| Edited by Kiichi Fukui. Minoru Murata: Corresponding author. E-mail: mmura@rib.okayama-u.ac.jp |
Telomeres are specialized nucleoprotein complexes of eukaryotic chromosome ends that are essential for chromosome stability and integrity (Blackburn, 1991; Greider, 1996; Muniyappa and Kironmai, 1998). The telomere consists of telomeric DNA repeats and associated proteins. The telomeric DNA contains specific G-rich repeat sequence, such as TTTTGGGG in Oxytricha, TTGGGG in Tetrahymena, TTAGGG in most vertebrates (Zakian, 1995) and TTTAGGG in most plants (Ganal et al., 1991; Kilian and Kleinhof, 1992; Richard and Ausubel, 1988; Wu and Tanksley, 1993). The G-rich strand of repeat sequences terminates as a single-stranded 3’ overhang (Makarov et al., 1997; Wright et al., 1997), and is thought to fold back into the duplex region of telomeric repeats to form a loop named “t-loop” in vivo (Griffith et al., 1999; Greider, 1999).
Telomere binding proteins are known to play important role in telomere functioning, and have been classified into two groups. The first group is the proteins interacting with double-stranded telomeric DNA, e.g. Rap1p of budding yeast (Lustig et al., 1990), Taz1p of fission yeast (Copper et al., 1997) and human TRF1 and TRF2 (van Steensel and de Lange, 1997; van Steensel et al., 1998), all of which have been shown to have Myb-like DNA-binding domains. The second group binds to single-stranded G-rich overhangs at the ends of chromosome and forms protective caps. In Oxytricha nova, the telomere end-binding protein (TEBP), a heterodimeric protein composed of 56 kDa α and 41 kDa β subunits, has been shown to bind to single-stranded telomeric DNA to form ternary structure (Fang and Cech, 1993; Gottschiling and Zakian, 1986; Gray et al., 1991). The α subunit has oligosaccharide/oligonucleotide-binding folds (OB-folds) for DNA binding. A monomeric protein similar to the α subunit of Oxytricha has also been reported to bind to single-stranded telomeric DNA in Euplotes crassus (Wang et al., 1992). The budding yeast Cdc13 also has been identified as a protein that binds to the most distal end of the chromosome and recruits telomerase (Evans and Lundblad, 1999; Gottschiling and Zakian, 1986; Penrock et al., 2001).
The protection of telomere 1 (Pot1) protein was identified in fission yeast and human based on the similarity to N-terminal regions of various single-stranded telomeric DNA binding proteins (Baumann and Cech, 2001). The fission yeast Pot1 (spPot1) is essential for protection of chromosome ends, and its deletion causes rapid telomere degradation and chromosome circularization. The human POT1 (hPOT1) has similar features, and is shown to be recruited to telomere by association with TRF1 (Baumann et al., 2002).
In plants, double-stranded telomeric DNA binding proteins have been isolated from Arabidopsis (Chen et al., 2001, Hwang et al., 2001), rice (Yu et al., 2000) and tobacco (Yang et al., 2003), and shown to bind their telomeric DNA via Myb-like DNA binding domain. In addition, new two telomere-binding proteins were recently found in A. thaliana (Schrumpfova et al., 2004). Both proteins had a Myb-like domain at the N-terminus, unlike many other telomere-binding proteins that have the domain at the C-terminus. These two Myb-like DNA binding proteins were shown to bind not only to double-stranded but also to single-stranded telomeric DNA in vitro. The single-strand specific DNA binding protein was found in the nuclear extract of Arabidopsis by affinity chromatography (Kwon et al., 2004), and in other species, e.g. rice (Kim et al., 1998), tobacco (Hirata et al., 2004), mung bean (Lee et al., 2000) and soybean (Kwon et al., 2004). However, no POT1-like protein was found.
Two POT1-like proteins (AtPOT1-1 and AtPOT1-2) have been identified in the Arabidopsis database based on the N-terminal similarities (Baumann and Cech, 2001), but neither of them has been characterized yet. In this study, therefore, we cloned the full-length cDNAs and characterized their gene structures and expression.
Plants of Arabidopsis thaliana (L.) Heynh. ecotype Columbia (Col-0) grown in a greenhouse were used to isolate RNA, crude protein and nuclei. The Arabidopsis cell suspension used in this study was derived from roots of A. thaliana ecotype Columbia, and maintained at 22°C in modified Murashige and Skoog medium supplemented with 2.0mg/ml 2,4-D (Mathur et al., 1998).
To amplify cDNA of two putative POT1-like proteins (AtPOT1-1 and AtPOT1-2), we designed two sets of primers for AtPOT1-1 (forward primer: 5’-GGGAGATGGCGAAGAAGAGAG-3’ and reverse primer: 5’-TTAATGAAGTAGTCTAGTACCAAA-3’) and for AtPOT1-2 (forward primer,: 5’-ATGGAGGAGGAGAGAAGAGATC-3’, and reverse primer: 5’-TCATGAAGCATTGATCCAAGTT-3’). Reverse transcription mediated PCR (RT-PCR) was made with OneStep RT-PCR Kit (QIAGEN) to obtain cDNA from total RNA as a template. Total RNA was isolated from buds, flowers, leaves, roots, and stems of Columbia plants, and cell suspensions using an RNeasy Plant minikit (QIAGEN).
To decide the sequences of full-length cDNA, 5’- and 3’-RACE-PCR were performed using a SMART RACE cDNA Amplification Kit (BD Bioscience), according to the manufacturer’s instruction. The primers used were as follows: AtPOT1-1 (5’-CGGTTGGAAAGCTGAGCAGTACAT-CTCTTG-3’ for 5’-RACE, 5’-CAGAATAGATGCTGGGGGTGCATCAGTC-3’ for 5’-nested PCR and 5’-TTCCTTTGCCTTGTTCGAAGGTGTGG-3’ for 3’-RACE) and AtPOT1-2 (5’-GGTTCAGAAGCTTCACGTGTTGACCA-3’ for 5’-RACE and 5’-AACTGAGATGCCACCCTGCAACATC-3’ for 3’-RACE).
All PCR products were purified with QIAquick PCR Purification Kit (QIAGEN), and cloned into a pGEM-T easy vector (Promega). The plasmid DNAs were then purified using MagExtractor Plasmid (TOYOBO), and their insert sequences were determined with Big Dye Terminator (Applied Biosystems) and ABI PRISM 310 Genetic Analyzer (PE Biosystems). The sequence data have been deposited in DDBJ/EMBL/Genebank database (Accession numbers NM_126547 and AB006700).
The total RNA (25 μg each for AtPOT1-1 and 10 μg each for AtPOT1-2) isolated from buds, leaves and cell suspensions were separated on 1% agarose gel containing 18% formaldehyde, blotted to positive-charged nylon membranes (BIODYNE PLUS, PALL), and then fixed with a UV Crosslinker (Pharmacia Biotech). Plasmid DNA containing a full-length cDNA was labeled with DIG High Prime (Roche) and purified with AutoSeq G-50 colums (Amersham Biosciences). Pre-hybridization and hybridization was performed using DIG-Easy Hyb (Roche) at 50°C. For detection, anti-Digoxigenin-AP, Fab fragments (Roche) and CDP-Star ready-to-use (Roche) were used. Luminescent signals were detected and analyzed by LAS 1000 plus (FUJI Film).
For protein extraction, leaves of 4 week-old plants were harvested, ground with a pestle and mortar in liquid nitrogen. The powdered leaves were then transfered into a 1.5 ml centrifuge tube with the sample buffer [10% glycerol, 5% 2-mercaptoethanol, 2.3% sodium lauryl sulfate (SDS), 62.5 mM Tris-HCl, pH 6.8, 0.01% bromophenol blue (BPB)] and vortexed. After boiling for 5 min, it was centrifuged for 5 min at RT at 15,000 rpm, and the supernatant was collected. For nucleus extraction, the powdered leaves were resuspended in 20 ml MSG buffer [0.5 M hexyleneglycol, 5 mM piperazine-1,4-bis (2-ethanesulfonic acid) (PIPES) (pH7.0), 5 mM MgCl2, 5 mM KCl, 0.2 M sucrose, 0.4% Triton X-100, 7 mM 2-mercaptoethanol, 0.4 mM phenylmethanesulfonyl fluoride (PMSF)] and filtered through the Miracloth (Calbiochem). The sample was then centrifuged for 10 min at 4°C at 100 xg and the supernatant was transferred into a new tube. The pellet was resuspended again with 20 ml MGS buffer, and centrifuged again. The supernatant was transferred into the tube containing the first supernatant. The mixed supernatant was loaded on 40% and 20% Percoll (Amersham Biosciences) in MGS buffer, centrifuged for 90 min at 4°C at 5,000 xg. The fraction under 40% Percoll was collected, and resuspended with an equivalent volume of MGS buffer. The sample was centrifuged for 10min at 4°C at 5,000 xg to pellet the nuclei. The pellet was then resuspended in 1 ml of MGS buffer.
The antibody was made by JBioS (Saitama, Japan) against the peptide (SLIGIVLEQREPKQC) synthesized based on the N-terminal 25–39 amino acid sequence of AtPOT1-1. The synthetic peptide was inoculated into rabbit, and the antibody was used after affinity purification.
The protein samples were boiled, mixed with loading buffer (125mM Tris-HCl, pH6.8, 4% SDS, 20% glycerol, 5% 2-mercaptoethanol) and loaded on a 12.5% SDS-polyacrylamide gel. Proteins were electro-blotted to a membrane (Immobilon-P, MILLIPORE) at 4°C at 1.5mA / cm2 for 1.5h. For detection, anti-rabbit IgG, horseradish Per-oxidase linked F(ab’)2 fragment (from donkey) (Amersham Biosciences) and ECL Western Blotting Detection Reagent (Amersham Biosciences) were used. Luminescent signals were detected and analyzed by LAS 1000 plus (FUJI Film).
By an independent search for the proteins homologous to hPOT1, we found two candidates, AAD29059 and BAB08953, which correspond to the previously reported, NP_178592 and NP_196249, respectively (Baumann et al., 2002). The former protein was encoded by the gene located on chromosome 2 (At2g05210), which was predicted to consist of seven exons and six introns. The putative coding sequence was 1095 bp in length. The gene of the latter protein was located on chromosome 5 (At5g06310), and consisted of ten exons and nine introns. The coding sequence was 1392bp in length. These two genes were designated here as AtPOT1-1 and -2, respectively (Fig. 1).
 View Details | Fig. 1. Predicted and determined gene structures and splicing variants of AtPOT1-1 (A) and AtPOT1-2 (B). The exons unpredicted are indicated as shaded boxes. Black boxes show 5’- and 3’-untranslated regions (UTRs). Asterisks show a putative stop codon. Open arrowheads indicate primer positions used for RT-PCR, and black arrowheads are the primers used for 5’- and 3’-RACE. |
To confirm the expression of AtPOT1-1 and -2 genes, we performed RT-PCR with total RNA from buds, flowers, leaves, roots, stems and cell suspension cultures as templates (Fig. 2). For both genes, distinct DNA amplification was observed in all tissues examined. Although the single 1095 bp fragment for AtPOT1-1 and the 1392 bp for AtPOT1-2 were expected to be amplified, two bands for both genes appeared in all tissues examined (Fig. 2). This result suggested the mis-prediction of the gene structures and/or occurrence of alternative splicing.
 View Details | Fig. 2. RT-PCR of AtPOT1-1 and AtPOT1-2. The RNA samples isolated from buds (1), flowers (2), leaves (3), roots (4), stems (5) and cell suspensions (6) were used as templates. |
To obtain full-length cDNA of AtPOT1-1 gene, we performed 5’- and 3’- RACE using total RNA isolated from cell suspensions as a template. Sequencing of amplified DNA fragments indicated that three different transcripts are being produced from the gene by alternative splicing. Although only seven exons and six introns were predicted from the database, the gene was thought to be composed of 12 exons and 11 introns from cDNA sequence of one (I) of the three transcripts (Fig. 1A). The first two exons were found in the 5’-untranslated region (UTR) and the start codon occurred in the middle of exon 3. Three other new exons (exon 6, 7 and 9) were identified in intron 3 and 4 of the predicted form. In the other two splicing variants (II and III), exons 1, 2 and the first half of exon3 in the 5’-UTR were not detected, but the sequences of exon 4, 5 and 8 to 11 were identical among the three variants. For exon 12, there was no difference between variant I and II, but variant III cDNA lacked the latter part of the sequence that included 3’-UTR. Striking differences among the three variants occurred between exons 6 and 7. In variant II, intron 6 was skipped, and the large exon, which is composed of exons 6, 7 and intron 6, was formed. In variant III, on the other hand, the donor splicing site (AG) of intron 6 shifted 28 bp toward 3’ direction, resulting in a slightly longer exon 4.
Variant I, which had a 2054-bp cDNA sequence, was the longest among the three splicing variants. Since the stop codon appears in the middle of exon 12, the coding region, 1404 bp in length, putatively encodes a 54 kDa protein with 467 amino acids. In variant II, a mis-splicing at intron 6 made the transcript longer. Since the primers for RT-PCR were designed between exons 3 and 12, the upper bands observed in Fig. 2 were thought to originate from variant II. However, the mis-splicing gave rise to a new stop codon in the combined exon (exon 6 + intron 6 + exon 7), so the coding sequence is only 598 bp in length, which putatively encodes a 23 kDa protein of 199 amino acids. Variant III and variant I were similar in sequence and size, and could not be distinguished by RT-PCR experiment (Fig. 2). The longer size of exon 6 in the variant III caused a frame shift, and a new stop codon appeared in exon 7. As a result, a 23 kDa protein of 199 amino acids was deduced from the 598 bp coding sequence.
Since the RT-PCR result (Fig. 2) suggested that the AtPOT1-2 gene also produces two splicing variants, we performed 5’- and 3’-RACE were performed using total RNA isolated from cell suspensions as a template and a set of primers designed from the predicted exons 5 and 6 of AtPOT1-2 (Fig. 1B). Sequencing analysis of the amplified DNA fragments indicated the presence of two splicing variants as expected. The gene was thought to be composed of ten exons and nine introns as shown in the database (Fig. 1B). However, exon 6 was 27bp shorter than predicted. As a result, the splicing variant I had a 1365 bp coding region in the 1499 bp cDNA, which encodes a putative 54 kDa protein of 455 amino acids (Fig. 3). In variant II, intron 3 was not spliced, resulting in a longer transcript (1655 bp in length). This result indicates that the large product observed in the RT-PCR experiment corresponds to variant II, whereas the smaller band to variant I. However, since the stop codon appears in intron 3, the amino acid sequence deduced from variant II is only 179 in length, much shorter than that of variant I. The RACE experiments indicated that both variants have the same size and sequence of 5’- and 3’- UTRs.
 View Details | Fig. 3. Alignment of deduced amino acid sequences from database of fission yeast Pot1, human Pot1 and two Arabidopsis POT1-like proteins. Amino acid residues of the homologs, identical and similar to those of Arabidopsis POT1-like proteins are shown in white on a black background and in black on a shaded background, respectively. – indicates a gap in introduced to maximize alignment. A white bar indicates putative OB fold shown by EBI Protein Database. A possible telomere end-binding protein α subunit, central domain is indicated as a black bar in AtPOT1-1, and a shaded bar in AtPOT1-2. |
To examine the transcription levels of AtPOT1-1 and -2, we performed northern blot hybridization for the RNA samples from floral buds, leaves and suspension cells (Fig. 4). As shown in the RT-PCR experiment (Fig. 2), two distinct bands were observed for both genes, indicating that the two splicing variants (I and II) of each gene are equally transcribed in all tissues examined. Variant III of AtPOT1-1 seemed to be transcribed at a low frequency, because fewer clones were obtained than the other two variants. The transcription level of AtPOT1-1 gene was much higher in floral buds than in leaves and suspension cells. However, such differentiation was not found in the AtPOT1-2 transcription.
 View Details | Fig. 4. Northern hybridization of AtPOT1-1 (A) and AtPOT1-2 (B). RNA of 25 μg for AtPOT1-1 and 10 μg for AtPOT1-2 from buds (1), leaves (2) and cell suspension (3) was used. The probes used were DIG-labeled AtPOT1-1 and AtPOT1-2 cDNA. |
Since three splicing variants were detected for AtPOT1-1, we examined whether they are translated into peptides or not. Western blot analysis with the antibody against the N-terminal amino acids of AtPOT1-1 revealed a single band in the crude protein sample from leaves, the molecular weight (MW) of which was almost equal to that expected from the cDNA sequence of variant I (Fig. 5). In the sample from isolated nuclei, two small bands were detected in addition to the large band of 50 kDa that appeared in the leaf sample. However, the MW of the two bands were larger than expected from the sequences of variant II and III (23 kDa), so they might have been post-translationally modified. Since no antibody was successfully raised, the translation of the AtPOT1-2 gene has not been investigated. We examined the subcellular localization of AtPOT1-1 was investigated by immunofluorescence labeling with anti-AtPOT1-1, but detected no signals (no data shown).
 View Details | Fig. 5. Western blotting analysis with the antibody against the N-terminal amino acids of AtPOT1-1. Crude protein samples from leaves (1) and from isolated nuclei (2) were loaded. |
Two POT1-like genes were found in the Arabidopsis genome, and we cloned their cDNAs to determine the gene structures and expression. The AtPOT1-1 gene is thought to consist of 12 exons and 11 introns, but three splicing variants were detected in all tissues examined. The antibody against the N-terminal 15 amino acids recognized three different bands in the nuclear extract of leaves by western blotting. This indicates that the three variants are being translated, but the largest one (ca. 54kDa) must be the major translated peptide. The other gene, AtPOT1-2, also had two splicing variants. Although their translated peptide has not been confirmed yet, the amino acid sequence of AtPOT1-2 (variant I) is similar to that of AtPOT1-1 (I) at the C-terminal as well as the N-termini (Fig. 3), indicating functional similarity between those two proteins.
The telomere end-binding proteins have been identified in various organisms, e.g. budding yeast Cdc13 (Lin and Zakian, 1996), fission yeast and human Pot1 (Bauman and Cech, 2001) and chicken Pot1 (Wei and Price, 2004), since the protein was first found in Oxtricha nova (Gottschling and Zakian, 1986). This protein family had a single oligosaccharide/oligonucleotide (OB) fold in the N-terminal region, which is a broadly conserved among eukaryotes as a structural element for binding single-strand telomeric termini (Mitton-Fry et al., 2002). The Protein Database in EBI (http://www.ebi.ac.uk/) revealed that the single OB occupies a position between 9 and 164 amino acids in all variants (I, II and III) of AtPOT1-1 and between 10 and 151 amino acids in AtPOT1-2 (I and II variants) (Fig. 3). In addition, we found the regions of 40–145 amino acids in AtPOT1-1 and 38–144 in AtPOT1-2 to be a possible telomere end-binding protein α-subunit, central domain. These results suggest that both the AtPOT1-1 and AtPOT1-2 proteins belong to the same family as other POT1 proteins do.
Like AtPOT1-1 and -2 genes, alternative splicing and translation of C-terminal truncated forms also occur in human Pot1 proteins (Baumann et al., 2002). Among the five splicing variants, four are ubiquitously expressed, but the remaining one is leukocyte-specific. However, this sort of differential expression was not detected in this study for AtPOT1 proteins. In human Pot1, the variant 2 that lacks the C-terminal half of variant 1 was shown to have a higher affinity to the telomeric DNA in vitro than other variants. This differential affinity remains unsolved in AtPOT1.
Immunofluorescence with the antibody against the N-terminal 15 amino acids of AtPOT1-1 was carried out to Arabidopsis and tobacco cultured cells to examine the subcellular localization, but detected no signals. This might have been caused by the low amount of the protein and/or the position at the bottom of the telomere, because in human, actively growing cells expressing V-5-epitope tagged hPot1 were used for this sort of experiment (Baumann et al., 2002). The t-loops are known to be notoriously fragile, and gentle chromatin extraction is required for maintaining the structure (Murti and Prescott, 1999). Special care might be necessary for visualizing the POT1-like proteins, particularly in plants that have rigid cell walls.
|
Index