ARTICLE

TOP ARTICLE LITERATURE CITED

In this paper, we report the widespread occurrence of the nucleic acid-binding cold shock domain (CSD) in plants and identify the first eukaryotic homologs that are nearly identical to bacterial cold shock proteins (CSP). Using Arabidopsis as a model system, we determined that its four unique CSD genes are differentially regulated in response to low temperature.

Prokaryotic response to low temperature has been extensively studied in Escherichia coli and is accompanied by a spectacular accumulation of nucleic acid-binding CSPs (Graumann and Marahiel, 1998; Yamanaka et al., 1998; Bae et al., 2000). CspA, the most prominent of the nine-member E. coli CSP family, accumulates up to 10% of total proteins during cold stress (Jiang et al., 1997). The three-dimensional structure of E. coli CspA forms a five-stranded -barrel structure (Newkirk et al., 1994; Schindelin et al., 1994) that contains two consensus RNA-binding motifs (RNP1 and RNP2), which facilitate nucleic acid recognition/binding (Schroder et al., 1995). CspA has been hypothesized to prevent RNA secondary structure formation (Jiang et al., 1997), thereby enhancing translation at low temperature.

The CSD, which encompasses bacterial CSPs, is the most conserved nucleic acid-binding domain and is capable of binding single-stranded DNA/RNA and double-stranded DNA (Graumann and Marahiel, 1996). The CSD is proposed to be an ancient structure that was present before the divergence of prokaryotes and eukaryotes (Graumann and Marahiel, 1998). It is interesting to note that cyanobacteria lack CSD proteins, however, they contain RNA-binding domain proteins (RBD; Sato, 1995). RBD proteins are thought to have evolved a similar three-dimensional functional surface for nucleic acid binding through convergent evolution (Graumann and Marahiel, 1996) and may have replaced CSD proteins (Graumann and Marahiel, 1998). Bacterial CSP homologs show high homology to the well-characterized eukaryotic Y-box proteins (Wolffe, 1993; Bae et al., 2000), which contain an N-terminal CSD and C-terminal auxiliary domains that facilitate a broad range of in vivo functions such as RNA masking and transcriptional and translational regulation (Sommerville, 1999). Surprisingly, within the plant kingdom, only four proteins are documented to contain a CSD. Arabidopsis (AtGRP2 and AtGRP2b), tobacco (Nicotiana tabacum; NtGRP; Kingsley and Palis, 1994), and wheat (Triticum aestivum; WCSP1; Karlson et al., 2002) contain an N-terminal CSD in addition to Gly-rich domains that are interspersed by CX₂CX₄HX₄C (CCHC) retroviral-like zinc fingers. As noted by Guy (1999), Arabidopsis and tobacco CSD proteins were not studied in any extent for relation to low temperature or nucleic acid binding. With our recent entry to this class (WCSP1), we provided the first evidence for cold regulation of a plant CSD protein and functionally characterized its nucleic acid-binding activity (Karlson et al., 2002).

In the present study, a comparative (tBLASTn) GenBank expressed sequence tag (EST) database search was conducted in an effort to identify novel plant sequences that contain CSDs. Highly conserved CSDs were identified within 19 genera that represent lower plants, monocots, dicots, and woody plants. Multiple homologs were found within individual species, which is indicative of small gene families. ESTs were placed into two groups based upon presence (Type-I) or absence (Type-II) of C-terminal auxiliary domains and multi-aligned with ClustalX software (Fig. 1, A and B, respectively). Because of the limited number of high-quality sequence data and incomplete open reading frames (ORFs), only putative amino acid sequences from N-terminal CSDs were used for multiple sequence alignment and phylogenetic analysis (Type-I; Fig. 1A). It is important to note that high-quality data from several Type-I ESTs extended well beyond the CSD, revealing Gly-rich domains and variable quantities of C-terminal CCHC zinc fingers (not shown). In Arabidopsis, AtGRP2 (At4g38680) and AtGRP2b (At2g21060) contain two Gly-rich regions and two CCHC zinc fingers, however, seven CCHC zinc fingers are interspersed within Gly-rich regions of two undesignated proteins (At4g36020 and At2g17870). Interestingly, C-terminal CCHC zinc fingers were not found in the lower plant EST sequences (Chlamydomonas reinhardtii and Ceratopteris richardii), and their Gly-rich domain composition appears to be different from higher plant CSD proteins.

View larger version (122K): [in this window] [in a new window]   Figure 1.   Multiple alignment of deduced amino acid sequences of CSD homologs. A, Alignment of Type-I N-terminal CSDs encoded by EST sequences. Four previously characterized plant CSD proteins (AtGRP2, AtGRP2b, NsGRP2, and WCSP1) and E. coli CspA are included as references. Note that these ESTs are not complete ORFs and include only N-terminal CSDs. B, Alignment of Type-II putative amino acid sequences encoded by complete ORFs. Note that these are nearly identical in size and homology to prokaryotic CSPs. EST sequences are listed with an abbreviated genus name, species name, and corresponding GenBank accession numbers. The abbreviations and corresponding genera are: A, Arabidopsis; B, Brassica; C, Ceratopteris; C, Chlamydomonas; G, Gly; G, Gossypium; H, Hordeum; L, Lycopersicum; L, Lotus; M, Medicago; M, Mesembryanthemum; O, Oryza; P, Pinus; S, Solanum; S, Sorghum; S, Secale; T, Triticum; and Z, Zea. Identical conserved consensus amino acids are indicated by asterisks, whereas conserved substitutions are indicated by colons and periods. Consensus regions corresponding to the five -sheets of E. coli CSPs are overlined, and critical core hydrophobic residues are circled in red. Homology plots are illustrated below multiple alignments.

Of outstanding interest was the discovery that plants also contain complete ORFs that encode putative CSD proteins that are nearly identical to prokaryotic CSPs in size and sequence (Type-II; Fig. 1B). Contrary to prokaryotes, all eukaryotic CSD proteins characterized thus far contain additional C-terminal auxiliary domains such as Arg-Gly repeats, CCHC zinc fingers, basic/aromatic islands, and additional CSDs (Salvetti et al., 1998; Graumann and Marahiel, 1998; Sommerville, 1999). It is important to note that the wheat and barley (Hordeum vulgare) ESTs encode putative proteins solely composed of a CSD. Furthermore, Type-II ESTs were detected only within wheat and barley and are not within the Arabidopsis genome.

Previous three-dimensional structural analyses identified residues that are critical for hydrophobic core formation in E. coli CSP five-stranded -barrel CSD structure (Yamanaka et al., 1998). As designated by red circles, these residues are almost completely conserved within all identified EST sequences (Fig. 1). Within Type-I ESTs, Brassica rapa contained a single exceptional amino acid in the fourth -strand region where Phe is present instead of Val. A major exception occurred within the third -strand region, where Leu was present in plant CSDs, whereas, Val is present in the E. coli consensus (Yamanaka et al., 1998; Fig. 1A). Because of the hydrophobic nature of Leu, it is possible that this highly conserved substitution does not compromise three-dimensional structure. Similar to bacteria, Type-II EST sequences contained a conserved Val residue in this same position (Fig. 1B). Because of the conservation of critical hydrophobic core residues, it is likely that the three-dimensional structure is conserved within both Type-I and II plant CSDs, thereby rendering them competent for putative nucleic acid-binding functions.

Yamanaka et al. (1998) previously reported that the loop region between 3- and 4-strand was the most diverse among E. coli CSPs and may determine specific in vivo function. Unlike bacteria, eukaryotic homologs typically contain four additional basic residues within this same region. However, plant CSDs are moderately conserved between 3- and 4-strands and do not contain additional residues (Fig. 1). This observation is similar to Caenohabditis elegans LIN-28, a eukaryotic CSD protein that also contains two C-terminal CCHC zinc fingers (Yamanaka et al., 1998). Type-I EST sequences showed the highest diversity within the N terminus and within putative turn regions between -strands 1-2, 2-3, and 4-5, the significance of which is unknown.

Phylogenetic analysis of the novel plant CSDs revealed general evolutionary trends, where monocots, dicots, and closely related genera (i.e. Brassica spp./Arabidopsis) and species (i.e. T. aestivum/Triticum turgidum) were similarly grouped. Type-II ESTs, which encode a complete CSD protein, were the most closely related to bacterial CspA (Fig. 2). Because of the limitations of EST sequence data, it is critical to note that the phylogenetic tree was generated as a comparison of N-terminal CSDs.

View larger version (32K): [in this window] [in a new window]   Figure 2.   Phylogenetic analysis of plant CSD homologs. The multiple alignment was analyzed by ClustalX with a bootstrapped neighbor-joining method and displayed with TreeViewPPC software. The phylogenetic tree was rooted with E. coli CspA as the outgroup, and individual branch lengths indicate evolutionary distance of the sequences. Two major groupings were detected within plant CSDs: CSDs that contain additional C-terminal auxiliary domains (Type-I) and sequences that are composed solely of a CSD (Type-II).

Using Arabidopsis as a model plant, we investigated the response of a complete plant CSD gene family to low temperature stress. Genome data analysis confirmed that Arabidopsis contains four unique CSD proteins (AtGRP2-At4g38680, AtGRP2b-At2g21060, At2g17870, and At4g36020). Plants were grown under continuous illumination in a controlled growth chamber (25°C) and were sampled before and 4, 12, 24, and 48 h subsequent to their transfer to a separate pre-equilibrated growth chamber (4°C). Total leaf RNA was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA), and 1 µg was used as a template for semiquantitative reverse transcriptase (RT)-PCR as described by (Cheng et al., 2002). Gene-specific primers were used to amplify individual genes from synthesized cDNA. Within the tested time frame of low temperature treatment, the transient increase of At2g17870 was similar to the positive cold-responsive control (Cor47; Gilmour et al., 1992; Fig. 3). These data contrasted the slower increase of AtGRP2 and At4g36020 and the apparent down-regulated response of AtGRP2b. Our RT-PCR data are consistent with E. coli CSPs, where individual CSPs are regulated differentially in response to low temperature (Yamanaka et al., 1998).

View larger version (31K): [in this window] [in a new window]   Figure 3.   Semiquantitative RT-PCR analysis of four Arabidopsis CSD genes in response to cold treatment. Total RNA was extracted from leaves harvested from plants before and subsequent to 4, 12, 24, and 48 h of 4°C treatment and used as template for gene-specific amplification of AtGRP2, AtGRP2b, At4g36020, At2g17870, Cor47, and AAc1. Inversed images from equally loaded ethidium bromide-stained gels revealed that AtGRP2, At4g36020, and At2g17870 increase in response to cold, whereas AtGRP2b is down-regulated in the same time course. Cor47 and actin 1 (AAc1) were used as positive controls for low temperature and constitutive responses, respectively.

Unlike heat shock, conserved responses to low temperature stress are largely unknown within prokaryotes and eukaryotes. It is interesting to consider the structural conservation of CSD proteins within prokaryotes and eukaryotes and to assess whether this is because of a convergent role for nucleic acid-binding function or for a similar in vivo functional role in relation to low temperature stress. Characterization and functional analyses of newly identified homologs will allow us to assess the importance of the CSD in plants response to low temperature stress. Our previous functional analysis of WCSP1 (Karlson et al., 2002) and the high conservation of critical amino acids within the CSD ESTs supports the supposition that plant CSDs are capable of binding nucleic acids. The responsiveness of WCSP1 and multiple Arabidopsis CSD genes to low temperature support the notion that common mechanisms for cold adaptation may exist within plants and bacteria.