Publication: Proteochemometric model for predicting the inhibition of penicillin-binding proteins
Issued Date
2015-01-01
Resource Type
ISSN
15734951
0920654X
0920654X
Other identifier(s)
2-s2.0-84922005039
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
Journal of Computer-Aided Molecular Design. Vol.29, No.2 (2015), 127-141
Suggested Citation
Sunanta Nabu, Chanin Nantasenamat, Wiwat Owasirikul, Ratana Lawung, Chartchalerm Isarankura-Na-Ayudhya, Maris Lapins, Jarl E.S. Wikberg, Virapong Prachayasittikul Proteochemometric model for predicting the inhibition of penicillin-binding proteins. Journal of Computer-Aided Molecular Design. Vol.29, No.2 (2015), 127-141. doi:10.1007/s10822-014-9809-0 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/35779
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Proteochemometric model for predicting the inhibition of penicillin-binding proteins
Other Contributor(s)
Abstract
© 2014 Springer International Publishing Switzerland. Neisseria gonorrhoeae infection threatens to become an untreatable sexually transmitted disease in the near future owing to the increasing emergence of N. gonorrhoeae strains with reduced susceptibility and resistance to the extended-spectrum cephalosporins (ESCs), i.e. ceftriaxone and cefixime, which are the last remaining option for first-line treatment of gonorrhea. Alteration of the penA gene, encoding penicillin-binding protein 2 (PBP2), is the main mechanism conferring penicillin resistance including reduced susceptibility and resistance to ESCs. To predict and investigate putative amino acid mutations causing β-lactam resistance particularly for ESCs, we applied proteochemometric modeling to generalize N. gonorrhoeae susceptibility data for predicting the interaction of PBP2 with therapeutic β-lactam antibiotics. This was afforded by correlating publicly available data on antimicrobial susceptibility of wild-type and mutant N. gonorrhoeae strains for penicillin-G, cefixime and ceftriaxone with 50 PBP2 protein sequence data using partial least-squares projections to latent structures. The generated model revealed excellent predictability (R <sup>2</sup> = 0.91, Q <sup>2</sup> = 0.77, Q <inf>Ext</inf> <sup>2</sup> = 0.78). Moreover, our model identified amino acid mutations in PBP2 with the highest impact on antimicrobial susceptibility and provided information on physicochemical properties of amino acid mutations affecting antimicrobial susceptibility. Our model thus provided insight into the physicochemical basis for resistance development in PBP2 suggesting its use for predicting and monitoring novel PBP2 mutations that may emerge in the future.