Dengue Virus NS1 Protein Modulates Cellular Energy Metabolism by Increasing Glyceraldehyde-3-Phosphate Dehydrogenase Activity

Abstract

Dengue is one of the main public health concerns worldwide. Recent estimates indicate that over 390 million people are infected annually with the dengue virus (DENV), resulting in thousands of deaths. Among the DENV nonstructural proteins, the NS1 protein is the only one whose function during replication is still unknown. NS1 is a 46- to 55-kDa glycoprotein commonly found as both a membrane-associated homodimer and a soluble hexameric barrel-shaped lipoprotein. Despite its role in the pathogenic process, NS1 is essential for proper RNA accumulation and virus production. In the present study, we identified that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) interacts with intracellular NS1. Molecular docking revealed that this interaction occurs through the hydrophobic protrusion of NS1 and the hydrophobic residues located at the opposite side of the catalytic site. Moreover, addition of purified recombinant NS1 enhanced the glycolytic activity of GAPDH in vitro. Interestingly, we observed that DENV infection promoted the relocalization of GAPDH to the perinuclear region, where NS1 is commonly found. Both DENV infection and expression of NS1 itself resulted in increased GAPDH activity. Our findings indicate that the NS1 protein acts to increase glycolytic flux and, consequently, energy production, which is consistent with the recent finding that DENV induces and requires glycolysis for proper replication. This is the first report to propose that NS1 is an important modulator of cellular energy metabolism. The data presented here provide new insights that may be useful for further drug design and the development of alternative antiviral therapies against DENV.

Importance: Dengue represents a serious public health problem worldwide and is caused by infection with dengue virus (DENV). Estimates indicate that half of the global population is at risk of infection, with almost 400 million cases occurring per year. The NS1 glycoprotein is found in both the intracellular and the extracellular milieus. Despite the fact that NS1 has been commonly associated with DENV pathogenesis, it plays a pivotal but unknown role in the replication process. In an effort to understand the role of intracellular NS1, we demonstrate that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) interacts with NS1. Our results indicate that NS1 increases the glycolytic activity of GAPDH in vitro. Interestingly, the GAPDH activity was increased during DENV infection, and NS1 expression alone was sufficient to enhance intracellular GAPDH activity in BHK-21 cells. Overall, our findings suggest that NS1 is an important modulator of cellular energy metabolism by increasing glycolytic flux.

FIG 1

Identification of GAPDH as an iNS1-interacting partner. (A) Extracts from mock-infected (Mock) and DENV2-infected (DENV) HUVEC-C cells were added to a column containing immobilized anti-NS1, and the eluted fractions were separated by SDS-PAGE prior to band excision and in-gel digestion. (B) The trypsinized peptides were subjected to Q-Tof MS/MS analysis under both mock infection and DENV infection conditions. MM, molecular mass. (C) The peak corresponding to the most abundant peptide was identified and selected for a second round of MS. The peptide sequence (with z being equal to 2) was then analyzed by use of the Mascot algorithm, using the sequences in the Swiss-Prot database for protein identification. The identified peptide corresponded to a unique fragment of the GAPDH protein, which was confirmed by a search of the NCBI database. The results presented here are representative of those from two independent experiments.

FIG 2

Confirmation of the NS1-GAPDH interaction by direct binding ELISA and a pulldown assay. (A) Microtiter plates were coated with purified GAPDH (10 μg/ml), and increasing amounts of recombinant DENV2 NS1 were added. The bound NS1 was detected using an anti-NS1 polyclonal antibody. BSA was used as a negative control. The error bars indicate the standard deviations from three independent experiments, and the asterisks indicate significant differences from the control using two-way ANOVA and the Bonferroni posttest. ***, P < 0.001. (B) For the pulldown assay, the rNS1 and GAPDH proteins at a concentration of 2.5 μM each were incubated overnight at 4°C with anti-NS1 polyclonal antibody covalently coupled to an amino-linked agarose resin. Elution was carried out with the elution buffer provided in the co-IP kit (Pierce). The control reaction was carried out following the same protocol described above, except that only GAPDH was incubated with the resin. The same procedure was performed using control IgG coupled to an amino-linked agarose resin. (C) As described in the legend to panel B, except that the GAPDH protein was replaced by the HPX protein, which does not interact with NS1. (D) Pulldown assay using extracts of mock- and DENV2-infected BHK-21 cells. IN, input; FT, flowthrough; E1 and E2, 1st and 2nd elution fractions, respectively; C+, positive control (rNS1 and GAPDH). The results presented here are representative of those from three independent experiments.

FIG 3

Confirmation of the NS1-GAPDH interaction by a cross-linking assay using the cross-linking reagent EGS. Different amounts of the rNS1 (1.5 or 3 μg) and GAPDH (3, 4, or 6 μg) proteins were incubated in the absence or presence of 10 mM EGS. The samples were incubated for 30 min at room temperature prior to analysis by Western blotting using anti-GAPDH (A) or anti-NS1 (B) antibodies. As a control, both GAPDH (6 μg) and rNS1 (3 μg) were incubated alone in either the absence or the presence of 10 mM EGS, followed by Western blot analysis with their respective antibodies. The results presented here are representative of those from two independent experiments.

FIG 4

Molecular docking of NS1 and GAPDH. The constructed model of dimeric DENV2 NS1 (blue) was used as the system's ligand, and both the GAPDH monomer (pink) (A) and the GAPDH tetramer (B) were set as the receptors. GAPDH catalytic residues Cys152 and His179 are marked in green. In both cases, the interaction interface was the same and did not obscure the GAPDH catalytic site. (C) Expanded view of the contact surface region between dimeric NS1 and GAPDH (shown in panel A) in which the amino acid residues of NS1 involved in the contact are shown in orange. (D) The same view as that in panel C, except that the dimeric NS1 structure is rotated 90° clockwise and the GAPDH structure is not represented for clarity. (E) Three-dimensional model of dimeric DENV2 NS1_mut (Asn10Ala, Lys11Ala, Gly161Asp, and Val162Asp), represented in blue, interacting with the monomeric state of GAPDH (orange). The structure of GADPH (pink) is in the same orientation as that which occurs when it interacts with wild-type NS1. The mutated residues are colored in green. (F) The same view as that in panel E with the structure rotated 90° counterclockwise and with deletion of the NS1_mut protein for clarity. Docking structures were generated by the web-based server ClusPro (version 2.0) and displayed with PyMOL.

FIG 5

Effects of NS1 binding on GAPDH activity. Different amounts of NS1 (0.5 or 2 μM) were used to evaluate the effect of NS1 binding on human GAPDH activity. The negative control was performed with 2 μM BSA. (A) GAPDH activity is represented as the decrease of the relative fluorescence intensity of NADH after its conversion to NAD⁺. The data were obtained in three independent experiments, and the bars indicate the standard errors. F/F₀ is the relative fluorescence intensity, in which F₀ and F correspond to the fluorescence intensity at the initial time point and each different time point, respectively. Measurements were performed without (W/o) substrate (circles), without NS1 (squares), and in the presence of 2 μM BSA (triangles), 0.5 μM NS1 (diamonds), or 2 μM NS1 (solid triangles). The reaction mixtures were incubated at 25°C for 20 min with 50 mM Tris-HCl (pH 7.4), 2 mM MgCl₂, 1 mM ATP, 1 mM EDTA, 0.25 mM β-NADH, 13 units/ml PGK, 33 0.4 nM human GAPDH, and 5 mM 3-phosphoglycerate. (B) The data in panel A were converted to percentages using a representative time point of each reaction (10 min) to better show the increase in GAPDH activity as a function of the NS1 concentration. The data were analyzed for statistically significant differences using a two-tailed, unpaired Student's t test. *, P < 0.05; **, P < 0.01.

FIG 6

Colocalization of iNS1, GAPDH, and calreticulin proteins visualized by confocal microscopy. Mock- or DENV2-infected BHK-21 cells and BHK-21 cells transfected with pcDNA3.1 or pcDNA3.1-NS1 were cultivated for 48 h prior to immunofluorescence analysis. The cells were fixed in formaldehyde prior to permeabilization with Triton X-100. After blocking, the cells were labeled with anti-GAPDH rabbit monoclonal (red stain), anti-GAPDH mouse monoclonal (green stain), anti-NS1 mouse polyclonal (green stain), and anticalreticulin rabbit monoclonal (calreticulin is an ER marker; red stain) antibodies. DAPI labeling was used to identify the nuclei. In the merged images, the yellow staining corresponds to the sites of iNS1 and GAPDH or GAPDH and calreticulin colocalization. All images are representative of those from three independent experiments. For better visualization, the cells identified by a white arrowhead were enlarged, and the images are displayed as insets. Bars, 20 μm.

FIG 7

Measurement of intracellular GAPDH activity. The kinetics of intracellular GAPDH activity were monitored by determination of the increase in absorbance at 340 nm every 10 s for 60 s, using 30 μg of the cell extract as the source of the enzyme. The conversion of absorbance units to NADH production was done by use of the molar absorptivity of NADH (6.22 mM⁻¹ cm⁻¹). Mock- or DENV2-infected and pcDNA- or pcDNA-NS1-transfected BHK-21 cells were cultivated for 24 h (A and D, respectively) or 48 h (B and E, respectively) prior to analyzing the enzyme activity. At the end of the curve, the concentration of NADH was converted into GAPDH activity, where 1 unit of enzyme activity corresponds to the reduction of 1 μM β-NAD/min. The GAPDH activity was determined for both the infected (C) and the transfected (F) BHK-21 cells. The data were analyzed for statistically significant differences using a two-tailed, unpaired Student's t test. *, P < 0.05. (G to J) The levels of NS1 and GAPDH expression from the transfected (G and I) or infected (H and J) BHK-21 cells were determined by Western blotting (WB) using an anti-NS1 antibody. Approximately 30 μg of the cell extract from each condition was used to determine iNS1 expression, and 15 μl of the culture supernatant was used to determine sNS1 levels in the transfected cells. The loading control for the Western blot assay was anti-β-actin. The Western blots are representative of those from three independent experiments, and the error bars indicate the standard deviations from three independent experiments.