Siglecs facilitate HIV-1 infection of macrophages through adhesion with viral sialic acids

Abstract

Background: Human immunodeficiency virus type 1 (HIV-1) infects macrophages effectively, despite relatively low levels of cell surface-expressed CD4. Although HIV-1 infections are defined by viral tropisms according to chemokine receptor usage (R5 and X4), variations in infection are common within both R5- and X4-tropic viruses, indicating additional factors may contribute to viral tropism.

Methodology and principal findings: Using both solution and cell surface binding experiments, we showed that R5- and X4-tropic HIV-1 gp120 proteins recognized a family of I-type lectin receptors, the Sialic acid-binding immunoglobulin-like lectins (Siglec). The recognition was through envelope-associated sialic acids that promoted viral adhesion to macrophages. The sialic acid-mediated viral-host interaction facilitated both R5-tropic pseudovirus and HIV-1(BaL) infection of macrophages. The high affinity Siglec-1 contributed the most to HIV-1 infection and the variation in Siglec-1 expression on primary macrophages from different donors was associated statistically with sialic acid-facilitated viral infection. Furthermore, envelope-associated sialoglycan variations on various strains of R5-tropic viruses also affected infection.

Conclusions and significance of the findings: Our study showed that sialic acids on the viral envelope facilitated HIV-1 infection of macrophages through interacting with Siglec receptors, and the expression of Siglec-1 correlated with viral sialic acid-mediated host attachment. This glycan-mediated viral adhesion underscores the importance of viral sialic acids in HIV infection and pathogenesis, and suggests a novel class of antiviral compounds targeting Siglec receptors.

Figure 1. A structural model of gp120 and associated glycans.

(A) Location of glycosylation sites on the IIIB strain of HIV-1 gp120 with complex (ψ) and high mannose or hybrid (ϕ) types. (B) Distribution of glycans on HIV-1 gp120 and gp41 trimer (cyan) in complex with CD4 (red). Coordinates for the gp120 trimer model with two-domain CD4 was kindly provided by P. Kwong . The four-domain CD4 model (pdb code 1WIO) is placed in the complex by superposition of the two N-terminal domains. The view is from host cell down to virus. The V1 and V2 loops are drawn in the figure, as both loops are missing in the crystal structure of the gp120 and CD4 complex (pdb code 1GC1). Subunits are color-coded cyan for gp41 and gp120, red for CD4, magenta for complex-type carbohydrates, and green for high mannose and hybrid type-carbohydrates. The V3 loop region is denoted by a red circle.

Figure 2. SPR experiments between gp120 and Siglecs.

(A) Binding of recombinant gp120 from 92US715 (0.65 µM), 92UG21-9 (0.49 µM), and 93UG037 (0.2 µM) HIV-1 isolates onto protein A-captured recombinant Siglec-1, -5, -7, and -9 Fc fusion proteins. The sensorgrams in each panel represent individual Siglecs captured in the order of decreasing surface densities between black, grey and dashed curves. Flow cell 1 has immobilized protein A but no captured receptor (blank). The binding curves of gp120 to captured Siglec-3 are shown in Figure S2. (B) Binding of 10 µM α(2,8)-sialyllactose-PAA (8′SA-PAA) onto protein A-captured recombinant Siglec-1 and -9 Fc fusion proteins in similar settings as (A). (C) Binding of serial dilutions of 92US715 gp120 between 0.43–0.054 µM or 8′SA-PAA between 0.062–20 µM onto immobilized Siglec-5 and -7 under high immobilization density. (D) Binding of 92US715 gp120 and 8′SA-PAA to protein A-captured Siglec-9 in the presence of α(2,6)-sialyllactose (6′-SL), and the effect of deglycosylation of 92US715 gp120 on the binding of protein A-captured Siglec-9. All sensorgrams are shown in response units (vertical axis) versus sample injection time (horizontal axis) in seconds. Dissociation constants are listed in Table 1.

Figure 3. Recombinant biotin-labeled gp120 from the 93MW959 isolate (R5-tropic subtype C) binding to Siglec-transfected CHO cells.

(A) Untransfected (open histogram) or Siglec-transfected CHO cells (grey histogram) were incubated with gp120 after being blocked with BSA and murine IgG, and then stained with PE-conjugated streptavidin. (B) The panels are similar to (A) except the cells were pretreated with neuraminidase (NA). (C) The Siglec-1 expression levels detected by an anti-Siglec-1 antibody (top panel) from clones 118 (open histogram) and 78 (grey histogram) of transfected CHO cells correlated with their binding to gp120 (bottom panel). (D) Binding of gp120 to Siglec-1 transfected CHO cells in the presence (grey histogram) or absence (open histogram) of 50 µg/mL human Siglec-1 blocking antibody (top) or pre-incubated with (grey histogram) or without (open histogram) 50 µg/mL recombinant soluble CD4 (bottom). Dotted lines correspond to staining with PE-conjugated streptavidin only. (E) Binding of gp120, untreated (open), NA or sodium periodate (NaIO₄) treated (grey), to Siglec-1, or Siglec-9-transfected CHO cells. Siglec-9-transfected cells were pretreated with NA. (F) Binding of gp120 to Siglec-1 or Siglec-9 transfected CHO cells in the presence of titrating α2,3-/α2,6-sialyllactose (grey) or lactose (white).

Figure 4. Siglec expression and recombinant gp120 binding to human monocytes or MDM.

(A) The expression of Siglecs (grey) on CD14⁺ monocytes. (B) The expression of Siglecs (grey) on CD14⁺ MDM. The controls were stained with PE-conjugated isotype-matched IgGs. All samples were blocked with mouse IgG prior staining. (C) Binding of biotinylated gp120 from 93MW959 (top) or 92UG21-9 (bottom) isolates to NA-treated (grey) or untreated (open) MDM in the presence (left) or absence (right) of a CD4-blocking antibody (OKT4). (D, E) Binding of biotinylated 93MW959 (top) or 92UG21-9 (bottom) recombinant gp120, untreated (open histogram) or treated (grey histogram) with NA (D) or NaIO₄ (E) to NA treated MDM in the presence of a CD4-blocking antibody (OKT4).

Figure 5. Single-round infection assays using R5- and X4-tropic HIV pseudoviruses.

Except for panel A, which is shown in Arbitrary Light Units (ALU), the levels of viral entry (grey bars) are shown as percentage compared to their respective controls (open bars). (A) R5 and X4-tropic HIV and VSV entry into MDM in the presence of (from left to right) 0, 0.1, 1, 10, or 50 mg/mL sialyllactose (SL). (B) Entry of JRFL and VSV into MDM in the presence of 50 mg/mL SL or lactose, 1 mM EDTA, or PBS. (C) R5- and X4-tropic HIV-1 and VSV entry into MDM treated with 50 mg/mL SL, 1 mg/mL asialofetuin, 1 mg/mL fetuin, or PBS. (D) Entry of JRFL and VSV into MDM in the presence of 100 µg/mL recombinant Siglec-1, Siglec-3, Siglec-9, 50 mg/mL SL, or PBS. (E) Infection of MDM using R5-(left side) and X4-(right side) tropic HIV pseudoviruses treated with 1 mM NaIO₄ (grey) or glycerol (white). Results are representative of one out of three experiments.

Figure 6. Effect of Siglec-specific compounds and blocking antibodies on pseudovirus and HIV-1_BaL infections of MDM.

(A) Entry of JRFL and VSV into MDM in the presence of 500 ng/mL CVN, 100 µg/mL recombinant Siglec-9, 50 mg/mL mannose, 1 mg/mL mannan, 1 mM EDTA, or PBS. (B) Infection of MDM with JRFL and VSV pseudoviruses in the presence of 20 µg/mL Siglec-1 and -3, 100 µg/mL of Siglec-9 blocking antibodies, or 100 µg/mL control IgG (from mouse or sheep, open bars). (C & D) Infections of MDM with HIV-1_BaL (125 TCID₅₀/10⁶ cells) were measured in p24 (ng/mL) at various days post-infection (DPI) in the presence of 50 mg/mL SL (black circles), 100 µg/mL T20 (black triangles), or PBS (open squares). Graphs A & B represent two infection experiments with different donor cells.

Figure 7. Host variation in pseudovirus and HIV-1_BaL infections of MDM.

(A) Infections of MDM from five separate donors with JRFL, SF33 or VSV pseudoviruses in the presence of 50 mg/mL SL or PBS. Levels of viral entry (grey bars) are shown as percentage compared to PBS control (shown as a dotted line at 100% level). (B) Infections of MDM from four separate donors with JRFL pseudovirus in the presence of 20–100 µg/mL anti-Siglec-1 antibody (grey bars) or control sheep IgG (white bar). 100 ug/mL α-Siglec-1 was used in experiments with donors 13, 14, and 15; 20 ug/mL α-Siglec-1 was used with donor 16. Results are representative of one out of three experiments. (C) Relative infection levels from seven separate donors compared to their PBS controls on days 12 or 14 post-infection with the HIV-1_BaL in the presence of 50 mg/mL SL (patterned bars) or PBS (dotted line at 100% level). (D–I) HIV-1_BaL relative infection levels plotted against host expression of Siglec-1 (D), Siglec-3 (E), Siglec-9 (F), CCR5 (G), CD4 (H), and CD14 (I). The error bars on donors 6, 9, and 11 are less than 1% and do not appear in the figure. p-values were generated using Spearman correlation.

Figure 8. A model displaying the involvement of Siglec receptors in HIV-1 infection of macrophages.