Comparative Analysis of Techniques Oriented on the Recognition of Ligand Binding Area in Proteins - Identification of Ligand Binding Site and Protein-Protein Interaction Area - page 80

Table 4.5 Characteristics of easy and hard predictable binding cavity. The values given are as
follows: volume, accessible vertices, buried vertices, average depth for cavities binding particular
ligand. Values given are calculated according to PDBSum data
FMN NAD +

Easy 1055.27 3376.99

72.36 74.21

11.48 12.90

10.15 12.89

Hard 1141.74 3423.98

41.35 69.78

6.43 13.00

8.10 13.00

Δ H cutoff threshold set at

Table 4.6 Best- and worst-case results for the FOD model with the

Δ H value (i.e. the difference between expected and observed
hydrophobicity) for particular residue is in excess of 80% of the peak value computed for the entire
protein. Such resid ue is suspected of involvement in binding pocket generation
NAD +

80%-this means that calculated

FMN

Best Worst Best Worst

FOD 3HL0 2PH5 3P7N 3BW2

2GWL 2WN7 3A3B 3CB0

Cutoff 80% 1BMD 1 H94 1QZU 1VYR

1U8X 2GR9 1F4P 2WQF

1XAH 1 AD3 1OBO 2Z6I

3JSA 1GIQ 2Z98 1N9L

1DHS 1TOX 2VZJ 2Z6D

1KYQ 1KYQ 3F2V 3B9O

1EBF 3KET 1MVL 1DNL

In the case of proteins depicted in Fig. 4.18 , NAD + binding occurs in close prox-
imity to catalytic residues (marked by red circles). The presented examples indicate
that local maxima (or the global minimum, in the case of 3F2V) point to residues
with catalytic properties. This is likely why the identification of catalytic residues,
as presented in Prymula et al. ( 2011 ), appears to be significantly more accurate
than identi fi cation of NAD + and FMN binding sites, which is the subject of this
discussion.

4.5.5 Summary

It should be noted that the programs presented in this chapter apply diverse tech-
niques to identify ligand binding sites. Low consistency of results (with the excep-
tion of the universally poor identification of FMN sites which appears to be restricted