Two Independent Regions of Human Telomerase Reverse Transcriptase Are Important for Its Oligomerization and Telomerase Activity*

Human telomerase reverse transcriptase (hTERT), the catalytic subunit of telomerase, contains motifs conserved among reverse transcriptases. Several nucleic acid-dependent polymerases that share a “fingers, palm, and thumb substructure” were shown to oligomerize. Here we demonstrate that hTERT also has this ability using partially purified recombinant hTERTs and mammalian cells co-expressing differently tagged hTERTs. Human template RNA (hTR), by contrast, has no effect on the structural oligomerization of hTERTs. Therefore, hTERT has an intrinsic ability of oligomerization in the absence of hTR. We identified two separate regions as essential for the oligomerization. The regions, amino acids 301–538 (amino-terminal region) and amino acids 914–928 (carboxyl-terminal region), are outside the fingers and palm substructure covering motif T to D and interact with each other in vivo . A substituted mutant of hTERT, hTERT-D712A-V713I, which was reported as a dominant negative form of hTERT, bound to the wild-type hTERT and inhibited its telomerase activity transiently expressed in telomerase-negative finite normal human fibroblast. The truncated forms of hTERT containing the binding region to the wild-type hTERT partially transfected cipitation FuGENE transfection manufactur-er’s Telomerase activity by two methods. First, telomerase repeat amplification protocol (TRAP) assay TeloChaser Co. of fractionated by electrophoresis on a polyacrylamide and then SYBR-Green (Molecular TRAP enzyme-linked immunosorbent (ELISA) was quantitatively measure

Telomeres are specialized structures positioned at the ends of linear eukaryotic chromosomes that provide a mechanism for maintaining chromosome length and stability. The termini of telomeric DNA cannot be fully replicated by the conventional replication machinery. Telomerase, a ribonucleoprotein complex composed of template RNA and several proteins, elongates telomeres as one means of end replication (1). Telomerase reverse transcriptase (TERT 1 ), the catalytic subunit of telomerase, is a specific type of reverse transcriptase that forms stable complexes with template RNA (TR) (2). Human TERT is the rate-limiting factor for telomerase activity both biologically and biochemically (3,4). Introduction of hTERT into normal human primary cells overcomes senescence and extends their lifespan (3,5). We recently reported that hTERT and hTR, are the minimum components required for telomerase activity reconstituted in vitro with purified forms (4).
TERT is part of a large family of nucleic acid-dependent nucleic acid polymerases that share a " fingers, palm, and thumb substructure" (2, 6 -8). Human TERT contains several motifs conserved among many reverse transcriptases, and additional motifs conserved only among TERTs from species ranging from budding yeasts to humans (9 -11). Some polymerases that share a fingers, palm, and thumb substructure, such as HIV reverse transcriptase, polio RNA-dependent RNA polymerases (RdRP), and hepatitis C virus RdRP, oligomerize (12)(13)(14)(15)(16). Oligomerization in these enzymes induces conformational changes, which provide active or open forms that are essential for catalytic functions.
In Saccharomyces cerevisiae, telomerase forms an active multimer in vivo that may contain two active sites. This suggests that Est2p can oligomerize (17). Recently, one group reported that the human telomerase complex also forms a homodimer that contains two template RNA molecules (18), and another group found that two separate, catalytically inactive TERT proteins can complement each other in trans to reconstitute catalytic activity (19). These results indicated that just one hTERT molecule and one hTR molecule alone could not reconstitute telomerase activity. In other words, hTERT and hTR molecules must form a multimer and reconstitute telomerase activity by working together. However, it was not clear whether disruptions to the oligomeric formation of hTERT reduce telomerase activity.
Here we demonstrate that two independent regions outside of motifs T to D have an important role in the oligomeric interaction of hTERT in vitro using purified recombinant hTERT in the absence of hTR, and in mammalian cells transiently co-expressing various tagged hTERTs. These two independent regions can interact. We also demonstrate that catalytically inactive truncated forms of hTERT, which contain the binding region can inhibit telomerase activity of the wild-type hTERT.

EXPERIMENTAL PROCEDURES
Plasmid Construction-Mammalian expression vectors: The plasmids pNKZ-FLAG, pNCZ-FLAG, and pNKZ-GST derived from pSG5UTPL, are mammalian expression vectors. pNKZ-FLAG or pNC-FLAG vector was used to express amino-or carboxyl-terminal FLAGtagged protein. pNKZ-GST was used to express amino-terminal GSTfused protein (4, 20, 21). The EcoRI-SalI fragment containing the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
hTERT cDNA was subcloned into the EcoRI-SalI sites of pNKZ-GST. pCI-Neo is derived from pCI-Neo-hTERT. To construct a carboxylterminal FLAG-tagged wild-type hTERT plasmid, the EcoRI-BamHI fragment containing the truncated hTERT cDNA was also subcloned into the EcoRI-BamHI site of pNCZ-FLAG. The BamHI-BamHI fragment containing the truncated hTERT cDNA of the carboxyl-terminal side, which was subcloned by PCR using appropriate primers, was subcloned into the BamHI site of the plasmid pNCZ-FLAG-hTERT containing the EcoRI-BamHI fragment.
All truncated hTERT constructs were subcloned by PCR using the appropriate primers, except ⌬CT1 and ⌬CT2. EcoRI-BamHI fragments containing truncated hTERT cDNA were subcloned into the EcoRI-BamHI sites of the mammalian expression vector pNKZ-FLAG. ⌬CT1 and ⌬CT2 were constructed by inserting the EcoRI-BamHI or EcoRI-XhoI fragment of the hTERT cDNA derived from pNKZ-FLAG-hTERT into the EcoRI-BamHI or EcoRI-SalI sites of the plasmid pNKZ-FLAG, respectively. All truncated forms of hTERT tagged with 2xHA epitopes at the carboxyl terminus were derived from pCI-Neo-hTERT-HA (9,22). Substitutions of aspartic acid with alanine and valine with isoleucine at positions 712 and 713, which was defective in substrate-binding and reported as a dominant negative form of hTERT, were introduced via PCR site-directed mutagenesis (23).
Baculovirus expression vectors: The amino-terminal GST-fused hTERT baculovirus expression vector pBKM-GST-hTERT was constructed by inserting the NotI-BglII fragment of the GST-hTERT cDNA derived from pNKZ-GST-hTERT into the NotI-BglII sites of the pVL1393 Baculovirus Transfer Vector (PharMingen) (4). The carboxylterminal FLAG-hTERT baculovirus expression vector pBKM-cFLAG-hTERT was constructed by inserting the EcoRI-BglII fragment of the carboxyl-terminal FLAG-hTERT cDNA derived from pNCZ-FLAG-hTERT into the EcoRI-BglII sites of the pVL1393 Baculovirus Transfer Vector. Other hTERT constructs for baculovirus expression were also constructed by inserting the NotI-BglII fragment of the hTERT cDNA derived from pNKZ-FLAG into the NotI-BglII sites of the pVL1393 Baculovirus Transfer Vector. These hTERT truncations contained a FLAG-epitope at the amino terminus.
Cells and the Generation of Recombinant Baculoviruses-Sf9 and High5 cells were cultured, and recombinant baculoviruses were prepared as described (4). COS-1 cells and TIG-3 cells were cultured by the standard method in Dulbecco;s modified Eagle's medium containing 10% fetal calf serum.
Immunoprecipitation-The hTERT constructs were transfected into COS-1 cells by calcium phosphate precipitation. COS-1 cells (1 ϫ 10 7 ) were sonicated in 600 l of buffer A and the cell lysate was stored at Ϫ80°C until use. The lysate was immunoprecipitated with 20 l of anti-HA monoclonal antibody (Santa Cruz Biotechnology, Inc.) immobilized on Protein A-Sepharose 4FF, rotated for 2 h at 4°C, and washed three times with washing buffer G (20 mM Tris-HCl, pH 7.5, 20% glycerol, 0.5% Nonidet P-40, 300 mM NaCl). The bound proteins were separated by SDS-PAGE then visualized by Western blotting.
The cell lysate was also incubated with 20 l of anti-FLAG M2 affinity gel for 1 h at 4°C and washed three times with washing buffer H (20 mM Tris-HCl, pH 7.5, 20% glycerol, 1.0% Nonidet P-40, 300 mM NaCl). The bound proteins were separated by SDS-PAGE then visualized by Western blotting. GST Pull-down Assays-Fifty microliters of glutathione-Sepharose (Amersham Biosciences, Inc.) was added to ϳ1 g of partially purified recombinant GST-fused hTERT or GST and FLAG-tagged hTERT proteins, and the mixtures were constantly rotated in binding buffer (20 mM Tris-HCl, pH 7.5, 20% glycerol, 0.05% Nonidet P-40, 0.25% MEGA-9, 150 mM NaCl, 1 mM dithiothreitol) for 1 h at room temperature. Thereafter, the beads were washed three times with binding buffer, then bound proteins were resolved on SDS-PAGE and visualized by Coomassie Brilliant Blue staining or Western blotting.
Telomerase Assay-The hTERT constructs were transiently transfected into telomerase-negative TIG-3 cells by calcium phosphate precipitation or with FuGENE 6 transfection reagent (Roche Molecular Biochemicals). Extracts were prepared from ϳ2.5 ϫ 10 4 transiently transfected TIG-3 cells using lysis buffers according to the manufacturer's protocol. Telomerase activity was measured by two methods. First, a PCR-based telomerase repeat amplification protocol (TRAP) assay was carried out with TeloChaser (TOYOBO Co. Ltd.). The products of the PCR were fractionated by electrophoresis on a 10% polyacrylamide gel and then visualized by staining with SYBR-Green I (Molecular Probes). Second, a TRAP enzyme-linked immunosorbent assay (ELISA) was used to quantitatively measure telomerase activity with a TRA-PEZE ELISA telomerase detection kit (Intergen Co. Ltd.).
Preparation of hTR RNA-hTR was prepared using the T7 in vitro transcription system as described for hTR-cDNA (pGRN164) (4).

Oligomeric Interaction of hTERT in Vitro and in Vivo-
To examine the homomeric interaction of hTERT, GST-fused and FLAG-tagged hTERT expressed in insect cells were purified using affinity chromatography (Fig. 1A, lanes 1-3), and the binding of FLAG-tagged hTERT to GST-hTERT was examined using the GST pull-down assay. Purified FLAG-hTERT was pulled-down by GST-hTERT in vitro (Fig. 1B, lane 5). This binding was specific, because GST alone could not bind FLAG-hTERT (Fig. 1B, lane 6). Importantly, the presence of human telomerase RNA (hTR) did not positively nor negatively affect this oligomeric interaction (Fig. 1C, lanes 7 and 8). We also confirmed that the oligomeric interaction occurred in vivo, because HA-and FLAG-tagged hTERT proteins transiently coexpressed in COS-1 cells were co-immunoprecipitated by anti-HA (Fig. 1D, lane 13) or by anti-FLAG M2 antibody (Fig. 1E, lane 20), respectively. These results indicate that hTERT proteins labeled with different tags interacted with each other in the oligomer form in vivo and in vitro. This interaction does not require hTR, strongly suggesting that hTERT has an intrinsic ability to oligomerize.
RNase Treatment Does Not Inhibit Homomeric Interaction of hTERT-To demonstrate that hTERT can oligomerize in vivo without template telomerase RNA, the lysate of COS-1 cells transiently co-transfected with HA-tagged hTERT and FLAGtagged hTERT was treated with RNase A, and then the binding of FLAG-hTERT to HA-hTERT was examined by co-immunoprecipitation using anti-HA antibody. Although telomerase activity of cell extract was diminished by RNase treatment (Fig.  2B), RNase treatment had no effect on homomeric interaction of hTERT ( Fig. 2A), indicating that intact telomerase RNA is dispensable in the structural oligomerization of hTERT but absolutely necessary for telomerase activity.
Two Regions Bind to the Wild-type hTERT in Vitro and in Vivo-If the homomeric interaction of hTERT does not require the presence of hTR, then the region(s) necessary for oligomeric interaction can be mapped using truncated forms of hTERT. Four truncations covering the amino-terminal region spanning the putative hTR-binding region (Fig. 3A, lanes 3, 4, 7, and 8), the carboxyl-terminal region harboring motif E, and the putative thumb domain (Fig. 3A, lanes 1, 2, 5, and 6), were constructed and the recombinant truncated proteins were purified from insect cells to examine the ability of these proteins to bind GST-hTERT by GST pull-down assay. Three constructs, CF1, NF1, and NF2, bound the wild-type GST-hTERT (Fig. 3B, lanes   11, 19, and 23), whereas CF2 had no binding ability in vitro (Fig. 3B, lane 15). These results indicate that at least the amino-and carboxyl-terminal regions can bind the wild-type hTERT in vitro. We then examined the binding abilities of the two regions in vivo using truncated versions of FLAG-hTERT in the presence of HA-hTERT in COS-1 cells (Fig. 4A, lanes  2-11), as well as the ability of the expressed FLAG-hTERT proteins to HA-hTERT by immunoprecipitation with anti-HA antibody. CF2 and PF did not bind the wild-type hTERT (Fig.  4A, lanes 19 and 22), whereas the other truncated constructs did (Fig. 4A, lanes 14 -18, 20, and 21). Under our experimental conditions, the expression levels of the truncated constructs of FLAG-hTERT proteins differed from those of HA-hTERT proteins. These mapping results in vivo and in vitro were consistent, indicating that the two independent regions are involved in the homomeric interaction. These two regions are located outside motifs T to D.
The Amino-and Carboxyl-terminal Regions of hTERTs Interact in Vivo-The interactions between the two regions of hTERT and the wild-type hTERT raised two possibilities. The truncated regions may only bind to the wild-type hTERT as a partner, or they may interact with various forms of hTERT. We therefore examined whether these two binding regions interact with each other. Truncated forms of FLAG-hTERT and HA-hTERT were transiently co-expressed in COS-1 cells (Fig. 5A,  lanes 1-5 and 11-15), then co-immunoprecipitated with anti-HA antibody. CF1, aa 914 -1132, could bind NF2, aa 301-534, indicating that the truncated amino-and carboxyl-terminal regions interact with each other (Fig. 5A, lane 9). CF2, aa 928 -1132, did not bind to any of the truncated proteins. These results strongly suggest that the amino acid sequence, aa 914 -927 including motif E, is critical for the binding. Fig. 5 shows no interaction between the differently tagged amino-or carboxylterminal regions (Fig. 5A, lanes 6, 7, 16, and 17; other data not shown). Thus, the homomeric interaction does not require the wild-type hTERT but does require the amino-and carboxylterminal regions. The oligomeric interaction of hTERT proteins may proceed in a head to tail fashion, because the amino-and carboxyl-terminal regions bound each other but not the homologous regions. Under these conditions, NF1, aa 201-534, bound weakly to CF1 compared with NF2 (Fig. 5A, lane 8). The discrepancy may be due to the limitations of using truncation mutants the structural integrity of which may be disrupted.
hTERT-D712A-V713I and Two Truncation Mutants, CF1 and NF2, Partially Inhibited the Telomerase Activity-To estimate the functional relevance of oligomerization of hTERT for telomerase activity in mammalian cells, finite normal human fibroblasts, TIG-3 cells, were transfected with HA-tagged wildtype hTERT in combination with inactive substituted mutation of hTERT-D712A-V713I at the VDV sequence, which is critical for substrate binding, or several truncated versions of hTERT. Using the lysate of these transfected cells, telomerase activity was measured by TRAP assay and TRAP ELISA. The telomerase activity of the wild-type hTERT was clearly inhibited by hTERT-D712A-V713I in TIG-3 cells compared with that of the wild-type hTERT alone in the TRAP assay (Fig. 6B, lanes 7 and  11) and TRAP ELISA (Fig. 6C). In the TRAP ELISA, the telomerase activity of the wild-type hTERT in TIG-3 cells was partially reduced in combination with hTERT-CF1 or hTERT-NF2, which could bind the wild-type, compared with that in combination with vector or CF2, which could not (Fig. 6C). In the TRAP assay, these differences were not clear among the presence of different mutant hTERTs (Fig. 6B, lanes 7-10). However, in TRAP ELISA experiments, telomerase activity of the wild-type hTERT was inhibited by the truncated mutants in a dose-dependent manner (Fig. 6D). This result strongly suggests that the truncation mutants, which can bind to the wild-type hTERT, have a negative effect on telomerase activity. DISCUSSION TERT is a unique enzyme among a family of nucleic acid-dependent polymerases harboring a fingers, palm, and thumb substructure, because it forms a tight complex with template RNA for the activity (2, 10, 24). Its long amino-and carboxylterminal parts outside of the fingers and palm (aa 525-928) might retain TERT-specific functions (11,22,(25)(26)(27)(28), because these parts are somewhat conserved only among TERTs (9 -11). We previously reported that hTERT and hTR are the minimum components required for telomerase activity when telomerase is reconstituted in vitro with two purified components (4). During this study we found that the catalytic activity of the purified hTERT was not detectable when concentrations of hTERT were low in the assay reaction (data not shown). This concentration dependence of hTERT reminded us of the template switching of telomerase previously reported in S. cerevi-
siae (17) and the oligomeric interactions of poliovirus (15) and hepatitis C virus RdRPs (29). Two groups recently reported the oligomeric role of telomerase using the different methods (18,19). However, it was not clear whether TERT forms multimer intrinsically or with help of template RNA. Here we show that the homomeric interaction of hTERT in vitro with partially purified differently tagged-hTERTs. Human TR seems to have no effect on the structural oligomerization of hTERT (Fig. 1C), and RNase treatment did not affect the homomeric interaction in vivo ( Fig. 2A) and in vitro (data not shown). These results indicate that hTERT has an intrinsic ability to oligomerize in the absence of intact hTR.
Two separate regions of hTERT (aa 301-538 and aa 914 -1132) can be mapped to bind the wild-type hTERT in vitro and in vivo. This result may be consistent with a previous report (25) in which some combinations of two different truncated hTERT mutants defective in telomerase activity reconstituted telomerase activity. Two regions we mapped are outside the fingers and palm substructure covering motifs T to D. The amino-terminal fragment, aa 301-538, overlaps the region important for hTR binding (25,26,30). The carboxyl-terminal fragment, aa 914 -1132, includes motif E and a putative thumb. The two separate regions can bind each other, but no homologous interaction of aa 301-538 or aa 914 -1132 was detected ( Fig. 4A and data not shown). The result seems to support a model that the homomeric interaction of hTERT occurs in a "head to tail" fashion. Our result cannot explain the previous result that the amino-terminal region (aa 1-300) and some truncated mutants (such as aa 301-1132) reconstituted telomerase activity (25). The region (aa 1-300) critical for telomerase activity (25, 26) may be not essential for the oligomerization but essential to interact with one or more critical factors such as hTR-binding proteins or Hsp90, which is recruited to telomerase by hTR or hTERT.
Functional oligomeric formations of telomerase have been proposed since two functional template RNAs in the telomerase complex having two active sites were found in S. cerevisiae and in humans (17,18). In our experiments, the substituted mutant hTERT-D712A-V713I, bound to the wild-type hTERT in vivo and in vitro (data not shown) and inhibited its telomerase activity (Fig. 6), suggesting that oligomeric formation of the wild-type and the mutant hTERTs is the reason of the inhibition. The D712A or D712A-V713I mutants at the VDV sequence, which is critical for substrate binding, have been previously described as dominant negative mutants that eliminate endogenous telomerase activity and cause telomere shortening and cell senescence due to lack of telomerase activity (23,31). However, the inhibitory effect of the mutant on the wild-type hTERT was not severe when transiently co-expressed in the telomerase-negative cells even if the amount of plasmid of the mutant was more than 10 times higher than the wild-hTERT (Fig. 6). The result seems to be consistent with the previous report by Beattie et al. (25) who observed a partial restoration of telomerase activity in the presence of hTERT-D712A and the truncated hTERTs harboring the intact pocket for active center but missing the amino-terminal region. These results suggest that the mutants defective in substrate binding (D712A or D712A-V713I) are not dominant negative enzymatically when these mutants oligomerized with the wild-type hTERT, although reconstituted telomerase activity of oligomers consisting of the wild-type and D712A or D712A-V713I hTERTs seems to be much weaker than that of the wild-type oligomer (Fig. 6) (25). The apparent dominant negative phenotype of D712A or D712A-V713I in the previous reports may be explained by a FIG. 6. The effect of hTERT mutants for telomerase activity of hTERT in vivo. A, lysate (100 g) prepared from the TIG-3 cells transiently transfected with the wild-type hTERT, hTERT-CF1, hTERT-CF2, hTERT-NF2, hTERT-D712A-V713I, or vector was assayed for telomerase activity by TRAP assay. B and C, cell lysate, 100 g (B) or 1 g (C), prepared from the TIG-3 cells transiently co-transfected both with 25 ng of the wild-type hTERT and with 500 ng of vector, hTERT-CF1, hTERT-CF2, hTERT-NF2, and hTERT-D712A-V713I were assayed for telomerase activity by TRAP assay (B) and TRAP ELISA (C). D, lysate (100 g) prepared from the TIG-3 cells transiently co-transfected both with 25 ng of the wild-type hTERT and with increasing amounts of hTERT-CF1 or hTERT-NF2 were assayed for telomerase activity by TRAP ELISA. Data are shown as the mean Ϯ S.D. huge difference in expression levels between endogenous hTERT and the ectopically expressed mutant hTERT, which may squelch out one or more critical factors for telomerase activity.
We expected that the truncated mutants, which have the hTERT-binding regions, would exhibit strong inhibitory effects on the wild-type hTERT by competing oligomerization of the wild-type hTERT. Rather weak inhibitory effects of the aminoterminal and the carboxyl-terminal binding regions may be due to weaker binding abilities to the wild-type hTERT or inefficient recruitment of the proteins to the wild-type hTERT.
Telomere maintenance is essential to the replicative potential of malignant cells, and inhibition of telomerase leads to telomere shorting and cessation of unrestrained proliferation (9,10,23,(31)(32)(33). The intrinsic property of hTERT to oligomerize may be an additional target to design specific inhibitors of telomerase. This strategy has been applied to HIV reverse transcriptase (14, 34 -36).