A Mycobacterial Gene Involved in Synthesis of an Outer Cell Envelope Lipid Is a Key Factor in Prevention of Phagosome Maturation

Bacterial strains and cells.Transposon mutants of a fish isolate of M. marinum (ATCC 927) were generated (31). M. marinum broth cultures and cultures on solid media were grown using Middlebrook 7H9 broth and 7H10 agar enriched with 10% oleic acid-albumin-dextrose complex at 30°C. The transposon mutants were cultured in these media containing 30 μg/ml kanamycin at 30°C. Human peripheral monocytes were isolated from the buffy coat by Ficoll gradient centrifugation. Isolated monocytes were allowed to differentiate into macrophages in RPMI 1640 supplemented with 5% fetal bovine serum in human serum-opsonized tissue-culture-grade petri plates with hydrophilic membranous bases (Lumox; Greiner Bio-One GmbH, Frickenhausen, Germany) for 7 days at 37°C in a 5% CO₂ atmosphere.

Screening for mutants not competent for retarding phagosome maturation.A bacterial inoculum was prepared from a pool of more than 4,000 individual mycobacterial transposon mutant colonies on Middlebrook 7H10 agar. A bacterial pellet was harvested by centrifugation and washed three times in phosphate-buffered saline (PBS). A small amount of glass beads (diameter, 0.2 mm) was added, and the suspension was shaken on a mill to disrupt the bacterial clumps. After sedimentation of the glass beads for 30 min, a fraction of the supernatant suspension was carefully removed and dispersed further by three passages through a 27-gauge needle.

Ten Lumox dishes with confluent human monocyte-derived macrophages (HMDMs) (3 × 10⁶ cells) were pulsed with 0.5 ml of colloidal iron dextran particles (Basic MicroBeads; catalog no. 130-048-001; Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) for 2 h, followed by a chase for 2 h. Then monolayers were washed with PBS to remove the excess beads. The macrophages were then pulsed with an M. marinum mutant library at a multiplicity of infection (MOI) of 10:1 for 2 h, rinsed with PBS to remove the excess bacteria, and chased overnight. The plates were then rinsed with RPMI once, equilibration buffer [50 mM piperazine-N,N′-bis(ethansulfonic acid) (PIPES) (pH 7.0), 50 mM KCl, 2 mM MgCl₂, 5 mM EGTA, 1 mM dithiothreitol, 10 μM cytochalasin B] was added, and the plates were incubated on ice for 20 min. Lysis buffer (50 mM PIPES [pH 7.0], 50 mM KCl, 2 mM MgCl₂, 5 mM EGTA, 220 mM mannitol, 68 mM sucrose, 1 mM dithiothreitol, 10 μM cytochalasin B) was added, and the cells were scraped off using a rubber policeman and collected in a tube. The macrophage lysate was passed 15 times through a 22-gauge needle for homogenization and then applied to a washing buffer-equilibrated Mini MACS column placed in a magnetic field. The flowthrough was collected for later analysis. The column was washed with wash buffer (2 mM EDTA and 0.5% bovine serum albumin in PBS), and the flowthrough was saved for analysis. The organelle fraction retained in the magnetic field in the column harboring the magnetic beads was eluted with elution buffer (10 mM triethanolamine, 10 mM acetic acid, 1 mM EDTA, and 0.25 M sucrose in PBS) after removal of the magnet. A small aliquot of the eluant was saved for analysis, while the remaining portion was centrifuged, resuspended in a small volume of Middlebrook 7H9 broth, and plated on 7H10 agar with kanamycin (30 μg/ml). Agar dishes were incubated at 30°C for 5 days, and then the colonies were scraped from the plates. The harvested colonies were pooled, washed three times, and dispersed to obtain a single-cell suspension for infection of HMDMs. The selection procedure was repeated two more times. Aliquots of each preparation obtained were frozen at −80°C.

Western blot analysis of the lysosomal fraction recovered by magnetic separation.HMDMs were pulsed with Basic MicroBeads (Miltenyi Biotec) for 1 h, and this was followed by a chase for 2 h or overnight. Control macrophages were incubated without beads. After the chase, macrophages were lysed as described above, and magnetic separation of subcellular compartments was performed using the procedure described above for selection of mutants that were not competent for retarding phagosome maturation. Aliquots of the macrophage lysate, the fraction not binding to the magnetic column (flowthrough), and the eluted organelle fraction containing the magnetic beads were stored for analysis by sodium dodecyl sulfate (SDS)-10% polyacrylamide gel electrophoresis (PAGE). Samples were solubilized in Laemmli sample buffer and separated by SDS-PAGE. Following electrophoretic transfer to a polyvinylidene difluoride membrane (GE Healthcare), the membrane was probed with early endosomal antigen 1 (EEA-1) antibody (BD Pharmingen, Heidelberg, Germany), lysosome-associated membrane protein 1 (LAMP-1) antibody (BD Pharmingen), and glyceraldehyde-3-phosphate dehydrogenase protein (GAPDH) antibody (Zymed Laboratories, San Francisco, CA) and developed by enhanced chemiluminescence (ECL; GE Healthcare).

Southern blot analysis.Portions (4 μg) of extracted chromosomal DNA of isolated M. marinum mutants and the wild type (WT) were digested with BamHI endonuclease. The restricted DNA was separated by electrophoresis and transferred to a nylon membrane. A probe specific for the aph gene was amplified from plasmid pUC4K (GE Healthcare) and labeled with digoxigenin using a DIG DNA labeling kit (Roche Applied Sciences, Mannheim, Germany). After prehybridization and hybridization the membrane was developed using the CDP Star luminescence detection reagent (Amersham Biosciences, Freiburg, Germany).

Identification of transposon insertion sites by sequencing.Genomic sequences flanking the transposon insertion sites were identified by using an arbitrary primed PCR and direct sequencing strategy as described previously (31). Briefly, both sides flanking the transposon insertion site in a mutant were amplified in two separate nested PCRs. The primers used in the first round were primers ARB1 (5′-GGCCACGCGTCGACTAGTACNNNNNNNNNN-3′) and RPCRa1 (5′-CTTGCTCTTCCGCTTCTTCTC-3′) or primers ARB1 and RPCRb1 (5′-CAGGCACGTCGAGGTCTTTC-3′). Each first-round PCR product was used as a template and subjected to a second PCR. The primers used in the second round were primers ARB2 (5′GGCCACGCGTCGACTAGTAC-3′) and RPCRa2 (5′ CTCTACACCGTCAAGTGCGAAGAG-3′) or primers ARB2 and RPCRb2 (5′-CTTTCAGATGGATGGCGTAG-3′). The product from the second round of PCR was purified using a QIAGEN PCR purification kit (QIAGEN, Hilden, Germany). DNA sequencing reactions were performed with primer RPCRa2 or RPCRb2 and Big Dye Terminator sequencing kits (Applied Biosystems, Foster City, CA). Reaction mixtures were analyzed with an ABI Prism 310 genetic analyzer. The sequences obtained were matched with the sequence of M. marinum ATCC BAA-535, and the homologies of the interrupted potential open reading frames (ORFs) with the “Omniome” protein database were analyzed using the BLASTX function available through the TIGR Comprehensive Microbial Resource at www.tigr.org .

Staining of mycobacteria with FITC or TRITC.WT or mutant M. marinum cells were suspended in 200 μl of 0.1 M sodium bicarbonate buffer (pH 9.0) containing 1 mg/ml fluorescein isothiocyanate (FITC) or 0.1 mg/ml tetramethylrhodamine-5-isothiocyanate (TRITC) (both dyes were obtained from Sigma-Aldrich, Germany) and incubated at 30°C for 30 min. After incubation the excess dye was removed by three washes with PBS. The bacteria were then resuspended in RPMI for macrophage infection.

Phenotype screening of mutants by fluorescence microscopy.Mutants and the WT were screened for phagosome maturation by fluorescence microscopy. Briefly, adherent cultures containing 3 × 10⁶ HMDMs were transfected with adenoviral expression vectors of the Rab5- or Rab7-green fluorescent protein (GFP) fusion protein. The transfected cells were infected at an MOI of 1:1 with TRITC-labeled WT or mutant bacteria for 2 h. After a chase for 2 h or overnight, cells were fixed with 4% paraformaldehyde. For lysosome-associated membrane protein staining infected HMDMs were permeabilized with 0.5% polyethylene glycol and marked for LAMP-1 using a mouse anti-human LAMP-1 antibody (BD Pharmingen), followed by a secondary goat anti-mouse immunoglobulin G antibody labeled with Alexa 594 (red) (Molecular Probes, Eugene, OR). The cells were washed, and the basal membranes of the Lumox dishes were excised and mounted on a slide. The slides were later viewed with an Olympus IX81 fluorescence microscope. Colocalization was quantified by examining a minimum of 100 individual phagosomes for each bacterial strain and each marker.

Radiolabeling and quantitative analysis of mutants reaching the phagolysosome. M. marinum [WT, P1 mutant, heat-killed WT, and complemented mutant P1(pGPC352)] were metabolically radiolabeled by growing them in Middlebrook 7H12 medium containing 1μCi [1-¹⁴C]palmitic acid (Becton Dickinson GmbH, Heidelberg, Germany). ¹⁴C-labeled bacteria were washed to remove unincorporated label and then dispersed to obtain single-cell suspensions prior to infection of HMDMs. The procedure described above for screening mutants that were not competent for retarding phagosome maturation was then employed to isolate phagosomes containing bacteria. The fractions of M. marinum WT, mutant P1, and heat-killed WT bacterial cells that were phagocytosed and delivered to phagolysosomes (PL) were quantified by measuring the radioactivities of the fractions with a β-scintillation counter.

Complementation of P1 mutant.A 3.8-kb region from M. marinum harboring the putative pmiA gene was PCR amplified using primers 5′-TGCGGCCGCTCTAGATGCGGTCAGGTATGTCAGCA-3′ and 5′-GGGGGATCCACTAGTCTATCGACGCTGGCGCAT-3′ and was cloned into plasmid pOLYG using a BD infusion kit (25) to obtain pGPC352 conferring resistance to hygromycin. Plasmid pGPC352 was transformed into mutant P1 by electroporation. P1(pGPC352) transformants were selected on 7H10 plates containing 30 μg/ml kanamycin and 50 μg/ml hygromycin. In addition, nested-deletion DNA fragments were generated from the 3,821-bp M. marinum fragment in pGPC352 using a combination of internal PCR primers. The amplified fragments were cloned in the same way into pOLYG to obtain pGPC369 to pGPC374 (see Fig. 6). These plasmids were transformed into the mutant for analysis of functional complementation. The P1 strains complemented with these plasmids were also labeled with ¹⁴C and chased along with PL markers (microbeads) as described above.

Macrophage infection.Differentiated HMDMs were infected with M. marinum WT, P1, and P1(pGPC352) at an MOI of 10:1 and incubated at 37°C for 2 h. Cells were washed thoroughly with RPMI to remove extracellular bacteria and then incubated further at 37°C. At different times (2 h, 24 h, 48 h, 72 h, and 96 h) cells infected with all three bacterial strains were lysed using 0.1% SDS for 10 min, neutralized using 20% bovine serum albumin, and plated on 7H10 agar plates for enumeration of CFU. The CFU count obtained after 2 h was considered the CFU count for time zero postinfection.

Infection of mice.Specific-pathogen-free C57BL/6 mice were infected intravenously with 4 × 10⁵ viable M. marinum WT, mutant P1, and complemented P1(pGPC352) cells. At 1 and 2 weeks after infection, five mice from each group were sacrificed, and their livers and spleens were excised aseptically. Small equal samples of each liver and spleen were fixed in 4% formaldehyde for histopathological analysis. The remaining liver and spleen samples were homogenized, and diluted aliquots were plated on 7H10 agar. The plates were incubated for 5 days, and the CFU were counted.

Statistical analysis.A statistical analysis of fluorescence microscopy colocalization data was performed by using a one-way analysis of variance (ANOVA) post hoc range test and pairwise multiple comparisons with Tamhane's T2 corrections, assuming nonequal variances using the statistical software package SPSS 11.0.4 for Mac OS X (SPSS Inc., Chicago, IL). The CFU data for livers and spleens of mice infected with the M. marinum WT, P1, and P1(pGPC352) strains were analyzed for five mice each at weeks 1 and 2 with the same statistical software package. An analysis was done using one-way ANOVA for livers and spleens and each week separately, using Bonferroni corrections for liver week 1 CFU values and Tanhame's T2 corrections for liver week 2 and spleen week 1 and 2 CFU values.

Histological analysis.The spleen and liver of each mouse were placed in 4% neutral formalin and processed for paraffin embedding and subsequent sectioning. Consecutive sections (2 to 5 μm) were mounted on glass slides, deparaffinated, stained with hematoxylin and eosin, and examined by a pathologist with no prior knowledge of sample identities.

Lipid extraction and analysis by TLC.Lipids of the M. marinum WT, mutant P1, and complemented P1(pGPC352) strains were isolated as described by Dobson et al. (11). Briefly, the outer highly nonpolar lipids were isolated using petroleum ether. After removal of the outer lipid coat, the cells were treated with chloroform-methanol-0.3% aqueous NaCl to obtain the nonpolar lipid component. The defatted cells were subjected to alkaline methanolysis and digestion with tetramethylammonium hydroxide, followed by conversion of the salts obtained to methyl esters using iodomethane. The lipids and mycolic acid esters obtained were analyzed on Silica Gel 60 F₂₅₄ precoated high-performance thin-layer chromatography (TLC) plates (Merck, Darmstadt, Germany). Solvent systems were chosen as described by Dobson et al. (11).

Screening for M. marinum mutants permitting phagosome maturation.To determine the mycobacterial genes that inhibit PM, a screen to isolate Tn5367 transposon mutants with defects in inhibition of PM was developed (Fig. 1). The screen design took advantage of our finding that M. marinum can endure the hostile phagolysosomal milieu for a longer time (24 h) (data not shown). It was also based on the presumption that mutants that cannot inhibit PM are enriched in the phagolysosomal compartment. Based on this presumption, transposon mutants were chased together with magnetic microbeads into the phagocytic pathway of HMDMs. After an overnight chase the plasma membrane of the HMDMs was lysed under carefully controlled conditions, preserving the intracellular organelles. The homogenate containing the subcellular components was then passed through MACS columns under a magnetic field. Next, the column was removed from the magnetic field, and the mutants were eluted along with the PL marker. The beta-galactosidase activity, as a lysosomal marker (5) of the eluted and flowthrough fractions, was determined (data not shown), and the results revealed that phagolysosomes were strongly enriched in the fraction recovered. The lysosomal fraction was plated on 7H10 agar and incubated for 5 days, after which the bacterial colonies were just visible. These colonies were scraped off the agar, pooled, and carefully dispersed to obtain a single-cell suspension, which was used to infect a fresh culture of HMDMs. The selection procedure was repeated three times for further enrichment of mutants that were not capable of inhibiting PM. After the final selection 100 individual colonies were picked for further investigation.

• Open in new tab
• Download powerpoint

FIG. 1.

Screening for mycobacterial mutants permitting phagosome maturation: phagosomal processing of mycobacterial mutants with defects in PM inhibition. EE, early endosome; LE, late endosome; PL, phagolysosome. WT bacteria remain in a Rab5-positive vacuole, whereas heat-killed bacteria and mutants with defects in PM are processed into a LAMP-1-positive, vATPase-positive, and cathepsin D-positive PL. An M. marinum transposon mutant library was chased together with magnetic microbeads to the PL, and the mutants proceeding to the PL were selected on a MACS column using a strong magnetic field.

Magnetic isolation of lysosomal fraction.In order to validate the conclusion that the microbeads in our screening procedure were indeed localized in lysosomes, we performed a control experiment without bacteria (Fig. 2). Macrophages without iron dextran microbeads, macrophages with microbeads after a 2-h chase, and macrophages with microbeads after an overnight chase were analyzed by SDS-PAGE and Western blotting to determine the presence of markers of the endosomal pathway, namely, LAMP-1 and EEA-1. The housekeeping gene product GAPDH was used as a control for equal loading of macrophage samples. For macrophages without microbeads we observed LAMP-1 and EEA1 staining in the lysate (Fig. 2, lane A) and flowthrough fractions (lane B), while only a minute amount was nonspecifically absorbed to the magnetic column (lane C). In contrast, for macrophages chased with iron dextran microbeads for 2 h or overnight there was significant isolation of LAMP-1-positive, EEA-1-negative, and GAPDH-negative lysosomal vesicles on the magnetic column (lanes F and I).

• Open in new tab
• Download powerpoint

FIG. 2.

Western blot analysis of lysosomal fraction recovered by magnetic separation. HMDMs were pulsed with Basic MicroBeads (Miltenyi Biotec) for 1 h, followed by a chase for 2 h (lanes D to F) or overnight (lanes G to I). Lanes A to C contained control macrophages without beads. After the chase, macrophages were lysed (lanes A, D, and G), and magnetic separation of subcellular compartments was performed by using the same procedure that was used for selection of M. marinum mutants that were incompetent for retarding phagosome maturation. Lanes B, E, and H contained a macrophage lysate fraction that did not bind to the magnetic column (flowthrough). Lanes C, F, and I show organelle fraction binding to the magnetic column in the absence (lane C) or presence (lanes F and I) of Basic MicroBeads. Following SDS-PAGE and blotting membranes were probed with LAMP-1, EEA-1, or GAPDH antibody and developed by enhanced chemiluminescence analysis.

Defining transposon insertion sites by sequence and Southern blot analysis.In order to define transposon-disrupted regions and to identify independent insertions in identical genes or loci, 100 mutants selected for their inability to prevent PM were analyzed by sequencing the transposon inserted locus. Southern blot analysis was also performed to confirm that each mutant had only a single transposon insertion and to identify mutants with transposon insertions in the same gene (data not shown). The sequences obtained were compared with the available M. marinum ATCC BAA-535 sequence database of the Sanger Institute (http://www.sanger.ac.uk/cgi-bin/BLAST/submitblast/m_marinum ) using BLAST. This analysis revealed that the insertions were not spread randomly across the genome. Instead, several regions with obvious accumulations of transposon insertions were identified (data not shown).

In general, our results describing the diversity of mycobacterial genes involved in retarding PM are in accordance with previous reports. Genes coding for FadD proteins and transporters and genes belonging to the PE/PPE gene family also appeared in our screen, as was the case in previous screens (26, 34). Interestingly, a mutant having a transposon insertion in a gene involved in isoprenol biosynthesis that was identified by Pethe et al. was also identified by our screen. Moreover, we took utmost care to disperse bacterial clumps to obtain single bacilli in order to apply the selective pressure to each single mutant. This enabled us to identify genes coding for membrane proteins and secretory proteins. Stewart et al. speculated that underrepresentation of such proteins in genetic screens has been due to cross presentation (34). Our screen also identified genes contributing to the resistance of mycobacteria to killing by macrophages, as reported independently by other workers (14, 19, 20, 23, 29, 32).

Surprisingly, transposon insertion sites of 15 mutants could not be mapped to the available M. marinum ATCC BAA-535 genome database, although similar genes in other mycobacterial genomes were identified, albeit with low probability scores. This indicated either that there is phylogenetic distance between M. marinum ATCC 927, a fish isolate, and ATCC BAA-535, a human isolate, or that unidentified mycobacterial prophages were present.

Immunofluorescence microscopy of HMDMs infected with mutants.We employed Rab5- and Rab7-GFP fusion protein and LAMP-1 antibody staining to characterize the phagosomes containing M. marinum WT or 6 mutants chosen randomly from the 100 mutants isolated after the three rounds of selection. Fluorescence micrographs revealed that phagosomes containing the WT indeed retained Rab5 (Fig. 3A) even after an overnight chase, whereas LAMP-1 (Fig. 3C) and Rab7 (Fig. 3E) were excluded. In contrast, phagosomes containing mutants were observed to colocalize with Rab7 and LAMP-1. Quantifying the colocalization events in more than 100 individual phagosomes for each mutant with each marker showed that one of the six randomly selected mutants (P1) exhibited the strongest reduction in the capacity to resist PM; for this mutant 73% ± 3% of the phagosomes acquired LAMP-1 and 69% ± 4% acquired Rab7. In contrast, for WT bacteria only 22% ± 2% of the phagosomes acquired LAMP-1 and 13% ± 2% acquired Rab7 (Fig. 4 A). The differences between the WT and complemented mutant P1(pGPC352) versus mutant P1 are statistically significant (Fig. 4B). The colocalization analysis indicated that the gene inactivated by transposon insertion in the P1 mutant indeed has an important function in preventing PM.

• Open in new tab
• Download powerpoint

FIG. 3.

Phagosome phenotypic characterization by fluorescence microscopy. (A) Fluorescence micrographs of HMDMs expressing Rab5-GFP fusion protein and infected with TRITC-labeled M. marinum WT (red). (B) HMDMs expressing Rab5-GFP fusion protein and infected with TRITC-labeled mutant P1 (red). (C) HMDMs infected with FITC-labeled M. marinum WT (green) and stained for LAMP-1 (red). (D) HMDMs infected with FITC-labeled mutant P1 (green) and stained for LAMP-1 (red). (E) HMDMs expressing Rab7-GFP fusion protein and infected with TRITC-labeled M. marinum WT (red). (F) HMDMs expressing Rab7-GFP fusion protein and infected with TRITC-labeled mutant P1 (red). Rab5- and Rab7-GFP fusion protein-expressing HMDMs were prepared by transfecting HMDMs with an adenoviral vector containing the corresponding genes. HMDMs were stained for LAMP-1 using mouse anti-human LAMP-1 and were tagged with a goat anti-mouse immunoglobulin G labeled with Alexa 594 (red).

• Open in new tab
• Download powerpoint

FIG. 4.

Analysis of colocalization data for WT and mutant M. marinum with endocytic markers. (A) Percentages of phagosomes containing M. marinum WT, P1, and complemented P1(pGPC352) colocalized with LAMP-1, Rab7, and Rab5. The data are the means of three independent experiments, and a minimum of 100 phagosomes were counted in each experiment for each sample. The error bars indicate standard deviations. (B) Statistical analyses of data for each time were performed by using a one-way ANOVA post hoc range test and pairwise multiple comparisons with Tamhane's T2 corrections assuming nonequal variances. P values are indicated.

Mutant P1 is attenuated in HMDMs.Our results show that WT M. marinum was able to multiply at 37°C until day 3 in HMDMs, as shown in Fig. 5, although the growth rate was low. For mutant P1 there was a steady decline in survival. These findings correlated with recent findings of Kent et al. (21). After 72 h a rapid increase in multiplication of the WT was observed. In contrast, the survival of mutant P1 in HMDMs was severely attenuated, as shown by the rapid decrease in survival with a lag phase of 24 h. In contrast, the mutant complemented strain P1(pGPC352) was virulent and had a growth pattern identical to that of the WT.

• Open in new tab
• Download powerpoint

FIG. 5.

Survival of M. marinum WT, mutant P1, and complemented mutant P1(pGPC352) in human monocyte-derived macrophages. Macrophages were infected at an MOI of 10:1, incubated at 37°C, and lysed for CFU counting at different times. The symbols indicate the means for two independent experiments in which duplicate determinations were made at each time. The error bars indicate the standard deviations.

Transcomplementation of P1 and coelution of ¹⁴C-labeled bacterial cells with lysosomal marker.The transposon insertion site of mutant P1 was mapped to a sequence that was very similar (97% sequence identity) to ORF MM3386 in the M. marinum ATCC BAA-535 genome. Neither the gene encoding the putative 203-amino-acid peptide sequence of M. marinum ATCC 927 nor MM3386 encoding a putative 197-amino-acid peptide was found to have significant similarity to genes in any other genome in the current databases (NCBI, EBI, and TIGR Comprehensive Microbial Resource). No signature patterns, domains, repeats, motifs, or features in the putative peptide sequence could be predicted with confidence using the latest SMART (Simple Modular Architecture Research Tool) at EMBL, the latest BLAST engines at the NCBI, or the latest releases of the PROSITE search engine at http://www.expasy.ch/tools/scanprosite/ , except for an RGD motif at amino acid positions 113 to 115 of the putative peptide sequence. The gene was given the provisional name phagosome maturation inhibition A (pmiA).

Even though the database queries could not suggest any function for the interrupted pmiA gene, the downstream genes on the complementary strand were identified as putative hydrolase (MM3387) and carboxylase (MM3388) genes. A sketch of the organization of these genes is shown in Fig. 6. The region spanning the transposon insertion site was cloned into a pOLYG shuttle vector, electroporated into P1, and selected on plates containing hygromycin and kanamycin. To examine whether reconstitution of the disrupted gene restored the WT phenotype, HMDMs were infected with [¹⁴C]palmitate-labeled bacteria [WT, heat-killed WT, P1, and P1(pGPC352)] and then pulsed and chased with marker for PL and selected on a MACS column under a magnetic field, as described above for the transposon mutant screen. Bacteria that coeluted with the PL marker were quantified with a β-scintillation counter. As shown in Fig. 7, the heat-killed M. marinum coeluted with the PL marker, whereas a minor fraction of the WT coeluted with PL. A significant fraction of P1 coeluted with the PL marker, whereas the level for the transcomplemented P1(pGPC352) strain was the same as the WT level. We inferred from these observations that the P1 mutant was not competent enough to resist PM and, when reconstituted with the appropriate gene, was able to regain the lost phenotype. To narrow the range of genes affected by the transposon insertion and responsible for the phenotypic change in P1, a set of nested deletions in the complementation plasmid was generated. The range of WT chromosomal sequences covered by the nested plasmids is shown in Fig. 6. Plasmid pGPC374 harboring just 1,332 bp of the WT sequence spanning pmiA and no other putative mycobacterial ORF was sufficient to restore the WT phenotype, as less than 5% of P1 bacterial cells harboring this plasmid coeluted with the PL marker (Fig. 7). These results confirmed that the transposon-interrupted pmiA gene itself was responsible for the inhibition of PM and eliminated the possibility of involvement of any polar effects on the neighboring genes.

• Open in new tab
• Download powerpoint

FIG. 6.

Physical map of pmiA (MM3386) in M. marinum ATCC 927 and the adjacent putative hydrolase (MM3387) and carboxylase (MM3388) genes. The numbering of the ORFs is based on the numbering for the M. marinum ATCC BAA-535 strain sequenced at the Sanger Institute. The M. marinum ATCC 927 chromosomal fragments cloned in this study for the complementation analysis of the mutation in mutant P1 are indicated in shaded bars below the gene graphs. The GenBank accession number is also indicated.

• Open in new tab
• Download powerpoint

FIG. 7.

Coelution of bacterial cells with lysosomal marker. M. marinum WT, mutant P1, and complemented mutant P1(pGPC352) were metabolically labeled with [¹⁴C]palmitate. HMDMs were infected with live strains and with heat-killed WT. Macrophages were pulsed and chased with a lysosomal marker (microbeads). Subcellular homogenates were applied to a MACS column under a strong magnetic field. After nonspecifically retained organelles were removed by washing, the phagosomes containing lysosomal markers were eluted. The fraction of bacteria that coeluted along with the PL marker was quantified with a β-scintillation counter. The values above the columns indicate the percentage of cells coeluting with the lysosomal marker.

In vivo survival of P1 in mice.C57BL/6 mice were intravenously challenged with WT, P1, and P1(pGPC352). One and two weeks postinfection mice were sacrificed, and the numbers of bacteria in spleens and livers were determined. Box plots of liver and spleen CFU are shown in Fig. 8A and B. Analysis of the number of CFU recovered from infected mice by one-way ANOVA for livers and spleens and for each week separately showed that the capacity to maintain viable bacterial cells in the infected organs was significantly diminished for the mutant P1 compared to both WT and P1(pGPC352); the P values were <0.004 for livers for the first and second weeks, <0.001 for spleens for the first week, and <0.011 for spleens for the second week.

• Open in new tab
• Download powerpoint

FIG. 8.

In vivo survival of M. marinum mutant P1 in mice. C57BL/6 mice were challenged intravenously with an inoculum consisting of 4 × 10⁵ CFU of M. marinum WT, P1, or P1(pGPC352) per mouse. After infection, livers and spleens were excised and homogenized, and each homogenate was diluted and plated on 7H10 agar. Box plots of the liver and spleen CFU recovered from livers and spleens of five mice infected with WT, P1, and P1(pGPC352) for each time are shown. The asterisk indicates an “extreme” value, and O′ indicates an “outlier.”

Histopathology of infected organs.Infection with M. marinum mutant P1 resulted in significantly reduced pathological changes in C57BL/6 mouse livers. The capacity of P1 to induce granuloma formation was clearly attenuated, as the number and extent of epitheloid granulomas were reduced with this mutant compared to the results obtained with both the WT and complemented mutant P1(pGPC352), especially in the second week. However, in P1(pGPC352)-infected animals the number of granulomas was more variable than the number in WT-infected animals. The majority of the granulomas in P1-infected animals developed in the lobules, and fewer granulomas developed in the portal tracts. Granulomas in P1-infected animals also showed more and stronger signs of inflammation around the granulomas (i.e., mainly lymphocytes and a few granulocytes). A striking feature of the liver sections of WT- and P1(pGPC352)-infected animals was the severe endothelialitis, whereas the endothelialitis in P1-infected animals was only marginal (Fig. 9; data for the first week not shown). One of the five WT-infected livers was entirely necrotic, with widespread hepatocyte damage. Necrosis in the granulomas was observed in none of the animals in the other groups. The pathological changes in the spleens were less pronounced. In the second week the red and white pulp of the spleens of two of five WT-infected animals had very few granulomas and there were no granulomas in two of five animals, and in one animal the entire spleen was necrotic. In the complemented mutant P1(pGPC352)-infected animals all five spleens had a few small granulomas in the white pulp. In contrast, the histology of the P1-infected animals was normal, and there were granulomas in only one animal.

• Open in new tab
• Download powerpoint

FIG. 9.

Representative micrographs showing the histopathology of infected organs. Magnification, ×400. The granuloma morphology in M. marinum-infected C57BL/6 mice is shown. Mice were intravenously infected with 4 × 10⁵ CFU of M. marinum WT, P1, or P1(pGPC352) and were sacrificed 2 weeks postinfection. Liver sections were stained with hematoxylin and eosin. (A and B) Large epithelioid granulomas in a WT-infected mouse and in a P1(pGPC352)-infected mouse. (C) Smaller, less-well-organized granulomas in a P1-infected mouse. (D and E) Endothelialitis in WT- and P1(pGPC352)-infected mice. (F) No sign of endothelialitis in a P1-infected mouse.

Noncording phenotype and lipid profile.Unexpectedly, P1 produced flat, smooth, transparent colonies when it was recovered from the organs of mice. The WT bacteria retained the rough colony morphology, and the morphology of the complemented P1(pGPC352) strain reverted back almost completely to the WT morphology (Fig. 10), confirming that the altered colony morphology of P1 was due to disruption of the pmiA gene. A comparison of the mycolic acid profiles of the WT, P1, and P1(pGPC352) strains revealed no obvious differences, as determined by TLC. Therefore, we used a systematic approach to study the lipid profiles, as described by Dobson et al. (11). When the highly hydrophobic outer layers of lipids were extracted using petroleum ether and analyzed by TLC, a lipid moiety in the WT was resolved which was missing in the P1 mutant and was present in the complemented P1(pGPC352) mutant (Fig. 11).

• Open in new tab
• Download powerpoint

FIG. 10.

Colony morphology of mutant bacterial cells: bright-field microscopy of M. marinum WT, P1, and P1(pGPC352) colonies 7 days after recovery from mouse organs and seeding onto Middlebrook medium. After passage through mouse organs, P1 produced flat, smooth, translucent colonies. The characteristics of P1(pGPC352) colonies were similar to the characteristics of WT colonies. Scale bars = 0.1 mm.

• Open in new tab
• Download powerpoint

FIG. 11.

Lipid profile of the mycobacterial cell wall. Nonpolar lipids noncovalently bound to the cell wall of M. marinum WT, P1, and P1(pGPC352) were isolated using petroleum ether and were separated by TLC. The chromatogram was developed by treating it with molybdate phosphoric acid in ethanol, followed by charring. The arrow indicates the position of a lipid fraction present in WT and P1(pGPC352) but not in the P1 mutant.