Vaults and Telomerase Share a Common Subunit, TEP1*

Vaults are large cytoplasmic ribonucleoprotein complexes of undetermined function. Mammalian vaults have two high molecular mass proteins of 193 and 240 kDa. We have identified a partial cDNA encoding the 240-kDa vault protein and determined it is identical to the mammalian telomerase-associated component, TEP1. TEP1 is the mammalian homolog of the Tetrahymena p80 telomerase protein and has been shown to interact specifically with mammalian telomerase RNA and the catalytic protein subunit hTERT. We show that while TEP1 is a component of the vault particle, vaults have no detectable telomerase activity. Using a yeast three-hybrid assay we demonstrate that several of the human vRNAs interact in a sequence-specific manner with TEP1. The presence of 16 WD40 repeats in the carboxyl terminus of the TEP1 protein is a convenient number for this protein to serve a structural or organizing role in the vault, a particle with eight-fold symmetry. The sharing of the TEP1 protein between vaults and telomerase suggests that TEP1 may play a common role in some aspect of ribonucleoprotein structure, function, or assembly.

Vaults are large cytoplasmic ribonucleoprotein complexes of undetermined function. Mammalian vaults have two high molecular mass proteins of 193 and 240 kDa. We have identified a partial cDNA encoding the 240-kDa vault protein and determined it is identical to the mammalian telomerase-associated component, TEP1. TEP1 is the mammalian homolog of the Tetrahymena p80 telomerase protein and has been shown to interact specifically with mammalian telomerase RNA and the catalytic protein subunit hTERT. We show that while TEP1 is a component of the vault particle, vaults have no detectable telomerase activity. Using a yeast three-hybrid assay we demonstrate that several of the human vRNAs interact in a sequence-specific manner with TEP1. The presence of 16 WD40 repeats in the carboxyl terminus of the TEP1 protein is a convenient number for this protein to serve a structural or organizing role in the vault, a particle with eight-fold symmetry. The sharing of the TEP1 protein between vaults and telomerase suggests that TEP1 may play a common role in some aspect of ribonucleoprotein structure, function, or assembly.
With a mass of 13 MDa, vaults are the largest ribonucleoprotein complexes known (1). Mammalian vaults are localized to the cytoplasm and are composed of a small RNA and three proteins of 100, 193, and 240 kDa in size (for review, see Refs. 2 and 3). The 100-kDa subunit, termed the major vault protein (MVP) 1 constitutes Ͼ70% of the particle mass. Its cDNA has been cloned from human, rat, Dictyostelium, and Discopyge ommata electric ray and their sequences are highly conserved both at the gene and protein level (4 -8). Likewise vault-associated RNA (vRNA) has been cloned from human, rat, mouse and bullfrog (9,10). vRNAs exhibit species-specific length variation ranging from 86 to 141 bases; however, all of the vRNAs can be folded into a similar secondary structure. Furthermore, vRNA is not a structural component of the vault particle as it makes up Ͻ5% of the vault mass, and its degradation does not result in the alteration of vault structure (1). All mammalian vaults contain two higher molecular weight vault proteins, p193 and p240. Recently, we have identified the p193 by its interaction with the MVP in a yeast two-hybrid screen and confirmed its identity by peptide sequence analysis (11).
Vaults have a unique barrel shape that is made up of two halves with each half vault capable of opening into a flower-like structure with eight petals surrounding a central ring. Each petal is formed by six copies of MVP, with 96 copies of MVP in the complete vault particle. Vaults are widely distributed throughout eukaryotes, and their morphology is highly conserved among these species, suggesting that their function is essential (12). Although the function of the vault particle has remained elusive, a portion of vaults have been localized to the cytoplasmic face of the nuclear membrane at or near nuclear pore complexes (13). Moreover, vault particle mass and symmetry are strikingly similar to the predicted mass of the putative central plug of the nuclear pore complex, leading to the proposal that vaults participate in nuclear-cytoplasmic transport (13). Recently, a reconstruction of the vault has been completed to 31-Å resolution indicating the hollow nature of the vault particle, consistent with a carrier/sequestration function (14). In support of this hypothesis, it has recently been discovered that, in response to estrogen, the estrogen receptor is found associated with vaults in a nuclear extract (15). Furthermore sea urchin vaults have been shown to localize to the nucleus and nucleolus in adult coelomocytes, the invertebrate equivalent of a macrophage (16). Vaults have also been determined to be up-regulated in some multidrug-resistant (MDR) cell lines, and one way vaults may function in MDR may be by regulating the amount of drug that reaches the nucleus (4,10). In Torpedo electric ray, high levels of vaults are found in the electric lobe that contains the cholinergic electromotor neurons innervating the electric organ (17). Vaults are localized in cholinergic nerve terminals in close proximity to synaptic vesicles. Thus, vaults are highly enriched in the electromotor system of electric rays where they are transported to the nerve terminal possibly along microtubules (18).
Here we demonstrate a direct connection between vaults and a well studied nuclear enzyme complex, telomerase. To complete our characterization of the protein components of vaults, we focused our attention on the higher molecular weight vault proteins, p193 and p240. Here we describe the identification of p240 as the mammalian telomerase-associated protein, TEP1. We demonstrate that TEP1 associates with vaults in biochemical fractionations and by immunoprecipitation analysis. In addition, we show that purified vaults do not have telomerase activity using the telomere repeat amplification (TRAP) assay. Using the yeast three-hybrid system, we show that several of the human vault RNAs interact with mouse TEP1 in a sequence-specific manner. The sharing of the TEP1 protein between two seemingly unrelated RNPs suggests that TEP1 may play a common role in some aspect of ribonucleoprotein structure, function, or assembly.

EXPERIMENTAL PROCEDURES
Peptide Sequence Analysis-Vaults were purified from monkey liver as described (14,19). Purified vaults were fractionated onto four 6% SDS-polyacrylamide gels and stained with copper, and the appropriate bands were excised. An estimated 24 pmol of the 240-kDa vault protein was sent to William S. Lane at the Harvard Microchemistry Facility. Peptide sequences were determined on a Finnigan TSQ-7000 Triple Quadrapole Mass Spectrometer with some residues being confirmed by Edman chemistry.
Biochemical Fractionation Analysis-HeLa cell extracts (S100 and P100) and discontinuous sucrose gradient fractionation of the P100 extracts were carried out as described (10). The 100,000 ϫ g pellet was resuspended by Dounce homogenization with a B pestle. Both S100 and P100 extracts and P100/sucrose gradient fractions were resolved by SDS-polyacrylamide gel electrophoresis and transferred to Hybond membrane (Amersham Pharmacia Biotech). The equivalent of 2 ϫ 10 6 cells were present in each lane of the S100 and P100 extract. Each of the sucrose gradient fractions also represented 2 ϫ 10 6 cells. The membrane was incubated with 0.1 g/ml affinity purified anti-TEP1 or a 1:500 dilution of affinity purified anti-vault antibody, followed by a horseradish peroxidase-conjugated secondary antibody and visualized by ECL (Amersham Pharmacia Biotech). For Northern analysis, equivalent aliquots of each of the sucrose gradient fractions were extracted for RNA and analyzed as described previously (10).
Immunoprecipitations-For immunoprecipitation from S100 and P100 extracts equivalent amounts of protein (ϳ800 g) were incubated with the following affinity purified antisera, anti-TEP1, anti-p240, or anti-vault for 2 h at 4°C. Protein A-Sepharose beads were added and further incubated for 2 h at 4°C. The beads were washed three times with buffer A (50 mM Tris-Cl (pH 7.4), 1.5 mM MgCl 2 , 75 mM NaCl) containing 1% Triton X-100, 1 mM DTT, and protease inhibitors and three times in phosphate-buffered saline. Precipitates were analyzed by immunoblotting (see above).
Antibodies-The specificity of the polyclonal anti-TEP1 antisera was described previously (25). Polyclonal anti-vault antisera were generated by injection of highly purified vaults into a rabbit as described previously (19). A fragment of p240 (amino acids 191-441) was expressed in the pET expression system (Novagen). The protein was purified on a His-bind column and injected into a rabbit. A p240 (191-441) containing glutathione S-transferase (GST) fusion protein was coupled to Affi-Gel 15 resin (Bio-Rad) to make an affinity column. Polyclonal antisera were affinity purified and used at 1:250 on Western blots.
Telomerase Activity Assays-Telomerase activity in purified vaults, S100, P100, and P100/sucrose gradient fractions was determined using the TRAP assay. The TRAPeze Telomerase Detection Kit (Intergen) was used as described by the manufacturer. About 10,000 cell equivalents of cellular extracts were examined. Telomerase extension was for 30 min at 30°C, and extended products were amplified by a two-step PCR (94°C for 30 s, 59°C for 30 s) for 22 cycles. Products were separated on 12.5% polyacrylamide gels and exposed to PhosphorImager screens (Molecular Dynamics). Because of the presence of a PCR inhibitor, telomerase extension of purified vaults (3 g) was followed by phenol:chloroform extraction and ethanol precipitation. The precipitated "extended" products were amplified and analyzed as described above.
Yeast Three-hybrid Plasmids and Analysis-Human vRNA genes (hvg1-4) were amplified by PCR. Digested, gel-purified PCR products were subcloned into the SmaI site in pMS2-1 downstream of the MS2 RNA (21). After sequence analysis the MS2-hvg (sense and antisense) hybrid genes were subcloned into pIIIEx426, a yeast expression vector (22). All other plasmids used, and the methods used in Fig. 5, are described in Ref. 20.

RESULTS
The 240-kDa Vault Protein Is Identical to TEP1-To identify and characterize the vault proteins, p193 and p240 (formerly referred to as p210), highly purified vaults from monkey liver were fractionated by SDS-polyacrylamide gel electrophoresis and stained, and the appropriate gel bands (p240) were excised and submitted to the Harvard Microchemistry Facility for peptide sequence analysis. Initially two peptide sequences were obtained (NQCLATLPD(I/L)K and FAQFDEYQLAK) and determined to be unique in a TBLASTN search of the nonredun-dant nucleotide sequence data base. A human cDNA library (kindly provided by Dr. Owen Witte, HHMI, UCLA) was screened with a 512-fold degenerate oligonucleotide to FAQFDEYQ. One clone of about 4 kb was identified and sequenced. A later BLAST search revealed that the p240 open reading frame sequence was identical to a newly identified human TEP1 sequence (initially identified as TP1 GenBank TM accession number U86136 (20)). Interestingly, sequence alignment of our truncated p240 open reading frame with TEP1 revealed the presence of a 1.5-kb intron in our clone at nucleotide 646 based on the TEP1 sequence. Similarly, the EST clone (H33937) encoding a portion of the rat TEP1 sequence was determined to contain an intron (23). Importantly, when we spliced out the 1.5-kb intron from our p240 clone, the remaining 2.5-kb open reading frame sequence was identical to that determined for the 5Ј-end of TEP1. Alignment of the p240 peptide sequences against the human TEP1 revealed the peptide locations to be amino acids 49 to 59 and 369 to 379. To exclude the possibility that only the amino-terminal portion of p240 was related to TEP1, further peptide sequence analysis of the p240 vault protein was carried out. This resulted in a peptide that aligned perfectly with amino acids 2553 to 2573 of TEP1 near the carboxyl terminus, confirming that p240 and TEP1 were identical.
TEP1 Associates with Vaults-To establish that this telomerase-associated subunit was indeed a vault particle component, vaults purified from rat liver were examined by immunoblot analysis. An anti-TEP1 antisera (25) recognized the p240 vault protein confirming its association with vaults, Fig.  1A. Previous studies using indirect immunofluorescence of vaults in fibroblast cells revealed a punctate cytoplasmic distribution, with subpopulations of vaults localized to the lamellopodia in freshly plated cells and at the tips of actin stress fibers (24). In subcellular fractions of HeLa cells, vaults are purified from the 100,000 ϫ g microsomal pellet (P100 (10, 19)). In contrast, most telomerase assays are routinely performed on the 100,000 ϫ g supernatant (S100) fraction. To determine where TEP1 fractionates, HeLa cells were fractionated into S100 and P100 extracts and analyzed by immunoblotting. TEP1 is present in both S100 and P100 fractions (Fig. 1B, top  panel, lanes 1 and 2). Further fractionation of the P100 extract on a discontinuous sucrose gradient revealed that the majority of the TEP1 is found in the 40/45% sucrose layer, coincident with the pattern observed for the major vault protein (MVP, Fig. 1B, both panels, lanes 5 and 6). These results suggest that the majority of the TEP1 in the P100 fraction is co-fractionating with the vault particle.
We have generated polyclonal anti-p240 antisera from a bacterially expressed fragment of the p240 (amino acids 191-441, see "Experimental Procedures"). The anti-p240 antisera recognized a single protein species of 240 kDa in P100 extracts from HeLa cells ( Fig. 2A) and in vitro translated recombinant TEP1 by immunoprecipitation analysis (data not shown). To further analyze the association of TEP1 with the vault particle, S100 and P100 extracts were immunoprecipitated using antivault, anti-p240, and anti-TEP1 antibodies, followed by immunoblot analysis with anti-TEP1 antisera. Immunoprecipitates from P100 extracts with an anti-vault antibody contain a 240-kDa protein recognized by the anti-TEP1 antisera (Fig. 2B,  lane 4). In contrast, immunoprecipitates from S100 extracts with an anti-vault antibody did not contain a 240-kDa protein (Fig. 2B, lane 8). As expected, immunoprecipitates from both S100 and P100 with anti-TEP1 and anti-p240 antisera contain a 240-kDa protein (Fig. 2B, lanes 2, 3, 6, and 7). Using the anti-vault, anti-TEP1, and anti-p240 antibodies, about equal amounts of the TEP1 protein are immunoprecipitated from the P100 extract (Fig. 2B, lanes 2-4). Our data indicate that the majority of the TEP1 in the P100 associates with the vault particle.
Vaults Do Not Have Telomerase Activity-Like telomerase, vaults are associated with a small RNA. In humans, there are multiple related vRNAs that vary from 86 to 99 bases in length. Four members of the human vRNA gene family have been identified to date; hvg1 is 96 bases, hvg2 and 3 are 86 bases in length (10), and hvg4 is 99 bases in length. hvg4 was identified by a search of the nonredundant nucleotide sequence data base and is located on chromosome X. We have determined that both hvg1 and hvg4 consistently copurify with vaults as evidenced by pelleting at 100,000 ϫ g and are present in all cell types examined. By contrast, the related vRNAs hvg2 and/or hvg3 are not present in all cell types examined, and when present FIG. 2. Coimmunoprecipitation of vaults and TEP1. A, immunoblotting of HeLa cell P100 extract with affinity purified anti-p240 antisera. B, the indicated antisera were used for immunoprecipitations from either P100 or S100 extracts. The precipitated proteins were then analyzed by immunoblotting with anti-TEP1 antisera (lanes 1-8). A nonspecific control IgG did not precipitate any cross-reactive protein (lanes 1 and 5). Lane 9 contains the P100 starting material. Protein markers are indicated on the left in kDa.

FIG. 3. Active telomerase complexes of variable sizes are present in discontinuous sucrose gradients.
A, Northern analysis of RNA extracted from S100, P100, and P100/sucrose gradient fractions hybridized with hTER (top panel) or hvg-specific probes (bottom panel). B, TRAP assays were performed on the indicated extracts (lanes 1-9). The lanes represent HeLa (total cells), S100, P100, P100/sucrose gradient fractions as indicated. Each lane corresponds to the activity present in 5,000 cells. The 36-bp internal TRAP assay standard (ITAS) is indicated.
FIG. 4. Vaults do not have telomerase activity. TRAP assays were performed on highly purified vaults from rat liver (3 g). The telomerase positive cells and TSR8 controls were supplied with the TRAPeze kit (Intergen). The telomerase positive control corresponds to the activity present in 500 cells. they do not consistently pellet at 100,000 ϫ g with vaults (10). However, all of the human vRNAs appear to be present in S100 extracts.
To determine whether the vRNAs and telomerase RNA (hTER) co-fractionate, total RNA was extracted from the S100, P100, and P100/sucrose gradient fractions and examined by Northern analysis. Like the vault proteins, the majority of the vault-associated RNAs (hvg1 and hvg4) are found in the 40/ 45% sucrose gradient fractions (Fig. 3A, bottom panel). However, significant amounts of all of the human vault RNAs are found in the S100. In contrast, the telomerase RNA fractionated throughout the sucrose gradient, with a slight peak in the 30% fraction (Fig. 3A, top panel). We next examined these fractions for telomerase activity using the TRAP assay (Fig.  3B). Telomerase activity was present in both the P100 and most of the P100/sucrose gradient fractions, suggesting that active telomerase complexes of varying size are present in these fractions and that they occasionally co-fractionate with vaults. Because telomerase activity partly cosedimented with vault particles in this partial purification, we next determined whether telomerase activity was also associated with highly purified vault particle preparations. Using the TRAP assay, we determined that purified vaults have no detectable telomerase activity (Fig. 4).
Vault RNAs Interact with TEP1-Mouse TEP1 has been shown to interact with mouse telomerase RNA (mTER) using a yeast three-hybrid assay (20). In this assay, the RNA of interest is fused to the hairpin RNA sequence of phage MS2 where transcriptional activation of the lacZ gene is mediated through a protein-RNA hybrid-protein interaction with the MS2 coat protein (21). Preliminary experiments indicated that the rat vRNA can be UV cross-linked in purified vaults to both the p193 and p240 vault proteins. 2 Using the yeast three-hybrid assay we have determined that the human vRNAs interact with TEP1 in a sequence-specific manner when cloned downstream of the MS2 RNA (Fig. 5). Coexpression of mTEP1 and the human vault RNAs (hvg1, hvg2, and hvg4) resulted in strong activation of the reporter genes HIS3 and lacZ and allowed growth on 3-aminotrizole medium and detection of ␤-galactosidase activity (Fig. 5, A and B). TEP1 also interacted with mTER but not with several control RNAs such as Drosophila tRNA arg , and antisense hvg1, hvg2, or hvg4 (Fig. 5A). Both sense and antisense hvg3 interacted with TEP1, therefore we could not determine whether the interaction of hvg3 with TEP1 was specific (data not shown). This assay indicates that mTER and three of the human vault RNAs specifically interact with TEP1. Because the three-hybrid assay cannot be used to examine binding of more than one RNA at a time, in vitro RNA-protein binding experiments will be required to determine whether the binding of hvg RNA and telomerase RNA compete for recognition by TEP1. DISCUSSION We have cloned the cDNA encoding the 240-kDa vault protein based on peptide sequence obtained from purified protein and determined it is identical to the telomerase-associated 2 V. Kickhoefer and L. Rome, unpublished data.
FIG. 5. Three-hybrid analysis of human vRNA interaction with murine TEP1. A, various MS2-tagged RNAs as indicated were cotransformed with a plasmid containing the putative RNA binding region of mTEP1 (the first ϳ870 amino acids, see Harrington et al. (20)) into the three-hybrid yeast strain L40-coat and patched on synthetic drop-out plates lacking uracil, leucine, and histidine and containing 10 mM 3-aminotriazole. An interaction between the tagged RNA and TEP1 allows growth on 3-aminotrizole. hvg1, hvg2, and hvg4 are the human vRNAs while gvh1, gvh2, and gvh4 refer to the antisense orientation of the human vRNAs relative to the MS2 hairpins. B, ␤-galactosidase assays on transformants shown in panel A and controls. Liquid ␤-galactosidase assays were performed in triplicate and quantified, with 10 Ϫ3 standard ␤-galactosidase units (420 nm) shown on the x axis. Error bars indicate standard deviation for each sample. Control cotransformations: mTER is wildtype mouse telomerase RNA; RETm is in the antisense orientation relative to the MS2 hairpins; tRNA is the Drosophila tRNA arg ; mTER/IRP, hvg1/IRP, tRNA/ IRP refers to cotransformations with the iron response RNA binding protein (21) instead of TEP1. protein, TEP1. Co-immunoprecipitation and subcellular fractionation, demonstrate that TEP1 is associated with the purified vault particle. The TEP1 protein contains a number of interesting domains including an amino-terminal repeat domain, an RNA binding domain, an ATP/GTP binding motif, and 16 WD40 repeats in the carboxyl terminus (20,23). We also demonstrate that purified vaults have no apparent telomerase activity and that several of the human vRNAs are capable of interacting with TEP1 in a yeast three-hybrid assay. The basis for the sharing of this protein between vaults and telomerase is unclear. It is tempting to speculate that the TEP1 protein may be acting as a chaperone aiding in the assembly of vault particles and possibly in the assembly of telomerase complexes. The role of TEP1 in the telomerase complex has not yet been defined (20,23). However, immunoprecipitates of TEP1 have been shown to contain telomerase activity and TEP1 associates with TERT and hTER in cell lysates (20,23,25,26). The exact role of TEP1 in the telomerase complex is unclear, as telomerase activity can be reconstituted in vitro with just TERT and hTER (27)(28)(29). As the minimal telomerase complex does not appear to be dependent on the presence of TEP1, perhaps it has a more general role in telomerase complex assembly. Possibly the binding of TEP1 to a specific RNA is what determines whether it associates with vaults or telomerase in vivo. Using a yeast three-hybrid system, murine TEP1 has been shown to interact specifically with mTER (20). Similarly, we demonstrate that several of the human vRNAs (hvg1, hvg2, and hvg4) specifically interact with TEP1 in a three-hybrid assay. More direct binding experiments will be required to determine whether the binding sites for the telomerase RNA overlaps or is distinct from the binding site for human vRNAs. TEP1 may also have a role in protecting the RNAs from degradation, thus maintaining an available pool of TEP1/TER or TEP1/vRNA for future assembly into either complex. It is also possible that TEP1/TER might have additional functions since both are present in cells that do not possess telomerase activity (data not shown).
The presence of the 16 WD40 repeats in TEP1 is a convenient number for this protein to serve a structural or organizing role in the vault, a particle with 8-fold symmetry. Heterotrimeric G proteins containing seven WD40 repeats have been shown to form a seven-blade propeller-like structure (30). Evidence from cryoelectron microscopy and stoichiometry calculations suggests that the 96 copies of the MVP form the barrel of the vault particle and that the barrel is hollow (14). The stoichiometry of the two higher molecular weight vault proteins is more difficult to predict from mass and cryoelectron microscopy data. Yeast two-hybrid analysis indicates that MVP interacts with the 193-kDa vault protein (11). We suggest that the potential 8-fold symmetry of the TEP1 WD40 repeats could support a role for TEP1 in the elaboration of the vault particle structural symmetry, possibly at the ends of the vault particle.
Although the function of the vault particle is not yet known, the sharing of the TEP1 protein with telomerase may aid in elucidating vault function. The sharing of protein subunits is not a new finding, and several examples have been identified (for review, see Ref. 31). Cytosolic aconitase and the iron regulatory binding protein have been shown to be the exact same protein with two mutually exclusive functions that are regulated by cellular iron levels (32). When cellular iron levels are high, the protein binds iron-sulfur clusters and acts as a cytosolic aconitase converting citrate to isocitrate. However, when iron stores are low, the protein is no longer able to bind iron and subsequently acts as an RNA binding protein, binding to the IRE element upstream of ferritin thus translationally regulating the iron storage protein levels. Crystallins are stable proteins required at high concentrations in the lens. There are two classes of vertebrate crystallins: the ubiquitous highly conserved ␣-, ␤-, and ␥-crystallins, and the species-specific ␦-, -, ⑀-, and -crystallins. All of the species-specific crystallins are also metabolic enzymes (for review, see Refs. 33 and 34). The term gene-sharing has been used to describe the enzyme/crystallin genes that encode a protein with two different functions. The sharing of the TEP1 protein between vaults and telomerase, two seemingly unrelated RNP complexes, suggests that TEP1 may be involved in some aspect of RNP regulation. It is not known what determines whether TEP1 associates with vaults or telomerase or whether there is a dynamic equilibrium between the association of TEP1 with vaults or telomerase. It is even possible that TEP1 is a part of other RNPs. TEP1 interaction with vaults might be via a TEP1-RNA, TEP1-protein, or through a combination requiring both TEP1-RNA and TEP1protein interactions. It is also possible that vaults themselves function as potential chaperones, possibly delivering TEP1 to the nucleus or aiding in the assembly of TEP1 with other telomerase components in the cytoplasm and then assisting in their delivery to the nucleus. The precise role of TEP1 in the de novo assembly of either of these RNPs will be the subject of further studies.