Roles of Multiple Promoters in Transcription of Ribosomal DNA: Effects of Growth Conditions on Precursor rRNA Synthesis in Mycobacteria

Abstract

The roles of multiple promoters in the synthesis of rRNA under different conditions of growth were investigated, using two mycobacterial species as model organisms. When Mycobacterium smegmatis was grown under optimal conditions, its two rRNA operons contributed equally, with two promoters, one from each operon, being responsible for most transcripts. In stationary-phase growth or balanced growth under carbon starvation conditions, one operon (rrnA_f) dominated and its three promoters contributed more equally to the generation of transcripts. Mycobacterium tuberculosis has a single operon with two promoters, one of which generated 80% of transcripts, at all stages of growth. We infer that each promoter functions independently according to its intrinsic strength when cells are growing slowly so that one operon with three promoters is roughly equivalent to three operons with one promoter; at high growth rates, occlusion effects reduce the efficiency of multiple promoters to that of a single promoter.

The arrangement of multiple promoters in tandem is a general feature of rrn operons. For example, Escherichia coli has seven rrn operons per genome, and each rrn operon has two promoters (5). Similarly, each of the 10 operons of the Bacillus subtilis genome has two promoters (17). Streptomyces coelicolor A3(2) has six rrn operons per genome (19), and at least one (rrnA) has four tandem promoters (20). The roles of these multiple promoters in controlling the synthesis of rRNA are not fully understood.

Studies with mycobacteria have revealed that the slow-growing pathogen Mycobacterium tuberculosis has a single rRNA operon per genome (2, 14, 18) designated rrnA_s (12). This operon (Fig. 1a) is regulated by two promoters arranged in tandem (10, 21). The fast-growing species Mycobacterium smegmatis has two operons (designated rrnA_f and rrnB_f) per genome (Fig. 1b); one of these operons (rrnA_f) has three promoters, and the other (rrnB_f) has one (10). Interestingly, some fast-growing species (for example, Mycobacterium abscessus and Mycobacterium chelonae) have a single rrnA-like operon but may have as many as five promoters by which they increase the potential for rRNA synthesis (11). The fact that mycobacteria have a minimal number of rrn operons suggests that these operons may be a useful tool for determining the roles of multiple promoters in rRNA synthesis; the presence of only one or two operons should greatly simplify the analysis of the precursor rRNA (pre-rRNA) fraction in order to reveal the contributions of different operons and the tactical usage of their promoters.

FIG. 1.

Promoters of rrn operons of M. tuberculosis and M. smegmatis. The single rrnA_s operon of M. tuberculosis (a) and the rrnA_f and rrnB_f operons of M. smegmatis (b) are shown. The transcription starting points () of the promoters (P1, etc.) and the binding site (◂—<) on the RNA-like strand of DNA for the oligomer (JY15) used in primer extension experiments are indicated. The UDP-N-acetyl glucosamine carboxyvinyl transferase (UNAcGCT) (EC 2.5.1.7) and tyrosyl tRNA synthetase (Y-tRNA Sythe) genes are indicated. Distances (in base pairs) between transcription starting points etc. are given beneath horizontal arrows.

The yield of RNA from mycobacteria is related to the growth rate.

The amount of RNA per cell is thought to be governed by the number of ribosomes, with rRNA accounting for most (approximately 83%) of the RNA fraction. There is also a correlation between the number of ribosomes per average cell and the growth rate (3). Thus, one would anticipate that the yield of RNA would be related to the growth rate; to determine if this is indeed the case for mycobacteria, yields of RNA from M. smegmatis grown under optimal and suboptimal (carbon starvation) conditions and for slow-growing M. tuberculosis were determined.

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 (4) or Kohn-Harris glucose medium (see below) containing 0.1% Tween 80. A seed culture of M. smegmatis for inoculation was grown in Lemco broth for 26 h. This culture was used to inoculate medium at the rate of 20 ml of inoculum per liter. The Kohn-Harris glucose medium was based (6) on the medium of Kohn and Harris (15), which uses glucose (2 g/liter) as the carbon source. Trace elements (5 ml/liter) were provided in the solution of Kelly and Clarke (13). M. smegmatis was grown under optimal conditions (generation time τ of approximately 2 h) in complete medium (Lemco broth) and under conditions of carbon starvation (τ ≈ 23 h) in Kohn-Harris glucose medium (see Table 1). The slow-growing M. tuberculosis (strain H37Rv) was grown in complete Dubos medium (7) containing 0.1% Tween 80 (τ ≈ 72 h) (see Table 1). Samples of cells were removed at appropriate intervals, and RNA fractions were isolated. Bacteria were collected and resuspended in 1-ml portions of guanidinium buffer (6 M guanidinium chloride, 0.1% [vol/vol] Tween 80, 10 mM EDTA, 1 mM 2-mercaptoethanol) and left at −20°C for 15 min. The suspension was added to half of 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 1-min pulses on the 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 RNA fraction was isolated as described previously (10).

TABLE 1.

Specific radioactivities after the primer extension assay of RNA isolated from mycobacteria at representative stages of growth

Species, medium, and sample	Time	E640	Stage of growth	Radioactivity/assay (total counts/μg of RNA)
M. smegmatis
Lemco broth (τ = 2 h)
i	0 h	0.034	Stationary (inoculum)	2,200a
a	3 h	0.090	Balanced growth	54,000
b	7 h	0.430	Balanced growth	35,000
c	11 h	1.000	Early stationary	15,000
d	16 h	1.350	Stationary	3,000
e	26 h	1.700	Stationary	2,000
f	31 h	1.860	Stationary	8,000
g	36 h	1.700	Stationary	8,000
Kohn-Harris glucose medium (τ = 23 h)
h	0 h	0.034	Stationary (inoculum)	2,200a
j	20 h	0.070	Balanced growth	5,000
k	36 h	0.240	Balanced growth	11,000
l	56 h	0.590	Early stationary	6,000
m	90 h	0.870	Stationary	4,000
n	121 h	1.230	Stationary	6,000
o	150 h	1.310	Stationary	4,000
M. tuberculosis
Dubos medium (τ = 72 h)
iTb	0 days	0.025	Stationary (inoculum)
r	5 days	0.103	Balanced growth	2,140
s	10 days	0.360	Balanced growth	1,520
tb	13 days	0.580	Early stationary	2,160
u	22 days	0.680	Stationary	2,260
v	40 days	1.250	Stationary	280

^aEstimated from the properties of the stationary-phase culture from which the inoculum was taken. 

^bDuplicate samples were grown. Values of E₆₄₀ were essentially the same (±0.02 E₆₄₀ units), and the yields of RNA were the same (62 ± 2 μg per 100 ml of culture). 

The highest yield of RNA (approximately 3 μg/E₆₄₀) was obtained for M. smegmatis grown in Lemco broth during early balanced growth (τ ≈ 2 h). A smaller yield (approximately 2 μg/E₆₄₀) was obtained for M. smegmatis grown under conditions of carbon starvation (τ ≈ 23 h). The lowest yield (approximately 0.8 μg/E₆₄₀) was obtained for M. tuberculosis (τ ≈ 72 h). Thus, as anticipated, there is a relationship between RNA content and growth rate.

The contributions of the two rrn operons of M. smegmatis to rRNA synthesis differ under different growth conditions.

To determine the roles played by the two M. smegmatis rrn operons, primer extension analysis was performed to identify the origins of the pre-rRNA transcripts. The assay was designed to facilitate the comparison of pre-rRNA transcripts of the rrnA_f and rrnB_f operons of M. smegmatis and the single (rrnA_s) operon of M. tuberculosis. First, a DNA primer was chosen so that its affinity for its target site was the same in each case. This was achieved by identifying a target common to all the pre-rRNA transcripts under scrutiny; the oligonucleotide primer 5′ CACACTATTGAGTTCTC 3′ (JY15) has a target site which is located approximately 150 nucleotides upstream from the 5′ end of the 16S rRNA sequence and is present in all three of the rrn operons studied (11) (Fig. 1). Reaction conditions were further standardized by keeping the amount of RNA per assay constant for each mycobacterial species (30 μg per assay for M. smegmatis and 10 μg per assay for M. tuberculosis). Thus, the numbers of extension products per assay should be directly proportional to the concentrations of pre-rRNA species. In addition, the DNA primers were labelled with ³²P at their 5′ ends. The use of end-labelled primers ensures that the radioactivity of each product is directly proportional to the number of molecules produced, irrespective of their length. This property allows direct comparison of the yields of pre-rRNA products originating from different promoters of the same operon and of products derived from different operons. The primer was end labelled with [γ-³²P]ATP by means of T4 polynucleotide kinase, and the primer extension was performed using the avian myeloblastosis virus reverse transcriptase primer extension system, as described previously (10). The extension products were separated on an 8% (wt/vol) polyacrylamide–8 M urea gel and visualized by autoradiography. Quantitative measurements of radioactivity were obtained with a PhosphorImager (model 4005; Molecular Dynamics, Chesham, Buckinghamshire, United Kingdom) using the software supplied with the instrument.

The results are interpreted with the assumption that, within the cell, pre-rRNA synthesis and processing are balanced so that a steady state is achieved, giving rise to a pool of pre-rRNA species and their early processing products. The primer extension assay measures both the size and composition of this pool. Six primer extension products were identified for M. smegmatis (see Fig. 2). RNase protection experiments (9) confirmed that four were transcription products (pre-rRNA_A) of the rrnA_f operon. Three products of 236, 136, and 60 nucleotides correspond to transcription products [designated pre-rRNA_A(P1), pre-rRNA_A(P2), and pre-rRNA_A(PCL1), respectively] originating from transcription start points of promoter sequences designated P1, P2, and PCL1 which have the classical −10 boxes of promoters requiring a sigma 70-like transcription factor (for discussion, see reference 11). A fourth product (product b) of 99 nucleotides, which does not correspond to the product of a readily recognizable promoter sequence, is possibly a product of early processing of the transcripts pre-rRNA_A(P1) and pre-rRNA_A(P2). Quantitative analysis revealed that product b corresponded to 23% ± 2% for samples a to g and 27% ± 3.5% for samples h to o of the sum of the radioactivities of pre-rRNA_A(P1) and pre-rRNA_A(P2). Two of the three promoters, namely, P1 and PCL1, are homologous with the two promoters of the single rrn (rrnA_s) operon of M. tuberculosis. The third promoter, P2, is a characteristic feature of fast growers (11). The spacing distances between promoters P1 and P2 and between P2 and PCL1 are 100 and 76 bp respectively, comparable with the distance (77 bp) between the P1 and PCL1 promoters of M. tuberculosis (10). We infer on the basis of data for E. coli promoters (16) that the mycobacterial promoters are separated sufficiently to allow the formation of an initiation complex between RNA polymerase and a promoter without interference from an initiation complex formed at an adjacent promoter.

FIG. 2.

Analysis of RNA fractions isolated from M. smegmatis by the primer extension assay. Pre-rRNA species were identified by the sizes of the products (Fig. 1) generated by the primer extension procedure using ³²P-labelled primer JY15 (see Materials and Methods). Lanes T, C, G, and A contain products of sequencing reactions (see Materials and Methods) performed with primer JY15 and a recombinant phagemid containing an appropriate rrnA_f or rrnB_f sequence. Pre-rRNA_A(P1) etc. denote primer extension products originating from the P1 promoter of the rrnA_f operon etc. Product a is thought to be an early product of pre-rRNA_B processing; product b is thought to be an early product of pre-rRNA_A(P1) and pre-rRNA_A(P2) processing, as discussed in the text. (−), no RNA. Samples a to g and f to o are defined in Table 1. Autoradiographs were obtained after exposure at −70°C for 16 h (a) and 96 h (b).

The remaining two primer extension products were found to be derived from transcription products (pre-rRNA_B) of the rrnB_f operon. The product of 160 nucleotides reveals a pre-rRNA_B(P1) transcript originating from a transcription start site of a classical promoter sequence (10, 11). The product of 126 nucleotides (product a) does not correspond to a transcript originating from the start point of a readily recognized promoter sequence and is regarded as a possible product of early processing of pre-rRNA_B(P1). Product a was present in each of the samples analyzed; the amount of product a was 14% ± 3% (samples a to g) and 18% ± 5% (samples h to o) of pre-rRNA_B(P1).

The relative abundancies of pre-rRNA transcripts and processing products within the pre-rRNA fraction provide a measure of the steady state established between pre-rRNA synthesis and processing. It was noted above that the relative abundancies of the putative products, products a and b, were similar for all the samples studied. This result shows that the balance between pre-rRNA synthesis and processing was very similar for all of the conditions investigated. Hence we infer that the steady-state level of the pre-rRNA pool provides a measure of the rate of pre-rRNA synthesis.

When M. smegmatis was grown in complete medium, the primer extension products were found to be most abundant during balanced growth (samples a and b [Fig. 2; see also Table 1]), indicating that pre-rRNA synthesis had taken place most rapidly at this stage of the growth cycle. The specific radioactivities (counts per microgram of RNA or counts/E₆₄₀) of the products were least for RNA samples harvested during mid-stationary phase (samples d and e [Fig. 2; see also Table 1]), indicating that pre-rRNA synthesis was diminished before a burst of synthesis in late stationary phase (samples f and g [Fig. 2; see also Table 1]). Pre-rRNA species were also found to be most abundant during balanced growth when M. smegmatis was grown in conditions of carbon starvation (samples j and k [Fig. 2 and Table 1]). However, the rate of pre-rRNA synthesis was lower for cells using glucose as a source of carbon (τ ≈ 23 h) than for cells grown in a rich medium (τ ≈ 2 h). In contrast, the rates of pre-rRNA synthesis in stationary phase were less dependent on growth conditions than cells in balanced growth. Thus, the rate of pre-rRNA synthesis and the rate of cell proliferation are generally correlated.

Growth conditions influenced not only the rate of synthesis of the pre-rRNA fraction but also its composition. Products derived from pre-rRNA_A and pre-rRNA_B were identified at all stages of growth. During balanced growth in complete medium, the majority of products were pre-rRNA_A(P2) and pre-rRNA_B(P1), in roughly equal proportions (Fig. 2 and 3A), and the contributions of the two operons were virtually identical. However, during slow growth in carbon starvation medium, the balance is changed such that the rrnA_f operon contributes approximately three times more to the pool of pre-rRNA than the rrnB_f operon (Fig. 2 and 3B).

FIG. 3.

Effects of growth conditions on the expression of rrn operons of M. smegmatis. M. smegmatis was grown in Lemco broth (A) (generation time ≈2 h) or Kohn-Harris glucose medium (B) (generation time ≈ 23 h). ■, pre-rRNA_A; , pre-rRNA_B. Radioactivity is shown in 10⁶ counts (Megacounts) per assay.

The efficiencies of different promoters of the rrnA_f operon are altered under different growth conditions.

The primer extension experiments enabled us to analyze the contributions made by the three tandemly arrayed promoters of the rrnA_f operon under different growth conditions. During balanced growth in complete medium, the average ratio of the products pre-rRNA_B(P1)/pre-rRNA_A(P1)/pre-rRNA_A(P2)/pre-rRNA_A (PCL1) is 1:0.01:0.8:0.11 (Fig. 4A). However in carbon-limiting medium at suboptimal growth rates, the rrnA_f operon was the major contributor to pre-rRNA synthesis (Fig. 4B). On average, the relative amounts of products pre-rRNA_B(P1), pre-rRNA_A(P1), pre-rRNA_A(P2), and pre-rRNA_A(PCL1) were obtained in the ratio 1.0:0.3:1.6:1.1.

FIG. 4.

Radioactivities of primer extension products derived from pre-rRNA species of M. smegmatis. The radioactivities of the products shown in Fig. 2 were measured with a PhosphorImager. The radioactivities of products originating from rrnA_f(P1) and rrnA_f(P2) include an appropriate proportion of the putative early processing product, product b (Fig. 2). The radioactivities of the products originating from rrnB_f(P1) include the early processing product, product a (Fig. 2). M. smegmatis was grown in Lemco broth (A) or Kohn-Harris glucose medium (B). The y axes for panels A and B are the same.

At low growth rates, low rates of initiation allow individual promoters of a tandem array to function independently of each other and hence their relative contributions reflect their intrinsic strengths; thus, the contributions illustrated in Fig. 4B probably reflect the intrinsic activity of the four promoters, with rrnA_f (P2) providing the largest contribution and rrnA_f (PCL1) and rrnB_f(P1) contributing equally. However, at high growth rates (Fig. 4A), we find that rrnB_f (P1) contributes more than the three rrnA_f promoters combined, suggesting that the three tandem promoters are acting at suboptimal efficiency at high rates of initiation; this would be compatible with steric effects such as promoter occlusion (1, 8).

Contributions of the two promoters of the rrnA_s operon to the pre-rRNA species of the slow-growing M. tuberculosis.

In order to address the question of how the two promoters (P1 and PCL1) of the single M. tuberculosis operon (rrnA_s) interact, M. tuberculosis H37Rv was grown in Dubos medium (7) containing 0.1% Tween 80. The generation time (approximately 72 h) was almost five times that found for optimal growth (approximately 15 h) of this species (22). The radioactivity of the RNA fraction after the primer extension assay varied little during growth, except for a reduction in late stationary phase (Table 1). Two primer extension products, pre-rRNA_A(P1) and pre-rRNA_A(PCL1) were identified in all samples (Fig. 5). In each case, pre-rRNA_A(PCL1) was found to be the more abundant species. Thus, for samples r to u, the radioactivity of pre-rRNA_A(P1) (the 134-nucleotide product) ranged from 120 to 600 counts/g of RNA fraction compared with 1,400 to 1,800 counts/μg of RNA fraction for pre-rRNA_A(PCL1) (the 57-nucleotide product). On average, the products derived from the P1 and PCL1 promoters were in the ratio 0.25:1. A similar ratio (0.27:1) was found (see above) for the products of the P1 and PCL1 promoters of the rrnA_f operon of M. smegmatis grown in Kohn-Harris glucose medium. According to our hypothesis, the above-mentioned ratios reflect the intrinsic strengths of the P1 and PCL1 promoters. It is inferred that the intrinsic strength of the PCL1 promoter is significantly higher than that of the P1 promoter, and the results obtained for the slow-growing M. tuberculosis support the notion that the two promoters of the rrnA_s operon are an asset because they increase the efficiency for pre-rRNA synthesis above the level of an operon regulated by a single promoter.

FIG. 5.

Analysis of RNA fractions isolated from M. tuberculosis by the primer extension assay at representative stages of growth. Pre-rRNA species were identified by the sizes of the products (Fig. 1) generated by the primer extension procedure using ³²P-labelled primer JY15 (see Materials and Methods). Lanes T, C, G, and A contain products of sequencing reactions (see Materials and Methods) performed with primer JY15 and a recombinant phagemid containing an appropriate rrnA_s sequence. Pre-rRNA_A(P1) and pre-rRNA(PCL1) denote primer extension products originating from the P1 and PCL1 promoters, respectively. (−), no RNA. Samples r to v are defined in Table 1. Autoradiography was obtained after exposure at −70°C for 16 h.

Concluding remarks.

Our results may be explained by a model which suggests that in tandemly arrayed promoters, each promoter functions independently when cells are growing slowly; thus, when growth is slow, one operon with three promoters is roughly equivalent to three operons with one promoter, maximizing the use of limited resources. However, when nutritional resources are not limiting, this advantage is diminished as the growth rate increases until multiple promoters are no more effective than a single promoter. The conclusions we have reached about the use of multiple promoters by the mycobacterial species studied may be relevant to other rrn operons whose transcription depends on tandem promoters.