INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Most knowledge about the regulation of bacterial cell division and replication of DNA stems from the analysis of only three species, Escherichia coli (43, 44), Caulobacter crescentus (35, 37), and Bacillus subtilis (25, 34). In some cyanobacteria these processes are reported to be under the control of a circadian clock (1, 5, 15, 18, 24, 39). However, studies directly concerning the diel expression of cell cycle-relevant genes in cyanobacteria are scarce (21).

The cell cycle of the marine cyanobacterium Prochlorococcus is characterized by a well-defined and discrete DNA synthesis phase, S (42). In the field, the cell cycle is highly synchronized by the daily alternation of night and day. DNA replication occurs in late afternoon, and cell division occurs at night (15, 20, 41, 42). It is not known at which stage or by which regulatory mechanism the linkage between cell cycle and environmental conditions is achieved. Experiments in which the time of light onset was changed suggested that the passage from darkness to light (equivalent to sunrise) might be involved in timing of DNA synthesis in Prochlorococcus (16a).

To elucidate potential components involved in the synchronization and diel control of cell cycle progression in Prochlorococcus, the genes dnaA and ftsZ were cloned and analyzed. The GTP-binding protein FtsZ is widely distributed among eubacteria, archaea, and plastids and is usually considered the key factor in the initiation of cell division by the formation of a ring-shaped structure that recruits several other proteins (FtsA, FtsQ, and FtsW) to the division site (8). FtsZ has also recently been found in a mitochondrion (2), where it is normally replaced by Dynamin (reviewed in reference 9). DnaA is a ubiquitous bacterial protein that acts as a helicase to initiate DNA replication in eubacteria. In E. coli, it recognizes asymmetric 9-bp AT-rich elements, called DnaA boxes, near the origin of replication, oriC. Furthermore, it acts as a repressor for its own expression and as a transcriptional regulator for other genes (23).

Laboratory cultures of Prochlorococcus are difficult to grow and to maintain axenically. Other obstacles are that average cell densities reached by Prochlorococcus in culture are lower than for most bacteria and that its cell size is particularly small (0.6 µm on average), leading to low biomass yields (27). Here, this issue was resolved using a large-volume turbidostat exposed to a dynamic light regime, with irradiance progressively varying in a bell curve-like fashion between 0 and about 1,000 quanta (micromoles meter² second¹) during the 12-h photoperiod (6). These conditions allowed us to simulate average light conditions found in the upper mixed layer of oceanic waters near the equator. Using this system, the expression of ftsZ and dnaA was monitored in synchronized cultures of Prochlorococcus sp. strain PCC 9511.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Culture conditions and sampling. Two replicate 10-liter turbidostat cultures of the axenic Prochlorococcus sp. strain PCC 9511 (31) were grown in PCR S11 medium (27) in 20-liter polycarbonate flasks, placed in a thermoregulated bath at 21 ± 1°C and under a cycle of 12 h of light and 12 h of dark (L/D) (light from 8:00 a.m. to 8:00 p.m.). In two related studies, these times were shifted by 2 h in order to have solar noon at 12:00 (6, 12). These papers report details about the turbidostat setup and light systems (6) as well as the expression of photosynthetic genes (12).

2

1

Preparation and analysis of DNA. Most DNA manipulations were carried out according to standard protocols (36). Prochlorococcus sp. strain PCC 9511 DNA was purified from freeze-dried cells as described previously (14). PCR was performed using 10 ng of DNA, 10 pmol of each primer, 250 µM concentrations of each deoxynucleoside triphosphate, 2 U of Taq DNA polymerase (Qiagen) or AmpliTaq Gold (Perkin-Elmer), and 1× Taq buffer supplemented with 2.5 mM MgCl₂. After initial denaturation at 93°C for 5 min, the reaction mixtures were heated at 93°C for 45 s. Annealing was performed for 45 s at 54°C for a 571-bp ftsZ fragment or 62°C for the amplification of a 540-bp PCC 9511 dnaA fragment. Elongation occurred at 72°C for 1 min. After 35 cycles, the final step at 72°C was extended for 5 min. Plasmid and PCR product purification was done using commercial kits (Qiagen and Genomed). Products of cycle sequencing reactions (Bigdye terminator cycle sequencing kit) were separated on an ABI 373 automatic sequencer (Applied Biosystems Inc., Perkin Elmer). The 571-bp PCR fragment of ftsZ obtained by primers FTF and TFR (3) was cloned, yielding pPCCftsZ571 (strains and plasmids are reported in Table 1). This plasmid served as a template to generate single-stranded RNA probes or as a probe in Southern hybridization to isolate the ftsZ coding region from a size-selected HindIII plasmid minibank in vector pBluescript SK(+). A genomic library of Prochlorococcus sp. strain PCC 9511 was established by the ligation of partially digested EcoRI fragments into ZAPII (Stratagene). This library was screened for dnaA by hybridization using the Prochlorococcus marinus SS120 dnaA gene as a probe (30). The RNase P probe was kindly provided by Astrid Schön, Würzburg, Germany. Primers for the generation of probes and sequencing are listed in Table 2. Preliminary sequence data for Prochlorococcus MED4 were obtained from the DOE Joint Genome Institute at http://spider.jgi-psf.org/JGI_microbial/html/.

RNA extraction and Northern hybridization. RNA was extracted from 400 ml of culture for each sampling point as described (10). Northern hybridization was carried out at 61.5°C for single-stranded RNA probes and 50°C for DNA probes in 120 mM sodium phosphate buffer (pH 7.2)250 mM NaCl7% sodium dodecyl sulfate (SDS)50% formamide. RNA probes were produced with 1.85 MBq of [-³²P]UTP (Amersham) using the Maxiscript transcription kit (Ambion). Plasmids pPCCftsZ571 and pPCCdnaA540 served as templates. RNase protection assays were performed in accordance with the manufacturer's instructions (RPAIII kit; Boehringer).

Immunology. For protein extraction, cells were collected by centrifugation, disrupted by adding 0.1% SDS, sonicated three times for 10 s each time at 4°C with a Sonopuls HD 60 set at 50% of maximum power, and incubated twice at 95°C for 5 min. Protein concentrations were measured using the Bio-Rad protein assay. A polyclonal antiserum against recombinant Anabaena sp. strain PCC 7120 FtsZ (courtesy of C.-C. Zhang) was used for expression analysis (19). Western blots were prepared from total proteins separated on SDS12% polyacrylamide gels (normalized to 1 µg per lane) and blotted on Hybond-C extra membranes (Amersham). Incubation with antisera was performed at titers of 1:1,000 (FtsZ antibody). Secondary antisera were conjugated with horseradish peroxidase, and blots were developed with the chemiluminescence substrate SuperSignal (Pierce). Signals were quantified using PCBAS 2.09 software.

Nucleotide sequence accession numbers. The sequences reported in this paper have been deposited in the EMBL database under the accession numbers AJ011025 and AF158628.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Organization of the genomic regions encoding DnaA and FtsQ to FtsZ in Prochlorococcus sp. strain PCC 9511. The gene arrangement around dnaA (Fig. 1A) is very unusual compared to gene arrangements in other eubacteria. However, it is apparently highly conserved among different Prochlorococcus strains, since almost the same arrangement is found in P. marinus SS120 (30). ORF1 shows pronounced similarity to the gene encoding YCF25 (AAC08078), a protein encoded in the plastid genome of Porphyra purpurea (29). The other hypothetical proteins are similar to the products of several ORFs in Synechocystis sp. strain PCC 6803 (17). Over the whole length, the amino acid sequence of strain PCC 9511 DnaA (463 amino acids) is 85% identical to that of SS120 (461 amino acids) and 49% identical to that of Synechocystis sp. strain PCC 6803. The ftsZ gene of Prochlorococcus sp. strain PCC 9511 is preceded by two genes highly similar to genes present in the cell wall and division gene cluster of E. coli. These genes encode a D-alanineD-alanine ligase (8) and FtsQ, an intermediate recruit to the division site (7). PCC 9511 lacks a homologue of ftsA, which is located in E. coli between ftsZ and ftsQ. Scanning of the total genome sequence of Prochlorococcus sp. strain MED4 confirmed the absence of an ftsA homologue in the genome of that strain. The gene downstream of ftsZ shows similarity to panB and is followed by a putative homologue of hemN. Both genes overlap by 8 bp at their 3' ends. In E. coli the product of hemN catalyzes the oxidative decarboxylation of coproporphyrinogen III to form protoporphyrinogen IX (40). Database searches show that FtsZ of Prochlorococcus sp. strain PCC 9511 has the highest identity to FtsZ from the marine Synechococcus strain WH8103 and a significantly lower similarity to three other cyanobacterial FtsZ proteins (Table 3). Both nucleotide sequences determined in this study (4,372 nucleotides [nt] for dnaA and 4,790 nt for ftsZ) are 100% identical to the respective DNA segments in the total genome of Prochlorococcus sp. strain MED4, indicating that these two strains might be genetically the same organism.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Genomic region containing ftsZ and dnaA in Prochlorococcus sp. strain PCC 9511. (A) The dnaA gene is framed by four putative genes, ORF1 to ORF4, close homologues of which reside at corresponding sites in P. marinus sp. strain SS120 (30). (B) Organization of the ftsZ locus. The numbers designate putative gene start and stop codons (GenBank accession no. AJ011025 for ftsZ and AF158628 for dnaA). Genetic symbols are as follows: ddlB, D-alanineD-alanine ligase; ftsQ, filamentous temperature-sensitive Q; ftsZ, filamentous temperature-sensitive Z; panB, 3-methyl-2-oxobutanoate hydroxymethyltransferase; and hemN, oxygen-independent coproporphyrinogen III oxidase. The locations of antisense RNA probes used in the expression analysis are indicated by arrows.

Synchronization by modulated L/D cycles. The turbidostat culture of PCC 9511 was maintained in exponential growth at an average density of 9.76 × 10⁷ ± 3.1 × 10⁷ cells ml¹ during the 4 days of sampling (data not shown). Flow cytometric analyses indicated that the cell cycle was highly synchronized and that the daily alternation of cell cycle phases was very similar every day throughout the experiment (Fig. 2). From the beginning to the middle of the light period, almost all cells of the population were in the G₁ phase. At the end of the day about 70% of the cell population had entered the S phase. Two hours after virtual sunset, this population proceeded through G₂, and in the middle of the night (6 h after the end of the light period), most cells had divided and were back in G₁.

View larger version (50K): [in this window] [in a new window]   FIG. 2.   Synchronization of Prochlorococcus sp. strain PCC 9511 by modulated L/D cycles. (A) Distribution of the cell population over the different cell cycle phases at each sampling point during four consecutive L/D cycles; (B) Flow-cytometric DNA fluorescence distributions for five representative time points (boxed). a.u., arbitrary units.

Transcript accumulation in synchronized Prochlorococcus sp. strain PCC 9511 cultures follows a diel rhythm. The steady-state level of ftsZ and dnaA mRNA showed considerable temporal variation and oscillated during the course of each cell cycle in a periodic way (Fig. 3). To detect minor differences in the total amount of RNA per lane, the RNA component of RNase P was used as an internal standard. The expression of ftsZ and dnaA peaked at the end of the light period (i.e., 10 h after light onset). This time corresponded to the S-phase maximum. In a parallel study using the same material, diel expression of several photosynthetic genes was shown but with maxima at time points very different from those for ftsZ and dnaA (12).

View larger version (50K): [in this window] [in a new window]   FIG. 3.   Transcript levels of dnaA (top) and ftsZ (bottom) in the turbidostat culture of Prochlorococcus sp. strain PCC 9511. Light (8 a.m. to 8 p.m.) and dark (8 p.m. to 8 a.m.) phases are shown by the black and white bars, respectively. Samples were taken at 4-h intervals. The RNA levels were determined by Northern hybridization (dnaA) or RNase protection assays (ftsZ). The size of the undigested ftsZ RNA probe was 621 nt, and that of the full-length ftsZ protected fragment was 571 nt. A DNA probe for rnpB was used to assess the amounts of RNA loaded per lane.

Immunodetection of FtsZ. Antibodies raised against the recombinant FtsZ of Anabaena sp. strain PCC 7120 detected a single protein band of about 50 kDa in the Prochlorococcus sp. strain PCC 9511 lysate (Fig. 4A). This size corresponds well to that obtained for Anabaena (19). To analyze FtsZ abundance at different cell cycle stages, total cell proteins from three consecutive days were immunoblotted using this serum. Although expression changed slightly from day to day, the overall pattern was similar during the three consecutive photocycles. FtsZ concentration reached a minimum during the light period (when the number of cells in G₁ is maximum) and increased during the S and the division phases (Fig. 4B). Two- and fourfold dilutions of the sample taken 26 h after the beginning of the experiment indicated that the amount of FtsZ varied by a factor of 2 to 4 during a 24-h L/D cycle (Fig. 4C). This variation did not result from loading variability, since immunostaining of the same membrane using an antiserum against the photosynthesis protein PsbO did not reveal a comparable drop during the light period (data not shown).

View larger version (37K): [in this window] [in a new window]   FIG. 4.   Detection of FtsZ by immunoblotting. (A) FtsZ level in Prochlorococcus sp. strain PCC 9511 during three L/D cycles. Light and dark phases are displayed by black and white bars, respectively. (B) Graphic representation of FtsZ expression. (C) Semiquantitative assessment of the relative amount of FtsZ. Samples of the second L/D cycle were blotted together with two- and fourfold dilutions of the first sample of that day (taken at 26 h [stars]).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The genome region surrounding ftsZ in Prochlorococcus sp. strain PCC 9511 is partially conserved compared to that in E. coli, where the genes are clustered in the order ddlB-ftsQ-ftsA-ftsZ. The apparent lack of ftsA, which is essential in E. coli (33), might be interpreted as an example of how genome minimization has been achieved during the evolution of the small genome of Prochlorococcus (38). The gene arrangement around dnaA is even less conserved than in E. coli, but it is similar to that previously found in Prochlorococcus SS120 (30). DnaA of the latter strain, expressed in vitro, recognized the oriC of E. coli and B. subtilis, suggesting a similar molecular basis for the initiation of replication in these eubacteria (30). The synteny of this genome region between the two Prochlorococcus strains is seemingly trivial. However, other markers are very different between these strains, e.g., multiple pcb genes (11) and a phycoerythrin gene cluster are present in strain SS120 (14), but not in strains MED4 and PCC 9511 (28, 31). The degree of 16S rRNA identity between the strains lies in the same range (98%) as that between members of different genera of enterobacteria.

We show here that ftsZ and dnaA mRNA levels covary and are maximally expressed during the S phase. A simultaneous expression of genes like dnaA and ftsZ might well constitute the molecular basis for coordinated timing between DNA synthesis and cell division. For E. coli, cell cycle-related variations in the amount of ftsZ mRNA have previously been demonstrated (13, 16, 32). However, the method used to achieve synchronization in our study (simulation of a natural light regimen) is completely different from those used for heterotrophic bacteria. The oscillation of ftsZ mRNA abundance is partially matched by changes at the protein level. Synthesis of FtsZ starts during the S phase, and the concentration reached a maximum at night, i.e., at a time at which the mRNA level has clearly dropped again. Although we only roughly determined the timing of FtsZ expression, the maximum FtsZ level seems to correlate well with the onset of cell division. For E. coli, a titration mechanism that triggers cell division once a certain amount of FtsZ per cell is reached has been postulated (26). As far as we know, the amount of FtsZ actually required has never been assessed. As shown here, such changes in FtsZ concentration might be rather small, given that in Prochlorococcus the decrease in FtsZ amount per unit of total cellular protein during the day was on the order of about 50 to 75% only.

The factors that coordinate the synchronous expression of genes such as the cell cycle genes dnaA and ftsZ will have to be further investigated. They could involve a circadian clock, as in the case of Synechococcus sp. strain PCC 7942 (18). Alternatively, they could be under the direct control of light through photoreceptors. Finally, these genes could be expressed when a specific cell constituent or cell property, such as size (4), reaches a critical threshold. The total genome sequences of three different Prochlorococcus strains to be available within the near future will become a powerful tool to elucidate these mechanisms in detail.