Wholecell screening

The simplest way to find new tuberculosis drugs is to screen for whole-cell antimycobacterial activity, much in the way that streptomycin and isoniazid were discovered. Realizing that it was impractical to do this with M. tuberculosis itself, researchers at Glaxo Wellcome employed a rapid-growing, non-pathogenic surrogate, Mycobacterium aurum A + (Chung et al 1995). Using this organism as a host for the Vibrio harveyi luxAB genes, we developed a simple bioluminescence assay to screen about 10 000 compounds/day synthesized in bead-based combinatorial libraries (G. A. C. Chung, P. Andrew, S. Polger, J. Silen, C. DeLuca-Flaherty & K. Duncan, unpublished work 1997). Other high throughput assays have been described that monitor viability with bioluminescence (Cooksey et al 1993, Arain et al 1996) or using Alamar Blue (Collins & Franzblau 1997).

Although several novel structural templates with activity against M. tuberculosis have been discovered this way, there are significant disadvantages to this approach. Many targets will be masked from the samples being tested, for example if they are intracellular, or are not expressed under the conditions used in the assay. In addition, it is often hard to improve upon the properties of a whole-cell active lead compound when there is no knowledge of the target. It is generally thought to be more effective to screen for agents that modulate the activity of a specific



Ttil/Th£ Immune Long-teím Switch stimulation immunity to inlection


tssenlial Gene prodiKts Gens products enzymes essentiaHbr sssantial toi

Ttil/Th£ Immune Long-teím Switch stimulation immunity to inlection intracellular survival in vivo surcar


BCÜ Vaccine






FIG. 2. Targets for novel tuberculosis drugs. BCG, Bacillus Calmette—Guerin; INH, isoniazid; PZA, pyrazinamide; RIF, rifampicin; Th, T helper.

target in vitro, and then to further modify any leads obtained to incorporate or improve upon the whole-cell activity. For this approach to be successful, significant effort must be applied to identifying, characterizing and validating appropriate targets. Targets identified by genetic means are not always suitable for screening; often the function of a gene product may not be known or an enzyme's substrate may not be available.

Ideally, one would wish to either hit multiple targets with a single new agent or hit a target that has a role in more than one growth phase. The advantage of the former is that it is highly unlikely that resistance will develop by a single mutation in the gene encoding the target, since the alternative target(s) would still be inhibited. Such a scenario can be envisaged for inhibition of the multiple arabinosyl transferases that recognize the substrate decaprenol-phosphoarabinose (Mikusova et al 1995, Lee et al 1995) or the three proteins that catalyse mycolyl transfer (Belisle et al 1997).

0 0

Post a comment