Background: Dipeptides and their analogs have multiple biological functions, such as
bitter taste, angiotensin-converting enzyme (ACE) inhibition, etc. That design and modification of
dipeptides guided by theoretical models are of great significance for research and development on
drug, especially in discovery and optimization of lead compound, which can reduce the experimental
expenditure and improve the hit rate.
Methods: We used conventional 3D-QSAR methods, Comparative Molecular Field Analysis
(CoMFA) and Comparative Molecular Similarity Indices Analysis (CoMSIA), and Topomer
CoMFA to construct Three-Dimensional Quantitative Structure-Activity Relationship (3D-QSAR)
models for four catalogues of dipeptides and their derivatives, which were bitter-tasting dipeptides,
ACE-inhibitory dipeptides, anti-cryptococcal and anti-microtubule dipeptide derivatives.
The CoMFA and CoMSIA models based on a training set were optimized through varying the
force fields and their combination. Topomer CoMFA experientially completed identification and
alignment of pose of the fragments.
Results and Conclusion: We have achieved theoretical 3D-QSAR models with good reliability of
prediction for correspondingly novel samples, which had good statistically significance with excellent
cross-validated correlation coefficients (Q2 >0.5) and conventional correlation coefficients
(R2 >0.9). The counter maps from the CoMFA and CoMSIA models could provide the direct information
about contributions of the force fields for activity, aiding the design and modification of
novel bioactive dipeptides and by which relationships between bitter dipeptides with ACEinhibitory
dipeptides were explored briefly and a set of novel anti-cryptococcal dipeptide derivatives
with predicted activities were designed. Topomer CoMFA models for bitter dipeptides, ACEinhibitory
dipeptides and anti-cryptococcal dipeptidomimetics were ideal, proved the above results
and could provide a tool for virtual screening.