Regulation of alternative polyadenylation by Nkx2-5 and Xrn2 during mouse heart development

  1. Keisuke Nimura  Is a corresponding author
  2. Masamichi Yamamoto
  3. Makiko Takeichi
  4. Kotaro Saga
  5. Katsuyoshi Takaoka
  6. Norihiko Kawamura
  7. Hirohisa Nitta
  8. Hiromichi Nagano
  9. Saki Ishino
  10. Tatsuya Tanaka
  11. Robert J Schwartz
  12. Hiroyuki Aburatani
  13. Yasufumi Kaneda  Is a corresponding author
  1. Osaka University Graduate School of Medicine, Japan
  2. Kyoto University Graduate School of Medicine, Japan
  3. Graduate School of Frontier Biosciences, Osaka University, Japan
  4. University of Houston, Unites States
  5. The University of Tokyo, Japan
7 figures and 1 additional file

Figures

Figure 1 with 9 supplements
Nkx2-5 deficiency increases transcription from regions downstream of transcription termination sites.

(A) Co-occupancies of each pair of factors and histone modifications are shown. White indicates a high correlation, and red indicates a low correlation. (B) Nkx2-5, Tbx5, and Gata4 were …

https://doi.org/10.7554/eLife.16030.003
Figure 1—source data 1

Overlap of peaks between transcription factors and between the results from this study and those from previousely published studies.

(A) Overlap of peaks between Nkx2-5, Tbx5, and Gata4 ChIPseq data in this study. (B) Overlap of Nkx2-5 peaks among E12.5 hearts (this study), HL1 cells with BirA-fused Nkx2-5 (HL1_BirA) (He et al., 2011), and Adult hearts (van den Boogaard et al., 2012). (C) Overlap of Tbx5 between E12.5 hearts (this study) and HL1 cells with BirA-fused Tbx5 (HL1_BirA) (He et al., 2011). (D) Overlap of Gata4 peaks between native ChIPseq in this study and previousely published crosslink ChIPseq (He et al., 2014).

https://doi.org/10.7554/eLife.16030.004
Figure 1—figure supplement 1
Transcription factors-associated proteins in E12.5 hearts.

This figure is related to Figure 1b.

https://doi.org/10.7554/eLife.16030.005
Figure 1—figure supplement 2
ChIPseq replicate correlations.

Scatterplots of pair-wise ChIPseq replicates and Pearson correlation are shown.

https://doi.org/10.7554/eLife.16030.006
Figure 1—figure supplement 3
In vivo transcription factor binding motif by native ChIPseq.

De novo motif analysis by Homer using all peaks in ChIPseq data. Obtained motifs are compared with the most matched known motif, respectively. The motif (T/C)GATTGG found in Gata4 peaks is similar …

https://doi.org/10.7554/eLife.16030.007
Figure 1—figure supplement 4
Validation of the antibodies used for ChIP-seq.

The indicated proteins were immunoprecipitated from nuclear extracts of E12.5 hearts using the corresponding antibodies. Arrowheads indicate immunoprecipitated proteins. Asterisks indicate the IgG …

https://doi.org/10.7554/eLife.16030.008
Figure 1—figure supplement 5
Genome browser representation at Tnnt2 and Atps2a2 loci.

Genome browser representation of the indicated histone modifications, transcription factors, and transcription factor-associated protein enrichment profiles in E12.5 hearts is shown for the Tnnt2

https://doi.org/10.7554/eLife.16030.009
Figure 1—figure supplement 6
Heatmap of factor occupancy and histone modification enrichment for 8 kb regions centred on TSSs (left panel) and TTSs (middle panel) are shown with reference to the RefSeq gene expression level (right panel).
https://doi.org/10.7554/eLife.16030.010
Figure 1—figure supplement 7
Average signal profiles over a 3 kb meta-gene including 3 kb upstream and 3 kb downstream.

The ChIP-seq data were analysed using CEAS. The lines correspond to genes with High, Middle, Low, and No expression and all RefSeq genes.

https://doi.org/10.7554/eLife.16030.011
Figure 1—figure supplement 8
Genome browser representation of strand-specific RNA-seq tag counts from eCMs transfected with the indicated siRNAs.

Red boxes indicate read-through RNAs. neg., negative strand., pos., positive strand. The arrow heads show polyadenylation sites.

https://doi.org/10.7554/eLife.16030.012
Figure 1—figure supplement 9
mRNA with long 3’UTR in Nkx2-5-knockout embryonic hearts.

(A) Genome browser representation of the read-through RNAs in Nkx2-5-knockout embryonic hearts. Red boxes indicate long 3’UTR. The arrow heads show polyadenylation sites. (B) Average profiles of …

https://doi.org/10.7554/eLife.16030.013
Figure 2 with 4 supplements
A link between Nkx2-5-dependent chromatin conformation and RNAPII.

(A) Chromatin conformation capture (3C) analysis of the TSSs and downstream regions of Tnnt2 and Atp2a2 in the indicated siRNA-treated eCMs. The corresponding BACs for the regions were used as …

https://doi.org/10.7554/eLife.16030.014
Figure 2—figure supplement 1
Quantification of 3C and western blotting data.

(A) Quantification of 3C data by qPCR at Figure 2A. Tnnt2, n = 5. Atp2a2, n = 4. *, p < 0.05. (B) Quantification of western blotting data at Figure 4H. n = 3. *, p < 0.05, compared to siControl.

https://doi.org/10.7554/eLife.16030.016
Figure 2—figure supplement 2
The chromatin conformation of Tnni1 is independent of Nkx2-5.

(A) Genome browser representation of strand-specific RNA-seq tag counts from eCMs transfected with the indicated siRNAs. neg., negative strand; pos., positive strand. (B) Chromatin conformation …

https://doi.org/10.7554/eLife.16030.017
Figure 2—figure supplement 3
siRNA knockdown efficiencies of three different siRNAs for each gene.

(A) eCMs were transfected with three different siRNAs for each gene, and the expression level of Nkx2-5, Gata4, and Tbx5 was measured by real-time PCR. Expression values were normalised against Rplp2

https://doi.org/10.7554/eLife.16030.018
Figure 2—figure supplement 4
Functional annotations of genes with increased and decreased mRNA with long 3’UTR.

(A) The knockdown of Nkx2-5, Gata4, and Tbx5 in eCMs was analysed by Western blotting. siNkx2-5, Nkx2-5 siRNA; siTbx5, Tbx5 siRNA; siGata4, Gata4 siRNA; WT, wild type. (B, C) Enriched gene …

https://doi.org/10.7554/eLife.16030.019
Nkx2-5 associates with the 5’-3’ exonuclease Xrn2.

(A) Co-immunoprecipitates derived using the indicated antibodies from nuclear extracts of E12.5 hearts and aliquots (6%) of the input proteins were analyzed by Western blotting. (B) Xrn2 and Nkx2-5 …

https://doi.org/10.7554/eLife.16030.020
Figure 4 with 4 supplements
Nkx2-5 functions together with Xrn2 to regulate APA.

(A) Xrn2 knockdown was analyzed by Western blotting. (B and C) qRT-PCR analysis of mRNAs expression of the long 3’ UTRs (B) and gene bodies (C) of Tnnt2 and Atp2a2 in Xrn2-knockdown eCMs, normalized …

https://doi.org/10.7554/eLife.16030.021
Figure 4—figure supplement 1
Knockdowns of Nkx2-5 or Xrn2 affect the expression of the long 3’UTR regions in chromatin-fractioned RNA.

(A) Genome browser representation of strand-specific chromatin-fractioned RNA-seq tag counts from eCMs transfected with the indicated siRNAs. (B) qRT-PCR analysis of mRNA expression of long 3’ UTRs …

https://doi.org/10.7554/eLife.16030.023
Figure 4—figure supplement 2
Knockdown of Nkx2-5 affects Xrn2-binding to Myl7.

Xrn2 binding in eCMs transfected with the indicated siRNAs was analyzed by ChIP-qPCR. The control values were set to 1.0. Error bars indicate the mean ± s.e.m. (n = 3). *, p < 0.05.

https://doi.org/10.7554/eLife.16030.024
Figure 4—figure supplement 3
Nkx2-5 functions together with Xrn2 to regulate APA.

The EtBr staining gels and the blottings of b-actin as an internal control are shown, related to Figure 4F.

https://doi.org/10.7554/eLife.16030.025
Figure 4—figure supplement 4
Correlation analysis of Nkx2-5 binding and long 3’UTR expression.

(A) The average profiles of the mRNAs with long 3’ UTRs in genes that Nkx2-5 binds to TTS (3997 genes) in eCMs transfected with the indicated siRNAs. The gray area indicates the coding region. …

https://doi.org/10.7554/eLife.16030.026
Knockdown of both Nkx2-5 and Xrn2 perturbs heart looping.

(A) Transfection of siRNA into embryonic hearts. GFP was used to detect transfected fields. (B) We discarded the embryos with low transfection efficiency. (C) Representative morphologies of heart …

https://doi.org/10.7554/eLife.16030.027
Figure 6 with 1 supplement
Nkx2-5 genetically interacts with Xrn2.

Histological analysis of Nkx2-5+/-and Xrn2+/- newborn hearts. Frontal sections from newborn hearts were stained with hematoxylin and eosin. ASD was observed in Xrn2+/-(n = 6 of 9) and Nkx2-5+/-Xrn2+/…

https://doi.org/10.7554/eLife.16030.028
Figure 6—figure supplement 1
Nkx2-5 genetically interacts with Xrn2.

(A) Scheme illustrating the targeting of exon 1 and 2 in Xrn2. Two gRNAs and the Cas9 RNA were injected into fertilized eggs. F, Forward primer; WT-R, reverse primer for the wild-type allele; MT-R, …

https://doi.org/10.7554/eLife.16030.029
Author response image 1
RNA-immunoprecipitation analysis of Xrn2 in eCMs.

Relative Xm2-binding to mRNA at the downstream regions (-3) of Tnnt2 (18.4 kb) and Atp2a2 (47.2 kb)was analyzed by RIP in eCMs transfected the indicated siRNAs. The mRNA of long 3’ UTRs of Tnnt2 and …

https://doi.org/10.7554/eLife.16030.031

Additional files

Supplementary file 1

Lists of mapping results, primers, and antibodies.

(A) Mapping results of ChIP-seq . (B) Mapping results of RNA-seq. (C) Primers used in this study. (D) Antibodies used in this study.

https://doi.org/10.7554/eLife.16030.030

Download links