|Home | About | Journals | Submit | Contact Us | Français|
Most mycobacterial species possess a full complement of genes for the biosynthesis of molybdenum cofactor (MoCo). However, a distinguishing feature of members of the Mycobacterium tuberculosis complex is their possession of multiple homologs associated with the first two steps of the MoCo biosynthetic pathway. A mutant of M. tuberculosis lacking the moaA1-moaD1 gene cluster and a derivative in which moaD2 was also deleted were significantly impaired for growth in media containing nitrate as a sole nitrogen source, indicating a reduced availability of MoCo to support the assimilatory function of the MoCo-dependent nitrate reductase, NarGHI. However, the double mutant displayed residual respiratory nitrate reductase activity, suggesting that it retains the capacity to produce MoCo. The M. tuberculosis moaD and moaE homologs were further analyzed by expressing these genes in mutant strains of M. smegmatis that lacked one or both of the sole molybdopterin (MPT) synthase-encoding genes, moaD2 and moaE2, and were unable to grow on nitrate, presumably as a result of the loss of MoCo-dependent nitrate assimilatory activity. Expression of M. tuberculosis moaD2 in the M. smegmatis moaD2 mutant and of M. tuberculosis moaE1 or moaE2 in the M. smegmatis moaE2 mutant restored nitrate assimilation, confirming the functionality of these genes in MPT synthesis. Expression of M. tuberculosis moaX also restored MoCo biosynthesis in M. smegmatis mutants lacking moaD2, moaE2, or both, thus identifying MoaX as a fused MPT synthase. By implicating multiple synthase-encoding homologs in MoCo biosynthesis, these results suggest that important cellular functions may be served by their expansion in M. tuberculosis.
Mycobacterium tuberculosis, the causative agent of tuberculosis, accounts for approximately 1.7 million deaths each year (13). The success of M. tuberculosis as a pathogen is attributable, at least in part, to flexibility in its metabolism, which allows the organism to adapt to the diverse conditions encountered during transmission, infection, and pathogenesis (54). Since the virulence of M. tuberculosis is inextricably linked to its physiology (54), understanding the metabolism and metabolic flexibility of this pathogen under conditions relevant to disease is of paramount importance. A key environmental condition to which M. tuberculosis must adapt in vivo is hypoxia (2, 40, 52). Enzymes that utilize molybdenum cofactor (MoCo) harness the redox properties of molybdenum to catalyze redox reactions in carbon, nitrogen, and sulfur metabolism and to reduce terminal electron acceptors for anaerobic respiration. A search of the M. tuberculosis H37Rv proteome revealed nine proteins with recognizable MoCo-associated domains (see Table S1 in the supplemental material). One of these is NarG, the catalytic subunit of the narGHJI-encoded, membrane-bound respiratory nitrate reductase (NR) (2, 48), suggesting a potentially important role for MoCo in the metabolism of M. tuberculosis in vivo (52).
The M. tuberculosis genome contains multiple genes involved in nitrate metabolism; of these, the narGHJI-encoded NR has both nitrate respiratory (47, 48) and assimilatory (25) functions, while a “fused” nitrate reductase, NarX, which has homology to parts of the NarG, NarH, and NarI proteins, has no apparent function (48). The DosR-regulated nitrate transporter, NarK2, is upregulated in response to hypoxia (32, 53) and nitric oxide (53), which results in a concomitant increase in nitrate reduction under these conditions (48). narGHJI is required for survival of M. tuberculosis under anaerobic conditions in vitro (2) and for persistence of Mycobacterium bovis BCG in the lungs, liver, and kidneys of immunocompetent mice (12, 56). However, a ΔnarG mutant of M. tuberculosis H37Rv showed no phenotype in mice, possibly reflecting the fact that mouse granulomas are not severely hypoxic (2, 52). As a result, the role of nitrate reduction in the pathogenesis of M. tuberculosis remains unclear. Another MoCo-dependent protein investigated previously in M. tuberculosis is NuoG, a subunit of the type I NADH dehydrogenase, which plays a role in inhibiting macrophage apoptosis and in virulence in the mouse model of tuberculosis infection (51). Little is known about the roles of other MoCo-dependent enzymes in M. tuberculosis. Genome-wide essentiality screens predict that most are dispensable for growth in vitro (see Table S1 in the supplemental material) (22, 42). However, a mutant of M. tuberculosis with a transposon mutation in modA, a component of the modABC-encoded ABC-type transport system for molybdenum (9), was found to be attenuated in mice (7), implicating molybdenum uptake in virulence, presumably through the function of one or more MoCo-dependent enzymes.
In the majority of molybdenum-containing enzymes, the metal is coordinated to the dithiolene group of molybdopterin (MPT) to form MoCo, with the exception of nitrogenases, in which molybdenum is coordinated in an iron-molybdenum cofactor (reviewed in reference 14). The biosynthesis of MoCo involves four steps (Fig. (Fig.1):1): the formation of precursor z (cyclic pyranopterin monophosphate) from GTP, the addition of two sulfur atoms to form MPT, the activation of MPT by adenylation, and the insertion of the metal to form MoCo (reviewed in references 26, 44, and 45). The dimethyl sulfoxide (DMSO) reductase family of enzymes requires the covalent addition of GMP (15) or CMP (16) to MoCo to form molybdopterin guanine dinucleotide (MGD) or molybdopterin cytosine dinucleotide (MCD), respectively, which are not interchangeable in Escherichia coli (30).
Multiple homologs of genes implicated in several steps in the MoCo biosynthetic pathway are found in the M. tuberculosis genome (9). In spite of this expansion, certain MoCo biosynthetic pathway genes were predicted to be essential for growth of M. tuberculosis in vitro (22, 42). The notion that MoCo biosynthetic homologs may be differentially required under certain conditions is supported by various lines of evidence: an M. tuberculosis moeB1 mutant is defective in arresting phagosome maturation (24), moaC1 and moaX mutants showed a reduced ability to parasitize macrophages (36), and a moaC1 mutant was attenuated for growth in primate lungs (11). In addition, the gene expression profile of M. tuberculosis in mice specifically identified moaB2 as part of a gene cluster which was highly expressed in vivo but not in vitro (50). Most recently, mutants with independent transposon mutations in moaC1 and moaD1 in the M. tuberculosis W-Beijing strain GC1237 were identified by high-content, phenotypic cell-based screening as being defective in the ability to arrest phagosome maturation, thus further implicating MoCo biosynthesis in biogenesis of the mycobacterial phagosome (4).
In the present study, we investigated the function of MoCo biosynthetic genes from M. tuberculosis, focusing primarily on those involved in the second step of the pathway, which is catalyzed by MPT synthase. We confirm the functions of various moaD and moaE homologs and provide direct evidence that moaX, which shares homology with both moaD and moaE, encodes a novel, fused MPT synthase with MoaD as well as MoaE activity. The demonstrated contribution by multiple moaD and moaE homologs to the biosynthesis of MoCo in M. tuberculosis suggests that important cellular functions are likely served by the expansion of these genes in M. tuberculosis and other members of the M. tuberculosis complex (MTBC).
MoCo biosynthetic genes were identified in mycobacterial genomes using the annotations for M. tuberculosis as a reference (http://tuberculist.epfl.ch). Homologs were identified by BLASTP analysis (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to determine sequence similarity, and ACT software (8) was used to determine genetic context.
The bacterial strains and plasmids used in this study are detailed in Table S2 in the supplemental material. All Escherichia coli strains were grown in Luria-Bertani broth or on Luria agar and incubated at 37°C. Unless otherwise indicated, Mycobacterium smegmatis strains were grown in Middlebrook 7H9 medium (Merck) supplemented with 0.085% NaCl, 0.2% glucose, 0.2% glycerol, and 0.05% Tween 80 with shaking or on Middlebrook 7H10 agar supplemented with 0.085% NaCl, 0.2% glucose, and 0.5% glycerol. M. tuberculosis strains were cultured standing in Middlebrook 7H9 medium supplemented with 0.2% glycerol, Middlebrook oleic acid-albumin-dextrose-catalase (OADC) enrichment, and 0.05% Tween 80. M. smegmatis recombinants ectopically expressing genes under the control of the hsp60 promoter were incubated at 30°C to reduce the toxicity observed at 37°C. When used, nitrate was added in the form of NaNO3 at a final concentration of 10 mM. For nitrate assimilation experiments, M. smegmatis strains were cultured in modified Mycobacterium phlei minimal medium (20) in which asparagine was replaced with NaNO3, whereas M. tuberculosis was grown in MB medium (25, 56) supplemented with NaNO3 (MB-nitrate). To monitor nitrite accumulation in culture, M. tuberculosis strains were grown aerobically in DTA medium (Dubos broth base containing 5% glycerol, 0.5% albumin, 0.75% dextrose, and 0.05% Tween 80) supplemented with NaNO3 (DTA-nitrate), and nitrite was measured using the Griess assay (55). A standard curve was generated using nitrite standards ranging from 10 to 50 nmol/ml. Cultures containing nitrite concentrations above this range were diluted appropriately prior to analysis. To monitor anaerobic NR activity in M. smegmatis strains, 10 ml of M. phlei minimal medium containing NaNO3 was inoculated with M. smegmatis at a starting optical density at 600 nm (OD600) of 0.05. Anaerobic conditions were generated using an AnaeroGen anaerobic jar (Oxoid) and monitored using an indicator strip, as previously described (56). Where appropriate, ampicillin (Ap), kanamycin (Km), and hygromycin (Hyg) were used in E. coli cultures at 200, 50, and 150 μg/ml, respectively, and Hyg, Km, and gentamicin (Gm) were used in mycobacterial cultures at 50, 25, and 10 μg/ml, respectively.
The regions upstream and downstream of the moaA1-moaD1 gene cluster in M. tuberculosis H37Rv were cloned from BAC-Rv48 (5) to generate the suicide plasmid p2ΔmoaABCD (see Table S2 in the supplemental material). Digestion of BAC-Rv134 with XmaI generated a 4,780-bp fragment, containing 2,188 bp of sequence upstream of M. tuberculosis moaD2 and 2,312 bp of sequence downstream of this gene. Using the 4,780-bp fragment as a template, the primers 5′-CGAATTCTTGACGTACTACCCCCTTTCG (forward primer; the EcoRI site is underlined) and 5′-GGTGTTGATCGACCAGTTGC (reverse primer) were used to amplify a 571-bp product, starting 31 bp from the 3′ end of moaD2 and extending 88 bp beyond the NcoI site downstream of moaD2. This PCR product was then used to replace the 793-bp region between the NcoI site and the EcoRI site within moaD2, thus creating a 222-bp in-frame deletion in the 4,780-bp fragment. This fragment and the hyg-lacZ-sacB cassette from pGOAL19 were cloned in p2NIL (31) to create p2ΔmoaD2. The regions flanking the narGHJI operon were cloned from BAC-Rv71 (5) to create pNARKO (see Table S2 in the supplemental material). The primers listed in Table S3 in the supplemental material were used to amplify regions flanking the moaD2 and moaE2 genes from M. smegmatis mc2155, and the resulting amplicons were used to construct the suicide plasmids p2ΔSMmoaD2 and p2ΔSMmoaE2 (see Table S2 in the supplemental material). Mutants of M. tuberculosis and M. smegmatis were constructed by two-step allelic exchange mutagenesis using suicide plasmids carrying the corresponding mutant alleles (31). A complementation vector carrying M. tuberculosis moaD2 was constructed by cloning the 4.8-kb XmaI fragment, which contains the Rv0870c-moaA2-moaD2 operon, in pTT1B (33). A vector carrying the narGHJI operon was cloned from BAC-Rv71 (5). moaD2 and moaE2 from M. smegmatis and moaD1, moaD2, moaE1, moaE2, and moaX from M. tuberculosis were amplified by PCR using the primers described in Table S4 in the supplemental material. PCR products were cloned downstream of the hsp60 promoter in pMhsp60 to produce the corresponding expression plasmids (see Table S2 in the supplemental material). The genotypes of all mutant strains were confirmed by Southern blot analysis (see Fig. S1, S2, and S3 and Table S5 in the supplemental material), and in the case of M. tuberculosis mutants carrying the ΔmoaD2 allele, the genotype was further confirmed by PCR analysis (see Fig. S2 in the supplemental material).
Conversion of MPT to dephospho-form A was performed using a previously described method (17), which was modified for extraction from mycobacteria. Briefly, 100 ml of an M. smegmatis culture was harvested and resuspended in 2 ml of extraction buffer (10 mM sodium ascorbate, 5 mM EDTA, pH 7.4). The cell suspension was transferred to two lysing matrix B tubes (IEPSA) and lysed in a Fastprep 120 cell disruptor (Bio 101 Inc.) for three cycles of 45 s (speed, 4.5) with cooling on ice for 1 min between pulses. The clarified supernatant was acidified and boiled with an iodine solution, excess iodine was removed by the addition of sodium ascorbate, and the pH of the extract was adjusted to 8.5 to 9.0 using 1 M NH4OH. After centrifugation, the extract was concentrated to 90 μl and dephosphorylation was performed overnight using calf intestinal phosphatase (NEB). An aliquot of this extract was separated by high-pressure liquid chromatography (HPLC) on a Dionex reverse-phase C18 column (150 by 4.6 mm; 5-μm particle size) using a Dionex Ultimate 2000 high-pressure liquid chromatography (HPLC) system with fluorescence monitoring (370-nm excitation and 450-nm emission wavelengths). The column was equilibrated with 90% buffer A (5 mM ammonium formate) and 10% buffer B (100% methanol) at a flow rate of 1 ml/min and was maintained under these conditions for 10 min to allow the dephospho-form A to elute. The column was then washed with 100% B for 5 min, returned to the original conditions, and equilibrated for 15 min before injection of the next sample. A dephospho-form A standard was prepared by treating 5 mg of xanthine oxidase (Sigma), which was diluted in a final volume of 500 μl of 50 mM Tris buffer (pH 7.5), as described above. The dephospho-form A peak was identified by comparing the extract before and after dephosphorylation.
RNA was extracted from cultures grown to logarithmic phase (OD600 of ~0.5 to 0.6). Gene expression analysis by reverse transcription-PCR (RT-PCR) or real-time, quantitative RT-PCR (qRT-PCR) was carried out using the primers listed in Table S6 in the supplemental material, according to previously described methods (18, 28).
The biosynthesis of MoCo from GTP occurs via a four-step pathway that involves a number of gene products (Fig. (Fig.1).1). Analysis of the complements and genomic arrangements of MoCo biosynthetic genes in mycobacterial species for which whole-genome sequences are available revealed significant differences both between and within species (Fig. (Fig.2;2; see Table S7 in the supplemental material). Most notable is the multiplicity, in members of the MTBC, of homologs of genes implicated in the first two steps of the pathway, namely, moaA, moaC, moaE, moaD, and moeB. Of the genome sequences analyzed in this study, the moaA1-moaB1-moaC1-moaD1 cluster with the proximal moeB2 gene is present in M. tuberculosis H37Rv and CDC1551 and in M. bovis AF2122 and BCG. These genes have been identified as part of a cluster acquired by horizontal gene transfer (49). Similarly, the horizontally acquired moaA3-moaB3-moaC3-moaX gene cluster (49) is present in M. bovis AF2122 and BCG and in M. tuberculosis CDC1551. Interestingly, in M. tuberculosis H37Rv, moaX and moaC3 are present, while the 5′ end of moaB3 and the entire moaA3 gene are absent owing to an IS6110-mediated deletion in the RvD5 region (6). Studies in India have shown that the presence of moaA3 varies among clinical isolates in that region, further confirming the polymorphic nature of this locus in M. tuberculosis (41, 46). The moaX gene located in this cluster is notable in that it comprises a fusion of moaD and moaE components, which encode the small and large subunits of MPT synthase, respectively (see Fig. S4 in the supplemental material). BLAST analysis identified similar, fused open reading frames in other organisms, but their functions have not been characterized. Further evidence for variability in MoCo biosynthesis gene complements among strains of M. tuberculosis is provided by moaE1, which is present in H37Rv but not CDC1551. The remaining genes are highly conserved across species and include two convergently arranged operons, Rv0863-moaC2-mog-moaE and Rv0870c-moaA2-moaD2, separated by rpfA. This cluster of genes is present in all mycobacterial species analyzed except M. leprae, which has three pseudogenes in this region (moaA2, moaC2, and mog). Of the MoCo biosynthetic gene complement, M. leprae retains intact copies of only moaB2, moeA1, and moeB1. The predicted inability of M. leprae to synthesize this cofactor is consistent with the absence of genes encoding MoCo-dependent enzymes in its genome (reference 10 and data not shown).
Given the multiplicity of homologs associated with the first two steps of the MoCo biosynthetic pathway, the in vitro essentiality of moaA1, moaC1, and moaD1 predicted by transposon site hybridization (TraSH) (42) was surprising. These genes are not part of the core, conserved MoCo biosynthetic gene complement but belong instead to a horizontally acquired cluster (49) found only in members of the MTBC (Fig. (Fig.2;2; see Table S6 in the supplemental material). To further investigate this, we constructed a mutant of M. tuberculosis H37Rv in which a region spanning from within the 5′ region of moaA1 to the 3′ region of moaD1 was deleted and replaced by a Hyg resistance marker (see Table S2 in the supplemental material). In contrast to the essentiality predicted from TraSH (42), the resulting mutant was viable during selection on Middlebrook 7H10 agar and displayed no growth defect when cultured in Middlebrook 7H9, DTA, or DTA-nitrate medium, thereby confirming that the entire moaA1-moaD1 gene cluster was dispensable under these conditions.
We reasoned that the activity provided by the MoCo-dependent NR, NarGHI, could be used as an indirect readout for the production of MoCo, as a reduced availability of its cofactor would affect the activity of this enzyme. The NarGHI enzyme has been shown to support both nitrate respiratory and assimilatory functions in M. tuberculosis (25, 56). The respiratory activity of NarGHI can be assayed directly by monitoring the accumulation of nitrite in cultures of M. tuberculosis grown in DTA-nitrate (56). In this case, NR activity is limited by nitrate diffusion into the cell under aerobic growth conditions and not by the amount of active NarGHI enzyme present (48). This activity is not required for growth of the organism, as the culture medium (DTA) contains other sources of nitrogen for this purpose (asparagine and pancreatic digest of casein). On the other hand, the assimilatory activity of NarGHI can be assayed by monitoring the ability of the organism to grow on MB-nitrate, a nitrogen-limiting medium that contains nitrate as the sole nitrogen source (25). In this case, the NR activity provided by NarGHI is essential for the growth of M. tuberculosis, as it constitutes the first step in the nitrite assimilation pathway. Under these conditions, nirBD is transcriptionally upregulated (25), which is likely to increase the rate of conversion of nitrate to nitrite. Therefore, a defect in MoCo biosynthesis which reduces the amount of active NarGHI present will limit the flux through this pathway and result in reduced growth on nitrate as a sole nitrogen source. As an NR-negative control, an unmarked deletion mutation was introduced in the narGHJI operon of M. tuberculosis H37Rv to produce a ΔnarGHJI null mutant. As expected (25, 48), this strain was unable to grow on nitrate as the sole nitrogen source (Fig. (Fig.3A)3A) or to produce nitrite when cultured aerobically in DTA-nitrate (Table (Table1),1), but both activities were restored by integration of a copy of the M. tuberculosis narGHJI operon at the attB site.
The effect of the ΔmoaA1-moaD1::hyg mutation on MoCo availability was then evaluated by assessing the respiratory and assimilatory NR activities of M. tuberculosis. The ΔmoaA-moaD1::hyg mutant showed an accumulation of nitrite in culture medium comparable to that for the parental wild-type strain (Table (Table1)1) but was significantly impaired for growth on nitrate as the sole nitrogen source (Fig. (Fig.3A).3A). Growth of the ΔmoaA1-moaD1::hyg mutant on MB-nitrate was restored to wild-type levels by integration of a fragment carrying the moaA1-moaD1 gene cluster at the attB site, confirming an association between the nitrate assimilation defect of this strain and loss of function of one or more genes in the moaA1-moaD1 cluster (Fig. (Fig.3A3A).
The observation that NR activity, and therefore MoCo biosynthesis, was not eliminated by deletion of the moaA1-moaD1 cluster implicated other homologs of genes in this cluster in the biosynthesis of MoCo in M. tuberculosis under the conditions tested. To investigate this possibility, we focused on the step catalyzed by MPT synthase, a heterotetrameric enzyme comprised of a central dimer of two MoaE subunits with a MoaD subunit on either side (37, 38). Since one of the moaD homologs (moaD1) had been deleted in the ΔmoaA1-moaD1::hyg mutant, an in-frame deletion in another homolog, moaD2, was introduced into the wild-type and ΔmoaA1-moaD1::hyg mutant strains to assess the effect of individual versus combined moaD homolog loss on MPT biosynthesis. The resulting mutants showed no growth defects when cultured in Middlebrook 7H9 or DTA-nitrate (data not shown). Deletion of moaD2 in the wild-type strain had no effect on nitrite production in aerobic culture (Table (Table1).1). The effect of this mutation on nitrate assimilation was minor as evidenced by the marginal growth impairment of the ΔmoaD2 strain on MB-nitrate compared to the parental wild type (Fig. (Fig.3B).3B). As expected from the profound nitrate assimilation defect of its parental ΔmoaA1-moaD1::hyg strain, the ΔmoaA1-moaD1::hyg ΔmoaD2 double mutant showed negligible growth on MB-nitrate, effectively phenocopying the ΔnarGHJI control in this assay (Fig. (Fig.3A).3A). However, respiratory NR activity in the ΔmoaA1-moaD1::hyg ΔmoaD2 mutant was reduced by approximately 70% compared to that in the ΔmoaA1-moaD1::hyg strain (Table (Table1),1), suggesting that moaD2 contributes to MoCo production under the conditions tested. Therefore, in the double mutant strain, the respiratory NR activity is likely to be limited by the level of holoenzyme rather than by nitrate diffusion into the cell, with the former being compromised by a deficiency in MoCo.
Unlike the ΔnarGHJI strain, which was completely devoid of respiratory and assimilatory NR activities, the ΔmoaA1-moaD1::hyg ΔmoaD2 strain retained an ability to reduce nitrate despite loss of both moaD1 and moaD2 (Table (Table1).1). The residual capacity for MoCo biosynthesis, inferred from the respiratory NR activity of the double mutant strain, suggested that MoaX may provide MoaD-like activity. To test this hypothesis, we designed a strategy to investigate M. tuberculosis moaX function by means of expression in a heterologous host defective in moaD and/or moaE function. The strategy exploited the comparatively simple MoCo biosynthetic gene complement in M. smegmatis mc2155, which includes single MPT synthase-encoding genes, moaD2 and moaE2 (Fig. (Fig.2;2; see Table S7 in the supplemental material).
moaD2 and moaE2 deletion mutants of M. smegmatis mc2155 were constructed and found to grow normally in standard broth culture, confirming the dispensability of these genes under these conditions (data not shown). To determine the impact of the mutations on MoCo biosynthesis in M. smegmatis, NR activity was again used as a readout. As observed by Weber et al. (56), no anaerobic NR activity was detected in M. smegmatis even though this organism possesses an intact narGHJI operon (http://www.tbdb.org). However, integration of M. tuberculosis narGHJI and its upstream region at the attB locus conferred anaerobic respiratory NR activity on M. smegmatis, as evidenced by nitrite production at a level of 491 ± 41 nmol/ml after 5 days of anaerobic culture of the narGHJI integrant (mc2155 attB::pMVnar) versus 3 ± 1 nmol/ml for the empty vector control integrant (mc2155 attB::pMV306H). In contrast, the level of nitrite in anaerobic cultures of the M. smegmatis ΔmoaD2 and ΔmoaE2 mutants carrying M. tuberculosis narGHJI was indistinguishable from the background level (≤3 nmol/ml after 5 days of culture for M. smegmatis ΔmoaD2 attB::pMVNar, ΔmoaD2 attB::pMV306H, ΔmoaE2 attB::pMVNar, and ΔmoaE2 attB::pMV306H), suggesting that the lack of anaerobic respiratory NR activity was attributable to MoCo deficiency in the ΔmoaD2 and ΔmoaE2 mutants. The MPT content of wild-type M. smegmatis, the ΔmoaD2 mutant, and a complemented derivative carrying M. smegmatis moaD2, expressed on an episomal plasmid under the control of the mycobacterial hps60 promoter (pMSmD2), was analyzed by HPLC-based detection of the derivative dephospho-form A (Fig. 4A to C). This MPT derivative was not observed in the ΔmoaD2 mutant (Fig. (Fig.4E)4E) but was detected in the wild-type and complemented strains (Fig. 4D and F).
In addition to NarGHJI, M. smegmatis contains a putative MoCo-dependent assimilatory-type nitrate reductase, NarB, which is predicted to enable growth on nitrate as a sole nitrogen source (19). Consistent with an inability to produce MoCo, the M. smegmatis ΔmoaD2 and ΔmoaE2 mutants were unable to assimilate nitrate (Fig. (Fig.5A).5A). Restoration of growth on nitrate thus provided a means of functionally assessing the various moaD and moaE homologs from M. tuberculosis, including moaX (Fig. (Fig.5;5; see Fig. S5 in the supplemental material). Ectopic expression of M. smegmatis moaD2 and moaE2 under the control of the hsp60 promoter reversed the nitrate assimilation defects of the ΔmoaD2 (Fig. (Fig.5B;5B; see Fig. S5 in the supplemental material) and ΔmoaE2 (Fig. (Fig.5C)5C) strains, respectively, thus validating the complementation strategy. Expression of M. tuberculosis moaD2 in M. smegmatis ΔmoaD2 (Fig. (Fig.5B;5B; see Fig. S5 in the supplemental material) and of M. tuberculosis moaE1 or moaE2 in M. smegmatis ΔmoaE2 (Fig. (Fig.5C)5C) similarly restored the ability of the mutants to grow on nitrate, confirming that these M. tuberculosis genes encode functional MPT synthase subunits. In contrast, M. tuberculosis moaD1 was unable to complement the growth defect of the M. smegmatis ΔmoaD2 mutant (Fig. (Fig.5B;5B; see Fig. S5 in the supplemental material). RT-PCR analysis confirmed expression of M. tuberculosis moaD1 in the M. smegmatis ΔmoaD2 strain, suggesting that the failure to complement was not due to a lack of expression of this gene (see Fig. S6 in the supplemental material). Importantly, expression of moaX fully complemented the nitrate assimilation defect of both M. smegmatis single mutants (Fig. 5B and C) as well as the ΔmoaD2 ΔmoaE2 double mutant (Fig. (Fig.5D),5D), whereas expression of M. tuberculosis moaD2 (Fig. (Fig.5D),5D), moaE1, or moaE2 (Fig. (Fig.5E)5E) had no effect on the growth defect of the double mutant. Together, these results confirm that moaX has both moaD and moaE activities.
To determine whether loss of moaD1 and/or moaD2 affected the expression of the remaining moaD homologs, transcript levels of moaD1, moaD2, and moaX were measured by qRT-PCR in logarithmic-phase cultures of the wild-type and ΔmoaA1-moaD1::hyg, ΔmoaD2 and ΔmoaA1-moaD1::hyg ΔmoaD2 mutant strains grown in DTA-nitrate (Table (Table2).2). All three genes were expressed in wild-type M. tuberculosis. However, no significant differences in transcript level were observed in the mutant strains.
The suggestion that MoCo biosynthesis is important in the physiology and pathogenesis of M. tuberculosis is supported by several lines of evidence (4, 11, 24, 36, 50). In this study, we assessed the contributions of various homologs in the MoCo biosynthetic pathway to synthesis of the cofactor in M. tuberculosis H37Rv by generating targeted gene deletions. The assimilatory and respiratory NR activities of the MoCo-dependent NarGHI were used as complementary tools to investigate the effects of these deletions on MoCo biosynthesis. The assimilatory NR assay was highly sensitive to mutations in MoCo biosynthetic genes, suggesting that it was able to detect relatively small changes in the level of the cofactor (Fig. (Fig.3).3). However, defects in MoCo biosynthesis are expected to affect the function of all MoCo-dependent enzymes, including NarGHI. While little is known about the roles that most of the predicted MoCo-dependent enzymes play in the physiology of M. tuberculosis, the loss/impairment of another MoCo-dependent enzyme function(s) may have contributed to the growth phenotypes observed for some of the MoCo biosynthesis gene mutants when cultured in MB-nitrate. In contrast, the respiratory NR assay, which directly measures nitrite production, was able to detect NR activity not discernible by the assimilatory assay (Table (Table1)1) but could not detect minor changes in cofactor levels, presumably because NR activity is limited by diffusion of nitrate rather than by levels of the enzyme during aerobic growth (48). Our results suggest that the levels of the holoenzyme are likely to become limiting for NR activity in mutants of M. tuberculosis in which MoCo levels are more severely reduced.
We were able to generate an M. tuberculosis strain lacking the moaA1-moaD1 gene cluster, confirming that these genes are dispensable under the conditions employed in this study. The finding that deletion of moaD2 in the ΔmoaA1-moaD1::hyg mutant decreased, but did not eliminate, nitrate reduction confirmed a role for moaD2 in MoCo biosynthesis in M. tuberculosis while simultaneously implicating MoaX in the provision of additional MoaD activity under the conditions tested. The latter conclusion was confirmed by the ability of moaX to complement the nitrate assimilation defect of the ΔmoaD2 mutant of M. smegmatis. In addition, the moaE1, moaE2, and moaX genes from M. tuberculosis were all able to restore growth of the M. smegmatis moaE2 mutant on nitrate, confirming the function of these genes and demonstrating that MoaX also provides an additional source of MoaE activity.
Although their relative contributions to MoCo biosynthesis have yet to be established, the demonstrated activity of multiple moaD homologs in M. tuberculosis raises the question of whether the interactions between the various MoaD and MoaE proteins are specific or promiscuous. Chimeras between subunits of human and E. coli MPT synthases are able to form active enzymes. However, the activities of the various complexes differ, with the bacterial enzyme being most active, followed by the chimera between human MoaD and bacterial MoaE, which is more active than the human enzyme, and the bacterial MoaD-human MoaE chimera (23). Therefore, if the moaD and moaE homologs are coexpressed in M. tuberculosis, various combinations of tetramers are likely to form. The reason for the inability of moaD1 to complement the nitrate assimilation defect of the M. smegmatis moaD2 mutant is not clear, but it may indicate the absence of an interacting partner. One possibility is that moeB2, which is located proximal to moaD1 in the M. tuberculosis genome but is not present in M. smegmatis, may be specifically required for the adenylation of MoaD1.
Our results provided direct evidence that M. tuberculosis moaX encodes a novel, fused MPT synthase with both MoaD and MoaE activities, consistent with conservation in MoaX of key MoaE residues required for catalysis and stabilization of the MoaD-MoaE interface (57) and of the C-terminal glycine in MoaD required for adenylation and transfer of sulfur to precursor z (43) (see Fig. S4 in the supplemental material). However, given the way in which the C terminus of MoaD interacts with both MoaE and MoeB in the E. coli system, the functional implications of the fused nature of MoaX are likely to be profound (21, 37, 38, 57). Notably, inclusion of an additional glycine residue at the C terminus of E. coli MoaD completely inhibited adenylation by MoeB and significantly reduced the rate of MPT synthesis, demonstrating that the position of this terminal residue is critical for MPT synthase function (43). Assuming that this mechanism is conserved, MoaX would have to be cleaved into its constituent MoaD and MoaE components in order to function as a canonical MPT synthase in mycobacteria. This possibility is currently under investigation in our laboratory.
Transcriptional analysis of moaD1, moaD2, and moaX in the wild-type and mutant strains revealed no significant change in the levels of the remaining genes when moaD1, moaD2, or both were deleted, indicating that these genes do not regulate one another's expression. Furthermore, while the nitrate assimilation studies suggested that the cofactor is limiting in the ΔmoaA1-moaD1::hyg and ΔmoaA1-moaD1::hyg ΔmoaD2 strains, the remaining moaD homolog(s) was not transcriptionally upregulated, suggesting that MoCo levels do not regulate transcription of these genes. In E. coli, the moaABCDE operon is positively regulated under anaerobiosis by the global regulator Fnr and under high internal molybdate concentrations by ModE (3). M. tuberculosis contains a member of the Fnr/Crp family of transcriptional regulators, Rv3676, whose loss results in growth defects in liquid culture, macrophages, and mice (35). A putative binding site for this regulator was identified upstream of moaX (1); however, expression of this gene was not affected by loss of Rv3676 (35). Instead, a recent study revealed that Rv3124 acts as a positive transcriptional regulator of the moaA1-moaD1 operon in M. tuberculosis (27). The upregulation of Rv3124 observed in the NRP-1 and NRP-2 stages in the Wayne model (29) suggests a possible mechanism for positively regulating MoCo biosynthesis in M. tuberculosis in response to reduced oxygen availability. However, MoCo biosynthetic genes, including the moa1-moaD1 operon, are not upregulated in the early response of M. tuberculosis to hypoxia (32, 53) or in NRP-1 and NRP-2 (29), and moeB1 is the only gene in the MoCo biosynthetic pathway that shows modest upregulation at a late stage of the enduring hypoxic response (39). Therefore, hypoxia may not be the principal trigger for Rv3124-dependent upregulation of moaA1-moaD1 expression, and the association, if any, between hypoxia and MoCo production remains unclear. These results, together with the lack of a ModE homolog (9) and riboswitch regulator (34) in M. tuberculosis, point to fundamental differences in the regulation of MoCo metabolism in M. tuberculosis compared to other organisms that warrant further investigation.
In summary, the results of this study have confirmed the function of multiple homologs of MPT synthase subunit-encoding genes from M. tuberculosis and implicated them in de novo biosynthesis of MoCo in vitro. These findings suggest that important cellular functions are likely served by the expanded complement of MoCo biosynthesis genes. They also underscore the need to further investigate the role of MoCo-dependent enzymes in the physiology of M. tuberculosis and assess the impact of MoCo biosynthesis defects on the growth and survival of this pathogen.
This work was funded by grants from the Medical Research Council of South Africa (to V.M.), the National Research Foundation (to V.M.), the University of the Witwatersrand, and the Howard Hughes Medical Institute (International Research Scholar's grant to V.M.). M.J.W. was supported by a postdoctoral fellowship from the Percy Fox Foundation.
We thank Chuck Sohaskey and Digby Warner for critically reviewing the manuscript and for many helpful discussions and suggestions; Bhavna Gordhan for technical assistance; Helena Boshoff, Adrie Steyn, and Stewart Cole for providing reagents; and members of the MMRU for advice and assistance.
Published ahead of print on 20 October 2010.
†Supplemental material for this article may be found at http://jb.asm.org/.