Oxygen depletion induced dormancy in Mycobacterium smegmatis

Abstract

We report here that the physiological behaviour of the fastgrowing saprophytic Mycobacterium smegmatis under in vitro oxygen-depletion and reactivation conditions is strikingly similar to the characteristics shown by the slowgrowing pathogenic M. tuberculosis. M. smegmatis died rapidly when shifted abruptly from aerobic to anaerobic conditions. In contrast to the lethal shock of abrupt oxygen depletion, the slow depletion through a selfgenerated oxygen gradient permitted an adaptation to a persistent state which showed increased resistance against the bactericidal effects of anaerobiosis. The anaerobic persistent culture did not synthesise DNA and showed synchronised division upon reactivation in oxygen rich medium, indicating that the persistent bacilli are uniformly arrested at a defined stage of the cell cycle. Upon reactivation the persistent culture started synthesising DNA only after the first cell division, suggesting that the persistent cells contain two chromosomes. Furthermore, the persistent culture developed sensitivity to metronidazole and resistance against ofloxacin. These results suggest that M. smegmatis might be useful as a fastgrowing non-pathogenic model for comparative molecular analyses of mycobacterial dormancy.

Mycobacterium smegmatis, Dormancy, Anaerobiosis

1 Introduction

Most individuals who are infected with Mycobacterium tuberculosis do not progress to active disease. However, the bacillus may remain in the host for decades in a dormant state, with the potential for revival and initiation of clinical disease [1]. Almost two billion people may be latently infected with the tubercle bacillus [2]. Mycobacteria in the dormant state are resistant against killing by the standard antimycobacterial drugs [1]. Almost half a century ago Wayne and collaborators showed that year-old blocked lesions from resected human lung tissues, i.e. lesions with little or no oxygen, contained viable tubercle bacilli [3]. This demonstrated a potential in M. tuberculosis for survival under anaerobic conditions. In pioneering work over the last decades Wayne and coworkers developed an in vitro system for mycobacterial dormancy and confirmed that oxygen depletion is a signal that triggers a dormancy response of the bacilli [1, 4].

This laboratory is interested in the mechanisms of mycobacterial dormancy and reactivation, which are largely unknown. The experimental disadvantages associated with the slow growth rate and the infectious nature of M. tuberculosis prompted us to investigate the possibility of using a fastgrowing non-pathogenic Mycobacterium as a model to uncover the molecular mechanism of dormancy. Thus, we asked the question whether the oxygen depletion induced dormancy response is conserved in the fastgrowing saprophyte Mycobacterium smegmatis.

2 Materials and methods

2.1 Strain and cultivation

All experiments were conducted with M. smegmatis mc²155 [5]. Cultures were carried out in 20×125-mm test tubes according to Wayne and Hayes [4] using Dubos Tween-albumin broth (Difco, Detroit, MI) at 37°C. Seventeen ml of culture was used throughout and, therefore, the ratio of head space air volume (8.5 ml) to liquid volume (17 ml) was always 0.5. Depending on the conditions of aeration desired, caps with latex liners were either loosely (‘unsealed culture’, unlimited oxygen supply) or tightly screwed down (‘sealed culture’, limited oxygen supply). Depending on the oxygen diffusion rate desired, tubed cultures were either incubated on a rotary shaker-incubator under vigorous shaking at 250 r.p.m. (‘shaken culture’, rapid diffusion of oxygen from head space into the liquid culture) or gently stirred using magnetic stirring bars at 170 r.p.m. (‘stirred culture’, slow diffusion of oxygen from the head space into the liquid culture). In some instances, tightly sealed caps with rubber septa were used to permit addition of reagents by needle without opening the container.

2.2 Monitoring of growth, survival, oxygen depletion and nucleic acid synthesis

Growth and survival of the bacterial populations were monitored by viable count measurements as described [4]. Oxygen depletion was monitored using the oxygen indicator dye methylene blue as described [4]. Syntheses of RNA and DNA were monitored by adding 37 kBq of [5,6-³H]uracil (1.81 TBq mmol⁻¹) per ml of culture medium and the uptake was measured as described [6]. Briefly, cell-bound substrates were collected by filtering 2-ml samples through 0.45-μm NC filter membranes. One sample was filtered directly and a duplicate sample was incubated for 24 h at 37°C in 0.3 M KOH before filtering to permit distinction between DNA and RNA. DNA was estimated from the amount of bound label remaining after alkali treatment of the bacilli to destroy the RNA; the RNA was estimated as the difference between total bound label and that remaining after alkali treatment of the bacilli. All results were corrected for zero-time absorption of label by the filter.

3 Results

3.1 Death upon abrupt shift to anaerobic conditions and survival upon slow depletion of oxygen

For the study of the growth, death and survival dynamics of M. smegmatis under various oxygen limiting conditions we first determined the generation time and growth limits of an aerated, i.e. unsealed shaken culture. Growth was monitored by counting c.f.u. Fig. 1 A shows that the culture grew with a generation time of 2.5 h. Medium limitation of growth was not detectable up to a cell density of ∼10⁹ c.f.u. ml⁻¹. After that, the growth deflected and the culture entered a death phase in which the viable cell count dropped 10-fold within ∼30 h. The death phase was followed by regrowth.

1

Growth, death and survival of M. smegmatis under various culture conditions. Log c.f.u. ml⁻¹ as a function of time under various culture conditions is shown. Aerobic logarithmic preculture was diluted to ∼10⁶ c.f.u. ml⁻¹ and incubated in an unsealed or sealed tube under vigorous shaking or gentle stirring conditions as described in Section 2. A: Growth, death and regrowth of an unsealed shaken culture. B: Growth and anaerobic death of a sealed shaken culture. C: Growth and increased anaerobic survival of a sealed stirred culture. f, d: fading and complete decolorisation of the oxygen indicator methylene blue, respectively. Mean values and standard deviations are shown from two experiments. The cultures were checked microscopically for any clumping of cells and for pH changes. Significant clumping was never observed and the pH of the cultures was constant.

Wayne and collaborators showed that an abrupt shift of an aerobic M. tuberculosis culture to anaerobic conditions led to the rapid death of the culture [4]. In contrast, a slow, gradual shift of an aerobic M. tuberculosis culture to anaerobic conditions did not lead to the death of the culture. Rather, the culture was able to adapt and survive anaerobiosis for an extended time by entering a non-replicating synchronised state of persistence, i.e. a state of dormancy [4]. The death and survival behaviour upon abrupt and gradual shift to anaerobic conditions, respectively, represents a key feature of the dormancy response of the tubercle bacillus in vitro. To determine whether the saprophytic M. smegmatis is capable of responding in a similar way, oxygen depletion experiments were carried out. In order to shift an aerobic M. smegmatis culture abruptly from aerobic to anaerobic conditions we applied conditions where a culture selfgenerates rapid oxygen depletion. This was accomplished by using a sealed culture, thus limiting the total amount of oxygen available in the head space of the tube, and incubation under vigorous shaking conditions to ensure rapid diffusion of the limited amount of oxygen from the head space into the medium. Oxygen depletion was monitored using the indicator dye methylene blue in control cultures. Reduction and decolorisation of this dye served as a visual indication of oxygen depletion. Fig. 1 B shows that an initial aerobic growth phase (2.5 h generation time) stopped abruptly at ∼2×10⁸ c.f.u. ml⁻¹. The deflection from aerobic growth was accompanied by a rapid depletion of oxygen, indicated by fading and complete decolorisation of the oxygen indicator methylene blue within 5 h (Fig. 1B). A death phase followed where the viable count dropped 10-fold within ∼50 h. These results show that an abrupt shift of an aerobic culture of M. smegmatis to anaerobic conditions induces the rapid death of the culture.

Next we asked whether a slow oxygen depletion would allow the culture to adapt to anaerobiosis by shifting down to a state with an extended life span. Conditions were chosen where the culture selfgenerates a gradual temporal oxygen gradient. This was accomplished by using a sealed culture, thus limiting the total amount of oxygen available, and gently stirred incubation. The gentle stirring did not disturb the liquid surface and thus kept the oxygen diffusion rate from the head space into the medium low. Fig. 1 C shows that an initial aerobic growth phase (2.5 h generation time) stopped abruptly at ∼2×10⁷ c.f.u. ml⁻¹. After a plateau phase of about 70 h a brief microaerobic growth phase was detectable that terminated at ∼5×10⁷ c.f.u. ml⁻¹. After that the culture entered a phase which is characterised by an extended life span indicated by the slow decline of the viable counts (10-fold reduction of the viable counts within about 300 h). A very slow depletion of the oxygen was observed after the deflection from the aerobic growth phase (Fig. 1C). Fading of methylene blue did not start until after termination of the brief microaerobic growth phase. Complete decolorisation took about an extra 80 h. Taken together, these results show that a slow gradual shift of an aerobic culture of M. smegmatis to anaerobic conditions leads to the adaptation and increased survival of the bacilli similar to M. tuberculosis.

3.2 Synchronous cell division upon reactivation

A key feature of the M. tuberculosis dormancy response is that the anaerobic persistent culture is arrested at a uniform stage of the cell cycle, i.e. the culture is non-replicating and synchronised [4, 6]. Pulse labelling experiments of anaerobic sealed stirred cultures (at 240 h) to monitor DNA synthesis of the persistent M. smegmatis culture showed no significant replication (Table 1). Low level RNA synthesis was detectable, suggesting that the non-replicating cells are still transcriptionally active (Table 1). To determine whether the persistent M. smegmatis culture is synchronised reactivation experiments were carried out. Anaerobic sealed stirred culture (at 240 h) was diluted 1:100 in fresh oxygen rich medium and incubated under vigorous shaking in an unsealed tube. Fig. 2 shows synchronous cell division upon reactivation after a lag phase of about 1 h. This result demonstrates that the M. smegmatis culture is not arrested in a random state but rather appears to have arrived in the anaerobic phase in a uniform stage of the cell cycle, similar to M. tuberculosis. The M. smegmatis culture during the plateau phase before the brief microaerobic growth phase was not synchronised.

1

Nucleic acid synthesis and antibiotic susceptibility of persistent and active cultures

Culture typea	Nucleic acid synthesisb (kc.p.m. ml−1)		Antibiotic susceptibilityc (% bacilli surviving)
	DNA	RNA	Metronidazole	Ofloxacin
Persistent culture	0±1	3±2	20±10	137±20
Active culture	36±5	128±16	95±9	10−4±10−4

^aPersistent culture: anaerobic 240-h-old sealed stirred culture containing 10⁷ c.f.u. ml⁻¹. Active culture: aerobic logarithmic unsealed shaken culture containing 10⁷ c.f.u. ml⁻¹.

^bIncorporation of label from [5,6-³H]uracil into DNA and RNA is shown. Persistent and active cultures were pulse labelled for 2.5 h with 37 kBq [5,6-³H]uracil ml⁻¹. DNA and RNA synthesis was estimated as described in Section 2. Mean values and standard deviations are shown from two experiments.

^cPercent survival after exposure to metronidazole and ofloxacin is shown. Persistent and active cultures were exposed to metronidazole (120 μg ml⁻¹) and ofloxacin (1 μg ml⁻¹), respectively. C.f.u. were determined after 24 h exposure to the drug. Percent survival is expressed in terms of drug-free control at the time samples were taken (persistent culture: 10⁷ c.f.u. ml⁻¹; active culture: 2×10⁷ c.f.u. ml⁻¹). Mean values and standard deviations are shown from two experiments.

2

Synchronous cell division after reactivation of anaerobic persistent M. smegmatis culture. Log c.f.u. ml⁻¹ as a function of time after reactivation of an anaerobic persistent culture is shown. 240-h-old sealed stirred culture was diluted 1:100 in fresh oxygen rich medium and incubated in an unsealed tube under vigorous shaking conditions. Mean values and standard deviations are shown from three experiments.

3.3 Replication start after the first cell division

Analyses of the DNA synthesis of reactivated M. tuberculosis culture indicated that the cells first undergo a round of cell division before they start DNA synthesis [6]. This suggests that the dormant bacilli contain two copies of the genome [6]. To determine whether the similarity of the dormancy response between M. smegmatis and M. tuberculosis includes this characteristic feature DNA labelling experiments were carried out. The results in Fig. 3 A show that DNA synthesis was only detectable after the first division. This suggests that the dormant M. smegmatis bacilli might arrest, like M. tuberculosis, with a duplicated chromosome. RNA synthesis could be detected immediately upon reactivation of the dormant organisms (Fig. 3B).

3

Incorporation of label from [5,6-³H]uracil into DNA and RNA during synchronised cell division after reactivation. A: kc.p.m DNA ml⁻¹ (•); B: kc.p.m RNA ml⁻¹ (•) as a function of time after reactivation of a dormant culture are shown. 240-h-old sealed stirred culture was diluted 1:10 in fresh oxygen rich medium containing 37 kBq [5,6-³H]uracil ml⁻¹. DNA and RNA were monitored by measuring the incorporation of label from [5,6-³H]uracil as described in Section 2. DNA was estimated from the amount of bound label remaining after alkali treatment of the bacilli to destroy the RNA; the RNA was estimated as the difference between the total bound label and that remaining after alkali treatment of the bacilli. kc.f.u. ml⁻¹ (◯) were determined from parallel unlabelled cultures and are shown in A and B. Mean values and standard deviations are from two experiments.

3.4 Metronidazole sensitivity and ofloxacin resistance of dormant bacilli

Anaerobic dormant tubercle bacilli are resistant against the common anti-mycobacterials but develop a sensitivity to metronidazole, a drug against anaerobes [4, 7]. To determine whether M. smegmatis develops sensitivity to metronidazole we exposed anaerobic dormant culture (at 240 h) to the drug and measured the survival rate. Table 1 shows that the dormant bacteria were killed by metronidazole whereas the actively growing aerobic bacteria were not affected by the drug. Table 1 also shows that dormant M. smegmatis culture was resistant against the anti-mycobacterial gyrase inhibitor ofloxacin [8], which killed the actively growing culture. These data suggest that the dormant M. smegmatis culture has a similar antibiotic susceptibility pattern as dormant tubercle bacilli.

4 Discussion

We report here that the physiological behaviour of M. smegmatis under oxygen depletion and reactivation conditions is strikingly similar to the characteristics shown by M. tuberculosis[4, 6, 7]. This suggests that the dormancy response to oxygen depletion is not a unique characteristic of the pathogenic mycobacteria. Rather it might be – at least in part – a conserved response in slowgrowing pathogenic as well as in fastgrowing saprophytic mycobacteria.

The demonstrated physiological similarities between M. smegmatis and M. tuberculosis are intriguing in the light of the recently started molecular analysis of anaerobic dormant M. tuberculosis[9–12]. Two genes were proposed to play a role in the dormancy response: the sigma factor sigF[9] and the chaperon alpha crystallin [10, 11]. Both genes are not present in M. smegmatis[9, 10]. Furthermore, a thickening of the cell wall was described for anaerobic M. tuberculosis but not for anaerobic M. smegmatis culture [11]. We propose two interpretations of the discrepancies between the similarities in the physiological dormancy response we have demonstrated and the differences at the molecular and morphological level between the two mycobacteria identified to date. Firstly, the molecular machinery involved in dormancy could be different between M. smegmatis and M. tuberculosis but lead to a physiologically similar dormancy response. Secondly, an interpretation we favour, there might be so far unidentified molecular mechanisms controlling the dormancy response which are similar in both closely related species. The molecular and morphological factors known to be different between the two mycobacteria might play important but perhaps not exclusive roles in the mycobacterial dormancy response. Taken together, we believe that the fastgrowing M. smegmatis might be useful as a non-pathogenic model for a comparative molecular analysis of mycobacterial dormancy.

Acknowledgements

We would like to thank Erik Böttger and Peter Sander, University of Hannover, for the ‘mycobacterial starter kit’ and the time they spent with one of us (T.D.) in their laboratory in the summer of 1996. We are grateful to Lawrence Wayne, University of California, Irvine, CA, for his thoughts on mycobacterial dormancy. We thank Bernd Hutter for discussion and Chris Tan, Institute of Molecular and Cell Biology (IMCB) for his support and his thoughts on tuberculosis. This study was supported by the IMCB.