QSAR MODELING OF COMPOUNDS DERIVED FROM 1,2,3- TRIAZOLOPIPERIDINE AS DPP-4 ENZYME INHIBITORS USING SEMIEMPIRICAL AM1

This study aims to model the derived compounds of 1,2,3-triazolopiperidine using semiempirical method AM1 and determine the further derivation with the better IC50 values against DPP-4 enzyme theoretically. This research employed ChemDraw Pro 12 software for for 2D structural drawing, Hyperchem 8.0 for 3D modelling, and MLR statistical analysis for modeling QSAR equations. The semiempirical method was likely to be the appropriate platform to apply because the correlation coefficient of H1 NMR chemical shift between theoretical and actual value is relatively close, 0.8891. The multilinear regression analysis produced 4 equation models where the best one is equation 4 as detailed below:” IC50 = 875.5116 + (-7400.27*qH35) + (-0.00133* Eat.is) + (-3230.72* qN23) + (3.30277* μ)” n = 25; r2 = 0.594; Adjusted r2= 0.486; PRESS = 1.2× 104. Finally, the theoretically promising substituent was -CN possessing IC50 value = 1.61 nM.


INTRODUCTION
QSAR (quantitative structure and activity relationship) has been designed to quantify a consistent relationship between molecular skeleton and its biological activity resulting an opportunity to evaluate the properties of new materials compared to a series of molecules being modeled [1,2]. The QSAR method can also allow chemists working in synthetic field to improve their desired products performance by modelling the molecules through appropriates approach [3,4]. Therefore, validating methods applied in a computational project is a vital step that mainly drive the accuracy of the output.
Generally, validation of computer-assisted modelling methods in such field is likely to be carried-out by comparing the computed data with the experimental observation [5]. One of the reliable approach is comparison the chemical shift value of nuclear magnetic resonance (NMR) between calculated and actual data. Through statistical analysis, the most appropriate method could be determined by observing the correlation coefficient (r 2 ) number [6].
QSAR seems to be powerful tool for predicting drug compounds and the efficacy for treating deadly diseases like type 2 diabetes mellitus (T2DM). Diabetes mellitus is of metabolic diseases characterized by hyperglycemia due to a deficiency of insulin secretion, insulin resistance, or both [7,8,9,10]. In T2DM case, the causes may vary from the dominant insulin resistance accompanied by relative insulin deficiency to the dominant decrease of insulin secretion accompanied by insulin resistance [8,9,10].
Currently, one of the most effective oral antidiabetic therapies is by inhibiting the enzyme dipeptidyl peptidase-4 (DPP-4), the main inactivator of the degradation of glucagon like peptide-1 (GLP-1) and glucosedependent insulinotropic peptide (GIP).
Scientific report shows that DPP-4 inhibitors appear to be effective in controlling blood glucose levels without causing weigh gain [6].
Sitagliptin, saxagliptin, vildagliptin and linagliptin are DPP-4 inhibitors available for the treatment of T2DM in Indonesia and many other countries [11,12,13]. However, those drugs could cause runny or stuffy nose, sore throat, headache, backache, stomachache, diarrhea, constipation, hypoglycemia, weakness, drowsiness, and blurred vision. To reduce these side effects, the development of 1,2,3triazolopiperidine compounds, the main skeleton of sitagliptin, needs to be done by modifying the structure with the QSAR assistance [11,12,13] Such research has been carried-out on the compounds through the correlation of their structure and descriptor properties with experimental activities.
Modeling the triazolopiperazin amide derivatives, similar compounds to 1,2,3-triazolopiperidine, has been accomplished using the semiempirical method [14]. In the term of time consuming, the method appears to be more efficient than ab-initio in the design of drug compounds.
The research suggests that further studies of DPP-4 inhibitor compounds need to be conducted especially for other compounds which inhibitory mechanism and structure resemble triazolopiperazin compounds.

METHODS
The structure of 1,2,3-triazolopiperidine derivatives and their activity (IC50) were taken from [12]. This study employed a computer

Calculation of 1 H NMR δ of the compounds with semiempirical AM1 method
The optimized structures were then analyzed using proton NMR chemical shift values (δ) and compared with experimental data obtained from [9] . In the program, the structure of 1,2,3-triazolopiperidine derivatives was left-clicked to select compute menu then followed by NMR invoke. In the new page, setup was clicked then select H to start the calculation.

QSAR analysis using multiple linear regression analysis (MLR)
The study used MLR analysis with forward method in SPSS 21 program. The independent variables were electronic and molecular parameters obtained from the file.log and QSAR properties, while the dependent variable was IC50 values of the 1,2,3-triazolopiperidine derivatives.

Determination of the best QSAR equation model using electronic and molecular parameters
The best model determined from the the QSAR analysis was gained by comparing its IC50 value to the experimental biological activity. This can be achieved from the correlation coefficient (r 2 ) of the curve.

QSAR analysis data
The resulted QSAR equation of AM1 is presented in Table 2

Designing new derivatives
As the best QSAR equation, model 4 was applied to design new compounds and predict their biological activity. The calculated parameters were qH35, Eat.is, qN23 and µ.
The IC50 values of the proposed compounds resulted from the model can be seen in the Table 3.  The new compounds designed from extensive, S2 subsite, S1 subsite, S01 subsite, and S02 subsite [22]. In details, the composition of them is described in Table 4.