A novel conserved isoform of the ubiquitin ligase UFD2a/UBE4B is expressed exclusively in mature striated muscle cells

Abstract

Yeast Ufd2p was the first identified E4 multiubiquitin chain assembly factor. Its vertebrate homologues later referred to as UFD2a, UBE4B or E4B were also shown to have E3 ubiquitin ligase activity. UFD2a function in the brain has been well established in vivo, and in vitro studies have shown that its activity is essential for proper condensation and segregation of chromosomes during mitosis. Here we show that 2 alternative splice forms of UFD2a, UFD2a-7 and -7/7a, are expressed sequentially during myoblast differentiation of C2C12 cell cultures and during cardiotoxin-induced regeneration of skeletal muscle in mice. UFD2a-7 contains an alternate exon 7, and UFD2a-7/7a, the larger of the 2 isoforms, contains an additional novel exon 7a. Analysis of protein or mRNA expression in mice and zebrafish revealed that a similar pattern of isoform switching occurs during developmental myogenesis of cardiac and skeletal muscle. In vertebrates (humans, rodents, zebrafish), UFD2a-7/7a is expressed only in mature striated muscle. This unique tissue specificity is further validated by the conserved presence of 2 muscle-specific splicing regulatory motifs located in the 3' introns of exons 7 and 7a. UFD2a interacts with VCP/p97, an AAA-type ATPase implicated in processes whose functions appear to be regulated, in part, through their interaction with one or more of 15 previously identified cofactors. UFD2a-7/7a did not interact with VCP/p97 in yeast 2-hybrid experiments, which may allow the ATPase to bind cofactors that facilitate its muscle-specific functions. We conclude that the regulated expression of these UFD2a isoforms most likely imparts divergent functions that are important for myogenisis.

Figure 1. A larger isoform of UFD2a is specifically expressed in striated muscle.

A). Protein lysates were prepared from normal human tissues and human salivary gland cells (HSG) that were untreated (-) or arrested in mitosis (+Noc) and subjected to SDS PAGE. After transfer to nitrocellulose, samples were analyzed by immunoblotting using a rabbit polyclonal antibody recognizing UFD2a. A 97-kDa band visible in many muscle lysates and some other tissues represented cross-reactivity to phosphorylase b (not shown). B). UFD2a-7/7a expression in cardiac muscle (Hrt) but not coronary artery (CA) from an explanted human heart (3 patient samples were analyzed; 1 is shown here). C). Human skeletal muscle (Sk Mu) lysates and lysates made from human HeLa cells arrested in mitosis (M) with nocodazole were treated with λprotein phosphatase (λppase). Interphase (I) and untreated mitotic (M) HeLa cells are shown for comparison; phos refers to the phosphorylated form of UFD2a. Top and bottom panels represent different exposures of the same Western blot to account for the difference in the UFD2a levels in total protein from cell cultures compared to human tissue. D).Total RNA from human skeletal muscle (Hu SkM), K562 human erythroleukemia cells, and primary human monocyte-derived macrophages (Hu Mac) were used as templates for cDNA reactions. RT-PCR analyses were performed using primers recognizing a portion from the 5′-end of UFD2a (upper panel), the 3′-end of UFD2a (middle), or from the housekeeping gene HPRT (lower). The 5′ UFD2a primers amplified the expected 830-bp fragment from other tissues and the novel 1217-bp and 1370-bp fragments in muscle.

Figure 2. Positions of the muscle-specific exons.

A). This map shows the locations of the 29 exons of the human UFD2a gene spread across approximately 200 kb near the telomere on the short arm of chromosome 1. Ubiquitous exons are shown in black. Exon 7, which has been found in fetal brain and striated muscle, is shown in blue. The striated muscle–specific exon 7a is show in red. B). Illustration of the UFD2a protein showing positions of the MPAC domain, catalytic U box, exons 7 and 7a, and the region conserved in lower eukaryotes (light green).

Figure 3. Expression of the muscle-specific alternative isoforms of UFD2a is conserved in mice and zebrafish.

A). Among 8 murine tissues, UFD2a-7/7a is expressed only in striated muscle. Protein lysates prepared from adult murine tissues analyzed by immunoblotting using a rabbit polyclonal antibody recognizing UFD2a. B). Schematic diagram depicting the exons of zebrafish ube4B (UFD2a ortholog) and the location of various primer sets used for RT-PCR. C). RNA was extracted from whole 7 pfd zebrafish embryos and subjected to RT-PCR analysis using primers F2 and R2. The 333-bp band represents the fragment expected from the amplification of exons 6, 8, and 9, and the 873-bp fragment includes amplification of the suspected exons 7 and 7a. D). RNA extracted from 4 tissues from multiple adult zebrafish was analyzed by RT-PCR using primers F2 (specific for exon 7) and R2 (exon 8).

Figure 4. UFD2a exon-7a is highly conserved across vertebrates and contains conserved motifs associated with muscle-specific alternative splicing.

A). Alignment of exon 7a amino acid sequence. Asterisks indicate complete conservation across all species, and colons indicate strong conservation. Negatively charged residues are shown in blue, and hydrophobic residues are shown in red. B). Alignment of the 3′-end of exon 7a plus the first 100 bp of the following intron. The splice donor site is shown in blue. The CTAAC putative muscle-specific alternative splicing signal is shown in red. The TGCATG Fox motif enriched in introns of striated muscle exons is shown in purple. Residues identical in all sequences are marked with an asterisk. The amino acid sequence of human exon 7a is shown above.

Figure 5. UFD2a-7 and UFD2a-7/7a mRNA and protein expression are upregulated during differentiation and regeneration.

A). C2C12 cells were plated in growth medium overnight and then transferred into differentiation medium (DM). Cell cultures were harvested for analysis after the indicated times. Equal total protein was loaded on SDS-PAGE (see Figure S1 for poneau S staining) and analyzed for expression of the indicated proteins. The UFD2a panel was performed on a large-format 6% bis-acrylamide gel for greater separation of the 2 larger UFD2a isoforms and probed with the 15041 polyclonal antibody, which recognizes all isoforms. B). C2C12 cells were plated in growth medium overnight and then transferred into DM. Cell cultures were harvested for analysis after the indicated times. Total RNA was isolated using TRIzol and used as a template for RT-PCR. PCR was performed on the resulting cDNA (made with or without reverse transcriptase (RT)) using a 5′ primer complementary to exon 6 and a 3′ primer specific for either the exon 7/exon 8 junction (top panel) to detect UFD2a-7 transcripts or the exon 7a/exon 8 junction (middle panel) to detect UFD2a-7/7a transcripts. Primers specific to exon 19 and 20 were used in the bottom panel to show total UFD2a transcripts. C). Total RNA was isolated with TRIzol from duplicate samples of cells used in panel A and subjected to RT-PCR. Quantitative PCR was performed in duplicate using TaqMan primers that recognized the indicated transcripts. Transcript expression data, normalized to the expression of HPRT, are expressed relative to the expression level at time zero. Error bars indicate standard deviation.

Figure 6. Expression of UFD2a alternative-splice forms occurs sequentially during regeneration and developmental myogenesis.

A). Mice received intramuscular injections of cardiotoxin (CTX) in the right tibialis anterior muscle. After the indicated number of days of recovery, the treated and untreated contralateral muscles were harvested and analyzed by immunoblotting for UFD2a, the mature muscle marker myosin heavy chain (MHC), the proliferating myoblast marker Myf5, myogenin, MyoD, and vinculin as a loading control. The inset at right shows the exposure from a low-bis gel in which the UFD2a-7 isoform of intermediate size is more clearly separated from UFD2a-7/7a at Day 5 and more clearly shows its absence at Day 12 (arrows point to the two isoforms). B). Skeletal muscle (upper panels; pooled gastrocnemius, soleus, and quadriceps muscles) or hearts (lower panels) from E14, E16, and E18 embryos and P1 and P7 pups were dissected from individual embryos or pups at each time point and analyzed by Western blotting for UFD2a (note a second set of mice are shown in Fig S2). Adult mouse heart (H), skeletal muscle (SkM), and spleen (Sp) were simultaneously analyzed for comparison. C). RNA was extracted from whole zebrafish embryos at 10, 12, 24, and 48 hpf, and nested PCR was performed using primers F3 and R3 on the 873- and 756-bp fragments produced from the initial RT-PCR.

Figure 7. The muscle-specific UFD2a-7/7a isoform does not interact with VCP/p97.

A). Yeast cells transformed with bait plasmids pDEST32-UFD2a, pDEST32-UFD2a-7/7a, or empty pDEST32 and prey plasmids pDEST22-VCP/p97 or empty pDEST22 were grown in liquid double dropout medium (-leu, -trp) to saturation and diluted to an OD₆₀₀ of 1. Serial dilutions of these cell cultures were prepared in 96-well plates and transferred to triple-dropout agar plates (-leu, -trp, -his) supplemented with 25 mM 3-AT. B). Isolated colonies of yeast cells transformed (as in panel A) were grown to saturation, diluted to OD₆₀₀ of 0.3, and grown to log phase (OD₆₀₀ = 0.8-1.5). Cells were disrupted with glass beads in Z buffer; ONPG substrate and β-mercaptoethanol were added to the resulting supernatants. After incubation at 37°C, supernatants were transferred to 96-well plates and read at 420 nm. B). Relative β-galactosidase activity is calculated from 6 independent experiments; error bars indicate standard error. The average activity of UFD2a+VCP/p97 was significantly higher than that of the other transformants (*p = 0.02). The activity of UFD2a-7/7a+VCP/p97 did not differ from that of the negative controls.