RPLP1, a Crucial Ribosomal Protein for Embryonic Development of the Nervous System

Ribosomal proteins are pivotal to development and tissue homeostasis. RP Large P1 (Rplp1) overexpression is associated with tumorigenesis. However, the physiological function of Rplp1 in mammalian development remains unknown. In this study, we disrupted Rplp1 in the mouse germline and central nervous system (Rplp1CNS Δ). Rplp1 heterozygosity caused body size reductions, male infertility, systemic abnormalities in various tissues and a high frequency of early postnatal death. Rplp1CNS Δ newborn mice exhibited perinatal lethality and brain atrophy with size reductions of the neocortex, midbrain and ganglionic eminence. The Rplp1 knockout neocortex exhibited progenitor cell proliferation arrest and apoptosis due to the dysregulation of key cell cycle and apoptosis regulators (cyclin A, cyclin E, p21CIP1, p27KIP1, p53). Similarly, Rplp1 deletion in pMEFs led to proliferation arrest and premature senescence. Importantly, Rplp1 deletion in primary mouse embryonic fibroblasts did not alter global protein synthesis, but did change the expression patterns of specific protein subsets involved in protein folding and the unfolded protein response, cell death, protein transport and signal transduction, among others. Altogether, we demonstrated that the translation “fine-tuning” exerted by Rplp1 is essential for embryonic and brain development and for proper cell proliferation.


Introduction
The assurance of proper ribosome functionality is essential for the development of all multicellular organisms. Dysfunctions in most RPs induce developmental defects ranging from general translation impairment-related defects to tissue-specific phenotypes [1,2]. More than 50 ribosomal protein (RP)-encoding loci were found to be mutated in a group of developmental abnormalities in Drosophila termed ''minutes'' [3][4][5][6]. ''Minutes'' are characterized by general developmental retardation, a reduced body size, short, thin bristles and diminished fertility, all of which are likely caused by a reduction in protein synthesis. In mammals, RP genes mutations are associated with tissue-specific abnormalities such as the mouse Tail-short (Ts), Tail-short shionogi (Tss) and Rabotorcido (Rbt) mutants, which present with skeletal abnormalities, short, kinky tails and neural tube defects such as exencephaly, spina bifida and cleft palate [7,8]. Recent studies have shown that the developmental defects in Ts mutants are caused by mutations in Rpl38, thus affecting the production of RPL38 [1], which regulates axial-skeletal patterning by modulating the translation of a subset of Hox mRNAs [1]. Similarly, Rpl24 hypomorphism causes Belly spot and tail (Bst) mutants, which present with white hind feet, kinky tails and ventral midline spots. Rpl24 mutations impair Rpl24 mRNA splicing and RPL24 production, thus affecting ribosome biogenesis, protein synthesis and the cell cycle [9].
The ribosomal stalk is a flexible lateral protuberance in the large ribosomal subunit that constitutes the specific elongation factor recognition motif [10][11][12]. In high eukaryotes, this stalk comprises RP Large P0 (P0) and two heterodimers formed by RP Large P1 (P1) and RP Large P2 (P2). This structure is known as the P complex or P proteins. P0 is connected to the rest of the ribosome through the ribosomal protein L12 and the 28S ribosomal RNA [13]. In contrast to other ribosomal proteins, P1 and P2 are not imported into the nucleus for their assembly, and constantly shuttle between the ribosome and the cytoplasm [14,15]. Interestingly, P1 and P2 stabilize each other in the cytoplasm [16,17]. Of importance is that ribosomal stalk protein alterations have been found in human tumors. For example, Rplp1 mRNA levels are increased by five-fold in colorectal cancer tissues [18]. Similarly, human lymphoid cell lines containing mutant P53 were shown to overexpress P1 [19]. We previously found that the RNA and protein levels of P0, P1 and P2 (P proteins) were significantly increased in gynecologic tumors [20]. From a therapeutic point of view, gonadotropin-releasing hormone (GnRH) analogues, which are used to treat breast, prostate and ovarian cancers, have been found to exert their anti-proliferative effects through the downregulation of P1 and P2 [21], suggesting the potential clinical implications of targeting ribosomal stalk proteins. Interestingly, previous work from our laboratory showed that P1 overexpression allows cells to bypass replicative senescence, likely due to cyclin E overexpression consequent to increased E2F1 promoter activity [22]. Moreover, P1 was found to cooperate with Ras Val12 in the transformation of murine NIH3T3 cells [22].
The requirement of ribosomal stalk proteins in proliferating cells differs vastly among species. S. cerevisiae homologs of P1 and P2 are dispensable for viability [23]. In contrast, the depletion of P2 in human cells does not affect viability, but rather impairs proliferation, likely by affecting the efficiency of eukaryotic translation initiation factor 5B (IF-2)-mediated ribosome assembly [16,24].
To date, the physiological function of Rplp1 has remained elusive. This study explored the role of Rplp1 in development and proliferation using Rplp1-deficient mouse models and derived cells. As proteins from the P complex are highly expressed in the fetal brain [25], we decided to explore the role of the P1 protein in brain development. We found that Rplp1 was essential for embryonic and brain development. Rplp1 deletion induced proliferation defects and apoptosis in vivo. Moreover, Rplp1 deletion caused a senescence-associated proliferation arrest in primary mouse embryonic fibroblasts (pMEFs). Overall, we propose that Rplp1 absence provokes a stress response associated with misfolded proteins that induces a different translation pattern, rather than a general disturbance of ribosomal function and/or protein synthesis.

Cell culture
Mouse embryonic fibroblasts (MEFs) were cultured in Dulbecco's Modified Eagle Medium (DMEM; Invitrogen Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal calf serum (FCS; Lonza, Basel, Switzerland), 2 mM L-glutamine (Invitrogen), 1 mM sodium pyruvate (Invitrogen, 11360-039), 100 units/mL of penicillin, 100 mg/mL of streptomycin (Pen/ Strep; Invitrogen) and 0.5 mM b-mercaptoethanol (Invitrogen). Mouse embryonic stem (ES) cells (blastocysts derived from mouse strain 129/Sv) were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Invitrogen) supplemented with 15% fetal calf serum (FCS) (Lonza), 2 mM L-glutamine (Invitrogen), 1 mM sodium pyruvate (Invitrogen), 1 mM MEM non-essential amino acids (Invitrogen), 100 units/mL of penicillin, 100 mg/mL of streptomycin (Pen/Strep; Invitrogen), 0.5 mM b-mercaptoethanol (Invitrogen) and 1000 U/mL LIF (Chemicon, Billerica, MA, USA). The culture medium was changed daily. Fresh medium was prepared every 2 to 3 days. All animal procedures followed Association for Assessment and Accreditation of Laboratory Animal Care guidelines and were approved by institutional Animal Care and Use Committee. Stable transfection was performed with 30 mg of retroviral vectors LMP shRNA (control) LMP shp53 , and LMP shp16INK4A into Phoenix cells. Phoenix cells were maintained in a Minimum Essential Media (Gibco) supplemented with 50 mg/mL penicillin-streptomycin and 10% FBS. After 48 hours, the Phoenix supernatant was added into MEF cells, which were selected with 1 mg/mL puromycin and further treated with OHT during 4 days in order to induce the Rplp1 knock-down. For the transient transfection 2610 5 MEF cells were seeded per well in 6-well plates. After 24 h they were transfected with Lipofectamine 2000 with the p16 INK4AsiRNA , p53 siRNA and Cy3 dye-labeled negative control (AM17010). Then, the cells were treated with OHT during four days in order to induce the Rplp1 knock-down. Stable cell lines were done according.

Primary MEF (pMEF) immortalization
Early passage pMEFs (P2) were immortalized by p19 ARF knockdown. A plasmid encoding p19 ARF shRNA was transfected into the Phoenix E retroviral packaging cell line (based on HEK293 cells). After 15 hours, the medium was changed and the cells were incubated at 32uC to facilitate virus production. Fortyeight hours later, the viral supernatant was collected, filtered through a 0.45-mm filter and mixed 1:1 with MEF medium and polybrene to a final concentration of 8 mg/ml. The mixture was incubated at RT for 5 minutes and then added to 30% confluent primary MEFs. The pMEFs were incubated at 32uC with 5% CO 2 . Twenty-four hours later, the medium was changed and the cells were transferred to a 37uC incubator. The cells were selected for 4 days post-infection with 2 mg/ml of puromycin.

Growth curve
The growth curves were performed with freshly isolated pMEFs. Briefly, 1.5610 5 cells per clone were seeded in triplicates into 6-well plates. The triplicates were treated with 1 mM 4hydroxytamoxifen (4-OHT) (Sigma-Aldrich, St. Louis, MO, USA) for 4 days, and the other triplicates were used as an untreated control. Every three days, the cells were trypsinized, counted and re-plated similarly. The growth curves were constructed by plotting the cumulative cell number versus the passage number.
Proliferation assay pMEFs were treated with 1 mM 4-OHT for 4 days. Three days later, the cells were incubated for 2 hours with 10 mM bromodeoxyuridine (BrdU). After incubation, the cells were washed with PBS 1X, harvested and fixed with ice-cold ethanol. Next, the cells were centrifuged at 10006g for 10 minutes. The supernatants were aspirated and the pellets resuspended in 3 mL of 0.08% pepsin in 0.1 M HCl, followed by an incubation at 37uC for 20 minutes with occasional mixing. The cells were then centrifuged as described above, and the pellets were resuspended in 1.5 mL of 2 M HCl. The mixtures were incubated at 37uC for 20 min. Next, 3 mL of 0.1 M sodium borate were added, and the samples were centrifuged. The pellets were resuspended in 2 mL of IFA/ Tween20 (0.5% Tween 20 in IFA) and centrifuged. The cells were subsequently stained with 75 mL of a FITC-conjugated anti-BrdU antibody (Becton Dickinson, Franklin Lakes, NJ, USA) at a 1:5 dilution in IFA (10 mM HEPES, pH 7.4; 150 mM NaCl; 4% fetal bovine serum; 0.1% sodium azide) for 30 minutes on ice. Next, 2 mL of IFA/Tween 20 were added, and the cells were centrifuged and resuspended in 0.25 mL of IFA. Finally, 0.25 mL of a 20 mg/mL propidium iodide solution (PI) were added. The percentages of BrdU-positive cells were determined on a flow cytometer (FACSCanto, BD Biosciences; software FACS Diva, San Jose, CA, USA).

Cell cycle
Confluent primary or immortalized MEFs were trypsinized and resuspended in 1 mL of PBS 1X and subsequently fixed with icecold absolute ethanol for at least one hour. The cells were then incubated with DNase-free RNaseA (100 mg/mL; Sigma-Aldrich) at 37uC for 30 minutes, after which 100 mL of 1 mg/mL propidium iodide (PI) (Sigma) were added. Finally, the cell cycle profiles were collected using a flow cytometer (FACSCanto, BD Biosciences; software, FACS Diva) (BD Biosciences).

Apoptosis assay
Primary or immortalized MEFs were trypsinized and washed twice with PBS 1X. The supernatant was also collected to include the late apoptotic cells. The cells were resuspended in 0.1 mL of binding buffer (10 mM HEPES, pH 7.4; 140 mM NaCl; 2.5 mM CaCl 2 ) containing Annexin V-FITC (1:100 dilution; BD Biosciences). The mixture was incubated for 15 minutes at RT in the dark. Next, 400 mL of DAPI-containing binding buffer (final concentration, 0.2 mg/mL) were added. The samples were kept on ice for 15 minutes and then analyzed by flow cytometry.

Senescence assay
Primary or immortalized MEFs growing in culture dishes or on coverslips were washed once with PBS 1X and fixed with 2% formaldehyde and 0.2% glutaraldehyde for 10 minutes at RT. Next, the cells were washed twice with PBS 1X and stained with a b-galactosidase staining solution (40 mM citric acid/sodium phosphate pH 6.0, 0.15 M NaCl, 2 mM MgCl 2 , 5 nM potassium ferrocyanide, 5 nM potassium ferricyanide, 1 mg/mL of X-Gal). The cells were incubated at 37uC in a dry incubator overnight. The reagents are included with the Senescence b-galactosidase staining kit (Cell Signaling, Danvers, MA, USA).

Protein synthesis
To measure protein synthesis, we used a system based on the azide-alkyne reaction (Click-it Protein synthesis kit) (Invitrogen). Briefly, newly synthetized proteins are labeled with an azide (or alkyne)-coupled methionine analog. A fluorophore or HRPconjugated alkyne (or azide) is then used for detection. In our study, we used a biotin-conjugated alkyne to perform western blotting. Rplp1 F/F ; Cre-ER T2 + and Rplp1 +/F ; Cre-ER T2 + immortalized MEFs were treated with 1 mM 4-OHT for 4 days or left untreated (control). Three days later, the cells were washed once with warm PBS and incubated for 1 hour with methionine-free DMEM (Invitrogen) to deplete the methionine reserves of the cells. Next, the medium was removed, and the cells were incubated for 3 hours in methionine-free DMEM with the methionine analog azide-homoalanine (AHA; Invitrogen). The cells were then washed once with PBS, followed by the addition of lysis buffer (1% SDS, 50 mM Tris-HCl pH 8, 1 mM Na 3 VO 4 , 10 mM NaF, 0.1 mM PMSF, 1X Roche Complete Protease Inhibitor Cocktail; Roche, Basel Switzerland). The cells were collected with a cell scraper and incubated for 30 minutes in lysis buffer. Next, the cell lysates were sonicated (5 30-second on/off cycles; BioruptorPlus; Diagenode, Seraing, Belgium) and centrifuged at 15,7006g for 15 minutes at 4uC. The supernatant was transferred to a fresh Eppendorf tube. A Bradford assay was used to measure the protein concentration as follows: 1 mL of protein extract was mixed with 999 mL of a 1:5 dilution of Bradford solution (Bio-Rad) in water. The absorbance at 595 nm was determined with a spectrophotometer (Eppendorf BioPhotometer; Hamburg, Germany). Subsequently, 50 mg of protein were transferred to a new tube, to which were added 100 mL of biotin-alkyne in 2x reaction buffer (Invitrogen) and water to a final volume of 160 mL. The following reagents were then added sequentially: 10 mL of reagent C (CuSO 4 ) (Invitrogen), 10 mL of reagent D (buffer additive 1) and 20 mL of reagent E (buffer additive 2). The mixture was shaken for 20 minutes, and then 600 mL of methanol, 150 mL of chloroform and 400 mL of water were added. The mixture was vortexed and centrifuged for 5 minutes at 13,2006g. The upper aqueous phase was discarded, and 450 mL of methanol were added. The protein solution was centrifuged as described above, and the methanol was discarded. The remaining protein pellet was washed again with methanol as described above, after which the pellet was air-dried. The pellet was resuspended in SDS-sample buffer 1X (see 2.6.2), vortexed for 10 minutes and then heated for 10 minutes at 70uC. A quarter of the mixture was separated by SDS-polyacrylamide gel electro-phoresis and then transferred to a PVDF membrane (Bio-Rad). The membrane was blocked for 30 minutes with 5% BSA in TBS-T and then for 30 minutes with an ABC peroxidase reagent (Vectastain Burlingame, CA, USA), which reagent contains a biotinylated horseradish peroxidase (HRP) that had been preincubated with avidin at a specific ratio to form large complexes. The membrane was washed 3 times for 5 minutes each at RT. Finally, the membrane was developed with an ECL kit (Pierce, Appleton, WI, USA). Cells that had been treated with 1 mg/mL of cycloheximide (Sigma), which inhibits the elongation step of protein synthesis, were used as a negative control. The PCR reaction conditions were as follows: For Rplp1, an initial denaturation step was performed for 3 minutes at 94uC, followed by 35 cycles of denaturation for 15 seconds at 94uC, primer annealing for 1 minute at 57uC and elongation for 2 minutes at 72uC and a final elongation step for 7 minutes at 72uC. For Cre, an initial denaturation step was performed for 3 minutes at 94uC, followed by 35 cycles of denaturation for 30 seconds at 94uC, primer annealing for 1 minute at 58.5uC and elongation for 2 minutes 30 seconds at 65uC and a final elongation step for 7 minutes at 65uC.

Southern Blot
Genomic DNA from cells or mouse tissues was digested either with AseI for the short arm and PflFI for the long arm or HindIII. The digested samples were separated on a 0.8% agarose gel in TAE buffer (40 mM Tris Base, 20 mM acetic acid, 1 mM EDTA). The DNA was fragmented by depurination in 250 nM HCl for 20 minutes and denatured in 500 mM NaOH, 1.5 M NaCl for 30 minutes. The DNA was then transferred to a Hybond XL membrane (GE Healthcare Life Sciences, Cleveland, Ohio). Next, the membrane was incubated for 20 minutes in 40 mM NaPi and baked for 1 hour at 80uC, followed by DNA crosslinking (UV Stratalinker; Stratagene, La Jolla, CA, USA). The membrane was then prehybridized in Church's buffer (500 mM NaPi pH 8, 7% SDS, 1 mM EDTA) for one hour at 65uC. Subsequently, the membrane was hybridized overnight with a dCTP a-32 P radio labeled probe in Church's buffer at 65uC. The sequences of the genomic probes were as follows: P3LA (for long arm): tcatttggctaggtggcaaggactgtgcctgcagattctccgtcactatcagttttataaattg gagatgctcgcagttgctcatgctactatttcatatacatgaaaaaagatttgtttaaaaagtcg aatgtgctatttaacttggtacagctggcatttgactggaaaacaggtagttttttattataggat acaagctgatttcatagtgtagtcttcgattcttggagcgtttaagaaaagtttggagaaaaaca ttgctggtatattttgtggaaatgtgttttgaaatatagaaatctttactgcaagtgctcagaggtg ctttcactgttgaggcacggtacatgtgaagtacctgttcctgtgttctacagggttctcctctacc agagtttcaacagcttgtactttttgatatggctgtatatgaagaatgttttctatttagatgcctca gacacacttaagcgttatattgatcctgctgatacttttaacagacctgctatttattacttcacaa agcagccataataattgggttttgatcaagaatgcgaaatggattatttgtcctgagctagccat aaacaggaagacctagagtcacgacaactgtaccctttgccctttgttgcaatggaagctccc aaacagactatctcttcttcatgctgttgtccaaagatgcctgttggttgttgactaagaagaga acgttgaatgattttcatgtttagtgtgggcaggtttgatttcggtgaacttgcattgtaaactgct gcacttctaatggaacctcaaagtaatcctgggttgtgcttgtttcccaat.
The following day, the membrane was washed twice in washing buffer (1% SDS, 30 mM NaPi) for 5 minutes at 55uC, once for 15 minutes at 65uC and once for 30-45 minutes at 65uC. Finally, the membrane was exposed to a phosphoimager overnight and subsequently to X-Ray film for several days.

Protein isolation
For protein isolation, the cells were trypsinized and washed twice with ice-cold PBS. After centrifugation at 2406g and supernatant removal, the cell pellet was resuspended in 50 L of RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 0.25% sodium deoxycholate, 1 mM Na 3 VO 4 , 10 mM NaF, 0.1 mM PMSF, 1% NP-40, 1X Roche Complete Protease Inhibitor Cocktail; Roche) and incubated for 30 minutes on ice. The lysate was then sonicated (30-second on/off cycles; Bioruptor Plus; Diagenode) and centrifuged at 15,7006g for 15 minutes at 4uC, followed by transfer to a fresh Eppendorf tube and storage at 280uC. A Bradford assay was used to measure the protein concentration as follows: 1 mL of protein extract was mixed with 999 mL of a 1:5 dilution of Bradford solution (Bio-Rad) in water.
The absorbance at 595 nm was determined with a spectrophotometer (Eppendorf BioPhotometer; Eppendorf). For protein isolation from mouse tissues, the tissues were previously snapfrozen in liquid nitrogen and stored at 280uC. A single tissue piece was minced with a sharp scalpel, followed by the addition of RIPA buffer. The tissue was subsequently homogenized with a tissue homogenizer (Polytron PT 2500E). The tissue lysate was incubated for 30 minutes on ice, followed by the above-described procedure.

Western blot
For western blot analysis, the protein samples were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE). After separation, the proteins were transferred to a PVDF membrane (Bio-Rad) that was subsequently blocked with 5% non-fat dried milk (NFDM) in TBS-T or in 5% BSA in TBS-T if antibodies against phosphorylated proteins were to be used. After blocking, the membrane was incubated overnight at 4uC with the primary antibody diluted in blocking solution. The membrane was then washed and incubated for 30 minutes to 2 hours at RT with the HRP-conjugated secondary antibody diluted in 5% NFDM in TBS-T. Finally, the membrane was washed and developed with an ECL kit (Pierce).

Two-dimensional gel electrophoresis
Protein extracts from three different Rplp1 iD pMEFs that had been treated with 4-OHT for four days or not (control) were separated by two-dimensional gel electrophoresis. The dysregulated proteins (p,0.01) in the Rplp1 iD pMEFs were selected and identified by ESI-MS as described in [26].

ROS assay
Control and inducible pMEFs were treated with OHT for 4 days. DCF-DA was then added to the cells for a 30 minutes incubation. The amount of DCF (the oxidized, fluorescent form of DCF-DA; Invitrogen) was then analyzed by flow cytometry.

Isolation, fixation and embedding of mouse tissues
To isolate mouse or mouse embryonic tissues, the mouse was first sacrificed by cervical dislocation. After cleaning with 70% ethanol, the abdominal wall and peritoneum were opened to expose the organs or embryos, which were removed with forceps and sharp scissors, washed in PBS and fixed ON in 4% PFA at 4uC or in Roti-Histofix (4% formaldehyde, Roth) at RT. To isolate the brains from E18.5 embryos, the heads were removed and immersed in PBS in a Petri dish. The skulls were carefully removed with fine forceps. This procedure was performed under a stereomicroscope (Zeiss Stemi 2000; Zeiss, Oberkochen, Germany). After fixation, the tissues were washed in PBS and incubated for 30 minutes at 4uC in 30% ethanol, followed by a transfer to 50% ethanol, processing and paraffin embedding. The paraffin-embedded tissues were then cut into 5-mm-thick slices with a microtome (Microm, Cavriago, Italy). The slices were mounted on glass slides and dried overnight at 55uC. Hematoxylin and eosin (H&E) staining or immunohistochemistry (IHC) were then performed. All of the procedures performed in this study were approved by the Ethical Committee of Animal Research at the Leibniz Institute for Age Research.

IHC and microscopy
To perform IHC analyses of the mouse embryonic brains, the paraffin sections were first deparaffinized. For antigen retrieval, the sections were incubated in 10 mM sodium citrate, pH 6.0 for 10 minutes at a sub-boiling temperature in a microwave. The sections were allowed to cool at RT for 30 minutes and were then washed three times with PBS for 5 minutes each at RT. Next, the sections were incubated in blocking solution (1% BSA, 5% goat serum, 0.4% Triton X-100 in PBS) to avoid non-specific antibody binding for 1 hour at RT in a humidified chamber. The blocking solution was then removed, and a primary antibody dilution in blocking solution was added. The sections were incubated overnight at 4uC in a humidified chamber with the primary antibody. Next, the sections were washed three times in PBS at RT and incubated with a fluorochrome-conjugated secondary antibody diluted in blocking solution. The sections were then washed three times in PBS for 5 minutes each and finally mounted with coverslips using a DAPI-containing mounting medium (ProLong gold; Invitrogen). The mounted sections were air-dried for several hours and stored temporarily at 4uC and long-term at -20uC. IHC images were acquired with a fluorescence microscope (Zeiss Axio Imager ApoTome or Zeiss Axio Imager M1; Zeiss) coupled to a charge-coupled device (CCD)-camera (Zeiss Ax-ioCamMRm) and AxioVision software (Carl Zeiss). Bright-field images were obtained with a Zeiss Axio Imager M1 microscope fitted with a CCD-camera (Zeiss AxioCam MRc5). The number of cells positive for a certain marker was determined with the ImageJ software program.
Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) of mouse brain sections Paraffin sections of different embryonic brain stages were deparaffinized and incubated in 10 mM sodium citrate, pH 6.0 for 10 minutes at a sub-boiling temperature. The sections were cooled at RT for 30 minutes and washed three times for 5 minutes each at RT with PBS. The sections were incubated with a TUNEL reaction mixture (1x buffer, Amersham; 0.3 units/mL of terminal deoxynucleotidyltransferase, Amersham; 6.66 mM biotin-linked dUTP, Roche Diagnostics; water; Amersham, GE Healthcare Life Sciences) for 1 hour at 37uC in a humidified chamber. The sections were then washed three times for 5 minutes each in PBS at RT and incubated with a 1:500 dilution of streptavidin-Cy3 (Sigma-Aldrich) in PBS with 1% BSA for one hour at RT. The sections were washed three times for five minutes each with PBS and mounted with coverslips, using a DAPI-containing mounting medium (ProLong gold; Invitrogen). The mounted sections were air-dried for several hours and stored temporarily at 4uC and longterm at 220uC. In vivo bromodeoxyuridine (BrdU)-proliferation assay Pregnant females at E13.5 were injected intraperitoneally with a single 50-mg/g mouse body weight dose of BrdU (Sigma). After 30 minutes, the females were sacrificed by cervical dislocation and the embryos were collected. The embryonic heads were removed and fixed in 4% PFA at 4uC overnight. Portions of the embryos were used for DNA isolation and further analysis by PCR. The embryonic heads were processed as described in the IHC section. After deparaffination, the sections were incubated in 10 mM sodium citrate, pH 6.0 for 10 minutes at a sub-boiling temperature, cooled at RT for 30 minutes and washed three times for 5 minutes each with PBS at RT. Next, the sections were incubated in 2 M HCl at 37uC for 30 minutes to denature the DNA. The sections were washed three times in PBS for 5 minutes each, followed by a trypsin incubation (Zytomed, ZUCO43-15) for 20 minutes at 37uC. The sections were then washed three times for 5 minutes each in PBS and incubated for one hour at RT in a humidified chamber with blocking solution (1% BSA, 5% goat serum, 0.4% Triton X-100 in PBS). Finally, immunostaining was performed with an anti-BrdU antibody (Abcam, Cambridge, UK). The antibodies used are described in Table S1.

Generation of Rplp1 knockout mice
To dissect the role of Rplp1 in vivo, we generated Rplp1 constitutive and conditional knockout mice. Briefly, the two homologous regions of the Rplp1 gene were amplified by PCR from Sv129/J background mouse genomic DNA. These PCR fragments were subcloned into the DTA vector, which contains a neomycin (neo) resistance gene to ampicillin and the DTA (Diphtheria Toxin A) gene. A single LoxP site and anFRT-Neo-FRT-LoxP cassette were inserted by homologous recombination upstream of exon 1 and downstream of exon 3, respectively ( Figure 1A). After gene targeting in embryonic stem (ES) cells, a southern blot analysis with 59-and 39-external probes confirmed the correct gene targeting event in the ES cells ( Figure 1A and 1B). The targeted ES clone (1H1) was injected into blastocysts of C57BL/6N mice to generate chimeras that gave rise to germline offspring carrying the Rplp1 targeted (T) allele. The germline offspring (Rplp1 +/T ) was confirmed by southern blotting (Figure 1C and 1D). The intercrossing of Rplp1 +/T mice with mice expressing Flp-recombinase (FLP) or Nestin-cre (NesCre) generated Rplp1

Rplp1 heterozygous mice exhibit severe developmental defects
We obtained Rplp1 +/D and Rplp1 F/D mice (Rplp1 heterozygous mice, Rplp1 Het ), which carry only one functional allele of Rplp1, from matings between Rplp1 +/T ; NesCre + mice and between Rplp1 +/F ; NesCre + mice. Rplp1 Het mice were born at a significantly reduced frequency (Figure 2A). Moreover, approximately 50% of the Rplp1 Het mice died during the early postnatal period. Rplp1 Het newborn mice were smaller than the control littermates ( Figure 2B) and exhibited a yellow abdominal coloration ( Figure 2C). A histological analysis of the intestinal sections revealed perforations in the small and large bowel. Yellow-colored meconium ileus was present in the lesions. The epithelial structure was destroyed and inflammatory cells were present around these lesions ( Figure 2D). Additional abnormalities were observed in the liver, where there was very little hematopoiesis and erythroblasts and granular cells were lacking ( Figure S1). The thymuses in Rplp1 Het mice were much smaller and exhibited abnormal organization. The brown fat adipoblasts were poorly differentiated ( Figure S1). Rplp1 Het mice exhibited rigid, shorter, kinky tails ( Figure 2E); a histological analysis of Rplp1 Het tail sections further revealed that the serial cartilage was disorganized and the long bones were abnormally enlarged ( Figure 2F). Furthermore, Rplp1 Het male mice were infertile, making the generation of Rplp1 knockout mice impossible. However, the histological analyses of Rplp1 Het testes sections revealed no differences between the Rplp1 Het and control mice ( Figure S1). The Rplp1 Het mouse phenotype is summarized in Figure 2G.

Rplp1 conditional deletion in the CNS causes perinatal lethality and reduced brain size
To study the role of Rplp1 in the CNS, Rplp1 conditional knockout mice (Rplp1 CNSD ) were generated by crossing Rplp1 floxed mice (Rplp1 F/F ) with Nestin-Cre transgenic mice. Intercrossing between Rplp1 +/F ; NesCre + mice did not yield any Rplp1 F/F ; NesCre + (Rplp1 CNSD ) offspring, indicating that Rplp1 CNSD mice died during development or the early postnatal period ( Figure 3A). We further analyzed the developmental stage at which the Rplp1 CNSD mice died. We found viable Rplp1 CNSD embryos at embryonic day (E) E12.5, E13.5, E15.5 and E17.5. Furthermore, a cesarean procedure at E18.5 allowed us to recover viable Rplp1 CNSD embryos. However, these were small, featured flat heads ( Figure 3B) and had dramatically reduced brain sizes ( Figure 3C). Overall, the mutant mice were unable to survive.
Because the brain defects at developmental stage E18.5 were dramatic, we analyzed the structure of the Rplp1 CNSD embryonic brain at earlier stages. An histological examination revealed that at E12.5, compared with wild-type embryos, Rplp1 CNSD embryos showed a reduced ganglionic eminence, a thinner neocortex and an enlarged right lateral ventricle ( Figure 3D). At E13.5, the mutant embryos exhibited enlarged ventricles, atrophy of the neopallial cortex and roof of the midbrain, a reduced ganglionic eminence and reduced cellularity in the thalamus. At E15.5 the cortex and the midbrain were very small and atrophic, suggesting that brain growth was arrested at E13.5 ( Figure 3D). The brain atrophy and arrested development explained the flattened shape of the head observed at later stages (e.g., E18.5; Figure 3B). At E15.5, the mutant embryos had a thinner cortex, which was associated with an abnormal cortical layer structure ( Figure 3E). Histological analysis showed that the mutant cortex had a thinner ventricular zone (VZ) and was devoid of subventricular zone (SVZ) and cortical plate (CP). Overall, the cellularity was greatly reduced. The efficient deletion of Rplp1 in Rplp1 CNSD embryonic brains was already evident at E13.5 ( Figure 3F). Notably, we found that P2 was also knocked-out in the Rplp1 CNSD brains, possibly because P1 and P2 stabilize each other in the cytoplasm [17].
Taken together, our data show that Rplp1 deletion in the CNS affected brain size and development, thus causing perinatal lethality ( Figure 3D).

Rplp1 deletion in the CNS causes increased apoptosis and proliferation arrest in progenitor cells
To study the causes of the decreased cellularity in the Rplp1 CNSD neocortex, we evaluated apoptosis. Terminal deoxynucleotidyl transferase UTP nick-end labeling (TUNEL) staining revealed increased apoptosis in the neocortexes of the E12.5, E13.5 and E15.5 brain sections ( Figure 4A and 4B). Concurrently, western blot analysis showed that caspase-3 activation was greatly increased in the E13.5 Rplp1 CNSD brains, as determined by its cleavage ( Figure 4C).
Next, we examined whether Rplp1 deletion caused a proliferation defect in the mouse embryonic brain. First, we performed an in vivo BrdU labeling assay by pulse-labeling the E13.5 embryos with BrdU for 30 minutes. We found that the number of BrdU + cells was significantly reduced in the Rplp1 CNSD embryonic brain, thus indicating a proliferation defect ( Figure 5A and 5B).
Next, immunostaining of the E13.5 brain sections with an anti-Ki67 antibody revealed that the number of Ki67 + cells was significantly reduced in the Rplp1 CNSD embryonic brain, compared with the controls (Figure 5C and 5D), suggesting the presence of fewer cycling cells. Staining of the Rplp1 CNSD embryonic brain sections with an antibody against phosphorylated histone-3 (PH 3) consistently revealed reduced mitosis in the Rplp1 CNSD neocortex ( Figure 5E and 5F). To further analyze the proliferation and apoptosis of Rplp1 CNSD neural progenitors, we performed a primary neurosphere formation assay and found that the Rplp1 CNSD neural progenitors formed very few neurospheres and the sizes of the formed neurospheres were dramatically reduced ( Figure 5I and 5J), indicating compromised proliferation and enhanced apoptosis.
Consequent to the reduced proliferation and increased apoptosis, postmitotic neurons (Tuj1+) were found to be greatly reduced in the mutant neocortex ( Figure 5G and 5H). These observations well explain the absence of the CP and intermediate zone (IZ) layers ( Figure 3E) and contribute to an explanation of the brain atrophy of Rplp1 CNSD mice.
To further understand the proliferation defects observed in the Rplp1 CNSD embryonic brains, the key cell cycle regulators in the Rplp1 CNSD E13.5 brain were analyzed. The cyclin A and cyclin E protein levels were reduced in the mutant brain ( Figure 6). Moreover, the protein levels of the CDK inhibitors p21 CIP1 and p27 KIP1 were reduced in the mutant brain ( Figure 6). Similarly, p53 expression was slightly downregulated in the Rplp1-deleted embryonic brain ( Figure 6) suggesting that p21 CIP1 and p27 KIP1 inhibition can be responsible for the apoptosis observed in the brain [27,28]. We conclude that Rplp1 deletion in the developing CNS caused perinatal lethality, reduced brain size and morphological brain defects consequent to greatly increased apoptosis in association with a proliferation defect.

Rplp1 deletion in MEFs causes a p16 INK4A /pRb pathwaymediated proliferation defect and increased senescence
To further understand the role of Rplp1 in cell proliferation and apoptosis, we generated Rplp1 F/F mice after crossing Rplp1+/F mice with Cre-ER T2 transgenic mice [29], and subsequently isolated the Rplp1 F/F CER + (Rplp1 iD ) and Rplp1 +/F CER + (control) pMEFs. Rplp1 was efficiently deleted in the Rplp1 iD pMEFs upon the addition of 1 mM 4-OHT for four days ( Figure 7A and 7B). First, we assessed proliferation in Rplp1 iD and control pMEFs that had been treated with 1 mM 4-OHT for four days. Rplp1 iD pMEFs showed a proliferation defect, with reduced cell numbers ( Figure 7C). Furthermore, BrdU labeling of Rplp1 iD and control pMEFs after four days of 1 mM 4-OHT revealed a considerable reduction of BrdU+ cells among the Rplp1 iD cells, compared with the controls (Figure 7D and 7E). However, when we analyzed the cell cycle profiles, we found no obvious differences in the cell cycle distribution between the control cells and those with a single Rplp1 allele deletion ( Figure 8A). Next, we investigated whether increased apoptosis contributed to the reduced numbers of Rplp1 iD pMEFs. Fluorescence-activated cell sorting (FACS) analysis revealed no obvious differences in the Annexin V+ cell populations of mutant and control cells that had been stained with Annexin V/ DAPI ( Figure 8B).
From the growth curve ( Figure 7C), we noticed a premature cessation of proliferation in Rplp1 iD pMEFs after passage 4. To follow the fates of these cells, we performed b-galactosidase (b-gal) staining and found that the frequency of b-gal + cells was higher among Rplp1 iD pMEFs than among the controls ( Figure 8C and 8D), indicating that Rplp1 deletion caused early senescence in the pMEFs. To investigate the cause of this senescence, we analyzed the levels of important cell cycle regulatory proteins and found higher levels of the CDK inhibitor p16 INK4A , a marker of senescence associated with the downregulation of pRb activity ( Figure 8E). However, the protein levels of p21 CIP1 and p27 KIP1 did not differ between the control and Rplp1 iD pMEFs ( Figure 8F). Because senescence can be caused by the accumulation of reactive oxygen species (ROS) [30,31] and the non-essential ribosomal proteins RPL1, RPL32 and RPL36 can prevent intracellular ROS generation [32], we measured the intracellular ROS levels. No differences were observed between the deleted and control cells ( Figure S2), suggesting that ROS was not the cause of senescence in Rplp1 iD pMEFs. Finally we expressed lentiviral vectors expressing p16 INK4A-shRNA and p53 shRNA in Rplp1 iD pMEFs which were further depleted for Rplp1 (see materials and methods). The inhibition of p16 INK4A in this context rescues, at least partially, the senescent phenotype while p53 inhibition does not seen to have any effect (data not shown). Moreover transient transfection strategy using siRNAs showed similar results ( Figure S3). Taken together, these experiments showed that Rplp1 deletion in pMEFs induced a p16 INK4A /pRb pathway-mediated proliferation defect and led to increased senescence.

Rplp1 deletion in MEFs does not affect global protein synthesis but alters the translation pattern
Rplp1 is important for proper ribosomal interactions with elongation factors, especially eEF-2 [33,34]. To assess whether Rplp1 deletion affected protein synthesis, Rplp1 iD and control immortalized MEFs were treated with 4-OHT for four days. Three days after the end of the treatment, the cells were labeled with a methionine analog (azide-homoalanine), and the de novo synthesized proteins were detected by western blotting ( Figure 9A) and quantified by densitometry ( Figure 9B). Proper Rplp1 deletion was assessed by western blotting ( Figure 9A). The experiment was performed with two pairs of control and Rplp1 iD immortalized MEFs. The protein synthesis level was unchanged when Rplp1 was deleted. Therefore, Rplp1 deletion in MEFs did not affect general protein synthesis.
Two-dimensional gel electrophoresis was performed to examine the protein translation pattern in Rplp1-deficient pMEFs. The proteins that were more strongly dysregulated in Rplp1 iD pMEFs were selected and identified by mass spectrometry. We identified fourteen proteins that were significantly dysregulated in Rplp1 iD pMEFs (p,0.01); these were associated with different cellular functions such as protein folding (70%), cell death (40%), gene expression (20%) and metabolism (20%) ( Figure 9C and Table S2). In agreement with the finding that protein folding was the most affected group, the upregulation of stress-related proteins such as the endoplasmic reticulum (ER) chaperones HSP74 and GRP78 was confirmed by western blot analysis ( Figure 8E).

Discussion
To study the function of P1 in vivo, we generated Rplp1 knockout mouse models. Rplp1 null mice could not be generated due to the infertility of Rplp1 Het mice. Surprisingly, Rplp1 et mice, which carry only one Rplp1 allele, presented with a severe phenotype that led to early postnatal death in approximately 50% of the offspring. Rplp1 Het mice exhibited rigid, kinky, short tails. This phenotype is consistent with the general notion that RP mutational haploinsufficiency is often associated with severe defects in mammals. For example, the phenotype of Rplp1 Het mice is reminiscent of that of Ts (Ts/+) mutant mice, which possess a spontaneous, dominant mutation in Rpl38. [1]. While homozygous Ts mutants die before implantation, the heterozygous mice present with a short, kinky tail and skeletal patterning defects [5]. Rpl38 regulates 80S-mRNA complex formation for specific Hox mRNAs [1]. Rpl24 heterozygosity in Bst/+ mice was also shown to cause kinked tails [9]. RPL24 regulates ribosomal subunit association and protein translation [2]. These observations suggest that defects in ribosomal subunit joining or assembly might induce skeletal defects. For other RPs, single-allele deletion leads to lethality during gastrulation, a period during which cell division and differentiation increase drastically, due to an error in ribosomal biogenesis (i. e.: RpS6 mutations) [2]. In contrast, heterozygous mutations in other RPs such as Rps19 [35], Rpl29 [36] and Rpl24 [9] are compatible with embryonic development. Moreover, some human patients who carry RPS19 mutations present with skeletal malformations [37].
Our results show that the specific knockout of Rplp1 in the CNS provoked perinatal lethality and morphological brain defects. The observed brain atrophy was caused by the impaired proliferation and increased apoptosis of neuroprogenitors. Rplp1 CNSD mice exhibited neocortical apoptosis, which also contributed to brain atrophy. Consistently, deficiencies in other RPs such as Rpl22 were shown to selectively arrest the development of a/b-T cells in a p53 activation and apoptosis-mediated manner [38]. RPL11 silencing was shown to induce p53-dependent apoptosis [39], and Rps19 deficiency impaired ribosomal biogenesis and activated p53 in zebrafish [40]. Upon the silencing of other RPs such as RPL23 [41], RPS9 [42] or RPS6 [43], p53 becomes activated and induces apoptosis. In contrast, Rplp1 deletion led to apoptosis independently of P53. This proliferation defect was accompanied by the downregulation of cyclin E, but appeared to be independent of p16 INK4A and p19 ARF . Surprisingly, the CDK inhibitors p21 CIP1 and p27 KIP1 were also downregulated. Although p21 CIP1 and p27 KIP1 exert direct cell cycle control at a nuclear level, these proteins can also exert other independent functions when cytoplasmically localized which could contribute to the observed abnormalities of the IZ and CP layers in the Rplp1 CNSD embryonic brain [44][45][46]. Moreover, the detrimental effect of a single-allele deficiency of Rplp1 on neurosphere formation might be associated with the downregulation of p21 CIP1 and p27 KIP1 , as they have previously been associated with cell differentiation and apoptosis [27,28,44].
In agreement with our in vivo observations, the proliferation defects were also observed in Rplp1 iD pMEFs. Interestingly, the proliferation defect in Rplp1 iD pMEFs induced p16 INK4A /pRb pathway-mediated premature senescence, which was also p53independent. Importantly p16 INK4A inhibition is able to bypass, at least partially, the senescence phenotype indicating that this pathway is crucial for Rplp1-deficiency-mediated senescence. This finding was consistent with finding that Rplp1 overexpression could bypass senescence in pMEFs [22]. A similar proliferative defect was also previously observed in P1-P2 knocked-down human cells [47]. In contrast to the Rplp1 CNSD embryonic brain, the p27 KIP1 , p21 CIP1 and cyclin E protein levels were unaltered in Rplp1 iD pMEFs, indicating that Rplp1 deletion has different consequences in specific tissues or cell types. This finding was also supported by the fact that, in strong contrast to proliferating cells, Rplp1 deletion in post-mitotic mature B cells and follicular dendritic cells in Rplp1 BD mice did not result in an obvious phenotype (Perucho et al., unpublished observations). Given that embryonic stem (ES) cells contain high levels of P1 ( [22] and data not shown), our results suggest that Rplp1 is crucial in embryonic stem cells and progenitor cells (e.g., neural system), but dispensable in certain differentiated cells such as mature B cells. Consistent with this assumption, Rplp1 expression was downregulated upon the differentiation of mouse ES cells to embryonic bodies [48]. It is also possible that in different cell types, the ability of the ribosomal subunits to compensate for losses in P1/P2 loss differs. For example, P0, P1 and P2 contain a common C-terminal domain that is responsible for the ribosomal interaction with EF-2 [33]. However, the C-terminal domain of P0 is not required for ribosomal activity when P1 and P2 are bound to the ribosome [49]. In other words, the C-terminal domain of P0 might be sufficient for ribosomal-translation factor interactions in some cell types or conditions, whereas in other cell types that require proliferation, P1 and P2 act as rate-limiting factors. For example, the proportions of ribosomal stalk proteins in the cytoplasm have been reported to vary from less than 1% [50] to 75% [51], possibly reflecting the different requirements among organisms and metabolic cell conditions [52].
On the other hand, Rplp1 deletion (and concurrently Rplp2 deletion) in the Rplp1 iD pMEFs did not affect global protein translation. This is in agreement with previous observations in human cells [47]. In contrast, the depletion of other RPs (i.e. Rps9) was shown to reduce the protein synthesis rate [42]. Interestingly, we identified 14 specific proteins that were dysregulated in Rplp1 iD pMEFs, most of which were ER chaperones. For example, most of the upregulated proteins in Rplp1 iD pMEFs are so in reaction with the ribosomal stress response (i.e.: GRP75, GRP78, SYAC and HSP74). In addition, some of the alterations observed in proliferation and apoptosis could be due to the reduced levels of Rab1b given its role in the control of cell growth [53][54][55]. The ER is home to an array of interlinked chaperone proteins upon which secreted proteins depend for correct folding, partner chain assimilation and final multimer assembly. The ER stress response constitutes a cellular process that is triggered by various conditions that disturb protein folding. Eukaryotic cells have developed an adaptive mechanism, the unfolded protein response (UPR), that acts to remove unfolded proteins and restore ER homeostasis [56]. In mammalian cells, imbalanced RP levels or the absence or malfunction of any RP has been shown to induce ribosomal stress, with severe consequences for cell survival [57]. Although this response has been linked to p53 [58], p53-independent mechanisms are also involved [59,60]. Interestingly, RP knockdown or Quantification of the amount of b-gal + cells for each analyzed clone analyzed. (E and F). Western blot analysis of protein extracts from Rplp1 iD and control pMEFs that were treated with 1 mM 4-OHT for 4 days or left untreated. doi:10.1371/journal.pone.0099956.g008 haploinsufficiency can provoke a ribosomal stress response that can be associated with senescence, UPR or other mechanisms [61,62]. The data presented herein link the ER-UPR with the senescence response in a single effector pathway downstream of ribosomal stress. Moreover, because protein folding is coupled with elongation, the absence of P1 and P2 might affect translation elongation by disrupting protein folding. EF-2 upregulation might reflect an attempt by the cell to compensate for the impaired stalk recognition, as P0, P1 and P2 comprise the EF-2 recognition motif [34]. For example, we cannot rule out the possibility that Rplp1 has additional extra-ribosomal functions, as P2 [63], at replication forks [64]. The relationship between the dysregulated proteins in Rplp1 iD pMEFs and the observed phenotype in the mutant mice remains to be fully characterized. Although the global protein level is unchanged, the absence of P1 might cause an increased translation error rate that would generate unfolded proteins.
Our results suggest that P1 has potential extraribosomal functions related to protein folding, a process that occurs outside the ribosome but within the ER compartment.
On the other hand, it has been hypothesized that Rplp1 might be involved in internal ribosomal entry site (IRES)-dependent translation [16]. Interestingly, both the p27 KIP1 and p53 mRNAs contain IRES [65]. This alternative hypothesis should be explored in future studies. Importantly, the P1 disturbance particularly affected proliferating cells (neuroprecursors), strongly suggesting the following: that P1 plays specific roles in the ribosome by ''finetuning'' translation accordingly to the cellular requirements and/ or that proliferating cells are equipped with an extra-sensitive ER stress response to preserve their stemness properties. Our work also suggests that Rplp1 impairment-mediated activation of the ER stress pathway translates into different response mechanisms (senescence in vitro and apoptosis in vivo) that are incompatible with cell survival and normal development. Overall, our work shows that Rplp1 is crucial for development and proliferation and proposes P1 as a novel factor of protein proper folding and translational ''fine-tuning''.  Table S2 Dysregulated proteins identified in Rplp1 iD pMEFs by two-dimensional gel electrophoresis and classification according to gene ontology (www.geneontology.org) in the different biological processes in which they are involved. Green indicates protein involvement in a biological process and red indicates noninvolvement. (XLSX)