Endocytic function is critical for influenza A virus infection via DC-SIGN and L-SIGN

The ubiquitous presence of cell-surface sialic acid (SIA) has complicated efforts to identify specific transmembrane glycoproteins that function as bone fide entry receptors for influenza A virus (IAV) infection. The C-type lectin receptors (CLRs) DC-SIGN (CD209) and L-SIGN (CD209L) enhance IAV infection however it is not known if they act as attachment factors, passing virions to other unknown receptors for virus entry, or as authentic entry receptors for CLR-mediated virus uptake and infection. Sialic acid-deficient Lec2 Chinese Hamster Ovary (CHO) cell lines were resistant to IAV infection whereas expression of DC-SIGN/L-SIGN restored susceptibility of Lec2 cells to pH- and dynamin-dependent infection. Moreover, Lec2 cells expressing endocytosis-defective DC-SIGN/L-SIGN retained capacity to bind IAV but showed reduced susceptibility to infection. These studies confirm that DC-SIGN and L-SIGN are authentic endocytic receptors for IAV entry and infection.

lung 13 . A growing body of literature indicates that DC-SIGN and/or L-SIGN recognize glycans expressed by a range of different viruses to promote attachment and infection, as well as the capture and sequestration of virus, which may then be passed on to other permissive cells (reviewed in 10 ).
The relevance of DC-SIGN in enhancing the infectious entry of IAV into primary human cells has been established in studies by Wang et al. 9 and Hilliare et al. 7 , with both reporting that infection of monocyte-derived DCs could be blocked by anti-DC-SIGN mAb. In a previous study we demonstrated that SIA-deficient Lec2 CHO cells were resistant to IAV infection, however expression of DC-SIGN or L-SIGN by Lec2 cells resulted in Ca 2+ -dependent IAV attachment and enhanced susceptibility to IAV infection 8 . DC-SIGN/L-SIGN recognize glycans expressed on the viral HA and the neuraminidase (NA) glycoproteins. While the degree of HA glycosylation is highly variable between strains, levels of NA glycosylation tend to be more conserved (reviewed in 14 ). Moreover, previous studies have demonstrated that the degree of glycosylation on the viral HA, rather than the NA, was critical in determining the ability of different viruses to infect Lec2 cells expressing either DC-SIGN or L-SIGN 8 . While we have some insight regarding the nature of IAV ligands recognized by CLR, the mechanisms by which DC-SIGN and L-SIGN promote IAV infection have not been defined. For example, DC-SIGN/L-SIGN may act as attachment factors, concentrating virions at the cell surface to promote interaction with other (unknown) receptors for infectious entry. Alternatively, recognition of IAV by DC-SIGN/L-SIGN may trigger direct receptor-mediated internalization, demonstrating their ability to act as bone fide entry receptors for IAV infection. The aim of this current study was to define the mechanisms by which DC-SIGN/L-SIGN act to enhance IAV infection.
Mutation of putative internalization domains in the N-terminal cytoplasmic tail of DC-SIGN (LL, YXXL and EEE) or L-SIGN (LL) has demonstrated the importance of the LL motif for efficient CLR-mediated endocytosis and trafficking 15,16 . Moreover, deletion of the entire cytoplasmic domain has been used to generate endocytosis-defective mutants of DC-SIGN/L-SIGN 16,17 . Herein, we demonstrate that Lec2 cells expressing DC-SIGN or L-SIGN with mutations (in which the dileucine motif (LL) was replaced by dialanine (AA)) or deletions (33 and 41 amino acids for DC-SIGN and L-SIGN, respectively) within the cytoplasmic tail bound IAV efficiently, but showed major defects in their ability to internalize monoclonal antibody (mAb) and virus, and in their susceptibility to IAV infection. Our studies confirm that both DC-SIGN and L-SIGN function as authentic receptors for IAV uptake and infection.

Infectious entry of IAV into SIA-deficient Lec2 cells expressing DC-SIGN/L-SIGN occurs independently of SIA and is pH-and dynamin-dependent.
Our previous studies have demonstrated that SIA-deficient Lec2 CHO (Lec2-ctrl) cells were resistant to infection by IAV, whereas Lec2 cells expressing human DC-SIGN or L-SIGN were highly susceptible 8 . These findings were confirmed using IAV strain BJx109 (Fig. 1a). Previous studies also demonstrated that IAV infection of Lec2 cells expressing DC-SIGN/L-SIGN was inhibited by mannan, but not by pre-treatment of cells with bacterial sialidase 8 , arguing that infection of Lec2 cells expressing CLR occurred independently of SIA. However, IAV strains differ markedly in their preference for particular SIA linkages and we have not formally addressed if this impacts on the ability of IAV to infect Lec2 cells via DC-SIGN/L-SIGN. IAV strain HKx31 shows HA preference for α 2,6-linked SIA whereas a horse-serum resistant (HS R ) mutant of HKx31 shows preference for α 2,3-linked SIA 18 , although both viruses express similar numbers of glycans on the viral HA. Despite differences in SIA specificity, HKx31 and HKx31 HS R infected Lec2-DC-SIGN and Lec2-L-SIGN cells to comparable levels ( Fig. 1b), consistent with the notion that glycosylation of the viral HA, rather than its specificity for particular SIA linkages, is the most important determinant of CLR-mediated infection.
As expression of a single CLR (either DC-SIGN or L-SIGN) restores the susceptibility of Lec2 cells to IAV infection, we can probe this unique experimental system to determine the mechanisms by which a single receptor can promote IAV infection. It is well established that IAV infection is pH dependent and that the acidic pH of the endosomes induces an irreversible conformational change in the viral HA, promoting membrane fusion and release of the viral capsid into the cytoplasm (reviewed in 19 ). Therefore, we used NH 4 Cl to examine the rate of infectious entry of IAV into CHO-ctrl, Lec2-DC-SIGN or Lec2-L-SIGN cells. In these experiments, cell monolayers were inoculated with BJx109 virus for 1 hr at 37 °C, washed and cultured for various times before the addition of 10 mM NH 4 Cl to retain virus in endosomes and prevent further infection. All wells were fixed at 8 hr post-inoculation and stained for expression of viral NP. As seen in Fig. 1c, inoculation of CHO-ctrl, Lec2-DC-SIGN and Lec2-L-SIGN with BJx109 (with no NH 4 Cl added = mock) resulted in 60-80% infection 8 hr post-exposure whereas addition of NH 4 Cl immediately after addition of virus (t = 0 hr) reduced infection levels to < 1%. Infectious entry of BJx109 into CHO-ctrl cells occurred during the first 30 min at 37 °C and reached maximum levels (i.e. equivalent to mock) by 2 hr. In contrast, entry via DC-SIGN or L-SIGN was minimal during the first 30 min and increased progressively, with maximum levels detected by 4-6 hr.
To gain insight regarding the mechanisms underlying infectious entry of IAV via DC-SIGN/L-SIGN, we included dynasore, a small molecule inhibitor of clathrin-and caveolin-mediated endocytosis 20 (Fig. 1d). Compared to mock (i.e. no dynasore), addition of 50 μM dynasore at the time of virus inoculation (i.e. t = 0 hr) reduced levels of IAV (BJx109) infection to < 5%, confirming that entry into all cell types occurs via dynamin-dependent mechanisms. Addition of dynasore 2 hr after inoculation had only minor effects on levels of IAV infectivity, confirming that the inhibitory effects of dynasore are not mediated at post-entry steps in the virus life cycle. Further, the observation that dynasore did not inhibit the ability of PIV3 to infect CHO-ctrl cells (Fig. 1d), confirmed the specificity of dynasore-mediated inhibition of IAV infection since PIV-3 infects cells in a pH-independent manner via direct fusion with the plasma membrane (reviewed in 21 ).  In some wells, 10 mM NH 4 Cl was added to the cells at the time of virus addition (t = 0 hr) or at various times thereafter to prevent further infection. All cells were incubated for 8 hr. Data represent the mean percent infected cells (± 1 SEM). *#^s ignificantly different compared to mock for CHO-ctrl, Lec2-DC-SIGN and Lec2-L-SIGN cells, respectively. p < 0.05, one-way ANOVA followed by Tukey's post-hoc test. (d) CHO-ctrl, Lec2-ctrl, Lec2-DC-SIGN and Lec2-L-SIGN cells were incubated with 10 7 PFU of BJx109 (MOI 50 PFU/cell) for 60 min at 37 °C and then washed (mock). In some wells, 50 μM of dynasore was added to the cells at the time of infection (t = 0 hr) and replaced after washing, or dynasore was added two hours after incubation with virus (t = 2 hr). Cells were incubated for a total of 8 hr. Additional wells incubated with 10 6 FFU of PIV-3 (MOI 5 FFU/cell) in the presence or absence of 50 μM dynasore were fixed and stained for expression of PIV-3 NP at 16 hr post-infection. Data represent the mean percent infected cells (± 1 SEM). *+#^s ignificantly different compared to mock for CHO-ctrl, Lec2-ctrl, Lec2-DC-SIGN and Lec2-L-SIGN cells, respectively. p < 0.05, one-way ANOVA followed by Tukey's post-hoc test. Data in (c,d) are pooled from 3 independent experiments. indicated that LL was the major motif contributing to the endocytic capacity of DC-SIGN and L-SIGN 15,16 . Therefore, we generated constructs expressing DC-SIGN/L-SIGN with the LL motif mutated to AA, or with the entire N-terminal intracellular domain of DC-SIGN (33 amino acids) or L-SIGN (41 amino acids) deleted (Fig. 2a). These constructs were used to generate stable Lec2 cell lines expressing wild-type (DC-SIGN-WT, L-SIGN-WT), mutated (DC-SIGN-AA, L-SIGN-AA) or deleted (DC-SIGN-DEL, L-SIGN-DEL) forms of each CLR. Cell surface expression of DC-SIGN and L-SIGN was confirmed using fluorescein-conjugated mAb 120612, which is cross-reactive between the two CLRs ( Fig. 2b). Note that Lec2 lines expressing WT or DEL forms of DC-SIGN and L-SIGN showed similar levels of cell-surface receptor whereas Lec2-DC-SIGN-AA and Lec2-L-SIGN-AA lines expressed higher levels of cell-surface CLR.
To determine if mutation or deletion within the N-terminal intracellular domain abrogated the endocytic capacity of DC-SIGN/L-SIGN, cell monolayers were incubated with anti-DC-SIGN/L-SIGN mAb at 4 °C to facilitate binding to cell-surface CLR, then moved to 37 °C to promote internalisation. Following incubation at 4 °C, mAb bound to the surface of Lec2 cells expressing -WT, -AA and -DEL forms of DC-SIGN (Fig. 2c, upper panels) or L-SIGN (Fig. 2c, lower panels). When cell monolayers were moved to 37 °C, the anti-DC-SIGN/L-SIGN mAb redistributed into intracellular compartments in cells expressing WT DC-SIGN or L-SIGN (Fig. 2c), however the majority of mAb remained cell-surface associated for cells expressing -AA and -DEL mutants of either CLR. Note that mAb internalized by Lec2-DC-SIGN-WT or Lec2-L-SIGN-WT cells at 37 °C colocalized with Rab5 + compartments ( Fig. 2d; Pearson's coefficient: Rr = 0.54 and 0.65 respectively), whereas little colocalization was evident in Lec2 cells expressing -AA or -DEL mutants ( Fig. 2d; Pearson's coefficient: Rr < 0.5 for all mutants). Together, these data confirm that cross-linking of DC-SIGN and L-SIGN expressed by Lec2 cells results in endocytosis and delivery to Rab5 + early endosomes whereas mutation (LL to AA) within or deletion of the intracellular cytoplasmic domain abrogates endocytic capacity of DC-SIGN and L-SIGN.

Lec2 cells expressing endocytosis-defective mutants of DC-SIGN/L-SIGN bind IAV efficiently, but exhibit defects in virus internalization and are resistant to IAV infection.
We compared the ability of Lec2 cells expressing -WT, -AA and -DEL forms of DC-SIGN and L-SIGN to bind to IAV in the presence or absence of Ca 2+ . Representative histograms in Fig. 3a confirm that cell lines expressing WT and mutant forms of DC-SIGN/L-SIGN all bound IAV strain BJx109 in a Ca 2+ -dependent manner, consistent with CLR-mediated recognition of virus. Analysis of the geometrical means of the fluorescence intensity from triplicate samples indicated that -WT and -DEL mutants exhibited similar capacity to bind IAV whereas -AA mutants had a tendency to bind increased amounts of virus (Fig. 3b), consistent with enhanced cell-surface expression of -AA compared to -WT and -DEL on the Lec2 lines used in these studies (Fig. 2b). Next, we compared the ability of BJx109 to infect Lec2 cell lines expressing WT and mutated forms of DC-SIGN/L-SIGN, as measured by immunofluorescent staining of viral NP at 8 hr post-infection. For Lec2 cells expressing DC-SIGN, the AA mutation was associated with a significant reduction in sensitivity to infection by BJx109 and infection was further reduced in Lec2-DC-SIGN-DEL cells (Fig. 4b(i)). For cells expressing L-SIGN, the AA mutation or deletion of the intracellular domain resulted in a profound reduction in susceptibility to BJx109 infection compared to cells expressing WT CLR (Fig. 4b(i)). The BJx109 strain expresses 4 potential sites of N-linked glycosylation on the head of the viral HA whereas the PR8 strain lacks glycosylation on the head of its HA glycoprotein 5,6,22 . Sialylated CHO-ctrl cells were susceptible to infection by PR8 whereas SIA-deficient Lec2-ctrl cells were not (Fig. 4b(ii)). Moreover, expression of -WT, -AA or -DEL forms of DC-SIGN or L-SIGN did not restore susceptibility to infection, confirming the critical role of HA glycosylation in determining sensitivity to CLR-mediated infection.

Discussion
It is well established that DC-SIGN and L-SIGN are mannose-specific CLRs that act as attachment factors and promote infection by a range of different viruses, including Ebola and Marburg viruses, dengue virus (DV), human immunodeficiency virus (HIV), hepatitis C virus (HCV) and phleboviruses (reviewed in 3,23 ). In most studies the mechanism/s underlying CLR-mediated infection enhancement have not been defined. However, it is clear that DC-SIGN can enhance infection of viruses by mechanisms that are dependent and independent of DC-SIGN-mediated endocytosis. For example, cells expressing WT or endocytosis-defective (LL and DEL) mutants of DC-SIGN showed similar sensitivity to DV infection 16 . Similarly, mutations within the cytoplasmic tails of DC-SIGN/L-SIGN that prevent efficient internalization of mAb-lectin complexes did not abrogate augmentation of Ebola virus infection 17 . Thus DC-SIGN and/or L-SIGN appear to act as attachment factors for these viruses, enhancing infection via cooperation with one or more additional cellular receptors. In contrast, cells expressing endocytosis-defective (LL) mutants of DC-SIGN were resistant to pH-dependent infection via arthropod-borne phleboviruses 24 , confirming that DC-SIGN can act as an authentic receptor for viral attachment and endocytosis, leading to infectious entry. Our findings that AA and DEL mutants of DC-SIGN/L-SIGN showed no defects in their ability to bind IAV but differed dramatically in their ability to facilitate endocytosis of mAb and virus, as well as their susceptibility to IAV infection, demonstrate that both DC-SIGN and L-SIGN represent attachment and internalization receptors for IAV. Hilliare et al. 7     of an anti-DC-SIGN mAb, indicating that CLR-mediated infection of primary cells can occur independently of cell surface SIA. We previously reported that pre-treatment of CHO cells with bacterial sialidase abrogated IAV infection but did not alter infection of Lec2 cells expressing either DC-SIGN or L-SIGN 8 . These studies confirm that high level expression of CLRs on SIA-deficient transfectants bypassed the requirement for SIA-mediated attachment. Herein, we confirm that DC-SIGN or L-SIGN expression on SIA-deficient Lec2 cells promotes CLR-mediated attachment, endocytosis and targeting of IAV to acidified endosomes, resulting in IAV infection, albeit at a somewhat slower rate than for sialylated CHO cells (Fig. 1c). Previous studies using epithelial cells have demonstrated that when one entry pathway is blocked (for example, clathrin-mediated, caveolin-mediated or macropinocytosis), IAV can enter or infect via another pathway, suggesting multiple routes of infectious entry into sialylated epithelial cells [25][26][27] . Therefore, sialylated parental CHO cells are likely to express multiple different attachment and entry receptors to facilitate infectious entry and this may be an important factor determining the more rapid infection of these cells compared to SIA-deficient CLR-expressing transfectants.
Identification and characterization of specific transmembrane glycoproteins that function as attachment factors and entry receptors for IAV has been complicated by the ubiquitous presence of SIA on the surface of mammalian cells. While interactions between IAV and epithelial cells induce signalling components implicated in virus uptake, such as phosphatidyl inositol 3-kinase (PI3K) 28 and protein kinase C (PKC) 29 , it is not clear how virus binding transmits signals across the plasma membrane. Recently, multivalent binding of IAV to cell-surface SIA was reported to generate a signalling platform at the plasma membrane, activating receptor tyrosine kinases (RTKs) and subsequently intracellular signalling molecules such as PI3K 30 . We demonstrate that in the absence of SIA, IAV binding to DC-SIGN or L-SIGN can also activate cellular signalling pathways to induce dynamin-dependent internalisation of virus. Although the specific signalling components are yet to be elucidated, our data clearly implicate the LL motif in the cytoplasmic tail of DC-SIGN and L-SIGN in receptor-mediated IAV entry and infection.
Cross-linking of cell-surface DC-SIGN with mAb has been shown to induce efficient endocytosis of DC-SIGN in DCs 31 , suggesting that IAV-induced clustering of DC-SIGN/L-SIGN may activate signalling pathways that induce dynamin-dependent uptake of the receptor cluster and associated virus. However, while ligation with mAb relocalizes DC-SIGN to late endocytic and lysosomal compartments 15 , DC-SIGN is retained in early endosomes following uptake of HIV-1 32 , suggesting that cargo-dependent signalling may determine the particular internalization pathway utilized. DC-SIGN-mediated uptake of soluble antigens is impaired in the presence of clathrin inhibitors 33  Viruses. The IAV strains used in this study were BJx109 (H3N2, a high-yielding reassortant of A/PR/8/34 (PR8, H1N1) with A/Beijing/353/89 (H3N2) that expresses the H3N2 surface glycoproteins) and PR8 (A/ PR/8/34, H1N1). HKx31 (also known as X-31, a reassortant of PR8 with A/Aichi/2/68 (H3N2), bearing the H3N2 surface glycoproteins) and horse-serum resistant HKx31 (HKx31 HS R ) 18 were also used. Viruses were grown in 10-day old embryonated eggs and titrated on Madin-Darby canine kidney (MDCK) cells by standard procedures and expressed as plaque forming units (PFU)/ml) 35 . Viruses were purified from allantoic fluid by rate zonal sedimentation on 25 to 75% (wt/vol) sucrose gradients, as described previously 35  DC-SIGN/L-SIGN, or cytoplasmic OVA (CHO-ctrl, Lec2-ctrl) were selected in the presence of 1 mg/ml geneticin (G418; Invitrogen). Transfected cells were screened for cell-surface CLR expression using a fluorescein-conjugated MAb (clone 120612; R&D Systems, Minneapolis, USA) which is cross-reactive between DC-SIGN/L-SIGN and single cells with high cell-surface CLR expression were sorted using a FACSAria cell sorter (BD Biosciences, CA, USA) and expanded in culture for use in experiments.

Binding of anti-DC-SIGN/L-SIGN mAb and IAV to cells. For analysis of mAb and IAV binding by flow
cytometery, adherent cell lines were detached from plastic flasks using 0.75 mM EDTA in phosphate-buffered saline (PBS). After washing, cells were incubated with fluorescein-conjugated anti-DC-SIGN/L-SIGN mAb (clone 120612, R&D Systems) as per manufacturer's instructions in FACs buffer (PBS supplemented with 1% FCS (vol/ vol)) at 4 °C for 40 min. After incubation, cells were washed and levels of mAb binding determined using flow cytometry. For IAV binding, detached cells were washed and all subsequent steps performed in lectin buffer (TBS containing 10 mM CaCl 2 and 1 mg of bovine serum albumin (BSA)/ml). Cells were incubated with 5 μg/ml of purified BJx109 in lectin buffer at 4 °C for 30 min. After washing, bound IAV was detected with 5 μg/ml biotinylated MAb C1/1 (a kind gift from Prof. Lorena Brown, Department of Microbiology and Immunology, University of Melbourne), followed by streptavidin-conjugated to allophycocyanin ((APC) Life Technologies, Eugene, USA) and flow cytometry. To determine if IAV binding to cells was Ca 2+ -dependent, CaCl 2 was omitted from lectin buffer and replaced with 5 mM EDTA.
Virus infection assays. Cells were seeded into 8-well chamber slides (Lab-Tek, Life Technologies) and infected with IAV as described previously 8 . Briefly, after overnight culture, slides with confluent cell monolayers were washed and incubated with IAV in serum-free α MEM for 1 hr at 37 °C (to allow virus binding and entry). After removal of virus inoculum cells were washed and incubated for a further 6-8 hr at 37 °C in serum-free α MEM. Slides were washed with PBS and then fixed with 80% (vol/vol) acetone. IAV-infected cells were stained using MAb MP3.10g2.1C7 (WHO Collaborating Centre for Reference and Research on Influenza, Melbourne, Australia), specific for the nucleoprotein (NP) of type A influenza viruses followed by FITC-conjugated goat anti-mouse Ig (Millipore, MA, USA). The percentage of infected cells was determined by co-staining with 4′ ,6-diamidoino-2-phenylindole (DAPI) and counting the total number of cells versus FITC positive cells under × 100 magnification. A minimum of four random fields were selected for counting, assessing at least 200 cells for each sample. In some experiments, cell monolayers were treated with 10 mM NH 4 Cl or 50 μM of dynasore (Sigma, St. Louis, USA) at various times as indicated.
For PIV-3, slides with confluent cell monolayers were washed and incubated with PIV-3 in serum-free α MEM for 1 hr at 37 °C. After removal of virus inoculum cells were washed and incubated for a further 16 hr at 37 °C in serum-free α MEM. Slides were washed with PBS and then fixed with 80% (vol/vol) acetone. PIV-3-infected cells were stained using a MAb specific for the HN protein (Abcam, Cambridge, UK) followed by FITC-conjugated goat anti-mouse Ig (Millipore).
Internalisation of CLR-specific mAb or IAV. Cells seeded into 8-well chamber slides were cultured overnight, washed and incubated with anti-DC-SIGN/L-SIGN mAb (clone 120612) or with 1 × 10 8 PFU/ml BJx109 (multiplicity of infection (MOI) 50 PFU/cell) in serum-free nutrient mixture F12K (Hams) media (Gibco-BRL) for 30 min at 4 °C (to allow mAb/IAV binding) or at 37 °C (to allow mAb/IAV entry). After this time, slides were washed with PBS and fixed in 80% (vol/vol) acetone. For binding and internalization of mAb, detection of mAb was enhanced using Alexa Fluor ® 488 chicken anti-mouse Ig (Life Technologies, Eugene, USA). For binding and internalization of IAV, detection was performed using 5 μg/ml of biotinylated MAb C1/1, followed by Alexa Fluor ® streptavidin-488 (Molecular Probes, Oregon, USA). All slides were co-stained with 4′ ,6-diamidino-2-phenylindole (DAPI) to detect cell nuclei and examined by confocal microscopy. Images were acquired with a Ziess LSM700 confocal microscope in conjunction with Zen2012 software. Rr values are indicative of colocalisation/correlation and were calculated using Image J software. In some experiments, cells were transfected with mRFP-Rab5 plasmid (#14437, Addgene, MA, USA) using Lipofectamine ® 2000 transfection reagent (Life Technologies) according to manufacturer's instructions. Cells were then cultured an additional 16 hours to allow for protein expression, washed and used in internalisation assays as described above.

Statistical Analysis.
Graphing and statistical analysis of data was performed using GraphPad Prism (GraphPad Software, San Diego, USA). An unpaired Student's t-test was used to compare two sets of data. When comparing three or more sets of values, the data were analysed by one-way analysis of variance (ANOVA; nonparameteric), followed by post-hoc analysis using Tukey's multiple comparison test. p values ≤ 0.05 were considered significant.