A ¹H n.m.r. study of the role of the glutamate moiety in the binding of methotrexate to Lactobacillus casei dihydrofolate reductase

Abstract

The binding of a series of amide derivatives of methotrexate to Lactobacillus casei dihydrofolate reductase has been studied by inhibition constant measurements and by 1H n.m.r. spectroscopy. Amide modification of the α‐carboxylate of methotrexate was found to prevent interaction of the γ‐carboxylate with the imidazole of His 28. Estimates of the contributions to the binding energy from the α‐carboxylate‐Arg 57 and γ‐carboxylate‐His 28 interactions have been made from a combination of inhibition and n.m.r. data.

NMR Studies of Ligand Binding to Dihydrofolate Reductase

“…For the g-amide MTX complex with DHFR, the 1 H chemical shifts of the His 28 imidazole C2 proton indicated that its pK value is not perturbed from its value in the ligandfree enzyme showing that an ion-pair interaction is not formed (see Figure 17). [116] The 1 H NMR spectrum of the a-amide MTX complex with DHFR also showed that the pK of His 28 was not perturbed even though the g-carboxylate group is available for interaction. Thus, modification of the a-carboxylate not only destroys the Arg 57 interactions, but also perturbs the overall structure such that the available free g-carboxylate can no longer form an ion pair with His 28.…”

“…These differences in chemical shift are consistent with a change in the orientation of the benzoyl ring, which results in changes in the ring-current shielding effects at neighboring protons. [116,133] …”

“…methotrexate g-amide (10); (~) methotrexate a-amide (9); (± ± ±) free enzyme. [116] (Reprinted with permission from Feeney. [23] )…”

“…In earlier simulation studies [115] using NMR-type distance constraints (classified as strong, medium, and weak) extracted from an X-ray structure of folate bound to human DHFR, [16] we found that molecules of this type require at least 50 ligand ± protein distance constraints to produce acceptable structures. In the case of the binding site determined for methotrexate (5), the pteridine ring part is very well-defined (see Figure 10), but the structure around the glutamate moiety needed to be improved by including restraints reflecting the previously determined [116,117] specific interactions between the glutamate acarboxylate group with Arg 57 and the g-carboxylate group with His 28 (see Section 5). The NADPH binding site appears to be essentially preformed in the binary complex and this Figure 10.…”

“…[12, 45, 133±144] NMR spectroscopy is proving to be an excellent alternative method to X-ray crystallography [140] for monitoring the design process by detecting specific interactions between a new inhibitor and the protein, and assessing whether or not the predicted interactions have taken place. [25,116,117,143] We have designed improved DHFR inhibitors by preparing analogues of trimethoprim (4 or brodimoprim 10, in which a bromo substituent replaces the 4'-OCH 3 group) that have side chains on the O3' position of the benzyl ring to make additional interactions with certain protein residues that are normally involved in substrate binding but not used for binding to 4. For example, when the 2,4-diaminopyrimidine ring of a brodimoprim analogue is modeled into its binding site in DHFR it is seen that the brodimoprim, unlike folate or methotrexate, cannot make any direct interactions with Arg 57 and His 28 ( Figure 18).…”