Expression of the telomeric repeat binding factor gene NgTRF1 is closely coordinated with the cell division program in tobacco BY-2 suspension culture cells.

Telomeres are vital for preserving chromosome integrity during cell division. Several genes encoding potential telomere-binding proteins have recently been identified in higher plants, but nothing is known about their function or regulation during cell division. In this study, we have isolated and characterized a cDNA clone, pNgTRF1, encoding a putative double-stranded telomeric repeat binding factor of Nicotiana glutinosa, a diploid tobacco plant. The predicted protein sequence of NgTRF1 (Mr = 75,000) contains a single Myb-like domain with significant homology to a corresponding motif in human TRF1/Pin2 and TRF2. Gel retardation assays revealed that bacterially expressed full-length NgTRF1 was able to form a specific complex only with probes containing three or more contiguous telomeric TTTAGGG repeats. The Myb-like domain of NgTRF1 is essential, but not sufficient, to bind the telomeric repeat sequence. The glutamine-rich extreme C-terminal region, which does not exist in animal proteins, was additionally required to form a specific telomere-protein complex. The dissociation constant (Kd) of the Myb motif plus the glutamine-rich domain of NgTRF1 to the two-telomeric repeat sequence was evaluated to be 4.5 +/- 0.2 x 10-9 m, which is comparable to that of the Myb domain of human TRF1. Expression analysis showed that NgTRF1 gene activity was inversely correlated with the cell division capacity of tobacco root cells and during the 9-day culture period of BY-2 suspension cells, while telomerase activity was positively correlated with cell division. In synchronized BY-2 cells, NgTRF1 was selectively expressed in G1 phase, whereas telomerase activity peaked in S phase. These findings suggest that telomerase activity and NgTRF1 expression are differentially regulated in an opposing fashion during growth and cell division in tobacco plants. The possible physiological functions of NgTRF1 in tobacco cells are also discussed.

Telomeres are specialized nuclear protein complexes located at the ends of linear eukaryotic chromosomes. They are essential and functional components for preserving chromosome integrity and for protection from exonucleolytic degradation and end-to-end fusion with other chromosomes (1,2). Telomeres are also crucial for chromosome organization in the nucleus, especially during cell division. The DNA sequence of telomeres is highly conserved in most eukaryotes and consists of tandem repeats of short, G-rich sequence elements, such as TTAGGG in vertebrates (3) or TTTAGGG in higher plants (4). This G-rich strand extends beyond the complementary C-rich strand and terminates as a singlestranded 3Ј-overhang in many divergent organisms (5)(6)(7). Telomeres are synthesized and maintained by telomerase, a ribonucleoprotein complex that contains a specialized reverse transcriptase activity that uses its own RNA subunit as a template (5,8,9). In mammalian cells, telomerase activity is tightly associated with cell proliferation, de-differentiation, senescence, and immortalization. For example, in humans, telomerase is highly expressed in tumor cells and germ-line and embryonic cells, but is not detected in most normal somatic tissues (10). Lack of telomerase activity in human somatic cells results in telomere shortening during differentiation and aging, while stable maintenance of telomere length occurs in germ-line and tumor cells, which have unlimited cell division capacity (10,11). These observations suggest that telomerase activity and telomerase-mediated stabilization of telomere length are intimately tied to the proliferative abilities of cells (12,13).
In addition to telomerase activity, associations between the telomere repeat sequences and specific binding proteins appear to be required for the integrity and proper functions of telomeres. Studies of telomere chromatin structure have suggested that telomeres are packaged into specialized nucleoprotein complexes (14 -16). Protein components of the telomere complex have been identified and characterized in several organisms, including ciliates, yeast, and humans. The telomerebinding proteins can be divided into two distinct groups. Members of the first group bind specifically to the single-stranded 3Ј-extension at the extreme termini of telomeres, which are necessary for chromosome capping and telomerase regulation (17)(18)(19). The other group of proteins interacts with the doublestranded telomeric repeats. For example, Rap1p from Saccharomyces cerevisiae specifically binds to yeast duplex telomeric DNA and plays a role in the regulation of telomere length maintenance (20,21). Taz1p was identified in Schizosaccharomyces pombe in a one-hybrid screen using double-stranded telomeric DNA as a target and found to be involved in telomere length regulation, repression of telomere adjacent genes, and the interactions between telomeres and the spindle pole body during meiotic prophase (22)(23)(24). In humans, two distinct Mybrelated proteins, TRF1 1 and TRF2, have been identified as double-stranded telomeric DNA-binding proteins (25)(26)(27). TRF1 is a suppressor of telomere elongation and involved in the negative feedback mechanism that stabilizes telomere length by inhibiting telomerase at the ends of individual telomeres (28). A dominant-negative allele of TRF2 induced end-to-end chromosome fusions in metaphase and anaphase cells, indicating that TRF2 plays a key role in the protective activity of telomeres in human cells (29). Recently, it has been reported that TRF2, along with TRF1, acts as a negative regulator of telomere length (30). Another telomeric protein, Pin2, was identified in HeLa cells (31). Pin2 is identical in sequence to TRF1, except for an internal deletion of 20 amino acids, suggesting that these two proteins may be derived from the same gene, PIN2/TRF1 (31). Most recently, the Pin2-interacting protein PinX1 was shown to bind the telomerase catalytic subunit hTERT and potently inhibit its activity (32).
In contrast to the extensive knowledge of yeast and mammalian telomeres, the understanding of plant telomeres is rudimentary. Recent studies have shown that telomere structure in higher plants is very similar to that in other eukaryotes (33,34). Gel mobility shift assays have detected double-stranded telomeric DNA-binding proteins in maize and Arabidopsis crude extracts (35,36), whereas proteins that bind to the G-rich single-stranded telomeric DNA have been characterized in nuclear protein extracts of both monocot and dicot plants, including Arabidopsis, rice, and mung bean (37)(38)(39). Interestingly, these studies indicate that the expression of plant telomerebinding proteins is developmentally regulated. For example, the Arabidopsis telomeric protein ATBP2 is transiently expressed during the onset of leaf senescence (38). Three kinds of telomere-protein complexes (A, B, and C) have been observed in mung bean seedlings; complex C was specifically found in very young radicles, while a faster-migrating complex A was detected only in the hypocotyl and root tissues (39). Recently, cDNAs encoding potential telomeric DNA-binding proteins have been isolated from rice and Arabidopsis (40 -42). Like human TRF1/Pin2 and TRF2, the predicted protein sequences of rice RTBP1 and Arabidopsis AtTRP1 and AtTBP1 contain a single Myb domain at their C termini, and they exhibit sequence similarity to Myb-related double-stranded telomeric DNA-binding proteins from other organisms. Gel retardation assays showed that the Myb domain of both the rice and Arabidopsis proteins was able to bind to double-stranded, but not single-stranded, plant telomeric DNA (40 -42). The mRNAs for rice RTBP1 and Arabidopsis AtTBP1 were ubiquitously present in all tissues examined (40,42). However, these studies did not provide any information about the physiological roles and developmental regulation patterns of these telomere-related genes. In addition, there is no evidence that the encoded telomeric proteins are localized to telomeres in planta.
We are interested in elucidating the molecular mechanisms of the regulation of genes for telomeric DNA-binding proteins in higher plants. As nothing is known about the regulation and physiological function of these genes in relation to the cell division, the particular aim of the present study was to identify the possible relationship between cell division and telomerebinding protein gene activity in tobacco BY-2 suspension cells. For this purpose, we isolated a full-length tobacco cDNA clone, pNgTRF1, which encodes a protein that shares significant se-quence identity with previously identified double-stranded telomeric repeat binding factors in mammalian cells. We found that NgTRF1 was able to interact specifically with the doublestranded plant telomeric repeat sequence (TTTAGGG) 4 . Expression analysis showed that the NgTRF1 transcript is present in tissues with negligible cell division activity. During the 9-day culture period for BY-2 suspension cells, the NgTRF1 gene was highly activated only in stationary phase cells that had no capacity for proliferation. Our results further showed that in synchronized BY-2 cells, NgTRF1 mRNA was selectively expressed in G 1 phase, which contradicts the idea that telomerases are usually active during S phase. These results indicate that the degree of NgTRF1 expression in tobacco is inversely correlated with the capacity for cell division. Since telomerase activity is usually associated with the cell division program, the present results suggest that an opposite regulatory mechanism exists for the control of telomerase activity and the expression of NgTRF1 during the cell cycle in tobacco plants.

EXPERIMENTAL PROCEDURES
Polymerase Chain Reaction-The first strand cDNA, synthesized from 1 g of poly(A) ϩ RNA isolated from mature leaves of 10-week-old Nicotiana glutinosa L., which are diploid tobacco plants, was PCRamplified using mixed oligonucleotide primers (CTGAATTCAA(G/A)A-A(G/A)GT(G/A/T/C)(C/A)G(G/A/T/C)GA(T/C)GA and CTGGATTCGT(T/ C)TTCCA(T/C)TT(G/A)TC(T/C)TT). These primer sequences corresponded to the amino acid sequences KKVRDD and KDKWKT, respectively, which are highly conserved in previously identified putative plant telomeric proteins (40 -42). EcoRI and BamHI sites were included at the 5Ј-ends of the sense and antisense primers, respectively, to facilitate subcloning of PCR products. PCR was performed in a total volume of 50 l containing 5 l of the first strand cDNA reaction products, 1 M primers, 10 mM Tris (pH 8.0), 50 mM KCl, 1.5 mM MgCl 2 , 0.01% gelatin, 200 M deoxynucleotides, and 2.5 units of Taq polymerase (Promega Biotech, Madison, WI). Thirty-five cycles were carried out, each consisting of 1 min at 94°C, 2 min at 48°C, and 2 min at 72°C in an automatic thermal cycler (PerkinElmer Life Sciences). PCR products were separated on an agarose gel, and an ϳ500-bp band was eluted and reamplified by PCR to increase the amount of DNA used for subsequent subcloning.
Screening of a Zap II Tobacco Leaf cDNA Library-Construction of the diploid tobacco N. glutinosa leaf cDNA library was described earlier (43). Total RNA was isolated from 10-week-old mature leaf tissues. Poly(A) ϩ RNA was then prepared from total RNA using oligo(dT)cellulose (Roche Applied Science) according to the manufacturer's instructions. The preparation of double-stranded DNA complementary to poly(A) ϩ RNA (5 g) and its ligation to Uni-Zap arms were performed according to the protocol of Stratagene (La Jolla, CA). The cDNA library was screened using the 500-bp PCR products described above as probes under low stringency hybridization and washing conditions (43).
Cloning and Expression of Full-length and Various Deletion Mutants of NgTRF1-The plasmid pGEX4T-1 (Amersham Biosciences) was used for the expression of GST-NgTRF1 fusion proteins. Escherichia coli BL21 (DE3) strains containing the plasmids were grown at 37°C in 100 ml of 2ϫ LB medium (10 g of trypton, 10 g of yeast extract, 5 g of NaCl per liter) supplemented with 1% glycerol as an additional carbon source and 70 g/ml ampicillin. Cells were grown for a further 4 h after induction with 1 mM isopropyl-1-thio-␤-D-galactopyranoside at an OD 600 of 0.6 -1.0. The pellet was resuspended in phosphate-buffered saline containing 1 mM phenylmethylsulfonyl fluoride. The suspension was sonicated on ice with Vibracell sonicator (Sonics and Materials Inc., Danbury, CT), and Triton X-100 was added to a final concentration of 1%. Various fusion proteins were purified by affinity chromatography using glutathione-Sepharose 4B from GST purification modules (Amersham Biosciences), and the GST tags were cleaved off by thrombin.
Gel Retardation Assay and Calculation of Dissociation Constants-DNA probes and competitors for gel retardation assays are described in Table I. To reduce nonspecific DNA-protein binding, purified NgTRF1 protein (0.5-1.0 g) was preincubated with 0.5 g of poly(dI-dC) and 0.5 g of nonspecific DNA oligonucleotide in 20 l of binding buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, 50 mM NaCl, and 5% glycerol) for 10 min on ice. End-labeled DNA probe (0.25 ng) was then added to the reaction mixtures. After incubating for 10 min on ice, the mixtures were loaded on an 8% nondenatured polyacrylamide gel. Be-fore loading, gels were prerun at 10 V/cm for 30 min, and electrophoresis was performed in 0.5ϫ TBE (54 mM Tris borate pH 8.3, 1 mM EDTA) for 2.5 h. For the competition experiments, varying amounts of cold competitor molecules were preincubated with NgTRF1 before the addition of radiolabeled probe. The gel was dried and autoradiographed. Binding activity was quantified with a PhosphorImager (Fuji).
Determination of the dissociation constants (K d ) was carried out by incubating a fixed amount of 12 kDa protein (2.0 ϫ 10 Ϫ8 M) with increasing amounts of probes (NgTR-2, M4, and M11) under the standard binding conditions. Probes were used at concentrations between 0.8 ϫ 10 Ϫ8 and 4.5 ϫ 10 Ϫ8 M depending on the protein-probe combination. After gel electrophoresis, bound and free probes were quantified with a scintillation counter (Beckman) and PhosphorImager. Two quantified results were compared with a calibration curve with known DNA concentrations. Dissociation constants were determined based on the equations established by Riggs et al. (44). For the reversible binding reaction of a 12-kDa NgTRF1 polypeptide (the Myb motif and the C-terminal glutamine-rich domain) to the two-telomeric repeat DNA site is is the concentration of DNAprotein complex, [P f ] that of unbound protein in the solution and [D f ] that of unbound DNA. The corresponding equilibrium dissociation constants were estimated from the slope of a Scatchard plot of the results.
Subcellular Localization of NgTRF1-The termination codon of the green fluorescent protein (GFP) cDNA was removed using PCR. The resulting fragment was then fused in-frame to the full-length pNgTRF1 coding region or to truncated pNgTRF1 mutants. Transient expression of GFP fusion constructs was performed by introducing the DNAs into onion epidermal cells using the particle bombardment method according to the manufacturer's protocol (Bio-Rad). Fluorescence photographs of onion cells were taken using a Zeiss (Jena, Germany) Axiophot fluorescence microscope fitted with fluorescein isothiocyanate filters (excitation filter, 450 -490 nm; emission filter, 520 nm; dichroic mirror, 510 nm) and Fuji 400 color film. The optimal exposure time was 1 s.
Isolation of Genomic DNA and Southern Blot Analysis-N. glutinosa leaf genomic DNA was isolated as described previously (43) with modifications. Each gram of N. glutinosa leaf pulverized under liquid nitrogen was suspended in 2.5 ml of extraction buffer (8.0 M urea, 50 mM Tris-Cl, pH 7.5, 20 mM EDTA, 350 mM NaCl, 2% sarkosyl, 5% phenol, and 20 mM 2-mercaptoethanol). After successive extractions with phenol/chloroform (1:1, v/v), the aqueous phase was concentrated by ethanol precipitation. The pellet was resuspended in 10 mM Tris-Cl (pH 7.5) and 1 mM EDTA, adjusted to a density of 1.5 g ml Ϫ1 by the addition of saturated CsCl, and the DNA was banded overnight in a vertical rotor at 200,000 ϫ g. The DNA band was collected, extracted with 1-butyl alcohol, and dialyzed extensively against 10 mM Tris-Cl (pH 7.5) and 1 mM EDTA. The N. glutinosa genomic DNA (10 g/lane) was digested with appropriate enzymes, separated by electrophoresis in a 0.7% agarose gel, and blotted to a nylon membrane filter (Bio-Rad). The filter was hybridized to the 32 P-labeled EcoRI/BamHI fragment of pNgTRF1 as described previously (43).
RNA Isolation and Northern Blot Analysis-Total RNAs were obtained from various tissues of N. glutinosa plants and BY-2 suspension culture cells as described previously (43). Total RNA was precipitated overnight at 4°C by the addition of 0.3 volumes of 10 M LiCl, and then precipitated in ethanol. Total RNA (20 g) was separated by electrophoresis in a 1% formaldehyde-agarose gel and blotted to a nylon membrane (Bio-Rad). To ensure equal loading of RNA, the gel was stained with ethidium bromide after electrophoresis. To confirm complete transfer of RNA to the membrane filter, both gel and membrane were viewed under UV light at the end of transferring. The filter was hybridized to the 32 P-labeled EcoRI/BamHI fragment of pNgTRF1 under high stringency hybridization and washing conditions. The blots were washed as described previously (43) and visualized by autoradiography at Ϫ80°C using Kodak XAR-5 film and an intensifying screen.
Tobacco Suspension Cell Culture and Synchronization-The tobacco BY-2 (Nicotiana tabacum L. cv. bright yellow-2) suspension cell line was maintained in Murashige-Skoog salt medium (Sigma) on a rotary shaker (120 rpm) at 27°C in the dark. Every week, 10 ml of stationary phase cells were added to 40 ml of fresh medium and cultured. To synchronize the tobacco cells in S phase, 10 ml of 7-day-old suspension cells were transferred to 40 ml of fresh Murashige-Skoog medium containing 10 mM hydroxurea and cultured. After 24 h, the cells were washed extensively with 600 ml of 3% (w/v) sucrose solution and then resuspended in 50 ml of fresh medium. Samples were taken at 2-h intervals for various analyses. To arrest the cell cycle progression at G 1 /S phase, actively dividing 3-day-old suspension culture cells were incubated with or without 10 M indomethacin, after which samples were taken at 4-h intervals for Northern analysis.
Preparation of Tobacco Cell Extracts-A portion (0.5 g) of synchronized cells at various phases in the cell cycle was ground with a mortar and pestle under liquid nitrogen and suspended in 2 ml of ice-cold extraction buffer W (50 mM Tris acetate, pH 7.5, 5 mM MgCl 2 , 100 mM potassium glutamate, 20 mM EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 0.6 mM vanadyl ribonucleoside complex, 1.5% (w/v) polyvinylpyrrolidone, and 10% glycerol). After centrifugation at 13,000 rpm for 15 min at 4°C, polyethylene glycol 8000 (Sigma) was added to the supernatant to a final concentration of 10%, mixed for 30 min at 4°C, and centrifuged at 13,000 rpm for 5 min. The resulting pellet was then resuspended in 0.5 ml of buffer W and centrifuged again at 13,000 rpm for 5 min at 4°C. The supernatant was either used immediately for the telomerase activity assay or stored at Ϫ80°C until use.
Telomerase Activity Assay-Telomerase activity was monitored using the telomere repeat amplification protocol (TRAP). The-184-bp DNA fragment derived from the multicloning site region of Bluescript SK was used for an internal standard (IS). The TRAP assay was conducted in 40 l of reaction mixture composed of 50 mM Tris acetate (pH. 8.3), 50 mM potassium glutamate, 0.1% Triton X-100, 1 mM spermidine, 1 mM dithiothreitol, 50 M each dNTP, 5 mM MgCl 2 , 10 mM EGTA, 100 ng l Ϫ1 bovine serum albumin, and 0.5 units of Taq polymerase (Promega). After the addition of tobacco cell extracts containing 1 g of total proteins and 50 ng of GG (21) forward primer (CACTATCGACTACGC-GATCGG), the telomerase reaction was allowed to proceed at 24°C for 45 min. For a negative control of the telomerase assay, some samples were pretreated with 10 ng of RNase A (Sigma) or heated to 95°C for 10 min. 50 ng of (C 3 TA 3 ) 3 reverse primer (CCCTAAACCCTAAAC-CCTAAA) and 1.5 ng of IS DNA containing GG (21) and (C 3 TA 3 ) 3 primer sequences at its 5Ј-and 3Ј-end, respectively, were then added, and the reaction mixture was covered with 50 l of paraffin oil and heated to 94°C. The products of the telomerase reaction were amplified with 30 cycles of PCR, each consisting of 30 s at 94°C, 30 s at 65°C, and 90 s at 72°C, with an additional 5 min at 72°C in an automatic thermal cycler (PerkinElmer Life Sciences). PCR products were separated on a 10% nondenaturing polyacrylamide gel, and the gel was stained in 0.01% SYBR Green I for 45 min and scanned on a FluorImager (Fuji). TRAP ladders were read from the first product excluding primer dimer up to below the IS band. Telomerase activities are expressed as the ratio of the intensity of the TRAP ladder to the IS signal.

RESULTS
Isolation and Characterization of NgTRF1 cDNA-As a first step to gain insight into the regulation of telomere-binding protein genes during the cell cycle, we proceeded to isolate cDNAs encoding homologs of the Myb-related double-stranded telomeric DNA-binding proteins that have been identified in plants and mammals. Poly(A) ϩ RNA was isolated from mature leaves of N. glutinosa L. Following the synthesis of the first strand cDNA from 1 g of poly(A) ϩ RNA, PCR was carried out with mixed oligonucleotides corresponding to the amino acid sequence KKVRDD for the upstream primer, KDKWKT for the downstream primer (see "Experimental Procedures" for sequences), and the first strand cDNA as the template. These primer amino acid sequences are highly conserved in rice RTBP1 and Arabidopsis AtTBP1 and AtTRP1 (40 -42). Total PCR products of about 500 bp were radioactively labeled and used as probes to screen an N. glutinosa leaf cDNA library under low stringency hybridization and washing conditions (43). Four putative clones containing inserts of about 0.7-2.5 kb in size were isolated. Subsequent restriction enzyme mapping and DNA sequencing analyses revealed that these clones represent a single group of overlapping sequences. Fig. 1A shows the restriction enzyme map of pNgTRF1 that contains the longest insert among isolated clones. The pNgTRF1 clone (GenBank TM accession no. AF543195) is 2459-bp long and comprises a 99-bp 5Ј-untranslated region, a 2046-bp coding region encoding 682 amino acids, and a 314-bp 3Ј-untranslated region. The predicted molecular mass of the polypeptide encoded by pNgTRF1 is 75 kDa, which is slightly larger than those of other putative plant telomere-binding proteins, including rice RTBP1 (70 kDa) and Arabidopsis AtTBP1 (70.6 kDa), and AtTRP1 (65 kDa) (40 -42). Isolation of the complete pNgTRF1 sequence allowed us to compare it to the telomeric DNA-binding proteins of other organisms and to analyze their structural relationships. Despite apparently conserved functions, telo-mere-binding proteins show a relatively low degree of sequence similarity. Tobacco NgTRF1 shares 35-44% identity with other putative plant telomere-binding proteins, with AtTRP1 being the most closely related, while NgTRF1 has a limited homology to human TRF1 (17% identity) and TRF2 (14% identity) (Fig.  1B). NgTRF1 is also homologous (44%) to a protein encoded by an Arabidopsis-expressed sequence tag clone (GenBank TM accession number NP190243) whose function has not been addressed. As was found in other double-stranded specific telomere-binding proteins, NgTRF1 possesses a single Myb-like domain near the C-terminal region and one potential nuclear localization signal. The Myb-like domain of NgTRF1 is strongly conserved (85-90%) compared with that of rice and Arabidopsis proteins, and is 27 and 24% identical to its corresponding domains in TRF1/Pin2 and TRF2, respectively, suggesting that this domain may function as a DNA-binding motif on its telomeric recognition site in tobacco cells (Fig. 1B). Human TRF1 and TRF2 have acidic and basic regions, respectively, at their N termini, which are followed by a dimerization domain (25)(26)(27). However, these domains are not found in NgTRF1 and in any of the rice RTBP1 and Arabidopsis AtTBP1 and AtTRP1. The D-like motif (RIFGDPN) in TRF1/Pin2, which is reminiscent of the destruction box (RXXXGDXXN) (31), is also absent in plant proteins. Instead, plant telomeric proteins have a region similar to the ubiquitin domain (UBD) (Fig. 1B). The UBD, also called UBL or UBQ, is defined by a stretch of 45-80 amino acid residues with significant sequence homology to ubiquitin (45). An emerging general property of UBD is its ability to bind to the 26 S proteasome. Finally, NgTRF1 has two serine residues at positions 347 and 349, which are possible CDK1 phosphorylation sites (Fig. 1B, indicated by star). Taken together, these structural conservations imply that NgTRF1 may play a role in telomere function in tobacco plants.
NgTRF1 Binds Specifically to Plant Double-stranded Telomeric Sequences in Vitro-To explore whether NgTRF1 can bind telomeric DNA, we expressed NgTRF1 in E. coli, and the purified protein was used in a gel retardation assay with a 32 P-labeled NgTR-4 (see Table I) possessing four plant duplex telomeric DNA repeats, (TTTAGGG) 4 . The full-length 75-kDa NgTRF1 gave rise to a single, discrete DNA-protein complex that migrated more slowly than the free probe ( Fig. 2A). The intensity of this shifted band increased upon the addition of increasing amounts (0.5-1 g) of NgTRF1. The DNA binding specificity of NgTRF1 was also confirmed by competition binding experiments, which showed that a 25-fold excess of cold NgTR-4 was enough to displace the labeled probe ( Fig. 2A). However, telomeric repeats of other organisms, such as human (HTR-4) and Caenorhabditis elegans (CTR-4), as well as nonspecific DNA (NS), were not capable of interacting with NgTRF1 ( Fig. 2A). In addition, a 50-fold molar excess of these unrelated, nonspecific cold DNAs failed to compete with the labeled NgTR-4, implying that NgTRF1 indeed binds specifically to double-stranded plant telomeric DNA in vitro. In con-trast, the expressed NgTRF1 did not exhibit any detectable DNA-binding capacity to single-stranded telomeric repeats (data not shown). To assess the minimum number of telomeric repeats required for binding to NgTRF1, we next performed gel shift assays using probes (NgTR-1, NgTR-2, NgTR-3, and NgTR-4) with different numbers of telomeric DNA repeats (Table I). As shown in Fig. 2B, full-length NgTRF1 can form complexes only with the probes comprising three or more contiguous TTTAGGG repeats. Thus, under our experimental conditions, the minimum length of a telomeric DNA bound by NgTRF1 in vitro spans at least three repeats. There was no significant difference in the complex formation pattern between NgTR-3 and NgTR-4 (Fig. 2B). This may indicate that NgTRF1 requires three TTTAGGG repeats for efficient complex formation. However, we could not rule out the possibility that probes with more than five TTTAGGG repeats have an enhanced binding activity to NgTRF1.
The isolated Myb motif of TRF1 binds specifically to human telomeric DNA (46). As the C-terminal Myb-like region of NgTRF1 is 27 and 24% identical to its corresponding domains of TRF1 and TRF2, respectively (Fig. 1B), we investigated whether it is responsible for the DNA binding specificity. Various NgTRF1 deletion mutants were expressed in E. coli, and the isolated proteins were tested for their DNA binding properties. The 67-kDa mutant, which lacks the C-terminal Myblike domain, demonstrated complete absence of binding activity to NgTR-4, confirming that this motif is essential for effective complex formation (Fig. 2C). The gel shift analysis, however, revealed that the Myb-like motif of NgTRF1 is not sufficient to form the telomere-protein complex, as evidenced by the results that neither the Myb domain (8 kDa) nor a GST-Myb fusion protein (34 kDa) interacted with NgTR-4 (Fig. 2C). In contrast, the truncated mutants composed of the Myb motif plus the extreme C-terminal region (the 38-and 12-kDa proteins in Fig.  2C) retained full DNA binding activities. As removal of the C-terminal portion abrogated the DNA binding activity, it seems most likely that, in addition to the Myb motif, this domain is also required for optimal binding to telomeric DNA. This extreme C-terminal polypeptide is rich in glutamine residues (Fig. 1B, indicated by dots) and found only in the plant telomere-binding proteins. Its precise role, however, remains to be elucidated. Fig. 2C also shows that polypeptides containing the Myb motif and the C-terminal glutamine-rich domain produced more than one specific complex when interacting with NgTR-4. This led us to repeat the gel mobility shift assay using various DNA sequences carrying one to four contiguous copies of the TTTAGGG repeat and the isolated DNA-binding domain of NgTRF1 to identify the recognition site within the telomeric sequence at which DNA-protein interactions occur. The results showed that a 26-kDa protein containing the Myb motif plus the glutamine-rich domain generated a single retarded band with the two-repeat sequence (NgTR-2) and two bands with the three-repeat sequence (NgTR-3), while it exhibited no DNA binding ability with the one-repeat sequence (NgTR-1) (Fig.  3A). When the NgTR-4 probe possessing four telomeric repeats was tested, a third, additional complex of weaker intensity and slower migration was resolved. These results indicate that the recognition site of the NgTRF1 DNA-binding domain resides within the two-telomere repeat sequence TTTAGGGTT-TAGGG. To further specify the target sequence recognized by NgTRF1, we synthesized a series of telomeric repeat mutants and analyzed them for their binding activities to the isolated 26-kDa NgTRF1 polypeptide. 14 single-base mutants of the 14-bp two-telomeric repeat sequence, TTTAGGGTTTAGGG, were prepared (Fig. 3B). Gel retardation analysis revealed that substitutions of any of the first 3 nucleotides (M1, M2, and M3 probes) or the terminal 3 nucleotides (M12, M13, and M14 probes) did not significantly affect the binding ability of the protein, while A 3 T substitutions in the 4th and 11th positions, respectively, reduced the binding activity by ϳ50% (Fig.  3B). In contrast, the DNA binding capacity was almost completely diminished with the M5, M6, M7, M8, M9, and M10 probes, which contained single-base mutations in the internal GGGTTT sequence (Fig. 3B). These results suggest that the internal GGGTTT sequence in the two-telomere repeat is crucial for binding of the Myb plus the glutamine-rich domains of NgTRF1.
Since the results described in Fig. 3 suggest that the internal GGGTTT core sequence is critical for binding of NgTRF1, we next went on to estimate its DNA binding affinity with different DNA substrates. The dissociation constant for binding of the 12-kDa protein containing the Myb-plus glutamine-rich domains (Fig. 2) to the two-telomeric repeat sequence was calculated from a quantitative gel mobility shift assay (Fig. 4) as described under "Experimental Procedures." The plot in Fig.  4B depicts the concentration of DNA-protein complex plotted against the concentration of complex divided by the concentration of free DNA (NgTR-2), and the dissociation constant was evaluated from the slope to be 4.5 Ϯ 0.2 ϫ 10 Ϫ9 M. This value is slightly higher than that of the Myb-like domain of human TRF1 (3.2 Ϯ 0.5 ϫ 10 Ϫ9 M) (46). To show that the affinity is specific for the plant telomeric repeat, we carried out a quantitative gel retardation analysis using M4 and M11, which contain A 3 T substitutions in the 4th and 11th positions of the two-telomeric repeat, respectively (Fig. 3B). Fig. 4 shows that the dissociation constants of the 12 kDa protein to M4 and M11 are 1.3 Ϯ 0.3 ϫ 10 Ϫ8 M and 8.4 Ϯ 0.6 ϫ 10 Ϫ9 M, respectively, indicating that M4 and M11 are still able to interact with the protein although they possess significantly lower affinity compared with NgTR-2. In contrast, as shown in Fig. 3B, the 12-kDa polypeptide failed to bind the DNA substrates (M5-M10), which have mutations in the internal GGGTTT core sequence, with any concentrations of those DNAs (data not shown). Thus, the results presented in Figs. 3 and 4 indicate that the internal GGGTTT sequence in the two-telomeric repeat DNA site is indeed important, and are consistent with the recent model that base-specific contacts between telomerebinding proteins and telomeric repeat sequences are made within the internal core sequence in both humans (46) and yeast (47).
Targeting of NgTRF1 to the Nucleus-Since the predicted sequence of NgTRF1 contains a Myb-like domain as well as a putative nuclear localization signal (KRRK, Fig. 1B), the protein is expected to localize to the nucleus. To confirm this, we performed an in vivo targeting experiment that employed an NgTRF1-fused green fluorescent protein (GFP) as a fluorescent marker in a transient transfection assay. The GFP gene was fused to the 5Ј-end of the pNgTRF1 coding region in-frame under the control of the cauliflower mosaic virus 35 S promoter (Fig. 5A), and the resulting construct was introduced into onion epidermal cells by the particle bombardment method (48). Localization of the fusion protein was then determined by visualization with a fluorescence microscope. As shown in Fig. 5B, the control GFP was uniformly distributed throughout the cell (panel a), while the GFP-NgTRF1 fusion protein was localized to the nucleus (panel b). This subcellular localization pattern was essentially identical to that of the GFP-NgTRF1 1-360 mutant protein in which the C-terminal-half was truncated (panel c). In contrast, the GFP-NgTRF1 361-681 protein, which lacked the N-terminal 360 amino acid residues, exhibited uniform accumulation inside the cell (panel d). These observations support the notion that the putative nuclear targeting sequence of NgTRF1 is sufficient and that no additional post-translational modification may be necessary for the NgTRF1 protein to be targeted to the nucleus.
Organization and Expression of the NgTRF1 Gene-From the results described above, it appears that NgTRF1 is a nuclear protein that specifically binds plant telomeric DNA in vitro. We therefore wanted to characterize the NgTRF1 gene in more detail at the molecular level. To assess the NgTRF1 gene copy number in the tobacco genome, genomic Southern blot analysis was carried out using pNgTRF1 as a probe. The genomic DNA isolated from mature leaves of N. glutinosa (diploid, 2n) and N. tabacum cv. Samsun NN (amphidiploid, 4ϫ) was digested with EcoRI, HindIII, or XbaI, and hybridized to the 32 P-labeled EcoRI/BamHI fragment of pNgTRF1 under normal stringency conditions (Fig. 1A). This hybridization detected only one clear band in N. glutinosa, whereas two major hybridizing bands were observed in N. tabacum (Fig. 6A). No additional fragments were visible in any of the digests, even with low stringency hybridization or longer exposure of the blot to x-ray film (data not shown). These results imply that the NgTRF1 gene is present in a single copy per haploid tobacco genome.
To examine the spatial and temporal expression pattern of the NgTRF1 gene, we measured the level of corresponding mRNA in different parts of tobacco plants by Northern blot analysis. Total RNAs isolated from leaves, roots, stems, and flowers of N. glutinosa were hybridized to the 32 P-labeled EcoRI/BamHI fragment of pNgTRF1 under high stringency conditions. A substantial level of the 2.5-kb transcript was detected in every tissue examined, indicating that NgTRF1 is constitutively expressed in mature tissues of tobacco plants (Fig. 6B). These results are in line with the previous observations that rice RTBP1 and Arabidopsis AtTBP1 are ubiquitously present in various parts of plants (40,42). To investigate whether NgTRF1 expression is correlated with tissue development, roots were serially dissected according to different developmental stages (sections 1-4), and total RNAs from each section were subjected to RNA gel blot analysis. Interestingly, the abundance of the message varied depending on the sections; the maximum amount of NgTRF1 transcript was found in the fully differentiated region (section 1 in Fig. 6C), but expression gradually decreased acropetally, with only background levels detectable in section 4, where actively dividing FIG. 3. Gel retardation assays of the DNA-binding domain (Myb-like motif plus C-terminal glutamine-rich region) of NgTRF1 to the different telomere repeats and to the mutated two-telomere repeats. A, gel retardation assays were carried out with radiolabeled NgTR-1 (lanes 1-3), NgTR-2 (lanes 4 -6), NgTR-3 (lanes 7-9), or NgTR-4 (lanes 10 -12). Each set of lanes contained 0, 0.5, and 1 g of isolated 26-kDa polypeptide, respectively, which consists of the Myb-like motif and the C-terminal glutamine-rich region of NgTRF1. B, a series of mutants of two-telomere repeats were prepared and analyzed by gel retardation assays. Each double-stranded probe contained a single nucleotide transition in the two-telomere repeat as shown in shaded boxes. The open box refers to the internal GGGTTT core sequence. Gel retardation assays were performed using 1 g of the 26-kDa polypeptide and radiolabeled NgTR-2 or mutated probes as indicated above each lane. meristemetic cells are present (Fig. 6C). Thus, these results suggest that the expression of NgTRF1 is differentially regulated in different stages of root development, raising the possibility that the tobacco telomere-binding protein gene is subject to control by a development-specific mechanism.
Cell Division-dependent Regulation of NgTRF1-Telomerase is a specialized RNA-directed DNA polymerase and is responsible for the synthesis of telomeric repeat DNA (49). As found in animal tissues, the telomerase activity and expression of mRNA for the telomerase reverse transcriptase subunit, At-TERT, are closely linked to cell division in Arabidopsis; the AtTERT transcript is abundantly present in callus and shoot apical meristems, but not in mature leaf tissue (50,51). Likewise, a high level of telomerase activity was detected in tobacco BY-2 suspension culture cells and roots, both of which contained a high proportion of actively dividing cells, while very low levels of activity were found in other mature tobacco tissues, including stems, leaves, and flowers (52). In synchronized BY-2 cells, telomerase activity is low during most phases of the cell cycle, but is strongly induced at the onset of S phase, indicating that it is regulated in a cell cycle-dependent fashion (53,54).
Because NgTRF1 has a distinct expression pattern in each stage of root development (Fig. 6C), we considered the possibility that, along with telomerase activity, NgTRF1 expression is also mediated in a cell division-specific manner in tobacco cells. To address this possibility, we monitored the change in steady-state NgTRF1 transcript levels as well as telomerase activity during a 9-day culture period in asynchronous tobacco BY-2 suspension cells. The results showed that telomerase activity, as measured by TRAP assay, was relatively low on day 1. This low, basal expression of activity was markedly enhanced on the second day, and reached a maximum at day 4 (Fig. 7B). This time point corresponded to the middle of the exponential growth phase of the cultures (Fig. 7A). Subsequently, telomerase activity declined, resulting in negligible activity in 9-day-old culture cells. In contrast, NgTRF1 showed a remarkably different expression pattern. The amount of NgTRF1 mRNA was very low during the logarithmic phase of growth in tobacco suspension cultures (Fig. 7C). The transcript began to accumulate at day 7, and attained a maximal level A, the GFP coding region was fused in-frame to the full-length pNgTRF1 coding region or to truncated pNgTRF1 mutants. Constructs were introduced into onion epidermal cells by the particle bombardment method and expressed under the control of the CaMV 35 S promoter. B, expression of the introduced genes was viewed after 12 h by fluorescence microscopy under dark field or light field. during days 8 and 9, when BY-2 cells have no capacity for proliferation. The CYM gene, which belongs to the family of mitotic cyclins, was included in the RNA expression experiments as a positive control (55), and its mRNA was specifically expressed only in actively dividing cells (i.e. within the exponential growth phase) (Fig. 7C). Therefore, telomerase activity and NgTRF1 expression levels demonstrated positive and negative correlations, respectively, with the program of cell proliferation in BY-2 suspension cultures.
To obtain additional evidence supporting the cell division-dependent expression of NgTRF1, telomerase activity and NgTRF1 transcript levels were measured at different time points in BY-2 cell suspension cultures following exposure to 10 mM hydroxyurea, which arrests the cell cycle at S phase (56). The proportion of synchronized cells was estimated by mitotic index. The maximum mitotic index was measured to be about 70% 12 h after release from hydroxyurea-induced S phase arrest (Fig. 8A). RNA gel blot analysis revealed that CYM mRNA, a marker for M phase (55), was barely detectable until 10 h after S phase blocking, accumulated to a high level at 12-16 h, and started to decrease at the 18 h point, suggesting that partially synchronous cell division occurred during 12-16 h (Fig. 8A). By contrast, the NgTRF1 transcript accumulated only at low levels during the progression through S and M phase. Subsequently, the level increased gradually, and a peak was observed 24 h after release from S phase, a time point corresponding to late G 1 phase (Fig. 8A). Consistent with previous results (54), the S phase-specific telomerase activity rapidly disappeared as the cell cycle progressed in BY-2 cells (Fig.  8B). Thus, these results suggest that NgTRF1 mRNA, in addition to telomerase activity, is expressed in a cell cycle-dependent fashion in tobacco BY-2 cells, with NgTRF1 being activated specifically in the cells that exit mitosis and enter interphase.
The prostaglandin inhibiting drug indomethacin was previously shown to arrest the cell cycle progression at G 1 /S phase of tobacco BY-2 cells (57). To address cell cycle-dependent expression of NgTRF1 in more detail, 10 M indomethacin was added to the actively dividing 3-day-old BY-2 culture cells, after which the levels of both NgTRF1 and CYM mRNAs were examined at different time points. The expression of the M phase marker CYM showed a G 1 /S block in indomethacin-treated cells; its expression began to decline at 8 h, and thereafter much lower level of transcript was detected (Fig. 9). In contrast, induction of NgTRF1 mRNA was clearly detected at 8-h treatment, and this G 1 phase-specific expression of the transcript continuously increased in the presence of indomethacin for at least 16 h after treatment (Fig. 9). These results suggest that the NgTRF1 gene is activated in the G 1 phase during the cell division of BY-2 cells. Such a finding, in conjunction with the data presented in Figs. 6C, 7, and 8, is consistent with the hypothesis that there exists a tight regulatory mechanism by which the activities of telomerase and telomere-binding protein are inversely modulated during tobacco cell division. On the other hand, it is worth noting that the relative abundance of the G 1 -specifc NgTRF1 mRNA was significantly lower than that in culture cells at the stationary growth phase (compare lanes 9 and 10 in Fig. 9) and that in various mature organs (data not shown). This may indicate that the NgTRF1 gene is more abundantly expressed in quiescent cells than in cycling cells of tobacco plants. DISCUSSION In humans, telomeres and telomerase have attracted much interest, since their structures and activities are closely associated with the program of cellular proliferation, differentiation, aging, and tumor growth. Although far less is known about plant telomeres and telomerase compared with animal systems, it is becoming increasingly apparent that telomere structure and patterns of telomerase regulation are largely conserved between animals and higher plants (33,34). agarose-formaldehyde gel. The gel was blotted onto a membrane filter, and the blot was hybridized to the 32 P-labeled probe for pNgTRF1. The equivalence of RNA loading among lanes was demonstrated by ethidium bromide staining of RNA on the gel. C, differential expression of the NgTRF1 gene in different parts of roots. Tobacco roots were serially dissected according to developmental stage (sections 1-4) and total RNAs (20 g) from each section were subjected to RNA gel blot analysis as described above.
Recent research on the organization of telomeric chromatin indicates that telomeres are packaged into specialized nucleoprotein complexes and that non-histone proteins are an integral component of and are responsible for the integrity and proper functioning of telomeres (14 -16). Telomere-binding proteins have been extensively characterized in several organisms, such as ciliates, yeast, and humans, and are classified into two functional groups; the single-stranded binding proteins that interact with the 3Ј extension of the extreme termini of telomeres (17)(18)(19), and the double-stranded-specific binding proteins (20 -27). Although these telomere-binding proteins exhibit a limited amino acid sequence identity, they share a unique domain that resembles the DNA-binding motif present in the vertebrate c-Myb family of transcriptional activators (46,58). For instance, yeast Rap1p contains two Myb-like domains that coordinate specific binding to DNA in a tandem orientation (59), whereas the DNA-binding domains of Taz1p, TRF1/ Pin2, and TRF2 possess a single Myb motif at their C-terminal ends (26,60,61). The isolated Myb motifs of TRF1 and Taz1p were found to bind specifically to duplex telomeric DNA, indicating that the Myb-like domain is responsible for specific telomeric DNA recognition (46,60,61).
In the present study, we have isolated and characterized a cDNA clone, pNgTRF1, that encodes a putative telomere-binding protein in N. glutinosa, a diploid tobacco plant. As expected, NgTRF1 contains a single C-terminal Myb-like domain that is FIG. 7. Growth phase-dependent telomerase activity and NgTRF1 mRNA expression in BY-2 suspension cells. A, growth curve of tobacco BY-2 cells in batch suspension culture. Stationary phase cells (9 days after previous subculture) were subcultured into fresh medium and incubated for 9 days. The mean cell weight was determined daily; the error bars represent the S.E. determined from three samples. B, telomerase activity during 9 days of culture for BY-2 suspension cells. The level of telomerase activity was estimated at 1-day intervals by TRAP assay. The reaction mixture of the TRAP assay contained 1 g of total extract protein, as described under "Experimental Procedures." The 184-bp internal standard DNA band (IS) is indicated. TRAP ladders were read from the first product excluding primer dimer up to below the IS band. Telomerase activities are expressed as the ratio of the intensity of the TRAP ladder to the IS signal. C, total RNAs (20 g) isolated from samples taken at the indicated time points were blotted and hybridized with the probes indicated, as described in Fig. 6. The CYM gene was included for comparison as a gene expressed during the exponential growth phase. The equivalence of RNA loading among lanes was demonstrated by ethidium bromide staining of RNA on the gel. significantly similar to its corresponding domains in rice and Arabidopsis proteins as well as human TRF1/Pin2 and TRF2, suggesting that it may function as a DNA-binding motif on its telomeric recognition site in tobacco cells (Fig. 1). Gel retardation analysis reveals that the expressed Myb-like domain of NgTRF1 is essential, but not sufficient, for specific binding to the plant double-stranded telomeric repeat sequence (Fig. 2). The extreme C-terminal region of NgTRF1, which consists of eight glutamine residues, is also required for telomere-protein complex formation. This glutamine-rich region is highly homologous among plant proteins, but is not found in human telomeric proteins (Fig. 1). In addition, the N-terminal acidic region, dimerization domain, and D-like motif, which are apparent in TRF1/Pin2 (31), do not exist in plant telomeric proteins, while all of the plant proteins identified to date contain a putative UBD, a possible interacting motif with the 26 S proteasome (Fig. 1). These different architectural properties raise the possibility that the mechanism by which the telomeric repeats and telomere-binding proteins form specific complexes might be different between plants and animals. Since none of the plant telomeric proteins have yet been localized to telomeres in vivo, further studies on the mode by which the telomeric repeat-binding proteins recognize telomeric DNA and on their physiological relevance in plant cells are required.
The enzyme telomerase is one of the most important components of the telomere complex. Although plants show developmental differences and a more plastic pattern of differentiation, several lines of evidence indicate that the expression pattern of telomerase activity in higher plants is similar to that of animals (33,34). As in mammalian cells, high levels of telomerase activity have been detected in actively dividing tissues, including young immature embryos, dedifferentiated callus, root tips, and immature floral buds of various plant species, but the activity is very low or undetectable in most mature vegetative tissues (52,(62)(63)(64)(65). Recently, regulation of the gene encoding the telomerase reverse transcriptase subunit has been studied on a molecular level in Arabidopsis (50, 51). The AtTERT mRNA highly accumulates in callus and shoot apical meristems, as opposed to mature leaf tissue, and this expression pattern of AtTERT parallels the activity of telomerase. In Arabidopsis telomerase-null mutants generated by a T-DNA insertion, telomere length decreased by about 500 bp per generation (51). Although these telomerase-deficient Arabidopsis plants could survive up to 10 generations, the last five generations showed severe developmental defects in both vegetative and reproductive organs, confirming that telomerase function is essential for the maintenance of telomere integrity in higher plants (66). By using synchronized tobacco BY-2 suspension culture cells as a model system, we have recently shown that telomerase activity is tightly coordinated with cell cycle progression. The amount of telomerase activity strikingly increases at early S phase, and this activity was further induced by auxin, a cell division promoting hormone, and inhibited by abscisic acid, a phytohormone known to induce cyclindependent protein kinase inhibitors (54). Based on these findings, we were tempted to presume that, in addition to telomerase activity, the expression of telomere-binding protein genes is also associated with the program of cell division in tobacco plants. This hypothesis was particularly interesting because nothing was known about the regulation of telomeric protein genes during cell division in higher plants. Our present data demonstrate that the activity of the NgTRF1 gene is inversely linked to the capacity of cell division in tobacco roots; high amount of its mRNA is present in the fully differentiated region but gradually declines acropetally, with the lowest level occurring in the root tips, which are primarily populated with actively dividing cells (Fig. 6C). Our results further show that, during the 9-day culture period of BY-2 cells, the NgTRF1 transcript appears in cells within the stationary growth phase, during which telomerase activity is negligible (Fig. 7). Thus, the opposite pattern of expression between the cell divisionassociated telomerase activity and NgTRF1 gene has been demonstrated in roots and during the growth phase of BY-2 suspension cells of tobacco plants.
In human HeLa cells, Pin2 protein levels are highly regulated during the cell cycle, with increased expression during G 2 FIG. 9. Induction of NgTRF1 and suppression of CYM by indomethacin in 3-day-old BY-2 cells. The actively dividing 3-day-old BY-2 suspension culture cells were incubated with or without 10 M indomethacin, which arrests the cell cycle progression at G 1 /S phase, then the mRNA levels of NgTRF1 and CYM were measured at different time points by Northern analysis as described in Fig. 6. The expression levels of transcripts in 9-day-old stationary growth phase cells were included for comparison. The equivalence of RNA loading among lanes was demonstrated by ethidium bromide staining of RNA on the gel. The relative levels of gene expression are shown as the ratio of intensities of corresponding mRNA and rRNA bands measured by PhosphorImager. and M phase and decreased expression during G 1 phase (31). Moreover, overexpression of Pin2 resulted in an accumulation of HeLa cells in G 2 ϩ M phase, suggesting that the Pin2 protein level may help regulate cell cycle progression (31). These results prompted us to examine if the activity of NgTRF1 fluctuates at distinct stages of the cell cycle in synchronized BY-2 cells. We found that NgTRF1 expression is tightly controlled in a cell cycle-dependent manner, with its mRNA level being selectively expressed only in G 1 phase, in contrast to the S phase-specific telomerase activity and M phase-specific CYM mRNA expression (Figs. 8 and 9). Taken together, all the results presented in Figs. 5-9 led us to propose that the regulatory mechanisms of telomerase activity and NgTRF1 gene are inversely associated by an intimate signaling network of as-yet unidentified cellular factors during cell division progression.
The critical question that remains to be unraveled is whether NgTRF1 indeed plays a physiological role in tobacco cells. The ability of NgTRF1 to specifically bind to double-stranded plant telomeric repeat sequences in vitro (Figs. 2-4) as well as its nuclear localization (Fig. 5) suggest that it may be involved in telomere function in vivo. However, localization of NgTRF1 to telomeres in vivo must be confirmed. Telomeric proteins have been shown to negatively regulate telomere length. Such proteins include Rap1p and Taz1p in yeast (20 -24) and TRF1/Pin2 and TRF2 in human (28,30). A novel Pin2-binding protein, PinX1, was isolated by yeast-two hybrid screening from human HeLa cells and shown to bind hTERT, thereby inhibiting telomerase activity and affecting tumorigenicity (32). As the expression of NgTRF1 is strictly and conversely modulated during cell division, it would be interesting to determine the biological significance of the cell cycle fluctuation of NgTRF1 mRNA levels in relation to telomerase activity and telomere length in BY-2 cells. It was previously reported that tobacco nuclei contained telomeric proteins that inhibited telomerase activity by reducing its accessibility to telomeric DNA and, hence, might be involved in regulation of telomere length (67). Thus, it is attractive to speculate that NgTRF1 participates in telomere length regulation by affecting telomerase activity during the tobacco cell division. Conversely, NgTRF1 may serve as a regulatory marker in cell cycle progression; its accumulation is needed for tobacco cells to exit M phase and enter G 1 phase (Figs. 8 and 9). In this regard, a high concentration of NgTRF1 could be a signal for cells to stop dividing and go into the interphase.
Both human TRF1 and Pin2 are phosphoproteins in vivo, and a cell cycle-dependent change in the level of Pin2 protein is controlled at the post-transcriptional level, possibly through a D-like motif related to the destruction box that mediates degradation of many mitotic proteins (31). Intriguingly, NgTRF1 contains putative CDK1-dependent phosphorylation sites as well as UBD, a potential binding domain for the 26 S proteasome (Fig. 1). Thus, further experiments are needed to define the possible role of phosphorylation and control of the NgTRF1 protein level in the course of the cell division in tobacco plants. By using PENT (primer extension/nick translation) assay, it has been recently reported that telomere architecture is developmentally different in all telomerase-negative tissues of the dicot plants Silene latifolia and A. thaliana; the fraction of telomeres with detectable G-overhang structures is substantially changed from 50% in seedlings to 35% in leaves (68). These results, along with the growth-and cell division-dependent expression of NgTRF1 and telomerase activity, are consistent with the idea that plant telomere structure is developmentally dynamic (34). Further functional studies of NgTRF1 and characteristics of the proteins it interacts with will be critical to the understanding of the relationship between cell cycle progression and such a dynamic function of telomeres and telomerase activity in higher plants.