The RNA Surveillance Factor UPF1 Represses Myogenesis via Its E3 Ubiquitin Ligase Activity

Abstract

UPF1 is an RNA helicase that orchestrates nonsense-mediated decay and other RNA surveillance pathways. While UPF1 is best known for its basal cytoprotective role in degrading aberrant RNAs, UPF1 also degrades specific, normally occurring mRNAs to regulate diverse cellular processes. Here we describe a role for UPF1 in regulated protein decay, wherein UPF1 acts as an E3 ubiquitin ligase to repress human skeletal muscle differentiation. Suppressing UPF1 accelerates myogenesis, while ectopically increasing UPF1 levels slows myogenesis. UPF1 promotes the decay of MYOD protein, a transcription factor that is a master regulator of myogenesis, while leaving MYOD mRNA stability unaffected. UPF1 acts as an E3 ligase via its RING domain to promote MYOD protein ubiquitination and degradation. Our data characterize a regulatory role for UPF1 in myogenesis, and they demonstrate that UPF1 provides a mechanistic link between the RNA and protein decay machineries in human cells.

Keywords: E3 ligase; RING domain; UPF1; myogenesis; nonsense-mediated decay; ubiquitin-proteasome.

Figure 1. UPF1 knockdown accelerates myoblast differentiation

(A) Schematic of a time course of human myoblast differentiation following UPF1 knockdown (KD). Two genetically distinct human myoblast cell lines (54-1 and MB135 cells) were transfected with two different siRNAs against UPF1, as well as two different non-targeting siRNAs as controls (Day -2). When the transfected cells reached full confluency (Day 0), differentiation was induced by switching from high-serum growth media to low-serum differentiation media (DM). 54-1 and MB135 cells respectively differentiate less and more rapidly, and so were differentiated for seven and four days, respectively. (B) Immunoblot for UPF1 protein from 54-1 cells following control or UPF1 KD immediately prior to (Day 0) or two days after (Day 2) induction of differentiation. Equal amounts of protein were loaded based on the BCA assay. α-tubulin, loading control. Bar plot, quantification of UPF1 protein levels relative to the loading control. (C) Immunofluorescence labeling of 54-1 cells with an antibody against BrdU (red) at the indicated time points, prior to significant cell fusion. At each time point, cells were fixed after one hour of BrdU labeling. Box plot, percentage of BrdU+ nuclei measured over 10 fields. Whiskers, max and min over the fields. */**/***, p < 0.05/0.01/0.001. (D-E) Immunofluorescence labeling of 54-1 or MB135 cells with antibodies against Myogenin (MYOG, red) and Myosin Heavy Chain (MHC, green) at the indicated time points. MB135 differentiate more rapidly than do 54-1 cells and so a shorter time course was used. Box plot, percentage of MYOG+ nuclei measured over 10 fields. Whiskers, max and min over the fields. */**/***, p < 0.05/0.01/0.001. See also Figure S1.

Figure 2. UPF1 overexpression slows myoblast differentiation

(A) Schematic illustrating a time course of human myoblast differentiation following the induction of UPF1 overexpression in MB135 myoblasts. Transgenic UPF1 was induced or not induced 12 hours prior to the induction of differentiation. (B) Levels of UPF1 mRNA at Day 0, 12 hours after the addition of Dox to induce UPF1 expression. Error bars, standard deviation. (C) Immunoblot for total and transgenic UPF1 protein at the same time point as in (B), measured using antibodies against UPF1 and FLAG. H3, loading control histone H3. (D) Immunofluorescence labeling with an antibody against BrdU (red) at Day 0. Cells were fixed and labeled after incubation with BrdU-containing media for one hour. Box plot, percentage of BrdU+ nuclei measured over eight fields. Whiskers, max and min over the fields. */**/***, p < 0.05/0.01/0.001. (E) Immunofluorescence labeling with antibodies against Myogenin (MYOG, red) and Myosin Heavy Chain (MHC, green) at Day 2. Box plot, percentage of MYOG+ nuclei measured over 10 fields. Whiskers, max and min over the fields. */**/***, p < 0.05/0.01/0.001.

Figure 3. UPF1 knockdown promotes a myogenic gene expression program, including MYOD-specific targets

(A) Schematic illustrating when RNA was collected for RNA-seq (Day 2) during differentiation of 54-1 myoblasts following control or UPF1 KD. (B) Gene Ontology (GO) enrichment analysis of genes that were up-regulated by ≥1.5-fold following UPF1 versus control KD at Day 2. (C) Relative mRNA levels of genes that are specifically activated by MYOD and not other myogenic factors (red) (Conerly et al., 2016; Ishibashi et al., 2005), as well as MYOD mRNA itself (purple), following UPF1 versus control KD at Day 2. The illustrated genes exhibited increases in expression of ≥1.5-fold. (D) Immunoblots for UPF1 and MYOD proteins one day following transfection with a control or UPF1-targeting siRNA (Day -1, or equivalently Hour -24), one day prior to the induction of differentiation. H3, loading control histone H3. Bar plots, quantification of UPF1 and MYOD relative to the loading control.

Figure 4. UPF1 knockdown induces MYOD protein in the absence of MYOD mRNA up-regulation

(A) Schematic illustrating time points for sample collection following UPF1 KD (Day -2 to 0, or equivalently Hour -48 to 0) in 54-1 myoblasts. (B) Immunoblot for MYOD protein in the 24 hours immediately following transfection with a control or UPF1-targeting siRNA (Day -2 to Day -1, or equivalently Hour -48 to -24). H3, loading control histone H3. Bar plot, quantification of MYOD protein relative to the loading control. Red arrow indicates when MYOD protein levels detectably increased (between Hour -30 and -27). (C) Relative levels of MYOD mRNA in the 48 hours immediately following transfection with a control or UPF1-targeting siRNA (Day -2 to Day 0, or equivalently Hour -48 to 0). Red arrow indicates when MYOD protein levels detectably increased (see B). */**/***, p < 0.05/0.01/0.001. (D) Estimates of MYOD mRNA half-lives at Day 0 (Hour 0) in control- or UPF1-KD cells. MYOD mRNA levels were measured 0, 0.5, 1, 2, 4, 8 and 12 hours after the addition of actinomycin D (ActD, 2.5μg/mL) to inhibit transcription. (E) Relative levels of MYOD pre-mRNA (left) and mature mRNA (right) at Day 0 (Hour 0) in control- or UPF1-KD cells, normalized to MYOD pre-mRNA or mature mRNA levels in control KD samples. Error bars, error propagation computed with the balanced repeated replication (BRR) method. */**/***, p < 0.05/0.01/0.001. See also Figure S2.

Figure 5. UPF1 promotes proteasome-dependent degradation of MYOD protein

(A) Schematic illustrating a time course of differentiation following the induction of UPF1 overexpression in MB135 myoblasts. Samples were collected during the 12 hours immediately following Dox treatment (Hours -12 to 0). (B) Immunoblot for total UPF1, FLAG-tagged transgenic UPF1, and MYOD proteins, measured during the 12 hours following Dox treatment. H3, loading control histone H3. Bar plot, quantification of total UPF1 and MYOD proteins relative to the loading control. (C) Immunoblot for total UPF1 and MYOD proteins, measured 12 hours after Dox treatment, in cells treated with the proteasome inhibitor MG132 (10 μM, eight-hour treatment). H3, loading control histone H3. Bar plot, quantification of MYOD protein relative to the loading control.

Figure 6. UPF1’s RING domain has E3 ubiquitin ligase activity

(A) Schematic of UPF1’s protein domain structure (Kadlec et al., 2006; UniProt Consortium, 2012). Orange, RING-like domain. Purple, RNA helicase domain. (B) Protein structure of UPF1’s RING domain 1 (orange, residues 121-172 from PDB 2WJV; (Clerici et al., 2009)), illustrating two zinc fingers coordinating zinc ions (grey spheres), the canonical E2-RING E3 interaction pocket formed by Loop1 and Loop2 (green) which faces inside, and the UPF2 (cyan) binding surface on the periphery. Structure was visualized with PyMOL (Schrödinger, LLC, n.d.). (C) Top, schematic diagram of UPF1’s RING domain 1 (orange, residues 121-172 from PDB 2WJV). The first zinc ion (grey circle) is held by a CCCH motif and the second zinc ion is held by a CSCH motif. Green, Loop1 and Loop2. Bottom, estimated amino acid sequence conservation of Loop1 and Loop2 (residues 123-140). Asterisks indicate the residues S124, N138, and T139 that we selected for mutagenesis. (D) Top, schematic of construct to enable Dox-inducible expression of the mutant UPF1^{S124A/N138A/T139A}. This construct was used to generate the clonal myoblast cell line MB135-Tet-UPF1^{S124A/N138A/T139A}. Bottom, immunoblot for total and transgenic UPF1 protein from MB135-Tet-UPF1^WT and MB135-Tet-UPF1^{S124A/N138A/T139A} cells 12 hours following transgene induction with Dox. H3, loading control histone H3. (E) Schematic of co-immunoprecipitation (co-IP) experiments with cell lysates from MB135-Tet-UPF1^WT or MB135-Tet-UPF1^{S124A/N138A/T139A} cells. Induced or uninduced cells were treated with MG132 for six hours to inhibit the proteasome. IP eluates from each pull-down were then probed with antibodies against UPF1 and MYOD. Ub, ubiquitin. (F) Immunoblots of Input and IP eluates from (E). Left, Input total lysates were probed for UPF1, MYOD and Ubiquitin. Right, from top to bottom, IP eluates from the anti-UPF1, anti-MYOD, and anti-Ubiquitin pull-downs were probed for UPF1 and MYOD. (G) As (E), but induced cells were treated or not treated with MG132 for six hours to inhibit the proteasome. IP eluates from the anti-MYOD pull-down were probed with antibodies against UPF1 and Ubiquitin. (H) Immunoblots of Input and IP eluates from (G). Left, Input total lysates were probed for UPF1, MYOD and Ubiquitin. Right, IP eluates from the anti-MYOD pull-down were probed for UPF1, MYOD and Ubiquitin. Bar plots, quantification of UPF1 (top) or MYOD-Ub (bottom) in co-IP with MYOD relative to the level of MYOD IP. See also Figure S3.

Figure 7. UPF1 RING domain mutations stabilize MYOD protein and promote myogenesis

(A) Immunoblot for MYOD protein, measured throughout a time course following cycloheximide (CHX, 100 μg/mL) treatment to inhibit translation. Identical numbers of induced or uninduced MB135-Tet-UPF1^WT or MB135-Tet-UPF1^{S124A/N138A/T139A} cells were used. H3, loading control histone H3, which has a long half-life (Toyama et al., 2013). Lower blot, levels of induced UPF1 ^WT and UPF1^{S124A/N138A/T139A} proteins. (B) Immunofluorescence labeling of induced or uninduced MB135-Tet-UPF1^WT or MB135-Tet-UPF1^{S124A/N138A/T139A} cells with antibodies against Myogenin (MYOG, red) and Myosin Heavy Chain (MHC, green) at Day 2. Box plot, percentage of MYOG+ nuclei measured over 10 fields. Whiskers, max and min over the fields. */**/***, p < 0.05/0.01/0.001. (C) Schematic of proposed interactions between UPF1 and MYOD (MD) during myogenesis. The E3 ligase activity of UPF1’s RING domain promotes proteasome-mediated degradation of MYOD protein in myoblasts. Mutating the E2-E3 binding pocket of UPF1 stabilizes MYOD protein, which in turn promotes myogenesis. DM, differentiation media. Green versus red loops within UPF1 indicate the wild-type versus mutated E2-E3 binding pockets. Black versus red dotted arrows indicate proposed RING domain activities for wild-type versus mutated UPF1, including recruiting an E2 conjugase and transferring ubiquitin (Ub) for the wild-type protein. See also Figure S4.