**3.2 Structure of DHDPR**

### **3.2.1 Subunit and quaternary structure of DHDPR**

The three-dimensional structure of DHDPR has been elucidated by X-ray crystallography from five diverse bacterial species, namely, *Bartonella henselae*, (PDB: 3IJP), *E. coli* (Scapin et al., 1995, 1997), *M. tuberculosis* (Cirilli et al., 2003)*, S. aureus* (Girish et al., 2011), and *T. maritima* (Pearce et al., 2008). DHDPR from *E. coli* (Fig. 8) was the first DHDPR enzyme to be extensively studied in terms of structure and function (Farkas & Gilvarg, 1965; Reddy et al., 1995; Scapin et al., 1995, 1997).

DHDPR is a tetrameric enzyme consisting of four identical monomers (Fig. 8). Each monomer is comprised of an N-terminal nucleotide binding domain and a C-terminal substrate binding domain (Fig. 9). In *E. coli* DHDPR, the nucleotide binding domain is formed by the first 130 and last 36 residues of the polypeptide chain, whereas the substrate binding domain is formed by residues 130-240. The nucleotide binding domain consists of four α-helices and seven βstrands, which are arranged to form a Rossmann (dinucleotide binding) fold. The substrate binding domain contains two α-helices and four β-strands, which form an open mixed βsandwich (Scapin et al., 1995). Interactions between the four subunits of the tetramer occur exclusively between residues of the substrate binding domain. A long loop (Leu182 to Gly204) also extends from the substrate binding domain and plays an important role in maintaining the quaternary structure of the enzyme. The four monomers interact by pairing the four βstrands on the substrate binding domain to form a 16-stranded, mixed, flattened β-barrel (Fig. 8). This central barrel is anchored by the four long loops (Leu182 to Gly204) that extend from the body of the substrate binding domain of each monomer and wrap around the mixed βsheet of the neighboring monomer. Residues 65-74 and 127-130 form flexible hinge regions between the nucleotide and substrate binding domains (Scapin et al., 1995).

Fig. 9. Structure of the *E. coli* DHDPR monomer bound to NADH and the substrate analogue, 2,6-PDC (PDB: 1ARZ).

#### **3.2.2 Substrate binding site**

236 Biochemistry

the body of the substrate binding domain of each monomer and wrap around the mixed βsheet of the neighboring monomer. Residues 65-74 and 127-130 form flexible hinge regions

between the nucleotide and substrate binding domains (Scapin et al., 1995).

Fig. 9. Structure of the *E. coli* DHDPR monomer bound to NADH and the substrate

Fig. 8. Structure of *E. coli* DHDPR (PDB: 1ARZ).

analogue, 2,6-PDC (PDB: 1ARZ).

The consensus sequence, E(L/A)HHXXKXDAPSGTA is found in the substrate binding domain of all known bacterial DHDPR enzymes (Pavelka et al., 1997). This sequence is thought to contain residues involved in binding of substrate and/or catalysis. Molecular modelling studies, using the apo form (enzyme in the absence of substrate) of *E. coli* DHDPR as a structural template, suggest a cluster of five basic residues are the key catalytic site residues (Scapin et al., 1997), namely His159, His160, Arg161, His162 and Lys163 (all contained within the consensus sequence). These residues are located in the loop connecting -strand B7 to -helix A5. Structural studies of *E. coli* DHDPR in complex with NADH and the substrate analogue and inhibitor, 2,6-pyridinedicarboxylate (2,6-PDC), show that 2,6- PDC is bound to the substrate binding domain of DHDPR, in a spherical cavity bordered by residues from both the nucleotide binding (Gly102-Phe106 and Ala126-Ser130) and substrate binding domains (Ile155-Gly175 and Val217-His220) (Scapin et al., 1997). The bound inhibitor makes several hydrogen bonding interactions with the atoms of the conserved E(L/A)HHXXKXDAPSGTA motif. Similar interactions are observed between 2,6-PDC and DHDPR from *M. tuberculosis* (Cirilli et al., 2003).

### **3.2.3 Nucleotide binding site**

The nucleotide binding domain of DHDPR adopts a Rossmann fold, which is typical of nucleotide-dependent dehydrogenases (Fig. 9). The consensus sequence (V/I)(A/G)(V/I)- XGXXGXXG located within this domain, is conserved in all NAD(P)H-dependent dehydrogenases, including DHDPR (Pavelka et al., 1997). Structural analyses of *E. coli*  DHDPR show that this motif extends from the C-terminal end of -strand B1 to the loop that connects B1 to -helix A1. An acidic residue (Glu38 in *E. coli* DHDPR) is located approximately 20 amino acids downstream of the conserved consensus sequence. The two hydroxyl groups from the adenine ribose are known to interact with the side-chain of Glu38 and also the backbone atoms of the glycine rich motif GXXGXXG. Several hydrophobic interactions exist between the adenine ring of NADH and the residues Arg39, Gly84 and His88. The pyrophosphate group of NADH is located over the -helix A1 and interacts with residues contained within the loop connecting -strand B1 and -helix A1 (Reddy et al., 1996; Scapin et al., 1997).

### **3.3 Inhibition of DHDPR**

The substrate analogue, 2,6-PDC, is a competitive inhibitor (*K*i = 26 M) of DHDPR (Scapin et al., 1995) (Fig. 10A). Other substrate analogues such as picolinic acid (Fig. 10B), isopthalic acid (Fig. 10C), pipecolic acid (Fig. 10D) and dimethyl chelidamate (Fig. 10E), are much weaker inhibitors, each displaying an IC50 > 10 mM (Hutton et al., 2003). A vinylogous amide that acts as a competitive inhibitor of DHDPR (*K*i = 32 M) has been described and is one of the most potent inhibitors of DHDPR reported to date (Caplan et al., 2000). Molecular modeling in tandem with conventional drug screening strategies has identified novel inhibitors, including sulfones and sulfonamides, with *K*i values ranging from 7-90 M (Caplan et al., 2000). However, a sub-micromolar inhibitor of DHDPR has not been discovered to date.

Fig. 10. Inhibitors of DHDPR.
