SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Mycobacteria have a unique outer membrane (OM) that is thicker than any other known biological membrane. Nutrients cross this permeability barrier by diffusion through porins. MspA is the major porin of Mycobacterium smegmatis. In this study we showed that three paralogues of MspA, namely MspB, MspC and MspD are also porins. However, only the mspA and mspC genes were expressed in the wild-type strain. None of the single deletion mutants displayed a significant OM permeability defect except for the mspA mutant. Deletion of the mspA gene caused activation of transcription of mspB and/or mspD in three independent strains by unknown chromosomal mutations. It is concluded that mspB and mspD provide backup porins for M. smegmatis. This also indicated that a minimal porin-mediated OM permeability is essential for survival of M. smegmatis. Electron microscopy in combination with quantitative image analysis of protein gels revealed that the number of pores per cell dropped from 2400 to 800 and 150 for the ΔmspA and ΔmspA ΔmspC mutant (ML10) respectively. The very low number of pores correlated well with the at least 20-fold lower channel activity of detergent extracts of the ML10 strain and its 15- and 75-fold lower permeability to nutrient molecules such as serine and glucose respectively. The amount of Msp porin and the OM permeability of the triple porin mutant lacking mspA, mspC and mspD was not altered. The growth rate of M. smegmatis dropped drastically with its porin-mediated OM permeability in contrast to porin mutants of Escherichia coli. These results show that porin-mediated influx of nutrients is a major determinant of the growth rate of M. smegmatis.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Many mycobacteria are ubiquitously found in soil and water, whereas a few members of this genus are major human and animal pathogens (Rastogi et al., 2001). Scientific interest in mycobacteria has not only been sparked by the paramount medical importance of Mycobacterium tuberculosis, but also by properties that distinguish mycobacteria from other bacteria. Their most prominent feature is their unique cell wall, which comprises an outer membrane (OM) thicker than any other known biological membrane (Paul and Beveridge, 1992). This OM plays a crucial role in extra- and intracellular survival and growth of mycobacteria under harsh environmental conditions or in drug resistance (Brennan and Nikaido, 1995; Barry, 2001). In addition, mycobacteria are known for their notoriously slow growth with doubling times between 2 and 6 h for saprophytes such as M. smegmatis and M. chelonae and about 24 h for the strictly pathogenic members of the M. tuberculosis complex (Cox, 2003). The record of bacterial slow growth is set by M. leprae, which lost nearly half of its metabolic functions by reductive evolution (Cole et al., 2001) and doubles approximately every 14 days. It is unknown whether slow growth is beneficial or even necessary for pathogenicity of mycobacteria and ‘why M. tuberculosis requires 24 h to divide when M. smegmatis only takes 2 h?’ (Jacobs, 2000). Thus, it is not surprising that many factors have been invoked to explain the slow growth of M. tuberculosis: (i) slow RNA synthesis (Harshey and Ramakrishnan, 1977); (ii) slow DNA elongation (Hiriyanna and Ramakrishnan, 1986) maybe attributed to the presence of the DNA-binding protein MDBP1 (Matsumoto et al., 2000); (iii) slow protein synthesis attributed to the lack of multiple copies of rRNA operons (Bercovier et al., 1986); and (iv) slow porin-mediated uptake of nutrients (Jarlier and Nikaido, 1990). So far, there is little evidence demonstrating that any of these factors really limits the growth rate of mycobacteria. Recently, it was proposed based on comparison of macromolecular compositions that M. tuberculosis should be capable of growing much faster than it actually does (Cox, 2004). We reported a small but significant acceleration of both glucose uptake and growth of the slow-growing M. bovis BCG in liquid culture upon expression of the porin MspA of M. smegmatis (Mailaender et al., 2004). The growth-promoting effect of MspA on M. bovis BCG was surprising because MspA levels were 40-fold lower compared with M. smegmatis indicating that the endogenous OM permeability might be much lower in M. bovis BCG compared with M. smegmatis. This also indicated that nutrient uptake across the OM might be one of the factors determining the generation time of slowly growing mycobacteria.

Mycobacterium smegmatis has a 100- to 1000-fold lower permeability for hydrophilic solutes than Escherichia coli (Jarlier and Nikaido, 1990; Trias and Benz, 1994; Chambers et al., 1995). This is at least partially attributed to the 45-fold lower number of porins as determined by electron microscopy (Engelhardt et al., 2002). Among those porins, MspA accounts for at least 75% of the permeability of M. smegmatis for glucose (Stahl et al., 2001). Loss of MspA increased resistance of M. smegmatis to hydrophilic β-lactam antibiotics 8- to 16-fold consistent with its prominent role for OM permeability (Stephan et al., 2004a). The structure of MspA is completely different from that of any other known porin: eight monomers form a goblet-like protein with a single central channel of 10 nm length (Faller et al., 2004). If diffusion of hydrophilic solutes through porins is really a determinant of the growth rate of mycobacteria, a lower number of porins should result in slower growth. However, this was not observed for the mspA mutant of M. smegmatis. The interpretation of the properties of the mspA mutant might have been complicated by the presence of three other potential porins of M. smegmatis, MspB, C and D, which are almost identical to MspA (2, 4 and 18 different amino acids out of 184) and might partially compensate for the loss of MspA (Stahl et al., 2001). However, nothing is known about the expression and the physiological functions of these proteins.

In this study, we examined the expression of the four known msp genes in M. smegmatis and analysed the changes of their expression profile in porin mutants. We further determined the number of porins, the rate of glucose uptake in single Δmsp mutants, in a ΔmspAΔmspC double mutant and a ΔmspAΔmspCΔmspD triple mutant. The latter two mutant strains showed a drastically reduced porin-mediated OM permeability and a much slower growth rate in rich and minimal media. The generation time increased drastically with decreasing number of porins demonstrating that porin-mediated nutrient uptake is a major determinant of the growth rate of M. smegmatis.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

MspB, MspC and MspD are porins of M. smegmatis

Previous work showed that mspA is the major porin gene of M. smegmatis and is important for glucose uptake (Stahl et al., 2001). Nothing is known about the physiological functions of the three putative porin genes mspB, mspC and mspD, which are located at positions different from mspA in the chromosome of M. smegmatis(Fig. 1). To analyse whether these genes have porin functions expression vectors were constructed using the strong mycobacterial promoter psmyc (Kaps et al., 2001). All msp genes were fused to the same mycobacterial consensus ribosome binding site (RBS) (5′-AAGGAGG-3′) to achieve similar expression levels of the corresponding proteins. This RBS was tested in fusions with gfp and gave rise to more fluorescence in M. smegmatis than the RBS of the gfp+ gene (C. Detsch and M. Niederweis, unpublished). These msp expression vectors (pMN041–pMN044, Table S1) restored cephaloridine uptake in the ΔmspA mutant of M. smegmatis to half of the wild-type level (not shown). Another transport experiment was based on the observation that the PhoA activity depended on the level of porin expression and was then used as a measure of OM permeability of M. smegmatis for the PhoA substrate p-nitrophenylphosphate (PNPP) (will be published elsewhere). All four msp porin genes complemented the permeability defects of the double porin mutant M. smegmatis ML10 for uptake of PNPP. Uptake of PNPP mediated by MspA and MspB was fourfold above background and dropped to three- and two-fold levels for MspC and MspD respectively. These results demonstrated that all Msp proteins have porin function in M. smegmatis.

image

Figure 1. Location of porin genes in the chromosome of M. smegmatis. The genome of M. smegmatis mc2155 was sequenced and annotated by The Institute for Genomic Research (http://www.tigr.org). The arrows indicate the length and transcriptional orientation of annotated genes and predicted open reading frames (ORF). The msp genes are highlighted in red, genes with homologues in other bacteria in blue and open reading frames in grey. Note that the genes are only denominated by their number. The complete names as assigned by TIGR include the prefix ‘msmeg’. Genes were given names instead of numbers by TIGR, if putative functions were assigned to ORFs with high confidence. Putative functions of genes in proximity of the msp genes are: mspA locus: msmeg0951 (transcriptional regulator of the TetR family), msmeg0958 (cytochrome P450), hemL (glutamate-1-semialdehyde-2,1-aminomutase); mspB locus: msmeg0505 (periplasmic sugar binding protein), msmeg0506 (ABC transporter), ugpE (ABC transporter), sugC (ABC transporter), msmeg0514 (transcriptional regulator); mspC locus: mscL (large conductance mechanosensitive channel), mog (molybdopterin biosynthesis protein), msmeg5466 (putative heat shock protein HtrA), msmeg5467 (sensor histidine kinase), msmeg5468 (DNA-binding response regulator); mspD locus: msmeg6015 (ABC transporter), msmeg6017 + msmeg6018 (transposase, IS1547), msmeg6020 (cation-transporting ATPase CtpC, E1-E2 family) and msmeg6022 (ABC transporter).

Download figure to PowerPoint

The porin genes mspA and mspC are expressed in wild-type M. smegmatis

To examine which of the msp porin genes are expressed in M. smegmatis, total RNA was prepared from the wild-type strain grown in standard 7H9 medium containing 0.2% glycerol. The RNA was reversely transcribed and the cDNA was amplified with primers specific for the four msp porin genes. DNA fragments of the correct size were obtained only for the mspA and mspC genes. No DNA was detected when the reverse transcription step was omitted (Fig. 2). This showed that the mspA and mspC, but not the mspB and mspD genes are expressed in wild-type M. smegmatis under those conditions. It should be noted that the reverse transcription polymerase chain reaction (RT-PCR) data cannot be interpreted quantitatively, as the efficiency of DNA amplification of the four different primer pairs (Table S2) is different and the final yield of DNA is not proportional to the initial amount of template anymore after 30 cycles.

image

Figure 2. Expression of porin genes in M. smegmatis wild-type and porin mutants. In total, 100 ng of total RNA of M. smegmatis strains SMR5 (wt), MN01 (ΔmspA) and ML10 (ΔmspAΔmspC) was used to generate cDNAs specific for the porin genes mspA (A), mspB (B), mspC (C) and mspD (D) by reverse transcription (RT), which was then amplified by PCR. The ‘+’ sign denotes samples, in which the RNA was added before the RT step, whereas the ‘–’ sign denotes samples, in which the RNA was added after the RT step to detect contaminations with chromosomal DNA. One-fifth of each PCR sample was loaded on a 1% agarose gel. The gels were stained with ethidium bromide and are shown as negative images to enhance the visibility of weak bands.

Download figure to PowerPoint

Construction of porin gene deletion mutants of M. smegmatis

To examine the functions of the porin genes mspB, mspC and mspD, we wanted to analyse the OM permeability of mutants lacking these genes. In addition, the msp porin genes should be deleted consecutively in one strain to assess the importance of the porin-mediated OM permeability for M. smegmatis in general. However, four efficient resistance genes are currently not available for mycobacteria. Therefore, the msp porin genes should be replaced by the recently constructed FRT-hyg-FRT expression cassette (Stephan et al., 2004b). Expression of the FLP recombinase will then specifically remove the hyg gene from the chromosome generating a marker-free deletion mutant and enable the re-use of the hyg gene as resistance marker. Constructions of the unmarked ΔmspA (ML02) and ΔmspC (ML06) mutants and the ΔmspAΔmspC mutant (ML10) were described by Stephan et al. (2004b) in detail (Table S3). The ΔmspB (ML04) and ΔmspD (ML08) mutants were constructed in a similar manner to examine whether the single mutants would reveal any phenotype (Table S3). Southern blot analysis of the mspB locus demonstrated the allelic exchange in the strain ML03 (ΔmspB::hyg) and the excision of the hyg gene from the chromosome of strain ML04 (ΔmspB) (Fig. 3A). Southern blot analysis of the mspD locus showed the allelic exchange in the strain ML07 (ΔmspD::hyg) and the excision of the hyg gene from the chromosome of strain ML08 (ΔmspD) (Fig. 3B). Thus, unmarked single deletion mutants for each of the known porin genes of M. smegmatis are now available.

image

Figure 3. Construction of unmarked porin gene deletion mutants of M. smegmatis. A. Southern blot analysis of the mspB locus using AatII-digested chromosomal DNAs of the strains SMR5 (wt), ML03 (ΔmspB::hyg) and ML04 (ΔmspB) revealed three fragments of 1140 bp, 1450 bp and of more than 1850 bp, respectively, consistent with the DNA sequence analysis. B. Deletion of the porin gene mspD. Southern blot analysis of the mspD locus using BspEI-digested chromosomal DNAs of the strains SMR5 (wt), ML07 (ΔmspD::hyg) and ML08 (ΔmspD) revealed three fragments of 1730, 630 and 1090 bp, respectively, consistent with the DNA sequence analysis. C. Construction of the triple porin gene mutant. Southern blot analysis of the mspD locus using SacII-digested chromosomal DNAs of the strains ML10 (ΔmspA,ΔmspC), ML13 (ΔmspA, ΔmspC, ΔmspD::pMN254, attB::pwmycmspA) and ML14 (ΔmspA,ΔmspC,ΔmspD::hyg, attB::pwmycmspA). In all experiments, the digested chromosomal DNA was separated on a 1% agarose gel, blotted onto nitrocellulose membrane and detected using probes labelled with digoxigenin, a secondary antibody coupled to an alkaline-phosphatase and the colour substrate nitroblue tetrazolium chloride (NBT)/5-bromo-4-chloro-3-indolyl phosphate (BCIP).

Download figure to PowerPoint

Various attempts to delete the mspB or mspD gene in the porin double mutant ML10 using the allelic exchange vectors that were used to construct the single gene deletion mutants, did not yield a triple knockout mutant. This may indicate that the number of porins in the ML10 mutant could not be further reduced without impairing the growth of M. smegmatis. To circumvent this problem the plasmid pML502 was integrated at the attB site of ML10 to yield the strain ML11 (Table S3). This plasmid provided an mspA expression cassette, which significantly increased the expression of Msp porins by the ML10 strain as observed by Western blot experiments (not shown). After removal of the FRT-flanked hygromycin cassette of the integrated plasmid by the Flp recombinase (ML12, Table S3), the mspD gene was replaced by a FRT-hyg-FRT cassette in two consecutive homologous recombinations using the suicide mspD deletion vector pMN254. Southern blot analysis of the mspD locus showed the single cross-over of the plasmid into the chromosome in the strain ML13 (ΔmspD::pMN254) and the second cross-over resulting in the loss of the plasmid from the chromosome of strain ML14 (Fig. 3C). Then, the hyg cassette was removed from the mspD locus by the Flp recombinase to yield the strain ML15 (Table S3). Surprisingly, the mspA gene could also be removed from the attB site by the Cre recombinase without any deleterious effect for the triple porin deletion mutant ML16 (Table S3). This indicated that expression of mspD was either very low or that loss of mspD was compensated for by increased expression of mspB or another porin gene. All strains were verified by analysis of the chromosomal DNA of the corresponding strains by PCR and Southern blots (not shown).

Expression of silent porin genes is activated upon deletion of the mspA gene

Reverse transcription polymerase chain reaction experiments were done with all porin mutants of M. smegmatis to confirm the absence of corresponding mRNAs. The ΔmspA mutant MN01 does not express the mspA gene anymore (Fig. 2A) as shown previously (Stahl et al., 2001), but expression of the two silent porin genes mspB and mspD is turned on upon deletion of the mspA gene (Fig. 2B and D). Thus, the ΔmspA mutant MN01 expresses three porin genes, whereas wild-type M. smegmatis expresses only two. Activation of transcription of mspB and mspD was also observed, after the mspA gene was deleted in the ΔmspC mutant to construct the double porin mutant ML10 (Fig. 2B and D). Transcription of the mspB but not of the mspC gene was activated in the newly constructed ΔmspA mutant ML02 (not shown). In conclusion, we observed activation of one or two silent porin genes after deletion of the mspA gene in three independent strains. This result underlined the importance of mspA for OM permeability and growth of M. smegmatis. To examine whether activation of transcription was caused by regulatory events following the reduction of OM permeability after loss of mspA, the ML02 strain was transformed with the mspA expression plasmid pMN014. RT-PCR showed that pMN014 restored mspA expression again, which, however, did not shut down mspB expression (not shown). This result suggested permanent changes in these mutants such as promoter-up mutations as the mechanism of activation of the silent porin genes. However, sequencing of approximately 400 bp upstream of the mspB and mspD genes after amplification of the chromosomal DNA of the strains SMR5, MN01 and ML02 did not reveal any differences.

Number of porins in the OM of porin mutants of M. smegmatis

It is essential to determine the number and the types of porins in the OM to be able to interpret results of permeability measurements of porin mutants on a reliable basis. To provide reference data, we relied on electron microscopy, which proved to be an easy method to directly count the number of pores in cell walls of M. smegmatis by negative staining with uranyl acetate. Earlier experiments revealed that cell wall fragments of wild-type M. smegmatis contained approximately 1000 randomly distributed pores per µm2 cell wall (Engelhardt et al., 2002). The number of visible pores was clearly reduced in the ΔmspA mutant MN01 resulting in an average of 333 ± 86 pores per µm2 for nine cell wall fragments (Fig. 4B). Whole cells of wild-type M. smegmatis and all porin mutants were selectively extracted in a buffer containing 0.5% isotridecylpolyethyleneglycolether (Genapol) at high temperatures as previously described (Heinz et al., 2003a). The Msp porins were the only proteins in these extracts observed in Coomassie-stained gels (Fig. 4C). The amount of extractable porins in all single porin mutants was determined relative to the wild-type by quantitative image analysis of both a Coomassie-stained protein gel and of a Western blot using the anti-MspA antiserum pAK MspA♯813 and calibration curves obtained with known amounts of MspA. Deletion of the mspB, mspC or mspD gene did not cause a significant reduction in staining intensity indicating that the number of extractable porins was not reduced in these mutants compared with wild-type M. smegmatis (not shown). By contrast, the amount of Msp porins in n-octyl polyethylene oxide (OPOE) extracts was reduced to 20% of the wild-type level in Coomassie-stained protein gels for the ΔmspA mutant and not detectable anymore for the double (Fig. 4C) and triple porin mutant (not shown). This is similar to the threefold reduced number of pores observed in cell-wall fragments of the ΔmspA mutant by electron microscopy. Quantitative image analysis of Western blots of OPOE extracts showed an approximately fivefold reduction of Msp porins extracts for both the ML10 and the ML16 strain compared with the ΔmspA mutant (Fig. 4D). Hence, taking the electron microscopy data as the more precise reference, the number of pores in the ΔmspAΔmspC mutant is approximately 60 MspB and MspD pores per µm2 cell wall. This corresponds to a 15-fold reduced number of porins compared with wild-type M. smegmatis. Surprisingly, the amount of porins in the triple mutant ML16 was similar to that of the double mutant ML10 (Fig. 4D) indicating that either mspD expression contributed little to the total amount of Msp porins in ML10 or that expression of the remaining porin gene mspB was upregulated to compensate for the loss of MspD.

image

Figure 4. Expression levels of porins in M. smegmatis and porin mutants. Electron microscopy of an isolated cell-wall fragments of M. smegmatis SMR5 (wild-type) (A) and MN01 (ΔmspA) (B) negatively stained with uranyl acetate. Cell wall pores are stain-filled and appear as black dots surrounded by a bright ring indicating the pore protein. Scale bar represents 100 nm. C. Equal amounts (10 µl) of whole-cell extracts from M. smegmatis SMR5 (wt), MN01 (ΔmspA) and ML10 (ΔmspAΔmspC) were separated in a 10% SDS-polyacrylamide gel, which was stained with Coomassie brilliant blue. Purified MspA was loaded as indicated to estimate the amount of porin in these samples. D. Equal amounts (1 µl) of whole-cell extracts from M. smegmatis SMR5 (wt), MN01 (ΔmspA), ML10 (ΔmspAΔmspC) and ML16 (ΔmspAΔmspCΔmspD) were separated in a 10% SDS-polyacrylamide gel. The samples were blotted on a nitrocellulose membrane and detected with the anti-MspA serum pAK MspA♯813, which cross-reacts with other Msp porins (Stahl et al., 2001).

Download figure to PowerPoint

The channel activity of extracts from theΔmspA ΔmspC mutant was drastically lower compared with wild-type M. smegmatis

Porins were selectively extracted from whole cells of wild-type M. smegmatis and the double porin mutant ML10 (ΔmspAΔmspC) as described earlier (Heinz and Niederweis, 2000) and above (Fig. 4C). No channels were detected when the extraction buffer was used in lipid bilayer experiments, whereas the membrane conductance increased by several orders of magnitude after addition of both cell extracts (Fig. 5A and B). This demonstrated the presence of porins in both extracts. More than 80% of the porins solubilized from wild-type cells had a single channel conductance of 4.8 nS (Fig. 5C) in 1 M potassium chloride consistent with earlier results (Niederweis et al., 1999; Stahl et al., 2001). The single channel conductances of the porins solubilized from cells of the ML10 mutant were identical with that of MspA (Fig. 5D) indicating that porins with a high similarity to MspA were expressed. This demonstrates that activation of transcription of the mspB and mspD genes indeed leads to a functional expression of these porins in the OM of the ML10 strain consistent with the finding that all Msp proteins increase the OM permeability of M. smegmatis to p-nitrophenylphosphate. The channel activities were determined by measuring a series of 10-fold dilutions of both extracts in the extraction buffer. While a 1:104 dilution of wild-type extract gave rise to reconstitution rates between 4.6 and 14 pores/min in eight diphytanoylphospatidylcholine membranes, no single channel was detected using the same dilution of extracts obtained from the double porin deletion mutant ML10. However, for a 1:103 dilution of the same extract, reconstitution rates between 0.2 and 2.3 pores min−1 were measured in 11 diphytanoylphospatidylcholine membranes. These data show that there is at least a 20-fold lower channel activity in extracts from the ML10 strain compared with wild-type M. smegmatis and is consistent with a 15-fold or larger reduced number of Msp porins.

image

Figure 5. Channel activity of detergents extracts of porin mutants of M. smegmatis. Single channel recordings of a diphytanoyl phosphatidylcholine membrane in the presence of detergent extracts from M. smegmatis SMR5 (A) and ML10 (B). Ten milligrams cells were extracted in 35 µl POP05 buffer and diluted 10 000-fold (SMR5) and 1000-fold (ML10), respectively, in the extraction buffer before adding 5 µl to both sides of the membrane. Probability P of conductance steps G observed in the single-channel recordings shown in A and B. A total of 131 and 98 single-channel events from 8 and 11 different membranes was analysed using extracts from wild-type M. smegmatis (C) and the ΔmspAΔmspC mutant (D) respectively. The main single-channel conductance was 4.8 nS for both samples.

Download figure to PowerPoint

A total of 71 pores of extracts of the porin triple mutant ML16 was analysed in 20 membranes using lipid bilayer experiments. The channel activity was very low compared with wild-type M. smegmatis, but not significantly different from that of the double mutant ML10 (not shown). This is consistent with the similar amount of Msp porins in detergent extracts of ML10 and ML16. Approximately 45% and 30% of all pores had single channel conductances of 2.3 nS and 4.6 nS respectively. This demonstrated that also the MspB protein forms open channels as shown above. The proportion of 2.3 nS channels increased from 10% (ML10) to 45% (ML16). As 2.3 nS is the major single channel conductance of recombinant MspA expressed in E. coli and also after extraction of porins from M. smegmatis using a mixture of chloroform and methanol, the pore with the smaller conductance may represent a different conformation of the Msp pore as it was assumed earlier (Niederweis et al., 1999).

The OM permeabilities of a double and a triple porin mutant of M. smegmatis are severely reduced for small and hydrophilic nutrient molecules

It was shown previously that the porin-mediated permeation through the OM is the rate-limiting step for uptake of glucose by M. smegmatis (Stahl et al., 2001). Therefore, the rate of glucose uptake was determined for the msp mutant strains to examine the physiological functions of the individual Msp porins in M. smegmatis. First, the uptake rates of all strains were compared at a glucose concentration of 20 µM, as differences in diffusion through porins are more pronounced at solute concentrations in the µM range. M. smegmatis SMR5 (wild type) took up glucose with a rate of 3.6 nmol min−1 per mg cells. This rate was not significantly changed in the ΔmspB and ΔmspC mutants (Fig. 6A). Both ΔmspA mutants (MN01 and ML01) showed an approximately two-fold reduced uptake rate of 1.5 and 1.9 nmol min−1 per milligram of cells respectively. This result is similar to results obtained earlier for M. smegmatis MN01 (Stahl et al., 2001) and confirmed that mspA is the major porin gene of M. smegmatis. Uptake of glucose by the double (ML10) and the triple porin mutant (ML16) was extremely slow with 0.3 nmol min−1 per milligram of cells and is consistent with the much lower number of Msp porins in their OMs (Fig. 6A, Fig. S1A). These porin mutants also took up 20 µM glycerol with identical rates (Fig. S1B). These results also emphasized the importance of the total number of Msp porins for the OM permeability of M. smegmatis.

image

Figure 6. Glucose uptake by porin mutants of M. smegmatis. A. Accumulation of [14C]glucose by M. smegmatis SMR5 (wild-type), the two ΔmspA mutants MN01 and ML01, the ΔmspB mutant ML03, the ΔmspC mutant ML05 and the ΔmspAΔmspC mutant ML10 was measured. The assay was performed at 37°C at a final glucose concentration of 20 µM. The uptake experiments were performed in triplicate and are shown with their standard deviations. The uptake rates were determined by regression analysis of the first four values of each strain for each strain. The dotted lines represent the regression lines. B. Michaelis–Menten analysis. A series of glucose uptake measurements was performed with glucose concentrations ranging from 1 to 20 µM for M. smegmatis SMR5 (wild-type), MN01 (ΔmspA) and ML10 (ΔmspAΔmspC). For each strain the uptake rates at different glucose concentrations were approximated by Michaelis–Menten functions, which are shown as regression lines.

Download figure to PowerPoint

A series of similar uptake experiments with glucose concentrations ranging from 1 to 20 µM was performed to determine the apparent permeability coefficients of M. smegmatis wild-type, the ΔmspA and the ΔmspA ΔmspC mutant for glucose (Fig. 6B). The triple porin mutant ML16 was not included in this experiment because there was no difference in the initial uptake experiments compared with the parent strain ML10 (Fig. S1). For all strains the data were fitted well by the Michaelis–Menten equation (Fig. 6B). Data analysis yielded vmax and Km values for the overall transport of glucose (Table 1), which were used to calculate minimal permeability coefficients according to Jarlier and Nikaido (1990). The minimal permeability coefficient of the ΔmspA ΔmspC mutant for glucose is 3.5 × 10−7 cm s−1 and is 50-fold lower than that of wild-type M. smegmatis (Table 1). Thus, the M. smegmatis double porin mutant has a similar permeability for glucose as M. chelonae, which is known for its very low OM permeability (Jarlier and Nikaido, 1990). It should be noted that in this case this minimal estimate is likely to be close to the true permeability coefficient for the transport across the OM of the ΔmspA ΔmspC mutant, because uptake of glucose did not reach an equilibrium rate within 15 min in contrast to wild-type M. smegmatis and the other porin mutants (Fig. 6A). A steady-state equilibrium between the glucose pool in the medium and in the periplasm is quickly reached if diffusion through the porins is fast as observed for the wild-type strain and the ΔmspB and ΔmspC mutants. This time is delayed for the ΔmspA mutants and no apparent equilibrium is observed for the ΔmspA ΔmspC mutant indicating that diffusion of glucose through the few remaining porins of the ML10 strain is outpaced by glucose transport across the inner membrane. Using the same method, minimal permeability coefficients were determined for serine (Table 1). Deletion of mspA and mspC reduced the permeability of M. smegmatis to serine 16-fold to a level as low as that of M. chelonae (Jarlier and Nikaido, 1990).

Table 1.  Kinetic parameters and permeability coefficients P of M. smegmatis and porin mutants for hydrophilic solutes.
  M. smegmatis M. chelonae E. coli
Wild-typeΔmspAΔmspAΔmspC
  1. Data were from: a.Stahl et al. (2001), b.Jarlier and Nikaido (1990), c.Bavoil et al. (1977), d.Nikaido et al. (1977). It should be noted in this regard that the permeability coefficients of M. smegmatis wt and of the mspA mutant for glucose are identical to that obtained earlier (Stahl et al., 2001).

  2. The vmax and Km values are measured with whole cells and reflect the overall transport across both inner and outer membrane. ND, not determined.

Serine
 vmax/nmol min−1 mg−11.2 0.9  0.06NDND
 Km/µM3.2 2.9  2.7NDND
 P/cm s−1d2.4 × 10−5 2.0 × 10−5  1.4 × 10−6NDND
Glucose
 vmax/nmol min−1 mg−12.6 1.4  0.5   4.76b107c
 Km/µM9.219.11321070b  3c
 P/cm s−11.8 × 10−5 4.6 × 10−6  2.4 × 10−7   2.8 × 10−8b  1.4 × 10−2c
Cephaloridine
 P/cm s−17.2 × 10−7a 8.4 × 10−8a  3.0 × 10−8   1.0 × 10−7b  9.3 × 10−6d

Using a modified Zimmermann–Rosselet assay (Stahl et al., 2001) we analysed the permeability of M. smegmatis ML10 for the hydrophilic zwitterionic β-lactam antibiotic cephaloridine. Cephaloridine at a concentration of 800 µM was hydrolysed with a rate of 24.3 ± 0.5 nmol min−1 mg−1 by lysed cells. This β-lactamase activity is identical to that determined earlier for M. smegmatis wild-type and the ΔmspA mutant MN01 (Stahl et al., 2001) demonstrating that deletion of the porin genes mspA and mspC did not affect the total β-lactamase activity. Intact cells of M. smegmatis ML10 hydrolysed 800 µM cephaloridine at a rate of 0.2 nmol min−1 mg−1, which is 120-fold lower than its total β-lactamase activity. This value was in the range of the β-lactamase activity detected in the supernatant of the same cells indicating that the loss of two porins strongly slowed down the uptake of cephaloridine by M. smegmatis. To rule out the possibility that this was attributed to an indirect effect, M. smegmatis ML10 was transformed with the mspA expression vector pMN014 featuring a psmycmspA fusion. Episomal expression of mspA fully restored the hydrolysis rate of cephaloridine to the wild-type level with 4.7 nmol min−1 mg−1 without changing the total β-lactamase activity. The minimal permeability coefficient of M. smegmatis ML10 for cephaloridine was determined to 3.0 × 10−8 cm s−1, which is 24-fold and 3-fold lower compared with M. smegmatis wild-type and the ΔmspA mutant MN01 respectively (Stahl et al., 2001).

Deletion of mspA decreased the permeation of all solutes significantly consistent with earlier experiments and with the conclusion that MspA is the major porin of M. smegmatis (Stahl et al., 2001). Additional deletion of the porin gene mspC caused a drastic reduction in the permeation rates of all three hydrophilic solutes analysed in this study demonstrating that Msp porins provide general diffusion pathways for small and hydrophilic solutes indispensable for M. smegmatis.

Deletion of porin genes reduces the growth rate of M. smegmatis

To examine the influence of porin-mediated nutrient uptake on the growth of M. smegmatis, the growth of porin mutants was examined on plates and in liquid culture. To use the sizes of colonies on agar plates as measure for the growth rate it is essential that they all start from a single cell. Therefore, small cultures of M. smegmatis SMR5 (wild type), MN01 (ΔmspA), ML10 (ΔmspAΔmspC) and ML16 (ΔmspAΔmspCΔmspD) were filtrated through 5 µm filters and appropriate dilutions were plated on 7H10 agar plates. Pictures were taken from single colonies for each strain on each day after colonies became visible. On day 2, small colonies were observed for SMR5 and MN01, but not for ML10 and ML16 (not shown). On day 4, it was obvious that the colony sizes were similar for SMR5 and MN01, but significantly smaller for ML10 and ML16 (Fig. 7A). A more subtle difference was that the fine structure of the colony, which was visible on day 4 for M. smegmatis wild type, was less elaborated for the ΔmspA mutant. The appearance of these structures was delayed for 2 days for the porin mutants ML10 and ML16. Furthermore, both porin mutants appeared to grow more in the vertical direction compared with the wild-type and MN01 strains (Fig. 7A, day 7). These results demonstrated that the porin mutants ML10 and ML16 exhibited a clear growth defect on agar plates compared with the wild-type and MN01 strains.

image

Figure 7. Growth of porin mutants of M. smegmatis. A. Growth on plates. M. smegmatis SMR5 (wt), MN01 (ΔmspA), ML10 (ΔmspAΔmspC) and ML16 (ΔmspAΔmspCΔmspD) were plated on 7H10 agar plates after filtration through a 5 µm filter to obtain single cells. The plates were incubated at 37°C and pictures were taken from single colonies of each strain on each day. Photographs of colonies were taken using the same magnification for all strains on a particular day (day 4: 10×, day 7: 6.5×). B and C. Growth in liquid culture. M. smegmatis SMR5 (filled squares), MN01 (filled triangles) and ML10 (open circles) were grown in 7H9 medium containing 0.2% (22 mM) glycerol (B) or in MMT minimal medium containing 10 µM glucose (C). The experiments were performed in triplicate. Data are shown with their standard deviations.

Download figure to PowerPoint

The growth of the three strains M. smegmatis SMR5, MN01 and ML10 was also examined in liquid culture. In the standard 7H9 medium containing 0.2% (22 mM) glycerol the generation times in the exponential phase were 3.3, 4.3 and 5.9 h respectively (Fig. 7B). Thus, deletion of the mspA gene alone significantly reduced the growth rate of M. smegmatis in 7H9 medium in contrast to earlier experiments (Stahl et al., 2001), in which the growth of M. smegmatis was probably influenced to a larger extent by clumping of the cells. Deletion of the two porin genes mspA and mspC further reduced the growth rate consistent with the phenotype of the ML10 strain on agar plates. In minimal medium containing 10 µM glucose (Fig. 7C) growth of all strains was clearly limited by the amount of carbon source, but growth differences between the strains SMR5, MN01 and ML10 were even more pronounced with generation times of 6.5, 10.7 and 12.7 h, respectively, in the exponential phase. Similar results were obtained in minimal medium containing 1 mM succinate (not shown).

Expression of mspA in the porin mutants MN01 and ML10 from the mspA expression vector pMN016 fully restored the amount of Msp porins, as determined by in protein gels stained with Comassie and in Western blot experiments using the anti-Msp serum pAK♯813 (not shown), and the growth rates of the strains to wild-type level (Fig. S2). Generation times were 5.2 h for SMR5/pMS2 (empty vector) and MN01/pMN016 and 4.6 h for ML10/pMN016 as determined from the exponential phase of the growth curves by regression analysis. The generation times were longer than those obtained for the same strains without plasmids indicating that plasmid replication and/or antibiotic selection significantly reduced the growth rate of M. smegmatis. Complimentation of the growth defect of ML10 by expression of mspA clearly demonstrates that the loss of porins and not an indirect effect is the cause of the slower growth of the porin mutants MN01 and ML10 of M. smegmatis.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The mspB and mspD genes provide backup porins for M. smegmatis

In this study, we demonstrated that all four msp genes encode for functional porins in M. smegmatis. Deletion of the mspA gene in three independent strains (MN01, ML02, ML10) activated transcription of one or two silent porin genes mspB and mspD in M. smegmatis. Expression of mspA at wild-type level in these mutants did not shut off mspB and/or mspD transcription indicating that one or several chromosomal mutations and not a regulatory mechanism triggered activation of these genes. An insertion element with homology to the IS1547 element of M. tuberculosis (Fang et al., 1999) was found 156 bp upstream of the mspD gene in wild-type M. smegmatis (Fig. 1). This situation is reminiscent of the quiescent porin gene nmpC of E. coli, whose expression is activated by loss of an insertion element (Highton et al., 1985). However, the IS1547-like element was also present in all mspA mutants as observed by PCR and DNA sequencing (not shown). Hence, the mechanism of activation of mspB and mspD transcription remains obscure. These results show that the mspB and mspD genes function as backup porins in M. smegmatis similar to several silent porin genes of E. coli (Highton et al., 1985; Blasband et al., 1986; Fajardo et al., 1998; Prilipov et al., 1998). These results indicate further that: (i) deletion of mspA may be detrimental for M. smegmatis unless expression of other porin genes compensates for the loss of MspA and (ii) the permeability defects of the mspA mutants are alleviated by additional expression of other porin genes and underestimate the real physiological role of MspA in wild-type M. smegmatis.

It is not clear whether the Msp porins have additional functions. MspC might have a different substrate specificity compared with MspA because of the N102E mutation, which adds a further negative charge on the periplasmic side of the constriction zone (Faller et al., 2004) and may therefore change the permeation rate of charged solutes. The three other amino acid exchanges of MspC compared with MspA are located in loop regions and are likely not to alter the permeability properties of this channel. A very intricate network to balance the OM permeability of E. coli by regulating porin gene expression is based on a rather subtle difference in channel size of the two main porins OmpF and OmpC (Ferenci, 1999; Liu and Ferenci, 2001). Thus, some of the msp genes might be expressed only, or at different levels, under certain conditions, e.g. nutrient limitation, which were not examined yet. The localization of mspB and mspD next to genes encoding putative inner membrane transporters for sugars (msmeg0505–0508 and msmeg6015/6022) might support this assumption.

The number of Msp porins determines the OM permeability of M. smegmatis for hydrophilic solutes

The 15-fold lower number of pores of the ML10 strain correlates well with the lower channel activity of detergents extracts and the 15- and 75-fold lower permeability compared with wild-type M. smegmatis to small and hydrophilic nutrient molecules such as serine and glucose respectively (Table 1). The OM permeability for each solute decreased with the number of porins of the M. smegmatis strains indicating that the Msp porins determine the OM permeability of M. smegmatis for small and hydrophilic solutes and hence the overall transport across the cell envelope under standard conditions. Considering the similarity of the cell envelopes in mycobacteria (Paul and Beveridge, 1992) it is assumed that this holds true for all mycobacteria. M. chelonae is the only Mycobacterium other than M. smegmatis, for which permeability coefficients are available. It has a 640-fold lower permeability for glucose than that of M. smegmatis, whereas the permeability of M. chelonae for cephaloridine is only sevenfold lower (Jarlier and Nikaido, 1990). M. chelonae appears to have one unidentified copy of an mspA-like gene (Niederweis et al., 1999), but porin mutants have not yet been described.

Diffusion through an MspA pore is less efficient than through OmpF/OmpC pores of E. coli

Extrapolation revealed that M. smegmatis would need approximately 1.8 × 106 MspA-like porins to achieve the same permeability for glucose as E. coli assuming that transport across the inner membrane of M. smegmatis would not become rate-limiting (Fig. 8, Table 1). A permeability of 0.014 cm s−1 for glucose is achieved by E. coli by a quasi two-dimensional crystalline layer of porins (Steven et al., 1977). This corresponds to approximately 1.7 × 104 porins per cell based on a cell surface area of 1.9 µm2 (see Experimental procedures) and of 114 nm2 for the surface area of a single OmpF molecule (Cowan et al., 1992). Thus, the porin pathway of M. smegmatis is approximately 100-fold less efficient than that of E. coli. As the major porin MspA of M. smegmatis constitutes only one central channel compared with three pores in each OmpF protein (Cowan et al., 1992; Faller et al., 2004), diffusion of glucose through the MspA pore is approximately 30-fold less efficient than through the OmpF pore. This might be caused by the 2.5-fold longer MspA channel, which significantly increases the chance of collisions of the solutes with the side walls compared with OmpF, and/or the very different constriction zones of the two pores. Further experiments are needed to distinguish between these mechanisms.

image

Figure 8. Correlation of permeability and growth rates with the number of porins in M. smegmatis. The permeability coefficients for glucose (blue circles; from Table 1) and growth rates in 7H9 medium containing 0.2% (22 mM) glycerol (full red triangles) and in MMT minimal medium containing 10 µM glucose (open red triangles; from Fig. 7B) are shown for the three strains M. smegmatis SMR5 (wt) and the porin mutants MN01 (ΔmspA) and ML10 (ΔmspAΔmspC). Regression analysis yielded the following functions for the dependence of three parameters on the number of pores: the straight line y = −7.2·10−7 + 7.7·10−9x (blue solid line) with a correlation coefficient of 0.997 for the permeability for glucose, the hyperbolic decay function y = (6.1·2474)/(2474 + x) (red dotted line) with a correlation coefficient of 0.977 for growth in 7H9 medium containing 0.2% glycerol (22 mM) and the straight line y = 13 − 2.7·10−3x (red solid line) with a correlation coefficient of 0.999 for growth in minimal medium. The blue dashed line represents the permeability coefficient of E. coli for glucose (0.014 cm s−1, Bavoil et al., 1977). The red dashed line represents the maximal doubling time of E. coli in rich medium (24 min, Bremer and Dennis, 1996).

Download figure to PowerPoint

The growth rate of M. smegmatis depends on sufficient porin-mediated nutrient influx

The number of porins per cell determines the influx rate of hydrophilic solutes across the OM and was calculated for wild-type M. smegmatis and the porin mutants MN01 (ΔmspA) and ML10 (ΔmspAΔmspC) to 2400, 800 and 150 pores, respectively, using a surface area of 2.4 µm2 for an exponentially growing cell (see Experimental procedures). A striking observation was that the growth rate of M. smegmatis, both in rich and minimal medium and on agar plates, dropped drastically with its porin-mediated OM permeability (Fig. 8). This is in contrast to porin mutants of E. coli and Salmonella typhimurium, which did not show a growth defect in rich media (Henning and Haller, 1975; Nurminen et al., 1976). Only under glucose limitation and in competition experiments, a growth advantage of wild-type E. coli over an ompF mutant was observed (Liu and Ferenci, 1998). Loss of both the OmpF and OmpC porins increased the doubling time of E. coli by 15% (Bavoil et al., 1977). Why do porin mutants of M. smegmatis show a severe growth defect, but porin mutants of E. coli do not? Bavoil and Nikaido found that a 1000-fold reduced residual porin level is sufficient to support growth of E. coli at maximal rate with 0.2% glucose (Bavoil et al., 1977). By contrast, in M. smegmatis a modest 3- and 15-fold reduced number of porins increased the generation time by 30% and 80% respectively. In minimal medium the growth rate of M. smegmatis declined more rapidly with the number of porins than in rich medium (Fig. 8) underlining the greater importance of porins at low nutrient concentration. This has been observed very early for E. coli (Bavoil et al., 1977) and is a consequence of the first Fick's law of diffusion. Higher influx rates are achieved at higher nutrient concentrations and less porins are needed to maintain a constant nutrient influx into the cell under those conditions. The data fit perfectly to a straight line (Fig. 8, open triangles) demonstrating that porin-mediated nutrient influx across the OM is the sole determinant of the growth rate of M. smegmatis porin mutants at low carbon concentrations. This was confirmed by complementation experiments, which restored both the OM permeability to hydrophilic solutes and the growth rate to wild-type levels. However, it is not clear whether the low efficiency of the porin pathway limits the growth rate of wild-type M. smegmatis. Experimental proof of this hypothesis would require to show that a larger number of porins in the wild-type strain would increase the growth rate. However, different mspA expression vectors using strong mycobacterial promoters only increased the initial amount of mspA transcript, but not the number of MspA porins in the OM (D. Hillmann and M. Niederweis, unpublished) indicating that unknown mechanisms limit integration of MspA into the OM. Acceleration of growth was observed after heterologous expression of mspA in M. bovis BCG (Mailaender et al., 2004; Sharbati-Tehrani et al., 2004). Taken together these results establish the slow porin-mediated uptake of hydrophilic nutrients is an important determinant of the slow growth of mycobacteria as proposed earlier by Nikaido and coworkers (Jarlier and Nikaido, 1990). It is likely that slow nutrient uptake across the OM is not the only growth rate-limiting step and that the rates of other cellular processes have been adapted. Indeed, it was observed for E. coli that the effective permeabilities of the inner and the outer membranes are well matched for galactosides for concentrations in the range of 100–200 µM (West and Page, 1984). It should be noted that an excess of porins may also be dangerous for M. smegmatis as its sensitivity to drugs (Stephan et al., 2004a) and to toxic compounds generated by immune cells (Sharbati-Tehrani et al., 2005) is increased in the wild-type compared with porin mutants. These observations underline the importance for M. smegmatis to balance its porin-mediated OM permeability between two mutually exclusive needs of rapid uptake of nutrients and protection from toxic compounds.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial strains and growth conditions

All bacterial strains used in this study are listed in Table S3. Mycobacterial strains were grown at 37°C in Middlebrook 7H9 liquid medium (Difco Laboratories) supplemented with 0.2% glycerol, 0.05% Tween 80 or on Middlebrook 7H10 agar (Difco Laboratories) supplemented with 0.2% glycerol unless indicated otherwise. E. coli DH5α was used for all cloning experiments and was routinely grown in LB medium at 37°C. The following antibiotics were used when required at the following concentrations: ampicillin (100 µg ml−1 for E. coli), kanamycin (30 µg ml−1 for E. coli; 10 µg ml−1 for M. smegmatis), hygromycin (200 µg ml−1 for E. coli, 50 µg ml−1 for M. smegmatis).

Construction of overexpression vectors for the msp porin genes of M. smegmatis

To construct expression vectors for the msp porin genes with identical expression signals, the genes were amplified by PCR using the plasmids pPOR6, pCS12 and pCS4 as template for mspA, mspC and mspD and genomic DNA from the M. smegmatis mspA mutant MN01 for mspB. A PacI (underlined) restriction site and a consensus mycobacterial ribosome binding site (bold) were introduced using the appropriate forward oligonucleotides mspASDfwd, mspBSDfwd, mspCSDfwd and mspDSDfwd and a SwaI restriction site using the reverse oligonucleotides mspASDrev, mspBSDrev, mspCSDrev and mspDSDrev (Table S2). The PCR fragments were digested PacI and SwaI, purified by preparative gel electrophoresis and ligated with the appropriately digested pMN016 DNA to give the porin expressing plasmids pMN041, pMN042, pMN043 and pMN044. All plasmid constructions were verified by restriction enzyme digestion and double-stranded DNA sequencing.

Construction of porin gene mutants of M. smegmatis

To construct single mutants of M. smegmatis lacking the porin genes mspB and mspD fragments of approximately 1000 bp of DNA up- and downstream of mspB and mspD were amplified via PCR using chromosomal DNA of M. smegmatis SMR5 as template and oligonucleotides listed in Table S2. The restriction sites for PmeI and SpeI were introduced into the upstream fragments as well as PacI and SwaI sites into the downstream fragments. The individual sequences were cloned to bracket the FRT-hyg-FRT cassette in the suicide vector pMN252 via the corresponding restriction enzymes. The final deletion vectors for the mspB and mspD genes were named pMN247 and pMN254 respectively. The cloned fragments were sequenced to ensure the absence of PCR errors. These plasmids were transformed into M. smegmatis SMR5. Direct selection for double cross-over recombinants yielded three clones after 5 days of incubation in each experiment. Chromosomal DNA was prepared from all six clones. Specific probes for mspB (963 bp) and mspD (398 bp) were amplified by PCR using chromosomal DNA of M. smegmatis SMR5 as a template. Oligonucleotides mspBdownf and mspBdownr were used for the mspB probe and CS4-02 and CS4-09 for the mspD probe (Table S2). The annealing temperatures were 62°C and 50°C respectively. The genomic DNAs of the six mspB and mspD deletion candidates were digested with AatII and BspEI, respectively, and Southern hybridization was performed as previously described (Stephan et al., 2004b). Locus-specific Southern blots revealed that all clones resulted from homologous recombination. One clone for each gene was selected and named ML03 (ΔmspB::hyg) and ML07 (ΔmspD::hyg). Excision of the hyg genes by the Flp recombinase was performed as described (Stephan et al., 2004b) and occurred with a frequency between 2.5% and 10%. The unmarked porin gene deletion mutants were named ML04 (ΔmspB) and ML08 (ΔmspD). PCR fragments amplified with the primers mspBupf/mspBdownr and CS4-10/CS4-12 (Table S2) and chromosomal DNA from M. smegmatis ML04 and ML08 as templates, respectively, were sequenced using the primers mspBkofwd2 and CS4-05 (Table S2) respectively. The sequences revealed that the hyg cassettes were precisely excised and that one FRT site was left on the chromosome.

A new strategy was devised to construct a triple porin mutant of M. smegmatis lacking the porin genes mspA, mspC and mspD. The mspA gene should be integrated at the attB site into the chromosome of the porin double mutant ML10 to provide additional porins for rescue of the triple mutant. For this purpose the plasmid pMN012 (Table S1) was digested using PmeI and XhoI to excise the pwmycmspA expression cassette. The fragment was then inserted between two loxP sites into the SmaI linearized vector pBS246 to yield pML500. The loxp-mspA-loxP fragment was excised from pML500 with NotI and inserted into the PmeI site of pML113 resulting in pML502, the integrative mspA expression vector. The strain M. smegmatis ML10/pML102 was transformed with pML502 and hygromycin-resistant clones were selected on plates. Chromosomal DNA was prepared from one clone. An upstream of attB-localized specific probe (293 bp) was amplified by PCR using chromosomal DNA of M. smegmatis ML10 as a template and oligonucleotides attBup-fwd and attBup-rev (Table S2). The annealing temperature was 56°C. A Southern blot hybridization of SacII-digested genomic DNA revealed that the selected clone was positive for integration of pML502 into the attB site and was designated M. smegmatis ML11 (Table S3). The hygromycin resistance gene hyg was excised using the Flp recombinase, as the hygromycin gene was flanked by FRT sites. Excision was performed using the Flp expression vector pMN234 (Table S1) as described previously (Stephan et al., 2004b) with an efficiency of 48%. One hygromycin-sensitive strain was named M. smegmatis ML12 (Table S3). The mspD deletion vector pMN254 (Table S1) was transformed into M. smegmatis ML12. Selection for single cross-over yielded 27 colonies. Chromosomal DNA was prepared from five clones. PCR with chromosomal DNA revealed that only one out of five clones underwent single homologous recombination. This strain was designated M. smegmatis ML13 (Table S3). A 4 ml culture of M. smegmatis ML13 was grown overnight and filtrated through a 5 µm filter. Then, the culture was diluted first to an OD600 of 0.05 and again 1:100 in 7H9 medium. A 100 µl aliquot of both dilutions was plated on hygromycin and streptomycin containing 7H10 agar plates to select for double cross-over events. A total of 201 colonies were obtained after 4 days. Chromosomal DNA was prepared from five clones and digested with SacII. Southern blot hybridization using the mspD specific probe revealed that three out of five candidates were positive for mspD deletion. One clone was named M. smegmatis ML14 (Table S3). Excision of the hygromycin gene using the Flp expression vector pMN234 (Table S1) was performed as described previously (Stephan et al., 2004b) with an efficiency of 27%. The chromosomal DNAs of two clones were analysed by PCR and by Southern blot hybridization using SacII-digested DNA. Both experiments confirmed excision of the hygromycin resistance gene in both clones (data not shown). One clone was named M. smegmatis ML15 (Table S3). To remove the rescue porin gene mspA from the attB site, the ML15 strain was transformed with the Cre recombinase expression plasmid pML183 (Table S1) to mediate site-specific recombination between loxP sites. One colony was grown in a 4 ml liquid culture medium and filtrated through a 5 µm filter. The culture was diluted in 7H9 medium and plated on 7H10 agar plates containing streptomycin for counter-selection against the cre expressing plasmid pML183. Chromosomal DNA was prepared from three clones. PCR and Southern blot hybridization using SacII-digested chromosomal DNA and the attB specific probe revealed that all clones were positive for mspA excision (data not shown). One clone was named M. smegmatis ML16 (Table S3). Liquid cultures for all strains were inoculated from a single clone from the selection plate and were used for both preparation of chromosomal DNA and storage of the strain in 20% glycerol at −80°C to minimize accumulation of secondary mutations.

RNA preparation and RT-PCR

Mycobacterium smegmatis was grown in 10 ml cultures to an OD600 of 0.8 and mixed with 5 ml killing buffer (20 mM Tris-HCl, 5 mM MgCl2, 20 mM NaN3). The cell suspension was incubated on ice for 5 min. Cells were harvested by centrifugation and lysed by agitation with glass beads (FastRNA Tubes-Blue) in a FastPrep© FP120 bead beater apparatus (Bio-101) for 20 s at level 6.5. RNA preparation was performed using the Nucleospin® RNAII kit (Macherey-Nagel) and sonication to render the chromosomal DNA accessible to degradation by DNaseI as described (Stephan et al., 2004c). The RT-PCR experiments were performed using an OneStep RT-PCR Kit (Qiagen). Total RNA 100 ng was transcribed by Omniscript reverse transcriptase and Sensiscript reverse transcriptase into cDNA for 35 min at 50°C. The following primers were used for the amplification of porin gene specific cDNA: mspA-FP2 and mspA-RP2 (mspA), mspB-FP and mspA-RP (mspB), mspC-FP2 and mspA-RP (mspC) and mspD-FP and mspD-RP (mspD) (Table S1). Thirty cycles (1 min at 96°C, 30 s at 60–65°C depending on the primer pair, 2 min at 72°C) were run to amplify the cDNA. The PCR products were analysed using 1% agarose gels, which were stained with ethidium bromide.

Preparation of detergent extracts from M. smegmatis and analysis by gel electrophoresis and lipid bilayer experiments

Porins were selectively extracted from whole cells of M. smegmatis as described (Heinz et al., 2003b). Briefly, 10 mg of M. smegmatis cells were washed with phosphate-buffered saline (PBS), resuspended in 35 µl POP05 buffer and heated to 100°C under stirring for 30 min. The cell suspension was cooled on ice for 10 min and centrifuged at 4°C for 15 min. The supernatant (crude extract) was used for further experiments (15 µl for Coomassie staining, 1 µl of a 10−2 dilution for immunoblot analysis). Quantitative image analysis of immunoblots and protein gels was performed using the software NIH Image 1.62 (http://rsb.info.nih.gov/nih-image).

Lipid bilayer measurements were performed as described previously (Niederweis et al., 1999). A series of 10-fold dilutions of both extracts in the extraction buffer was used to determine the channel activity. A 5 µl aliquot of the extracts was added to both sides of the membrane and the membrane current was measured in an aqueous solution of 1 M KCl after application of a transmembrane potential of 10.2 mV. The temperature was kept at 20°C throughout the experiment.

Electron microscopy of cell wall fragments of M. smegmatis

Staining of cell-wall fragments of M. smegmatis with uranyl acetate and electron microscopy was performed as described (Engelhardt et al., 2002).

Calculation of cell surface areas for M. smegmatis and E. coli

Electron microscopy showed that a cylinder with hemispherical polar caps is a good model to approximate the rod-like shape of a M. smegmatis cell. The cell length L and the diameter (2R) of M. smegmatis were determined to 1.5 µm (Greendyke et al., 2002; Dziadek et al., 2003) and to 0.5 µm (Cunningham and Spreadbury, 1998; Dziadek et al., 2003) respectively. Then, the surface area A of M. smegmatis is 2πRL = 2.4 µm2. The same model describes also the cell shape of E. coli (Grover et al., 2004). Its surface area was calculated to 1.9 µm2 based on the carefully determined length of 1.36 ± 0.29 µm and diameter of 0.447 ± 0.034 µm of an average E. coli cell growing at steady state in a glucose containing minimal medium with a doubling time of 42 min (Grover and Woldringh, 2001).

Measurement of the OM permeability for cephaloridine and glucose

The measurement of permeation of cephaloridine through the OM of M. smegmatis was exactly performed as previously described (Stahl et al., 2001). The permeability coefficient P was calculated using a Km of 146 µM for the β-lactamases of M. smegmatis (Trias et al., 1992) and an approximation of 132 cm2 mg−1 (dry weight) as the surface area : weight ratio for mycobacteria (Jarlier and Nikaido, 1990).

Glucose uptake measurements were carried out as previously described (Stahl et al., 2001). To reduce aggregation and clumping all M. smegmatis cells were filtered through a 5 µm pore size filter (Sartorius) and regrown for 2 days at 37°C before inoculating 100 ml cultures (Stephan et al., 2004a). The cells were harvested at an OD600 of 0.5 by centrifugation (1250 g at 4°C for 10 min), washed once in 2 mM PIPES (pH 6.5), 0.05 mM MgCl2 and resuspended in the same buffer. Radio-labelled [14C]glucose and non-labelled glucose were mixed and added to the cell suspension to obtain final concentrations of 1, 2.5, 3.3, 5, 10 and 20 µM. The mixtures were incubated at 25°C and 1 ml samples were removed at the indicated times. The cells were filtered through a 0.45 µm pore size filter (Sartorius), washed with 0.1 M LiCl and counted in a liquid scintillation counter. All experiments were performed in triplicate. The mean dry weight of the cells in these samples was 0.6 ± 0.1 mg. The uptake rate was expressed as nmol per milligram of cells. Uptake rates at the individual concentrations, Km and vmax values for the overall transport and a minimal estimate of the permeability coefficient were determined as described previously (Jarlier and Nikaido, 1990; Stahl et al., 2001).

Growth experiments

Mycobacterium smegmatis SMR5, MN01 and ML10 were incubated on 7H10 agar plates for 3 days at 37°C until the colonies showed a smooth appearance. Then, 4 ml 7H9 medium containing 0.2% glycerol and 0.05% Tween 80 were inoculated with cells from these plates and incubated for 12–20 h at 37°C using a roller drum at 110 r.p.m. This procedure almost completely abolished the occurrence of clumps, which might have affected the growth rates of the strains differently. An appropriate volume of the precultures was taken to inoculate a 100 ml culture to a final OD600 of 0.01. Before inoculating MMT minimal medium (6 g l−1 Na2HPO4, 3 g l−1 KH2PO4, pH 7.4, 0.5 g l−1 NaCl, 1 g l−1 NH4Cl, 2 mM MgSO4, 0.1 mM CaCl2, 0.05% Tween 80) containing 10 µM glucose or 1 mM succinate as a carbon source, cells were harvested by centrifugation (10 min at 3200 g and 4°C), washed twice and then resuspended in 4 ml of the same minimal medium used for the final culture. Growth of the strains in the cultures was followed by measuring the OD600 in triplicate. Growth rates in the exponential phase were determined by regression analysis. The condition of the bacteria in the cultures was followed by staining of cell aliquots with the LIVE/DEAD stain using the protocol of the manufacturer (Molecular Probes) and fluorescence microscopy. During the exponential phase of the cultures only very few clumps were observed. However, in stationary phase both the number and the size of cell clumps increased drastically so that clumping became visible without a microscope. Up to half of the cells in these clumps emitted red fluorescence light indicating that these cells were dead.

For the complementation experiments the M. smegmatis strains MN01 and ML10 were transformed with the mspA expression vector pMN016 (Table S3). The wild-type strain SMR5 was transformed with pMS2 to compare the growth in hygromycin-containing media. Growth curves were recorded as described above.

Morphology of M. smegmatis strains on plates

Cultures of 4 ml 7H9 medium were inoculated with M. smegmatis SMR5, MN01, ML10 or ML16 and grown overnight at 37°C. The cultures were then filtrated through a 5 µm pore size filter to ensure that all colonies would arise from single cells. The optical density of the filtrate was measured at 600 nm. The filtrates were initially diluted to an OD600 of 0.05 using 7H9 medium and then 10−4, 10−5 and 10−6-fold dilutions were made. For each dilution, 100 µl were plated on 7H10 agar plates. The plates were wrapped with parafilm and incubated at 37°C. Plates with a larger number of single colonies were chosen and pictures of colonies were taken at different magnifications over 8 days using a Stemi 2000-C stereomicroscope (Zeiss) connected to a AxioCam MRc camera (Zeiss).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Iris Kaps for initial RT-PCR experiments, Kristin Hasselt for measuring the growth rate of complemented porin mutants in liquid culture, Dr Xiuan Bay for constructing the plasmid pMN016, Jonathan Lowery for improving the language of the manuscript and Dr Wolfgang Hillen for support. Preliminary sequence data for M. smegmatis was obtained from The Institute for Genomic Research web site at http://www.tigr.org. This work was mainly supported by the Deutsche Forschungsgemeinschaft (NI 412).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1. Plasmids used in this work. Up- and downstream homologous sequences of genes are subscripted as up and down. ?Origin? means origin of replication. The annotations AmpR, HygR and KanR indicate that the plasmid confers resistance to ampicillin, hygromycin and kanamycin, respectively.

Table S2. Oligonucleotides used in this work. Restriction sites are underlined. The consensus ribosome binding site in the oligonucleotides for the msp expression vectors is marked in bold. Oligonucleotides mspBdownf and mspBdownr were used for the mspB probe and CS4-02 and CS4-09 for the mspD probe.

Table S3. Strains used in this work. The annotations SmR, GenR and HygR indicate that the strain is resistant to the antibiotics streptomycin, gentamicin and hygromycin, respectively. It should be noted that all strains are derivatives of SMR5 and are therefore resistant to streptomycin. This is not indicated in the table except for SMR5.

Fig. S1. Uptake of glucose and glycerol by the porin triple mutant M. smegmatis ML16. The accumulation of 20?μM [14C]glucose (A) and 20?μM [14C]glycerol (B) by the porin double mutant M.?smegmatis ML10 (ΔmspA ΔmspC; black circles) and the porin triple mutant M.?smegmatis ML16 (ΔmspA ΔmspC ΔmspD; white circles) was measured at 37°C for 16 minutes. The experiments were done in triplicate. Data are shown with standard deviations.

Fig. S2. Complementation of the growth defect of the porin double mutant M.?smegmatis ML10. M.?smegmatis SMR5/pMS2 (open circles), MN01/pMN016 (closed circles) and ML10/pMN016 (closed squares) were grown in 7H9 medium containing 0.2% glycerol (22?mM). The experiments were done in triplicate. Data are shown with standard deviations. Generation times were determined by regression analysis of the exponential phase (OD600 ? 1 to 4) for all strains.

FilenameFormatSizeDescription
MMI_4878_sm_FigS1.zip65KSupporting info item
MMI_4878_sm_FigS2.zip43KSupporting info item
MMI_4878_sm_TablesS1-S3.doc59KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.