Siglec-1 Is a Novel Dendritic Cell Receptor That Mediates HIV-1 Trans-Infection Through Recognition of Viral Membrane Gangliosides

The novel dendritic cell receptor Siglec-1 binds sialyllactose moieties on HIV-1 membrane gangliosides, thereby enhancing HIV-1 transinfection.


Introduction
HIV-1 can infect CD4 + cells of the lymphoid and myeloid lineage with a strong preference for CD4 + T cells. Myeloid DCs exhibit innate resistance against HIV-1 infection. HIV-2, on the other hand, efficiently infects myeloid DCs due to its accessory viral protein Vpx, not present in HIV-1, which is able to counteract the myeloid restriction factor SAMHD1 [1,2]. Despite lacking these additional target cells, HIV-1 exhibits a higher pathogenicity than HIV-2 and has dominated the global AIDS pandemic. Indeed, myeloid DCs can contribute to the spread of HIV-1 through trans-infection of CD4 + T cells [3,4]. This process requires HIV-1 binding to the DC surface, viral capture and release of trapped viruses at the infectious synapse, a cell-to-cell contact zone between uninfected DCs and interacting CD4 + T cells, which facilitates infection by locally concentrating virus and viral receptors [5].
Classical myeloid DCs patrol submucosal surfaces, where they capture and internalize microbial pathogens through various cell surface receptors. Pioneering studies suggested that HIV-1 is trapped by immature DCs (iDCs) in mucosal tissues through binding of its envelope glycoproteins to the C-type lectin DC-SIGN, with subsequent transfer of infectious particles to secondary lymphoid tissues, where trans-infection occurs [4,6]. Later reports indicated, however, that HIV-1 captured by iDCs is rapidly degraded [7][8][9], arguing against this original ''Trojan horse'' hypothesis. Conversely, maturation of DCs with lipopolysaccharide (LPS), a microbial product significantly augmented in the plasma of HIV-1-infected individuals [10], markedly enhanced the capacity of DCs to capture HIV-1 and mediate trans-infection [5,7,9]. These results suggested that HIV-1 capture by LPSmatured mDCs (LPS mDCs) plays an essential role in HIV-1 pathogenesis, facilitating viral spread in the densely populated lymphoid tissue, where many uninfected T cells contact viruspresenting mDCs.
Other receptors besides DC-SIGN have been identified as binding factors for HIV-1 but do not explain why LPS mDC capture of HIV-1 is independent of viral glycoproteins [9]. Instead, HIV-1 capture is markedly sensitive to reductions in viral sphingolipid content [11] and relies on HIV incorporation of membrane gangliosides [12,13]. Furthermore, we recently showed that sialyllactose in gangliosides serves as the viral attachment factor for LPS mDCs [13]. Since HIV-1 and cellular secreted vesicles, termed exosomes, use the same pathway for mDC capture [11], HIV-1 may have hijacked a pre-existing cellular route for vesicle capture to facilitate efficient transfer to multiple target cells.

Siglec-1 Is Up-Regulated in Highly Trans-Infecting LPS mDCs
To identify the molecule on DCs that mediates HIV-1 and exosome capture, we performed transcriptome analysis on differentially matured DCs with a highly divergent capacity to capture and transmit HIV-1. We used efficiently trans-infecting LPS mDCs and compared them to DCs matured in the presence of the clinical grade cocktail ITIP (ITIP mDCs), which exhibit strongly reduced HIV-1 capture and trans-infection capacity ( Figure 1A) [14]. We focused our analysis on the Siglec family (including CD83) because these type I transmembrane proteins have an amino-terminal V-set domain that had been shown to interact with sialylated ligands [15]. Most members of the family were equally expressed in LPS mDCs and ITIP mDCs, and this was also observed for the maturation marker CD86 ( Figure 1B). DC-SIGN, SIGLEC7, and SIGLEC14 were slightly up-regulated in LPS mDCs, but this difference was not statistically significant for DC-SIGN and marginally significant for SIGLEC14 and SIGLEC7, respectively (p = 0.03 and p = 0.04). In contrast, SIGLEC1 expression was strongly up-regulated in LPS mDCs compared to ITIP mDCs with genome-wide significance (p = 3.5610 24 ; Figure 1B). Furthermore, SIGLEC1 ranked 20 th of all differentially regulated genes in comparative transcriptome analysis. The differential expression of Siglec-1 in LPS and ITIP mDCs was confirmed by quantitative real-time PCR (qRT-PCR; Figure 1C) and Fluorescence Activated Cell Sorting (FACS; Figure 1D). Comparison with iDCs also revealed a significantly higher expression level and surface density of Siglec-1 in LPS mDCs ( Figure 1C,D).

Siglec-1 Expressed in LPS mDCs Capture Distinct Ganglioside Containing Vesicles, Such as HIV-1 Viral-Like Particles, Liposomes, and Exosomes
To test whether Siglec-1 is the surface molecule on LPS mDCs responsible for the capture of vesicles and viruses that carry sialyllactose-containing gangliosides in the outer leaflet of their membrane, we used a previously established FACS assay [11,13]. This assay makes use of HIV-1 virus-like particles lacking the viral envelope glycoproteins and carrying a fusion of the viral structural protein Gag with eGFP (VLP HIV-Gag-eGFP ). These fluorescent VLPs follow the same trafficking route as wild-type HIV-1 in LPS mDCs [11]. VLP capture of LPS mDCs was evaluated in the presence of antibodies (Abs) against different Siglecs or mannan, a C-type lectin inhibitor blocking the HIV-1 interaction with DC-SIGN. Besides Siglec-1, we included Abs against CD83, highly expressed in ITIP and LPS mDCs ( Figure 1B); Siglec-7, moderately up-regulated in LPS mDCs ( Figure 1B); and Siglec-5/14 too, due to their high homology to the V-set domain of Siglec-1. VLP capture was almost completely abolished when LPS mDCs were pre-treated with the a-Siglec-1 monoclonal Ab (mAb) 7D2 (Figure 2A; p,0.0001). However, pretreatment with Abs against other Siglec family members or blockade of DC-SIGN with mannan had no effect ( Figure 2A).
We have previously shown that Texas Red (tRed) labeled Large Unilamellar Vesicles (LUV) mimicking the size and lipid composition of HIV-1 and containing the ganglioside GM1 (LUV HIV-tRed ) follow the same trafficking route as VLP HIV-Gag-eGFP in LPS mDCs. Binding and capture in both cases depends on the recognition of sialyllactose exposed in gangliosides of the vesicle membrane [13]. Accordingly, capture of GM1-containing LUV HIV-tRed by LPS mDCs was efficiently and specifically inhibited by the a-Siglec-1 mAb 7D2 ( Figure 2B; p,0.0001). The residual capture by 7D2-treated LPS mDCs was similar to that exhibited by untreated LPS mDCs capturing LUV HIV-tRed containing GM1 without the sialic acid group (Asialo GM1), confirming that sialic acid in the vesicle membrane is crucial for Siglec-1 recognition ( Figure 2B; p,0.0001). We extended this analysis to cellular exosomes, which also carry sialyllactosecontaining gangliosides in their membrane [16] and can be internalized by LPS mDCs [11]. Fluorescent exosomes were efficiently captured by LPS mDCs, and this capture was almost abolished by mAb 7D2 treatment ( Figure 2C; p,0.0001).
Titration of the a-Siglec-1 mAb 7D2 revealed a dose-dependent inhibition of VLP capture ( Figure 2D). Specificity of the mAb 7D2-mediated inhibition was confirmed by pre-incubation of this mAb with different Siglec proteins. Pre-incubation with purified

Author Summary
Mature dendritic cells (mDCs) capture and store infectious HIV-1 and subsequently infect neighboring CD4 + T cells in lymphoid organs. This process, known as trans-infection, is a key contributor to HIV pathogenesis, but the precise mechanism and the identity of the receptor on the mDC surface that recognizes viral particles remain controversial. Although the interaction of HIV-1 envelope glycoproteins with the C-type lectin DC-SIGN has been suggested to mediate HIV-1 capture and trans-infection, later studies revealed an envelope glycoprotein-independent virus capture mechanism in mDCs. Here, we identify Siglec-1 as the surface receptor on mDCs that boosts their uptake of HIV-1 and their capacity to trans-infect CD4 + cells, leading in turn to HIV-1 disease progression. Siglec-1 captures the virus by interacting with sialyllactosecontaining gangliosides exposed on viral membranes. This indicates that Siglec-1 functions as a general binding molecule for any vesicle carrying sialyllactose in its membrane, including exosomes and other viruses. We suggest that this natural pathway through mDC, which would normally lead to antigen processing and presentation, has been subverted by HIV-1 for its own storage and transmission.
If Siglec-1 serves as a recognition receptor on DCs, its surface expression should correlate with their respective VLP capture ability. Capture was low in iDCs and stable over time ( Figure 2F, left graph), while VLP capture was strongly enhanced following LPS treatment ( Figure 2F, right graph). This increased VLP capture ability directly correlated with a strong up-regulation of Siglec-1 surface expression on LPS mDCs ( Figure 2F, right graph). We also performed quantitative FACS analysis to determine the absolute number of Siglec-1 Ab Binding Sites (ABS) on ITIP mDCs, iDCs, and LPS mDCs ( Figure 2G). The VLP capture capacity of these distinct DC subtypes was strongly correlated with the mean number of Siglec-1 ABS expressed per cell (r = 0.9695; Figure 2G). Furthermore, Siglec-1 expression also correlated with the relative VLP capturing capacity of LPS mDCs derived from the same donor ( Figure S2). These experiments show a direct correlation between Siglec-1 expression on the DC surface and their respective VLP capture capacity.

Siglec-1 Captures HIV-1 and Traffics with the Virus to the same Sac-Like Compartment
To extend these observations to authentic virus, we performed similar experiments with infectious HIV-1. Again, LPS mDCs captured significantly more virus than iDCs or ITIP mDCs ( Figure 3A). The a-Siglec-1 mAb 7D2 inhibited HIV-1 capture of LPS mDCs by 80% ( Figure 3A; p = 0.0019), while pre-treatment with mannan had no effect. Noteworthy, a-Siglec-1 mAbs also blocked binding of HIV-1 to LPS mDCs at 4uC ( Figure S3). Similarly, pre-treatment of iDCs with the mAb 7D2 reduced HIV-1 capture by 60% ( Figure 3A; p = 0.0005), indicating that even at lower surface expression levels of Siglec-1 on iDCs ( Figure 1D), this receptor still constitutes an important capture moiety. Consistently, capture inhibition by mAb 7D2 was much weaker on ITIP mDCs ( Figure 3A; p = 0.001), which exhibited the lowest Siglec-1 surface expression ( Figure 1D). The effect of the mAb 7D2 on HIV-1 capture was dependent on blocking cell surface Siglec-1, as addition of the inhibitor after virus exposure had no effect ( Figure 3B). Importantly, primary blood myeloid DCs exposed to LPS also up-regulated Siglec-1 expression levels ( Figure  S4) and showed increased HIV-1 capture capacity that could be blocked by pretreatment with a-Siglec-1 mAb 7D2 ( Figure 3C; p = 0.0022). These results strongly suggest that Siglec-1 is the molecule responsible for HIV-1 capture by DCs, especially upon triggering of Siglec-1 expression by LPS.
Next, we investigated whether Siglec-1 traffics together with sialylated ligands, such as ganglioside-containing liposomes, VLPs, or HIV-1, reaching the same sac-like compartment where these particles are stored [11,13]. LPS mDCs were pulsed with these different fluorescent particles and subsequently stained with the a-Siglec-1 Alexa 488 mAb 7-239 ( Figure 3D). Confocal microscopy revealed extensive co-localization of Siglec-1 with GM1-containing LUV HIV-tRed , VLP HIV-Gag-Cherry , and HIV-1 Cherry in the same compartment ( Figure 3D, Movies S1, S2, S3, and Figure S5). We then assessed whether binding of a-Siglec-1 mAb 7D2 to LPS mDCs would be sufficient to internalize Siglec-1 into a similar compartment. Following incubation for 4 h at 37uC, most of the bound a-Siglec-1 mAb 7D2 was indeed found within a sac-like compartment ( Figure 3E). Hence, binding of mAb 7D2, probably causing Siglec-1 cross-linking at the cell surface, is sufficient to induce Siglec-1 internalization.

Siglec-1 Mediates HIV-1 Trans-Infection to Target Cells and Accumulates at the Infectious Synapse
To assess the relevance of Siglec-1 for HIV-1 trans-infection, we pulsed distinct DCs with equal amounts of infectious virus in the presence or absence of blocking reagents and cocultured them with a CD4 + reporter cell line ( Figure 4A). Controls performed with the protease inhibitor saquinavir, which abolishes production of infectious virus, demonstrated that this assay measured only trans-infection of reporter cells by DC-captured virus without a contribution from potentially de novo infected DCs ( Figure 4A-B, last bars). Pretreatment of LPS mDCs with the a-Siglec-1 mAb 7D2 inhibited HIV-1 trans-infection by 85% ( Figure 4A; p = 0.0052), while blocking DC-SIGN through mannan had no effect. Analogously, pretreatment of iDCs with 7D2 reduced HIV-1 trans-infection by 55% ( Figure 4A; p = 0.0091). In contrast, ITIP mDC-mediated trans-infection was not affected by 7D2 but was blocked by mannan ( Figure 4A; p = 0.0014). Addition of any of the inhibitors tested after DC viral pulse had no significant effect on trans-infection ( Figure 4B), except for the mAb 7D2 in LPS mDCs (p = 0.0069). This latter inhibitory effect could not be explained by differences in viral capture ( Figure 3B) but is most likely attributed to the cell-to-cell adhesion function of Siglec-1 [19], where establishment of infectious synapses may be partially impaired when Siglec-1 is blocked in LPS mDCs. Indeed, when we analyzed infectious synapse formation between HIV-1 Cherry pulsed LPS mDCs cocultured with CD4 + T cells, Siglec-1 polarized towards the site of the cell-to-cell contact zone where viruses were also concentrated ( Figure 4C). The importance of Siglec-1 for HIV-1 trans-infection was also confirmed for blood myeloid DCs. LPS stimulation strongly enhanced their potential for trans-infection ( Figure 4D; p,0.0001), and this increase could be abolished by pre-incubation with mAb 7D2 ( Figure 4D; p,0.0001).

SIGLEC1 Silencing Blocks Viral Capture and Trans-Infection, While De Novo Expression of SIGLEC1 Rescues It
To verify the essential role of Siglec-1 during HIV-1 capture and trans-infection, we applied two complementary experimental strategies: RNA interference to reduce Siglec-1 expression levels in LPS mDCs and transfection of Siglec-1 into cells devoid of this receptor. In the first approach, we transduced DCs with lentiviral particles coding for different shRNAs by co-infection with vpxexpressing lentiviruses to counteract the restriction factor SAMHD1 and facilitate DC productive infection. Transduction of two different SIGLEC1-specific shRNAs, but not of a nontarget shRNA control, led to a drastic decrease in Siglec-1 surface expression and a concurrent loss of VLP HIV-Gag-eGFP capture ( Figure 5A,B). Furthermore, transduction of a SIGLEC1-specific shRNA, but not of a control shRNA, decreased LPS mDC capacity for HIV-1 trans-infection to a reporter CD4 + cell line ( Figure 5C). We next assessed whether endogenous Siglec-1 expression in cells lacking this molecule on their surface could rescue their capacity for HIV-1 capture and trans-ifection. This was first attempted for the monocytic cell line THP-1, but could not be pursued since transfection with any of the plasmids tested up-regulated Siglec-1 expression, probably through TLR signaling ( Figure S6A, top panels). Thus, we chose Raji B cell line instead, which lack endogenous expression of Siglec-1 and could be efficiently transfected without unspecific up-regulation of Siglec-1 ( Figure S6A, bottom panels, and S6B). Transfection of a Siglec-1 expression vector significantly enhanced VLP HIV-Gag-eGFP capture in the Siglec-1-positive cell population, and this effect was abolished by pretreatment with the a-Siglec-1 mAb 7D2      Figure 5D,E). No increased capture was seen in the Siglec-1-negative population of Siglec-1 transfected cells or following transfection of Siglec-5 or Siglec-7 expression plasmids ( Figure 5D,E). Pre-incubation with sialyllactose also blocked VLP capture in Siglec-1 transfected Raji cells ( Figure S7). Accordingly, transfection of a Siglec-1 expression vector into Raji cells significantly increased their capacity for HIV-1 trans-infection to a reporter CD4 + cell line ( Figure 5F), and this effect was again abolished by pre-incubation of transfected cells with the mAb 7D2 (p,0.0001; Figure 5F). Equivalent results were obtained when T cell, shown in a 3D volumetric x-y-z data field. (D) HIV-1 transmission to reporter cells from distinct blood myeloid DCs that had been pre-incubated with 10 mg/ml of the indicated mAbs for 30 min before viral exposure as in (A  Siglec-1 transfected HEK-293T cells were analyzed ( Figure S8). We finally verified that as opposed to DC-SIGN, Siglec-1 viral capture does not rely on the recognition of envelope glycoproteins ( Figure S9). Transfection of a Siglec-1 expression vector in Raji cells allowed for efficient capture of HIV-1 with or without envelope glycoproteins, whereas Raji DC-SIGN cells only captured viruses bearing glycoproteins ( Figure S9). The complementary approaches of SIGLEC1 knockdown and de novo expression on heterologous cells strongly support our conclusion that Siglec-1 is a central molecule mediating HIV-1 capture and trans-infection.

Discussion
Three lines of evidence identify Siglec-1 as a novel DC receptor for HIV-1 capture and trans-infection: (i) Siglec-1 expression correlates with viral capture and trans-infection capacity of DCs, (ii) mAbs against Siglec-1 specifically inhibit HIV-1 capture in a dosedependent manner, and (iii) SIGLEC1 knockdown reduces viral capture and trans-infection, while heterologous de novo expression of Siglec-1 enhances HIV-1 capture and trans-infection. An important role for Siglec-1 in HIV-1 infection is in line with previous studies reporting increased expression of Siglec-1 on CD14 + monocytes and macrophages in HIV-1 infection [20][21][22]. However, these studies analyzed Siglec-1 interactions with sialylated viral envelope proteins, while our results clearly show that HIV-1 capture depends on sialyllactose on viral membrane gangliosides interacting with Siglec-1, but does not require viral glycoproteins.
DC-SIGN was initially proposed as the HIV-1 attachment factor concentrating virus particles on the surface of DCs [4], but later studies showed a variable contribution of DC-SIGN to HIV-1 capture and trans-infection [23]. Our results indicate that both DC-SIGN and Siglec-1 contribute to trans-infection by iDCs, while HIV-1 capture by highly trans-infecting LPS mDCs is independent of DC-SIGN and requires Siglec-1. Hence, although Siglec-1 viral binding via sialyllactose recognition does not discriminate between infectious or noninfectious HIV-1 particles, the greater the expression of Siglec-1, the greater the amount of virions captured and transmitted by DCs, diminishing the relative contribution of DC-SIGN gp120-mediated viral capture to trans-infection. Given that lectins such as DC-SIGN and Siglec-1 generally achieve highavidity binding by clustering of both receptor and ligand [15], recognition of thousands of sialyllactose containing gangliosides in the viral membrane by Siglec-1 should be clearly superior to the interaction of DC-SIGN with only 1467 envelope trimers per virion [24]. Siglec-1 is the only Siglec family member tested that mediated HIV-1 capture, although all Siglecs interact with sialic acid through their respective V-set domains. This could be caused by different specificities, but is most likely due to Siglec-1 containing the largest number of Ig-like C2-type domains of all Siglecs; these domains act as spacers separating the ligand-binding site from the cell surface. Therefore, Siglec-1 extends beyond the glycocalix of the cell, and is thus available for interaction with external ligands, while other family members mainly bind ligands in cis [15].
Although Siglec-1 expression is restricted to myeloid cells, particularly to tissue macrophages found in secondary lymphoid tissues [17,25], its expression can be rapidly induced and upregulated once myeloid cells are activated [26]. Indeed, DCs exhibit a characteristic mature phenotype in HIV-1 viremic patients [27], and up-regulation of Siglec-1 on mDCs is therefore likely to play an important role in HIV-1 dissemination in lymphoid tissues, thus contributing to HIV-1 disease progression. DC maturation is probably not directly induced by HIV-1 [28], but is more likely a consequence of factors released upon HIV-1 infection. Circulating LPS has been shown to be significantly augmented in HIV-1 patients due to the increased translocation of microbial products from the gastrointestinal lumen once infection is established [10]. Thus, LPS may facilitate HIV-1 progression by local and systemic stimulation of DCs, which (i) leads to Siglec-1 up-regulation and enhanced viral spread and (ii) creates the proinflammatory milieu associated with HIV-1 infection and immune activation.
This work together with several other recent reports indicates that HIV-1 uses a highly sophisticated strategy to evade DC immune surveillance and facilitate disease progression. Viral capture through Siglec-1 on the mDC surface is beneficial for viral spread through trans-infection, but could also be detrimental for the virus if leading to successful antigen presentation. However, captured HIV-1 do not appear to reach the endolysosomal compartment of LPS mDCs [29], where antigen processing occurs. Furthermore, interaction of HIV-1 with DC-SIGN can cause down-regulation of MHC class II molecules and interferon genes, impairing antiviral immune responses while triggering infectious synapse formation [30]. If productive fusion of the viral and cellular membrane occurs, HIV-1 replication is blocked by the myeloid-specific restriction factor SAMHD1 [1,2], thus preventing viral antigen production. On the other hand, if DC resistance to infection is bypassed, the interaction of newly synthesized HIV-1 proteins with a cell-intrinsic sensor elicits antiviral immune responses, not typically engaged owing to the absence of DC infection [31].
Siglec-1 captures HIV-1 through its interaction with sialyllactose-containing gangliosides exposed on viral membranes, and therefore functions as a general recognition receptor for vesicles carrying sialyllactose in their membrane. These include exosomes [16] and probably other sialyllactose-containing viruses. Gangliosides have been observed in the membrane of, for example, vesicular stomatitis virus (VSV), semliki forest virus, or murine leukemia virus [32,33], but are likely to be present in the membrane of many other enveloped viruses as well. Siglec-1dependent viral capture may be important for direct infection of DCs in these cases, but may also enhance immune recognition, thus benefiting the host. Accordingly, Siglec-1 has been shown to efficiently capture VSV in vivo and facilitate antiviral responses and prevent viral neuroinvasion [34,35]. The observation that Siglec-1 also captures cellular microvesicles suggests that this pathway normally leads to antigen presentation through exosomes [36] and has been hijacked by HIV-1 for infectious virus storage and spread. The discovery of the role of Siglec-1 in capturing sialylated viruses expands our understanding of HIV-1 transmission mechanisms and warrants novel therapeutic approaches aimed to prevent viral dissemination.

Ethics Statement
The institutional review board on biomedical research from Hospital Germans Trias i Pujol approved this study.

Primary Cell Cultures
Peripheral blood mononuclear cells (PBMCs) were obtained from HIV-1-seronegative donors, and monocyte populations or myeloid DCs were isolated and cultured as described in [9]. Monocyte-derived mature DCs were differentiated for 48 h with 100 ng/ml of LPS (Sigma-Aldrich) or ITIP (300 IU/ml IL-1b, 1,000 IU/ml IL-6, 1,000 IU/ml TNF-a, all from CellGenix, and 1 mg/ml PGE2 from Sigma-Aldrich). LPS myeloid DCs were differentiated for 24 h with 100 ng/ml of LPS. Autologous and heterologous CD4 + T cells were enriched from PBMCs using the RossetteSep a-CD8 + cocktail (Stem cell) and maintained in RPM1 with 10% fetal bovine serum (FBS, Invitrogen) supplemented with 10 IU/ml of IL-2 (Roche).

Transcriptome Analysis
DCs (3610 6 ) were centrifuged and resuspended in RNAlater solution (Ambion). After lysate homogenization using QIAshredder spin columns (Qiagen), total RNA isolation was performed with the RNeasy Mini Kit (Qiagen), including a 15-min DNAse I treatment step. Affymetrix GeneChip Human Gene 1.0 ST arrays were processed with R using affy and limma Bioconductor packages [37,38]. After robust multichip average and quantile normalisation, differential expression was computed using moderated paired t test. Adjusted p values were computed with the Benjamini & Hochberg method [39], and a 0.05 cutoff was applied to select significant genes.

Comparative Gene Expression Analysis by qRT-PCR
In total, 1 mg of RNA obtained as in the previous section was reverse transcribed using the TaqMan reverse transcription reagents (including multiscribe reverse transcriptase and random hexamers; Applied Biosystems). Predesigned TaqMan gene expression assays and the comparative Ct (DDCt) method [40] were used to determine relative SIGLEC1 gene expression. SIGLEC1 mRNA quantification (FAM dye-labeled probe) was normalized using the endogenous control gene Beta Glucuronidase (VIC/TAMRA dye labeled probe) in multiplex qPCR experiments run on the Applied Biosystems 7500/7500 Fast Real-Time PCR System and analyzed with the 7500 Software v2.0.4. A cDNA sample from PBMCs was used as a reference for all relative quantification results.

Siglec-1 Surface Expression Analysis by FACS
DCs were blocked with 1 mg/ml of human IgG (Baxter, Hyland Immuno) and stained with a-Siglec-1-PE 7-239 mAb (AbD Serotec) following the manufacturer's instructions at 4uC for 20 min. Samples were analyzed with FACSCalibur (Becton-Dickinson) using CellQuest and FlowJo software to evaluate collected data. Raji DC-SIGN B cell line (kindly provided by Y. van Kooyk) was maintained in RPMI with 1 mg/ml of G418 (Invitrogen). All media contained 10% FBS, 100 IU/ml of penicillin, and 100 mg/ ml of streptomycin (all from Invitrogen). VLP HIV-Gag-eGFP and VLP HIV-Gag-Cherry were obtained as previously described [11]. HIV NL4-3 was obtained following transfection of the molecular clone pNL4-3 (NIH AIDS Research and Reference Reagent Program from M. Martin). HIV NL4-3-Cherry was obtained following cotransfection of pCHIV and pCHIV mCherry in a 1:1 ratio [41]. HIV NL4-3 lacking the envelope glycoprotein was obtained as described elsewhere [9]. The p24 Gag content of the viral stocks and VLP was determined by ELISA (Perkin-Elmer) or by a quantitative Western blot [13]. HIV NL4-3 used in infectious assays was titrated employing the TZM-bl reporter cell line as described in [42].

Production of Liposomes and Exosomes
Large unilamellar vesicles (LUVs) were prepared as in [13], and exosomes were isolated from Jurkat cells as described in [11].
VLP, Liposome, Exosome, and HIV-1 Capture Assays LPS mDCs (2610 5 ) were pre-incubated at 16uC for 30 min with 10 mg/ml of a-Siglec-1 mAb 7D2 (HSn 7D2, Abcam), IgG1 isotype control mAb (107.3, BD Bioscience), a-Siglec-7 celladhesion neutralizing pAb (R&D Systems), a-Siglec-5/14 celladhesion neutralizing mAb (194128; R&D Systems, which recognizes both Siglec-5 and Siglec-14, sharing 99% of amino acid homology in the three extracellular distal domains) or a-CD83 mAb (HB15e; R&D Systems) or with 500 mg/ml of mannan from Saccharomyces cerevisiae (Sigma-Aldrich). Capture experiments were performed maintaining compound concentration and pulsing mDCs in parallel applying either 200 mM of the respective LUV HIV-tRed formulations or 150 ng of VLP HIV-Gag-eGFP Gag per 2610 5 cells for 30 min at 37uC. Exosome DiI capture was performed pulsing 1610 5 pretreated LPS mDCs with 150-250 mg of exosomes for 4 h at 37uC. After extensive washing, positive DCs were acquired by FACS. To test for potential crossreactivity of a-Siglec-1 mAb 7D2, 2.2 mM of the mAb were preincubated or not with more than 100-fold molar excess of recombinant human protein Siglec-1, and more than 200-fold molar excess of Siglec-7, Siglec-5/14, or CD83 (all from R&D Systems) 30 min at RT prior addition to the LPS mDCs. After incubation with mixes, LPS mDCs were pulsed with VLPs as indicated earlier. Fab fragments were generated from a-Siglec-1 7D2 and Isotype mAbs using the Fab Micro Preparation kit (Pierce) according to the manufacturer's instructions. Quality of Fab preparations was assessed with SDS-PAGE and Coomassie staining. Titration of a different a-Siglec-1 mAb was performed with functional grade clone 7-239 (AbD Serotec). DCs were also assessed for VLP capture for 1 h as described above but starting 5 d after isolation (when LPS was added to LPS mDCs) and continuing 6, 24, and 48 h after LPS addition. In parallel, DCs were labeled with a-Siglec-1-PE 7-239 mAb and a-HLA-DR-PerCP (clone L243, BD Biosciences). The mean number of Siglec-1 Ab binding sites per cell was obtained with a Quantibrite kit (Becton Dickinson) at day 7 as previously described for DC-SIGN [9].
HIV NL4-3 capture was assessed pre-incubating 2.5-3610 5 distinct monocyte-derived DCs or blood myeloid DCs at 16uC for 30 min with 10 mg/ml of the a-Siglec-1 mAb 7D2, the isotype control, or 500 mg/ml of mannan. Subsequently, DCs were pulsed with HIV NL4-3 at an MOI of 0.1 (50-80 ng of p24 Gag estimated by ELISA) for 5 h at 37uC. In parallel, untreated DCs equally pulsed with HIV NL4-3 were exposed to inhibitors right after viral capture. After extensive washing, cells were lysed with 0.5% Triton X-100 to measure p24 Gag antigen content by ELISA. HIV NL4-3 binding was performed pre-incubating LPS mDCs with the indicated mAbs, but maintaining cells at 4uC during viral pulse. Cells were lysed to detect p24 Gag or stained with Siglec-1-Alexa 488 7-239 mAb (Ab Serotec) to confirm arrested endocytosis of Siglec-1 at 4uC as compared to cells exposed to the virus at 37uC by FACS.

Trans-Infection Assays
DCs were treated and pulsed with HIV NL4-3 as described above. After extensive washing, cells were co-cultured with the TZM-bl CD4 + target cell line to measure trans-infection. Pulsed monocytederived DCs or myeloid DCs were co-cultured in quadruplicate or duplicate at a ratio of 1:1 or 5:1, respectively. Cells were assayed for luciferase activity 48 h later (BrightGlo Luciferase System; Promega) in a Fluoroskan Ascent FL luminometer (Thermo Labsystems). Background values consisting of non-HIV-1-pulsed co-cultures or reporter CD4 + cells alone were subtracted for each sample. To detect possible productive infection of pulsed cells or re-infection events, some DCs were cocultured in the presence of 0.5 mM of the protease inhibitor Saquinavir.

Transduction of DCs
VSV-G-Pseudotyped SIV3 lentivector (kindly provided by A. Cimarelli) was produced as in [43]. Isolated monocytes (5610 5 ) were infected with SIV3 particles and transduced with two different SIGLEC1-specific or one nontarget shRNA control MISSION Lentiviral Transduction Particles (Sigma-Aldrich) at an MOI = 50. Transduced monocytes were differentiated into LPS mDCs and assessed for VLP capture and HIV-1 trans-infection as described above. Adequate phenotypic maturation of DCs was evaluated as in [9]. Lentiviral transduction particles carrying the GFP reporter gene cloned in the same pLKO.1-puro vector backbone (MISSION TurboGFP Control Transduction Particles) were used to evaluate transduction efficiency by FACS (estimated 75%-98% at day 7, when cells were employed).

Statistical Analysis
Statistics were performed using paired t test (considered significant at p#0.01) or Spearman correlation with GraphPad Prism v.5 software.