The loss of human lives for tuberculosis (TB) remains unhampered as a result of the synergy with the AIDS pandemic and the emergence of drug resistant strains. Each year, there are almost 2 million TB-related deaths worldwide and nearly 9 million people become infected. Moreover, one third of the world's population is estimated to be infected with TB [1]. The available therapies target only a few bacterial functions and the growing numbers of extensively drug-resistant strains urges for the identification of novel pathways and new drug targets. While promising antimycobacterials have been recently synthesized, they are still far from being available for inclusion in TB therapies [2–4].

Maltose-1-phosphate was initially identified in Mycobacterium bovis BCG cell extracts [5]. An enzyme synthesizing maltose-1-phosphate from maltose and ATP was later identified in Actinoplanes missouriensis and named maltokinase (Mak) [6]. Apparently, it was constitutively expressed, regardless of the sugars present in the growth medium, thus unlikely to play an essential role in maltose catabolism [7]. A recombinant Mak from Streptomyces coelicolor was also partially characterized [8]. Maltose-1-phosphate could also be synthesized without the requirement of ATP in E. coli, where it was proposed to be involved in the regulation of maltose metabolism [9]. The maltokinase gene (mak) is present in most of the available mycobacterial genomes and it was considered essential for the growth of M. tuberculosis H37Rv [10]. Moreover, mak is always linked with the trehalose synthase gene (treS) and in some organisms fused into one gene. This highly conserved sequential arrangement is found in many bacterial groups and suggests a shared metabolic route.

Trehalose is a ubiquitous disaccharide found in bacteria, archaea, fungi, plants and in some invertebrates, where it plays a multitude of biological roles [11, 12]. In mycobacteria and related organisms, trehalose is synthesized by three different pathways and serves a functional role as a compatible solute under osmotic or thermal stress, and a structural role in cell wall components [13–15]. Mycobacterium smegmatis mutants defective in the three pathways are unable to grow unless trehalose in supplemented to the growth medium [14]. However, each pathway seems to have a specific role and hierarchy in closely related actinobacterial species. While the OtsA/OtsB pathway seems to be essential for the survival of M. tuberculosis in infected mice, the inactivation of the TreY/TreZ did not cause obvious phenotypic effects under the examined conditions [16]. However, the inactivation of TreS promoted increased survival of the mutant in mice, when compared to the wild-type strain [16]. In Corynebacterium glutamicum, the TreY/TreZ pathway dominates, the OtsA/OtsB plays only a minor accessory role and the TreS pathway seems to be involved in trehalose degradation [17, 18]. Single mutations targeting each of the pathways in M. smegmatis resulted in no apparent phenotypic defects indicating that the three pathways are redundant in this organism [14].

The TreS from M. smegmatis is capable of interconverting trehalose and maltose [19]. TreS was also shown to catalyze the formation of trehalose from glycogen via its amylase activity and with maltose as the intermediate [20]. Moreover, the synthesis of glycogen from trehalose has also recently been reported in M. smegmatis [21]. In this proposed three-step pathway, TreS and Mak sequentially convert trehalose into maltose and further into maltose-1-phosphate while a final maltosyltransferase transfers maltose from maltose-1-phosphate to glycogen. Although M. smegmatis has an essential trehalase to degrade trehalose directly into glucose [22], it has been suggested that it is energetically advantageous to convert trehalose to glycogen rather than to hydrolyze it to glucose [21], which cannot accumulate in cells.

Despite of the increasing numbers of genome sequences showing that mak genes are widely distributed throughout the bacterial world, there is still a tremendous lack of biochemical data on the corresponding enzymes as only the Mak from A. missouriensis and Streptomyces coelicolor have been biochemically characterized [7, 8]. Therefore, we cloned and successfully expressed in E. coli the mak gene from M. bovis BCG, the organism where maltose-1-phosphate was firstly detected. We purified the recombinant bioactive enzyme and determined its biochemical and kinetic properties to further understand its role in mycobacterial physiology.

Maltose-1-phosphate was identified for the first time in Mycobacterium bovis BCG [5] but a maltose-1-phosphate synthesizing enzyme (maltokinase, Mak) was only identified much later in Actinoplanes missouriensis [6]. Although the corresponding gene (mak) can now be identified in many bacterial genomes, including those of most mycobacteria, only the Mak from A. missouriensis and S. coelicolor have, so far, been characterized [7, 8]. Interestingly, the M. tuberculosis H37Rv mak gene (Rv0127) was considered essential for the pathogen's growth [10]. This information and the absence of biochemical data on mycobacterial maltokinases prompted us to study the enzyme from the slow-growing species M. bovis BCG.

Our results show that, in contrast to the A. missouriensis and S. coelicolor enzymes, which use ATP as the sole phosphate donor (GTP and UTP are used at less than 3% by the A. missouriensis Mak and not used by the S. coelicolor enzyme), the M. bovis BCG maltokinase was able to use ATP, GTP and UTP with comparable efficiency. This substrate flexibility may reflect an absolute requirement for maltose-1-phosphate, corroborating the proposed essentiality of the mak gene in M. tuberculosis [10]. However, lower flexibility of the Mak for NTP substrates may occur in vivo. Similar behavior has been detected between native and recombinant forms of enzymes involved in trehalose metabolism [23, 24]. Interestingly, some unprecedented residual activity (< 7%) could also be detected with ADP and GDP as phosphate donors. Like the enzymes from A. missouriensis and S. coelicolor, only maltose served as the acceptor substrate. Although the K _m values for maltose were similar for the A. missouriensis and M. bovis BCG Mak (no kinetic data is available for the S. coelicolor Mak), the latter had a K _m value for ATP (0.74 mM) slightly higher than that measured for the A. missouriensis Mak (0.54 mM).

A major obstacle for the characterization of the M. bovis BCG maltokinase was its very low stability. Our attempts to stabilize the enzyme with glycerol failed as it severely inhibited the activity. Unlike the native Mak from A. missouriensis, which could be stabilized by maltose, this disaccharide only moderately stabilized the mycobacterial Mak while trehalose was ineffective. Only the addition of 50 mM NaCl to the preparation stabilized the enzyme for characterization. The mycobacterial Mak was, like the A. missouriensis and the S. coelicolor enzymes, a monomer in solution. While no data is available for the cation dependence of the S. coelicolor Mak, both the A. missouriensis and the M. bovis BCG enzymes were dependent on Mg²⁺ ions for maximal activity. The cations Co²⁺ or Mn²⁺ could partially replace Mg²⁺ while Zn²⁺ was inhibitory. Divalent cations seem to play an essential role in the activity of kinases, namely in the stabilization of the negatively charged phosphate donor groups [7]. The temperature profile for activity of the recombinant Mak from M. bovis BCG was similar to that of the native Mak from A. missouriensis, with maximum near 60°C. The S. coelicolor recombinant Mak had maximal activity at about 45°C. All three enzymes had optimal activity between pH 7 and 9 [7, 8], which is also the optimum pH range for the activity of other kinases [7]. The Mak from M. bovis BCG had no detectable aminoglycoside phosphotransferase activity, which is in agreement with the data obtained with the S. coelicolor Mak [8]. It is likely that many of the putative aminoglycoside phosphotransferases with significant amino acid identity with maltokinases (>40%) are incorrectly annotated and may instead have maltokinase activity.

The frequent association of the mak gene with the trehalose synthase (treS) gene, either as a bicystronic unit or as a fused gene, strongly suggests a synergistic action and a common biochemical pathway. While such a pathway has been previously proposed [8], very recent research [21] corroborate this hypothesis by demonstrating that M. smegmatis converts excess trehalose into glycogen through the sequential action of TreS, Mak and of a maltosyltransferase (Fig. 6). The latter enzyme was shown to use maltose-1-phosphate as the substrate for the transfer of maltose to glycogen [21]. The genetic organization in A. missouriensis lend additional support to this hypothesis since a glycogen-branching enzyme gene (glgB) is located immediately downstream the contiguous treS and mak genes. In Rubrobacter xylanophilus, a bacterium that constitutively accumulates high levels of trehalose, a glgB gene is also located immediately downstream a fused treS-mak bifunctional gene and a similar function may be anticipated. Interestingly, the M. smegmatis TreS was also found to possess amylase activity and capable to catalyze the synthesis of trehalose from glycogen, with maltose as intermediate [20]. Curiously, Mycobacterium leprae, in which glycogen has not been reported, lacks both treS and mak homologues [25]. However, the reason for the missing glycogen may be explained by the absence of other genes involved in glycogen synthesis like glgB or glgP, which are pseudogenes in M. leprae. On the other hand, both M. gilvum and M. vanbaalenii possess a mak but lack a treS gene. Unfortunately no data on trehalose or glycogen synthesis is available for these organisms. Nevertheless, among the mycobacterial enzymes involved in this metabolic circuit (trehalose → maltose → maltose-1-phosphate → glycogen → maltose → trehalose) (Fig. 6), all enzymes but the maltokinase have been biochemically studied [13, 19, 21].

Mycobacteria possess three different pathways for the synthesis of trehalose [13, 16]. This disaccharide is constitutively present in the cytoplasm of mycobacterial cells, an essential structural component of cell walls and its levels increase during osmotic and thermal stress [13–15]. Although the TreS has much higher affinity for maltose than for trehalose [19], in most cases it seems to be involved in trehalose degradation [16] and has only been implicated in trehalose biosynthesis in M. smegmatis [14]. M. smegmatis possesses an essential trehalase that converts trehalose into glucose [22] (Fig. 6). Elbein and coworkers argue that it is energetically favorable to convert trehalose to glycogen by the recently discovered TreS-Mak-maltosyltransferase pathway (Fig. 6) [21]. However, normal ATP intracellular concentrations within the 1 to 10 mM range [26] are inhibitory of the maltosyltransferase (GMPMT) activity [21] but not of the Mak activity. These authors also propose that the role of this pathway is to convert excess trehalose into glycogen during stress conditions. Therefore, under non-stressing conditions, in which the ATP levels rise, maltose-1-phosphate cannot be the substrate for glycogen, which may be synthesized via ADP-glucose [21]. Since mycobacteria contain other alpha-1,4-glucans and polysaccharides [25], it is tempting to speculate that maltose-1-phosphate may also be used as donor of maltose units for those macromolecules, under specific conditions (Fig. 6). Moreover, the constitutive expression of Mak in A. missouriensis [7] and the suggested essentiality of the M. tuberculosis mak gene (Rv0127) indicate alternative roles for maltose-1-phosphate in mycobacterial physiology. This is also supported by the fact that only mak, but not the treS gene (Rv0126), was considered essential for growth. Moreover, the presence of a putative promoter within the terminal part of the treS gene indicating independent regulation of mak expression, also suggests an alternative role for the Mak and maltose-1-phosphate in mycobacteria. Since trehalose-6-phosphate is an important signaling molecule in yeast and plant metabolism [12, 27], maltose-1-phosphate could, hypothetically, play a similar role in mycobacteria. In fact, the involvement of maltose-1-phosphate in the regulation of sugar metabolism in E. coli has been proposed earlier but not confirmed experimentally [9].

Several questions remain unanswered and the role(s) of maltose-1-phosphate in mycobacterial metabolism only now begins to be unraveled. While further studies are required to answer these questions, the biochemical properties of the maltokinase from M. bovis BCG, the first of this bacterial group to be characterized, provide an important step towards the clarification of its role in mycobacterial metabolism.

Although the maltokinase gene was considered essential for M. tuberculosis growth [10], implying that the corresponding enzyme might be a suitable target for the design/development of anti-mycobacterial drugs, no mycobacterial maltokinase had been characterized to date. The fast and straightforward method for high-level production and purification of bioactive maltokinase from M. bovis BCG, a major bottleneck in mycobacterial research, and the determination of its biochemical properties, some of which are unique among these type of enzymes, contribute to the knowledge of these enzyme's abilities and provide insights about its unexplored role in mycobacterial physiology.