We also compared the resistance to SDS and replication in mice of the Rv2951c mutant, deficient in synthesis of phthiocerol dimycocerosates but producing phthiodiolone dimycocerosates, wi
Trang 1between phthiocerol and phthiodiolone dimycocerosates and their critical role in the virulence and permeability of Mycobacterium tuberculosis
Roxane Sime´one, Patricia Constant, Wladimir Malaga, Christophe Guilhot, Mamadou Daffe´ and Christian Chalut
De´partement Me´canismes Mole´culaires des Infections Mycobacte´riennes, Institut de Pharmacologie et de Biologie Structurale, Toulouse, France
Mycobacteria are the agents of several important
human diseases, including tuberculosis and leprosy,
and remain an important cause of mortality and
morbidity worldwide According to the World
Health Organization (http://www.who.int/en/index.html),
Mycobacterium tuberculosis, the causative agent of tuberculosis in humans, is responsible for more than
8 million new cases and kills 2 million people every year Little is known about the molecular mechanisms
of mycobacterial pathogenicity, but accumulated data
Keywords
ketoreductase; Mycobacterium tuberculosis;
phenolic glycolipids; phthiocerol
dimycocerosates; tuberculosis
Correspondence
C Chalut, Institut de Pharmacologie et de
Biologie Structurale, 205 route de Narbonne,
31077 Toulouse Cedex, France
Fax: +33 5 6117 5994
Tel: +33 5 6117 5473
E-mail: Christian.Chalut@ipbs.fr
(Received 12 January 2007, revised 9
February 2007, accepted 14 February 2007)
doi:10.1111/j.1742-4658.2007.05740.x
Phthiocerol dimycocerosates and related compounds are important mole-cules in the biology of Mycobacterium tuberculosis, playing a key role in the permeability barrier and in pathogenicity Both phthiocerol dimyco-cerosates, the major compounds, and phthiodiolone dimycodimyco-cerosates, the minor constituents, are found in the cell envelope of M tuberculosis, but their specific roles in the biology of the tubercle bacillus have not been established yet According to the current model of their biosynthesis, phthiocerol is produced from phthiodiolone through a two-step process in which the keto group is first reduced and then methylated We have previ-ously identified the methyltransferase enzyme that is involved in this pro-cess, encoded by the gene Rv2952 in M tuberculosis In this study, we report the construction and biochemical analyses of an M tuberculosis strain mutated in gene Rv2951c This mutation prevents the formation of phthiocerol and phenolphthiocerol derivatives, but leads to the accumula-tion of phthiodiolone dimycocerosates and glycosylated phenolphthiodio-lone dimycocerosates These results provide the formal evidence that Rv2951c encodes the ketoreductase catalyzing the reduction of phthiodio-lone and phenolphthiodiophthiodio-lone to yield phthiotriol and phenolphthiotriol, which are the substrates of the methyltransferase encoded by gene Rv2952
We also compared the resistance to SDS and replication in mice of the Rv2951c mutant, deficient in synthesis of phthiocerol dimycocerosates but producing phthiodiolone dimycocerosates, with those of a wild-type strain and a mutant without phthiocerol and phthiodiolone dimycocerosates The results established the functional redundancy between phthiocerol and phthiodiolone dimycocerosates in both the protection of the mycobacterial cell and the pathogenicity of M tuberculosis in mice
Abbreviations
CFU, colony-forming unit; CI, competition index; DIM, diester of phthiocerol or phthiodiolone; DIM A, phthiocerol dimycocerosates; DIM B, phthiodiolone dimycocerosates; Hyg, hygromycin; Km, kanamycin; PGL, phenolglycolipid; PGL-tb, phenolglycolipid from M tuberculosis.
Trang 2strongly suggest that the mycobacterial cell envelope
plays a key role in both the pathogenesis and
resist-ance to the various hostile environments encountered
by pathogenic mycobacteria during the infection
pro-cess The mycobacterial cell envelope is a complex and
unusual structure [1] A key feature of this structure is
its very high lipid content, consisting of up to 60% of
the dry weight of the bacteria Among these lipids, the
diesters of phthiocerol and phenolic glycolipids (PGLs)
have attracted attention for years They are specifically
found in slow-growing species that include the patho-genic species Mycobacterium leprae, Mycobacterium ulcerans, Mycobacterium marinum and members of the
M tuberculosis complex [2] In M tuberculosis, the diesters of phthiocerol and phthiodiolone, called DIMs, are composed of a mixture of long chain b-diols that are esterified by multimethyl-branched fatty acids named mycocerosic acids [2] (Fig 1) The chemical structures of PGLs are very similar to those of DIMs, except that they harbor a phthiocerol chain
Fig 1 Proposed biosynthetic pathway leading to DIM A and PGL-tb from DIM B and glycosylated phenolphthiodiolone dimycocerosates, respectively The keto, hydroxyl and methoxyl groups are boxed in rectangles p, p¢ ¼ 2–4; n, n¢ ¼ 16–18; m2 ¼ 15–17; m1 ¼ 20–22;
R ¼ C 2 H5or CH3.
Trang 3x-terminated by an aromatic nucleus, the so-called
phenolphthiocerol, which in turn is glycosylated [2]
(Fig 1) In M tuberculosis the major constituents,
phthiocerol dimycocerosates (DIM A) and glycosylated
phenolphthiocerol dimycocerosates (PGL-tb), are
usu-ally accompanied by minor structural variants, called
phthiodiolone dimycocerosates (DIM B) and
phenol-phthiodiolone dimycocerosates, which contain a keto
group in place of the methoxy group at the terminus
of the b-diols (Fig 1)
Several laboratories have shown that both DIMs
and PGLs contribute to the permeability barrier
formed by the cell envelope of M tuberculosis and to
virulence [3–6] Nevertheless, their precise molecular
mechanisms of action are still unknown, and the
speci-fic roles of the various members of the DIM family,
e.g DIM A and DIM B, in these functions have never
been investigated The recent developments in our
understanding of the biosynthetic pathway of DIMs
have provided the means to investigate the
contribu-tion of these compounds to the biology of M
tubercu-losis Considerable efforts, started more than 40 years
ago, have been devoted to deciphering the complex
biosynthetic pathways of DIMs and PGLs These
efforts have led to the identification of more than 20
proteins required either for the formation or for the
translocation of these compounds [7] Remarkably,
most of these genes are clustered in a 70 kb region of
the M tuberculosis chromosome [3,4,8] Several lines
of evidence indicate that DIM A is formed from
DIM B in a two-step process [9]: the reduction of the
keto function carried by DIM B into a hydroxyl
group, to form phthiotriol dimycocerosates, followed
by the methylation of the hydroxyl group, catalyzed
by a methyltransferase, to yield DIM A (Fig 1) This
model is supported by the characterization of the two
enzymes responsible for these enzymatic reactions
Indeed, we previously showed that the protein encoded
by Rv2952 catalyzes the methylation of phthiotriol and
phenolphthiotriol to form phthiocerol and
phenol-phthiocerol, respectively [10] Subsequently, Onwueme
et al [11] identified Rv2951c as being the gene
enco-ding the phthiocerol ketoreductase that catalyzes
the reduction of the keto moiety in phthiodiolone
during DIM biosynthesis However, the role of
Rv2951c has been investigated in M ulcerans and
Mycobacterium kansasii by gene complementation
studies, and no formal evidence that Rv2951c has the
same function in M tuberculosis has been reported to
date
In this article, we describe the contruction of an
Rv2951c M tuberculosis mutant Biochemical analyses
of this strain demonstrated that Rv2951c catalyzes
the reduction of the keto moiety of both DIM B and phenolphthiodiolone dimycocerosates in M tuber-culosis.Comparison of the phenotypes of this Rv2951c mutant with those of the wild-type strain and a mutant deficient in DIM production demonstrates that DIM A and DIM B fulfil redundant functions regarding both the resistance of M tuberculosis to SDS and its viru-lence in the mouse model
Results
Disruption of the Rv2951c gene in M tuberculosis H37Rv and biochemical characterization of the Rv2951c gene-disrupted mutant
It has recently been proposed that Rv2951c encodes an oxydoreductase involved in the reduction of the keto group of DIM B to yield phthiotriol dimycocerosates, which in turn would be methylated to give DIM A [11] This proposal was based on the observation that some M ulcerans and M kansasii strains naturally produce DIM B but not DIM A, and that M ulcerans harbored a mutation within the Rv2951c ortholog Gene complementation studies with a multicopy plas-mid carrying a functional Rv2951c ortholog from
M marinum partially restored DIM A synthesis in the recombinant strains [11] Although these studies strongly suggested that Rv2951c catalyzes the reduc-tion of the keto group of DIM B, the role of Rv2951c remains to be demonstrated
To formally establish the function of this enzyme
in the synthesis of DIM in M tuberculosis, we construc-ted an Rv2951c knockout M tuberculosis H37Rv mutant strain, named PMM74, by replacing the wild-type allele of Rv2951c with a kanamycin (km)-dis-rupted allele using the temperature-sensitive⁄ sacB proce-dure [12] (supplementary Fig S1A) Lipids were then extracted from the PMM74 mutant and analyzed by TLC As shown in Fig 2A, the disruption of Rv2951c
in M tuberculosis selectively abolished the production
of DIM A but not that of DIM B When equivalent amounts of lipids were loaded on TLC plates, we observed that the mutant cells accumulated more DIM B than did the wild-type cells To further quantify the lipids produced by the various strains, cells were labeled with [1-14C]propionate, a precursor known to be incorporated into methyl-branched fatty acyl-contain-ing lipids, includacyl-contain-ing DIMs Analysis of the labelacyl-contain-ing (Fig 2B) confirmed that there were no traces of DIM A
in the DRv2951c::km mutant, and revealed that the amounts of DIM B accumulated by the DRv2951c::km mutant (46% of the labeled lipids) corresponded to those of DIM A + DIM B produced by the wild-type
Trang 4cells (47% of the labeled lipids) Complementation of the Rv2951c mutation by the introduction of a wild-type allele Rv2951c in the PMM74 mutant fully restored the production of DIM A (Fig 2A), indicating that the phenotypic differences observed between the mutated and the wild-type strains relied solely on the disruption
of the Rv2951c gene Together, these results esta-blished that Rv2951c is involved in the biosynthesis of DIM A and that DIM B is a precursor of DIM A in
M tuberculosis
To further characterize the structure of the DIM-like substances produced by the PMM74 mutant, lipids exhibiting Rf values similar to those of DIM A and DIM B were purified by preparative TLC and analyzed by MALDI-TOF MS The mass spectrum of the purified lipids exhibiting a TLC mobility similar
to that of DIM B showed a series of pseudomolecular ion (M + Na+) peaks at m⁄ z 1346, 1360, 1374, 1388,
1402, 1416, 1430, 1444, 1458, 1472, and 1486 (Fig 3A) These values were identical to those observed in the mass spectrum of DIM B purified from the wild-type strain (data not shown), confirm-ing that PMM74 still produced DIM B In sharp con-trast, no pseudomolecular ion peaks that would correspond to DIM A were seen in the mass spectrum
of purified lipids from PMM74 whose Rf value was similar to that of DIM A (Fig 3B) Indeed, the expected pseudomolecular ion (M + Na+) peaks were observed in the mass spectrum of the DIM A from the wild-type strain at m⁄ z 1362, 1376, 1390, 1404,
1418, 1432, 1446, 1460, 1474, 1488, and 1502 (Fig 3C) The m⁄ z value for each peak in this series
is 16 mass units higher than that of DIM B, due to the reduction of the keto function of DIM B, fol-lowed by the methylation of the resulting hydroxyl group Thus, inactivation of the Rv2951c gene in
M tuberculosis abolishes the production of DIM A but does not affect that of DIM B
Because the biosynthetic pathways of DIMs and PGLs are known to be closely related, we next focused
on the role of Rv2951c in PGL-tb biosynthesis Most
of the M tuberculosis strains, such as H37Rv, are devoid of PGL-tb, due to a frameshift mutation in the pks15⁄ 1 gene, and production of PGL-tb can be restored by introducing a functional pks15⁄ 1 gene into these strains [13] Accordingly, both the PMM74 and wild-type strains were transformed with plasmid pPET1 carrying a functional M bovis BCG pks15⁄ 1 gene, and the structure of the expected PGL-tb was determined after lipid extraction The PMM74:pPET1 strain produced a major glycoconjugate exhibiting slightly higher TLC mobility than PGL-tb pro-duced from H37Rv:pPET1 (Fig 2C) The glycolipids
B
C
A
D
Fig 2 TLC analyses of lipids extracted from the M tuberculosis
H37Rv DRv2951c::km and DppsE::km mutant strains (A) TLC
ana-lysis of DIMs from M tuberculosis H37Rv, the PMM74
(DRv2951c::km) mutant, and the PMM74:pRS01-complemented
strains Lipids extracts were dissolved in CHCl 3 , loaded onto the
TLC plate, and run in petroleum ether ⁄ diethylether (90 : 10, v ⁄ v).
DIMs were visualized by spraying the plate with 10%
phosphomo-lybdic acid in ethanol, followed by heating The positions of DIM A
(arrows) and DIM B (arrowheads) are indicated (B) TLC analysis of
radiolabeled DIMs from M tuberculosis H37Rv and the PMM74
mutant strains Lipids were visualized by using a PhophorImager
system (Molecular Dynamics, Sunnyvale, CA, USA) The positions
of DIM A (arrow) and DIM B (arrowheads) are indicated (C) TLC
analysis of glycolipids extracted from M tuberculosis H37Rv,
H37Rv:pPET1, and PMM74:pPET1 Lipids were dissolved in CHCl3
and run in CHCl3⁄ CH 3 OH (95 : 5, v ⁄ v) Glycoconjugates were
visu-alized by spraying the TLC plate with 0.2% anthrone (w ⁄ v) in
con-centrated H 2 SO 4 , followed by heating The position of PGL-tb
(arrow) is indicated (D) TLC analysis of radiolabeled DIMs extracted
from M tuberculosis H37Rv and the PMM56 (DppsE::km) mutant
strains Lipids were visualized by using a PhophorImager system.
The positions of DIM A and DIM B are indicated.
Trang 5produced by the PMM74:pPET1 and H37Rv:pPET1
strains were also analyzed by MALDI-TOF MS
following purification The mass spectrum of the
PGL-like compound from the PMM74:pPET1 strain showed
a series of pseudomolecular ion (M + Na+) peaks
16 mass units lower than for PGL-tb from the wild-type strain (series of M + Na+peaks at m⁄ z 1864,
1878, 1892, 1906, 1920, 1934, 1948, 1962, 1976, 1990,
2004, and 2018) (Fig 4) Because the Rv2951c gene was shown to be involved in the modification of the keto group of DIM B to yield DIM A, we speculated that the glycolipid produced by the PMM74:pPET1 mutant might be a glycosylated phenolphthiodiolone dimycocerosate To confirm this hypothesis, this glyco-lipid was further analyzed by 1H-NMR (Fig 5) All the proton signal resonances typical of PGL-tb were detected, with the notable exception of those corres-ponding to the methoxyl group of the phenolphthio-cerol dimycocerosate portion of PGL-tb expected at 3.32 p.p.m (singlet, 3H) [14,15] This observation was consistent with the absence of the proton resonance at 2.85 p.p.m corresponding to that of the methine pro-ton of the carbon bearing the methoxyl group [14,15] Furthermore, the proton resonances of the methyl group b (Fig 5) was observed at 1.05 p.p.m., instead
of 0.87 p.p.m for the corresponding methyl resonances
in phthiocerol and phenolphthiocerol dimycocerosates [16] Together, these results clearly demonstrated that the PGL produced by the PMM74 mutant strain is a triglycosylated phenolphthiodiolone dimycocerosate, most likely a tri-O-methyl-fucosyl-(a1–3)-rhamnosyl-(a1–3)-2-O-methyl-rhamnosyl-a-phenolphthiodiolone dimycocerosate, and therefore that Rv2951c is also implicated in the biosynthesis of PGL-tb in M tuber-culosis by catalyzing the reduction of the keto group
of phenolphthiodiolone
Construction of a DIM-less mutant of
M tuberculosis H37Rv by disruption of the ppsE gene
The construction of a M tuberculosis mutant unable
to synthesize DIM A but proficient at DIM B produc-tion, namely the PMM74 mutant strain, prompted us
to address the question of the specific role of DIM A and DIM B in the biology of M tuberculosis We chose to compare the phenotypes of the DRv2951c::km mutant (DIM A–, DIM B+) with those of the wild-type strain (DIM A+, DIM B+) and those of a DIM-less mutant The last of these was constructed by insertion⁄ deletion within the ppsE gene which encodes
a polyketide synthase required for the formation of the b-diol chain [17,18] (supplementary Fig S1B) Bio-chemical analyses of the resulting mutant, named PMM56, confirmed that it was unable to synthesize either DIM A or DIM B (Fig 2D)
A
B
C
Fig 3 MALDI-TOF mass spectra of purified lipids exhibiting TLC
mobilities similar to those of DIM B (A) and DIM A (B) from
M tuberculosis PMM74 and of DIM A (C) from M tuberculosis
H37Rv (wild type).
Trang 6Effect of the absence of DIM A on the
susceptibility of M tuberculosis to SDS
In a previous study, we demonstrated that DIMs are
involved in the resistance of the tubercle bacillus to
detergent, a feature related to the cell envelope
per-meability [4] To determine the contribution of DIM A
and DIM B to this resistance, we compared the
sensi-tivity to SDS of three M tuberculosis strains: the
wild type (DIM A+, DIM B+), the PMM74 mutant
(DIM A–, DIM B+), and the PMM56 mutant
(DIM A–, DIM B–) The three strains were incubated
with 0.1% SDS for 1, 4 and 8 days, and their survival
was then evaluated (Fig 6) The DIM-less mutant
(PMM56) was much more sensitive to SDS than the
wild-type strain: after 1 day of exposure to the
deter-gent, the number of colony-forming units (CFUs) was
20-fold lower for PMM56 than for the wild-type
strain, confirming our previous observation [4] In
sharp contrast, no difference in CFU number was
detected between the wild-type strain and the DIM A
mutant (PMM74) that was still able to synthesize
DIM B The survival curves of the wild type and
PMM74 were almost superimposable for the first two
time points, and the number of CFUs was even higher for PMM74 than for the wild type after 8 days of incubation with SDS These data indicate that the lack
of DIM A in PMM74 did not induce an important structural cell wall modification
Effect of the absence of DIM A on the virulence
of M tuberculosis in mice DIM deficiency has been shown to be associated with virulence attenuation of M tuberculosis in mice [3,4,8]
To investigate the role of DIM A and DIM B in this phenotypic modification, we performed in vivo compe-tition assays in mice to compare the virulence of PMM56 (DIM-less) and PMM74 (DIM A-less) mutant strains to that of the wild-type strain Mice were infected intranasally either with a mixture of strains H37Rv:pMV361H and PMM74 or with a mix-ture of strains H37Rv:pMV361H and PMM56 The infectious dose used for each mouse was around
5· 103CFU, in a ratio (CFUs of mutant inocula-ted⁄ CFUs of wild type inoculated) of 0.722 for the PMM74⁄ H37Rv:pMV361H inoculum and 0.944 for the PMM56⁄ H37Rv:pMV361H inoculum, as deter-mined by growth on 7H11 plates containing either kanamycin (Km) or hygromycin (Hyg) Mice were killed 1 or 21 days postinfection, and the mutant and wild-type loads in lungs and spleens were determined
on the basis of growth on selective media
As expected, 21 days postinfection, we observed a marked difference between the number of CFUs of the wild-type strain and that of the PMM56 mutant strain in lungs and in the spleen (Fig 7A) Indeed, both strains multiplied in lungs, but whereas an aver-age of 5.92· 106CFU was recovered from lungs for the wild-type strain, only 3.91· 104CFU were recov-ered for the PMM56 mutant strain We also observed
a growth defect for the DIM-less mutant in the spleen: on day 21, an average of 8.57· 103CFU was recovered for the wild-type strain against 1.24· 102CFU for the PMM56 strain A competition index (CI) was determined by calculating the ratio of mutant to wild-type bacteria after correcting for the ratio of these strains in the inoculum It appeared that, 21 days after infection, the ratio of PMM56 to wild-type bacteria was diminished more than 100-fold (CI ¼ 9.28 · 10)3) in lungs and 18-fold (CI¼ 5.39· 10)2) in the spleen, relative to the initial infect-ing ratio (Fig 7B) In contrast, mycobacteria deficient
in DIM A but synthesizing DIM B (the DRv2951c::km mutant) did not show significantly attenuated growth
in comparison to the wild-type strain in mice Indeed,
21 days postinfection, averages of 1.41· 107CFU and
A
B
Fig 4 MALDI-TOF mass spectra of the purified glycolipid from
M tuberculosis PMM74:pPET1 (A) and of PGL-tb from M
tubercu-losis H37Rv:pPET1 (B).
Trang 7of 1.50· 104CFU, respectively, were recovered from
the lungs and the spleen for the wild-type strain,
against 1.10· 107CFU and 1.50· 104CFU for
the PMM74 mutant strain These CFU counts
esta-blished that mice infected with PMM74 had similar
ratios of mutant to wild-type bacteria in both lungs
and spleen after 21 days as compared to the ratio of
bacteria in the initial inoculum (CI¼ 1.04 and 1.44)
(Fig 7B)
These mixed-infection experiments confirmed that
DIMs are major determinants for the pathogenicity of
M tuberculosis Moreover, the ability of the PMM74 mutant strain to replicate and persist in the lungs and spleens of mice clearly indicates that DIM A is not required for full virulence of M tuberculosis when DIM B is produced in the mycobacterial cell at the same level as DIM A in the wild-type cell
Discussion
The main objectives of this study were: (a) to fur-ther characterize the biosynthetic pathways of two
Fig 5 The 1 H-NMR spectrum of the purified glycolipid from PMM74:pPET1 The structure of the analyzed compound is shown above the spectrum (p, p¢ ¼ 2–4; n, n¢ ¼ 16–18; m ¼ 15–17; R ¼ C 2 H 5 ) The two doublets at 6.97 and 7.08 p.p.m (g, h) are assigned to phenolic proton resonances Three anomeric proton resonances are seen at 5.50 p.p.m (1H, i) and 5.15 p.p.m (2H, i¢) The four singlets at 3.5–3.6 p.p.m (j) are assigned to sugar-linked methoxyl proton resonances The multiplet centered at 4.83 p.p.m (a) is attributed to methine resonances of esterified b-diol The doublet at 1.15 p.p.m (f) corresponds to the resonance of a methyl group in the a position of the fatty acyl residues The signals that correspond to the resonance of the methine proton of the a carbon (d) and that of the carbon bearing the methyl group near the keto group (d¢) are observed at 2.55 p.p.m Signals of several terminal methyl proton resonances are seen at 0.8–1.0 p.p.m (e), consistent with the presence of multimethyl branched fatty acyl residues The resonance of the protons of the methyl group adjacent to the keto group is seen as a signal at 1.05 p.p.m (b) The two arrows show the proton signal resonance positions corres-ponding to the methoxyl group (expected at 3.32 p.p.m.) and the methine proton of the carbon bearing the methoxyl group (expected at 2.85 p.p.m.) of the phenolphthiocerol dimycocerosate portion in PGL-tb.
Trang 8important virulence factors for M tuberculosis, i.e.
DIMs and PGL-tb; and (b) to address the question of
the function of the two related molecules DIM A and
DIM B According to the accepted model of DIM and PGL biosynthesis [17], DIM A and PGL-tb are expec-ted to be derived from DIM B and glycosylaexpec-ted phe-nolphthiodiolone dimycoserosates, respectively, after two enzymatic steps: the reduction of the keto group
to give phthiotriol and glycosylated phenolphthiotriol dimycocerosates, and the methylation of the hydroxyl group by the previously identified methyltransferase encoded by Rv2952 [10] Onwueme et al [11] have demonstrated, in a recent study, that the lack of a functional Rv2951c ortholog in some M ulcerans and
M kansasii strains was responsible for diacyl phthioc-erol deficiency in these strains In addition, they have shown that complementation of M ulcerans and
M kansasii with a functional Rv2951c gene from
M marinum leads to the accumulation of diacyl phthiotriols [11], indicating that Rv2951c encodes the phthiodiolone ketoreductase that catalyzes the forma-tion of phthiotriol, the substrate of the Rv2952 methyl-transferase in these strains
In the present study, we constructed and biochemi-cally characterized an M tuberculosis strain harboring
a mutation within the Rv2951c gene Upon transfer of
Fig 6 Susceptibility to SDS of the M tuberculosis wild-type
(DIM A + , DIM B + ) (circle), PMM74 (DIM A – , DIM B + ) (triangle) and
PMM56 (DIM A–, DIM B–) (square) mutant strains Strains were
inoculated in 7H9 supplemented with ADC to which 0.1% SDS
was added The number of viable bacteria was evaluated by plating
serial dilutions of the different cultures onto 7H11 supplemented
with OADC and incubation at 37 C The values shown are the
means ± standard deviations for three independent experiments.
B
A
*
Fig 7 Competition between mutant and wild-type strains in infected mice (A) Numbers of CFUs recovered from the lungs and the spleen
of mice infected with a mixture of the type H37Rv:pMV361H and DIM-less PMM56 strains (left panel) or with a mixture of the wild-type H37Rv:pMV361H and DIM A-less PMM74 strains (right panel), at 1 day (J1) or 21 days (J21) after infection The CFU numbers were determined by plating dilutions of homogenized tissues on 7H11 media containing either Km (for CFU counts of the mutant strains) or Hyg (for CFU counts of the wild-type strain) White, black and gray bars represent the numbers of CFUs corresponding to the H37Rv:pMV361H, the PMM56 and the PMM74 strains, respectively Values are means ± standard deviations (error bars) of CFU counts for five infected mice.
*When spleen homogenates from the five mice infected with the H37Rv::pMV361H ⁄ PMM56 mixture were plated on 7H11 plates contain-ing Km, no colonies were obtained for four mice, indicatcontain-ing that fewer than 50 bacteria of the PMM56 mutant strain were present in the spleen of these mice For the determination of the number of PMM56 bacteria present in the spleen of these four mice, we thus chose a value of 50 CFU per spleen The average number of CFUs recovered from the spleen of mice infected with the H37Rv::pMV361 ⁄ PMM56 mixture is therefore overestimated for the PMM56 mutant (B) CI for the DIM-less PMM56 (white bars) and the DIM A-less PMM74 (gray bars) mutant strains in the lungs and spleen of the infected mice after 21 days of infection CI is defined as the ratio of mutant to wild-type CFUs in the organ divided by the mutant to wild-type bacterial ratio in the inoculum.
Trang 9plasmid pPET1, this strain accumulated both DIM B
and glycosylated phenolphthiodiolone dimycoserosates,
but was unable to synthesize DIM A and PGL-tb
This mutant strain did not produce the intermediates
phthiotriol and glycosylated phenophthiotriol
dimyco-cerosates found in the DRv2952::km mutant strain,
demonstrating that the products of Rv2952 and
Rv2951c are not involved in the same enzymatic step
of the DIM A and PGL-tb pathway This phenotype
was not due to a polar effect on a downstream gene,
as the transfer of a functional Rv2951c gene carried on
a mycobacterial plasmid fully reversed this biochemical
phenotype Thus, our data extend the results reported
by Onwueme et al to M tuberculosis, and further
demonstrate that the ketoreductase encoded by
Rv2951c is involved in the biosynthetic pathway of
PGL-tb by catalyzing the formation of
phenolphthiot-riol dimycoserosates
Previous reports have established that DIMs are
important virulence factors of M tuberculosis and
contribute to the cell envelope permeability barrier
[3–5,19] However, these reports were based on the
phenotypic analysis of mutants deficient in both
DIM A and DIM B biosynthesis, and did not allow
the precise definition of the role of each of the
mole-cules With our DIM-less and DIM A-less mutants
derived from the same parental strain, we first
addressed the question of the specific role of DIM A
and DIM B in the resistance of cells to SDS We
demonstrated that the occurrence of comparable
amounts of DIM B can complement the absence of
DIM A with regard to the SDS resistance, suggesting
that the two molecules fulfill redundant functions
regarding the protection of the mycobacterial cell
against environmental attack These results also
sug-gest that the methoxyl group located at the terminal
end of the phthiocerol chain does not significantly
contribute to the structural organization of the
myco-bacterial cell envelope We next used the mouse
infec-tion model to investigate the effect of the lack of
DIM A production on virulence attenuation of the
resulting bacteria We found that, unlike the
DIM-less mutant, the mutant deficient in DIM A replicated
as well as the wild-type strain in mice, indicating that:
(a) DIM A is not required per se for full virulence of
M tuberculosis; and (b) DIM B can contribute to the
same extent as DIM A in pathogenicity when this
compound is overproduced in the mycobacterial cell
Interestingly in this experiment, the DIM-less mutant
(PMM56) was clearly defective for growth in both
lung and spleen, which is consistent with previous
findings by Rousseau et al [19] These data are in
contrast with those of Cox et al [3], who found that
DIMs are required for optimal growth in the lungs but not in the spleen of mice The discrepancy between these results might be explained by the dif-ferent genetic backgrounds of the M tuberculosis strains used in these studies or⁄ and by different experimental conditions Indeed, Cox et al [3] ted mice by intravenous injection with a high infec-tious dose (106CFU), whereas Rousseau et al used, like us, the intranasal route of infection and a low infectious dose (104 CFU)
The precise molecular mechanisms by which DIMs act in the course of infection are still unclear It is possible that DIMs contribute to pathogenicity pas-sively, by protecting the tubercle bacillus against the antimicrobial responses of the host during infection Indeed, several lines of evidence, including our SDS experimental data, indicate that these molecules con-tribute to the structural organization of the myco-bacterial cell envelope and are involved in the cell wall permeability barrier of mycobacteria [4] The growth defect observed for the DIM-less mutant in mice could therefore result from altered cell wall per-meability In contrast, the PMM74 mutant strain may behave like the wild-type strain in mice because this mutant has an unaffected cell envelope organiza-tion Alternatively, the external localization of DIMs
in the cell envelope raises the possibility that these compounds act in vivo by interacting with some com-ponent of the host cell and by modulating the host immune response to contain the infection This hypothesis is supported by recent data suggesting that DIMs modulate the host immune response in the very early steps of infection [19] In that situ-ation, it can be inferred from our results that DIM B is able to fulfill the same function as DIM A
in this immunomodulation activity, suggesting that the methoxyl group carried by DIM A is not involved in this process
Our study provides the first structure–function ana-lysis of DIMs in the pathogenicity of the tubercle bacillus Nevertheless, more experiments are required
to clarify the precise roles played by DIMs in the cell envelope architecture and in virulence The generation
of M tuberculosis mutants producing DIM derivatives, such as the PMM74 mutant, could be very useful for addressing this issue Indeed, the biochemical charac-terization of these mutants and the analysis of their growth characteristics in various cellular and animal models may lead to the identification of the structural motif(s) of DIMs involved in pathogenesis, and thereby provide clues to decipher the mechanism by which these compounds contribute to the pathogenesis
of tuberculosis
Trang 10Experimental procedures
Bacterial strains, growth media and culture
conditions
Plasmids were propagated at 37C in Escherichia coli DH5a
in LB broth or LB agar (Invitrogen, Cergy Pontoise, France)
supplemented with either Km (40 lgÆmL)1) or Hyg
(200 lgÆmL)1) M tuberculosis H37Rv, PMM56 and
PMM74 strains were grown at 37C in Middlebrook 7H9
broth (Invitrogen) containing ADC (0.2% dextrose, 0.5%
BSA fraction V, 0.0003% beef catalase) and 0.05%
Tween-80 when necessary, and on solid Middlebrook 7H11 broth
containing ADC and 0.005% oleic acid (OADC) For
bio-chemical analyses, mycobacterial strains were grown as
sur-face pellicles on Sauton’s medium When required, Km and
Hyg were used at concentrations of 40 lgÆmL)1 and
50 lgÆmL)1, respectively Sucrose 2% (w⁄ v) was used to
sup-plement 7H11 for the construction of the PMM74 mutant
General DNA techniques
Molecular cloning experiments were performed using
stand-ard procedures The cloning vectors used were pGEM-T
(Promega, Lyon, France) and pPR27 [20] Mycobacterial
genomic DNA was extracted from 5 mL of saturated
cul-tures as previously described [21] PCR experiments for
plasmid constructions or genomic analysis were performed
with standard conditions on a GeneAmp PCR system 2700
thermocycler (Applied Biosystems, Courtaboeuf, France)
PCR was performed in a final volume of 50 lL containing
2.5 units of Pfu DNA polymerase (Promega)
Construction of M tuberculosis H37Rv
gene-disrupted mutants
A 2749 bp DNA fragment containing the Rv2951c gene
of M tuberculosis H37Rv flanked by 760 and 843 bp at the
5¢- and 3¢-end, respectively, was amplified by PCR from genomic DNA using oligonucleotides 2951A and 2951B (Table 1), and cloned into pGEM-T to give pCG163 An internal Rv2951c fragment of 801 bp was removed by a ClaI digestion and substituted by a km resistance cassette [22] to yield pCG167 This plasmid was subsequently digested by PmeI, and the 4542 bp fragment that contained the disrup-ted Rv2951c gene and its flanking regions was purified and inserted at the XbaI site of pPR27, a mycobacterial thermo-sensitive suicide plasmid harboring the counterselectable marker sacB, to give plasmid pCG175 This vector was elec-trotransformed in M tuberculosis H37Rv, and transform-ants were selected on 7H11 supplemented with OADC and
Km at 32C [12] Two clones were selected and grown in
5 mL of 7H9 medium containing Tween-80 and Km at
32C for 3 weeks Several dilutions of these cultures were then plated onto 7H11 agar plates containing OADC, Km and 2% sucrose, and incubated at 39C PCR screening for disruption of Rv2951c was performed with a set of specific primers (2951C; 2951D; 2951E; res1; res2) (Table 1) after extraction of the genomic DNA from several Km- and sucrose-resistant colonies One clone giving the correspond-ing pattern for disruption of Rv2951c was selected for fur-ther analyses and named PMM74 (supplementary Fig S1A)
To construct a DIM-less mutant of H37Rv, we chose to inactivate the ppsE gene, one of the genes shown to be involved in the formation of the phthiocerol backbone [17,18] We used the strategy described by Bardarov et al [23] A 2660 bp fragment of the ppsE gene was amplified using M tuberculosis genomic DNA and primers ppsE1 and ppsE2 (Table 1) in a final volume of 50 lL containing 2.5 units of Taq DNA polymerase (Roche Molecular Bio-chemicals, Meylan, France) The PCR fragment was inserted within the vector pGEM-T to give pWM39 The km resist-ance cassette was then inserted between the KpnI and BglII sites of the ppsE gene fragment generating a 523 bp deletion
to yield pWM40 The PmeI fragment from pWM40 was then cloned within the cosmid vector pYUB854 [23] The resulting
Table 1 Oligonucleotides used in this study.