Temporal separation of cell division and diazotrophy in the marine diazotrophic cyanobacterium Trichodesmium erythraeum IMS101


  • Present address: Beatriz Díez, Departament de Biologia Marina i Oceanograa, Institut de Ciències del Mar (CMIMA, CSIC), E-08003 Barcelona, Spain.

  • Editor: Hermann Bothe

Correspondence: Gustaf Sandh, Department of Botany, Stockholm University, S-106 91 Stockholm, Sweden. Tel.: +46 8 163 918; fax: +46 8 165 525; e-mail: sandh@botan.su.se


Examination of the diurnal patterns of basic cellular processes in the marine nonheterocystous diazotrophic cyanobacterium Trichodesmium revealed that the division of cells occurred throughout the diurnal cycle, but that it oscillated and peaked at an early stage in the dark period. Transcription of the early cell division gene ftsZ and the occurrence of the FtsZ protein showed a similar diurnal rhythmicity that preceded the division of cells. DNA replication (dnaA gene transcription) occurred before the transcription of ftsZ and hetR, the latter encoding the key heterocyst differentiation protein. Transcription of ftsZ and hetR in turn preceded the development of the nitrogen-fixing diazocytes and nifH transcription, and were at the minimum when diazotrophy was at the maximum. The nifH gene transcription showed a negative correlation to the circadian clock gene kaiC. Together, the data show a temporal separation between cell division and diazotrophy on a diurnal basis.


The marine nonheterocystous diazotrophic cyanobacterium Trichodesmium is a globally important organism, as it sequesters large quantities of carbon dioxide (via photosynthesis) and atmospheric nitrogen gas (via nitrogen fixation) into the biogeochemical cycles of warm`er oligotrophic oceans (Karl et al., 2002). To accomplish this, Trichodesmium shows a unique physiological and developmental adaptation to accommodate the nitrogen fixation process while at the same time performing oxygenic photosynthesis. The genus Trichodesmium fixes nitrogen in the light period (Saino & Hattori, 1978), although this is the supposed exclusive norm for heterocystous cyanobacteria. Both natural and cultured Trichodesmium spp. (including the genus Katagnymene; Lundgren et al., 2005) synthesize the nitrogen-fixing enzyme nitrogenase in a specific cell type, the diazocyte, a strategy to protect the oxygen-sensitive nitrogenase enzyme complex (see Fredriksson & Bergman, 1997; Berman-Frank et al., 2001). The frequency of diazocyte cells varies on a diurnal basis (Fredriksson & Bergman, 1995; Lin et al., 1998), in a pattern that positively correlates with the diazotrophic activity in Trichodesmium.

Cell division generally precedes cell differentiation and a connection between the processes was recently proposed for a heterocystous cyanobacterium (Sakr et al., 2006a). As for Trichodesmium, many cyanobacteria, under natural conditions, time their physiological activities to distinct periods during light/dark cycles. The nondiazotrophic unicellular genera Prochlorococcus and Synechococcus often show a highly synchronized cell division that takes place either late in the light period or at the beginning of the dark period (Jacquet et al., 2001; Asato, 2003). A contrasting pattern is seen in the marine unicellular diazotrophic cyanobacterium Cyanothece ATCC 51142, where both the cell division and the transcription of the cell division gene ftsZ peak early in the light period (Stöckel et al., 2008; Toepel et al., 2008). To our knowledge, the only filamentous cyanobacterium studied so far is the heterocystous Anabaena flos-aquae, which showed a nonsynchronized cell division that was confined to the light period (Lee & Rhee, 1999).

Diurnal patterns in cell division and related molecular regulatory mechanisms are unknown in filamentous, nonheterocystous cyanobacteria. This is also the case for any coupling between cell division and development of diazocytes in the cyanobacterium Trichodesmium. As cell division and diazocyte development in Trichodesmium may relate to its profound global ecological success in the oligotrophic oceans (Karl et al., 2002) and to its unique behavior as a diazotroph, these processes were examined at the structural and molecular level over the diurnal light/dark cycle.

Materials and methods

Material and growth conditions

Cultures of Trichodesmium erythraeum IMS101 were grown in YBCII media (Chen et al., 1996) in 2-L polycarbonate culture flasks (Nalgene) with constant aeration. The cultures were maintained at 25–27 °C with a 12-h/12-h light/dark cycle, with a constant irradiance of 70–80 μmol photons m−2 s−1. Growth was monitored by measuring chlorophyll a concentrations (Meeks & Castenholz, 1971). A minimum of three biological replicas were used in each subsequent analysis.

Diurnal sampling

Trichodesmium samples were initially collected every 2 h (light period) and 4 h (dark period) over the 24-h interval, while every 2 h in later experiments (light and dark). The samples were placed in sterile 50-mL Falcon tubes. From these, 15 and 30 mL were used for RNA and protein extraction, respectively, and 1 mL for light microscopy (LM) analyses. The cell samples were filtered onto 5 μm (for RNA and protein extractions) or 8 μm (for LM analyses) Nuclepore polycarbonate filters (Whatman). The RNA samples were quickly immersed in RLT buffer (RNeasy kit, Qiagen) and frozen in liquid nitrogen and kept at −80 °C until further processing. The protein samples were filtered and frozen in liquid nitrogen and stored at −80 °C.


To increase the contrast when identifying and quantifying cell division in Trichodesmium filaments, samples were collected at each sample point (as above) and directly fixed/stained in a 5% (v/v) acidic iodine–potassium iodine solution according to Lugol (Carl Roth). Cell division patterns were examined using bright-field LM (Olympus BX60 microscope, Olympus). Quantifications of the frequency of dividing cells were based on 500–1000 cells per sample. The percentage of diazocytes in the cell population was also counted at each time point (≥1000 cells per sample).

Transmission electron microscopy

Filaments were harvested from an actively growing aerated Trichodesmium culture 4 h into the dark period (during a 12-h/12-h light/dark cycle) and were immediately fixed for 2 h in 2.5% (w/v) glutaraldehyde dissolved in YBCII medium. Subsequent dehydration and embedding steps were performed according to Lundgren et al. (2001). Ultrathin sections were obtained using an ultramicrotome (2088 Ultrotome V, LKB Bromma, Sweden) and the sections were placed on Cu grids and poststained with lead citrate and uranyl acetate. The sections were viewed in a Zeiss EM 906 transmission electron microscope at 80 kV.

Fluorescence in situ immunolocalization

Filaments from Trichodesmium were collected by filtration onto an 8-μm Nuclepore membrane filter (Whatman) at mid-day and the filters were submerged in ice-cold absolute ethanol. The cells were stored at −20 °C until further processing. Samples were next transferred onto glass slides, air dried and incubated in 200-μL drops of 0.5% dimethyl sulfoxide for 15 min. Following 3 × 2 min washes in phosphate-buffered saline (PBS), the samples were incubated for 1 h with a polyclonal rabbit anti-NifH Rhodospirillum rubrum antibody, diluted 1 : 100 in PBS. The samples were then washed 3 × 2 min in PBS and the secondary goat anti-rabbit antibody conjugated to a fluorescent marker (Alexa 350, Molecular Probes) diluted 1 : 200 was added. The incubation lasted for 45 min at room temperature. After 3 × 2 min washes in PBS, an antifade reagent (Prolong Gold, Molecular Probes) was added. The blue fluorescent-labeled cells were identified using a Zeiss Axiovert M-200 microscope (Zeiss) equipped with filter set 49 (Zeiss, G 365, FT 395, BP 445/50).

Immunogold localization

Trichodesmium filaments were harvested at mid-day and were fixed in 3% (w/v) freshly made paraformaldehyde for 2 h. The cells were then dehydrated and embedded in LR-White (TAAB Laboratories Equipment), and ultrathin sections were placed on Ni grids according to the protocol described in Lundgren et al. (2001). The grids with sections were blocked in PBS containing 10% (w/v) bovine serum albumin for 1 h, followed by incubation in the primary antibody, a polyclonal rabbit anti-NifH R. rubrum antibody, diluted 1 : 100 in PBS, at 4 °C overnight. The samples were then washed 3 × 10 min in PBS. The secondary antibody, a polyclonal goat anti-rabbit antibody conjugated to 5 nm colloidal gold (GE Healthcare), was diluted 1 : 20 in PBS. Drops of this solution were placed to cover the sections on the grids and the incubation lasted for 1 h at room temperature. After subsequent washing of the sections with PBS (2 × 10 min) and dH2O (2 × 10 min), the samples were air dried and viewed as above.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting

Cell-free protein extracts of Trichodesmium were prepared by adding 200 μL sample buffer [2% (w/v) SDS, 2.5% (v/v) 2-mercaptoethanol, 10% (v/v) glycerol, 25 mM Tris-HCl, pH 6.8, and 0.1 tablet mL−1 protease inhibitor cocktail; Complete Mini, Roche] to the filtered filaments, followed by boiling for 5 min. The lysate was cleared from cell debris by centrifugation for 10 min at 15 800 g at 4 °C, and the supernatant was collected. Equal loading was achieved by determining the protein concentrations using a protein assay kit (RC DC, Bio-Rad) before loading. SDS-PAGE was performed using 10% (v/v) (29 : 1) polyacrylamide gels. The proteins were transferred from the gels and immunolabeled according to Klint et al. (2007) using a polyclonal rabbit anti-FtsZ Anabaena PCC 7120 diluted 1 : 500 in PBS-T and a secondary goat anti-rabbit antibody conjugated to horse radish peroxidase (Dako), diluted 1 : 5000 in PBS-T. Detection of the secondary antibody was achieved using a chemiluminescent reagent (ECL Plus, GE Healthcare) according to the manufacturer's instruction and the staining was visualized using a ChemiDoc XRS system (Bio-Rad).

RNA extraction

RNA was isolated from the Trichodesmium cells using a commercial extraction kit (RNeasy Mini, Qiagen). After purification, the RNA samples were treated with an additional step of DNAse (0.17 U μL−1) in 1 × DNAse buffer (200 mM Tris-HCl, pH 7.5, 10 mM EDTA and 75 mM MgCl2) at 37 °C for 30 min, followed by 68 °C for 10 min. The samples were then ethanol precipitated and spun down for 10 min at 15 800 g before being resuspended in 30 μL RNAse-free dH2O, and stored at −80 °C. The concentration of RNA was measured spectrophotometrically, and real-time reverse transcriptase (RT)-PCR analyses were performed with RNA as a template to test for and avoid possible DNA contamination.

DNA extraction

Genomic DNA was isolated from 200 mL of actively growing Trichodesmium culture. After filtration onto a 5-μm Nuclepore polycarbonate filter (Whatman), DNA was extracted using a commercial kit (DNeasy Tissue Kit, Qiagen).

Real-time RT-PCR

Real-time RT-PCR was carried out in a two-step reaction. A cDNA pool was generated from each RNA sample using the iScriptcDNA Synthesis Kit (Bio-Rad) according to the manufacturer's protocol. The real-time-PCR reaction was then carried out on an iCycler iQ (Bio-Rad) using iQSYBR® Green Supermix (Bio-Rad). The gene-specific PCR primers designed for the second step are presented in Table 1; primers for hetR and nifH were taken from El-Shehawy et al. (2003). The reaction mixture was prepared according to the manufacturer's protocol and the reaction was performed as follows: 95 °C for 5 min, 40 cycles of 94 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s. Fluorescence levels were monitored at 72 °C and the relative quantification of cDNA was performed using serial dilutions of a known concentration of Trichodesmium genomic DNA and normalized to the abundance of 16S rRNA gene. Melt curve analysis and sequencing of the products were used to confirm the specificity of the PCR reactions and the identity of the PCR products (data not shown).

Table 1.   Gene-specific primers used for real-time RT-PCR, presented in the 5′→3′ mode
GenesForward primersReverse primers


Filaments of T. erythraeum IMS101 (hereafter referred to as Trichodesmium), grown under controlled (12-h/12-h light/dark cycles) nitrogen-fixing (combined nitrogen-free medium) conditions, showed a doubling time of c. 2.4 days. Dividing cells were identified by the following criteria: (1) longer cells (5–10 μm); (2) development of division septa (Fig. 1, arrows); and (3) Lugol staining, being particularly pronounced at the cell walls connecting two vegetative cells (Fig. 1a). Division never took place synchronously in the whole filament, but was restricted to small groups of cells that spread along the filaments.

Figure 1.

 Characteristics of cell division in Trichodesmium erythraeum IMS101. (a) Light micrograph of a filament stained with 5% Lugol's solution. Note the prominent, lighter stained, zone of diazocytes in the center of the filament. Scale bar=10 μm. (b) Group of dividing cells (arrows) in a filament stained with Lugol's solution. Scale bar=5 μm. (c) Transmission electron microscopy micrograph of a longitudinally sectioned filament depicting a dividing cell with developing septa (lower part). Also seen are electron-dense cyanophycin granules (cg), medium electron-dense polyhedral carboxysomes (cb), thylakoid membranes (th) and electron-transparent gas vacuoles (gv). Scale bar=1 μm.

Cell division in Trichodesmium occurred throughout the diurnal cycle, although at varying relative quantitative levels (Fig. 2a). The proportion of dividing cells, out of the total number of cells, varied from 5% to 20%. A period of high cell division frequency was apparent a few hours into the dark period. This was always preceded by a considerable decline in cell division during the later half of the light period (Fig. 2a).

Figure 2.

 Diurnal variations in cell division and diazocyte abundance in filaments of Trichodesmium erythraeum IMS101. (a) Frequency of dividing cells within filaments. (b) Diurnal variations in diazocyte frequency given as a proportion of the total cell number. The cells were fixed/stained with 5% Lugol's solution and examined using LM.

As found earlier, groups of lighter cells (more transparent) were apparent in the Trichodesmium filaments when subject to Lugol staining (Fig. 1a; Bryceson & Fay, 1981; El-Shehawy et al., 2003). An almost identical distribution of nitrogenase-containing cells was also detected using in situ immunolocalization (Fig. 3a), and similar to the situation in natural populations (Fredriksson & Bergman, 1997), the nitrogenase-containing cells showed distinct subcellular rearrangements, including reduced amounts of gas vacuoles (Fig. 3b and c). The lighter cells after Lugol staining match both the pattern and the subcellular appearance of nitrogenase-containing cells along filaments of Trichodesmium, confirming a weaker labeling of diazocytes by Lugol staining (Figs 1a and 3a–c).

Figure 3.

 Immunolocalization of nitrogenase (NifH) in filaments of Trichodesmium erythraeum IMS101. (a) Fluorescence in situ immunolocalization of NifH in an intact filament of Trichodesmium. Scale bar=10 μm. (b) Longitudinally sectioned Trichodesmium filaments illustrating the subcellular ultrastructure of 5-nm immunogold-labeled diazocytes (left) and vegetative cells lacking immunogold label (right). Scale bar=2 μm. (c) A diazocyte zone border with an immunogold-labeled diazocyte (right) and vegetative cells lacking an immunogold label but resembling the typical diazocyte subcellular ultrastructure. Scale bar=1 μm.

The lighter-stained diazocytes constituted on average 22% of the total cell number (Fig. 2b), and the frequency ranged from 15% to 27% on a diurnal basis (Fig. 2b), being the highest in the light and the lowest in the dark period. The transcription of nifH, encoding one of the proteins in the nitrogenase complex, was the lowest when the frequency of diazocytes was at minimum (Fig. 5a).

Figure 5.

 Diurnal transcription profiles in Trichodesmium erythraeum IMS101 of genes involved in nitrogen fixation, DNA replication, cell differentiation and the circadian clock. The relative transcription levels of (a) nifH, (b) dnaA, (c) hetR and (d) kaiC are normalized to 16S rRNA transcription levels. Representative graphs are shown, coinciding with data from at least three additional experiments.

Diurnal transcription levels of ftsZ (Fig. 4a) and the occurrence of the corresponding cell division protein, FtsZ (Fig. 4b), were next examined in Trichodesmium using real-time RT-PCR and immunoblotting, respectively. The ftsZ gene was transcribed throughout the diurnal light/dark cycle and showed a small increase late in the light period and a pronounced peak during the first half of the dark period (Fig. 4a). A polyclonal anti-Anabaena PCC 7120 FtsZ antibody (Klint et al., 2007) revealed fourfold fluctuations in the FtsZ protein levels over the diurnal cycle, with high FtsZ levels being apparent towards the middle of the dark period (Fig. 4b). The synthesis of the FtsZ protein followed the ftsZ transcription pattern with an c. 2-h delay and correlated positively with the cell division patterns and the appearance of diazocytes (Fig. 2b).

Figure 4.

 Diurnal expression profiles of the early cell division gene ftsZ in Trichodesmium erythraeum IMS101. (a) The relative transcription levels of ftsZ over a diurnal cycle normalized to 16S rRNA gene transcription levels. A representative graph is shown, coinciding with data from at least three additional experiments. (b) Immunoblot analysis of the diurnal occurrence of FtsZ, using a polyclonal anti-Anabaena 7120 FtsZ antibody. One band with an Mr of c. 50 000 is found throughout the diurnal cycle. The relative abundance of the FtsZ protein is given as bars in the figure below (pixel intensities; imagej program), with the average FtsZ protein level over the 24-h period set to 1.

Diurnal transcription patterns of the DNA replication gene dnaA were also examined in the same cultures of Trichodesmium. As seen in (Fig. 5b) dnaA transcription correlated positively with ftsZ transcription, and was at the maximum c. 2 h before ftsZ in replicate experiments. The transcription of hetR (Fig. 5c), encoding the master gene for heterocyst differentiation and potentially involved in diazocyte development as suggested earlier (El-Shehawy et al., 2003), peaked together with the transcription of ftsZ in Trichodesmium. The larger peak also coincided with the development of diazocytes (Fig. 2b).

As the basic components of a cyanobacterial circadian clock (Dong & Golden, 2008) are present in the genome of Trichodesmium (http://www.jgi.doe.gov), the transcription of kaiC, encoding the master clock protein KaiC, was followed. The transcript levels of the kaiC gene were highest in the middle of the dark period, followed by an approximately sixfold decline in expression towards the light period (Fig. 5d). The expression of the kaiC gene showed a pattern reciprocal to the transcription of nifH (Fig. 5a).


As shown previously, both photosynthesis and nitrogen fixation are confined to the light/day period in Trichodesmium, while here we show that cell division, including ftsZ transcription, FtsZ synthesis and septum formation, is primarily confined to the dark period. It is also clear that patches of cells along the filaments of Trichodesmium divide nonsynchronously, although a small proportion divides throughout the diurnal period. However, most importantly, a distinct time period of enhanced cell divisions was discovered in the dark, and cell division consistently appeared just after DNA replication. Likewise, cell division has also been shown to be coupled to DNA replication in the unicellular cyanobacterium Microcystis (Yoshida et al., 2005). Moreover, the nonsynchronized cell division pattern found in the Trichodesmium filaments, with small groups of cells dividing at a given time point, may be due to different positions in the cell division cycle of the individual cells, perhaps in combination with nutrient gradients caused by the semi-regular distribution of groups of nitrogen-fixing diazocytes along the filaments.

Compared with cell division, a reciprocal pattern was seen in diazocyte abundance and nifH transcription, the latter coinciding with earlier data (Chen et al., 1998; Lin et al., 1998; El-Shehawy et al., 2003). Hence, key physiological processes, such as photosynthesis and nitrogen fixation, take place in the light (our data, Berman-Frank et al., 2001), probably as a consequence of their high energy requirements, while cell division in Trichodesmium is confined to the dark hours. A similar but inverted strategy with respect to diazotrophy and cell division is apparent in the unicellular diazotrophic cyanobacteria Cyanothece ATCC 51142 (Stöckel et al., 2008; Toepel et al., 2008) and Gloeothece PCC 6909 She (Peschek et al., 1991). In these unicellular diazotrophs, cell division is confined to the light and nitrogen fixation to the dark period.

Cell division is known to be required for the initiation of heterocysts differentiation in Anabaena PCC 7120 (Sakr et al., 2006a, b). The timing of cell division in Trichodesmium shows a tight sequential coupling to the emergence of diazocytes and ftsZ coexpresses with the heterocyst differentiation regulator hetR, previously implicated in diazocyte differentiation (El-Shehawy et al., 2003). It therefore appears that cell division may be required for diazocyte development in Trichodesmium. As opposed to heterocysts, diazocytes retain their division capacity (Janson et al., 1994; Fredriksson & Bergman, 1997). Therefore, changes in diazocyte abundance (Fig. 2b; Fredriksson & Bergman, 1995; Lin et al., 1998) may originate from either dividing diazocytes and/or from dividing and differentiating vegetative cells.

Many essential metabolic processes are known to be regulated by the circadian clock, composed in cyanobacteria of the core proteins KaiA, KaiB and KaiC (Johnson et al., 2008). The presence of a clustered kaiABC operon and several of the input and output regulators of the circadian clock (http://www.jgi.doe.gov), together with the oscillating pattern of kaiC presented here, supports the operation of a circadian clock in Trichodesmium and strengthens data obtained previously (Chen et al., 1998). Although kaiC was expressed at low levels, increased levels of kaiC coincided in time with a higher expression of ftsZ, dnaA and hetR, while there seems to be a negative correlation between kaiC and nifH expression. Further studies will, however, be needed before linking the circadian clock to the regulation of basic cellular processes, for example cell division, DNA replication and cell differentiation, in Trichodesmium.

In summary, our data show for the first time that cell division in the globally successful cyanobacterium Trichodesmium takes place in the dark period on a diurnal basis, and that there exists a clear diurnal separation between cell division and cell development, on the one hand, and nitrogen fixation and photosynthesis, on the other. This separation may represent an important means to optimize the energy-demanding daytime diazotrophic potential of this marine primary producer.


Financial support from the Swedish Research Council (to B.B.) and from K. Wallenberg and A. Wallenberg's Foundation is gratefully acknowledged, as are research stipends from Hierta-Retzius Foundation and Stockholm Marine Research Center (to G.S.). S. Lindwall is acknowledged for excellent experimental assistance.