autoBUSTER Example Application: forcing a zinc ion site to be tetrahedral using utility distance and angle restraints


These are preliminary notes on an initial examination in January 2008. They need to be worked up into a proper tutorial.


2a0f seems a likely target with problematic zinc sites

  • structure of a Aspartate carbamoyltransferase at 2.9 Å resolution.
    • pdb quotes Rwork21.7%, free 27.9%, P321 a=b=120.45Å c=155.24Å
    • higher resolution related structure with deposited SF 1d09 at 2.4Å similar cell. Molprobity %Rama Out=1.63%, score 3.21.
    • Also 2air 2.0Å resolution H3 cell Molprobity %Rama Out=2.51%, score 3.32
  • initial refinements show this is a fairly awful structure. B factors of B, C and D chains high. Correlation much worse for this than the A chain?
run Rwork/Rfree Molprobity %Rama Out/score
initial pdb 0.250/0.250(1) 8.96%/4.39
2a0f_001_control.log 0.2236/0.3269 11.39%/4.70
2a0f_002_autoncs.log 0.2743/0.2966 5.86%/4.39
2a0f_003_autoncs_target1d09.log 0.2843/0.2881 2.43%/3.75
2a0f_004_autoncs_break_SSBOND.log 0.2521/0.3302 5.97%/4.53
2a0f_005_autoncs_break_SSBOND_metal_restraints.log 0.2563/0.3259 6.42%/4.50
2a0f_006_autoncs_target1d09_break_SSBOND.log 0.2785/0.2970 2.32%/3.62
2a0f_007_autoncs_target1d09_break_SSBOND_metal_restraints 0.2785/0.2964 2.31%/3.61
  • Using the restraints produces marginally improved refinement stats.

(1) no free set in data

  • Observation automated weight procedure mucked by the target cutins - reason produces big movements that effect bond length. Need to advise setting constant X-ray weight until structure stabilizes.
run final X-ray weight
2a0f_001 1234
2a0f_002 738
2a0f_003 100
2a0f_004 2025
2a0f_005 1620
2a0f_006 184
2a0f_007 184

  • Tuesday 28-01-2008
    • look at the zinc sites for the runs 004, 005 - with autoncs but no target.
  • 004- breaking spurious disulfide and refinement sorts out a not to bad tetrahedral site around B site zinc even in the absence of restraints. Distances of SG 2.49,2.49,2.78,3.20.
    • problem in B site is a nonsense water W296 that gets to within 2.34 of zinc. This starts in 2a0f with 2.15 of B141 N and with five atoms within 2.7Å
    • Restraints have more effect in D site pulling D141 CYS into coordination - density really poor around this this though.
    • May be that D chain improperly placed?

  • Water positions in 2a0f very suspect and likely to prevent sensible movement -

for the metal sorting example B site is mucked up by W296.

  • so do a further set of runs same as before with a constant X-ray weight of 1200 and no water:
run Rwork/Rfree Molprobity %Rama Out/score
initial pdb 0.258/0.263(1) 8.96%/4.39
2a0f_nohoh_001_control 0.2345/0.3413 13.94%/4.81
2a0f_nohoh_002_autoncs 0.2532/0.3239 7.52%/4.55
2a0f_nohoh_003_autoncs_target1d09 0.2769/0.2894 2.10%/3.71
2a0f_nohoh_004_autoncs_break_SSBOND 0.2501/0.3351 10.51%/4.70
2a0f_nohoh_005_autoncs_break_SSBOND_metal_restraint 0.2516/0.3338 10.84%/4.68
2a0f_nohoh_006_autoncs_target1d09_break_SSBOND 0.2686/0.2981 2.88%/3.73
2a0f_nohoh_007_autoncs_target1d09_break_SSBOND_metal_restraints 0.2635/0.3027 2.32%/3.52
  • Produce picture of metal sites for runs 000 (original pdb map), 004, 005:

2a0f_znb_orig.png 2a0f_znb_fix_no_rest.png 2a0f_znb_fix_rest.png

    • contoured at 1.5 sigma
    • for the record look at ligand-metal-ligand angles for no restraints or restrained:
ligand B114 SG B138 SG B141 SG
B109 SG 113,110 112,108 118,115
B114 SG 95,98 107,108
B138 SG 111,116
  • For comparison record density around ZN site in D chain - very weak density, whole chain not correctly placed or disordered? 2a0f_znd_fix_rest.png

(Impression of 2a0f is that A chain is well placed this is next to ZN binding domain from B chain - reasonable but as one works along C chain and D chain things get worse and worse. Chains are not placed at all reasonably.) Maybe worth looking to see whether it is possible to do properly someday.


Page by Oliver Smart original version 22 June 2008. Address problems, corrections and clarifications buster-develop@globalphasing.com