The Association of the Human PM/Scl-75 Autoantigen with the Exosome Is Dependent on a Newly Identified N Terminus*

The exosome is a complex of 3 (cid:1) 3 5 (cid:1) exoribonucleases that functions in a variety of cellular processes, all con-cerning the processing or degradation of RNA. Paradox-ically, the previously described cDNA for the human autoantigenic exosome subunit PM/Scl-75 (Alderuccio, F., Chan, E. K., and Tan, E. M. (1991) J. Exp. Med. 173, 941–952) encodes a polypeptide that failed to interact with the exosome complex. Here, we describe the cloning of a more complete cDNA for PM/Scl-75 encoding 84 additional amino acids at its N terminus. We show that only the longer polypeptide is able to associate with the exosome complex. This interaction is most likely mediated by protein-protein interactions with two other exosome subunits, hRrp46p and hRrp41p, one of which was confirmed in a mammalian two-hybrid system. In addition we show that the putative nuclear localization signal present in the C-terminal region of PM/Scl-75 is sufficient, although not essential for nuclear localization of the protein. Moreover, the deletion of this element ab-rogated the nucleolar accumulation of PM/Scl-75, although its association with the exosome was not dis-turbed. This suggests that this basic element turer’s procedure in a total volume of 25 (cid:3) l, containing (cid:1) 1 (cid:3) g of circular plasmid DNA (pCI-neo; Promega), containing the coding sequence of the protein, and in the presence of [ 35 S]methionine. Reverse Transcriptase-PCR— RNA was isolated from cells using Trizol reagent (Invitrogen) according to the manufacturer’s instruc-tions. Synthesis of cDNA from 1 (cid:3) g of RNA was performed using a reverse transcription system (Promega) with 10 pmol of PM/Scl-75-specific primer. The reaction mixture was incubated for 10 min at room temperature, followed by 90 min at 42 °C. For RT-PCR 1 (cid:3) l of cDNA was added to 49 (cid:3) l of the reaction mixture containing 75 m M Tris-HCl, pH 8.8, 20 m M (NH 4 ) 2 SO 4 , 0.01% (v/v) Tween 20, 2 m M MgCl 2 , 0.2 m M dNTPs, 20 pmol of forward and reverse primer, and 8 units of Taq DNA polymerase. PCR was performed using a T3 thermocycler (Biometra): 2 min at 94 °C; 35 cycles of 30 s at 94 °C, 30 s at 60 °C, and 30 s at 72 °C; and 5 min at 72 °C. 10 (cid:3) l of PCR product was analyzed on a 1.5% agarose gel. The following primers were used: RT primer, 5 (cid:2) -CAGGT- GTAGAAACAACTTCTGAAGGAGG-3 (cid:2) ; PM/Scl-75 reverse primer, 5 (cid:2) -CACCAAGAGATCTGACTGCCTGCCAG-3 (cid:2) ; PM/Scl-75a forward primer, 5 (cid:2) -AGATCTCGAG CCTGTATGGGCGGGCTGGTTAGGATTC-3 (cid:2) ; and PM/Scl-75c forward primer, 5 (cid:2) -AGATCTCGTCGACCGAATTC- l of Renilla luciferase substrate solution (containing a quencher for the firefly luciferase activity) was added, and again the luminescence was determined to monitor the transfection efficiency.

ases (8), the PM/Scl complex was shown to be related to the yeast exosome, a complex containing ϳ10 exoribonucleases (9).
The exosome functions in a variety of processes involving the 3Ј 3 5Ј processing or degradation of RNA. Among these processes are the maturation of 5.8 S rRNA (10 -12), the processing of many small nuclear and nucleolar RNAs (13)(14)(15), and the turnover of different types of mRNAs (16,17), especially the AU-rich element containing mRNAs (18,19). PM/Scl-75 has been suggested to be an AU-rich element binding protein, involved in the recruitment of the exosome to this class of mRNAs (19).
Like five other exosome proteins, PM/Scl-75 contains an RNase PH domain, which is homologous to the prokaryotic 3Ј-5Ј exoribonuclease RNase PH. Based on mutual interactions between exosome components and structural similarity with the bacterial protein polynucleotide phosphorylase (PNPase), a component of the bacterial degradosome, recently a model for the structure of the human exosome was generated. In this model the six proteins containing an RNase PH domain form the core of the exosome, which adopts a hexameric ring structure (20).
Cloning of a cDNA encoding the human PM/Scl-75 protein revealed an acidic polypeptide sequence of only 39.2 kDa. Its migration at 75 kDa in SDS-PAGE was explained by its highly charged C-terminal half (1). In immunofluorescence experiments the protein showed mainly nucleolar localization (1), although it was also found to be present in the cytoplasm and nucleoplasm of human cells (19). The N-terminal half of the protein shows homology to the bacterial RNase PH protein (8), and the C-terminal half is characterized by a region of ϳ100 amino acids highly enriched in acidic residues. Close to the C terminus a putative nuclear localization signal (NLS) is located consisting of seven basic residues (KRRKKKR) (1). This cDNA of PM/Scl-75 has been used in a number of transfection experiments to study its subcellular localization and association with the exosome. Although similar experiments with most other exosome components were successful, EGFP fusion proteins of PM/Scl-75 were unable to enter the nucleolus, and PM/Scl-75 failed to interact with other exosome components in a mammalian two-hybrid system (20).
Here, we show that the previously described protein sequences of PM/Scl-75 are most likely incomplete. We have cloned and characterized additional cDNAs encoding PM/ Scl-75 that in comparison with the original sequence encode 84 additional amino acid residues at its N terminus. In contrast to the previously described PM/Scl-75 polypeptide, this longer variant appeared to be able to associate with the exosome and to accumulate in the nucleolus.

EXPERIMENTAL PROCEDURES
cDNA Cloning and Construction of Deletion Mutants-All cDNAs used were cloned into pACT and pBIND vectors (Promega) or suitable * This work was supported in part by the Council for Chemical Sciences of the Netherlands Organization for Scientific Research. 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.
Bioinformatics-Predictions of functional domains and putative NLSs were done using the Pfam data base of Hidden Markov models (21) and the PredictNLS server (22), respectively. Sequence data base searching was performed using BLAST on GenBank TM data bases containing either ESTs or genomic sequence information.
Immunoblot Analysis-For immunoblot analysis, the proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. To visualize the proteins, the blots were incubated with autoimmune patient or rabbit sera diluted 5000-and 500-fold, respectively, in blocking buffer (4% skimmed milk, phosphate-buffered saline (PBS), 0.1% Nonidet P-40). As secondary antibody, horseradish peroxidase-conjugated rabbit-anti human IgG or swine anti-rabbit IgG (Dako Immunoglobulins) were used, 5000-fold diluted in blocking buffer. Visualization was performed by chemiluminescence.
In Vitro Translation-Radioactively labeled proteins were produced using a reticulocyte lysate system (Promega) according to the manufacturer's procedure in a total volume of 25 l, containing ϳ1 g of circular plasmid DNA (pCI-neo; Promega), containing the coding sequence of the protein, and in the presence of [ 35 S]methionine.
Transient Transfection of HEp-2 Cells and Direct Immunofluorescence-For transfection, cDNAs were cloned into suitable pEGFP vectors (Clontech), allowing expression of the proteins fused to the C terminus of the EGFP protein. HEp-2 (human epithelioma, ATCC number CCL-23) cells were grown to 80% confluent monolayers by standard tissue culture techniques in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal calf serum. For immunoprecipitation, ϳ10 ϫ 10 6 cells were transfected with 20 -30 g of DNA in 1600 l of Dulbecco's modified Eagle's medium containing 10% fetal calf serum by electroporation, which was performed at 270 V and 950 microfarad using a Gene-Pulser II (Bio-Rad). After transfection, the cells were seeded in 75-cm 2 culture flasks and cultured overnight. After washing twice with PBS, the cells were resuspended in 500 l of lysis buffer (25 mM Tris-HCl, pH 7.5, 100 mM KCl, 2 mM EDTA, 1 mM dithioerythritol, 0.5 mM phenylmethylsulfonyl fluoride, and 0.05% Nonidet P-40) and homogenized by sonication. For fluorescence microscopy, ϳ2 ϫ 10 6 cells were transfected with 10 -20 g of DNA in 800 l of Dulbecco's modified Eagle's medium containing 10% fetal calf serum by electroporation, as described above. After transfection, the cells were seeded onto coverslips and cultured overnight. The cells were washed twice with PBS, fixed with 4% paraformaldehyde in PBS for 20 min, washed in PBS twice, briefly rinsed in acetone, dried, and finally mounted with PBS/ glycerol. The expressed EGFP-tagged proteins were visualized by fluorescence microscopy.
Immunoprecipitation-Polyclonal antibodies from rabbits and patients were coupled to protein A-agarose beads (Biozym) in IPP500 (500 mM NaCl, 10 mM Tris-HCl, pH 8.0, 0.05% Nonidet P-40) by incubation for 2 h at room temperature. The beads were washed twice with IPP500 and once with IPP150 (150 mM NaCl, 10 mM Tris-HCl, pH 8.0, 0.05% Nonidet P-40). For each immunoprecipitation, the cell extract was incubated with the antibody-coupled beads for 1 h at 4°C. Subsequently, the beads were washed three times with IPP150, and the co-immunoprecipitated proteins were analyzed by immunoblotting.
Mammalian Two-hybrid Analysis-All of the interactions were analyzed using the CheckMate mammalian two-hybrid system (Promega), essentially according to the manufacturer's protocol. Briefly, 3-4 ϫ 10 5 COS-1 cells seeded in one well of a 6-well plate were transfected with three vectors (1 g each), pACT and pBIND, either with or without insert, and the pG5luc reporter vector using 5 l of FuGENE Transfection Reagent (Roche Applied Science), as described by the manufacturer. After 40 -48 h of growth, the cells were harvested using 500 l of Passive Lysis Buffer (Promega), and the activity of both the firefly luciferase and the control Renilla luciferase were determined using a dual luciferase reporter assay system (Promega) on a Berthold Lumat LB 9507 Luminometer. In brief, 100 l of firefly luciferase substrate solution was added to 20 l of cell extract, and the luminescence was measured to determine the efficiency of the interaction. Next, 100 l of Renilla luciferase substrate solution (containing a quencher for the firefly luciferase activity) was added, and again the luminescence was determined to monitor the transfection efficiency.

RESULTS
Amino Acid Sequence of PM/Scl-75-Several amino acid sequences for the human PM/Scl-75 polypeptide have been reported previously. Besides the first sequence ever described (here referred to as PM/Scl-75a-␣, accession number M58460) (1), a splicing variant containing the sequence encoded by an additional exon (PM/Scl-75a-␤, accession number HSU09215) and a sequence containing 68 additional amino acid residues at the N terminus (PM/Scl-75b-␣, accession number Q06265) have been reported. Moreover, the sequence reported for Mus musculus PM/Scl-75 (accession number Q9JHI7) contains, compared with the human PM/Scl-75a-␣, 84 additional amino acids residues at the N terminus (23). The reported heterogeneity in sequences for the N-terminal region of the human PM/Scl-75 prompted us to screen the human EST data bases with the available human and mouse PM/Scl-75 cDNA sequences. These analyses demonstrated that in humans PM/Scl-75 sequences are expressed that fully correspond to the N-terminal region of the mouse protein. Thus, in comparison with PM/Scl-75a the human protein may also contain 84 additional amino acids. Based upon these data extended open reading frames for the human PM/Scl-75 were generated (PM/Scl-75c-␣ and PM/Scl-75c-␤) and submitted to the EMBL data base under accession numbers AJ505989 and AJ517294. Fig. 1a shows the sequence of PM/Scl-75c, in which the differences with previously described PM/Scl-75 sequences are indicated. In Fig. 1b a schematic overview of the human polypeptides and their mRNAs is shown. It is important to note that PM/Scl-75a lacks a significant part of the RNase PH domain as predicted by Pfam (21). The RNase PH domain is expected to be important for the interactions of PM/Scl-75 with other exosomal proteins in the core of the exosome (20).
When the cDNA sequences are compared with the human genome, it is clear that the difference between the sequences is due to the use of an apparent alternative promotor in the PM/Scl-75 gene or to alternative splicing of the PM/Scl-75 pre-mRNA, as is illustrated in Fig. 1c. The 5Ј end of the cDNAs of PM/Scl-75c is encoded by an alternative first exon and lacks the first 234 nucleotides of the first exon of PM/Scl-75a. The protein sequence of PM/Scl-75b was derived from a cDNA very similar to that of PM/Scl-75a. The PM/Scl-75b cDNA contains a stretch of 8 instead of 7 thymidine residues at position 395-402 (which may be due to a sequencing error of the original clone), leading to a longer open reading frame (23). To investigate which of the corresponding mRNAs are actually expressed in human cell lines, RNA isolated from HEp-2, HeLa, 293, Jurkat, and MOLT-4 cells was analyzed by RT-PCR using primers specific for the PM/Scl-75a/b and PM/Scl-75c mRNAs. Both mRNAs were found to be present in all cells lines (results not shown). To discriminate between PM/Scl-75a and PM/Scl-75b, the RT-PCR products were cloned and sequenced. The results showed that the T-stretch consisted of 8 thymidine residues in all cases, indicating that the expressed mRNA corresponds to PM/Scl-75b. Strikingly, however, no evidence for the existence of mRNAs encoding either PM/Scl-75a or PM/Scl-75b was found in the EST data bases, strongly suggesting that PM/Scl-75c is the most abundantly expressed isoform of the protein. To generate cDNAs encoding PM/Scl-75c, we performed PCR on human teratocarcinoma and placenta cDNA libraries using primers specific for the 5Ј region of PM/Scl-75c. The PCR products obtained were sequenced and found to be identical to the sequence of the cDNA encoding PM/Scl-75c as derived from the EST data bases.
The polypeptides encoded by the PM/Scl-75a-␣ and PM/Scl-75c-␣ cDNAs were produced by in vitro transcription/translation, and their migration in SDS-PAGE gels was compared with that of PM/Scl-75 from a cytoplasmic HeLa cell extract, which was detected by immunoblotting using anti-PM/Scl-75 rabbit antibodies (Fig. 2). The results showed that the migration of PM/Scl-75c was identical to that of HeLa cell PM/Scl-75. The migration of in vitro translated PM/Scl-75a was clearly different, in agreement with previous observations (1).
To investigate the occurrence of the extra exon (10*) that is present in the ␤ variant of PM/Scl-75 (Fig. 1d, exon a), the GenBank TM EST data base was screened for sequences containing the exon-exon junctions 10*-11 and 10 -11. In total, 52 ESTs encompassing this region were identified, 10 of which contained exon 10*, indicating that both splice variants of PM/Scl-75 are expressed but that the isoform lacking exon 10* may be more abundant.
Taken together, these results indicate that PM/Scl-75c-␣ is the predominant isoform of PM/Scl-75 but that at least four splice variants exist that differ by their N terminus and the presence or absence of 17 amino acids encoded by exon 10*.
The Subcellular Localization of PM/Scl-75-To investigate the subcellular localization of PM/Scl-75 polypeptides, constructs encoding PM/Scl-75a-␣, PM/Scl-75c-␣, and PM/Scl-75c-␤ tagged with EGFP were generated. The fusion proteins were expressed in transiently transfected HEp-2 cells and as a control the endogenous PM/Scl-75 protein in HEp-2 cells was visualized by immunofluorescence using rabbit anti-PM/Scl-75 antibodies. Whereas the highest concentration of the endogenous PM/Scl-75 protein was found in the nucleoli (Fig. 3A), EGFP-PM/Scl-75a-␣ failed to enter the nucleoli and accumulated in the nucleoplasm (Fig. 3B). In contrast, EGFP-tagged PM/Scl-75c-␣ and PM/Scl-75c-␤ efficiently entered the nucleo-lus of HEp-2 cells (Fig. 3, C and D). These data indicate that the N-terminal 84 amino acids of PM/Scl-75c are important for nucleolar accumulation and that the 17 amino acids encoded by exon 10* do not affect this process. To investigate whether the N-terminal region of PM/Scl-75c is sufficient for nucleolar targeting, the N-terminal 98 amino acids of PM/Scl-75c were fused to the N terminus of EGFP, and the subcellular localization of this fusion protein was analyzed in transfected HEp-2 cells. This fusion protein distributed throughout HEp-2 cells, similar to EGFP alone (data not shown). Previously, a sequence element (KRRKKKR) with similarity to NLSs has been reported to reside at the C-terminal end of PM/Scl-75 (1). To investigate the role of this element in nuclear and nucleolar entry, mutants of PM/Scl-75a-␣ and PM/Scl-75c-␣ lacking the C-terminal 39 amino acids were generated (⌬NLS). As can be seen in Fig. 3E, PM/Scl-75a-␣ ⌬NLS failed to enter the nucleus. Remarkably, PM/Scl-75c-␣ ⌬NLS still was transported to the nucleus but was unable to enter the nucleolus (Fig. 3F). Finally, an EGFP fusion protein was expressed containing only the C-terminal 39 amino acids of PM/Scl-75. Fig. 3G shows that this element is sufficient for transportation of EGFP to the nucleus. These data show that both N-and C-terminal elements of PM/Scl-75 are involved in its nuclear entry and that the C-terminal elements (also) play a role in nucleolar targeting.
The Association of PM/Scl-75 with the Exosome-In addition to elements directly involved in subcellular transport processes, the association with the exosome may also play a role in the subcellular localization. To investigate the effect of the mutations on the association with the exosome, immunoprecipitations with anti-EGFP antibodies were performed using lysates of HEp-2 cells transfected with constructs encoding the EGFP-tagged PM/Scl-75 mutants. The immunoprecipitated material was analyzed by Western blotting, using a human serum (Ven96) reactive with several exosome proteins (including PM/Scl-100, PM/Scl-75, hRr4p, hRrp41p, and hRrp42p). The patient serum stained the different PM/Scl-75 fusion proteins very efficiently, which shows that all variants/mutants were expressed at similar levels. The results showed that many endogenous exosome proteins (although PM/Scl-100 was not  detected in the immunoprecipitates) co-precipitated with both PM/Scl-75c-␣ and PM/Scl-75c-␣⌬NLS but not with PM/Scl-75a-␣ (Fig. 4a). In a similar type of experiment, we demonstrated that both the ␣ and ␤ form of PM/Scl-75c can be incorporated in the exosome complex (Fig. 4b).
Previously, we demonstrated that PM/Scl-75a-␣ was not able to interact directly with any other exosome component in a mammalian two-hybrid system, although its yeast counterpart (Rrp45p) is known to interact with yeast Rrp41p (20). The full-length cDNAs encoding PM/Scl-75a-␣, PM/Scl-75c-␣, and 10 (putative) human exosome components (hRrp4p, hRrp40p, hRrp41p, hRrp42p, hRrp46p, PM/Scl-100, hCsl4p, hRrp44p/ hDis3p, OIP2, and hMtr3p) were cloned in both the pACT (in-frame with the sequence encoding the VP16 transcription activation domain) and pBIND (in-frame with the sequence encoding the GAL4 DNA-binding domain) vectors of the Checkmate mammalian two-hybrid system. COS-1 cells were cotransfected with each pair of these constructs, and after 40 -48 h the luciferase activity of extracts prepared from these cells was determined. This revealed that, in contrast to PM/Scl-75a-␣, PM/Scl-75c-␣ fused to the GAL4 DNA-binding domain was able to interact with hRrp41p fused to the VP16 activation domain (Fig. 5). This interaction was confirmed by a complementary experiment in which the fusion domains were exchanged (data not shown). The interaction of PM/Scl-75c with hRrp41 adds further evidence to the model for the human exosome, in which PM/Scl-75 is flanked by hRrp41p and hRrp46p (20).
Alternative Splicing of PM/Scl-75-As can be seen in Figs. 3 and 4, the subcellular localization and exosome association of the PM/Scl-75c-␤ isoform, containing the extra exon 10*, in living cells was indistinguishable from that of the PM/Scl-75c-␣ isoform. The data from the EST libraries (see above) suggested that the ␣ isoform is more abundantly expressed than the ␤ isoform. To shed more light on this issue, we studied their expression in several cell lines by Western blotting. The ratio between the amounts of the ␣ and ␤ isoforms appeared to vary significantly between different cell lines, although in most cells analyzed PM/Scl-75c-␣ was at least as abundant as PM/Scl-75c-␤. Fig. 6 shows that in Jurkat, HeLa, and HEp-2 cells (lanes 1-3) the amount of both isoforms is comparable. In COS1 and 293 cells (lanes 4 and 5) the ␣ isoform of PM/Scl-75 appears to be the most abundant one. In HeLa cells, the relative level of PM/Scl-75-␤ was significantly higher in nuclear extracts than in cytoplasmic extracts (lanes 6 and 7), suggesting a predominant nuclear localization for PM/Scl-75c-␤. DISCUSSION In this study we have shown that the previously reported cDNA and amino acid sequences for the human polymyositis/ scleroderma autoantigen PM/Scl-75 are probably incomplete. We cloned a cDNA with an extended N terminus, and only the longer form of PM/Scl-75 was able to associate with the exosome complex, to interact with at least one other exosome subunit, and to enter the nucleolus, three functional activities that are shared with other exosome subunits. The presence of exon 10* in the coding sequence of PM/Scl-75 (observed in a previously reported splice variant) did not affect the behavior of PM/Scl-75 with regard to its subcellular localization and exosome association. In addition, PM/Scl-75 was shown to contain both N-and C-terminal elements involved in its nuclear localization, the latter of which also appeared to be responsible for nucleolar accumulation of this protein.
The Amino Acid Sequence of PM/Scl-75-The cDNA of PM/ Scl-75a-␣ (1) lacks the first exon compared with the cDNA of PM/Scl-75c-␣, and as a consequence the open reading frame starts at another methionine. Note that the 5Ј end of the PM/Scl-75b-␣ cDNA is identical to that of PM/Scl-75a-␣, with the exception of one additional nucleotide in the former, leading to a longer open reading frame in the 5Ј region (23). The expression of mRNAs corresponding to these two types of cDNAs is most likely due to the existence of alternative promoters and transcription initiation sites in the gene of this protein (Fig. 1c). In contrast to PM/Scl-75c-␣ the expression of PM/Scl-75a-␣ and PM/Scl-75b-␣ mRNAs is not supported by entries in the EST data bases. A TBLASTN search of the human EST data base with the N-terminal 80 amino acids of the mouse sequence resulted in 49 of 50 hits that fully matched the N terminus of the additional sequence. No ESTs were observed encoding the sequence corresponding to the N terminus of PM/Scl-75a or PM/Scl-75b. Nevertheless, the latter was detected in several cell lines by RT-PCR, followed by sequencing of the products.
The lack of expression of PM/Scl-75a-␣ is supported by the observation that the corresponding in vitro translated protein migrates faster than the PM/Scl-75 protein from HeLa cells (Fig. 2) (1). However, immunoblotting of primate cell extracts with rabbit anti-PM/Scl-75 antibodies revealed several minor polypeptide bands, including one with an apparent molecular mass of ϳ60 kDa (Figs. 2 and 6), which is similar to the migration of PM/Scl-75a-␣. Taken together, these data suggest that the N-terminal region of the major PM/Scl-75 isoform corresponds to the sequence encoded by PM/Scl-75c-␣ and that as a result of an alternative promoter PM/Scl-75b-␣ is pro-duced. The expression of mRNAs either containing or lacking exon 10* (Fig. 1d) is supported by a series of data base entries and thus is consistent with the expression of two splicing variants (designated ␣ and ␤). Western blot analyses of extracts of several primate cell lines indeed show a doublet for PM/Scl-75, in agreement with the simultaneous expression of these splice variants in a variety of cells. Interestingly, the results of HeLa cell fractionation experiments suggested that the ␤ isoform might accumulate somewhat more efficiently in the nucleus than the ␣ isoform.
Transport of PM/Scl-75 to Nucleus and Nucleolus-Previously, the C-terminal element of PM/Scl-75, rich in basic amino acids, was proposed to be a nuclear localization sequence. Here, we have shown that this element indeed can direct a reporter protein to the nucleus. However, this element did not appear to be essential for nuclear import of PM/Scl-75, as long as PM/ Scl-75 was able to interact with the rest of the exosome complex. On the other hand, the basic sequence element of PM/ Scl-75 is important for nucleolar accumulation of this protein.
Based upon the presence of putative NLS sequences in at least two other exosome proteins, hRrp41p (amino acids 85-90, ERKRRP) and PM/Scl-100 (amino acids 752-758, AKKRERA), two mechanisms for nuclear entry of the exosome can be envisaged. First, the nuclear import of the completely assembled exosome may be mediated by the concerted action of several signals. Second, partially assembled parts of the exosome may enter the nucleus and assemble into a complete exosome in the nucleoplasm or nucleolus. All of these basic elements may also be involved in nucleolar targeting of the exosome, because elements rich in basic residues have been demonstrated to play an essential role in nucleolar accumulation (24).
Association of PM/Scl-75 with the Exosome-The direct interaction between PM/Scl-75c and hRrp41p supports the model for the human exosome that we recently proposed, as illustrated in Fig. 7 (20). The minor isoform of the protein, PM/Scl-75b, most likely is also capable of interacting with hRrp41p and the exosome, because it contains the complete RNase PH domain (Fig. 1b). Another interaction predicted by that model, between PM/Scl-75 and hRrp46p (Fig. 7), could not be detected in the mammalian two-hybrid system. PM/ Scl-75 represents together with hRrp41p the equivalent of one PNPase subunit in the structurally related PNPase trimer, and the two RNase PH domains of PNPase interact with FIG. 5. Two-hybrid interactions between PM/Scl-75a-␣, PM/ Scl-75c-␣, and other exosomal proteins. COS-1 cells were co-transfected with constructs encoding PM/Scl-75a-␣, PM/Scl-75c-␣, and other exosomal proteins, fused to either the VP16 transcription activation domain (pACT constructs) or the GAL4 DNA-binding domain (pBIND constructs), simultaneously with a reporter plasmid. The resulting luciferase activity is depicted in relative luminescence units (RLU). The activity observed for the combination of hRrp42p in pBIND and hCsl4p in pACT was defined as 100 relative luminescence units (26). each other via their most C-terminal sequences. The failure of PM/Scl-75 to interact with its putative neighbor hRrp46p, which is predicted to be mediated by the N-terminal regions of their RNase PH domains, might be due to the fact that the hRrp46p cDNA that was used is also incomplete. Very recently an alternative sequence for hRrp46p was published (25) that contains an N terminus with 33 additional amino acids. Although the complete RNase PH domain is present in the original sequence of hRrp46p, it is possible that some flanking amino acids might be required for the interaction with PM/Scl-75, either by making direct contacts or by stabilizing the proper conformation of the RNase PH domain. Another explanation might be the possible interference of the fusion parts of the mammalian two-hybrid constructs, which are both attached to the N terminus of the proteins to be analyzed. Whether the AU binding properties of PM/Scl-75 (19) are also influenced by the extra N-terminal sequence remains to be investigated, although this activity is most likely mediated by the C-terminal part of the protein, because the RNase PH domain is responsible for the interaction of the protein with the exosome complex. Because no functional differences were found between the splice variants PM/Scl-75-␣ and PM/Scl-75-␤, and both PM/Scl-75b and PM/Scl-75c contain the complete RNase PH domain, the function of this multiplicity of PM/Scl-75 splice variants remains to be iden-tified, although the function of the ␤ form might be specific for the nucleus given its increased nuclear accumulation.