Send Orders for Reprints to reprints@benthamscience.ae Current Computer-Aided Drug Design, 2014, 10, 303-314 303 Theoretical Modeling of HPV: QSAR and Novodesign with Fragment Approach Girinath G. Pillai 1,2 , Lauri Sikk 1 , Tarmo Tamm 3 , Mati Karelson 1 , Peeter Burk 1 and Kaido Tämm *,1 1 Institute of Chemistry, University of Tartu, Ravila 14a, Tartu 50411, Estonia 2 Department of Chemistry, University of Florida, Gainesville, FL, 32611, USA 3 Institute of Technology, University of Tartu, Nooruse 1, Tartu50411, Estonia Abstract: Structure-activity relationships in a data set of HPV6-E1 helicase ATPase inhibitors were investigated based on two different sets of descriptors. Statistically significant four parameter Quantitative Structure-Activity Relationships (QSAR) models were constructed and validated in both cases (R 2 =0.849; R 2 cv =0.811; F=52.20; s 2 =0.25; N=42). A Fragment based QSAR (FQSAR) approach was applied for developing a fragment-QSAR equation, which enabled the construction of virtual structures for novel ATPase inhibitors with desired or pre-defined activity. Keywords: FQSAR, Fragment approach, HPV6, Papillomavirus, SAR, QSAR. INTRODUCTION Papillomaviruses (PV) are highly diverse, and occur in most mammals and birds. Hundreds of “PV types” have been detected in humans, the only intensively studied host. PVs cause benign tumors (warts, papillomas) in their natural host and occasionally in related species. Papillomas are induced in the skin and mucosal epithelia, often at specific sites of the body. Some papillomatous proliferations induced by specific types of PVs bear a high risk for malignant progression. The infection frequently leads to microlesions, which are barely or not at all visible without optical aid. PVs seem to coexist with their host over long periods of time. Many PVs appear to occur preferentially in a latent life cycle, because a wide variety of different types can be detected at random sites of healthy skin of humans and animals [1]. Papillomavirus isolates are traditionally described as “types”. PV types have been detected in all carefully examined mammals and birds, with the possible exception of laboratory mice. Human papillomaviruses (HPV) are a heterogeneous group of circular, double- stranded DNA viruses that have been identified throughout the human host. The viruses infect and replicate in the cutaneous or mucosal epithelia of humans (and other mammals). There are hundreds of genotypes of human papillomaviruses, which cause conditions ranging from plantar (HPV1) and genital warts (HPV6 and -11) to cervical cancer (HPV16, -18, and -31). HPV6 and -11 are responsible for laryngeal papillomatosis, which is an infection of the respiratory tract. HPV6 is also responsible for the majority of genital warts. Antiviral agents capable of specifically inhibiting PV replication could play an important role in the treatment of these diseases, but unfortunately no such antiviral agent *Address correspondence to this author at the Institute of Chemistry, University of Tartu, Ravila 14a, Tartu 50411, Estonia; Tel: +3727375257; Fax: +3727375264; E-mail: karu@ut.ee exists at present. The recent progress toward the identification and characterization of specific molecular targets offers the prospect of effective HPV antiviral compounds [2, 3]. Current work applies both standard and fragment based Quantitative Structure-Activity Relationships ((F)QSAR) methodology to the analysis of HPV, and is based on the experimental work done by White et al. [3], on a series of small molecules inhibiting the ATPase (Adinosine Tri- Phosphatase) activity of HPV6-E1 helicase. E1 is the only PV protein that possesses enzymatic activity and is the most highly conserved of the PV proteins. Thus, the E1 helicase has been considered the most attractive molecular target for the development of antiviral agents. QSARs are routinely used in the development of predictive models for drugs and drug candidates to estimate their activities and possible side effects. Recent QSAR approaches and reviews show promising results in several diseases, such as tuberculosis [4], HIV [5, 6] and central nervous system disorders [7, 8]. The key challenge for QSAR modelers is to find a relevant set of molecular descriptors in compliance with the statistical algorithm, which sets the correlation with the target activity of a molecular series. The whole principle is based on the assumption that structural features are responsible for the behavior of the whole molecule as a drug candidate and that there is a direct link between the internal characteristics and the biological activity of the molecular structure. METHODOLOGY Several free and commercial pieces of QSAR software and toolboxes are available for the development of models or for the prediction of properties and activity [9]. In the present study, two alternative sets of molecular descriptors from Codessa Pro [10] and PaDel-Descriptor 17-/14 $58.00+.00 © 2014 Bentham Science Publishers