1 PBS/DMEM media containing fluorescent probes. Cells were placed at 37°C for the specified time (5, 10, 15, 30 or 45 minutes), and harvested via cell scraper as described above. Fluorescence data was collected using LSRII or Fortessa flow cytometer and analyzed using FlowJo 9.8.8.

Electron microscopy

Cells were grown on Nunc permanox slide chambers (Thermofisher), washed in PBS, infected with 15 MOI of IAV PR8 for 1-hour as described above, and fixed for one hour in a solution of 0.5% glutaraldehyde and 4% paraformaldehyde buffered with 0.1 M sodium cacodylate, pH 7.3. Following fixation, cells were washed in 0.1 M sodium cacodylate buffer and post fixed overnight in buffered 1% osmium/1.5% potassium ferrocyanide. After fixation, cells were rinsed with buffer, dehydrated in a graded series of ethanol, and embedded in EMbed 812 (Electron Microscopy Sciences). A diamond knife mounted in a Porter-Blum MT-2B ultramicrotome was used to cut 70-90 nm thin sections. The sections were mounted on 200 mesh copper grids and stained with 2% aqueous uranyl acetate and Reynold's lead citrate and examined in a JEOL JEM 1400 transmission electron microscope.

GST-PAK1-PBD pull-down assays

Pull down assays were performed using the RAC1 Activation Assay Biochem Kit from Cytoskeleton, Inc. (Cat # BK035) per manufacturer’s directions. Briefly, cells were cultured to 80-90% confluency in 100 mm dishes and all subsequent steps unless specified were carried out on ice. Cells were wash with ice cold PBS, and lysed directly in dishes with 250 μL ice-cold Cell Lysis Buffer (Cytoskeleton, Inc. Cat # CLB01) supplemented with Protease Inhibitor Cocktail (Cytoskeleton, Inc., Cat # PIC02). Lysates were immediately centrifuged at 10,000 x g for 1 minute at 4°C and clarified supernatants snap frozen in liquid nitrogen and stored at −80°C. For the assay, 28 μL of 10x Loading Buffer was added to cell extracts followed by 2.5 μL 100X guanine diphosphate (GDP) or 2.5 μL 100x guanosine 5-O-triphosphate (GTPyS). The reaction was carried out for 15 minutes at room temperature and then stopped by adding 31 μL ice-cold 10x STOP buffer on ice.

Mass spectrometry

GDP or GTPγS treated cell extracts were incubated with GST-tagged p21 binding domain PAK1-PBD beads (Cytoskeleton, Inc., Cat # PAK02) for 1 hour at 4°C. Bound protein was collected by centrifugation and washing of the beads which were then digested using a trypsin/LysC enzyme mix. Released peptides were cleaned using CDS Empore SDB-RPS tips, and peptides run on a Bruker TIMS TOF Flex in-line with Pepsep PSC15-75-1.9 UHP nC Reprosil column using a NanoElute LC. Peptides were analyzed and matched to protein IDs using a DDA PASEF short and a Shotgun PASEF analysis method at the mass spectrometry core of Penn State College of Medicine.

Statistics

GraphPad Prism 9.5.1 software (San Diego, CA) was utilized for graphs and statistical comparisons by 2-way ANOVA with unpaired and paired comparisons via t-test corrected by the Holm-Sidak method. P-values of <0.05 were considered significant.

Results

SP-R210 isoforms modulate infectivity and affect the anti-viral response in macrophages

Our recent studies showed that infection of macrophages with dominant negative (DN) depletion of SP-R210_L (SP-R210_L(DN)) results in robust activation of IRF7 but not IRF3, indicating preferential activation of endosomal Toll-like receptors [28]. To pursue this finding further, we monitored expression of nuclear protein (NP) by flow cytometry as an indirect measure of viral genome replication to assess the impact of SP-R210_L depletion on IAV infection in macrophages with different levels of SP-R210 isoforms (Figure 1A). Figure 1B shows bimodal rightward shift of NP fluorescence over time in SP-R210_L-replete control cells incubated with either H3N2 or H1N1 strains of influenza cells compared to the mostly unimodal rightward distribution of NP fluorescence in (SP-R210_L(DN) cells. Western blot analysis indicates that expression of both viral NP and non-structural protein 1 were attenuated (Supplemental Figure S1). Furthermore, infection with a recombinant strain carrying a green fluorescent protein (GFP)-tagged NS1 gene segment PR8 strain (PR8-GFP), which is expressed after endosomal release of the viral genome [56, 57], was also attenuated. NS1 is expressed after infection but not incorporated into viral progeny and thus not part of incoming virus. Hence, inhibition of both NP and NS1-GFP indicated a block in either a post-nuclear or an endosomal arm of the viral life-cycle in SP-R210_L(DN) cells. Optical rendering of hemagglutinin staining of control and SP-R210_L(DN) cells infected with the PR8-GFP IAV strain by confocal microscopy showed vesicular distribution of IAV 1-hour after infection in both cell lines. Expression and membrane localization of hemagglutinin (HA) 5-hours after infection, however, was attenuated in SP-R210_L(DN) cells compared to controls (Supplemental Figure S2), supporting an endosomal block. Human monocytic cell lines exhibit low to no expression of SP-R210_L [29, 58] (Figure 1A). NP expression was also attenuated in the human monocytic KG-1 cell line, with SP-R210_S as the most abundant isoform (Figure 1A, C), whereas human AMs (hAMs) with SP-R210_L as the more abundant isoform were readily infected supported synthesis of NP (Figure 1A, C), indicating a conserved mechanism of restrictive infection in SP-R210_L-deficient macrophages.

Figure 1. SP-R210_L mediates viral genome expression and suppression of antiviral response in macrophages.

Cells were infected with 2 MOI of IAV strains Phil82 (B,C), PR8 NS1-GFP (B), PR8 (B,C, and D-G), and Aichi (C) for 10-hours and analyzed by flow cytometry utilizing an influenza nuclear protein (NP) antibody or an NS1-GFP reporter strain to monitor uptake and expression of NP or NS1 over time as described in Methods section. (A) Western blotting comparison shows differential expression of SP-R210_L vs. SP-R210_S in human AMs, KG-1 cells and similar expression in RAW264.7 control cell extracts consistent with previous studies [27-29, 37, 48, 115]. (B, C) Expression of the IAV genome is suppressed in SP-R210_L(DN) and KG-1 cells compared to control and human alveolar macrophages, respectively. In (B), NP was monitored at 6 (faint grey histogram), 12 (dashed line histogram), and 24 (black line histogram)-hrs after infection using either an NP antibody (FITC-NP) or the NS1-GFP reporter. In (C) flow cytometry was performed 12-hrs after infection. (D) Disruption of SP-R210_L enhances expression of IFNβ in response to PR-8 infection. IFNβ was measured by ELISA in media harvested over time. Data shown are means±SD from n=2 independent experiments performed in duplicate. **p<0.01 compared to t=0, #p<0.01 compared to compared to control cells at each time point. (E) Polyclonal SP-R210 antibodies targeting the unique carboxy-terminal domain of SP-R210 suppress IAV infection in a concentration-dependent manner. Control and SP-R210_L(DN) cells treated with increasing concentrations of anti-SP-R210 antibody (5, 10, 20, 40, 100, and 200 μg/mL) for 5-hours, upon which cells were washed and 2 MOI IAV was added to the cells in 1:1 DMEM and the appropriate concentration of anti-SP-R210 antibodies or IgG control. Infection was allowed to proceed for a total of 12-hours and then cells were harvested and stained against IAV NP as an indirect measure for IAV uptake and replication. SP-R210 antibody treatment of RAW264.7 cells resulted in a dose-dependent inhibition of IAV infection in macrophages, suggesting that SP-R210 may be involved in the macrophage response to IAV. Data shown are means±SD from N=2 independent experiments performed in duplicate. *p<0.05 compared to 200 μg/mL IgG pre-immune control. (F) SP-R210 antibodies induce and attenuate secretion of IFNβ in response to IAV infection in control and SP-R210_L(DN) cells, respectively. Data shown are means±SD from N=2 independent experiments performed in duplicate. **p<0.01 and *p<0.05 compared to pre-immune IgG, #p<0.01 compared to compared to control cells at respective time points. (G) SP-R210 antibodies enhance secretion of IFNβ in response to IAV infection in in human AMs. Data shown are means±SD from N=2.

To address the biological significance of the above findings, we measured IFNβ secretion in IAV infected cells. Figure 1D shows that IAV did not elicit type I IFNβ secretion in control cells compared to robust induction in SP-R210_L(DN) cells. Pre-incubation with a polyclonal antibody recognizing the carboxy-terminal domain of both isoforms inhibited infection in both control and SP-R210_L(DN) cells [29] (Figure 1E). Preincubation of control cells with SP-R210 antibodies, however, stimulated IFNβ secretion (Figure 1F), consistent with the partial inhibition of infection (Figure 1E). SP-R210 antibodies also enhanced IFNβ secretion in human alveolar macrophages (hAMs) (Figure 1G). On the other hand, pre-incubation of SP-R210_L(DN) cells with SP-R210 antibodies suppressed both IFNβ secretion and percentage of NP+ cells, indicating activation of additional infection control mechanism(s) in SP-R210_L(DN) cells (Figure 1E, F).

SP-R210 isoforms modulate differential activation of endosomal toll-like receptor (TLR) agonists and response to IAV infection

Endosomal TLRs, TLR3 and TLR7, mediate innate sensing of IAV infection by recognizing viral nucleic acid released following uptake of dead infected cells or during transit of the virus through the endolysosomal compartment in macrophages, respectively [59-61]. To determine whether SP-R210 isoforms modulate endosomal TLR activation, we measured secretion of IFNβ and TNFα after treatment of control, SP-R210_L(DN), and SP-R210(KO) cells with either poly IC (pIC), a TLR3 agonist, or imiquimod (IM), a TLR7 agonist (Figure 2). Incubations with pIC and IM were carried out using the same protocol design as IAV infection withdrawing the stimulus after 2-hour incubation in serum free medium. Both pIC and IM treatment resulted in robust stimulation of IFNβ secretion in SP-R210_L(DN) cells compared to untreated cells (Figure 2A and B). In contrast, both control and SPR210(KO) cells were unresponsive compared to baseline. Both pIC and IM, however, induced secretion of TNFα in both SP-R210_L(DN) and SP-R210(KO) cells, whereas IM induced TNFα only in control cells (Figure 2C and D). TNFα levels in IM-treated control cells plateaued 40-50% lower than in SP-R210_L(DN) and SP-R210(KO) cells (Figure 2D). Basal cytokine levels in untreated cells were similar in all three cell lines (Figure 2A-D). IFNβ secretion declined after 8-hours of agonist addition (Figure 2A and B), consistent with stimulus withdrawal at the 2-hour time point. TNFα secretion plateaued after 4-hours in pIC-treated SP-R210(KO) cells (Figure 2C) and in all IM-treated cells (Figure 2D). This biphasic TNFα kinetic was delayed by 2-hours in pIC-treated SP-R210_L(DN) cells (Figure 2C). These findings suggest that SP-R210 isoforms have differential roles in inhibition and activation of endosomal TLR signaling.

Figure 2. Differential cytokine responses in control and SP-R210-deficient macrophages.

Control, SP-R210_L(DN), and SP-R210(KO) cells cultured overnight in FBS replete DMEM were washed and placed in 0.5 mL of 1:1 ratio of serum-free DMEM to PBS in the presence or absence of 20 μg/mL Poly(I:C) or 5 μg/mL Imiquimod. Media were collected for the 2-hour time point and then replaced with 0.5 mL DMEM/10% FBS and media harvested at 2, 6, and 22-hours. The concentration of IFNβ (A,B) and TNFα (C,D) was quantitated by ELISA. N=3. All data are shown as means±SD. (A) **p<003 and ***p<0.0005 for SP-R210_L(DN) vs SP-R210_L(DN)+pIC at t=4- and 8-hours, respectively. (B) **p<006 and ***p<0.0001 for SP-R210_L(DN) vs SP-R210_L(DN)+IM at t=4- and 8-hours, respectively. (C) *p<0.015; **p<002, ***p<0.0002 for pIC treated SP-R210_L(DN) and SP-R210(KO) cells at t=4-8-hours. (C) *p<0.015; **p<002, ***p<0.0002 for pIC treated SP-R210_L(DN) and SP-R210(KO) vs untreated cells at t=4-8-hours. (D) *p<0.05; **p<0.006 for IM treated SP-R210_L(DN) and SP-R210(KO) vs untreated cells at t=4-8-hours.

Differential cytokine response kinetics in SP-R210-deficient cells also occurred after IAV infection (Figure 3 and 1D). The rate of IFNβ increased 8- and 10-hours after infection of SP-R210_L(DN) and SP-R210(KO) cells, whereas IFNβ secretion increased slowly near basal levels in control cells throughout the time course (Figure 3A and 3A inset). The rate of TNFα secretion increased 8-hours after infection in all cell lines, although significantly lower and near basal secretion in control cells (Figure 3B and 3B inset). The kinetics of IFNβ (Figure 1D and 3A) and TNFα (Figure 3B) secretion, however, showed a plateau phase after 10-hours in SP-R210_L(DN) cells, whereas both cytokines continued to increase in SP-R210(KO) cells (Figure 3A and 3B). This unmitigated cytokine secretion reflects an initial delay and then accelerated expression of NP and frequency of infected SPR210(KO) cells compared to controls (Figure 3A and B). In contrast, the infection rate in SP-R210_L(DN) cells increased slowly after 8 hrs (Figure 3A and B). These findings indicate that SP-R210 isoforms couple intracellular fate and recognition of infection in macrophages.

Figure 3. Differential cytokine kinetics in control and SP-R210-deficient cells.

Control, SP-R210_L(DN), and SP-R210(KO) cells cultured overnight in FBS replete DMEM were washed, placed in infection medium for 2-hours with no infection (NI) or with 4 MOI IAV PR8. Media were collected for the 2-hour time point and then replaced with 0.5 mL DMEM/10% FBS and media harvested at 2, 8, 10, and 24-hours. The concentration of IFNβ (A) and TNFα (B) was quantitated by ELISA. N=2 experiments performed in duplicate. All data are shown as means±SD. In (A) ⁺⁺⁺⁺, ****p<0.001 SP-R210_L(DN)-IAV vs SP-R210(KO)-IAV and Control cells-IAV and no infection controls at t=10-hours, ^##p<0.01 SP-R210_L(DN)-IAV vs SP-R210(KO)-IAV at t=24-hours, and ⁺⁺⁺⁺, ****p<0.001 ⁺⁺⁺⁺, ****p<0.001 SP-R210(KO)-IAV compared to Control cells-IAV and no infection controls at t=24-hours.

SP-R210 isoforms modulate endosomal trafficking of IAV

We then performed experiments to evaluate differences in attachment vs. uptake of IAV (Figure 4). Attachment of IAV to control and SP-R210_L(DN) cells at 4°C as monitored by flow cytometry was similar (Figure 4A). Binding of IAV to isolated cell membranes from each of the three cell lines was similar (Figure 2D). Furthermore, depletion of SP-R210 isoforms did not affect α2,3-sialic acid receptors of IAV PR8 hemagglutinin [62] (Figure 2E). Expression of α2,6-sialic acid receptors, however, was significantly higher in SP-R210_L(DN) cells (Figure 2E). This difference did not affect uptake of the α2,6 sialic acid binding H3N2 Phil82 strain (Figure 1A). The intracellular fate of IAV, however, was different. In control cells, the NP antibody stained cytosolic vesicles at early time points and then nuclei with intense staining in 50% of control cells by 4-hours after infection compared to no or sparse nuclear staining in SP-R210_L(DN) cells (Figure 4C). NP staining over time showed spatial co-occurrence and colocalization with markers of early (Rab5) and late endosomes (Rab7), and endolysosomes (LAMP1) over the first 60-minutes of infection and then in nuclei of control (Figure 4F) and SP-R210(KO) cells (Supplemental Figure S3a) at 4-hours, indicating endosomal escape and then replication of the viral genome in the nucleus. The NP antibody did not stain uninfected cells (Figure 4F, Supplemental Figure S3aA, S3aF, and S3bA) and staining with secondary antibody only was also negative (Supplemental Figure S4). In contrast to control and SP-R210(KO) cells, NP staining in SP-R210_L(DN) cells distributed from the cell periphery (Figure 3G, Supplemental Figure S3b) localizing initially with Rab5+ vesicles and then with Rab7+ and LAMP1+ vesicles over the time course (Figure 4G, white arrowheads, Supplemental Figure S3a and S3bC-D blue arrows) but not in the nucleus. These cells indicate that SP-R210_L(DN) cells restrict internalized IAV in the endo-lysosomal compartment.

Figure 4. Disruption of SP-R210_L but not both isoforms attenuates endosomal escape and expression of the viral genome.

(A and B) Selective depletion of SP-R210_L leads to sustained attenuation of IAV viral genome expression, whereas depletion of both isoforms results in temporary suppression of IAV genome expression. *p-value<0.05, **p-value<0.005, ***p-value<0.0005 and ****p-value<0.00005 for control vs. SP-R210_L(DN) cell comparisons; ^##p-value<0.005, ^####p-value<0.00005 for SP-R210(KO) vs. SP-R210_L(DN) cell comparisons. +p-value<0.05 for control vs SP-R210(KO) cell comparison. N=4 independent experiments performed in duplicate. Cells were infected with 4 MOI IAV PR8. Cells were stained with a biotinylated anti-NP antibody and NP stained cells visualized using a PE-Cy5 secondary antibody. (C) Lack of SP-R210_L limits expression and nuclear accumulation of NP but not uptake of the virus in SP-R210_L(DN) cells. Control and SP-R210_L(DN) cells were cultured on 12 mm glass coverslips placed in 12-well dishes overnight at a density of 150,000 cells per well. Cells were washed in DPBS and infected with 10 MOI IAV PR8 at 4°C for 1.5-hours and then either harvested for flow cytometry analysis of bound virus on non-permeabilized cells, or returned at 37°C and infection allowed to proceed for 1.5- or 4-hours. Coverslips were the washed, fixed permeabilized in blocking buffer, and stained with DAPI to visualize nuclei or TRITC-conjugated anti NP antibodies. Stained cells were washed, and coverslips mounted on slides, and visualized on a Nikon epi-fluorescent microscope. Cells were infected with 10 MOI IAV PR8. (D) Lack of SP-R210 isoforms does not affect binding of IAV PR8 to isolated cell membranes. Binding of IAV PR8 to isolated cell membranes from control, SP-R210_L(DN), SP-R210(KO) cells was by ELISA using anti-hemagglutinin antibodies as described recently [48]. N=2 independent experiments performed in duplicate. (E) Lack of SP-R210 isoforms does not deplete α2,3-sialic receptors for IAV PR8 hemagglutinin, whereas SP-R210_L depletion results in increased α2,6-sialic acid residues on SP-R210_L(DN) cells. MAA lectin and SNA lectin were used to stain control, SP-R210_L(DN), and SP-R210(KO) to determine α2,3- and α2,6- on the cell surface by flow cytometry, respectively. **p-value<0.005 comparing WT to SP-R210_L(DN) cells; ^##p-value<0.005 comparing SP-R210(KO) to SP-R210_L(DN) cells. N=3 independent experiments. (F,G) Cells were cultured on coverslips and infected with PR8 IAV for 15-, 30-, 60-, or 240-minutes with 10 MOI IAV PR8 as described above. Cells were then fixed and co-stained with NP and Rab5 antibodies (upper row), Rab7 (middle rows), or LAMP1 antibodies (bottom row), and counterstained with DAPI to visualized nuclei. Internalized IAV transitioned through early (Rab5+) and late endosomes (Rab7+) endosomes (white arrowheads) and then nuclear entry in control cells (F) compared to accumulation in perinuclear Rab7+ vesicles (white arrowheads) and spatial overlap with LAMP1+ vesicles in SP-R210_L(DN) cells over time (G). Anti-mouse FITC-conjugated and anti-rabbit TRITC-conjugated secondary antibodies were used to visualize staining with NP and endolysosomal marker antibodies Images were acquired using a LEICA confocal microscope using IMARIS software.

To further assess the impact of SP-R210 isoform depletion in endocytic trafficking, we monitored endocytic uptake using Dextran or transferrin fluorescently-labeled with either FITC or the pH-sensitive dye pHrodo by flow cytometry (Figure 5). Dextran-10,000 is internalized by both clathrin-mediated pinocytosis and clathrin-independent macropinocytosis, whereas transferrin is internalized transiently through early macropinosomes before transport to recycling endosomes and through clathrin-coated pit endocytosis and predominantly recycled instead of sorting to lysosomes for degradation via the transferrin receptor [12-15, 63]. Uptake of FITC-labeled Dextran by SP-R210_L(DN) and SPR210(KO) cells exhibited a transient peak at 5-minutes after addition of the fluorophore followed by a plateau compared to saturation kinetics in control cells (Figure 5A). The kinetics of acid-induced fluorescence of pHrodo-labeled dextran, however, indicates expansion of acidified compartments in SP-R210_L(DN) but not in control or SP-R210(KO) cells (Figure 5B). On the other hand, SP-R210(KO) cells accumulated more FITC-transferrin (Figure 5C), suggesting that depletion of SP-R210 isoforms alters trafficking of the transferrin receptor. The kinetics of pHrodo transferrin, however, were similar in all cell lines (Figure 5D). Tracking of fluorescent IAV uptake showed similar uptake kinetics of PR8 labeled with Alexa Fluor 488 (AF488) (Figure 5E). The acid-induced fluorescence of PR8 labeled with the pH-sensitive dye pHrodo, however, increased earlier in control cells compared to both SP-R210_L(DN) and SP-R210(KO) cells between 20 and 40-minutes of infection (Figure 5F), suggesting faster endolysosomal transit. We then utilized a monoclonal HA antibody recognizing a pH-sensitive epitope that is exposed upon conformational transition of the HA trimer to the fusion competent form at the acidic pH of endosomes [52]. The antibody also recognizes the HA monomer and can be used to monitor newly synthesized HA during transit in the endoplasmic reticulum [64]. This experiment revealed transient exposure of the pHHA epitope with a peak at 30 min after infection (Figure 5G, I) followed by decline till 1-hour after infection and then robust expression of new HA after 90-minutes in control cells (Figure 5G). In contrast, the pHHA epitope exhibited two peaks of enhanced intensity at 30-and 90-minutes after infection, followed by slow decline over the next 140-minutes in SP-R210_L(DN) cells (Figure 5G). This second peak was absent in SP-R210(KO) cells. Expression of new HA in SP-R210(KO) cells, however, increased slowly compared to control cells. A monoclonal antibody recognizing a non-pH-dependent epitope on the HA trimer, however, showed similar uptake of the virus with an earlier 15-minute peak (Figure 5H,I) than the 30-minute peak with the pHHA stain (Figure 5G) in both control and SP-R210_L(DN) cells (Figure 5H), consistent with similar attachment and uptake. Synthesis of new HA trimer was seen 150-minutes after infection and abundant on the cell surface by 5-hours in control but not in SP-R210_L(DN) cells (Figure 5H, Supplemental Figure S2). These results indicate that SP-R210_L predominates over the SP-R210_S-mediated endolysosomal sorting of IAV in macrophages.

Figure 5. Depletion of SP-R210_L or both isoforms alters vesicular trafficking of endosomal markers and IAV.

(A-B) Increased uptake of Dextran in acidified vesicles in SP-R210_L(DN) cells compared to control and SP-R210(KO) cells. Cells were incubated with 125 μL of 50 μg/mL FITC- or 30 μg/mL pHRodo-labeled Dextran and cells harvested at indicated time points in this and following experiments were analyzed by flow cytometry. *p-value<0.05, and***p-value<0.0005 for WT vs. SP-R210_L(DN) cell comparisons; (C-D) Depletion of both SP-R210 isoforms enhances endocytosis (C) but not transport of transferrin (D) to acidified endolysosomes. Cells were incubated with 125 μL of 50 μg/mL FITC-Transferrin or 125 μL of 30 μg/mL pHRodo-Transferrin in 1:1 DMEM and uptake of transferrin assessed by flow cytometry. ^##p-value<0.005 between SP-R210_L(DN) and SP-R210(KO) cells; ⁺⁺p-value<0.005, and ⁺⁺⁺p-value<0.0005 for WT vs. SP-R210(KO) cell comparisons. (E) SP-R210-deficient cells exhibit similar uptake rate of IAV PR8. Cells were incubated with 15 MOI of AlexaFluor488-PR8 at 37°C in with 1:1 DMEM/PBS and uptake of FITC-PR8 assessed by flow cytometry over time. N=3 independent experiments performed in duplicate. (F) Lack of SP-R210_L delays transport of IAV to acidified vesicles. Cells were incubated with 15 MOI of IAV PR8 at room temperature for 15-minutes, washed and incubated with 1:1 DMEM/PBS and at 37°C for the indicated times. ⁺p-value<0.05 and⁺⁺p-value<0.005 for WT vs.SP-R210(KO) cell comparisons. **p-value<0.005 between WT and SP-R210_L(DN) cells. N=4 independent experiments performed in duplicate (G-H) Depletion of SP-R210_L but not SP-R210_S leads to retention of IAV in acidified vesicles, whereas depletion of both isoforms appears to delay endosomal exit and replication of the viral genome at early stage of infection. Control, SP-R210_L(DN), and SP-R210(KO) cells were infected with 10 MOI IAV PR8 for 15-, 35-, 60-, 90- and 240-minutes. Cells were harvested, fixed, permeabilized, and stained with the pH-sensitive anti-hemagglutinin (HA) pHHA antibody clone (G),pH-insensitive HA antibody recognizing the HA trimer (H) or with secondary antibody only to subtract background signal . ^#p-value<0.05 for SP-R210_L(DN) and SP-R210(KO) cell comparisons; ⁺⁺⁺⁺p-value<0.00005 between WT and SP-R210(KO) cells and ****p-value<0.00005 for WT and SP-R210_L(DN) cell comparisons at each time point. N=8 independent experiments in experiments shown in G and n=2 and H, respectively. Data shown are means±SD. (I) Experimental design for experiments shown in G and H

SP-R210 isoforms modulate macropinocytic entry and activation of IAV infected macrophages

Electron microscopic visualization of IAV revealed vesicles containing IAV in the process of fusion in control but not in SP-R210_L(DN) cells (Figure 6A). Internalization of several IAVs in macropinosomes were found in both control (Figure 6B-C) and SP-R210_L(DN) cells (Figure 6D-E). Vesicles containing partially degraded viral particles were found in SP-R210_L(DN) cells (Figure 6F). Structures indicating formation of virus ribonucleoprotein (vRNP) coating vesicles [65] or budding and transiting through the ER were seen in control and SP-R210(KO) cells (Supplemental Figure S5A-D and E-G), indicating that viral genome replication and transcription are not impaired in these cells. Whether these cells produce infectious virus requires future investigation. IAV excludes uptake of the macropinocytic marker TMR-Dextran-70,000 consistent with multiple virus particles taken up in the same vesicle (Supplemental Figure S6).

Figure 6. Macropinocytosis and macrophage activation in response to IAV infection are coupled in SP-R210_L(DN) but not control cells.

(A-F) Electron microscopic visualization revealed internalization of IAV in macropinocytic vesicles 1-hour after infection in control and SP-R210_L(DN) cells and in the process of engulfment by membrane ruffles (white arrows). IAV in the process of fusion in endosomes (A, open arrow) was seen in control cells and degradation in SP-R210_L(DN) cells (F, open arrow). En: endosome; Mp: macropinosome; MR: membrane ruffles; MPc: macropinosome cup. (G-O) Incubation with macropinocytic inhibitors of NHE1 (G; EIPA), PAK1 (H: IPA3.0), and RAC1 (I:EHT1864) attenuated inhibited IAV infection in both control and SP-R210_L(DN) cells in a concentration-dependent manner. The inactive PAK-1 analog PIR3.5 (J) and GEF Trio NSC23766 (K) had no effect. Inhibition of macropinocytosis by EIPA and EHT1864 blocked secretion of IFNβ in SP-R210_L(DN) cells (L, N) and TNFα in both cell lines (N,O). Inhibitors were added 30-minutes prior to infection with IAV PR8 and maintained for 2-hours during infection. Cells were then washed and infection allowed to proceed for 12-hours in DMEM/10%FBS. Media were harvested at the end of the incubation period for cytokine measurements by ELISA and cells harvested for flow cytometry analysis of infection. N=3 independent experiments in duplicate. *p<0.05, **p<0.005 for % NP+ cells and cytokine comparison for each cell line. ^$$p<0.005 for cytokine comparisons between cell lines at t=0. ^#p<0.05 and ^##p<0.005 for TNFα comparisons to t=0 in control cells.

The use of inhibitors supports macropinocytic uptake of IAV in Raw macrophages. At the highest concentration tested, the macropinocytic inhibitor EIPA suppressed IAV infection by 60% and >95% in control and SP-R210_L(DN) cells (Figure 6G), respectively. EIPA blocks macropinocytosis by inhibiting submembrane pH and activation of Rho-family (Rho) GTPases ras-related C3 botulinum toxin substrates 1 and 2 (RAC1/2) and cell division cycle 42 (CDC42) [66]. These small Rho-family GTPases mediate spatiotemporal actin-driven macropinosome formation and closure in macrophages [67, 68]. IPA3.0 an inhibitor of p21 activated kinase 1 (PAK1) [69] and a downstream effector of RAC1, also blocked infection in control cells, although less effectively than EIPA in SP-R210_L(DN) cells (Figure 6H). The direct RAC1 inhibitor EHT1864 [70], which displaces the guanine nucleotide from RAC1, however, blocked IAV infection in both control and SP-R210_L(DN) cells in a concentration-dependent manner (Figure 6I) to similar degree as EIPA (Figure 6G). PIR3.5, an inactive analog of IPA3.0, had no effect (Figure 6J). The RAC1-GEF Trio interaction inhibitor NSC23766 [71] also had no effect (Figure 6K). Stimulation of INFβ (Figure 6L and N) and TNFα (Figure 6M and O) in SP-R210_L(DN) cells was sensitive to low concentrations of EIPA, IPA3.0, and EHT1864 compared to gradual decline in macropinocytic uptake (Figure 6G, H, and I), indicating that inhibition of RAC1 alone is sufficient to abolish macropinosomal signaling in SP-R210_L(DN) cells.

RhoGTPases cycle between inactive GDP and active GTP-bound forms that depend on guanine exchange factors (GEFs) and GTPase activating proteins (GAPs). GEFs exchange GDP for GTP to enable binding of RhoGTPases to effector proteins, whereas GAPs augment RAC1 GTPase catalytic activity and return to the inactive GDP-bound form. These events occur at the plasma membrane where RhoGTPases are attached through a carboxy-terminal geranyl lipid moiety [67, 72, 73]. To this end, pull-down experiments in whole cell extracts using the GST-tagged p21 binding domain (PBD) of PAK1 showed that non-hydrolysable GTPγS induced co-precipitation of RAC proteins with SP-R210 isoforms in control and SP-R210_L(DN) cells (Supplemental Figure S7A-C). Mass spectrometry data (Supplemental Figure S7D-F) show that GTPγS enhanced association with RAC1, 2, and CDC42 in all cell lines. The mass spectrometry analysis also indicated differential pull-down with the guanine exchange factors rho guanine exchange factor 1 (ARHGEF1) [74], and vav guanine exchange factor 3 (VAV3) [75, 76], and the viral restriction factor interferon-inducible transmembrane protein 3 (IFITM3) [9, 77] (Supplemental Figure S7G-I). VAV1 and 3 are GEFs for Rac1/2 but not CDC42 [75, 76, 78, 79]. Western blot analysis (Figure 7A,B) of isolated cell membranes, however, showed that selective depletion of SP-R210_L resulted in >4-fold increase in relative expression of SP-R210_S in SP-R210_L(DN) cells, and depletion of both isoforms SPR210(KO) cells resulted in differential enrichment of RAC1 (Figure 7C) but not RAC2 (Figure 7D,E) in both SP-R210_L(DN) and SPR210(KO). CDC42 was enriched in SPR210(KO) cell membranes (Figure 7D,F). ARHGEF1, however, was selectively enriched in SP-R210_L(DN) cells and partially depleted in SP-R210(KO) cells (Figure 7D,G), whereas the membrane association of both VAV1 and VAV3 were reduced in SP-R210(KO) cells (Figure 7H-I). Together these results suggest that SP-R210_S recruits different effectors of RAC1/2 activation and endosomal function during IAV internalization when relative expression of SP-R210_L is decreased.

Figure 7. Differential RhoGTPases and guanine nucleotide exchange factors in SP-R210-deficient macrophages.

(A-C) Western blotting (A) and densitometric analysis of SP-R210 isoforms (B) and RAC1 (C) on isolated cell membranes from control, SP-R210_L(DN), and SP-R210(KO) macrophages. *p<0.01, ****p<0.0001. (D-I) Western blotting (D) and densitometric analysis of RAC2 (E), CDC42 (F), ARHGEF1 (G), VAV1 (H), and VAV3 (I). *p<0.01, *p<0.001. ***p<0.0001. ****p<0.00001, n=3 experiments performed in triplicate. All data are shown as means±SD.

IFITM3 contributes to endolysosomal restriction of IAV in SP-R210_L(DN) macrophages

Given the above findings implicating IFITM3 in SP-R210_L(DN) cells, we investigated whether IFITM3 restricts infection in these cells. IFITM3 limits endo-lysosomal viral escape through a variety of mechanisms [19, 20, 22-24, 80]. Given the above findings implicating IFITM3 in SP-R210_L(DN) cells, we investigated whether IFITM3 restricts infection in these cells. Western blot analysis showed increased basal expression of IFITM3 in uninfected SP-R210_L(DN) cells (Figure 8A). IAV infection enhanced expression of IFITM3 after 6-hours in SP-R210_L(DN) cells, whereas it had no effect on expression of IFITM3 in control cells (Figure 8A), consistent with differences in induction of IFNβ shown above. mRNA sequencing data also showed increased Ifitm3 expression in SP-R210_L(DN) cells [28]. Knockdown of IFITM3 with siRNA (Figure 8B) resulted in significant increase in percentage of NP expressing SP-R210_L(DN) cells (Figure 8C), whereas IFITM2 had no effect (Figure 8D).

Figure 8. IFITM3 is increased and contributes to endosomal restriction of the IAV in SP-R210_L(DN) cells.

(A) Basal and IAV-induced expression of IFITM3 in SP-R210_L(DN) cells. Control and SP-R210_L(DN) cells were infected with PR8 IAV at 4 MOI, samples were collected at 3-, 6-, 12-, and 24-hours post infection, and processed for Western blotting and densitometry analysis of IFITM3 expression. GADPH was probed as loading control. (B) Control and SP-R210_L(DN) cells were transiently transfected with IFITM3 siRNA and knockdown confirmed by western blotting. (C) IFITM3 siRNA knockdown partially restores expression of IAV viral genomes in SP-R210_L(DN) cells, without effect on WT cells. *p-value<0.05 as indicated, n=3 experiments. (D) Knock down of IFITM2 had no effect on IAV infection. Infection of siRNA transfected cells was assessed 24-hours after addition of virus,

Given that SP-R210_L(DN) cells express higher constitutive levels of IFITM3 compared to controls (Figure 8A above ), we expected increased localization of IFITM3 with endosomal markers in uninfected cells. The studies on Figure 9A-F, however, show that this was not the case. The strength of colocalization between IFITM3 and endosomal markers was assessed using both Pearson’s correlation coefficient (PCC) (Figure 9B and E) and intensity correlation quotient (ICQ) (Figure 9C and F) statistical analyses of confocal imaging data. We expected sequential decline and increase in acquisition of Rab5 and Rab7 reflecting endosomal maturation over time [81]. Both PCC and ICQ analysis of uninfected cells showed similar but moderate constitutive colocalization of IFITM3 and endosomal markers with differences in colocalization kinetics after infection. PCC analysis showed decline in IFITM3/Rab5 colocalization for the first 15-minutes and after 30-minutes in SP-R210_L(DN) and control cells (Figure 9B), respectively. IFITM3/rab5 colocalization returned to basal levels after 15-minutes in SP-R210_L(DN) cells but was unchanged at both 15- and 30-minutes in SP-R210(KO) cells. The ICQ analysis showed similar trends as PCC (Figure 9C), although less pronounced, indicating the involvement of a small fraction of Rab5+ vesicles. The IFITM3/Rab7 colocalization trended lower (Figure 9E) or transiently lower (Figure 9F). The co-localization of IFITM3/Rab7 in SP-R210_L(DN) cells (Figure 9E,F) followed similar kinetics as the IFITM3/Rab5 colocalization (Figure 9B,C). In SPR210(KO) cells, however, IFITM3+ vesicles acquired Rab7 over time consistent with endosomal maturation (Figure 9E,F) and the kinetics of the FITC-transferrin endosomal marker shown above (Figure 5C above). The IFITM3/Rab7 co-localization in SP-R210(KO) cells was significantly different than control and SP-R210_L(DN) cells at 15-minutes after infection (Figure 9E-F), suggesting differences in endosomal trafficking of IFITM3 in SP-R210(KO) cells. Additional studies, however, are needed to determine if these colocalizations represent virus containing endosomes or vesicles that have not fused with virus containing macropinosomes. These results, together with the trafficking and inhibitor studies above (Figures 4F-G, 5G, and 6L-O), indicate that both SP-R210_S and IFITM3 are needed to couple anti-viral responses and restrict infection in macropinosome and endo-lysosome compartments of macrophages (Working model on Figure 10).

Figure 9. Depletion of both SP-R210 isoforms enhances endosomal localization of IFITM3 during IAV infection.

Control, SP-R210_L(DN), and SP-R210(KO) cells cultured on glass coverslips were infected with 4 MOI of PR8 for 15- and 30-minutes. Cells were fixed, permeabilized, and stained for IFITM3 with either Rab5 (A) or Rab7 (D) to visualize early and late endosomes, respectively. Images were captured using a Nikon Confocal microscope. Colocalization analysis of fluorescent images of 20 random cells was performed using Pearson’s correlation coefficient (PCC) (B, E) or intensity correlation quotient (ICQ) (C,F). Significant differences were observed only for IFITM3/Rab7 colocalization at 15-minutes after infection in SP-R210(KO) cells compared to WT and SP-R210_L(DN) cells with transient declines in the latter (D,E,F). *p<0.05, ^$$p<0.01 at 15-minutes between SP-R210KO vs control and SP-R210_L(DN), respectively, based on 20 cells per cell line in random microscopic fields. Data are means±SD.

Figure 10. Working model of SP-R210 isoform-mediated macropinocytosis of IAV.

Binding of IAV to cell membrane and SP-R210 isoforms elicits macropinocytic membrane projections and macropinocytic cup formation in a RAC-dependent manner leading to internalization of the virus in macropinosomes. SP-R210_L inhibits acquisition of TLR7+ endosomes during macropinosome acidification in an IFITM3-dependent manner inhibiting virus sensing in control cell macropinosomes or endosomes. Viral genome exits and then replication and transcription of the viral genome proceeds unimpeded in control cells. On the other hand, SP-R210_S triggers macrophage activation during macropinocytic entry and IFITM3-dependent containment of the virus in TLR7+ macropinosomes of increasing acidity where degradation of the virus facilitates virus sensing and amplification of the anti-viral response in macrophages. Image was generated using BioRender (www.BioRender.com).

Discussion

We report that SP-R210 isoforms modulate RAC-dependent macropinocytosis and sorting of IAV in macrophages under study. Macropinocytic entry in SP-R210_L(DN) cells is coupled to induction of TNFα and IFNβ in SP-R210_L(DN) cells and restriction of the virus in endolysosomes. SP-R210_L limits TLR activation in response to endosomal TLR agonists and infection, and contributes to viral trafficking that enables endosomal escape and subsequent replication of the viral genome in control cells. Depletion of both SP-R210 isoforms relieved the block in viral recognition resulting in continuous rise in secretion of both IFNβ and TNFα, indicating a persistent inflammatory storm in response to the virus that was ineffective at containing viral proliferation in SP-R210(KO) cells. It remains to be determined, however, whether lack of SP-R210_S contributed to temporal delay in IFNβ production allowing the virus to escape and suppress downstream interferon signaling. Studies utilizing pHrodo-Dextran showed that selective depletion of SP-R210_L expression leads to expansion of acidified endosomes. Tracking of pH-induced conformational change of influenza hemagglutinin (HA) indicate transport and retention of IAV in acidified endolysosomes with peak activities 30- and 90-minutes after infection and diminished endosomal escape of viral genomes. The latter peak does not form in either control cells or SP-R210(KO) cells with diminished expression of both isoforms. Uptake of IAV into acidified endosomes, however, was reduced in both SP-R210_L(DN) and SP-R210(KO) cells reflecting the reduced synthesis of new HA in the latter at early infection stage. Studies on IFITM3 colocalization with endosomal markers in control and SPR210(KO) cells suggests that endosomal IFITM3 is not sufficient to prevent viral genome escape in control and SP-R210(KO) cells, although additional studies are needed to assess colocalization with the virus. In this regard, additional studies are needed to characterize progression of macropinosomes to Rab7+IFITM3+SP-R210_S+Rab5-SP-R210_L- vesicles in the endolysosomal system. It is possible, however, that in control cells IFITM3 plays an anti-inflammatory role in a SP-R210_L-mediated trafficking route while IAV escapes through SP-R210-independent endosomes. In this case, macropinocytic inhibitors suppressed IAV uptake by 60%, suggesting additional entry routes. In this regard, IFITM3 inhibits virus-induced cytokine production in dendritic cells by targeting proteasomal degradation of Nogo-B [8]. Depletion of both isoforms uncoupled infection from cytokine response regulation. It is worthy to note, however, that the SP-R210(KO) cells contain diminished expression but similar ratio of the two isoforms as control cells; these become detectable in isolated membranes or after immunoprecipitation. Alternative splicing or exon deletion of the target gene can escape CRISPR-Cas9 deletion [82, 83], although we cannot exclude the presence of endogenous small exon splice variants of the Myo18A gene that were not targeted [30]. The present studies support the concept that IFITM3 and SP-R210_S work together to direct endolysosomal trafficking and restriction of IAV in SP-R210_L(DN) cells via a novel macropinocytic sorting mechanism (Working model on Figure 10).

Macrophages and dendritic cells have adapted macropinocytosis for pathogen surveillance, antigen presentation, signal transduction, and metabolism of oxidized lipids through constitutive and inducible mechanisms [17, 84]. Macropinosomes undergo luminal acidification, shrinkage, and then progression through acquisition of endolysosomal membranes [63, 85]. A number of viruses including influenza exploit macropinocytosis for entry in a variety of cell types, although there is a lack of studies on macropinocytosis of IAV in macrophages [2]. Inhibition of IAV uptake by sodium-hydrogen antiporter 1 (NHE1), RAC and PAK1 inhibitors and SP-R210 antibodies in the present studies indicate that IAV binding stimulates macropinocytic uptake through SPR210/RAC-dependent mechanisms. Constitutive macropinocytosis depends on the calcium sensing receptor CASR and membrane phosphatidic acid. In contrast to constitutive macropinocytosis, stimulated macropinocytosis involves signaling by cell-surface receptors including the receptor for M-CSF and the chemokine CXCL12 receptor CXCR4, and LPS through TLR4 and CD14 [27, 67, 86, 87]. Constitutive and stimulated macropinocytosis can be distinguished by differential sensitivity to inhibitors of the calcium-sensing receptor CASR and sodium-hydrogen exchange antiporter NHE1, respectively [16, 88, 89]. Both forms of macropinocytosis, however, depend on generation of phosphoinositide intermediates and activation of the small RhoGTPase RAC1 at the cell membrane driving remodeling of the actin cytoskeleton, membrane ruffles, and formation of the macropinocytic cup. Macropinosome closure and internalization depends on the downstream RAC1 effector kinase PAK1. RhoGTPases RAC2 and CDC42 upstream of PAK1 activation are also involved in the temporal events that drive macropinosome closure and internalization, although less-well studied [66, 67, 84, 90-93]. Mass spectrometry and Western blot data indicate differential formation of protein complexes containing RAC1, RAC2, CDC42, and GEFs in control and SP-R210-deficient cells. In this regard, MYO18Aα interacts indirectly with CDC42 binding kinase MRCK through PDZ domain-dependent protein complexes and actin through its amino-terminal KE motif [94], which allows it to modulate actomyosin contractility and lamellipodial dynamics [95]. The SP-R210_L isoform, however, appears to have a specific role restricting basal membrane localization of RAC1 as demonstrated by enrichment of RAC1 in isolated membranes of both SP-R210_L(DN) and SPR210(KO) cells, yet the fate of IAV in the two cell lines was markedly different.

The SP-R210_L(DN) cells, however, are distinguished by cell membrane enrichment of ARHGEF1, and differential pull-down of VAV3 and IFITM3 in whole cell extracts compared to depletion of VAV3 in SPR210(KO) cells. In this context, CASR couples to multiple G proteins for signaling including G_α13 (GNA13) [96]. ARHGEF1 links G-protein coupled receptor and RhoGTPase signaling by acting as a GAP for GNA13 but GEF for RhoGTPases like RHOA and RAC1 [16, 96, 97]. VAV3 activates RAC1 and CDC42 through different domains [98]. The functions of both VAV3 and IFITM3 are modulated by tyrosine phosphorylation [99-102]. In the latter case, tyrosine phosphorylation controls IFITM3 membrane localization, endocytic trafficking and degradation [100], which may contribute to differential sorting of the virus in endosomal vesicles.

The polybasic carboxy-terminal CAAX motif of RAC1 mediates GDP-RAC1 and GTP-RAC1 nanoclusters and macropinocytic activity through interaction with phosphatidic acid and phosphatidylinositol-3,4,5-phosphate (PI(3,4,5)P3) at the plasma membrane [103, 104]. Inactive and active RAC1 nanoclusters distribute between PI(4,5)P2 and PI(3,4,5)P3 enriched membrane microdomains in complex with RAC1 GEFs and GAPs depending on cell type [104, 105]. MYO18A co-isolated with protein complexes bound to PI(3,4,5)P3 but not PI(4,5)P2 or PI(3,5)P2 affinity columns [106]. PI(3,4,5)P3 in particular acts as a rheostat for the endosomal antiviral activity of IFITM3 [107]. Topologically, the macrophage SP-R210 isoforms assume a type II membrane protein orientation based on flow cytometry analysis using antibodies targeted to the unique carboxy-terminal and coiled-coil SP-A binding domains of SP-R210 isoforms (Supplemental Figure S8A) on macrophages [35, 36, 58], NK cells [40], and activated T lymphocytes [58]. In agreement, computational prediction [108] indicates a transmembrane helix within the myosin-like motor domain and either large (Supplemental Figure S8B) or truncated intracellular domains encompassing or lacking the PDZ domain in SP-R210_L or SP-R210_S, respectively. Cell-surface localization, however, does not occur in lung epithelial A549 cells where MYO18A is only detectable after permeabilization (Supplemental Figure S8C), consistent with submembrane localization and in shuttling between ER and Golgi of MYO18Aα in epithelial cell types [38, 94]. Direct and allosteric interactions of PDZ domains with plasma membrane and intracellular PIs, cholesterol, and protein ligands have been described [109-113]. Future studies are needed to address the function of SP-R210 isoforms in organization and trafficking of PI lipid enriched membranes driving discrete forms of RAC-dependent macropinocytic sorting to permissive and restrictive macropinosomes. Differential expression of SP-R210 isoforms in macrophages could result in SP-R210 isoform homo- and heterodimers affecting recruitment of different signaling scaffolds to membrane microdomains at the cell surface.

Conclusions

We report that SP-R210 isoforms modulate abortive and restrictive forms of IAV infection in a model macrophage cell line [28]. Our findings support the notion that changes in stoichiometric composition and levels of SP-R210 isoforms influence IAV endocytic fate and endocytic properties of macrophages. IAV infection in control and SP-R210_L(DN) cells results in differential phosphorylation of IRF3 and IRF7, respectively [28]. This suggests differential sorting and recognition of IAV in TLR3+ and TLR7+ endosomes, respectively. On the other hand, differential phosphorylation at serine 536 of the NFκB p65 subunit may contribute to survival of infected SP-R210_L-deficient macrophages [28]. Alveolar macrophages maintain high levels of the SP-R210_L isoform through paracrine modulation by SP-A [27]. Lack of SP-A suppresses expression of SP-R210_L on alveolar macrophages while expression of SP-R210_S is not affected. SP-A limits virus recognition at extracellular and endosomal levels [48]. Deletion of SP-A enhances clearance of IAV infection in connection with enhanced expression of IFNβ [48]. Accordingly, SP-A suppresses IFNβ by SP-R210_L(DN) cells but not SPR210(KO) cells depending on infection severity (Supplemental Figure S9). The role of the SP-A/SPR210 receptor isoform pathway in the evolution of the inflammatory immune response to IAV infection in vivo awaits further investigation.