|dc.description.abstract||Tuberculosis continues to be one of the major causes of morbidity and mortality worldwide. Several mycobacterial species such as M. tuberculosis and M. africanum are responsible for causing this disease in humans. Reports of high cAMP levels in mycobacterial species (as compared to other bacteria such as E. coli) suggested that this second messenger may play an important role in the biology of mycobacteria. Further, it was reported that infection with mycobacteria led to an increase in the cAMP levels within the host macrophage. More recent studies have shown that this cAMP increase may be due to bacterially derived cAMP, hinting at a role for cAMP in mycobacterial pathogenesis. Given this background, the study of cAMP in mycobacteria proves to be an interesting field of research.
Signalling through cAMP involves an interaction of this cyclic nucleotide with a cAMP-binding protein. These proteins typically contain a cyclic nucleotide-binding domain (CNB domain) linked to another (effector) domain. The CNB domain is thought to allosterically control the activity of the effector domain, thus mediating cellular responses to altered cAMP levels. For example, in the case of eukaryotic protein kinase A (PKA), binding of cAMP to the CNB domain results in relieving the inhibitory effects of the regulatory subunit on the catalytic subunit. The catalytic subunit then phosphorylates its target substrates, eliciting a variety of cellular responses.
This work involves the characterisation of novel cAMP-binding proteins from mycobacteria, in an attempt to better understand cAMP signalling mechanisms in these organisms. The genome of M .tuberculosis H37Rv is predicted to code for ten CNB domain-containing proteins. One of these genes is Rv0998 (KATmt). KATmt was found to contain a GCN5 related N-acetyltransferase (GNAT) domain linked to a CNB domain. KATmt finds orthologues throughout the genus Mycobacterium, thereby suggesting its role in the basic physiology of these organisms. In addition, such a domain fusion is unique to mycobacteria and hence promises to deliver insights into the biology of this medically important genus. Presented here are the biochemical and functional characterisation of KATmt and its orthologue from M. smegmatis, MSMEG_5458 (KATms). Recombinant KATms bound cAMP with high affinity, validating the functionality of its CNB domain. Mutational and analogue-binding studies showed that the biochemical properties of the CNB domain were similar to mammalian protein kinase A and G-like CNB domains. The substrate for the GNAT acetyltransferase domain was identified to be a universal stress protein from M. smegmatis (MSMEG_4207). MSMEG_4207 was acetylated at a single lysine residue (Lys 104) by KATms in vitro. Further, cAMP binding to KATms increased the initial rate of acetylation of MSMEG_4207 by 2.5-fold, suggesting allosteric control of acetyltransferase activity by the CNB domain. To ascertain that KATms acetylated MEMEG_4207 in vivo, an in-frame deletion of the KATms gene was generated in M. smegmatis (ΔKATms). MSMEG_4207 was immunoprecipitated from wild-type M. smegmatis and the ΔKATms strains, followed by mass spectrometric analysis. Acetylated MSMEG_4207 was only present in the wild-type strain, confirming that KATms and MSMEG_4207 is an in vivo enzyme-substrate pair. Key biochemical differences were observed between KATms and KATmt. KATmt had an affinity for cAMP in the micromolar range, close to three log orders lower than that of KATms. In addition, KATmt showed strictly cAMP-dependent acetylation of MSMEG_4207. This demonstrates that orthologous proteins often evolve under varied selective pressures, resulting in divergent properties.
Using a combination of bioluminescence resonance energy transfer (BRET) and amide hydrogen/deuterium exchange mass spectrometry (HDXMS), the conformational changes that occur upon cAMP binding to the CNB domain of KATms were monitored. A BRET-based conformation sensor was constructed for KATms by inserting KATms between GFP2 (green fluorescent protein) and Rluc (Renilla luciferase). An increase in BRET upon cAMP binding to the sensor was observed. HDXMS analysis revealed that
besides the CNB domain, the only other region that showed conformational changes in KATms upon cAMP-binding was the linker region. To confirm that the linker region was important in propagating the effects of cAMP-binding to the acetyltransferase domain, an additional construct for BRET analysis encompassing the CNB domain and the linker region was generated. The magnitude of the increase in BRET was similar to the full length BRET-based sensor, validating the crucial role of the linker region in propagating cAMP-mediated conformational changes. A ‘PXXP’ motif found in the linker region, showed maximum exchange in HDXMS analysis. Mutation of both these proline residues to alanine in KATms, as well as KATmt, resulted in decoupling of cAMP-binding and allosteric potentiation of acetyltransferase activity. In contrast to the intricate parallel allosteric relays observed in other CNB domain-containing proteins, the CNB domain in KATms functions as a simpler cyclic nucleotide binding-induced switch involving stabilization of the CNB and linker domain alone. Therefore, KATms is an example of a primordial CNB domain where conformational changes are a consequence of binding-induced ordering alone.
Using a computational approach, putative substrate proteins of KATmt from M. tuberculosis were identified. The substrate specificity of lysine acetyltransferases is determined loosely by a consensus sequence around the lysine residue which is acetylated. Using this property of protein acetyltransferases, the genome of M. tuberculosis H37Rv was mined for proteins harboring lysine residues in a similar sequence context as seen in MSMEG_4207. In vitro biochemical analysis of some of the predicted substrates helped confirm a subset of enzymes belonging to the fatty acyl CoA synthetase (FadD) class as substrates of KATmt. The acetylation of FadDs by KATmt was cAMP-dependent. In each of the four proteins tested, acetylation was found to occur at a single conserved lysine residue. To confirm that FadDs were acetylated by KATmt in vivo, BCG_1055, the orthologue of KATmt in M. bovis BCG, was deleted using the specialised transduction method. FadD13, one of the FadDs acetylated by KATmt in vitro, was immunoprecipitated from wild-type M. bovis and the ΔBCG_1055 strains using
a FadD13-specific polyclonal antibody. Acetylated FadD13 was almost completely absent in ΔBCG_1055 but substantial amounts of acetylated FadD13 were present in the wild-type strain, indicating that FadD13 was indeed an in vivo substrate of KATmt. The functional consequences of acetylation of FadDs were analysed using an in vitro fatty acyl CoA synthetase assay. The activities of FadD2 and FadD13 were inhibited on acetylation with KATmt, while acetylation of FadD5 resulted in the formation of a novel product. Therefore, modification of the highly conserved lysine residue in these enzymes by acetylation led to loss or alteration of their enzymatic activity, suggesting that acetylation may be used as a regulatory mechanism to modulate the activities of some of the FadDs by KATmt in a cAMP-dependent manner. Given the extensive role of FadDs in cell wall biosynthesis and lipid degradation in mycobacteria, it seems possible that post-translational control by KATmt in a cAMP-dependent manner constitutes a novel mechanism utilised by these bacteria to regulate these pathways.
This direct regulation of protein lysine acetylation by cAMP appears to be unique to mycobacteria, as orthologues of KATmt are not found outside this genus. In addition, the biochemical differences between KATmt and its orthologue from M. smegmatis KATms, indicate species specific variation, on a common theme. This study is the first report of protein lysine acetylation in mycobacteria. In addition to the identification of several proteins subject to this post-translational modification, the effect of acetylation on the enzymatic activities of some of them has been elucidated.||en_US