Gene expression levels of imp genes in M. tuberculosis The relative contributions of the IMPase homologues will be proportional to their activity, and their level of expression. We therefore carried out RTq-PCR experiments to determine the levels of expression of impA, suhB, cysQ and impC mRNA in exponential cultures of M. tuberculosis. Expression levels were normalized to those of sigA mRNA which remains constant.

The level of cysQ was the highest, almost equal to sigA (Table 5). impA and impC were expressed at approximately 40% of this level, while suhB was lowest, at 12% of the cysQ level. Table 5 mRNA levels Gene mRNA level normalised to sigA* impA 0.41 (0.3- 0.5) suhB 0.11 (0.096- 0.13) impC 0.36 (0.27- 0.46) cysQ 0.95 (0.76- 1.18) *To ensure equal amounts of cDNA were used each value was standardized Compound C to sig A to generate unit-less values (95% confidence interval) Discussion To investigate how M. tuberculosis synthesises inositol, we carried out a genetic analysis

of four IMPase homologues in M. tuberculosis. The impA and suhB genes were shown to be dispensable, with no phenotype detected in terms of the levels of mycothiol, PIMs, LM or LAM. CysQ is also dispensible, although isolating the mutant ARN-509 purchase proved more difficult, requiring introduction of the M. smegmatis mspA porin gene for mutant isolation, but not for subsequent survival. It cannot be excluded, however, that the cysQ mutants that were eventually obtained had acquired a suppressor mutation, which had allowed deletion of cysQ or had allowed growth of the mutant on media lacking inositol and preventing cell

death. In contrast to these three genes, we were only able to inactivate Chlormezanone impC by introducing a second copy of the gene. The TraSH mutagenesis H 89 protocol which provides a genome-wide indication of essentiality [47] supports our data, with only impC of these four genes being reported as putatively required for optimal growth in vitro. Inositol production is likely to be essential for mycobacterial growth, because of the essentiality of both classes of mycobacterial inositol-containing molecules, namely phospholipids [8] and mycothiol Our previous work showing that a PI synthase mutant is an inositol auxotroph [23] is consistent with this. Both SuhB and CysQ have been shown to have IMPase activity [41, 48] and we have shown that the M. smegmatis ImpA has IMPase activity (unpublished data). However, none of the three mutants constructed are auxotrophic for inositol, indicating that there is potential redundancy of function between the homologs and deletion of three or four genes might be required to see sufficient loss of activity to cause auxotrophy. A recent report suggests that CysQ is likely to play a role in sulphur metabolism, as its activity as 3′-phosphoadenosine-5′-phosphatase is several orders of magnitude higher than as an inositol phosphatase [49]. However, it may still contribute to the redundancy in inositol phosphatase activity.