ZHOU Hai-Yan LI Yuan-Yuan LI Jing

(School of Biology and Biological Engineering, South China University of Technology, Guangzhou 510006, China)

ABSTRACT To obtain useful information for identifying inhibitors of urate transporter 1 (URAT1), three-dimensional quantitative structure-activity relationship (3D-QSAR) analysis was conducted for a series of lesinurad analogs via Topomer comparative molecular field analysis (CoMFA). A 3D-QSAR model was established using a training set of 51 compounds and externally validated with a test set of 17 compounds. The Topomer CoMFA model obtained (q2 = 0.976, r2 = 0.990) was robust and satisfactory. Subsequently, seven compounds with significant URAT1 inhibitory activity were designed according to the contour maps produced by the Topomer CoMFA model.

Keywords: 3D-QSAR, Topomer CoMFA, URAT1 inhibitors;

1 INTRODUCTION

Gout is the most common inflammatory arthritis affected by the deposition of monosodium urate (MSU) in joints and surrounding tissues, charac- terized by repeated swelling, redness and severe pain in the joints[1]. Hyperuricemia, commonly caused by overproduction and/or under-excretion of urate, is a condition necessary for MSU formation and deposition[2]. Reports to date show that more than 90% of patients with hyperuricemia are due to insufficient uric acid secretion, supporting the regulation of uric acid excretion as a potentially effective therapeutic strategy[3]. In humans, the level of sUA is primarily controlled by a complex transporter system expressed in the proximal tubules of the renal. In short, uric acid is freely filtered in glomerulus, and ＞90% of the urate that passed through the kidney is reabsorbed back into the bloodstream, while ＜10% of the remained is excreted in urine[4]. The reabsorption process is primarily mediated by uric acid transporter 1 (URAT1, SLC22A12), also referred to as urate transporter 1 or urate-anion exchanger. URAT1 was firstly certified as a 12-transmembrane protein in 2002, which is mainly expressed in the apical side of renal cortical proximal tubular epithelial cells, and is considered to be an important target for the management of hyperuricemia[5].

Benzbromarone was the first URAT1 inhibitor approved by US FDA in 1976 for the treatment of gout but subsequently withdrawn from the market in 2003 due to its hepatotoxicity[5]. Next, lesinurad (RDEA594, developed by Ardea Biosciences) was granted restricted approval by the US FDA in 2015 (Fig. 1)[6]. The drug bears a relatively rigid molecular scaffold structure consisting of naphthalene and 1,2,4-triazole rings as well as carboxylic acid, a functional group commonly present in URAT1 inhibitors. Due to its outstanding potency and favorable toxicology profile, diverse lesinurad analogs, for example, verinurad (RDEA3170, developed by AstraZeneca), have advanced through clinical trials[7-12]. Cai et al.[13]introduced substi- tuents into the CH2linker that resulted in a marked decrease in bioactivity. Tian and co-workers reported that the S atom could be replaced with CH2for the enhancement of URAT1 inhibitory activity[14]while studies by Wei et al.[15]suggested that tetrazole is a satisfactory bioisostere of carboxylic acid in lesinurad.

However, currently, there is no report of URAT1 protein structure, which is not conducive to the drug design based on receptor structure. Docking studies have been recently performed on a homology model of hURAT1[16]. As hURAT1 is a flexible transmem- brane protein, the docking results are not reliable enough to design new compounds because of the uncertain binding conformations. On the other hand, the wide application of Topomer comparative molecular field analysis (CoMFA) with the objective of establishing generic structure-activity relation- ships (SAR) can be used for virtual screening due to its rapid and high efficiency[17-19]. To our knowledge, Topomer CoMFA was reported for the study of SARs of URAT1 inhibitors, which is mainly studied on benzbromarone and its analogues to explain whether the methylene linker between triazole and naphthalene can increase the inhibitory activity of those compounds to hURAT1[20].

For our studies, the structures we collected were targeted lesinurad with the structural skeleton of naphthyltriazole in order to find URAT1 inhibitors with better activity. The data obtained may be effectively applied to clarify the mechanisms underlying URAT1 inhibition by these compounds and provide critical information for designing novel inhibitors with expected activities.

Fig. 1. Representative URAT1 inhibitors and current design strategies

2 MATERIALS AND METHODS

2. 1 Data sets

In total, 68 inhibitors targeting URAT1 were identified from the literature[14,15]. The inhibitory activities (IC50, μM) were converted to reciprocal logarithmic values (pIC50= -log(IC50/106) ranging from 4.44 to 8.03. All analyses of URAT1 inhibitor compounds were carried out using SYBYL-X 2.0 (Tripos Software, Saint Louis, MO, USA) and MOE 2010.10 software.

2. 2 Molecular modeling

Based on the pIC50values and structural diversity, the compounds were grouped into two sets at a ratio of ～3:1, among which 51 compounds are the training set and 17 are the test set (Table 1)[21].

To generate the most stable conformation, energy minimization was conducted using a Tripos force field with a distance-dependent dielectric and the energy gradient convergence criterion and maximum iterations set to 0.005 kcal/mol ? and 100, respectively[22-24]. The independent variables were the Topomer CoMFA descriptors and the dependent variables were the pIC50values[25,26].

Table 1. Chemical Structures and IC50 Values of URAT1 Inhibitors

7* -I images/BZ_47_1328_290_1357_321.png8.80 5.06 images/BZ_47_1330_392_1358_423.png8 -OCH2CH3 images/BZ_47_1389_493_1409_516.png9.10 5.04 9* -OCH2CH3 -Br images/BZ_47_1328_680_1357_711.png11.20 4.95 images/BZ_47_1330_782_1358_813.png-H -Br 6.87 5.16 10 11 -CH3 -Br 6.16 5.21 12 -CH2CH3 -Br 5.37 5.27 13 -(CH2)2CH3 -Br 2.03 5.69 14 -CH(CH3)2 -Br 2.21 5.66 -Br 3.52 5.45 15* 16 -OCH3 -Br 8.90 5.05 17 -OCH2CH3 -Br 11.49 4.94 -H -Br images/BZ_47_1266_2431_1293_2460.png0.0338 7.47 18 images/BZ_47_1351_2479_1378_2508.pngimages/BZ_47_1383_2479_1405_2507.pngimages/BZ_47_1352_2576_1380_2605.pngimages/BZ_47_1384_2576_1406_2604.pngimages/BZ_47_1268_2626_1295_2655.png19* -CH3 -Br images/BZ_47_1266_2697_1293_2724.pngimages/BZ_47_1350_2742_1376_2769.pngimages/BZ_47_1380_2742_1402_2769.png0.0267 7.57 images/BZ_47_1351_2834_1377_2862.pngimages/BZ_47_1382_2835_1403_2861.pngimages/BZ_47_1269_2881_1295_2909.png

20 -CH2CH3 -Br images/BZ_48_1267_295_1292_322.pngimages/BZ_48_1348_339_1374_365.pngimages/BZ_48_1378_339_1400_365.png0.0274 7.56 images/BZ_48_1350_428_1375_455.pngimages/BZ_48_1380_429_1401_455.pngimages/BZ_48_1269_474_1295_501.png21 -CH(CH3)2 -Br images/BZ_48_1266_552_1293_579.pngimages/BZ_48_1350_597_1376_624.pngimages/BZ_48_1380_597_1402_624.png0.0259 7.59 images/BZ_48_1351_689_1377_717.pngimages/BZ_48_1382_690_1403_716.pngimages/BZ_48_1269_736_1295_764.png22 -C(CH3)3 -Br images/BZ_48_1266_812_1293_839.pngimages/BZ_48_1350_857_1376_884.pngimages/BZ_48_1380_857_1402_884.png0.0211 7.68 images/BZ_48_1351_949_1377_977.pngimages/BZ_48_1382_950_1403_976.pngimages/BZ_48_1269_996_1295_1024.png-Br images/BZ_48_1266_1072_1293_1099.png0.0165 7.78 23 images/BZ_48_1350_1117_1376_1144.pngimages/BZ_48_1380_1117_1402_1144.pngimages/BZ_48_1351_1209_1377_1237.pngimages/BZ_48_1382_1210_1403_1236.pngimages/BZ_48_1269_1256_1295_1284.png-F -Br images/BZ_48_1266_1323_1293_1353.png0.0251 7.60 24 images/BZ_48_1351_1372_1378_1402.pngimages/BZ_48_1383_1372_1405_1401.pngimages/BZ_48_1352_1472_1380_1502.pngimages/BZ_48_1384_1472_1406_1501.pngimages/BZ_48_1268_1523_1295_1553.png-Cl -Br images/BZ_48_1266_1592_1293_1619.png0.0228 7.64 25* images/BZ_48_1350_1637_1376_1664.pngimages/BZ_48_1380_1637_1402_1664.pngimages/BZ_48_1351_1729_1377_1757.pngimages/BZ_48_1382_1730_1403_1756.pngimages/BZ_48_1269_1776_1295_1804.png-Br -Br images/BZ_48_1266_1846_1293_1875.png0.0197 7.71 26 images/BZ_48_1351_1894_1378_1923.pngimages/BZ_48_1383_1894_1405_1922.pngimages/BZ_48_1352_1991_1380_2020.pngimages/BZ_48_1384_1991_1406_2019.pngimages/BZ_48_1268_2041_1295_2070.png-I -Br images/BZ_48_1266_2103_1293_2133.png0.029 7.54 27 images/BZ_48_1351_2152_1378_2182.pngimages/BZ_48_1383_2152_1405_2181.pngimages/BZ_48_1352_2252_1380_2282.pngimages/BZ_48_1384_2252_1406_2281.pngimages/BZ_48_1268_2303_1295_2333.png28 -OH -Br images/BZ_48_1266_2363_1293_2393.pngimages/BZ_48_1351_2412_1378_2442.pngimages/BZ_48_1383_2412_1405_2441.png0.0341 7.47 images/BZ_48_1352_2512_1380_2542.pngimages/BZ_48_1384_2512_1406_2541.pngimages/BZ_48_1268_2563_1295_2593.png29 -OCH3 -Br images/BZ_48_1266_2632_1293_2659.pngimages/BZ_48_1350_2677_1376_2704.pngimages/BZ_48_1380_2677_1402_2704.png0.0192 7.72 images/BZ_48_1351_2769_1377_2797.pngimages/BZ_48_1382_2770_1403_2796.pngimages/BZ_48_1269_2816_1295_2844.png30 -OCH2CH3 -Br images/BZ_48_1266_2883_1293_2913.pngimages/BZ_48_1351_2932_1378_2962.pngimages/BZ_48_1383_2932_1405_2961.png0.024 7.62 images/BZ_48_1352_3032_1380_3062.pngimages/BZ_48_1384_3032_1406_3061.pngimages/BZ_48_1268_3083_1295_3113.png

31 -OCH(CH3)2 -Br images/BZ_49_1266_283_1293_313.pngimages/BZ_49_1351_332_1378_362.pngimages/BZ_49_1383_332_1405_361.png0.0275 7.56 images/BZ_49_1352_432_1380_462.pngimages/BZ_49_1384_432_1406_461.pngimages/BZ_49_1268_483_1295_513.png32* -OC(CH3)3 -Br images/BZ_49_1266_546_1293_575.pngimages/BZ_49_1351_594_1378_623.pngimages/BZ_49_1383_594_1405_622.png0.0283 7.55 images/BZ_49_1352_691_1380_720.pngimages/BZ_49_1384_691_1406_719.pngimages/BZ_49_1268_741_1295_770.png-Br images/BZ_49_1266_806_1293_835.png0.0316 7.50 33 images/BZ_49_647_849_675_881.pngimages/BZ_49_1351_854_1378_883.pngimages/BZ_49_1383_854_1405_882.pngimages/BZ_49_1352_951_1380_980.pngimages/BZ_49_1384_951_1406_979.pngimages/BZ_49_1268_1001_1295_1030.png-CN -Br images/BZ_49_1266_1066_1293_1095.png0.0209 7.68 34 images/BZ_49_1351_1114_1378_1143.pngimages/BZ_49_1383_1114_1405_1142.pngimages/BZ_49_1352_1211_1380_1240.pngimages/BZ_49_1384_1211_1406_1239.pngimages/BZ_49_1268_1261_1295_1290.png35* -H -Br images/BZ_49_1206_1365_1235_1396.pngimages/BZ_49_1266_1319_1294_1351.pngimages/BZ_49_1356_1372_1384_1404.pngimages/BZ_49_1389_1373_1413_1403.png0.0307 7.51 images/BZ_49_1356_1478_1385_1510.pngimages/BZ_49_1390_1479_1413_1510.pngimages/BZ_49_1256_1525_1285_1557.png36 -F -Br images/BZ_49_1206_1625_1235_1656.pngimages/BZ_49_1266_1579_1294_1611.pngimages/BZ_49_1356_1632_1384_1664.pngimages/BZ_49_1389_1633_1413_1663.png0.0241 7.62 images/BZ_49_1356_1738_1385_1770.pngimages/BZ_49_1390_1739_1413_1770.pngimages/BZ_49_1256_1785_1285_1817.png37 -Cl -Br images/BZ_49_1206_1885_1235_1916.pngimages/BZ_49_1266_1839_1294_1871.pngimages/BZ_49_1356_1892_1384_1924.pngimages/BZ_49_1389_1893_1413_1923.png0.0173 7.76 images/BZ_49_1356_1998_1385_2030.pngimages/BZ_49_1390_1999_1413_2030.pngimages/BZ_49_1256_2045_1285_2077.png38* -Br -Br images/BZ_49_1206_2145_1235_2176.pngimages/BZ_49_1266_2099_1294_2131.pngimages/BZ_49_1356_2152_1384_2184.pngimages/BZ_49_1389_2153_1413_2183.png0.0209 7.70 images/BZ_49_1356_2258_1385_2290.pngimages/BZ_49_1390_2259_1413_2290.pngimages/BZ_49_1256_2305_1285_2337.png39 -I -Br images/BZ_49_1206_2405_1235_2436.pngimages/BZ_49_1266_2359_1294_2391.pngimages/BZ_49_1356_2412_1384_2444.pngimages/BZ_49_1389_2413_1413_2443.png0.0294 7.53 images/BZ_49_1356_2518_1385_2550.pngimages/BZ_49_1390_2519_1413_2550.pngimages/BZ_49_1256_2565_1285_2597.png40 -OH -Br images/BZ_49_1206_2665_1235_2696.pngimages/BZ_49_1266_2619_1294_2651.pngimages/BZ_49_1356_2672_1384_2704.pngimages/BZ_49_1389_2673_1413_2703.png0.0295 7.53 images/BZ_49_1356_2778_1385_2810.pngimages/BZ_49_1390_2779_1413_2810.pngimages/BZ_49_1256_2825_1285_2857.png41* -OCH3 -Br images/BZ_49_1206_2925_1235_2956.pngimages/BZ_49_1266_2879_1294_2911.pngimages/BZ_49_1356_2932_1384_2964.pngimages/BZ_49_1389_2933_1413_2963.png0.0177 7.75 images/BZ_49_1356_3038_1385_3070.pngimages/BZ_49_1390_3039_1413_3070.pngimages/BZ_49_1256_3085_1285_3117.png

42 -OCH2CH3 -Br images/BZ_50_1206_325_1235_356.pngimages/BZ_50_1266_279_1294_311.pngimages/BZ_50_1356_332_1384_364.pngimages/BZ_50_1389_333_1413_363.png0.0206 7.69 images/BZ_50_1356_438_1385_470.pngimages/BZ_50_1390_439_1413_470.pngimages/BZ_50_1256_485_1285_517.png43 -OCH(CH3)2 -Br images/BZ_50_1206_585_1235_616.pngimages/BZ_50_1266_539_1294_571.pngimages/BZ_50_1356_592_1384_624.pngimages/BZ_50_1389_593_1413_623.png0.0231 7.64 images/BZ_50_1356_698_1385_730.pngimages/BZ_50_1390_699_1413_730.pngimages/BZ_50_1256_745_1285_777.png44 -OC(CH3)3 -Br images/BZ_50_1206_845_1235_876.pngimages/BZ_50_1266_799_1294_831.pngimages/BZ_50_1356_852_1384_884.pngimages/BZ_50_1389_853_1413_883.png0.0258 7.59 images/BZ_50_1356_958_1385_990.pngimages/BZ_50_1390_959_1413_990.pngimages/BZ_50_1256_1005_1285_1037.png-Br images/BZ_50_1266_1059_1294_1091.png0.0262 7.58 45 images/BZ_50_647_1109_675_1141.pngimages/BZ_50_1206_1105_1235_1136.pngimages/BZ_50_1356_1112_1384_1144.pngimages/BZ_50_1389_1113_1413_1143.pngimages/BZ_50_1356_1218_1385_1250.pngimages/BZ_50_1390_1219_1413_1250.pngimages/BZ_50_1256_1265_1285_1297.png46 -CN -Br images/BZ_50_1206_1365_1235_1396.pngimages/BZ_50_1266_1319_1294_1351.pngimages/BZ_50_1356_1372_1384_1404.pngimages/BZ_50_1389_1373_1413_1403.png0.0140 7.85 images/BZ_50_1356_1478_1385_1510.pngimages/BZ_50_1390_1479_1413_1510.pngimages/BZ_50_1256_1525_1285_1557.png47* -H -Br images/BZ_50_1206_1623_1235_1654.pngimages/BZ_50_1267_1581_1295_1611.pngimages/BZ_50_1356_1633_1384_1663.pngimages/BZ_50_1390_1634_1413_1663.png0.0328 7.48 images/BZ_50_1206_1741_1235_1772.pngimages/BZ_50_1355_1735_1383_1766.pngimages/BZ_50_1388_1736_1412_1765.pngimages/BZ_50_1265_1786_1293_1816.png48 -CH3 -Br images/BZ_50_1206_1883_1235_1914.pngimages/BZ_50_1267_1841_1295_1871.pngimages/BZ_50_1356_1893_1384_1923.pngimages/BZ_50_1390_1894_1413_1923.png0.0257 7.59 images/BZ_50_1206_2001_1235_2032.pngimages/BZ_50_1355_1995_1383_2026.pngimages/BZ_50_1388_1996_1412_2025.pngimages/BZ_50_1265_2046_1293_2076.png49 -CH2CH3 -Br images/BZ_50_1206_2143_1235_2174.pngimages/BZ_50_1267_2101_1295_2131.pngimages/BZ_50_1356_2153_1384_2183.pngimages/BZ_50_1390_2154_1413_2183.png0.0282 7.55 images/BZ_50_1206_2261_1235_2292.pngimages/BZ_50_1355_2255_1383_2286.pngimages/BZ_50_1388_2256_1412_2285.pngimages/BZ_50_1265_2306_1293_2336.png50 -CH(CH3)2 -Br images/BZ_50_1206_2403_1235_2434.pngimages/BZ_50_1267_2361_1295_2391.pngimages/BZ_50_1356_2413_1384_2443.pngimages/BZ_50_1390_2414_1413_2443.png0.0215 7.67 images/BZ_50_1206_2521_1235_2552.pngimages/BZ_50_1355_2515_1383_2546.pngimages/BZ_50_1388_2516_1412_2545.pngimages/BZ_50_1265_2566_1293_2596.png51 -C(CH3)3 -Br images/BZ_50_1206_2663_1235_2694.pngimages/BZ_50_1267_2621_1295_2651.pngimages/BZ_50_1356_2673_1384_2703.pngimages/BZ_50_1390_2674_1413_2703.png0.0167 7.78 images/BZ_50_1206_2781_1235_2812.pngimages/BZ_50_1355_2775_1383_2806.pngimages/BZ_50_1388_2776_1412_2805.pngimages/BZ_50_1265_2826_1293_2856.png-Br images/BZ_50_1267_2881_1295_2911.png0.0131 7.88 52 images/BZ_50_1206_2923_1235_2954.pngimages/BZ_50_1356_2933_1384_2963.pngimages/BZ_50_1390_2934_1413_2963.pngimages/BZ_50_1206_3041_1235_3072.pngimages/BZ_50_1355_3035_1383_3066.pngimages/BZ_50_1388_3036_1412_3065.pngimages/BZ_50_1265_3086_1293_3116.png

*Compounds belonging to the test set

2. 3 Partial least-squares analysis

The cross-validated correlation coefficient (q2) and the non-cross-validated correlation coefficient (r2) values were used to optimize the numbers of components in the partial least-squares (PLS) models, which were obtained by the leave-one-out (LOO) cross-validation procedure[27]. The internal quality of the models was measured usingq2[28]obtained according to Eq. (1):

WhereYpred,Yexp, andYmeanrepresent the predicted, experimental and mean biological activities of the training set, respectively.

The predictive ability of the 3D-QSAR models was evaluated using the predicted correlation coefficientr2pred[29]calculated from Eq. (2):

Where SD represents the total of the squared deviation scores between the bioactivity of the test set and the average bioactivity of the training set, and PRESS is the total of squares between the predicted and experimental bioactivities of the test set[30].

The optimal 3D-QSAR model was graphically represented by field contour maps, and coefficients were formed by the CoMFA StDev * Coeff field type.

3 RESULTS AND DISCUSSION

3. 1 3D-QSAR modeling

Empirically, the naphthalene ring was used as the common skeleton and a compound was randomly selected as the cut template to acquire Ra (red), Rb (blue) and core (black) groups, as presented in Fig. 2. Other training molecules were certified automatically and cut as above. Theq2andr2values were deter- mined as 0.976 and 0.990, respectively. Statistical analyses validated the precise predictive predicta- bility of the QSAR model of Topomer CoMFA. The root mean square error (RMSE) values calculated using Excel 2007 were 0.111 and 0.168 for the training and test sets, respectively. The difference in RMSE between the two sets was not significant, indicative of good external prediction ability of the model and no overfitting phenomenon.

A graphical plot of the predicted versus experi- mental pIC50values for all the compounds is presented in Fig. 3, and the corresponding data are specified in Table 2. Our results exhibited that the predicted values of the molecules were consistent with the experimental ones. Based on comparison of the fragment contribution values of Ra and Rb, we speculate that substitution of Ra has a greater impact on URAT1 inhibitory activity than the substitution of Rb.

Fig. 2. Splitting of the template molecule into three sections

Fig. 3. Scatter plot of experimental versus predicted pIC50 values of the training and test sets in our 3D-QSAR model using Topomer CoMFA

Table 2. Experimental and Predicted pIC50 Values of URAT1 Inhibitors with Ra and Rb Fragment Contributions

* Compounds belonging to the test set

Furthermore, the parameterrm2proposed by Roy et al.[31]and the Golbraikh-Tropsha statistical method[24]were applied for external verification of the test set. The relevant formulae are as follows: Eq. (3):

(R2: correlation coefficient between the activity test and predicted values)

Eq. (4):

(k:slope of the regression line between the experimental activity value (y-axis) and predicted value (x-axis) when the intercept is 0)

Eq. (5):

(k': slope of the regression line between the predicted activity value (y-axis) and experimental value (x-axis) when the intercept is 0)

Eq. (6):

(R02: correlation coefficient between the experimental activity value (y-axis) and predicted value (x-axis))

Eq. (7):

(R'02: correlation coefficient between the predicted activity value (y-axis) and experimental value (x-axis))

Eq. (8):

(rm2: statistics of the external prediction ability of the evaluation model proposed by Roy)

Eq. (9):

Yobs: experimental value of the test sets

Ypred: predictive value of the test sets

All statistical parameters were within the ranges of the reference values, indicating that the model has sufficient prediction ability for external samples (Table 3).

Table 3. Statistical Parameters for External Validation of the QSAR Model

3. 2 3D contour map analysis

To facilitate analyses, a molecule was randomly chosen as a reference for the 3Dcoefficient contour map. The results obtained using Topomer CoMFA models were graphically interpreted by field contribution maps. The calculated Topomer CoMFA steric and electrostatic contour maps are presented in Fig. 4.

In the steric field, the green contours located at Ra showed a degree of closure, suggesting that only moderate bulky substituents would be favorable, and the yellow contours indicated that bulky substituents would not be tolerated (Fig. 4a). These features could account for the higher activity of compound 58 (pIC50= 7.83) with -OCH3as Ra and the relatively lower activities of compounds 59 (pIC50= 7.74), 60 (pIC50= 7.67) and 61 (pIC50= 7.62) with -OCH2CH3, -OCH(CH3)2and -OC(CH3)3Ra groups, respectively, despite the same Rb for these compounds. In the electrostatic field (Fig. 4b), the advantage of electron-withdrawing groups is expressed by red contours of the R1group and that of electron- donating substituents by blue contours. This pheno- menon could explain the higher activity of compounds 64 (pIC50= 8.03) and 63 (pIC50= 7.94) with -NO2and -CN as Ra, respectively, and lower activity of compound 52 (pIC50= 7.88) with a -cyclopropyl Ra group.

As shown in Fig. 4c, Rb can be divided into two regions, specifically, a substituted triazole ring (triazole-R2) and long chain carboxyl group (R3). CH2-, S atom, sulphinyl and sulphonyl groups are commonly used for linkage between the triazole and carboxyl groups. The yellow contours located at the S atom of R3suggested that oversized bridge linkage groups (larger than the sulphonyl group) would not be tolerated here. In fact, the activity of sulphonyl groups was generally higher than that of sulphinyl and single S atom groups, as observed for compounds 58 (pIC50= 7.83), 41 (pIC50= 7.75) and 29 (pIC50= 7.72). Finally, the large green area and blue contours around the end of R3indicate that the substituent of two carboxylic acids would be favorable in this region, as observed for compounds 23 (pIC50= 7.78) and 6 (pIC50= 5.62).

Fig. 4. Topomer CoMFA contour maps for URAT1 inhibitors. (a) and (b) depict steric and electrostatic field maps of the Ra fragment, respectively; (c) and (d) depict steric and electrostatic field maps of the Rb fragment

The SARs for URAT1 inhibitors summarized in Fig. 5 provide useful references for the optimization of current molecular activities and the design of effective inhibitors with novel scaffolds.

Table 4. Structures of Newly Designed Molecules a～g

3. 3 Design of novel URAT1 inhibitors

Based on intimate contour analyses of 3D-QSAR models and further SAR assessments, pivotal structural features of ligands beneficial for URAT1 inhibitory activity were included in the “naphthyl triazole” scaffold to design seven new URAT1 inhibitors a～g (Table 4). In these compounds, R3was retained as “sulfinyl-two carboxyl” or “sulfonyl-two carboxyl” groups. The main alteration was substitution at the R1group. For example, N(CH3)3- was introduced as an electron-donating group and CF3-, CN-, NO2-, Cl- and furan were introduced as electron-withdrawing groups. Similar modifications were introduced at the R2group. The predictedIC50values of the novel compounds using the Topomer CoMFA model and ADMET properties predicted with Discovery Studio 3.5 software are shown in Table 5. Notably, the predicted pIC50values of these compounds were not markedly different from those of compound 64, supporting their potential as URAT1 inhibitors. Moreover, compounds a and c displayed higher water solubility than compound 64, and the absorption of all compounds, except c and f, was better. The efficacy of these compounds as URAT1 inhibitors will be validated in future experiments.

Table 5. Predicted IC50 Values and ADMET Properties of the Newly Designed Molecules a～g

4 CONCLUSION

In this study, 3D-QSAR was employed to identify the structural determinants of URAT1 inhibitors. The high statistical values (q2= 0.976;r2= 0.990) support the Topomer CoMFA model with strong reliability and predictive ability. The results obtained using the 3D-QSAR models indicate that medium-sized electron-withdrawing groups are suitable for introduction into R1and “sulfinyl-two carboxyl” or “sulfonyl-two carboxyl” groups appro- priate for R3. Furthermore, seven novel compounds were designed with the predicted pIC50and ADMET properties. In summary, the models developed using 3D-QSAR provide valuable information that should facilitate the design of efficient innovative URAT1 inhibitor compounds.