Figures and data
![](https://prod--epp.elifesciences.org/iiif/2/90762%2Fv1%2Fcontent%2F487425v2_fig1.tif/full/max/0/default.jpg)
Composition of the secreted NS1 from DENV-infected Vero cells.
DENV 2 WT (Chan et al., 2019) cell culture supernatant was filtered, supplemented with protease inhibitor cocktail and 0.05% sodium azide, concentrated using a 100 kDa MWCO Vivaflow cassette and purified using 56.2 anti-NS1 antibody immunoaffinity chromatography as described previously (Gagnon et al., 2012). The eluted isNS1wt was dialysed against PBS, concentrated, and stored at –80°C until further use. (a) Schematic of isNS1 purification to illustrate the samples used for gel analyses. % NS1 is measured by the total amount of NS1 (quantified using the anti-NS1 ELISA kit (Bio-rad) as a percentage of total protein (quantified using the Bradford assay) found in each sample. Details of the % enrichment in NS1 along the purification process is as shown in Supplementary Fig S1 B. (b) Coomassie blue detection of proteins from Crude, Wash and Elute immunoaffinity fractions for isNS1wt, with the recombinant sNS1 (rsNS1) obtained from Shu et al (2022) as a positive control, after separation on a 10% Native-PAGE gel (left). The Crude and Elute fractions contain 1 µg of total protein. The Wash fraction contains approximately 100 ng of total protein in maximum well volume of the gel. The same set of samples were also subjected to a western blot detection of NS1 using 56.2 anti-NS1 antibody after separation on a 10% Native- PAGE (right). The Crude and Elute fractions contain 500 ng of total protein. The Wash fraction contains approximately 100 ng of total protein in maximum well volume of the gel. (c) Coomassie blue detection of proteins from Crude, Wash and Elute immunoaffinity fractions for isNS1wt and rsNS1 (Shu et al., 2022), after separation on a 4-20% reducing SDS-PAGE gel. The Crude and Elute fractions contain 1 µg of total protein. The Wash fraction contains approximately 100 ng of total protein in maximum well volume of the gel. Similarly, the same set of samples were also subjected to a western blot detection of NS1 and ApoA1 using 56.2 anti-NS1 antibody or ApoA1 antibody (Biorbyt, orb10643) respectively, after separation on a 4- 20% reducing SDS-PAGE (right). (d) In-gel protein identification of the purified isNS1wt by liquid chromatography mass spectrometry (LC-MS). Proportion of NS1, ApoA1 and other unidentified proteins quantified in total ion intensity, obtained from the following samples: Elute in solution (boxed in blue), 250 kDa gel band (boxed in purple), 50 kDa gel band (boxed in red) and 25 kDa gel band (green). The boxed gel bands are from representative gels showing the different protein species found while the actual gel bands used for protein identification by LC- MS are as shown in Supplementary Fig 2a-b.
![](https://prod--epp.elifesciences.org/iiif/2/90762%2Fv1%2Fcontent%2F487425v2_fig2.tif/full/max/0/default.jpg)
CryoEM analysis of secreted NS1 in complex with antibody Ab56.2. (a)
Size exclusion chromatography was run on a Superdex 200 increase 3.2/300 GL column connected to the ÄKTA purifier with a flow rate of 0.075 mL/min in PBS (pH 7.4) for purified isNS1wt (gray), Ab56.2 (faded red), Fab56.2 (faded blue), isNS1wt:Ab56.2 (red) and isNS1wt:Fab56.2 (blue). A slight leftward shift in elution volume was observed for isNS1wt upon complexing with Ab56.2 and Fab56.2. 2D class averages of (b) isNS1wt:Ab56.2 and (c) isNS1wt:Fab56.2 showing representative sub-class of the Fab56.2:isNS1:HDL particles with black scale bar, 50 Å. The corresponding number of particles and percentages are listed below the respective boxes. Red dashed line boxes highlight two rare views consisting of 1033 particles (3.5%) only seen in isNS1wt:Fab56.2 sample. (d) Model of isNS1wt:Fab56.2 predicted structures rigid body fitted in the CryoEM map of Fab56.2:isNS1:HDL (grey, contoured at 0.14) with correlation value of 0.75 to the fitted regions (map simulated from atoms at 5 Å). isNS1wt is coloured by its three domains, namely the β-roll (orange), wing (blue), and β-ladder (cyan). Fab56.2 is coloured by its heavy chain (dark green) and light chain (light green).
![](https://prod--epp.elifesciences.org/iiif/2/90762%2Fv1%2Fcontent%2F487425v2_fig3.tif/full/max/0/default.jpg)
Secreted NS1 forms free dimers in complex with antibody Ab56.2. (a)
Size exclusion chromatography was run on a Superdex 200 increase 3.2/300 GL column connected to the AKTA purifier with a flow rate of 0.075mL/min in PBS (pH 7.4) for purified isNS1ts (gray), Fab56.2 (faded blue) and isNS1ts:Fab56.2 (red). A similar leftward shift in elution volume wa also observed for isNS1ts upon complexing with Fab56.2. (b) 2D class averages of isNS1ts:Fab56.2 dataset with 326,548 particles picked and separated into 3 distinct populations, HDL spheres, Fab56.2:isNS1:HDL, and free isNS1:Fab56.2. Black scale bar, 50 Å, as indicated. The corresponding number of particles and percentages are listed below the respective boxes. (c) Model of isNS1ts dimer and Fab56.2 predicted structures rigid body fitted in the isNS1ts:Fab56.2 density map (grey, contoured at 0.14) with correlation value of 0.53 (overall, map simulated from atoms at 5 LJ). isNS1ts is coloured by its three domains, namely the β-roll (orange), wing (blue), and β-ladder (cyan). Fab56.2 is coloured by its heavy chain (dark green) and light chain (light green). (d) Density map fitting between isNS1ts:Fab56.2 (grey, contoured at 0.14) to Fab56.2:NS1ts:HDL (yellow, contoured at 0.1) with correlation value of 0.53 (overall) and 0.7137 (on D2NS1 map region only). Inset shows the difference in the pose of Fab from the free NS1 form to the HDL-bound form. (e) Density map fitting between Fab56.2:isNS1wt:HDL (purple, contoured at 0.05) to Fab56.2:NS1ts:HDL (yellow) with correlation value of 0.72 (overall).
![](https://prod--epp.elifesciences.org/iiif/2/90762%2Fv1%2Fcontent%2F487425v2_fig4.tif/full/max/0/default.jpg)
Interaction sites of NS1:ApoA1 complex identification by crosslinking mass spectrometry. (a)
SDS-PAGE analysis of isNS1wt with or without the addition of DSSO crosslinker. (b) The identified crosslinks are visualised on the NS1 and bovine ApoA1 constructs. The intramolecular (NS1:NS1 and ApoA1:ApoA1) crosslinks are in magenta. The intermolecular NS1:ApoA1 crosslinks are in green. (c) The isNS1 residues that are involved in NS1:NS1 interactions are visualised on the NS1 dimer model. The NS1 cartoon model i coloured by its three domains, namely the β-roll (orange), wing (blue), and β-ladder (cyan) with the intramolecular (magenta) and intermolecular (green) crosslinking sites depicted as spheres. (d) The overall model interpretation of NS1:ApoA1 complex within the crosslinker theoretical distance cut-off at =< 30 Å as depicted. ApoA1 dimer cartoon model with its conserved helices as labelled coloured in intervals of grey and light purple. (e) The NS1:ApoA1 dimer model with validated crosslinks were fitted into the cryoEM envelope. (f) SDS-PAGE analysis of crosslinked rsNS1 alone or with human HDL (Lanes 3-5). Non-crosslinked rsNS1 and Human HDL are the control (Lanes 1-2). The crosslinked rsNS1 can be seen in higher oligomers (Lane 3). (g) Identified crosslinks are mapped on the NS1 and ApoA1 constructs, coloured as per panel B.
![](https://prod--epp.elifesciences.org/iiif/2/90762%2Fv1%2Fcontent%2F487425v2_fig5.tif/full/max/0/default.jpg)
Interaction interface of rsNS1c-Fab562 complex characterized by HDX-MS. (a)
Woods differential plot showing the differences in deuterium exchange (y-axis) between Fab562-bound and free NS1c protein across various residues (x-axis) at 10 min labeling timepoint. Negative differences indicate protection against deuterium exchange across NS1c peptides in the presence of Fab562, as compared to free NS1c. A p-value <0.05 is considered as significance threshold, which identified 6 non-significant peptides (grey lines) and 48 protected peptides (blue lines). (b) Cartoon representation of NS1c dimer model showing differences in deuterium exchange at 10 min labeling as indicated in key. Dashed line distinguishes the two monomers in the dimer. Peptide spanning residues 300-310 showing the highest protection are indicated. (c, d) Woods differential plot comparing the differences in deuterium exchange of Fab562 in the presence and absence of NS1c for various peptides of (c) heavy chain, and (d) light chain. A p-value <0.05 is considered as significance threshold, which identified deprotected (red lines), non-significant, and protected peptides as indicated. (e) Cartoon representation of Fab562 model showing deuterium exchange differences at 10 min mapped for Fab562-NS1c complex, as per key. Plots were generated using Deuteros 2.0, while cartoon structures were generated using PyMoL.
![](https://prod--epp.elifesciences.org/iiif/2/90762%2Fv1%2Fcontent%2F487425v2_fig6.tif/full/max/0/default.jpg)
sNS1 associates with ApoA1 in DENV infected mouse and human serum. (a)
AG129 mice (n = 10) were infected with DENV2 NS1 T164S mutant virus and the pooled infected sera collected on day 4 post-infection was subjected to NS1 immunoaffinity purification using anti- NS1 56.2 coupled resin as in Supplementary Fig. 1. 2 mg of the purified eluate was then subjected to Western blot analysis after separation and transfer from Native PAGE for detection of ApoA1 and NS1. ApoA1 was detected using the mouse monoclonal anti-ApoA1 clone 513 (Invitrogen, MIA1404) (left panel) and the oligomeric NS1 was detected using anti-NS1 56.2 IgG clone (right panel). (b) Protein AG resin (Pierce) pre-cleared DENV1 infected patient serum (n = 1) from the CELADEN trial (Low et al., 2014) was immunoprecipitated with 10 or 50 mg of rabbit polyclonal anti-ApoA1 antibody (Biorbyt, orb10643) to detect association between ApoA1 and NS1 by ELISA. The amount of ApoA1 and NS1 in the immunoprecipitated sample was determined by human ApoA1 (Abcam, ab189576) and PlateliaTM NS1 Ag (Bio-Rad) ELISAs.