Isolation and Proteomic Characterization of Human Parvulin-associating Preribosomal Ribonucleoprotein Complexes* 210

Human parvulin (hParvulin; Par14/EPVH) belongs to the third family of peptidylprolyl cis-trans isomerases that exhibit an enzymatic activity of interconverting the cis-trans conformation of the prolyl peptide bond, and shows sequence similarity to the regulator enzyme for cell cycle transitions, human Pin1. However, the cellular function of hParvulin is entirely unknown. Here, we demonstrate that hParvulin associates with the preribosomal ribonucleoprotein (pre-rRNP) complexes, which contain preribosomal RNAs, at least 26 ribosomal proteins, and 26 trans-acting factors involved in rRNA processing and assembly at an early stage of ribosome biogenesis. Since an amino-terminal domain of hParvulin, which is proposed to be a putative DNA-binding domain, was alone sufficient to associate in principle with the pre-rRNP complexes, the association is probably through protein-RNA interaction. In addition, hParvulin co-precipitated at least 10 proteins not previously known to be involved in ribosome biogenesis. Coincidentally, most of these proteins are implicated in regulation of microtubule assembly or nucleolar reformation during the mitotic phase of the cell. Thus, these results, coupled with the preferential nuclear localization of hParvulin, suggest that hParvulin may be involved in ribosome biogenesis and/or nucleolar re-assembly of mammalian cells.

Peptidylprolyl cis-trans isomerases (PPIases) 1 catalyze the rotation about the peptide bond preceding proline; a step that can be rate-limiting for the folding of newly synthesized protein (1). PPIases also have the ability to bind many proteins and to act as chaperones, thus they are believed to regulate folding, assembly, and trafficking and controlling activity of proteins in the cell (2,3). PPIases are most familiar as the targets of the immunosuppressive drugs cyclosporin A and FK506, which bind to cyclophilin (CyP) and FK506-binding protein (FKBP), respectively (4,5). Both CyPs and FKBPs are ubiquitous, highly expressed and conserved from bacteria to human. A number of CyP and FKBP homologues have been identified in almost all organisms and in almost every cellular compartment of the cell (6). Their ubiquitous presence strongly suggests that PPIases have some general cellular functions as protein foldases and chaperones. However, all mutants with a defect in single and multiple PPIase (CyPs and FKBPs) genes were viable with no remarkable phenotype in yeast (7,8). Thus, the general role of PPIases as protein foldases and chaperones in the cell has been difficult to elucidate in vivo and is still controversial.
Despite this uncertainty of the role of PPIase in protein folding, the unique functions in higher organisms have been reported for some of the CyPs and FKBPs. For example, CyPA, which is the most abundant cytosolic PPIase, modulates human HIV-1 infectivity through controlling Gag processing, in which CyPA regulates cleavage of Gag polypeptide by HIV-1 protease through its ability to catalyze cis-trans interconversion of the proline-peptide bond around the cleavage site of the Gag polypeptide (9,12,13). NinaA, which is one of the isoforms of CyPA, has been definitively shown to be involved in the trafficking and maturation of specific isoforms of rhodopsins in Drosophila photoreceptor cells (2,11). In the case of FKBP12, it is involved in heart formation in mice through its ability to regulate the activity of calcium ion channels (14).
The third family of PPIases, the parvulins, is also conserved from bacteria to human. Parvulin, which was found initially in Escherichia coli cells, has only 92 residues, and is the smallest member in the PPIase families (15). E. coli parvulin was also shown to catalyze prolyl isomerization not only in short peptides, but also in protein folding reactions in vitro (16). Homologues of the E. coli parvulin are present in fungi (Ssp1) (17), yeast (Ess1/Ptf1) (18), Drosophila (Dodo) (19), and human (Pin1) (20). All of these homologues have a unique WW proteinprotein interaction domain extending amino-terminal to the catalytic PPIase domain. The WW domain is able to bind the specifically phosphorylated Ser/Thr-Pro motif (21), and the catalytic PPIase domain preferentially isomerizes proline residues preceded by phosphorylated Ser or Thr with more than a 1000fold selectivity compared with unphosphorylated peptides (22). Although none of the CyPs and FKBPs have been reported to be essential in any of the species so far analyzed, at least yeast Ess1 is essential (23) and its defect is complemented by its homologues from other species including human (20). By a combination of genetic and biochemical experiments, Ess1 was shown to function in transcription machinery by associating to the carboxyl-terminal domain of RNA polymerase II; it was proposed that Ess1 together with a CyP control the recruitment of transcriptional regulatory factors, such as Sin3-Rpd3 histone deacetylase, via their interactions with the carboxyl-terminal domain (24,25). In addition, depletion of human Pin1 from HeLa cells was reported to induce mitotic arrest and its overexpression arrest in the G 2 phase of the cell cycle (20). Thus, Pin1 is an essential PPIase that regulates mitosis in the human cell. However, the knockouts of the dodo and Pin1 genes did not show the lethal phenotype in Drosophila and mice, respectively, indicating functional redundancy in higher eukaryotes (26,27).
In addition to Pin1, a second PPIase (hParvulin, Par14/ EPVH) belonging to the parvulin family has been identified in all human tissues so far analyzed (28,29). This hParvulin consists of 131 amino acid residues, and has an amino-terminal tail of 35 amino acid residues, which does not show sequence homology to a WW domain of Pin1. The carboxyl-terminal domain with 96 amino acid residues has 34.2% sequence homology to the PPIase domain of Pin1. Although hParvulin has been reported to have substrate specificity with preference not for phosphorylated Thr or Ser, but for positively charged residues preceding proline, its rate constant of prolyl cis to trans isomerization is at least 1000-fold lower than those of CyPs and FKBPs (29). The NMR solution structural analysis showed that hParvulin folds into a ␤␣3␤␣␤2 structure that was essentially the same as that of human Pin1, and contained an unstructured 35-amino acid basic tail amino-terminal to the PPIase catalytic core that replaces the WW domain of human Pin1 homologues (30,31). The structural model provided the molecular basis for preferential substrate specificity to positively charged residues preceding proline, and for the putative role of the amino-terminal 35-amino acid tail as a DNA-associating segment. However, the cellular function of hParvulin is not understood.
Here we report proteomic analysis of the nuclear complexes bound to hParvulin. We found that the hParvulin-associating complexes isolated contained pre-rRNAs and ribosomal proteins, as well as other proteins, many of which are known or expected to be involved in rRNA processing and assembly of ribosomal proteins. These findings coupled with the other results suggest that hParvulin is a component of the pre-rRNP complexes formed at an early stage of ribosome biogenesis.
Preparation of Glutathione S-Transferase (GST)-fused hParvulin-Drs. F. Fujimori and T. Uchida kindly provided the expression vector (pGEX-4T-1) containing the gene coding for hParvulin (Par14) that is fused amino-terminal to GST through a short sequence with a thrombin cleavage site. E. coli BL21(DE3) cells were transfected with the pGEX-4T-1 vector and were grown in LB medium (1.0% Bacto-tryptone, 0.5% Bacto-yeast extract, 1.0% NaCl, and 1 mM NaOH), containing ampicillin (50 g/ml). After the induction of protein expression with 0.1 mM isopropyl-␤-D-thiogalactopyranoside at A 600 ϭ 0.7, BL21(DE3) cells were grown for a further 4 h at 37°C and were harvested by centrifugation at 8,000 rpm for 30 min at 4°C. The cells were lysed in sodium phosphate-buffered saline (PBS, pH 7.4), containing 2 mM EDTA, 0.1 mM phenylmethanesulfonyl fluoride (PMSF), and 0.1% 2-mercaptoethanol, sonicated, and centrifuged at 16,000 rpm for 30 min at 4°C. The cleared lysates were affinity-purified with glutathione-Sepharose beads and eluted with reduced glutathione as described in the manufacturer's instruction manual (Amersham Biosciences). The GST-hParvulin eluate was fractionated with ammonium sulfate, and was used for further experiments after dialysis against PBS (pH 7.4) containing 0.1% 2-mercaptoethanol.
Preparation of Nuclear Extract from Culture Cells-Mammalian cells were grown in RPMI 1640 or Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal calf serum, 2 mM Lglutamine, 160 units/ml penicillin G, and 0.1 mg/ml streptomycin at 37°C in an incubator with 5% CO 2 . Cultured cells were lysed in buffer A (10 mM HEPES, pH 7.8, containing 10 mM KCl, 2 mM MgCl 2 , 0.1% IGEPAL CA-630, 1 mM dithiothreitol, 50 mM NaF, 40 mM glycerol phosphate, 10 mM Na 3 VO 4 , 5 g/ml PMSF, 2 g/ml each of aprotinin and pepstatin A), and incubated on ice for 15 min. The cytosolic fraction was obtained as the supernatant by centrifugation of the lysate at 4°C for 3 min at 1,000 rpm, and the precipitate was used as the nuclear pellet. The cytosol fraction and the nuclear pellet were stained with 4Ј,6-diamidino-2-phenylindole. After confirmation of the purity of the nuclear pellet under fluorescent microscopy BX50 FLA (KS Olympus, Tokyo Japan), it was sonicated on ice in buffer B (10 mM HEPES, pH 7.8, 10 mM KCl, 1 mM dithiothreitol, 50 mM NaF, 40 mM glycerol phosphate, 10 mM Na 3 VO 4 , 5 g/ml PMSF, 2 g/ml aprotinin, 2 g/ml pepstatin A, plus 1% IGEPAL CA-630, 1% sodium deoxycholate), and centrifuged at 16,000 rpm for 5 min at 4°C. The supernatant was used as nuclear extract.
Isolation of GST-hParvulin-associating Complexes-The nuclear ex- tract and cytosolic fraction were preincubated with Sepharose CL-6B for 1 h at 4°C and centrifuged at 14,000 rpm for 5 min at 4°C, respectively. The supernatant was filtered with a filtrating membrane with a 0.45-m pore size, and was incubated with glutathione-Sepharose 4B beads (50 l) that were preincubated with GST-hParvulin for 1 h at 4°C. The glutathione-Sepharose beads were washed five times in buffer C (50 mM Tris-HCl, pH 8.0, 0.2 M NaCl, 0.5 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 1 mM dithiothreitol, 50 mM NaF, 40 mM glycerol phosphate, 10 mM Na 3 VO 4 , 5 g/ml PMSF, 2 g/ml aprotinin, 2 g/ml pepstatin A), and subsequently twice with buffer D (buffer C except PMSF and aprotinin) before thrombin treatment. The hParvulin-associating complexes were released from the glutathione-Sepharose beads by cleaving between GST and hParvulin with 5 units of thrombin in 50 l of buffer D for 2 h at 16°C. As a control, GST bound to glutathione-Sepharose 4B beads was also pulled down with the nuclear extract and cytosolic fraction, respectively.
Protein Identification by Mass Spectrometer Analyses-Peptide preparation after SDS-PAGE was done according to the method described by Yanagida et al. (32). The peptide mixture was mixed with an equal volume of 10 mg/ml ␣-cyano-4-hydroxycinnamic acid in 50% acetonitrile, 0.1% trifluoroacetic acid and aliquots of 0.5 l were applied onto a target disk and allowed to air dry. Peptide mass fingerprint was obtained using a PE Biosystems MALDI-TOF mass spectrometer (MS) (Voyager DE-STR, Foster City, CA). The data base-fitting program

FIG. 3. SDS-PAGE analyses of the hParvulin-associating proteins pulled down from the fractionated cell extracts.
A, mouse L929 cells: lanes 1 and 2, pulled down from the cytoplasmic fraction with GST and GST-hParvulin (Par14), respectively; lanes 3 and 4, pulled down from the nuclear extract with GST and GST-hParvulin, respectively. B, human HeLa cells: lanes 1 and 2, pulled down from the nuclear extracts with GST and GST-hParvulin, respectively. C, human 293EBNA cells: lanes 1 and 2, pulled down from the nuclear extract with GST and GST-hParvulin, respectively. Molecular weights (MW, kDa) are given at the left side. The elution positions of thrombin and GST are also given at the right side. All SDS-PAGE analysis was carried out for a fraction that was eluted from glutathione-Sepharose by thrombin cleavage. MS-Fit available at the World Wide Web site at the University of California, San Francisco (prospector.ucsf.edu/ucsfhtml3.4/msfit.htm), was used to interpret MS spectra of protein digests (33). A protein was considered identified when the spectrum of its measured peptide masses met the previously established criteria for positive identification of proteins using MALDI-TOF mass spectrometry and automated data base analysis (33). First, to distinguish a valid match from a false positive, a minimum of five measured peptide masses must match tryptic peptide masses calculated for an individual protein in the data base with Ͻ50-ppm average deviation in mass between measured and calculated values. Second, the peptides identified by these matches must provide at least 15% sequence coverage of the identified protein.
Incomplete tryptic cleavage and peptide modifications that may alter the peptide masses, such as oxidized methionine or carbamidomethyl cysteine, were calculated for the putatively identified protein and compared with the measured masses. The modified peptides identified in the search were added to the list to increase the number of matching peptides and sequence coverage. Protein that revealed the highest scored peptide mass fingerprint matching by data base search was retrieved as the identified protein.
RNase Treatment of GST-hParvulin-associating Complexes-The nuclear extract was incubated with GST-hParvulin bound to glutathione-Sepharose beads at 4°C for 1 h. The hParvulin-associating complex on the glutathione-Sepharose beads was washed 5 times with buffer C and incubated with 5 g/ml RNase A in buffer C at 37°C for 5 min. After washing twice with buffer D, GST-hParvulin on the Sepharose beads was cleaved with thrombin, and the eluate was subjected to SDS-PAGE analysis. The GST-hParvulin was also pulled down with the cell lysate pretreated with 5 g/ml RNase A or ethidium bromide at 37°C for 5 min.
Construction of the Expression Vectors for Direct hParvulin, Its GSTfused Deletion Mutants and Epitope-tagged hParvulin-For construction of a direct expression vector for hParvulin, a cDNA fragment containing the entire hParvulin gene was amplified by PCR using a primer set, 5Ј-ATCGGATCCGCGCACATTCTATGTGAAAAA-3Ј and 5Ј-CGGAATTCTTATCTGACCTTTACTGCATTG-3Ј from the GST-fused hParvulin expression plasmid as a template, and was inserted into the BamHI and EcoRI sites of pET-3a plasmid (Novagen, Madison, WI). The protein expression was done with the BL21(DE3) strain.
cDNA fragments for hParvulin deletion mutants comprising the amino-terminal 41-residue domain and the carboxyl-terminal PPIase domain were amplified by PCR using primer sets, 5Ј-ATCGGATCCGCG-CACATTCTATGTGAAAAA-3Ј and 5Ј-CGAATTCTTATTTTCTTCCTT-CGACCATAA-3Ј, and 5Ј-CTGGATCCATGCCGCCCAAAGGAAAAAG-TG-3Ј and 5Ј-CGGAATTCTTATCTGACCTTTACTGCATTG-3Ј, respectively, from the full-length cDNA of the hParvulin gene as a template. The amplified cDNA fragments were digested with restriction enzymes whose sites were tagged with primers, BamHI and EcoRI, and subcloned downstream of a sequence encoding the NH 2 -terminal-tagged GST in a pGEX-2T-based vector (Fig. 1).
For construction of epitope-tagged expression plasmids, a DNA fragment encoding human parvulin with the FLAG tag at its amino termini was amplified by PCR, by using a human placenta cDNA library (ORI-GENE) as a template. A primer set used for PCR was 5Ј-ATATAGCT-AGCGCCACCATGGACTACAAGGACGACGACGACAAGCCGCCCAA-AGGAAAAAGTGGT-3Ј/5Ј-TATATGGATCCTTATTTTCTTCCTTCGA-CCAT-3Ј. The resulting fragment was introduced between the NheI and BamHI sites of the expression vector pcDNA3.1 (Invitrogen) (Fig. 1).
Northern Blot Hybridization Analysis of Pre-rRNA and rRNA Species-RNAs were prepared from the hParvulin-associating complexes by the RNAgents Total RNA Isolation System, followed by phenolchloroform extraction and isopropyl alcohol precipitation. The RNAs obtained were fractionated on 1.0% agarose gels containing 1.1 M formaldehyde and were transferred to Hybond N ϩ nylon membranes as described (34).
An oligodeoxyribonucleotide 5Ј-TCAGACAGGCGTAGCCCCGG-GAGGAACCCG-3Ј, which is complementary to nucleotides 6996 -7025 of the mouse rDNA sequence, was labeled at the 5Ј-end by T4 polynucleotide kinase with 50 Ci of [␥-32 P]ATP (3000 Ci/mmol) (Amersham Biosciences), and was used as a probe for 5.8 S rRNA. DNA fragments corresponding to nucleotides 901-1100, 4434 -4633, 7201-7410, and 9257-9455 of the mouse rDNA were amplified by PCR, in which 30-base oligodeoxyribonucleotides of both ends of each fragment were used as primers, and were used for 5ЈETS, 18S, ITS2, and 28S probes, respectively (Fig. 2). These probes were labeled by PCR using a reaction mixture that contained 1 ng of the amplified fragments described above  The proteins identified in the hParvulin-associating pre-rRNP complexes are indicated in the first column. Each entry contains the identified protein name with NCBI accession numbers in parentheses. Origin of the proteins identified is indicated in the second column (M, mouse; H, human). The protein bands corresponding to the staining protein-band number in Fig. 5 are indicated in the third column and yeast homologues are in the fourth column. The proteins that are shown to be associated with yeast Nop7p and Nug1p, are indicated by the circle in the fifth and the last columns, respectively. salmon sperm DNA, at 42°C for 2 h, the RNA blots were hybridized to the probes at 42°C overnight in the prehybridization solution containing 5 ϫ 10 5 cpm. The blots were washed three times with 2 ϫ SSPE, 0.1% SDS at room temperature for 5 min and twice with 0.2 ϫ SSPE, 0.1% SDS at 56°C for 15 min. They were exposed to the Phos-phorScreen and were analyzed by STORM TM (Molecular Dynamics Japan Inc., Tokyo, Japan).
Immunocytochemistry-Human 293EBNA cells were cultured in 8-well culture slides (Biocoat, BD PharMingen) and transfected with the expression plasmid DNA containing the Flag-hParvulin gene using Lipo-fectAMINE. The transfected cells were fixed with 3.7% formaldehyde in PBS. After washing with PBS, the cells were treated with 3% (w/v) skim milk in PBS at room temperature. The cells were incubated with 10 g/ml mouse anti-FLAG IgG as the first antibody overnight at 4°C. After another washing with PBS the cells was further incubated with Alexa Fluor 488-conjugated anti-mouse IgG as the second antibody for 1 h at room temperature. Fluorescent images were visualized with a confocal laser-scanning microscope Fluoview (Olympus, Tokyo, Japan).

RESULTS AND DISCUSSION
Despite the availability of information from the three-dimensional structure at atomic resolution and the enzymatic substrate specificity of hParvulin (29 -31), its cellular function is entirely unknown. Therefore, we tried to search its associating proteins by using a currently developed proteomic approach with the hope of identifying possible functional linkage to the proteins whose cellular functions are well established. For this purpose, we used an affinity tag, GST with a thrombin cleavage site, incorporated into hParvulin at the NH 2 terminus to precipitate hParvulin-associating proteins from the fractionated cell extracts of mouse L929 cells. Since hParvulin was reported to be present in the nucleus as well as the cytoplasm (28,29), we expected to precipitate associating proteins from both the cytosolic fraction and the nuclear extract. However, hParvulinassociating proteins were obtained mostly from the nuclear extract of L929 cells (Fig. 3A). We recognized at least 65 staining bands as the hParvulin-associating proteins on SDS-PAGE gels (Fig. 3A). The presence of hParvulin-associating proteins in nuclear extract was also ascertained by use of the fractionated cell extracts from human HeLa and 293EBNA cells (Fig. 3,  B and C). Thus, hParvulin has the ability to pull down proteins present in the nuclear extracts of mammalian cells so far examined, suggesting that it plays a major role in the nucleus of mammalian cells. This is supported by its preferential presence in the nucleus of 293EBNA cells examined by using the FLAGtagged hParvulin molecules (Fig. 4).
We could retrieve 62 unique proteins for the hParvulinassociating proteins by use of the data base-fitting program MS-Fit after in-gel digestion and MALDI-TOF mass analysis of protein staining bands excised from SDS-PAGE gels (Fig. 5). Of these, 15 ribosomal proteins and 23 non-ribosomal proteins were identified directly by mass data searching as known proteins (Supplementary Tables I and II Table II).
The remaining proteins retrieved (24 proteins) were putative, hypothetical, or unnamed protein products. Therefore, we performed computer analysis with the program FASTA (version 3.2t09; Ref. 62) to search their sequences against over 679 million residues in 698,000 library sequences in the DNA sequence data base available by November 2001 at DDBJ (Mishima, Japan). By this analysis we identified 11 more ribosomal proteins; thus, a total of 26 proteins pulled down by hParvulin from the nuclear extract of L929 cells are counted as ribosomal proteins (Supplementary Table I). We also found that 10 of the other putative or unnamed proteins have high homology with the trans-acting proteins involved in ribosome biogenesis (Supplementary Table  II). In addition, although the protein named GTP-associating protein NGB (AF348208) was found to have homology with Saccharomyces cerevisiae Nog1p (Ypl093wp), which is an essential nucleolar GTP-associating protein. Thus, Nog1p is considered to be a trans-acting factor in ribosome biogenesis. Two other named proteins are also considered to be the trans-acting factors: i.e. pescadillo homologue 1 (PES1, AF289539, homologue of yeast Nop7p) and nucleolar protein mNIFK (AB056870, homologue of yeast Nop15p) (Supplementary Table II). Therefore, combined with the direct identification by mass data searching analysis, we considered a total of 26 proteins in the isolated hParvulin-associating proteins to be trans-acting factors involved in ribosome biogenesis (Table I).
Thus, three-quarters (52 of 62) of the hParvulin-associating proteins obtained from the nuclear extract of L929 cells were either ribosomal proteins or the trans-acting factors involved or thought to be involved in ribosome biogenesis. In addition, the isolated hParvulin-associating proteins showed RNA integrity for association with hParvulin as judged by RNase treatment (Fig. 6), indicating that hParvulin holds its associating proteins in an RNA-protein complex. We confirmed the presence of pre-ribosomal RNAs (pre-rRNAs) by Northern blot analysis (Fig. 7). Altogether these results strongly suggested that hParvulin interacted specifically with pre-rRNP complexes present in the nucleolus of L929 cells. Furthermore, this specific association of hParvulin to the pre-rRNP complexes was validated by the use of the NH 2 -terminal domain of hParvulin (⌬C-Par), which alone could pull down mainly the hParvulin-associating rRNP complexes (Fig. 8, A and B, for details, see Supplementary Results). We also subjected some of the protein bands selected randomly on SDS-PAGE gels of the hParvulin-associating complexes obtained from human 293EBNA and HeLa cells to MALDI-TOF and mass data searching analysis, respectively (Fig. 3, B and C). The results confirmed that all of the proteins analyzed are consistent with those identified in the complexes obtained from L929 cells (Supplementary Tables I  and II). Thus, we believe that the association with the rRNP complexes in the nuclear extracts from mammalian cells is a general characteristic of hParvulin.
We found a number of trans-acting factors involved in the early stages of ribosome biogenesis in the hParvulin-associating complexes, including fibrillarin, nucleolin, Nop5/58, and Nop56, which participate in the beginning of pre-rRNA processing at sites A 0 to A 2 present in the 5Ј-external transcribed spacer of pre-rRNA (35). Thus, we expected that the isolated hParvulin-associating complexes contain pre-rRNAs, like 45S, 41S, and 32S. As expected, in addition to the matured 28S, 18S, and 5.8S rRNAs, we could detect 45S, 34S, 32S, and 12S pre-rRNA species in the isolated complexes, which were assigned by the cDNA probes for 5ЈETS, ITS2, 18S, 28S, and 5.8S (Figs.  2 and 7). Although the matured rRNA species were detected, we excluded the possibility that the isolated complexes were contaminated by the matured ribosomal particles present in the cytoplasm, because we detected hParvulin-associating proteins and rRNAs almost exclusively in the nuclear extract but not in the cytosolic fraction of the cells used, in which the matured ribosomal particles are enriched. Therefore, the isolated hParvulin-associating complexes are those formed not necessarily in the nucleolus, but in the nucleus of the cell. However, since our procedure required 1 h for incubation of hParvulin with the nuclear extract before Northern analysis pre-rRNA species may have been degraded during the isolation process from the nuclear extract. However, no matter to what extent the pre-rRNA species are degraded or not, we conclude that hParvulin could associate with pre-rRNP complexes that contain many trans-acting factors involved in rRNA processing and assembly at an early stage of ribosome biogenesis.
The association of hParvulin with the rRNP complexes is quite striking in terms of their protein constituents. We have reported previously the isolation of a pre-rRNP complex present in the nucleolus of human 293EBNA cells by use of nucleolin as affinity bait (33). Nucleolin is an abundant nucleolar protein of exponentially growing eukaryotic cells and is directly involved in the regulation of ribosome biogenesis and maturation. We showed that the nucleolin-binding rRNP complex contained more than half of the ribosomal proteins present in the matured cytoplasmic ribosome particle and at least four transacting factors, such as nucleophosmin (B23) and RNA helicases, but not the other trans-acting factors such as fibrillarin and Nop5/58 that are involved in pre-rRNA processing at early stages of ribosome biogenesis. Based on those results, we proposed that the nucleolin-associating rRNP complexes we iso- lated are those formed at a very late stage of ribosome biogenesis in the nucleolus of the mammalian cell. In addition to this, two other groups reported the isolation of preribosomal particles from yeast cells by use of nuclear GTPase Nug1p and nucleolar Nop7p as affinity bait, respectively (36,37). Nug1p, which is required for ribosomal subunit export from the nucleus to the cytoplasm, co-precipitated with proteins of the 60 S subunit, late precursors to the 25S and 5.8S rRNAs, and at least 21 non-ribosomal proteins that include those implicated in 60S subunit export and nuclear pore complexes. This rRNP complex was proposed to be the transport intermediate for 60S subunit export in yeast cells (36). On the other hand, Nop7p, which is an essential nucleolar protein necessary for production of 60S ribosomal subunits, was associated with 66S preribosomes containing either 27SB or 25.5S plus 7S pre-rRNAs, and about 20 non-ribosomal proteins, of which 8 proteins were newly identified proteins that are involved in the assembly processes of ribosome biogenesis (37). However, in both cases, the pre-rRNP complexes did not contain the trans-acting factors participating in rRNA processing at an early stage of ribosome biogenesis, such as fibrillarin, Nop56, and Nop5/58. Thus, those yeast pre-rRNP complexes isolated by two different groups are also formed at a relatively late stage of ribosome biogenesis. Nonetheless, we noticed that 7 trans-acting factors corresponding to yeast Erb1p, Nog1p, Nop7p, Spb1p, Has1p, Mrt4p, and Nsa3p (Cic1p), that are present in the hParvulinassociating complexes are found in both pre-rRNP complexes associated with Nug1p and Nop7p, and 6 other factors corresponding to Drs1p, Ebp2p, Nop2p, Rrp1p, Nop15p, and Ytm1p are in either of the complexes (Table I). The rest of the transacting factors we assigned are shown to be uniquely present in the hParvulin-associating rRNP complexes. These include homologues of yeast Dbp9p, Nop1p (fibrillarin), Nop4p, Nop56p, Nop5/58p, Nsr1p (nucleolin), Rrs1p, Mak21p, Nog2p, and 4 RNA helicases. Many of these are trans-acting factors that are involved in pre-rRNA processing and assembly at a very early stage of ribosome biogenesis. Thus, the hParvulin-associating rRNP complexes we isolated are characterized uniquely by the presence of these early trans-acting factors.
Among those early trans-acting factors identified, fibrillarin (Nop1p), Nop5/58, and Nop56 are known to be associated with the box C ϩ D small nucleolar RNAs (35,38). In addition, in yeast cells, 2Ј-O-methyltransferase was predicted to be present in the small nucleolar RNP complex, and Spb1p, which is reported to be associated with Nop1p and Nop58p, is proposed to be one of the candidates (39 -41). We found that the hParvulin-associating complexes also contain a homologue of yeast Spb1p (Table I). Thus, these results indicate that the isolated complexes contained major components of the known box C ϩ D small nucleolar RNP complex that plays an essential role in pre-rRNA processing at an early stage of ribosome biogenesis. Although the H/ACAbox RNP complexes that are required for the site-specific pseudouridylation of rRNA are also well known trans-acting factors involved in an early stage of ribosome biogenesis (42), so far we could not find any components of the H/ACA-box small nucleolar RNP complex in the hParvulin-associating complexes.
In addition to the trans-acting factors, by use of hParvulin as an affinity bait we have co-purified 10 more non-ribosomal proteins of previously unknown function in ribosome biogenesis along with the rRNP complexes (Table II and Supplementary  Table III). Of these at least 3 proteins are reported or expected to be present in the nucleolus or nucleus. Among these the nucleolar protein includes the p160 Myb-binding protein (U63648) (43,54,55). The proteins known to be present in the nucleus include the homologue of targeting protein for Xenopus kinesin-like protein 2 (Xklp2) (TPX2, AF244547) (44 -46, 56 -58) and the p32cdc2-like PITSLRE protein kinase ␣ SV9 isoform (AF080683) (47,60). The other non-ribosomal proteins we identified are fibronectin (48,49), tubulin ␣1, putative (AK01960, ␤-tubulin isotype M-␤ 5), growth arrest specific 11 (50,51), putative (AK007869, eukaryotic translation elongation factor 1) (52,53,59), putative (AK010776), and GAJ (Supplementary Table III). All of these non-ribosomal proteins except GAJ and putative (AK010776) can be categorized coincidentally into protein factors participating or expected to be participating in microtubule assembly or nucleologenesis TABLE II Non-ribosomal proteins of previously unknown function in ribosome biogenesis The proteins identified in the hParvulin-associating pre-rRNP complexes are indicated in the first column. Each entry contains the identified protein name with NCBI accession number in parentheses. Origin of the proteins identified is indicated in the second column (M, mouse; H, human). The protein bands corresponding to the staining protein-band number in Fig. 5  occurring during the M-phase of the cell cycle (Table II and  Supplementary Table III). Thus, based on these observations we propose that the hParvulin-associating rRNP complexes isolated represent those formed during postmitotic nucleolar reformation before rDNA transcription or premitotic nucleolar disassembly. Accordingly, hParvulin may play a role in those processes. The two known biochemical characters of hParvulin are as the amino-terminal domain is suggested to interact with DNA and/or RNA and as the carboxyl-terminal domain has the PPIase activity with a preferential substrate specificity for positively charged residues preceding proline. These seem to fit its proposed function very well in events such as those occurring in ribosome biogenesis, rDNA transcription, and remodeling of the nucleolus, because very basic proteins such as ribosomal proteins and histones are major participants in those events. Currently, a number of protein constituents of the nucleolus and interchromatin granule clusters (IGC) have been reported for human and mouse cells, respectively (61, 62). As we expected, most of the ribosomal proteins and the trans-acting factors, p160 Myb-binding protein, and tubulins that are identified in the hParvulin-binding rRNP complexes are listed in the reported protein constituents of the human nucleolus. On the contrary, we could not find any of the reported protein components of IGC in the hParvulin-associating rRNP complexes, except some ribosomal proteins which have been considered to be contaminants during isolation of IGC by Mintz et al. (10). Thus, the isolated hParvulin-associating rRNP complexes may not be a constituent of IGC. However, if the ribosomal proteins found in IGC are also minor constituents of IGC we cannot rule out the possibility that the hParvulin-associating rRNP complexes may also be minor constituents of IGC. The detailed proteomic analysis of each highly organized nuclear structure will be needed to clarify how DNA replication, transcription, pre-mRNA processing, ribosome biogenesis, and RNA transport are coordinated in the nucleus.