Distinct Activities of Myf5 and MyoD Indicate Separate Roles in Skeletal Muscle Lineage Specification and Differentiation

Abstract

Most transcription factor families contain highly related paralogs generated by gene duplication, and functional divergence is generally accomplished by activation of distinct sets of genes by each member. Here we compare the molecular functions of Myf5 and MyoD, two highly related bHLH transcription factors that regulate skeletal muscle specification and differentiation. We find that MyoD and Myf5 bind the same sites genome-wide but have distinct functions: Myf5 induces histone acetylation without Pol II recruitment or robust gene activation, whereas MyoD induces histone acetylation, recruits Pol II, and robustly activates gene transcription. Therefore, the initial specification of the muscle lineage by Myf5 occurs without significant induction of gene transcription. Transcription of the skeletal muscle program is then achieved by the subsequent expression of MyoD, which binds to the same sites as Myf5, indicating that each factor regulates distinct steps in gene initiation and transcription at a shared set of binding sites.

Figure 1. Myf5 Has Weak Transcriptional Activity Compared with MyoD

(A) Phase contrast images of M&M MEFs transduced with either 1x lenti-MyoD or 1x lenti-Myf5 for 1 day and then switched to DM for 3 days. (B) M&M MEFs were transduced with a range of lenti-MyoD or lenti-Myf5 titers for 24 hr then switched to DM for 20 hr. Expression of Myog and Cdh15 was assayed by RT-qPCR, using mRpl17a as an internal control. (C) RNA-Seq data from M&M MEFs transduced with either 1x lenti-MyoD or 1x lenti-Myf5 for 24 hr then switched to DM for 20 hr. Data points depict the log₂ fold change over non-transduced cells. Genes showing a dose-response to increasing amounts of lenti-GFP were removed from the analysis to eliminate confounding viral-response effects. The red line represents the fitted trend line comparing MyoD and Myf5 while the black line represents a slope of 1. (D) RNA-seq reads over the myogenin locus. All tracks have the same scale. (E) Genes upregulated at least 4-fold by MyoD Myf5 with p values less than 10⁻¹⁰. Genes identified as “viral response” (Figure S3) have been removed from the analysis. See also Figures S1–S3.

Figure 2. MyoD and Myf5 Bind the Same Subset of E-Boxes

(A) ChIP-seq reads over the Cdh15 promoter. The mock ChIP samples are from cells transduced with lenti-GFP and immunoprecipitated with antibodies against Myf5 or MyoD. The tracks scales have been normalized using the average peak height to account for differences in the two antibody efficiencies. (B) Overlap of MyoD and Myf5 ChIP-seq peaks plotted by peak coverage. (C) Rank comparison of the top 50,000 MyoD and Myf5 peaks. Each box in the plot represents the overlap of two ranked peak lists. The bins are ordered from the origin as follows: top 5,000 peaks, top 10,000 peaks, etc. The color scale indicates the fraction of loci in the bin bound by both MyoD and Myf5. (D) Motifs enriched under MyoD or Myf5 peaks. (E) The ratio of the number of peaks containing the motif compared with the number of control regions containing the motif.

Figure 3. MyoD and Myf5 Induce Comparable Levels of Local Histone H4 Acetylation

(A and B) H4Ac ChIP-seq reads at the Fibronectin-1 promoter (positive control locus) (A), and an intergenic region of Chr12 surrounding a common MyoD/Myf5 peak (B). M&M MEFs were transduced for 24 hr and then switched into DM for an additional 20 hr. (C–E) Average H4Ac around MyoD (C), Myf5 (D), or control (E) peak lists from M&M MEFs transduced with lenti-GFP, lenti-MyoD, or lenti-Myf5. “0” denotes the center of the MyoD or Myf5 meta-peaks and the plots extend 500 bp to either side. Control E-boxes are E-boxes bound by MyoD in P19 cells but not in MEFs. The graphic below represents theoretical locations of positioned (solid lines) and non-positioned nucleosomes (dotted lines) surrounding E-box binding sites.

Figure 4. MyoD and Myf5 Have Similar Binding Kinetics

(A and B) EMSA: in vitro translated MyoD, Myf5, or Chimera was mixed with E12 and hot probes containing a single GG E-box (A), or a pair of GG E-boxes (B). The cold competitor is the same sequence as the hot probe. (C and D) EMSA band intensity is plotted as a function of the time incubated with the cold competitor.

Figure 5. Addition of a Strong Acidic Activation Domain to the N Terminus Imbues Myf5 with Robust Gene Activation Potential

(A) Schematics depicting MyoD and Myf5 domains, the Myf5-Chimera, and the Gal4 fusion constructs. AD, activation domain; H/C, HisCys region (required for interaction with Pbx); H3, helix 3. (B) Myod- or Myf5-Gal4 fusion constructs were co-expressed in differentiating C2C12s with a Gal4-Luc reporter. The data represent the average of three biological replicates and the error bars represent the SD from the mean. (C and D) RNA-Seq data from M&M MEFs transduced with either lenti-MyoD or lenti-Chimera for 24 hr and then switched to DM for an additional 20 hr. Genes showing a dose-response to increasing amounts of lenti-GFP were removed from the analysis. The black line indicates a linear trend line fitted to the data. Genes that were induced 2-fold or more (log fold change of 1) by MyoD (C) or all genes (D) are displayed. See also Figure S4.

Figure 6. Pol II Recruitment to Myogenic Loci by MyoD, Myf5, and the Chimera

(A and B) Total Pol II (N-20), unphosphorylated Pol II (8WG16), and elongating Pol II (Ser2P) occupancy at the 5′ and 3′ ends of the Myog gene (A), and the Chrng gene (B). A single biological replicate was done for each ChIP and data are represented as the fold change over the Ch7 negative control region to minimize technical differences between the three antibodies. (C–F) Triplicate ChIPs from each sample using the 8WG16 antibody. Error bars represent the SD of the three ChIPs. A region of chromosome 7 with no annotated genes or transcribed retroelements was used as a negative control (E), and the 5′ end of the Fn1 gene was used as a positive control (F). See also Figure S5.