Tenotomy decreases reporter protein synthesis via the 3′-untranslated region of the β-myosin heavy chain mRNA

William W. Ashley Jr., Brenda Russell

Abstract

We tested the hypothesis that the β-myosin heavy chain (β-MHC) 3′-untranslated region (UTR) mediates decreased protein expression after tenotomy of the rat soleus. We also tested the hypothesis that decreased protein expression is the result of RNA-protein interactions within the 3′-UTR. β-MHC was chosen for study because of its critical role in the function of postural muscles such as soleus. Adult rat soleus muscles were directly injected with luciferase (LUC) reporter constructs containing either the β-MHC or SV40 3′-UTR. After 48 h of tenotomy, there was no significant effect on LUC expression in the SV40 3′-UTR group. In the β-MHC 3′-UTR group, LUC expression was 37.3 ± 4% (n = 5, P = 0.03) of that in sham controls. Gel mobility shift assays showed that a protein factor specifically interacts with the β-MHC 3′-UTR and that tenotomy significantly increases the level of this interaction (25 ± 7%,n = 5, P = 0.02). Thus the β-MHC 3′-UTR is directly involved in decreased protein expression that is probably due to increased RNA-protein binding within the UTR.

  • mechanical signal transduction
  • ribonucleic acid binding protein
  • translational control
  • muscle atrophy

tenotomy is an unloading model of muscle atrophy that results in rapid decrease of muscle cell size and protein content (9). Tenotomy results in a 34% decrease in wet weight after 1 wk (9), and, specifically, decreases the β-myosin heavy chain (β-MHC) protein content of postural muscles such as the soleus (12). It is well known that chronic adaptation is mediated by changes at the level of transcription and protein degradation (14), which do not have significant effects until about 1 wk of decreased load (3, 5). Acute regulation, however, is less well understood, but translational control may be one of the earliest adaptive events of unloading. Protein synthesis can be quickly adjusted by control of initiation, elongation, or altering message stability (5, 13,19).

Regulation that takes place during the initiation phase of translation occurs by altering the rate at which ribosomes attach to the 5′-untranslated region (UTR) of mRNA (18). Regulation can also occur during the elongation phase as ribosomes move along the mRNA to translate protein. Analysis of polysomal mRNA distribution on sucrose density gradients suggests that protein synthesis after tenotomy is translationally regulated during the elongation phase (15). The mechanisms of such regulation are unclear, but recent work suggests that the 3′-UTR may be important. For example, in primary cultured cardiac myocytes, α-MHC is translationally regulated via its 3′-UTR after contractile arrest (7). In another study, translation of myocyte enhancer factor 2A was shown to be repressed via sequences in its 3′-UTR (4). One interpretation of the role of the 3′-UTR is that highly conserved sequences within this region may help cells respond to environmental stresses such as hypoxia or heat shock (8,17). Mechanically transduced changes, such as the severely decreased load after tenotomy, could also be considered a major stress. In this study, we investigated the possibility that the 3′-UTR of β-MHC, a major contractile protein, plays a role in the mechanotransduction pathway that leads to atrophy after tenotomy.

Translational regulation is usually mediated by the interaction oftrans-acting protein factors with the 5′- or 3′-UTR of mRNA. These interactions, which depend on both the mRNA primary sequence and its secondary structure, can initiate a chain of events that leads to altered protein expression. For example, altered protein-3′-UTR binding has a role in regulating expression of cholesterol 7α-hydroxylase in liver (1). More importantly, Booth and Kirby (5) have shown that proteins specifically interact with activity-responsive elements in the cytochrome c 3′-UTR of skeletal muscle. We have tested whether a similar mechanism may be involved in the regulation of β-MHC protein expression also involving the 3′-UTR.

Our hypothesis is that the β-MHC 3′-UTR mediates decreased protein synthesis after 48 h of tenotomy in rat soleus muscle. Furthermore, we hypothesize that this regulation of protein expression is the result of specific and differential interactions oftrans-acting protein factors with the β-MHC 3′-UTR that effectively downregulate protein synthesis after tenotomy.

METHODS

Animals

Proper animal handling protocols were conducted at all times. Age- and weight (>250 g)-matched female Sprague-Dawley rats were used 1 wk postpartum, after pups had been collected for use in other experiments. Animals were provided with food and water ad libitum.

DNA Construct

The full-length β-MHC 3′-UTR (kindly donated by Dr. G. Goldspink) that concluded with the poly(A) signal was subcloned into the BamH I/EcoR I sites of pBluescript II SK (Stratagene, La Jolla, CA). The Xba I/EcoR V fragment was then isolated and inserted into the Xba I andHpa I sites of vector pGL3-control (Promega, Madison, WI) after the 3′ end of the luciferase coding sequence (Fig.1 A). The pGL3-control vector containing the SV40 3′-UTR and poly(A) signal (Fig.1 B) was used as a control because it is not an endogenous muscle UTR and is unlikely to be regulated by changes in muscle load. For more sequence details and projected secondary structures, please refer below.

Fig. 1.

A schematic diagram of the fusion reporter constructs used to assay for the role of the β-myosin heavy chain (β-MHC) 3′-untranslated region (UTR) in translational control. The constructs are identical except for the 3′-UTR. The luciferase coding sequence fused to either the β-MHC (A) or SV40 (B) 3′-UTR is driven by an SV40 promoter and allows one to specifically assess the role of the 3′-UTR. Luc gene, luciferase gene.

DNA Injection

Rats were anesthetized by ether inhalation and subsequent intramuscular injection of ketamine-xylazine (0.1 ml/g). A small incision (<2 cm) was made in the hindlimb and the soleus exposed by blunt dissection. The left and right soleus muscles of each animal were injected with pGL3-SV40-3′-UTR or pGL3-β-MHC-3′-UTR parallel to the fibers at three discrete points (∼1 mm apart) along the muscle. Each animal received a total of 75 μl of DNA (1.3 μg/μl) per leg combined with a small amount of india ink to facilitate identification of the injection sites. The incision was closed using a sterile technique. The animals were allowed to recover, and tenotomy was performed 1 day after injection.

Tenotomy

Rats were again anesthetized by ether inhalation and subsequent intramuscular injection of ketamine-xylazine (0.1 ml/g). To induce the atrophic process, the right calcaneus tendon was severed with a sterile scalpel. The left tendon was exposed but left intact to serve as the sham control. Wounds were sealed with Nexaband tissue sealant (Veterinary Products, Phoenix, AZ). The rats were awake and mobile by 3 h posttenotomy. Rats were killed 48 h after tenotomy. The left and right soleus muscles were carefully removed, flash frozen in liquid nitrogen, and stored at −80°C for further processing.

Immunohistochemistry

Tissue was mounted on cork, frozen, and serial sectioned into 10-μm slices. Sections were taken from areas near the injection site as identified by india ink marking. The sections were washed in PBS, fixed with 4% paraformaldehyde for 10 min, and incubated in primary anti-luciferase antibody (Promega) for 45 min. The sections were then incubated in BSA for 15 min to reduce nonspecific binding of the secondary antibody. Finally, sections were incubated in fluorescein-labeled secondary antibody for 30 min, washed in PBS, and mounted on glass slides. The sections were analyzed for luciferase protein expression using digital microscopy (Nikon Microphot FXA microscope fitted with Photometrics-cooled charge-coupled device camera).

Tissue Processing

Each soleus was ground to a fine powder in liquid nitrogen using a mortar and pestle and placed in precooled, preweighed plastic tubes for use in luminometry experiments, the bicinchoninic acid assay (BCA; Pierce, Rockford, IL), and RNA-protein gel mobility shift assays. Cell lysis buffer containing 0.5% Nonidet P-40 (NP-40), 10 mM HEPES, 3 mM MgCl2, 40 mM KCl, 2 mM dithiothreitol (DTT), 0.5 mM phenylmethylsufonyl fluoride, and Sigma protease inhibitor cocktail (Sigma, St. Louis, MO) was added (3.5 μl/mg tissue), and the tissue was homogenized on ice at high speed (30,000 rpm) for 60 s (3 × 20 s) using a Tissue Tearor homogenizer (Biospec). To eliminate large cell debris, this homogenate was centrifuged at 3,000 rpm at 4°C for 10 min. The supernatant fraction was then transferred to a new tube and stored at −80°C.

Luminometry Assay for Translated Luciferase

Soleus extract was thawed and 100 μl added to a glass test tube. Luciferase (500 μl) enzyme substrate (Promega) was added to the test tube, carefully mixed, and the ensuing luminescence reaction was quantified using an EG & G Berthold Lumat LB luminometer. Light production was expressed in relative luminometry units (RLU).

Total Protein Measurement

For measurement of total protein, tissue was ground and homogenized in cell lysis buffer as above. The BCA assay was then used to measure protein concentration in each sample. Briefly, 10 μl of sample was added to 200 μl of working reagent and incubated at 37°C for 30 min. The colorimetric reaction was analyzed at 550 nm and compared with a protein standard prepared from known concentrations of BSA. This information was used to assess changes in total protein concentration after tenotomy and to standardize the amount of protein used in the RNA-protein gel mobility shift assay.

RNA-Protein Gel Mobility Shift Assay

Probes.

Sense and antisense [α-32P]CTP-labeled RNA probes containing the complete β-MHC, α-MHC, or SV40 3′-UTR were transcribed in vitro from cDNA linearized vectors. Sense probes were transcribed using SP6 RNA polymerase, and antisense probes were transcribed using T7 RNA polymerase. Both radiolabeled probes were transcribed by using 500-μM cold ATP, GTP, and UTP and 12-μM cold CTP and were labeled with 25 μCi [α-32P]CTP by incubation at 37°C for 1 h. For nonradiolabeled probes, 500-μM cold ATP, CTP, GTP, and UTP were combined and incubated with polymerase at 37°C for 1 h. RNase-free DNase I (1 unit) was added to digest DNA, and the unincorporated nucleotides were removed using Sephadex spin columns (ProbeQuantTM G-50 microcolumns; Pharmacia Biotech, Piscataway, NJ). Radiolabeled probes were analyzed by PAGE and radioactivity determined on a scintillation counter. Nonradiolabeled probes were analyzed by agarose/urea gel electrophoresis. The average riboprobe RNA concentration was ∼50 μg/ml as determined spectrophotometrically by excitation at 260 and 280 nm. Probe folding pattern was estimated using the mFold 2.3 folding program (Dr. M. Zuker, Washington Univ., St. Louis, MO) (22).

Native Gels.

To determine gross RNA-protein interactions, soleus muscle homogenate was combined with the appropriate volume of incubation buffer (0.5% NP-40, 10 mM HEPES, 3 mM MgCl2, 40 mM KCl, and 2 mM DTT), RNase inhibitor (50 units), and radiolabeled sense or antisense probe (40,000 cpm) for a final volume of 21 μl (55 mg protein/ml). The mixture was incubated for 12 min at room temperature to allow RNA-protein hybridization to occur. To verify that shifted bands were the result of protein interaction with RNA, some samples were combined with proteinase K (50 or 100 μg/ml) and then incubated at 37°C for 30 min before hybridization with radiolabeled probe. After hybridization, 10 μl of loading buffer (8% sucrose, 0.025% bromphenol blue, 0.025% xylene cyanol, and 1× Tris base EDTA boric acid) was added and the mixture electrophoresed on a 5% polyacrylamide gel. The samples were run for ∼4 h at 4°C at 35–40 mA.

For competition experiments, soleus muscle homogenate was combined with the appropriate volume of incubation buffer and RNase inhibitor (50 units) as above. Then, before radiolabeled sense probe was added, nonradiolabeled sense probes (10×, 50×, or 100× vs. RNA concentration of the radiolabeled probe) were added to the mixture. After mixing and incubation for 1–3 min, radiolabeled sense probe was added (40,000 cpm) for a final volume of 21 μl (55 mg protein/ml). The mixture was incubated for 12 min at room temperature to allow RNA-protein hybridization to occur. RNA-protein complex formation was assayed by nondenaturing PAGE as above.

Denaturing gels.

For denaturing gels, samples were prepared as above (see Native Gels). After hybridization, the RNA-protein complexes were covalently cross-linked for 9 min (8,400 μJ), 3 cm from the source. Four microliters of 5× SDS loading buffer (95% formamide, 0.025% xylene cyanol, 0.025% bromphenol blue, 0.5 mM EDTA, and 0.025% SDS) were added and samples were then heated at 95°C for 4 min. Samples were loaded onto a 7% SDS polyacrylamide separating gel with a 4% stacking gel. Gels were run for about 3 h at room temperature at 35–40 mA.

All RNA-protein interactions were visualized by exposure of gel to Kodak Biomax MS radiosensitive film (Eastman Kodak, Rochester, NY) and analyzed by densitometry (Molecular Dynamics, Sunnyvale, CA).

Statistics

Data are presented as means ± SE, where nequals the total number of rats used in a given experiment. Unless otherwise noted, paired Student's t-test was used on raw data to determine statistical significance at the P < 0.05 level.

RESULTS

3′-UTR Reporter Constructs Are Expressed After Direct Injection In Vivo

Soleus muscle successfully incorporated and expressed luciferase protein after transfection by direct injection of the construct shown in Fig. 1 A. Transfected muscle (Fig.2 A) stained positively for luciferase protein indicating successful local transfection. We found relatively homogenous transfection in an area local to the injection site and variable transfection at more distant locations. Untransfected muscle (Fig. 2 B) showed no staining, confirming that confirms that secondary antibody does not bind the tissue in the absence of luciferase protein.

Fig. 2.

Immunofluorescent staining of frozen sections of soleus.A: muscle cells 48 h after injection with luciferase constructs. Positive staining for luciferase indicates successful transfection. Luciferase is detected using fluorescein-labeled anti-luciferase antibody. B: untransfected control muscle. Calibration bar is 20 μm.

Tenotomy Does Not Alter Total Protein Concentration by 48 h

To assess the gross short-term effects of tenotomy, the BCA assay was used to determine total protein concentration before and after tenotomy. As expected, total protein concentration in tenotomized soleus was unchanged at 104 ± 5% (n = 9,P = 0.5) of the concentration measured in sham controls. Thus 48 h posttenotomy is too early for gross changes in protein content to be detected.

β-MHC 3′-UTR Mediates Decreased Luciferase Protein Expression After Tenotomy

Luciferase expression in tenotomized soleus transfected with the pGL3-SV40–3′-UTR construct was 100.90 ± 5% (n = 5, P = 0.82) of that in sham controls (Fig. 3). Thus tenotomy had no significant effect on luciferase expression in soleus transfected with the SV40 3′-UTR construct. However, in soleus transfected with the pGL3-β-MHC-3′-UTR, luciferase expression after tenotomy was 37.3 ± 4% (n = 5, P = 0.02) of that in sham controls. Therefore, luciferase expression after tenotomy was significantly reduced only in the pGL3-β-MHC-3′-UTR group. The raw individual measurements (Fig. 3 B) from each group vary within a small range, indicating that plasmid transfection and expression levels are consistent from rat to rat. Thus any significant differences seen across groups is biologically significant. The two outliers were greater than two standard deviations above the mean, so they were not used in the data analysis. However, the data are still significant when these outliers are included.

Fig. 3.

A: effect of tenotomy on the expression of the luciferase reporter constructs. Luciferase expression (in RLU, relative luminometry units) is significantly decreased only in tenotomized (Tenot) soleus injected with the pGL3-β-MHC-3′-UTR construct (P < 0.02). B: raw data for luciferase expression in RLU after transfection. Raw individual measurements from each group vary within a small range, indicating that plasmid transfection and expression levels are consistent. Therefore, any significant differences seen across groups are biologically significant. * P = 0.02.

The Sense and Antisense β-MHC 3′-UTRs Are Homologous in Shape

The sequence of the full β-MHC 3-UTR (sense) was folded into its most energetically favorable (−27.4 kJ) structure (Fig.4 A) as projected by mFold 2.3 folding software (22). It forms a long stem with several loops and bulges that are potential sites for protein binding. Furthermore, it possesses a 5′ palindromic sequence (AUCCCAACCCUA) and an AU-rich region at its 3′ end, both of which could have regulatory roles. The β-MHC 3′-UTR sense and antisense (Fig. 4 B) riboprobes appear to be quite similar in shape. Therefore, an RNA binding protein that interacts with mRNA based solely on its structure could interact similarly with both sense and antisense β-MHC 3′-UTR probes. The α-MHC and SV40 3′-UTR riboprobes (Fig. 4, Cand D), however, are both sequentially and structurally unrelated to the β-MHC 3′-UTR probes. Therefore, if protein binding is specific to the β-MHC 3′-UTR, then hybridization of the α-MHC or SV40 3′-UTR riboprobe with soleus extract will result in a completely different pattern of binding than hybridization with either of the β-MHC 3′-UTR riboprobes.

Fig. 4.

Sequence and projected secondary structures at 25°C of the UTRs in this study. There is little sequential homology shared between any of the 3 UTRs. A: the sense β-MHC 3′-UTR folds into a stable stem structure with several loops and bulges that are potential sites of translational regulation. It also possesses a 5′ palindromic sequence (1) and a 3′ AU rich region (2). B: the antisense β-MHC 3′-UTR also folds into a stable stem structure. It shares a high degree of structural homology with the sense β-MHC 3′-UTR. C: the sense α-MHC 3′-UTR folds into a more complicated structure. It is very different from either of the β-MHC 3′-UTRs. D: the sense SV40 3′-UTR is also very different from the β-MHC 3′-UTRs. Its sequence is as follows: CTAGAGTCGG GGCGGCCGGC CGCTTCGAGC AGACATGATA AGATACATTG ATGAGTTTGG ACAAACCACA ACTAGAATGC AGTGAAAAAA ATGCTTTATT TGTGAAATTT GTGATGCTAT TGCTTTATTT GTAACCATTA TAAGCTGCAA TAAACAAGTT AACAACAACA ATTGCATTCA TTTTATGTTT CAGGTTCAGG GGGAGGTGTG GGAGGTTTTT TAAAGCAAGT AAAACCTCTA CAAATGGGTA AAATCGATAA G.

A Trans-Acting Protein Factor Interacts With the β-MHC 3′-UTR in a Sequence-Specific Manner

To test the hypothesis that the decreased protein expression via the 3′-UTR is the result of protein interaction with the UTR, gel mobility shift assays were performed using protein extract hybridized to 32P-labeled sense and antisense riboprobes. Representative nondenaturing 5% polyacrylamide gels are shown (Fig.5). Free β-MHC 3′-UTR sense and antisense probes are 163 and 154 nucleotides (nt), respectively, and run to the appropriate position (Fig. 5 A,lanes 2 and 3). After hybridization with extract, a band appeared at ∼500 nt, and the density of the free probe band decreased. Note there was no measurable difference in the binding pattern between extracts from sham-operated vs. tenotomized soleus when analyzed densitometrically under nondenaturing conditions (control is 100%, tenotomized is 92.2 ± 8%; n = 4,P = 0.4).

Fig. 5.

A: a representative nondenaturing 5% polyacrylamide gel. Lanes are as follows: lane 1, L, RNA ladder; lane 2, β, free radiolabeled β-MHC 3′-UTR sense probe; lane 3, , free radiolabeled β-MHC 3′-UTR antisense probe; lane 5, βE, radiolabeled β-MHC 3′-UTR sense probe + protein extract (55 mg/ml); lane 6, βE 10× cold, 10× cold β-MHC 3′-UTR sense probe + protein extract (55 mg/ml) + radiolabeled β-MHC 3′-UTR sense probe; lane 7, βE 50× cold, 50× cold β-MHC 3′-UTR sense probe + protein extract (55 mg/ml) + radiolabeled sense probe; and lane 8, AβE, radiolabeled β-MHC 3′-UTR antisense probe + protein extract (55 mg/ml). All probes are used at 40,000 cpm. β-MHC 3′-UTR sense and antisense probes are 163 and 154 nt, respectively. Lanes 1 and 2 show probes running to the correct size. As indicated by shifted band at ∼500 nt, lane 5 shows that there is RNA-protein interaction. Lanes 6 and 7 suggest specific RNA-protein interaction as 500-nt band is competed away by addition of homologous cold probe, and free probe band returns to levels near that seen in probe alone lane. As indicated by completely different shifted band shown in lane 8, antisense probe does not interact with same protein. B: a representative nondenaturing 5% polyacrylamide gel. Lanes are as follows: lane 1, βE, radiolabeled β-MHC 3′-UTR sense probe + extract (55 mg/ml);lane 2, βE pk50, radiolabeled β-MHC 3′-UTR sense probe + extract (55 mg/ml) + proteinase K (50 μg/ml);lane 3, βE pk100, radiolabeled β-MHC 3′-UTR sense probe + extract (55 mg/ml) + proteinase K (100 μg/ml); andlane 4, αE, radiolabeled α-MHC 3′-UTR sense probe + extract (55 mg/ml). Lane 1 shows formation of an RNA-protein complex as seen in A, lane 5. Lanes 2and 3 show that RNA-protein complex seen in lane 1 can be eliminated by digesting protein with proteinase K.Lane 4 shows that migration pattern of radiolabeled α-MHC 3′-UTR sense probe + extract is completely different from that of radiolabeled β-MHC 3′-UTR sense probe + extract. C: a representative nondenaturing 5% polyacrylamide gel. Lanes are as follows: lane 1, L, RNA ladder; lane 3, β, free radiolabeled β-MHC 3′-UTR sense probe; lane 4, SV, free radiolabeled SV40 3′-UTR sense probe; lanes 5 and6, βE, radiolabeled β-MHC 3′-UTR sense probe + protein extract (55 mg/ml); and lanes 8–11, SVE, radiolabeled SV40 3′-UTR sense probe + protein extract (55 mg/ml). All probes are used at 40,000 cpm. β-MHC and SV40 3′-UTR sense are 163 and 314 nt, respectively. As seen in lanes 5 and6, a shifted band appears at ∼500 nt when radiolabeled β-MHC 3′-UTR sense probe is hybridized with extract. However, no such band appears in lanes 8–11 when radiolabeled SV40 3′-UTR sense probe is hybridized with extract. Thus SV40 3′-UTR probe does not interact with same protein factor. D: a representative nondenaturing 5% polyacrylamide gel. Lanes are as follows: lane 1, L, RNA ladder; lane 3, β, free radiolabeled β-MHC 3′-UTR sense probe; lane 4,β 100× cold α, 100× cold α 3′-UTR sense probe + radiolabeled β-MHC 3′-UTR sense probe; lane 6, βE, protein extract (55 mg/ml) + radiolabeled β-MHC 3′-UTR sense probe; lane 7, βE 50× cold α, 50× cold α 3′-UTR sense probe + protein extract (55 mg/ml) + radiolabeled β-MHC 3′-UTR sense probe; and lane 8, βE 100× cold α, 100× cold α 3′-UTR sense probe + protein extract (55 mg/ml) + radiolabeled β-MHC 3′-UTR sense probe. All probes are used at 40,000 cpm. As shown in lanes 6–8, increasing amounts of cold heterolgous probe does not alter appearance of band at 500 nt. This, along with Fig. 5 A (lanes 5–7), indicates that the RNA-protein interaction at 500 nt is specific to the β-MHC 3′-UTR.

The mobility pattern of sense probe plus extract was completely different from that observed with antisense probe plus extract (Fig.5 A, lane 8). This difference indicates that the interaction seen in lane 5 is sequence specific. The binding is also completely different when the same muscle extract is hybridized with the radiolabeled α-MHC 3′-UTR probe (Fig.5 B, lane 4). Furthermore, no band appears at 500 nt when the radiolabeled SV40 3′-UTR sense probe is hybridized with extract (Fig. 5 C, lanes 8–11). Thus the SV40 3′-UTR probe does not interact with the same protein factor as the β-MHC 3′-UTR probe.

Competition with cold sense probe was used as a further test of binding specificity. Hybridization of sense probe to muscle extract was allowed to occur in the presence of 10× or 50× nonradiolabeled homologous sense probe. The interaction at ∼500 nt slightly decreased in the 10× cold probe lane (Fig. 5 A, lane 6), and the free probe band increased. In the 50× cold probe lane, the 500 nt RNA-protein interaction was almost completely competed away, and the free probe band at 163 nt returned to levels approaching those observed in the probe alone lane (Fig. 5 A, lane 7). However, cold heterologous (α-3′-UTR) probe does not compete away the interaction at 500 nt (Fig. 5 D, lanes 6–8). This indicates that a specific RNA-protein interaction does occur between the β-MHC mRNA 3′-UTR and a protein binding factor.

Samples were incubated with proteinase K to digest the proteins to verify that the band observed at 500 nt was the result of mRNA interaction with a protein and not the result of nucleic acid complex formation or double-stranded RNA. As shown in Fig. 5 B(lanes 2 and 3), incubating extract with proteinase K for 30 min at 37°C (50 or 100 μg/ml) eliminates the shifted band. This confirms that it is indeed a protein that is binding to the sense β-MHC 3′-UTR riboprobe.

Specific Protein Binding Factor Interaction With the β-MHC 3′-UTR Is Increased After Tenotomy

To examine the physiological significance of the binding seen under native conditions, RNA-protein interactions in sham and tenotomized soleus were examined under denaturing conditions. A representative SDS polyacrylamide gel is shown (Fig.6). The band at 116 kDa confirms that there is interaction occurring in both the sham and tenotomized muscles (Fig. 6, lanes 1 and 2) when extract is hybridized with radiolabeled β-MHC 3′-UTR sense probe. Interestingly, this band is much darker in the tenotomized lane due to greater binding of protein to the labeled riboprobe. No binding occurs at 116 kDa when radiolabeled β-MHC 3′-UTR antisense probe is hybridized to extract (lanes 3 and 4). Densitometric analysis revealed that, under denaturing conditions, the 116-kDa band in the sense-tenotomized lane is (25 ± 7%, P = 0.02,n = 5) more dense than the same band in the sense sham-operated lane (Fig. 7). Thus the protein binding factor interaction with the β-MHC mRNA 3′-UTR is significantly increased after tenotomy.

Fig. 6.

A representative denaturing SDS polyacrylamide gel. The lanes are as follows: lane 1, βE Sham, radiolabeled β-MHC 3′-UTR sense probe + sham-operated soleus extract (55 mg/ml); lane 2, βE Tenot, radiolabeled β-MHC 3′-UTR sense probe + tenotomized soleus extract (55 mg/ml); lane 3, AβE Sham, radiolabeled β-MHC 3′-UTR antisense probe + sham-operated soleus extract (55 mg/ml); and lane 4, AβE Tenot, radiolabeled β-MHC 3′-UTR antisense probe + tenotomized soleus extract (55 mg/ml). A shifted band appears at 116 kDa in the sense probe + extract lanes (lanes 1 and 2). This band appears to be more dense in the tenotomized lane. No binding occurs at 116 kDa when radiolabeled β-MHC 3′-UTR antisense probe is hybridized to extract (lanes 3 and 4).

DISCUSSION

Our findings support the hypothesis that atrans-acting protein factor interacts with the β-MHC 3′-UTR to induce decreased protein expression within 48 h of unloading. This mechanism offers muscle a way to reduce mass quickly to minimize energy demands.

Translational Control of Protein Synthesis After Tenotomy

The results of the present investigation showed that tenotomy is a negative regulator of protein expression in vivo and acts via a 3′-UTR-dependent mechanism. The decrease in luciferase expression after tenotomy may be attributed to translational control in the form of altered translational rates or altered mRNA stability. We think that this 3′-UTR-mediated effect is the result of translational block and not decreased mRNA stability. This is consistent with other studies on muscle. For example, analysis of polysomal mRNA distribution on sucrose density gradients showed that atrophy after tenotomy is translationally regulated during the elongation phase (15). Furthermore, in our laboratory, primary cultured skeletal myocytes were transfected with the same pGL3-SV40–3′-UTR and PGL3-β-MHC-3′-UTR constructs used in this study and then treated with EGTA-AM to induce translational block (6). We showed that transfection and EGTA-AM treatment had no effect on mRNA stability (2). Likewise, mechanical arrest in cultured cardiac myocytes transfected with a similar pGL3-α-MHC-3′-UTR chimeric construct also had no effect on mRNA stability (10). mRNA stability is associated with the length of the poly(A) tail or the presence of destabilizing elements such as AAUAAA (AREs) within the 3′-UTR (11). Poly(A) binding proteins (PABP), which interact with poly(A) tails and destabilize the mRNA, are known to be general in their effects. Accordingly, we would expect a simple poly(A)-mediated destabilization to affect most cellular mRNAs equally. Our results showed that tenotomy only affected luciferase expression in soleus transfected with pGL3-β-MHC-3′-UTR and not pGL3-SV40–3′-UTR. Furthermore, it has been shown that the β-MHC mRNA seems to demonstrate unusual stability both in hypertrophy and atrophy models (16, 19). Thus we rule out PABP-mediated mRNA destabilization. Moreover, although the β-MHC 3′-UTR has one AU rich region (Fig. 4 A), it does not contain any sequence known to be a destabilizing element. These data lend strong support to the hypothesis that the decreased protein expression observed after tenotomy is the result of translational block. However, we cannot completely rule out the possibility that decreased mRNA stability plays a minor role in decreased reporter protein expression. Whether the effect is entirely translational block or also changes in stability, it is clear that the β-MHC 3′-UTR is directly involved in the decreased protein expression observed after tenotomy.

Though our data showed that no change can be detected in total protein levels after 48 h of tenotomy, it is well established that tenotomy triggers a cascade that eventually leads to atrophy (9). However, we did show that expression of luciferase from a transfected construct changed after 48 h of tenotomy. Remembering that the effects of protein degradation are not significant until about 1 wk of unloading, this suggests that expression of endogenous β-MHC protein would be translationally blocked after tenotomy. Furthermore, the 25% decrease in total MHC levels observed by Jakubiec-Puka et al. (12) after 2 days of tenotomy could be explained entirely by decreased de novo translation of MHC protein in the unloaded muscle.

RNA-Protein Interaction Within the β-MHC 3′-UTR

Native gel mobility shift assays indicate that atrans-acting protein factor specifically interacts with the β-MHC 3′-UTR in vivo. Our data also suggest that protein binding relies on mRNA primary sequence and/or secondary structure, but does not rely on structure alone. Furthermore, native proteins retain tertiary and quaternary structure. Therefore, although the binding observed under native conditions could be the result of a single protein interacting with the 3′-UTR, it is more likely the result of a large protein complex interacting with the 3′-UTR. This could explain why, under native conditions, we could not resolve differences in the RNA-protein interactions in the β-MHC 3′-UTRs of sham vs. tenotomized soleus. Indeed, native conditions may not be sensitive enough to resolve the subtle gel mobility differences that would occur if a single protein component of a large multiprotein complex is enriched. By denaturing the protein, we eliminated tertiary and quaternary structure. This allowed us to examine the mobility of RNA complexed with individual proteins. Under denaturing conditions, we observed an increased RNA-protein interaction within the β-MHC 3′-UTR.

Our discovery of this mechanism is novel for β-MHC in vivo, but similar findings have been reported in vitro. Yan et al. (21) showed an increase in contractile activity of skeletal muscle leads to a decrease in RNA-protein interaction in the cytochrome c 3′-UTR. In addition, other data from our laboratory showed that decreased contractile activity leads to increased binding of proteins with the α-MHC 3′-UTR in cardiac myocytes (10). Moreover, our interpretation fits well with a model offered by Ku and Thomason (15) that suggests that translational block occurs as a result of ribosomes stalled on the mRNA strand. Protein interaction with the 3′-UTR could cause such a stall by physically blocking ribosome movement, by altering the structure of the mRNA, or by simply masking the mRNA (20). It is possible that protein interaction with the 3′-UTR also increases nuclease activity that effectively destabilizes the mRNA.

Model of Translational Control After Unloading

We have shown that tenotomy results in decreased protein expression mediated by the β-MHC 3′-UTR. In addition, we have shown that tenotomy leads to increased RNA-protein interaction in the β-MHC 3′-UTR. These coincident events suggest a model (Fig.8) in which the reduced protein expression observed soon after tenotomy is mediated by increased interaction of a trans-acting protein factor with the β-MHC 3′-UTR. One possibility is that under normal loading conditions, ribosomes move freely along the complete mRNA coding sequence translating protein as they move from codon to codon. Upon reaching the termination codon, the ribosomes detach from the mRNA and release the nascent protein. During unloading, the affinity of atrans-acting factor for the β-MHC 3′-UTR significantly increases. Increased interaction of this factor with the 3′-UTR could physically prevent the movement of the ribosomes along the mRNA. As a result, ribosomes are stalled on the mRNA, translation is blocked, and completed protein is never released. Thus interaction of this factor with the 3′-UTR effectively represses the translation of protein. Alternatively, interaction of this factor with the 3′-UTR could help stabilize nuclease activity on the mRNA by anchoring the nuclease close to the mRNA. As a result, mRNA would be less stable and protein expression would decrease.

Fig. 7.

Denaturing gels were analyzed densitometrically. Data are plotted as the optical density (OD) of the 116-kDa shifted band as a percent of the sham (n = 5). The OD of the 116-kDa band in the tenotomized lane is significantly (25%) higher than in the sham lane.

Fig. 8.

A model of how mechanical unloading may affect translation. Under normal loading conditions, the 3′-UTR is unbound and ribosomes are able to track freely along the entire coding sequence and release the nascent protein when the termination signal is reached. In unloaded conditions, such as tenotomy, a trans-acting protein factor interacts with the 3′-UTR. Such an interaction may physically prevent the ribosomes from moving along the entire mRNA coding sequence. As a result, complete protein is not translated and released from the ribosomes. The result is decreased protein expression after unloading.

Our study confirms earlier work that suggests that the effects of unloading are translationally controlled (5,14) and proves that translational control, including alterations in mRNA stability, can be directly mediated by the β-MHC mRNA 3′-UTR. β-MHC is critical for proper function of postural muscle and is the primary protein affected during unloading-related atrophy. The mechanism suggested here would allow a muscle to regulate β-MHC protein expression quickly and reversibly.

Acknowledgments

We thank to Dr. G. Goldspink for providing the β-MHC 3′-UTR, Dr. G. Nikcevic for expert help in creating the luciferase constructs and careful reading of the manuscript, and Dr. R. J. Solaro and David Montgomery for critical reading of the work.

Footnotes

  • This research was supported by National Institutes of Health Grants HL-40880 (B. Russell) and AR-08405 (W. Ashley).

  • Address for reprint requests and other correspondence: B. Russell, Dept. of Physiology and Biophysics, Univ. of Illinois at Chicago, 835 S. Wolcott Ave., Chicago, IL 60612-7342 (E-mail:Russell{at}uic.edu).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

REFERENCES

View Abstract