Archives

  • 2018-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2019-12
  • 2020-01
  • 2020-02
  • 2020-03
  • 2020-04
  • 2020-05
  • 2020-06
  • 2020-07
  • 2020-08
  • 2020-09
  • 2020-10
  • 2020-11
  • 2020-12
  • 2021-01
  • 2021-02
  • 2021-03
  • 2021-04
  • 2021-05
  • 2021-06
  • 2021-07
  • 2021-08
  • 2021-09
  • Next the BE obtained for the

    2021-07-20

    Next, the BE obtained for the different complexes were evaluated. From the BE obtained from our MD simulations, a very good binder can be differentiated from a very weak binder (−11.85 kcal/mol for MTX vs. −6.74 and −3.61 kcal/mol for compounds 14b and 15c, respectively) but ligands with similar binding affinities cannot be easily differentiated. In function of the IC50 values we were expecting exactly the opposite values for compounds 14b and 15c. Similar to these unexpected results, there are others that might be observed in Table 1S in Supplementary Material. In addition, it is important to note that the value of R2 obtained for this correlation is very low (0.49) (Fig. 3). This result clearly indicates that by using this approximation is not possible to discriminate between compounds with similar binding affinities; in addition the low correlation obtained indicates a very poor predictive power of the method. This result was not surprising since we obtained similar results in our previous work using this approach [5]. Considering that MD simulations might neglect or poorly approximate terms that are playing determinant roles (such as lone pair directionality in hydrogen bonds, explicit π … π stacking polarization effects, hydrogen bonding networks, induced fit, and conformational entropy), we cannot expect to detect clear differences between compounds possessing relatively similar BE. In the next step of our study, reduced model systems were optimized using combined semiempirical, ab initio and DFT calculations. To perform these calculations, reduced system models were employed whose design is explained in the calculation methods section. PM6 optimizations were performed considering all receptor Pentylenetetrazole that might interact after initial positioning of the ligands against Glu30 residue. Next, RHF/6-31G(d) and DFT (PBE1PBE/6-31G(d)) single point calculations were carried out for each complex optimized from PM6 computations. It is important to note that the L-R interaction is a dynamic process and therefore in order to have a more accurate description of such situation, four different snapshots for each complex were considered. This resulted in different energy values and such variation can be observed in the error bars which are shown in Fig. 4, Fig. 5, Fig. 6. Once the BE of the different complexes were obtained from the theoretical calculations, the different correlations between these theoretical calculations and our experimental data (Table 1) were calculated. The Fig. 4 shows that semi-empirical calculations (PM6) gave a correlation between the BE and the inhibitory activity with an R2 value of 0.49, which does not improve those results obtained with the LIE method (Fig. 3). However, the results obtained from RHF/6-31G(d) (Fig. 5) and PBE1PBE/6-31G(d) (Fig. 6) were significantly better with R2 values of 0.77 and 0.76, respectively. Nevertheless it is important to note that not only the value of R2 is important in a correlation but also how the distribution of the various points along the line is. Regarding Fig. 5, Fig. 6 it is evident that the different points are clustered into two well-defined groups. This would indicate that although these correlations allow us to differentiate between a very active compound with respect to a compound with low affinity, however there is room for doubt whether it is possible to distinguish between two compounds having similar affinities. To corroborate this assumption we removed the three classical type inhibitors (MTX, compounds 1 and 2) from the series and a new correlation was obtained with this new series. Such as we expected the new correlations gave very low R2 values (0.64 and 0.61 for ab initio and DFT calculations, respectively). These results show the severe limitations of these approaches in order to correctly predict the inhibitory activity between compounds with similar affinities. This result was somewhat disappointing for us so we seek a new parameter or kind of molecular descriptor that might be able to predict the inhibitory activity between compounds with similar BE in this series.