Uptake of Sulfate but Not Phosphate by Mycobacterium tuberculosis Is Slower than That for Mycobacterium smegmatis

Abstract

Knowledge of the metabolic pathways used by Mycobacterium tuberculosis during infection is important for understanding its nutrient requirements and host adaptation. However, uptake, the first step in the utilization of nutrients, is poorly understood for many essential nutrients, such as inorganic anions. Here, we show that M. tuberculosis utilizes nitrate as the sole nitrogen source, albeit at lower efficiency than asparagine, glutamate, and arginine. The growth of the porin triple mutant M. smegmatis ML16 in media with limiting amounts of nitrate and sulfate as sole nitrogen and sulfur sources, respectively, was delayed compared to that of the wild-type strain. The uptake of sulfate was 40-fold slower than that of the wild-type strain, indicating that the efficient uptake of these anions is dependent on porins. The uptake by M. tuberculosis of sulfate and phosphate was approximately 40- and 10-fold slower than that of M. smegmatis, respectively, which is consistent with the slower growth of M. tuberculosis. However, the uptake of these anions by M. tuberculosis is orders of magnitude faster than diffusion through lipid membranes, indicating that unknown outer membrane proteins are required to facilitate this process.

INTRODUCTION

In 2009, Mycobacterium tuberculosis caused 9.4 million cases of tuberculosis, resulting in the deaths of approximately 1.7 million people (53). After inhalation, M. tuberculosis is phagocytosed by alveolar macrophages and resides in nutrient-limited phagosomes. M. tuberculosis can prevent the phagosome acidification and influx of many toxic compounds into the phagosome by blocking fusion with late endosomes and lysosomes. This mechanism is critical for the survival of M. tuberculosis in macrophages and is a key feature of its virulence in the host environment (34). However, it is obvious that the acquisition of essential nutrients also is required for the replication of M. tuberculosis in macrophages. Recently, carbon metabolism has received increased attention due to its importance for the virulence of M. tuberculosis (5, 23), but knowledge about the metabolism of other essential nutrients is scarce (25). On a molecular level, the uptake of nutrients precedes any intracellular metabolism and often is the target of regulatory mechanisms.

In this study, we examined the uptake of inorganic anions, in particular nitrate, sulfate, and phosphate, by M. tuberculosis and Mycobacterium smegmatis. A possible source of nitrate for M. tuberculosis in vivo is the oxidation of nitric oxide, which is generated in large amounts within macrophages and restricts the growth of M. tuberculosis (13). The activity of nitrate reductase drastically increases upon the entry of M. tuberculosis into the dormant state (49, 50), indicating that M. tuberculosis uses nitrate as an alternative terminal electron acceptor under anaerobic conditions. Sohaskey demonstrated that nitrate enhances the survival of M. tuberculosis during a sudden shift from aerobic to anaerobic respiration (39). NarK2 is a putative nitrate/nitrite transporter of M. tuberculosis, which is required to reduce nitrate anaerobically (41). In macrophages, M. tuberculosis is exposed to reactive nitrogen and oxygen intermediates which, among other reactions, oxidize and nitrosylate cysteines. Genes involved in cysteine biosynthesis are upregulated in dormancy models for M. tuberculosis, which is consistent with the need for the replacement of these damaged proteins (37, 48). Cysteine synthesis requires the availability of sulfur, e.g., by the uptake of sulfate by the ABC transporter composed of CysW, CysT, and CysA (52). Genes encoding inner membrane phosphate transport systems such as Pst are essential for the survival of M. tuberculosis in macrophages and mice (33, 36), indicating that phosphate inside phagosomes of macrophages is indeed limited.

In this study, we examined the utilization and uptake of these inorganic anions by M. tuberculosis and compared these processes to those of M. smegmatis. For a long time, the focus has been on inner membrane transporters of M. tuberculosis, because they can be easily recognized by similarities to known transporters of other bacteria. However, it has recently been established that mycobacteria have an outer membrane (12), which constitutes the first and primary permeability barrier which has to be overcome for the transport of any nutrient molecule. To this end, the outer membrane porin MspA and its Msp paralogues are required for phosphate uptake by M. smegmatis (14, 51). It has been proposed that M. tuberculosis also uses outer membrane pore proteins for the uptake of inorganic anions (25). Indeed, porin activity has been demonstrated in both M. tuberculosis (16) and M. bovis BCG (19), but the identity of these proteins is unknown and experimental evidence for their physiological function is lacking. We used Msp porin mutants to show that these outer membrane channel proteins are required for nitrate utilization and sulfate uptake by M. smegmatis. The quantitative analysis of the permeability for inorganic anions indicates that M. tuberculosis uses porins to enable the transport of inorganic anions across its outer membrane.

MATERIALS AND METHODS

Chemicals and enzymes.

Hygromycin B was purchased from Calbiochem. All other chemicals were obtained from Merck, Amersham, Roche, or Sigma at the highest purity available.

Bacterial strains and growth conditions.

M. smegmatis and M. tuberculosis (Table 1) were grown at 37°C in Middlebrook 7H9 medium (Difco), Dubos medium (7.4 mM KH₂PO₄, 17.7 mM Na₂HPO₄, 4.3 mM sodium citrate, 2.4 mM MgSO₄, 15 mM asparagine, 0.2% [wt/vol] Casamino Acids, 0.1% Tween 80, 0.2% glycerol [vol/vol], and 4% Dubos medium albumin [pH 7.2; Becton Dickinson]), or Hartmans-de Bont (HdB) medium [30 μM EDTA, 500 μM MgCl₂, 7 μM CaCl₂, 0.8 μM NaMoO₄, 1.68 μM CoCl₂, 5.49 μM MnCl₂, 6.95 μM ZnSO₄, 20 μM FeSO₄, 0.8 μM CuSO₄, 6 μM K₂HPO₄, 6 μM NaH₂PO₄, 15 mM (NH₄)₂SO₄, pH 6.9] supplemented with 0.2% glycerol and 0.05% Tween 80 (38). To control the concentration of the nitrogen source, (NH₄)₂SO₄ in the HdB medium was replaced by Na₂SO₄. NaNO₃, NaNO₂, asparagine, glutamic acid, or arginine then was added as the sole nitrogen source as indicated. Similarly, to control the concentration of the sulfur source, HdB medium was modified by replacing all sulfate salts with chloride salts. Na₂SO₄ then was added as the sole sulfur source as indicated. Hygromycin was used when required at a concentration of 50 μg/ml.

Table 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant genotype or description	Source or reference
Strains
M. smegmatis SMR5	mc2155 derivative; Smr (rpsL*)	35
M. smegmatis ML16	SMR5 derivative; ΔmspA ΔmspC ΔmspD	44
M. tuberculosis	H37Rv	ATCC 27294
Plasmids
pMS2	PAL5000 origin, ColE1 origin, hyg	15
pMN016	PAL5000 origin, ColE1 origin, hyg psmyc-mspA	15

Growth experiments in liquid media.

M. smegmatis SMR5/pMS2, ML16/pMS2, and ML16/pMN016 were incubated on HdB agar plates at 37°C until the colonies showed a smooth appearance. Five ml HdB medium containing 0.2% glycerol and 0.05% Tween 80 then was inoculated with cells from these plates and incubated for 12 to 20 h at 37°C with shaking (200 rpm). This procedure significantly reduced the occurrence of clumps, which may otherwise differentially affect the growth rates of the strains. Cells were harvested by centrifugation (3,000 × g at 4°C for 10 min), washed three times, and then resuspended in 5 ml of the HdB medium supplemented with 30 or 0.1 mM NaNO₃ as the sole nitrogen source or in 0.01 mM Na₂SO₄ as the sole sulfur source. Precultures was used to inoculate 100 ml HdB medium supplemented with 0.1 mM NaNO₃ as the sole nitrogen source or with 0.01 mM Na₂SO₄ as the sole sulfur source to a final optical density at 600 nm (OD₆₀₀) of 0.01. Growth in these cultures was monitored by measuring the OD₆₀₀ every 12 h.

M. tuberculosis H37Rv was inoculated into 20 ml 7H9 medium containing 10% oleic acid-albumin-dextrose-catalase supplement (7H9-OADC) (Becton Dickinson), 0.2% glycerol, and 0.05% Tween 80 and incubated at 37°C to an OD₆₀₀ of 3.0. Cells were harvested by centrifugation (3,000 × g for 10 min), washed three times, resuspended in 5 ml of sterile Millipore water, and inoculated into 200 ml HdB medium containing 10 mM (NH₄)₂SO₄, 30 or 10 mM NaNO₃, 10 mM NaNO₂, 10 mM asparagine, 10 mM glutamic acid, or 10 mM arginine as the sole nitrogen source to a final OD₆₀₀ of 0.01. The growth of the strains in the cultures was monitored by measuring the OD₆₀₀ in triplicate every 2 days. The exponential growth phase was fitted to the equation n = n₀ × e^kt (1). In this formula, the number of bacteria (n; measured according to the optical density of the culture) is an argument of the function of growth time t. The parameters n₀ and k are the initial number of bacteria and the specific growth rate, respectively. The generation time (t_g) was calculated using the equation t_g = ln(2)/k.

Nitrate- and sulfate-dependent growth of M. smegmatis on plates.

Cultures of 5 ml HdB medium were inoculated with M. smegmatis SMR5/pMS2, ML16/pMS2, and ML16/pMN016 and grown overnight at 37°C. The cultures were centrifuged at 3,000 × g, washed 3 times with sterile Millipore-Q water, and filtrated through 5-μm-pore-size filters to remove clumps of bacterial cells and to improve dispersed growth. The optical density of the filtrate was measured at 600 nm. The filtrates were initially diluted to an OD₆₀₀ of 0.05 using sterile Millipore-Q water. Dilutions (10⁻⁴-, 10⁻⁵-, and 10⁻⁶-fold) then were made. For each dilution, 100-μl aliquots were plated on HdB agar plates containing different concentrations of NaNO₃ or Na₂SO₄ (ranging from 0 to 100 mM) as the sole nitrogen or sulfur source, respectively. The plates were wrapped with parafilm and incubated at 37°C. Pictures of single colonies were taken at 16-fold magnifications for 4 or 5 days using a Stemi 2000-C stereomicroscope (Zeiss) equipped with a digital camera (Zeiss AxioCam MRc). For drop assays, 5 μl cells at OD₆₀₀ dilutions ranging from 10⁻² to 10⁻⁶ were dropped on appropriate agar plates.

Measurement of sulfate and phosphate uptake by M. smegmatis and M. tuberculosis.

M. smegmatis and M. tuberculosis H37Rv cells were grown in 7H9 and Dubos media, respectively. Cells were harvested at an OD₆₀₀ of 0.6 by centrifugation (3,000 × g at 4°C for 10 min), washed twice in uptake buffer [2 mM piperazine-N,N′-bis(2-ethanesulfonic acid), 0.05 mM MgCl₂, 0.05% Tween 80, pH 6.8], and resuspended in the same buffer. Radioactively labeled Na₂³⁵SO₄ and KH₂³²PO₄ were mixed with unlabeled Na₂SO₄ or KH₂PO₄ and added to cell suspensions of M. smegmatis (kept on ice) and M. tuberculosis (kept at room temperature) to obtain final sulfate or phosphate concentrations of 1, 2.5, 5, 10, and 20 μM, respectively. The cells were incubated with their substrates at 37°C for 5 min. Samples of 150 μl were taken at the indicated times and mixed with 300 μl killing buffer (0.1 M LiCl in 10% formalin). Cells were filtered by centrifugation through a 0.45-μm-pore-size Spin-X tube (Costar) at 12,000 × g for 1 min. The Spin-X tube filters were washed once with 600 μl killing buffer by centrifugation, and their radioactivity was measured in a liquid scintillation counter (Beckman). All experiments were performed in triplicate. The amount of accumulated solutes is expressed in nmol per mg cells (dry weight). Uptake rates were determined by fitting data obtained for the first two or three time points (up to 2 or 5 min). The Michaelis-Menten constant (K_m), maximal uptake velocity (V_max) for the overall transport, and a minimal estimate of the permeability coefficient were determined as described previously (44).

As a control, the uptake of inorganic ions was inhibited by the following procedure. M. smegmatis SMR5 and M. tuberculosis H37Rv cells were harvested at an OD₆₀₀ of 0.6 by centrifugation, washed twice, and resuspended in uptake buffer as described above. The cells were incubated at 37°C as described previously (52), followed by incubation for 1 h with 0.01 mM carbonyl cyanide 3-chlorophenylhydrazone (CCCP) (1) or 0.15 mM N,N′-dicyclohexylcarbodiimide (DCC) (52). Samples were taken at 0.5, 1, 2, 4, 8, and 16 min after the addition of the inhibitors as described above to determine the uptake of sulfate and phosphate at a solute concentration of 20 μM.

Growth of M. smegmatis on plates in the presence of inhibitors.

M. smegmatis SMR5 (wild type [wt]) and M. tuberculosis H37Rv cells were plated on HdB or 7H10-OADC agar plates with or without 0.01 mM CCCP and 0.15 mM DCC. The plates were incubated at 37°C for 3 to 4 days (for M. smegmatis) or 3 weeks (for M. tuberculosis). Pictures of single colonies were taken at 16-fold magnification using a Zeiss stereomicroscope Stemi 2000-C. Plates were scanned using a digital scanner.

Computer models of the MspA pore with solutes.

To examine whether the size and charge of glucose, glycerol, serine, NO₃⁻, SO₄²⁻, and PO₄³⁻ fit within the MspA pore, the three-dimensional structure of MspA (1UUN) and the structures of the chemical compounds of glucose (CID5793), glycerol (CID753), serine (CID617), NO₃⁻ (CID943), SO₄²⁻ (CID1117), and PO₄³⁻ (CID1061) were obtained from http://ncbi.nlm.nih.gov. The fit of these solutes into the constriction zone of the MspA pore was evaluated using the UCSF Chimera software.

RESULTS

M. tuberculosis utilizes nitrate as the sole nitrogen source.

Nitrogen often is acquired by bacteria in the form of nitrate. While it was shown a long time ago that M. tuberculosis assimilates nitrate (6, 11), it also was reported that nitrate is not used as the sole nitrogen source by M. tuberculosis, possibly due to inhibitory effects of nitrite and other intermediates (21, 47). Another study showed that the avirulent M. tuberculosis H37Ra strain utilizes nitrate efficiently, in contrast to the virulent strain H37Rv (11). Thus, to clarify the role of nitrate in nitrogen utilization by M. tuberculosis, we first examined whether nitrate can be utilized as a sole nitrogen source by M. tuberculosis H37Rv. To this end, wt M. tuberculosis H37Rv and M. smegmatis SMR5, as a positive control (17), were grown in modified Hartmans-de Bont (mHdB) minimal medium with 30 mM nitrate as the sole nitrogen source. These experiments showed that both M. smegmatis SMR5 (Fig. 1A) and M. tuberculosis H37Rv (Fig. 1B) utilized nitrate as the sole nitrogen source, whereas no growth was observed in medium without a nitrogen source.

Fig 1.

Growth of M. smegmatis and M. tuberculosis with various sole nitrogen sources. The cell density (OD₆₀₀) of M. smegmatis SMR5 (A) and M. tuberculosis H37Rv (B) grown in liquid HdB medium with 30 mM NaNO₃ as the sole nitrogen source (filled circles) or without any nitrogen source (open circles) was determined using a spectrophotometer. (C) Growth of M. tuberculosis H37Rv in HdB liquid medium containing 10 mM one of the following sole nitrogen sources: (NH₄)₂SO₄ (filled circles), NaNO₃ (open circles), NaNO₂ (filled triangles), asparagine (open triangles), glutamic acid (filled squares), or arginine (open squares). The experiments were performed in triplicate.

Asparagine is generally regarded as the preferred nitrogen source for the growth of M. tuberculosis (22). In addition, glutamic acid is frequently used in rich media, such as Middlebrook media (24). Growth experiments in HdB minimal medium showed that asparagine and glutamate are indeed preferably utilized by M. tuberculosis as nitrogen sources (Fig. 1C). M. tuberculosis also grows on other amino acids, such as arginine, as well as on small compounds, such as ammonia (NH₄⁺), as sole nitrogen sources. However, these compounds do not support the growth of M. tuberculosis to the same extent as glutamic acid and asparagine. No growth was observed in the presence of 10 mM sodium nitrite (NaNO₂) (Fig. 1C), presumably due to its toxic effect, as observed earlier (11). The growth rate of M. tuberculosis in HdB minimal medium was the highest for glutamate (generation time, 1.2 days) as the sole nitrogen source and decreased for asparagine (1.3 days), ammonia (1.8 days), arginine (2.1 days), and nitrate (2.8 days). These results indicate that M. tuberculosis preferentially uses glutamate and asparagine as nitrogen sources.

Porins are required for the efficient growth of M. smegmatis on nitrate as the sole nitrogen source.

To examine nitrate uptake mechanisms across the outer membrane, we employed the well-characterized porin triple mutant ML16 of M. smegmatis (44). Agar dilution experiments on minimal medium agar plates containing NaNO₃ as the sole nitrogen source showed a reduction in the number of CFU by three to four orders of magnitude, which was caused by the loss of porins in M. smegmatis ML16 at nitrate concentrations of 0.1 to 100 mM (see Fig. S1 in the supplemental material). One millimolar was the minimal nitrate concentration that did not impair the growth of wt M. smegmatis on agar plates (see Fig. S1). In liquid minimal medium, 1 mM NaNO₃ did not support the maximal growth rate of the porin mutant ML16 (Fig. 2A). Further, colonies of the porin triple mutant ML16 were significantly smaller than those of wild-type SMR5 for all nitrate concentrations (see Fig. S2 in the supplemental material). These growth defects of the porin mutant under nitrate-limiting conditions were completely reversed by the expression of the porin gene mspA (Fig. 2A; also see Fig. S1 and S2), underlining the importance of porins for nitrate utilization by M. smegmatis.

Fig 2.

Growth of M. smegmatis in nitrate- and sulfate-limited medium depends on Msp porins. (A) Growth in HdB medium supplemented with nitrate as the sole nitrogen source. M. smegmatis SMR5/pMS2 (filled circles), ML16/pMS2 (filled triangles), and ML16/pMN016 (open circles) were grown in HdB medium containing 1 mM NaNO₃ as the sole nitrogen source. (B) Growth in HdB medium supplemented with sulfate as the sole sulfur source. M. smegmatis SMR5/pMS2 (filled circles), ML16/pMS2 (filled triangles), and ML16/pMN016 (open circles) were grown in HdB medium containing 0.01 mM Na₂SO₄ as the sole sulfur source. The experiments were performed in triplicate. Data are shown with their standard deviations.

Background growth on unknown nitrogen-containing contaminants was observed for all three strains. However, the growth of the porin mutant ML16 was more affected by the lack of nitrogen sources than that of wild-type M. smegmatis and of the complemented mutant (see Fig. S2 in the supplemental material). Furthermore, the optimal growth rate was achieved by ML16 between 10 and 100 mM nitrate, while this concentration was approximately 10-fold lower for wild-type M. smegmatis and the complemented porin mutant (see Fig. S2). The nitrogen concentration provided as ammonium sulfate in standard 7H10 Middlebrook medium is 7.5 mM, which is in the same range as the nitrate concentration required for optimal growth, indicating that both nitrogen sources can be utilized efficiently by wild-type M. smegmatis. These experiments indicate that porins are required for the efficient utilization of nitrate by M. smegmatis.

Porins are required for efficient growth of M. smegmatis with sulfate as the sole sulfur source.

Sulfate is considered the primary source of sulfur for most bacteria and can be taken up by M. tuberculosis by using the inner membrane transporter composed of CysW, CysT, and CysA (52). However, as for nitrate, it is unknown how sulfate is taken up across mycobacterial outer membranes. We hypothesized that mycobacteria take up sulfates through porins in a manner similar to that of phosphates (51). Dilution experiments on HdB minimal medium agar plates containing Na₂SO₄ as the sole sulfur source showed a reduction in the number of CFU by two to three orders of magnitude caused by the loss of porins in M. smegmatis ML16 at sulfate concentrations of 0.01 to 100 mM (see Fig. S3 in the supplemental material). Interestingly, wt M. smegmatis grew on sulfur contaminations in the medium containing no added Na₂SO₄, while the porin mutant ML16 did not (see Fig. S3 and S4 in the supplemental material), indicating that the utilization of these sulfur contaminants is highly porin dependent. The optimal growth rate was achieved by ML16 at 10 mM sulfate, while this concentration was at least 10-fold lower for wild-type M. smegmatis and the mspA-complemented porin mutant. A severe growth defect of the porin mutant ML16 also was observed in liquid minimal HdB medium containing 0.01 mM Na₂SO₄ (Fig. 2B), which was the minimal sulfate concentration that did not impair the growth of wt M. smegmatis on agar plates (see Fig. S3). Further, colonies of the porin triple mutant ML16 were significantly smaller than those of wt SMR5 at all sulfate concentrations (see Fig. S4). Very high sulfate concentrations reduced the growth of M. smegmatis independently of the presence of porins (see Fig. S3 and S4), defining an upper threshold of sulfate concentrations for efficient utilization. The growth defects of the porin mutant under sulfate-limiting conditions were completely reversed by the expression of the porin gene mspA (Fig. 2B; also see Fig. S3 and S4), underlining the importance of porins for sulfate utilization by M. smegmatis.

It should be noted that the porin mutant M. smegmatis ML16 initially grew very slowly in medium with sulfate as the sole sulfur source, but the growth rate increased significantly after ∼60 h (Fig. 2B). This likely is not due to a chromosomal mutation that compensates for the loss of porins, because the same growth rate was observed in three independent cultures. Our never observing any compensatory mutation in other growth experiments with porin mutants of M. smegmatis further supports this argument (43, 44, 51). However, the exponential growth of the porin mutant ML16 was greatly delayed compared to that of wt M. smegmatis. It is possible that the low-level expression of the porin gene mspB in ML16 (44) is sufficient to support and sustain the early onset of growth on nitrate but not on sulfate (Fig. 2). The fact that no growth of ML16 was observed on agar plates without added sulfate, in contrast to wt M. smegmatis, indicates that Msp porins also play a role in the utilization of unknown sulfur-containing contaminants (see Fig. S4 in the supplemental material). Taken together, these experiments indicate that porins are required for the efficient utilization of sulfate by M. smegmatis.

Msp porins are required for uptake of sulfate by M. smegmatis.

The limited growth of the porin mutant ML16 on NaNO₃ and Na₂SO₄ as sole nitrogen and sulfur sources (Fig. 2; also see S1, S2, S3, and S4 in the supplemental material), respectively, indicated that Msp porins are used by M. smegmatis to enable the diffusion of nitrates and sulfates across the outer membrane. To provide direct evidence for this assumption, we wanted to directly examine the role of porins in the uptake of these solutes by employing radiolabeled substrates. However, the longest-lived radioactive nitrogen isotope, ¹³N, has a half-life of less than 10 min and is not suitable for these experiments, whereas the ³⁵S sulfur isotope has a half-life of 87 days and can be effectively used to measure sulfate uptake kinetics. To this end, wt M. smegmatis and the porin mutant ML16 were grown in Middlebrook 7H9 medium containing 4 mM Na₂SO₄. Uptake rates were measured at Na₂SO₄ concentrations ranging from 1 to 20 μM. Figure 3A shows the kinetics of sulfate uptake at 20 μM sodium sulfate as an example. The apparent uptake rates of 1.22 and 0.02 nmol · min⁻¹ · mg⁻¹ cells for M. smegmatis SMR5 (wt) and ML16, respectively, were obtained by fitting the first three time points (Fig. 3A). The uptake rates at different sulfate concentrations followed Michaelis-Menten kinetics (Fig. 3B) and yielded V_max values of 1.3 and 0.13 nmol · min⁻¹ · mg⁻¹ cells and K_m values of 2.1 and 36.4 μM for wt M. smegmatis and the mutant ML16, respectively. Thus, apparent permeability coefficients were 7.7 × 10⁻⁶ and 1.9 × 10⁻⁷ cm/s for wt M. smegmatis and the triple porin mutant ML16, respectively (see also Table S1 in the supplemental material). The 40-fold reduced permeability of the porin mutant ML16 strongly indicates that inorganic sulfate diffuses through Msp porins in M. smegmatis.

Fig 3.

Uptake of sulfate by M. smegmatis and M. tuberculosis. Accumulation of ³⁵S-labeled sulfate by M. smegmatis SMR5 (wild-type), ML16 (ΔmspA ΔmspC ΔmspD) (A and B), and wild-type M. tuberculosis H37Rv (D and E) were measured at 37°C at a final concentration of 20 μM. The uptake rates were determined by the regression analysis of the first 2 min and calculated to be 1.22, 0.025, and 0.25 nmol · min⁻¹ · mg⁻¹ cell dry weight for wild-type M. smegmatis, the mutant ML16 (A), and M. tuberculosis (D), respectively. The dotted lines represent regression lines. A series of sulfate uptake measurements of M. smegmatis SMR5, the mutant ML16 (B), and M. tuberculosis (E) were performed at sulfate concentrations ranging from 1 to 20 μM. The dotted line represents the approximation of the data using the Michaelis-Menten equation. Data analysis yielded V_max values of 1.3 and 0.13 nmol · min⁻¹ · mg⁻¹ and K_m values of 2.1 and 36.4 μM for wild-type M. smegmatis and the porin mutant ML16, respectively. The uptake of sulfate at a final concentration of 20 μM by M. smegmatis (C) and M. tuberculosis (F) at 37°C in the presence of the inhibitor CCCP (10 μM) or DCC (150 μM) also is shown.

To examine whether the increased detection of cell-associated radiolabeled sulfate over time was the result of uptake rather than adsorption to the cell surface, uptake experiments were performed in the presence of 150 μM N,N′-dicyclohexylcarbodiimide (DCC), an inhibitor of the proton-translocating F₁F₀-ATPase, or 10 μM carbonyl cyanide m-chlorophenyl hydrazone (CCCP), a proton-gradient uncoupler. It should be noted that DCC and CCCP do not directly inhibit diffusion through outer membrane porins but inhibit uptake across the inner membrane by transporters that utilize ATP or a proton gradient, respectively, as energy sources. In the absence of uptake across the inner membrane, solutes accumulate rapidly in the periplasm, which has a small volume compared to that of the cytoplasm. Rapid solute equilibration across the outer membrane halts net porin-mediated diffusion. These principles were established by Nikaido and coworkers for E. coli (28, 29). The fact that porin mutants take up many solutes more slowly shows that porin-mediated diffusion across the outer membrane is also the rate-limiting step in M. smegmatis under the conditions of these experiments (43, 44, 51). Inhibitor concentrations of 150 μM DCC and 10 μM CCCP were selected because they did not cause nonspecific growth defects in M. smegmatis on agar plates (see Fig. S5A and B in the supplemental material). The addition of DCC completely eliminated the uptake of sulfate, in contrast to only mild effects of CCCP (Fig. 3C), indicating that sulfate was indeed taken up rather than absorbed on the cell surface and is predominantly transported across the inner membrane by an ATP transporter in mycobacteria. This interpretation is consistent with the DCC sensitivity of the M. tuberculosis sulfate transporter composed of CysA, CysW, and CysT (52). Genome analysis revealed that a similar transporter exists in M. smegmatis.

Uptake of sulfate and phosphate by M. tuberculosis.

A quantitative transport analysis is necessary to understand the acquisition of nutrient molecules by M. tuberculosis. Therefore, we sought to determine sulfate and phosphate uptake kinetics for M. tuberculosis. To this end, transport experiments with sulfate and phosphate at concentrations ranging from 1 to 20 μM were performed with M. tuberculosis H37Rv. The uptake of sulfate was much slower by M. tuberculosis than M. smegmatis, by factors ranging from 5- to 30-fold (Fig. 3D and E; also see Table S1 in the supplemental material). For example, in M. tuberculosis the uptake rates were 0.01 and 0.25 nmol · min⁻¹ · mg⁻¹ cells at sulfate concentrations of 1 and 20 μM, respectively. While uptake rates at higher sulfate concentrations approached an asymptotic value in M. smegmatis (Fig. 3B), they were still in a linear range for M. tuberculosis (Fig. 3E). This made it impossible to determine Michaelis-Menten parameters for sulfate uptake in M. tuberculosis. Similarly to M. smegmatis, 150 μM DCC and 10 μM CCCP did not inhibit the growth of M. tuberculosis on agar plates (see Fig. S5C in the supplemental material). However, DCC completely inhibited the uptake of sulfate by M. tuberculosis, whereas CCCP yielded partial inhibition (Fig. 3F). Taken together, these experiments indicated that sulfate was actively taken up by M. tuberculosis and was not adsorbed on the cell surface.

Similar experiments with ³²P-labeled phosphate showed that phosphate was taken up by M. tuberculosis with rates of 0.01 and 0.33 nmol · min⁻¹ · mg⁻¹ cells at concentrations of 1 and 20 μM, respectively (Fig. 4A). Phosphate uptake rates increased linearly with the solute concentration for M. tuberculosis, which is similar to the observations for sulfate (Fig. 4B). As was the case for sulfate, the addition of inner membrane transport inhibitors inhibited phosphate uptake by M. tuberculosis (Fig. 4C), indicating that M. tuberculosis actively takes up phosphate. Unlike sulfate, however, CCCP completely inhibited uptake, whereas DCC had only mild effects, indicating that phosphate uptake is driven by a symporter which couples phosphate and proton transport across the inner membrane rather than an ABC transporter, as in the case of the sulfate uptake system composed of CysA, CysW, and CysT (52).

Fig 4.

Uptake of phosphate by M. tuberculosis. (A) Accumulation of ³²P-labeled phosphate by M. tuberculosis H37Rv was measured at 37°C at a final concentration of 20 μM. The uptake rate was determined by regression analysis to be 0.33 nmol · min⁻¹ · mg⁻¹ cell dry weight. The dotted line represents the regression line. (B) Michaelis-Menten analysis. A series of phosphate uptake measurements of M. tuberculosis were performed with phosphate concentrations ranging from 1 to 20 μM. The dotted line represents the approximation of the data using the Michaelis-Menten equation. (C) Uptake of phosphate at a final concentration of 20 μM by M. tuberculosis at 37°C in the presence of the inhibitor CCCP (10 μM) or DCC (150 μM).

DISCUSSION

M. tuberculosis can utilize nitrate but prefers amino acids as nitrogen sources.

In this study, we established that M. tuberculosis can utilize nitrate as the sole nitrogen source, albeit at lower efficiency than asparagine, glutamate, and arginine. The beneficial effects of asparagine and glutamate for the growth of M. tuberculosis were reported long ago, and they have been incorporated as nitrogen sources in standard media for mycobacteria (24). The capacity of M. tuberculosis to release ammonia from asparagine and glutamate is crucial for rapid adaptation to acidic conditions in vitro (42) and may play a role in the survival of M. tuberculosis in vivo. These findings also indicate that M. tuberculosis has uptake systems for these amino acids which enable their utilization as nitrogen sources. The uptake of these amino acids by M. tuberculosis must involve both inner and outer membrane proteins (26); however, the identities of these proteins are unknown.

Nitrate and sulfate uptake depends on porins in M. smegmatis.

The growth delay of the porin triple mutant M. smegmatis ML16 in media with limiting amounts of nitrate and sulfate as sole nitrogen and sulfur sources indicates that the efficient uptake of these solutes is dependent on porins. This was confirmed by sulfate uptake experiments (Fig. 3). In previous experiments, a porin mutant of M. smegmatis showed much slower growth on low phosphate concentrations, and the uptake of phosphate was strongly reduced in this strain, indicating that phosphate uptake across the outer membrane of M. smegmatis also is dependent on porins (51). This was a surprising finding because the major porin MspA and the other very similar Msp porins have a highly negatively charged constriction zone (7). The observation that the uptake of sulfate by M. smegmatis at each substrate concentration is faster than that of phosphate (Fig. 5) also is consistent with the diffusion of these solutes through Msp pores due to the lower negative charge of sulfate compared to that of phosphate, resulting in less electrostatic repulsion in the negatively charged constriction zone of the Msp pores. A space-filling model of the MspA channel based on its crystal structure shows that nitrate, sulfate, and phosphate fit into the constriction zone of the pore (Fig. 6). Taken together, these results show that small, inorganic anions use MspA and other Msp porins to enter cells of M. smegmatis. These results also indicate that M. smegmatis does not have an anion-specific porin, in contrast to many Gram-negative bacteria (2, 10, 18).

Fig 5.

Comparison of sulfate and phosphate uptake by M. smegmatis and M. tuberculosis. The sulfate (A) and phosphate (B) uptake measurements of M. smegmatis and M. tuberculosis were performed with isotope concentrations ranging from 1 to 20 μM. The dotted line represents the approximation of the data using the Michaelis-Menten equation. The phosphate uptake data for M. smegmatis were taken from Wolschendorf et al. (51).

Fig 6.

Structural models of the MspA pore with small, hydrophilic solutes. Shown is the permeation of chemical compounds through the MspA pore viewed from the top (A) and bottom (B). Glucose is shown inside the MspA constriction zone because it represents the largest molecule. These models were generated using the UCSF Chimera software. Negatively and positively charged amino acids are shown in red and blue, respectively. Other amino acids are shown in gray.

How are nitrate and sulfate taken up by M. tuberculosis?

This study and other studies show that nitrate and sulfate are taken up by M. tuberculosis. While the transporters NarK2 (40) and that composed of CysT, CysW, and CysA (52) have been identified as inner membrane uptake systems for nitrate and sulfate, respectively, it is unclear how these solutes cross the outer membrane of M. tuberculosis. It is obvious that the slow growth of M. tuberculosis with a generation time of about 24 h requires the uptake of less nutrients per time unit than faster growing mycobacteria, such as M. smegmatis, with a generation time of 3 to 4 h. In this study, we found that uptake rates for sulfate by wt M. tuberculosis are much slower than those for M. smegmatis (Fig. 5A) and are rather similar to those of the porin triple mutant M. smegmatis ML16 (compare Fig. 3B and E), which had an overall permeability for sulfate of 1.9 × 10⁻⁷ cm/s (see Table S1 in the supplemental material). This indicates that the number of porins and their permeability for sulfate in M. tuberculosis rather resembles that of the M. smegmatis porin mutant. This is consistent with early findings that the amount of porin proteins which can be extracted from M. tuberculosis cells is rather low (16). However, despite this slow sulfate uptake compared to that of M. smegmatis and other bacteria, it is important to note that the permeability of M. tuberculosis for sulfate is still orders of magnitude faster than that of other inorganic anions for model lipid membranes. For example, the permeability coefficient of the smaller and less charged chloride anion for model lipid membranes is 1.5 × 10⁻¹¹ cm/s (46). In addition, the mycobacterial outer membrane has an extremely low fluidity due to the extraordinary length of the mycolic acids (20), their covalent attachment to the arabinogalactan-peptidoglycan network (3), and the unusual outer membrane architecture (12, 26). Since the fluidity of the mycobacterial outer membrane is much lower than that of a model lipid membrane made of lipids with C₁₆ fatty acids, the direct diffusion of sulfate across the outer membrane of M. tuberculosis should be much slower and therefore cannot account for the observed sulfate uptake. A similar argument can be made for nitrate, which has a permeability coefficient of 2 × 10⁻⁹ cm/s for model lipid membranes (32), indicating that the uptake of both nitrate and sulfate across the outer membrane requires porin-like channel proteins in M. tuberculosis.

Uptake rates for phosphate in M. tuberculosis and M. smegmatis are similar.

This study showed that the uptake of phosphate by M. tuberculosis is almost as fast as that by M. smegmatis (Fig. 5), which has a permeability coefficient of 2 × 10⁻⁶ cm/s (51). This is a surprising finding, because mspA is the most highly expressed gene in M. smegmatis (30) and provides an efficient diffusion pathway for many hydrophilic solutes (14, 27, 43, 44), while M. tuberculosis does not have MspA homologs (43). However, we have shown earlier that phosphate diffusion through the MspA and MspC pores is rather inefficient, in contrast to that of uncharged or zwitterionic solutes such as glucose or serine. For example, the apparent permeability coefficient of phosphate for M. smegmatis is more than 8-fold lower than that of glucose (51). Further, the loss of the porins MspA and MspC resulted in only a 2-fold decrease of the apparent permeability coefficient of phosphate for M. smegmatis, in stark contrast to the 75-fold reduced permeation of glucose (44, 51). While these results show that MspA and MspC provide the main pores for the diffusion of phosphate across the outer membrane of M. smegmatis, the diffusion of phosphate through the Msp pores is much less efficient than that of glucose. This finding has been rationalized by the highly negatively charged constriction zone of MspA, which is constituted by aspartates 90 and 91 and thought to impede the diffusion of anions more than uncharged solutes. In contrast, the uptake of glucose at a concentration of 20 μM by M. tuberculosis is almost 100-fold slower than that by M. smegmatis (42) and is so slow that a permeability coefficient cannot be reliably calculated. In fact, phosphate is taken up by M. tuberculosis almost 10 times faster than glucose (0.3 versus 0.04 nmol · mg⁻¹ · min⁻¹) (Fig. 5). Since M. tuberculosis contains several copies of the high-affinity Pst phosphate uptake system, the transport of phosphate across the inner membrane might be more efficient than that of glucose, for which no transporter is known (25). An alternative explanation is that diffusion across the outer membrane is the rate-limiting step. The extremely low permeability of phosphate across model lipid membranes of 10⁻¹² to 10⁻¹³ cm/s (4) is six orders of magnitude higher than the permeability coefficient for M. tuberculosis cells as deduced from comparisons to M. smegmatis (Fig. 5; also see Table S1 in the supplemental material), also indicating that M. tuberculosis utilizes outer membrane porins for the uptake of phosphate. Indeed, porin activity was observed in detergent extracts of M. tuberculosis (16) and M. bovis BCG (19), which could account for a preferential uptake of phosphate by M. tuberculosis. However, these porins have not been identified yet.

Conclusions.

In this study, we determined the permeability of M. tuberculosis for sulfate and phosphate. The uptake of these solutes by M. tuberculosis is several orders of magnitude faster than that of model lipid membranes, strongly indicating that it is mediated by proteins. While sulfate and phosphate transporters in the inner membrane of M. tuberculosis are known (8, 9), their counterparts in the outer membrane have yet to be identified. It was shown that phosphate uptake is required for the virulence of M. tuberculosis (31). The finding that exogenous nitrate protects hypoxic mycobacteria from acid stress also suggests an important role of nitrate uptake in vivo (45). Taken together, these findings suggest that the proteins which enable the diffusion of small, inorganic anions across the outer membrane are required for the survival of M. tuberculosis in vivo.