A p53/lnc‐Ip53 Negative Feedback Loop Regulates Tumor Growth and Chemoresistance

Abstract Acetylation is a critical mechanism to modulate tumor‐suppressive activity of p53, but the causative roles of long non‐coding RNAs (lncRNAs) in p53 acetylation and their biological significance remain unexplored. Here, lncRNA LOC100294145 is discovered to be transactivated by p53 and is thus designated as lnc‐Ip53 for lncRNA induced by p53. Furthermore, lnc‐Ip53 impedes p53 acetylation by interacting with histone deacetylase 1 (HDAC1) and E1A binding protein p300 (p300) to prevent HDAC1 degradation and attenuate p300 activity, resulting in abrogation of p53 activity and subsequent cell proliferation and apoptosis resistance. Mouse xenograft models reveal that lnc‐Ip53 promotes tumor growth and chemoresistance in vivo, which is attenuated by an HDAC inhibitor. Silencing lnc‐Ip53 inhibits the growth of xenografts with wild‐type p53, but not those expressing acetylation‐resistant p53. Consistently, lnc‐Ip53 is upregulated in multiple cancer types, including hepatocellular carcinoma (HCC). High levels of lnc‐Ip53 is associated with low levels of acetylated p53 in human HCC and mouse xenografts, and is also correlated with poor survival of HCC patients. These findings identify a novel p53/lnc‐Ip53 negative feedback loop in cells and indicate that abnormal upregulation of lnc‐Ip53 represents an important mechanism to inhibit p53 acetylation/activity and thereby promote tumor growth and chemoresistance, which may be exploited for anticancer therapy.


Introduction
The tumor suppressor p53 is a central regulator in cell cycle and apoptosis, and functional loss of p53 remains the most common event in cancers, which contributes to tumor growth and chemoresistance. [1,2] In response to various cellular stresses, including DNA damage, oxidative stress, or oncogene activation, p53 is upregulated/ activated, then acts as a transcription factor to transactivate downstream target genes, like cyclin-dependent kinase inhibitor 1A (CDKN1A), p53-upregulated modulator of apoptosis (PUMA) and BCL2 associated X (BAX), thereby triggers cell cycle arrest and/or apoptosis. Because of its importance in physiological and pathological processes, p53 activity is tightly regulated by multiple mechanisms, especially post-translational modifications. [2] Acetylation is an indispensable modification mechanism to activate p53 during stress responses. The acetylation state of a protein is controlled by histone acetyltransferases (HAT) and histone deacetylases (HDAC), enzymes that catalyze the addition and removal of an acetyl group from a lysine residue, respectively. Specially, p53 is acetylated by E1A binding protein p300 (p300)/CREB binding protein (CBP) at lysines 164, 370, 372, 373, 381, 382, and 386, by MOF/Tat interactive protein 60 kDa (TIP60) at lysine 120, and by monocytic leukemia zinc finger (MOZ) at lysines 120 and 382, and it is deacetylated by HDAC1 and sirtuin 1 (SIRT1). [2] Acetylated p53 displays decreased ubiquitylation, enhanced DNA binding activity, and thus increased transcriptional activity. [3,4] Loss of acetylation at all eight lysines completely abolishes p53-mediated cell cycle arrest and apoptosis. [5] The p53 KQ/KQ knock-in mice that express acetylation-mimicking form of p53 show neonatal lethality with substantial p53 activation and induction of different p53 target genes, [6] further confirming the importance of acetylation for p53 activation in vivo.
Here, we identify a lncRNA that is induced by p53 (named lnc-Ip53) and then inhibits p53 acetylation by upregulating HDAC1 expression and attenuating p300 activity, which forms a negative feedback loop. Both in vitro and in vivo studies show that aberrant expression of lnc-Ip53, which is frequently observed in hepatocellular carcinoma (HCC) and other cancer types, promotes tumor growth and chemoresistance by inhibiting p53 acetylation/ activity.
We characterized lnc-Ip53 as a 3094-nt polyadenylated RNA without protein-coding capacity ( Figure S1G-I, Supporting Information). Both chromatin immunoprecipitation (ChIP)sequencing data from ENCODE ( Figure S1H, Supporting Information) and our luciferase reporter assay ( Figure S1J, Supporting Information) revealed that the 1500-bp region upstream of lnc-Ip53 exhibited promoter activity, which was reduced by sip53 and enhanced by overexpressing wild-type but not mutant p53 (Figure 1D; Figure S1K, Supporting Information). A p53-responsive element (p53RE) was predicted within this region and its deletion (p-∆p53RE) diminished the lnc-Ip53 promoter activity (Figure 1E; Figure S1L, Supporting Information). Consistently, in-sertion of the wild-type but not mutant p53RE upstream of the promoter of "firefly" luciferase gene increased luciferase activity, which was further enhanced by p53 overexpression ( Figure 1F; Figure S1M, Supporting Information). Electrophoretic mobility shift assay (EMSA) and antibody-supershift assay revealed an in vitro interaction between p53 and p53RE ( Figure 1G,H). ChIP further confirmed their in vivo interaction ( Figure 1I). These results suggest that p53 transactivates lnc-Ip53 transcription by directly binding to the p53RE in the lnc-Ip53 promoter.
Acetylation of p53, which enhances DNA binding activity, is essential for its activity to transactivate downstream target genes, such as CDKN1A and PUMA. [3][4][5] ChIP revealed that the amount of CDKN1A or PUMA promoters that were precipitated by antibodies against acetylated or total p53 significantly increased in lnc-Ip53-silencing cells ( Figure 2F) but reduced in lnc-Ip53overexpressing cells ( Figure 2G). Moreover, silnc-Ip53 significantly increased the luciferase activity of PG13-Luc reporter that carried wild-type p53 binding sites, but had no effect on the activity of MG15-Luc that contained mutant p53 binding sites ( Figure 2H; Figure S2H, Supporting Information). Consistently, silnc-Ip53 enhanced the mRNA levels of CDKN1A and PUMA in p53-wild-type cells, but not in p53-null cells ( Figure 2I; Figure  S2I,J, Supporting Information), whereas lnc-Ip53 overexpression attenuated the Dox-, etoposide-or H 2 O 2 -induced upregulation of CDKN1A and PUMA (Figure 2J-L; Figure S2K, Supporting Information). Altogether, lnc-Ip53 may impede p53 acetylation and then inhibit its activity.
We then investigated whether lnc-Ip53 influenced the p53regulated cell activities, such as cell cycle and apoptosis. As shown, Dox-induced G2/M arrest was promoted by silnc-Ip53 ( Figure 3A; Figure S3A, Supporting Information) but was diminished by lnc-Ip53 overexpression ( Figure 3B; Figure S3B, Supporting Information). In addition, silnc-Ip53 promoted both basal and Dox-induced apoptosis ( Figure 3C; Figure S3C,D, Supporting Information), whereas lnc-Ip53 overexpression attenuated Dox-induced apoptosis ( Figure 3D). Consistently, the growth and colony formation of tumor cells were significantly suppressed by silnc-Ip53 ( Figure 3E; Figure S3E, Supporting Information) but were promoted by lnc-Ip53 overexpression (Figure 3F; Figure S3F, Supporting Information). Furthermore, sip53 antagonized the stimulatory effect of silnc-Ip53 on the expression Figure 1. Lnc-Ip53 is a transcriptional target of p53. A) Various p53 activators induced the expression of lnc-Ip53 along with CDKN1A and p53. Cells were treated for 12 h with 0.5 µM doxorubicin (Dox), 50 µM etoposide (Eto), 10 µM nutlin-3a (Nut), or vehicle control (Ctrl: PBS as control for Dox, DMSO as control for Eto and Nut). B) Overexpression of wild-type p53 (p53wt) but not R175H mutant p53 (p53mut) increased the lnc-Ip53 level. Cells stably expressing p53wt, p53mut, or control vector (Vec) were used. C) Silencing p53 (sip53) decreased the lnc-Ip53 level. Cells were transfected with the indicated RNA duplexes for 32 h, then incubated with PBS (Ctrl) or 0.5 µM Dox for 12 h. For (A-C), the mRNA level of CDKN1A was used as a positive control. D) sip53 reduced the activity of the lnc-Ip53 promoter (left), whereas overexpressing p53wt but not p53mut increased its activity (right). E) Deletion of p53RE reduced the activity of the lnc-Ip53 promoter. F) p53 enhanced the activity of pGL3-promoter reporter containing wild-type but not mutant p53RE. G,H) EMSA and antibody-supershift assay verified the in vitro interaction of p53 with p53RE in the lnc-Ip53 promoter. The biotin-labeled DNA-protein complexes are indicated by arrow. Cold N.S., nonspecific scrambled oligonucleotide. Nuclear extracts were from SK-HEP-1 cells. I) p53 interacted with the lnc-Ip53 promoter in vivo. Cells were incubated with 0.5 µM Dox for 6 h before ChIP. The antibody-precipitated DNAs were amplified by real-time quantitative PCR (qPCR, left) and semi-quantitative PCR for 35 cycles (right). The promoters of CDKN1A and GAPDH were used as positive and negative controls, respectively. + or −, cells with (+) or without (−) the indicated treatment. Data are shown as mean ± SEM of at least three independent experiments; p-values were determined by unpaired Student′s t-test; NS, not significant.

Lnc-Ip53 Promotes Tumor Growth and Chemoresistance by Impeding p53 Acetylation In Vivo
The in vivo biological significance of lnc-Ip53 was then explored using human tumor tissues and mouse xenograft models. Analysis of our HCC study cohort and the transcriptome data from The Cancer Genome Atlas (TCGA) revealed that lnc-Ip53 was upregulated in different malignancies, including HCC ( Figure 4A; Figure S4A, Supporting Information). The Kaplan-Meier survival analysis revealed an association between high lnc-Ip53 level and short recurrence-free survival (RFS) of HCC and this association was more pronounced among patients who carried wild-type p53 ( Figure 4B). Furthermore, the levels of acetylated p53 significantly decreased in HCC tissues ( Figure 4C; Figure S4B, Supporting Information) and were inversely correlated with lnc-Ip53 levels ( Figure 4D), suggesting that upregulation of lnc-Ip53 may contribute to reduction of acetylated p53 in cancers.
We further examined the in vivo effect of lnc-Ip53 using mouse xenograft models. As shown, silencing lnc-Ip53 significantly suppressed xenograft growth ( Figure 4E; Figure S4C, Supporting Information) and increased the levels of acetylated p53 and CDKN1A/PUMA in xenografts ( Figure 4F,G). On the other hand, enhanced lnc-Ip53 expression allowed xenografts to grow faster and be more resistant to Dox treatment ( Figure 4H Figure S4D, Supporting Information).
Collectively, upregulation of lnc-Ip53 may confer tumor with growth advantage and chemoresistance by inhibiting p53 acetylation.
Taken together, we disclose a novel p53/lnc-Ip53 negative feedback loop, in which p53 is activated by various stresses, then binds to the lnc-Ip53 promoter and transactivates lnc-Ip53 expression, whereas lnc-Ip53 directly interacts with HDAC1 to prevent its degradation and associates with p300 to attenuate its activity, leading to abrogation of p53 acetylation/activity, which consequently promotes tumor development and chemoresistance (Figure 8).

Discussion
p53 is a critical tumor suppressor that responds to diverse stresses by orchestrating specific cellular processes, including cell cycle arrest and apoptosis. Acetylation is a vital mechanism to regulate the activity of p53, [2] and lncRNAs are emerging as important regulators for protein modifications and activities. [7] Nevertheless, the roles of lncRNAs, especially p53-responsive lncR-NAs, in p53 acetylation and their biological significance remain unexplored. We reveal that lncRNA lnc-Ip53 is transactivated by p53 and in turn represses p53 acetylation and subsequent p53 activation by stabilizing HDAC1 protein and inhibiting p300 activity, thereby forming a negative feedback loop. Abnormal upregulation of lnc-Ip53, which is detected in different types of cancers, promotes tumor growth and chemoresistance by inhibiting p53 acetylation.
Acetylation is indispensable for p53 activation during stress responses. Enhanced acetylation increases the transcriptional activity of p53, whereas total p53 level is uncoupled from p53 activity. [18,19] Loss of acetylation at all eight lysines completely abolishes the p53-mediated cell cycle arrest and apoptosis. [5] Moreover, the p53 KQ/KQ knock-in mouse, which expresses an acetylation-mimicking mutant p53 with lysine-to-glutamine (KQ) substitutions, exhibits a p53 hyperactive phenotype in multiple tissues, confirming the significance of acetylation for p53 activation in vivo. [6] Thus, inhibition of p53 acetylation may represent a rapid and effective mechanism to inactivate p53.
It has been reported that p53 can be acetylated by p300/CBP, and deacetylated by HDAC1. [2] p300 and CBP share an overall 91% homology in their HAT domains and show similar function. [16] Recent reports show that lncRNAs, such as CASC9 and SATB2-AS1, may interact with p300/CBP and promote acetylation of histone 3, resulting in enhanced gene expression. [20][21][22][23] Nevertheless, it is still unknown whether lncRNAs may affect p53 acetylation by interacting with p300 or HDAC1. Here we disclosed that lnc-Ip53 bound to p300 and decreased p300 activity, thus impaired p53 acetylation. The 1195-1815-aa region of p300 (p300-3) contains the cysteine-histidine-rich region 3 (CH3) and HAT domains, which are required for p300 to bind to p53 and to exert its acetyltransferase function, respectively. [3,17] Considering that lnc-Ip53 can interact with p300-3 and inhibit p300-3mediated acetylation of p53, we therefore speculate that lnc-Ip53 may compete with p53 for binding to p300, and consequently inhibit p300-mediated p53 acetylation. The HDAC family members including HDAC1 are overexpressed in multiple cancer types and several HDAC inhibitors have been approved for cancer treatment. [24] HDAC inhibitors, including trichostatin A (TSA), valproic acid (VPA), sodium butyrate, and SAHA, can upregulate acetylated p53 and induce growth arrest and apoptosis, [3,[25][26][27] suggesting that HDACmediated deacetylation acts as an important mechanism for p53 inactivation and tumor growth. The protein level of HDAC1 is modulated by ubiquitination-dependent degradation [28,29] and microRNA. [30,31] However, the lncRNA that regulates HDAC1 expression is not identified yet. We disclosed that lnc-Ip53 directly bound to the 2-150-aa region of HDAC1 and stabilized HDAC1 protein, leading to accumulation of HDAC1 protein and subsequent inhibition of p53 acetylation and abrogation of p53 activity. Given that the 2-150-aa region of HDAC1 contains the lysine residue 74 (K74), whose mutation hinders HDAC1 ubiquitination, [29] we therefore suppose that binding of lnc-Ip53 may mask the K74 site of HDAC1 and thereby prevent the ubiquitination and degradation of HDAC1.
Loss of p53 function contributes to tumor growth and chemoresistance. [1] A few lncRNAs have been shown to regulate p53 function. For instance, lncRNAs P53RRA and PSTAR are downregulated in tumor tissues. P53RRA interacts with GT-Pase activating protein (SH3 domain) binding protein 1 (G3BP1) and thus releases p53 from the G3BP1 complex, which promotes p53 nuclear translocation, cell cycle arrest, apoptosis and ferroptosis; enhanced P53RRA expression represses tumor growth. [32] PSTAR interacts with heterogeneous nuclear ribonucleoprotein K (hnRNP K) to enhance its SUMOylation, thus promotes the binding of hnRNP K to p53 and stabilizes p53 protein; silencing PSTAR inhibits cell cycle arrest, promotes HCC cells proliferation and tumorigenesis. [33] Lnc-HUR1, which is upregulated by hepatitis B virus, binds to p53 and inhibits its transcriptional activity; enhanced lnc-HUR1 expression promotes tumor growth and DEN-induced HCC. [34] Nevertheless, it remains unknown whether lncRNA may regulate tumor growth and/or chemoresistance by modulating p53 acetylation. Here, we revealed that lnc-Ip53 was upregulated in different cancer types and silencing lnc-Ip53 suppressed the growth of mouse xenografts with wildtype p53, but not those with acetylation-resistant p53. Consistently, lnc-Ip53 overexpression promoted xenograft growth and chemoresistance, which was attenuated by SAHA that promoted p53 acetylation. Further analyses of acetylated p53 and its target genes in mouse xenografts confirmed that lnc-Ip53 inhibited p53 acetylation/activity in vivo. Moreover, increased level of lnc-Ip53 was associated with decreased level of acetylated p53 in human HCC and was correlated with poorer survival of HCC with wildtype p53. Our results suggest that lnc-Ip53 may promote tumor growth and chemoresistance by repressing p53 acetylation.
As a central hub of the cellular stress response, p53 signaling is amplified or terminated by positive or negative feedback loops, as exemplified by p53-MDM2 negative feedback circuit. Deregulation of p53-MDM2 loop resulting from abnormal enhanced MDM2 activity is observed in some cancer cells. [35] Though a few lncRNAs have been shown to form feedback loops with p53, [9,11,12] only p53-PURPL regulatory loop was reported to be deregulated in cancer cells. PURPL, which is transactivated by p53, reduces the level of p53 protein by preventing MYBBP1Amediated p53 stabilization, and overexpression of PURPL promotes growth of colorectal cancer. [12] Here, we characterize a novel p53-lnc-Ip53 negative feedback circuit and elucidate its biological significance in cell cycle progression and apoptosis, as well as in tumor development and chemoresistance.
To our knowledge, this is the first report concerning lnc-Ip53. The findings disclose the regulatory mechanisms of a new lncRNA in the function of p300/CBP and HDAC1 and the acetylation/activity of p53, identify a novel p53/lnc-Ip53 feedback loop, and implicate lnc-Ip53 as a potential target for anticancer therapy.
Human Tissues: Human HCC and adjacent non-tumor liver tissues were collected from patients who underwent HCC resection and confirmed histologically at Sun Yat-sen University Cancer Center. No local or systemic treatment had been conducted before surgery. Tissues were immediately snap-frozen in liquid nitrogen until use. Informed consent was obtained from each patient, and the protocol was approved by the Institutional Research Ethics Committee at Sun Yat-sen University Cancer Center.   A) The HDAC1 level was reduced by silnc-Ip53 (left) and increased by overexpressing lnc-Ip53 (right). Cells were transfected with the indicated RNA duplexes for 48 h (left) or infected with lentiviruses for 96 h (right) before immunoblotting. MOCK, cells exposed to Lipofectamine RNAiMAX but not RNA duplexes. B) The xenografts with lnc-Ip53 silencing displayed reduced HDAC1 level. n = 8 mice per group from Figure 4E. C) The xenografts with lnc-Ip53 overexpression showed increased HDAC1 level. n = 4 mice per group from Figure 4H. D) Spearman′s correlation coefficient analysis revealed a positive association between HDAC1 and lnc-Ip53 levels in human samples. The RNA levels of lnc-Ip53 in 19 paired T and N tissues are from Figure 4A and the protein levels of HDAC1 are from Figure S4B in Supporting Information. E) HDAC1 overexpression abrogated silnc-Ip53-induced p53 acetylation. Cells expressing Flag or Flag-HDAC1 were transfected with the indicated RNA duplexes for 42 h, then incubated with 5 µM SAHA for 6 h, followed by IP using IgG or antibody against acetyl lysines (anti-Ac) and then immunoblotting for p53. F) HDAC1 overexpression abrogated silnc-Ip53-stimulated CDKN1A/PUMA expression. Cells expressing Flag or Flag-HDAC1 were transfected with the indicated RNA duplexes for 52 h before qPCR. G,H) HDAC1 overexpression attenuated silnc-Ip53-induced G) G2/M arrest and H) apoptosis. Cells expressing Flag or Flag-HDAC1 were G) transfected with the indicated RNA duplexes for 39 h and then incubated with 0.1 µM Dox for 9 h before cell cycle analysis, or H) transfected with the indicated RNA duplexes for 72 h before DAPI staining. HDAC1 was analyzed by immunoblotting (A,B,D) or IHC (C). + or −, cells with (+) or without (−) the indicated treatment. Data are shown as mean ± SEM of examined samples (C) or three independent experiments (F-H); p-values were determined by unpaired Student′s t-test; NS, not significant. nonhomologous to any human genome sequences. The sequences of oligos are listed in Table S2, Supporting Information.
Rapid Amplification of cDNA Ends: The 5′-end and 3′-end of the lnc-Ip53 transcript were characterized using 5′-full rapid amplification of cDNA ends (RACE Kit; TaKaRa, Kyoto, Japan) and 3′RACE System (Invitrogen, Carlsbad, CA, USA), respectively. See also Supporting Information.
Cell Transfection: RNA duplex at a final concentration of 10 nm and plasmid DNA were transfected using Lipofectamine RNAiMAX and Lipofectamine 2000 (Invitrogen), respectively.
Lentivirus Production and Infection: To produce lentiviruses, human embryonic kidney cells expressing SV40 large T antigen (HEK293T) were co-transfected with lentivirus expression vector and packaging vectors (Lenti-X HTX Packaging Mix; Clontech, Palo Alto, CA, USA). The lentiviral supernatant was harvested and used to infect target cells. See also Supporting Information.
Analysis of Gene Expression: The gene levels were detected by real-time quantitative PCR (qPCR), immunoblotting (IB), immunohistochemistry (IHC), and northern blotting. See also Supporting Information.
Luciferase Reporter Assay: Luciferase activity was measured using dualluciferase reporter assay system (Promega). Renilla luciferase expressed by pRL-TK (Promega) was used as internal control to correct the differ-ences in both transfection and harvest efficiencies. See also Supporting Information.
Isolation of Subcellular Fraction: The cytoplasmic and nuclear extracts were isolated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, Waltham, MA, USA) and validated by immunoblotting using GAPDH and lamin B2 as controls for cytoplasmic and nuclear extracts, respectively.
EMSA and Antibody-Supershift Assay: EMSA was performed using LightShift Chemiluminescent EMSA kit (Thermo Scientific). For competition assay, unlabeled p53 consensus binding oligonucleotides were co-incubated with nuclear extracts and labeled probe. For antibodysupershift assay, nuclear extracts were preincubated with anti-p53 or isotype-matched IgG. See also Supporting Information.
ChIP Assay: ChIP assay was used to investigate the interaction between protein and gene promoter. Cells (3-5 × 10 6 ) were cross-linked with 1% formaldehyde at room temperature (RT) for 10 min, incubated with 125 mm glycine at RT for another 3 min and washed twice with icecold PBS. The cells were collected with 2 mL of DTT solution (100 mm Tris-HCl at pH 9.5, 10 mm DTT) and incubated at RT for 10 min followed by centrifugation at 5000 g and 4°C for 5 min. The cell pellets were resuspended in 150 µL of SDS lysis buffer (50 mm Tris-HCl at pH 8.0, 2 mm EDTA and 1% SDS) supplemented with protease inhibitor cocktail (Bimake, Houston, TX, USA), sonicated at 4°C for 4 min (30 s on and 30 s off) on Bioruptor (Diogenode, Liege, Belgium) under the high-power model, then centrifuged at 13 000 g and 4°C for 10 min. The supernatants were mixed with twofold volume of dilution buffer (20 mm Tris-HCl at pH 8.0, 200 mm NaCl, 2 mm EDTA, 0.1% sodium deoxycholate, 1% Triton X-100, and protease inhibitor cocktail). The cross-linked chromatin complexes in the supernatants were immunoprecipitated at 4°C overnight with 2 µg of antibodies against p53 (sc-126; Santa Cruz Biotechnology, Dallas, TX, USA), K382-acetylated p53 (ab75754; Abcam, Cambridge, MA, USA), isotype-matched mouse control IgG (A7028; Beyotime) or rabbit control IgG (A7016; Beyotime), then collected by incubation with 10 µL of protein A/G magnetic beads (Bimake) at 4°C for 2 h. The beads were washed five times with ice-cold IP lysis buffer (25 mm Tris-HCl at pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% NP-40, 5% glycerol and protease inhibitor cocktail), then eluted in 400 µL of elution buffer (0.1 m NaHCO 3 , 1% SDS) and rotated at RT for 1 h, followed by addition of 8 µL of 0.5 m EDTA and 16 µL of 1 m Tris-HCl at pH 6.5 and 60 µg of proteinase K (TaKaRa). The eluted chromatin complexes were incubated at 65°C overnight to reverse DNA-protein crosslink. The DNA was then purified and analyzed by qPCR or semi-quantitative PCR using specific primers listed in Table S2, Supporting Information. The promoters of CDKN1A and GAPDH were used as positive and negative controls, respectively.
Immunoprecipitation Assay: IP assay was performed as previously described [36] with modifications. Briefly, cells (2 × 10 6 ) were resuspended in 250 µL of NP-40 lysis buffer (20 mm Tris-HCl at pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% NP-40, 10% glycerol and protease inhibitor cocktail) and incubated at 4°C for 30 min, followed by centrifugation at 13 000 g and The ubiquitylated-HDAC1 level was C) increased by silnc-Ip53 but D) decreased by lnc-Ip53 overexpression. Cells were C) transfected with the indicated RNA duplexes for 42 h and then incubated with 20 µM MG132 for 6 h, or D) cells overexpressing lnc-Ip53 or vector control (Vec) were incubated with 20 µM MG132 for 12 h, followed by IP using anti-HDAC1 and immunoblotting by ubiquitin antibody (anti-Ub). E) Lnc-Ip53 interacted with HDAC1 in vivo. Cells expressing Flag or Flag-HDAC1 were subjected to RIP using anti-Flag affinity gel, followed by qPCR for lnc-Ip53 and negative control genes (18S rRNA, U6). F) GST pulldown assays verified the direct interaction between HDAC1 and lnc-Ip53 in vitro. The purified GST or GST-HDAC1 proteins were incubated with lnc-Ip53 or its antisense RNA (lnc-Ip53-AS), followed by GST resin-precipitation and RNA detection by qPCR. G-I) RNA pulldown assays revealed the direct interaction between the 1950-2550-nt region of lnc-Ip53 (Ip53-core) and the 2-150-aa domain of HDAC1. Biotin-labeled RNA was incubated with purified GST or GST-tagged proteins, followed by biotin-streptavidin pulldown and protein detection by immunoblotting (IB) using anti-GST antibody. In (H), the indicated GST-tagged HDAC1 fragments were incubated with full-length lnc-Ip53 or lnc-Ip53-AS. In (I), the indicated lnc-Ip53 fragments were incubated with full-length HDAC1. J) The role of lnc-Ip53 in increasing HDAC1 level was attenuated when Ip53-core was deleted. Cells overexpressing vector control (Vec) or the indicated lnc-Ip53 fragments were subjected to immunoblotting. + or −, cells with (+) or without (−) the indicated treatment. Data are shown as mean ± SEM of three independent experiments; p-values were determined by two-way ANOVA (A) or unpaired Student′s t-test (E,F); NS, not significant.
RIP Assay: RIP assay was used to explore the interaction between protein and RNA and performed as previously described [37] with modifications. Briefly, cells (2 × 10 7 ) expressing Flag or Flag-tagged proteins were cross-linked, resuspended in 0.5 mL of ice-cold RIPA lysis buffer (50 mm Tris-HCl at pH 8.0, 150 mm KCl, 5 mm EDTA, 0.1% SDS, 1% Triton X-100, 0.5% sodium deoxycholate, 0.5 mm DTT, 500 U mL −1 RNase inhibitor [Promega] and protease inhibitor cocktail), sonicated at 4°C for 4 min (30 s on and 30 s off) on Bioruptor under the low-power model, then centrifuged at 13 000 g and 4°C for 10 min. The supernatants were mixed with equal volume of RIP binding/wash buffer (25 mm Tris-HCl at pH 7.5, 150 mm KCl, 5 mm EDTA, 0.5% NP-40, 0.5 mm DTT, 500 U mL −1 RNase inhibitor and protease inhibitor cocktail). The cross-linked RNA-protein complexes in the supernatants were immunoprecipitated by incubation with 20 µL of anti-Flag affinity gel (Bimake) at 4°C for 4 h. The affinity gel was washed three times with RIP binding/washing buffer, then resuspended in 50 µL of IP lysis buffer (25 mm Tris-HCl at pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% NP-40 and 5% glycerol) supplemented with 1 µL of RNase inhibitor and 60 µg of proteinase K, followed by incubation at 55°C for 1 h and at 70°C for another 45 min to reverse RNA-protein crosslink. The RNA was extracted by TRIzol and analyzed by qPCR. The sequences of primers are listed in Table S2, Supporting Information.
Purification of GST-Fusion Proteins: Escherichia coli (E. coli) BL21 (DE3; TaKaRa) carrying vectors expressing GST or GST-fusion proteins were grown in LB (Sangon Biotech, Shanghai, China) at 37°C with shaking until OD 600 was about 0.6, followed by addition of 0.1 mm IPTG (Sangon Biotech) and incubation at 22°C for another 8 h. The E. coli were resuspended in ice-cold PBS supplemented with protease inhibitor cocktail and 1 mm DTT, sonicated at 4°C for 60 min (5 s on and 15 s off) on Bioruptor under the high-power model, then centrifuged at 15 000 g and 4°C for 30 min. GST-fusion proteins in the supernatants were purified by GST-sefinose (TM) resin (Sangon Biotech) according to the manufacturer's instructions. All proteins were concentrated in BC100 buffer (20 mm Tris-HCl at pH 8.0, 0.5 mm EDTA, 100 mm KCl, 20% glycerol, 0.5 mm DTT, 0.5 mm PMSF [Beyotime], and protease inhibitor cocktail) by Amicon Ultra-15 Centrifugal Filter Devices (UFC901008; Millipore, Billerica, MA) and stored at −80°C. The protein concentration was measured by BCA Protein Assay Kit (Beyotime).
In Vitro Transcription: For GST pulldown, RNA pulldown and in vitro acetylation assays, biotin-labeled RNAs and unlabeled RNAs were in vitro transcribed from the relevant PCR products using Biotin RNA Labeling Mix (Roche, Mannheim, Germany) and NTP Mix (Invitrogen) together with T7 RNA polymerase (Promega), respectively, then purified by LiCl precipitation. The purified RNAs were incubated in annealing buffer (Beyotime) at 90°C for 2 min, placed on ice for 2 min and kept at RT for 20 min to form secondary structure. The sequences of primers are listed in Table S2, Supporting Information.
GST Pulldown Assay: GST pulldown assays were performed as previously described [38] with modifications. To examine in vitro interaction between HDAC1 and lnc-Ip53, 2 µg of purified GST or GST-HDAC1 proteins were incubated with 0.5 µg of in vitro transcribed unlabeled RNA in 500 µL of RIP binding/wash buffer (25 mm Tris-HCl at pH 7.5, 150 mm KCl, 5 mm EDTA, 0.5% NP-40, 0.5 mm DTT, 500 U mL −1 RNase inhibitor and protease inhibitor cocktail) at 4°C for 2 h, and collected by incubation with 100 µL of BeyoGold GST-tag purification resin (Beyotime) at 4°C for 6 h. The resin was washed five times with ice-cold RIP binding/wash buffer. The RNA in the resin-precipitates was extracted by TRIzol and analyzed by qPCR. The sequences of primers are listed in Table S2, Supporting Information.