Sitagliptin does not inhibit the M1 alanyl aminopeptidase from Plasmodium falciparum

The M1 alanyl aminopeptidase from Plasmodium falciparum has been shown to be an essential hemoglobinase enzyme, catalyzing the final stages of hemoglobin break-down within intra-erythrocytic parasites 1. Recently there has been much interest in this protease as a potential drug target for the development of novel antimalarials [1-13]. In a recent report, Krishnamoorthy and Achary propose that Sitagliptin may serve as a potent competitive inhibitor of the M1 alanyl aminopeptidase enzyme from Plasmodium falciparum [14]. The molecule of interest, the M1 alanine aminopeptidase or PfA-M1, has been well studied and characterized by our group and others (For examples of recent papers, see [1-13]). To date, multiple published X-ray crystal structures and models are available of inhibitors bound to the active site of PfA-M1. Krishnamoorthy and Achary report a docking analysis of the enzyme with “about 100 low molecular weight protease inhibitors ...”. The docking results were validated by inclusion of specific substrate (Ala-β-naphthylamide) from which they calculated an in silico Km value that was closely correlated with experimental data. Unfortunately, this correlation with experimental data was not observed with the selected positive inhibitor control, Bestatin. The interaction / inhibition of PfAM1 by Bestatin has been reported previously [1, 9, 12, 15] and the dipeptide analog has an in vitro Ki in the nM range for PfAM1. The in silico value calculated was ~ 100 μM, indicating that the parameters defined for docking were likely inadequate for the cation-occupied active site. The article states that Sitagliptin is the most potent in silico inhibitor with a Ki(avg) of 2 uM (however only 8/100 docking results were provided in Supplementary Table 1). We have completed an in vitro analysis of the effect of Sitagliptin on the activity of both PfA-M1 and the second neutral aminopeptidase, PfA-M17. Aminopeptidase activity assays were carried out in 200 l total volume in 50 mM Tris-HCl pH 8.0, at 37 oC (with the addition of 2 mM CoCl2 for PfA-M17). Following a 10 min incubation of enzyme and Sitagliptin (0 – 0.5 mM), reactions were initiated by addition of fluorigenic substrate (L-Leucine-7-amido-4-methylcoumarin). Progress curves were monitored using a spectrofluorimeter until a final steady-state velocity was reached. We determined the inhibitory kinetics via Ki values from Dixon plots of 1/vs versus inhibitor concentration when [S]<<KM. We report here that this compound had no effect on the aminopeptidase activity of either enzyme. In silico drug docking suffers from several problems, including modeling the physics of the system, solvent effects, dynamics, and the difficulty in accurately ranking the docked results, and therefore relies critically on validation by experiment. It is vital that we rigorously test hypotheses generated from webservers to ensure that these algorithms continue to improve in their accuracy and hence usefulness. Blanket statements about efficacy from untested in silico studies will confound the literature and waste precious resources by following up on false positives identified in poorly controlled in silico studies. Computational biology has an important role to play in research and drug design; however, it is absolutely vital that we apply critical evaluation of results obtained to ensure that it becomes a robust method in the future.