Autoregulation of RNA Helicase Expression in Response to Temperature Stress in Synechocystis sp. PCC 6803

RNA helicases are ubiquitous enzymes whose modification of RNA secondary structure is known to regulate RNA function. The pathways controlling RNA helicase expression, however, have not been well characterized. Expression of the cyanobacterial RNA helicase, crhR, is regulated in response to environmental signals that alter the redox poise of the electron transport chain, including light and temperature. Here we analyze crhR expression in response to alteration of abiotic conditions in wild type and a crhR mutant, providing evidence that CrhR autoregulates its own expression through a combination of transcriptional and post-transcriptional mechanisms. Temperature regulates crhR expression through alteration of both transcript and protein half-life which are significantly extended at low temperature (20°C). CrhR-dependent mechanisms regulate both the transient accumulation of crhR transcript at 20°C and stability of the CrhR protein at all temperatures. CrhR-independent mechanisms regulate temperature sensing and induction of crhR transcript accumulation at 20°C and the temperature regulation of crhR transcript stability, suggesting CrhR is not directly associated with crhR mRNA turnover. Many of the processes are CrhR- and temperature-dependent and occur in the absence of a correlation between crhR transcript and protein abundance. The data provide important insights into not only how RNA helicase gene expression is regulated but also the role that rearrangement of RNA secondary structure performs in the molecular response to temperature stress. We propose that the crhR-regulatory pathway exhibits characteristics similar to the heat shock response rather than a cold stress-specific mechanism.


Introduction
The rearrangement of RNA secondary structure, required for numerous crucial cellular functions, is catalyzed by a variety of enzymes including RNA helicases. RNA helicases belong to a gene superfamily, SF2, members of which are encoded in essentially every organism from a variety of viruses to humans. SF2 is comprised of a number of protein families, with the DEAD-box proteins comprising the largest family [1]. RNA helicases function as molecular motors utilizing ATP hydrolysis to catalyze rearrangement of RNA and RNP structure with individual helicases performing specific functions potentially affecting all aspects of RNA metabolism [2][3][4]. Cellular pathways requiring RNA helicase activity include not only housekeeping functions such as translation initiation, ribosome biogenesis and RNA splicing and turnover but also developmental and stress pathways and small RNA metabolism [5][6][7]. Once RNA helicase expression is induced, the resulting RNA helicase activity has the potential to regulate expression of downstream genes required for the developmental or stress response.
Organisms utilize a variety of pathways to respond to temperature shift, the best characterized is heat shock while the mechanisms regulating the cold shock response are less defined. Cold stress induces expression of a limited number of genes, however, unlike heat stress, a sigma factor or two-component signal transduction system functioning as a global regulator of the response has not been identified [8]. This indicates that alternative mechanism(s) regulate gene expression in response to a temperature downshift in bacteria.
Frequently, RNA helicases are associated with the cold stress response in bacteria and higher organisms [9][10][11][12][13][14][15]. In prokaryotes, extensive analysis associates RNA helicase activity with ribosome biogenesis, RNA turnover and cold stress. Of the five Escherichia coli DEAD-box RNA helicases, srmB and dbpA are required for ribosome biogenesis, rhlB and rhlE with RNA degradation as a component of the degradosome and csdA is associated with both functions [11][12][16][17][18][19]. Frequently, these functions are observed in response to cold stress, for example, at low temperature csdA and rhlE are degradosome-associated [16][17]. The association of RNA helicase expression and function with low temperature is also observed in photosynthetic Gram-negative cyanobacteria in which two DEAD-box RNA helicases have been studied, crhC and crhR [20][21][22][23][24][25]. crhC is expressed solely in response to low temperature [20][21] while crhR is regulated by the redox status of the electron transport chain, with expression increased by conditions that enhance reduction of the chain, for example low temperature or salt stress and dark-light transition [22,26]. Despite the presence of numerous cold-induced RNA helicases in prokaryotes, the mechanism(s) by which their expression is regulated by temperature shift is not well defined.
At the physiological and morphological levels, crhR inactivation has profound effects on cellular metabolism at 30uC which are exacerbated at 20uC [27]. crhR mutants are cold sensitive with respect to both growth and photosynthetic activity, a phenotype resulting primarily from a defect in photosynthetic carbon fixation. These physiological effects are manifested morphologically by the crhR mutant progressively accumulating cellular damage at 20uC, including a reduction in the level and organization of carboxysomes and thylakoid membranes with the concomitant accumulation of membrane vesicles within the cells [27].
Here, we comprehensively investigate the molecular regulation of crhR expression in response to abiotic stress. In particular, we identify that temperature regulation of transcript and protein stability contributes significantly to crhR expression involving a unique combination of CrhR-dependent and CrhR-independent pathways. The results provide evidence that CrhR RNA helicase activity is required for transcriptional and post-transcriptional mechanisms that result in autoregulated expression.

Bacterial strains and growth conditions
Wild type Synechocystis sp. strain PCC 6803 and the partial deletion mutant, DcrhR [27], were maintained on BG11-agar Northern analysis. Total RNA (5 mg) was probed with a 93 bp HincII-SacII internal fragment of crhR. Synechocystis cells were grown at 30uC and stressed by transfer to the dark or 20uC for 1 h before harvesting at the indicated temperature. Shown below is the stripped blot probed with the Synechocystis rnpB as a control for RNA loading. Transcript abundance was quantified using Image J software Version 1.45 S (NIH, USA) [32], crhR transcript abundance was normalized using corrected rnpB levels with basal transcript abundance observed in illuminated wild type cells grown at 30uC set to 1.0 serving as a reference for the fold-change values shown (Methods S1). (B) Western analysis. Soluble protein (25 mg) isolated from the cells used for Northern analysis above was probed with anti-CrhR antibody. The anti-CrhR antibody detects a 55 kDa polypeptide in wild type cells and a ,27 kDa truncated polypeptide in the DcrhR mutant. Relative protein abundance is provided below each lane, in comparison with the abundance detected in illuminated wild type cells grown at 30uC set to 1.0. doi:10.1371/journal.pone.0048683.g001 Synechocystis were grown to mid-log phase at 30uC at which time the cultures were transferred to 20uC for the indicated times before harvesting. crhR transcript was detected in total RNA probed with a 93 bp HincII-SacII internal fragment of crhR. The blots were stripped and probed with the Synechocystis rnpB gene as a control for RNA loading. C) Quantification of crhR transcript levels. Transcript levels were quantified as described in Figure 1 and expressed as accumulation relative to crhR abundance observed in illuminated wild type cells grown at 30uC (set to 1.0) serving as a reference for the fold-change values shown (Method S1  [20][21]28]. Media for crhR mutant growth was supplemented with spectinomycin and streptomycin, both at 50 mg/mL [27]. Temperature and dark stress were induced by transferring aliquots of cultures grown at 30uC to the indicated condition for 1 h. Cells for protein analysis were harvested at the stated growth temperature and cell pellets flash frozen in liquid nitrogen. For RNA extraction, cells and RNases were rapidly inactivated by addition of an equal volume of ice-cold ethanol-phenol buffer (ethanol25% buffer saturated phenol) directly to the cell culture at the stated growth temperature. For temperature and half-life time course experiments, 400 mL cultures grown at 30uC were transferred to the indicated conditions and 50 mL aliquots removed at the indicated times for RNA and protein extraction. All experiments were repeated a minimum of two times with representative data shown.

RNA manipulation
RNA was extracted using glass bead lysis in the presence of phenol followed by extensive phenol-chloroform extraction and lithium chloride precipitation. Northern analysis, using a 32 P-  Synechocystis were grown at 30uC to mid-log phase (0 min) at which time the cultures were cold stressed at 20uC for 1 h to induce maximal crhR transcript abundance. de novo RNA synthesis was subsequently inhibited by the addition of rifampicin (400 mg/ml) and one-half of the culture transferred back to 30uC. Aliquots for RNA extraction were harvested at the indicated times. crhR transcript was detected in total RNA probed with a 93 bp HincII-SacII internal fragment of crhR. The blots were stripped and probed with the Synechocystis rnpB gene as a control for RNA loading. Quantification of relative transcript abundance (%) at each time point is provided below each lane, normalized for the level of rnpB detected in each lane as described in Methods S1. Open circles, DcrhR; open boxes, wild type. doi:10.1371/journal.pone.0048683.g004 labelled riboprobe corresponding to a 93 bp HincII-SacII internal fragment of the crhR ORF, was performed in formamide buffer at 65uC, as described previously [20][21]28]. Transcript half-life was estimated in the presence of rifampicin (400 mg/mL), added to cultures immediately prior to transfer to the new temperature. Accumulation of the Synechocystis rnpB transcript was utilized as a control for RNA loading. The rnpB transcript abundance is extensively utilized for this purpose in cyanobacteria [29][30][31]. Transcript size was estimated using Fermentas RiboRuler TM RNA markers. Transcript levels were quantified using the Image J software Version 1.45 S (NIH, USA) [32].

Protein manipulation
Protein extraction and immunoblot analysis were performed essentially as described previously [21,28]. Soluble protein was extracted using glass bead lysis, 25 mg resolved by 10% SDA-PAGE, electro-transferred to Hybond ECL membrane and probed with the indicated antibody. Antibody complexes were detected on X-ray film using the Amersham ECL Western Blotting Detection kit. Polyclonal antiserum against Synechocystis CrhR or E. coli Rps1 was used at a dilution of 1:5000. Rps1 levels were used as an internal control for protein loading. Protein halflife was determined in cells grown to mid-log phase at 30uC, transferred to 20uC for 2 h to achieve maximum accumulation of Figure 5. CrhR protein half-life. Wild type (A) and DcrhR (B) Synechocystis were grown to mid-log phase at 30uC (0 min) at which time the cultures were transferred to 20uC for 2 h to achieve maximum CrhR and DCrhR abundance. Chloramphenicol (250 mg/ml) was added to inhibit de novo protein synthesis and one-half of each culture was transferred back to 30uC. Samples for soluble protein extraction were harvested at the indicated times. Blots were simultaneously probed with antibodies against CrhR and E. coli ribosomal protein S1 (Rps1), used as a control for protein loading. Quantification of relative CrhR and DCrhR protein abundance at each time point is provided, below each lane, normalized for the level of Rps1, as described in Methods S1. doi:10.1371/journal.pone.0048683.g005 CrhR at which time chloramphenicol (250 mg/ml) was added to inhibit de novo protein synthesis and half of each culture transferred to 30uC. Protein concentration was quantified using the Bradford assay (Bio-Rad) with BSA as the standard. Protein levels were quantified using Image J software Version 1.45 S (NIH, USA) [32].
ImageJ analysis X-ray films were scanned using a UMAX PowerLook 2100XL scanner with the resolution set to 800 dpi. Scans were saved as tif files and imported into the ImageJ software package, available from http://imagej.nih.gov/ij/ [32]. The density of transcript and protein signals were plotted, corrected for background and integrated to give area values. Calculations normalizing for loading, correction of detected signals and determination of accumulation and abundance to quantify transcript and protein levels are described in Methods S1.

Induction of crhR expression in response to temperature and light-dark stress
We have extended our previous observations indicating that crhR expression is regulated in response to light-induced alteration of the redox poise of the electron transport chain [22] by investigating expression in response to temperature and light-dark transitions in wild type and crhR mutant Synechocystis (Fig. 1). In the crhR mutant, the DCrhR peptide is biochemically inactive, as it does not unwind dsRNA or anneal complementary ssRNAs (Chamot and Owttrim, unpublished). The fate of the truncated mRNA and protein products can therefore be investigated in the absence of biochemically active CrhR. In wild type cells grown at 30uC, a basal level of crhR transcript accumulation is observed, a level which decreases significantly in response to dark treatment for 1 h (Fig. 1A). crhR mutation altered this basal level as crhR transcript accumulation is enhanced 2.6-fold under standard growth conditions at 30uC in illuminated cells (Fig. 1A). However, regulation of transcript accumulation at 30uC was not completely lost in the DcrhR mutant as transcript abundance, while elevated, was not increased to the levels observed under cold stress (Fig. 1A). While a predominant ,1.5 knt crhR transcript was observed in wild type cells other stable, low abundance transcripts are also detected. crhR inactivation altered the transcript pattern, with four prominent stable transcripts of 2.3, 1.5, 1.3 and 0.75 knt accumulating in mutant cells. Although in wild type cells, the crhR probe detects multiple transcripts, they do not accumulate to the levels observed in the mutant. In response to temperature stress at 20uC for 1 h, crhR transcript levels increase and decrease substantially in the light and dark, respectively (Fig. 1A). However, similar responses to low temperature in both wild type and DcrhR cells were observed, increasing significantly and, importantly, to approximately the same degree with respect to the basal levels observed at 30uC (Fig. 1A). Thus, crhR mutation altered the basal transcript abundance at 30uC but not the magnitude of the initial response to low temperature stress. Unexpectedly, accumulation of the rnpB transcript, coding for the functional RNA, RNase P, and Figure 6. Temperature gradient of crhR expression. Wild type and DcrhR Synechocystis were grown to mid-log phase at 30uC and divided into 6 aliquots, each of which was incubated at the indicated temperature for 1 h and samples were harvested for RNA and protein extraction. (A) Total RNA on a northern blot was probed with a 93 bp HincII-SacII internal fragment of crhR, stripped and probed with the Synechocystis rnpB gene as a control for RNA loading. crhR transcript levels were quantified at each temperature as described in Methods S1 and the fold change in crhR accumulation compared to the abundance in illuminated, wild type cells grown at 30uC (set to 1.0) is provided below each lane. Ethidium bromide staining of the RNA present on the Northern blot indicates that essentially equal quantities of RNA were loaded in each lane ( Figure S1). (B) Western blots were probed with anti-CrhR antiserum and subsequently probed with antibodies against E. coli Rps1 as a control for protein loading. Protein levels were quantified at each temperature and the fold change in CrhR or DCrhR compared to the abundance in illuminated, wild type cells grown at 30uC (set to 1.0) is provided below each lane, as described in Methods S1. doi:10.1371/journal.pone.0048683.g006 used as a control for RNA loading, is marginally (8-10%) but consistently reduced in the DcrhR mutant under all conditions tested (Fig. 1A). CrhR protein abundance corresponds with transcript accumulation in wild type cells with a basal level observed in the light at 30uC increasing significantly at 20uC and decreasing in response to dark treatment (Fig. 1B). In contrast, temperature regulation of protein accumulation was significantly altered in the DcrhR mutant, in which the 27 kDa truncated CrhR polypeptide (DCrhR) was constitutively present at an elevated level, irrespective of temperature or light-dark stress (Fig. 1B). Interestingly, the enhanced abundance of the DCrhR peptide observed under all conditions was essentially identical to that observed in wild type cells at 20uC. Thus, in wild type and DcrhR cells, CrhR protein levels do not correspond to transcript abundance and CrhR protein accumulates to a maximal level, irrespective of either transcript abundance or temperature.

Time course of crhR transcript accumulation
Alteration of the cellular response to temperature and light-dark stresses associated with CrhR inactivation prompted investigation of the kinetics of transcript accumulation. An essentially identical initial response in transcript accumulation is observed in both wild type and DcrhR cells exposed to 20uC, transcript increasing in a linear fashion from a basal level at 30uC to a maximal level within 20 min ( Fig. 2A and 2B). crhR transcript accumulation occurs transiently in wild type cells, decreasing to the basal level observed at 30uC within 6 h of exposure to cold stress ( Fig. 2A and  2C). The transient expression occurring in wild type cells is not observed in the absence of functional CrhR, maximal crhR transcript levels accumulate after 20 min and remain consistently elevated for the duration of the experiment (Fig. 2B and 2C). Again, differential accumulation of four stable crhR transcripts was observed between wild type and mutant cells. This suggests that there is a defect in crhR transcript processing and/or degradation of the processed transcripts in the absence of CrhR activity ( Fig. 2A  and 2B).

Time course of CrhR protein accumulation
CrhR protein accumulation in response to growth at 20uC also did not reflect the kinetics of crhR transcript accumulation in either cell type. In wild type cells, CrhR protein abundance corresponded with transcript levels during the early stages of the low temperature response but with delayed kinetics, reaching a maximum within 2 h of exposure. This correspondence did not hold at longer exposure times, with protein levels remaining elevated for the course of the experiment while transcript decreased significantly (compare Fig. 2A and Fig. 3A). As observed in Figure 1, temperature-regulated expression of the DCrhR polypeptide was absent in the DcrhR mutant. DCrhR protein was expressed at relatively constant levels at all time points, with only a slight increase observed with extended low temperature exposure ( Fig. 3B and 3C). Thus, again, crhR mutation resulted in a constant  Table S1. (B) Western blots were simultaneously probed with antibodies against CrhR and E. coli Rps1 that served as a control for protein loading. The relative fold change in CrhR and DCrhR accumulation, corrected for Rps1 levels, as described in Methods S1, is given below each lane. Protein samples were resolved on separate gels, hence normalization was performed independently for the wild type and mutant, with abundance observed in wild type and mutant at 30uC set to 1.0. doi:10.1371/journal.pone.0048683.g007 level of CrhR protein accumulation irrespective of transcript level or temperature.

crhR transcript and CrhR protein half-life
The observed alterations in transcript and protein accumulation may reflect changes in macromolecular half-life. Indeed, crhR transcript abundance was regulated by temperature, but in a similar manner in both wild type and DcrhR cells. Transcript stability was significantly enhanced at low temperature in wild type Synechocystis with half-lives of .30 and ,10 min at 20uC and 30uC, respectively (Fig. 4A). Half-life was not altered by crhR mutation as similar values are observed in the DcrhR cells, .30 min at 20uC and ,13 min at 30uC (Fig. 4 B). These results suggest that the enhanced accumulation of crhR transcript or the lack of transient accumulation observed in DcrhR cells in response to low temperature do not result from a defect in CrhR-dependent degradation of the crhR transcript.
Temperature regulation of protein half-life similarly controls CrhR accumulation in wild type cells (Fig. 5A). At 30uC, CrhR exhibited a relatively short half-life of ,3 h which was increased significantly to .48 h at 20uC. In the DcrhR mutant, DCrhR peptide levels remained elevated over the entire time course, not altering significantly in response to either chloramphenicol or temperature (Fig. 5B). Thus, while transcript half-life is not affected by crhR mutation, peptide half-life is significantly altered at both temperatures but more dramatically at 30uC, a response that is CrhR-dependent. The results indicate that crhR inactivation does not affect crhR transcript turnover but significantly affects CrhR protein turnover.

Temperature gradient induction of crhR transcript and protein accumulation
It was of interest to determine if the temperature induction was a gradual or an all-or-none process. This was investigated by analyzing steady-state transcript and protein levels in Synechocystis exposed to temperatures ranging from 35uC to 10uC for 1 h (Fig. 6). At the transcript level, a basal level of crhR transcript accumulated at 35 and 30uC, which progressively increased to a maximum at 15uC (Fig. 6A). At 10uC, transcript accumulation was enhanced marginally, crhR transcript levels increasing 1.7-fold in comparison with the basal level detected at 30uC (Fig. 6A). CrhR protein expression in wild type cells mimicked transcript levels with the basal level of CrhR protein expression observed at 35 and 30uC increasing to a maximum at 20uC and subsequently decreasing to the basal level at 10uC (Fig. 6B). In the DcrhR mutant, transcript abundance increases in response to decreasing temperature but reaches a maximum more rapidly and accumulates above basal levels at 10uC compared to wild type (Fig. 6A). As observed in Figure 3B, DCrhR protein levels in the crhR mutant remain essentially constant over the entire temperature gradient, increasing marginally in response to a 1 h exposure to temperatures #20uC (Fig. 6B).

Time course of crhR transcript and protein accumulation at 10uC
The observation that crhR transcript levels were not enhanced in response to growth at 10uC indicates that a 5uC downshift in temperature dramatically affected cellular ability to respond to low temperature. This could result from either an inability to sense and/or respond at the transcriptional level or simply reflect a temperature-induced reduction in overall transcriptional activity. Figure 8. Schematic summary of crhR expression and regulation. crhR expression is controlled by a complex interaction between temperature-regulation of both transcript accumulation and protein degradation. crhR transcript. crhR transcript half-lives are equal and short in both wild type and DcrhR Synechocystis cells which contributes to a basal level of transcript accumulation at 30uC. A temperature downshift to 20uC rapidly induces crhR transcript accumulation associated with enhanced half-life and with similar kinetics in both cell types, suggesting that CrhR is not required for temperature sensing or induction of its own transcript accumulation. At low temperature, crhR is transiently accumulated in wild type cells whereas conversely, transcript remains elevated at 20uC in the DcrhR mutant, suggesting that CrhR activity is required for the transient expression. Although crhR half-life is influenced by temperature, being significantly longer at 20uC, it is identical in both cell types. This suggests that CrhR is not directly involved in degradation of its own transcript but functional CrhR is directly associated with repression of crhR transcript accumulation, most likely through another mechanism, possibly altered regulation of transcription. CrhR protein. CrhR protein levels correspond to transcript levels in wild type Synechocystis, accumulating to a basal level at 30uC and increasing significantly at 20uC. This is distinctly not the case in the DcrhR mutant in which CrhR protein remains elevated and constant at both temperatures. CrhR protein accumulates to the level observed in wild type Synechocystis at 20uC, irrespective of transcript accumulation or temperature. Combined, these results suggest that the reduced level of CrhR at 30uC is caused by proteolytic degradation which is temperature-and CrhR-dependent as this process is inactive in the DcrhR mutant. In addition, there appears to be a maximal level of CrhR that can accumulate in cells, irrespective of the transcript level, a process that is CrhR independent. This accumulation appears to be a default level as it was never observed to increase above the level detected in wild type cells at 20uC. doi:10.1371/journal.pone.0048683.g008 To further investigate this phenomenon, crhR transcript and protein levels were analyzed in response to prolonged exposure to 10uC. Overall, the time course of both crhR transcript and protein accumulation in wild type cells resembled that observed at 20uC (Fig. 2 and 3) however the kinetics were delayed at 10uC. In wild type cells, maximal transcript accumulation occurred within 3 h, plateauing and then decreasing up to 24 h (Fig. 7A). Conversely, CrhR protein accumulated progressively over the course of the experiment, even subsequent to crhR transcript decline (Fig. 7A), as also observed in Figures 2 and 3. In contrast, crhR transcript in the DcrhR mutant was constitutively observed at all time points at 10uC (Fig. 7B, Table S1). The pattern of transcript accumulation differed from that observed at 20uC (Fig. 6A) with the 2.3 and 0.75 knt transcripts predominating at 10uC (Table S1). In addition, although the total accumulation of the four detected transcripts was not altered (Table S1), the relative accumulation of the two transcripts shifted in response to extended exposure to 10uC, with the 0.75 knt predominating at shorter and the 2.3 knt predominating at longer exposures, respectively (Fig. 7B, Table S1). This was not a result of alteration in total transcript, as quantification indicates that abundance of all four transcripts was constant over the experiment (Table S1). This is reflective of the results shown in Figures 2B and C, in both cases crhR transcript accumulation in the mutant is maximal within 1 h of exposure to 20uC and remains constant thereafter. These results imply that there is a defect in crhR transcript processing in the absence of CrhR RNA helicase activity. Again, DCrhR peptide levels remain relatively constant under all conditions tested, with only a marginal increase observed after prolonged exposure to 10uC, irrespective of RNA abundance (Fig. 7B). Similar to the results shown in Fig. 1A were rnpB transcript levels are consistently reduced under all conditions at 20uC, rnpB levels decreased progressively in response to extended growth at 10uC in the DcrhR mutant (Fig. 7B). The observed decrease in rnpB accumulation would normally be an indication of unequal RNA loading, however analysis of total RNA in each lane, visualized by ethidium bromide staining, indicated that RNA loading was approximately equal in each lane (Fig. S1).
crhR expression was also evaluated in response to exposure of a single culture to a 5uC decrease in growth temperature for one hour, progressively from 35 to 10uC. In this scenario, a significant increase in crhR transcript level was detected in wild type cells in response to the temperature decrease from 15 to 10uC. This implies that wild type cells possess the capacity to respond to 10uC if they are cold adapted for one hour (data not shown). In agreement, CrhR protein levels also increase at 10uC in response to this temperature regime. As expected from experiments presented above, the level of the DCrhR polypeptide detected in mutant cells remained relatively constant, increasing slightly at temperatures below 20uC. Again, rnpB levels were observed to decrease at 10uC in the single culture exposed to the progressive 5uC decreases in temperature for 1 h intervals (data not shown).

Discussion
The lack of a common global regulator of cold stress indicates that a variety of transcriptional and post-transcriptional mechanisms control gene expression in response to a temperature downshift [8]. The data presented here provide evidence that expression of the cyanobacterial RNA helicase, crhR, is controlled by a complex network of regulatory checkpoints involving both CrhR-dependent and CrhR-independent pathways, as outlined in Figure 8. Determination if CrhR is directly or indirectly involved in the autoregulation process and the mechanism by which the autoregulation functions are crucial questions. RNA helicases have well defined direct roles in translation initiation, RNA turnover and ribosome biogenesis in prokaryotic and eukaryotic systems [4,19]. An indirect role for CrhR in autoregulatory mechanisms may be associated with one of these cellular pathways. For example, transient expression and proteolysis may originate indirectly from CrhR-regulated translation of transcripts whose protein products are required for these processes. An indirect role for CrhR in the observed regulation involving small RNA metabolism, as observed for RNA helicases in abiotic stress responses in eukaryotic systems [6] or on transcription cannot be ruled out. These may not reflect an indirect effect as p68 and p72 RNA helicases that catalyze similar biochemical reactions as CrhR [25], function as transcriptional co-regulators [33].
Temperature plays a profound role in crhR expression, enhancing both crhR transcript and protein half-life, although CrhR performs divergent roles in both mechanisms. CrhRindependent pathways are associated with temperature sensing, signal transduction and subsequent molecular response at the initial stages of cellular response to low temperature. The CrhRindependence of the initial response generating crhR transcript accumulation is related to the enhanced reduction of the electron transport chain at low temperature that is required for crhR transcription [22]. A major aspect regulating crhR transcript abundance is controlled by temperature alteration of crhR transcript half-life, unstable at 30uC and significantly stabilized at 20uC, as observed for other cold-regulated bacterial transcripts [21,[34][35]. The level of crhR transcript accumulation at different temperatures can potentially be explained by the temperature alteration of transcript stability, although we cannot rule out effects of CrhR regulation of transcription. The pathway conferring transcript stability is CrhR-independent, as crhR transcript half-life is not altered by crhR mutation. This observation suggests that CrhR is not directly involved in the RNA degradation pathway responsible for turnover of its own transcript. This is an important observation as RNA helicases frequently function in RNA turnover in prokaryotic systems, switching of RNA helicase composition in the RNA degradosome occurring in response to low temperature, resulting in formation of a coldadapted degradation complex [36][37][38]. It is possible that CrhR is indirectly involved in degradation of its own transcript through regulation of translation initiation or another aspect of RNA metabolism associated with expression of the required RNase.
Transient accumulation of crhR transcript at 20uC and CrhR protein stability at all temperatures is CrhR-dependent. Transient transcript accumulation is a common characteristic of coldinduced genes in prokaryotic systems, although the mechanism is not well characterized [34,[39][40]. While transient accumulation in the absence of functional CrhR can also be interpreted as evidence that CrhR functions in the degradation of its own transcript, the lack of crhR transcript half-life alteration in the crhR mutant suggests that this is not the case. The kinetics of crhR transcript and protein accumulation more closely fit a scenario in which CrhR binding to the crhR transcript directs the CrhR-crhR complex to a CrhR-independent RNA degradation pathway. This proposal is similar to the direct mechanism by which proteolysis of the RNA chaperone CspC at high temperature regulates transient expression of heat shock mRNAs in E. coli [41], except that CrhR protein levels remain elevated continuously during temperature stress. Thus it appears that the CrhR-dependent transient accumulation of crhR transcript involves a direct association with RNA helicase activity, a role that does not involve RNA degradation.
CrhR peptide accumulation depends on a complex interplay between temperature regulation of both transcript and protein stability. The CrhR-dependent enhancement of CrhR peptide half-life at low temperature occurs in the absence of a correlation between protein and transcript levels. CrhR protein continues to accumulate in the absence of significant transcript accumulation at low temperature, similar to that observed for Rbp proteins in cyanobacteria [39]. This is in contrast to other cold-induced proteins where both transcript and protein transiently accumulate [34]. Changes in CrhR abundance could result from either altered translational efficiency or proteolysis. The increase in CrhR peptide abundance in the absence of functional CrhR at 30uC suggests that CrhR is not required for translation of its own transcript but is more likely indirectly involved to produce a component required for activity of the proteolytic degradation machinery. The mechanism does not appear to involve a general protein degradation pathway, as Rps1 levels are not significantly altered in these experiments. In addition, truncated CrhR polypeptide levels did not exceed those present in wild type cells at 20uC, suggesting that a CrhR-independent mechanism limits CrhR accumulation irrespective of transcript level or temperature. This implies that there is a maximum, default level to which CrhR can accumulate in Synechocystis, a process that is CrhR-and temperature-independent.
Proteolysis regulates the level of numerous proteins and regulatory circuits in bacterial systems, a prime example being the heat shock response [42][43][44]. In a natural environment, Synechocystis would normally experience temperatures below 25uC, conditions under which crhR expression would be constitutively induced. A temperature upshift to 30uC would enhance the proteolytic machinery that degrades CrhR. We therefore suggest that the results presented here resemble the regulation observed in response to heat stress rather than a cold stress specific phenomenon [44].
Regulation of prokaryotic gene expression is primarily thought to occur at the transcriptional level, with limited examples of control at the level of transcript or protein stability. The findings reported here, summarized in Figure 8, indicate that crhR expression is regulated in an unexpectedly complex manner, involving both CrhR-dependent and -independent pathways. CrhR-dependent pathways contribute to crhR expression via a novel combination of transcriptional and post-transcriptional mechanisms including autoregulation of the transient expression of crhR transcript at low temperature and CrhR protein accumulation at all temperatures. In contrast, aspects of temperature sensing, signal transduction, the initial increase in crhR accumulation and temperature-regulation of crhR transcript half-life involve CrhR-independent pathways. Frequently, these mechan-isms regulate crhR expression in the absence of a correlation between protein and transcript levels. CrhR modulation of these divergent pathways coordinates cyanobacterial response to temperature fluctuation.
The data provide unique insights into the complexity of pathways regulating RNA helicase expression associated with bacterial response to temperature stress at the molecular level. Moreover, the research highlights the importance of RNA helicase remodeling of RNA secondary structure on downstream gene expression and the physiological implications for bacterial adaptation to temperature change. Figure S1 Ethidium bromide stained gel corresponding to the crhR induction time course at 10uC. Total RNA (5 mg) extracted from wild type and DcrhR Synechocystis cells was separated on a 1.2% formaldehyde agarose gel at 100 V for 2.5 h. The gel was stained with ethidium bromide and imaged using a LKB 2011 Macrovue UV transilluminator equipped with a Kodak EDAS DC 290 camera and processed using Kodak 1D 3.6 imaging software. The ethidium fluorescence indicates that essentially equal amounts of RNA were loaded in each lane. Therefore, variations in rnpB transcript levels are not related to unequal RNA present in each lane.

(TIF)
Table S1 Quantification of crhR transcript hybridization detected in the DcrhR mutant shown in Figure 7B. (DOCX) Methods S1 Quantification of RNA and protein abundance. The methodology and calculations used to quantify the relative transcript and protein levels using Image J [32], utilizing rnpB and Rps1 accumulation as controls for transcript and protein loading, respectively, are provided. (DOCX)