Figure 2. Model Overview.

(A) Quaternary architecture of the model shows that three copies of gp120 and gp41 form the central part of a trimer (3-fold cyclic symmetry). Also three copies of CD4 and 17b are attached to the three gp120s. A small opening through the center along the symmetry axis is also visible. (B) shows the model from a direction orthogonal to the symmetry axis. gp120 is rendered using smooth surface representation. The core is colored pale green and the variable loops are highlighted using different colors. The gp41 model occupies the center of the trimer (all three copies are shown). Notice that all the variable regions (except V3) are away from the central area. (C) provides a different view of the same. (D) shows the fitting of our model into EM density map EMD5020 (Liu et al., 2008). Our model includes the variable loops that are missing in previously reported x-ray structures of the same complex (e.g. 1GC1 (Kwong et al., 1998) as shown in the inset). (E) compares the gp120 model from 1GC1 with our model. Note that the core as well as the V1/V2 stub align almost perfectly. Finally, (F) compares the gp41 model from 4NCO (Julien et al., 2013a). Although we used the same models for the two heptad-repeats (HR1 and HR2), to improve the fitting with EMD5020 while maintaining a favorable interface with gp120, our algorithm moved the HR1 part slightly closer to the center of the trimer. But there is still an opening as seen in (A).