Correspondence: Asim K. Bej, Department of Biology, University of Alabama at Birmingham, 1300 University Blvd., CH464, Birmingham, AL 35294-1170, USA. Tel.: +1 205 934 9857; fax: +1 205 975 6097; e-mail: email@example.com
A cold shock domain (CSD)-containing protein, CspD, of molecular mass ∼7.28 kDa in a psychrotolerant Antarctic Janthinobacterium sp. Ant5-2 (ATCC BAA-2154) exhibited constitutive expression at 37, 22, 15, 4 and −1 °C. The cspD gene encoding the CspD protein of Ant5-2 was cloned, sequenced and analyzed. The deduced protein sequence was highly similar to the conserved domains of the cold shock proteins (Csps) from bacteria belonging to the class Betaproteobacteria. Its expression was both time- and growth phase-dependent and increased when exposed to 37 °C and UV radiation (UVC, dose: 1.8 and 2.8 mJ cm−2). The results from the electrophoretic mobility shift and subcellular localization study confirmed its single-stranded DNA-binding property. In silico analysis of the deduced tertiary structure of CspD from Ant5-2 showed a highly stable domain-swapped dimer, forming two similar monomeric Csp folds. This study established an overall framework of the structure, function and phylogenetic analysis of CspD from an Antarctic Janthinobacterium sp. Ant5-2, which may facilitate and stimulate the study of CSD fold proteins in the class Betaproteobacteria.
Microorganisms isolated from Antarctica are suitable candidates to study physiological and genetic mechanisms for the adaptation to cold and subzero temperatures. It is notable that a number of bacteria that have been isolated from the Antarctic subzero temperature environment are psychrotolerant, exhibiting optimum growth temperatures at or above 22 °C (Morita, 1975). Accumulation of proteins of the cold shock domain (CSD) family and the regulation of their corresponding genes is one of the adaptive responses to cold temperatures that has been described in both mesophilic and psychrotolerant bacteria including Escherichia coli (Phadtare et al., 1999), Bacillus subtilis (Schindler et al., 1999), Arthrobacter globiformis (Berger et al., 1996), Pseudomonas putida (Gumley & Inniss, 1996), Salmonella spp. (Jeffreys et al., 1998), Rhodococcus spp.(Bej et al., 2000) and Pseudomonas sp.30-3 (Panicker et al., 2002). The CSD has been reported to be an evolutionarily conserved nucleic acid-binding domain of ancient origin found in eubacteria. It is also homologous to the CSD in human Y-box protein YB-1 and to other eukaryotic Y-box proteins (Graumann & Marahiel, 1998).
The structures of cold-shock proteins (Csps) from different bacteria have been determined by either X-ray crystallography or nuclear magnetic resonance, for example E. coli CspA (Newkirk et al., 1994; Schindelin et al., 1994), B. subtilis CspB (Schnuchel et al., 1993), Bacillus caldolyticus Csp (Mueller et al., 2000), Thermotoga maritima Csp (Kremer et al., 2001) and Neisseria meningitidis Csp (Ren et al., 2008). All of them share a common OB (oligonucleotide/oligosaccharide-binding) fold consisting of five β-barrel sheets with two consensus RNA-binding domains (RNP1 and RNP2) placed side by side on separate β-sheets, comprising a high proportion of basic and aromatic residues. The binding of B. subtilis CspB and B. caldolyticus Csp with hexathymidine (dT6) involves stacking interactions between phenylalanine residues and the thymidine base, together with hydrogen bonds between the side chains of polar amino acids and pyrimidine bases (Max et al., 2007).
Escherichia coli CspA family of proteins consist of nine homologs to the major cold-shock protein CspA (CS7.4) (Phadtare et al., 1999) and they either function as a RNA chaperones by minimizing the secondary structure formation in mRNAs to allow efficient translation at low temperatures or as transcription regulators and transcription antiterminators (Bae et al., 2000). Escherichia coli CspA, CspB, CspG and CspI are cold inducible, whereas CspC and CspE are constitutively expressed and have been shown to function as suppressors of the temperature-sensitive mukB106 mutation. The mukB gene is involved in the chromosome partitioning during cell division in E. coli (Yamanaka et al., 1994). The expression of E. coli cspF and cspH has not been associated with any particular growth condition or phenotype (Giaquinto et al., 2007). Non-cold-inducible E. coli CspD functions as a DNA replication inhibitor during the stationary growth phase. Its expression is inversely dependent upon the growth rate and induced upon glucose starvation at 37 °C (Yamanaka & Inouye, 1997). In another study, the function of the CspD in Halobacterium NRC-1 has been attributed to protection against UV damage of DNA, although the exact mechanism of such functions has not been elucidated (Baliga et al., 2004). It has been shown that under antibiotic stress, E. coli CspD causes arrest in growth and reduced viability under the direct regulation of a motility and quorum-sensing regulator toxin–antitoxin pair, MqsR/MqsA, that possesses RNase activity and influence cell physiology (e.g. biofilm formation and motility) leading to the establishment of the persister population (Kim et al., 2010; Kim & Wood, 2010).
It is apparent from the above findings that although all Csps contain highly conserved CSD, their physiological roles may vary depending upon the organisms, proteins and environment. Most structural and functional analyses were performed on Csps from mesophiles, but not from psychrophiles or psychrotolerant bacteria. Among all the classes of phylum Proteobacteria, the Csps from bacteria belonging to class Betaproteobacteria have not been characterized yet. Therefore, the major objective of this study was to elucidate the structure and physiological role of a cold-shock protein, CspD from a psychrotolerant Janthinobacterium sp. Ant5-2 isolated from Antarctica that has to survive extreme cold environment.
Materials and methods
Bacterial strains and growth media
Janthinobacterium sp. strain Ant5-2 (ATCC BAA-2154) was isolated from Lake Podprudnoye a small proglacial lake located in central Schirmacher Oasis, East Antarctica (latitude −70.766758°, longitude 11.610249°). Samples were collected aseptically in November 2008 from shallow water near the shoreline when ice-cover was absent and transported to the lab frozen and processed shortly after arrival. For all experiments in this study, Ant5-2 was grown in 1 : 2 (v/v) trypticase soy broth (TSB) (BD, Franklin, NJ). Escherichia coli BL21(DE3)pLysS (F - ompT hsdSB(rB -, mB -) galdcm (DE3) pLysS (CamR) (Novagen, WI) was grown at 37 °C in Luria–Bertini (LB) medium.
Growth kinetics and freeze tolerance
For growth and cold tolerance, 250 mL of 1 : 2 (v/v) TSB medium was inoculated with 1 mL of overnight culture of Ant5-2 and incubated at 22 °C until the OD450 nm reached 0.2. Then, 50 mL aliquots were transferred to incubator shakers set at 37, 22, 15, 4 and −1 °C. The OD was determined at various time intervals using a Lambda 2 spectrophotometer (Perkin Elmer, Norwalk, CT). To determine freeze tolerance, Ant5-2 culture was grown to exponential phase and frozen at −20 °C. The frozen culture was then subjected to one freeze–thaw cycle and viable plate count was determined before and after freezing as described previously (Panicker et al., 2002).
Radiolabeling of total cellular proteins and immunoblot
Total cellular proteins of Ant 5-2 cultures were radiolabeled using EasyTag™ Express 35S methionine Protein Labeling Mix (11 μCi μL−1) (Perkin Elmer) and transferred immediately to −1, 4, 15 and 22 °C, respectively, and incubated for 1.5 h as described previously (Panicker et al., 2002). The cell pellets were collected and proteins denatured at 98 °C for15 min in 2 × sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel loading buffer (BioRad, Hercules, CA). The protein concentration was estimated colorimetrically using BCA Protein assay kit (Pierce, Rockford, IL); 50 μg proteins from each sample was separated in 12% (w/v) denaturing SDS polyacrylamide gel; the gel was dried and autoradiographed. For immunoblot assays, total proteins from Ant5-2 cultures incubated at −1, 4, 15, 22 and 37 °C and cultures exposed to different doses of UVC (290–100 nm) were extracted, transferred and immunoblotted with CapB antibody from Pseudomonas sp. 30/3 in duplicate as described previously (Panicker et al., 2010). The membranes were scanned and the relative amount of CspD expressed in Ant5-2 was analyzed with imagej 1.40 software (http://rsb.info.nih.gov/ij/index.html).
Analysis of cold shock protein (Csp)
The cold shock gene was PCR amplified using primers designed from Csp of Janthinobacterium lividum (accession no. DQ074977), F-cspD-Ant5-2 (5′-dTTAGATTGGCTGAATGTTCGAAGCTTGC-3′) forward and R-cspD-Ant5-2 (5′-dATGGCAACTGGCATCGTAAAATGG-3′) reverse primers. The PCR amplification of the homologs of E. coli cspA on Ant5-2 genomic DNA was also attempted using a set of universal degenerate primers for Enterobacteriaceae CSPU5 (5′-dCCCGAATTCGGTAHAGTAAAATGGTTYAACKC-3′) and CSPU3 (5′-dCCCGAATCCGGTTACGTTASCWGCTKSHGGDCC-3′) (Francis & Stewart, 1997). The cycling parameters, cloning and sequencing method used was as described previously (Panicker et al., 2010). The deduced amino acid sequence of CspD from Ant5-2 was aligned with homologous sequences obtained from blast p and with protein sequences in NCBI database using clustal x (http://www.clustal.org/) and the phylogenetic trees were generated by neighbor joining method and mega version 4.0 (http://www.megasoftware.net/) software (Tamura et al., 2007).
In situ localization of Csp (Ant5-2) by immunofluorescence staining
Ant5-2 culture was grown in 1 : 10 (v/v) TSB growth medium at 22 °C until OD450 nm=0.3; then aliquots of 50 mL were exposed to 4, 15 and 22 °C, respectively, for 1 h. The fixation, permeabilization and staining of cells with rabbit anti-CapB antiserum (1 : 1500 dilutions in bovine serum albumin–phosphate-buffered saline (PBS), followed by HiLyte Fluor 488-labeled goat anti-rabbit IgG secondary antibody (AnaSpec, San Jose, CA) along with 4′-6-diamidino-2-phenylindole (DAPI) (Sigma, St. Louis, MO) was performed as described previously (Panicker et al., 2010). Slides were mounted in PBS-glycerol (pH 7.2) solution and observed under a Leica™ fluorescence microscope (Bannockburn, IL).
Electrophoretic mobility shift assay (EMSA)
cspD was PCR amplified using forward F-BamH1-CspD-Ant5-2 (5′-dGCCAGGATCCATGGCAACTGGCATC-3′) and reverse R-XhoI-CspD-Ant5-2 (5′-dGCCACTCGAGTTAGATTGGCTGAATG-3′) primers from Ant 5-2 cells and cloned into pET45b(+) plasmid (Novagen, WI). The CspD from Ant5-2 protein was expressed in E. coli BL21 (DE3) by inducing with 1 mM isopropyl-β-d-thiogalactopyranoside for 6 h and then purified using the His.Tag kit (Novagen). DNA-binding assay was performed by incubating purified CspD of Ant5-2 with 0.72 ng of 34-mer randomized oligonucleotide sequence (5′-dGCCAGAATTCTTAGATTGGCTGAATGTTCGAAGC-3′) and double stranded (ds)-DNA (PCR product) using fluorescence-based EMSA kit (Molecular Probes) according to the manufacturer's instructions.
Homology modeling and docking
The primary and secondary structure analyses were performed using the protparam tool on ExPASy server (Bairoch et al., 2005) and psipred (Jones, 1999), respectively. The tertiary structure of CspD from Ant5-2 was generated by the modeller software from hhpred alignments on HHpred servers (Soding et al., 2005; Eswar et al., 2006). The significance of the protein structure similarity was measured by TM-score calculated by T-align, a more sensitive method than root-mean-square-deviation (Zhang & Skolnick, 2005). The protein–protein docking was performed using the hex 5.1 software according to its manual (Ritchie & Venkatraman, 2010).
Growth kinetics and freeze tolerance of Ant5-2
During the initial 4 h, Ant5-2 cultures exhibited faster growth rates at 22 and 37 °C than at 15 °C (Fig. 1). Within 20 h of incubation at 15, 22 and 37 °C, the cultures grew exponentially and reached stationary phase. However, the culture at 37 °C exhibited a decline in growth after 48 h of incubation. In contrast, the cultures maintained at −1 and 4 °C did not show any significant cell proliferation for 24 and 4 h, respectively. Thereafter, the culture at 4 °C exhibited exponential growth and reached stationary phase within 72 h. A slow but steady exponential growth of the culture at −1 °C was noticed after incubation for 24 h. After one freeze–thaw cycle, increased survival was observed when Ant5-2 was exposed to 4 °C before freezing. The cultures transferred to 4 °C showed 94.1% survival compared with the 48.9% survival of cultures incubated at 22 °C (Supporting Information, Fig. S1).
Identification and expression of CspD in Ant5-2
The autoradiogram exhibited expression of a ∼7.28-kDa Csp (CspD) at all temperatures (Fig. S2). The immunoblot results showed that the expression of CspD in Ant5-2 was both time- and growth phase-dependent (Fig. 2a). Its expression increased at 37 °C and UVC exposure (Fig. 2b and c).
A 204-bp DNA fragment encompassing the entire ORF of the cspD gene (accession no. HQ873479) was PCR amplified. The deduced amino acid sequence exhibited 100% identity with the cold shock transcription regulator of J. lividum DSM 1522 and >98% sequence identity with the RNA chaperone, transcription antiterminator of Herminiimonas arsenicoxydans and with Csps from different bacteria belonging to class Betaproteobacteria (Figs 3 and Fig. S3). The CspD from Ant5-2 showed highest identity and similarity to E. coli CspE (67/83%) and E. coli CspD (56/74%) when compared with Csps from E. coli, B. subtilis and Pseudomonas sp. 30/3 (Fig. 3). PCR amplification of the cspA family of genes in Ant5-2 genomic DNA using CSPU5 and CSPU3 universal primers (Francis and Stewart, 1997) resulted in negative outcome (Fig. S4).
Phylogenetic analysis of CspD from Ant5-2
The phylogenetic analysis revealed that the Csps from bacteria belonging to the class Betaproteobacteria including CspD from Ant5-2 group together in a single cluster with low bootstrap value, i.e. 40% (Fig. 3). Csps from E. coli and B. subtilis also grouped separately with bootstrap values of 50% and 42%, respectively, with the exception of E. coli CspD, which aligned more closely to the Betaproteobacteria node with a low bootstrap value of 37%. DEAD-box RNA helicase containing CSD from Archaea Methanococcoides burtonii (AAF89099) was used as an outgroup,
Subcellular localization and DNA-binding activity of CspD in Ant5-2
Immunofluorescence staining was used to localize CspD using the anti-CapB rabbit-antiserum at different temperatures to determine the possible cellular role of CspD in Ant5-2. The cellular location of the nucleoid was confirmed by DAPI staining (Fig. 4a, c, and e). At 4 °C, a dense accumulation of the anti-CapB antibody immunoconjugated with the green Hilyte Fluor 488-labeled goat anti-rabbit IgG secondary antibody was observed in and around the nucleoid region (Fig. 4aand b). At 15 and 22 °C, the green fluorescence was dispersed in the cytosol as well as in the nucleoid region (Fig. 4c–f). The purified CspD protein from Ant5-2 (Fig. S5) exhibited binding affinity with single stranded (ss)-oligonucleotides with increasing concentration (Fig. 5) and not with dsDNA (PCR product) (data not shown).
Molecular structure of CspD from Ant5-2 in silico
Based on the amino acid residues and use of the homology modeling approach, the secondary and the tertiary structures of CspD from Ant5-2 indicated that the aromatic residues are conserved and three of the eight aromatic residues were docked on the nucleic acid-binding surface, F15 (F12), F17 (F20), and F28 (F31) (amino acid numbering on E. coli CspA is indicated in parentheses) (Feng et al., 1998). CspD from Ant5-2 has five basic and three acidic residues on the nucleic acid-binding surface. Its calculated theoretical isoelectric point (pI) was 5.6. Five β-strands and one α-helix were identified by the secondary-structure prediction (Fig. 6a). The solvent-exposed basic amino acids were K7 in β1 strand, K13 in L1, H30 in β3, K40 in L3 and K57 in L4 located on the nucleic acid-binding surface (Fig. 6b). The tertiary structure was designed with N. meningitidis CSD protein (Nm-Csp) (PDB reference: 3CAM) using the template provided by hhpred and modeller software (Soding et al., 2005; Eswar et al., 2006). The structure of the monomer of CspD from Ant5-2 consists of two subdomains of similar length separated by a long loop. Subdomain 1 includes β-strands 1–3 and subdomain 2 contains a β-ladder comprising strands 4 and 5 (Fig. 6a and b). The TM-score of the predicted structure was calculated to be 0.96738. It has been reported that the Nm-Csp form a dimer in the crystallographic asymmetric unit consisting of two five-stranded β-barrels (Ren et al., 2008). Because protein pairs with a TM-score >0.5 are mostly in the same fold (Xu & Zhang, 2010), we tested whether CspDAnt5-2 form a dimer-like Nm-Csp by docking monomer pairs with the hex 5.1 software (Ritchie & Venkatraman, 2010). The software predicted that the two modeled CspD monomers from Ant5-2 form a highly stable dimer like Nm-Csp with E-total −1278.08 and E-shape: −1107.25, the lowest interaction free energy and the site of interaction, respectively (Fig. 6c and d).
This study encompasses structural, functional and molecular phylogeny-based evolutionary analysis of the cold shock protein, CspD, from an Antarctic bacterium Janthinobacterium sp. Ant5-2 belonging to the CSD family of proteins in Betaproteobacteria. Janthinobacterium sp. Ant5-2 is a psychrotolerant bacterium, which can tolerate a wide range of temperature ranging between 37 and −1 °C as evident from the growth curve (Fig. 1). The slow growth rate at −1 °C could be possibly due to the decreased cellular metabolic activities at cold temperatures. It was found that Ant5-2 culture exposed to an intermediate temperature of 4 °C have a survival advantage upon freezing followed by thawing compared with the untreated cultures. In this study, the cspD gene sequence in Ant5-2 was identified by PCR amplification using oligonucleotide primers based on a similar gene sequences from its closest relative J. lividum (accession no. DQ074977). The complete genome sequences of yet another close relative of Ant5-2, i.e. Janthinobacterium sp. Marseille (accession no. NC009659) consist of only two copies of cspD, i.e. cspD1 (accession no. YP001353208) and cspD2 (accession no. YP001354206), and the fosmid library sequences of J. lividum (accession no. DQ074977) consist of one csp (accession no. AAZ39217). No other sequences similar to the other Csp proteins have been identified on the genome of these Janthinobacterium spp. Furthermore, we have shown that PCR amplification of the Ant5-2 genomic DNA using the universal oligonucleotide primers CSPU5 and CSPU3 targeting the csp family of genes resulted in negative result (Fig. S3). This suggests that CspD is most likely the only protein from CSD family of proteins present in Ant5-2.
The cold-shock responses in mesophilic and psychrophilic bacteria are found to be different, including the lack of repression of housekeeping protein synthesis and the presence of cold-acclimation proteins in psychrophiles (D'Amico et al., 2006). Many of the Csp family of proteins in mesophiles are constitutively expressed and function as cold acclimation proteins in psychrophiles (D'Amico et al., 2006). Our results demonstrate that unlike E. coli and like psychrophiles, the CspD in Ant5-2 is constitutively expressed at cold temperatures (Fig. 2a). Its subcellular localization in and around the nucleoid region at 4 °C and its binding affinity with ss-oligonucleotide probes that possessed randomized sequences indicated that CspD in Ant5-2 may bind through hydrophobic interaction to ssDNA without apparent sequence specificity (Figs 4 and 5), further supporting its possible role as a RNA chaperone at cold temperatures. In a previous complementation study, the Csp and CSD fold proteins of Archaea was able to function effectively to rescue a growth defect in a cold-sensitive E. coli strain, BX04 (Giaquinto et al., 2007). It illustrates that Csps and CSD fold proteins have retained a high degree of functional similarity. In addition we observed that CspD expression in Ant5-2 increased at 37 °C and upon UV exposure (Fig. 2b and c), and as described previously (Yamanaka et al., 2001; Kim & Wood, 2010), the cells also become elongated at 37 °C (data not shown), indicating that the CspD in Ant5-2 is a stress-inducible protein. Because CspD in Ant5-2 shares structural similarity with E. coli CspD, it might retain the same function as DNA replication inhibitor at 37 °C. It has also been reported that PprM, a homolog of Csp and a homodimer like E. coli CspD, is involved in the expression of many protein(s) that are important for the radioresistance of Deinococcus radiodurans (Ohba et al., 2009).
In this study, we have shown that the overall fold of the predicted CspD monomer from Ant5-2 did not closely resemble those of other bacterial cold-shock proteins. In both E. coli CspA and Bs-CspB, each chain is folded into an independent three-dimensional biological unit whereas the predicted Ant5-2 CspD dimer is composed of the N-terminal residues 1–36 from one chain and the C-terminal residues 37–67 from the other chain. The stable dimer prediction was performed with the help of the hex 5.1 docking software, which is considered to be a more reliable platform for ‘protein–protein’ compared with ‘protein–ligand’ docking. The predicted CspD dimer from Ant5-2 was formed by the exchange of two β-strands between protein monomers, but formed a symmetric unit of 2 five-stranded β-barrels unlike Nm-Csp that form two asymmetric five-stranded β-barrels. Despite differences, the predicted CspD dimer in Ant5-2 had significant structural similarities with the Nm-Csp and Bs-CspB dimers (Ren et al., 2008), sharing the same folds as that of monomeric Csps. This implies that it binds to ssDNA in a similar fashion. As evident from the electrostatic properties, the only DNA-binding region in the predicted tertiary structure of CspD dimer of Ant5-2 is the side of the β-barrel, which corresponds to the DNA-binding site in OB-fold proteins such as E. coli CspA and B. subtilis CspB. Although its theoretical pI is 5.6, the attractive potential for nucleic acids is created by the solvent-exposed basic amino acids located on the nucleic acid-binding surface. The solvent-exposed aromatic residues on the surface of these molecules also bind and melt nucleic acid secondary structure to facilitate transcription and translation at low temperatures (Phadtare et al., 2004).
The phylogenetic relationship of the CspD from Ant5-2 and Csps from three classes of phylum Proteobacteria, i.e. Betaproteobacteria, Gammaproteobacteria and Firmicutes, revealed that they distinctly form orthologous protein groups indicating a speciation event at each node except E. coli CspD. It should be noted that the low bootstrap values obtained in this study might be due to the highly conserved nature of CSD among different bacteria and also different Domains of life and obtaining high bootstrap values may require a relatively large amount of data (Berbee et al., 2000). Unlike Nm-Csp, CspD from Janthinobacterium sp. Ant5-2 is the only representative Csp from class Betaproteobacteria whose structure and function analyses illustrate its role in cold adaptation. Because N. meningitidis are commensal organisms that reside in the human upper respiratory tract, the role of Nm-Csp could not be described in context of cold adaptation (Ren et al., 2008).
In conclusion, we have described the growth characteristics, expression and overall structure and function of CspD in a psychrotolerant Antarctic bacterium Janthinobacterium sp. Ant5-2. Our principal finding is that CspD appears to undergo domain swapping to form stable dimeric structures and possess ssDNA-binding activity essential for cold and UV adaptations in extreme Antarctic environment.
We thank Col. (IL) J. N. Pritzker ARNG (Retired), Tawani Foundation (Chicago), for supporting the Tawani 2008 International Antarctic Scientific Expedition; Marty Kress, VCSI, Inc./NASA); NASA's Astrobiology program, Art Mortvedt, Selby Wilderness, Alaska; Dr Rasik Ravindra, Director, NCAOR; 2008–2009 Maitri and Novolazarevskaya Station staffs; Maitri Cdrs. A. Chaturvedi, and Dr P. Malhotra; geologist A. Swain; personnel at and Dr R. Fischer, Biology, UAB for the support. Thanks to R. B. Hoover (MSFC, NASA) for helping us with the identification of the strain.