¹³C NMR Quantitative Spectrometric Data-Activity Relationship (QSDAR) Models of Steroids Binding the Aromatase Enzyme

Abstract

Five quantitative spectroscopic data-activity relationships (QSDAR) models for 50 steroidal inhibitors binding to aromatase enzyme have been developed based on simulated (13)C nuclear magnetic resonance (NMR) data. Three of the models were based on comparative spectral analysis (CoSA), and the two other models were based on comparative structurally assigned spectral analysis (CoSASA). A CoSA QSDAR model based on five principal components had an explained variance (r(2)) of 0.78 and a leave-one-out (LOO) cross-validated variance (q(2)) of 0.71. A CoSASA model that used the assigned (13)C NMR chemical shifts from a steroidal backbone at five selected positions gave an r(2) of 0.75 and a q(2) of 0.66. The (13)C NMR chemical shifts from atoms in the steroid template position 9, 6, 3, and 7 each had correlations greater than 0.6 with the relative binding activity to the aromatase enzyme. All five QSDAR models had explained and cross-validated variances that were better than the explained and cross-validated variances from a five structural parameter quantitative structure-activity relationship (QSAR) model of the same compounds. QSAR modeling suffers from errors introduced by the assumptions and approximations used in partial charges, dielectric constants, and the molecular alignment process of one structural conformation. One postulated reason that the variances of QSDAR models are better than the QSAR models is that (13)C NMR spectral data, based on quantum mechanical principles, are more reflective of binding than the QSAR model's calculated electrostatic potentials and molecular alignment process. The QSDAR models provide a rapid, simple way to model the steroid inhibitor activity in relation to the aromatase enzyme.

Towards understanding aromatase inhibitory activity via QSAR modeling

“…Over the course of the next five years (2001-2005) (Gironés and Carbó-Dorca, 2002[34]; Beger and Wilkes, 2002[12]; Beger et al, 2001[10]; Polanski and Gieleciak, 2003[71]; Beger et al, 2004[11]; Leonetti et al, 2004[50]; Cavalli et al, 2005[17]), it can be seen that studies employed higher number of descriptors as well as made use of more descriptor types (e.g. molecular fields, spectral, molecular surface and quantum chemical) were observed in seven publications where the only CV method applied on the datasets was the LOO-CV (Figure 4(Fig.…”

“…In addition, Beger et al (2001[10]) built quantitative spectroscopic data-activity relationship (QSDAR) models for 50 steroidal AIs developed based on data collected via simulated 13 C nuclear magnetic resonance (NMR). The models were based on comparative spectral analysis (CoSA) and comparative structurally assigned spectral analysis (CoSASA).…”

“…As summarized in Table 4(Tab. 4) (References in Table 4: Nagy et al, 1994[58]; Recanatini, 1996[75]; Oprea and García, 1996[67]; Sulea et al, 1997[92]; Recanatini and Cavalli, 1998[76]; Cavalli et al, 2000[18]; Beger et al, 2001[10]; Gironés and Carbó-Dorca, 2002[34]; Beger and Wilkes, 2002[12]; Polanski and Gieleciak, 2003[71]; Leonetti et al, 2004[50]; Beger et al, 2004[11]; Cavalli et al, 2005[17]; Bak and Polanski, 2007[8]; Castellano et al, 2008[16]; Nagar et al, 2008[55]; Mittal et al, 2009[53]; Gueto et al, 2009[36]; Dai et al, 2010[24]; Roy and Roy, 2010[78]; Roy and Roy, 2010[77]; Nagar and Saha, 2010[57]; Nagar and Saha, 2010[56]; Narayana et al, 2012[65]; Nantasenamat et al, 2013[64]; Nantasenamat et al, 2013[61]; Kishore et al, 2013[44]; Worachartcheewan et al, 2014[103]; Worachartcheewan et al, 2014[101]; Nantasenamat et al, 2014[63]; Dai et al, 2014[25]; Awasthi et al, 2015[7]; Xie et al, 2015[105]; Shoombuatong et al, 2015[85]; Xie et al, 2014[102]; Kumar et al, 2016[48]; Ghodsi and Hemmateenejad, 2016[32]; Song et al, 2016[91]; Prachayasittikul et al, 2017[72]; Adhikari et al, 2017[1]; Lee and Barron, 2018[49]; Pingaew et al, 2018[70]; Barigye et al, 2018[9]), it can be observed that prior to 2010, MLR and PLS models, also known as white-box approaches, were the most popular and yet simple learning algorithms used for QSAR modeling of AIs. …”

Comparative Structural Connectivity Spectra Analysis (CoSCoSA) Models of Steroid Binding to the Corticosteroid Binding Globulin

“…8 These comparative spectral analysis (CoSA) models using simulated 13 C NMR data yielded higher cross-validated correlations than were seen with comparative molecular field analysis (CoMFA) methods and comparative structure assigned spectra analysis (CoSASA). 6,7 Our previous QSDAR CoSA model for the 30 corticosterone binding globulin steroids was significantly robust with an r 2 of 0.80 and a q 2 of 0.78, 6 but we believed we could do better if we were able to add structural information to the one-dimensional CoSA models. We added structurally assigned chemical shift to the backbone of steroids, and we obtained a CoSASA model with an r 2 of 0.80 and a LOO cross-validation q 2 of 0.73.…”

NMR QSAR Model for the Analysis of 4‐(5‐Arylamino‐1,3,4‐thiadiazol‐2‐yl)benzene‐1,3‐diols

“…These techniques were used for discovering high-affinity ligands that bind to the FK506 binding protein (FKBP) and nonpeptide inhibitors of stromelysin [12,13]. Beger and co-workers demonstrated that 13 C-NMR data can be used to produce reliable quantitative spectrometric data-activity relationship (QSDAR) models of binding to the corticosterone binding globulin, aromatase enzyme, and aryl hydrocarbon receptor [14][15][16][17]. Comparative structural connectivity spectra analysis (CoSCoSA) models of steroids binding to the corticosteroid were also presented [18].…”