Functional interaction between RNA helicase II⁄Gua and ribosomal protein L4 Hushan Yang, Dale Henning and Benigno C. Valdez Department of Pharmacology, Baylor College of Medicine, Houston, Texas, USA Ribosome biogenesis is a complicated cellular process which occurs in the nucleolus [1]. The entire scenario begins at the end of mitosis and includes ribosomal DNA transcription, pre-ribosomal RNA (pre-rRNA) modifications and processing as well as assembly of rRNAs and ribosomal proteins into preribosome sub- units which are then exported to the cytoplasm to form the mature ribosomes [2]. Errors in this process are reported to be associated with several diseases [3–8]. The availability of genetic manipulations makes ribosome biogenesis much better studied in yeast than in higher eukaryotes such as mammalian and frog sys- tems, resulting in the identification of more than 80 yeast ribosomal proteins and numerous trans-acting elements including small nucleolar RNAs (snoRNAs) as well as nonribosomal proteins. However, in mam- malian systems, ribosome biogenesis is far from being thoroughly understood due to the increased complex- ity. To date, only a few nucleolus-localized nonribo- somal proteins have been implicated in pre-rRNA processing in mammalian cells and include B23 ⁄ NO38 ⁄ NPM [9], C23 ⁄ nucleolin [3], fibrillarin [10,11], p120 [12], EBP1 [13], Bop1 [14] and p19 Arf [15]. No bona fide RNA helicase has been implicated in this process in higher eukaryotes except RNA helicase II ⁄ Gua [16,17]. RNA helicase II ⁄ Gua is a multifunctional nucleolar protein with in vitro RNA-dependent ATPase activity, ATP-dependent RNA helicase activity and GTP-stimu- lated RNA foldase activity [17–19]. The presence of both RNA unwinding and RNA folding activities in two distinct domains of the same protein highly sug- gests a role of Gua in rRNA biogenesis [19]. Using antisense oligodeoxynucleotide and siRNA to down- regulate Gua expression in Xenopus oocytes [16] and mammalian cells [17], respectively, we demonstrated that Gua is important for 18S and 28S rRNA produc- tion in both systems. In addition, Gua was also showed to participate in other major cellular activities such as cell growth and differentiation [19,20], regulation of Keywords ribosomal protein; ribosomal RNA biogenesis; RNA helicase; nucleolus Correspondence B. C. Valdez, Department of Pharmacology, Baylor College of Medicine, Houston, TX 77030, USA Fax: +1 713 798 3145 Tel: +1 713 798 7908 E-mail: bvaldez@bcm.tmc.edu (Received 11 April 05, revised 19 May 05, accepted 9 June 05) doi:10.1111/j.1742-4658.2005.04811.x RNA helicase II ⁄ Gua is a multifunctional nucleolar protein involved in ribosomal RNA processing in Xenopus laevis oocytes and mammalian cells. Downregulation of Gua using small interfering RNA (siRNA) in HeLa cells resulted in 80% inhibition of both 18S and 28S rRNA production. The mechanisms underlying this effect remain unclear. Here we show that in mammalian cells, Gua physically interacts with ribosomal protein L4 (RPL4), a component of 60S ribosome large subunit. The ATPase activity of Gua is important for this interaction and is also necessary for the func- tion of Gua in the production of both 18S and 28S rRNAs. Knocking down RPL4 expression using siRNA in mouse LAP3 cells inhibits the pro- duction of 47 ⁄ 45S, 32S, 28S, and 18S rRNAs. This inhibition is reversed by exogenous expression of wild-type human RPL4 protein but not the mutant form lacking Gua-interacting motif. These observations have sug- gested that the function of Gua in rRNA processing is at least partially dependent on its ability to interact with RPL4. Abbreviations aa, amino acid; GST, glutathione S-transferase; Gua, RNA helicase II ⁄ Gua; HA, hemagglutinin; IPTG, isopropylthio-b- D-galactoside; NLS, nuclear localization signal; NoLS, nucleolar localization signal; RPL4, Ribosomal Protein L4; rDNA, ribosomal DNA; rRNA, ribosomal RNA; RNP, ribonucleoprotein; siRNA, small interfering RNA; snRNA, small nuclear RNA; snoRNA, small nucleolar RNA. 3788 FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS c-Jun-mediated gene expressions [21] and in vitro clea- vage by PIAS1 [22]. It is unclear whether the effect of Gua on cell proliferation is due to its involvement in ribosomal RNA production, c-Jun-mediated gene expression or, as yet, other undiscovered mechanisms. RNA helicases are believed to function through regulation of RNA structure rearrangement and RNA- RNA, RNA-protein or protein–protein interactions [23,24]. To date, at least 18 putative ATP-dependent RNA helicases have been suggested to contribute to ribosome production in yeast S. cerevisiae [25]. It is hypothesized that RNA helicases interact with other trans-acting protein factors in preribosomal particles during pre-rRNA processing and ribosome assembly to modulate specific intracellular RNA structures [26–28]. To test the hypothesis that Gua may function by interacting with other nucleolar proteins in rRNA production, we did immunoprecipitation and identified ribosomal protein L4 physically interacting with Gua in an RNA-independent manner. It was also demon- strated that the interaction is necessary for the function of Gua in 28S rRNA production in mammalian cells. Results Ribosomal protein L4 physically interacts with RNA helicase II ⁄ Gua In our search for Gua-interacting proteins, we used isopropylthio-b-d-galactoside (IPTG) to induce a stable LAP3 clone expressing FLAG epitope-tagged mouse Gua protein. Two days after induction, the cells were lysed, treated with RNase A to rule out any protein that may be in the complex due to binding to the same RNA, and subjected to immunoprecipitation using anti-FLAG Ig resin (Sigma, St Louis, MO). The immunoprecipitates were resolved on a sodium dodecyl sulfate–polyacryl- amide (10%) gel and analyzed by silver staining. Several stained bands were detected in the FLAG-Gua lane but not in the control lane (Fig. 1A). One of the major bands with molecular weight of approximately 50 kDa was recovered and sent for mass spectrometry sequen- cing. It turned out to be ribosomal protein L4. There were other faster-migrating bands in the FLAG-Gua lane which were of greater abundance than those in the control lane (Fig. 1A). We did not sequence these bands, but we suspect they represent other ribosomal proteins and ⁄ or trans-acting factors of a large nucleolar complex essential for ribosome biogenesis. The yeast two-hybrid system was used to prove the direct interaction between Gua and RPL4. We sub- cloned human Gua and RPL4 into pGBKT7 and pGADT7 yeast expression vectors, respectively. The growth of yeast cells containing both Gua and RPL4 in a triple drop-out medium that lacks tryptophan, leucine, and histidine indicates interaction of the two proteins (Fig. 1B, right). The specificity of Gua–RPL4 interaction is shown by the inability of the yeast clones that harbor RPL4 and p68, a DEAD-box helicase implicated in RNA splicing and export [29], or RPL4 and p53, to grow in a triple drop-out medium (Fig. 1B, right). The growth of yeast cells containing the above expression constructs in a double drop-out medium that lacks tryptophan and leucine indicates that these constructs were expressed (Fig. 1B, left). Because protein–protein interactions shown by yeast two-hybrid system are not always direct, we performed an in vitro pull down assay using bacterially expressed GST-RPL4 and untagged Gua mixed together and pulled down with GSH-resin. Gua was pulled down with GST-RPL4, which further supported the direct interaction between Gua and RPL4 (Fig. 1C). The in vivo association of Gua with RPL4 was shown in both human HeLa cells and mouse LAP3 cells. We cotransfected either FLAG-tagged Gua and protein A-tagged RPL4, or FLAG-tagged RPL4 and protein A-tagged Gua into HeLa cells and did immu- noprecipitation using anti-FLAG resin. Figure 1D shows that both overexpressed (proA-RPL4) and endogenous RPL4 interact with Gua. Figure 1E shows that both overexpressed (proA-Gua) and endogenous Gua interact with RPL4. We tried using either anti- Gua or anti-RPL4 Ig to do similar experiments to show an association of endogenous Gua and endo- genous RPL4. However, neither antibody worked for immunoprecipitation although they performed well in western blot analyses. To determine the expression and cellular localization of RPL4 protein, anti-FLAG Ig was used for indirect- immunofluorescence of HeLa cells transfected with FLAG-tagged RPL4. GFP-tagged Gua was cotrans- fected as a control to show the positions of nucleoli. The localization of RPL4 to the nucleolus further indi- cates the role it may play in rRNA processing and ribosome assembly (Fig. 1F). RPL4 is a component of 60S ribosome large subunit with proposed cytoplasmic localization. However, we did not observe strong cyto- plasmic fluorescent signal for FLAG-RPL4. This observation has been shown with other reported ribo- somal large subunit proteins such as L23 [30]. A DEVD (Asp-Glu-Val-Asp) mutant of RNA helicase II ⁄ Gua does not interact with RPL4 The DEVD motif of RNA helicases is critical for their ATPase activity which is necessary for RNA helicase- H. Yang et al. Gua–RPL4 interaction in mammalian rRNA production FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS 3789 mediated reorganization of RNA ⁄ protein structure [23]. We previously showed that the DEVD motif of Gua is important to 18S and 28S rRNA production [17]. Since we surmised that Gua–RPL4 interaction is necessary for Gua to function in pre-rRNA processing, we sought to determine if the DEVD motif is also important to the Gua–RPL4 interaction. DEVD was mutated to ASVD with reported abolishment of both ATPase and helicase activities [16]. An SAT mutant, in which the SAT motif was mutated to LET, was used as a control. This mutant still retains ATPase activity but not RNA helicase activity [19]. We cotransfected FLAG-tagged human RPL4 with hemagglutinin (HA)- tagged wild-type, DEVD mutant or SAT mutant form of human Gua into HeLa cells and did immunopreci- pitation with anti-FLAG resin. Figure 2A shows that FLAG-RPL4 was pulled-down efficiently in all three precipitates. However, only wild-type Gua was also present in the precipitate, suggesting the relevance of the DEVD and SAT motifs in its association with F D C A B E FLAG-vector FLAG-Guα Supernatant IP: anti-FLAG IP: anti-FLAG WB: anti-FLAG FLAG-Guα Guα FLAG-RPL4 ProA-Guα ProA-RPL4 RPL4 WB: anti-Guα WB: anti-FLAG Input Sup’t Wash IP Input Sup’t Wash IP WB: anti-RPL4 Pellet Mock GST-RPL4 GST Mock GST-RPL4 GST Guα Guα RPL4 Other proteins Double drop-out (No Trp, No Leu) Triple drop-out (No Trp, No Leu, No His) RPL4 Guα IgG-H Fig. 1. Gua interacts with RPL4. (A) Stable LAP3 clones were induced with 2 mM IPTG for 48 h to express FLAG-tagged mouse Gua. RNase A-treated lysates were used in immunoprecipitation using anti-FLAG resin. Silver staining shows the precipitation of  50-kDa protein (RPL4) in cells expressing mouse Gua but not in cells expressing vector alone. (B) Yeast two-hybrid analysis showing the interaction of human RPL4 with human Gua and Gub. Yeast clones were grown on selection media. Growth in the absence of tryptophan and leucine would indi- cate presence of the appropriate vectors used to clone RPL4 and its candidate partner. Presence of colonies in the triple drop-out medium (no tryptophan, no leucine, no histidine) would indicate interaction between RPL4 and the other protein. (C) In vitro interaction of RPL4 with Gua. Purified GST, GST-RPL4 or blank control was mixed with purified untagged Gua in a binding buffer prior to addition of GSH-resin. Cen- trifugation separated the supernatant from the resin. Both the supernatant and resin were analyzed by western blot analysis using anti-Gua Ig. (D) Overexpressed Gua interacts with both endogenous and overexpressed RPL4. Extracts from HeLa cells cotransfected with FLAG- tagged Gua and protein A-tagged RPL4 were immunoprecipitated using anti-FLAG resin and probed with the indicated antibodies. (E) Over- expressed RPL4 interacts with both endogenous and overexpressed Gua in HeLa cells. Extracts from HeLa cells cotransfected with FLAG-tagged RPL4 and protein A-tagged Gua were immunoprecipitated with anti-FLAG resin and probed with indicated antibodies. (F) HeLa cells transfected with FLAG-tagged human RPL4 were stained by indirect immunofluorescence using anti-FLAG Ig. Anti-mouse IgG coupled to rhodamine was used as secondary antibody. GFP-tagged human Gua was cotransfected and visualized directly under microscope, as a control showing the position of nucleoli. Gua–RPL4 interaction in mammalian rRNA production H. Yang et al. 3790 FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS RPL4. This experiment further proved the specificity of Gua–RPL4 interaction since B23, an abundant nucleolar phosphoprotein which is also implicated in ribosomal RNA processing [31], was not pulled-down by RPL4 (Fig. 2A, bottom). In a similar experiment using mouse cell line, Xpress-tagged mouse RPL4 was transfected into a sta- ble LAP3 clone which expressed either IPTG-induced FLAG-tagged wild-type, DEVD mutant or SAT mutant form of mouse Gua. Immunoprecipitation was again carried out using anti-FLAG resin and the pre- cipitate was analyzed by western blot analysis. Figure 2B shows that all three forms of Gua pulled- down Xpress-tagged mouse RPL4 but with different efficiencies. Wild-type Gua interacted with RPL4 with the highest efficiency, while the DEVD mutant showed the least. This is demonstrated by comparing the signal intensities of the precipitates with those of the original inputs (Fig. 2B). The ratio (input : IP) is approxi- mately 1 : 5 for wild-type, 1 : 1 for SAT mutant and 5 : 1 for DEVD mutant. The observed difference in the sensitivity of the two experiments (Fig. 2A,B) might be attributed to difference in the levels of expression of FLAG-Gua and FLAG-RPL4. The sig- nal intensity of the FLAG-Gua (Fig. 2B) is greater than FLAG-RPL4 (Fig. 2A), which is possibly due to higher expression level of FLAG-Gua in the LAP3 stable cell line that was induced with IPTG compared to FLAG-RPL4 that was expressed by transient trans- fection. The DEVD motif is important for the function of Gua in both 18S and 28S rRNA production We have demonstrated that in Xenopus oocytes, wild- type Gua can reverse the aberrant rRNA processing pattern while the DEVD mutant cannot [16], highlight- ing the importance of the DEVD motif to the function of Gua in both 18S and 28S rRNA production in Xenopus. In the mammalian system, we were able to demonstrate that an SAT mutant, which lacks helicase activity, can restore 28S but not 18S rRNA production in mouse LAP3 cells [17], which suggests that the SAT motif is important in 18S but not 28S rRNA produc- tion. Because the helicase activity is dependent on the presence of the ATPase activity of Gua [18,19], it is reasonable to expect that mutation of the DEVD motif would consequently result in defects of 18S matur- ation. However, whether or not the DEVD motif is necessary for 28S production in mammalian cells is unknown. Here, a rescue experiment was performed exactly as described [17] to address this issue. Briefly, a stable LAP3 clone was induced with IPTG to over- express a DEVD mutant form of the human Gua, after which the cells were treated with si935, an effect- ive siRNA that specifically targets mouse Gua mRNA but not human Gua mRNA. Figure 3 shows that treat- ment of the cells with si935 effectively inhibited the production of both 18S and 28S rRNAs (lane 3, com- pared with lanes 1 and 2), which conforms to our pre- vious results [17]. However, in this experiment the expression of a DEVD mutant form of human Gua protein did not restore 18S nor 28S rRNA (Fig. 3. lane 4) as the wild-type did [17]. Thus, we conclude that the DEVD motif is indispensable for the function of human Gua in both 18S and 28S rRNA production, consistent with our results in the Xenopus oocyte [16]. Amino acids 264–333 of human RPL4 is important to its interaction with Gua Human RPL4 has not been extensively studied after its cloning [32]. Human and mouse RPL4 are 90% A B Input Sup’t IP Input IP Input IP Input IP Input Sup’t IP Input Sup’t IP WB: anti-FLAG FLAG-RPL4 HA-Gu (WT) FLAG-RPL4 HA-Gu (SAT-M) FLAG-RPL4 HA-Gu (DEVD-M) IP: anti-FLAG IP: anti-FLAG WB: anti-HA WB: anti-B23 WB: anti-FLAG WB: anti-Xpress WB: anti-B23 FLAG-Guα Xpress-RPL4 Xpress-RPL4 FLAG-Guα (WT) Xpress-RPL4 FLAG-Guα (SAT-M) Xpress-RPL4 FLAG-Gu (DEVD-M) B23 FLAG-RPL4 HA-Guα B23 Fig. 2. The DEVD motif of Gua is important to Gua–RPL4 interac- tion. (A) HeLa cells were cotransfected with FLAG-tagged human- RPL4 and plasmids encoding HA-tagged wild-type (WT), SAT mutant (SAT-M) or DEVD mutant (DEVD-M) form of human Gua. Whole cell extracts were immunoprecipitated using anti-FLAG resin and probed with anti-FLAG, anti-HA or anti-B23 Ig. (B) LAP3 cells were transfected with Xpress-tagged mouse RPL4 and induced with 2 m M IPTG for 48 h to express FLAG-tagged wild-type, SAT mutant or DEVD mutant of mouse Gua. Whole cell extracts were immunoprecipitated using anti-FLAG resin and probed with anti- FLAG, anti-Xpress or anti-B23 Ig. H. Yang et al. Gua–RPL4 interaction in mammalian rRNA production FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS 3791 homologous in their cDNA-derived amino acid sequences. The 98 amino acid C-terminal of human RPL4 protein has little homology with its mouse homologue. However, their N-termini (amino acids 1– 333) are 99% identical with differences in only three amino acid residues. In higher eukaryotes, other than the proposed involvement in ribosome assembly, RPL4 has been implicated in cell proliferation and differenti- ation during rat neurogenesis [33] with unknown mech- anism. In yeast, ribosomal protein RPL2 is most homologous to human RPL4. The yeast RPL2 has two copies, RPL2A and RPL2B [34]. A decrease in the expression of RPL2A leads to reduced production of 60S large subunits and mature ribosomes, which conse- quently results in slower growth rates [34]. These data suggest the relevance of RPL4 in lower and higher eukaryotes. We hypothesize that the function of RPL4 is partly regulated by protein–protein interactions. To determine the RPL4 domains involved in Gua interaction, we generated FLAG-tagged human RPL4 deletion mutants (Fig. 4A,D), expressed them in HeLa cells and tested their ability to bind with Gua via immunoprecipitation. Analysis of the overexpressed proteins by indirect immunofluorescence showed that wild-type RPL4 (amino acids 1–428), N1 mutant (amino acids 1–264) and C1 mutant (amino acids 131– 428) predominantly localize to the nucleolus but the C2 mutant (amino acids 264–428) is dispersed within the nucleus but not in the nucleolus (Fig. 4B), indica- ting the region of amino acids 131–264 probably con- tains both the nuclear (NLS) and nucleolar (NoLS) localization signals while amino acids 264–428 may harbor another NLS but no NoLS. Figure 4C reveals that RPL4 C1 and C2 mutants, but not its N1 mutant form, coimmunoprecipitate with Gua, suggesting the Gua-interacting domain resides in amino acids 264–428 of RPL4. We speculated that if Gua–RPL4 interaction is important to cellular func- tions, then the chance should be high that the Gua- interacting domain in RPL4 would be in a conserved region. As the region of amino acids 333–428 is not highly conserved among different species, we focused on amino acids 264–333 as a possible Gua-interacting motif in RPL4. This hypothesis was proved to be cor- rect by coimmunoprecipitation of three mutants har- boring amino acids 264–333 (Fig. 4F, M3, M5, M6). The other three RPL4 mutants that lack amino acids 264–333 did not coimmunoprecipitate with Gua (Fig. 4F, M1, M2, M4). We observed that two bands are recognized by the anti-FLAG Ig in mutant M1. The lower band should be the correct deletion mutant expression product according to its expected molecular size. The identity of the upper band remains to be determined. Localizations of M2 (amino acids 131–264) and M3 (amino acids 131–333) mutants to the nucleolus are in accordance with the finding that both NLS and NoLS are within amino acids 131–264. Mutant M1 (amino acids 131–196) is dispersed within the whole cell but with stronger signal intensity in the cytoplasm than in the nucleus, suggesting that both the NLS and NoLS should be in the region of amino acids 196–264. Because both M4 (amino acids 204–264) and M5 (amino acids 204–333) mutants localize to the nucleo- plasm but not to the nucleolus, it would follow that the major NoLS for human RPL4 is within amino acids 196–204. For mutant M4, the fluorescent signal is mainly in the nucleoplasm, however, a significant proportion was also found in the cytoplasm. The locali- zations of M5 and M6 mutants, consistent with that of C2 mutant, are predominantly in the nucleus excluding the nucleolar region (Fig. 4E). Thus, we suspect that a strong NLS is within amino acids 264–333 while a IPTG si935 47S/45S 32S 28S 28S 18S 12 34 18S – – + – – + + + Fig. 3. The DEVD motif of Gua is important to both 18S and 28S rRNA production. LAP3 cells were induced with 2 m M IPTG to express DEVD mutant of human Gua. Cells were then treated with si935 for 48 h followed by pulse-labeling with [ 32 P]orthophosphate for 1.5 h and a chase for 3 h with normal growth medium. Total RNAs were extracted, resolved on a 1.2% agarose-formaldehyde gel and blotted onto a membrane for phosphorimager analysis. Gua–RPL4 interaction in mammalian rRNA production H. Yang et al. 3792 FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS weak NLS may be within amino acids 196–264. Com- bined with the NoLS, this weak NLS is capable to cause most RPL4 molecules to enter the nucleolus. It is not uncommon to have more than one nuclear local- ization signal within a protein [35,36]. Downregulation of RPL4 inhibits rRNA production in mouse LAP3 cell line RPL4 is a component of the 60S ribosome large sub- unit. To date, there is no report showing a direct involvement of RPL4 in pre-rRNA processing. Ribo- somal proteins are always produced in the cytoplasm, and then imported into the nucleoli to participate in preribosome assembly. The ribosomes are then expor- ted back into the cytoplasm where they direct protein production [37]. As we hypothesize the interaction between RPL4 and Gua is important to the function of Gua in rRNA production, it will aid to our hypothesis if we could determine whether downregulation of RPL4 has any effect on this process. A sequence near the 3¢ end of mouse RPL4 was used to design a small interfering RNA (si-L4-M1), which targets mouse but not human RPL4 mRNA (Fig. 5A). Downregulation effects were examined at both mRNA and protein lev- els, using RT-PCR and western blot analysis, respect- ively. Treatment of LAP3 cells with 100 nm si-L4-M1 for 48 h resulted in a decrease of the mouse RPL4 mRNA level by 70% (Fig. 5B, lanes 7 and 8). This decrease was dose-dependent. When 5 nm or 10 nm si-L4-M1 was used, the mRNA level decreased by about 42% or 55%, respectively (Fig. 5B, lanes 3–6). C F E D A B WT 1 1 131 Constructs Guα-binding Localization 264 WT anti-FLAG IP: anti-FLAG WB: anti-FLAG FLAG-C1 FLAG-N1 FLAG-C2 HA-Guα FLAG-M3 FLAG-M5 FLAG-M2 FLAG-M6 FLAG-M4 FLAG-M1 HA-Guα WB: anti-HA WB: anti-FLAG WB: anti-HA IP: anti-FLAG Hoechst Phase anti-FLAG Hoechst Phase N1 C1 C2 N1 C1 C2 M1 M2 M3 M4 M5 M6 M1 M2 M3 M4 M5 M6 264 333 428 131 131 131 196 264 333 264 204 204 264 – – + – + + Nucleoplasm and cytoplasm Nucleoli Nucleoli Nucleoplasm and cytoplasm Nucleoplasm Nucleoplasm 333 333 428 428 + – + + Nucleoli Nucleoli Nucleoli Nucleoplasm N1 C1 C2 M1 M2 M3 M6 M5 M4 Constructs Guα-binding Localization Fig. 4. Mapping of Gua-binding domain in human RPL4. (A) Schematic representation of wild-type and mutant forms of human RPL4. The open bar represents regions conserved between human and mouse RPL4. The shaded bar represents nonconserved regions. (B) Cellular localization of human RPL4 mutants. HeLa cells transfected with FLAG-tagged human RPL4 and various mutants were stained by indirect immunofluorescence using anti-FLAG Ig. Anti-mouse IgG coupled to FITC was used as secondary antibody. Nuclei were visualized by Hoe- chst stain. The phase images show dark phase nucleoli. (C) Whole cell extracts from HeLa cells cotransfected with HA-tagged human Gua and FLAG-tagged human RPL4 deletion mutants were immunoprecipitated using anti-FLAG resin and blotted as indicated. (D) Schematic rep- resentations of human RPL4 mutants M1 to M6 and their (E) cellular localization. (F) Whole cell extracts from HeLa cells cotransfected with HA-tagged human Gua and FLAG-tagged human RPL4 deletion mutant shown in (D) were immunoprecipitated using anti-FLAG resin and blotted as indicated. H. Yang et al. Gua–RPL4 interaction in mammalian rRNA production FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS 3793 Lacking an antibody against mouse RPL4 protein, we instead used an indirect way to determine the effect of si-L4-M1 on the protein level of mouse RPL4. We cotransfected HeLa cells with si-L4-M1 and an Xpress-tagged mouse RPL4 construct. In this experi- ment, the remaining exogenously expressed mouse RPL4 protein level after si-L4-M1 treatment could be measured with anti-Xpress Ig. Figure 5C shows that mouse RPL4 mRNA level was decreased by 83% after treatment of the cell with 100 nm si-L4-M1 for 48 h while the exogenously expressed mouse RPL4 protein level was downregulated by 72%. The siRNA did not have significant influence on mRNA levels of human RPL4 and U1C, and the protein levels of human RPL4 and B23. We hypothesized that downregulation of RPL4 would lead to an aberrant rRNA processing pattern if Gua–RPL4 interaction is necessary for the function of Gua in rRNA production. After 48 h of si-L4-M1 treatment, LAP3 cells were pulse-labeled with [ 32 P]- orthophosphate and chased with growth medium for 3 h. As a negative control, a nonrelated siRNA (si934Scr) was included [17]. Total RNA was extracted, resolved on a 1.2% agarose-formaldehyde gel and transferred to a hybond-N nitrocellulose filter for phos- phorimager analysis. We found that all four main visible species of rRNA (47 ⁄ 45S, 32S, 28S and 18S) were dramatically decreased in samples treated with si-L4-M1 (Fig. 5D). However, the decreases in 47 ⁄ 45S rRNAs were not as great as those in mature 28S rRNA (Fig. 5D), indicating that only part of the decrease in 28S was due to less precursors while the remaining changes resulted from the influence by downregulation of RPL4 on other pathways involved in rRNA produc- tion. Ethidium bromide-stained gel (Fig. 5D, bottom) is shown to indicate equal loading of the RNA. Wild-type RPL4 but not its mutant form which lacks the Gua-interacting domain reverses inhibition of rRNA production We constructed a deletion mutant of human RPL4, D264-333, which lacks the Gua-interacting domain amino acids 264–333 and another mutant D204-264 as a control (Fig. 6A). We had already localized the NoLS of RPL4 to amino acids 196–204, which was supported by the nucleolar localization of both mutants (Fig. 6B). The immunoprecipitation experi- ment shows their Gua-binding activity (Fig. 6C), and supports our earlier findings (Fig. 4F). Because the region including amino acids 264–333 seems to be the RPL4-Gua-interacting domain, the D204-264 mutant, A B C D Human 1120 Mouse 1129 si934Scr mRPL4 mU1C Relative mRPL4 Relative mL4 si-L4-M1 47S/45S si934Scr si-L4-M1 32S 28S 18S 18S 28S RT-PCR Western hU1C hRPL4 mRPL4 Xpress-mRPL4 hRPL4 hB23 100 94 65 52 40 49 35 100 17 100 2226 1 2345678 si934Scr si934Scr 20 nM si-L4-M1 5 nM si-L4-M1 100 nM si-L4-M1 100 nM si-L4-M1 100 nM si-L4-M1 _ _ _ __ _________ ___ Fig. 5. Downregulation of mouse RPL4 using si-L4-M1 resulted in aberrant rRNA processing. (A) Comparison of human and mouse RPL4 partial cDNA sequences containing the si-L4-M1 region. Underscored nucleotides differ from human to mouse. (B) siRNA-mediated down- regulation of mouse RPL4 mRNA. LAP3 cells were transfected with increasing concentrations of si-L4-M1. Total RNA was isolated after 48 h and analyzed by RT-PCR to determine the mRNA levels of mouse RPL4 and mouse U1C. (C) siRNA-mediated downregulation of mouse RPL4 protein. HeLa cells were cotransfected with Xpress-tagged mouse RPL4 and si-L4-M1. Total RNA was isolated after 72 h using TRIzol Reagent (Invitrogen) and analyzed by RT-PCR to determine the mRNA levels of mouse RPL4, human RPL4 and human U1C (left panel). A parallel experiment was done to analyze changes in the protein levels of mouse RPL4, human RPL4 and human B23 after si-L4-M1 treat- ment. (D) LAP3 cells were treated with 100 n M si-L4-M1 for 48 h. Total 32 P-labeled RNA was analyzed as described in the legend to Fig. 3. Ethidium bromide staining of both 18S and 28S rRNA is shown at the bottom. Gua–RPL4 interaction in mammalian rRNA production H. Yang et al. 3794 FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS but not the D264-333 mutant form, coimmunoprecipi- tated with Gua (Fig. 6C). To determine the significance of Gua–RPL4 inter- action in rRNA biogenesis, we first used si-L4-M1 to downregulate endogenous mouse RPL4 expression and then exogenously expressed the human orthologue and looked for reversal of inhibition of rRNA production. Figure 6(D) (lanes 1 and 2) shows that si-L4-M1 effectively inhibited production of all four species of rRNAs, which is consistent with the results in Fig. 5(D). This effect was reversed by the exogenous expression of human wild-type RPL4, suggesting that the human orthologue can functionally replace mouse RPL4 (Fig. 6D, lane 3). However, the expression of mutant human RPL4 lacking the Gua-interacting domain could not reverse the aberrant rRNA process- ing pattern as effectively as the wild-type while the mutant lacking amino acids 204–264 had a similar effect to that of the wild-type, indicating that the Gua–RPL4 interaction is important to the function of Gua in rRNA processing (Fig. 6D, lanes 4 and 5). Human RPL4 associates with 28S but not 18S rRNA To determine if RPL4 associates with 18S or 28S rRNA, we performed RNA immunoprecipitation using HeLa cells transiently transfected with either FLAG vector or FLAG-tagged human RPL4. RNA–RPL4 complexes in the nucleolar extracts were immunopre- cipitated with anti-FLAG resin, and the RNA compo- nents were resolved on a 1.2% agarose-formaldehyde gel, blotted onto a nitrocellulose membrane and sub- jected to northern blot analysis. Figure 7 shows that A BC D ∆204-264 ∆204-264 anti-FLAG Hoechst Phase IP: anti-FLAG WB: anti-FLAG WB: anti-FLAG 1 st transfection si934Scr FLAG vector FLAG vector FLAG-RPL4 (WT) FLAG-RPL4 (∆204-264) FLAG-RPL4 (∆264-333) si-L4-M1 si-L4-M1 si-L4-M1 si-L4-M1 2 nd transfection 18S 28S 18S 28S 32S 47S/45S 28S± SE 49±4 54±5 84±10 84±10 73±6 83±5 42±7 47±6 1 234 5 100 100 Total± SE WB: anti-HA 204 264 428 428 333 264 333 ∆264-333 ∆264-333 ∆204-264 ∆264-333 FLAG-∆204-264 FLAG-RPL4 (WT) FLAG-RPL4 (∆204-264) FLAG-RPL4 (∆264-333) FLAG-∆264-333 HA-Guα 1 1 Fig. 6. Reversal of inhibition of rRNA production. (A) Schematic representation of human RPL4 deletion mutants D204-264 and D264-333. The open bar represents regions conserved between human and mouse RPL4. The shaded bar represents nonconserved regions. (B) Indi- rect immunofluorescence showing both D204-264 and D264-333 are localized to nucleoli. (C) Whole cell extracts from HeLa cells cotransfect- ed with HA-tagged human Gua and FLAG-tagged human RPL4 deletion mutant were immunoprecipitated using anti-FLAG resin and blotted as indicated. (D) LAP3 cells were transfected with either 100 n M si934Scr or 100 nM si-L4-M1 as indicated. After 48 h, cells were next trans- fected with FLAG-vector, FLAG-tagged wild-type human RPL4 or FLAG-tagged human RPL4 mutants as indicated. After an additional 48 h, cells were pulse-labeled with [ 32 P]orthophosphate for 1.5 h and chased with cold medium for 3 h. Total RNA was extracted and analyzed as described in Fig. 3. Ethidium bromide staining for 18S and 28S rRNAs is shown in the middle panel. The lowest panel shows expression of the FLAG-tagged human wild-type RPL4 (WT) and its mutant forms (D204-264 and D264-333) by western blot analysis using anti-FLAG Ig. The numbers below the upper panel correspond to the amount of 28S rRNA or total rRNA ± standard error relative to samples in lane 1 (set at 100) calculated with IMAGE-QUANT software. Results were average of three independent experiments ± SE. RNA-Total 28S probe 28S rRNA 18S rRNA 18S probe 12 34 65 RNA-IP- FLAG-vector RNA-IP- FLAG-RPL4 Fig. 7. Human RPL4 associates with 28S but not 18S rRNA. HeLa cells were transfected with either FLAG-vector only or FLAG- tagged human RPL4. After 48 h, cells were collected and RNA- RPL4 complexes were immunoprecipitated from nucleolar extracts using anti-FLAG resin as described under Experimental procedures. RNA components were isolated and resolved in a 1.2% agarose- formaldehyde gel and blotted onto a nitrocellulose membrane, which was subjected to northern blot analysis as described under Experimental procedures. H. Yang et al. Gua–RPL4 interaction in mammalian rRNA production FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS 3795 overexpressed RPL4 pulled-down 28S but not 18S rRNA in a dose-dependent manner (lanes 5 and 6). We did not observe any signal from 47 ⁄ 45S, 36S and 32S pre-rRNAs, the precursors of 28S rRNA. Based on our previous experience [16], a single oligodeoxy- nucleotide probe we used to detect 28S could not detect higher molecular weight pre-rRNAs under our hybridization conditions for unknown reasons. As we used nucleolar extract as the starting material for the immunoprecipitation experiment, the 28S rRNA pulled down by RPL4 should be newly produced. Discussion RNA helicase II ⁄ Gua is the first nucleolar RNA heli- case shown to be directly involved in rRNA processing in both the metazoan and mammalian systems [16,17]. There are several other nonhelicase nucleolar proteins which have been demonstrated to also function in rRNA processing in higher eukaryotes including B23 ⁄ nucleophosmin, C23 ⁄ nucleolin, Bop1, p120 and p19 Arf . Each of these proteins was found to function at least partially through RNA–protein or protein– protein interactions [12,14,38–42]. Among the 18 RNA helicases which have been directly implicated in yeast ribosome biogenesis, at least nine were demonstrated to functionally interact with other protein factors [25,37,43]. Based on the bona fide helicase activity of Gua and its demonstrated role in rRNA processing, it is conceivable that Gua will be shown to have partners that facilitate its function in the ribosome biogenesis pathway. In this paper, we report the identification of ribo- somal protein L4 as a Gua-interacting partner through immunoprecipitation in mouse LAP3 cells (Fig. 1A). We noticed that several other fast migrating bands were also pulled down by anti-FLAG resin (Fig. 1A, other proteins), which we suspect to be proteins associ- ated with either Gua or RPL4. The high concentration of RNase A (200 lgÆmL )1 ), which was used in previ- ous reports to isolate specific target-associated proteins [31,44], suggests that these additional interactions might not be RNA-mediated. The Gua–RPL4 inter- action was further confirmed by immunoprecipitation from HeLa cells (Fig. 1D,E), yeast two-hybrid analysis (Fig. 1B) and in vitro binding assay (Fig. 1C). It is noteworthy that Gub also interacts with RPL4 as shown by the two-hybrid analysis (Fig. 1B, lower right). As a paralogue of Gua,Gub also possesses in vitro ATPase and helicase activities, but no RNA foldase activity [45]. The current data suggest that both paralogues arose through gene duplication but the resulting genes are differentially regulated and might possess different functions [46]. Overexpression of Gub in mouse LAP3 cells leads to inhibition of total rRNA production, suggesting contrasting roles for Gub and Gua [17]. It would be valuable to deter- mine whether the inhibitory effect of Gub on rRNA biogenesis is through its competitive interaction with RPL4. Indirect immunofluorescence showed a predom- inant localization of newly produced FLAG-tagged RPL4 protein to the nucleolus (Fig. 1F) which is con- sistent with the published report that most newly formed ribosomal proteins are highly concentrated in the nucleolus [47]. Burial of FLAG epitope within the highly structured mature ribosome subunit might account for the absence of strong fluorescent signal in the cytoplasm. The distribution of RPL4 in the nucleo- lus seems more localized compared with the more dis- persed localization of Gua throughout the entire nucleolus (Fig. 1F). This subtle discrepancy between the localizations of the two proteins may indicate that Gua interacts with other partners in different sub- nucleolar regions, which is consistent with the presence of other additional bands shown in Fig. 1A. Moreover, our previous immunoelectron microscopy experiments showed that rat Gua is localized to the dense fibrillar component (DFC) and granular component (GC) within the nucleolus [16]. An interesting finding was the importance of the DEVD motif in Gua–RPL4 interaction (Fig. 2). We previously showed that in Xenopus, the DEVD motif of Gua was important for both 18S and 28S rRNA production [16]. However, in mammalian cells, we were only able to prove that SAT motif is necessary for 18S maturation [17]. As the unwinding activity is dependent on the ATPase activity, we speculated that the DEVD motif of Gua is also necessary for 18S pro- duction in mammalian cells. In this report, we showed the conserved importance of the DEVD motif of Gua to 28S maturation in mouse LAP3 cells (Fig. 3). Because DEVD is important for 28S production as well as Gua–RPL4 interaction, and because RPL4 is a component of the ribosome large subunit which con- tains 28S but not 18S rRNA, it is reasonable to sus- pect that RPL4 might be involved in the function of Gua in 28S rRNA production. Through a series of deletion mutants of RPL4 used in the immunoprecipitation and indirect immunofluo- rescence experiments, we identified the NLS and NoLS as well as the Gua-interacting domains in RPL4 (Figs 4 and 5). However, it is worth mentioning that the use of deletion mutants may not accurately reflect the exact functional states of protein inter- actions since the possibility exists that the shortened proteins may be unfolded and thus nonfunctional. Gua–RPL4 interaction in mammalian rRNA production H. Yang et al. 3796 FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS Subtle point mutations in the identified Gua-interacting domain might lend more support to our conclusions. Downregulation of RPL4 via si-L4-M1 resulted in the inhibition of production of all four rRNA species (Figs 5D and 6D lane 2), strongly suggesting a general mechanism whereby RPL4 modulates rRNA biogen- esis through rDNA transcription, rRNA turn over, ribosome production rate, ribosome stability or rRNA degradation. It is possible that the amount of RPL4 in the cell correlates with the assembly or stabilization of pre-rRNA processing machineries or preribosomal par- ticles. Perhaps RPL4 is actually a component of the pre-rRNA processing machinery. If so, a dramatic change in the amount of RPL4 protein level might lead to disassembly of the specific machineries or parti- cles, which would send feedback signals to RNA polymerase I to advance more slowly or even to rRNA- degrading complexes to degrade the unincorporated mature rRNA [17]. This hypothesis might help to interpret the involvement of other nucleolar proteins in pre-rRNA production such as p19 Arf [15]. It might also explain why there is a decrease of 18S rRNA. More- over, inhibition of the 28S pathway might concomit- antly result in the reduction of 18S rRNA through an unidentified mechanism. For example, many yeast mutants with 25S rRNA production defects also show an inhibition in 35S pre-rRNA cleavages which lead to decrease in 18S biogenesis [48]. Other than the proposed general function of RPL4 in overall rRNA production, we also hypothesize that RPL4 plays a direct role in 28S production through its interaction with Gua. Several arguments and lines of evidence support this: (a) our rescue experiment showed that wild-type human RPL4 reversed the aber- rant rRNA processing pattern (Fig. 6D, compare lanes 2 and 3) but the mutant lacking the Gua-interacting domain had no effect (Fig. 6, compare lanes 2 and 5); (b) inhibition of 28S rRNA production is more signifi- cant than that of 47 ⁄ 45S when RPL4 is downregulated (Fig. 5D, lanes 1 and 2); (c) RPL4 is an important component of the 60S ribosome subunit which con- tains the 28S but not the 18S rRNA; (d) RNA immu- noprecipitation revealed coprecipitation of RPL4 with 28S but not with 18S rRNA (Fig. 7); (e) ATPase activ- ity of Gua is important for both 28S production and Gua–RPL4 interaction. These five lines of evidence support a more direct role for the Gua–RPL4 inter- action in 28S production than a possible general mechan- ism, although it is likely that both mechanisms coexist. It is not uncommon for a nucleolar protein to function in different pathways. For example, the function of C23 in ribosome biogenesis is reflected in almost all steps of the process including rDNA transcription, pre-rRNA processing, preribosome assembly and nucleocytoplasmic transport [39]. What then could be a mechanism whereby the Gua– RPL4 interaction facilitates 28S rRNA biogenesis? The fact that Gua and RPL4 have been identified in ribo- nucleoprotein (RNP) particles [31,49] indicates that their interaction might cause them to be localized into pre-rRNA processing machineries essential for pre60S ribosome particles. It is known that interruption of early assembly steps results in disassembly of the parti- cles and destabilization of pre-rRNAs [43]. Moreover, we did observe several other bands in the immunopre- cipitation assay (Fig. 1A) coimmunoprecipitating with Gua, which may represent other proteins in the same processing machinery as Gua. Once Gua has been incorporated into the RNP particle, it might function in early rRNA processing steps such as regulating interactions between guide snoRNAs and pre-rRNAs, helping the endo- and exo-nucleases in removing inter- nal or external transcribed spacer sequences as well as modulating the numerous trans-acting factors and ribosomal proteins in the pre60S particles through regulation of RNA-RNA, RNA–protein and protein– protein interactions [43]. In yeast S. cerevisiae, involve- ment of ATP-dependent RNA helicase has been implicated in each of these possible roles [50–54]. The functional diversity of an RNA molecule is based on its extreme flexibility. With the help of proteins, RNA retains or loses its active configuration in response to various different cellular signals [55]. RNA helicase is a candidate to function in an energy-dependent manner in this process. In addition, Gua has another activity, GTP-stimulated RNA folding activity which resides within a domain separate from the ATPase ⁄ helicase activity. Proteins utilizing GTP as an energy source have recently been found to participate in ribo- some biogenesis [56,57]. In addition, it was recently reported that GTP-binding state might influence the nucleolar targeting of nucleostamin, a nucleolar pro- tein which shuttles between nucleoplasm and nucleolus with suspected roles in cell cycle and cell proliferation regulation [58–60]. Interestingly, the C-terminal FRGQR-containing region of Gua has also been reported to be critical for both GTP-stimulated RNA foldase activity and nucleolar localization [20,61]. It remains to be identified if the FRGQR region of Gua is relevant to GTP–binding or RPL4 interaction. It would not be surprising if the Gua–RPL4 interaction is found to be important in all three roles mentioned considering the complexity of ribosome assembly, which highly demands versatile energy producers and consumers. The multifunctional property of Gua makes it a good candidate. H. Yang et al. Gua–RPL4 interaction in mammalian rRNA production FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS 3797 [...]...Gua–RPL4 interaction in mammalian rRNA production In summary, the data in this report suggest that RPL4 functions in a general manner in rRNA biogenesis whereas the Gua–RPL4 interaction is more direct in the production of 28S rRNA and maturation of pre60S particles More questions arise such as whether RPL4 recruits Gua or vise versa, what is the exact mechanism of Gua ⁄ RPL4 involvement in rRNA processing,... Bertwistle D, Sugimoto M & Sherr CJ (2004) Physical and functional interactions of the Arf tumor suppressor protein with nucleophosmin ⁄ B23 Mol Cell Biol 24, 985–996 32 Bagni C, Mariottini P, Annesi F & Amaldi F (1993) Human ribosomal protein L4: cloning and sequencing of the cDNA and primary structure of the protein Biochim Biophys Acta 1216, 475–478 FEBS Journal 272 (2005) 3788–3802 ª 2005 FEBS H Yang... BC, Henning D, Perumal K & Busch H (1997) RNA- unwinding and RNA- folding activities of RNA helicase II ⁄ Gu: Two activities in separate domains of the same protein Eur J Biochem 250, 800–807 3800 H Yang et al 20 Ou Y, Fritzler MJ, Valdez BC & Rattner JB (1999) Mapping and characterization of the functional domains of the nucleolar protein RNA helicase II ⁄ Gu Exp Cell Res 247, 389–398 21 Westermarck... silver staining The protein bands showing only in the Guaexpressed samples were excised and analyzed by mass spectrometry Yeast two-hybrid analysis The cDNAs for human Gua and human RPL4 were subcloned into pGBKT7 and pGADT7 yeast expression vectors, respectively Protein protein interaction was determined in yeast exactly as previously described [62] In vitro binding assay Recombinant proteins were expressed... is required for 60S ribosomal subunit assembly RNA 4, 566–581 52 Liang WQ, Clark JA & Fournier MJ (1997) The rRNA-processing function of the yeast U14 small nucleolar RNA can be rescued by a conserved RNA helicase- like protein Mol Cell Biol 17, 4124– 4132 53 Weaver PL, Sun C & Chang TH (1997) Dbp3p, a putative RNA helicase in Saccharomyces cerevisiae, is required for efficient pre-rRNA processing predominantly... The DEXD ⁄ H-box RNA helicase RHII ⁄ Gu is a co-factor for c-Jun-activated transcription EMBO J 21, 451–460 22 Valdez BC, Henning D, Perlaky L, Busch RK & Busch H (1997) Cloning and characterization of Gu ⁄ RH-II binding protein Biochem Biophys Res Commun 234, 335–340 23 Jankowsky E, Gross CH, Shuman S & Pyle AM (2001) Active disruption of an RNA protein interaction by a DExH ⁄ D RNA helicase Science... Takahashi N (2002) Isolation and proteomic characterization of human parvulinassociating preribosomal ribonucleoprotein complexes J Biol Chem 277, 23773–23780 50 O’Day CL, Chavanikamannil F & Abelson J (1996) 18S rRNA processing requires the RNA helicase- like protein Rrp3 Nucleic Acids Res 24, 3201–3207 51 Daugeron MC & Linder P (1998) Dbp7p, a putative ATP-dependent RNA helicase from Saccharomyces... synthase, and Nhp2p, a putative RNA- binding protein, are present together with Gar1p in all H BOX ⁄ ACA-motif snoRNPs and constitute a common bipartite structure RNA 4, 1549–1568 27 Venema J & Tollervey D (1995) Processing of pre -ribosomal RNA in Saccharomyces cerevisiae Yeast 11, 1629–1650 28 Venema J, Bousquet-Antonelli C, Gelugne JP, Caizergues-Ferrer M & Tollervey D (1997) Rok1p is a putative RNA helicase. .. mouse U1C, antisense RT-PCR primer, human and mouse L4, sense RT-PCR primer, human L4, antisense RT-PCR primer, mouse L4, antisense Northern blot analysis probe, for 18S rRNA Northern blot analysis probe, for 28S rRNA examined with a Nikon Eclipse TE2000-U inverted microscope equipped with a Coolsnap digital color camera Metabolic labeling of cells and subsequent RNA analysis The procedure is similar to... Cruz J, Kressler D & Linder P (1999) Unwinding RNA in Saccharomyces cerevisiae: DEAD-box proteins and related families Trends Biochem Sci 24, 192–198 25 de la Cruz J, Lacombe T, Deloche O, Linder P & Kressler D (2004) The putative RNA helicase Dbp6p functionally interacts with Rpl3p, Nop8p and the novel trans-acting factor Rsa3p during biogenesis of 60S ribosomal subunits in Saccharomyces cerevisiae . signal; RPL4, Ribosomal Protein L4; rDNA, ribosomal DNA; rRNA, ribosomal RNA; RNP, ribonucleoprotein; siRNA, small interfering RNA; snRNA, small nuclear RNA; . Functional interaction between RNA helicase II⁄Gua and ribosomal protein L4 Hushan Yang, Dale Henning and Benigno C. Valdez Department