In Mycobacterium tuberculosis, the genes hsaD (2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid hydrolase) and nat (arylamine N-acetyltransferase) are essential for survival inside of host macrophages. These genes act as an operon and have been suggested to be involved in cholesterol metabolism. However, the role of NAT in this catabolic pathway has not been determined. In an effort to better understand the function of these proteins, we have expressed, purified and characterized TBNAT (NAT from M. tuberculosis) and HsaD (2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid hydrolase) from M. tuberculosis. Both proteins demonstrated remarkable heat stability with TBNAT and HsaD retaining >95% of their activity after incubation at 60 °C for 30 min. The first and second domains of TBNAT were demonstrated to be very important to the heat stability of the protein, as the transfer of these domains caused a dramatic reduction in the heat stability. The specific activity of TBNAT was tested against a broad range of acyl-CoA cofactors using hydralazine as a substrate. TBNAT was found to be able to utilize not just acetyl-CoA, but also n-propionyl-CoA and acetoacetyl-CoA, although at a lower rate. As propionyl-CoA is a product of cholesterol catabolism, we propose that NAT could have a role in the utilization of this important cofactor.
- arylamine N-acetyltransferase (NAT)
- heat stability
- 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid hydrolase (HsaD)
- Mycobacterium tuberculosis
Despite over 100 years of research, Mycobacterium tuberculosis, a causative agent of tuberculosis, remains one of the leading causes of death by bacterial infection. While the pathogenesis of the bacterium is extremely complex, the success of this organism has largely been attributed to its ability to reside intracellularly in macrophages, which allows the bacteria to avoid the threat of the host immune response [1–3]. Unable to suppress the bacterial infection, the host immune system encases the infected macrophages into dense granuloma structures that are believed to be metabolically barren environments with low oxygen and poor nutrient availability . Yet, despite these harsh conditions, M. tuberculosis has been demonstrated to survive over 40 years within a granuloma [5,6]. During this long-term survival, the bacterium appears to be metabolically active, suggesting that M. tuberculosis must utilize alternative nutrients . As such, further study of nutrient metabolism in M. tuberculosis would greatly assist the understanding of long-term persistence and bacterial pathogenesis.
NAT (arylamine N-acetyltransferase) is a cytosolic protein that transfers an acetyl group from acetyl-CoA to an arylamine species . NAT enzymes have been identified in numerous eukaryotic and prokaryotic species, including Mycobacterium smegmatis, Mycobacterium bovis BCG (Bacille Calmette–Guérin) and M. tuberculosis [9,10]. Within M. tuberculosis, NAT is found in a five-gene operon that contains the enzymes hsaA (Rv3570c), hsaD (2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid hydrolase) (Rv3569c), hsaC (Rv3568c), hsaB (Rv3567c) and nat (Rv3566) (Figure 1) . Individual genetic knockouts of hsaD, hsaA and nat are unable to survive within macrophages, highlighting the importance of this operon [12,13]. The expression of this operon is controlled by the TetR-like repressor KstR, which is believed to be involved in lipid metabolism . Supporting this role, homologues of the genes hsaA, hsaD, hsaC and hsaB in Rhodococcus RHA1 have been demonstrated to be involved in the meta-cleavage of cholesterol (Figure 2) . The role of NAT in cholesterol degradation remains unresolved, with no direct links between the established enzymatic reactions of NAT and cholesterol metabolism. Although the exact role of intracellular cholesterol in vivo is unknown, recent studies have demonstrated that M. tuberculosis can utilize cholesterol as a carbon source and that cholesterol could be important for bacterial persistence [16,17]. Furthermore, it has been reported in humans and mice with late-stage tuberculosis that large amounts of cholesterol can become concentrated in alveoli and that many bacteria will closely associate with lipid droplets in the necrotic foci of caseating granulomas [18–21].
HsaD (2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid hydrolase) is a cytosolic carbon–carbon hydrolase that can catalyse the cleavage of the cholesterol metabolite 4,9-DSHA [4,5-9,10-diseco-3-hydroxy-5,9,17-trioxoandrosta-1(10),2-dien-4-oic acid] into 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid and 2-hydroxyhexa-2,4-dienoic acid [15,22]. While the enzyme was originally annotated as being involved in biphenyl metabolism, recent work has demonstrated that HsaD has preferential selectivity for 4,9-DSHA as a substrate over the biphenyl metabolite HOPDA (2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid) . Furthermore, in the structure of HsaD, there is a considerably larger active site than that found in homologous enzymes that are involved in biphenyl degradation, which may provide the necessary space to fit the considerably larger cholesterol metabolite 4,9-DSHA .
To better understand the role of the proteins encoded by this essential operon, we have produced recombinant HsaD and NAT from M. tuberculosis and in the present study describe their characterization.
MATERIALS AND METHODS
Unless specifically stated, all chemicals and reagents were purchased from Sigma–Aldrich (Poole, Dorset, U.K.). All enzymes were purchased from Promega (Southampton, U.K.) or New England Biolabs (Hitchin, U.K.).
Chemical synthesis of HOPDA
HOPDA was produced synthetically using a chemical route modified from that reported by Speare et al. . The experimental protocol and chemical products are described in the Supplementary Online Data at http://www.BiochemJ.org/bj/418/bj4180369add.htm. The molar absorption coefficient (ε) for HOPDA was calculated in 100 mM phosphate buffer (pH 7.4) with 20 mM NaCl and was determined to be 13.2 mM−1·cm−1. The calculated molar absorption coefficient (ε) is similar to previous published values .
Heterologous expression and purification of M. tuberculosis NAT, Mycobacterium marinum NAT and chimaeric proteins
NAT enzyme activity measurement
The rate of CoA formation was determined spectrophotometrically using the colorimetric agent DTNB [5,5′-dithiobis-(2-nitrobenzoic acid); Ellman's reagent] as previously described . Briefly, the production of the coloured TNB (5-thio-2-nitrobenzoic acid) was assayed at 405 nm following the reaction of DTNB with the free thiol-CoA that is formed during the enzymatic reaction. The specific activity of TBNAT with various acyl-CoA species was determined by incubating TBNAT (0.75 μg) in the presence of 500 μM hydralazine and different acyl-CoA species (400 μM) in assay buffer (20 mM Tris/HCl, pH 8.0) with 0.5% DMSO at 37 °C. Apparent kinetic parameters were calculated using variable concentrations of acyl-CoA (20–640 μM) with a fixed concentration of hydralazine (500 μM) and NAT enzyme (0.5 μg/ml NAT in 20 mM Tris/HCl, pH 8.0 with 0.5% DMSO). The time was adjusted such that a linear initial rate was measured which corresponded to a change in attenuance (D) of at least 0.2 absorbance unit at 405 nm. The kinetic constants were calculated by hyperbolic fit to the initial rate.
Cloning, heterologous expression and purification of HsaD
The gene encoding HsaD (rv3569c) from M. tuberculosis H37Rv was PCR-amplified from gDNA (For 5′-TTCATATGATGACAGCTACCGAGGAA-3′ and Rev 5′-GCGGATCCTCATCTGCCACCTCC-3′), digested with NdeI/BamHI and cloned into pET28b(+). Following sequencing, the hsaD insert was digested with XbaI/HindIII from pET28b(+)-hsaD, removing the gene with a hexahistidine tag. This fragment was gel-purified and cloned into pVLT31 to yield the plasmid pVLT31-hsaD_H6. After confirmation of the sequence, the plasmid was electroporated into electrocompetent Pseudomonas putida strain KT2442.
Ps. putida containing pVLT31-hsaD_H6 was grown at 30 °C in LB (Luria–Bertani) medium supplemented with 10 μg/ml rifampin and 12.5 μg/ml tetracycline to D600 1.0 and induced with 1 mM IPTG (isopropyl β-D-thiogalactoside) overnight. The culture was harvested (6000 g for 20 min at 4 °C) and the pellet was resuspended in 25 ml of lysis buffer (100 mM phosphate buffer, pH 7.4, 25 mM imidazole, 25 μg/ml DNase I, 200 μg/ml lysozyme and 1 mM pefabloc). The bacteria were lysed by sonication at 4 °C and centrifuged at 16000 g for 20 min. The lysate was filtered through a 0.2 μm filter and run on an Ni-NTA (Ni2+-nitrilotriacetate) column (2 ml of resin/l of bacterial culture). The column was washed sequentially with 2 column volumes (4 ml) of 100 mM phosphate buffer (pH 7.4) containing the following concentrations of imidazole: 25, 50, 100 and 250 mM. The fractions were collected and a sample of each (5 μl) was subjected to SDS/PAGE (12% gel). The fractions that contained HsaD purified to apparent homogeneity were dialysed into 100 mM sodium phosphate buffer (pH 7.4) and 20 mM NaCl and concentrated with an Amicon Ultra concentrator (Millipore) to give a final yield of 10–20 mg of pure HsaD/l of bacterial culture.
The enzymatic activity of HsaD was measured by incubating 45 μl of 1.3 μM of purified HsaD in 100 mM of phosphate buffer (pH 7.4) with 20 mM NaCl with 5 μl of HOPDA (45–460 μM) and monitoring the decrease in D450 of HOPDA over 180 s (Sunrise, Tecan). The Km (app) and Vmax (app) were determined by hyperbolic fit to the initial rate.
Physical properties of the enzymes
Purified HsaD (0.5 mg/ml) was centrifuged at 4000, 5500, 7000 and 10000 rev./min for 10 h at 20 °C and then scanned at 280 nm across the centrifuge rotor (Beckman Optima XL-A). After the completion of the first scan, the sample was centrifuged for an additional 2 h and then scanned at 280 nm across the centrifuge rotor to confirm that the sample had reached equilibrium. The attenuance versus radii was fitted to a first-order exponential curve using the software suite Ultraspin (http://ultraspin.mrc-cpe.cam.ac.uk).
Purified HsaD (5 μM) in 100 mM sodium phosphate buffer (pH 7.4) with 20 mM NaCl was incubated at a range of temperatures (20–85 °C as determined by the internal temperature probe) for 240 s. After incubation, the CD of the sample was measured from 200 to 260 nm on a Jasco J720 spectropolarimeter. The temperature of the sample was measured with an internal temperature probe.
Effect of heat on protein activity
Each of the protein samples [HsaD, TBNAT, MMNAT, MM12TB3 (NAT with domains 1 and 2 from M. marinum and domain 3 from M. tuberculosis) and TB12MM3 (NAT with domains 1 and 2 from M. tuberculosis and domain 3 from M. marinum)] were incubated at temperatures from 20 to 85 °C for 30 min, placed on ice for 5 min, centrifuged at 12000 g for 5 min and then assayed for activity.
Purification and characterization of TBNAT with respect to the acyl donor
TBNAT was expressed as a His-tagged protein and purified with a Ni-NTA resin to apparent homogeneity. Previous studies have demonstrated that TBNAT can acetylate hydralazine at a greater rate than any other substrate tested [27,29]. As such, various CoA derivatives were tested as cofactors of TBNAT using hydralazine as a substrate (Figure 3). The hydrolysis of the thiol ester was readily measurable with n-propionyl-CoA and acetoacetyl-CoA but at a lower rate than with acetyl-CoA. The Km (app) was determined and found to be 0.56±0.03 mM for acetyl-CoA and 0.29±0.02 mM for n-propionyl-CoA. The Km (app) for acetoacetyl-CoA was found to be greater than 0.6 mM and could not be determined accurately due to the low rate of reaction.
HsaD production and purification
Numerous attempts to express HsaD using Escherichia coli and T7 polymerase expression plasmids have been unsuccessful. Therefore an alternative expression system was utilized, in which the gene encoding HsaD (with a hexahistidine tag) was digested from the previously generated pET28b(+)-hsaD construct and cloned into the broad-host expression plasmid pVLT31 . This newly generated construct was then cloned into Ps. putida KT2442 and the expression conditions were optimized. The highest yields of HsaD were obtained when bacteria were grown in LB medium with 12.5 μg/ml tetracycline and 10 μg/ml rifamycin at 30 °C to D600 1.0 and then induced with 1 mM IPTG overnight. The presence of the hexahistidine tag introduced from pET28b(+) allowed for one-step purification using Ni-NTA resin, purifying the enzyme to apparent homogeneity as determined by SDS/PAGE (Figure 4). The molecular mass of the protein was determined by SDS/PAGE to be 33.8 kDa, which agrees with the predicted molecular mass of HsaD with a hexahistidine tag. The identity of HsaD was confirmed by MALDI–TOF-MS (matrix-assisted laser-desorption ionization–time-of-flight MS) (MASCOT score >8300). This expression protocol produced 10–20 mg of purified HsaD/l of bacterial culture.
Characterization of recombinant HsaD
The molecular mass of HsaD in solution was determined by analytical ultracentrifugation equilibrium sedimentation (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/418/bj4180369add.htm). The molecular mass of HsaD was calculated to be 132±9 kDa (n=16), suggesting that HsaD (33.75 kDa) forms a tetramer in solution. Using increasing concentrations of HOPDA the enzyme kinetics of HsaD was studied. From this, the Km (app) of HOPDA was calculated to be 0.23±0.02 mM, the Vmax (app) 3.5±0.2 mM·min−1·mg−1, the Kcat 0.25 s−1 and the Kcat/Km (app) 1129 s−1·M−1. The Km (app) of HOPDA was not significantly different between HsaD with (0.23±0.02 mM) and without a hexahistidine tag (0.24±0.05 mM).
Both TBNAT and HsaD demonstrated very high heat stability (Figures 5A and 6A). With TBNAT, the enzyme begins to lose activity after incubation at temperatures higher than 60 °C and was completely inactivated at 75 °C (Figure 5A). This heat stability is not found in any other bacterial or mammalian NATs previously characterized [31–33]. A lower heat stability was found in NAT from M. marinum, with the enzyme losing all activity after incubation at 50 °C (Figure 5B). The first and second domains of TBNAT appear to strongly influence the heat stability of the protein (Figure 5B). The transfer of domains 1 and 2 from MMNAT to TBNAT (MM12TB3) considerably decreased the enzymatic heat stability (Figure 5B). In contrast with the chimaeric protein in which only the third domain from MMNAT is transferred to TBNAT (TB12MM3), the temperature stability profile is very similar to TBNAT itself (Figure 5A).
The effect of temperature on HsaD stability was assayed by studying both enzymatic activity and CD. When the protein was heated, the CD spectrum associated with the protein's α-helices (as identified by the signal at 220 nm) diminished dramatically across a temperature range of 65–70 °C (Figure 6A). When the protein was incubated at 65±0.4 °C (as determined by the sample probe), the α-helices unfolded in a linear time-dependent manner (Figure 6B). The co-operative unfolding of the protein occurred over the same temperature range when assessed by enzymatic activity as when assessed by loss of CD signal at 220 nm (Figure 6A). Hydrolytic activity was significantly decreased after incubation at 70 °C but was only marginally affected by incubation at 60 °C for 30 min (Figure 6A). Increasing concentrations of NaCl did not alter the heat stability of HsaD (Figure 6C). The heat stability of HsaD was not altered by the presence or absence of a hexahistidine tag (results not shown).
Previous work has demonstrated that the five predicted open reading frames (Rv3566c–Rv3570c), which include the genes encoding HsaD and TBNAT, are expressed as a single mRNA transcript (Figure 1) . As these genes act as an operon, it is likely that they support a co-ordinated function. In order to better understand the role of this operon in M. tuberculosis, we have cloned, expressed and characterized HsaD and TBNAT.
Attempts to express soluble HsaD using T7 polymerase expression systems were unsuccessful . It was only when the expression plasmid pVLT31  was used that soluble HsaD was expressed. The introduction of a hexahistidine tag into the expression plasmid greatly assisted in protein purification. HsaD was demonstrated by analytical ultracentrifugation to form a tetramer in solution, which is consistent with the behaviour of other meta-cleavage hydrolases [35–37] and agrees with the predicted oligomerization state from the crystal structure of HsaD .
The high enzymatic heat stability of TBNAT has not been found in the NAT enzymes from any other species, including the closely related organisms M. smegmatis  and M. marinum (Figure 5B). NAT from M. smegmatis  and M. marinum  have been used as models of the more difficult to express TBNAT. The first and second domains of NAT from M. tuberculosis  appear to contribute strongly to the thermal stability of the protein, with the transfer of domains 1 and 2 from MMNAT to TBNAT causing a dramatic reduction in the heat stability (Figure 5B). In an effort to identify residues that may be important to the heat stability of TBNAT, the primary sequence of the first and second domains of NAT from M. smegmatis and M. marinum was compared with the first and second domains of TBNAT (Figure 7A). A total of 14 residues were identified that were present in NAT both from M. smegmatis and M. marinum but not TBNAT. The locations of these amino acid differences were mapped on to the structure of MMNAT (Figure 7B). Interestingly, 12 of these 14 amino acid residues were found to be surface-exposed on MMNAT (Figures 7C and 7D) . In TBNAT, ten of the changed residues involve a change to a more hydrophobic residue, two involve replacement of an alanine residue by proline (residues 54 and 106) and a further replacement of a glutamine residue with proline is found in TBNAT (residue 148). In the two cases, a negatively charged residue has been replaced by an uncharged residue (Glu55 replaced by glutamine, Glu63 replaced by leucine and Glu172 replaced by threonine) where the residues are the same in both M. marinum and M. smegmatis NAT proteins (Figure 7A). Previous studies of cold-shock proteins from Bacillus species have demonstrated that just a few surface-exposed residues can play a major role in the protein's heat stability .
HsaD displayed a strong correlation between protein denaturation and a decrease in enzymatic activity (Figure 6A). In addition, the temperature stability of HsaD was not significantly affected by increasing ionic strengths (Figure 6C), suggesting that hydrophobic interactions may be important in stabilizing the protein. This is supported with evidence from the crystal structure of HsaD, with the highest concentration of surface-exposed hydrophobic amino acids in a region where the protein forms a tetrameric core. The maximal temperature that HsaD can tolerate before there is any significant loss of activity is 60 °C. This is comparable with several meta-cleavage hydrolases expressed in mesophilic bacteria, including BphD from Rhodococcus RHA1 (65 °C) , ThnD from Sphingomonas macrogoltabidus strain TFA (65 °C)  and CarC from Pseudomonas resinovorans strain CA10 (58 °C)  (Table 1). In contrast, a second group of meta-cleavage hydrolases, including TodF from Ps. putida (45 °C)  and CumD from Pseudomonas fluorescens IPO1 (48 °C) , have a lower maximal temperature that can be tolerated before there is a loss of activity (Table 1). Comparison of the primary sequences of these two groups of meta-cleavage hydrolases does not identify any specific amino acids that could contribute to the heat stability of the protein. This may well be expected as these meta-cleavage hydrolases share a relatively low sequence similarity (23–32% identity with HsaD) (Figure 8A) and yet have a very similar structural fold (root mean square 1.29–1.81 Å (1 Å=0.1 nm) alignment of HsaD to BphD, CarC and CumD respectively) (Figure 8B).
The genes that encode HsaD and TBNAT have been demonstrated to be critical for intracellular survival of M. tuberculosis in host macrophages [12,13]. During the initial bacterial infection, there is a marked increase in internal temperature as a defence mechanism against the pathogen. The enzymatic heat stability of HsaD and TBNAT may prevent the inactivation of these proteins during the initial infection as both proteins are completely active at temperatures greater than 50 °C. While the intracellular temperature, even during the macrophage respiratory burst, is unlikely to reach 45 °C, many NAT enzymes will lose >50% of their activity at this temperature. In contrast, NAT and HsaD from M. tuberculosis will be completely active at this temperature, creating a ‘margin of safety’ for these essential enzymes.
The proposed pathway of cholesterol degradation contains two independent metabolic routes, where one pathway is involved in the removal and oxidation of the C17 acyl chain, while another degrades the four-ring cholesterol skeleton (Figure 2) [15,22,45,46]. Previous studies have demonstrated that products of this operon (HsaA, HsaB, HsaC and HsaD) are likely to have a role in the oxidation and meta-cleavage of the cholesterol skeleton . However, the role of TBNAT, which is encoded in the same operon, does not easily fit into the proposed degradation pathway of the four-ring cholesterol skeleton. Interestingly, the second metabolic pathway involved in the degradation of the C17 acyl chain will yield one molecule of acetyl-CoA and two molecules of n-propionyl-CoA per cholesterol molecule (Figure 2) [15,46,47]. Given that TBNAT can utilize both n-propionyl- and acetyl-CoA donors (Figure 3), it is proposed that TBNAT could have a role in the utilization and regulation of these CoA species. The utilization of n-propionyl-CoA is doubly important to bacterial survival, as n-propionyl-CoA can be toxic to the bacteria at high concentrations and yet acts as a key precursor in lipid biosynthetic pathways . The Km (app) of acetyl-CoA and propionyl-CoA with TBNAT are of the same order of magnitude, although the rate of the enzymatic reaction with propionyl-CoA is considerably less. It remains to be determined whether this mechanism is effective in vivo, but it provides a hypothesis for the functional integration of the operon. Thus, as indicated in Figure 1, TBNAT may act to promote propionyl-CoA and acetyl-CoA usage and homoeostasis for cell wall synthesis, with cholesterol providing the fuel both via the degradation of the acyl side chain and the meta-cleavage of the sterol rings. Given the importance of cholesterol for intracellular survival of M. tuberculosis [16,17] these enzymes are an important focus for further studies.
We are grateful to The Wellcome Trust for financial support. N.A.L. is funded by the Natural Science and Engineering Council of Canada and Linacre College, Oxford, U.K.
We thank Lindsay Eltis (Department of Biochemistry and Molecular Biology, University of British Columbia, Vancouver, Canada) for donating the plasmid pVLT31.
Abbreviations: BCG, Bacille Calmette–Guérin; 4,9-DSHA, 4,5-9,10-diseco-3-hydroxy-5,9,17-trioxoandrosta-1(10),2-dien-4-oic acid; DTNB, 5,5′-dithiobis-(2-nitrobenzoic acid); HOPDA, 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid; HsaD, 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid hydrolase; LB medium, Luria–Bertani medium; NAT, arylamine N-acetyltransferase; MMNAT, NAT from Mycobacterium marinum; MM12TB3, NAT with domains 1 and 2 from M. marinum and domain 3 from Mycobacterium tuberculosis; Ni-NTA, Ni2+-nitrilotriacetate; TB12MM3, NAT with domains 1 and 2 from M. tuberculosis and domain 3 from M. marinum; TBNAT, NAT from M. tuberculosis
- © The Authors Journal compilation © 2009 Biochemical Society