Transcriptome changes and cAMP oscillations in an archaeal cell cycle

Abstract

Background: The cell cycle of all organisms includes mass increase by a factor of two, replication of the genetic material, segregation of the genome to different parts of the cell, and cell division into two daughter cells. It is tightly regulated and typically includes cell cycle-specific oscillations of the levels of transcripts, proteins, protein modifications, and signaling molecules. Until now cell cycle-specific transcriptome changes have been described for four eukaryotic species ranging from yeast to human, but only for two prokaryotic species. Similarly, oscillations of small signaling molecules have been identified in very few eukaryotic species, but not in any prokaryote.

Results: A synchronization procedure for the archaeon Halobacterium salinarum was optimized, so that nearly 100% of all cells divide in a time interval that is 1/4th of the generation time of exponentially growing cells. The method was used to characterize cell cycle-dependent transcriptome changes using a genome-wide DNA microarray. The transcript levels of 87 genes were found to be cell cycle-regulated, corresponding to 3% of all genes. They could be clustered into seven groups with different transcript level profiles. Cluster-specific sequence motifs were detected around the start of the genes that are predicted to be involved in cell cycle-specific transcriptional regulation. Notably, many cell cycle genes that have oscillating transcript levels in eukaryotes are not regulated on the transcriptional level in H. salinarum. Synchronized cultures were also used to identify putative small signaling molecules. H. salinarum was found to contain a basal cAMP concentration of 200 microM, considerably higher than that of yeast. The cAMP concentration is shortly induced directly prior to and after cell division, and thus cAMP probably is an important signal for cell cycle progression.

Conclusion: The analysis of cell cycle-specific transcriptome changes of H. salinarum allowed to identify a strategy of transcript level regulation that is different from all previously characterized species. The transcript levels of only 3% of all genes are regulated, a fraction that is considerably lower than has been reported for four eukaryotic species (6%-28%) and for the bacterium C. crescentus (19%). It was shown that cAMP is present in significant concentrations in an archaeon, and the phylogenetic profile of the adenylate cyclase indicates that this signaling molecule is widely distributed in archaea. The occurrence of cell cycle-dependent oscillations of the cAMP concentration in an archaeon and in several eukaryotic species indicates that cAMP level changes might be a phylogenetically old signal for cell cycle progression.

Figure 1

Optimized synchronization of H. salinarum cultures. The synchronization was performed as described in Experimental Procedures using the DNA polymerase inhibitor aphidicolin. The average cell length and its standard deviation was calculated from the lengths of 50 cells that were determined microscopically with an ocular micrometer. The cell density was determined microscopically with a Neubauer counting chamber. Times of addition and removal of the inhibitor are indicated. The time of inhibitor removal was set to zero to allow a direct comparison of the times shown in this and additional Figures (2 – 6, see Additional file 1). The box in this and additional Figures denotes the only time interval in which dividing cells with visible constrictions could be observed. Microscopic images of dividing cells and the intracellular DNA localization have been published previously [22].

Figure 2

Average transcript level profiles of seven clusters of co-regulated genes. Most of the cell cycle-regulated transcripts were sorted into seven clusters of co-regulated genes (compare text and Table 1). The average transcription profiles of all seven clusters and the standard deviations are shown. Genes that share an identical profile of induction and repression do not necessarily share the same degree of induction/repression, therefore the transcript profiles of all genes were normalized to their highest value (= 100%) before calculation of averages and standard deviations. Gene identifiers, names, and functional classes are summarized in Table 1.

Figure 3

Comparison of results obtained by DNA microarray analysis and by Northern blot analysis. 13 genes were selected for verification of the microarray results with an independent method, i.e. Northern blot analysis. They represent all clusters of co-regulated genes as well as unregulated control genes. The transcript level profiles obtained for individual genes by microarray analysis are shown on the left side (average of three biological replicates). On the right side the results of Northern blot analysis are shown (one typical experiment). Gene identifier [23] and the gene No. in Table 1 are included.

Figure 4

Comparison of results obtained by DNA microarray analysis and by Northern blot analysis. 13 genes were selected for verification of the microarray results with an independent method, i.e. Northern blot analysis. They represent all clusters of co-regulated genes as well as unregulated control genes. The transcript level profiles obtained for individual genes by microarray analysis are shown on the left side (average of three biological replicates). On the right side the results of Northern blot analysis are shown (one typical experiment). Gene identifier [23] and the gene No. in Table 1 are included.

Figure 5

Comparison of results obtained by DNA microarray analysis and by Northern blot analysis. 13 genes were selected for verification of the microarray results with an independent method, i.e. Northern blot analysis. They represent all clusters of co-regulated genes as well as unregulated control genes. The transcript level profiles obtained for individual genes by microarray analysis are shown on the left side (average of three biological replicates). On the right side the results of Northern blot analysis are shown (one typical experiment). Gene identifier [23] and the gene No. in Table 1 are included.

Figure 6

The conserved motif around the start site of cluster five genes. The program MEME was used to identify a conserved motif in the 400 nt region around the translational start point of cluster five genes (compare Table 2). A. The region is shown schematically, and the numbering refers to the translational start point. Because the majority of haloarchaeal transcripts are leaderless, the transcriptional and the translational start points often nearly coincide. The basal promoter elements "transcription factor B recognition element" (BRE) and "TATA box" are indicated at positions they would have for leaderless transcripts. Gene identifiers [23] are shown to the left. The conserved motif and its direction are indicated by arrows which are drawn to scale. B. The sequence logo of the conserved motif of cluster five genes (generated with MEME).

Figure 7

Cell cycle-dependent changes of the cAMP level. The cAMP level was determined in synchronized cultures, and it was revealed that concentration changes might occur in the time period between 1.5 and 3.5 hours (data not shown). Therefore this period was characterized with a higher time resolution of 15 minutes. The box in A and B denotes the time period of cell division (compare Fig. 1). A. One out of three biological replicates is shown. The cAMP concentration was determined in duplicate measurements, and the cell density was determined with a Neubauer counting chamber. B. Average cAMP level changes in three biological replicates. The absolute cAMP values (fmol cAMP/10⁹ cells) had a somewhat large variance, therefore the highest value in every independent experiment was set to 100% and the relative values were used to calculate the average and the standard deviation.

Figure 8

Overview of characterized cell cycle-dependent processes in H. salinarum. The Figure summarizes schematically the results obtained in this study and previously published results [22].