The ω Subunit of RNA Polymerase Is Essential for Thermal Acclimation of the Cyanobacterium Synechocystis Sp. PCC 6803

The rpoZ gene encodes the small ω subunit of RNA polymerase. A ΔrpoZ strain of the cyanobacterium Synechocystis sp. PCC 6803 grew well in standard conditions (constant illumination at 40 µmol photons m−2 s−1; 32°C; ambient CO2) but was heat sensitive and died at 40°C. In the control strain, 71 genes were at least two-fold up-regulated and 91 genes down-regulated after a 24-h treatment at 40°C, while in ΔrpoZ 394 genes responded to heat. Only 62 of these heat-responsive genes were similarly regulated in both strains, and 80% of heat-responsive genes were unique for ΔrpoZ. The RNA polymerase core and the primary σ factor SigA were down-regulated in the control strain at 40°C but not in ΔrpoZ. In accordance with reduced RNA polymerase content, the total RNA content of mild-heat-stress-treated cells was lower in the control strain than in ΔrpoZ. Light-saturated photosynthetic activity decreased more in ΔrpoZ than in the control strain upon mild heat stress. The amounts of photosystem II and rubisco decreased at 40°C in both strains while PSI and the phycobilisome antenna protein allophycocyanin remained at the same level as in standard conditions. The phycobilisome rod proteins, phycocyanins, diminished during the heat treatment in ΔrpoZ but not in the control strain, and the nblA1 and nblA2 genes (encode NblA proteins required for phycobilisome degradation) were up-regulated only in ΔrpoZ. Our results show that the ω subunit of RNAP is essential in heat stress because it is required for heat acclimation of diverse cellular processes.


Introduction
DNA-dependent RNA polymerases (RNAPs) catalyze the transcription of genetic information from DNA to RNA. The core of the multi-subunit RNAP is conserved throughout all cellular life forms [1]. The RNAP core of the majority of eubacteria, contains a catalytic center consisting of b and b9 subunits [2], two identical a subunits that enhance transcription efficiency and participate in promoter recognition [3], and a small v subunit. In cyanobacteria, however, the RNAP core consists of six subunits because b9 has been split into two parts, an N-terminal c subunit and a C-terminal b9 subunit [4]. For promoter recognition and transcription initiation, the bacterial RNAP core recruits a s factor. Bacteria encode one essential primary s factor and varying number non-essential s factors [5]. Different s factors favor different promoters thus orchestrating the transcriptional efficiencies of different genes.
The v subunit of the RNAP core is encoded by the rpoZ gene. Knock out strains of the v subunit have been constructed in the proteobacterium Escherichia coli [6], the actinobacteria Mycobacterium smegmatis [7], Streptomyces coelicolor [8] and Streptomyces kasugaensis [9], and in the cyanobacterium Synechocystis sp. PCC 6803 [10], indicating that rpoZ is not an essential gene. Studies in E. coli have revealed that the v subunit acts as a molecular chaperone for the b9 subunit [11], suggesting that the v subunit has a similar role as the essential eukaryotic RPB6 subunit of RNAP [12]. We have recently shown that in the DrpoZ strain of Synechocystis, recruitment of the primary s factor, SigA, by the RNAP core occurs less frequently than in the control strain, and as a consequence, many highly expressed genes are down-regulated in DrpoZ [10].
The optimum temperature for Synechocystis is 30-32uC but cells grow for a few days even at 43uC [13][14][15]. Pretreatment of Synechocystis cells in mild heat stress leads to acquired thermotolerance allowing survival in otherwise lethal temperatures up to 50uC [16][17][18]. Photosynthesis is a heat-sensitive process [19], and photosystem II (PSII) is the most vulnerable component, for which it takes hours to fully acclimate to an elevated temperature [20]. Transcriptomics and proteomics studies have revealed that heat treatment induces expression of many heat shock genes and numerous genes with unknown functions [20,21].
Previous studies have shown that group 2 s factors play roles in acclimation to elevated temperatures. The group 2 s factor gene sigB is rapidly up-regulated upon a heat shock [22,23] and the SigB factor, in turn, up-regulates especially the expression of the small heat shock protein HspA [14] and some other heat shock proteins [24]. Although SigC does not regulate heat shock genes, it is essential for heat acclimation processes as it is important for sustained functional photosynthesis in elevated temperatures [15,25]. Upstream of the s factors in the signaling cascades are histidine kinases (Hiks). For heat stress, Hik34 has been recognized as an important regulator, negatively controlling the expression of some heat shock genes like the htpG gene and the groESL1 operon [26]. Furthermore the CIRCE/HrcA system has been shown to regulate the expression of some heat shock genes including the groESL1 operon and the groEL2 gene [27].
In the present study, the v subunit of the RNAP core was found to be essential for the survival of cells even under mild heat stress. The results show that mild heat treatment at 40uC induces decrease of the RNAP content in the control strain but not in the DrpoZ strain. Furthermore, twice as many genes responded to heat treatment in DrpoZ than in the control strain (CS), and 80% of the heat-responsive genes were unique to DrpoZ. Mild heat stress induced reduction of light-saturated photosynthetic activity in both strains but this reduction was more prominent in DrpoZ than in CS. According to our results, many aspects of heat acclimation occurred differently in DrpoZ than in CS, and a combination of inappropriate responses in several cellular functions, rather than a deficiency in the expression of a single gene or operon, was the reason for the heat lethal phenotype of DrpoZ.

Results and Discussion
The DrpoZ strain has difficulties in acclimation to elevated temperature In our standard growth conditions, continuous light at the photosynthetic photon flux density (PPFD) of 40 mmol m 22 s 21 , and 32uC, the DrpoZ strain grows like CS [10]. At 40uC, CS grows essentially like it grows at 32uC (Fig. 1A), the doubling times during the first day being 11.660.2 h (Fig. 1) and 11.460.3 h [10] at 40uC and 32uC, respectively. The DrpoZ strain grew more slowly than CS during the first day at 40uC (Fig. 1A), with a doubling time of 18.562.0 h. A survival test indicated that the DrpoZ strain contained only 3.5610 2 60.4610 2 colony forming units (CFUs) after 24-h growth at 40uC while CS contained almost a hundred thousand times more CFUs, 3.4610 7 60.1610 7 . Transfer of cells back to the standard conditions did not rescue DrpoZ cells after two days of incubation at 40uC, but cells died. The initial growth of DrpoZ was slow at 38uC, with the doubling times for the first day of 12.160.3 h and 15.260.8 h for CS and DrpoZ, respectively ( Fig. 1B). At 38uC, however, the DrpoZ cells were able to acclimate, and similar doubling times, 25.460.5 h for CS and 25.261.6 h for DrpoZ, were measured after the second day (Fig. 1B). The DrpoZ+rpoZ strain, in which the rpoZ gene has been re-introduced to the genome under the strong psbA2 promoter [10], grew similarly as CS at 40uC (Fig. 1A). This indicates that the heat-sensitive phenotype of DrpoZ is due to the lack of the v subunit.
A DNA microarray analysis in standard conditions revealed that many genes involved in carbon concentrating mechanisms (CCM) and carbon fixation are down-regulated in the DrpoZ strain compared to CS [10]. Because temperature rise decreases the availability of inorganic carbon (the equilibrium concentration of dissolved CO 2 at 40uC is only 82% of that at 32uC), we tested if growth can be rescued by improving the availability of soluble inorganic carbon by increasing the pH of the growth medium to 8.3. Alkaline conditions have been previously shown to rescue many mutants with deficiencies in carbon metabolism. The growth of the heat-sensitive s factor mutant DsigC can be rescued by improving the availability of soluble inorganic carbon at 43uC by rising the pH of the growth medium from 7.5 to 8.3 [15,25]. Furthermore, Synechocystis strains DNdhB, lacking a functional NAD(P)H dehydrogenase complex, and DNdhD3/NdhD4, with an inactivated CO 2 uptake system, are able to grow at pH 8.3, but not at pH 7.5 [28], and even a mutant deficient of the main carboxysome operon can be grown in alkaline conditions [29]. In contrast to mutants with deficiencies in carbon concentrating mechanisms, the growth of DrpoZ cells at 40uC was not rescued at pH 8.3 (Fig. 1C), indicating that the heat-lethal phenotype of DrpoZ is probably not only caused by deficiencies in CCM. The growth rates of CS and DrpoZ were 10.660.2 h and 11.260.3 h, respectively, when cells were grown in BG-11 medium without added bicarbonate in standard conditions, confirming that DrpoZ cells are able to cope with low carbon conditions. Furthermore, the similarity of the growth rates in the presence and absence of added bicarbonate suggest that the bicarbonate addition to BG-11 has a negligible effect on the inorganic carbon content of the medium in growth experiments performed under ambient air.
Since DrpoZ survived only for a limited time at 40uC, all subsequent experiments were done by growing cells first in standard conditions to OD 730 ,1, and then transferring the cells to 40uC for 24 h. The 24-h heat treatment was selected because drastic difference between growth of mutant and CS was obvious after the first 24-h (Fig. 1A). Both strains grew during the 24-h treatment at 40uC (OD 730 increased from 1.0 to 1.5 in CS and to 1.4 in DrpoZ, respectively), suggesting that dense DrpoZ cultures might tolerate high temperature better than dilute cultures.

Comparison of gene expression of the control and DrpoZ strains at 40uC
To get a more comprehensive picture on why DrpoZ is not able to acclimate to mild heat stress, gene expression changes were studied by DNA microarray analysis. For DNA microarray analysis, CS and DrpoZ were grown in standard conditions and then treated at 40uC for 24 h before RNA was isolated. In addition, the results from standard growth conditions [10] were used as controls. All microarray data are available in GEO (accessions GSE59451). In the control strain, 71 genes were at least two-fold up-regulated upon heat treatment and 91 genes were down-regulated to one half or less ( Fig. 2A). Complete lists of upregulated (Table S1) and down-regulated (Table S2) genes in CS are included as supplemental material. In DrpoZ, the heat treatment induced up-regulation of 200 genes ( Fig. 2A, Table  S3) and down-regulation of 194 genes ( Fig. 2A, Table S4). Thus, 2.4 times more genes responded to mild heat treatment in the mutant strain than in CS ( Fig. 2A).
The differentially expressed genes were assigned to functional categories according to Cyanobase (Fig. 2B), and a heat map was constructed to further facilitate comparison between strains (Fig. 3). The heat map includes genes that were up or down regulated upon mild heat stress in DrpoZ, in CS or both, and in addition transcript levels of these heat-responsive genes were compared in DrpoZ and CS in standard growth conditions. All results included in Fig. 3 are collected in Table S5.
Only 33 genes were down-regulated upon mild heat-treatment in both strains ( Fig. 2A, Table S6). Nearly 40% of them encode hypothetical or unknown proteins (Fig. 2B, Table S6); genes with an assigned name are included in Fig. 2A. For the genes with known function, the decrease in the expression of the desB gene, encoding an acyl-lipid desaturase, is most probably an acclimation response compensating for temperature-induced increase in membrane fluidity. Up-regulation of the desB gene in low temperatures and adjustment of lipid saturation are well known responses to low and high temperature [30,31]. The heat shock genes have been shown to be rapidly but only transiently upregulated upon heat shock [21]. Up-regulation of heat shock genes typically occurs within minutes and transcripts disappear during the first hours of heat treatment. Accordingly, none of the heat shock genes was up-regulated after a 24-h treatment at 40uC. The hspA gene was up-regulated in DrpoZ in standard conditions [10] but this difference between the strains disappeared after the mild heat treatment. The groES heat shock gene was down-regulated in both strains and in addition the htpG heat shock gene was downregulated in CS (Table S2) and the dnaJ heat shock gene was down-regulated in DrpoZ (Table S4). In addition to heat shock proteins, some other proteins have been suggested to affect heat responses. The clpB1 gene encoding a protease, and slr1674 (a hypothetical protein) have shown to affect rapid heat responses, whereas hik34 (encoding a histidine kinase) and hypA1 (encoding a hydrogenase formation protein) affect sustained thermotolerance of PSII, and cpcC2 (encoding a phycobilisome rod linker polypeptide) is essential for both responses [32]. The slr1674, hypA1 and clpB1 genes were up-regulated in DrpoZ compared to CS at 40uC, whereas cpcC2 was 1.5 fold down-regulated.
The vast majority of genes up-regulated upon a mild heat treatment in the control or DrpoZ strains belonged to functional categories hypothetical or unknown (Figs. 2B and 3, Table S6). The other large group of up-regulated genes was transport and binding proteins comprising 20 and 21 genes in CS and DrpoZ, respectively (Figs. 2B and 3). Many of them, including ammonium/methylammonium permeases, ABC-type basic amino acid and glutamine transporter, a permease protein for urea transporter and a manganese transporter (Table S6), were up-regulated in both strains. However, some transporters were up-regulated in one strain only, like nitrate/nitrite transporter genes, which were among the most highly up-regulated genes in DrpoZ, but were not up-regulated in CS. Some other differences in central nitrogen metabolism genes were detected in addition. The nblA1 and nblA2 genes encoding phycobilisome degradation proteins [33,34] were up-regulated only in DrpoZ while glnB, encoding the nitrogen metabolism regulator protein PII [35], was up-regulated only in CS. Interestingly, Rre37, controlling some sugar catabolism genes in parallel with SigE mainly during nitrogen starvation [36], was up-regulated upon heat stress in both strains, but upregulation of its target genes glgP and glgX was only detected in DrpoZ. Differential regulation of several genes involved in nitrogen metabolism may suggest that acclimation of nitrogen metabolism to elevated temperature fails to occur normally in DrpoZ.
Seven genes showed opposite expression change in DrpoZ and CS upon mild heat stress. Five genes were down-regulated in DrpoZ and up-regulated in CS, but only one of these genes, trmD encoding tRNA (guanine-N1-)-methyltransferase, has an assigned function ( Fig. 2A). On the other hand, two genes were upregulated in DrpoZ and down-regulated CS; these genes were gcvP encoding glycine dehydrogenase and ccmR, which encodes a repressor protein regulating many genes involved in carbon concentrating mechanisms [37]. In standard growth conditions, the ccmR gene is down-regulated simultaneously with the downregulation of its target genes and operons [10] indicating complex regulation of carbon concentrating mechanisms in DrpoZ.
According to DNA microarray results, more than 80% that showed up or down regulation in DrpoZ did not respond similarly to a mild heat treatment in CS ( Fig. 2A). Up-regulation of photosynthetic and respiratory genes was more common in DrpoZ than in CS (Figs. 2B and 3). Furthermore, many genes for biosynthesis of amino acids and cofactors, prosthetic groups and carriers were up-regulated upon heat stress in DrpoZ strain but only few in CS (Fig. 2B).
Although DrpoZ grew well in standard conditions, the DNA microarray analysis revealed that 187 genes were at least two-fold up-regulated and 212 genes down-regulated in DrpoZ cells compared to CS in standard growth conditions [10]. Our next question was whether the genes showing different response to mild-heat treatment in DrpoZ and CS were similarly or differently expressed in the standard conditions. The heat map reveals that numerous genes up-regulated upon heat stress in DrpoZ were actually down-regulated in DrpoZ compared to CS in standard conditions (Fig. 3). For example, genes encoding NADH dehydrogenase subunits that were shown to be down-regulated in DrpoZ in standard conditions [10] were up-regulated in DrpoZ upon heat treatment but not in CS (Fig. 3). Furthermore, many genes that were down-regulated upon mild heat stress in DrpoZ were found to be up-regulated in DrpoZ compared to CS in standard conditions (Fig. 3). In standard conditions we showed that recruitment of the primary s factor SigA occurs less frequently in DrpoZ than in CS, which leads to down-regulation of many highly expressed genes in DrpoZ [10]. The physiological experiments using group 2 s factor mutant strains have revealed that SigB and SigC factors are important for high temperature acclimation responses [14,15,25] and thus is tempting to speculate that the v subunit not only affects the recruitment of SigA but also the recruitment of the other s factors.
RNA polymerase and total RNA contents decrease in mild heat stress more in CS than in DrpoZ Next we analyzed the RNAP content of cells in mild heat stress. The cells were grown under standard conditions and then transferred to 40uC for 2, 6 or 24 h. Western blots showed a clear decrease of RNAP during the high temperature treatment in CS; after one day treatment, cells had lost 45% of the RNAP core subunits a and b (Fig. 4AB). On the contrary, the DrpoZ strain lost less than 10% of RNAP core subunits a and b (Fig. 4AB). In addition, the amount of the primary s factor, SigA, decreased in heat stress; after 24-h heat treatment 45% and 17% of SigA was lost in CS and DrpoZ, respectively (Fig. 4C). The v subunit decreased similarly in CS as the other RNAP core subunits (Fig. 4D). In accordance with decrease of RNAP in CS, the total RNA content of CS cells decreased from 1.2 mg/ml in cultures with OD 730 = 1 [10] to 0.8 mg/ml after a 24-h treatment at 40uC (Fig. 4E). In the DrpoZ strain, the RNA content was similar as in CS in standard conditions [10]. The RNA content of DrpoZ decreased only 17% during the 24-h heat treatment (Fig. 4E) suggesting that the higher RNAP content of DrpoZ keeps transcription in DrpoZ more active than in CS in mild heat stress. However, the stability of transcripts is known to vary according to environmental cues [38] and we cannot rule out the possibility that the RNA contents of CS and DrpoZ are affected by RNA stability at high temperatures.
More than 90% of the total RNA in cells consists of rRNA, and analysis of total RNA by agarose gel electrophoresis revealed that the rRNA content of CS was lower than that of the DrpoZ strain (Fig. 4E). In E.coli, severe heat stress has shown to disturb ribosome assembly [39] and on the other hand, ribosomes form inactive 100S dimers when cells enter a non-growth mode in stationary phase [40,41]. In our mild heat stress conditions, CS grew as well as in standard conditions, indicating that translation remained fully active although the rRNA content of the cells decreased. Increase in temperature speeds up enzyme reactions and a lower amount of ribosomes might provide fully active translation in a slightly elevated temperature. In the case of DrpoZ, further experiment are required to find out whether a high rRNA content directly affects ribosome content and whether all ribosomes are translationally active or not.
We used total RNA samples in DNA microarray analysis, and the decrease in the RNA content of the cells during mild heat stress might affect the DNA microarray results, as we do not know whether the mRNA/rRNA ratio remained similar in all samples. However, overall signal intensities in the DNA microarray raw data did not reveal any systematic differences between the treatments or the strains, suggesting that the mRNA/rRNA ratio was not drastically different between samples. The method used for data normalization was found to be important when time series samples were analyzed [42]; in pairwise comparisons, performed in the present study, the quantile method is regularly used.

Photosynthetic capacity of DrpoZ decreased in mild heat stress
Photosynthesis is known to be a heat sensitive process. A 60-min heat treatment at 42uC was shown to reduce photosynthetic activity by 15% [43], and many parts of photosynthetic reactions, including carbon fixation by Rubisco and photosynthetic light Figure 3. Comparison of heat stress responsive genes in CS and DrpoZ. The left panel shows genes whose expression was at least two-fold up-regulated or down-regulated either in the control or DrpoZ strains or both upon heat treatment when the gene expression was compared to the expression of the same strain under standard growth conditions, and in addition gene expression of DrpoZ and CS were compared in standard growth conditions. The heat maps show log 2 fold change values (P,0.05) on the scale from 22 (blue) to 2 (red); values bigger than 2 are also shown in red and values smaller than 22 are blue. If the P value was $0.05, the fold change was given the value 0. Genes were arranged to categories according to Cyanobase, letters on the left indicating the same categories as in Fig. 2B. On the right, magnification of differently regulated genes in photosynthesis (top), regulatory functions (middle), and transport and binding proteins (bottom) is shown. doi:10.1371/journal.pone.0112599.g003 reactions, especially the oxygen evolving complex of PSII, are known to be heat sensitive [44]. As many genes belonging to the category ''photosynthesis and respiration'', showed differential response to heat in CS and DrpoZ (Figs. 2B and 3), we studied the acclimation of the photosynthetic machinery. To measure heatinduced changes, cells were grown in standard conditions and thereafter treated at 40uC for 24 h under constant illumination, PPFD 40 mmol m 22 s 21 .
We detected the amounts of different photosynthetic complexes during the 24-h treatment at 40uC with western blotting. A clear decrease of PSII (measured using an antibody against the PSII core protein CP43) and 10 to 15% decrease of Rubisco (measured using an antibody against RbcL) occurred in both strains (Figs. 5A and 5B), while the PSI content (antibody against PSI reaction center protein PsaB) remained at the same level as in standard growth conditions (Fig. 5C). In CS, the phycobilisome antenna proteins phycocyanin and allophycocyanin remained constant during the 24-h heat treatment at 40uC (Figs. 5D and 5E). However, in DrpoZ the phycocyanin content of the cells decreased (Fig. 5D) although allophycocyanin (Fig. 5E) remained at the same level as in the control conditions. Interestingly, heat treatment induced up-regulation of the nblA1 and nblA2 genes (encoding the phycobilisome degradation proteins NblA1 and NblA2, respectively) in DrpoZ but not in CS (Fig. 3). NblA1 and NblA2 proteins form a heterodimer [33] that acts as an adaptor guiding the Clp protease to phycobilisomes [45]. These findings suggest that upregulation of NblA proteins in DrpoZ upon heat stress induces degradation of phycobilisome rods that consist of phycocyanin.
After the 24-h treatment at 40uC, the light-saturated photosynthetic activity of CS, measured by oxygen evolution, was 92% of that measured in standard conditions (Fig. 6). In standard conditions, light-saturated photosynthetic activity of DrpoZ was circa 20% lower than in CS (Fig. 6A) and it further decreased in mild heat stress being only 68% of that measured in CS after 24-h treatment at 40uC (Fig. 6). The light-saturated PSII activities of the cells grown in mild heat stress, measured using a quinone electron acceptor, were 2.0260.08 and 1.3860.16 mmol O 2 / OD 730 /h in the control and DrpoZ strains, respectively, indicating that PSII of the DrpoZ strain was vulnerable to heat-treatment.

Conclusions
The heat-lethal phenotype of DrpoZ strain emphasizes the view that the small v subunit of RNAP is an important core polypeptide although cells can survive without it in optimal laboratory conditions. The total RNA content of the cells remains higher in DrpoZ than in CS during heat stress, and therefore the heat-lethal phenotype of DrpoZ is probably not caused by a decrease in active RNAP due to the proposed chaperone-like activity of the v subunit. Instead, our data suggest that numerous heat acclimation processes malfunction in DrpoZ. As summarized in Fig. 7, these acclimation processes include adjustment of transcription, photosynthesis and nitrogen metabolism. Gene expression respond differently in DrpoZ and CS, and the data indicate that the small v subunit affects expression of specific genes not only in standard growth conditions but also during heat stress.

Strains, growth conditions and growth measurements
The glucose tolerant control strain of Synechocystis sp. PCC 6803 [46], the v subunit inactivation strain DrpoZ and the complementation strain DrpoZ+rpoZ [10] were grown in BG-11 medium supplemented with 20 mM Hepes pH 7.5. The OD 730 of liquid cultures was set to 0.1 (0.35 mg of chlorophyll (Chl) a/ml), and the cells were grown (30 ml of cell culture in a 100-ml Erlenmeyer flask) at 32uC, 38uC or 40uC in air level CO 2 under constant illumination at the PPFD of 40 mmol m 22 s 21 and shaking at 90 rpm. In some experiments, as indicated, BG-11 medium was supplemented with 20 mM Hepes, pH 8.3. Samples of dense cultures were diluted with BG-11 before the optical density was measured, so that OD 730 did not exceed 0.4, and the dilutions were taken into account when the final results were calculated. All measurements were conducted on at least three independent biological replicates.

Survival rates at 40uC
OD 730 was set to 0.1, and cells were grown at 40uC for 24 h. The OD 730 was measured, cells were diluted with fresh BG-11 medium to OD 730 = 0.1. Then culture was serially diluted to 1:10, 1:100, 1:1000 and 1:10 000; and twenty drops containing 10 ml of the dilutions were spotted onto BG-11 plates. Plates were grown in standard conditions for one week, the colonies were counted and CFUs were calculated as CFUs/1-ml cell culture with OD 730 = 0.1.

DNA microarray analysis
For DNA microarray studies, OD 730 was set to 0.1, and the cells were grown in standard growth conditions for three days. Then the samples from standard conditions (OD 730 = 1, 40 ml) were harvested by centrifugation at 4500 g for 6 min at 4uC in precooled centrifuge tubes [10] or cells were treated at 40uC under continuous illumination, PPFD 40 mmol m 22 s 21 , for 24 h before harvest. The RNA was isolated using the hot-phenol method as described in [47], and further purified with RNeasy Mini Kit (Qiagen) to remove DNA contaminations. A 8615 K custom Synechocystis sp. PCC 6803 array (Agilent) was used in microarray experiments [48], and hybridizations and data collection were done as described previously [49]. The data were normalized using the quantile method and the t-test was used to identify differentially expressed genes. A gene was considered differentially regulated if log 2 of the fold change was $1 (at least two-fold upregulated) or #21 (down-regulated to one half or less) and P, 0.05. Gene expression data were visualized with a heat map drawn with the open source software Multiple Experiment Viewer [50].
Total RNA content of the cells Cells were first grown in standard growth conditions and then treated at 40uC under continuous illumination, PPFD 40 mmol m 22 s 21 , for 24 h before harvest. Total RNA was isolated with the hot-phenol method [47] from 1-ml of cell culture with OD 730 = 1, and suspended in 12 ml of water. RNA concentration was measured with NanoDrop spectrophotometer and RNAS were visualized by running 5-ml samples on 1.2% agarose gels and staining the gels with ethidium bromide.

Photosynthetic activity
Light-saturated photosynthetic activity in vivo was measured (1 ml sample, OD 730 = 1) with a Clark type oxygen electrode (Hansatech Ltd.) at 32uC in BG-11 medium supplemented with 10 mM NaHCO 3 . The light-saturated PSII activity was measured using 0.7 mM 2,6-dichloro-p-benzoquinone as an artificial electron acceptor, and samples were also supplemented with 0.7 mM ferricyanide to keep the electron acceptor in oxidated form.

Supporting Information
Table S1 Genes at least two fold up-regulated in the control strain after a 24-h treatment at 406C. (PDF)     Fig. 3 and their expression data. (PDF)