Patent Application
20090170142
The cell envelope plays a major role in the physiology of Mycobacterium tuberculosis, the causative agent of tuberculosis in human, a disease still responsible for more death than any other single infectious agent. First, this complex structure provides a strong resistance to degradation by host enzymes and a permeability barrier to antibiotics and to toxic molecules produced by the host. Second, it contains components which exert an active effect to facilitate uptake of the bacterium and to modulate the host immune response (Daffé and Draper, 1998). The mycobacterial envelope is characterized by a very high lipid content (60% of the dry weight) and the occurrence of lipids with unusual structures (Daffé and Draper, 1998). The major lipid constituents of this cell wall are the mycolic acids. These molecules are specifically found in the Corynebacterineae suborder including corynebacteria, mycobacteria, nocardia or rhodoccoci where they exist either as esters of trehalose or esterified to the arabinogalactan core of the bacteria cell wall (Daffé and Draper, 1998; Daffé, 2005). All mycolic acids consist of α-alkyl β-hydroxy branched fatty acids but their size and their structure differ according to the bacterial species (Asselineau et al., 2002). For instance mycobacterial mycolic acids are made of very long carbon chains (C60-C90) containing additional motifs such as oxygen functions, cyclopropyl ring or methyl branches whereas Corynebacteria spp produce a mixture of saturated and insaturated corynomycolic acids which typically range in size from 30 to 36 carbons. Mycolic acids are key structural components of the cell envelope and their biosynthesis pathway is the target of the major antituberculous drug, isoniazid (Banerjee et al., 1994). The structure of mycolate has been found to be critical for initial replication and persistence in vivo (Takayama et al., 2005). In slow growing mycobacteria, mycolic acids are associated with a number of extractable lipids containing methyl-branched fatty acids (Minnikin et al., 2002). Various studies have shown that some of these compounds, such as phthiocerol dimycocerosates (DIM) and phenolic glycolipid (PGL-tb), contribute to pathogenicity of M. tuberculosis (Camacho et al., 1999; Reed et al., 2004).
The biosynthesis of the various and unusual lipids of M. tuberculosis involves the combined action of fatty acid synthase (Fas) systems and type-I polyketide synthases (Pks). For instance, the formation of mycolates required two Fas systems and one Pks: Fas-I, a multifunctional protein, is dedicated to the production of short C16,18 fatty acids; Fas-II, a complex of monofunctional proteins elongates fatty acids generated by Fas-I to yield long chain fatty acids ranging from C48 to C64 in length; Pks13 catalyses the condensation of two fatty acids to form mycolic acids (Takayama et al., 2005; Portevin et al., 2004). Biosynthesis of DIM and PGL-tb requires Fas-I and seven Pks (Onwueme et al., 2005). Finally, Fas-I and Pks2 or Pks3/4 are involved in the formation of the various multimethyl-branched fatty acids found in the trehalose derived lipids specific of M. tuberculosis (Sirakova et al., 2001; Dubey et al., 2002). Overall, the genome of M. tuberculosis encodes more than eighteen type-I Pks and two Fas systems (Cole et al., 1998). These enzymes are the key players which endow M. tuberculosis with the unique ability to produce an impressive variety of lipids of unique structure.
To be functional, the acyl carrier protein (ACP) domains of Fas and Pks need to be converted from their inactive apo-forms to their functional holo-forms by the covalent attachment of a 4′-phosphopantetheine (P-pant) group to an hydroxyl group of an invariant serine residue (Walsh et al., 1997; Keating and Walsh, 1999). The role of this flexible prosthetic arm is to provide an attachment site for chain-extension intermediates and to shuttle the growing chains between the different catalytic sites of the synthase complexes (Cane and Walsh, 1999). This feature is shared by another class of enzymes, the non-ribosomal peptide synthases (NRPS), which are involved in the production of siderophores in M. tuberculosis (De Voss et al., 1999). This posttranslational modification is catalyzed by an 4′-phosphopantetheinyl transferase (PPTase) which transfers the P-pant group from coenzyme A (CoA) to the ACP (Lambalot et al., 1996). PPTases have been identified and biochemically characterized in a number of microorganisms and have been classified in three groups based on primary sequence similarity and substrate specificity (Wiessman et al., 2004 and references therein). Members of the first group exemplified by the Holo-(acyl carrier protein) synthase (AcpS) of Escherichia coli, are about 120 residues in size and act as homotrimers (Parris et al., 2000). These AcpS-type PPTases have a narrow substrate specificity limited to the ACP of type-II Fas and Pks systems (Mootz et al., 2001). The second group comprises PPTases ressembling the Sfp (Surfactin phosphopantetheinyl transferase) protein, a PPTase required for the production of the antibiotic surfactin in Bacillus subtilis (Quadri et al., 1998a). These Sfp-type PPTases are about twice the size than those of the first group (220-240 residues) and exist under monomeric form (Mofid et al., 2004). They are often associated with genes encoding the production of secondary metabolites but exhibit very broad substrate specificities and are usually able to modifying both type-I and type-II ACP and peptidyl carrier protein (PCP) domains (Weissman et al., 2004; Quadri et al., 1998a). PPTases of the third group are incorporated as a catalytic domain in type-I Fas and allow self-phosphopantetheinylation of the ACP domain of the protein (Fichtlscherer et al., 2000).
Usually bacteria contain more than one PPTase dedicated to one or several P-pant dependant pathways. For instance, E. coli has three PPTases (Lambalot et al., 1996), AcpS involved in fatty acid synthesis, EntD a Sfp-type PPTase involved in the biosynthesis of the siderophore enterobactin and the product of gene yhhU which has an unknown physiological function. It has been shown that AcpS is essential for cell viability but not EntD (Flugel et al., 2000). In contrast, in Bacillus subtilis which has two PPTases, Sfp can complement the activity of AcpS and sustain fatty acids biosynthesis after inactivation of acpS (Mootz et al., 2001).
In their initial analysis of M. tuberculosis genome, Cole et al. (1998) identified Rv2523c as a gene encoding a putative PPTase related to AcpS. At the same time, a second PPTase was discovered in M. tuberculosis by Quadri et al (1998b) who found that Rv2794c encodes a Sfp-type PPTase, renamed PptT, that was able to activate in vitro, two NRPS required for the assembly of siderophore mycobactin. However, in spite of the importance of the Fas-I system and type-I Pks for the biology of M. tuberculosis, no data concerning the posttranslational modification of these enzymes have been reported in mycobacteria. Two PPTases have been identified which would be responsible for the activation of more than 20 proteins in M. tuberculosis but their respective repertoire of substrates, their putative redundancy and their importance for the mycobacterial biology have not been previously investigated. Answering these questions is crucial to understand the biosynthesis of lipids in M. tuberculosis.
The inventors have now found that orthologs of the two previously identified M. tuberculosis PPTases are found in other Corynebacterineae species. They demonstrated that both PPTases are essential for growth of mycobacteria and display identical functions in mycobacteria and corynebacteria: AcpS is responsible for the posttranslational modification of Fas-I and PptT, the Sfp-type PPTase, is involved in the activation of the condensing enzyme Pks13. In addition, they showed that the various type-I Pks required for the formation of lipid virulence factors are activated by PptT in M. tuberculosis. The involvement of PptT of M. tuberculosis in several P-pant dependent pathways essential for growth or virulence makes it an attractive potential target for the development of new antimycobacterial drugs.
A first aspect of the present invention is hence the use of a PptT protein, as a target for screening compounds for identifying those having an antibiotic activity. According to this invention, said PptT protein is preferably from a pathogenic bacterium containing mycolic acids, and more preferably a corynebacterium, a mycobacterium, or a nocardia. For example, the PptT protein is from a bacterium selected in the group consisting of Corynebacterium diphtheriae, Corynebacterium minutissimum, Corynebacterium pseudotuberculosis, Mycobacterium africanum, Mycobacterium avium, Mycobacterium chelonae, Mycobacterium fortuitum, Mycobacterium kansasii, Mycobacterium leprae, Mycobacterium malmoense, Mycobacterium marinum, Mycobacterium microti, Mycobacterium paratuberculosis, Mycobacterium scrofulaceum, Mycobacterium simiae, Mycobacterium szulgai, Mycobacterium tuberculosis, Mycobacterium ulcerans, Mycobacterium xenopi, Nocardia asteroids, Nocardia brasiliensis, Nocardia farcinica, Nocardia nova, and Nocardia otitidiscaviarum. In particular, PptT proteins of SEQ Nos: 2 to 9 can be used according to the invention.
Of course, fusion proteins corresponding to a PptT protein as described above, fused to a peptidic moiety that enables its easier purification without modifying its activity, are herein also considered as PptT proteins.
The present invention also pertains to an in vitro screening process for identifying compounds having an antibiotic activity, by measuring the activity of a PptT protein in the presence or absence of said compounds.
In a process according to the invention, the PptT activity can be measured by (i) incubating an isolated PptT with a Pks protein or subunit thereof comprising an acyl carrier protein domain, and with a labelled Acetyl-CoA, (ii) precipitating said Pks protein or subunit thereof, and (iii) measuring the level of labelling of said precipitate. Any PptT protein as above-described can be used in such a test.
The present invention is illustrated by the experimental results and the figures described below.
Bacterial Strains and Growth Conditions
E. coli strains DH5α and C600 used for cloning experiments were grown in Luria Bertani (LB) Broth (Difco). C. glutamicum was cultured in Brain Heart Infusion (BHI) medium (Difco) containing 0.05% of Tween 80 to prevent aggregation. The C. glutamicum ΔacpS::km mutant was grown in BHI medium supplemented with 0.03% Tween 40 and 0.03% (weight/volume) sodium oleate (Sigma). M. smegmatis was grown in LB Broth supplemented with 0.05% Tween 80. When required ampicillin (Amp), kanamycin (Km), chloramphenicol (Clm), streptomycin (Str), Hygromycin B (Hyg) and sucrose were used at a final concentration of 100 g/ml, 40 μg/ml (for E. coli) or 25 g/ml (for M. smegmatis and C. glutamicum), 15 μg/ml, 25 μg/ml, 200 μg/ml (for E. coli) or 50 g/ml (for M. smegmatis) and 5% (weight/volume), respectively.
Construction of Plasmids for the Production of the Recombinant PptT, Pks13, Mas and PpsA-D Proteins of M. tuberculosis in E. coli.
The pks13 (Rv3800), mas, ppsA, ppsB, ppsC and ppsD genes were amplified by PCR from M. tuberculosis H37Rv total DNA and the resulting fragments were inserted into the pET26b E. coli expression vector (Novagen) under the control of the T7 promoter to give plasmids pWM35, pETMas, pETA, pETB, pETC, pETD, respectively. These vectors allow expression of recombinant Pks13, Mas and PpsA-D proteins fused to a polyhistidine tag peptide at their C-terminal ends in the BL21ΔentD E. coli strain. Plasmid pWM35γ originates from pWM35 following site-directed mutagenesis of the two codons corresponding to the two catalytic residues Ser55 and Ser1266 responsible for the attachment of the P-pant moiety of CoA on the N- and C-terminal ACP domains of Pks13. In pWM35γ, these codons have been substituted by alanine encoding codons.
To coproduce PptT and the various Pks of M. tuberculosis in E. coli, pptT (Rv2794c) was amplified from M. tuberculosis H37Rv genomic DNA and inserted into plasmid pET26b downstream of the T7 promoter. The pptT gene plus 108 bp upstream of the start codon, a region carrying the T7 promotor region was reamplified by PCR and cloned into the BclI site of plasmid pLysS (Novagen) to give pLSfp. A derivative vector of pLSfp, named pLSfpA producing a non-functional truncated PptT protein was also made by digesting pLSfp by EcoRI which cut within the pptT gene. DNA ends were filled-in with Klenow fragment and the plasmid religated. All these constructs were checked by DNA sequencing (Genome Express, Grenoble, France).
The strains and plasmids used in the present study are summarized in Table 1 below.
E. coli
E. coli BL21(DE3)ΔentD
M. smegmatis
M. smegmatis mc2155 harboring
M. smegmatis mc2155 harboring
M. smegmatis mc2155 ΔacpS::km
M. smegmatis mc2155 ΔpptT::km
M. smegmatis mc2155 ΔacpS::km
M. smegmatis mc2155 ΔpptT::km
C. glutamicum
C. glutamicum ATCC13032
C. glutamicum ATCC13032
E. coli expression vector
M. tuberculosis
M. tuberculosis
M. tuberculosis
M. tuberculosis
M. tuberculosis
M. tuberculosis
E. coli vector
M. tuberculosis
C. glutamicum, AmpR,
C. glutamicum, , AmpR,
E. coli/C. glutamicum
M. smegmatis, ClmR
M. smegmatis, KmR,
M. smegmatis, KmR,
E. coli/mycobacteria
M. smegmatis, StrR
M. smegmatis, StrR
Construction of the C. glutamicum ΔacpS::km and ΔpptT::km Mutants and Complementation Vectors.
The C. glutamicum ΔacpS::km and ΔpptT::km mutants were generated as previously described (Portevin et al., 2004). Two DNA fragments overlapping the acpS gene (NCgl2405) at its 5′ and 3′ extremities were amplified by PCR from C. glutamicum strain ATCC13032 total DNA and inserted, flanking a Km resistance cassette, into the vector pMCS5 (MoBiTec, Göttingen, Germany) to give pMCS5ΔacpS. Similarly, two DNA fragments overlapping the pptT gene (NCgl1905) at its 3′ and 5′ extremities were amplified and cloned into plasmid pMCS5. The Km resistance cassette was inserted between these two fragments to yield plasmid pMCS5ΔpptT.
These two plasmids were transferred into C. glutamicum by electroporation and transformants were selected on plates containing either Km (for pMCS5ΔpptT) or Km, 0.03% sodium oleate (Sigma) and 0.03% Tween 40 (for pMCS5Δacps) (Sigma). Transformants in which allelic exchange had occurred between the WT chromosomal acpS or pptT gene and the mutated-borne alleles were characterized by PCR using various primers combination. Two recombinant strains, C. glutamicum ΔacpS::km and C. glutamicum ΔpptT::km named CGL2039 and CGL2035, respectively, were selected for further studies.
To construct the complementation plasmids, the acpS and pptT genes of C. glutamicum were amplified by PCR from C. glutamicum genomic DNA and the pptT gene of M. smegmatis was amplified by PCR from plasmid pC-pptTms. PCR products were inserted into a modified pCGL482 (Peyret et al., 1993) under the control of the cpsB promoter to give plasmids pCGL-acpScg, pCGL-pptTcg and pCGL-pptTms, respectively.
Construction of the M. smegmatis ΔacpS::km and ΔpptT::km Conditional Mutants and Complementation Vectors.
The construction of the M. smegmatis conditional mutant strains were performed as previously described (Portevin et al., 2004). Two fragments containing the acpS gene (ortholog of Rv2523c of M. tuberculosis H37Rv) and the pptT gene (ortholog of Rv2794c of M. tuberculosis H37Rv) flanked by their 5′- and 3′-ends were amplified by PCR from chromosomal DNA of M. smegmatis mc2155 and cloned into pGEM-T (Promega) to yield pG2523S and pG2794S. Each plasmid was linearized with a restriction enzyme which cut at a unique site within either the acpS or the pptT (BclI for pG2523S and MscI for pG2794S) and ligated with a Km resistance cassette to give plasmids pG2523SK and pG2764SK. The 4.76 kb NdeI-ApaI fragment of pG2523SK and the 4.44 kb NdeI-ApaI fragment of pG2794SK that contained the disrupted allele of acpS and pptT, respectively, were inserted between the SmaI and ApaI sites of pJQ200 (Quandt and Hymes, 1993), a mycobacterial suicide plasmid harboring the counterselectable marker sacB to yield pJ2523SK and pJ2794SK, respectively. These constructs were electrotransferred into M. smegmatis mc2155 and transformants were plated on LB plates supplemented with Km at 37° C. For each transformation, several transformants were selected and characterized by PCR using various primers. Two strains harbouring a pattern corresponding to the insertion of pJ2523SK and pJ2794SK in the chromosome as a result of single recombination event were named PMM68 and PMM70, respectively and used for further studies (
For the construction of the thermosensitive complementation plasmid pC-acpSms, the acpS gene of M. smegmatis mc2 155 was amplified by PCR and cloned into a derivative of the E. coli-mycobacteria shuttle plasmid pMIP12 (Le Dantec et al., 2001) harboring a NdeI site downstream of the pBlaF* promoter. The resulting construct was then digested with PacI and NheI and the 1.78 kb fragment overlapping the pBlaF* promoter, the acpS gene and the terminator region of pMIP12 was subsequently cloned into the vector pCG76 (Guilhot et al., 1994) which harbours a thermosensitive mycobacterial replicon and a streptomycin resistance to give pC-acpSms. The pptT gene of M. smegmatis, the acpS gene of C. glutamicum (NCgl2405) and the pptT gene of C. glutamicum (NCgl1905) were PCR-amplified from genomic DNA. Each PCR product was digested with NdeI and SpeI and ligated independently between the NdeI and SpeI restriction sites of pC-acpSms to yield pC-pptTms, pC-acpScg and pC-pptTcg respectively.
Vectors pC-acpSms and pC-acpScg were introduced in strain PMM68 and pC-pptTms and pC-pptTcg in strain PMM70 by electrotransformation. Transformants were selected on LB plates containing Km and Str at 30° C. For each transformation, one colony was resuspended in liquid media and grown at 30° C. before plating onto LB plates containing Km, Str and 5% sucrose at 30° C. to induced the second crossover event at either the acpS chromosomal locus (for PMM68 transformed with pC-acpSms or pC-acpScg) or the pptT chromosomal locus (for PMM70 transformed with pC-pptTms or pC-pptTcg). For each construct, several clones were screened by PCR using various primers (
Biochemical Characterization of Fatty Acids of C. glutamicum Strains
Cultures of C. glutamicum were grown to exponential growth and fatty acids were prepared from cells and separated by analytical TLC on Durasil 25 according to Laval et al (2001). For GC and GC-MS analyses, trimethylsilyl derivatives of fatty acids were obtained and analysed as described previously (Constant et al., 2002).
Labeling of the 4′-phosphopantetheinylated Proteins in C. glutamicum
The C. glutamicum WT, ΔpptT::km and Δpks13::km strains were grown to exponential growth phase in 5 ml of CGXII minimal medium (Keilhauer et al., 1993) supplemented with 0.05% Tween 80 at 30° C. 5 μl of β-[β-14C]alanine (49 mCi/mmol; Sigma) were then added to each culture and cells were incubated for an additional 12 hours. Bacterial cells were harvested by centrifugation, washed twice with Phosphate Buffered Saline (PBS) and resuspended in 500 μl of PBS. 500 μl of glass beads were added to each bacterial suspension and cells were disrupted by agitation for 3 min in a Mini BeadBeater. Cellular extracts were centrifuged at 12000 rpm for 10 min at 4° C. and proteins of the supernatants were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The polyacrylamide gel was stained with Coomassie blue, dried and the radiolabelled proteins were detected by exposure to Kodak X-Ray films.
Labeling of the Recombinant Mycobacterial Pks in E. coli.
Expression vectors pWM35, pWM35γ, pETMas, pETA, pETB, pETC and pETD were cotransferred with either pLSfp or pLSfpA into E. coli BL21ΔentD. For each transformation, one bacterial colony was inoculated in 2 ml LB medium supplemented with Km and Clm and incubated overnight at 37° C. 200 μl of the preculture was recovered and centrifugated for 2 min at 4000 rpm. Bacterial cells were then washed twice with 500 μl of M9 medium (Sambrook and Russell, 2001), diluted in ml of M9 medium with Km and Clm and grown at 30° C. When the optical density at 600 nm reached 0.5, 5 μl of 1M isopropyl-(3-D-thiogalactopyranoside (IPTG) and 5 μl of β-[β-14C]alanine were added to the culture and incubation was continued for 4 additional hours at 30° C. 1 ml of culture was then centrifuged, cells were washed with 1 ml of 50 mM Tris-HCl pH8.0 and finally resuspended in 100 μl of 50 mM Tris-HCl pH8.0 containing 0.1% triton X-100. For labeling analysis, 101 of cells suspension were incubated with 10 μl of 2× denaturing buffer (Sambrook and Russell, 2001) at 95° C. for 5 min before separation of proteins by SDS-PAGE. Polyacrylamide gels were stained by Coomassie blue before drying and exposition to autoradiography.
M. tuberculosis produces a number of proteins which must be converted from their inactive apo-forms to their functional holo-forms by transfer of the P-pant moiety of CoA to a conserved serine on their ACPs or PCPs domains/or subunit. These include proteins AcpM and Fas-I involved in fatty acids biosynthesis but also a number of type-I Pks and NRPS proteins that are required for the biosynthesis of mycolic acids, cell wall associated lipids and siderophores. Two genes encoding proteins with similarities to PPTases were previously identified on the genome of M. tuberculosis H37Rv: Rv2523c, also called acpS because of the similarities of the encoded 130 amino acids protein with AcpS of various bacterial origin, and Rv2794c, which was previously shown to encode a Sfp-type PPTase of 227 amino acids, named PptT, responsible for the modification of a set of NRPS involved in the biosynthesis of mycobactins (Quadri et al., 1998b). In order to look for additional PPTase genes, we searched the M. tuberculosis H37Rv genome with conserved motifs of PPTases as probes (Weissman et al., 2004). No novel gene was identified suggesting that M. tuberculosis contains just two PPTases responsible for the activation of 20 protein substrates.
We then addressed the question of whether these two PPTase genes were conserved in the various mycobacterial species and in the closely related corynebacterial species. In all the analysed genome (including M. tuberculosis strains H37Rv and CDC151, M. bovis, M. bovis BCG, M. leprae, M. avium, M. smegmatis, M. microti, M. marinum, C. glutamicum, C. diptheriae and C. efficiens), we found orthologs of acpS and pptT. The levels of proteins similarities were more than 80% for the mycobacterial proteins and around 40% for the corynebacterial proteins with their M. tuberculosis counterparts.
Therefore these bioinformatic analyses showed that the related mycobacterial and corynebacterial species contain only two highly conserved PPTases with different catalytic properties. However, none of these two proteins had been characterized in mycobacteria or corynebacteria.
In mycobacteria, there is a large number of protein which needed to be activated for being functional: for instance, two Fas systems, at least three NRPS and eighteen type-I Pks in M. tuberculosis. However, no information was available both on the repertoire of protein substrates recognized by each PPTases and on the putative redundancy of these two proteins. As several of the proteins requiring activation are involved in biological processes essential for the mycobacterial viability, we first addressed the question whether any of the two PPTase genes can be disrupted in the model mycobacterial strain M. smegmatis. Two non replicative plasmids carrying the counterselectable marker sacB and a disrupted allele of either acpS or pptT of M. smegmatis were constructed and inserted into the chromosome by single crossover to give strains PMM68 and PMM70, respectively (
To investigate the role of acpS and pptT in mycobacterial growth, cultures of recombinant strains PMM77, PMM78 and the WT strain of M. smegmatis were grown at 30° C. and streaked on LB agar plates. When plates were incubated at 30° C., PMM77 and PMM78 strains grew as well as the WT. However, when plates were incubated at 42° C., a nonpermissive temperature for plasmid replication, both recombinant strains exhibited growth inhibition in contrast to the WT (
Among the PPTase substrates, those responsible for the synthesis of fatty acids and mycolic acids are known to be essential for growth of M. smegmatis in laboratory conditions. We thus anticipated that both PPTases were involved in at least one of these metabolic pathways. Therefore, to study the role played by AcpS and PptT in fatty acids and mycolic acids biosynthesis, we switched to C. glutamicum, another bacterial model. Indeed, this bacterial species exhibits a cell envelope closely related to the one of mycobacteria, with a similar cell wall core (Daffé, 2005). However, we demonstrated in a previous work that this strain is more tolerant than the mycobacterial strain to mutation in genes involved in lipid metabolism (Portevin et al., 2004). To investigate the role of the two PPTases, we attempted to knockout the corresponding genes in C. glutamicum. Two alleles of acpS and pptT, disrupted by a Km resistant gene, were transferred on a non-replicative plasmid in C. glutamicum and transformants were selected on Km. For pptT, several Km resistant clones were screened by PCR and one clone that gave an amplification pattern consistent with the insertion of the Km cassette within pptT was retained for further analysis. Interestingly this ΔpptT::km mutant displayed phenotypic changes similar to those observed for the C. glutamicum Δpks13 mutant which lacks an essential enzyme for the production of corynomycolates (Portevin et al., 2004). On BHI agar plates, the ΔpptT::km mutant exhibited small and rough colonies instead of large and shiny colonies for the WT. In addition this mutant grew much slower than the WT in liquid media and cells aggregated strongly whereas, under the same conditions WT cells grew well dispersed (data not shown). These phenotypic modifications were completely reverted when the mutant strain was transformed with plasmid pCGL-pptTcg which contains a functional copy of pptT from C. glutamicum, indicating that the observed phenotype of the mutant strain were due to the deletion of the pptT gene.
For acpS, the same strategy gave ΔacpS::km mutants, only when the growth medium was supplemented with fatty acids. A phenotypical analysis confirmed that these C. glutamicum ΔacpS::km mutants were unable to grow on solid or liquid BHI medium except when medium were supplemented with sodium oleate and Tween 40. Under these conditions, no phenotypical differences were observed between the WT and ΔacpS::km recombinant strain (data not shown). This auxotrophy for oleic acid suggested that disruption of acpS in C. glutamicum affects the biosynthesis of fatty acids. The transfer of pCGL-acpScg, a corynebacterial plasmid carrying a functional acpS gene from C. glutamicum, into the ΔacpS::km mutant strain restored the capacity for the bacteria cells to grow on solid or liquid medium without oleate indicating that the phenotypic changes observed between the mutated and the WT strains relied solely on the disruption of the acpS gene.
Altogether these experiments showed that disruption of acpS and pptT in C. glutamicum led to strains exhibiting different phenotypes. The oleic acid auxotrophy of the ΔacpS::km mutant suggested that acpS is involved in fatty acids biosynthesis. In contrast the pptT gene was not required for the viability of C. glutamicum but the morphological alterations of the ΔpptT::km mutant indicated cell envelope modifications consistent with absence of mycolate production.
To further characterize the phenotypes induced by the disruption of acpS and pptT, we carried out biochemical characterization of fatty acids and corynomycolates produced by the two C. glutamicum recombinant strains.
Cultures of C. glutamicum WT, ΔacpS::km mutant and ΔacpS::km:pCGL-acpScg complemented strain were grown to exponential phase (in presence of oleic acid and Tween 40 for the mutant strain) and fatty acids were released from bacteria by saponification of whole cells. Thin layer chromatography (TLC) analysis showed that the ΔacpS::km mutant synthesized corynomycolates at the same level than the WT or the complemented strain (
Unlike mycobacteria which possess two fatty acid synthase systems (Fas-I and Fas-II), C. glutamicum lacks the Fas-II system but has two Fas-I proteins (Fas-IA and Fas-IB). It was shown that Fas-IA is essential for growth but not Fas-IB (Radmacher et al., 2005). The fact that disruption of acpS in C. glutamicum affects the biosynthesis of fatty acids means that the Fas-IA enzyme is inactive in the ΔacpS::km mutant strongly suggesting that AcpS is responsible for the 4′-phosphopantetheinylation of this enzyme. These results also established that PptT cannot complement the lack of AcpS to perform this reaction in C. glutamicum providing a strong support to the model that the two PPTases are not redundant. Moreover the ability for the recombinant strain to sustain corynomycolates synthesis in presence of exogenous fatty acids implies that the AcpS PPTase is not required for the posttranslational modification of the condensing enzyme Pks13.
We next investigated whether phenotypic changes observed with the ΔpptT::km mutant reflect cell wall modifications. TLC and GC analyses revealed that disruption of pptT in C. glutamicum abolished the production of corynomycolates (
Taken together, these experiments established that the AcpS and PptT display different functions in C. glutamicum, the former being involved in the Fas-I activation and the latter in the modification of Pks 13.
The phenotypic and biochemical analysis of the two C. glutamicum mutants provide indirect evidences that AcpS and PptT activate Fas-I and Pks13 respectively. In order to directly demonstrate the specificity of the two PPTases, we designed an experiment to label the P-pant arm and to visualize its transfer onto the various protein substrates. The various C. glutamicum WT and recombinant strains were grown on CGXII minimal medium in presence of β-[β-14C]alanine, a precursor of CoA. The principle of this experiment was to generate a pool of CoA harbouring a 14C radiolabeled 4′-phosphopantetheine prosthetic group within the bacteria. This pool may then serve as substrate for the two PPTases to specifically label the 4′-phosphopantetheinylated proteins. After growth, cells were harvested, lyzed and proteins from cellular extracts were separated by SDS-PAGE before Coomassie blue staining and autoradiography to visualize the radiolabelled proteins.
When this experiment was performed with the WT C. glutamicum, two protein bands were labeled exhibiting apparent molecular weight of 170 kDa and higher than 220 kDa (
When the same experiments was performed with ΔpptT strain, Pks13 was no longer labelled (
Therefore, these experiments established for the first time the repertoire of protein substrate for the two PPTases of C. glutamicum: AcpS and PptT have a strict substrate specificity for the Fas-I enzymes and Pks 13 respectively.
The above-described experiments established the function played by each PPTase in corynebacteria cells. To study if these proteins display similar functions in mycobacteria, we tested whether the disruption of either the pptT or the acpS gene in M. smegmatis could be complemented by the expression of the corresponding orthologs of C. glutamicum. To address this issue, two mycobacterial temperature-sensitive plasmids, named pC-acpScg and pC-pptTcg, harbouring the acpS and the pptT gene of C. glutamicum, respectively, were constructed. pC-acpScg was transferred in M. smegmatis PMM68 which is merodiploid for the acpS gene and pC-pptTcg in M. smegmatis PMM70, the merodiploid for the pptT gene. Several transformants were grown in LB at 30° C. before plating on solid medium containing sucrose to induce the second recombination event between the two chromosomal alleles at either the pptT or the acpS loci. Several colonies, in which the WT chromosomal copy of acpS or pptT were replaced by the mutated allele, were selected for further analyses and named PMM84 (ΔacpS::km:pC-acpScg) and PMM85 (ΔpptT::km:pC-pptTcg).
PMM84 and PMM85 exhibited WT phenotype identical to the one obtained with mutants PMM77 (ΔacpS::km:pC-acpSms) and PMM78 (ΔpptT::km:pC-pptTms) which expressed AcpS and PptT of M. smegmatis, respectively. Both mutants strains sustained normal growth in LB liquid medium and on LB agar plates at 30° C. but were unable to support growth on solid medium at 42° C., a non permissive temperature for plasmid replication (
A similar cross-complementation was observed when the PptT of M. smegmatis was expressed in the C. glutamicum ΔpptT::km. In contrast to C. glutamicum ΔpptT::km cells, the strain complemented with pptT from M. smegmatis showed no aggregation in liquid medium and grew as smooth shiny colonies on agar plates. In addition, analysis of fatty acids by TLC and GC after saponification of whole cells revealed that the production of corynomycolates had been restored at the same level than in the WT strain (
We demonstrated that AcpS is involved in the activation of Fas-I in corynebacteria and mycobacteria whereas PptT catalyzes the posttranslational modification of Pks13. This protein is the only type-I Pks encoded by the corynebacterial genome. However, the repertoire of type-I Pks is much larger in mycobacteria. These additional proteins are involved in the formation of extractable lipids, such as DIM or PGL-tb in M. tuberculosis, which are key virulence factors. We then wonder whether PptT is also involved in the activation of the other type-I Pks of M. tuberculosis.
To address this question, various M. tuberculosis Pks were coproduced in E. coli with PptT of M. tuberculosis and we looked at the activation of these proteins using radiolabeling with β-[β-14C]alanine. In primarily experiments, we observed that the E. coli PPTase EntD was responsible for partial activation of some M. tuberculosis Pks (our unpublished data). To overcome this problem, we constructed an E. coli BL21(DE3) entD-disrupted strain by allelic exchange and carried out the same experiments in the new strain. The experiment was first performed with Pks13 because we already knew that it is a substrate of PptT (see results above). Results revealed that Pks13 was labelled in cells coproducing Pks13 and PptT from M. tuberculosis but not in cells producing Pks13 alone (
Once the test functional, we then looked at the activation of other type-I Pks by PptT. Five different type-I Pks were independently coproduced with PptT in E. coli BL21 ΔentD and in all cases, transfer of the radiolabeled P-pant was detected (
According to this, it can be concluded that PptT is not only responsible for the activation of Pks13 in M. tuberculosis but is also required for the modification of other type-I Pks involved in the biosynthesis of lipids required for virulence.
Discussion
Given the crucial role of cell envelope lipids in the biology of M. tuberculosis, tremendous efforts have been made during the last decades to decipher the cellular processes leading to the production and translocation of these components. From these efforts, the concept has emerged that the unique lipid structures found in mycobacteria are synthesized by the combined action of Fas systems and type-I Pks: both classes of enzymes which have to be converted from inactive apo-forms to functional holo-forms. In this study, we have examined the role of two PPTases in the posttranslational modification of these biosynthetic enzymes. We provided direct evidence that the two PPTases activate each a defined subset of protein substrates in mycobacteria and corynebacteria and are both essential for the viability of mycobacteria. These results have important implications for our understanding of the lipids metabolism in mycobacteria and related bacteria. They demonstrate the central role played by the two PPTases in the biology of these microorganisms, defining new promising drug targets for fighting tuberculosis.
Concerning the lipid metabolism in mycobacteria, we propose a model in which AcpS is dedicated to the posttranslational modification of Fas-I and the AcpM subunit of Fas-II whereas PptT activates the numerous type-I Pks and NRPS of M. tuberculosis (
Our results have also important implications for our understanding of the biology of mycobacteria and especially M. tuberculosis. Indeed, M. tuberculosis contains more than 20 proteins which have to be 4′-phosphopantetheinylated. Ours findings provided a definition of the substrate repertoire of each PPTase and showed that both enzymes are required for the formation of essential components for the viability of mycobacteria. Hence, the enzyme, Fas-I, which is activated by AcpS, catalyses the synthesis of C16-C18 fatty acids which are incorporated in the various lipid constituents of the plasma membrane. Synthesis of short fatty acids by Fas-I is also one of the first steps of the long biosynthesis pathway leading to the formation of mycolates, which are key structural elements of the mycobacterial cell wall skeleton (Daffé and Draper, 1998). This pathway includes also two other proteins AcpM and Pks13 activated by AcpS and PptT, respectively: AcpM is a subunit of the Fas-II system synthesizing the long meromycolate chain and Pks13 is the condensase catalysing the last condensation step of mycolate formation. Therefore, the two mycobacterial PPTases are required for mycolates formation. As a consequence, they are both essential for viability of mycobacteria.
In addition, these two enzymes are also required for the production of important virulence factors. For instance, the enzymes MbtB and MbtD-F activated by PptT are involved in the assembly of mycobactin siderophores that are required for growth within human macrophages (de Voss et al., 2000). Along the same line, Fas-I and seven type-I Pks, activated by AcpS and PptT respectively, are involved in the formation of DIM and PGL-tb, two complex lipids produced by a very limited number of mycobacterial species and two important virulence factors of M. tuberculosis. Indeed, DIM-less mutants are affected in their capacity to multiply within the host and to cause diseases (Cox et al., 1999). Similarly, a clinical isolates of M. tuberculosis exhibiting a hypervirulence phenotype in various animal model was shown to be attenuated by mutation in Pks15/1, an enzyme required for PGL-tb formation (Reed et al., 2004).
Thus the two PPTases appear to have central roles for the biology of the pathogen M. tuberculosis being required both for viability and pathogenesis. Other mycobacterial pathogens such as M. leprae, M. ulcerans or M. avium also possess orthologs of AcpS and PptT and all produce mycolates and lipid virulence factors. Therefore, the two mycobacterial PPTases are very promising targets for the development of drugs for fighting mycobacterial infections.
The PPTase activity consists in the transfer of the 4′-phosphopantetheine group (P-pant) from coenzyme A onto the acyl carrier protein (ACP) domain of Pks. The PptT activity is assayed by using a radioactive assay method as described previously for various PPTases including Sfp from Bacillus subtilis or AcpS from E. coli (Lambalot, Gehring et al. 1996; Quadri, Weinreb et al. 1998a; Mootz, Finking et al. 2001) and is adapted to the mycobacterial PptT PPTase. This method measures the incorporation of the 3H-labeled 4′ phosphopantetheine group from (3H) coenzyme A into apoenzymes.
Typically, reaction mixtures containing MgCl2, the PptT protein, the protein substrate (apoenzyme) and the (3H) CoenzymeA cosubstrate is incubated for 30 min at 37° C. Reactions are quenched by addition of 10% trichloroacetic acid (TCA) and BSA added as a carrier. Precipitated proteins are collected by centrifugation and the resulting pellets are washed with TCA and dissolved in 1M Tris base. The redissolved proteins are mixed with liquid scintillation cocktail and the amount of radioactivity incorporated into the protein substrate quantified using a liquid scintillation analyzer.
The PptT protein used in these assays is fused to the MBP (Maltose Binding Protein) protein at its N-terminal extremity. This MBP-PptT fusion protein is overexpressed in E. coli strain BL21 and partially purified by affinity chromatography using an amylose resin.
Different protein substrates (apoenzymes) consisting of either whole mycobacterial Pks proteins or ACP domains of these Pks are tested in order to select the most suitable substrate for the test. As examples of Pks proteins that can be used in this test, Pks13 and Mas (mycocerosic acid synthase, which is implicated in the synthesis of mycocerosic acids in mycobacteria producing DIMs and PGLs) can be cited. They can be produced fused to a carboxy-terminal His tag in a engineered E. coli strain deleted for the entD gene which encodes an E. coli PPTase to avoid background activation (production of apo forms). Such His tagged-proteins are purified by affinity chromatography (Nickel column) followed by size exclusion chromatography.
Once the test is functional, the (3H) CoenzymeA is substituted by CoenzymeA analogs harboring either a fluorescent or a biotinylated phosphopantetheine group to design an enzymatic assay suitable for high throughput screening of libraries of compounds. Several studies have demonstrated that these modified substrates may be efficiently transferred on various protein substrates by Sfp (La Clair, Foley et al. 2004; Yin, Liu et al. 2004).
| Number | Date | Country | Kind |
|---|---|---|---|
| 05292610.2 | Dec 2005 | EP | regional |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/IB2006/004075 | 12/8/2006 | WO | 00 | 11/10/2008 |
