The rotavirus enterotoxin NSP4 directly interacts with the caveolar structural protein caveolin-1

Abstract

Rotavirus nonstructural protein 4 (NSP4) is known to function as an intracellular receptor at the endoplasmic reticulum (ER) critical to viral morphogenesis and is the first characterized viral enterotoxin. Exogenously added NSP4 induces diarrhea in rodent pups and stimulates secretory chloride currents across intestinal segments as measured in Ussing chambers. Circular dichroism studies further reveal that intact NSP4 and the enterotoxic peptide (NSP4(114-135)) that is located within the extended, C-terminal amphipathic helix preferentially interact with caveola-like model membranes. We now show colocalization of NSP4 and caveolin-1 in NSP4-transfected and rotavirus-infected mammalian cells in reticular structures surrounding the nucleus (likely ER), in the cytosol, and at the cell periphery by laser scanning confocal microscopy. A direct interaction between NSP4 residues 112 to 140 and caveolin-1 was determined by the Pro-Quest yeast two-hybrid system with full-length NSP4 and seven overlapping deletion mutants as bait, caveolin-1 as prey, and vice versa. Coimmunoprecipitation of NSP4-caveolin-1 complexes from rotavirus-infected mammalian cells demonstrated that the interaction occurs during viral infection. Finally, binding of caveolin-1 from mammalian cell lysates to Sepharose-bound, NSP4-specific synthetic peptides confirmed the yeast two-hybrid data and further delineated the binding domain to amino acids 114 to 135. We propose that the association of NSP4 and caveolin-1 contributes to NSP4 intracellular trafficking from the ER to the cell surface and speculate that exogenously added NSP4 stimulates signaling molecules located in caveola microdomains.

FIG. 1.

Linear schematic of RV SA11 full-length NSP4 and deletion mutants with forward and reverse primers. PCR fragments were generated with the forward and reverse primers listed to the right of each construct and were cloned into DBD plasmid pD22 and AD plasmid pD32 of the ProQuest yeast two-hybrid system with Gateway technology. The NSP4 full-length clone of 528 nt was subjected to specific deletions to construct two 3′ deletion mutants, nt 1 to 321 and nt 1 to 450; three 5′ deletion mutants, nt 135 to 528, nt 304 to 528, and nt 336 to 528; and two 5′ and 3′ deletion mutants, nt 240 to 420 and nt 270 to 450.

FIG. 2.

Transient expression of NSP4 in mammalian cells. BSC-1, Caco-2, and MDCK cells were transiently transfected with pcDNA3.1-NSP4 as described in Materials and Methods. Equal protein concentrations of lysates were separated by 12% SDS-PAGE and electrotransferred to nitrocellulose for Western blot analyses. A specific band for full-length, fully glycosylated NSP4 at ∼28 kDa was observed in all three cell lines when probed with rabbit anti-NSP4_120-147 (lanes 2, 4, and 6). The Caco-2 cells also showed monoglycosylated and unglycosylated, full-length NSP2 (lane 4). Untransfected cell lysates did not show any NSP4-specific bands (lanes 1, 3, and 5).

FIG.3.

Confocal imaging and colocalization of NSP4 and caveolin-1. (A) Detection of NSP4 in transiently transfected mammalian cells. Caco-2 (left), MDCK (center), and BSC-1 (right) cultured cells were transfected with pCDNA3.1-NSP4 and probed sequentially with rabbit anti-NSP4_120-147 and goat anti-rabbit IgG-FITC. The cells were viewed by LSCM with a 63× Apochromat oil immersion objective. NSP4 was detected surrounding the nucleus, in vesicular structures in the cytosol, and at the cell periphery. Primary and secondary antibody controls lacked background fluorescence (far left). The BSC-1 cells (far right) are magnified twofold over the Caco-2 and MDCK cells to more easily view the subcellular staining distribution. (B) Colocalization of NSP4 and caveolin-1. Caco-2 cells were transfected with pcDNA3.1-NSP4, grown for 36 h, and fixed with methanol-acetone. Mouse anti-NSP4_120-147 and rabbit anti-caveolin-1 were used to probe for NSP4 and caveolin-1, respectively, in the same cells. Goat anti-mouse IgG conjugated to FITC and goat anti-rabbit IgG conjugated to Texas Red were used to detect the primary antibodies. Cells were visualized by LSCM with a laser source of 488 and 568 nm. The leftmost panel shows the detection of caveolin-1 (Texas Red) with the Em/Filter = HQ598/40, iris = 2; the next panel shows detection of NSP4 (FITC) with Em/Filter = 523/DF35. The third panel depicts the merged Texas Red and FITC fluorescence images, and the rightmost panel shows the colocalization of caveolin-1 and NSP4 as determined with Universal Imaging Metamorph software, version 3.6. Fluorograms were generated (not shown) and revealed ∼65% colocalization of NSP4 and caveolin-1. (C) RV-infected MDCK cells were similarly probed and analyzed for caveolin-1 and NSP4 colocalization. As in panel B, caveolin-1 fluorescence is shown on the left, followed by NSP4 and then the merged and colocalized images. For visualization of the infected cells, the settings were a 598/40-nm filter, iris = 2.5, for Texas Red and a 530/40-nm filter, iris = 2.2, for FITC. Fluorograms generated by Metamorph software, version 3.6, disclosed that ∼73% of the caveolin-1 colocalized with NSP4 in RV-infected cells (data not shown).

FIG.3.

Confocal imaging and colocalization of NSP4 and caveolin-1. (A) Detection of NSP4 in transiently transfected mammalian cells. Caco-2 (left), MDCK (center), and BSC-1 (right) cultured cells were transfected with pCDNA3.1-NSP4 and probed sequentially with rabbit anti-NSP4_120-147 and goat anti-rabbit IgG-FITC. The cells were viewed by LSCM with a 63× Apochromat oil immersion objective. NSP4 was detected surrounding the nucleus, in vesicular structures in the cytosol, and at the cell periphery. Primary and secondary antibody controls lacked background fluorescence (far left). The BSC-1 cells (far right) are magnified twofold over the Caco-2 and MDCK cells to more easily view the subcellular staining distribution. (B) Colocalization of NSP4 and caveolin-1. Caco-2 cells were transfected with pcDNA3.1-NSP4, grown for 36 h, and fixed with methanol-acetone. Mouse anti-NSP4_120-147 and rabbit anti-caveolin-1 were used to probe for NSP4 and caveolin-1, respectively, in the same cells. Goat anti-mouse IgG conjugated to FITC and goat anti-rabbit IgG conjugated to Texas Red were used to detect the primary antibodies. Cells were visualized by LSCM with a laser source of 488 and 568 nm. The leftmost panel shows the detection of caveolin-1 (Texas Red) with the Em/Filter = HQ598/40, iris = 2; the next panel shows detection of NSP4 (FITC) with Em/Filter = 523/DF35. The third panel depicts the merged Texas Red and FITC fluorescence images, and the rightmost panel shows the colocalization of caveolin-1 and NSP4 as determined with Universal Imaging Metamorph software, version 3.6. Fluorograms were generated (not shown) and revealed ∼65% colocalization of NSP4 and caveolin-1. (C) RV-infected MDCK cells were similarly probed and analyzed for caveolin-1 and NSP4 colocalization. As in panel B, caveolin-1 fluorescence is shown on the left, followed by NSP4 and then the merged and colocalized images. For visualization of the infected cells, the settings were a 598/40-nm filter, iris = 2.5, for Texas Red and a 530/40-nm filter, iris = 2.2, for FITC. Fluorograms generated by Metamorph software, version 3.6, disclosed that ∼73% of the caveolin-1 colocalized with NSP4 in RV-infected cells (data not shown).

FIG.3.

Confocal imaging and colocalization of NSP4 and caveolin-1. (A) Detection of NSP4 in transiently transfected mammalian cells. Caco-2 (left), MDCK (center), and BSC-1 (right) cultured cells were transfected with pCDNA3.1-NSP4 and probed sequentially with rabbit anti-NSP4_120-147 and goat anti-rabbit IgG-FITC. The cells were viewed by LSCM with a 63× Apochromat oil immersion objective. NSP4 was detected surrounding the nucleus, in vesicular structures in the cytosol, and at the cell periphery. Primary and secondary antibody controls lacked background fluorescence (far left). The BSC-1 cells (far right) are magnified twofold over the Caco-2 and MDCK cells to more easily view the subcellular staining distribution. (B) Colocalization of NSP4 and caveolin-1. Caco-2 cells were transfected with pcDNA3.1-NSP4, grown for 36 h, and fixed with methanol-acetone. Mouse anti-NSP4_120-147 and rabbit anti-caveolin-1 were used to probe for NSP4 and caveolin-1, respectively, in the same cells. Goat anti-mouse IgG conjugated to FITC and goat anti-rabbit IgG conjugated to Texas Red were used to detect the primary antibodies. Cells were visualized by LSCM with a laser source of 488 and 568 nm. The leftmost panel shows the detection of caveolin-1 (Texas Red) with the Em/Filter = HQ598/40, iris = 2; the next panel shows detection of NSP4 (FITC) with Em/Filter = 523/DF35. The third panel depicts the merged Texas Red and FITC fluorescence images, and the rightmost panel shows the colocalization of caveolin-1 and NSP4 as determined with Universal Imaging Metamorph software, version 3.6. Fluorograms were generated (not shown) and revealed ∼65% colocalization of NSP4 and caveolin-1. (C) RV-infected MDCK cells were similarly probed and analyzed for caveolin-1 and NSP4 colocalization. As in panel B, caveolin-1 fluorescence is shown on the left, followed by NSP4 and then the merged and colocalized images. For visualization of the infected cells, the settings were a 598/40-nm filter, iris = 2.5, for Texas Red and a 530/40-nm filter, iris = 2.2, for FITC. Fluorograms generated by Metamorph software, version 3.6, disclosed that ∼73% of the caveolin-1 colocalized with NSP4 in RV-infected cells (data not shown).

FIG. 4.

Representative growth patterns of cotransformed yeast in the reverse yeast two-hybrid assay. MaV203 cells were cotransformed with pDest22caveolin-1 (DBD-caveolin-1) and pDest32NSP4 or pDestNSP432 mutants (AD-NSP4), and growth phenotypes were compared to those of three yeast control strains, negative, weak interaction (1+), and moderate interaction (2+) (top three rows). Individual transformants, pD32NSP4 80-140 (NSP4 nt 240 to 420), pD32NSP4 90-150 (nt 270 to 450), and pD32NSP4 1-175 (nt 1 to 528), were grown on CSM-Leu⁻ Trp⁻ (far left column). Control and test colonies were replica plated onto (i) CSM-Leu⁻ Trp⁻ His⁻ plus 3AT at 12.5, 50, or 100 mM; (ii) CSM-Leu⁻ Trp⁻ Ura⁻; and (iii) CSM-Leu⁻ Trp⁻ plus 0.2% 5FOA. Growth on CSM-Leu⁻ Trp⁻ and CSM-Leu⁻ Trp⁻ His⁻ with ≤50 mM 3AT with little to no growth on CSM-Leu⁻ Trp⁻ plus 0.2% 5FOA was considered a weak positive interactor. A moderate positive interaction was indicated by growth on CSM-Leu⁻ Trp⁻ Ura⁻, decreased growth on CSM-Leu⁻ Trp⁻ plus 0.2% FOA (far right panels), and inhibition of growth on CSM-Leu⁻ Trp⁻ His⁻ with ≥50 mM 3AT. The stronger the interaction, the greater the growth on the CSM-Leu⁻ Trp⁻ Ura⁻ plates while the growth on CSM-Leu⁻ Trp⁻ plus 0.2% FOA was significantly decreased.

FIG. 5.

Western blot analyses of yeast cell lysates cotransformed with full-length NSP4- or NSP4 mutant-caveolin-1 fusion protein. To verify the expression of both the NSP4- and caveolin-1 fusion proteins, the cotransformed MaV203 yeast cell lysates were separated by 12% SDS-PAGE, transferred to nitrocellulose membranes, and probed with rabbit anti-NSP4_2-22 (lanes 1 and 2), anti-NSP4_150-175 (lanes 3 to 7), or anti-NSP4_113-149 (lanes 8 to 10) for NSP4 or NSP4 mutants and anti-caveolin-1_2-32 (lanes 11 to 17) for caveolin-1. Untransformed controls are shown in lanes 1, 3, 6, 8, and 11.

FIG. 6.

Coimmunoprecipitation of caveolin-1 and NSP4 from RV-infected MDCK cells. (A) RV-infected (lane 4) and uninfected (lane 2) MDCK cell lysates were reacted with rat anti-caveolin-1_161-178 and protein G-agarose. The infected MDCK cell lysates were also reacted with rat preimmune sera, which served as a negative control (lane 1). All samples were separated by SDS-PAGE, transferred to nitrocellulose, and blotted with anti-NSP4_150-175. (A) The absence of NSP4-specific bands from infected MDCK cell lysates precipitated with preimmune serum and uninfected cell lysates reacted with rat anti-caveolin-1_161-178 are shown in lanes 1 and 2, respectively. Lane 3 contains RV-infected cell lysates alone and serves as the positive control. Full-length, fully glycosylated, monoglycosylated, and unglycosylated NSP4 are present in lanes 3 and 4. Additional bands at ∼18 and 15 kDa are reactive with the NSP4 peptide antisera and likely are cleavage products (lanes 3 and 4). A background band at ∼30 kDa is present in the uninfected negative control (lane 2). (B) To verify that the ∼26- and 24-kDa bands correspond to the monoglycosylated and fully glycosylated forms of NSP4, infected and uninfected MDCK cell lysates were endo H digested, separated by SDS-PAGE, transferred, and blotted with rabbit anti-NSP4_150-175. Lanes 1 and 3 correspond to the undigested negative and positive controls, respectively. Comparison of lanes 3 and 4 shows the effects of endo H treatment, with the fully glycosylated and monoglycosylated bands shifting to ∼20 kDa, equivalent to unglycosylated NSP4.

FIG. 7.

Refinement of the location of the caveolin-1 binding domain to NSP4 residues 114 to 135. In vitro binding assays were performed with NSP4 peptides 113 to 149 (A), 114 to 135 (B and D), and 2 to 22 (C) attached to cyanogen bromide-activated Sepharose 4B beads. MDCK (A, B, and C) and FRT (C and D) cell lysates were incubated with the immobilized peptide overnight, pelleted, and subjected to Western blot analysis with anti-caveolin-1 antibody. To make sure that the antibody was reactive in the Western blot assays, MDCK cell lysates alone were included in each panel (far right lane of each panel). To ensure the absence of nonspecific binding to the beads, MDCK cell lysates were reacted with peptide-free Sepharose 4B beads (lane 1 in panels A and B and lane 2 in panel C). FRT cell lysates were reacted with beads only (panel C, lane 1) and with both NSP4_2-22- and NSP4_114-135-bound beads (lane 4 in panel C and lane 2 in panel D, respectively) and served as an additional negative control. Test lanes for NSP4 peptide-caveolin-1 binding included lane 2 in panels A, B, and D and lanes 3 (MDCK cell lysates) and 4 (FRT cell lysates) in panel C.

FIG. 8.

Alignment of previously reported NSP4 domains with results from the yeast two-hybrid and in vitro capture assays. Selected known functional domains of NSP4 are shown, with H1, H2, and H3 indicating the three N-terminal hydrophobic domains. Glycosylation sites (CHO) are shown at aa 8 and 18. Most of H2 and a small part of H1 traverse the ER membrane such that residues 1 to 23 are localized in the ER lumen and aa 44 to 175 are cytoplasmic. The amphipathic α-helical region that folds as a coiled-coil domain (AAH CCD, aa 95 to 137), the enterotoxic peptide (aa 114 to 135), and the cleavage site at residue 112 are indicated in the extended cytoplasmic domain. The NSP4 binding sites for VP4, VP6, and tubulin are delineated at aa 112 to 148, 161 to 175, and 129 to 175, respectively. The NSP4 deletion mutants are linearly depicted, with the yeast two-hybrid results shown on the right. The alignment of the positive interactions localized the caveolin-1 binding site to NSP4 residues 112 to 140. The results from the peptide capture assays are represented in the lower box, which further defined the caveolin-1 binding site to NSP4 aa 114 to 135.