Increased ANKRD1 Levels in Early Senescence Mediated by RBMS1-Elicited ANKRD1 mRNA Stabilization

Abstract

Cellular senescence is a dynamic biological process triggered by sublethal cell damage and driven by specific changes in gene expression programs. We recently identified ANKRD1 (ankyrin repeat domain 1) as a protein strongly elevated after triggering senescence in fibroblasts. Here, we set out to investigate the mechanisms driving the elevated production of ANKRD1 in the early stages of senescence. Our results indicated that the rise in ANKRD1 levels after triggering senescence using etoposide (Eto) was the result of moderate increases in transcription and translation, and robust mRNA stabilization. Antisense oligomer (ASO) pulldown followed by mass spectrometry revealed a specific interaction of the RNA-binding protein RBMS1 with ANKRD1 mRNA that was confirmed by ribonucleoprotein immunoprecipitation analysis. RBMS1 abundance decreased in the nucleus and increased in the cytoplasm during Eto-induced senescence; in agreement with the hypothesis that RBMS1 may participate in post-transcriptional stabilization of ANKRD1 mRNA, silencing RBMS1 reduced, while overexpressing RBMS1 enhanced ANKRD1 mRNA half-life after Eto treatment. A segment proximal to the ANKRD1 coding region was identified as binding RBMS1 and conferring RBMS1-dependent increased expression of a heterologous reporter. We propose that RBMS1 increases expression of ANKRD1 during the early stages of senescence by stabilizing ANKRD1 mRNA.

Introduction

Cellular senescence is characterized by indefinite growth arrest and altered gene expression programs triggered in response to sublethal damage.¹ Senescent cells display enlarged and flattened cell morphology, increased expression of traditional senescence marker proteins such as p16 (CDKN2A), p21 (CDKN1A), and p53 (TP53), increased lysosomal activity with a senescence-associated β-galactosidase (SA-β-gal) active at pH 6, and a senescence-associated secretory phenotype (SASP).^2–5 Due to the dynamic nature of senescence, different genes are expressed at different times,⁶ leading to altered protein expression patterns as senescence progresses. Although senescent cells have been associated with exacerbation of many age-associated diseases and decline,^5,7,8 they are indispensable for morphogenesis during development, tumor suppression in young individuals, and wound repair. Accordingly, there is escalating interest in elucidating many aspects of this developmental program, including the dynamic changes of expressed proteins during different phases of senescence.

ANKRD1 (ankyrin repeat domain 1) was identified as a protein highly elevated in senescent cells.^9–11 Also known as cardiac ankyrin repeat protein (CARP), ANKRD1 is prominently expressed in cardiovascular and muscle diseases, playing a vital role as a component of the muscle sarcomere.^12,13 Previous reports have indicated that ANKRD1 functions as a transcriptional co-activator of p53 or a transcriptional corepressor of NF-κB.^14,15 Moreover, ANKRD1 was found to be highly induced during the initial stages of skin wound healing in mice.¹⁶ Studies utilizing ANKRD1 knockout (KO) revealed that ANKRD1 was essential for timely wound closure following cutaneous wounding.¹⁷ Linking these observations is the fact that a transient accumulation of cell senescence was found to promote wound repair,¹⁸ and ANKRD1 mRNA expression levels increased in replicative senescent fibroblasts.¹¹

As the mechanisms driving ANKRD1 production in senescence are unknown, here we set out to identify them. In human diploid fibroblasts that were induced to senescence through exposure to low doses of DNA damage or by replicative exhaustion, we observed significant increases in ANKRD1 expression levels during the early stages of senescence, followed by a subsequent decline. Although some regulation was mediated by increased transcription of ANKRD1 mRNA, much regulation occurred at the post-transcriptional level. Through antisense oligomer (ASO) pulldown of ANKRD1 mRNA followed by mass spectrometry (MS) analysis, we found that the RNA-binding protein RBMS1 associated with ANKRD1 mRNA during senescence triggered by etoposide (Eto). RBMS1 has been reported to participate in cancer progression by regulating the stability of mRNAs encoding cancer-related proteins,¹⁹ and RBMS1 controls the translation efficiency of SLC7A11 mRNA suggesting a potential role of RBMS1 in post-transcriptional regulation.²⁰ We found that during Eto-triggered senescence, RBMS1 increased in the cytoplasm, where it was found to interact with ANKRD1 mRNA and rendered it more stable. These findings were confirmed by RBMS1 silencing and overexpression experiments, and were extended to a heterologous reporter, where a segment from the ANKRD1 3′UTR interacting with RBMS1 conferred increased reporter activity. Taken together, our findings suggest that RBMS1 promotes the expression of ANKRD1 in the early stages of senescence by enhancing ANKRD1 mRNA stability.

Results

ANKRD1 expression levels increase with cell senescence

In a previous study, we found that ANKRD1 mRNA was significantly elevated in human diploid fibroblasts rendered senescent by replicative exhaustion when compared to proliferating fibroblasts.¹¹ To confirm this finding and extend it to other models of senescence, we investigated the expression of ANKRD1 protein in human WI-38 fibroblasts that were rendered replicatively senescent (RS) by culture to population doubling level (PDL) ∼55, compared to proliferating (P) fibroblasts at ∼ PDL23. In addition, we analyzed WI-38, IMR-90, and BJ fibroblasts that were rendered senescent by treatment with etoposide (Eto; 25 μM Eto for BJ cells and 50 μM Eto both for WI-38 and IMR-90 cells) and collected 6 days later. Senescence was confirmed in all four cell models by visualizing the presence of SA-β-gal activity (phase contrast micrographs), and by quantifying the percentages of SA-β-gal-positive cells (graphs) (Figure 1A). Senescence was also monitored by measuring cell proliferation levels in each senescence model using the MTS assay (Materials and Methods) (Figure 1B). Western blot analysis confirmed the increased expression levels of ANKRD1 in each senescent cell model, along with the classic senescence markers p21 (CDKN1A) and DPP4 (as described),²¹ which increase with senescence, as well as the decreased levels of the proliferation marker PLK1 (as described)²² in senescent cells (Figure 1C). Furthermore, reverse transcription (RT) followed by quantitative (q) PCR analysis indicated a significant increase in ANKRD1 mRNA levels in the senescent populations (Figure 1D).

Figure 1.

ANKRD1 upregulation in senescent cells. (A–D) Human diploid fibroblasts WI-38 that were either proliferating (P, ∼PDL23) or underwent replicative senescence (RS, ∼PDL55), as well as WI-38, IMR-90, and BJ human diploid fibroblasts that were either proliferating or rendered senescent by exposure to Etoposide (Eto, 50 µM for 6 days for WI-38 or IMR-90 cells; 25 µM for 6 days for BJ cells) were evaluated. Cells were assessed for SA-β-gal activity using light microscopy (A, left) and by calculating percentages of SA-β-gal-positive cells relative to the total population (A, right); proliferation was assessed by the MTS assay (B); expression of ANKRD1 along with protein markers that increased (DPP4, p21) or decreased (PLK1) with senescence, and loading control protein ACTB (β-actin) by Western blot analysis (C); and ANKRD1 mRNA levels (normalized to the levels of GAPDH mRNA, encoding the housekeeping protein GAPDH) by RT-qPCR analysis (D). (E) Visualization of ANKRD1 levels by fluorescence microscopy in WI-38 cells rendered senescent by Eto treatment for 24 h; DAPI staining was used to identify the nucleus. Size bar, 100 μm. (F) Western blot analysis of ANKRD1 levels in cytoplasmic extracts (CE) and nuclear extracts (NE) prepared by fractionating WI-38 cells that had been treated with DMSO or Eto for 24 h; the senescence-associated proteins p21 and p53, the cytoplasmic marker TUBA1A, and the nuclear marker LMNB1 were also assessed as control proteins. Ponceau S staining of the transfer membrane was used to evaluate the even loading and transfer of samples. All experiments were performed at least three times. The data in panels A, B, and D represent the means ± SD. Statistical significance was determined using Student’s t test (**P < 0.01; ***P < 0.001). Protein molecular weights were assessed relative to marker proteins (kDa, kilodaltons).

Immunofluorescence (IF) analysis revealed that ANKRD1 was primarily localized in the cytoplasm of WI-38 cells that had been rendered senescent by treatment with Eto for 24 h (Figure 1E). These findings were confirmed by Western blot analysis of cytoplasmic extracts (CE) and nuclear extracts (NE) prepared from cells treated as described in Figure 1E. As shown (Figure 1F), ANKRD1 levels increased with Eto treatment and were generally higher in CE than in NE, similar to the levels of the senescence protein marker p21; the nuclear proteins p53 (TP53) and LMNB1 (Lamin B1), known to increase and decrease with senescence, respectively, and the cytoplasmic protein TUBA1A (α-tubulin), were assessed to verify the fractionation efficiency. Taken together, our results indicate a robust and transient increase in ANKRD1 levels in senescent cells, primarily in the cytoplasm.

Given the dynamic nature of cellular senescence,⁶ we investigated the expression levels of ANKRD1 over time as cells progressed towards senescence. Western blot analysis of WI-38 whole-cell lysates collected at 2, 6, 24, 48, and 72 h after Eto treatment showed an initial increase in ANKRD1 levels (2, 6, and 24 h), followed by a decline, indicating early expression of ANKRD1 in senescence (Figure 2A), while after treatment with another chemical inducer of senescence, doxorubicin (Dox), ANKRD1 levels peaked at earlier time points (2 and 6 h) and declined afterward (Figure 2B). Eto-treated IMR-90 and BJ fibroblasts also showed an early increase in ANKRD1 levels followed by a drop after 24 h (Figure 2C and D). The different patterns of ANKRD1 abundance across these senescence paradigms reflect the fact that senescence programs vary depending on factors like cell type, metabolic status of the cell, and dose of damaging agent triggering senescence.

Figure 2.

Early induction of ANKRD1 during senescence. (A,B) Western blot analysis of ANKRD1 and p21 levels in WI-38 cells treated with either etoposide (Eto) (A) or doxorubicin (Dox) (B) for the times indicated. The control protein ACTB was assessed to identify differences in sample loading and transfer. (C,D) IMR-90 (C) and BJ (D) fibroblasts were treated with Eto (50 μM and 25 μM, respectively) for the indicated times, whereupon whole-cell lysates were prepared and the levels of ANKRD1, senescence marker p21, and loading control ACTB were assessed by Western blot analysis. Protein molecular weights were assessed relative to marker proteins (kDa).

In response to Eto-induced senescence, ANKRD1 levels rise through increased transcription, mRNA stabilization, and translation

We then investigated the regulatory mechanisms responsible for the increased levels of ANKRD1 during senescence. Analysis of the change in ANKRD1 mRNA levels over time following exposure to Eto by RT-qPCR analysis revealed a pattern of expression (Figure 3A) that mirrored the changes in protein (Figure 2A). To evaluate if this increase was due to induced transcription, we designed three different primer pairs directed at different exon-intron junctions on ANKRD1 precursor mRNA (pre-mRNA; Figure 3B, left). Amplification of the ANKRD1 pre-mRNA by RT-qPCR analysis was chosen as a surrogate method to assess ongoing transcription, as this approach quantifies the levels of pre-spliced ANKRD1 mRNA. As shown (Figure 3B, right), ANKRD1 pre-mRNA levels increased during Eto-induced senescence with a pattern similar to that of ANKRD1 mRNA (Figure 3A), although the magnitude of induction was substantially lower than that of ANKRD1 mRNA or ANKRD1 protein, suggesting that post-transcriptional regulatory mechanisms were likely involved.

Figure 3.

Regulation of ANKRD1 production at the levels of transcription, mRNA stability, and translation following Eto treatment. (A,B) RT-qPCR analysis of levels of mature ANKRD1 (A) or pre-ANKRD1 mRNA (B) using three different primer pairs spanning different exon-intron junctions (B, left), in WI-38 cells treated with Eto (50 μM) for 2, 6, 24, 72, and 144 h (B, right). (C) RT-qPCR analysis of the stability of ANKRD1, MYC, and GAPDH mRNAs in WI-38 cells treated with DMSO or Eto (50 μM) for 24 h, and then incubated in actinomycin D (2 μg/mL) for the times indicated. RNA levels in each sample were normalized to 18S rRNA levels. Half-lives for ANKRD1 mRNA were estimated as the time required for the mRNA to reach one-half (50%, discontinuous line) of its abundance at time 0 h. (D) Cytoplasmic lysates obtained from WI-38 cultures processed as in (A) were fractionated through sucrose gradients to assess the polysome distribution profiles (D, left). Unbound RNA (fractions 1, 2), ribosomal subunits and monosomes (fractions 3–5), low-molecular-weight polysomes (LMWPs, fractions 6–8) and high-molecular-weight polysomes (HMWPs, fractions 9–12), are indicated in the polysome traces. The relative distribution of ANKRD1 mRNA and ACTB mRNA (encoding the housekeeping protein ACTB) was measured by RT-qPCR analysis of RNA in each of the 12 gradient fractions, and represented the values in each fraction as a percentage of the total in that gradient; representative gradient data are shown (D, right). (E) WI-38 cells treated with DMSO or Eto as explained in (A) were further incubated with the inhibitor of protein synthesis cycloheximide (CHX, 50 μg/mL) for the indicated times, after which Western blot analysis was performed to assess the remaining levels of ANKRD1, the housekeeping stable protein ACTB, and the labile senescent markers p21 and p53 (E, top). The intensity of the bands on the Western blots was measured using ImageJ and represented in the graph (ANKRD1 relative to ACTB, bottom). All experiments were performed three or more times. The data in panels A-C represent the means ± SD. Statistical significance (*P < 0.05; **P < 0.01; ***P < 0.001) was assessed with Student’s t test.

We thus investigated if the stability of ANKRD1 mRNA changed in response to Eto treatment. We measured the half-life of ANKRD1 mRNA by blocking its synthesis using actinomycin D, a drug that inhibits RNA polymerase II (responsible for transcribing mRNAs), and measuring the time required for the existing ANKRD1 mRNA to reach 50% of its abundance at time 0, before actinomycin D was added. As shown in Figure 3C, while the half-life of ANKRD1 mRNA in control (DMSO-treated) cells was approximately 4.9 h, it increased to >8 h in Eto-treated cells, indicating greater stability for ANKRD1 mRNA after Eto treatment. As controls, we evaluated the levels of the stable transcript GAPDH mRNA and the levels of the labile transcript MYC mRNA following addition of actinomycin D, neither of which was substantially affected by Eto treatment (Figure 3C). These results supported the view that increased stability of ANKRD1 mRNA after Eto treatment contributed to the rise in ANKRD1 mRNA levels (Figure 3A).

We then investigated if the robust elevation of ANKRD1 protein levels observed at early time points following exposure to Eto (Figure 2A) were due to changes in the synthesis or degradation of ANKRD1. To test whether ANKRD1 mRNA translation changed in response to Eto treatment, we analyzed any shifts in distribution of ANKRD1 mRNA in polysomes that were fractionated across a sucrose gradient (Materials and Methods). Cytoplasmic lysates of untreated (DMSO) and Eto-treated (Eto) WI-38 cells were collected; after fractionation by ultracentrifugation through linear 10–50% sucrose gradients, unbound RNA was recovered in fractions 1 and 2, ribosome subunits (40S, 60S) and monoribosomes (80S) in fractions 3–5, low-molecular-weight polysomes (LMWP) in fractions 6–8, and high-molecular-weight polysomes (HMWP) in fractions 9–12. After isolating RNA from the fractions, we quantified the relative abundance of ANKRD1 mRNA in each fraction by RT-qPCR analysis. Inspection of the polysome traces (Figure 3D, left) revealed that Eto treatment caused an increase in total RNA in nontranslating fractions (∼1–4) and an overall decrease in total RNA in polysomes of all sizes (6–12). Importantly, however, ANKRD1 mRNA levels showed a modest rightward shift in distribution across the polysomes in Eto-treated cells, as well as higher peaks of abundance in high-molecular-weight polysome fractions, compared to DMSO-treated cells (Figure 3D). While both groups (Eto and DMSO) peaked at fraction 9, the shift in distribution suggests that translation of ANKRD1 mRNA may be moderately increased after Eto treatment. ACTB mRNA, encoding the housekeeping protein ACTB (β-actin), did not display this shift, despite some differences in abundance in fraction 8 (Figure 3D).

We also examined if the stability of ANKRD1 was altered in cells treated with Eto compared to DMSO-treated cells. We used cycloheximide (CHX), an inhibitor of nascent translation, to measure ANKRD1 protein stability. Western blot analysis to evaluate the decline in ANKRD1 levels after halting translation with CHX indicated that ANKRD1 protein was quite unstable, declining ∼5-fold in CHX-treated cells, whether they were treated with Eto or left untreated (Figure 3E). In contrast, the control stable protein ACTB remained highly abundant in both populations, and the relatively labile proteins p21 and p53 only showed moderate decline by 30 min (Figure 3E). Together, these results suggest that ANKRD1 is upregulated during early senescence through modest increases in transcription and translation, and a substantial increase in mRNA stability.

Identification of RNA-binding proteins associated with ANKRD1 mRNA

We then investigated the post-transcriptional regulatory mechanisms that control the expression of ANKRD1. We specifically focused on RNA-binding proteins (RBPs) capable of interacting with ANKRD1 mRNA, as RBPs are widely involved in regulating gene expression at the post-transcriptional level.^23,24 To identify these proteins, we first designed biotinylated antisense oligos (ASOs) to capture endogenous ANKRD1 mRNA and identify associated proteins; LacZ ASOs were assayed in control reactions (Figure 4A).

Figure 4.

Identification of ANKRD1 mRNA-associated proteins. (A) Schematic of the ANKRD1 mRNA ASO pulldown followed by mass spectrometry analysis to identify proteins interacting with ANKRD1 mRNA (created using BioRender). Briefly, WI-38 cells treated with Eto (50 µM for 24 h) were lysed and incubated with ANKRD1 ASOs along with LacZ ASOs in control reactions. After pulling down the complexes with streptavidin beads, the samples were analyzed by mass spectrometry (Materials and Methods), to identify ANKRD1 mRNA-bound proteins. (B) RT-qPCR analysis to assess the enrichment of ANKRD1 mRNA in ANKRD1 ASO pulldown samples compared to LacZ ASO pulldown samples. (C) Heatmap representation of ANKRD1 mRNA-interacting proteins from the mass spectrometry analysis in panel (A), comparing LacZ ASO and ANKRD1 ASO pulldown samples. Scale indicates PSM (peptide-spectrum match). (D) WI-38 cells were transfected with siRNAs directed at the mRNAs that encode PARP1, PABP1, PTBP1, HNRNPC, MSI2H, RALY, RBMS1, HNRNPM, HNRNPU, or IGF2BP3; 72 h later, cells were treated with Eto (50 μM), and 24 h after that, the levels of ANKRD1 in each transfected population were evaluated by Western blot analysis. The levels of the senescence marker p21 and loading control ACTB were assessed in the same samples (D, top). Red, silencing groups in which ANKRD1 levels increased; blue, silencing groups in which ANKRD1 levels decreased. The silencing efficiency of each RBP was measured by RT-qPCR analysis of the mRNAs encoding the corresponding protein in the same cultures. Data are represented as percent relative to the levels of each RBP mRNA in the siCTRL population (100%, discontinuous line) (D, bottom). (E) Measurement by RT-qPCR analysis of the enrichment in ANKRD1 mRNA in RIP reactions using antibodies recognizing RBMS1, HNRNPC, or RALY compared to IgG RIP reactions, in Eto-treated WI-38 cells. (F) Western blot analysis to assess the levels of RBMS1 in ANKRD1 ASO compared to LacZ ASO pulldown samples, along with input lysate aliquots. GAPDH and HuR served as negative controls, as they were not identified in the mass spectrometry analysis. All experiments were performed more than three times. The data in panels B, D, and E represent the means ± SD. Statistical significance (*P < 0.05; **P < 0.01; ***P < 0.001) was assessed with Student’s t test. Protein molecular weights were assessed relative to marker proteins (kDa).

Using total RNA collected from WI-38 fibroblasts (∼PDL23) treated for 24 h with Eto, we first validated the efficiency of the pulldown of ANKRD1 mRNA by RT-qPCR analysis; as shown (Figure 4B), ANKRD1 mRNA was strongly enriched in the ASO pulldown samples. We then identified the associated proteins by mass spectrometry (MS); peptides corresponding to the proteins most highly abundant in the ANKRD1 ASO pulldown are indicated (Figure 4C). To study if one or more of these RBPs regulated the abundance of ANKRD1 mRNA, we individually silenced each RBP using siRNAs directed at each encoding mRNA. As shown, silencing HNRNPC or RALY increased ANKRD1 protein levels, while silencing RBMS1 decreased ANKRD1 protein levels, as revealed by Western blot analysis (Figure 4D). The efficiency of silencing each RBP was evaluated by RT-qPCR analysis of the levels of mRNAs encoding each RBP (Figure 4D; the discontinuous line represents the levels of each RBP mRNA before silencing).

Further validation of the binding of ANKRD1 mRNA to specific RBPs was carried out by ribonucleoprotein immunoprecipitation (RIP) analysis. Using lysates prepared from WI-38 fibroblasts that had been treated with Eto for 24 h, we used antibodies that recognized RBMS1, HNRNPC, or RALY, as well as control antibodies (IgG), to carry out RIP analysis. Changes in the enrichment of ANKRD1 mRNA in each IP (relative to IgG IP) were normalized to changes in the levels of GAPDH mRNA present in the same IP samples, as GAPDH mRNA (encoding the housekeeping protein GAPDH) was a surrogate measure of differences in sample input. The results indicated that ANKRD1 mRNA was most enriched in RBMS1 RIP relative to the other two RBPs (Figure 4E). Furthermore, RBMS1 was robustly enriched in the ANKRD1 mRNA ASO pulldown compared to the LacZ ASO pulldown; control proteins HuR and GAPDH did not associate with either ASO (Figure 4F). In sum, RBMS1 was identified as a key protein capable of binding ANKRD1 mRNA and required for ANKRD1 steady-state abundance.

Mobilization of RBMS1 to the cytoplasm during Eto-induced senescence

We next studied whether RBMS1 levels changed in WI-38 cells undergoing senescence in response to Eto treatment. As shown by Western blot analysis (Figure 5A), total RBMS1 levels did not change across the times studied, while the levels of p21 did.

Figure 5.

Increased cytoplasmic localization of RBMS1 and binding to ANKRD1 mRNA after Eto treatment. (A) Western blot analysis of the levels of RBMS1, p21, and ACTB in WI-38 cells at the indicated times of Eto treatment. (B) Western blot analysis of the levels of RBMS1 in cytoplasmic extracts (CE) and nuclear extracts (NE) prepared from cells treated as in (A). The senescence marker p21, the cytoplasmic marker TUBA1A, and the nuclear marker LMNB1 were also analyzed. Ponceau S staining of the transfer membrane was performed to monitor loading and transferring of the samples. (C) RIP assay to measure ANKRD1 mRNA enrichment in RBMS1 IP at the indicated times after adding Eto compared to untreated samples (time 0); enrichment was represented as the abundance of ANKRD1 mRNA in RBMS1 IP relative to IgG IP at each time point. GAPDH mRNA levels were used to normalize all RT-qPCR data. All experiments were performed at least three times. The data in panel C represent the means ± SD. Statistical significance was determined using Student’s t test (**P < 0.01). Protein molecular weights were assessed relative to marker proteins (kDa).

Given that translocation to the cytoplasm is important for the function of some RBPs,^25–28 we investigated the subcellular distribution of RBMS1 as a function of time after induction of senescence. As shown, RBMS1 was more abundant in nuclear extracts (NE) in proliferating WI-38 cells, but its abundance in cytoplasmic extracts (CE) increased after time in Eto, particularly by 72 h, suggesting that a fraction of RBMS1 translocated to the cytoplasm during Eto-induced senescence. The cytoplasmic marker TUBA1A, the nuclear marker LMNB1, and the senescence marker p21 were also included to monitor sample processing; membranes were stained with Ponceau S to ensure equal loading and transfer of the samples (Figure 5B). Although the lysates used for RIP included both nuclear and cytoplasmic contents, RIP analysis generally mirrored the increased abundance of RBMS1 in the cytoplasm, displaying increased enrichment in ANKRD1 mRNA levels after exposure to Eto (see ANKRD1 mRNA in RBMS1 IP relative to IgG IP at each time point; Figure 5C). These results suggest that RBMS1 associated with ANKRD1 mRNA mainly in the cytoplasm. In sum, RBMS1 presence in the cytoplasm as well as its binding to ANKRD1 mRNA increased during Eto-induced senescence.

RBMS1 enhances the stability of ANKRD1 mRNA

To investigate the consequences of RBMS1 binding ANKRD1 mRNA, we first examined the impact of RBMS1 silencing on the levels of ANKRD1 mRNA and ANKRD1 protein in populations that were untreated (DMSO) or were treated with Eto. WI-38 cells were transfected with siRNAs directed at RBMS1 mRNA (siRBMS1), using siCTRL in control transfections; 72 h later, cells were treated with Eto or left untreated (DMSO), and both protein and RNA were analyzed 24 h later. As shown, silencing RBMS1 led to marked reductions in the levels of both ANKRD1 and ANKRD1 mRNA, particularly after Eto treatment (Figure 6A). We then analyzed the half-life of ANKRD1 mRNA in cells that were processed as explained in Figure 6A, expressing normal or silenced levels of RBMS1, and subsequently treated with Eto. Using actinomycin D to shut off transcription as explained earlier (Figure 3C), silencing RBMS1 shortened the half-life of ANKRD1 mRNA in Eto-treated from t_1/2>8 h (estimated t_1/2∼10.7 h) to a shorter t_1/2∼6 h. MYC and GAPDH mRNAs were included again as labile and stable controls (Figure 6B). These findings suggested that RBMS1 promoted the stability of ANKRD1 mRNA in cells undergoing Eto-induced senescence.

Figure 6.

RBMS1 positively regulates ANKRD1 mRNA stability. (A) WI-38 cells were transfected with siRBMS1 or siCTRL; 72 h later, the cells were treated with Eto or DMSO, and 24 h after that, ANKRD1 levels were measured by Western blot analysis (along with RBMS1, p21, and ACTB) (left) and ANKRD1 mRNA levels (normalized to GAPDH mRNA levels) were measured by RT-qPCR analysis (right). (B) The stabilities of ANKRD1 mRNA, labile control MYC mRNA, and stable control GAPDH mRNA were calculated in cells transfected and treated with Eto as explained in (A). Actinomycin D was added to the cultures (as explained in Figure 3C) for the times indicated and mRNA levels were measured at the times indicated; the half-life (t_1/2) of ANKRD1 mRNA in each group is shown. (C) WI-38 cells were infected with a control lentivirus (LentiCTRL) or a lentivirus that overexpressed a myc-tagged RBMS1 protein (LentiRBMS1); 24 h later, cells were treated with Eto and 24 h after that, cultures were assessed. The levels of RBMS1, myc tag, ANKRD1, p21 and ACTB were measured by Western blot analysis (left) and the levels of ANKRD1 mRNA were measured by RT-qPCR analysis (normalized to GAPDH mRNA levels; right). (D) Cultures that were processed as described in panel (C) and treated with Eto were used to evaluate the half-life (t_1/2) of ANKRD1 mRNA using actinomycin D as explained in panel (B). (E) In WI-38 cells that were treated as explained in panel (A), the levels of ANKRD1 pre-mRNA were analyzed using the same three primer pairs shown in Figure 3B. (F) Polysome analysis was carried out as explained in Figure 3D. WI-38 cytoplasmic lysates from cells processed as explained in panel (A) were fractionated through sucrose gradients to assess polysome distribution profiles (F, left). Unbound RNA (fractions 1, 2), ribosomal subunits and monosomes (fractions 3–5), LMWPs (fractions 6–8), and HMWPs (fractions 9–12), are indicated. The relative distribution of ANKRD1 mRNA and ACTB mRNA (encoding the housekeeping protein ACTB, β-Actin) was measured by RT-qPCR analysis in each of the 12 gradient fractions and is presented as a percentage of the values in each fraction as a percentage of the total mRNA in that gradient; representative gradient data are shown (F, right). All experiments were performed three or more times. In panels A to E, the data represent the mean ± SD. Statistical significance (*P < 0.05; **P < 0.01; ***P < 0.001) was assessed with Student’s t test. Protein molecular weights were assessed relative to marker proteins (kDa).

We also performed the converse experiment, that is, to overexpress RBMS1 and study the effect on ANKRD1 expression in proliferating and Eto-treated cultures. For this intervention, we employed lentiviral vectors (LentiCTRL and LentiRBMS1) to infect fibroblasts for efficient RBMS1 overexpression. As shown (Figure 6C), 24 h after infection with LentiRBMS1 and selection (24 additional hours, Materials and Methods), RBMS1 and RBMS1 mRNA were markedly elevated in both untreated cells (DMSO) and cells that were treated with Eto for 24 h. The increases in ANKRD1 and ANKRD1 mRNA levels were due at least in part to stabilization of ANKRD1 mRNA, as overexpression of RBMS1 increased the half-life of ANKRD1 mRNA from t_1/2∼8 h to an estimated t_1/2∼15.3 h (Figure 6D). The half-lives of control transcripts (GAPDH and MYC mRNAs) were not affected by RBMS1 overexpression.

Since the levels of ANKRD1 were also modestly elevated by Eto treatment via increased transcription and translation (Figure 3B and D), we also investigated whether RBMS1 influenced these processes. As shown, silencing RBMS1 in DMSO and Eto-treated cultures and employing the same primer pairs that were used in Figure 3B, we found that reducing RBMS1 levels did not affect the expression of pre-ANKRD1 mRNA, indicating that RBMS1 did not appear to influence the transcription of the ANKRD1 gene (Figure 6E). Similarly, analysis of fractionated polysomes revealed that RBMS1 silencing did not have an impact on the translation of ANKRD1, as measured by monitoring the sizes of polysomes associated with ANKRD1 mRNA. Likewise, ACTB mRNA, measured as a control, did not show changes as a function of RBMS1 levels (Figure 6F). Taken together, these results indicate that RBMS1 positively regulates ANKRD1 expression by stabilizing ANKRD1 mRNA.

RBMS1 associates with specific ANKRD1 3′UTR segments to stabilize ANKRD1 mRNA

To identify the region in ANKRD1 mRNA that interacts with RBMS1, a pulldown experiment was conducted using biotinylated RNAs transcribed in vitro spanning the entire ANKRD1 mRNA (termed F1-F7; Materials and Methods). Out of the seven fragments tested, RBMS1 was found to interact specifically with biotinylated fragments F5 and F7, both located in the 3′UTR region of ANKRD1 mRNA (Figure 7A). To further examine if RBMS1 influenced ANKRD1 mRNA levels by interacting with ANKRD1 3′UTR, luciferase reporter constructs derived from the dual reporter psiCHECK2 were generated in which the Renilla luciferase (RL) coding region was fused to each of the three fragments of ANKRD1 3′UTR (Figure 7B). The reporter plasmid, psiCHECK2, also expresses firefly luciferase (FL), which serves as an internal control for normalizing differences in transfection efficiency (Figure 7B). We first silenced RBMS1 by transfecting cells with siRBMS1 or siCTRL; 24 h after that, we added Eto and transfected reporter constructs psiCHECK2(Empty), psiCHECK2-ANKRD1 3′UTR(F5), psiCHECK2-ANKRD1 3′UTR(F6), and psiCHECK2-ANKRD1 3′UTR(F7). Twenty-four hours later, measurement of RL luciferase activity relative to internal control FL luciferase activity (RL/FL) revealed that RL activity from a chimeric RL mRNA bearing 3′UTR F5 (but not other chimeric RL mRNAs), was significantly elevated in control cells and significantly reduced when RBMS1 was silenced (Figure 7C). We then performed the converse experiment, overexpressing RBMS1 using LentiRBMS1 (LentiCTRL in control infections) as explained in Figure 6C, and 24 h later we added Eto and transfected the same reporter constructs (Figure 7B); after 24 h, we measured RL and FL activities. As observed, RBMS1 overexpression only increased RL/FL ratios in the study group in which chimeric RL was fused to F5, while Empty, F6, and F7 did not change (Figure 7D). These findings support the notion that RBMS1 promotes ANKRD1 production by associating with the ANKRD1 3′UTR F5 region.

Figure 7.

RBMS1 binds a fragment of ANKRD1 3′UTR that promotes expression of a heterologous reporter. (A) Schematic representation of the ANKRD1 mRNA, including the 5′UTR, coding region, and 3′UTR, as well as the biotinylated RNA fragments (F1 through F7) that were synthesized for biotin pulldown analysis. The presence of RBMS1 in the pulldown reactions was assessed by western blot analysis. (B) Schematic of the psiCHECK2 dual Renilla luciferase (RL) and internal control firefly luciferase (FL) reporter constructs used to identify the region(s) of ANKRD1 3′UTR regulated by RBMS1. (C,D) WI-38 fibroblasts were transfected with either CTRL siRNA or RBMS1 siRNA and used 24 h for analysis (C), or infected with LentiCTRL or LentiRBMS1 and 24 h later they were selected in puromycin for an additional 24 h (D). Cultures were then split into 24-well plates and treated with Eto (50 μM) and transfected with each of the four reporter vectors in panel (B); 24 h later, cells were collected, luciferase activities were measured as described (Materials and Methods), and RL/FL ratios were determined. (E) Proposed model. In proliferating fibroblasts, the levels of ANKRD1 are low and ANKRD1 mRNA is unstable; in the first 24 h after treating with Eto (during early senescence), ANKRD1 mRNA is stabilized partly due to the action of RBMS1, and ANKRD1 levels increase. In panels C and D, the data represent the mean ± SD. Statistical significance (*P < 0.05; **P < 0.01; ***P < 0.001) was assessed with Student’s t test. Protein molecular weights were assessed relative to marker proteins (kDa).

In sum, this study revealed a significant increase in ANKRD1 levels during the early stages of Eto-induced senescence. In this paradigm, RBMS1 positively regulated ANKRD1 expression by interacting with the ANKRD1 3′UTR, and enhancing ANKRD1 mRNA stability and ANKRD1 protein levels (Figure 7E).

Discussion

ANKRD1 has been found to be abundantly expressed in various cancers, skeletal muscle, and the heart, but its role in senescence has not yet been directly investigated.^{12,13,29–31} ANKRD1 was also reported to increase early during skin wound healing in mice, while ANKRD1 KO mouse models displayed delayed wound healing.^16–18 We and others previously reported that ANKRD1 was highly elevated in senescence.^9–11 Given the crucial roles of senescent cells in wound healing, we set out to study the regulation of ANKRD1 expression during senescence in molecular detail. We first confirmed that the levels of both ANKRD1 mRNA and ANKRD1 protein were significantly increased in multiple senescence paradigms (including exposure to ionizing radiation, not shown), and further observed that the elevation occurred early and transiently after Eto treatment, suggesting a potential specific function of ANKRD1 in response to early DNA damage.

We established that transcription of the ANKRD1 gene increased modestly (∼3-fold) in response to DNA damage, evidenced by a rise in the levels of ANKRD1 pre-mRNA. Although we did not investigate the transcriptional mechanisms leading to an increase in transcription, YAP-TEAD may be involved, as it was shown to activate transcription of the ANKRD1 gene.³² In the senescence paradigm studied here, a major driver of the increase in ANKRD1 protein levels was the stabilization of ANKRD1 mRNA in response to senescence-inducing Eto doses, and therefore we focused on this novel level of regulation of ANKRD1 production. By employing ASOs for pulldown followed by proteomic analysis with mass spectrometry, we identified RBMS1 as a key regulator of ANKRD1 in response to Eto treatment. A previous study reported that RBMS1 enhanced the translation of SLC7A11 mRNA by interacting with the translation initiation factor eIF3D,²⁰ while another study found that RBMS1 stabilized the glycosyltransferase B4GALT1 mRNA in breast cancer cells.³³ Crosslinking followed by immunoprecipitation and sequencing (CLIP-seq) analysis uncovered several mRNA targets of RBMS1 in colon cancer and further established that RBMS1 bound the 3′UTR regions of these mRNAs.¹⁹

We found a relative increase in the levels of cytoplasmic RBMS1 during Eto-induced senescence, while the overall expression of RBMS1 remained unchanged. In addition, the association of RBMS1 with ANKRD1 mRNA increased over time. Whether the increased binding by RBMS1 to ANKRD1 mRNA is influenced by post-translational modifications of RBMS1 (e.g., by phosphorylation, as has been reported for other RBPs),^34–36 remains to be investigated. The transport mechanisms responsible for mobilizing RBMS1 across the nuclear envelope also await investigation. In addition, whether RBMS1 works in conjunction with other RBPs, microRNAs or other regulatory RNAs to control ANKRD1 mRNA stability also remains undetermined. In particular, the possibility that RBPs RALY and/or HNRNPC might contribute to lowering ANKRD1 mRNA stability in proliferating cells deserves further study.

Experiments employing biotinylated RNA to map the site where RBMS1 binds ANKRD1 mRNA revealed preferential binding to the 3′UTR. The RBPmap prediction database (https://rbpmap.technion.ac.il/),³⁷ similarly suggested several putative binding sites on ANKRD1 mRNA converging in the 3′UTR (not shown). Fragment F5 (∼290 nucleotides long) appeared to mediate the regulatory effects of RBMS1 on ANKRD1 expression, as adding F5 to the 3′ end of the Renilla coding region rendered the chimeric reporter responsive to the levels of RBMS1. Supporting this notion, the luciferase reporter bearing F5 displayed reduced activity after silencing RBMS1, and conversely, it displayed increased activity after overexpressing RBMS1 (Figure 7C and D). Although RBMS1 CLIP-seq analysis in colon cancer cells was reported previously,¹⁶ ANKRD1 mRNA was not found among the targets. The availability of anti-RBMS1 antibodies superior to those available today will enable CLIP-seq analysis of RBMS1 binding to the fibroblast transcriptome in the future. Efforts to narrow down the region of RBMS1 that binds ANKRD1 mRNA, and the identification of possible post-translational modifications driving RBMS1 export to the cytoplasm and RNA-binding activity also warrant future study.

In light of the influence of RBMS1 on ANKRD1 expression, it will also be interesting to test if ablation of RBMS1 in mice reduces the levels of ANKRD1 and might, in turn, impair wound repair. Although the body-wide, constitutive RBMS1 knockout is embryonic lethal,³⁸ tissue-specific, inducible knockout mice, as well as mice in which RBMS1 is overexpressed, will shed light on the role of RBMS1 in different developmental and disease processes. The impact of RBMS1 on the expression of other mRNAs and proteins implicated in senescence, wound repair, cancer, skeletal muscle homeostasis, and physiology and disease in general, remains to be studied systematically.

In closing, the molecular function of ANKRD1 in senescence remains largely unknown. Previous studies proposed that nuclear ANKRD1 may serve as a transcription cofactor to facilitate p53 activation, suppress NF-κB activity in muscle cells, and antagonize the production of MMP13.^14,15,39 Our investigation revealed that ANKRD1 is primarily localized in the cytoplasm, suggesting the presence of undiscovered and unexplored functions for ANKRD1 in senescent fibroblasts. As the role of ANKRD1 in senescence-associated processes comes fully into view, knowledge of the molecular regulators of ANKRD1 production will lead to the discovery of attractive targets for therapeutic intervention.

Materials and Methods

Cell culture, senescence induction, and cell number measurements

Human WI-38 diploid fibroblasts (HDFs, Coriell Institute for Medical Research: ID AG06814), IMR-90 (ATCC), and BJ fibroblasts (ATCC) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco) supplemented with 20% fetal bovine serum (FBS, Gibco), antibiotics penicillin and streptomycin (Gibco), and non-essential amino acids (Gibco). All cells were maintained in a 5% CO₂ incubator at 37 °C. Replicative senescence (RS) was induced in WI-38 fibroblasts by passaging proliferating cells with a population doubling (PDL) of ∼23 until they ceased to grow (∼PDL55). Etoposide-induced senescence was triggered by treating cells continuously with a single dose of etoposide (Eto) at concentrations of 50 μM for WI-38 and IMR-90 cells, and 25 μM for BJ cells, as described.⁴⁰ Doxorubicin-induced senescence was triggered by treating cells continuously with a single dose of doxorubicin (Dox) at concentrations of 250 nM for WI-38 cells; DMSO was used in control treatments, as described.⁴¹

To assess population size, cells were counted with a cell counter (Bio-Rad, TC20 automated cell counter), and average cell numbers per milliliter were represented in the bar graphs. Proliferating WI-38, IMR-90, and BJ fibroblasts were seeded in 96-well plates (10⁴ cells per well), treated with either DMSO or Eto for 6 days, and analyzed by the MTS assay (Promega) following the manufacturer’s protocol.

Transfection, transduction, and treatment

Proliferating WI-38 fibroblasts (∼PDL23) were seeded at 2.0–2.5 × 10⁵ cells/mL in a 6-well plate to reach ∼50% confluency by the following day. Cells were transfected using Lipofectamine 2000 (ThermoFisher Scientific) in 1 mL Opti-MEM (Gibco) following the manufacturer’s protocol. Control siRNA (siCTRL), siPARP, siPABP1, siPTBP1, siHNRNPC, siMSI2H, siRALY, siRBMS1, siHNRNPM, siHNRNPU, and siIGF2BP3 were transfected at 50 nM and cells were collected 72 h later. The sequences of siRNAs are listed (Supplementary material, Table S1).

WI-38 fibroblasts were transduced with lentiviruses that either expressed RBMS1 or were empty (LentiRBMS1, LentiCTRL, respectively; Origene) at 10 MOI (multiplicity of infection) for 24 h; infected cultures were selected by treatment with puromycin (1 μg/mL) for an additional 24 h and the selection medium was then removed before initiating further treatments.

To measure mRNA stability, WI-38 cells were treated with either DMSO or Eto for 24 h, or transfected with either siCTRL or siRBMS1, or transduced with either LentiCTRL or LentiRBMS1 in the presence of Eto treatment for 24 h, whereupon they were exposed to either DMSO (control vehicle) or the inhibitor of RNA polymerase II actinomycin D (ActD, Cell Signaling) (2 µg/mL), and RNA was collected at 0, 2, 3, 4, 6, 8, or 9 h for analysis. To measure protein stability, WI-38 cells were treated with either DMSO or Eto for 24 h, and then exposed to either DMSO (control vehicle) or cycloheximide (CHX, Sigma) (50 µg/mL) for 0, 5, 15, or 30 min, whereupon protein samples were collected for analysis.

Senescence-associated (SA)-β-gal activity

We assessed SA-β-gal activity following the manufacturer’s protocol (Cell Signaling Technology). Briefly, fibroblasts were washed twice with 1× phosphate-buffered saline (PBS), fixed for 15 min at 25 °C, and stained in a freshly prepared solution at pH 6.0. Images were captured using a Nikon Digital Sight camera system mounted on a Nikon Eclipse TS100 microscope. SA-β-gal-positive cells were counted using ImageJ and the percentages of stained cells in three different fields were calculated.

Immunofluorescence

WI-38 fibroblasts were washed with PBS and then fixed in 4% formaldehyde (ThermoFisher Scientific) in PBS for 10 min. Cells were then washed with PBS, permeabilized with 0.2% Triton-X-100 for 5 min and washed again with PBS. To block nonspecific binding, the cells were incubated with 10% goat serum (ThermoFisher Scientific) for 1 h at 37 °C. Primary antibodies were diluted in normal goat serum and incubated with the fixed cells for 18 h at 4 °C. After multiple washes with PBS, the cells were incubated with secondary antibodies for 1 h at 37 °C. Following extensive washes, the cells were incubated with DAPI (1:5000, ThermoFisher Scientific) for 15 min at 25 °C in the dark. Fluorescent images were acquired using a Keyence microscope.

Protein isolation, subcellular fractionation, and Western blot analysis

To isolate proteins, cells were first washed twice in cold PBS, and the supernatant was discarded. Cells were then harvested by scraping and cell pellets were lysed using RIPA buffer supplemented with protease and phosphatase inhibitors (ThermoFisher Scientific). After 10 min of incubation on ice, the lysate was centrifuged (18,506 × g) for 20 min to remove insoluble fractions. The supernatant was collected, and protein concentration was measured using the Bradford assay (Bio-Rad). Protein lysates were mixed with 2× SDS Laemmli buffer (Bio-Rad) and heated at 95 °C for 5 min.

To separate nuclear and cytoplasmic proteins, we used the NE-PER kit (ThermoFisher Scientific) following the manufacturer’s instructions. The resulting protein lysates were then size-separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes (Bio-Rad). Membranes were blocked with 5% non-fat milk in TBST for 1 h at 25 °C. Membranes were then incubated with primary antibodies at 4 °C for 16 h. A list of antibodies used in this study is provided in Supplementary material, Table S2. The membranes were washed with 1× TBST and incubated with secondary antibodies in 5% non-fat milk for 1 h at 25 °C. After washing, the membranes were incubated with enhanced chemiluminescence (ECL) solution (Kwik Quant), and the chemiluminescent signals were detected using a ChemiDoc system (Bio-Rad).

RNA isolation and RT-qPCR analysis

Total RNA was isolated using TriPure™ Isolation Reagent from Sigma-Aldrich following the manufacturer’s instructions with phenol-chloroform extraction. Briefly, cells were harvested and homogenized in TriPure™ Isolation Reagent, and RNA was extracted by adding chloroform, followed by centrifugation. The aqueous phase was collected, and RNA was precipitated with isopropanol, washed with 75% cold ethanol, and dissolved in RNase-free water. The quantity and quality of RNA were assessed using a NanoDrop™ spectrophotometer (ThermoFisher Scientific).

For cDNA synthesis, 1 μg of total RNA was reverse transcribed using Maxima Reverse Transcriptase (ThermoFisher Scientific) following the manufacturer’s instructions. The synthesized cDNA was subjected to quantitative PCR analysis using specific primer pairs and SYBR green master mix (Kapa Biosystems) on a QuantStudio 5 Real-Time PCR System (ThermoFisher Scientific). The primers (Integrated DNA Technologies, IDT) and are listed in Supplementary material, Table S3. Relative mRNA levels were calculated using the 2^-ΔΔCT method and normalized to either ACTB mRNA, GAPDH mRNA, B2M mRNA, or 18S rRNA.

Polyribosome fractionation

To perform polyribosome fractionation, cells were treated with 100 μg/mL of CHX (Sigma) for 10 minutes at 37 °C in a 5% CO₂ incubator. Subsequently, the cytoplasmic lysates were isolated using polysome extraction buffer (PEB; 20 mM Tris–HCl pH 7.5, 100 mM NaCl, 5 mM MgCl₂ and 0.5% NP-40) and fractionated by ultracentrifugation through 10–50% linear sucrose gradients. Samples were then divided into 12 fractions for RT-qPCR analysis, as described.⁴²

Antisense oligo (ASO) pulldown and mass spectrometry (MS)

For the affinity pulldown of endogenous human ANKRD1 mRNA, total lysates from WI-38 cells that had been treated with Eto for 24 h were prepared using RIPA buffer supplemented with protease inhibitors (ThermoFisher Scientific) and RNase inhibitor (ThermoFisher Scientific). The lysates were incubated with 1 μg of biotin-labeled DNA oligomers (ASOs) complementary to ANKRD1 overnight at 4 °C. The biotinylated ASOs are listed in Supplementary material, Table S4. RNA complexes were washed with TENT buffer (20 mM Tris-HCl pH 8.0, 2 mM EDTA pH 8.0, 500 mM NaCl, 1% v/v Triton X-100) and isolated with Dynabeads M-280 streptavidin beads (ThermoFisher Scientific), as previously described.⁴³

Two samples were sent for mass spectrometry (MS) analysis (performed by Poochon Scientific). The MS analysis of the samples was carried out using a Thermo Scientific Orbitrap Exploris 240 Mass Spectrometer and a Thermo Dionex UltiMate 3000 RSLCnano System. Based on the PSM # (peptide spectrum match counts, an indicator of abundance), the relative abundance of a protein was determined for comparison between two samples.

Ribonucleoprotein immunoprecipitation (RIP)

RIP assay was performed as described.⁴⁴ Briefly, to assess the direct interactions between RBMS1 and ANKRD1 mRNA, we conducted RIP using Dynabeads Protein G (ThermoFisher Scientific) coated with either control IgG (Cell Signaling Technology) or RBMS1 antibody (Proteintech). Total lysates were prepared using RIPA lysis buffer supplemented with protease and phosphatase inhibitors (ThermoFisher Scientific) and RNase inhibitor (ThermoFisher Scientific), and incubated with Dynabeads (ThermoFisher Scientific) overnight at 4 °C. The beads were then washed with NT2 buffer (50 mM Tris–HCl [pH 7.5], 150 mM NaCl, 1 mM MgCl₂, 0.05% NP-40) five times. After treatment with 20 units of DNase I (20 min at 37 °C) and 0.1% SDS/0.5 mg/mL proteinase K (20 min at 55 °C), RNA was isolated and used to determine the enrichment of ANKRD1 mRNA in the immunoprecipitation (IP) materials.

Luciferase assay

Briefly, specific primers (Supplementary material, Table S5) were utilized to amplify 3′UTR fragments of ANKRD1 mRNA, and the fragments were inserted into the psiCHECK2 plasmid, downstream of the Renilla luciferase (RL) open reading frame (ORF). The vector also contained a constitutively expressed internal control, Firefly luciferase (FL). In the reporter assays, cells were transfected with RBMS1 siRNA along with control siRNA and 48 h later, the cells were split into 24 wells. Plasmids psiCHECK2(Empty), psiCHECK2-ANKRD1 3′UTR(F5), psiCHECK2-ANKRD1 3′UTR(F6), and psiCHECK2-ANKRD1 3′UTR(F7) were transfected using Lipofectamine 2000 (ThermoFisher Scientific). Twenty-four hours later, the cells were washed with cold PBS and lysed using passive lysis buffer (Promega). The activities of the reporter RL and internal control FL were analyzed using a dual-luciferase assay kit (Promega) following the manufacturer’s protocol.

Biotin pulldown assay

To synthesize biotinylated transcripts, PCR fragments were prepared using the primers indicated in Supplementary material, Table S6. Subsequently, the biotinylated transcripts were synthesized utilizing the MaxiScript T7 kit (Ambion). Cells were lysed with PEB buffer and incubated with purified biotinylated transcripts overnight at 4 °C. Following this incubation, the complexes were isolated using streptavidin-coupled Dynabeads (ThermoFisher Scientific). The level of RBMS1 present in the pulldown material was determined by Western blot analysis, as described.⁴⁵

Statistical analysis and graphs

Unless otherwise noted, experiments were repeated a minimum of three times. Quantitative data are represented as the means ± SD and compared statistically by unpaired Student’s t test using Prism GraphPad (9.0). Statistical significance was denoted as follows: *P < 0.05; **P < 0.01; ***P < 0.001.

Funding Statement

This work was supported in its entirety by the National Institute on Aging Intramural Research Program, NIH.

Authors’ Contributions

CHS, MR, CA, JLM, EJ, RM, JHY, AP, and DT performed experiments; CHS, KMM, and KA analyzed data; JLM, RM, and XY prepared and contributed reagents; CHS, KA, and MG wrote the paper; all authors provided input during the writing of the manuscript.

Data availability statement

The data that support the findings of this study are openly available in Figshare at https://figshare.com/search?q=10.6084/m9.figshare.25681281.

Disclosure statement

No potential conflict of interest was reported by the author(s).