Escherichia coli K-12 contains nine paralogs of CspA, CspA–CspI, collectively known as CspA family of cold-shock proteins (CSPs). Members of this family are small proteins consisting of a nucleic acid-binding domain called the cold-shock domain (CSD), one of the most evolutionarily conserved domains found in various life forms including bacteria, plants and animals (Wolffe 1994; Graumann and Marahiel 1998). CspA and some of its homologs share nucleic acid melting ability (Jiang et al. 1997; Phadtare et al. 2002) and transcription anti-termination activity (Bae et al. 2000), the two functions important for cold adaptation (Phadtare et al. 2002). In spite of the high degree of similarity among themselves, only five (cspA, B, E, G and I) are induced during cold shock (Gualerzi et al. 2003; Uppal et al. 2008). cspC, involved in regulation of growth (Rath and Jawali 2006), is constitutively expressed at 37 °C, while cspD is induced during starvation (Yamanaka and Inouye 1997; Uppal et al. 2014). The transcription from both cspE and cspD promoters is activated by cyclic AMP (cAMP) receptor protein (CRP), a global metabolic regulator (Uppal et al. 2011, 2014). CRP, also known as catabolite activator protein (CAP), is responsible for regulation of genes involved in various metabolic pathways in E. coli (Kolb et al. 1993). The absence of CRP leads to a growth defect at low temperature (Uppal et al. 2014).

In the present study, we show that cspB, cspG and cspI, the members of cspA family and induced in response to cold shock (reviewed by Gualerzi et al. 2003), are regulated by CRP. To investigate the role of CRP in regulating the cold-inducible csps, pcspA-gfp, pcspB-gfp, pcspG-gfp and pcspI-gfp, the gfp promoter fusions [Open Biosystems Inc. (Zaslaver et al. 2006)] for cspA, cspB, cspG and cspI, respectively, were used. Both M182 [E. coli K12 Δ (lacIPOZY) × 74galK galU strA] and M182Δcrp strains were transformed with the transcriptional fusions. Freshly diluted cultures were grown in 96-well plate without shaking. Both GFP-mut2 fluorescence (excitation: 480 nm and emission: 510 nm) and the absorbance of the culture at 600 nm were measured (OD600) in a multiwell plate reader (Infinite200, Tecan, UK). The specific fluorescence intensity (SFI), defined by fluorescence signal per cell, was calculated for growth at 37 °C (Fig. 1). Since GFP-mut2 variant is highly stable and non-toxic to the cells, the accumulation of the fluorescence per cell (SFI) serves as the reporter of the promoter activity (Zaslaver et al. 2006). For all the genes, the SFI levels showed continuous increase during growth at 37 °C in the wild type cells (Fig. 1a–d). For cspA, the SFI levels followed an almost similar pattern in both crp knockout cells and the wild type cells indicating no regulation by CRP (Fig. 1a). For cspB, the SFI levels were significantly lower (maximum ~7 fold) in crp knockout cells compared to the wild type cells (Fig. 1b) suggesting a positive regulation of cspB by CRP. In contrast to cspB, both cspG and cspI showed an increase in the SFI level in the knockout cells as compared to the wild type cells (Fig. 1c, d) indicating that both cspG (~1.9 fold) and cspI (~1.8 fold) are negatively regulated by CRP. These results suggest that CRP up-regulates cspB and down-regulates cspG and cspI. This was further confirmed by complementing M182Δcrp cells with pDCRP (West et al. 1993), a multicopy plasmid having crp and pDU9, plasmid control, respectively (Fig. 1e). The presence of multicopy crp marginally decreased cspA expression, while both cspI and cspG showed a ~1.9 and ~2.3 fold decrease thereby confirming that both cspG and cspI are down regulated by CRP. In contrast, cspB level did not show any change after complementation with multicopy crp. It is possible that cspB promoter responds differently to the multiple copies of crp gene in trans on a plasmid as compared to the single copy of crp present on the chromosome.

Fig. 1
figure 1

Plot of specific fluorescence intensity (SFI), expressed in arbitrary units (A.U.), against time for M182 and M182Δcrp cells harboring the GFP fusions for the respective genes a pcspA-gfp for cspA, b pcspB-gfp for cspB, c pcspG-gfp for cspG and d pcspI-gfp for cspI. The specific fluorescence intensity is calculated as the ratio of fluorescence and OD600. The upstream region present in the fusion plasmids pcspA-gfp, pcspB-gfp, pcspG-gfp and pcspI-gfp is spanning −213 to +214 for cspA; −203 to +234 for cspB; −205 to +239 for cspG; and −292 to +217 for cspI, respectively. e A bar diagram showing SFI levels after 14 h of inoculation in a plate assay for M182, M182Δcrp, M182Δcrp cells harboring pDU9 (control) and pDCRP (WT CRP) for cspA, cspB, cspG and cspI, respectively. The plate assay was carried out essentially as described in Uppal et al. (2011). f SFI levels from individual transcriptional GFP fusions (for cspA, cspB, cspG and cspI) in M182 and M182Δcrp cells after 4, 24 and 48 h of inoculation during continuous growth at 15 °C. The cultures were grown in a microtitre plate at 15 °C with continuous shaking. All the experiments were repeated thrice and the SFI data corresponds to the mean of three independent repeats with error bars showing the standard error

Full size image

Though above results confirm that, among the four cold-induced csps, only cspB, cspG and cspI are regulated by CRP at 37 °C, the role of CRP in regulating their expression at low temperature still needs to be addressed. For this, the effect of crp mutation on their expression during growth at 15 °C was analyzed. Both M182 and M182Δcrp strains harboring respective transcriptional fusions (for cspA, B, G, I) were grown continuously at 15 °C. The SFI levels were measured at 4, 24 and 48 h after inoculation (Fig. 1f). For cspB, crp deletion resulted in decreased SFI levels during initial stage of growth only, while cspG levels were consistently higher in crp deletion strain. For cspI, crp deletion resulted in higher SFI levels at 4 and 48 h, while the levels at 24 h did not show any difference as compared to the wild type. Interestingly, crp deletion caused two to fourfold increase in cspA levels at all stages of growth at 15 °C. These results indicate that CRP regulates cspA, cspB, csp G and cspI during different stages of growth at low temperature.

The CRP-mediated regulation could be either by direct binding of CRP to the gene promoter [e.g., cspE (Uppal et al. 2011) and cspD (Uppal et al. 2014)] or indirect, where CRP might regulate a primary regulator of the gene being analyzed. Sequence analysis of the promoter region corresponding to cspA, B, G and I revealed putative CRP target sites (Fig. 2a). Interestingly, the putative CRP target sites in cspA, cspG and cspI, the genes negatively regulated by CRP (Fig. 1), overlap with −10 element, −35 element and transcription start site of the gene promoter (Fig. 2b), respectively. In case of cspB and cspI, two CRP target sites were predicted. In order to find out whether CRP binds to the promoter region of cspA, B, G and I, electrophoretic mobility shift assay (EMSA) was carried out essentially as described in Uppal et al. (2011). The DNA fragments spanning the upstream regions (corresponding to the upstream region present in the respective transcriptional fusions) of these genes were amplified using MG1655 chromosomal DNA as a template. The primers used (Table 1) were as follows: SU87-A-F and SU88-A-R for cspA spanning positions −213 to +214, SU89-B-F and SU90-B-R for cspB spanning positions −203 to +234; SU91-G-F and SU92-G-R for cspG spanning positions −205 to +239; SU93-I-F and SU94-I-R for cspI spanning positions −292 to +217; SU82-D-F and SU81-D-R for cspD spanning positions −274 to +1; and SU40-E-F and SU43-E-R for cspE spanning positions −88 to +12 in the upstream region. These fragments with and without CRP (~200 nM) were electrophoresed on an 8 % polyacrylamide gel in the presence of cAMP (200 μM). As CRP is known to bind to cspE (Uppal et al. 2011) and cspD (Uppal et al. 2014), these genes were also included in the analysis as positive controls. In line with previous results (Uppal et al. 2011, 2014), CRP showed binding to both cspD and cspE (Fig. 2c). However, in spite of showing a match to the consensus CRP site (Fig. 2a), the putative target sites in cspA, cspB, cspG and cspI did not show binding to CRP. It may be possible that the putative sites in these genes have become degenerate and lost the binding function along the evolutionary process. Alternatively, the putative sites may have very week binding affinity for CRP and may only function in the presence of other hitherto unknown factor(s). Nevertheless, the above results indicate that the regulation of cspA, cspB, cspG and cspI by CRP is likely to be through an indirect mechanism. This indirect regulation could be due to a transcriptional regulator whose activity is controlled by CRP or could be via CRP-activated small non-coding RNAs (e.g., CyaR) which are induced as a result of nutritional depletion (DeLay and Gottesman 2009). One cannot also rule out the regulation due to changes in DNA topology as CRP is known to regulate gyrA levels in the cell (Gomez-Gomez et al. 1996).

Fig. 2
figure 2

a The 22 b consensus CRP target site sequence and the putative CRP target sites in cspD, cspE, cspA, cspB, cspG and cspI are aligned against each other. CRP target site is a palindromic sequence, where two conserved core motifs, having the five-base-pair sequence TGTGA, are separated by a six-base-pair spacer. The numbering of the nucleotide positions in the target site is essentially same as used by Ebright et al. (1984). The first conserved core motif is from nucleotide positions 4–8 and the second core motif is from 15 to 19. Nucleotide positions 9–14 is termed as spacer region in the CRP target site. The position of the dyad axis of the target sites is mentioned. The residues identical to the consensus are shown in bold and the two core motifs are underlined. The position of −10 element, −35 element and transcription start site overlapping with the CRP target sites in cspA (−11.5), cspG (−28.5) and cspI (+7.5), respectively, are shown by a rectangular box. b A pictorial representation of the positions of the CRP sites with respect to the gene promoter in the upstream region sequence (same as present in gfp transcriptional fusions) of cspA, cspB, cspG and cspI (figure not to scale). c Electrophoretic mobility shift assays (EMSA) probing binding of CRP (~200 nM) to the upstream region of cspA, cspB, cspG, cspI, cspD and cspE, genes using an 8 % polyacrylamide gel

Full size image
Table 1 Primers used in this study
Full size table

E. coli adapts to low temperature by inducing a cold-shock response which, in addition to many other classes of genes, includes induction of some of the cspA family members (cspA, cspB, cspE, cspG and cspI) (Gualerzi et al. 2003). The RNA chaperone activity and nucleic acid melting ability of Csps help to facilitate translation/transcription/anti-termination by resolving nucleic acid secondary structures at low temperature (Jiang et al. 1997; Bae et al. 2000; Phadtare et al. 2002). In a previous study (Uppal et al. 2011), we have shown that CRP activates cspE and the presence of CRP is critical for continuous growth at 15 °C indicating an important role for CRP during growth at low temperature. The present study reinforces and widens this role of CRP by showing that CRP regulates cold-induced cspA family members during growth at 37 °C (cspB, cspG, and cspI) as well as at 15 °C (cspA, cspB, cspG and cspI). It will be pertinent to mention here that CRP regulates cspA expression selectively during growth at low temperature. Similar to this, H-NS-mediated repression of virF happens selectively at low temperature, presumably due to altered level of DNA supercoiling, in Shigella flexneri (Drlica and Rosenthal 1999). The mechanism of the low temperature specific CRP-mediated regulation of cspA is not yet known.

Overall, this study demonstrates that CRP regulates multiple cold-induced cspA family members, albeit indirectly, during normal growth and at low temperature growth. A balanced level of these proteins seems to be important for optimal growth at low temperature (Xia et al. 2001). Also, prolonged expression of CSPs is known to adversely affect growth at low temperature (Yamanaka and Inouye 2001). Based on our results (present study; Uppal et al. 2011), we speculate that the growth defect conferred by crp deletion (Uppal et al. 2011) might be a result of altered transcription level of csp genes. As CRP regulates more than 70 genes encoding transcription regulators (Shimada et al. 2011), it is difficult to identify the factor(s) downstream of CRP mediating the indirect regulation of csps. Nonetheless, this study corroborates several earlier reports which have indicated a cross-talk between the cellular metabolic status and cold-shock response (Beaufils et al. 2007; Farewell and Neidhardt 1998; Jones et al. 1992; VanBogelen and Neidhardt 1990; Uppal et al. 2011). In addition, global gene expression profiling has indicated that many cold-induced genes may be subject to catabolite repression (Gosset et al. 2004; Phadtare and Inouye 2004). It would be interesting to explore if CRP facilitates a cross-talk between cellular metabolic status and cold-shock response.