SWI/SNF chromatin-remodeling complexes function in noncoding RNA-dependent assembly of nuclear bodies

Abstract

Paraspeckles are subnuclear structures that form around nuclear paraspeckle assembly transcript 1 (NEAT1) long noncoding RNA (lncRNA). Recently, paraspeckles were shown to be functional nuclear bodies involved in stress responses and the development of specific organs. Paraspeckle formation is initiated by transcription of the NEAT1 chromosomal locus and proceeds in conjunction with NEAT1 lncRNA biogenesis and a subsequent assembly step involving >40 paraspeckle proteins (PSPs). In this study, subunits of SWItch/Sucrose NonFermentable (SWI/SNF) chromatin-remodeling complexes were identified as paraspeckle components that interact with PSPs and NEAT1 lncRNA. EM observations revealed that SWI/SNF complexes were enriched in paraspeckle subdomains depleted of chromatin. Knockdown of SWI/SNF components resulted in paraspeckle disintegration, but mutation of the ATPase domain of the catalytic subunit BRG1 did not affect paraspeckle integrity, indicating that the essential role of SWI/SNF complexes in paraspeckle formation does not require their canonical activity. Knockdown of SWI/SNF complexes barely affected the levels of known essential paraspeckle components, but markedly diminished the interactions between essential PSPs, suggesting that SWI/SNF complexes facilitate organization of the PSP interaction network required for intact paraspeckle assembly. The interactions between SWI/SNF components and essential PSPs were maintained in NEAT1-depleted cells, suggesting that SWI/SNF complexes not only facilitate interactions between PSPs, but also recruit PSPs during paraspeckle assembly. SWI/SNF complexes were also required for Satellite III lncRNA-dependent formation of nuclear stress bodies under heat-shock conditions. Our data suggest the existence of a common mechanism underlying the formation of lncRNA-dependent nuclear body architectures in mammalian cells.

Keywords: chromatin-remodeling complex; long noncoding RNA; nuclear bodies; ribonucleoprotein assembly.

Fig. 1.

Paraspeckle localization of SWI/SNF complex components. (A and B) Localization of BRG1, BRM, and BAF155 to paraspeckles in human and mouse cells. The SWI/SNF components were detected by immunocytochemical analyses, and the paraspeckles were visualized by RNA FISH analyses of NEAT1 lncRNA. The experiments were performed in HeLa cells (A) and other human cell lines (A549 and HEK293), as well as a mouse cell line (NIH 3T3) (B). (Scale bars: 10 μm.) (C) Relocalization of BRG1 and BRM to perinucleolar caps upon transcriptional arrest caused by treatment of HeLa cells with actinomycin D (+Act D). The indicated PSPs and SWI/SNF components were detected by immunocytochemical analyses, and the paraspeckles were visualized by RNA FISH analyses of NEAT1 lncRNA. NONO, which is known to relocalize to perinucleolar caps, was used as a positive control. COIL relocates to another cap structure. (Scale bars: 10 μm.) (D) I-EM detection of BRG1 (Upper) in normal (control) paraspeckles and enlarged paraspeckles induced by the treatment of HeLa cells with MG132. Chromosome localization was detected by using an anti-histone H3 antibody (Lower). The arrows indicate paraspeckles. Cyt, cytoplasm; IG, interchromatin granule clusters; Nu, nucleus. (Scale bars: 0.5 µm.)

Fig. 2.

SWI/SNF subunits interact directly with essential paraspeckle components. (A) Detection of direct interactions between SWI/SNF components (BRG1 and BRM) and NEAT1 lncRNA by UV cross-linking IP. After UV cross-linking, the coimmunoprecipitated RNAs were quantified by quantitative RT-PCR, and the IP efficiencies (as percentages) were determined (n = 3). The “UV−” lanes are negative controls without UV cross-linking. IgG was used as a negative control for IP. (B) Interactions between BRG1 and five essential PSPs (Left) detected by co-IP with an anti-BRG1 antibody (αBRG1) in the presence (+) or absence (−) of RNase A (Left) or ethidium bromide (EtBr; Right). GAPDH was detected as a negative control.

Fig. 3.

SWI/SNF complexes are essential for paraspeckle formation. (A) The effects of RNAi-mediated knockdown of the indicated SWI/SNF components on paraspeckle formation. The effects were monitored by analyzing paraspeckle integrity by using RNA FISH detection of NEAT1 and immunofluorescent detection of PSPC1. The siRNAs used in these experiments are listed in Table S3. Because BRG1 and BRM are functionally redundant, these proteins were concomitantly knocked down with a combination of two siRNAs (BRG1#2/BRM#1 and BRG1#4/BRM#2). The results of single RNAi-mediated knockdown of BRG1 or BRM are shown in Fig. S3. (Scale bars: 10 μm.) (B) Quantification of the number of paraspeckle-positive cells after RNAi-mediated knockdown of the indicated SWI/SNF components (as shown in A). Data are represented as the mean ± SD of three replicates. The cell numbers for the control, BRG1#2/BRM#1, BRG1#4/BRM#2, BAF170, BAF155, BAF57, and BAF47 groups were 435, 229, 266, 165, 149, 198, and 227, respectively (*P < 0.05 and **P < 0.02 by Student t test). (C) The lack of paraspeckles in adrenal cortex adenocarcinoma SW13 cells (Upper). RNA FISH was used to detect NEAT1, and immunocytochemistry was used to detect RBM14 in SW13 and HeLa (control) cells. (Scale bars: 10 μm.) Nuclear speckles in SW13 cells as detected by RNA FISH analyses of MALAT1 and immunofluorescent analyses of SRSF2 (Lower). (Scale bars: 10 μm.) (D) Paraspeckle formation does not require the canonical nucleosome-remodeling activity of SWI/SNF complexes. Schematic representations of WT and two ATPase mutants (M1 and M2) of BRG1 are shown on the left. HAP1 cells were treated with siBRM to abolish the background expression of BRM. The essential lysine residue (marked as “K”) in WT BRG1 and the artificially inserted proline residue (marked as “P”) in the M2 mutant are shown in red and blue, respectively. The paraspeckles detected in WT, M1, and M2 HAP1 cells are shown on the right. (Scale bar: 10 μm.)

Fig. 4.

SWI/SNF complexes are required for proper interactions between PSPs. (A) RNA IP of NEAT1 with six essential PSPs from control and ΔSWI/SNF cells. The amounts of immunoprecipitated RNAs were quantified by quantitative RT-PCR using the NEAT1 and NEAT1_2 primer pairs shown in Table S2. The ratio of immunoprecipitated RNA from ΔSWI/SNF cells to that from the control cells was calculated. The expression level of GAPDH was measured as a control. Data are represented as the mean ± SD of three replicates. (B and C) Interactions between the essential PSPs in control (−) and ΔSWI/SNF (+) cells. Co-IP was performed by using antibodies against HNRNPK, RBM14, and SFPQ, and the expression levels of the indicated PSPs were monitored by immunoblotting (B). The interactions that were diminished by <40% in ΔSWI/SNF cells are indicated by I–IV, and the quantified co-IP ratios (ΔSWI/SNF/control) in I–IV are plotted on the graph shown in C. (D) Interactions between RBM14 and essential PSPs in control (−) and paraspeckle-depleted ΔFUS (+) cells. Co-IP was performed by using an anti-RBM14 antibody, and the expression levels of the indicated PSPs were monitored by immunoblotting. (E) Interactions between BRG1 and the essential PSPs in control (−) and NEAT1-depleted (+) cells.

Fig. 5.

SWI/SNF complexes are required for the formation of nSBs. (A) Immunostaining of scaffold attachment factor B (magenta) to detect the formation of nSBs before and after heat shock (HS) of control HeLa cells, SW13 cells, and BRG1/BRM-specific siRNA-treated HeLa cells. (Scale bars: 10 μm.) (B) Quantification of the numbers of nSB-positive cells shown in A. Data are presented as the mean ± SD of three replicates. The cell numbers for HeLa(HS−), SW13(HS−), HeLa(HS+), SW13(HS+), C(HS−), ΔBRG1/ΔBRM(HS−), C(HS+), and ΔBRG1/ΔBRM(HS+) groups were 276, 228, 303, 202, 341, 356, 332, and 403, respectively (**P < 0.02 by Student t test). (C) I-EM detection of BRG1 in nSBs. Dashed line (Left) indicates a single nSB. (Scale bars: Left, 0.5 μm; Right, 100 nm.) (D) RNA IP of Sat III lncRNA from heat-shock–treated HeLa cells using the αBRG1 antibody. The expression level of β-actin was used as a control. RT, reverse transcription. The sequences of the PCR primers used are shown in Table S2. (E) RT-PCR analyses of Sat III lncRNA expression in control HeLa cells, SW13 cells, and BRG1/BRM-specific siRNA-treated HeLa cells. HS, heat shock at 42 °C for 1 h. (F) A model of lncRNA-dependent nuclear body assembly. SWI/SNF complexes may be common assembly factors in mammalian cells. Nuclear bodies act to sequestrate specific regulatory proteins to control the expression of specific genes (9).