Abstract: The study describes SAR and QSAR of inhibition of Reverse Transcriptase-1 by HEPT derivatives using both classical and non-conventional physicochemical parameters along with hydrophobic parameter and indicator parameters. The set of HEPT derivatives Studied contains 85 compounds with different substitution at various positions. Application of multiple linear regression analysis indicated that combination of classical physicochemical parameters with indicator parameters yielded statistically significant model for modeling inhibitory activity (log1/C) against Reverse Transcriptase-1. Final selection of the potential HEPT derivative for the inhibition of Reverse Transcriptase-1 is made with the help of molecular modeling parameters.
Introduction
The Acquired Immuno Deficiency Syndrome (AIDS) is one of the most rapid sprayed diseases, which is caused by infection of human immuno deficiency virus (HIV). During the studies of effective therapies used to inhibits the compassing of HIV, Reverse Transcriptase (RT) has been distinguish as promising target, because Reverse Transcriptase (RT) is not required for normal host cell replication (Clercq, 1995) and Conversion process of the single-stranded viral RNA genome into double-stranded proviral DNA ahead of its adherence into the host genomic DNA is performed by the reverse transcriptase (Young, 1993). However, a serious puzzle with the reverse transcriptase inhibitors specifically HEPT derivatives is the outburst of viral strains that have point mutations in the region encoding HIV-1 RT which check these drugs from inhibiting RT (Hannongbua et al., 1996a, b; Kireev et al., 1997).
Inspired by the behavior of HEPT derivatives in inhibition of RT-1 and in continuation
to our earlier studies (Thakur, 2005; Balaban et al., 2005; Thakur et
al., 2004a-e, 2005), our objective in present study is to made SAR and QSAR
analysis of inhibition of reverse transcripase-1 by HEPT derivatives using three
different sets of molecular and hydrophobic descriptors consisting of some non-conventional
physicochemical parameters like approximate surface area (ASA), surface area
grid (SAG) and Hydration Energy (HE), hydrophobic parameter as logP and some
classical physicochemical properties like Molar Refractivity (MR), Molar Volume
(MV), Parachor (Pc), refractive index (η), Surface Tension (ST) and density
(d) in addition to indicator parameters. For QSAR modeling we have used maximum
R2 method and followed step-wise regression analysis (Chaterjee and
Hadi, 2000). To model the most potent HEPT derivative we optimized the molecules
using molecular mechanics method, applying MM+ force field. The parent
structure of the HEPT derivatives is presented as Fig. 1.
| Fig. 1: | Parent structure for HEPT derivatives |
Materials and Methods
Activity
The inhibitory action against RT’s of the HEPT derivatives were adopted
from the literature (Hannongbua et al., 2001).
Physicochemical Parameters
The classical physicochemical parameters Molar volume(MV), Parachor(Pc),
Molar Refractivity (MR), Refractive Index (η), Surface Tension (ST), Density
(D) and Polarizability (α), for the set of HEPT derivatives were calculated
from ACD Lab software [www.acdlabs.com].
While the non-conventional physicochemical parameters, ASA, SAG and hydration
energy were calculated with the help of Hyperchem7 (Demo-version).
Indicator Parameters
These are the dummy parameters sometimes used for accounting those structural
feature not covered in any molecular descriptor used. They assumed only two
values 1 or 0. If the assumed structural feature is present; then the indicator
parameters are 1 otherwise it is 0.
Molecular Modeling
Molecular optimization and modeling were performed applying MM+
force field for this purpose and for the calculation of modeling parameters
Hyperchem7 (Demo-version) [www.hyper.com]
were used.
Statistical Analysis
Maximum R2 method together with step-wise regression (Chaterjee and
Hadi, 2000) was carried for arriving at statistically significant models. In
present study linear mathematical models are developed to study Quantitative
Structure/Property- Activity Relationship (QSAR). Multiple linear regression
is used to develop these models.
Results and Discussion
The set of 85 HEPT derivatives and their adopted inhibition values for HIV-1 RT expressed as log1/C are presented in Table 1. Non-conventional physicochemical parameters are recorded in Table 2 while the classical physicochemical parameters are shown in Table 3. Hydrophobic parameter logP along with indicator parameters are recorded in Table 4. The inter correlation of classical and non-conventional physicochemical parameters and hydrophobic parameter logP are presented in Table 5 in form of correlation matrix.
The inter correlatedness among molecular descriptors as well as with the activity
shows that the mutual correlation exists between classical physicochemical parameters,
while this is not so with the other non-conventional physicochemical parameters
used.
| Table 1: | Sub-stituents and Biological activity (Observed)[] of HEPT derivatives used in present study |
| Table 1: | Continued |
| Table 2: | Non conventional physicochemical parameters used in present study |
| * HE = Hydration Energy, ASA = Approximate Surface Area, SAG = Surface Area Grid | |
| Table 3: | Classical physicochemical properties* of HEPT derivatives used in present study |
| Table 3: | Continued |
| Table 3: | Continued |
| *MR = Molar Refractivity, MV = Molar Volume, Pc = Parachor, η = Index of refraction, ST = Surface Tension, D = Density, α = Polarizability | |
| Table 4: | logP values and indicator parameters for HEPT derivatives used in present study |
| Table 4: | Continued |
| * logP = Partition coefficient (Octenol/Water), ITC = 1 if Cyclic structure at terminal of the chain at R3 position, 0 otherwise, IAC = 1 if Cyclic structure at alternate atoms, 0 otherwise, I6 = 1 if S atom attached at R2 position, 0 otherwise, ISP = 1 if substituted Phenyl ring present at the atom attached on R2, 0 otherwise, IOH = 1 if OH present in the chain at R3 position, 0 otherwise, I2 = 1 if S atom present at X position, 0 otherwise | |
| Table 5: | Correlation between various parameters and biological activity in form of correlation matrix |
| Table 5: | Continued |
| Table 5: | Continued |
| Table 5: | Continued |
Furthermore, data presented in Table 5 shows that none of the molecular descriptors, including non-conventional physicochemical parameters, hydrophobic parameters and classical physicochemical parameters correlate well with the activity (log1/C). From this we conclude that these descriptors can be combined to yield statistically significant multi-parametric model for modeling the activity. Initial regression analysis indicated that out of the 12 molecular descriptors Surface Area Grid (SAG), logP, Molar Volume (MV), Parachor (Pc) and index of refraction (η) in combination of indicator parameters plays the dominating role in modeling the activity.
In the case of non-conventional physicochemical parameters, from the perusal of Table 5 non-of the non-conventional physicochemical parameter shows the significant univariate correlation. For the improvement in the modeling potential we test bi, tri and tetra-parametric combinations of non-conventional physicochemical parameters and indicator parameters. Best model obtained from the penta-parametric combination of SAG, indicator parameters I2, I6, ISP and IOH. Model obtained from these parameters is as below:
log1/C = 0.0055(±0.0021) SAG + 1.2431(±0.2303) I2
- 1.2962(±0.2405) I6 + |
(1) |
The statistics obtained from model demonstrate the role of volumetric parameters in the modeling of activity log1/C. Positive coefficient of indicator parameters I2 and ISP in Eq. 1 exhibits the enhancement in the activity with the substitution at 2nd position and presence of substituted phenyl ring. While the negative coefficient of indicator parameters I6 and IOH shows the inverse relationship between biological activity log1/C and presence of substitution at 6th position and OH substitution, respectively.
Similarly in case of hydrophobic parameter the best result is obtained from
the penta parametric combination of logP and 4 indicator parameters I2,
I6, ISP and IOH. The model obtained from above
variable is shown below:
log1/C = 0.2748(±0.0991)logP + 1.1612(±0.2392) I2
− 1.2247(±0.2396) I6 + |
(2) |
Statistics generated by Eq. 2 express the enhancement in the activity log1/C with the increase in hydrophobicity of compounds i.e, increase in lipophilicity in these compounds enhance the inhibitory action. Equation also exhibits the role of indicator parameters similar to the previous Eq. 1.
In case of classical physicochemical parameters only Molar Volume (MV) shows
the significant univariate correlation (r = 0.50). In case of bi, tri and tetra-parametric
correlation, few significant results are shown by the different combination
of MV, Pc, ST, η and MR along with the indicator parameters. The best result
is obtained from the combination of Molar Volume (MV), Parachor (Pc) and index
of refraction (η) along with indicator parameters I2 and ISP
and the model obtained is as below:
log1/C = 0.1618(±0.0253)MV − 0.0557(±0.0095)Pc
+ 22.4781(±6.53) η + |
(3) |
Equation 3 expresses the domination of volumetric parameters in modeling of inhibitory activity of NNRT’s against the reverse transcriptase1. The equation also demonstrates the role of steric parameters in modeling the activity. Equation 3 again shows the similar behavior of indicator parameters I2 and ISP as Eq. 1 and 2.
Comparison of all three equations demonstrates that the highest Regression (R) value is obtained from the Eq. 3 this exhibits significant role of classical physicochemical parameters in modeling log1/C. This also demonstrates the dominating role of substitution at 2nd position and presence of substituted phenyl ring on the parent moiety. Comparison also expresses the domination of volumetric parameters over the hydrophobic and steric parameters in modeling the inhibitory activity of the compounds against RT-1.
In view of this above, we have concentrated on the results given by Eq.
3. Further regression analysis indicated that the model expressed by Eq.
3 has ten outliers in three different steps (compounds 2, 14, 18, 45, 46,
78, 79, 81, 82 and 83), the deletion of which give the following models with
excellent statistics:
log1/C = 0.1507(±0.0217)MV − 0.0509(±0.0082)Pc
+ 19.4523(±5.6290) η + |
(4) |
log1/C = 0.1563(±0.0203)MV − 0.0530(±0.0076)Pc
+ 19.8637(±5.1838) η + |
((5) |
log1/C = 0.1508(±0.0200)MV − 0.0511(±0.0075)Pc
+ 18.1677(±5.1166) η + |
(6) |
Comparison of Eq. 4-6 shows that the model obtained for the set of 75 compounds gives the better statistics and most suitable for the prediction of inhibitory activity of the compounds against reverse transcriptase-1. It is obvious that reduction in size of data set increases the regression value, but in present case significant lowering of Se and a large improvement in the F-statistics along with the improvement in the value of R2A from Eq. 3-6 justify the improvement in statistics and deletion of the compounds. The suitability of Eq. 6 for the prediction of log1/C expressed graphically in Fig. 2. Also, the observed and calculated log1/C along with residual values are presented in Table 6
At this stage, it is worthy to comment on R2A values.
We observed that as we passes from the model obtained for 85 compounds (Eq.
1-3) to model obtained for 75 compounds (Eq.
6) there is consistent increase in the value of R2A.
| Fig. 2: | Graph obtained between Obs. and Calc. log1/C values from Eq. 6 |
| Table 6: | Observed and calculated biological activity (log1/C) using Eq. 6 for HEPT derivatives |
| *Data point not included in calculation | |
The values increasing from 0.7022 to 0.8301, as we passes from Eq. 3-6. Such an increase in R2A values indicate that the deleted compounds have the unfair share in the modeling of respective activity and also showing exceptional behavior from their parent series. The value of R2A will decrease if the deletion of the compounds does not reduce the unexplained variation in the model enough to off set the loss of degree of freedom (Agrawal et al., 2001, Khadikar et al., 2002a,b; Thakur et al., 2003, 2004).
Based on the magnitude of residue from Eq. 6 we have selected
compounds 7, 9, 10, 27, 33, 52, 59 and the compound 73 for further molecular
modeling. This we have done to find out which HEPT derivative has the highest
correlative and predictive potential. We have, therefore, attempted molecular
modeling, using Hyperchem software[www.hyper.com].
The molecular modeling is demonstrated in Fig. 3-10,
respectively for compounds 7, 9, 10, 27, 33, 52, 59 and 73. The corresponding
molecular modeling parameters are presented in Table 7. In
order to resolve our problem of selecting out the HEPT derivative with the best
quality and desired potential; we have carried out further regression analysis
using the molecular modeling parameters from Table 7.
| Fig. 3: | Optimized structure of Comp. No. 7 |
| Fig. 4: | Optimized structure of Comp. No. 9 |
| Fig. 5: | Optimized structure of Comp. No. 10 |
| Fig. 6: | Optimized structure of Comp. No. 27 |
| Fig. 7: | Optimized structure of Comp. No. 33 |
| Fig. 8: | Optimized structure of Comp. No. 52 |
| Fig. 9: | Optimized structure of Comp. No. 59 |
| Fig. 10: | Optimized structure of Comp. No. 73 |
| Fig. 11: | Graph obtained between calculated and observed log1/C values from Eq. 7 |
| Table 7: | Molecular modeling parameters for compounds having minimum residue |
| * TE = Total Energy, DpM = Dipole Moment, RMSg = Root Mean Square Gradiant | |
Non-of the modeling parameter shows the significant univariate correlation
but result obtained from bivariate correlation is significant and model is shown
below:
log1/C = 2.222(±1.4249) RMSg - 0.1155 (±0.0621) TE +
6.7910(7) |
(7) |
Equation 7 demonstrates that the compound having the minimum Total Energy (TE) and highest RMSg is favorable for the inhibition activity against RT’s for the HEPT derivatives. Graphical representation of the model (Fig. 11) shows that the compound no 27 has the maximum predictive potential and most suitable for the modeling.
Conclusions
From the result and discussion made above we conclude that the classical physicochemical parameters can be used successfully for modeling inhibition of reverse transcriptase-1 by HEPT derivatives and that for the present set of HEPT derivatives MV is find to be prominent. The results also indicate that combination of classical physicochemical parameters and molecular (3D) modeling can be used for select the compound with potential activity.
Acknowledgment
Author’s thanks are due to Prof. Padmakar V Khadikar for his excellent guidance to perform this study.