2.1 Cell Cultures

293T (human, kidney) and A549 (human, lung) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 110 mg/L sodium pyruvate, and 4.5 g/L d-glucose. Flp-In T-REx 293 cells were purchased from Invitrogen and maintained in DMEM supplemented with 10% FBS, 10 µg/mL blasticidin (Invitrogen), and 100 µg/mL Zeocin (Invivogen). A549 and Flp-In T-REx293 TMEM106B^KO cell lines were cultured in DMEM supplemented with 10% FBS, 1 µg/mL puromycin (Invivogen). All cell lines were incubated at 37°C and 5% CO₂ in a humidified atmosphere.

2.2 Plasmids

For the construction of S-expressing plasmids, a plasmid expressing Lassa fever virus (LASV) glycoprotein (GP) was described previously [12]. The acid amino sequences of the SARS-CoV-2 S protein were retrieved from the GISAID (XBB.1.5, EPI_ISL_18095904; BA.2.86, EPI_ISL_18096761). According to the method described by Babcock et al. [13], the codon-optimized S genes were synthesized and cloned into the pSecTag2 vector. Plasmids encoding D614G, E484D, BA.2, and JN.1 S protein were constructed by PCR site-directed mutagenesis. Plasmid encoding TMEM106B was purchased from OriGene (Cat. No. RC206257) and cloned into the pCAGGS vector, then it was used as a template to produce glycosylation site mutations. All plasmid constructs used in this study were confirmed by Sanger sequencing.

For the construction of type II transmembrane serine proteases (TTSPs) expressing plasmids, TMPRSS2 (GenBank accession number, NM_005656), TMPRSS3 (GenBank accession number, NM_024022.3), TMPRSS4 (GenBank accession number, NM_019894.4), TMPRSS10 (GenBank accession number, NM_006587.4), TMPRSS11A (GenBank accession number, NM_182606.2), TMPRSS11B (GenBank accession number, NM_182502.3), TMPRSS11D (GenBank accession number, NM_004262.3), TMPRSS11E (GenBank accession number, NM_014058.4), TMPRSS11F (GenBank accession number, NM_207407.2), TMPRSS13 (GenBank accession number, NM_001077263.3) were synthesized and cloned into the pcDNA5/FRT-derived vector with a C9 tag at the C-terminus.

2.3 Generation of Knockout Cell Lines

Standard cloning protocol was followed to clone sgRNAs targeting the gene into pLentiCRISPRv2 plasmid (Addgene 52961) for knockout of TMEM106B. TMEM106B^KO cells were generated by transducing cells with a pool of two sgRNAs targeting TMEM106B and selecting with puromycin (1 µg/mL) for 3 days. Monoclonal cells were made by seeding a dilution series of cells and selecting wells containing a single cell colony. Genomic DNA was isolated from cells with a QIAamp DNA mini kit and identified by Sanger sequencing to verify the knockout.

2.4 Immunoblotting Assay

Cell monolayers were washed once with phosphate-buffer saline (PBS) and lysed in PBS buffer containing 0.3% n-decyl-β-d-maltopyranoside (Anatrace) and protease inhibitors (Roche). An aliquot of cell lysate was separated on a NuPAGE Novex 4 to 12% Bis-Tris Gel (Invitrogen) and electrophoretically transferred onto a nitrocellulose membrane (Invitrogen). After being blocked by 5% milk for 1 h at room temperature, the membrane was incubated with a primary antibody overnight at 4°C: TMEM106B (Proteintech, 60333-1-Ig), C9 (Santa Cruz, sc-57432), SARS-CoV-2 Spike S2 (Sino Biological, 40590-T62), GAPDH (Proteintech, 60004-1-Ig), β-actin (Sigma, A2228). The blot was washed three times with washing buffer (0.05% Tween-20 in TBS), followed by incubation with the horseradish peroxidase (HRP) conjugated secondary antibody for 1 h at room temperature (Cell Signaling Technology, #7076S). After washing, the signal was detected with chemiluminescent substrates and imaged with Amersham ImageQuant 800.

2.5 Production of Pseudotyped Viruses

293T cells in six-well plates were transfected with plasmids encoding various SARS-CoV-2 S protein or LASV GP by lipofectamine 2000 (Invitrogen). After incubation at 37°C for 6–8 h, DMEM supplemented with 10% FBS was added and transduced with a replication-deficient VSV vector that contains expression cassettes for firefly luciferase instead of the VSV-G open reading frame, VSV*ΔG-fLuc [14]. The supernatants containing pseudotyped viral particles were harvested at 48 and 72 h posttransfection, mixed, and centrifuged at 3000 rpm for 5 min to remove the cell debris. Then, the supernatants were filtered through a 0.45 µm PES syringe filter (Millipore) and stored at −80°C for subsequent studies The prepared pseudotyped viral particles (pp) were titrated in T-REx 293/hACE2 cells as previously described equivalent amount of pseudotyped viral particles (13 000 TCID₅₀/mL) were used in this study [15].

2.6 Viral Entry Assay

A549 TMEM106B^KO cells were seeded into 96-well plates for 1 day, and then infected with the desired pseudotyped viral particles at 37°C overnight. At 24 h postinfection (HPI), the medium was removed, and the cells were lysed with 30 µL/well of cell lysis buffer (Promega) for 15 min, followed by adding 50 µL/well of luciferase substrate (Promega), the luciferase activity was analyzed by Turner BioSystems.

2.7 Effect of TMPRSS2 or Other TTSPs on Cellular Entry

A549 TMEM106B^KO cells in six-well plates were cotransfected with the pCAGGS-TMEM106B plasmid and pCAGGS-TMPRSS2 (or other TTSPs) or an empty pCAGGS plasmid as a negative control by lipofectamine 3000 (Invitrogen). At 24 h posttransfection, cells were trypsinized and seeded into black-walled 96-well plates for overnight culture at 37°C. Following this incubation period, the adherent cells were infected with pseudotyped viral particles (D614Gpp, E484Dpp, BA.1pp, BA.2pp, XBB.1.5pp, BA.2.86pp, JN.1pp, and LASVpp). Luciferase activity was quantitatively assessed at 24 h postinfection (hpi) using a GloMax luminometer (Promega).

2.8 GFP-Split Fusion Assay

For cell–cell fusion assays with S-expressing cells, 293T cells in 12-well plates were cotransfected with GFP1-10 and various SARS-CoV-2 spike plasmids. For targeting cells, Flp-In T-REx293 TMEM106B^KO cells in 12-well plates were cotransfected with GFP11, pCAGGS-TMEM106B plasmids, and pCAGGS-TMPRSS2 or an empty pCAGGS plasmid as control by lipofectamine 3000 (Invitrogen). At 24 h post-transfected, 293T-GFP1-10, and T-REx293-GFP-11 cells were cocultured at a 1:1 ratio for 16–18 h and then imaged by Nikon ECLIPSE Ts2 with at least three fields for each well. The GFP area was quantified on ImageJ.

2.9 Flow Cytometry

Cell surface expression levels of WT TMEM106B and G2A TMEM106B were assessed by flow cytometry. A549 TMEM106B^KO cells were transfected with pCAGGS empty vector, WT TMEM106B, or G2A TMEM106B for 48 h. Following transfection, cells were immunostained using a rabbit anti-TMEM106B antibody (Proteintech, Cat. No. 20995-1-AP) and a Goat Anti-Rabbit IgG H&L secondary antibody (Alexa Fluor 488, Abcam, Cat. No. 150077). Flow cytometric analysis was performed using a Cytoflex flow cytometer (Beckman Coulter, Brea, CA, USA), and data were analyzed with FlowJo software.

2.10 Immunofluorescence

To visualize the subcellular localization of wild-type and mutant TMEM106B proteins, Overexpression of WT TMEM106B or G2A TMEM106B in T-REx 293 TMEM106B^KO cells were fixed with 4% paraformaldehyde for 20 min, followed by permeabilization with 0.1% Triton X-100 for 10 min. The cells were then incubated with blocking solution (1× PBS containing 5% FBS) for 1 h. Subsequently, the cells were incubated with mouse monoclonal antibodies specific to TMEM106B (Proteintech 60333-1-Ig) and rabbit monoclonal antibodies against LAMP1 (Cell Signaling 9901), EEA1 (Cell Signaling 2411), and Rab9 (Cell Signaling 5118s). The bound primary antibodies were detected using Alexa Fluor 647-conjugated (red) and Alexa Fluor 488-conjugated (green) secondary antibodies. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Imaging was conducted using a confocal fluorescence microscope (Zeiss LSM 900; Carl Zeiss, Thornwood, NJ).

2.11 Statistical Analysis

All the experiments were repeated at least three times. Data are mean ± SEM of three biological repeats. Statistical differences between the two groups were analyzed using the two-sided unpaired t-test with Welch's correction. Multiple group comparisons were determined using one-way analysis of variance (ANOVA) with Dunnett's multiple comparisons test by GraphPad Prism (v.9). p values less than 0.05 were considered statistically significant.

3.1 TMEM106B Supports Viral Entry of Multiple SARS-CoV-2 Omicron Variants

To systematically investigate whether multiple variants of SARS-CoV-2, including BA.2.86 and JN.1 variants of the BA.2 lineage, utilize TMEM106B to invade cells, we constructed Flp-In T-REx 293 and A549 TMEM106B^KO cell lines that have little or no ACE2 expression, respectively (Figure 1A,B). Furthermore, the CRISPR/Cas9 knockout of TMEM106B was verified by genome sequencing and immunoblotting assay (Figure 1C, D). The knockout of TMEM106B significantly reduced the susceptibility of T-REx 293 cells or A549 cells to the infection by VSV-based pseudotyped virus of multiple SARS-CoV-2 variants, including the D614G variant, E484D variant, BA.1 variant, BA.2 variant, XBB.1.5 variant, BA.2.86 variant, and JN.1 variant (Figure 1E,F). While the efficiency of E484D spike-mediated infection was reduced most significantly, the infection driven by the spike proteins from other tested Omicron strains, such as BA.2.86 and JN.1 were all reduced in TMEM106B^KO cells, which indicated multiple SARS-CoV-2 Omicron variants can utilize TMEM106B to infect cells.

3.2 TMPRSS2 Promotes Syncytial Formation of Multiple SARS-CoV-2 Omicron Variants in TMEM106B-Expressing Cells

Given the critical role of TMPRSS2 in priming the viral S protein cleavage and cell–cell fusion triggered by ACE2, we next investigated whether TMPRSS2 contributes to the syncytia formation mediated by the interaction between the S protein of multiple SARS-CoV-2 variants with the TMEM106B receptor. We conducted a dual fluorescent complementary system to test syncytia formation. As shown in Figure 2B, HEK293T cells cotransfected with plasmids encoding SARS-CoV-2 E484D spike and GFP1-10 (donor cells) were cocultured with T-REx 293 TMEM106B^KO cells cotransfected with plasmids expressing TMEM106B and GFP11 (target cells), the GFP-positive syncytia were formed after 16–18 h. We found that TMEM106B could promote cell–cell fusion mediated by SARS-CoV-2 E484D S protein, although its effect was less efficient than that mediated by ACE2. Notably, the syncytia formation enhanced by TMPRSS2 was only observed when the protease was overexpressed on the target cells, but not donor cells (Figure 2A–D). We also found that TMEM106B could trigger the syncytia formation mediated by S proteins from Omicron variants, including BA.2.86 and JN.1 strains, which was significantly enhanced by TMPRSS2 overexpression on the target cells (Figure 2E,F).

We then collected the adherent syncytia and examined the cleaved S species by western blot analysis assay, using a rabbit polyclonal antibody specifically detecting the S2 subunit. This experiment revealed that, in addition to full-length S and auto-cleaved S2 fragments (~80 kDa), another S2 species of approximately 65 kDa was detected only when TMEM106B was overexpressed on target cells, which was enhanced by TMPRSS2 expression (Figure 2G). This result indicates that the binding of S protein to TMEM106B could trigger the cleavage at S2′ site located in S2 subunit and generate a resultant band of S2′ fragment at ∼65 kDa. Furthermore, the overexpression of TMPRSS2 enhances the S2′ cleavage of the SARS-CoV-2 S variant upon its binding of TMEM106B receptor. Collectively, the above findings indicate that TMEM106B could trigger the S2′ proteolytic cleavage of the S proteins from multiple SARS-CoV-2 Omicron variants and induce syncytia formation, which was enhanced by TMPRSS2 overexpression.

3.3 TMPRSS11F and TMPRSS13 can Also Enhance Viral Entry and Cell–Cell Fusion Through TMEM106B

To investigate the role of other TTSPs except for TMPRSS2 in the cellular invasion of SARS-CoV-2 using TMEM106B, we constructed ten TTSPs with a C9 tag. The TTSPs expression was verified by western blot analysis (Figure 3A). The effect of TTSPs on viral entry was assessed by co-expressing TTSPs and TMEM106B, followed by infection with pseudotyped SARS-CoV-2 E484D, BA.2.86, and JN.1 viruses. In line with the findings from a previous study [10], TMPRSS2 overexpression significantly enhanced infection mediated by the E484D spike. Moreover, TMPRSS2, TMPRSS11F, and TMPRSS13 overexpression also enhanced infection mediated by BA.2.86 and JN.1 S proteins, while other TTSPs showed no significant effect (Figure 3B). We subsequently examined the effects of TMPRSS11F and TMPRSS13 in syncytium formation, and found that many large and multinucleated GFP⁺ cells were detected in the TMPRSS11F- or TMPRSS13-expressing cells, suggesting that S protein-induced cell–cell fusion is enhanced by these TTSPs (Figure 3C,D). Our results collectively indicate that the selected TTSPs can enhance Omicron viral entry and syncytia formation mediated by the TMEM106B receptor.

3.4 The N-Glycosylation Motif Is Critical for TMEM106B to Support SARS-CoV-2 S Protein-Mediated Entry

As a type II transmembrane protein, TMEM106B contains a short N-terminal intracellular tail, a transmembrane domain, and a long C-terminal extracellular structural domain, with five glycosylation sites (Figure 4A). Studies have shown that mutations in these glycosylation sites can alter lysosomal localization and impact transport [16], raising the question of whether these mutations affect the receptor function of TMEM106B. Meanwhile, the two SNPs (T185S and D252N) in TMEM106B have been associated with neurological disorders [17-19], prompting the need to examine their potential impact on its receptor activity. To this end, we constructed six mutations at N-glycosylation sites (TMEM106B N145A, N151A, N151S, N164A, N183A, and N256A) and two SNPs variant (T185S and D252N) to test their receptor activity to support E484D spike-mediated infection, and a double mutation (M210A/F213A) with disrupted receptor function identified in previous study serve as a control [10]. Protein expression was verified by immunoblotting, which showed no significant change in the molecular weight, although expression levels were slightly reduced (Figure 4B). Our results revealed that the N151A and N256A mutations significantly reduced SARS-CoV-2 E484D variant entry and syncytia formation compared to wild-type TMEM106B, while the M210/F213A mutant abolished TMEM106B receptor function, consistent with prior studies (Figure 4C,D). Interestingly, the disease-associated mutations T185S and D252N did not significantly alter receptor function. As a control, mutations at these sites did not affect the entry of LASV (Figure 4C). Notably, a SNP of N151S (rs761775508) disrupting the glycosylation site reduced TMEM106B receptor function of supporting viral entry and syncytia formation. This result implies this variant may potentially protect individuals against TMEM106B mediated SARS-CoV-2 infection (Figure 4C–E).

3.5 TMEM106B G2A Mutation in Myristoylation Site Enhances SARS-CoV-2 E484D Spike-Mediated Infection

Recent structural predictions suggest that TMEM106B may be a part of the lipid transfer protein family [20], and its N-terminus contains the predicted myristoylation signal (¹MGxxxS⁶) [21], which facilitates the anchoring of TMEM106B on the cell membrane. Interestingly, one SNP G2A (rs765967370) alters this putative myristoylation motif [22]. To investigate whether the acylation site mutation affects the receptor function of TMEM106B, we constructed the TMEM106B wild-type (WT) and G2A mutant and tested their expression in A549 TMEM106B^KO cells. Western blot analysis showed significantly higher expression of G2A compared to WT (Figure 5A), and the increased surface expression of G2A on the cell membrane was confirmed with FACS analysis (Figure 5B). Then, we observed that the G2A mutation enhanced viral entry and syncytium formation of SARS-CoV-2 E484D compared to wild-type TMEM106B. Moreover, in the presence of TMPRSS2, viral infection and syncytium formation mediated by G2A were further increased, although its efficacy was much less than that of the ACE2 (Figure 5C,D). We also observed that G2A mutation did not obviously alter TMEM106B subcellar location (Figure 5E). Overall, these data suggest that a TMEM106B SNP, G2A, may promote infection by increasing its cell surface expression and enhance the virus spread to ACE2-negative host cells.