Figure 1.
(a) Model 1 from 1KBH, the structure of NCBD in complex with ACTR [40]; (b) Chain C from 1ZOQ, the structure of NCBD in complex with IRF-3 [43] (the asymmetric unit comprises two NCBD molecules and two IRF-3 molecules); (c) Model 1 from 2KKJ, the core unbound conformer of NCBD [46]. In this work, we choose these conformations to represent the three PDB structures. Helical regions I, II and III are indicated for each structure. Beneath is the amino acid sequence of NCBD. We follow reference [46] in numbering the residues of NCBD sequentially from 1 to 59. This corresponds to residues 2059–2117 of mouse CBP in 2KKJ and 1KBH (also numbered 48–106 in 1KBH). Structure 1ZOQ contains residues 2065–2111 of human CBP, whose sequence matches residues 2066–2112 of 2KKJ and 1KBH; therefore, 1ZOQ contains residues 8–54 according to the numbering system that we use.
Figure 2.
Helix propensity, expressed as the probability of a given residue being part of a sequence of three consecutive residues whose dihedral angles place them in the -helical region of conformational space.
The helical region is defined by and
[70], [36]. The positions of the helices in three PDB structures are indicated beneath the graph: 2KKJ (the core conformer of the unfolded ensemble [46]), 1KBH (the ACTR-bound structure [40]) and 1ZOQ (the IRF-3-bound structure [43]).
Figure 3.
Distribution of root-mean-squared deviations of regions I, II and III, from the 2KKJ conformation.
The top left panel shows the probability distributions for the three regions individually. Probabilities that rmsds are below 0.3 nm (vertical dashed line) are given as percentages. Remaining panels show the two-dimensional distributions of pairs of regions together; probabilities that both rmsds are below 0.3 nm (area bounded by dashed lines) are given as percentages.
Figure 4.
Eight conformations visited during the simulation.
Panels on the left mark these in red, at positions given by the single-region rmsd measures used in Figure 2. The conformations were drawn from four continuous trajectories, also plotted on the panels (only the 304 K parts of each trajectory are shown). A: conformation at 59.36 ns (trajectory 13); B: 178.08 ns (trajectory 13); C: 159.73 ns (trajectory 35); D: 180.67 ns (trajectory 35); E: 72.98 ns (trajectory 37); F: 236.63 ns (trajectory 37); G: 60.39 ns (trajectory 42); H: 215.58 ns (trajectory 42).
Figure 5.
Contacts in 1KBH structure: NCBD bound with ACTR [40].
The top panel contains intra-protein contacts of NCBD: contacts shown in blue involve helix II, while contacts in green do not. The bottom panel contains the inter-protein contacts. Contacts shown in red involve helix II of NCBD, while those in solid orange involve helix I or helix III. Contacts shown in outline involve only non-helical regions of NCBD. A pair of residues is considered to be in contact if any pair of non-hydrogen atoms from the two residues are within 4.5 Å of one another.
Figure 6.
Probability distribution for of NCBD, computed with a G
model.
The top panel shows results for isolated NCBD, while the bottom panel shows results for NCBD in the presence of ACTR. Grey shadow: baseline model; blue lines: weakened intrahelical interactions in helices I and III; green lines: weakened intrahelical interactions in helix II; red lines: weakened interactions between helix II and the other helices.
Figure 7.
Cartoon (a) shows how the vectors relate to the molecule: unit vectors ,
and
follow the principal axes of the
atoms in each region, whether or not helices are formed. Cartoon (b) shows how the coordinate system is derived from the vectors. A plane is defined by its perpendicular
. The elevation of
above this plane, deemed to be positive in the direction of the perpendicular, is denoted by
, while
is the azimuth of
, with the reference direction being
and with positive sense given by rotation from
towards
by the shortest route. The main panel shows the probability distribution of tertiary structure as a function of
and
. The ACTR-bound structure is marked by +, the IRF-3-bound structure by
, and the core unfolded conformation by
. Since
points to the surface of a unit sphere, the bins cover a larger range of
at high and low
, to maintain approximately constant bin size. This leads to a stretched appearance for these bins in this equirectangular projection.
Figure 8.
(Left panel) Proportion of NMR constraints satisfied (within a certain tolerance), as a function of tolerance. Blue lines represent medium range constraints (2–4 residues); orange lines represent long range constraints (greater than 4 residues). In each case, the simulation results (solid lines) can be compared with the initial randomised set of conformations (dotted lines). (Right panels) Chemical shifts calculated with SPARTA+ from the simulations (green) and the 2KKJ structure (red) are compared with the experimental values (black) [46].
Figure 9.
Comparison of simulation results with SAXS data.
Black dots are the experimental data from reference [46]. The solid red line shows the mean of results computed from twenty randomly chosen frames of the simulation using CRYSOL [62]; dashed red lines indicate the confidence interval for the mean. Since we are interested in relative rather than absolute intensities, an offset has been added to the calculated intensities, to best fit the experimental data.