MKRN3-mediated ubiquitination of Poly(A)-binding proteins modulates the stability and translation of GNRH1 mRNA in mammalian puberty

Abstract

The family of Poly(A)-binding proteins (PABPs) regulates the stability and translation of messenger RNAs (mRNAs). Here we reported that the three members of PABPs, including PABPC1, PABPC3 and PABPC4, were identified as novel substrates for MKRN3, whose deletion or loss-of-function mutations were genetically associated with human central precocious puberty (CPP). MKRN3-mediated ubiquitination was found to attenuate the binding of PABPs to the poly(A) tails of mRNA, which led to shortened poly(A) tail-length of GNRH1 mRNA and compromised the formation of translation initiation complex (TIC). Recently, we have shown that MKRN3 epigenetically regulates the transcription of GNRH1 through conjugating poly-Ub chains onto methyl-DNA bind protein 3 (MBD3). Therefore, MKRN3-mediated ubiquitin signalling could control both transcriptional and post-transcriptional switches of mammalian puberty initiation. While identifying MKRN3 as a novel tissue-specific translational regulator, our work also provided new mechanistic insights into the etiology of MKRN3 dysfunction-associated human CPP.

Figure 1.

MKRN3 forms complex with poly(A)-binding protein (PABP) family members. (A) Identification of MKRN3-interacting proteins through co-immunoprecipitation followed by mass spectrometry analysis. Volcano plots depicting the proteins enrichment from co-immunoprecipitation experiments, and the red circles represent proteins enriched >8-fold by MKRN3 compare to empty vectors. HEK293T cells were ectopically expressing 3xFlag-tagged MKRN3 or Flag tagged empty vectors. Forty-eight hours later, cells lysates treated with or without RNaseA (50 μg/ml), were incubated with anti-Flag affinity gels, and the co-immunoprecipitates were subjected to trypsin digestion followed by mass spectrometry analysis. MKRN3, PABP family members including PABPC1, PABPC3 and PABPC4, and previously reported MKRN3 interacting protein MBD3 indicated in blue colour. Three samples each group. (B) Domain structures of the PABP family members, including PABPN1, PABPC1, PABPC3 PABPC4 and PABPC5, which commonly share RNA-binding RRM (RNA recognition motif). PABC, poly(A)-binding protein C-terminal domain. (C–F) 3XMyc tagged MKRN3 protein was co-immunoprecipitated with Flag-tagged PABPC1 (C), PABPC3 (D) or PABPC4 (E) when co-expressed in HEK293T cells. Cell lysates treated with or without RNaseA (50 μg/ml) were subjected to co-immunoprecipitation assay using anti-Myc antibody, followed by immunoblotting with anti-Myc or anti-Flag antibody. (F) Endogenous Mkrn3 and PABPC1 formed a complex in the hypothalamus of wild-type mouse. Endogenous PABPC1 proteins of hypothalamic lysates from wild-type male mice at postnatal day 15 treated with or without RNaseA were immunoprecipitated using anti-PABPC1, followed by immunoblotting with indicated antibodies.

Figure 2.

MKRN3 directly interacts with the PABC domain of PABP family members. (A–D) Recombinant MKRN3 directly interacted with PABPC1 (A), PABPC3 (B), or PABPC4 (C), but not PABPC5 (D), as revealed by GST pull-down assay. GST pull-down assays were performed with recombinant Gst-tagged MKRN3 and His6-tagged PABPC1, PABPC3, PABPC4 or PABPC5. The recovered Gst tagged proteins were resolved in SDS-PAGE and visualized with Commassie Blue gel staining, while the PABP proteins were detected by immunoblotting with anti-His antibody. PD, GST pull-down. (E) The PABC domain of PABPC1 directly interacted with the middle region (126–295aa) of MKRN3. The residue numbers were denoted underneath each schematic structural region of the proteins. GST pull-down assays were performed with recombinant Gst-MKRN3 or its fragments, and PABPC1-His6 or its fragments. PD, GST pull-down.

Figure 3.

MKRN3 ubiquitinates PABP family members that contain PABC domain and CPP-associated mutations of MKRN3 comprised this ubiquitination. (A) MKRN3 ubiquitinated PABPC1, PABPC3 and PABPC4, but not PABPC5 in vitro. An in vitro ubiquitination assay was carried out using the recombinant Flag tagged PABPC1, PABPC3, PABPC4 or PABPC5, His6-tagged UBA1 (E1) and UBCH5A (E2), and Gst-tagged MKRN3 (E3) or the E3 ligase dead mutant MKRN3 (C340G), together with the indicated components. Usp2cc, the catalytic core of human deubiquitinase Usp2; PABPCX represents PABPC1, PABPC3, PABPC4 or PABPC5. (B) Flag-tagged PABPC1 was efficiently ubiquitinated by 3XMyc-tagged MKRN3. MKRN3^−/− HEK293T cells were co-transfected with indicated plasmids, and cell lysates were immunoprecipitated with anti-HA affinity gels, treated with Usp2cc or not, followed by immunoblotting with anti-Flag antibody to detect ubiquitinated PABPC1. (C) The ubiquitination of PABPC1 was reduced in Mkrn3^m+/p− mouse hypothalamus at postnatal day 15, compared to those from the age-matched wild-type littermates (n = 3). Endogenous PABPC1 proteins of hypothalamic lysates from wild-type or Mkrn3^m+/p− mice were immunoprecipitated using anti-PABPC1 antibody, followed by immunoblotting with anti-Ub or other indicated antibodies. (D) CPP-associated mutations of MKRN3 compromised its E3 ligase activity towards PABPC1 in MKRN3^−/− HEK293T cells. Cells were transfected with HA-tagged Ub, Flag-tagged PABPC1, and 3xMyc-tagged wild-type MKRN3 or the indicated mutants (CPP-associated mutations, C340G, R365S, P417I and H420C). Cell lysates were immunoprecipitated with anti-HA affinity gels, followed by immunoblotting with anti-Flag antibody to detect ubiquitinated PABPC1 and other antibodies indicated.

Figure 4.

MKRN3 mediates non-proteolytic ubiquitination of PABPC1 at multiple sites. (A) Schematic distribution of the eight sites (Lys residues) for MKRN3-mediated ubiquitination on human PABPC1 identified by mass spectrometry analysis. PABPC1 proteins were recovered from the in vitro ubiquitination assay and subjected to trypsin digestion, followed by mass spectrometry analysis to map the ubiquitination sites. (B) Four Lys residues (K312, K512, K620 and K625, shown red colour in A) were shown to be the major sites for MKRN3-mediated ubiquitination of PABPC1. Lysates of MKRN3^−/-HEK293T cells ectopically expressing HA tagged Ub, 3XMyc tagged MKRN3, and Flag tagged PABPC1 or the indicated K-to-R mutants. Cell lysates were immunoprecipitated with anti-HA affinity gels, followed by immunoblotting analysis using anti-Flag to detect the ubiquitinated PABPC1 and other antibodies indicated. (C) Simultaneous four K-to-R mutation at the major ubiquitination sites almost totally abolished the ubiquitination of PABPC1 mediated by MKRN3. MKRN3^−/-HEK293T cells were ectopically expressing HA tagged Ub, 3XMyc tagged MKRN3 and Flag-tagged PABPC1 or the mutants bearing the indicated K-to-R mutations. Cell lysates were immunoprecipitated with anti-HA affinity gels, followed by immunoblotting using anti-Flag to detect the ubiquitinated PABPC1. 2KR, K312–512R; 3KR, K312–512-620R; 4KR, K312–512-620–625R. (D) MKRN3 ubiquitinated PABPC1 with non-proteolytic K27 and k29 ubiquitin linkages. MKRN3^−/–HEK293T cells were ectopically co-expressing 3xMyc-tagged MKRN3, Flag-tagged PABPC1 and HA tagged Ub (WT, or mutated at Lys6, Lys11, Lys27, Lys29, Lys33, Lys48 or Lys63 only) in the indicated combinations. Cell lysates were were immunoprecipitated with anti-HA affinity gels, followed by immunoblotting using anti-Flag to detect the ubiquitinated PABPC1 and other antibodies indicated.

Figure 5.

The ubiquitination of PABPs by MKRN3 attenuates the stability of GNRH1 mRNA. (A) Volcano plots depicting MKRN3 up-regulated and down-regulated genes detected by RNA-seq. Hypothalamus-derived GT1–7 cells were transfected with empty or MKRN3 expressing vectors, and RNA-seq were performed 48 h later, three samples each group. The red circles represent genes that are up-regulated by MKRN3 more than eight times, the blue circles represent genes that are down-regulated by MKRN3 more than 8 times and the green circles represent histone mRNAs. The yellow circles represent the top five up-regulated and down-regulated genes. (B) MKRN3 destabilized Gnrh1 mRNA in GT1–7 cells. Cells were transfected with empty vector, or vector expressing MKRN3 or MKRN3 (C340G), and treated with Actinomycin D (5 μg/mL) to block de novo transcription of Gnrh1 at different time points (0, 2, 4 or 6 h) after 36 h transfection. Total mRNAs were extracted and subjected to quantitative PCR (qPCR) with Gnrh1 signals normalized to that of Gapdh. Each sample was normalized to 100% at time zero, and the half-lives of Gnrh1 mRNA (t_1/2) were calculated. Data were presented as mean ± SD, three independent experiments. MKRN3 versus vector, *P < 0.05, significant difference; **P < 0.01, very significant difference. (C) Luciferase reporter assay to detect the relative activities of human GNRH1 (UTR). A luciferase reporter that contains 3′- and 5′-UTR (untranslated region) of human GNRH1 were constructed, named as PGL3-GNRH1(UTR)-Luc, and its activities were detected in wild-type or MKRN3^−/−HEK293T cells that ectopically expressing empty vector, MKRN3 or MKRN3 (C340G). Data were presented as mean ± SD, one-way ANOVA with Bonferroni post-hoc test. **P < 0.01, very significant difference, three independent experiments. (D) CPP-associated mutations compromised the suppressive effect of MKRN3 on GNRH1 (UTR) luciferase activity. MKRN3^−/– HEK293T cells were co-transfected with plasmids encoding GNRH1 (UTR) luciferase, PABPC1, wild-type MKRN3 or the indicated mutants. Data were presented as Mean ± SD, one-way ANOVA with Bonferroni post-hoc test. **P < 0.01, very significant difference; NS, no significant difference, three independent experiments. (E) MKRN3-mediated ubiquitination impaired PABPC1-activated GNRH1 (UTR) luciferase activity. MKRN3^−/− HEK293T cells were co-transfected with vectors encoding GNRH1 (UTR) luciferase, MKRN3 and PABPC1 or PABPC1_4KR. Data were presented as mean ± SD, one-way ANOVA with Bonferroni post-hoc test. **P < 0.01, very significant difference; NS, no significant difference, three independent experiments. (F) Gnrh1 mRNA was more stable in the neurons of Mkrn3^m+/p− mouse than that of wild-type. Neurons were isolated from the hypothalamus of wild-type or Mkrn3^m+/p− fetal male mice at embryonic day 18, subsequently cultured for 15 days, and treated with Actinomycin D (5 μg/ml) at different time points (0, 2, 4 or 6h). Total mRNAs were extracted and subjected to qPCR with Gnrh1 signals normalized to that of Gapdh. Each sample was normalized to 100% at time zero, and the half-lives of Gnrh1 mRNA (t_1/2) were calculated. Data were presented as mean ± SD, three independent experiments. **P < 0.01, very significant difference.

Figure 6.

The ubiquitination of PABPs by MKRN3 comprises its binding to the poly(A) of mRNA and the formation of translation initiation complex (TIC). (A) More PABPC1 and PABPC4 proteins were recovered with poly(A)-tailed mRNAs in the hypothalamus of Mkrn3^m+/p- mice than those from age-matched wild-type littermates. The total ploy(A)-tailed mRNAs were recovered with poly (U) pull-down assay, in which lysates of hypothalamic tissues were precipitated using poly(U) agarose beads (see Materials and methods for details). The poly(A)-bound PABPC1 and PABPC4 were detected by immunoblotting using indicated antibodies, three mice were used in each group. (B) MKRN3 mediated the ubiquitination of PABPC1 compromised its binding to (A)₃₀ RNA as detected by RNA EMSA. MKRN3^−/− HEK293 cells were co-transfected with HA-tagged Ub, Flag-tagged PABPC1, and 3xMyc-tagged MKRN3 or not. 24 hours later, cell lysates were immunoprecipitated using anti-Flag affinity gels, treated with Usp2cc or not, and eluted with Flag peptides before subjected to RNA EMSA analysis (see Materials and methods for details). Usp2cc, the catalytic core of human deubiquitinase Usp2. (C) MKRN3-mediated ubiquitination compromised the affinity of PABPC1 to (A)₃₀ RNA. EMSA assay was performed using (A)₃₀ RNA probe that labeled with γ-³²P. Recombinant PABPC1 was subjected to in vitro ubiquitination under varying indicated conditions, then immunoprecipitated with anti-Flag affinity gels before subjected to the EMSA assay. (D) Ablation of endogenous Mkrn3 promoted the binding of endogenous PABPC1 to Gnrh1 mRNA in the hypothalamus of Mkrn3^m+/p− mice compared to that of wild-type ones. RNA immunoprecipitation assay was done with the hypothalamus of wild-type and Mkrn3^m+/p− male mice at post-natal day 15, and PABPC1-bound mRNAs were enriched with anti-PABPC1 antibodies followed by qPCR assays using primers f1+r1. Data were presented as Mean ± SD, two-tailed unpaired t-test, n = 5. **P < 0.01, very significant difference. (E) The poly(A) tail-length of Gnrh1 mRNAs in the hypothalamus of Mkrn3^m+/p− mice were longer than those in wild-type. Total RNAs of hypothalamic tissues were prepared from male mice at post-natal day 15, followed by addition of a limited number of guanosine and inosine residues to the 3′-ends of poly(A)-containing RNAs, to allow reverse transcription-PCR (RT-PCR) amplification using primer R2 (see Materials and methods for details). S, gene-specific amplicons to tell the abundance of Gnrh1 mRNA, using primers F+R1; A, poly(A)-tail-containing amplicons to tell the poly(A) tail-length of Gnrh1 mRNA, using primers F+R2.*, gene-specific bands; **, poly(A) tail-length bands, three samples in each group. (F) Ablation of endogenous Mkrn3 promoted PABPC1-EIF4G1 interaction in the hypothalamus of mice. Co-immunoprecipitation assay were performed with the hypothalamus of wild-type and Mkrn3^m+/p− mice at post-natal day 15 using anti-IgG or anti-PABPC1 antibody, treated with RNaseA (50 μg/ml) or not, followed by immunoblotting with indicated antibodies, three samples in each group. (G) A model depicting how the MKRN3-PABPC1 axis controls the post-transcriptional switch of mammalian puberty initiation through regulating the stability and translation of mRNAs (including GNRH1) in hypothalamic cells.