Super resolution microscopy reveals that caveolin-1 is required for spatial organization of CRFB1 and subsequent antiviral signaling in zebrafish

Abstract

Understanding spatial distribution and dynamics of receptors within unperturbed membranes is essential for elucidating their role in antiviral signaling, but conventional studies of detergent-resistant membrane fractions cannot provide this information. Caveolae are integral to numerous signaling pathways and these membrane domains have been previously implicated in viral entry but not antiviral defense. This study shows, for the first time, the importance of spatio-temporal regulation of signaling receptors and the importance of the regulation of clustering for downstream signaling. A novel mechanism for virus evasion of host cell defenses is demonstrated through disruption of clusters of signaling molecules organized within caveolin-rich domains. Viral infection leads to a downregulation in Caveolin-1b (Cav-1b), disrupting clusters of CRFB1, a zebrafish type I interferon receptor (-R) subunit. Super-resolution microscopy has enabled the first single-molecule imaging of CRFB1 association with cav-1b-containing membrane domains. Strikingly, downregulation of Cav-1b, the major protein component of caveolae, caused CRFB1 clusters to disperse. Dispersal of CRFB1 clusters led to a suppressed antiviral immune response both in vitro and in vivo, through abrogation of downstream signaling. This response strongly suggests that CRFB1 organization within cav-1b-containing membrane domains is critical for IFN-mediated antiviral defense and presents a previously undescribed antiviral evasion strategy to alter IFN signaling and the antiviral immune response.

Figure 1. Cav-1b colocalizes with the zebrafish homolog of IFN-R and is positively correlated.

ZFL cells (n≥10) were transfected with Cav-1b-PAmCherry (red) and with CRFB1-dendra2 (green). For all images, 60×/1.2 NA magnification. Scale bars, 1 µm. Shown is the plasma membrane of one cell representative of the experiment (A) and a magnification (B) of the region marked by the white box in A. The image shows that Cav-1b and CRFB1 colocalize in the cell membrane. (C) Measurements of Cav-1b and CRFB1 show a positive pair correlation value g(r) greater than one, confirming that the two species are colocalized together. Pair correlation calculations were performed as described in Methods; briefly, g(r) >1 indicates positive correlation/clustering, and g(r) = 1 indicates a random distribution. (D) Pair correlation measurements of CRFB1 were calculated for the receptor, control morpholino (MO), and Cav-1b MO. CRFB1 is more prone to random distribution when Cav-1b is knocked down. Error bars SEM (n ≥8 cells).

Figure 2. Cav-1b expression is modulated during virus infection, and Cav-1b knockdown leaves morphants susceptible to infection.

A) Quantitative RT-PCR results revealed fold changes in the expression levels of Cav-1b in infected embryos when compared to uninfected embryos. Zebrafish were exposed seven dpf to 1×10⁶ TCID₅₀/mL virus. Total RNA was isolated from at 12, 24, and 48 hours post infection and reverse transcribed to cDNA (n = 20 fish per time point). All expression values have been normalized to the zebrafish β-actin gene. Error bars represent SEM of three replicates. B) Zebrafish embryos that were injected with Cav-1b morpholino (MO) to knock down the expression of Cav-1b or control MO were infected 48 hpf with 1×10⁶ TCID₅₀/ml virus and monitored for mortality. Results are representative of three separate experiments. Statistical analysis (Wilcoxon test) of the Kaplan-Meier curve was performed (*, p = 0.008). C) Zebrafish embryos that were injected with Cav-1b MO to knock down the expression of Cav-1b or control MO were infected by static immersion 48 hpf with 1×10⁶ TCID₅₀/ml virus. The graph indicates that early in infection (0–12 hpi), there is no difference in viral burden between Cav-1b morphants and controls. However, by 24–48 hpi, Cav-1b morphants have a higher viral burden. Figure is representative of three experiments; error bars are standard error of the mean (*, p<0.05). D) Western blot showing efficacy of MO knockdown in zebrafish. Zebrafish embryos from Control and Cav-1b MO, and Control MO with SHRV infection were compared for cav-1b expression at the 72 hpf developmental stage. At this time, infected fish were 24 hpi. Membranes were re-probed with antibody against β-actin to control for protein loading.

Figure 3. Decrease of Cav-1b expression negatively affects the IFN pathway.

A) Stat1 gene expression was assessed by qRT-PCR in Control MO and Cav-1b MO embryos that were either SHRV infected or uninfected. Total RNA was extracted from 10 fish per treatment, cDNA synthesized and Stat1 mRNA expression assessed by qRT-PCR 24 hpi. The data are representative of three individual experiments and error bars indicate SEM. Each bar represents the mean fold induction of SHRV-infected embryos over corresponding controls. All expression values were normalized to zebrafish 18s. B) ISRE promoter activity is dampened in Cav-1b knockdown ZFL cells upon SHRV infection. ZFL cells were transfected with 250 ng of zISRE-luc construct along with 250 ng of cav-1b MO or control MO. Twenty four hours post transfection the ZFL cells were infected with SHRV at an MOI of 0.01. Cells were harvested for luciferase measurements 24 hpi. The graph shows relative luminescence units of control uninfected cells compared to cav-1b MO or control infected cells. Error bars are representative of SEM for two experiments. (**, p<0.001).

Figure 4. CRFB1 becomes dispersed as a result of whole virus infection in vitro.

ZFL cells were infected 24 h post transfection and fixed prior to imaging. For all images, 60×/1.2 NA magnification. Scale bars, 1 µm. Shown for each part is the surface of one cell representative of the experiment. A) Uninfected cells overexpressing CRFB1 demonstrate that the receptor exists in clustered patches indicative of caveolae. B) Cells infected with SHRV demonstrate that CRFB1 becomes dispersed as a result of virus infection by 24 hpi. C) Pair correlation analysis confirms that compared to uninfected cells, CRFB1 becomes dispersed after infection. Values of g(r) in cells with SHRV infection are considered to be random in comparison to values in cells that remain uninfected (n≥8 cells per treatment).

Figure 5. Crosslinking CRFB1 keeps receptor molecules clustered despite caveolin depletion.

ZFL cells were co-transfected with MO and expression plasmid via nucleofection and allowed to recover/adhere to cell culture plates for ∼6 hr prior to addition of crosslinking reagent. The crosslinking reaction was performed according to the manufacturer’s procedures. Cells were subsequently replenished with media and returned to the incubator for 24 hr post crosslinking. Scale bars, 1 µm. A) Cells transfected with Cav-1b MO/CRFB1 without crosslinking treatment show dispersed receptor molecules. B) Cells transfected with Control MO/CRFB1 with crosslinking clearly show clustered receptor molecules. C) Cells transfected with Cav-1b MO/CRFB1 with crosslinking. This demonstrates that despite depletion of Cav-1b, receptor molecules remain clustered. D) Pair correlation analysis confirms that with crosslinking, CRFB1 remains clustered despite Cav-1b depletion. Values of g(r) in cells with Cav-1b KD are similar to that for Controls (n≥8 cells per treatment).

Figure 6. MxA expression is retained with rescue of Cav-1b depletion.

ZFL cells were transfected and rescued as described in Methods. Shown is the fold difference in gene expression of MxA, an interferon stimulated gene. MxA transcript levels in Cav-1b depleted cells treated with crosslinking reagent (A) or rescued with cav-1b plasmid (B) show that rescuing caveolar disruption negates the depletion of caveolae which keeps CRFB1 molecules clustered under normal conditions. When Cav-1b is depleted and caveolae are not maintained with crosslinking reagent or cav-1b plasmid, minimal MxA expression is measured. The knockdown of CRFB1, CRFB2, and CRFB5 is the negative control; data indicate that there is low induction of MxA without IFN receptor subunits. When Cav-1b is depleted and CRFB1 is kept clustered, MxA expression remains at the same level as in the controls, demonstrating that the clustering of the receptor is essential for downstream signaling and that caveolin-1 plays a critical role in keeping the receptor clustered. Representative of 3 experiments; error bars indicate SEM (*, p<0.05). These results indicate dampening of MxA transcript production between control MO and Cav-1b MO prior to rescue of caveolae domains, and no significant difference of MxA transcript levels between Control MO and Cav-1b MO after rescue of caveolae domains.