Mouse-derived GC-1spg cells and GC-2spd cells (a gift from Dr. Fei Sun, Nantong University, China) were cultured in DMEM (Life Technologies, United States) supplemented with 10% (v/v) FBS (Life Technologies, United States).

GC-2spd cell proliferation was detected using CCK-8 reagent (Dojindo Laboratories, Kumamoto, Japan). The absorbance of the cells treated with CCK-8 was measured by a microplate reader (51119000, Thermo Scientific, United States) at 450 nm.

Mice (C57BL/6, male) were maintained in an animal room with 12-h light/12-h dark cycles and free access to food and water at the laboratory animal center of Xiamen University. All experimental procedures were performed according to the approved guidelines from the Animal Care and Use Committee of Xiamen University (201507), and all mice were handled following the “Guide for the Care and Use of Laboratory Animals” and the “Principles for the Utilization and Care of Vertebrate Animals.”

Spermatogonia were obtained from the testes of thirty 7-day-old mice. Preleptotene, leptotene/zygotene, and pachytene spermatocytes were obtained from the testes of five 17-day-old mice. The germ cell mixture was isolated by a two-step enzymatic digestion protocol as described previously (Ogawa et al., 1997). In brief, the testes were decapsulated and then incubated with approximately 10 volumes of HBSS solution containing 1 mg/ml collagenase IV (17104–019, Gibco, United States) and 1 mg/ml DNase I (DN25, Sigma, United States) at 37°C with gentle shaking. After washing, the testes were incubated in HBSS containing 1 mM EDTA, 0.25% trypsin, and 500 mg/ml DNase I at 37°C. When most cells were dispersed, the action of trypsin was terminated by adding a 10% volume of fetal bovine serum (FBS). Then, the cell suspension was filtered through nylon mesh with a 40-μm pore size to remove large clumps of cells. The filtrate was centrifuged at 1,000 g for 5 min. After removing the supernatant, cells in the pellet were resuspended and cultured in F12/DMEM with 5% fetal bovine serum for 4–6 h for Sertoli cell adherence. Then, the supernatant cells were collected and centrifuged at 1,000 g for 5 min to obtain the germ cell mixture sediment. Next, the purified germ cells were isolated using gravity sedimentation with an STA-PUT device based on cytological classification and morphological analysis, as described previously (Bellvé, 1993). The purity of the four types of germ cells was assessed by observing cell diameters and morphology under a microscope bright field and the levels of marker genes and proteins.

Total RNA was isolated from separated spermatogonia, preleptotene spermatocytes, leptotene/zygotene spermatocytes, and pachytene spermatocytes using an RNeasy Mini Kit (74104, Qiagen, Germany) following the manufacturer’s protocol. DNase I in column digestion was performed to remove genomic DNA. Ribosomal RNA was eliminated by an Epicentre Ribo-zeroTM rRNA Removal Kit (Epicentre, USA) according to the manufacturer’s instructions. Then, sequencing libraries were generated using the NEBNext® UltraTM Directional RNA Library Prep Kit for Illumina® (NEB, USA) following the manufacturer’s instructions. RNA-seq was performed with Illumina’s HiSeq 2,500 system at RiboBio Co., Ltd. (Guangzhou, China). Sequences of samples were aligned to the UCSC musculus reference genome (mm10) using TopHat with default settings (Kim et al., 2013). RefSeq-annotated gene expression was calculated using Cuffdiff software (Trapnell et al., 2010). Genes with FPKM below 0.5 in all samples were filtered, and only genes with a fold change (FC) larger than 1.5 were counted as up- or downregulated genes. RNA-seq was deposited in the Gene Expression Omnibus database under accession GSE183678.

cDNA was produced with the PrimeScript^™ 1st Strand cDNA Synthesis Kit (6110A, TAKARA, Japan) following the manufacturer’s instructions. RT–qPCR was performed in an ABI PCR system (7500, Applied Biosystems, United States). mRNA expression levels were analyzed by the comparative threshold cycle method. We selected Nudcd3 as the normalizing gene because it exhibited similar expression levels in the four cell populations in our RNA-seq data. Primer information is presented in Supplementary Table S1.

Nuclear spreading of spermatogenic cells was performed as described previously (Scherthan et al., 2000). In brief, the spermatogenic cell suspension was placed on a glass slide and mixed with 1% Triton X-100 (9002–93–1, Solarbio Life Sciences, China). Spermatocyte swelling and spreading were monitored by phase-contrast microscopy. When cells obtained an opaque appearance, 3.7% formaldehyde and 0.1 M sucrose were added to the slide and gently mixed by tilting. Slides were then air-dried at 37°C and stored at −20°C until use.

Cellular fractionation was performed as described previously (Wang et al., 2021). Briefly, cells were washed with ice-cold PBS, collected, spun down and resuspended in ice-cold buffer I (10 mM Hepes, pH 8.0, 1.5 mM MgCl2, 10 mM KCl, 1 mM DTT) supplemented with protease inhibitor cocktail, followed by incubation for 15 min on ice to allow cells to swell. Igepal-CA630 was then added at a final concentration of 1% (10% stock solution) followed by vortexing for 10 s. Nuclei were collected by centrifugation for 2∼3 min at maximum speed (∼21,100 × g). The resulting supernatant was the cytosolic fraction. Nuclei were then lysed in ice cold buffer II (20 mM Hepes, pH 8.0, 1.5 mM MgCl2, 25% (v/v) glycerol, 420 mM NaCl, 0.2 mM EDTA, 1 mM DTT) supplemented with protease inhibitor cocktail followed by vigorous rotation at 4°C for 30 min and centrifugation for 15 min at maximum speed. The resulting supernatant was the nuclear fraction. Both cytosolic and nuclear RNAs were extracted by a phenol-chloroform-isoamyl alcohol mixture (77618, Sigma, United States).

In brief, the cell is cleaved and the total protein is extracted. Appropriate antibodies and beads were added to the protein samples and incubated overnight at 4°C. Centrifuge 14000 g for 5s the next day, and rinse with cold PBS (800 μl each) for 3 times. The supernatant was collected for 8–12% SDS-Page and the protein bands were detected with specific antibodies.

Cell apoptosis was determined using an Annexin V‐FITC Apoptosis Detection Kit (Beyotime, China). Cells were trypsinized and then terminated by adding an equal volume of medium, followed by centrifugation at 4°C for 2 min (200 × g). Next, cells were washed twice with ice-cold PBS. Then, the cells were incubated in darkness for 10 min in 100 μl ice-cold Annexin V binding buffer, and then 5 μl propidium iodide was added for another 5 min. Eventually, cell apoptosis was analyzed with flow cytometry.

Prepared sections were blocked with 3% BSA for 1 h at room temperature and then incubated with primary antibodies overnight at 4°C. After washing with PBST, the samples were incubated with the secondary antibodies at a 1:200 dilution for 1 h at 37°C. The slides were subsequently mounted with Vectashield containing DAPI (H-1200, Vector Laboratories, United States). The primary antibodies included Scp1 (ab15090, Abcam, United States), Scp3 (ab97672, Abcam, United States), cleaved caspase-3 (#9661, CST, USA), and Ptbp2 (ab154787, Abcam, United States). The secondary antibodies were Alexa Fluor® 594-conjugated goat anti-rabbit IgG (ZF-0516, ZSGB-BIO, China) and Alexa Fluor® 488-conjugated goat anti-mouse IgG (ZF-0512, ZSGB-BIO, China).

Western blotting was performed as previously described (Hu et al., 2015). In brief, proteins were extracted and separated by 10% (w/v) SDS-PAGE, then transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore, United States). The membrane was blocked in 5% skimmed milk, and then incubated with primary antibody overnight at 4°C. After washing with TBST (10 mM Tris. HCl, pH 7.6, 150 mM NaCl, Tween 20 0.1%), the membranes were incubated with Goat anti-Rabbit IgG Secondary Antibody, HRP (31460, ThermoFisher, United States) or Goat anti-Mouse IgG Secondary Antibody, HRP (31430, ThermoFisher, United States) for 1 h and washed twice with TBST. The membranes were developed using the ECL (Enhanced chemiluminescence) kit (K-12045-D50, Advansta, United States), visualized, and recorded with the use of an Imagequant LAS 4000 mini machine (GE Healthcare Life Sciences, United States). Primary antibodies: Plzf (sc-28319, Santa Cruz Biotechnology, United States), Stra8 (ab49405, Abcam, United States), Scp3 (ab97672, Abcam, United States), Ptbp2 (ab154787, Abcam, United States), and ubiquitin (#3936, CST, United States).

An siRNA for Rbakdn and the negative control were purchased from Guangzhou RiboBio Co., Ltd. Target sequence information of the siRNAs: #1, GGT​GGA​CAC​CAT​TAT​CCT​A; #2, CGT​ACT​CGA​AGA​AAT​CCA​A; #3, GCT​CCA​CAC​TCC​TCC​CAA​A. Lipofectamine 2000 (11668019, Invitrogen, USA) was used to transfect GC-2 cells according to the manufacturer’s protocol.

Regarding the Rbakdn-knockdown plasmid, the siRNA for Rbakdn targeted the sequence 5′-GGT​GGA​CAC​CAT​TAT​CCT​A-3′, and the control scrambled sequence was 5′-CCT​AAG​GTT​AAG​TCG​CCC​TCG-3′. The annealed double-stranded fragment (shRNA) was cloned into the lentiviral vector pSicoR. Plasmids were transfected into 293T cells to produce lentivirus using a Turbofect transfection reagent as previously described (Tao et al., 2016). To knock down Rbakdn, GC-2spd cells were infected with lentivirus expressing shRNA targeting Rbakdn.

RNA-FISH was conducted as previously described. GC-2spd cells seeded on cover glass were transfected with siRNAs for 48 h before fixation with fixation buffer (4% formaldehyde, 10% acetic acid, 1 × PBS) for 10 min. Cells were then permeabilized in 70% ethanol overnight and rehydrated in 2 × SSC buffer with 50% formamide. Hybridization was carried out in the presence of 30 ng of probes at 37°C overnight. Biotin-labeled probes were incubated with Streptavidin-Cy3™ (S6402, Sigma Aldrich, USA) in 2 × SSC buffer with 8% formamide, 2 mM vanadyl-ribonucleoside complex, and 0.2% RNase-free BSA at 37°C for 1 h in the dark. Nuclei were counterstained with DAPI and then washed twice with 2 × SSC with 8% formamide at room temperature for 15 min. Images of each cover glass were taken with a Carl Zeiss laser confocal microscope. The probe sequences for Rbakdn were as follows: 5′-GTG​GAG​AGA​AGA​CAA​CCC​GT-3′ and 5′-CGG​AGG​AGA​CCA​GGG​ATT​TT-3′.

Testes of 5-days postpartum mice were harvested and digested by a two-step enzyme digestion method as described previously (Liu et al., 2020). Briefly, testes were dispersed with 1 mg/ml collagenase type IV at 37°C for 10 min, followed by digestion in 0.25% trypsin/1 mM EDTA at 37°C for 10 min. Spermatogenesis was induced in vitro as described previously (Zhou et al., 2016). A single-cell suspension was obtained after filtration through a 40-mm cell strainer, and cells were collected by centrifugation. Then, the testicular cells were spread into laminin-coated dishes. From Day 0 to Day 6, cells were cultured in MEM α (Thermo Scientific, USA) supplemented with 10% KnockOut™ Serum Replacement (KSR, Thermo Scientific, USA), BMP-2/4/7, retinoic acid, and activin A. From Days 7–12, cells were cultured in aMEM containing 10% KSR, testosterone, FSH, and bovine pituitary extract (BPE, Millipore, USA). The medium was changed every 2 days. Cells were cultured in 5% CO₂ at 37°C.

The lentiviral vectors were packaged and injected into the testes of 4-week-old male mice as described previously (Zhao et al., 2013). In brief, under a microscope, the testes of pentobarbital sodium-anesthetized mice were pulled out, and 5 μl of fresh high-titer lentivirus containing 10 μg/ml polybrene (TR-1003, Sigma–Aldrich, USA) was injected into the seminiferous tubules through the efferent duct using a sharp glass capillary with a tip diameter of 50 μm. The testes were then returned to the abdominal cavity, and the abdominal wall and skin were closed with sutures.

RNA-IP was performed as previously described with minor modifications (Peritz et al., 2006). Briefly, cells grown in a 10-cm dish were lysed in polysome lysis buffer (100 mM KCl, 5 mM MgCl₂, 10 mM HEPES, pH 7.0, 0.5% NP-40, 1 mM DTT, 100 U/ml RNasin RNase inhibitor (Promega, N2511), 2 mM vanadyl ribonucleoside complex solution (Sigma, 94742), and 25 μl/ml protease inhibitor cocktail for mammalian tissues (Sigma, P8340)) and then subjected to IP followed by washing with polysome lysis buffer four times and then polysome lysis buffer plus 1 M urea four times. RNA was released by adding 150 μl of polysome lysis buffer with 0.1% SDS and 45 μg proteinase K (Ambion, AM2548) and incubated at 50°C for 30 min. RNA extracted with a phenol-chloroform–isoamyl alcohol mixture (Sigma, 77618) was recovered by adding 2 μl GlycoBlue (15 mg/ml, Ambion, AM9516), 36 μl 3 M sodium acetate and 750 μl ethanol followed by incubation at −20°C overnight. Precipitated RNAs were washed with 70% ethanol, air dried, and resuspended in RNase-free water followed by DNase I (Promega, M6101) treatment to remove genomic DNA. The resulting RNAs were subjected to RT–qPCR analysis.

Statistical analysis was performed using GraphPad Prism version 6.0 (GraphPad Software Inc., La Jolla, CA, USA). Two groups or data points were compared by a two-tailed t-test. Multiple comparisons were analyzed by two-way analysis of variance (ANOVA). p < 0.05 was considered indicative of statistical significance.

To detect the role of long noncoding RNAs in meiosis, we used the STA-PUT method and separated spermatogenic cells in different stages of meiosis, including spermatogonia (SG), preleptotene spermatocytes (PlpSCs), leptotene/zygotene spermatocytes (L/ZSCs), and pachytene spermatocytes (PaSCs). The separated spermatogenic cells were identified and tested for purity according to their morphology and the expression levels of marker genes ( Supplementary Figures S1A,B ). The results showed that the purity of the spermatogenic cells of each stage that we isolated was above 80% ( Supplementary Figure S1C).

After high-throughput RNA sequencing, we obtained a large number of lncRNAs that are dynamically expressed during meiosis. When we sorted the stage-specific lncRNAs according to expression levels, we found that the spermatogenic cells of each meiosis phase showed strong regularity in the expression of lncRNAs. Each cell type contained dynamically and highly expressed lncRNAs ( Supplementary Table S2 ). For example, Malat1 and H19 are dynamically and highly expressed in SG; Gm15441 and 5730408K05Rik are dynamically and highly expressed in PlpSCs; 1700020N01Rik and 1700100I10Rik are dynamically and highly expressed in L/ZSCs; and 1700028J19Rik and Gm2762 are dynamically and highly expressed in PaSCs ( Supplementary Table S2 ).

At the same time, we observed gradual increases in a large number of lncRNAs with meiosis progression, and Rbakdn was the most significantly upregulated lncRNA during meiosis ( Figure 1A ). We predicted its protein-coding ability by CPC (coding potential calculator; http://cpc.cbi.pku.edu.cn/) and found that it has a predicted score of only −0.896121, suggesting almost no potential to encode protein. Rbakdn may play an important regulatory role in the progression of meiosis.

We next verified the RNA-seq results and confirmed that Rbakdn began to be expressed in PlpSCs, and its expression level gradually increased during meiosis, as shown by RT–PCR ( Figure 1B). Furthermore, the mouse ENCODE transcriptome data indicate that Rbakdn is dynamically and highly expressed in mouse testes but rarely expressed in other organs (https://www.ncbi.nlm.nih.gov/gene/100042605). We detected Rbakdn expression in brain, liver, heart, lung, kidney, and testis tissues isolated from adult mice and confirmed that Rbakdn was highly and specifically expressed in the testes ( Figure 1C ). Rbakdn is therefore an lnRNA with expression restricted to the testes. Furthermore, we isolated testis tissues from 5- to 60-day-old mice and detected Rbakdn expression. The results showed that Rbakdn expression began on the 10th day postpartum and gradually increased during meiosis ( Figure 1D ).

To study the effect of Rbakdn on meiosis, we established an in vitro induction system to simulate meiosis in mouse spermatocytes ( Supplementary Figure S2A ). We collected cultured germ cells at various time points in the induction process and then detected the expression of molecular markers at each stage of spermatogenic cells. The expression of Plzf, a marker of spermatogonia, gradually decreased, indicating that spermatogonia gradually underwent the process of meiosis and lost their stem cell properties under the action of retinoic acid. The expression of Stra8, a marker of meiosis initiation, gradually increased under retinoic acid induction and reached its highest level at D6, indicating that spermatogonia initiated meiosis induced by Type I culture medium. Then, the cell culture medium was changed to Type II culture medium, Stra8 expression gradually decreased, and Scp3, a marker of the meiotic synaptonemal complex, began to be expressed. Scp3 expression gradually increased as induction progressed, indicating that the germ cells entered the meiotic process ( Supplementary Figure S2B )

We added siRNA specifically targeting Rbakdn on the third day of induction to test the role of Rbakdn in meiosis. In the control group, substratum Sertoli cells proliferated and formed a feeder layer. The supernatant germ cells proliferated to form clones and grew with culture ( Figure 2A, upper panel). In the group with Rbakdn knockdown, Sertoli cells proliferated and formed a feeder layer, suggesting that Rbakdn inhibition does not affect the proliferation of Sertoli cells. However, the size of the colonies formed by germ cell proliferation and differentiation was obviously reduced compared with that of the control after inhibiting Rbakdn ( Figure 2A, lower panel).

To further explore the effect of Rbakdn on the meiosis process, cultured germ cells were costained with SCP1 and SCP3. In the control group, during the induction process, spermatocytes of each stage of meiosis appeared sequentially. Specifically, almost no positive SCP1 and SCP3 signals were observed in the germ cells at Day 3 (D3), indicating that the germ cells were still in the spermatogonium stage. On the 6th day (D6), the germ cells showed a dotted positive SCP3 signal, indicating that the germ cells had started meiosis and entered the preleptotene stage. Then, linear SCP3 signals were observed in the germ cells at the 8th day (D8), indicating that the germ cells had entered the leptotene stage; subsequently, linear SCP3 in the germ cells gradually thickened, and punctate SCP1 also appeared at Day 10 (D10), suggesting that the germ cells had reached the zygotene stage. Finally, linear SCP1 and SCP3 were coexpressed in the germ cells at Day 12 (D12), showing that the germ cells had reached the pachytene stage ( Figure 2B, upper panel). After inhibiting Rbakdn, the germ cells showed a dotted positive SCP3 signal at D6, indicating that the germ cells had started meiosis and entered the preleptotene stage; however, the germ cells still showed a dotted positive SCP3 signal from D6 to D12, indicating that even though the germ cells had started meiosis and entered the preleptotene stage, they could not continue the further meiosis process ( Figure 2B, lower panel). These in vitro results indicated that Rbakdn knockdown inhibits the process of meiosis.

To further evaluate and verify the effect of Rbakdn on the development of spermatocytes, we used germ cell lines to study the effect of knocking down Rbakdn on the proliferation and apoptosis of spermatocytes.

Measurement of mRNA levels revealed that Rbakdn was expressed at low levels in GC-1spg cells (corresponding to a stage between type B spermatogonia and primary spermatocytes) but was highly expressed in GC-2spd cells (mouse spermatocyte cell line) ( Figure 3A ), verifying that Rbakdn is specific to testicular meiotic cells. Then, we performed nuclear-cytoplasmic separation and detected Rbakdn expression in the cytoplasm and nucleus. After separation of the cytoplasm and nucleus, the qRT–PCR results showed that the cytoplasmic marker protein β-actin was mainly expressed in the cytoplasm, and the nuclear marker protein U6 was mainly expressed in the nucleus. These control results demonstrated that the cytoplasm and nucleus were successfully separated. The results showed that approximately 60% of Rbakdn was expressed in the cytoplasm, and the rest was expressed in the nucleus. These results suggested that Rbakdn is mainly expressed in the cytoplasm ( Figure 3B ).

We designed three siRNAs targeting Rbakdn to knock down its expression in GC-2spd cells. The results showed that all three siRNAs effectively knocked down Rbakdn expression after transfection. Among them, Rbakdn-siRNA1 had the best effect ( Figure 3C ). Therefore, we used Rbakdn-siRNA1 (Rbakdn-siRNA) to conduct follow-up experiments. To further verify the distribution of Rbakdn in GC-2spd cells, we designed fluorescent probes of Rbakdn and performed FISH detection. The results showed that Rbakdn was expressed in both the cytoplasm and nucleus but mainly in the cytoplasm. After transfection with Rbakdn-siRNA, Rbakdn expression in GC-2spd cells was reduced ( Figure 3D ).

Then, we detected the effect of Rbakdn-siRNA on the growth of GC-2spd cells by CCK-8 and found that compared with the that in the blank control or siRNA negative control group, the number of GC-2spd cells with Rbakdn siRNA transfection began to decrease on the third day and significantly decreased on the fourth and fifth days ( Figure 3E ). We further used flow cytometry to detect the apoptosis of GC-2spd cells after Rbakdn knockdown and found that the apoptotic level of GC-2spd cells was significantly increased after transfection with Rbakdn-siRNA ( Figure 3F ). These results indicated that knockdown of Rbakdn in GC-2spd cells promoted the apoptosis of GC-2spd cells, supporting the observation in primary germ cells.

To further confirm the function of Rbakdn in the process of meiosis, we constructed a lentiviral vector expressing an shRNA targeting Rbakdn (Rbakdn-shRNA), which could effectively reduce Rbakdn expression ( Supplementary Figure S3 ). We then injected the lentivirus expressing Rbakdn-shRNA into the testes of one-month-old wild-type mice. Each mouse was injected with a control lentivirus in the left testis and the Rbakdn-shRNA lentivirus at an equal titer in the right testis. The testes injected with lentivirus were collected at D3, D7, D10, D14, and D28 after injection to investigate the effect of Rbakdn knockdown on spermatogenesis. Compared with that in control subjects, the volume of the testes injected with Rbakdn-shRNA showed no difference at D3 and D7. However, the volume of the testes injected with Rbakdn-shRNA decreased from D10 and decreased more obviously at D14 and D28 ( Supplementary Figure S4A ). The qRT–PCR results showed that Rbakdn-shRNA could effectively reduce Rbakdn expression in the testes ( Supplementary Figure S4B ).

HE staining results showed that the control subjects injected with the control virus displayed normal spermatogenesis, and no vacuoles were observed in the seminiferous tubules. In contrast, after injection with Rbakdn-shRNA, as indicated by red asterisks, vacuoles began to appear in the seminiferous tubules at D7, which were more obvious and serious at D14 and D28 ( Figure 4A ). Periodic acid-Schiff staining results further confirmed this defect in the testes injected with Rbakdn-shRNA ( Supplementary Figure S4C ). Immunofluorescence staining of Scp3 showed normal spermatogenesis in testes injected with the control virus. However, the number of Scp3-positive spermatocytes in testes injected with Rbakdn-shRNA was obviously reduced at D14 and D28 ( Supplementary Figure S4D ). We then dissected the mouse testes at D3, D7, D10, D14, and D28 after injection and detected the expression of the apoptosis-related protein cleaved caspase-3 by immunofluorescence. The results showed only a basic level of positive staining signals in testes from mice injected with the control virus, but spermatocyte apoptosis in the seminiferous tubules appeared at D7 and increased obviously on D10 after injection. On the 28th day, as indicated by white asterisks, many spermatocytes in the seminiferous tubules were apoptotic, and vacuolization was more severe in testes injected with Rbakdn-shRNA ( Figure 4B ). These observations suggested that Rbakdn knockdown disturbed the process of meiosis and induced spermatocyte apoptosis in vivo.

To explore the mechanism of Rbakdn, we performed predictive analysis of the interacting microRNAs and RNA-binding proteins using related lncRNA databases (http://starbase.sysu.edu.cn/rbpClipRNA.php?source=lncRNA&flag=none&clade=mammal&genome=mouse&assembly=mm10&RBP=all&clipNum=1&target=rbakdn) and found its potential RNA-binding protein, Ptbp2, which is an essential alternative splicing factor for spermatogenesis (Zagore et al., 2015; Hannigan et al., 2017). Then, we detected Ptbp2 expression in testes injected with Rbakdn-shRNA using immunofluorescence staining and found that Ptbp2 was highly expressed in spermatocytes in control subjects. In addition to the decrease in the number of spermatocytes, we observed a significant decrease in PTBP2 expression in the remaining spermatocytes after knockdown of Rbakdn ( Figure 5A ). We used the Ptbp2 antibody to pull down the RNA that binds to Ptbp2 and confirmed that Rbakdn could bind to Ptbp2 in GC-2spd cells ( Figure 5B). We found that Rbakdn knockdown did not affect the mRNA expression level of Ptbp2 in GC-2spd cells with transient (Rbakdn-siRNA) or stable (Rbakdn-shRNA) knockdown of Rbakdn ( Figures 5C,D). We assessed the protein level of Ptbp2 by Western blotting and observed that the protein expression of Ptbp2 was reduced in GC-2spd cells after Rbakdn knockdown ( Figure 5E ). These results both in vivo and in vitro indicated that knockdown of Rbakdn reduced the protein expression of Ptbp2.

We speculated that knockdown of Rbakdn might affect the stability of the Ptbp2 protein in spermatocytes, thus inducing degradation of the Ptbp2 protein. To confirm this hypothesis, we assessed Ptbp2 protein degradation in Rbakdn-knockdown GC-2spd cells treated with the proteasome inhibitor MG-132. After Rbakdn knockdown, Ptbp2 protein expression was reduced compared with that in the control group. Accordingly, Ptbp2 protein expression in GC-2spd cells increased after MG-132 treatment, and the ubiquitination level of Ptbp2 protein also increased ( Figure 5F). These results showed that Rbakdn could maintain the stability of Ptbp2 by inhibiting its ubiquitination-mediated degradation.