Pharmaceutical Chemistry

In this work, we present a new GIAO-HDFT universal scaling factor (mPW1PW91/3- 21G//PM7(I)) and a comparative study in which is investigated its ability to predict NMR 13C chemical shifts (δ) with high cost-effectiveness ratio. A set of 22 small molecules providing 27 different 13C δ determined in the gas phase was used for all scaling factors protocols: B3PW91/ccpVDZ//B3PW91/cc-pVDZ (II), B3PW91/cc-pVTZ//B3PW91/cc-pVDZ (III), B3LYP/6- 311+(2d,p)//B3LYP/6-31G(d) (IV), mPW1PW91/6-31G(d)//PM7(V) , mPW1PW91/6- 31G(d)//mPW1PW91/6-31G(d) (VI). Despite the calculation approximations the δ calculated at the

GIAO-mPW1PW91/3-21G//PM7 using a simple relationship (δscal = 1.14 . δcalc – 4.7, where δcalc and δscal are the calculated and the linearly scaled values of the 13C δ, respectively) were able to yield MAD and RMS errors as small as those obtained with other GIAO-HDFT with bigger basis sets

(protocols (II) to (VI)). The robustness of the new protocol and its applicability to practical problems was evaluated by the calculation of the δ for two natural compounds with synthesis, biological and therapeutic interest: tryptanthrin (indolo[2,1-b]quinazoline-6,12-dione) and (-)- loliolide (7aR)-6-hydroxy-4,4,7a-trimethyl-6,7-dihydro-5H-1-benzofuran-2-one. For both compound, the 6 protocols presented good agreement with experimental data. Moreover, for the second compound, the new protocol performs even better than the 5 others. In conclusion, GIAOmPW1PW91/ 3-21G//PM7 linear regression obtained by using the experimental and the calculated data, is a very attractive tool as an alternative to more computationally demanding approaches, which are usually applied in order to achieve 13C NMR δ calculations.