Correspondence: Aleksey M. Anuchin, Bach Institute of Biochemistry, Russian Academy of Sciences, Leninsky prospect, 33/2, Moscow, Russia. Tel.: +7 495 954 4047; fax: +7 495 954 2732; e-mail: email@example.com
The role of histone-like protein (Hlp) in the development of a dormant state in long-incubated stationary-phase Mycobacterium smegmatis cells was studied in two models: (1) adoption of ‘nonculturable’ (NC) state, which is reversible due to resuscitation with proteinaceous resuscitation-promoting factor (Rpf) and (2) the formation of morphologically distinct, ovoid resting forms. In the first model, inactivation of the hlp gene resulted in prolongation of culturability of starved cells followed by irreversible nonculturability when mycobacterial cells were unresponsive to resuscitation with Rpf. In the second model, M. smegmatis strain with the inactivated hlp gene was able to form dormant ovoid cells, but they were less resistant to heating and UV radiation than those of wild-type strain. The susceptibility of ovoid cells produced by Δhlp mutant to these damaging factors was probably due to a less condensed state of DNA, as revealed by fluorescent microscopy and DAPI staining. Evidently, Hlp is essential for cell viability at a later stage of NC dormancy or provides a greater stability of specialized dormant forms.
One of the most important strategies adopted by bacteria to cope with unfavorable factors is the ability to enter a dormant state in which cells preserve viability for a long time, acquire stress resistance and shut down metabolic activity (Lewis, 2007). Mechanisms responsible for the acquiring and maintenance of dormancy in spore formers are well established, but not much is known for nonsporulating bacteria – causative agents of infectious diseases, in particular tuberculosis. Recently, a sensational study on endospore formation in Mycobacterium marinum has been published (Ghosh et al., 2009); however, this claim was not confirmed in a later study (Traag et al., 2010). According to WHO, one-third of the world's population is latently infected with Mycobacterium tuberculosis (MTB) (Inge & Wilson, 2008), which likely persist as dormant cells in the human organisms, posing a significant problem due to resistance to chemotherapy (Mitchison, 1980). Although dormancy is the commonly accepted explanation of latent mycobacterial infection (Young et al., 2005), limited information has been available about persisting bacterial forms and molecular mechanisms behind their stability and resistance to stressful factors.
Among the known mechanisms responsible for the adoption of stress resistance of bacterial cells, it is worth considering the role of histone-like proteins, which bind DNA, changing its topology (Dorman & Deighan, 2003) and making it more stable against damage caused, for example, by γ or UV radiation (Boubrik & Rouvière-Yaniv, 1995). In Escherichia coli, histone-like proteins HU, H-NS, FIS also play an important role in transcription, recombination and replication (Thanbichler et al., 2005 and references therein). Histone-like protein, Hlp, is present in Mycobacterium smegmatis and contains the N-terminal domain, homologous to HU and the C-terminal domain with the mycobacterial specific PAKKA motif (Mukherjee et al., 2008). Regarding the physiological function of Hlp, it is worthwhile to note the significant increase in its level during transition of M. smegmatis cells to a nonreplicating state under microaerophilic conditions in the Wayne dormancy model. However, the viability of cells of M. smegmatis strain with inactivated hlp gene was not clearly distinct from that of wild-type strain (Lee et al., 1998) in the same dormancy model. We may reason that Hlp has no significant role in the transition to dormancy in the relatively short-term Wayne model but may be essential for developing dormancy in nonreplicating cells at later stages. Indeed, many genes, different from those expressed in cells undergoing starvation in the Wayne model, are upregulated at late stages (>24 h) in M. tuberculosis cells subjected to hypoxia (enduring response) (Rustad et al., 2008).
The objective of the present study is to clarify the role of Hlp in dormancy in M. smegmatis cells obtained in two experimental models after incubation in a prolonged stationary phase. We found that Hlp was essential for survival of NC cells or for a greater stability of specialized dormant forms, likely due to DNA condensation.
Materials and methods
Strains and growth conditions
Strains and plasmids used in this study are listed in Table 1. Mycobacterium smegmatis strain MC2 155 was routinely maintained on solid (1.5% agar) NB medium (HiMedia Laboratories, India); only fresh 3–5-day-old colonies served as inoculum to produce starter cultures. Cells were grown in the standard Sauton medium or liquid NB (nutrient broth) medium, as well as in the modified Hartman-de-Bont medium (Shleeva et al., 2004) or modified SR-1 medium (Anuchin et al., 2009) as described below. In standard CFU assays, aliquots of decimally diluted cell suspensions were plated on solid NB medium. The Δhlp strain and recombinant strains (Table 1) were maintained on the NB medium with 10 μg mL−1 kanamycin and grown under the same conditions as the Wt strain.
Table 1. Strains and plasmids used in the study
Kanamycin-resistant strain with inactivated hlp gene
The hlp gene was amplified by PCR using the forward primer 5′-GTGGATCCTGGAAATCAGTGGTCACAG-3′ and the reverse primer 5′-ATCTGCAGCCTCCCGACGAGAAGTAACG-3′ (BamHI and PstI restriction sites are in bold). The purified PCR product was ligated into the pGEM-T vector (Promega), resulting in pGEM-hlp, which was introduced into E. coli strain DH5α. Transformed clones were selected and examined by PCR. Thereafter, pGEM-hlp and pMind were digested with restriction enzymes BamHI and PstI and the hlp fragment was ligated into the pMind vector. The ligated product pMind-hlp was introduced into E. coli strain DH5α, and the sequence of the cloned gene was confirmed. All vectors were introduced into E. coli by electroporation according to the BioRad protocol; to incorporate the vectors in M. smegmatis cells, we used the procedure as described elsewhere (Parish & Stoker, 1998).
Isolation and purification of recombinant resuscitation-promoting factor (Rpf) protein
A truncated form of Micrococcus luteus Rpf, named RpfSm, served as an additive in resuscitation medium. RpfSm contained the conserved Rpf domain followed by 20-aa fragment of variable domain: ATVDTWDRLAECESNGTWDINTGNGFYGGVQFTLSSWQAVGGEGYPHQASKAEQIKRAEILQDLQGWGAWPLCSQKLGLTQADADAGDVDATE. The truncated gene was amplified by PCR from the pET-19b-Rpf (Mukamolova et al., 1998), using the T7 promoter primer: GCGAAATTAATACGACTCACTAT and the reverse primer: CGACGGATCCTCACTCGGTGGCGTCACGT (the BamH1 restriction site is marked in bold). The purified PCR product was digested with XbaI and BamH1, purified and ligated into pET19b vector, which was introduced in E. coli DH5α. The construct, containing the truncated rpf gene (rpfSm), was sequenced and used to transform E. coli HSM174 (DE3). RpfSm was purified from 350 mL cultures of E. coli producer strain grown at 37 °C in the rich medium (HiMedia) with ampicillin (100 μg mL−1) to OD600 nm 0.65–0.8. After induction with 1 mM IPTG, growth was continued for 2 h at room temperature. Cells were harvested by centrifugation at 3000 g for 15 min and frozen in binding buffer (BB) (20 mM Tris-HCl, pH 8.0; 0.5 M NaCl; 5 mM imidazole). Thawed cell suspensions in 10 mM MgSO4 were treated with RNAse and DNAse at concentrations 10 μg mL−1 each and then with 8 M urea. After sonication, the crude extract was centrifuged at 6000 g for 30 min to remove cell debris, and supernatant was applied onto a 2-mL Ni2+-chelation column (Sigma) equilibrated with BB. The Biological LP system (BioRad) was used for elution and refolding of the protein on the Ni-column: first, a series of washing steps with the BB, containing 8 M urea, were used. The second, refolding step included washing with BB at linearly decreasing urea concentrations (from 8 to 0 M). The protein was eluted with a linear gradient of imidazole from 5 to 500 mM. Protein was collected at 0–250 mM imidazole concentrations in a total volume of 4–5 mL. Rpf-containing fractions (30–50 μg mL−1) were dialyzed against 50 mM citric acid–sodium citrate buffer (pH 6.0). Protein samples were stored at +4 °C for 1 week without a significant loss in its activity.
Formation and resuscitation of dormant M. smegmatis cells
Myñobacterium smegmatis strains were grown under the conditions that favored the entering wild-type strain to ‘nonculturable’ (NC) state (inability to produce colonies on solid media) in stationary phase after cultivation of mycobacteria in the modified Hartman-de-Bont medium, lacking K+, at 37 °C for 120 h under aeration (Shleeva et al., 2004). In the other model, the strains under study were incubated for 4.5 months after growth in N-limited SR-1 medium to produce morphologically distinct ovoid cells (Anuchin et al., 2009).
The ability of ‘NC’ cells to resuscitate in liquid medium was estimated using the most probable number (MPN) assays in triplicate repeats with inoculation of 0.1 mL cell suspensions to 0.9 mL of the modified Sauton medium in plastic 48-well microplates (Corning) as described previously (Downing et al., 2005). The Sauton medium that served for resuscitation contained (L−1): KH2PO4, 0.25 g; MgSO4·7H2O, 0.25 g; l-asparagine, 2 g; glycerol, 6 mL; ferric ammonium citrate, 0.025 g; sodium citrate, 1 g; 1% ZnSO4, 0.05 mL; ± recombinant RpfSm protein, 5 μg (pH 7.0).
Light and fluorescence microscopy
Cell suspensions were examined under a microscope Eclipse E4000 (Nikon, Japan) in the phase-contrast and epifluorescence modes after staining with propidium iodide (3 μM) to detect injured/dead cells or with 4′-6-diamidino-2-phenylindole (DAPI) (2 μg mL−1) bound to double-helix DNA. Excitation was at 510 and 330 nm, and emission was at >560 and >380 nm for propidium iodide and DAPI, respectively.
Heat and UV treatment
One-milliliter aliquots were taken from stationary-phase (48 h) cultures in NB medium or from cultures stored for 4.5 months in N-limited SR-1 medium and were transferred into Petri dishes with 4 mL of liquid NB medium and then subjected to UV irradiation (BUV-30 lamp, 254 nm) as described elsewhere (Vorobjeva et al., 1995). Samples from the same cultures were also heated at 60–80 °C for 10 min. Cells after UV or heat treatment were plated onto solid NB medium for CFU assays.
Results and discussion
As already demonstrated, after cultivation for 68–70 h in the modified Hartman-de-Bont medium without K+ sources, stationary-phase wild-type M. smegmatis cells entered a dormant NC state and lost the ability to form colonies on the nutrient agar. NC cells of the wild-type strain were resuscitated in a liquid medium supplemented with Rpf. Similarly, the isogenic strain (Wt∷rpf) that harbors a plasmid containing the M. luteus rpf gene, also adopted the NC state, but resumed growth when transferred to the appropriate liquid medium without added Rpf (Shleeva et al., 2004). The previously designed approach to produce and resuscitate NC cells was applied to study M. smegmatis strain with inactivated hlp gene. Our experiments revealed that M. smegmatisΔhlp strain and its derivatives developed in the modified Hartman-de-Bont medium (without K+) at similar growth rates compared with the wild-type strain (data not shown). At the stationary phase, the Δhlp strain entered the NC state (0 CFU) later than the Wt-pMind strain with the empty plasmid (90–96 vs. 68–70 h) (Fig. 1a) or wild type (not shown). Complemented strain Δhlp∷hlp harboring the hlp gene on the plasmid entered the NC state only 2 h later than the Wt-pMind strain (data not shown). Next, NC cells of Wt-pMind and Δhlp strains were tested for their ability to resuscitate in the presence of recombinant M. luteus RpfSm protein, which appeared to be more active and stable during storage compared with full-length M. luteus Rpf (unpublished data). Contrary to Wt-pMind, NC cells of the strain Δhlp failed to resuscitate in the liquid medium with RpfSm (Fig. 2). NC cells of the complemented strain Δhlp∷hlp were partially resuscitated by RpfSm (Fig. 2). Therefore, the lack of Hlp resulted in transition of mycobacterial cells to an irreversibly NC (or likely, moribund) state under chosen conditions (K+ depletion) but not to a resuscitatable dormancy. Taken together, these data led to the conclusion that Hlp does not affect the onset of the transition to nonculturability but seems to be essential for cell viability at a later stage of dormancy.
It was noteworthy that the strain Δhlp∷rpf, lacking the hlp gene and harboring the plasmid-carried rpf gene, substantially differed from the Δhlp in culturability when cultivated in the modified Hartman-de-Bont medium. In particular, Δhlp∷rpf cells maintained the ability to produce colonies even during a prolonged (180 h) stationary phase, as opposed to Δhlp and Δhlp-AGH strains (Fig. 1b). However, insertion of the rpf gene to wild-type M. smegmatis (Wt-AGR) caused the development of NC state of cells incubated in the same medium and conditions (Fig. 1b), as reported previously (Shleeva et al., 2004). To ensure that the maintenance of plateability in long-stored Δhlp∷rpf cultures was indeed due to Rpf production, we studied the Δhlp∷rpf mut strain (Table 1) with the rpf gene disrupted by site-directed mutagenesis (Mukamolova et al., 2006). Our experiments demonstrated that mutations in the rpf gene restored the ability to adopt the NC state under the given cultivation conditions (Fig. 1b). Thus, the combined action of a downshift of Hlp and an upshift of Rpf greatly prolonged the ability of M. smegmatis to endure starvation in a completely culturable state, revealing opposite effects of two proteins on the formation of ‘nonculturability’. Earlier we found that M. smegmatis culture in stationary phase contained a proportion of viable cells which are capable of cryptic growth (Shleeva et al., 2004). We speculate that Rpf as a growth factor (Mukamolova et al., 1998) promotes multiplication of a similar population of viable cells as presented in a moribund Δhlp culture. This would result in dynamic equilibrium between cell death and growth and CFU, maintaining a stable level. Analogously, the delay in transition to NC state by Wt∷rpf strain, harboring the rpf gene (Fig. 1b), may reflect the Rpf-mediated growth stimulation of some cells in the population. The significantly different behavior of Δhlp∷rpf and Δhlp strains may be discussed from the point of view of the dual mode of Rpf action: growth-supportive with respect to debilitating populations (as with Δhlp strain) or per se resuscitative to nonplateable dormant cells produced by Wt or Δhlp∷rpf strains. Taken together, our results suggest that Hlp plays a role in the adoption of reversible NC in M. smegmatis at later stages of cultivation in the appropriate medium.
In the second set of experiments with Δhlp strain, we used the approach previously developed to obtain morphologically distinct ovoid dormant cells of Wt M. smegmatis after cultivation in the N-limited SR-1 medium. Ovoid dormant cells survived for several months and possessed a low metabolic activity level and elevated resistance to heating and antibiotics. Long-stored cultures of these cells contained a large proportion of NC cells that resumed growth in liquid media (Anuchin et al., 2009). Growth rates of Δhlp cells in the Sauton and modified SR-1 media were the same as those of the Wt strain (data not shown). When cultivated in SR-1 medium, Δhlp cells also produced ovoid dormant forms, like the wild-type strain (Fig. 3). However, ovoid forms of Δhlp strain were considerably less stable to elevated temperature or UV exposure than were dormant forms of Wt-pMind strain (Figs 4 and 5). Complemented strain Δhlp∷hlp revealed intermediate sensitivity to elevated temperature (Fig. 4). Similarly, Δhlp∷hlp demonstrated partial restoration of stability to UV treatment (1.3±0.75%, 0.2±0.097%, 0.02±0.014% of initial CFU mL−1 after 44, 97 and 146 J m−2 irradiation dose, respectively).
Hence, we may conclude that, despite the ability of mycobacterium with inactivated hlp gene to produce ovoid dormant cells, Hlp confers their resistance to stress conditions, consistent with published results as discussed below. An extreme increase was shown in the Hlp level in M. smegmatis cells subjected to cold shock (0 °C) and the inability of the strain with the inactivated hlp gene to grow at 10 °C (Shires, 2001). As to the action mechanism, it is possible that Hlp serves as a physical shield against stress factors that impair DNA, as in the case of another histone-like protein, Lsr2, in M. tuberculosis, which protects DNA from reactive oxygen intermediates (ROI) in vitro and during macrophage infection (Colangeli et al., 2009).
Examinations of DAPI-stained ovoid cells under epifluorescence microscope revealed clearly distinguishable areas of compact DNA in dormant ovoid Wt-pMind and Δhlp∷hlp cells, absent in Δhlp-pMind strain (Fig. 5). Evidently, Hlp caused changes in the nucleoid architecture in dormant M. smegmatis cells, similar to the DNA condensation in E. coli cells demonstrated to be the result of binding to Hlp (Mukherjee et al., 2008).
Another histone-like protein, Hc1, is responsible for nucleoid condensation in specialized dormant forms (reticular bodies) of chlamydia. A reverse process of DNA decondensation due to Hc1 dissociation in chlamydial dormant cells is controlled by the ispE gene product, an enzyme of nonmevalonic pathway of isoprenoid synthesis (Grieshaber et al., 2004, 2006). In this line, we have demonstrated self-reactivation of stationary-phase M. smegmatis NC cells due to ispE hyperexpression (Goncharenko et al., 2007).
Notwithstanding the significant increase of Hlp level in M. smegmatis cells under hypoxia conditions in the Wayne dormancy model inactivation of the hlp gene caused no phenotypic changes, as judged from ability of Δhlp strain to develop a nonreplicating state (Lee et al., 1998). In contrast to models used in the present study, the Wayne model reflects adaptation of cells to oxygen starvation when cells remain fully culturable and do not produce morphologically distinct dormant forms (Cunningham & Spreadbury, 1998). The results obtained in our study, exemplified by M. smegmatis, clearly show the significance of Hlp protein for the formation and stress resistance of two types of deeply dormant mycobacterial cells. Hlp (or other histone-like proteins) may be engaged in mechanisms responsible for prolonged persistence and stability of tubercle bacilli; however, further experiments are required to verify this possibility for MTB cells.
We thank Brian Robertson for providing the pMind plasmid, Thomas Dick for Δhlp strain and Galina Mukamolova for pAGH, pAGR and pAGRmut plasmids. This work was supported by the Programme ‘Molecular and Cellular Biology’ of the Russian Academy of Sciences and NM4TB EU project.