SWI/SNF remains localized to chromatin in the presence of SCHLAP1

Abstract

SCHLAP1 is a long noncoding RNA that is reported to function by depleting the SWI/SNF complex from the genome. We investigated the hypothesis that SCHLAP1 affects only specific compositions of SWI/SNF. Using several assays, we found that SWI/SNF is not depleted from the genome by SCHLAP1 and that SWI/SNF is associated with many coding and noncoding RNAs, suggesting that SCHLAP1 may function in a SWI/SNF-independent manner. © 2018, The Author(s), under exclusive licence to Springer Nature America, Inc.

Full text

Brief CommuniCation
https://doi.org/10.1038/s41588-018-0272-z
1Department of Genetics, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA. 2Department of Pharmacology and Lineberger Comprehensive
Cancer Center, University of North Carolina, Chapel Hill, NC, USA. 3Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill,
Chapel Hill, NC, USA. *e-mail: tmagnuson@unc.edu
SCHLAP1 is a long noncoding RNA that is reported to func-
tion by depleting the SWI/SNF complex from the genome.
We investigated the hypothesis that SCHLAP1 affects only
specific compositions of SWI/SNF. Using several assays,
we found that SWI/SNF is not depleted from the genome by
SCHLAP1 and that SWI/SNF is associated with many coding
and noncoding RNAs, suggesting that SCHLAP1 may function
in a SWI/SNF-independent manner.
The long noncoding RNA (lncRNA) second chromosome locus
associated with prostate cancer 1 (SCHLAP1) is a promising bio-
marker for metastatic prostate cancer1,2. SCHLAP1 is proposed to
function by antagonizing the SWI/SNF complex through direct
interaction, leading to complete disruption of SWI/SNF genomic
occupancy1. Evidence for this mechanism comes from the reported
loss of SMARCB1 occupancy measured by chromatin immunopre-
cipitation followed by sequencing (ChIP–seq)1. SWI/SNF is a large
multi-subunit chromatin remodeling complex that can be combina-
torially assembled to yield hundreds to thousands of biochemically
unique complexes3–5. We investigated the alternative hypothesis that
distinct forms of SWI/SNF are affected by SCHLAP1 expression.
However, using a varriety of biochemical and genomics assays, we
demonstrate that SWI/SNF occupancy is unaffected by SCHLAP1
expression, in contrast to results reported previously1. We show that
SWI/SNF binds coding and noncoding RNA, raising the possibility
that SCHLAP1 function is SWI/SNF-independent.
Consistent with the report by Prensner et al.1, we observed an
interaction between SMARCB1 and SCHLAP1 (Fig. 1a). We next
generated SCHLAP1-overexpressing benign prostate epithelial cells
(RWPE1;SCHLAP1 cells) or control cells (RWPE1;LACZ cells)1
(SCHLAP1 gift of A. Chinnaiyan). This model is the same as
that originally used to suggest global depletion of SMARCB1 by
SCHLAP1 (ref. 1). We then confirmed the phenotype of the cells
with respect to SWI/SNF expression and growth (Fig. 1b and
Supplementary Figs. 1,2). In addition, we confirmed the key result
that SCHLAP1 increased cell invasion (Fig. 1c).
To investigate which SWI/SNF subunits were depleted from chro-
matin upon SCHLAP1 expression, we fractionated RWPE1;LACZ
and RWPE1;SCHLAP1 cells based on subcellular localization or salt
extraction. Surprisingly, all SWI/SNF subunits assayed remained
strongly enriched in the chromatin or high salt fractions (Fig. 1d
and Supplementary Figs. 3–5). Consistent with these biochemical
experiments, we found that SMARCA4 and SMARCB1 localization
was not affected in RWPE1;SCHLAP1 cells by immunofluorescence
(Supplementary Fig. 3b). Immunoprecipitation of SMARCA4 or
SMARCB1 demonstrated that the SWI/SNF complex remains intact
in the presence of SCHLAP1 (Supplementary Figs. 3c,6). Finally, we
used a malignant rhabdoid tumor cell line with inducible SMARCB1
that, when expressed, causes growth arrest (gift of B. Weissman)6.
We reasoned that, if SCHLAP1 disrupted SMARCB1 chromatin
occupancy, then overexpression of SCHLAP1 should allow G401
cells to proliferate following induction of SMARCB1. However,
SMARCB1 induction led to growth arrest in a dose-dependent man-
ner (Fig. 1e and Supplementary Figs. 7,8). Together, these results
demonstrate that SCHLAP1 does not induce changes to SWI/SNF
composition or its association with chromatin.
We next performed ChIP–seq for three SWI/SNF subunits
(SMARCB1, SMARCA2, and SMARCA4) in RWPE1;SCHLAP1
cells. In contrast to a previous report1, we identify robust binding
for all three subunits in RWPE1 cells expressing SCHLAP1 (Fig. 1f).
In RWPE1;SCHLAP1 cells, we identified 6,490, 22,185, and 51,505
peaks for SMARCB1, SMARCA2, and SMARCA4, respectively
(Supplementary Table 1). This large number of peaks is in con-
trast to the previous report1, which identified approximately 6,500
SMARCB1 peaks in RWPE1;LACZ cells and close to no peaks in
the SCHLAP1-expressing cells. The numbers of peaks are consis-
tent with previous work from our laboratory, which showed 30,000–
45,000 SMARCA4 peaks4. In addition, we and others have reported
a large number of SWI/SNF peaks for a variety of subunits3,4,7,8.
The majority of SMARCA2 peaks overlapped a SMARCA4 peak
(Supplementary Fig. 9a), and SWI/SNF peaks were predominantly
located at promoters (45–75%; Supplementary Fig. 9b)4. SWI/SNF
binding was most prominent at highly expressed genes, with little to
no occupancy at non-expressed genes (Fig. 1g; expression data from
GSE98898 (ref. 9)). These results demonstrate that SCHLAP1 does
not function by disrupting SWI/SNF occupancy genome-wide, and
raises the question of how SCHLAP1 functions to promote cell
invasion and progression to metastatic disease.
To investigate whether SCHLAP1 expression induces chromatin
changes, we performed assay for transposase-accessible chromatin
with high-throughput sequencing (ATAC-seq) on RWPE1;LACZ
and RWPE1;SCHLAP1 cells10. We identified 273 and 3,167 sites
that open and close, respectively (Supplementary Table 2 and
Supplementary Fig. 10). The sites that open were more likely to
be located distally or in introns of genes (Supplementary Fig. 10).
Sites that open upon SCHLAP1 expression were enriched for dis-
tinct motifs compared to those that close (Supplementary Fig. 11
and Supplementary Tables 3, 4). Open sites were enriched in motifs
for TEAD and AP1 transcription factors, which are known to have
a role in defining oncogenic enhancers (Supplementary Fig. 11 and
Supplementary Tables 3, 4)11. To test whether these sites became
SWI/SNF remains localized to chromatin in the
presence of SCHLAP1
JesseR.Raab1, KeriaynN.Smith1, CamarieC.Spear1, CarlJ.Manner1, J.MauroCalabrese2 and
TerryMagnuson 1,3*
NATURE GENETICS | VOL 51 | JANUARY 2019 | 26–29 | www.nature.com/naturegenetics
26
Content courtesy of Springer Nature, terms of use apply. Rights reserved