|Home | About | Journals | Submit | Contact Us | Français|
All mycobacteria studied to date have an rRNA operon, designated rrnA, located downstream from a single copy of the murA gene, which encodes an enzyme (EC 18.104.22.168) important for peptidoglycan synthesis. The rrnA operon has a promoter, P1(A), located within the coding region of murA, near the 3′ end. Samples of RNA were isolated from Mycobacterium tuberculosis at different stages of the growth cycle and from Mycobacterium smegmatis grown under different conditions. RNase protection assays were used to investigate transcripts of both murA and rrnA. Transcription of murA was found to continue into the 16S rRNA gene, as if murA and rrnA form a hybrid (protein coding-rRNA coding) operon. During the growth of M. tuberculosis, the hybrid operon contributed approximately 2% to total pre-rRNA. Analysis of M. smegmatis RNA revealed that the level of murA RNA depended on the growth rate and that the patterns of expression during the growth cycle were different for murA and rrnA. M. smegmatis has a second rRNA operon, rrnB, located downstream from a single copy of the tyrS gene, encoding tyrosyl-tRNA synthetase. Transcription of tyrS was found to continue into the 16S rRNA gene rrnB. The hybrid tyrS-rrnB operon contributed 0.2 to 0.6% to rrnB transcripts. The pattern of tyrS expression during the growth cycle matched the pattern of rrnB expression, reflecting the essential role of TyrS and rRNA in protein biosynthesis.
All the mycobacteria studied to date have either one or two rrn operons, designated rrnA and rrnB. The rrnA operon, which is present in all mycobacteria, is located downstream from the murA gene, which encodes the enzyme UDP-N-acetylglucosamine 1-carboxyvinyl transferase (EC 22.214.171.124), or MurA (10). This enzyme catalyzes the transfer of the enol ether from phosphoenolpyruvate to the 3′ OH of UDP-N-acetylglucosamine during the early stages of peptidoglycan synthesis. Mycobacterium smegmatis, which is representative of many fast-growing mycobacteria (10), has a second rrn operon (rrnB), located downstream from the tyrS gene, encoding the enzyme tyrosyl-tRNA synthetase (EC 126.96.36.199), or TyrS (8, 15). TyrS is essential for protein biosynthesis; its role is the attachment of l-tyrosine to the ribose moiety of the 3′-terminal adenosine residue of tyrosyl-tRNA.
Each of the known rrnA operons has a promoter, P1(A), which has a transcription start point (tsp) situated within the coding region of murA or no more than 3 nucleotides downstream from the stop codon (10) (Fig. (Fig.1).1). Thus, a region of the murA gene near the 3′ end is transcribed both as part of the murA gene and as part of the rrnA operon, raising the possibility that murA and rrnA are transcribed as a single unit. Similarly, it is possible that tyrS and rrnB are transcribed as a single unit because of the absence of a transcription terminator in the 54-bp region downstream from the 3′ end of the tyrS coding region and the tsp for rrnB (8, 15).
In this study, the RNA fraction was isolated from Mycobacterium tuberculosis and M. smegmatis at different stages of growth and from M. smegmatis grown in two different media. RNase protection assays were used to investigate the extents of coordination of the expression of murA with rrnA and of tyrS with rrnB.
A Sequenase (U.S. Biochemicals) sequencing kit was supplied by Cambridge Biosciences. [α-35S]dATP was obtained from Amersham. A GeneClean kit was obtained from Bio 101. Oligonucleotide primers were prepared with an automated DNA synthesizer (model 370A; Applied Biosystems).
M. smegmatis NCTC 8159 (National Collection of Type Cultures) was maintained on Löwenstein-Jensen slopes and grown at 37°C with vigorous shaking in either Lemco broth (3) or Kohn-Harris glucose medium (described below) containing 0.1% Tween 80. Seed cultures of M. smegmatis for inoculation were grown in Lemco broth for 26 h. These cultures were used to inoculate medium at the rate of 20 ml of inoculum/liter. Kohn-Harris glucose medium was based (6) on the medium of Kohn and Harris (13) with glucose (5 g/liter) as the carbon source and trace elements provided at 5 ml/liter as described by Kelly and Clarke (12). M. tuberculosis H37Rv was grown in Dubos medium (7) containing 0.1% Tween 80.
Cells were collected, resuspended in 1 ml of guanidinium buffer (6 M guanidinium chloride, 0.1% [vol/vol] Tween 80, 10 mM EDTA, 1 mM 2-mercaptoethanol), and kept at −20°C for 15 min. The suspension was added to half the volume of heat-sterilized glass beads (0.15-mm diameter) contained in a 2-ml screw-cap microcentrifuge tube. Mycobacteria were ruptured by three pulses of 1 min each on a Mini-BeadBeater device (Biospec Products); debris and beads were sedimented by centrifugation (10,000 × g for 3 min), and the cleared lysate was retained. The pellet of beads and mycobacterial residue was briefly reextracted on the Mini-BeadBeater (30-s pulse) with 300 μl of fresh guanidinium buffer, and the resulting extract was pooled with the first. The lysate was extracted three times with 2 volumes of chloroform–3-methyl-1-butanol (24:1, vol/vol). The RNA fraction was precipitated by the addition of ethanol and redissolved in an appropriate volume of morpholinepropanesulfonic acid (MOPS) buffer, and residual DNA was removed by digestion with RNase-free DNase. The integrity of the RNA was checked by electrophoresis through formaldehyde gels.
Regions upstream from the 5′ end of the 16S rRNA coding region of rrnA of M. tuberculosis and rrnB of M. smegmatis were amplified by PCR with the primers shown in Table Table11 used to amplify the sequences illustrated in Fig. Fig.1.1. Appropriate plasmids for PCR amplification for M. tuberculosis and M. smegmatis were described previously (8). The procedures used resulted in a T7 promoter at the 3′ end of the minigene to provide a probe for RNase protection assays.
Radiolabelled transcripts were obtained by use of a Riboprobe kit (Promega, Madison, Wis.) according to the instructions provided. Each reaction mixture contained 50 μCi of [α-32P]CTP; transcription was carried out at 37°C for 1 h, and the reaction was terminated by heating the sample at approximately 90°C for 3 min. DNase was added to the cooled reaction mixture, which was then kept at 37°C for 15 min. The products were purified by electrophoresis through a 6% (wt/vol) polyacrylamide–8 M urea sequencing gel. An appropriate band was excised from the gel, and radiolabelled RNA was recovered after the gel fragment was soaked overnight in elution buffer (0.5 M ammonium acetate, 1 mM EDTA, 0.2% [wt/vol] sodium dodecyl sulfate). The eluate was desalted and then made 0.3 M in sodium acetate, and RNA was precipitated with ethanol.
Hybridizations were carried out with either formamide buffer [80% (vol/vol) formamide, 0.2 M sodium acetate, 40 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES) (pH 6.4)] at 50°C for 16 h as described by Sambrook et al. (16) or PES buffer (1 M NaCl, 1 mM EDTA, 25 mM PIPES [pH 6.8]) at 90°C for 70 min (14). Each sample contained radiolabelled probe (1 × 105 to 4 × 105 cpm) and 10 μg of M. tuberculosis RNA or 30 μg of M. smegmatis RNA.
After the hybridization step, samples were treated with RNase A and RNase T1 and then protein was removed by treatment with phenol-chloroform. RNA was precipitated, redissolved, and analyzed by electrophoresis through 6% polyacrylamide–8 M urea gels. The gels were calibrated by use of either HaeIII-digested X174 DNA molecular size markers (Promega, Madison, Wis.) or the products of sequencing reactions generated by the methods described below. After separation by electrophoresis, radioactive products were located by autoradiography at either approximately 20°C or −70°C with an intensifying screen.
DNA sequences were determined by the dideoxy chain termination procedure with [α-35S]dATP as described by Ji et al. (11) and with primer JY15.
Quantitative measurements of radioactivity were obtained with a PhosphorImager (model 4005; Molecular Dynamics, Chesham, Buckinghamshire, United Kingdom) and the software supplied with the instrument.
In order to investigate the coordinated expression of murA and rrnA of both M. tuberculosis and M. smegmatis and tyrS and rrnB of M. smegmatis, RNase protection assays were carried out. The constructs used to generate two sets of appropriate radioactive probes are illustrated in Fig. Fig.1.1. One set was designed to detect the 3′ end of M. tuberculosis murA RNA (Fig. (Fig.2A)2A) and the 3′ ends of murA RNA and tyrS RNA of M. smegmatis (Fig. (Fig.3A).3A).
The second set was designed to examine the expression of murA and tyrS and to relate the abundance of their transcripts to the abundance of pre-rRNA.
The data shown in Fig. Fig.2B2B (for M. tuberculosis murA) and 3C (for M. smegmatis murA and tyrS) provide evidence for large transcriptional units which extend through the protein coding region and into the 16S rRNA coding region. These transcripts comprise two regions, one coding for protein (MurA or TyrS) and the other coding for rRNA. For this reason, these RNA species are considered to be hybrid transcripts derived from hybrid operons. The hybrid operon identified in M. tuberculosis (Fig. (Fig.2)2) comprises murA at the 5′ end with the rrnA operon located downstream. The hybrid operons identified in M. smegmatis (Fig. (Fig.3)3) comprise murA at the 5′ end with the rrnA operon located downstream and tyrS at the 5′ end with the rrnB operon located downstream.
Previously (10), we described two categories of promoters that are present in M. tuberculosis in particular and in mycobacterial rrn operons in general (Fig. (Fig.1)1) and that are dedicated to pre-rRNA synthesis. One category (P1) is located within the coding region of murA, near the 3′ end; the second category (PCL1) is located within the hypervariable multiple-promoter region which extends from the 3′ end of murA to a conserved sequence motif, CL2, which lies upstream from the 16S rRNA gene and within the leader region of pre-rRNA transcripts (10). PCL1 promoters are associated with another conserved sequence motif, (CL1) (10, 11). A third category of promoter (PmurA), which lies upstream from murA, has now been identified. Previously, we identified a single (P1) promoter of the rrnB operon of M. smegmatis, located between tyrS and the 16S rRNA gene (8, 9). We have now identified a second promoter upstream from tyrS; this promoter is responsible for transcription of the hybrid operon.
M. tuberculosis has a single copy of murA per genome (4). Restriction enzyme digests of M. smegmatis genomic DNA revealed murA and tyrS to be present in single copies per genome (results not shown). Thus, murA, which forms part of a hybrid operon with rrnA, is the sole source of MurA in both M. tuberculosis and M. smegmatis. Similarly, tyrS, which forms part of a hybrid operon with rrnB in M. smegmatis, is the sole source of TyrS in M. smegmatis. We therefore carried out RNase protection studies using the probes described in Fig. Fig.4A,4A, A,5A,5A, and and5B5B in order to study murA RNA and tyrS RNA synthesis under different conditions of mycobacterial growth and to relate their synthesis to pre-rRNA synthesis.
The RNA fraction comprises mainly (approximately 83%) rRNA (2) and much lower levels of mRNA. The abundance of a particular mRNA within the RNA fraction represents a steady-state value which reflects both the rate of synthesis and the rate of degradation or processing. Comparative values of radioactivity per assay provide a measure of the relative number of transcripts at the steady state. The abundance of murA RNA within the RNA fraction was related to the abundance of transcripts of rrnA originating from tsp one (tsp1) and designated pre-rRNAA(P1) as shown in Fig. Fig.44 and Table Table22 for M. tuberculosis and in Fig. Fig.5A,5A, A,5B,5B, and and6A6A for M. smegmatis.
The results for M. tuberculosis (Fig. (Fig.4B4B and Table Table2)2) show that murA was expressed at all stages of growth, that is, not only during balanced growth but also during the stationary phase. For each assay, the number of murA transcripts was related first to the number of pre-rRNAA(P1) transcripts. The results obtained for samples r, s, t, and u (Table (Table2)2) show that there were 50 to 360 copies of murA RNA per 1,000 copies of pre-rRNAA(P1). Previously, we showed (9) that pre-rRNAA(P1) comprises approximately 20% of pre-rRNA. These estimates suggest that, on average, there were 10 to 70 copies of murA RNA for every 1,000 copies of pre-rRNA. The low level of pre-rRNA in the late log phase (sample v), originating from the P1 and PCL1 promoters and reported earlier (9), accounts for the observed high ratio of murA RNA to pre-rRNAA(P1) shown in Table Table22.
The data obtained for M. smegmatis (Fig. (Fig.5C)5C) show that murA was expressed during the stationary phase as well as during balanced growth. A comparison of the radioactivities of murA RNA- and pre-rRNAA(P1)-protected fragments revealed that more copies of murA RNA per copy of pre-rRNAA(P1) were found during exponential growth in Lemco broth than during the stationary phase (Fig. (Fig.6A,6A, panel i) or when growth conditions were less favorable, as in Kohn-Harris glucose medium (Fig. (Fig.6A,6A, panel ii).
Previously, it was shown that the contribution of pre-rRNAA(P1) to all transcripts (pre-rRNAA) of rrnA depended on the growth rate; pre-rRNAA(P1) accounted for approximately 1% of pre-rRNAA during early growth (e.g., Table Table3,3, samples a and b) in Lemco broth; this value rose to approximately 15% during the stationary phase (e.g., Table Table3,3, samples e, f, and g) or in Kohn-Harris glucose medium (9). During early growth in Lemco broth, the numbers of murA transcripts were found to exceed the numbers of transcripts of pre-rRNAA(P1) (Fig. (Fig.6A,6A, panel i). On the basis of the above-mentioned data, we estimate that, during culturing in Lemco broth, for every 1,000 copies of pre-rRNAA, we detected approximately 15 copies of murA RNA during the early stages of growth and 60 to 100 copies during the stationary phase. However, the overall pattern of murA RNA synthesis does not reflect the pattern of pre-rRNAA synthesis. The burst of pre-rRNAA synthesis [approximately 100-fold that of pre-rRNAA(P1)] that is known to take place during early growth in Lemco broth (9) has no counterpart in murA RNA synthesis. We infer that although the stimuli for murA and rrnA transcription may have features in common, they are not identical.
The results show that tyrS RNA was detected at all stages of growth of M. smegmatis (Fig. (Fig.5D).5D). When the mycobacterium was grown in Lemco broth, tyrS RNA was found to be most abundant during the early stages of growth and least abundant during the early and late stationary phases. When growth took place in Kohn-Harris glucose medium, the abundance of tyrS RNA was scarcely higher than the value found during the stationary phase in Lemco broth. In general, the profiles of tyrS RNA synthesis were found to correlate with the profiles of transcripts (pre-rRNAB) of rrnB (Fig. (Fig.5D);5D); for example, the burst in the synthesis of tyrS RNA noted during early growth in Lemco broth was also seen for the synthesis of pre-rRNAB. The abundance of tyrS RNA transcripts was related to the abundance of transcripts of pre-rRNAB (Fig. (Fig.6B);6B); ratios within the range of 1 to 6 copies of tyrS RNA for every 1,000 copies of pre-rRNAB were obtained. The highest ratio of tyrS RNA to pre-rRNAB, which was found during early growth in Lemco broth, is thought to reflect the need of the cells for a higher ratio of TyrS per ribosome during a period of rapid protein biosynthesis.
Here we have shown that the rrn operons of mycobacteria may form parts of larger hybrid operons in which both protein encoding and rRNA encoding DNAs are transcribed as single units. The starting point for the transcription of the hybrid operon of M. tuberculosis is likely to be between the insertion sequences IS1557, which lies immediately upstream of murA (4), and the 5′ end of the coding region of murA. No similar map is available for M. smegmatis; therefore it is not certain that murA is the first gene of the putative hybrid operon, and it is not known whether it is preceded by other genes. The RNase protection experiments show that transcription in both species continues beyond the 5′ end of the 16S rRNA coding region, suggesting that transcription will continue until an intrinsic terminator downstream from the 5S rRNA gene is reached because of the operation of particular (antitermination) mechanisms that ensure complete transcription of the rrnA operon (5). Initiation of murA transcription from the PmurA promoter ensures that there is one copy of pre-rRNA for each copy of murA RNA derived from that promoter. However, the contribution of the hybrid operon to pre-rRNA synthesis is small (approximately 1 to 4% [Table 2]) because of the presence of additional promoters dedicated to pre-rRNA synthesis and located between the end of the murA gene and the beginning of the 16S rRNA gene (8).
The second rRNA operon of M. smegmatis, rrnB, also may form part of a hybrid operon, in this case including the tyrS gene. In the genome of M. tuberculosis, tyrS is preceded by the Rv1688 gene, which probably codes for 3-methylpurine DNA glycosylase; the two genes appear to be transcribed as a single unit (4), as judged by the size and composition of the sequence separating them. Should a similar arrangement occur in the M. smegmatis genome, then the homologue of the Rv1688 gene would be part of the hybrid operon.
Within the bacterial cell, transcription and translation are coupled so that mRNA is translated as it is being synthesized. If, like mRNA translation, both ribosome assembly and rRNA processing take place as transcription proceeds, then it is likely that only a small proportion of hybrid transcripts will remain intact after transcription is complete. Early pre-rRNA processing events would have the effect of releasing mRNA into the cytosol; a transcriptional terminator also has this effect. Thus, the hybrid operon appears to make economical use of cell resources by using a minimum number of initiation and termination steps.
Our analysis of the pre-rRNA fractions of both M. tuberculosis and M. smegmatis has shown that in normal growth, transcripts originating from promoters dedicated to the expression of rRNA gene sequences represent the majority. The str operon of Escherichia coli provides another example of promoters located within an operon but dedicated to the expression of a particular gene within the operon. This operon (19, 20) comprises four genes in the order rpsL, rpsG, fus, and tufA. In addition to Pstr, the principal promoter for the operon, promoters dedicated to the expression of tufA (encoding elongation factor Tu) are thought to lie within fus (encoding elongation factor G).
Results obtained for E. coli rrn operons help to place our findings in a wider context. The rrnG operon, which does not appear to form part of a hybrid operon, is located downstream from clpB, a gene coding for E. coli heat shock protein F84.1 (18). An intrinsic terminator for clpB is thought to be located 31 bp downstream from the 3′ end of the coding region and 83 bp upstream from the −35 box of promoter P1 of rrnG (17), suggesting that clpB and rrnG are transcribed independently. In contrast, transcripts of the open reading frame upstream from the rrnB operon of E. coli were shown to continue into the 16S rRNA coding region; this finding was the first evidence for a hybrid operon (1). The open reading frame, which is regulated by two promoters (1), encodes glutamate racemase, an enzyme that catalyzes the synthesis of d-glutamate, an essential component of bacterial peptidoglycan.
We thank our colleague Andrew Lane for helpful discussions. We thank Simon A. Cox for help in the preparation of the manuscript.
J. A. G.-y-M. received financial support from COFAA and EDD, IPN, Mexico. This work was supported as part of the European Commission Science Research and Development Programme (contract ERBIC 18CT 9720253).