Short RNA half-lives in the slow-growing marine cyanobacterium Prochlorococcus

Abstract

Background: RNA turnover plays an important role in the gene regulation of microorganisms and influences their speed of acclimation to environmental changes. We investigated whole-genome RNA stability of Prochlorococcus, a relatively slow-growing marine cyanobacterium doubling approximately once a day, which is extremely abundant in the oceans.

Results: Using a combination of microarrays, quantitative RT-PCR and a new fitting method for determining RNA decay rates, we found a median half-life of 2.4 minutes and a median decay rate of 2.6 minutes for expressed genes - twofold faster than that reported for any organism. The shortest transcript half-life (33 seconds) was for a gene of unknown function, while some of the longest (approximately 18 minutes) were for genes with high transcript levels. Genes organized in operons displayed intriguing mRNA decay patterns, such as increased stability, and delayed onset of decay with greater distance from the transcriptional start site. The same phenomenon was observed on a single probe resolution for genes greater than 2 kb.

Conclusions: We hypothesize that the fast turnover relative to the slow generation time in Prochlorococcus may enable a swift response to environmental changes through rapid recycling of nucleotides, which could be advantageous in nutrient poor oceans. Our growing understanding of RNA half-lives will help us interpret the growing bank of metatranscriptomic studies of wild populations of Prochlorococcus. The surprisingly complex decay patterns of large transcripts reported here, and the method developed to describe them, will open new avenues for the investigation and understanding of RNA decay for all organisms.

Figure 1

Distribution of RNA decay rates and RNA half lives using the two phase decay step or the twofold decay step method. (a) RNA decay rates. (b) RNA half-lives. Time rates were binned in 1-minute increments. RNAs with stabilities of more than 60 minutes are not shown. The insets show the results for transcripts with decay rates of ≤10 minutes.

Figure 2

Comparison of global half-lives and cell doubling time of selected organisms. For all organisms the global median half-life is presented except for Plasmodium falciperum, for which only mean half-lives were available. Values were obtained from the following sources: Halobacterium salinarum [12], Sulfolobus solfactaricus and Sulfolobs acidocaldarius [11], E. coli [34], P. falciperum [54,55], Saccharomyces cerevisiae [56,57], Arabidopsis thaliana [9,58], Bacillus subtilis [6,59], and Prochlorococcus marinus (this study).

Figure 3

Expression profiles of 12 clusters determined by Mfuzz. In red are genes that are well supported within the cluster (that is, high fuzziness score) and in grey genes with weak support. Cluster 6 contains genes with the shortest half-lives and decay rates and cluster 11 highly expressed genes with long half-lives. Clusters 2 and 4 are highly enriched in genes coding for tRNAs, rRNAs and ncRNAs.

Figure 4

RNA decay profiles of all ribosomal protein transcripts. Genes that are transcribed as monocistrons or represent the first gene of the operon are shown as dark blue lines (single genes/1. gene in operon). All other genes are organized in operons and are localized up to 1.2 kb (light blue lines), between 1.3 and 2.4 kb (green lines), between 2.5 and 4.5 kb (orange lines), and ≥4.6 kb (red lines) downstream of the start codon of the first gene of the respective operon. The microarray signal intensity (expression) was normalized to time 0 h. Numbers in parentheses indicate the position within the operon. Genes without numbers in parentheses are monocistronic.

Figure 5

RNA decay profiles of single probes of glsF (ferredoxin-dependent glutamate synthase) - the longest gene (4.6 kb) in Prochlorococcus MED4. Single microarray probes are localized up to 1.2 kb (light blue lines), between 1.3 and 2.4 kb (green lines) and between 2.5 and 4.5 kb (orange lines) downstream of the start codon. The microarray signal intensity (expression) was normalized to time 0 h. Only probes with an expression value above 100 at time 0 h are shown.

Figure 6

RNA decay profiles of type I and type II operons. Both type I (left panel) and type II (right panel) operons have delayed decay profiles that are more pronounced with distance from the promoter. Type I operons are characterized by a plateau in transcript levels prior to decay whereas transcript levels in type II operons increase with time prior to decay and this increase is greater with distance from the promoter. The order of genes within each operon is indicated by numbers in parentheses. The microarray signal intensity (expression) was normalized to time 0 h.

Figure 7

A possible mechanism of transcriptional delay shown for the type II ATPase operon. A physical block (red ellipse), which might be built by proteins, congestion of polymerases or convergent polymerases, decelerates the polymerase velocity (0 minutes). After a certain time the block is disintegrated and stalled polymerases can continue with elongation of mRNA (10 minutes and 20 minutes), leading to a relative increase of mRNAs as a function of time and distance. TSS is the transcriptional start site of the operon. The insert on top shows gene expression over time of all genes of the ATPase operon starting with atp1 (dark blue line) and ending with PMM1447 (conserved hypothetical in light blue). For better visualization the operon was plotted in three separate graphs. The microarray signal intensity (expression) was normalized to time 0 h.