Homology Modelling of KS Domains

In an effort to rationalise the observed specificity of the KS domains, homology models of each KS were constructed using the CPHmodel server as described in Sect. 2.2.15. Models used a previously published structure of the DEBS KS3-AT3 didomain (PDB 2Q03) as a template, and homology models generated 

The same homology modelling approach was applied to the three psymberin KS domains. PsyA KSl and KS2 both harboured Ala residues at the X positions, yielding a binding pocket similar to that of BaeL KS6 (Fig. 3.8). Interestingly, 

The Met residue highlighted in red is postulated to provide steric bulk in the binding pocket to prevent acylation by P- branched substrates 

Comparison of the KS binding sites from homology models tallies well with the experimental data for these KS domains. In order to test the hypothesis that a bulky residue at the X-Cys position precludes acylation by fl-carbon branched substrates, a M237A mutant was proposed. Simple alteration to the homology model suggested that a Met to Ala mutation would indeed provide additional space in the binding pocket for a f)-branched substrate (Fig. 3.9). 

---