Cell Cycle Characteristics of Crenarchaeota: Unity among Diversity^▿

Strains.

A. pernix DSM 11879, P. aerophilum DSM 7523, Sulfolobus solfataricus DSM 1617, Sulfolobus acidocaldarius DSM 639, and S. tokodaii DSM 16993 were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) (Braunschweig, Germany). P. calidifontis VA1 (1) and A. hospitalis (6) were kindly provided by Haruyuki Atomi and David Prangishvili, respectively.

Media and growth conditions.

P. aerophilum was grown under anaerobic conditions at 100°C in basal salts medium supplemented with 0.1% KNO₃ and 0.05% yeast extract (30). P. calidifontis was grown aerobically at 90°C in TY medium (1% tryptone, 0.1% yeast extract, 0.3% sodium thiosulfate) (1). S. acidocaldarius and S. solfataricus were grown aerobically at 79°C in modified Allen mineral base medium supplemented with 0.2% tryptone (11). S. tokodaii was cultivated in Allen medium supplemented with 0.1% tryptone, 0.1% yeast extract, and 0.2% sucrose. A. pernix was grown aerobically at 90°C in Aeropyrum-JXT medium (DSMZ medium 820). A. hospitalis was grown at pH 3.5 in a medium containing 0.2% tryptone, 15 g/liter sulfur, 0.3 mM Ca(NO₃)₂, 1 mM MgCl₂, 26 mM (NH₄)₂SO₄, 2.2 mM K₂HPO₄, 1.4 mM KCl, 0.7 g/liter glycine, 4.5 μM MnCl₂, 5.9 μM Na₂B₄O₇, 0.08 μM ZnSO₄, 0.15 μM CuCl₂, 0.45 μM Na₂MoO₄, 0.06 μM VOSO₄, 0.02 μM CoCO₄, 0.02 μM NiSO₄, 20 μg/liter biotin, 20 μg/liter folic acid, 100 μg/liter pyridoxamine hydrochloride, 50 μg/liter thiamine hydrochloride, 50 μg/liter riboflavin, 50 μg/ml nicotinic acid, 50 μg/liter dl-calcium pantothenate, 1 μg/liter cyanocobalamine, 50 μg/liter p-aminobenzoic acid, and 50 μg/liter lipoic acid. Growth was monitored by counting cells using a Thoma chamber (depth, 0.02 mm) or by determining the optical density at 600 nm for all species except P. calidifontis, for which the optical density at 660 nm was determined. The cultures were grown for up to 10 cell generations before sampling at mid-exponential phase, as well as in stationary phase.

Flow cytometry.

The cells were prepared for flow cytometry as described previously (5), and measurements were obtained with an A40 analyzer (Apogee Flow Systems). The relative lengths of the cell cycle periods were calculated from the flow cytometry DNA content distributions, using Winflow 2.0 software kindly provided by Flemming Hansen, with correction for the exponential age distribution of the populations (33). Flow cytometry reference samples from exponentially growing M. thermautotrophicus DSM 1053 cultures were obtained from experiments performed as described by Majernik et al. (21). Flow cytometry reference samples from rifampin-treated Escherichia coli MG1655 seqA::Tn10 cultures (23) were kindly provided by Jan Olsson.

Microscopy.

Samples for microscopy analysis were collected like the samples used for flow cytometry, and cell mounting was performed as described previously (24). Microscopy was performed with a Nikon Optiphot-2 epifluorescence microscope.

Cell cycle characteristics.

To generate an overview of cell cycle features in the phylum Crenarchaeota, the following species belonging to three different orders were analyzed: A. hospitalis and S. tokodaii belonging to the Sulfolobales, A. pernix belonging to the Desulfurococcales, and P. aerophilum and P. calidifontis belonging to the Thermoproteales. These species represent branches that are widely separated in the phylogenetic tree and have different cell morphologies, pH requirements, optimal growth temperatures, generation times, and genome sizes (Table 1).

Flow cytometry revealed that in exponential phase, cells with two fully replicated genome copies (postreplicative cell cycle phase) dominated the growing population for all species analyzed (Fig. 1). The peak representing the population of cells with a single genome copy (prereplicative) was small and overlapped extensively with the peak representing the population that contained between one and two genome equivalents (replicative phase), showing that in all cases the G₁ (B) period was short. The DNA content distributions were used to calculate the relative lengths of the cell cycle phases (see Materials and Methods), and the values obtained are shown in Table 1.

Stationary phase.

In stationary-phase P. calidifontis and S. tokodaii cultures, most cells contained two fully replicated genome copies (Fig. 1L and 1P). Peaks corresponding to four genome copies were also observed at a low frequency. This probably resulted either from incomplete cell separation after division or from cell aggregation, since no single cells in which the fluorescence signal was significantly higher than the average signal could be detected by quantitative epifluorescence microscopy (not shown). For P. calidifontis, a small population with either one genome copy or between one and two genome copies was still observed long after the optical density had ceased to increase, indicating that chromosome replication did not run to completion in all cells under the conditions used. In contrast to the other organisms, the proportion of cells with a single chromosome increased in stationary-phase A. pernix cultures (Fig. 1H) compared to exponential phase (Fig. 1F), indicating that there was preferential arrest in the prereplicative phase, although cells with a higher DNA content also remained in the culture. For A. hospitalis, cells with a DNA content of less than one chromosome equivalent appeared in stationary phase (Fig. 1D), indicating either that DNA degradation took place or that chromatin restructuring in nongrowing cells resulted in a reduced dye binding capacity (5, 24).

The average light scatter decreased for S. tokodaii in stationary phase (compare Fig. 1O to Fig. 1M). No large changes were apparent for A. hospitalis, A. pernix, and P. calidifontis (compare Fig. 1C, 1G, and 1K to Fig. 1A, 1E, and 1I, respectively). However, light scatter reflects not only cell size, since this parameter also is affected by cell morphology and composition. Between species there was an obvious lack of correlation between light scatter and cell size; A. pernix (Fig. 1E and 1G) and P. calidifontis (Fig. 1I and 1K) displayed much lower light refraction than A. hospitalis (Fig. 1A and 1C) and S. tokodaii (Fig. 1M and 1O), although they were in the same size range. In the cases where DNA content distributions indicated incomplete replication and/or chromosome degradation (see above), changes in light scatter may have occurred as a result of cell shrinkage or swelling as a consequence of loss of cell envelope integrity in stationary phase under the extreme growth conditions used (3, 12).

Genome size estimation.

In flow cytometry, the fluorescence signal from stained chromosomes is linearly proportional to the amount of intracellular DNA, although the G+C content and chromatin structure may affect the absolute signal strength (5, 9). Thus, for the species with unknown genome sizes, estimates could be obtained by calibration with samples from organisms with known genome sizes and DNA contents (A. pernix, E. coli seqA::Tn10, M. thermautotrophicus, S. acidocaldarius, S. solfataricus, and S. tokodaii), as previously described (4, 21). The genome size was estimated to be 1.8 Mb for both P. calidifontis and A. hospitalis (Fig. 2 and Table 1).

Nucleoid structure and localization in P. aerophilum.

The nucleoids in P. aerophilum were usually clearly separated and preferentially distributed toward the cell poles (Fig. 3B, 3D, and 4B). Small P. aerophilum cells contained a single nucleoid, whose length gradually increased with cell length, until two separated nucleoids could be distinguished (Fig. 4A and 4B). Two nucleoids were present in 48% of the cells (Table 2), and the cells were between 3.5 and 7.5 μm long (Fig. 4A). The minimal nucleoid length in cells with two nucleoids (1.6 μm) (Fig. 4C) was slightly less than one-half the minimal length in cells in which two separated nucleoids could be detected (3.5 μm) (Fig. 4D). Nucleoid length appeared to increase further after the formation of two distinct fluorescence foci (Fig. 4B and 4E). P. calidifontis displayed similar characteristics, with 38% of the cells containing segregated nucleoids during exponential growth (not shown).

P. aerophilum cell division.

In most prokaryotes, cell division occurs by gradual invagination of the cytoplasmic membrane and the surrounding cell envelope, resulting in a constriction. However, no constrictions were observed in P. aerophilum. Instead, microscopy analysis revealed the presence of bent cells (Fig. 5A), with a bend angle between 0° and 45°. The bending occurred in cells longer than 5 μm with clearly segregated nucleoids. Divided cells that had not separated completely were also detected (Fig. 5B), indicating that the bent cells split in a rapid process. The split was not always centered, so that the daughter cell sizes in unseparated pairs ranged from 2.3 to 3.4 μm. An unusual feature was the presence of short particles that had the same diameter as regular cells but were devoid of DNA (Fig. 5C). The lengths of these particles ranged from 0.7 to 2 μm (average length for 70 short cells measured, 1.1 μm).

Prereplication period.

The prereplication period is short (<5% of the cell cycle) in S. acidocaldarius and S. solfataricus (5). Here, we found that this period is also short in A. hospitalis, A. pernix, P. calidifontis, and S. tokodaii. Thus, replication is initiated shortly after cell division in all Crenarchaeota studied to date. It is possible that completion of cell division could provide a positive signal for replication initiation or could release a replication block that allows subsequent initiation.

Replication period and replication rate.

The replication period accounted for between 13 and 31% of the doubling time, depending on the species, allowing estimation of replication fork progress rates. The in vivo replication rates of S. solfataricus and S. acidocaldarius, both of which contain three replication origins in a single chromosome, are on the order of 80 to 110 bp/s (18). The origin recognition boxes associated with the origins in these species (26, 27) are also conserved in the S. tokodaii genome sequence (15), and it is therefore likely that this species also harbors three origins, resulting in a replication rate of around 70 bp/s. A. pernix has been shown to contain two origins (26), yielding an in vivo replication rate of 110 bp/s. We have shown that flow cytometry is a convenient tool for genome size determination (4, 21), even for the eukaryote Giardia lamblia. Estimation of the genome size for A. hospitalis and P. calidifontis (Table 1) suggested that these organisms synthesize DNA at rates of 210 and 390 bp/s, respectively, assuming that there is bidirectional replication from a single origin. It remains to be determined whether these species also may contain additional origins, which would reduce these estimated rates to values similar to those for the other species.

Postreplication period.

The G₂ period, the time interval between termination of replication and initiation of genome segregation, is long in Sulfolobus species (12, 24). The crenarchaeal species characterized here also have a long postreplication period, and we consider it likely that these organisms, with the exception of Pyrobaculum (see below), also go through an extensive G₂ period, in agreement with the low frequency of cells displaying clearly segregated nucleoids. Furthermore, since very few dividing cells were observed in exponentially growing S. tokodaii cultures, the cell division period is also short (≤5% of the doubling time) in this species, similar to the situation in S. acidocaldarius and S. solfataricus (5).

Chromosome segregation and cell division in Pyrobaculum.

The Pyrobaculum species displayed a long postreplicative phase, similar to the organisms representing the Desulfurococcales and Sulfolobales. However, a G₂-like period appeared to be absent since approximately 48% of the cells in actively growing populations contained two clearly separated nucleoids (Table 2), indicating that genome segregation occurred during or shortly after replication. Furthermore, a correlation between cell length and distance between nucleoids was evident, and nucleoid length appeared to increase with cell length, indicating that genome segregation and nucleoid extension continued during the postreplicative period. It is, therefore, possible that cell growth contributes to nucleoid segregation in this organism, e.g., through nucleoid association with the inner cell surface. Addition of cell wall material only at the midsection could contribute to nucleoid separation but would not cause elongation of the nucleoids. Importantly, cell growth-coupled partitioning does not rule out the possibility that there is an active genome segregation mechanism that operates during or shortly after replication.

No constrictions were visible in actively growing P. aerophilum cells. Rapid splitting in the cell center after bending has been described for Thermoproteus tenax (13) and may be a common feature of Thermoproteales, including Pyrobaculum species. In T. tenax, invagination of the cytoplasmic membrane in the cell center, accompanied by new S layer synthesis, eventually results in two daughter cells apparently connected by a small compartment (31). It is possible that the DNA-free minicells detected in P. aerophilum cultures could correspond to this central region, released as a by-product of cell division.

Concluding remarks.

The overall cell cycle features were found to be similar in hyperthermophilic organisms representing three different orders in the Crenarchaeota, despite considerable evolutionary distance and vast differences in cellular properties and growth preferences. In contrast, organisms belonging to the phylum Euryarchaeota display more variable cell cycle architectures (19). It will be of significant interest to investigate to what extent the conservation of crenarchaeal cell cycle features extends to low-temperature species (17).