1. Structural Biology and Molecular Biophysics
Download icon

Structural analyses at pseudo atomic resolution of Chikungunya virus and antibodies show mechanisms of neutralization

  1. Siyang Sun
  2. Ye Xiang
  3. Wataru Akahata
  4. Heather Holdaway
  5. Pankaj Pal
  6. Xinzheng Zhang
  7. Michael S Diamond
  8. Gary J Nabel
  9. Michael G Rossmann  Is a corresponding author
  1. Purdue University, United States
  2. National Institute of Allergy and Infectious Diseases, National Institutes of Health, United States
  3. Washington University School of Medicine, United States
Research Article
Cite this article as: eLife 2013;2:e00435 doi: 10.7554/eLife.00435
7 figures and 9 tables

Figures

Structure of the CHIK VLPs.

(A) Surface-shaded figure of ectodomain (left) and surface-shaded figure of nucleocapsid (right), colored according to the radial distance from the center of the virus. White triangles indicate one icosahedral asymmetric unit. (B) Cross-section of the virus showing density above 1.5 σ also colored according to the radial distance from the center of the virus. (C) Resolution of β-strands in the E1 domain III.

https://doi.org/10.7554/eLife.00435.003
Diagrammatic representation of contacts between amino acid residues in adjacent subunits.

(A) Diagrammatic organization of the E1 and E2 subunits according to T = 4 icosahedral symmetry. The white numbers show the sequence in which the four independent subunits were generated. The black capital letters indicate the sequence of generating seven of the icosahedral asymmetric units. Icosahedral symmetry elements are shown as filled triangles and ellipses. Quasi-symmetry elements are shown as red outlined triangles and ellipses. (B) Contacts between the E1 (blue) and E2 (green) molecules. (C) Contacts between NCPs. (B) and (C) show the number of contacts between the indicated molecules. The quasi T = 4 related positions #1, #2, #3 and #4 are indicated, prefixed by their icosahedral symmetry identification A, B, C and D. The center of the “i3” icosahedral spike is indicated by a filled red triangle and the center of the “q3” quasi-3-fold spike is indicated by an outlined red triangle.

https://doi.org/10.7554/eLife.00435.005
Fit of the atomic structure into the cryoEM density.

(A) Fit of the E1E2 heterodimer, (B) Fit of the capsid protein. The quasi equivalent subunit closest to the icosahedral 5-fold axis was chosen for display.

https://doi.org/10.7554/eLife.00435.006
CHIKV VLP Fab complexes.

Right: “Difference maps” showing the surface structure of the Fab molecules. The difference maps were produced by setting all density within 1.7 Å of an atom in the VLP structure to zero. Left: “Road maps” showing the projected surfaces of the VLP-Fab complexes for one (triangular) icosahedral asymmetric unit. The coloring is according to the radial distance of the surface from the center of the VLP. The footprints of the Fabs are shown in yellow. The radial projection of the whole Fab molecules onto the surface of the VLP is shown with white contours representing the height of the projected density. The mosaic background shows the amino acids that form the viral surface.

https://doi.org/10.7554/eLife.00435.014
Inhibition of CHIKV by IgG and Fab fragments.

Neutralizing activity of intact MAb and Fab fragments against CHIKV is shown for (A) CHK-9, (B) CHK-152, (C) m10 and (D) m242, as determined by the reduction in the number of focus-forming units (FFU) on Vero cells. MAbs and Fab fragments were mixed with 100 FFU of infectious CHIKV (strain La Reunion 2006 OPY-1) for one hour at 37°C before infecting Vero cells. Each data point is the average of three independent experiments performed in triplicate. Error bars represent standard deviations.

https://doi.org/10.7554/eLife.00435.015
Fit of the Fab structures (blue) into the cryoEM density “difference” maps (hatched grey surface) calculated as described for Figure 4.

Shown also is how the Fab molecules bind to the E1(red)-E2 (green) heterodimer.

https://doi.org/10.7554/eLife.00435.017
“Road maps” showing footprint of four neutralizing Fabs on the VLP surface at position #3 as defined in Figure 2.

In order to differentiate between amino acids in different quasi 3-fold related subunits, their identity is defined as the amino acid sequence number in E1 + 2000, 3000, and 4000, and in E2 + 2500, 3500, and 4500 for positions #2, 3, and 4, respectively (see Figure 2). The surface is colored according to radial distance from the center of the VLP. The A, B, and D domains of E2 are bounded by a black, white and dashed white line, respectively. Residues in the putative receptor binding site on domain A of E2 are bounded by a yellow dashed line. The footprint of the Fabs onto the VLP surface is outlined in yellow.

https://doi.org/10.7554/eLife.00435.018

Tables

Table 1

Quality of fitting the atomic structural fragments into the cryoEM density

https://doi.org/10.7554/eLife.00435.004
A) Average height of the densities at the atomic positions (sumf) (Rossmann et al., 2001) on fitting the E1E2 heterodimer CHIKV structure into the cryoEM map at four quasi equivalent positions
All atomsT = 4 fittingIndependent domain fitting
#1#2#3#4Average#1#2#3#4Average
E1*I16.115.817.017.716.618.217.018.519.918.5
II15.717.218.816.717.018.018.619.018.518.6
III15.714.215.218.015.718.715.918.220.718.3
E2*A15.518.019.017.517.418.817.819.919.418.8
B10.09.49.69.69.810.911.311.712.411.7
C15.315.419.019.116.821.019.920.222.420.5
D13.316.415.816.115.416.417.216.016.816.5
Average14.915.817.016.816.117.817.318.219.018.1
Main chain atoms onlyT = 4 fittingIndependent domain fitting
#1#2#3#4Average#1#2#3#4Average
E1*I20.518.421.522.920.823.221.622.626.323.4
II17.020.923.020.220.321.723.323.522.822.8
III18.915.317.522.318.522.317.922.626.622.4
E2*A16.521.122.219.419.822.020.323.422.622.1
B10.79.49.910.110.010.811.213.013.412.1
C18.719.527.426.923.127.727.127.430.828.3
D13.418.718.518.517.319.419.819.119.019.3
Average16.517.620.020.018.521.020.221.723.121.5
B) Average height of the densities at the atomic positions (sumf) (Rossmann et al., 2001) on fitting the CHIKV homology model of the capsid protein into the cryoEM map at the four quasi equivalent positions
All atomsT = 4 fittingIndependent domain fitting
#1#2#3#4Average#1#2#3#4Average
D114.913.313.916.114.616.114.615.417.315.8
D217.017.418.118.117.618.319.117.419.018.6
Average16.015.416.017.216.117.417.116.518.217.3
Main chain atoms onlyT = 4 fittingIndependent domain fitting
#1#2#3#4Average#1#2#3#4Average
D118.516.517.818.917.920.416.919.222.719.8
D221.321.521.423.221.922.824.220.425.323.2
Average19.919.019.621.119.921.821.019.924.121.6
C) Average height of the densities at the atomic positions (sumf) (Rossmann et al., 2001) on fitting the CHIKV model of the transmembrane protein into the cryoEM map at the four quasi equivalent positions
All atomsT = 4 fittingIndependent domain fitting
#1#2#3#4Average#1#2#3#4Average
E113.714.613.218.615.013.914.414.717.515.1
E214.712.811.817.514.215.114.113.718.915.5
Average14.413.512.317.914.614.514.214.218.215.3
Main chain atoms onlyT = 4 fittingIndependent domain fitting
#1#2#3#4Average#1#2#3#4Average
E116.217.616.024.118.515.717.316.723.918.4
E216.715.013.620.716.517.716.715.823.718.5
Average16.516.014.422.017.516.717.016.223.818.4
  1. *

    Domain definition: E1 I (residues 1-36, 132-168, 273-293), II (residues 37-131, 169-272), III (residues 294-393). E2 A (residues 16-134), B (residues 173-231), C (residues 269-342), D (residues 7-15, 135-172, 232-268).

  2. Domain definition: D1(residues 119 to 183), D2(residues 184 to 267).

  3. Domain definition: E1 (residues 394-439), E2(residues 343-423).

Table 2

Distances in Å between N and C termini of fitted domains before regularization

https://doi.org/10.7554/eLife.00435.007
Quasi equivalent positions
Domains#1#2#3#4
E1I 36 to II 374.65.65.53.5
II 131 to I 1324.84.14.44.1
I 168 to II 1692.86.45.45.1
II 272 to I 2735.84.73.92.1
I 293 to III 2945.46.93.75.0
E2D 15 to A 163.86.35.75.9
A134 to D1353.15.04.42.3
D172 to B1733.48.83.05.7
B231 to D2326.27.75.97.2
NCPD268 to C2695.15.24.14.0
(D1)183 to (D2)1841.22.21.61.6
Table 3

Hinge angle change between domains in different heterodimers

https://doi.org/10.7554/eLife.00435.008
Hinge angle change in E1 between domain I and domain II (in degrees)
Position #1Position #2Position #3Position #4
Crystal structure-14.4-6.67.5-8.3
Position #18.09.36.8
Position #23.12.1
Position #3-3.3
Hinge angle change in E1 between domain I and domain III (in degrees)
Position #1Position #2Position #3Position #4
Crystal structure-7.3-8.0-9.42.1
Position #1-3.510.37.7
Position #213.09.2
Position #37.6
Hinge angle change in E2 between domain A and domain B (in degrees)
Position #1Position #2Position #3Position #4
Crystal structure-8.410.422.2-22.3
Position #1-14.125.5-29.7
Position #226.6-28.6
Position #3-26.3
Hinge angle change in E2 between domain A and domain C (in degrees)
Position #1Position #2Position #3Position #4
Crystal structure-11.5-10.4-7.6-17.1
Position #1-4.8-8.6-7.3
Position #27.28.3
Position #3-14.8
Table 4

RMS deviation in Å between quasi equivalent Cα atoms

https://doi.org/10.7554/eLife.00435.009
E1#1#2#3#4Xtal
#10.001.211.251.081.69
#20.001.281.271.52
#30.001.101.68
#40.001.17
Xtal0.00
E2#1#2#3#4Xtal
#10.002.181.791.901.50
#20.002.342.382.17
#30.002.281.84
#40.001.92
Xtal0.00
E1&E2#1#2#3#4Xtal
#10.001.941.651.682.28
#20.002.162.142.24
#30.001.872.37
#40.001.88
Xtal0.00
Capsid#1#2#3#4Xtal
#10.000.490.730.710.48
#20.001.090.940.71
#30.000.630.56
#40.000.52
Xtal0.00
Table 5

Correlation between amino acid sequence and cryoEM density

https://doi.org/10.7554/eLife.00435.010
A) Correlation between cryoEM densities in the four quasi equivalent positions of the E1 and E2 glycoproteins.
All atomsT = 4 fittingIndependent domain fitting
1234B1234B
11.000.170.230.080.221.000.400.440.400.21
21.000.280.380.301.000.410.300.28
31.000.400.281.000.360.28
41.000.191.000.15
B1.001.00
Main chain atoms onlyT = 4 fittingIndependent domain fitting
12#34B1234B
11.000.110.200.110.191.000.400.410.410.22
21.000.310.300.301.000.360.320.26
31.000.410.271.000.370.26
41.000.161.000.12
B1.001.00
B) Correlation between cryoEM densities of the four quasi equivalent capsid proteins.
All atomsT = 4 fittingIndependent domain fitting
#1#2#3#4#1#2#3#4
#11.000.410.180.171.000.200.190.30
#21.000.260.251.000.450.42
#31.000.131.000.41
#41.001.00
Main chain atoms onlyT = 4 fittingIndependent domain fitting
#1#2#3#4#1#2#3#4
#11.000.370.160.261.000.220.170.33
#21.000.310.271.000.340.43
#31.000.151.000.31
#41.001.00
C) Correlation between cryoEM densities of the four quasi equivalent E1&E2 TM and endodomain regions.
All atomsT = 4 fittingIndependent domain fitting
#1#2#3#4#1#2#3#4
#11.000.430.170.411.000.370.340.35
#21.000.240.381.000.440.34
#31.000.201.000.46
#41.001.00
Main chain atoms onlyT = 4 fittingIndependent domain fitting
#1#2#3#4#1#2#3#4
#11.000.280.080.271.000.230.110.26
#21.000.170.431.000.400.36
#31.000.141.000.50
#41.001.00
Table 6

Inter-molecular contacts (See methods for descriptions)

https://doi.org/10.7554/eLife.00435.011
Spike positionTotalHydrophobicPossible H bondsPossible salt bridges
A) Number of atom-to-atom contacts between E1-E2 in heterodimer
i3#123499 (42%)34 (13%)0
q3#2254112 (44%)25 (10%)0
#317959 (33%)33 (18%)0
#4302118 (39%)39 (13%)4 (1%)
Crystal861 (1%)32 (37%)1 (1%)
B) Number of atom-to-atom contacts of E2 with other E2s within a spike
i3#122946 (20%)30 (13%)82 (36%)
q3#2708 (11%)10 (14%)33 (47%)
#34714 (30%)6 (13%)10 (21%)
#474118 (23%)24 (46%)7 (9%)
C) Number of atom-to-atom contacts between glycoprotein spikes
q3 to q3A#3 - D #412342 (34%)21 (17%)7 (5%)
A#4 - D#413052 (40%)18 (14%)1 (1%)
i3 to q3A#2 - B#15827 (47%)7 (12%)0
A#1 - A#28829 (33%)19 (22%)0
A#1 - A#3118 (73%)2 (8%)0
D) Number of atom-to-atom contacts between E1 and E2 within the spikes.
q3A#2 - A#33016 (34%)5 (17%)0
A#3 - A#4208 (40%)3 (14%)0
i3A#4 - A#25625 (47%)11 (12%)0
A#1 - B#1224112 (50%)21 (9%)22 (10%)
E) Number of atom-to-atom contacts between capsid proteins*
A#2 - B#17920 (25%)13 (16%)15 (19%)
A#1 - A#3113 (27%)1 (9%)3 (27%)
A#3 - D#218351 (28%)22 (12%)38 (21%)
A#4 - D#4195 (26%)2 (11%)8 (42%)
A#1 - B#1209 (45%)2 (10%)2 (10%)
F) Number of atom-to-atom contacts between E1 & E2 in TM region
#110060 (60%)3 (3%)0
#28243 (52%)8 (10%)0
#32012 (60%)1 (5%)0
#42824 (86%)00
  1. *

    See Figure 2.

Table 7

Spike radii of gyration

https://doi.org/10.7554/eLife.00435.012
Distance from particle center (Å)Radii of gyration (Å)# of atoms
i3q3SINV* crystali3q3SINV* crystal
25558.361.063.3423462
26553.557.354.9131126118
27541.146.942.6125122101
28530.133.130.1849080
29533.234.233.8858091
30533.331.931.3908982
31530.731.224.5708092
32526.725.94147
33531.029.011
Overall40.542.441.9669669626
  1. *

    Sindbis virus (SINV)

Table 8

Comparison between VEEV and CHIK VLPs.

https://doi.org/10.7554/eLife.00435.013
E1-E2TMCapsid
A1A2A3A4A1A2A3A4A1A2A3A4
d in Å2.63.53.52.24.82.32.13.09.67.79.09.7
κ in degrees2.33.63.73.45.48.210.410.09.78.010.011.7
dCα in Å2.72.52.42.12.62.22.52.01.11.21.21.2
num69668070971411510099101149149149149
% identityE1 = 56%, E2 = 34%TM = 24%Capsid = 62%
  1. Differences between the positions of the centers of mass (d) and orientation (κ) relative to the icosahedral symmetry axes of the (E1-E2) heterodiners, the trans membrane (TM) helices and capsid proteins at positions A1, A2, A3 and A4 (see Figure 2). The RMS distances (dCα) are given between the number (num) of equivalent Cα positions of the superimposed molecules A1 to A4. The percentage of identical amino acids in the aligned proteins is given on the last line of the table

Table 9

Data collection and refinement statistics for the m242 and CHK-9 Fab molecules

https://doi.org/10.7554/eLife.00435.016
Fab m242Fab CHK9
Data collection
X-ray source23-ID-B23-ID-B
Wavelength (Å)1.031.03
Resolution (Å)3.13
Space groupP21212C2
Unit cell (Å)a = 137.0, b = 89.1, c = 94.4a = 87.9, b = 57.7, c = 118.1
Unique reflections21,11535,876
Average redundancy3.75.8
I/σ*17.5 (4.9)25.0 (5.0)
Completeness (%)98.3 (96.4)99.6 (99.9)
Rmerge (%)10.1 (23.4)5.5 (39.2)
Refinement
Resolution (Å)3.11.8
Rworking (%)29.3 (34.5)18.7 (23.3)
Rfree (%)§33.1 (34.3)22.0 (27.7)
rmsd bonds (Å)0.0050.01
rmsd angels (°)0.850.91
# of residues Ramachandran disallowed01
  1. *

    Values in parentheses throughout the table correspond to the outermost resolution shell

  2. Rmerge = Σ| I - <I> | / Σ I, where I is the measured intensity of reflections

  3. Rworking = Σ||Fobs| -|Fcalc|| / Σ|Fobs|

  4. §

    Rfree has the same formula as Rworking except that calculation was made with the structure factors from the test set

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)