OUP user menu

Cholesterol oxidase is required for virulence of Mycobacterium tuberculosis

Anna Brzostek, Bozena Dziadek, Anna Rumijowska-Galewicz, Jakub Pawelczyk, Jaroslaw Dziadek
DOI: http://dx.doi.org/10.1111/j.1574-6968.2007.00865.x 106-112 First published online: 1 October 2007


Recent reports have indicated that cholesterol plays a crucial role during the uptake of mycobacteria by macrophages. However, the significance of cholesterol modification enzymes encoded by Mycobacterium tuberculosis for bacterial pathogenicity remains unknown. Here, the authors explored whether the well-known cholesterol modification enzyme, cholesterol oxidase (ChoD), is important for virulence of the tubercle bacillus. Homologous recombination was used to replace the choD gene from the M. tuberculosis genome with a nonfunctional copy. The resultant mutant (ΔchoD) was attenuated in peritoneal macrophages. No attenuation in macrophages was observed when the same strain was complemented with an intact choD gene controlled by a heat shock promoter (ΔchoDPhspchoD). The mice infection experiments confirm the significance of ChoD in the pathogenesis of M. tuberculosis.

  • cholesterol oxidase
  • tuberculosis
  • pathogenesis
  • mycobacteria


Mycobacterium tuberculosis is a principal bacterial pathogen of humans and it is able to survive and propagate inside mononuclear phagocytes. It has developed multiple strategies to circumvent the normal fate of phagocytosed organisms. Uncontrolled M. tuberculosis growth in its human host is associated with extensive lung damage that ultimately results in death by suffocation due to insufficient oxygen. Tubercle bacilli cause the damage without classical virulence factors known in other pathogens (e.g., toxins of Corynebacterium diphtheriae, Shigella dysenteriae or Vibrio cholerae). However, a number of genes have been identified as important for progression of tuberculosis, including some surface and secreted proteins (GlnA, SodA, Erp), general cellular metabolism enzymes (Icl, LipF), metal uptake (MgtC, IdeR), anaerobic respiration and oxidative stress proteins (NarG, KatG, AhpC) or transcriptional regulators (alternative sigma factors) (for review see Smith, 2003).

An attribute responsible for success of M. tuberculosis as a pathogen is its extremely lipid-rich cell wall, which is involved in the pathogenesis or progression of tuberculosis (Cox et al., 1999; Glickman et al., 2000; Smith, 2003). Steroid metabolism of mycobacteria is unusually rich: mycobacteria code for 250 genes involved in fatty acid metabolism and for counterparts of all known lipid and polyketide biosynthetic and degradation pathways.

A wide range of microorganisms, including fast-growing mycobacteria, can metabolize cholesterol (5-cholesten-3β-ol) and use it as a sole carbon and energy source (Martin, 1977; Sedlaczek, 1988). The degradation of cholesterol is achieved through a complex metabolic pathway involving many enzymatic steps: this starts with the oxidation of the 17-alkyl side chain and the steroid ring system, ultimately degrading the entire molecule to CO2 and H2O (Mahato & Garai, 1997). Key reactions in this pathway include oxidation of 3β-hydroxy-5-ene to 3-keto-4-ene driven by cholesterol oxidase (ChoD), desaturation of ring A of 3-ketosteroids by Δ1-dehydrogenase (KsdD) and 9α-hydroxylation in ring B (Sedlaczek, 1988).

ChoD is one of the most widely used enzymes for the enzymatic transformation of cholesterol and for the determination of serum cholesterol levels in clinical laboratories. This enzyme has been isolated and biochemically characterized from a number of microorganisms. ChoD has been identified in fast-growing mycobacteria (Wilmanska et al., 1995). More recently, ChoD was identified as an important cytosolic factor in Rhodococcus equi, a primary pathogen of horses (Navas et al., 2001). The M. tuberculosis genome also appears to contain an ortholog of the choD gene. Fusion of green fluorescent protein (GFP) to the putative leader sequence of this gene resulted in the excretion of GFP to the culture supernatants, suggesting that ChoD of M. tuberculosis is an extracellular enzyme (Cowley & Av-Gay, 2001). In this paper, it is shown that intact ChoD is important for the growth of M. tuberculosis in peritoneal macrophages and lungs of mice.

Materials and methods

Bacterial strains and culture conditions

The following bacterial strains were used: Escherichia coli Top10 (Invitrogen), Mycobacerium smegmatisΔksdD (Brzostek et al., 2005) and M. tuberculosis H37Ra. The mycobacterial strains were cultured in Middlebrook 7H9 broth or 7H10 agar plates supplemented with albumin-dextrose-sodium chloride and with kanamycin (25 µg mL−1), or hygromycin (10 µg mL−1), when required. For steroid bioconversion experiments, M. smegmatis strains were cultured in NB broth: nutrient broth (Difco), 8.0 g L−1; glucose, 10.0 g L−1; supplemented with Tween 80, 0.2%, (pH 6.0–6.2).

Gene cloning strategies

Standard molecular biology protocols were used for all cloning protocols (Sambrook & Russell, 2001). All PCR products were obtained using thermostable ExTaq polymerase (Takara, Japan) and cloned initially into a TA vector (pGEM-Teasy, Promega), and then released by digestion with appropriate restriction enzymes before cloning into the final vectors. Restriction enzymes recognition sites were incorporated into the sequence of primers. The plasmid and primers used in this work are listed in Tables 1 and 2, respectively.

View this table:

Plasmids used in this study

Cloning vectors
pGemTEasyT/A cloningPromega
pMV306Mycobacterial integrating vector, KanRMed-Immune Inc.
p2NILRecombination vector, non-replicating in mycobacteria, KanRParish & Stoker (2000)
pGoal17The source of PacI cassette, AmpRParish & Stoker (2000)
pMV261Mycobacterial Escherichia coli shuttle vector, carrying heat shock (Phsp) promoter, KanRMed-Immune Inc.
Vectors used for gene replacement
pAB14choDTb PstI–HindIII fragment including 3′ end and its downstream region (2219 bp) in p2NIL, KanRThis study
pAB18choDTb BamHI–HindIII fragment including 5′ end and its upstream region (1976 bp) in pAB14, KanRThis study
pAB20pAB18 with PacI cassette from pGoal17, KanRThis study
Over-production and complementation vectors
pAB2choDTb under Phsp promoter in pMV261, KanRThis study
pAB54choDTb under Phsp promoter in pMV306, KanRThis study
View this table:

Primer sequences used for PCR amplification

Amplified regionPrimer namePrimer sequence
Primers used to amplify DNA for targeted gene replacement
choDTb 5′ flanking region – sensechoTb-GR15′-cgggatccgccctggtcgacgccggag-3′
choDTb 5′ flanking region – reversechoTb-GR25′-ccaagcttccacgacgtcttggcgaactc-3′
choDTb 3′ flanking region – sensechoTb-GR35′-ccaagcttatcggcaaccaggtcaccc-3′
choDTb 3′ flanking region – reversechoTb-GR45′-cccacgctcggcgaacagg-3′
Primers used to clone genes for expression or complementation experiments
choDTb geneTbcho-s5′-gaagatctatgaagccggattacgacg-3′
choDTb geneTbcho-r5′-ggaattctcagcccgcgttgctgaccg-3′
  • Underlined regions represent additional sequences, which include restriction sites used during cloning steps.

Targeted gene replacement

To perform unmarked deletions in the choD gene of M. tuberculosis, a suicidal recombination delivery vector was constructed. The vector contained the 5′ end of choD (147 plus 941 bp of upstream region), which was connected to the 3′ end of the gene (420 plus 1827 bp of downstream region). The 5′ and 3′ fragments of the gene in the final vector were ligated out of frame, resulting in the expression of nonfunctional protein. The protocol of Parish & Stoker (2000) was used to disrupt choDTb at its native locus on the chromosome. The plasmid DNA (pAB20) was treated with NaOH (0.2 mM) and integrated into the M. tuberculosis chromosome by homologous recombination. The resulting SCO (single crossover recombinant) mutant colonies were blue, KanR and sensitive to sucrose. The site of recombination was confirmed by PCR and Southern hybridization. The SCO strains were further processed to select for double-crossover (DCO) mutants that were white, KanS and resistant to sucrose (2%). PCR and Southern hybridization were used to distinguish between the wild-type and DCO mutants. Probes used to hybridize to each gene were generated by PCR, by the labeling of a nonradioactive primer extension system (DIG-labelling system, Amersham) (Fig. 1).


Replacement of the wild-type choD of Mycobacterium tuberculosis with a mutant sequence. The chromosomal localization of choD is represented by the gray arrow. The restriction sites used for digestion of chromosomal DNA are denoted by single letters (K — KpnI; S — SmaI). The restriction DNA fragment and the size of the internal deletion in the mutated copies are shown. The genomic DNA designated for Southern hybridization was isolated from: wild-type M. tuberculosis; single crossover homologous recombination mutants (SCO); double crossover homologous recombination mutants carrying the wild-type choD gene (wt-DCO) or choD gene with the internal deletion (mut-DCO), exclusively.

Complementation and overexpression constructs

The choD gene of M. tuberculosis was PCR amplified and cloned into the pMV261 vector downstream from the Phsp promoter. The resultant construct, named pAB2, was used for overproduction of ChoD in M. smegmatis. Next, the cloned gene and promoter (PhspchoD) were relocated into the pMV306 integration vector. The resultant construct (pAB54) was used for the complementation of M. tuberculosisΔchoD.

Growth of M. smegmatis strain expressing choD of M. tuberculosis on cholesterol

The biotransformation of cholesterol and detection of cholestenone was performed as described previously (Brzostek et al., 2005).

Mycobecterium tuberculosis virulence assessment in mice

Mycobacterium tuberculosis wild-type strain, ΔchoD mutant strain and ΔchoD-PhspchoD complementation strains were grown to log phase, aliquoted and frozen at −70°C. The number of viable bacteria was controlled by plating appropriate dilutions on agar plates. C57BL/6 female mice were infected i.v. at c. 8–12 weeks of age with 0.2 mL aliquots of mycobacterial suspensions containing 1 × 106 CFU (with at least 95% viable bacteria). Ten weeks postinfection, mice were euthanized and the number of CFU in the spleen and lung was determined. Organs were aseptically removed and homogenized in Middlebrook 7H9 broth using an Ultra-Turrax T25 homogenizer (IKA Labortechnik) and than sonicated to break up clumps of bacteria. To enumerate CFU values, the aliquots of 10-fold serial dilutions of homogenates were plated onto Middlebrook 7H10 agar medium and colonies were counted after 3 weeks of incubation at 37°C. The number of CFUs per entire organ of at least 10 individual mice from each group was reported as the log10 value.

In vitro infection of mouse peritoneal macrophages

Mouse peritoneal macrophages were obtained using 10% thioglycolate broth medium (TMB, Gibco) as described previously (Park & Rikihisa, 1991). The number of macrophages in isolated peritoneal cell suspensions was evaluated by microscope examination of Giemsa staining slides and reached 80–90%. Macrophages were washed with Iscove's medium (PAA-Cytogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS; PAA-Cytogen), 2 mM l-glutamine (Sigma), 0.05 mM 2-mercaptoethanol (Sigma) and plated in 24-well plates at 106 cells well−1. Following 24 h-incubation in a humidified 10% CO2 atmosphere at 37°C, the cells were washed twice with prewarmed Iscove's medium to remove nonadherent cells. Stock vials of mycobacterial strains were thawed, diluted in 7H9 broth and used for infection at a ratio of one bacillus per 10 cells. After 6 h, cells were treated for 2 h with 200 µg mL−1 amikacin in Iscove's medium to eliminate extracellular mycobacteria (Mehta et al., 2006). The cells were washed again three times with fresh medium and then incubation was continued. Every 24 h, 200 µL of fresh media was supplemented to each well, to keep macrophages in a good condition. Six and 8 days postinfection, 0.05% Triton X-100 was added to each well to release any intracellular bacilli. Ten-fold serial dilutions of the mycobacterial suspension (sonicated to break up clumps of bacteria) were plated on Middlebrook 7H10 agar plates and incubated at 37°C for 3 weeks before counting of CFU. Additionally, the monolayers of macrophages incubated for 10 days with mycobacteria were examined by conventional microscopy on a Nikon EclipseTE2000-S microscope with ELWD × 40/0.60 Plan Fluor objective.


Mycobacterium tuberculosis choDTb coding for an active protein

The sequencing of mycobacterial genomes revealed the presence of putative choD orthologs in M. tuberculosis (Cole et al., 1998), Mycobacterium bovis (Garnier et al., 2003), Mycobacterium leprae (Cole et al., 2001) and M. smegmatis (TIGR, database). Moreover, both M. smegmatis and M. tuberculosis genomes contain putative genes coding for other enzymes of cholesterol catabolism. At the amino-acid level, ChoD of M. tuberculosis (Rv3409c, choDTb) has 60% identity with ChoD of Streptomyces coelicolor (NP628939), but only limited (about 20%) similarity to ChoD of R. equi (CAC44897) or Brevibacterium sp. (JQ1193). The identity between ChoDTb and its orthologs of other mycobacteria is very high (100%M. bovis — MB3747; 88%M. leprae — ML0389; 83%M. smegmatis — MSMEG1602). Interestingly, although 50% of the M. leprae genome contains truncated genes, ML0389 is complete. However, some other genes of sterols catabolism in M. leprae are encoded by pseudogenes.

Microbial ChoD (EC catalyses the oxidation and isomerization of cholesterol to cholestenone (4-cholesten-3-one), which is an initial step in the cholesterol degradation process. The resulting cholestenone might be further catabolised by a number of enzymes to final inorganic compounds (Fig. 2). It has been previously shown that neither wild-type M. smegmatis strains nor M. smegmatisΔksdD mutant strains (which accumulates intermediates of the cholesterol biodegradation process — androstendion) growing in the presence of cholesterol are able to accumulate cholestenone in amounts that are detectable by GC (Brzostek et al., 2005). To verify the enzymatic activity of the putative ChoD enzyme of M. tuberculosis, choDTb was cloned under control of the heat shock promoter (Phsp65, choDTb) and introduced into M. smegmatisΔksdD, which is able to transform cholesterol to androstendion. The host strain and resulting mutant that overproduces ChoDTb (M. smegmatisΔksdD-PhspchoDTb) were cultured in the presence of cholesterol. The degradation of cholesterol and accumulation of the steroid intermediates were monitored by GC within 72 h. The investigated strains were able to use cholesterol and accumulate androstendion: however, the temporary accumulation of cholestenone was observed exclusively in the strain that overexpressed ChoDTb, (Fig. 3). Moreover, choDTb was cloned in an E. coli expression system under the control of an isopropyl-b-d-thiogalactopyranoside (IPTG)-inducible promoter. The cell extracts of IPTG induced and uninduced E. coli were assessed for enzymatic activity of ChoD using a colorimetric assay as described previously (Kiatpapan et al., 2001). IPTG induction resulted in higher activity (about six times) of ChoD detected in cell extracts (data not shown), confirming that the M. tuberculosis choD paralog codes for an active ChoD protein.


Proposed pathway for cholesterol degradation (based on Martin, 1979; Sedlaczek, 1988). ChoD, cholesterol oxidase; KsdD, 3-ketosteroid dehydrogenase; AD, 4-androstene-3,17-dione; 9αAD, 9α-hydroxy-4-androstene-3,17-dione; ADD, 1,4-androstadiene-3,17-dione; 9αADD 9α-hydroxy-1,4-androstadiene-3,17-dione.


Monitoring of cholesterol degradation (open symbols) and cholestenone (filled symbols) accumulation by Mycobacterium smegmatisΔksdD-1 (circles), and M. smegmatisΔksdD-1 overproducing ChoD of M. tuberculosis (triangles). Results are representative of three independent experiments.

Mycobacterium tuberculosisΔchoDTb is attenuated in mouse peritoneal macrophages

To assess the role of ChoD in the pathogenesis process of M. tuberculosis, a strain carrying an internal deletion in the choD gene was prepared using the two-step recombination protocol of Parish & Stoker (2000), as described in the ‘Materials and methods’. The mutant strain was verified by PCR and Southern hybridization (Fig. 1). Moreover, a control strain carrying the ΔchoD gene complemented with intact choD controlled with the heat shock promoter (PhspchoDTb) was constructed. The wild-type M. tuberculosis strain and two mutant strains (ΔchoD and ΔchoD-PhspchoD) grew with the same doubling time in Middlebrook 7H9/OADC broth (data not shown) and were used to infect in vitro mouse peritoneal macrophages at a ratio of 0.1 : 1. The number of viable bacteria used for infection and recovered from macrophages 6 and 8 days postinfection was analyzed by CFU (Fig. 4). Additionally, the morphology of macrophages was examined 10 days after infection using a microscope (data not shown). The dramatic decrease in the number of viable bacilli was observed in the case of the ΔchoD mutant, but not in the wild-type or complemented strains. The macrophages monolayer was damaged significantly only by strains carrying intact choD gene, but not with those that were ΔchoD.


Intracellular growth of Mycobacterium tuberculosis wild-type strain and choD-mutant strains. The peritoneal macrophages for each experiment were collected from three individual mice. (a) The number of viable bacteria was assessed by CFU 6 and 8 days postinfection. The data are expressed as the geometric means±SD of the log10 CFU counts obtained from five wells per group per time point. Mtb-wt represents the M. tuberculosis wild-type strain, Mtb-choDmut represents M. tuberculosisΔchoD and Mtb-choDcomp represents M. tuberculosisΔchoD complemented with an intact choD gene (ΔchoD-PhspchoD). Results are representative of three independent experiments.

Mycobacterium tuberculosisΔchoDTb is attenuated in lungs and spleens of C57BL/6 mice

Mycobacteria introduced i.v. into mice disseminate rapidly in the organism and can be identified in spleens and lungs. The number of viable bacteria isolated from these organs shows a progression of infection. Three M. tuberculosis strains were used to infect i.v. C57BL/6 mice with 106 tubercle bacilli per mouse: (1) wild type; (2) ΔchoD and (3) ΔchoD-PhspchoD. At least 10 mice of each group were euthanized 10 weeks postinfection and the number of viable bacteria in isolated spleens and lungs was determined by CFU (Fig. 5). The wild-type bacilli and complemented mutant carrying an intact choD were maintained in vivo over this 10-week period, with about 105 bacilli isolated from lungs and 104 from spleens. The ΔchoD strain was cleared out from the lungs of the majority of mice and only about 103ΔchoD-bacilli were recovered from the spleens of the infected animals. The experimental infection of C57BL/6 mice with the mycobacterial strains listed above was also performed for 7 and 13 weeks (three mice in the group) with similar results being obtained (data not shown).


Experimental infection of C57BL/6 mice with wild-type Mycobacterium tuberculosis (diamonds), M. tuberculosisΔchoD (squares) and M. tuberculosisΔchoD-PhspchoD (triangles). The number of viable bacteria was determined in isolated organs (lungs and spleens) by CFU 10 weeks postinfection. The results are expressed as the log10 CFU counts from at least 10 individual mice. The differences between wild type and ΔchoD, or ΔchoD and ΔchoD-PhspchoD were statistically significant using a Mann–Whitney test (in lungs: P=0.00002, P=0.0003 and in spleens: P=0.00006, P=0.00007, respectively). The median for each group is indicated by the arrow head. The first and third quartiles of each group are connected by open rectangles.


The purpose of this study was to assess whether ChoD plays a role in the survival of tubercle bacilli during infection. An unmarked mutation in the gene encoding the ChoD was used for infection of peritoneal macrophages and mice. The results clearly showed the significance of ChoD in the pathogenesis process of M. tuberculosis. ChoD was previously described as a major membrane-damaging factor of R. equi, a primary pathogen of horses and opportunistic pathogen in humans with clinical and pathological characteristics similar to pulmonary tuberculosis in immunocompromised patients (Navas et al., 2001). The mutation in cholesterol oxidase (choE) was associated with a loss of cooperative (CAMP-like) hemolysis with sphingomyelinase-producing bacteria (Navas et al., 2001). On the other hand, the analysis of R. equiΔchoE mutant performed recently by another group showed no difference between the mutant and parent strain in cytotoxic activity for macrophages or in intramacrophage multiplication (Pei et al., 2006). The opposite effect of choD disruption observed in this study might have been due to the differences between the species analyzed or methodology used. The mice were infected with R. equi for a period of 4 days only and CFU was calculated for bacteria recovered from the liver. Moreover, the macrophages used were not activated before infection and bacteria were incubated for 24 h only (Pei et al., 2006). The global analysis of the genes required for the adaptation and survival of M. tuberculosis in macrophages showed that a disruption of choD partially attenuated the resultant mutant in IFN-γ-activated, but not resting, macrophages (Rengarajan et al., 2005). It is likely that genes important for pathogenesis are overproduced in vivo. Recently, it has been shown by a promoter trap study that choD of M. tuberculosis is expressed more intensively in mouse lungs compared with broth culture (Dubnau et al., 2005).

As cholesterol is an essential component of mammalian membranes, an important issue is whether M. tuberculosis, the human intracellular pathogen, is able to utilize cholesterol. It was previously shown that M. tuberculosis was not able to grow on minimal medium with cholesterol as a sole source of carbon and energy and that cholesterol-supplementing media was still present in M. bovis BCG culture media after 4 weeks of growth (Av-Gay & Sobouti, 2000). More recently, cholesterol-dependent growth was described for M. bovis BCG encoding the ring-degrading enzymes upregulated during this utilization (van der Geize et al., 2007). The available evidence suggests that cholesterol is essential for phagocytosis of the bacilli by the macrophage and the inhibition of phagosome maturation (Gatfield & Pieters, 2000; de Chastellier & Thilo, 2006). The uptake and metabolism of cholesterol are important for M. tuberculosis to be able to persist in the macrophages for longer periods of time (van der Geize et al., 2007). However, the likely role of ChoD in each of these processes needs to be elucidated by further studies.


The work was supported by grants from the State Committee for Scientific Research (KBN, contract no. 3P05A14024) and ICGEB (contract CRP/POL07-01). Dr T. Parish is thanked for the p2NIL/pGOAL17 recombination system. Dr Richard Bowater is thanked for critical reading of this manuscript.


  • Editor: Roger Buxton


View Abstract