Novel paradigms for drug discovery: computational multitarget screening.

An established paradigm in current drug development is (i) to identify a single protein target whose inhibition is likely to result in the successful treatment of a disease of interest; (ii) to assay experimentally large libraries of small-molecule compounds in vitro and in vivo to identify promising inhibitors in model systems; and (iii) to determine whether the findings are extensible to humans. This complex process, which is largely based on trial and error, is risk-, time- and cost-intensive. Computational (virtual) screening of drug-like compounds simultaneously against the atomic structures of multiple protein targets, taking into account protein-inhibitor dynamics, might help to identify lead inhibitors more efficiently, particularly for complex drug-resistant diseases. Here we discuss the potential benefits of this approach, using HIV-1 and Plasmodium falciparum infections as examples. We propose a virtual drug discovery 'pipeline' that will not only identify lead inhibitors efficiently, but also help minimize side-effects and toxicity, thereby increasing the likelihood of successful therapies.

VarDetect: a nucleotide sequence variation exploratory tool.

BACKGROUND: Single nucleotide polymorphisms (SNPs) are the most commonly studied units of genetic variation. The discovery of such variation may help to identify causative gene mutations in monogenic diseases and SNPs associated with predisposing genes in complex diseases. Accurate detection of SNPs requires software that can correctly interpret chromatogram signals to nucleotides. RESULTS: We present VarDetect, a stand-alone nucleotide variation exploratory tool that automatically detects nucleotide variation from fluorescence based chromatogram traces. Accurate SNP base-calling is achieved using pre-calculated peak content ratios, and is enhanced by rules which account for common sequence reading artifacts. The proposed software tool is benchmarked against four other well-known SNP discovery software tools (PolyPhred, novoSNP, Genalys and Mutation Surveyor) using fluorescence based chromatograms from 15 human genes. These chromatograms were obtained from sequencing 16 two-pooled DNA samples; a total of 32 individual DNA samples. In this comparison of automatic SNP detection tools, VarDetect achieved the highest detection efficiency. AVAILABILITY: VarDetect is compatible with most major operating systems such as Microsoft Windows, Linux, and Mac OSX. The current version of VarDetect is freely available at http://www.biotec.or.th/GI/tools/vardetect.

Identification of potential HIV-1 targets of minocycline.

Minocycline, a broad spectrum antibiotic, has been discovered to have inhibitory activity against HIV-1 in vitro, but the targets inhibited are unknown. We used a docking with dynamics protocol developed by us to predict the binding affinities of minocycline against seven active sites of five HIV-1 proteins to putatively identify the potential target(s) of minocycline. The results indicate that minocycline has the highest predicted binding affinity against HIV-1 integrase.

Virtual screening of HIV-1 protease inhibitors against human cytomegalovirus protease using docking and molecular dynamics.

The clearance of cytomegalovirus viraemia in HIV-1-infected patients may partly result from the inhibition of cytomegalovirus protease by HIV-1 protease inhibitors contained in highly active antiretroviral therapy. We used a computational method to calculate the binding affinity of six HIV-1 protease inhibitors to cytomegalovirus protease based on its X-ray crystallography structure. The calculations showed that amprenavir and indinavir occupy the substrate-binding site of the cytomegalovirus protease with high affinity, and may be implicated in alleviating cytomegalovirus infection.

Structural insights into the cellular retinaldehyde-binding protein (CRALBP).

Cellular retinaldehyde-binding protein (CRALBP) is an essential protein in the human visual cycle without a known three-dimensional structure. Previous studies associate retinal pathologies to specific mutations in the CRALBP protein. Here we use homology modeling and molecular dynamics methods to investigate the structural mechanisms by which CRALBP functions in the visual cycle. We have constructed two conformations of CRALBP representing two states in the process of ligand association and dissociation. Notably, our homology models map the pathology-associated mutations either directly in or adjacent to the putative ligand-binding cavity. Furthermore, six novel residues have been identified to be crucial for the hinge movement of the lipid-exchange loop in CRALBP. We conclude that the binding and release of retinoid involve large conformational changes in the lipid-exchange loop at the entrance of the ligand-binding cavity.

Heptad-repeat-2 mutations enhance the stability of the enfuvirtide-resistant HIV-1 gp41 hairpin structure.

Enfuvirtide (T20) is a peptide-based fusion inhibitor derived from the heptad repeat 2 (HR2) region of HIV-1 glycoprotein 41 (gp41). The inhibitor binds to the gp41 heptad repeat 1 (HR1) region, thereby blocking viral HR1/HR2 association. Mutations in HR1 have been reported to cause enfuvirtide resistance and reduce viral fitness. In this study, we first showed that scores obtained by a residue-specific all-atom probability discriminatory function (RAPDF) may be used as a reliable predictor of structural stability of gp41 mutants by comparing it to experimentally determined melting temperatures, and as a reliable indicator of enfuvirtide resistance by comparing it to experimentally determined fusion inhibition and viral fitness levels. We then generated an initial set of 28 theoretical structures of the HR1/HR2 hairpin complex where each structure consists of one mutation on HR1 known to cause enfuvirtide resistance and a wild-type amino acid at the corresponding HR2 residue. Mutations were then introduced in the corresponding HR2 residue of each structure where the wild-type amino acid was changed to each of the other nineteen amino acids. The enfuvirtide-resistant HR1 mutants with compensatory mutations at the corresponding HR2 residues had better RAPDF scores than those HR1 mutants with wild-type HR2. This indicates that mutations in HR2 improve structural stability of the HR1/HR2 hairpin complex and may lead to enhanced enfuvirtide resistance when present with resistant HR1 mutations. Modification of the amino acid side chains that contribute to enfuvirtide resistance using the RAPDF scores as a guide may help design of a second generation of fusion inhibitors against the enfuvirtide-resistant strains.

Prediction of HIV-1 protease inhibitor resistance using a protein-inhibitor flexible docking approach.

Emergence of drug resistance remains one of the most challenging issues in the treatment of HIV-1 infection. Here we focus on resistance to HIV-1 protease inhibitors (PIs) at a molecular level, which can be analysed genotypically or phenotypically. Genotypic assays are based on the analysis of mutations associated with reduced drug susceptibility, but are problematic because of the numerous mutations and mutational patterns that confer drug resistance. Phenotypic resistance or susceptibility can be experimentally evaluated by measuring the amount of free drug bound to HIV-1 protease molecules, but this procedure is expensive and time-consuming. To overcome these problems, we have developed a docking protocol that takes protein-inhibitor flexibility into account to predict phenotypic drug resistance. For six FDA-approved Pls and a total of 1792 HIV-1 protease sequence mutants, we used a combination of inhibitor flexible docking and molecular dynamics (MD) simulations to calculate protein-inhibitor binding energies. Prediction results were expressed as fold changes of the calculated inhibitory constant (Ki), and the samples predicted to have fold-increase in calculated Ki above the fixed cut-off were defined as drug resistant. Our combined docking and MD protocol achieved accuracies ranging from 72-83% in predicting resistance/susceptibility for five of the six drugs evaluated. Evaluating the method only on samples where our predictions concurred with established knowledge-based methods resulted in increased accuracies of 83-94% for the six drugs. The results suggest that a physics-based approach, which is readily applicable to any novel PI and/or mutant, can be used judiciously with knowledge-based approaches that require experimental training data to devise accurate models of HIV-1 Pl resistance prediction.

Simple linear model provides highly accurate genotypic predictions of HIV-1 drug resistance.

Drug resistance is a major obstacle to the successful treatment of HIV-1 infection. Genotypic assays are used widely to provide indirect evidence of drug resistance, but the performance of these assays has been mixed. We used standard stepwise linear regression to construct drug resistance models for seven protease inhibitors and 10 reverse transcriptase inhibitors using data obtained from the Stanford HIV drug resistance database. We evaluated these models by hold-one-out experiments and by tests on an independent dataset. Our linear model outperformed other publicly available genotypic interpretation algorithms, including decision tree, support vector machine and four rules-based algorithms (HIVdb, VGI, ANRS and Rega) under both tests. Interestingly, our model did well despite the absence of any terms for interactions between different residues in protease or reverse transcriptase. The resulting linear models are easy to understand and can potentially assist in choosing combination therapy regimens.

Improved prediction of HIV-1 protease-inhibitor binding energies by molecular dynamics simulations.

BACKGROUND: The accurate prediction of enzyme-substrate interaction energies is one of the major challenges in computational biology. This study describes the improvement of protein-ligand binding energy prediction by incorporating protein flexibility through the use of molecular dynamics (MD) simulations. RESULTS: Docking experiments were undertaken using the program AutoDock for twenty-five HIV-1 protease-inhibitor complexes determined by x-ray crystallography. Protein-rigid docking without any dynamics produced a low correlation of 0.38 between the experimental and calculated binding energies. Correlations improved significantly for all time scales of MD simulations of the receptor-ligand complex. The highest correlation coefficient of 0.87 between the experimental and calculated energies was obtained after 0.1 picoseconds of dynamics simulation. CONCLUSION: Our results indicate that relaxation of protein complexes by MD simulation is useful and necessary to obtain binding energies that are representative of the experimentally determined values.

Prevalence of genotypic HIV-1 drug resistance in Thailand, 2002.

BACKGROUND: The prices of reverse transcriptase (RT) inhibitors in Thailand have been reduced since December 1, 2001. It is expected that reduction in the price of these inhibitors may influence the drug resistance mutation pattern of HIV-1 among infected people. This study reports the frequency of HIV-1 genetic mutation associated with drug resistance in antiretroviral-treated patients from Thailand. METHODS: Genotypic resistance testing was performed on samples collected in 2002 from 88 HIV-1 infected individuals. Automated DNA sequencing was used to genotype the HIV-1 polymerase gene isolated from patients' plasma. RESULTS: Resistance to protease inhibitors, nucleoside and non-nucleoside reverse transcriptase inhibitors were found in 10 (12%), 42 (48%) and 19 (21%) patients, respectively. The most common drug resistance mutations in the protease gene were at codon 82 (8%), 90 (7%) and 54 (6%), whereas resistant mutations at codon 215 (45%), 67 (40%), 41 (38%) and 184 (27%) were commonly found in the RT gene. This finding indicates that genotypic resistance to nucleoside reverse transcriptase inhibitors was prevalent in 2002. The frequency of resistant mutations corresponding to non-nucleoside reverse transcriptase inhibitors was three times higher-, while resistant mutation corresponding to protease inhibitors was two times lower than those frequencies determined in 2001. CONCLUSION: This study shows that the frequencies of RT inhibitor resistance mutations have been increased after the reduction in the price of RT inhibitors since December 2001. We believe that this was an important factor that influenced the mutation patterns of HIV-1 protease and RT genes in Thailand.

The effects of antiretroviral dose modification on the re-emergence of HIV-1 wild-type strains.

Four human immunodeficiency virus type 1 (HIV-1) treatment-naïve Thai patients began antiretrovival therapy with a triple drug regimen -zidovudine plus lamivudine plus indinavir; this regimen was modified at week 20 of therapy because of drug toxicity. The virus in all patients was suppressed to lower than 400 copies/ml while they were taking the triple antiretroviral drug regimen. However, suppression was lost after changing the antiretroviral regimen. A comparison of HIV-1 DNA sequences taken from the baseline (day 0) and week 24 showed no significant overgrowth in HIV-1 drug-resistant strains. There was no difference in the protease and reverse transcriptase (RT) mutation profiles. Resistant variants did not emerge, even after sub-therapeutic levels of antiretroviral drugs had been introduced to these patients for 4 weeks. These findings may have clinical implications for long-term treatment strategies.

Prevalence of HIV-1 drug resistance in antiretroviral-naive pregnant Thai women.

HIV-1 drug resistance may limit the use of antiretrovirals when attempting to reduce the vertical transmission rate. Establishing the prevalence of the HIV-1 mutations associated with antiretroviral resistance in pregnant women will enable clinicians to maximize the chances of preventing vertical transmission. In order to determine the prevalence of HIV-1 resistant strains among antiretroviral-naive pregnant Thai women, the nucleotide sequences of the HIV-1 polymerase (pol) gene were evaluated. The plasma samples were collected from the women during the 34th week of pregnancy: numerous secondary mutations could be found in the reverse transcriptase (RT) and protease gene, while no primary mutations in the pol gene were found. The result also showed that by detecting the delta32bp deletion within the CCR 5 locus, it was evident that none of HIV-1 infected individuals had homozygous or heterozygous delta32bp deletions of the CCR5 gene; moreover, no CCR5 gene mutations were found in any individual.

Prevalence of HIV-1 polymerase gene mutations in pre-treated patients in Thailand.

To determine the prevalence of drug resistance-conferring mutations in human immunodeficiency virus type 1 (HIV-1), 83 HIV-1 infected Thai patients who had been treated with any antiretroviral drug were studied. HIV-1 RNA was reverse transcribed and amplified by RT-PCR. The direct sequencing of HIV-1 reverse transcriptase (RT) and protease was then performed. Changes in nucleotide and amino acid sequences were determined by comparison with a pNL4-3 reference sequence. Data on mutations associated with resistance to antiretroviral drugs were obtained from literature. The mutations associated with lamivudine resistance (M184V/I) were found most often (in 45.7% of individuals). Zidovudine-resistant mutants: T215Y/F (36%), M41L (28%) and K70R (25.3%) were common; but mutations linked to didanosine (L74V) and multinucleoside-resistant genotypes (Q151M) were rarely recognized (2.4% and 3.6%, respectively). The stavudine-resistant mutant (V75T) and T69 insertions were not found. All subjects who had a significant exposure to antiretroviral drugs and current virological failure in the past carried drug-resistant genotypes. Genotypic resistance to zidovudine, lamivudine, zalcitabine, indinavir and ritonavir appeared in more than one third of the samples, which suggested that the prevalence of the HIV-1 resistance-conferring genotype resisting reverse transcriptase inhibitors and/or protease inhibitors was high in treatment experienced patients.

Homology modeling of mosquito cytochrome P450 enzymes involved in pyrethroid metabolism: insights into differences in substrate selectivity.

BACKGROUND: Cytochrome P450 enzymes (P450s) have been implicated in insecticide resistance. Anopheles minumus mosquito P450 isoforms CYP6AA3 and CYP6P7 are capable of metabolizing pyrethroid insecticides, however CYP6P8 lacks activity against this class of compounds. FINDINGS: Homology models of the three An. minimus P450 enzymes were constructed using the multiple template alignment method. The predicted enzyme model structures were compared and used for molecular docking with insecticides and compared with results of in vitro enzymatic assays. The three model structures comprise common P450 folds but differences in geometry of their active-site cavities and substrate access channels are prominent. The CYP6AA3 model has a large active site allowing it to accommodate multiple conformations of pyrethroids. The predicted CYP6P7 active site is more constrained and less accessible to binding of pyrethroids. Moreover the predicted hydrophobic interface in the active-site cavities of CYP6AA3 and CYP6P7 may contribute to their substrate selectivity. The absence of CYP6P8 activity toward pyrethroids appears to be due to its small substrate access channel and the presence of R114 and R216 that may prevent access of pyrethroids to the enzyme heme center. CONCLUSIONS: Differences in active site topologies among CYPAA3, CYP6P7, and CYP6P8 enzymes may impact substrate binding and selectivity. Information obtained using homology models has the potential to enhance the understanding of pyrethroid metabolism and detoxification mediated by P450 enzymes.

Binding capacity of ER-α ligands and SERMs: comparison of the human, dog and cat.

The estrogen molecule is the major risk factor related to mammary gland tumors, with estrogen receptor alpha (ER- α) as the important target stimulating growth. Therefore one alternative approach to treatment of breast cancer is to use selective estrogen receptor modulator (SERM), hormonal therapy. In this study, the structures of ER- α in humans, dogs and cats were predicted using the amino acid sequencing data bank and corrected for general protein structures, receptor sites and docking by adding 2,344 ligands with 15 SERMs into the database and calculating estimated inhibition constants (Ki). Thereby, ranking of best ligands of SERMs in humans, dogs and cats could be achieved. The results show that the shapes of ER- α differ between species but the major pocket sites are the same. Bazedoxifene, a new SERM proved to be the best estrogen antagonist and ER- α inhibitor in all species (human, dog, cat) with the lowest Ki. The other good ligands for dogs and cats are Neohesperidin, Dihydrochalcone, and Schreiber2. The differences in these protein structures may explain why there are only a few SERMs or other ligands which can be used as anti-cancer drugs.

Viral entry inhibitors block dengue antibody-dependent enhancement in vitro.

Severe dengue virus (DENV) disease symptoms, including dengue hemorrhagic fever and dengue shock syndrome, have been correlated with the presence of pre-existing antibodies that enhance rather than neutralize infections in Fc receptor bearing cells. These antibodies can originate from previous infection with a different serotype of dengue, or from waning antibody titers that occur in infants and young children as they are weaned from breast milk that contains protective dengue-specific antibodies. Despite the apparent importance of this antibody dependent enhancement (ADE) effect, there has been no description of any specific inhibitors of this process. We explored DENV entry inhibitors as a potential strategy to block ADE. Two different peptide entry inhibitors were tested for the ability to block antibody-mediated DENV-2 infection of human, FcRII bearing K562 cells in vitro. Both peptides were able to inhibit ADE, showing that entry inhibitors are possible candidates for the development of specific treatment for severe DENV infection.