Site Map   Home   Java Mage   Clash report

All-Atom Clash Report for sulfate-binding protein (1sbp)

As described in the Model Assessment section, Reduce was run on this file to add hydrogens and correct NQH orientations, and then Probe was run to produce the kinemage below, with contacts shown only for the serious atomic clashes that overlap more than 0.4 Å (red spikes). Drag with the mouse to rotate in 3D, and hit the "h" key to toggle a help list on and off (active only after you click somewhere in the Java window).

Although relatively old (1993), this structure is a bit better than average for 1.7 Å resolution, with only 3 really bad places (around a Lys and two Thr) and a scattering of smaller ones just over the threshold. Turn on the "details" button and use the Views pulldown menu to see closeups of the 3 problem areas. Lys 269 has a bad rotamer that clashes with the backbone. Both Thr look as though they were fit backwards into what was probably rather straight, bar-shaped electron density, since they both have eclipsed Chi1 angles and their methyls are in positions that would be more appropriate for the OH groups. [You can also use the pickcenter tool and zoom in closer to look at specific areas. For instance, turn on "waters" in the overview and zoom in on the clash (top center) of HOH 350 with a methylene H of Arg 261; perhaps in this case the density is actually due to an alternate Arg conformation rather than a water.]

Scroll down to see the Java kinemage below with a fuller analysis of one of the Thr cases, including a proposed correction of the problem.

Repair of clashing Thr 32 in 1sbp:

Animate to see the effects of the proposed change -- the original environment which is much too tight and polar for the Thr methyl (red clash spikes) is just right for the Og, with two Hbonds (pale green lenses of dots) and good contacts otherwise. Because even the non-H atoms clashed in the original conformation, refinement optimized them as well as it could by distorting bond angles; the N-Ca-Cb angle and the Og-Cb-Cg angle deviate from tetrahedral by more than 10°. Such problems are large enough to spot by eye (try looking down on the Ca-Cb bond), but are confirmed by choosing the "measure" tool* and clicking on the 3 atoms of the bond. Try it. Note also that this is a very bad rotamer, with Chi1 almost exactly eclipsed. Temporarily turn on both original and fixed versions together, to see that they occupy similar regions in space and could both approximately fit a suboptimal region of electron density.

In order to look for a better conformation, Prekin was called remotely from Mage to construct an ideal-geometry Thr 32 with rotatable angles and selectable rotamers; then Probe was called remotely to calculate contacts while the rotamers were tried in turn and the best one was further optimized by the user. [That method is described in Word et al. 2000, Protein Sci. 9, 2251-9; it requires the regular Mage running in Unix or Linux, but the results are shown here in Java Mage, in the second animation state.] Alternatively, the sidechain can be refit by the crystallographer in O or XtalView, with Probe called remotely to show the clashes. In either case, the new coordinates would be put back into refinement for validation and further optimization. Someone else analyzing this structure, however, even without the structure factors, can be fairly sure that the new conformation is more likely to be right.

* For those unfamiliar with the measure tool in Mage, it works as follows: At all times, the information line at the bottom of the graphics window shows the pointID of the point (or atom) last picked and the distance between the last two points picked. If "measures" is turned on, then the third field on the information line is the angle defined by the last 3 points picked, and the final field is the torsion angle defined by the last 4 points picked (with markers shown on those points). Thus, one can proceed along the backbone, picking successive atoms and reading out the phi, psi, and omega angles. Remember, however, that the torsion angle is the one around the CENTRAL bond of the 4 atoms that define it, so that the readout lags behind where you last picked.

To include Java kinemages in your own webpages, see the MageJava instructions.

Kinemages

Java Mage

Software:

Validation

Rotamers

FTP

2D Gallery

Lab Info

Display

Contacts

Utility