Introduction

Most organisms including bacteria respond to environmental changes and induce the expression and accumulation of a specific subset of proteins for adaptation (Barria et al. 2013; Graumann and Marahiel 1996; Thieringer et al. 1998). Likewise, cyanobacteria also respond to changes in environmental temperatures and adapt to temperature stress, such as cold stress. RNA-binding proteins (Rbps) were identified as cold-shock proteins in cyanobacteria 20 years ago (Sugita and Sugiura 1994; Sato 1995). They contain a single RNA recognition motif (RRM) (Kenan et al. 1991). In the unicellular cyanobacteria Synechococcus elongatus strains PCC 6301 and PCC 7942, two Rbps (Rbp1 and Rbp2) were shown to express differently in cold-treated cells (Mutsuda et al. 1999; Sugita et al. 1999). Rbp1 is barely detected at 30 °C while the production of Rbp1 is dramatically induced upon a downshift of temperature to 10 °C. In contrast, Rbp2 accumulates at similar levels during 25 h after this temperature downshift (Mutsuda et al. 1999). Similar cold-inducible expression of rbp genes was found in the multicellular cyanobacteria Anabaena variabilis M3 (Maruyama et al. 1999; Ehira et al. 2003), Anabaena sp. PCC 7120 (Ehira et al. 2003) and the bloom-forming cyanobacterium Microcystis sp. (Tan et al. 2011). Cyanobacterial Rbps are considered to be the counterparts of major cold-shock protein (Csp) in Escherichia coli or Bacillus subtilis (Yamanaka 1999; Graumann and Marahiel 1998).

We previously reported that rbp1-disrupted mutants of S. elongatus PCC 7942 grew very poorly at 20 °C, while the rbp2 disrupted grew well at 20 °C (Sugita et al. 1999). Therefore, Rbp1 is considered to have an essential role in cellular function at cold temperatures. Based on nucleotide sequence alignment of the upstream region from the translation initiation codon of rbp genes of A. variabilis M3, putative cis-elements, RBS, Boxes I, II and III, for cold-inducible expression of rbp were proposed (Ehira et al. 2003, 2005). However, it is still unknown whether such sequence elements are responsible for cold-inducible expression of rbp genes.

In this paper, we identified a new member of the rbp genes, rbp3, by a similarity search in the complete genome sequence of S. elongatus PCC 6301 (Sugita et al. 2007). Such rbp3 genes exist in cyanobacterial species whose genome sequences were completely determined. We carried out a comparative study to assess how three rbp genes, rbp1, rbp2 and rbp3, are expressed under cold temperatures in S. elongatus PCC 6301. Moreover, we used rbp1-luxAB transcriptional fusions or translational fusions with deletions to identify the cis-acting sequence elements involved in the accumulation of rbp1 transcript at cold temperatures. For this analysis, the transformable strain S. elongatus PCC 7942 was used. The rbp genes ant their upstream sequences are identical between S. elongatus strains PCC 6301 and PCC 7942 (Sugita et al. 1999).

Materials and methods

Culture conditions

Synechococcus elongatus PCC 6301 and PCC 7942 cells were grown at 30 °C in BG-11 medium (Sugita and Sugiura 1994) supplemented with 2 mM NaHCO3 under constant light (15 µE m−2s−1) with shaking. For cold stress experiments, cells grown at 30 °C were adjusted to 20, 15 or 10 °C rapidly in an ice-water bath and were then cultured in the temperature-adjustable incubator. Cells grown under the different temperature conditions were collected and frozen in liquid nitrogen.

RNA extraction, northern and dot blot analysis

RNA was extracted by Isogen (Nippon gene, Japan) according to the manufacturer’s protocol. The cells (20 mg) were lysed in 200 µl of hot Isogen (50 °C) and incubated at 50 °C for 10 min. The cell lysate was treated with chloroform, and RNA was precipitated with isopropanol. To purify RNA, the RNA preparation was treated with DNase I and was then treated with acid phenol and chloroform (1:1, v/v). For northern blot analysis, RNA (2 µg) was separated on a 1 % agarose gel containing formaldehyde and blotted on Hybond N+ membranes (GE Healthcare Life Sciences). For dot blot analysis, RNA (1 µg) was spotted and fixed on a membrane. Hybridization probes for rbp1, rbp2, rbp3, trnV-UAC and 16S rRNA genes were amplified by the polymerase chain reaction (PCR) from the S. elongatus genomic DNA using gene-specific primers (Table S1). The probe for luxAB was amplified by PCR from pAM977 (Kondo et al. 1993) with luxAB-F and luxAB-R primers (Table S1). DNA probes were labeled with [α-32P]dCTP, and northern or dot blot hybridization was carried out as previously described (Sugita and Sugiura 1994). Autoradiographic images were analyzed and quantified using a Fuji bio-imaging analyzer BAS-2000.

Construction of plasmids

Plasmid pALK9 was generated by insertion of a 2.1 kb promoter-less Vibrio harveyi luxAB gene (from the nucleotide position −125 with respect to the luxA initiation codon to +66 with respect to the luxB termination codon, derived from pAM977, Kondo et al. 1993) into the EcoRV and SalI sites of cyanobacterial transformational vector pAM1573, which contained the S. elongatus PCC7942 chromosomal neutral sites (NS) II, a chloramphenicol resistance gene cartridge (CmR) and a trpA terminator (Andersson et al. 2000). To construct the transcriptional fusion plasmids Prbp1::luxAB and Prbp2::luxAB, the promoter regions of rbp1 (nucleotide positions −153 to +127 relative to the transcription initiation site +1) and rbp2 (−126 to +247) were amplified by PCR and were then inserted into the pALK9 at StuI and XbaI sites.

To construct rbp1 promoter deletion series, Prbp1, F93, F63, F33 and F3, the upstream regions of rbp1 were amplified by PCR. Each PCR fragment was inserted into pALK9 at StuI and XbaI sites. To construct a translational fusion plasmid Trbp1, the rbp1 promoter region (−153 to +127) fused in-frame with the luxAB coding region (+16 from the initiation codon) was inserted into pAM1573. To construct rbp1 5′-UTR deletion series, TD1, TD2, TD3 and TD4, were generated by site directed mutagenesis of plasmid Trbp1 by PCR. All PCR primers are provided in Table S1.

Measurement of bioluminescence

To measure the luminescence from Synechococcus cells carrying rbp1-luxAB fusions, 100 µl of cell culture was transferred to a micro titer plate and mixed with 10 µl of a 0.05 % n-decanal (n-decyl aldehyde, Sigma) emulsion. Bioluminescence of the cell suspension was measured with a luminometer (ATTO Luminescencer JNR AB2100, Japan) immediately after the addition of n-decanal. The intensity of bioluminescence was expressed in counts of photons per minute per OD730. Cell culture without the luxAB reporter was used to measure the background and subtracted from sample measurements.

Results and discussion

Cyanobacterial rbp gene families consist of three types of members

The complete genome sequence of S. elongatus PCC 6301 revealed an open reading frame (syc0747_d) homologous to rbp1 and rbp2 (Sugita et al. 2007). syc0747_d (142 residues) showed 41 and 40 % amino acid identity with Rbp1 (106 amino acids) and Rbp2 (99 amino acids), respectively, and was therefore designated as Rbp3 and the gene as rbp3 (Fig. 1). The identical gene (Synpcc7942_0790) was found in S. elongatus PCC 7942 (http://www.genome.jp/kegg/catalog/org_list.html). Cyanobacterial Rbps can be classified into roughly three types I–III, based on the phylogenetic analysis of RRMs of Rbp proteins and the existence of C-terminal characteristic amino acid sequences (Fig. 1). Type I Rbps have a C-terminal glycine (Gly)-rich region. Type II proteins, including Rbp2, are small proteins without a C-terminal extension. Type III proteins have a non-Gly-rich C-terminal conserved sequence Pro-Asp-Pro-Arg-Trp-Ala ‘PDPRWA’ (Hamano et al. 2004). Type I proteins can be further divided into two subtypes Ia and Ib. Type Ia and Ib have a short (5–15 residues) and a long (30–80) Gly-rich region, respectively. Rbp1 belongs to Type Ia. Type 1b Rbps are distributed only among marine cyanobacterial species.

Fig. 1
figure 1

Schematic structure of cyanobacterial Rbp proteins. The filled, striped, dotted and gray regions represent RNA recognition motif (RRM), glycine-rich domain, the conserved sequence ‘SGWEDRSY’ unique to marine cyanobacterial Rbps and the conserved sequence ‘PDPRWA’ found in Rbp3, respectively

Full size image

rbp1 transcript highly accumulated at cold temperatures while rbp2 and 3 transcript levels largely decreased

To examine the effect of cold temperatures on the expression of the three rbp genes, S. elongatus PCC 6301 cells grown at 30 °C were transferred to various low temperatures and grown for 3 h. The rbp1 transcript accumulated at substantial levels at 20 °C and at highest levels at lower temperatures (10 and 15 °C). In contrast, the rbp2 and rbp3 transcripts decreased remarkably after a temperature shift from 30 to 20 °C or lower (Fig. 2a). rbp2 transcript (550 nt) is longer than rbp1 (510 nt) because it has a longer 5′UTR. The transcript level of the tRNAVal gene (trnV-UAC) located upstream of the rbp1 gene was not affected by low temperature shifts.

Fig. 2
figure 2

Steady-state transcript levels of rbp genes after different temperature shifts. RNA (2 µg) was subjected to northern blot analysis using each rbp-specific probe. a rbp transcript levels 3 h after the indicated temperature shift. b Changes in the accumulation of rbp transcripts after a temperature shift from 30 to 15 °C relative to time (0, 10, 30 min, 1 and 3 h)

Full size image

Next, we investigated the kinetics of changes in rbp transcript levels after this temperature shift. The rbp1 transcript level rapidly increased within 30 min and peaked at 1 h after the temperature shift from 30 to 15 °C (Fig. 2b). By contrast, the rbp2 transcript level largely decreased after 60 min following a downshift in temperature and appeared at a negligible level after 3 h. The level of rbp3 transcript also declined after 60 min after a temperature downshift but was maintained at half the level of the control (i.e., before cold shock) (Fig. 2b). This suggests that rbp3 may play a certain role in the cellular function at cold temperatures.

The conspicuous changes to rbp transcript level after exposure to low temperatures can be due to changes in either transcriptional activity or mRNA stability. To investigate this possibility, the region upstream from each rbp coding region was fused to the luxAB genef and was integrated into the PCC 7942 genome neutral site NSII (Fig. 3a). The temperature shift from 30 to 15 °C induced bioluminescence in rbp1-luxAB and the intensity increased tenfold (Fig. 3b, left panel). The intensity of bioluminescence decreased rapidly after a temperature upshift from 15 to 30 °C. Such a cold response was not observed for rbp2-luxAB (Fig. 3b). This indicates that the luxAB can be used as a reporter assay for investigating the regulatory mechanism of rbp1 gene expression under cold temperatures.

Fig. 3
figure 3

Detection of cold-shock expression of rbp genes by a luxAB reporter assay. a A reporter vector pALK9, which contained the luxAB gene, the chloramphenicol resistance gene (CmR), trpA terminator (T) and genome neutral site (NSII) for homologous recombination, was used throughout the present study. Each promoter region of the rbp1 or rbp2 genes was inserted into pALK9. b Bioluminescence from the reporter cells was measured at 0, 2, 4, 6, 8, 24 h after they were transferred from 30 to 15 °C (closed circles). Dashed lines indicate the point of temperature shift from 15 to 30 °C, and bioluminescence was measured at 2 and 4 h after this temperature shift. Control luminescence of each reporter cell under constant temperature (30 °C) is indicated as open squares

Full size image

Involvement of the upstream region from the transcription initiation site of rbp1 in cold-shock induction

To address how the accumulation of rbp1 transcript is regulated under cold temperatures, we carried out in vivo reporter assays of S. elongatus PCC 7942 cells carrying a series of deletion mutants of the rbp1 promoter region. The cells were grown at 30 °C and then transferred to 15 °C for 24 h. Bioluminescence from these cells was then measured. The bioluminescence intensity for Prbp1::luxAB containing the −153 to +127 region peaked at 15 °C and showed a 4.5-fold increase at 15 °C relative to 30 °C (Fig. 4a). As a control, Prbp2::luxAB did not exhibit any increase in bioluminescence or luxAB transcript levels. Prbp1 carrying the region from −153 to +97 gave 30 % lower intensity than Prbp1::luxAB at both temperatures 30 and 15 °C. This suggests that the +97 to +127 region containing the first 15-bp rbp1 coding region is probably required for high expression of rbp1 (Fig. 4a, c, d). Bioluminescence levels at 15 °C gradually decreased after further deletion of the upstream region from the transcription initiation site (F93, F63 and F33) but were maintained at about 5–8-fold higher level than at 30 °C. The luxAB transcript levels were 1.5- to 3-fold higher at 15 °C than at 30 °C in all the transformed cells (Fig. 4b, right panel). In contrast, F3 exhibited a background level of bioluminescence intensity even at 30 or 15 °C. Negligible levels of luxAB transcripts also accumulated at both temperatures (Fig. 4b). These results suggest that the region from −33 to −3 may contain an essential sequence for basal transcription of the rbp1 gene and that the 120-bp region (−34 to −153) from the transcription initiation site of rbp1 does not contain critical cis-elements required for cold-shock induction (Fig. 4c). However, we cannot exclude the possibility that the region from −33 to −3 is involved in not only basal transcription but also cold-inducible transcription of the rbp1 gene.

Fig. 4
figure 4

Deletion analysis of the upstream sequence of the transcription initiation site of the rbp1 gene. a rbp1-luxAB transcriptional fusions with various truncated upstream sequences of the transcription initiation site were constructed. Each construct was introduced into S. elongatus PCC 7942 cells. Bioluminescence was measured 24 h after a temperature shift from 30 to 15 °C. The data represent an average of five independent experiments. The relative level at 30 to 15 °C for each construct is shown on the right of the graph. b Steady-state levels of 16S rRNA, rbp1 and luxAB transcripts at 0 or 3 h after the temperature downshift were analyzed by dot blot hybridization. c The nucleotide sequence of the Prbp1::luxAB construct is shown. Nucleotide numbers are indicated and that of the transcription initiation site, represented by a bold capital ‘C,’ is +1. Capital letters start from ATG of the rbp1 coding region. Two underlines represent a putative −10 sequence and a Shine–Dalgarno (SD)-like sequence, respectively. d The nucleotide sequence downstream from the transcription initiation site of Prbp1, F93, F63, F33 and F3 construct is indicated. The upstream sequence of each construct is not shown

Full size image

Rbp1 mRNA is stabilized at low temperatures

The stability of cspA mRNA is known to play an important role in the cold-shock activation of cspA (Fan et al. 1997). To assess the possibility of posttranscriptional regulation of rbp1 expression, we tested rbp1 mRNA stability in cells treated by an inhibitor of RNA polymerase, rifampicin. Rifampicin (50 µg ml−1) was added to the cold (15 °C)-treated wild-type PCC 7942 cells, immediately after the temperature upshift, and then RNA was extracted at different time points. As shown in Fig. 5, rbp1 mRNA was extremely unstable at 30 °C and more than 95 % of the rbp1 mRNA was rapidly degraded within 10 min after the temperature shift from 15 to 30 °C, whereas rbp1 mRNA was not degraded at 15 °C. This result indicates that rbp1 mRNA stability is markedly altered by different growth temperatures.

Fig. 5
figure 5

Stability of rbp1 mRNA after temperature shift. Cultures were incubated at 15 °C and then shifted to 30 °C or maintained at 15 °C after the addition (+) or no addition (–) of rifampicin (50 µg ml−1). RNA was then extracted at different time points as indicated, and northern blot hybridization was carried out as described in “Materials and methods” section

Full size image

Role of rbp1 5′-UTR in cold-shock induction

The upstream sequences that are conserved among the cold-regulated rbp genes in Anabaena PCC 7120 and A. variabilis M3 were suggested to be involved in the regulation of cold-inducible expression (Maruyama et al. 1999; Ehira et al. 2003, 2005). Similar conserved sequences, Boxes I, II and III, are also present in the 5′-UTR of the S. elongatus rbp1 gene. To investigate the role of these sequences in rbp1 expression under cold temperatures, we performed in vivo reporter assays of S. elongatus PCC 7942 cells carrying a series of a 25-bp deletion within the rbp1 5′-UTR (Fig. 6a, b). Each deletion construct carries the rbp1 translated region (15 bp) fused in-frame to the translated region of luxAB (Fig. 6b, c). Trbp1 cells exhibited over 70-fold higher bioluminescence at 15 °C than at 30 °C. To our surprise, TD2 cells only emitted high bioluminescence (over 30-fold higher level at 15 °C than at 30 °C) while TD1, TD3 and TD4 exhibited very low bioluminescence. Besides, luxAB mRNA was detected at a low level in TD1, TD3 and TD4 cells both at 30 and 15 °C (Fig. 6b, right panel). This suggests that rbp1-luxAB mRNAs lacking SD-like sequence or Boxes II and III may be unstable in TD1, TD3 and TD4 cells. Further experiments, however, are needed to demonstrate how 5′-UTR is involved in stabilization of rbp1-luxAB mRNA. In contrast, rbp1-luxAB mRNA accumulated at a threefold higher level in TD2 cells than Trbp1 cells at 30 °C. This suggests that the region (from +63 to +87) containing Box I may be involved in destabilizing rbp1 mRNA at 30 °C. In TD2 cells, the bioluminescence levels significantly increased at 15 °C although the rbp1 mRNA levels were not largely changed at both at 30 and 15 °C under our reporter assay conditions. This suggests that the 5′-UTR of rbp1 mRNA controls the translation.

Fig. 6
figure 6

Deletion analysis of the 5′-UTR of rbp1 transcript, a The region from −153 to +127 of the rbp1 gene was translationally fused to the luxAB coding sequence, which was inserted into pAM1573 to generate Trbp1. b For deletion analysis of the 5′-UTR, a 25 bp each in the 5′-UTR of Trbp1 was deleted to construct TD1 to TD4. SD-like, Box I, Box II and Box III, are presented as different colored boxes, respectively. Each construct was introduced into S. elongatus PCC 7942 cells and bioluminescence was measured as described in Fig. 4. Steady-state levels of the 16S rRNA and luxAB transcript at 0 or 3 h after the temperature downshift were analyzed by dot blot hybridization. c The nucleotide sequence of the Trbp1 construct is shown. SD-like, Box I, Box II and Box III, are presented as different colored boxes, respectively (color figure online)

Full size image

In A. variabilis M3, an rbpAI transcript that lacks the entire 5′-UTR did not accumulate both at high (38 °C) and at low (22 °C) temperatures (Ehira et al. 2005). Taken together with our result and that of others (Ehira et al. 2005), we speculate that rbp1 mRNA stabilization at 15 °C requires the regions containing the SD-like sequence, Boxes II and III. In 5′-UTR of rbp1 mRNA, a stable secondary structure (minimum free energy −19.5 kcal/mol) could be formed at 30 °C and could hide the SD sequence preventing the entry of ribosomes, as shown in Fig. 7. The 5′ half of the 5′-UTR containing Boxes II and III might be involved in the formation of a stable secondary structure. Alteration of the secondary structure at different temperatures might require some unknown factor(s).

Fig. 7
figure 7

Possible secondary structures of the 5′-UTR of rbp1 mRNA. Secondary structures of the rbp1 transcript at different temperatures were predicted using MFOLD Web Server (http://unafold.rna.albany.edu/?q=mfold/rna-folding-form). SD-like, Box I, Box II and Box III are indicated by different colored thick lines. The outlined ATG indicates the translation start codon of the rbp1 coding region. The triangle represents the position of a 25-nt deletion (+63 to +87) (color figure online)

Full size image

E. coli and B. subtilis Csp proteins are considered to play a role in RNA stability at cold temperatures (Goldenberg et al. 1996; Fan et al. 1997; Giuliodori et al. 2010). Regulation of the cspA gene of E. coli has been extensively studied and posttranscriptional mechanisms play a major role in cold adaptation (Goldenberg et al. 1996; Brandi et al. 1996; Gualerzi et al. 2003). Moreover, the cspA mRNA itself plays a role as a thermosensor to modulate translation of the cold-shock protein CspA (Giuliodori et al. 2010). In the context of posttranscriptional regulation by changing mRNA stability, the expression of S. elongatus PCC 7942 rbp1, A. variabilis M3 rbpAI and E. coli cspA genes show a similar pattern in response to cold stress but the detailed mechanism, including cis-elements and trans-acting factors, involved in this cold regulation might differ RNA structural rearrangement of rbp1 mRNA exposed to cold temperatures may require other cold shock-induced protein(s), such as RNA helicase (Chamot et al. 1999; Rosana et al. 2012). This possibility remains to be further addressed.