Strategies Used by Pathogenic and Nonpathogenic Mycobacteria To Synthesize rrna

Transcription

1 JOURNAL OF BACTERIOLOGY, Nov. 1997, p Vol. 179, No /97/$ Copyright 1997, American Society for Microbiology Strategies Used by Pathogenic and Nonpathogenic Mycobacteria To Synthesize rrna J. A. GONZALEZ-Y-MERCHAND, 1 M. J. GARCIA, 2 S. GONZALEZ-RICO, 1 M. J. COLSTON, 1 AND R. A. COX 1 * Division of Mycobacterial Research, National Institute for Medical Research, The Ridgeway, Mill Hill, London NW7 1AA, United Kingdom, 1 and Departamento de Medicina Preventiva, Facultad de Medicina, Universidad Autonoma de Madrid, St. Arzobispo Morcillo 4, Madrid, Spain 2 Received 17 June 1997/Accepted 16 September 1997 One rrna operon of all mycobacteria studied so far is located downstream from a gene thought to code for the enzyme UDP-N-acetylglucosamine carboxyvinyl transferase (UNAcGCT), which is important to cell wall synthesis. This operon has been designated rrna f for fast-growing mycobacteria and rrna s for slow growers. We have investigated the upstream sequences and promoter activities of rrna f operons of typical fast growers which also possess a second rrn (rrnb f ) operon and of the rrna operons of the fast growers Mycobacterium abscessus and Mycobacterium chelonae, which each have a single rrn operon per genome. These fast growers have a common strategy for increasing the efficiency of transcription of their rrna operons, thereby increasing the cells potential for ribosome synthesis. This strategy involves the use of multiple (three to five) promoters which may have arisen through successive duplication events. Thus we have identified a hypervariable multiple promoter region (HMPR) located between the UNAcGCT gene and the 16S rrna coding region. Two promoters, P1 and PCL1, appear to play pivotal roles in mycobacterial rrna synthesis; they are present in all of the species examined and are the only promoters used for rrna synthesis by the pathogenic slow growers. P1 is located within the coding region of the UNAcGCT gene, and PCL1 has a characteristic sequence that is related to but distinct from that of the additional promoters. In fast-growing species, P1 and PCL1 produce less than 10% of rrna transcripts, so the additional promoters found in the HMPR are important in increasing the potential for rrna synthesis during rapid growth. In contrast, rrnb operons appear to be regulated by a single promoter; because less divergence has taken place, rrnb appears to be younger than rrna. Mycobacteria belong to the high guanosine-plus-cytosine branch of gram-positive bacteria. They are characterized by a complex cell envelope which includes many unusual glycolipids and mycolic acids, constituting a highly hydrophobic structure (22). The genus Mycobacterium can be conveniently subdivided into two categories based on growth rate. However, they are all very closely related as judged, for example, by the high levels of similarity (94% or more) between their 16S rrna gene sequences (26, 28, 33). The slow growers include the human pathogens Mycobacterium leprae and Mycobacterium tuberculosis, which have the ability to survive and grow within host cells. The mechanisms employed by these slow-growing pathogens to regulate growth rates are not understood. It is believed that growth requires strict control of a cell s capacity to synthesize proteins. For example, the number of ribosomes per cell is related to growth rate, as has been shown for M. tuberculosis (39). However, the molecular mechanisms used by mycobacteria to relate ribosome synthesis to growth rate are not known in detail. The use of conventional genetic approaches for the analysis of mechanisms of growth control is particularly difficult because of the slow growth rate. Comparison of homologous sequences from closely related mycobacterial species and from mycobacterial species which differ markedly in their growth rates provides a * Corresponding author. Phone: Fax: Permanent address: Departamento de Microbiologia, Escuela Nacional de Ciencias Biologicas, IPN, Mexico D.F , Mexico. Permanent address: Universidad Central De Venezuela, Instituto De Medicino Tropical, Dr. Felix Pifano C, Caracas, Venezuela. complementary approach (see, for example, references 14 and 15). The underlying principle is that highly conserved sequence motifs are likely to be important to function; in other words, the functional requirements of a motif may limit evolutionary divergence. Pathogenic and closely related slow-growing mycobacteria have a single rrna (rrn) operon of classical structure per genome, as shown previously for M. leprae (23, 24, 31), Mycobacterium bovis (35), and M. tuberculosis (18). These operons are considered to be members of one family, the rrna s operons, the subscript denoting slow growth (14). The fast-growing species Mycobacterium smegmatis has two rrn operons per genome (3), one of which is homologous to the rrna s operon and is designated rrna f (the subscript denoting fast growth); in all species examined to date, the rrna operon is located downstream of a gene encoding the enzyme UDP-N-acetylglucosamine carboxyvinyl transferase (UNAcGCT), which is involved in cell wall synthesis (11, 12). The second operon of M. smegmatis, designated rrnb f, was found downstream from an open reading frame (ORF) coding for tyrosyl-trna synthetase (12, 27). Thus one mechanism by which growth rates and ribosome synthesis are linked is gene dosage, with fast growers having two rrna operons and slow growers having only one. However, the possession of more than one operon per genome is not essential for rapid growth. For example, each of the fast growers Mycobacterium abscessus and Mycobacterium chelonae has a single rrn operon per genome (7). Production of rrna is also known to be regulated at the level of gene expression; thus although the rrna operons are homologous, the rrna f operon 6949

2 6950 GONZALEZ-Y-MERCHAND ET AL. J. BACTERIOL. of M. smegmatis has three promoters whereas the rrna s operon of M. tuberculosis has two (12). In the present study we have investigated the upstream sequences and promoter activities of three fast growers having two rrn operons per genome, Mycobacterium phlei, Mycobacterium fortuitum, and Mycobacterium neoaurum, and two fast growers having a single rrn operon per genome, M. abscessus and M. chelonae (7), and compared them with the archetypal slow grower M. tuberculosis and the most extensively studied fast grower, M. smegmatis. This has enabled us to further clarify the strategies used by mycobacteria to regulate rrna synthesis. MATERIALS AND METHODS Materials. All chemical reagents, primers, and commercial kits were described previously (10, 12). Bacterial strains. M. abscessus ATCC T (T denotes type strain), M. chelonae ATCC T, M. fortuitum ATCC 6841 T, M. neoaurum NCTC 10818, M. fortuitum NCTC 10394, M. phlei NCTC 8151, and M. smegmatis NCTC 8159 were maintained on Löwenstein-Jensen slopes at 4 C for short-term storage and grown for use in Lemco broth containing 0.1% (vol/vol) Tween 80. Isolation of DNA and RNA. Plasmid DNA (30), genomic DNA (11), and total RNA (12) were isolated by methods described previously. Construction of an M. chelonae minilibrary. M. chelonae DNA was prepared as described elsewhere (10). Standard techniques were used for cloning, plasmid isolation, restriction enzyme analysis, and colony blots (30). Plasmid puc18 (40) was used for cloning and sequencing of mycobacterial DNA. M. chelonae DNA was digested with restriction enzyme PstI. A minilibrary using puc18 as a vector was prepared with fragment sizes ranging from 2.5 to 3.5 kbp, and standard colony blot techniques were used for screening the minilibrary for a 16S rrna fragment. Amplification of the upstream regions of rrna operons by PCR. Bacterial DNA (1 to 100 ng) of M. phlei, M. fortuitum, and M. neoaurum was amplified by PCR (29) as described previously (16). The upstream regions of rrna operons, FIG. 1. Potential transcription start sites of rrn operons of typical fast growers. Both the primer extension and sequencing reactions were carried out with primer JY15 (see Materials and Methods). T, C, G, and A refer to sequencing reactions with ddttp, ddctp, ddgtp, and ddatp, respectively. tsa f, product attributed to transcripts directed by the indicated promoter of the rrna f operon; tsb f (P1), product of the rrnb f operon of M. smegmatis directed by the P1 promoter; tsb f (P1)?, possible product of the P1 promoter of an rrnb f operon; r, artifact of autoradiography; a and b, alternative starting points for PCL1 promoters. See the text for definitions of the CL1 region and PCL1 promoters. (a) Products corresponding to putative promoter elements of rrna f operons (see Fig. 3 to 5). Lane 1, M. phlei (M.ph) (autoradiography for 10 h); lanes 2 and 3, M. fortuitum (M.fo) (autoradiography for 16 h and 1 week, respectively) (the product adjacent to that designated tsa f (P3) is thought to originate from an alternative start site for the P3 promoter); lanes 4 and 5, M. neoaurum (M.ne) (autoradiography for 16 h and 1 week, respectively). (b) Direct comparison of nucleotide sequences and primer extension products. The products of the primer extension and sequencing reactions were separated by electrophoresis with a single gel. A section of the gel was chosen to highlight the presence in all three species of a product (shown by arrows) corresponding in size to the product [tsb f (P1)] directed by the P1 promoter of the rrnb f operon of M. smegmatis. Lane 1, primer extension product of M. smegmatis (M.sm) (autoradiography for 6 h); lane 2, M. phlei (autoradiography for 4 days); lane 3, M. fortuitum (autoradiography for 6 h); lane 4, M. neoaurum (autoradiography for 16 h). comprising approximately 114 codons of the gene for UNAcGCT, the promoter regions, and part (approximately 360 nucleotides) of the 16S rrna coding region, were synthesized with primer JG7 (5 CTG CAG CCG ATG GCT ATC GCT TG 3 ) in combination with RAC8 (5 CAC TGG TGC CTC CCG TAG G3 ). This combination was used previously to amplify the upstream regions of rrna operons of M. smegmatis and M. tuberculosis (12). The targets for primer JG7 correspond to the sequences complementary to positions 1 to 24 of the sequence upstream from the rrna operon of M. tuberculosis (12, 18). The targets for RAC8 correspond to positions 339 to 357 of the 16S rrna coding region of the rrna of M. tuberculosis (18). The upstream region of the rrna operon of M. abscessus was synthesized with the combination of primers RAC1 (5 TCG ATG ATC ACC GAG AAC GTG TTC 3 ) and RAC8. The binding site for RAC1 is 46 nucleotides downstream from the binding site for JG7. Comparison of the available sequences revealed that the region coding for the peptide SMITENVF is conserved and is thus a preferred target for PCR amplification. Cloning and sequencing. Escherichia coli DH5 was used as a host for the cloning of M. chelonae DNA. Competent cells were transformed by electroporation. The PCR products were ligated into pcr II and transformed into One Shot (INV F ) competent cells, as described previously (12). Transformant cells were selected on Luria-Bertani medium with the addition of ampicillin (100 g ml 1 ) in the presence of X-Gal IPTG (5-bromo-4-chloro-3-indolyl- -D-galactopyranoside isopropyl- -D-thiogalactopyranoside). DNA sequences were determined by the dideoxy chain termination procedure, as described previously (16), with universal primers (United States Biochemical Corp. and Pharmacia) and other appropriate primers (for examples, see reference 12). Where PCR products were used as a source of DNA, at least seven colonies were sequenced.

3 VOL. 179, 1997 MYCOBACTERIAL rrna OPERONS 6951 Primer extension studies. Oligonucleotide primer JY15 (5 CAC ACT ATT GAG TTC TC 3 ) has a target site close to the BoxA L motif, which is part of the CL2 region, described below, and which is present in all mycobacterial rrn operons studied so far. This primer was end labelled with [ - 32 P]ATP by T4 polynucleotide kinase, and the primer extension was carried out with avian myeloblastosis virus reverse transcriptase as described previously (12). Data bank searches and alignment of sequences. Computer-aided searches of data banks for the occurrence of nucleotide sequences from the leader region (CL2) of the 16S rrna gene and of sequences from the conserved region C of the gene coding for the enzyme UNAcGCT were carried out with the BLASTN program (1). Promoter sequences were aligned with the PILEUP program, which is part of the Genetics Computer Group sequence analysis software package (6). Nucleotide sequence accession numbers. The EMBL and GenBank accession numbers for the sequences determined in this study are as follows: M. phlei, X99776; M. fortuitum, X99775; M. neoaurum, X99777; M. chelonae, Y13911; M. abscessus, Y RESULTS Isolation and sequencing of upstream regions. The region upstream from the single rrn operon of M. chelonae was obtained by treatment of the genomic DNA with PstI endonuclease, separation of the fragments, and cloning of the appropriate 3-kbp fragment fraction in plasmid puc18 (see Materials and Methods). The region upstream from the single rrn FIG. 2. Potential transcription start sites of rrna operons of M. abscessus and M. chelonae. Primer extension and sequencing reactions were carried out with primer JY15 (see Materials and Methods). T, C, G, and A refer to sequencing reactions with ddttp, ddctp, ddgtp, and ddatp, respectively. ts, products attributed to pre-rrna transcripts of the rrna i operon directed by the indicated promoter. l, m, n, o, p, and q indicate products not attributable to promoters having consensus 10 and 35 boxes. a, a, b, and b indicate alternative starting points for the P4 promoter. Lanes: 1 and 3, control reaction mixture minus added RNA; 2, M. abscessus (M.ab) RNA; 4, M. chelonae (M.ch) RNA. (a) Electrophoresis for 1.5 h (8% [wt/vol] acrylamide). The autoradiograph was overexposed to reveal ts(p1). (b) Electrophoresis for 3 h (6% [wt/vol] acrylamide). In both panels, the products were separated with a single gel. operon of M. abscessus was obtained by PCR amplification with primers and methods described previously (11). The upstream regions of the rrna operons of M. phlei, M. fortuitum, and M. neoaurum were amplified by PCR and cloned as described above for M. abscessus (see Materials and Methods). The cloned upstream regions were sequenced by a strategy previously described for M. smegmatis (12). The identity of each species investigated was confirmed by analysis of the 16S rrna gene sequences. The variable V2 regions (positions approximately 138 to 220) are useful for this purpose (18). Analysis of transcription starting points. The transcription start sites of the rrna f and rrnb f operons of M. smegmatis were established by primer extension (12) and RNase protection assays (10a). These data provide a frame of reference for the assignment of possible promoter elements to primer extension products obtained for the other typical fast growers studied (Fig. 1). Primer JY15 yields products with transcripts of both the rrna f and rrnb f operons. Products corresponding to putative promoter elements (consensus 10 and 35 boxes) were identified as probable transcription starting points. Three starting points were identified in M. phlei, four in M. neoaurum, and five in M. fortuitum. In each case, one major product was found to correspond in size to the product of the single promoter of rrnb f of M. smegmatis (Fig. 1b). It was not possible to design

4 6952 GONZALEZ-Y-MERCHAND ET AL. J. BACTERIOL. FIG. 3. Summary of locations of the transcription starting points (tsp) of rrna operons. The hatched regions are defined in line a. The conserved sequences CL1 and CL2 and the variable VL and V2 regions are defined in the text. The HMPR extends from the 3 end of the UNAcGCT gene to the 5 end of the CL2 region. tsp indicated by solid arrows were established by primer extension studies (Fig. 1 and 2) and are numbered in the order (largest to smallest) of the sizes of the primer extension products. In each case the binding site for primer JY15 is located within the CL2 region. The tsp indicated by the broken arrow was identified by sequence similarities. The same scale, shown by the bar, was used throughout. primers which would distinguish between the rrna f and rrnb f operons. However, the correspondence in size with the product of rrnb f of M. smegmatis and the lack of correspondence with a recognizable promoter sequence in the rrna operon strongly suggest that these products are also derived from the rrnb f operon. The remaining products correspond to transcripts of the rrna f operons. The primer extension products obtained for M. chelonae and M. abscessus (Fig. 2) revealed five starting points in each case. Organization and analysis of rrna operons. The organization and analysis of rrna operons are summarized in Fig. 3. In all five species an ORF coding for a protein significantly similar to UNAcGCT was identified upstream from the 16S rrna gene, confirming that each operon is a member of the rrna family. Comparison of the sequences revealed two conserved elements, designated CL1 and CL2, located between the 3 ends of the ORFs and the 5 ends of the 16S rrna genes. The 19-bp CL1 element (consensus, 5 GGC AGG GTT GCC CCG AAA C 3 ) was first identified in M. tuberculosis, M. leprae, and six other slow growers (14). The CL2 element (5 TGT TGT TTG AGA ACT CAA TAG TGT GT 3 ) has the 12-bp motif coding for the putative BoxA L element, thought to be involved in ensuring complete transcription of the operon by diminishing premature termination (5), at the 5 end. This motif corresponds to the first 26 bp of the 31-bp invariant region which is present in all eight of the slow growers investigated by Ji et al. (14) and in the four typical fast growers included in this study, namely, M. phlei, M. smegmatis, M. fortuitum, and M. neoaurum. The CL2 motif appears to be confined to members of the genus Mycobacterium. A search of the databases revealed 50 related bacterial sequences. Fifteen perfect matches were identified, and all of them were found in the leader regions of mycobacterial rrn operons. The remaining 35 sequences, of approximately 65% similarity, were also found to be mycobacterial in origin and to be centered on the BoxA S motif of the spacer 1b region separating the 16S rrna and 23S rrna genes. The BoxA L and BoxA S sequences are closely related (15, 16). The number of transcription starting points per operon was found to correlate with the distances between the 3 ends of the ORFs and the 5 ends of the CL2 motifs (Fig. 3); these sections were designated hypervariable multiple promoter regions (HMPRs) (see below). The number of transcription starting points per operon, the number of operons per genome, the sizes of the HMPRs, and evolutionary distances are summarized in Table 1.

5 VOL. 179, 1997 MYCOBACTERIAL rrna OPERONS 6953 TABLE 1. Comparison of growth rates of mycobacterial species with evolutionary (Hamming) distances, the numbers of rrn operons per genome, the numbers of promoters per rrna operon, and the sizes of the HMPRs Species a Hamming distance (16S rrna) b No. of rrn operons/ genome No. of promoters/ rrna operon Size of HMPR (bp) Growth rate on solid medium c M. phlei M. smegmatis d M. fortuitum M. neoaurum M. chelonae M. abscessus M. tuberculosis d a All are fast growers except M. tuberculosis. b Based on the assumption that the 16S rrna sequences of M. bovis and M. tuberculosis are identical (32). The M. abscessus sequence is used for reference. c Values are days until colonies become visible. Data for all species except M. tuberculosis are from Wayne and Kubica (38). Data for M. tuberculosis is from Colston (4a). d See reference 12. Putative promoter sequences. The process of initiation of transcription is believed to involve a sequence extending from 55 to 25 bp from the start of transcription, a region of approximately 80 bp (21) as judged by the properties of typical E. coli promoters (for reviews, see references 9 and 20). Mycobacterial rpo (13) and sigma factors (8) are known to be homologous with their E. coli counterparts; hence the E. coli model is expected to be a guide to mycobacterial promoter functions (Fig. 4a). Having identified the starting points of transcription (see above), it was possible to identify putative promoter sequences, which are compared in Fig. 4 and 5. It is possible that some of the primer extension products may have resulted from RNA processing or degradation and that not all rrna promoters require the same sigma factor and hence may not have easily recognizable sequence features. Thus these should be considered putative promoters until further functional analysis has been carried out. However, evidence for the P1 and PCL1 promoters of M. tuberculosis is also based on previously published experiments using a promoterless reporter gene (14, 37). Two sets of promoters were found to be present in all of the species studied and were equivalent to the two promoters of the rrna s operon previously characterized for M. tuberculosis (12). One set, the P1 promoters, located furthest upstream from the 5 ends of the 16S rrna genes, lies within the coding regions of the genes thought to code for UNAcGCT. In all cases except M. chelonae and M. abscessus, the transcription starting points are located immediately downstream of the ORFs. In M. chelonae and M. abscessus, the transcription starting points associated with the P1 promoter are located within the ORFs (Fig. 3). In all cases, typical 10 boxes were identified but no typical 35 box was found (Fig. 4b). Each promoter of the second set, the PCL1 promoters (Fig. 4c), is characterized by having the conserved CL1 motif contiguous with the 10 box. These promoters were designated P3 in our earlier paper (12). The set of 10 and 35 boxes is typical of promoters requiring the initiation factor sigma 70 or its equivalent. The location of PCL1 promoters is not fixed. For example, in M. chelonae and M. abscessus these promoters are close to the genes thought to code for UNAcGCT, whereas in each of the other species the promoter is located further downstream (Fig. 3). The two promoters (P1 and PCL1) of M. tuberculosis are separated by 77 bp (Table 2). The promoter sequences (Fig. 4) do not overlap, indicating that each promoter can form an initiation complex with rpo independently of the other. However, there may be an indirect influence of one promoter on the other. For example, an elongation complex formed at P1 will transcribe the PCL1 sequence, and the passage of the rpo complex will inactivate the promoter, thereby reducing the time available for it to form initiation complexes with rpo. Conversely, the engagement of PCL1 in the formation of an initiation complex is expected to block the passage of upstream elongation complexes. These steric effects become more pronounced as the rate of formation of initiation complexes increases. Additional promoters of fast-growing species giving rise to the HMPR. As illustrated in Fig. 3, the rrna operons of all of the fast growers studied have more than the two promoters mentioned above. M. phlei, M. smegmatis, and M. neoaurum have one promoter (P2), and M. fortuitum has two promoters (P2 and P3) located between promoters P1 and PCL1. The extended promoter sequences of the additional P2 promoters are compared in Fig. 5a. The P3 promoter of M. fortuitum resembles the PCL1 promoter (Fig. 4c) more than it does the P2 promoter (Fig. 5a). Interestingly the two fast-growing species which have a single rrn operon, M. chelonae and M. abscessus, each have five promoters, giving rise to extended HMPRs. In addition to P1 and PCL1, M. chelonae has three promoters, P2, P3, and P4, between the two conserved elements, CL1 and CL2 (Fig. 3). There is high homology (80% or more) among P2, P3, P4, and PCL1, suggesting that sequence replication has occurred. The four repeat (r) sequences (r1 to r4), each of approximately 130 bp, were aligned in order of their similarities with the PILEUP program, and the sequences of the three promoters located within the repeats (corresponding to transcription starting points 3, 4, and 5) are compared in Fig. 5b in the order identified by the PILEUP program. The rrna operon of M. abscessus also has five promoters. The sequence data reveal five repeat sequences (r1 to r5), each of approximately 130 bp. The replicate sequence, r2, located between transcription starting points 2 and 3 (Fig. 3) was found to lack promoter activity even though it closely resembles the other promoter sequences (Fig. 5c). The replicate r2 has hexamers which each differ from the 10 and 35 boxes of the other promoters, namely, 5 TATAAC 3 compared with 5 TACAGT 3 and 5 GTGACT 3 compared with 5 TTGACT 3 (Fig. 5c). Comparison of Fig. 5b and c reveals that the replicate promoter sequences of M. chelonae and M. abscessus are very similar, as might be expected because the two species are known to be very closely related (Table 1). The ten additional promoters presented in Fig. 5 have conserved sequence motifs other than their 10 and 35 boxes in common. One motif (5 RACCRG 3 ) is centered close to position 46 (e.g., 48 to 43). Another motif (5 CR 2 7 CG) is centered on approximately position 12 (e.g., 9 to 15). In the PCL1, P2, P3, and P4 promoters of M. chelonae and M. abscessus, the doublet TG, which is found in positions 15 and 14, may be important to function (4, 17, 19). The features common to all of the additional promoters, which span approximately 70 bp, are summarized in Fig. 5d. Promoter P3 of M. fortuitum (Fig. 4b) also conforms to this consensus promoter (Fig. 5d). The 35 boxes and 10 boxes of the additional promoters, like PCL1 promoters, are separated by 18 bp in most cases. Thus in 17 promoters the separation is 18 bp, in one promoter the separation is 17 bp, and in one other promoter the separation is 16 bp.

6 6954 GONZALEZ-Y-MERCHAND ET AL. J. BACTERIOL. FIG. 4. Comparison of promoters present in all rrna operons investigated. The numbers denote nucleotide positions given in the appropriate GenBank entry. Dots indicate identical nucleotides; dashes indicate deletions. Asterisks indicate transcription starting points (tsp) identified by the primer extension method (Fig. 1 and 2). Crosses indicate additional (usually minor) tsp. N x denotes a sequence of base pairs of the length specified by the subscript. The locations of the tsp are shown in Fig. 3. Putative mycobacterial promoter elements are boxed. The promoters of M. tuberculosis were identified by Gonzalez-y-Merchand et al. (12). (a) Functional regions of putative mycobacterial promoters based on the E. coli model (9, 20, 21). The regions of sigma factor 70 believed to bind to the 35 and 10 boxes are indicated (9). Nucleotide positions are measured from the tsp. Uncertainty about the rpo binding region is indicated by a broken line; for an explanation see the text. USR, upstream region; DSR, downstream region. (b) P1 promoters. Uncertainty about the location of 35 boxes is shown by broken lines. The termination codons of the putative genes for UNAcGCT are underlined. (c) PCL1 promoters. The P3 promoter (corresponding to tsp3) of M. fortuitum, which possesses a CL1-like motif, is included for comparison. Abbreviations: M.ph, M. phlei; M.sm, M. smegmatis; M.fo, M. fortuitum; M.ne, M. neoaurum; M.ch, M. chelonae; M.ab, M. abscessus; M.tu, M. tuberculosis. Relative activities of promoters. The primer extension data (Fig. 1 and 2) suggest that the three categories of promoters, P1, PCL1 and P2/3/4, differ in their activities. The number of transcripts originating from a particular starting point is, depending on stability, directly proportional to the radioactivity of the end-labelled primer extension product. The relative activities of the promoters of rrna operons growing exponentially in rich media are summarized in Table 2. In general the P1 and PCL1 promoters, which are present in all mycobacterial rrna operons, were found to be less active than the additional promoters, P2, P3, and P4. The VL region. The VL region, which extends from the 3 end of the CL2 motif to the 5 end of the 16S rrna gene (Fig. 3), ranges in size from 66 bp (M. chelonae) to 159 bp (M. neoaurum). The differences in the sizes of the VL regions are reflected in their capacities to form stem-loop structures. Possible structures that may have been formed by the time RNA polymerase had proceeded to the 5 ends of the 16S rrna genes are compared in Fig. 6. A feature common to all mycobacterial rrn operons studied to date (see also references 14 to 16) is helix L1, which has a pyrimidine-rich loop of at least eight residues that includes the conserved motif 5 YYYG 3 (Fig. 6). The VL regions of M. chelonae and M. abscessus do not have the capacity to form stem-loop structures downstream from helix L1; in contrast, the representative fast grower, M. smegmatis, can form four additional stem-loop structures, and the pathogen M. tuberculosis can form two. The secondary structure of component motifs of pre-rrna may be altered as transcription proceeds. For example, helix 1 (Fig. 6) is believed to be modified in mature 16S rrna (18) through base pairing with more distant residues, namely, between residues 29 to 39 and 537 to 545 and between residues 19 to 21 and 908 to 910. Once transcription has proceeded beyond the 3 end of the 16S rrna gene, a sequence of approximately 50 bp, which is believed to form a bihelical stem structure with the CL2 motif and part of the VL region, is synthesized. The 5 end of the CL2 motif marks the beginning and the 5 end of helix L1 marks the end of this stem (16). We infer that because helix L1 is highly conserved it plays an essential functional role which has yet to be identified. A transient molecular scaffold function has been proposed for at least part of the VL region of the rrnb operon of E. coli (36); that is, early in the transcription of the operon the VL region binds

7 VOL. 179, 1997 MYCOBACTERIAL rrna OPERONS 6955 TABLE 2. Separation and relative activities of the promoters of rrna operons Species a TSP b Promoter (separation) c M. phlei 1 P1 2 P2 (96) CL1 PCL1 (85) proteins needed for the formation of the 30S subunit which are absent from the mature subparticle. DISCUSSION Relative abundance of transcripts d M. smegmatis 1 P1 2 P2 (94) CL1 PCL1 (76) M. fortuitum 1 P1 2 P2 (86) 3 P3 (78) CL1 PCL1 (76) M. neoaurum 1 P1 2 P2 (76) CL1 PCL1 (67) M. chelonae 1 P1 CL1 PCL1 (130) 2 P2 (126) 3 P3 (127) 4 P4 (128) M. abscessus 1 P1 CL1 PCL1 (130) 2 P2 (255) 3 P3 (130) 4 P4 (147) M. tuberculosis 1 P1 CL1 PCL1 (77) a M. phlei, M. smegmatis, M. fortuitum, and M. neoaurum are fast growers with two rrn operons per genome. M. chelonae and M. abscessus are fast growers with one rrn operon per genome. M. tuberculosis is a slow grower with one rrn operon per genome. b TSP, transcription starting point. c Separation is expressed as base pairs between a given promoter and the promoter immediately above it. d Data are for exponential growth in rich media (see Materials and Methods). Relative abundance of transcripts is directly proportional to radioactivity, since each transcript was labelled only at the 5 end. The rate at which a cell can synthesize rrna determines the rate at which it can assemble ribosomes and synthesize proteins. The maximum rate at which rrna synthesis can proceed sets an upper limit on the rate at which a cell can grow. Mycobacteria have evolved strategies for maximizing the efficiency and flexibility of rrna synthesis with a minimal number of rrn operons. Slow growers, such as M. tuberculosis, have a single rrn operon with multiple promoters. Typical fast growers have two (rrna and rrnb) operons per genome, and at least one of these (rrna) has multiple promoters. Interestingly, some species, such as M. chelonae and M. abscessus, have single rrn operons but are nevertheless classified as fast growers; here we show that these species appear to have acquired additional promoters by a process of sequence duplication. Thus mycobacteria have at least two levels at which rrna synthesis is regulated. In this study we have mapped the transcription starting points of the rrna operons of several species of mycobacteria. Analysis of sequences immediately upstream of these starting points has revealed promoter-like sequences. Although functional studies of mycobacterial promoters are relatively few, some general principles are beginning to emerge. While the 10 consensus sequence of E. coli appears to be conserved in mycobacterial promoters, a larger variety of sequences can be accommodated in the 35 region (2, 19, 25). Organisms of the closely related genus Streptomyces, also belonging to the high G C branch of gram-positive bacteria, can also tolerate a large variety of sequences in the 35 regions of their promoters (34). The principal sigma factors of M. smegmatis, M. tuberculosis, and M. leprae are nearly identical to the principal sigma factor of Streptomyces aureofaciens; while they are also nearly identical to the principal sigma factor of E. coli (RpoD) in the region responsible for binding to the 10 box, they differ substantially in the region involved in binding to the 35 box (2). In this study we have also found that while 10 consensuslike sequences could readily be identified, this was not always the case for the 35 sequence. The sequences of 25 promoters identified in this study are presented in Fig. 4 and 5. Promoters P1 and PCL1 of mycobacterial rrna operons appear to play a pivotal role in rrna synthesis because they are present in all of the slow growers and fast growers studied thus far (Fig. 3). These two promoters regulate the expression of the rrna s operons of M. tuberculosis and other slow growers, whereas additional promoters are present in the rrna f operons of fast growers (Fig. 3). A feature of the P1 promoter is its location within the coding region of a gene coding for an enzyme (UNAcGCT) important to cell wall synthesis, which suggests the possibility that regulation of cell wall synthesis and that of rrna synthesis are linked. The P1 promoters appear to be weaker than those of other categories, such as PCL1 and the additional promoters (Table 2), and features such as 35 boxes, upstream regions, and downstream regions are more difficult to identify and to relate to the model promoter (Fig. 4a). Compared with M. tuberculosis, each of the typical fast growers has an increased potential for pre-rrna synthesis which is achieved by two strategies. First, each species has two operons (rrna and rrnb) per genome; second, the rrna (rrna f ) operon has at least one strong promoter (Fig. 5a) in addition to promoters P1 and PCL1. M. fortuitum has two additional promoters; one of them (P3) has more in common (45 of 70 identities) with promoter PCL1 than with promoter P2 (33 of 70 identities). In light of the evidence for promoter duplication within the rrn operons of M. abscessus and M. chelonae (Fig. 5b and c; also see below), we suggest that P3 may have originated by duplication of PCL1 and subsequently diverged. The expression of both rrn operons of M. smegmatis was studied previously (12). The rrnb operon, which is located downstream from a gene coding for tyrosyl trna synthetase (27), was found to be regulated by a single promoter (P1). This promoter has features in common with the P2 promoter of the rrna operons of typical fast growers, especially with the P2 promoter of the rrna operon of M. smegmatis (Fig. 5e). The shared features include 14 of 16 identities in the sequences immediately upstream from the 35 boxes and near identical 10 boxes; three A TorT A base pairs immediately downstream from each 10 box; and very similar (6 of 7 identities) putative downstream regions. There is no sequence corresponding to the CL1 motif in the rrnb operon. Control of the expression of the rrnb operon by a single promoter is in contrast to the multiple promoters identified in rrna operons and may reflect the fact that the rrnb operon is much younger than rrna and that promoter duplication has not yet occurred.

8 6956 GONZALEZ-Y-MERCHAND ET AL. J. BACTERIOL. FIG. 5. Fast growers have promoters in addition to P1 and PCL1. For explanations of symbols and abbreviations, see the legend to Fig. 4. Note that the rrna operon of M. tuberculosis has no promoters in addition to P1 and PCL1. In sections b and c, sequences were aligned by using PILEUP (see Materials and Methods). USR, upstream region; DSR, downstream region. In sections a to c, the consensus sequence is based on the set of mycobacterial sequences shown. (a) P2 promoters of fast growers with two rrn operons per genome. (b) Promoters P2, P3, and P4 of M. chelonae. These promoters, including PCL1, are located within the repeats (r1 to r4) which are found in the region of positions 278 to 762. (c) Promoters P2, P3, and P4 of M. abscessus. The repeats (r1 to r5) are found in the region of positions 277 to 891. Repeat r2 appears to be inactive as a promoter (Fig. 2). (d) Summary sequence for promoters (see above) in addition to P1 and PCL1. The extended sequence spans the region of a promoter believed to be engaged in the formation of an open complex with rpo (21). N, either A, C, G, or T; R, either A or G; Y, either C or T; N, nucleotide present in some but not all promoters. (e) Comparison of the P1 promoter of the rrnb operon of M. smegmatis with the P2 promoter (see section a). The numbers for the P1 promoter refer to the nucleotide positions in the sequence cited by Ji et al. (16). The broken underlining denotes the putative downstream region. M. abscessus and M. chelonae are unusual in that they each have a single rrn operon (7) and yet are classified as fast growers. Thus their potential for increased pre-rna synthesis is achieved by the acquisition of additional promoters within the rrna operon. Because of their distinctive features (Fig. 3), the rrn operons of M. abscessus and M. chelonae are designated members of the rrna i subgroup; the subscript denotes an intermediate position between the rrna s operons typified by M. tuberculosis and the rrna f operons typified by M. smegmatis, etc. (Fig. 3). The rrna i promoters are characterized by more than two promoters (for example, P1, PCL1, P2, P3, and P4), by a different location for P1 within the upstream ORF, by greater distances between promoters (approximately 130 bp, compared with 77 bp for M. tuberculosis), and by an abbreviated VL region (Fig. 3 and 6). In the case of both M. chelonae and M. abscessus, the additional promoters P2, P3, and P4 are very similar in sequence to each other (Fig. 5); they are also related to PCL1 (Fig. 4c). These sequence similarities suggest that the additional promoters may have arisen through successive replication events, starting with the replication of PCL1. Promoters P2, P3, and P4, which closely resemble one another, have approximately equal strengths, and they initiate the majority of pre-rrna transcripts (Fig. 2; Table 2). The additional promoters have the principal features of strong E. coli promoters (compare Fig. 4a and 5d). The slow-growing M. tuberculosis relies entirely on the activities of promoters P1 and PCL1. Fast growers have additional promoters with much greater relative activities than P1 and PCL1. For example, in the fast-growing M. smegmatis, P1 and PCL1 contribute only about 7.5% of the total number of pre-rrna transcripts (10a). This observation suggests that under optimum conditions for balanced growth a slow grower such as M. tuberculosis will synthesize its complement of ribosomes at less than 1/10 of the rate observed for the fast growers studied. However, M. tuberculosis has approximately half the number of ribosomes possessed by M. smegmatis (39). M. smegmatis, when grown in a rich medium, was found to have a generation time of about 2 h (10a). Thus, pro rata, M. tuberculosis would be expected to double its complement of ribosomes in approximately 13.4 h; the reported generation time (38) is 14 to 15 hours, indicating that a high proportion of the cell s capacity for ribosome synthesis is used under these conditions of growth. Thus the HMPRs of the rrna operons of mycobacteria con-

9 VOL. 179, 1997 MYCOBACTERIAL rrna OPERONS 6957 FIG. 6. Possible secondary structures of the regions of pre-rrnas close to the 5 ends of 16S rrnas. Possible interactions with proteins such as antitermination factors are not considered. The nucleotide positions of the leader region refer to the appropriate GenBank entry; the nucleotide positions of mature 16S rrna are designated 10m, etc. The 5 ends of 16S rrnas are boxed. Features of the leader region are identified by the prefix or subscript L. Lowercase letters indicate nucleotides that are highly conserved among both fast and slow growers; bold italic letters denote nucleotides conserved in the fast growers M. phlei, M. smegmatis, M. fortuitum, and M. neoaurum. The conserved sequences CL1 and CL2 are defined in the text (see also Fig. 3). The BoxA L element, which is enclosed by broken lines, is located at the 5 end of the CL2 region. In helix L1, the lowercase letters indicate the sequence YYYG, which is found in all mycobacteria studied so far (16). (a) M. smegmatis. Structures homologous with those of M. phlei, M. fortuitum, and M. neoaurum can be observed. (b) M. abscessus. A very similar structure can be derived for M. chelonae. (c) M. tuberculosis. Similar structures can be derived for all of the other slow growers investigated (14). tain a wealth of phylogenetic and functional information. In this study we have demonstrated that growth rates and ribosome synthesis may be linked by a number of different strategies, including gene dosage and differential gene regulation, and that evolutionary changes within the HMPRs have contributed to increases in the efficiency and flexibility of rrna synthesis. Thus the fast growers have developed ways of increasing their capacity for rrna synthesis. An understanding of how mycobacteria utilize this flexibility in response to different conditions and stimuli should provide new insights into how this important group of bacteria respond to a diverse range of environmental conditions. ACKNOWLEDGMENTS M.J.G. received financial support from the Consejeria de EyC Comunidad de Madrid, Spain. J.A.G.y.M. received financial support from COFAA and EDD, IPN, Mexico. This work is supported as part of the European Commission Science Research and Development Programme, proposal number ERBIC 18PL REFERENCES 1. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman Basic local alignment search tool. J. Mol. Biol. 215: Bashyam, M. D., D. Kaushal, S. K. Dasgupta, and A. K. Tyagi A study of the mycobacterial transcriptional apparatus: identification of novel features in promoter elements. J. Bacteriol. 178: Bercovier, H., O. Kafri, and S. Sela Mycobacteria possess a surprisingly small number of ribosomal RNA genes in relation to the size of their genome. Biochem. Biophys. Res. Commun. 136: Chan, B., and S. Busby Recognition of nucleotide sequences at the Escherichia coli galactose operon P1 promoter by RNA polymerase. Gene 84: a.Colston, M. J. Unpublished data. 5. Condon, C., C. Squires, and C. L. Squires Control of rrna transcription in Escherichia coli. Microbiol. Rev. 59: Devereux, J., P. Haeberli, and O. Smithies A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12: Domenech, P., M. C. Menendez, and M. J. Garcia Restriction fragment length polymorphism of 16S rrna genes in the differentiation of fast-growing mycobacterial species. FEMS Microbiol. Lett. 116: Doukhan, L., M. Predich, G. Nair, D. Dussurget, I. Mandic-Mulec, S. T. Cole, D. R. Smith, and I. Smith Genomic organization of the mycobacterial sigma gene cluster. Gene 165: Fassler, J. S., and G. N. Gussin Promoters and basal transcription machinery in eubacteria and eukaryotes: concepts, definitions and analogues. Methods Enzymol. 273: Garcia, M. J., C. Guilhot, R. Lathigra, C. Menendez, P. Domenech, C. Moreno, B. Gicquel, and C. Martin Insertion sequence IS1137, a new IS3 family element from Mycobacterium smegmatis. Microbiology 140:

10 6958 GONZALEZ-Y-MERCHAND ET AL. J. BACTERIOL. 10a.Gonzalez-y-Merchand, J. A., M. J. Colston, and R. A. Cox. Unpublished data. 11. Gonzalez-y-Merchand, J. A., I. Estrada-Garcia, M. J. Colston, and R. A. Cox A novel method for the isolation of mycobacterial DNA. FEMS Microbiol. Lett. 135: Gonzalez-y-Merchand, J. A., M. J. Colston, and R. A. Cox The rrna operons of Mycobacterium smegmatis and Mycobacterium tuberculosis: comparison of promoter elements and of neighbouring upstream genes. Microbiology 142: Honoré, N. T., S. Bergh, S. Chanteau, F. Doucer-Populaire, K. Eiglmeier, T. Garnier, C. Georges, P. Launois, T. Limpaiboon, S. Newton, K. Niang, P. del Portillo, G. R. Ramesh, P. Reddi, P. R. Ridel, N. Sittisombut, S. Wu-Hunter, and S. T. Cole Nucleotide sequence of the first cosmid from the Mycobacterium leprae genome project: structure and function of the Rif-Str regions. Mol. Microbiol. 7: Ji, Y.-E., M. J. Colston, and R. A. Cox Nucleotide sequence and secondary structures of precursor 16S rrna of slow-growing mycobacteria. Microbiology 140: Ji, Y.-E., K. Kempsell, M. J. Colston, and R. A. Cox Nucleotide sequences of the spacer-1, spacer-2 and trailer regions of the rrn operons and secondary structures of precursor 23S rrnas and precursor 5S rrnas of slow-growing mycobacteria. Microbiology 140: Ji, Y.-E., M. J. Colston, and R. A. Cox The ribosomal RNA (rrn) operons of fast-growing mycobacteria: primary and secondary structures and their relation to rrn operons of pathogenic slow-growers. Microbiology 140: Keilty, S., and M. Rosenberg Constitutive function of a positively regulated promoter reveals new sequences essential for activity. J. Biol. Chem. 262: Kempsell, K. E., Y.-E. Ji, I. C. E. Estrada-Garcia, M. J. Colston, and R. A. Cox The nucleotide sequence of the promoter, 16S rrna and spacer region of the ribosomal RNA operon of Mycobacterium tuberculosis and comparison with Mycobacterium leprae precursor rrna. J. Gen. Microbiol. 138: Kenney, T. J., and G. Churchward Genetic analysis of the Mycobacterium smegmatis rpsl promoter. J. Bacteriol. 178: Knaus, R., and H. Bujard Principles governing the activity of E. coli promoters. Nucleic Acids Mol. Biol. 4: Krummel, B., and M. J. Chamberlin RNA chain initiation by Escherichia coli RNA polymerase. Structural transitions of the enzyme in early ternary complexes. Biochemistry 28: Lederer, E., A. Adam, R. Ciorbaru, J. F. Petit, and J. Wietzerbin Cell walls of mycobacteria and related organisms: chemistry and immunostimulant properties. Mol. Cell. Biochem. 7: Liesack, W., C. Pitulle, S. Sela, and E. Stackebrandt Nucleotide sequence of the 16S rrna from Mycobacterium leprae. Nucleic Acids Res. 18: Liesack, W., S. Sela, H. Bercovier, C. Pitulle, and E. Stackebrandt Complete nucleotide sequence of the Mycobacterium leprae 23S and 5S rrna genes plus flanking regions and their potential in designing diagnostic oligonucleotide probes. FEBS Lett. 281: Movahedzadeh, F., M. J. Colston, and E. O. Davis Determination of DNA sequences required for regulated Mycobacterium tuberculosis RecA expression in response to DNA-damaging agents suggests that two modes of regulation exist. J. Bacteriol. 179: Pitulle, C., M. Dorsch, J. Kazda, and E. Stackebrandt Phylogeny of rapidly growing members of the genus Mycobacterium. Int. J. Syst. Bacteriol. 42: Predich, M., L. Doukhan, G. Nair, and I. Smith Characterization of RNA polymerase and two sigma-factor genes from Mycobacterium smegmatis. Mol. Microbiol. 15: Rogall, T., J. Wolters, T. Flohr, and E. C. Böttger Towards a phylogeny and definition of species at the molecular level within the genus Mycobacterium. Int. J. Syst. Bacteriol. 40: Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Sharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Erlich Primer directed enzymatic amplification of DNA with a thermostable polymerase. Science 239: Sambrook, J., E. F. Fritsch, and T. Maniatis Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 31. Sela, S., and J. E. Clark-Curtiss Cloning and characterization of the Mycobacterium leprae putative ribosomal RNA promoter in Escherichia coli. Gene 98: Springer, B., E. C. Böttger, P. Kirschner, and R. J. Wallace, Jr Phylogeny of the Mycobacterium chelonae-like organism based on partial sequencing of the 16S rrna gene and proposal of Mycobacterium mucogenicum sp. nov. Int. J. Syst. Bacteriol. 45: Stahl, D. A., and J. W. Urbance The division between fast- and slow-growing species corresponds to natural relationships among the mycobacteria. J. Bacteriol. 172: Strohl, W. R Compilation and analysis of DNA sequences associated with apparent streptomycete promoters. Nucleic Acids Res. 20: Suzuki, Y., A. Nagata, Y. Ono, and I. Yamada Complete nucleotide sequence of the 16S rrna gene of Mycobacterium bovis BCG. J. Bacteriol. 170: Theissen, G., L. Thelen, and R. Wagner Some base substitutions in the leader of an Escherichia coli ribosomal RNA operon affect the structure and function of ribosomes. Evidence for a transient scaffold function of the rrna leader. J. Mol. Biol. 233: Verma, A., A. K. Kringer, and J. S. Tyagi Functional analysis of transcription of the Mycobacterium tuberculosis 16S rdna-encoding gene. Gene 148: Wayne, L. G., and G. P. Kubica The mycobacteria, p In P. H. A. Sneath, N. S. Mair, M. E. Sharpe, and J. G. Holt (ed.), Bergey s manual of systematic bacteriology, vol. 2. The Williams & Wilkins Co., Baltimore, Md. 39. Winder, F. G., and S. A. Rooney Effects of nitrogenous components of the medium on the carbohydrate and nucleic acid content of Mycobacterium tuberculosis BCG. J. Gen. Microbiol. 63: Yanisch-Perron, C., J. Vieira, and J. Messing Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and puc19 vectors. Gene 33: