Oral Antiviral Defense: Saliva- and Beverage-like Hypotonicity Dynamically Regulate Formation of Membraneless Biomolecular Condensates of Antiviral Human MxA in Oral Epithelial Cells

Abstract

The oral mucosa represents a defensive barrier between the external environment and the rest of the body. Oral mucosal cells are constantly bathed in hypotonic saliva (normally one-third tonicity compared to plasma) and are repeatedly exposed to environmental stresses of tonicity, temperature, and pH by the drinks we imbibe (e.g., hypotonic: water, tea, and coffee; hypertonic: assorted fruit juices, and red wines). In the mouth, the broad-spectrum antiviral mediator MxA (a dynamin-family large GTPase) is constitutively expressed in healthy periodontal tissues and induced by Type III interferons (e.g., IFN-λ1/IL-29). Endogenously induced human MxA and exogenously expressed human GFP-MxA formed membraneless biomolecular condensates in the cytoplasm of oral carcinoma cells (OECM1 cell line). These condensates likely represent storage granules in equilibrium with antivirally active dispersed MxA. Remarkably, cytoplasmic MxA condensates were exquisitely sensitive sensors of hypotonicity-the condensates in oral epithelium disassembled within 1-2 min of exposure of cells to saliva-like one-third hypotonicity, and spontaneously reassembled in the next 4-7 min. Water, tea, and coffee enhanced this disassembly. Fluorescence changes in OECM1 cells preloaded with calcein-AM (a reporter of cytosolic "macromolecular crowding") confirmed that this process involved macromolecular uncrowding and subsequent recrowding secondary to changes in cell volume. However, hypertonicity had little effect on MxA condensates. The spontaneous reassembly of GFP-MxA condensates in oral epithelial cells, even under continuous saliva-like hypotonicity, was slowed by the protein-phosphatase-inhibitor cyclosporin A (CsA) and by the K-channel-blocker tetraethylammonium chloride (TEA); this is suggestive of the involvement of the volume-sensitive WNK kinase-protein phosphatase (PTP)-K-Cl cotransporter (KCC) pathway in the regulated volume decrease (RVD) during condensate reassembly in oral cells. The present study identifies a novel subcellular consequence of hypotonic stress in oral epithelial cells, in terms of the rapid and dynamic changes in the structure of one class of phase-separated biomolecular condensates in the cytoplasm-the antiviral MxA condensates. More generally, the data raise the possibility that hypotonicity-driven stresses likely affect other intracellular functions involving liquid-liquid phase separation (LLPS) in cells of the oral mucosa.

Keywords: barrier immunity; biomolecular condensates; environmental hypotonic stress; human myxovirus resistance protein (MxA); interferon-λ/IL-29; macromolecular crowding; membraneless organelles (MLOs); oral/gingival epithelium; osmoregulation; regulated volume decrease (RVD).

Figure 1

IFN-λ1 (IL-29), but not IFN-α2, induces expression of MxA in oral epithelial cells (OECM1 cell line). (A,B) 35 mm cultures of OECM1 and A549 cells were left untreated or treated with human IFN-α2 or human IFN-λ1 (50 ng/mL for 2 days) followed by fixation (4% PFA for 1 h at 4 °C) and immunofluorescence imaging for MxA; (C) OECM1 cultures treated with IFN-λ1 for 2 days were exposed to 5% 1,6-hexanediol in PBS, or left untreated, and then fixed and imaged for MxA; (D) Western blot of extracts (30 µg/lane) prepared from parallel plates as in (A); (E) OECM1 cultures were transfected with pGFP-MxA expression vector and imaged in PBS 2 days later, followed by treatment with 5%-Hex and imaging of the same cells 5 min later. All scale bars = 10 µm.

Figure 2

Single-cell antiviral phenotype (protected against VSV; white arrows) of OECM1 cells expressing GFP-MxA mainly in condensates (A) or mainly in the dispersed phase (B). Cultures in 35 mm plates were transiently transfected with pGFP-MxA vector, and two days later were challenged with VSV (moi > 10 pfu/cell) and fixed 24 h after the start of infection [31]. VSV replication was assessed by immunostaining for the VSV nucleocapsid (N) protein (in red) [19,24]. White arrows point to GFP-containing cells with reduced VSV-N. Scale bars = 20 µm. (C) Schematic highlighting the dynamic equilibrium between GFP-MxA in condensed vs. dispersed phase (note from Figure 4B, and additional Figures below, that even in cells with visually “mainly” condensed GFP-MxA, there is 15–25% of GFP-MxA in the dispersed phase, which can be still antivirally active as in (B) above).

Figure 3

Reversible osmosensing by GFP-MxA condensates in oral epithelial cells—focus on beverage-like hypotonicity. (A) Sequential live-cell imaging of the same OECM1 cells expressing GFP-MxA condensates 2 days after transient transfection first in isotonic PBS (300 mOsm) and then after shifting to hypotonic ELB (40 mOsm). (B) the same cells as in (A) were sequentially imaged after further shifting back to isotonic PBS (300 mOsm). White arrows, formations of vacuole-like dilations [29] prior to condensate formation. (C) IFN-λ1 (50 ng/mL for 2 days) treated OECM1 cultures were fixed after washing with PBS, or after 5 min in ELB, or after 5 min in ELB, and then returned back to PBS for 5 min. Cultures were fixed using 4% PFA and immunostained for MxA. All scale bars = 10 µm.

Figure 4

Spontaneously reversible osmosensing by GFP-MxA condensates in oral epithelial cells-focus on saliva-like hypotonicity. (A) Sequential live-cell imaging of the same OECM1 cell expressing GFP-MxA condensates 2 days after transient transfection first in full-culture medium (330 mOsm) and then after shifting to hypotonic of one-third tonicity (110 mOsm; full medium diluted 1:2 with water) for the next 8–10 min. Scale bar = 10 µm. (B) Quantitation of GFP-MxA in condensates on a % per-cell basis in the images shown in (A). This quantitation was carried out using the small object subtract Filter in Image J [20,24].

Figure 5

Rapid disassembly of GFP-MxA condensates by drinking water, tea, and coffee. OECM1 cultures expressing GFP-MxA condensates (2 days after transfection) were first imaged in full medium (approx. 330 mOsm) and then shifted to one-third tonicity saliva-like medium (100 mOsm) for 1–2.4 h to allow completion of the disassembly and reassembly cycle as shown in Figure 4. Single live cells in the respective cultures were then imaged and the imaging continued upon shifting the cultures to drinking water (A), tea (B), coffee (C). Scale bars = 10 µm.

Figure 6

Testing a biophysical basis for hypotonicity sensing by GFP-MxA condensates in oral epithelial cells using calcein quenching as a reporter for macromolecular crowding. (A) OECM1 cells were preloaded with calcein-AM (2 µM in PBS) for 15 min, washed 4× with PBS and then imaged. The culture was then shifted to hypotonic ELB for 2 min and imaged immediately using the same fluorescence settings. The culture was then shifted to isotonic PBS for 9 min and cells imaged. Areas within white dashed boxes are shown at higher magnification in the lower panels. Scale bar = 10 µm. (B) Calcein fluorescence on a per-cell basis (in arbitrary units) is depicted. ****, p < 0.0001; ns, not significant; VLD, vacuole-like dilatations; n, number of cells quantitated.

Figure 7

Testing a biochemical basis for hypotonicity sensing by GFP-MxA condensates in oral epithelial cells. (A,B), OECM1 cells were either exposed to CsA (25 µM) or DMSO alone for 20 min in full-culture medium, and then shifted to one-third tonicity medium in the continued presence of CsA. Live-cell imaging was carried out as indicated. Scale bar = 10 µm. (C) Quantitation of % GFP-MxA per cell in condensates (in the same cells shown in (A,B)).

Figure 8

Testing a biochemical basis for hypotonicity sensing by GFP-MxA condensates in oral epithelial cells. (A,B) OECM1 cells were either kept in full-culture medium or exposed to TEA (20 mM) in full-culture medium for 20 min. Cultures were then shifted to one-third tonicity medium in the continued presence of TEA. Live-cell imaging was carried out as indicated. Scale bar = 10 µm. (C) Quantitation of % GFP-MxA per cell in condensates (in the same cells shown in (A,B)).

Figure 9

Hypothesis: Overview of the biophysical and biochemical mechanisms possibly involved in the cell-volume-driven dynamic regulation of the formation, disassembly, and reassembly of MxA condensates in oral epithelial cells subjected to saliva- and beverage-like hypotonicity. 2-DG, 2-deoxyglucose; KCC, potassium-chloride cotransporter channels 1–4; MLO, membraneless organelle, TEA, tetraethylammonium chloride; RVD, regulated volume decrease; VLD, vacuoele-like dilatations; WNK kinase, “With no lysine” kinase family members 1–4.