Ethics statement

Our research complies with all relevant ethical regulations in accordance with the ethics committee of the Medical Faculty of Heidelberg University (S024/2022). Human peripheral blood was received from anonymous donors without any information on sex and age. Gender was not determined since sex-specific differences were not the focus of this study.

Cells

293T cells (ATCC, CRL-3216) and TZM-bl reporter cells (NIH AIDS Reagent Program, courtesy of Dr. John C. Kappes, Dr. Xiaoyun Wu and Tranzyme Inc. (ARP-8129)) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco) supplemented with 10% heat-inactivated fetal calf serum (FCS, Capricorn) and 1% penicillin/streptomycin (Gibco). Huh 7.5 cells were maintained in complete DMEM medium supplemented with non-essential amino acids.

Primary CD4 + T cells and MDMs

Human peripheral blood of healthy, HIV-negative donors was obtained from the blood bank HD, according to regulation by local ethics committee (S024/2022). CD4 T cells were isolated from human peripheral blood of healthy, HIV-negative donors using the RosetteSep Human CD4 T cell enrichment kit (15062, StemCell Technologies) according to the manufacturer’s protocol. The cells were then used a 3×3 activation as described previously⁷ for 72 h and cultured in Roswell Park Memorial Institute Medium (RPMI, Gibco) supplemented with 10% heat-inactivated FCS, 1% penicillin–streptomycin, and 10 ng/ml interleukin 2 (IL-2, Biomol). For Monocyte-derived macrophages (MDMs), peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats by Biocoll (Merck Biochrom) density gradient centrifugation. CD14⁺ monocytes were then isolated from PBMCs by positive selection using magnetic beads (CD14 MicroBeads; 130-050201, Miltenyi Biotech) and an AutoMACS Pro Separator (Miltenyi Biotech). 1 × 10⁵ monocytes per well were then seeded in glass-bottom 96 well plates that were previously coated with fibronectin (2 µg/cm²) and maintained at 37 °C with 5% CO₂ in complete RPMI in the presence of 5% human AB serum (H4522, Sigma-Aldrich) for differentiation into MDMs for 10–14 days⁴⁸.

Viruses

Virus stocks of replication competent HIV-1 and HIV-2 strains (pNL4.3 WT, pNL4.3 ΔNef, pNL4.3 R5, ADA, CH077.t, CH058.c, CH0198, RHGA, CH0167, CH0293, HIV-2 Rod9-GFP) were generated by transfecting 293 T cells (sub-confluent 15 cm² dishes) with 25 µg of proviral constructs alongside 75 µl of linear polyethyleneimine (PEI, Sigma Aldrich) in Optimem medium (Gibco). Viruses containing Vpr-BlaM, Vpr.mRuby2 or Vpr.Int.GFP fusion proteins were similarly produced by PEI co-transfection of 293 T cells using 25 µg of pNL4.3 provirus with 7.5 µg of Vpr fusion constructs. For single-round HIV-1 virus production (pNL4.3ΔEnv VSVg, pNL4.3 ΔEnv HIV-1 Env), 22 µg of pNL4.3 ΔEnv were used, complemented with 3 µg of the respective glycoprotein expression vectors^43,48. Two to three days post-transfection, supernatants were harvested, filtered (0.45 µm) and ultracentrifuged through a 20% (w/v) sucrose cushion. Virus pellets were then resuspended in sterile-filtered PBS 0.1% BSA, aliquoted and stored at −80 °C. All replication-competent virions were handled in a BSL-3 containment laboratory.

Lentiviral pseudotyping

Lentiviral stocks were prepared by PEI transfection of sub-confluent 293 T cultures. Briefly, 22.5 µg of pWXPL-GFP lentiviral backbone was co-transfected alongside 2.3 µg of pAdvantage, 15 µg packaging vector psPax2 and 8 µg of envelope glycoprotein encoding plasmids (VSVg, pcDNA Con1, pcDNA JFH1, NL4.3 Env) or pCDNA3.1+ for ΔEnv particles. Lentiviral vectors carrying Vpx_mac239 were produced in 293 T cells by co-transfection of pWPI, pcDNA.Vpxmac239, pΔR8.9 NSDP, and VSV-G at a molar ratio of 4:1:3:1⁴⁸. Cell culture supernatants were harvested and filtered 3 days post-transfection, and further concentrated through a 20% (w/v) sucrose cushion ultracentrifugation. Aliquots were stored at -80 °C. Virus titers were determined by SG-PERT analysis⁹².

Plasmids

The proviral plasmids pNL4.3 WT and ΔNef have been previously described⁹³. The pNL4.3-R5 WT & ΔNef, containing 7 point mutations in NL4.3 Env were previously described^46,48. The pNL4.3 ΔEnv, and HIV-2 Rod9-GFP proviruses were previously described⁹⁴, provirus encoding the CCR5 tropic ADA provirus were obtained via the NIH AIDS reagent program (ARP-416)⁹⁵. All Transmitted/Founder strains were obtained via the NIH AIDS reagent program (pCH077.t (ARP-11742)⁹⁶ pCH058.c (ARP-11856)⁹⁶, pCH198⁹⁷, and chronic strains (pSTCO⁹⁸, pRHGA(ARP-12421)⁹⁸, pCH167(ARP-13544), pCH293(ARP13539)⁹⁷). The following expression plasmids were used: pMM311 (Vpr.BlaM)⁴⁴, the pmRuby2.Vpr vector was previously described⁹⁹ and kindly provided by Dr. Tom Hope. The Vpr.Int.GFP expression plasmid was kindly provided by Anna Cereseto. The pcDNA3.1 Vpx SIVmac239-Myc and the pΔR8.9 packaging vector harboring a Vpx binding motif were previously described⁹⁴.

Reagents and antibodies

The following dyes were used: Concanavalin A-488 (C7642, Sigma Aldrich) was used a concentration of 50 µ/ml to stain TZM-bl cell membrane for 10 min at RT in the dark, Alexa Fluor 488 NHS Ester (A20000, Invitrogen) was used to fluorescently label loose collagen gels as described below, Fluoromount DAPI (00-4959-52, Invitrogen) was used to mount coverslips and stain cell nuclei, the CCF2-AM dye (K1023, Thermo Scientific) was used to perform HIV-1 entry assays, and WGA-647 (W32466, Invitrogen) was used according to manufacturer protocol to label MDM plasma membranes. The following reagents were used: PEI (Sigma Aldrich), Pur-A-Lyzer™ Mega 3500 kit (PURG35010, Sigma Aldrich), recombinant human TLR2 (2616-TR-050, R&D systems), and RNEasy micro kit (74004, Qiagen).

The following antibodies were used: Zombie Violet dye (Biolegend), anti-p24 KC-57 antibody (Beckman Coulter), rabbit anti-HIV-1 p24 (Kindly provided by Prof. Dr. Barbara Müller), rabbit anti-HIV-1 gp120 (Valerie Bosch, DKFZ, Heidelberg), mouse anti-TLR8 (67317, Proteintech), mouse anti-EEA1 (68065, Proteintech), mouse anti-CD63 (556019, BD), sheep anti-TGN46 (AHP500GT, BioRad), mouse anti-GPP130 (923801, Biolegend), mouse antiLAMP1 (ab25630, Abcam), mouse anti-TfR (13-6800, Invitrogen), anti-HIV-1 gp120 used to probe Env epitope accessibility after collagen priming (PG16: ARP-12150; 3BNC117: ARP-12474; 10-1074: ARP-12477; NIH45-46: ARP-12174; 17b: ARP-4091) and sCD4-PE (CD4HP2H8, ACRO Biosystems). Secondary antibodies used were anti-mouse, anti-sheep or anti-human IgG1 AlexaFluor-568 antibodies (Invitrogen), as well as goat anti-rabbit IRDye700/800 conjugated antibodies (926-32211, Rockland), and goat anti-rabbit IgG-HRP (31460, Invitrogen).

Virus titer determination

One step Syber Green-based Product Enhanced Reverse Transcriptase assay (SG-PERT) was used to assess HIV-1 virus titers as described previously¹⁰⁰. Briefly, concentrated virus stocks were first diluted in PBS, or culture supernatants were directly lysed in 2x lysis buffer (50 mM KCL, 100 mM Tris-HCl pH 7.4, 40% glycerol, 0.25% Triton X-100) supplemented with 40 mU/µl Rnase inhibitor for 10 min at room temperature. Lysed samples were then exported from BSL-3 and further diluted 1:10 in dilution buffer (5 mM (NH₄)₂SO_4, 20 mM KCL, 20 mM Tris-HCl pH 8). In parallel, 10 µl per well of a dilution series of a virus standard (pCHIV, 8.09 × 10⁸ pURT/µl) were also lysed for 10 min. All lysed samples were then incubated with 10 µl of 2x reaction buffer (1x dilution buffer, 10 mM MgCl₂. 2x BSA, 400 µM each dATP, dTTP, dCTP, dGTP, 1pmol of each RT forward and reverse primers, 8 ng MS2 RNA, SYBR Green 1:10,000) supplemented with 0.5U of GoTaq HotStart Polymerase. RT-PCR reactions were carried out and read in a real-time PCR detector (CFX 96, Biorad) using the following program: (1) 42 °C for 20 min, (2) 95 °C for 2 min, (3) 95 °C for 5 s, (4) 60 °C for 5 s, (5) 72 °C for 15 s, 80 °C for 7 s, repeat steps 3–6 for 40 cycles with a melting curve read out as a final step. Sequence of the used primers are as follows: fwd RT primer (TCCTGCTCAACTTCCTGTCGAG) and rev RT primer (CACAGGTCAAACCTCCTAGGAATG).

Virus infectivity determination

To assess the infectivity of virus stocks, TZM-bl reporter cells were infected as previously described in the absence of DEAE dextran²⁸. In brief, TZMbl cells stably expressing HIV-1 entry receptors and containing luciferase and β-galactosidase genes under the control of the HIV-1 LTR promoter were infected using dilution series of concentrated virus in triplicates. 72 h post-infection, cells were fixed in 3% PFA, and incubated with a substrate solution (β-Gal supplemented with 200 µg/ml X-Gal) for 3 h at 37 °C. Blue cells were counted to determine infectious virus titers in the form of Blue Cell Units (BCUs). For the determination of the infectivity of virions after culture in suspension or in collagen, the amount of infectious units on TZM-bl cells was determined for each virus stock. The volume containing 10⁵ BCUs was then used for incubation in suspension or embedding in collagen. Supernatants of these cultures were then analyzed by the SG-PERT assay to determine their relative RT activity. Equivalent amounts of RT units were then used to infect TZM-bl cells for assessment of the impact of the culture condition on the relative infectivity of virus particles. TZM-bl cells were infected in triplicate and lysed 3 days post-infection using 1x lysis buffer (Promega) for 10 min. Lysates were then incubated with Luciferase substrate (Promega), and luciferase activity was measured for 5 s at a Tecan Infinity luminometer.

For saturation experiments, 10⁵, 10⁶, or 10⁷ BCUs were used for seeding or embedding in collagen matrices of equivalent volumes. The culture supernatants were then processed by SGPERT. TZM-bl cells were infected using equivalent amounts of RT units for all conditions.

Virus infectivity after seeding or embedding at different concentrations was then determined by measuring the average luciferase activity from cell lysates 3 days post-infection after subtraction of the background luciferase signal from uninfected cell lysates. In additional experiments, HIV-1 NL4.3 virus was incubated on top of 2D collagen matrices or in suspension in presence or absence of increasing amounts of lentiviral particles carrying Env (Pax2+Env particles) as determined by p24 ELISA for 16 h. TZM-bl cells were then infected with equivalent amounts of RT units for 48 h prior to lysis and determination of infectivity by luciferase assay.

The relative infectivity of the respective supernatants was calculated as the average luciferase activity divided by the average amount of RT units measured from the culture supernatant.

Virus embedding in 3D matrices

Type I collagen matrices of different densities were polymerized as previously described⁷. In brief, dense collagen gels (3 mg/ml) were generated by mixing 10X MEM medium with 7.5% NaHCO₃ (both Gibco) and highly concentrated rat tail collagen I (354236, Corning) at a 1:1:8 ratio on ice. Concentrated virus in complete DMEM was then added to the neutralized and chilled collagen solution at a 1:1 ratio (see “Virus infectivity determination” for determination of amount of virus used for embedding). 100 µl of virus-containing collagen solutions were then distributed in each well in a 96 well-plate and allowed to polymerize within 15 min at 37 °C. Loose collagen gels (1.7 mg/ml) were generated by mixing 10X MEM medium with 7.5% NaHCO₃ (both Gibco) and PureCol bovine skin collagen I (5005, Advanced Biomatrix) at a 1:1:8 ratio on ice. Concentrated virus was then added to the neutralized and chilled collagen solution at a 1:1 ratio and and the mixture was allowed to pre-polymerize for 10 min at 37 °C.

100 µl per well was then transferred to a 96-well plate. Gels were allowed to polymerize within 45 min at 37 °C.

Human lyophilized placental type III collagen (5019, Advance Biomatrix) was reconstituted using a chilled 2 mM CH₃COOH solution on ice to reach 6 mg/ml. Type III collagen matrices (3 mg/ml) were generated by mixing 10X MEM with 7.5% NaHCO₃ (both Gibco) with the reconstituted type III collagen solution at a 1:1:8 ratio on ice. Concentrated virus in complete DMEM was then added to the neutralized and chilled collagen solution at a 1:1 ratio. Gels were allowed to polymerize within 30 min at 37 °C.

Matrigel Growth Factor Reduced Basement Membrane (354230, Corning) gels (5 mg/ml) were generated according to manufacturer protocol. In brief, concentrated Matrigel aliquots (8.9 mg/ml) were thawed at 4 °C overnight. Thawed Matrigel was then combined with complete DMEM containing concentrated virus at a 3:2 ratio on ice. 100 µl of mixture was transferred to a 96-well plate, gels were allowed to polymerize within 45 min at 37 °C.

Agarose gels were prepared by preparing a 0.8% agarose solution in PBS. The dissolved agarose solution was combined 1:1 with complete DMEM containing concentrated viruses. 100 µl of the mixture was transferred to a 96-well plate, gels were allowed to polymerize at 4 °C within 15 min.

All polymerized gels were then overlaid with 100 µl pre-warmed complete DMEM and incubated at 37 °C. Culture supernatants were harvested at the indicated time points and processed for relative infectivity determination.

Western blotting

For detection of virion associated gp120, sTLR2 and p24 from supernatants of suspension and collagen cultures, supernatants were concentrated by ultracentrifugation through a 20% sucrose cushion (TLA-100 rotor, 108,000 g, 45 min) using an Optima LE-80k ultracentrifuge (Beckman Coulter), and lysed in 2× SDS sample buffer (10% glycerol, 6% SDS, 130 mM Tris Hcl pH 6.8, 10% β-Mercaptoethanol) and boiled for 5 min at 95 °C. Protein samples were resolved with SDS-PAGE electrophoresis and blotted onto nitrocellulose membranes. Membranes were blocked in 4% milk/TBST for 1 h and were incubated with primary antibodies overnight. We used secondary antibodies conjugated to IRDye700/800 (1:20,000, Rockland) for fluorescent detection with Licor (Odyssey).

Fluorescent collagen gels

Fluorescently labeled LC gels were generated as described previously³⁸. In brief, 5 mg of Alexa Fluor 647 NHS Ester dye (A20006, Invitrogen) was dissolved in 0.5 ml DMSO. The dissolved dye was then combined with PureCol bovine skin collagen (3 mg/ml) at a 1:100 ratio on ice and stirred overnight at 4 °C. After labeling, the excess dye was removed by dialysis using a 3.5 kDa molecular weight cut off Pur-A-Lyzer™ Mega 3500 kit (PURG35010, Sigma-Aldrich), placed in 1 L of acetic acid solution (0.02 N, pH 3.9) and stirred for 1 week at 4 °C. The resulting collagen solution was kept at 4 °C until further use. Fluorescently labeled collagen gels were polymerized as described above, by combining labeled and unlabeled collagen solutions at a 1:10 ratio.

Confocal reflection microscopy

Collagen gels were imaged by confocal reflection microscopy as previously described^7,101. Briefly, different collagen gels were generated in 15-well angiogenesis µ-slides (Ibidi) as described above. Point laser scanning confocal microscopy was then performed on a Leica SP8 microscope using an HC PL APO CS2 63×/1.4 N.A. oil immersion objective. Images were acquired using PMT detectors in reflection mode with a laser excitation at 567 nm, and a spectral detection window set between 550 nm and 570 nm wavelengths. Fluorescently stained collagen matrices were additionally imaged using a 488 nm laser.

Widefield imaging

Viruses were embedded in stained or unstained LC matrices as described above. The culture supernatants were then harvested and ultracentrifuged through a 20% (w/v) sucrose cushion. Virus pellets were resuspended in 3% PFA for 90 min. The fixed virus particles were then seeded on 0.01% poly-L-lysine coated coverslips for 30 min and mounted on microscopy slides with Fluoromount DAPI (Invitrogen). Samples were then imaged using an epifluorescence microscope (Olympus IX81 S1F-3) under a 100× oil objective (PlanApo, N.a. 1.40). Quantification of double-positive Vpr-mRuby2 and Alexa Fluor 647 signals was performed using the Spot Detector plugin of the Icy image analysis software.

TIRF microscopy

To exclude the possibility that collagen fragments remain attached at the surface of virions after contact with collagen fibers, HIV-1 NL4.3 R5 Vpr.Int.GFP virions were embedded in fluorescently labeled collagen for 16 h. The supernatants were then harvested and transferred to 0.01% poly-L-lysine coated µ-Slide 8 Well (80821, Ibidi). TIRF microscopy of the samples was performed on a Zeiss Axio Observer microscopy stand using an alpha PlanAPO 100×/ 1.46 N.A. oil objective and a Teledyne Prime-BSI sCMOS camera (6.5 µm × 6.µm pixel size). The sample was excited with a 561 nm or a 488 nm Visitron VS-laser control (Visitron Systems GmbH) and emission light was collected using Chroma TIRF Quad Line Beamsplitter with Quad Line Rejectionband 405/488/561/640 or Laser Beamsplitter zt 488 with 525/50 band-pass filter, respectively. The TIRF illumination was realized by a full 360-degree laser beam rotation at the back focal plane using 2D galvo scanners (Ring-TIRF). The evanescent field penetration depth was adjusted to 200 nm. Laser intensities, shutter, and camera were controlled using the VisiView program (Visitron Systems GmbH). The presence of AlexaFluor 647 labeled collagen fibers was assessed at the surface of Vpr.Int.GFP spots using the Icy software.

Harvesting of collagen fibers for cryo-ET analysis

To assess the morphology of single collagen fibers by cryo ET, DC or LC gels were prepared as previously in 1.5 ml Eppendorf tubes. Polymerized gels were disrupted by sonication on a Sonorex super RK 102 H sonicator for 10 s on ice. Disrupted individual fibers were resuspended in PBS and further processed for cryo-ET.

Plunge freezing

Collagen fibers and supernatant used for plunge freezing were collected as described above. Holey carbon grids (Cu 200 mesh, R2/1, Quantifoil®) were plasma-cleaned for 10 s in a Gatan Solarus 950 (Gatan). Samples were mixed with 10× concentrated 10 nm protein A gold (Aurion) prior plunge freezing. A total volume of 3 µl was used for plunge freezing into liquid ethane using an automatic plunge freezer EM GP2 (Leica). The ethane temperature was set to −183 °C and the chamber to 24 °C with 80% humidity. Grids were blotted from the back with Whatman™ Type 1 paper for 3 s. Grids were clipped into AutoGrids™ (Thermo Fisher Scientific).

Cryo-electron tomography and tomogram reconstruction

Cryo-electron tomography was performed using a Krios cryo-TEM (Thermo Fisher Scientific) operated at 300 keV and equipped with a post-column BioQuantum Gatan Imaging energy filter (Gatan) and K3 direct electron detector (Gatan) with an energy slit set to 15 eV. As a first step, positions on the grid were mapped at 8700× (pixel spacing of 10.64 Å) using a defocus of approximately -65 µm in SerialEM¹⁰² to localize collagen fibers or HIV particles. Tilt series were acquired using a dose symmetric tilting scheme¹⁰³ with a nominal tilt range of 60° to −60° with 3° increments with SerialEM. Tilt series were acquired at target focus −4 μm, with an electron dose per record of 3 e⁻/Ų and a magnification of 33,000× (pixel spacing of 2.671 Å). Beam-induced sample motion and drift were corrected using MotionCor2¹⁰⁴. Tilt series were aligned using AreTomo¹⁰⁵ and tomograms were reconstructed using R-weighted back projection algorithm with doseweighting filter and SIRT-like filter 5 in IMOD¹⁰⁶. Tomograms were used to measure the diameter of HIV particles in IMOD. For visualization, tilt series were aligned using protein A gold as fiducials in IMOD. Tomograms in Fig. 2 were reconstructed using R-weighted back projection algorithm with 3DCTF, dose-weighting filter and SIRT-like filter 10. In IMOD, 15 slices of the final tomogram were averaged and Fourier filtered. The diameter of HIV particles was measured from the outer leaflet of the viral membrane in IMOD. The average diameter of two measurements per particle was used in the graph.

Infectivity enhancement experiments

HIV-1 virions were cultured in suspension or in collagen for 16 h. After performing an SG-PERT analysis from the culture supernatants, equivalent amounts of reverse transcriptase units were then incubated with the specified infectivity enhancers for 20 min at 37 °C prior to infection of TZM-bl cells. Concentrations of infectivity enhancers used: EF-C (15 µg/ml)⁴⁰ or RM-8 (15 µg/ml)⁴¹. The relative infectivity of the treated virions was assessed as previously described.

Binding assay

NL4.3 Vpr mRuby2 virions were cultured in suspension or in collagen for 16 h. Viral titers in the supernatants were quantified by SG-PERT. Equivalent amounts of RT units were then used to infect TZM-bl cells seeded on glass coverslips in 24-well plates 24 h prior to infection. After 2 h at 4 °C to prevent virion internalization, the cells were then washed with PBS, and the plasma membrane was stained with Concanavalin A Alexa Fluor-488 (Invitrogen) according to manufacturer’s protocol. The samples were then fixed using 3% PFA for 90 min, and mounted on microscopy slides using Fluoromount-DAPI (Invitrogen). Spinning disk confocal microscopy was performed on a PerkinElmer UltraVIEW VoX microscope equipped with a Yokogawa CSU-X1 spinning disk head and a Nikon TiE microscope body. An Apo TIRF 60×/1.49 N.A. oil immersion objective and a Hamamatsu C9100-23B EM-CCD camera were used. Images were acquired using solid state lasers with excitation at 405 nm, 488 nm and 561 nm with matching emission filters. Z-stacks were acquired with a z-spacing of 0.5 µM steps. The percentage of cells with bound virus was then analyzed using the Imaris software (Oxford Instruments), in which Z-stacks were reconstructed in 3D. Cell membranes and virions were segmented as individual surfaces, and statistic values (surface volumes, relative distance of surfaces) were retrieved from the software.

Vpr.BlaM entry assay

TZM-bl reporter cells were infected with an MOI of 0.1 using HIV-1

NL4.3 viruses containing Vpr-BlaM after treatment in suspension or collagen for 16 h. A concentrated virus that was not seeded in medium prior to infection was used as positive control, virus particles lacking Env were used as negative control. Fusion of HIV-1 particles was allowed to proceed for 4 h at 37 °C. Cells were then washed twice in PBS and stained with 2 mM CCF2-AM dye (Invitrogen) supplemented with 2.5 mM Probenecid in Fluorobrite DMEM 2% FCS for 6 h at 11 °C to prevent particle fusion during the staining process according to manufacturer instructions. The same workflow was used to assess fusion of S or DC virions after incubation with or without EF-C as previously described. Where indicated, T20 was added to block fusion. Cells were then trypsinized fixed in 3% PFA in PBS at 4 °C overnight and analyzed by FACS (see Sup. Figure. 9a for gating strategy).

Primary CD4 + T cell infection

Activated primary CD4 + T cells were infected in 96-well plates in triplicate with suspension and collagen culture supernatants containing NL4.3 R5 as previously described⁷. Briefly, equivalent RT units were used between the different conditions, as quantified by SG-PERT. The amounts of RT units used between experiments were defined as previously, and amounted to 3.25 × 10¹⁰ + /- 1.8×10¹⁰ pURT per well. The cells were spininfected at 600 g for 90 min in the presence or absence of 3 µg/ml reverse transcriptase inhibitor Efavirenz (EFZ), then cultured at 37 °C for 3 days. The samples were then stained with a fixable Zombie Violet dye (Biolegend), fixed in 3% PFA for 90 min, then stained with an anti-p24 KC57 FITC antibody (Beckman Coulter) in 0.1% Triton 100-X for 30 min at 4 °C according to manufacturer protocol. Samples were then measured by flow cytometry (see Sup. Fig. 9b for gating strategy).

Primary MDM infection

Differentiated macrophages were spin transduced at 37 °C in a preheated centrifuge with lentiviral vectors containing Vpx_mac239 for 1 h at 20 g. Within 16 h post-transduction, cells were infected in triplicates with equivalent amounts of RT units as measured by SG-PERT from suspension and collagen culture supernatants in a 96-well plate format, in the presence or absence of EFZ. Virus input for infection was determined as for CD4 + T cells. MDMs were also treated with 50 ng/ml LPS as positive control. Infected MDM culture supernatants were harvested 3 days post-infection and further processed for cytokine analysis.

5 days post-infection, cells were harvested by trypsinization, stained with a fixable Zombie

Violet dye (Biolegend), fixed in 3% PFA for 90 min, then stained with an anti-p24 KC-57 FITC antibody (Beckman Coulter) in 0.1% Triton 100-X for 30 min at 4 °C according to manufacturer protocol. Infection rates were then determined by flow cytometry (see Sup. Fig. 9b for gating strategy).

Cytokine quantification

The amounts of cytokines and chemokines present in cell culture supernatants were determined by Eve Technologies Corporation using the Discovery Assay®: Human Cytokine Array/Chemokine Array 48-Plex. Samples were exported out of BSL-3 containment after treatment with 0.5% Triton 100-X for 30 min prior to shipping. Results are in pg/ml of cytokines/chemokines according to the company protein standard. Cell-free supernatants were also analyzed for levels of Il-6, Il-8 and TNF by enzyme-linked immunosorbent assay (OptEIA ELISA kits; BD Biosciences) according to manufacturer’s instructions.

Endotoxin level determination

To exclude potential endotoxin contamination of our agarose, collagen and matrigel stocks (batches from the same lot were used throughout the study), supernatants harvested from agarose, type I or III collagens, as well as matrigel matrices (polymerized in the absence of virus) were used for endotoxin testing. Endotoxin levels were quantified using a Pierce Chromogenic Endotoxin Quant Kit (ThermoScientific, A39553) according to manufacturer protocol using the low standard procedure.

Modeling

Mathematical modeling has been used extensively to reveal and quantify the dynamics of viral infections by connecting experimental data and mechanistic descriptions of the assumed processes. The mathematical model that we use here has been developed previously to estimate the contribution of cell-free and cell-to-cell transmission to HIV-1 spread given different environmental conditions and has been explained in detail within⁷. In brief, the model describes the turnover and dynamics of (un-)infected CD4 + T cells, CD8 + T cells and the viral load within different culture conditions by systems of ordinary differential equations.

Uninfected CD4 + T cells proliferate and die with specific proliferation, λ, and death rates, δ, respectively, with their proliferation capacity affected by the simultaneously proliferation of CD8 + T cells accounting for resource competition. To account for the various transmission modes, CD4 + T cells can get infected by either cell-free transmission, proportional to the viral load and regulated by the cell-free transmission rate β_f, or direct cell-to-cell transmission, proportional to the number of neighboring infected cells and dependent on the cell-free transmission rate β_c. The complete set of mathematical equations and detailed descriptions, as well as pre-defined and obtained parameter values used within the analyses, are given within⁷. In the original publication, the model was fitted simultaneously to the data of co-transfer experiments of infected and uninfected CD4 + T cells into 2D suspension, and 3D loose and dense collagen environments, to estimate the parameters controlling viral infection, i.e., β_f and β_c, and infer the relative contribution of both transmission modes to overall infection. Thereby, a single transmission rate for cell-free infection was assumed across all environments, with environmental restriction reducing the infectivity of cell-free virions within collagen, i.e., the transmission parameter β_f, to only 14% of the effectivity considered within suspension, i.e., β_f,loose = β_f,dense = ηβ_f,sus with η = 0.14. In this study, we re-performed the analysis done within⁷ by varying the factor accounting for reduced infectivity in collagen η between 0 and 1 within steps of 0.1, and also considering η = 0.275, as experimentally determined for primary target cells. Fitting was performed as described within⁷ using the optim-function within the R language of statistical computing. Posterior distributions of parameter estimates were obtained by performing ensemble fits for each value of η based on different starting values for the unknown parameters. By using a stepwise approach incorporating subsequent filtering steps that excluded unreasonable model predictions in terms of CD4 + T cell counts and viral loads (see Sup. Fig. 2b, c), we ensured convergence of parameter estimates, with posterior distributions for parameter estimates relying on ~110-145 successful fits for each value of η.

PRR inhibitor treatments

To determine the sensing pathway involved in the collagen-mediated sensitization of virus particles for innate immune recognition, MDMs were pretreated with the different inhibitors prior to infection. 5 µM of cGAS inhibitor (G140, Invivogen), 8 µM of TLR 1/2 inhibitor (Cu-CPT22, Selleckchem), 49 µM of TLRs 4& 2/6 inhibitor (GIT27, Tocris), 2 µM of TLR4 inhibitor (CLI-095, Invivogen), 100 µM of TLR2 inhibitor (TL2-C29, Invivogen) or 10 µM of TLR8 inhibitor (Cu-CPT9a, Invivogen) were incubated with MDMs 3 h prior to infection. The MyD88 inhibitor (Pepinh-MYD, Invivogen) was used at 20 µM and incubated with MDMs 4 h prior to infection. MDMs were also pretreated for 1 h with a TLR3/dsRNA complex inhibitor (Merck Millipore) 5 µM final concentration. MDMs were also treated with 100 µM of T20 HIV-1 fusion inhibitor (Roche) or 3 µg/ml Efavirenz (SML0536, Sigma). The different inhibitors were supplemented again during virus inoculation. The cells were then cultured for 3 days at 37 °C, and supernatants were harvested for cytokine analysis.

RNA-seq of MDMs challenged with suspension or collagen-primed HIV-1 virions

MDMs from three separate donors were challenged with equivalent RT units of HIV-1 NL4.3 R5 virions harvested from suspension or collagen cultures incubated for 16 h at 37 °C for 6 h or 24 h prior to RNA isolation (RNeasy Micro kit, Qiagen). As controls, untreated cells, and cells treated with LPS (50 ng/ml) were cultured for 24 h at 37 °C prior to RNA isolation. RNA samples were sent to BMKgene (Beijing, China). After verification of RNA purity, concentration and integrity using NanoDrop, Qubit 2.0 and Agilent 2100, qualified RNA was processed for library construction. The library was then sequenced by Illumina NovaSeq X high throughput sequencing. HISAT2 software was used to map clean reads to the reference genome (Homo_sapiens.GRCh38_release95.genome.fa.). StringTie contrast pairs were then used to assemble the mapped reads for subsequent analysis. Differential gene expression between samples was analyzed by DESeq2 using thresholding criteria of fold changes (FC) ≥ 1.5 and a p-value 107. In brief, a ranked list of differentially expressed genes for each comparison was used to perform a GSEA analysis using the MSigDB, with a specific analysis using the GO Biological Processes (C5) database⁹². From this analysis, normalized enrichment scores were calculated for each GO term, as well as the Benjamin-Hochberg false discovery rate (FDR). Raw data can be accessed from the Gene Expression Omnibus (GEO) database under the accession number GSE307977.

HIV-1 colocalization with TLR and sub-cellular compartments

MDMs were challenged with NL4.3 R5 Vpr.Int.GFP virions for 4 h at 37 °C. The plasma membrane of MDMs was then stained using WGA-647 (Invitrogen) for 10 min at 4 °C to avoid internalization of the dye, and samples were then fixed using 3% PFA for 90 min. After export out of BSL3 containment, the cells were permeabilized using 0.1% Triton 100X for 10 min at room temperature, incubated for 30 min with blocking buffer (PBS, 5% BSA) and then with primary antibodies for 45 min at room temperature. After incubation with appropriate secondary antibodies for 30 min at room temperature, the coverslips were mounted onto microscopy slides using Fluoromount-DAPI (Invitrogen). The samples were then imaged using a Leica SP8 point laser scanning confocal microscope using an HC PL APO CS2 63×/1.4 N.A. oil immersion objective. Multichannel images were acquired sequentially using HyD detectors or PMTs in the lightning mode. Z-stacks were acquired with 500 nm steps. Deconvoluted images were then used to segment virions and TLR/compartments using Imaris (Oxford instruments) while enabling object-object statistics. From the resulting data, the percentage of virus segments overlapping with TLR/compartment segments were calculated, using a threshold of 0.01 µm³ overlap to define colocalization.

Probing HIV-1 Env epitope accessibility

To assess the overall conformation of Env at the surface of virions, we adapted a workflow from Staropoli et al. (2019). In brief, HIV-1 virions from various strain containing Vpr.Int.GFP fusion proteins were cultured in suspension or in collagen matrices of different densities for 16 h. The supernatants were then harvested, and the viral titers were determined by SG-PERT analysis. Equivalent amounts of pURT were then purified by ultracentrifugation over a sucrose cushion using a tabletop ultracentrifuge (TL100, Beckman Coulter). The purified virus pellet was then resuspended in filtered PBS (0.1 µm) and distributed to a 96 well-plate at 1 × 10¹¹ pURT/well. The virions were then incubated with the different antibodies (anti-HIV-1 gp120 used to probe Env epitope accessibility after collagen priming (PG16: ARP-12150; 3BNC117: ARP-12474; 10-1074: ARP-12477; NIH45-46: ARP12174; 17b: ARP-4091)) (final concentration: 15 µg/ml) or sCD4-PE (1:50 dilution as per manufacturer instructions) for 1 h at 4 °C in staining buffer (RMPI with 20% FCS filtered through a 0.1 µm filter). After the first incubation with broadly or non-neutralizing antibodies, the samples were then incubated with a secondary, anti-human IgG1(H + L) AlexaFluor 568 antibody (1:200 dilution; Life Technologies, in staining buffer) for 1 h at 4 °C. The samples were then fixed using 32% PFA (Thermo Sci) to reach a final concentration of 4%. After export out of BSL-3 containment, the samples were then seeded on poly-L-lysine-coated coverslips for 1 h, then mounted on microscopy slides. Samples were then imaged using an epifluorescence microscope (Olympus IX81 S1F-3) under a 100X oil objective (PlanApo, N.a. 1.40). Quantification of double-positive Vpr-mRuby2 and Alexa Fluor 568 signals was performed using the Spot Detector plugin of the Icy image analysis software. For each primary/secondary antibody pair or sCD4-PE, appropriate signal thresholds were arbitrarily defined to exclude background signals. Background fluorescence thresholds were set using the secondary antibody-only condition as a setpoint for unspecific signals. The same thresholds were used to define true positive signals when comparing virions retrieved from suspension or collagen cultures for each antibody.

sTLR2 binding to collagen-primed virions

The association of sTLR2 with HIV-1 NL4.3 R5 or HIV-1 NL4.3 ΔEnv virions retrieved from suspension or collagen cultures was assessed by performing a western blot analysis. Virus-containing supernatants from suspension or collagen cultures were harvested after 16 h of incubation at 37 °C, and the viral titers determined by SGPERT analysis. For each virus, 10¹¹ pURT were then either incubated alone, or with increasing concentrations (0.5 or 2 µg/ml) of sTLR2 in complete RPMI medium for 30 min at 37 °C. The samples were then overlaid on top of a 20% (w/v) sucrose cushion and ultracentrifuged using a tabletop ultracentrifuge (TL100, Beckman Coulter) for 45 min at 108,000 g at 4 °C. To determine the background pelleting of sTLR2, the recombinant protein alone was incubated in the absence of virus at 0.5 or 2 µg/ml prior to ultracentrifugation. After removal of the supernatants, viral and/or protein pellets were lysed in 2x SDS sample buffer (10% glycerol, 6% SDS, 130 mM Tris Hcl pH 6.8, 10% β-Mercaptoethanol) and boiled for 5 min at 95 °C, prior to export out of BSL-3 containment. The samples were processed for SDS-PAGE and the membranes were blotted for p24, TLR2 and gp120 detection using a Licor Odyssey system. The ratio of TLR2/p24 was calculated using Fiji, after subtraction of the background TLR2 signal obtained from the sTLR2 only condition to obtain a corrected ratio.

Flow cytometry

Samples were measured by flow cytometry in BD FACS Celesta with BD FACS Diva Software. Compensation controls were added for each experiment. Gating was performed using FlowJo software 10.4.2 and data were processed in GraphPad Prism 8.4.3 software.

Statistical analysis

Statistical analysis of datasets was carried out using Prism version 8.4.3 (GraphPad). For each dataset, normal distribution was tested using the Shapiro-Wilk test, and the appropriate test was used accordingly as indicated in the figure legends. Unless otherwise specified, all statistical tests were two-tailed. Statistical significance was calculated using one- or two-way ANOVA tests, as well as paired or unpaired t-tests, and Wilcoxon matched paired tests. Correction for multiple comparisons are indicated in figure legends. Significant differences are indicated with the corresponding p-values. Absence of p-values indicates a lack of statistical significance.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.