Emerman AB and Blower MD (2018), The RNA-binding complex ESCRT-II in Xenopus lae...

XB-ART-55012

J Biol Chem 2018 Aug 10;29332:12593-12605. doi: 10.1074/jbc.RA118.003718.

Show Gene links Show Anatomy links

The RNA-binding complex ESCRT-II in Xenopus laevis eggs recognizes purine-rich sequences through its subunit, Vps25.

Emerman AB , Blower MD .

???displayArticle.abstract???
RNA-binding proteins (RBP) are critical regulators of gene expression. Recent studies have uncovered hundreds of mRNA-binding proteins that do not contain annotated RNA-binding domains and have well-established roles in other cellular processes. Investigation of these nonconventional RBPs is critical for revealing novel RNA-binding domains and may disclose connections between RNA regulation and other aspects of cell biology. The endosomal sorting complex required for transport II (ESCRT-II) is a nonconventional RNA-binding complex that has a canonical role in multivesicular body formation. ESCRT-II was identified previously as an RNA-binding complex in Drosophila oocytes, but whether its RNA-binding properties extend beyond Drosophila is unknown. In this study, we found that the RNA-binding properties of ESCRT-II are conserved in Xenopus eggs, where ESCRT-II interacted with hundreds of mRNAs. Using a UV cross-linking approach, we demonstrated that ESCRT-II binds directly to RNA through its subunit, Vps25. UV cross-linking and immunoprecipitation (CLIP)-Seq revealed that Vps25 specifically recognizes a polypurine (i.e. GA-rich) motif in RNA. Using purified components, we could reconstitute the selective Vps25-mediated binding of the polypurine motif in vitro Our results provide insight into the mechanism by which ESCRT-II selectively binds to mRNA and also suggest an unexpected link between endosome biology and RNA regulation.

???displayArticle.pubmedLink??? 29903915
???displayArticle.pmcLink??? PMC6093225
???displayArticle.link??? J Biol Chem
???displayArticle.grants??? [+]

Species referenced: Xenopus laevis
Genes referenced: ctr9 isyna1 snf8 vps25 vps28 vps36
GO keywords: ESCRT II complex [+]

???attribute.lit??? ???displayArticles.show???

	Figure 1. ESCRT-II is an RNA-binding complex in Xenopus laevis eggs A) Ethidium bromide-stained agarose gel demonstrating the composition of total RNA and RNA that co-immunoprecipitates with anti-ESCRT-II antibodies or nonspecific rabbit IgG (mock). Equal masses of total RNA and ESCRT-II RNA samples and equal volumes of ESCRT-II and mock RNA samples were loaded. B) Scatterplot demonstrating the normalized sequencing reads of RNA isolated from an ESCRT-II immunoprecipitation (IP) or total RNA from one replicate. Points highlighted in red are two-fold enriched in the ESCRT-II IP over total egg extract RNA with a p-value<0.01 (calculated using two biological replicates and the edgeR exact test, based on the quantile-adjusted conditional maximum likelihood model). Blue line depicts equal values in both samples. C) Histogram of ESCRT-II enrichment. Blue line indicates mean. D) Correlation of ESCRT-II enrichment (RPKM in ESCRT-II IP/Total) across two biological replicates. The Pearson correlation coefficient is 0.92. Red points are mRNAs that are significantly enriched in ESCRT-II IPs. Blue line depicts equal values in both samples. E) RTqPCR validation of ESCRT-II enrichment (abundance in ESCRT-II immunoprecipitations/abundance in total RNA) of transcripts identified by RIP-seq to be ESCRT-II associated. 28s rRNA or transcripts predicted by RIP-seq to be underrepresented in ESCRT-II immunoprecipitations were used as negative controls. Error bars represent SEM from three biological replicates.
	Figure 2. ESCRT-II binds directly to RNA through Vps25 in Xenopus egg extracts A) Western blot of ESCRT-II immunoprecipitations (IPs) or mock IPs (nonspecific rabbit IgG) performed under native (-) or denaturing (+) conditions. Vps22 is not detectable by Western blot with our ESCRT-II polyclonal antibody. H.C.=heavy chain. B) Autoradiograph of a UV crosslinking and immunoprecipitation (CLIP) experiment from Xenopus egg extract under high RNase conditions (0.1 mg/mL). A radioactive band consistent with the molecular weight of Vps25 (denoted by the red asterisk) is observed, while no bands at the molecular weights of Vps22 or Vps36 are apparent. The expected migrations of the ESCRT-II subunits are indicated to the left of the gel. T4 RNA ligase forms a covalent intermediate with pCp (used to radiolabel the RNA fragments) and appears in every lane. The bands above and below the Vps25 band (black asterisks) are nonspecific, as they appeared in the IgG control in some replicates of this experiment. H.C.= heavy chain. C) Autoradiograph of a CLIP experiment from Xenopus egg extract performed as described in (B), except under denaturing immunoprecipitation conditions. The same polyclonal ESCRT-II antibody was used for (A-C), but under denaturing conditions this antibody only immunoprecipitates Vps25.
	Figure 3. CLIP Analysis of ESCRT-II bound RNAs A) Genome browser views of CLIP read clusters for two transcripts. Inset illustrates a higher resolution view of aligned CLIP reads. 114 transcripts in the ESCRT-II CLIP-seq library had overlapping unique reads. B) Cumulative distribution plot demonstrating the fraction of RIP-seq transcripts with ≥2 CLIP-tags (red) or the entire ESCRT-II RIP-seq library as a function of their enrichment in the ESCRT-II RIP-seq library relative to total RNA. P-value for the difference between the two distributions is <0.0014 using a two-sided, unpaired t-test. C) RT-qPCR validation of ESCRT-II enrichment (abundance in ESCRT-II immunoprecipitation/abundance in total RNA) of transcripts containing CLIP-tags compared to a negative control RNA (28S rRNA). Two biological replicates are shown. D) MEME motif analysis (36) was used to analyze the sequence motifs present in ESCRT-II CLIP tags. The top-scoring motif is shown. E) Cumulative distribution plot demonstrating the fraction of RIP-seq transcripts containing a GArich motif (a 30 nt stretch of at least 90% purines) (red) or the entire ESCRT-II RIP-seq library (black) as a function of their enrichment in the ESCRT-II RIP-seq library. P-value for the difference between the two distributions is <1 x 10-12 using a two-side unpaired t-test.
	Figure 4. Analysis of ESCRT-II/RNA binding in vitro. A) Coomassie blue-stained gel of the recombinant Xenopus (Xen) and human (Hu) ESCRT-II complexes used in the in vitro RNA-binding assays. ∆MBD lacks the membrane-binding domains of human ESCRT-II. B-D) Autoradiographs of UV-crosslinked, in vitro binding reactions with: B) Xenopus ESCRT-II and 5’ end-labeled total egg RNA, C) Xenopus ESCRT-II and individual 5’ end-labeled in vitro transcribed RNAs that are underrepresented in ESCRT-II immunoprecipitations, or D) HuESCRT-II (FL) or HuESCRT-IIΔMBD (ΔMBD) and a bodylabeled, in vitro transcribed GA-rich CLIP-tag (a region of the ctr9 mRNA). B-C) A covalent intermediate of PNK and [γ-32P]-ATP (used to radiolabeled the RNA fragments) is indicated. D) Folch fraction liposomes were included in the binding reactions at the indicated concentrations. A fluorescent western blot (WB) of the same nitrocellulose membrane shown in the autoradiograph is shown as a loading control. Asterisk represents a nonspecific band. A-D) The expected migrations of the ESCRT-II subunits are indicated. E) Quantification of the autoradiograph in Figure (D) and two additional, independent replicates, depicting the fraction of RNA bound by each ESCRT-II subunit at the indicated concentrations of Folch fraction liposomes relative to binding with no liposomes present. Error bars are the SEM.
	Figure 5. ESCRT-II specifically binds GA-rich RNAs in vitro. A) Sequences of transcripts used for in vitro binding reactions. Nucleotides highlighted in red are mutations made to the ctr9 CLIP-tag. B) Autoradiograph of UV-crosslinked in vitro binding reactions as in Figure 4, performed with recombinant HuESCRT-II (FL) or HuESCRT-II ∆MBD (∆MBD) and the in vitro transcribed, body-labeled ctr9 CLIP-tag or the ctr9 CLIP-tag with the adenosines mutated. C) RNA EMSA with ESCRT-II ∆MBD at the indicated concentrations and the ctr9 CLIP-tag or a negative control sequence of the same length (a region of the smu1 transcript). D) Quantification of three independent RNA EMSA experiments using ESCRT-II ∆MBD and the ctr9 CLIP-tag. Error bars are SE. E) RNA EMSA reactions with ESCRT-II ∆MBD at the indicated concentrations and the ctr9 CLIP-tag containing or lacking mutations in the majority of guanosines or adenosines. The fraction of RNA bound by ESCRT-II ∆MBD is indicated below each lane.

References [+] :

Alam, Structural basis for ubiquitin recognition by the human ESCRT-II EAP45 GLUE domain. 2006, Pubmed