Supporting informationGhosh et al. 10.1073/pnas.0904104106SI Text

Materials and MethodsFlow Cytometry. Cells were harvested at different growth times,fixed, stained, and run in the flow cytometer as describedpreviously (1). Briefly, the cells were fixed in 70% (final con-centration) ice-cold ethanol and stored at 4° C. For flow cytom-etry, the fixed cells were washed with particle-free (passedthrough 0.2 �m nitrocellulose filters) 10 mM Tris-HCl (pH 7.9)buffer supplemented with 10 mM MgCl2. The cells were subse-quently resuspended into 0.1 � vol of the same buffer. Washedcells were stained with ethidium bromide (0.4 �g/mL, finalconcentration) and mithramycin A (16 �g/mL, final concentra-tion) and kept on ice for 30 min before flow cytometry (CompactA440 analyser Unit, Apogee Flow Systems and BioRad BryteHS). At least 20,000 cells were counted for each flow histogram.

Serial Dilution of Pure Spores to Isolate Single Spore Particle. Theidea of serial dilution to isolate a single spore was based on aprevious report (2). A batch of isolated spores was split into 2fractions; one was plated freshly, and after a week of growth thecolonies were counted to estimate the number of viable sporeparticles in the spore preparation. This number was �103/mL.The other fraction was serially diluted 104-fold, such that therewould be no more than 1 viable spore in each tube. Theso-obtained spore solutions were plated on 7H10 plates withnecessary supplements and the plates were incubated at 30° C.

To address the possibility that the spores were contaminations,we also extracted DNA from vegetative M. marinum (Mm) cells,spores, and Bacillus subtilis (Bs) (or from independent clonesoriginating from isolated single-spore particles) (Fig. S1B) byboiling in water for 5 min followed by centrifugation for 5 minat 11,000 � g in an Eppendorf centrifuge. The DNA was used astemplates in PCR amplification reactions of regions of the 16SrRNA gene, rnpA (which encodes the RNase P protein C5) andrnpB (which encodes the RNase P RNA). The results are shownin Fig. S1. The choice of using rnpB is based on that it haspreviously been used to differentiate closely related bacteria(3–8).

Detailed Description of Extraction of Total RNA from Mm and RNA DotBlot Analysis. Cells were broken in a 1-mL solution of TRIzol(Invitrogen) and 0.3 mL of chloroform with 0.1 mm silica beads(Q-Biogene, Lysing matrix B) using a BIO101 Savant FastPrepFP120 machine (6.5 speed, 30 sec pulse, 4 cycles). After 10 minof centrifugation at 11,000 � g, the aqueous phase was reex-tracted with 300 �L of chloroform and transferred to a tubecontaining 250 �L of isopropanol. Total RNA was precipitatedovernight at 4° C and washed with 1 mL of a 75% ethanol beforeresuspension in RNase free water. Contaminating DNA wasremoved by digestion with RQ1 DNase (Ambion) following themanufacturer’s instructions.

Dot blot and probing of RNA were done as described previ-ously (9). DNase-treated RNA samples prepared from 1-, 3-, 5-,and 7-day-old Mm cultures were denatured at 90° C for 5 min andthen spotted (10 �g in each spot) on Hybond-n � nylonmembranes (Amersham Biosciences) and air dried. The RNAwas immobilized by UV cross-linking by using a Bio-Rad cross-linker, followed by hybridization with specific 32P-5�-end-labeledoligonucleotides at 42° C. The membranes were washed asdescribed previously (9). After probing for the mRNAs, themembranes were stripped with hot 0.1% SDS, and then rehy-

bridized with 5S rRNA probes and washed as above. Hybrid-ization signals were analyzed using a Phosphor Imager (Image-Quant, Molecular Dynamics). Escherichia coli K12 stationaryphase RNA samples were also spotted on the same membranesand probed with the same oligonucleotides to calculate thebackground signals of each dot arising from unspecific binding,which were very low in our hybridization conditions. The back-ground values were subtracted from the corresponding values ofthe mRNA and 5S rRNA signals. Oligonucleotide probes werelabeled at their 5�-ends with �[32]ATP using T4 polynucleotidekinase (PNK) according to standard procedures. The nucleotidesequences of the oligonucleotide probes are listed in Table S6(PDF). To probe for Mm 5S rRNA, the oligonucleotide 5�-GCTGACAGGCTTAGCTTCCG was used.

Dipicolinc Acid (DPA) Assay. Mm cultures were harvested after 1day (vegetative cells), 7 weeks, and at the second month (bothare expected to contain a mixture of vegetative cells and spores).The cells and purified sample of spores from the 2-month-oldculture (2.0–10.0 OD600) were resuspended in physiologicalsaline and broken by grinding with glass beads followed byautoclaving at 15 lb/in2 for 15 min. The lysate was clarified bycentrifugation after treatment with acetic acid (0.1 mL of 1Nacetic acid added to the lysate) for 1 h at room temperature. Thiswas followed by the addition of freshly prepared ferrous ammo-nium sulfate hexahydrate and the color was measured at 440 nm(10). A calibration curve, plotted by using standard solutions ofpure DPA (Sigma-Aldrich) in 0.15 M NaCl, was used to calculatethe amount of DPA in the clear lysates.

Percoll-Sucrose Gradient Centrifugation Analysis of Fresh and Old MmCultures. We used a modified version of Percoll-sucrose densitygradient (11, 12) in an attempt to separate putative sporeparticles from vegetative cells in Mm cultures. A mycobacterialexpression vector carrying the gene for GFP expression, pG13(13), was used to follow the Mm strain through gradient frac-tionation and subsequent treatments with GFP-fluorescence tag.Briefly, equal numbers of cells (with similar OD600 values) from3- and 30-day-old plate cultures were resuspended into PBS and1 mL of the fresh and old Mm (with the pG13 plasmid)suspensions were layered on top of 25 mL preformed Percoll-sucrose gradients (70% Percoll with 7.8% sucrose in 30 mMNaCl centrifuged in a Ti50.2 rotor on Beckman-Coulter XL-100ultracentrifuge at 20,000 � g at 4° C for 30 min). The cellsuspensions were centrifuged at 38,000 � g at 4° C for 15 min,1-mL fractions were collected from bottom and GFP fluores-cence was measured on a Spectrofluorimeter (Hitachi F7000).The relative fluorescence intensities were plotted as shown inFig. S3A. To detect the presence of spore particles, the fractionswere individually wet-heat-treated at 65° C for 30 min and thenspread on 7H10 plates supplemented with 25 �g/mL kanamycin.This eliminated all vegetative cells (3-day cells) as well as thebackground acquired during the centrifugation and fractionationprocess. As shown in Fig. S3B, the 3-day-old cells did not containany survivors whereas in the gradient for 30-day-old culture, apeak of heat-resistant survivors appeared in a band at a slightlyheavier density (fractions 12 and 13) than the main peak. Thecolonies from the heat-resistant fractions all emitted greenfluorescence due to the expression of GFP, confirming theiridentity as Mm (carrying the pG13 plasmid).

Ghosh et al. www.pnas.org/cgi/content/short/0904104106 1 of 5

BioinformaticsSequence Data. Protein sequences from Bs were downloadedfrom the SubtiList web server (14–16). Sporulation-associatedproteins were obtained according to the SubtiList functionalcategory 1.8. In addition, protein sequences for sporulation-related factors SigE, SigF, SigG, Spo0A (TrcR in Fig. 5), andSpoVT were also downloaded. Amino acid sequences of iden-tified ORFs from Mycobacterium tuberculosis CDC1551, M.tuberculosis H37Rv, and Streptomyces coelicolor (Sc) A3 (2) weredownloaded from the Comprehensive Microbial Resource (17–20). Amino acid sequences of the initial gene predictions for Mmwere downloaded from the Sanger Institute FTP server (21).Proteins involved in Sc sporulation were obtained by queryingthe StrepDB database (26) for genes with annotations contain-ing any of the words ‘‘sporulation,’’ ‘‘whiA,’’ ‘‘whiB,’’ ‘‘whiD,’’‘‘whiE,’’ ‘‘whiG,’’ ‘‘whiH,’’ ‘‘whiI,’’ or ‘‘whiJ,’’ as well as genes withgene names matching ‘‘whi*’’ (‘*’ corresponding to any letter). Inaddition, the Sc gene for sigF was included.

Reciprocal Best-Hit Ortholog Search. The Washington UniversityBLAST (release date 10-May-2005; ref. 22) was used in a

reciprocal best hit approach (RBH) (23) to identify putativeorthologs for sporulation-associated Bs and Sc proteins in themycobacteria species listed above. We consider a pair of genes,a and b from genomes A and B, respectively, as putativeorthologs if a search with a against B identifies b as the mostsignificant match and a search with b against A results in a as themost significant match. The blastp program was used with defaultparameters and an expect cutoff at 0.001. Putative orthologs arelisted in Tables S1–S4. In total, 169 sporulation-associated Bsproteins were used as queries and 25, 21, 23, and 28 putativeorthologs were identified in Mm, Mtb CDC1551, Mtb H37Rv,and Sc, respectively, corresponding to 35 unique Bs sporulation-associated genes. Similarly, of 63 Sc sporulation-associatedgenes, 16, 14, 14, and 10 putative orthologs were identified inMm, Mtb CDC1551, Mtb H37Rv, and Bs, respectively, corre-sponding to 22 unique Sc genes. In addition, top-scoring hits forthe Bs transcription factors Spo0A, SpoIIAB, and SigF wereincluded. These hits did not return the initial Bs query proteinas the top-scoring hit in a reciprocal search and are thereforeconsidered as weak hits.

1. Maisnier-Patin S, Dasgupta S, Krabbe M, Nordstrom K (1998) Conversion to bidirec-tional replication after unidirectional initiation from R1 plasmid origin integrated atoriC in Escherichia coli. Mol Microbiol 30:1067–1079.

2. Lister J (1878) On the lactic fermentation and its bearings on pathology. Trans PatholSoc London 29:425–467.

3. Svard SG, Mattsson JG, Johansson KE, Kirsebom LA (1994) Cloning and characterizationof two mycoplasmal RNase P RNA genes. Mol Microbiol 11:849–859.

4. Herrmann B, Winqvist O, Mattsson JG, Kirsebom LA (1996) Differentiation of Chla-mydia spp by sequence determination and restriction endonuclease cleavage of theribonuclease P RNA gene. J Clin Microbiol 34:1897–1902.

5. Herrmann B, Pettersson B, Everett KDE, Mikkelsen NE, Kirsebom LA (2000) Character-ization of the rnpB gene and the RNase P RNA in the order of Chlamydiales. Int J SystEvol Microbiol 50:149–158.

6. Tapp J, Thollesson M, Herrmann B (2003) Phylogenetic relationships and genotyping ofthe genus Streptococcus by sequence determination of the RNase P RNA gene, rnpB.Int J Syst Evol Microbiol 53:1861–1871.

7. Innings A, Krabbe M, Ullberg M, Herrmann B (2005) Identification of 43 Streptococcusspecies by pyrosequencing analysis of the rnpB gene. J Clin Microbiol 43:5983–5991.

8. Rubin CJ, Thollesson M, Kirsebom LA, Herrmann B (2005) Phylogenetic relationshipsand species differentiation of 39 Legionella species by sequence determination of theRNase P RNA gene rnpB. Int J Syst Evol Microbiol 55:2039–2049.

9. Sambrook J, Fritsch EF, Maniatis T (2001) in Molecular Cloning, A Laboratory Manual(Cold Spring Harbor Laboratory Press, New York), 3rd ed.

10. Janssen FW, Lund AJ, Andersson LE (1958) Colorimetric assay for dipicolinic acid inbacterial spores. Science 127:26–27.

11. Trun NJ, Gottesman S (1990) On the bacterial cell cycle: Escherichia coli mutants withaltered ploidy. Genes Dev 4:2036–2047.

12. Tisa LS, Koshikawa T, Gerhardt P (1982) Wet and dry bacterial spore densities deter-mined by buoyant sedimentation. App Environ Microbiol 43:1307–1310.

13. Barker LP, Procella SF, Wyatt RG, Small PLC (1999) The Mycobacterium marinum G13promoter is a strong sigma 70-like promoter that is expressed in Escherichia coli andmycobacteria species. FEMS Microbiol Lett 175:79–85.

14. The SubtiList WWW Server, Institut Pasteur. Available at: http://genolist.pasteur.fr/Subtilist/. Accessed: 12/8/2007.

15. Moszer I, Glaser P, Danchin A (1995) SubtiList: a relational database for the Bacillussubtilis genome. Microbiology 141:261–268.

16. Moszer I (1998) The complete genome of Bacillus subtilis: From sequence annotationto data management and analysis. FEBS Lett 430:28–36.

17. The Comprehensive Microbial Resource (CMR), J. Craig Venter Institute. Available at:http://cmr.jcvi.org/cgi-bin/CMR/CmrHomePage.cgi. Accessed: 12/8/2007.

18. Peterson JD, Umayam LA, Dickinson TM, Hickey EK, White O (2001) The comprehensivemicrobial resource. Nucleic Acids Res 29:123–125.

19. Cole ST, et al. (1998) Deciphering the biology of Mycobacterium tuberculosis from thecomplete genome sequence. Nature 393:537–544.

20. Fleischmann RD (2002) Whole-genome comparison of Mycobacterium tuberculosisclinical and laboratory strains. J Bacteriol 184:5479–5490.

21. The Wellcome Trust Sanger Institute FTP Server: ftp://ftp.sanger.ac.uk/pub/pathogens/mm/. Accessed: 12/8/2007.

22. Gish W (1996–2004), The Washington University of St. Louis. Available at: http://blast.wustl.edu. Release date: 10 May 2005.

23. Rivera MC, Jain R, Moore JE, Lake JA (1998) Genomic evidence for two functionallydistinct gene classes. Proc Natl Acad Sci USA 95:6239–6244.

24. The Wellcome Trust Sanger Institute Web Service. Available at: http://www.sanger-.ac.uk/. Accessed: 12/8/2007.

25. Cerdeno-Tarraga AM et al. (2003) The complete genome sequence and analysis ofCorynebacterium diphtheriae NCTC13129. Nucleic Acids Res 31:6516–6523.

26. StrepDB The Streptomyces Annotation Server, John Innes Centre. Available at: http://strepdg.streptomyces.org.uk/. Accessed on: 12/8/2007.

27. Flardh K, Buttner MJ (2009) Streptomyces morphogenetics: Dissecting differentiationin a filamentous bacterium. Nat Rev Microbiol 7:36–49.

Ghosh et al. www.pnas.org/cgi/content/short/0904104106 2 of 5

Fig. S1. Purified spores’ germination into vegatative M. marinum cultures and formation of spores at late stationary phase. (A) Life cycle of Mm from purifiedspores by microscopy. 0h, Purified spore suspension; 6h, germinating spores; 8h, germinated cells; 1d, germinated cells; 2d and 3d, same as 1d; 5d-1 and 5d-2,appearance of endospores; 1w, appearance of mature spore; 2w, mature spores with cell debris. Note that spores, germinating spores, and endospores appearas bright reflecting particles. (Scale bars: 5 �m.) (B) Life cycle of Mm from purified spores by flow cytometry. An Mm culture was started from a purified sporesuspension and followed through germination, exponential growth to sporulation in late stationary phase by flow cytometry. The left and right columns showDNA content (fluorescence calibrated in chromosome numbers) and cell-size (light scattering in arbitrary units) distributions, respectively. The top frames referto stationary phase LB cultures of E. coli used as calibration standards for DNA content and cell size. The 9 lower frames correspond to Mm samples taken at 0,6, 8, 24, 48, 72, 120, 168, and 336 h after suspension of the purified spores into the growth medium. Fixation and staining procedures were as described in Materialsand Methods, and in Fig. 1A legend. The narrow size distribution and the shift of the peak as a whole with little widening implied synchronous transition ofspores into vegetative cells. At late stationary stage (after 120 h of growth), the cell-size distribution implies emergence of mature spores from the narrowingof the peak and its shift toward smaller size (right column: 168h and 336h). The fluorescence profiles show the onset of DNA synthesis by 6 h after addition ofgrowth medium; replication continues for 72 h, after which nonreplicative cell division is followed by sporulation.

Ghosh et al. www.pnas.org/cgi/content/short/0904104106 3 of 5

Fig. S2. Identity of spores as originating from M. marinum. (A) Development of single Mm colony from isolated single spore. Morphologic comparison of Mmcolonies from cultures germinated from single, purified (see Materials and Methods), independent spore particles (i–iv) with that from an exponentially growingculture inoculated from an old stock (v). (Scale bar: 1 cm.) (B) RNase P RNA gene sequences from independent clones originating from isolated single-sporeparticles. Alignment of residues 89–315 of the RNase P RNA gene from the chromosomes of clonally isolated colonies. (i–iv) Sequences from genomic DNA isolatedfrom single colonies grown out of isolated single-spore particle. (v) Sequence from RNase P RNA gene of Mm cloned in plasmid pIGn. (vi) Mm RNase P RNA genesequence obtained by blast search in the Sanger database using the M. tuberculosis H37Rv RNaseP RNA gene sequence (24) as reference. Sequences i–v areidentical to those reported in the database except for positions 93 and 158. These discrepancies could be due to the difference between our lab strain (ATCC927) and the strain used to build the database (ATCC BAA-535) or an error in the database. Oligos used to PCR-amplify region 89–315 were as follows. Forward:5�-ATCGAGGAAAGTCCGGACTTCACA-3� (comprising region 65–88); reverse: 5�-ACCCGCAGACTCGGGCGAGCAG-3� (complementary to region 294–315). Theforward oligo was used to obtain the sequences i–v. (C) Comparison of 16S rRNA gene sequences originating from Mm vegetative cells and spores. Alignmentof residues 179–476 of the 16S rRNA gene amplified from Mm vegetative cells and spores as indicated. (i) Mm 16S rRNA gene sequence obtained from the Sangerdatabase. (ii) Sequence from genomic DNA isolated from vegetative cells. (iii) Sequence from genomic DNA isolated from spores. Oligos used to PCR-amplify the179–476 region were: Forward: 5�-AGGACCACGGGATTCATGTCCTG-3� (comprising region 179–201); reverse: 5�-CCGTCAATCCGAGAAAACCCGAA-3� (comple-mentary to region 454–476). Both oligos were used to obtain the 16S rRNA sequences. (iv) For comparison, we included the corresponding 16S rRNA region fromCorynebacterium diphtheriae (25), which belongs to the order Actinomycetales (the same order as Mycobacterium spp.), in the alignment. (D) Comparison ofthe rnpA gene sequences originating from Mm vegetative cells and spores. Alignment of residues 110–246 of the rnpA gene amplified from Mm vegetative cellsand spores as indicated. (i) Mm rnpA gene sequence obtained from the Sanger database. (ii) Sequence from genomic DNA isolated from vegetative cells. (iii)Sequence from genomic DNA isolated from spores. Oligos used to PCR-amplify region 110–246 were as follows. Forward: 5�-AGGAATGCGCTGCAGACAGCAG-3�(comprising region 110–131); reverse: 5�-CACGGGTGAAGGTTCATCAGC-3� (complementary to region 226–246). Both oligos were used to obtain the rnpAsequences. (iv) For comparison, we included residues 110–246 of the rnpA gene from C. diphtheriae (25; see description of C) in the alignment. Together withthe difference comparing Mm and C. diphtheriae16S rRNA sequences (see description of C), this excluded contamination by Corynebacterium. (E) Amplificationof parts of the 16S rRNA gene, rnpA and rnpB from Bs, vegetative Mm cells, and Mm spores. DNA was extracted from Bs, Mm vegetative cells, and spores asoutlined above. The so-obtained DNA preparations were used as templates in PCR amplification reactions using the following primers. Set 1, complementaryto the Bs rnpB gene (14); forward: 5�-ATCGCTGCAGATCTTGAATC-3� (comprising region 19–38); reverse: 5�-CTCCGTTGAATTCCGTTAAG-3� (complementary toregion 272–291). Set 2, complementary to the Mm rnpA gene; forward: 5�-AGGAATGCGCTGCAGACAGCAG-3� (comprising region 110–131); reverse: 5�-CACGGGTGAAGGTTCATCAGC-3� (complementary to region 226–246). Set 3, complementary to the Mm 16S rRNA gene; forward: 5�-AGGACCACGGGATTCAT-GTCCTG-3� (comprising region 179–201); reverse: 5�-CCGTCAATCCGAGAAAACCCGAA-3� (complementary to region 454–476). Set 4, complementary to the MmrnpB gene; forward: 5�-GCGCCGAATCGAGGAAA-3� (comprising region 58–74); reverse: 5�-CACCTCGGATTGCTGTTAGCA-3� (complementary to region 101–121).For PCR amplification, we used the following settings: denaturation for 3 min at 95° C, 30 cycles at 95° C (15 sec), 60° C (15 sec), and 72° C (30 sec). Afteramplification, the samples were kept at 72° C for 7 min. The PCR-amplified products were run on 2% agarose gels in 50 mM Tris-Borat, pH 8.3, 1.2 mM EDTAbuffer. Lanes 1, 6, 11, 16, and 21, 100 base pair DNA ladder; lanes 2–5, no template added; lanes 7–10, DNA extracted from vegetative Mm cells; lanes 12–15,DNA extracted from Mm spores; lanes 17–20, DNA extracted from Bs culture. The following primer sets (see above) were used: lanes 2, 7, 12, and 17, set 1 (rnpBBs; product in lane 17); lanes 3, 8, 13, and 18, set 2 (rnpA Mm; products in lanes 8 and 13; however, we also detected products in lane 18 which is most likely dueto the fact that the rnpA primers used are not specific); lanes 4, 9, 14, and 19, set 3 (16S rRNA gene Mm; products in lanes 9 and 14); lanes 5, 10, 15, and 20, set4 (rnpB Mm; products in lanes 10 and 15). On the basis that Sc is an exosporulating bacteria (27), and that Sc spores are morphologically different compared withmycobacteria spores, we can also conclude that our Mm cultures were not contaminated with Sc.

Ghosh et al. www.pnas.org/cgi/content/short/0904104106 4 of 5

Other Supporting InfomationTable S1 (PDF)

Table S2 (PDF)

Table S3 (PDF)

Table S4 (PDF)

Table S5 (PDF)

Table S6 (PDF)

Table S7 (PDF)

Fig. S3. Fractionation of spore particles from fresh (3 days) and old (30 days) Mm (carrying the pG13 plasmid) cultures on Percoll-sucrose gradients. (A) Relativefluorescence intensity of each fraction in arbitrary units are plotted; open and solid symbols represent 3- and 30-day-old cultures, respectively. All fractions weresubjected to wet-heat treatment at 65° C for 30 min and plated on 7H10 medium supplemented with 25 �g/mL kanamycin and incubated at 30° C. (B) Fractionsthat contained survivors after wet-heat treatment. Only the 30-day-old culture shows the presence of heat-resistant survivors: fractions 12 (97 colonies), 13 (255colonies), 15 (5 colonies), and 20 (3 colonies). Note that the position of the band that corresponds to the wet-heat-resistant particles shows a slightly heavierdensity compared with the main peak. The colonies were analyzed by using a Nikon SMZ1500 stereomicroscope with Epi-fluorescence attachment. An exposuretime of 400 ms was used, and images were captured by Nikon NIS Elements D3.0 software. This allowed us to scan for the presence or absence of fluorescenceoriginating from the expression of GFP in the entire colony, as shown in C and D. A 4-week-old Mm culture without pG13 was used as a negative control. (C)Colonies after wet-heat treatment of fraction 12. (D) One colony after wet-heat treatment. Together, C and D confirm the identity as the GFP plasmid-containingMm strain. Note that all wet-heat-resistant colonies from fractions 12, 13, 15, and 20 showed green fluorescence.

Ghosh et al. www.pnas.org/cgi/content/short/0904104106 5 of 5