Nitric oxide is generated in skeletal muscle with activity and decreases Ca2+ sensitivity of the contractile apparatus, putatively by S-nitrosylation of an unidentified protein. We investigated the mechanistic basis of this effect and its relationship to the oxidation-induced increase in Ca2+ sensitivity in mammalian fast-twitch (FT) fibers mediated by S-glutathionylation of Cys134 on fast troponin I (TnIf). Force-[Ca2+] characteristics of the contractile apparatus in mechanically skinned fibers were assessed by direct activation with heavily Ca2+-buffered solutions. Treatment with S-nitrosylating agents, S-nitrosoglutathione (GSNO) or S-nitroso-N-acetyl-penicillamine (SNAP), decreased pCa50 ( = −log10 [Ca2+] at half-maximal activation) by ~−0.07 pCa units in rat and human FT fibers without affecting maximum force, but had no effect on rat and human slow-twitch fibers or toad or chicken FT fibers, which all lack Cys134. The Ca2+ sensitivity decrease was 1) fully reversed with dithiothreitol or reduced glutathione, 2) at least partially reversed with ascorbate, indicative of involvement of S-nitrosylation, and 3) irreversibly blocked by low concentration of the alkylating agent, N-ethylmaleimide (NEM). The biotin-switch assay showed that both GSNO and SNAP treatments caused S-nitrosylation of TnIf. S-glutathionylation pretreatment blocked the effects of S-nitrosylation on Ca2+ sensitivity, and vice-versa. S-nitrosylation pretreatment prevented NEM from irreversibly blocking S-glutathionylation of TnIf and its effects on Ca2+ sensitivity, and likewise S-glutathionylation pretreatment prevented NEM block of S-nitrosylation. Following substitution of TnIf into rat slow-twitch fibers, S-nitrosylation treatment caused decreased Ca2+ sensitivity. These findings demonstrate that S-nitrosylation and S-glutathionylation exert opposing effects on Ca2+ sensitivity in mammalian FT muscle fibers, mediated by competitive actions on Cys134 of TnIf.
- muscle fatigue
- skinned muscle fiber
- contractile apparatus
nitric oxide (NO) is generated in skeletal muscle at rest, primarily by neuronal nitric oxide synthase (nNOS), and production increases markedly with contractile activity (5, 20, 30). Endogenous production of NO modulates submaximal force production in skeletal muscle without altering maximum force, with NOS inhibitors and NO scavengers found to increase submaximal force (20) and applied NO donors depressing force (3, 16) (for review see refs. 30, 31, 34). Much of the depressing effect of NO on force production is independent of cGMP and is thought to be the result of a direct action in which NO [or an NO intermediate such as S-nitrosoglutathione (GSNO)] decreases the Ca2+ sensitivity of the contractile apparatus. Application of NO donors S-nitroso-N-acetylcysteine and nitroprusside to intact fast-twitch muscle fibers of the mouse was found to reduce Ca2+ sensitivity by ~16% (3) (i.e., decrease the pCa50, the pCa at 50% force, by ~−0.063 pCa units). It is generally presumed that NO mediates its effect by S-nitrosylation of cysteine residues on particular contractile or regulatory protein, but the protein(s) and residues involved have not been identified (31, 34). Many hundreds of cysteine residues in skeletal muscle have been found to be able to undergo S-nitrosylation, including residues on most of the contractile and associated regulatory proteins (37). Interestingly, in the mammals examined to date (rodents and rabbits), NO donors decrease Ca2+ sensitivity in fast-twitch fibers (3, 11, 16) but not in slow-twitch fibers (35).
We have previously demonstrated that combined treatment with a cysteine-specific oxidant (2,2′-dithiodipyridine, DTDP) and then glutathione (GSH), or with oxidized glutathione (GSSG) alone, induces a large increase in Ca2+ sensitivity in mammalian fast-twitch (i.e., type II) muscle fibers but not in mammalian slow-twitch (type I) fibers, nor in type II fibers from chicken or toads (21, 24, 25). The effect was evidently due to S-glutathionylation of Cys134 on fast troponin I (TnIf) (24), which is present only in the mammalian fast-twitch fibers. The Ca2+ sensitivity increase could be induced by the oxidant-GSH treatment in slow-twitch fibers after insertion of fast-twitch troponin. There are a total of four cysteine residues in the three proteins of the fast-twitch troponin complex, three on TnIf and one on troponin C (TnCf), but only Cys134 on TnIf is readily accessible and reactive to oxidants in the troponin complex in situ (8, 15) (see note in materials and methods on updated cysteine numbering).
In the present study we first directly confirm by mass spectroscopy that Cys134 on TnIf is indeed S-glutathionylated by the oxidation-GSH treatment, and use the biotin-switch technique to show that Cys134 of TnIf can also undergo S-nitrosylation, consistent with mass spectroscopy findings (37). We consequently hypothesized that NO decreases Ca2+ sensitivity in skeletal muscle fibers by S-nitrosylation of Cys134 on TnIf. We tested this by examining 1) whether the Ca2+ sensitivity changes in various mammalian and non-mammalian muscle fibers are in accord with the presence of Cys134, 2) whether S-nitrosylation and S-glutathionylation treatments show competitive effects, similar sensitivity to irreversible block by N-ethylmaleamide (NEM), and a reciprocal ability to protect Cys134 and the Ca2+ sensitivity changes of the opposing treatment from block by NEM, and 3) whether NO treatment results in reduced Ca2+ sensitivity in slow-twitch mammalian fibers following exchange of fast-twitch troponin. The findings provide strong evidence that the NO decreases Ca2+ sensitivity in skeletal muscle by inducing S-nitrosylation of Cys134 on TnIf, and that this effect competitively antagonizes the ability of oxidant-induced S-glutathionylation to increase Ca2+ sensitivity, which would adversely affect skeletal muscle performance in various circumstances.1
MATERIALS AND METHODS
Ethical approvals and muscle fibers and muscle biopsies.
All animal experiments were carried out in accordance with the Australian National Health and Medical Research Council’s “Australian code of practice for the care and use of animals for scientific purposes,” and with approval of the La Trobe University Animal Ethics Committee. Male Long-Evans hooded rats (34 in total, 4 to 10 mo old) and male Sprague-Dawley (6 in total, 4 to 6 mo old) were supplied respectively by breeding facilities at Monash University Animal House, Melbourne Victoria, and the Animal Resources Centre, Canning Vale Western Australia; there was no apparent difference between results obtained from the two rat strains (e.g., pCa shift with DTDP-GSH and GSNO treatments ~+0.22 and −0.06 pCa units, respectively, in type II fibers in both strains) nor over the age range examined. The rats were housed (2 or 3 animals per cage) in the Central Animal House of La Trobe University and kept under controlled temperature (22°C) and a 12:12 h light-dark cycle, with food and water provided ad libitum. They were killed by overdose of isoflurane (4 % vol/vol) in a glass chamber, and then the EDL and soleus muscles were removed by dissection. Two tropical cane toads (Bufo marinus) (a pest species in Australia, caught by a Queensland supplier) were maintained at 15°C to lower their activity and then stunned and killed by pithing, and the iliofibularis muscle removed. Two female chickens (~24 wk, Hy-Line brown layer strain) from a commercial supplier were killed by overdose of intravenous phenobarbitone and a segment of pectoralis major muscle removed.
All procedures on human subjects were approved by the Human Research Ethics Committee at Victoria University. Informed consent was obtained in writing from all subjects and the studies conformed to the standards set by the Declaration of Helsinki. Skinned fiber experiments were performed with fibers from vastus lateralis muscle biopsies from three young male subjects (age 23 ± 6 yr; height, 174 ± 9 cm; body mass, 75 ± 5 kg, means ± SD). All subjects were healthy and participated in regular physical activity but were not specifically trained in any sport. After injection of a local anesthetic (1% lidocaine) into the skin and fascia, a small incision was made in the middle third of the vastus lateralis muscle of each subject and a muscle sample taken using a Bergstrom biopsy needle by an experienced medical practioner, as described previously (22). The excised muscle sample was rapidly blotted on filter paper to remove excess blood and placed in paraffin oil (Ajax Chemicals, Sydney, Australia) for fiber dissection (see below).
Preparations and force recording.
Skinned fiber and stock solutions.
All chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise stated. As previously described (21), the “relaxing” solution contained (in mM): 50 EGTA2−; 8 total ATP; 10 creatine phosphate (CrP); 36 Na+; 126 K+; 8.5 total Mg2+; 90 total HEPES; pH 7.1 and pCa >9. The maximum Ca2+-activating solution “max” contained 50 mM CaEGTA and had a pCa ~4.7, with total Mg2+ adjusted to maintain 1 mM free (see ref. 36 for apparent affinity constants). These two solutions were mixed in appropriate ratio to produce solutions with pCa in the range 6.7 to 4.7. All solutions had an osmolality of 295 ± 5 mosmol/kg. A similar strontium-based solution (with pSr = −log10[Sr2+] of 5.2) was made by mixing relaxing solution with a Sr-EGTA solution similar to the maximum Ca2+-activating solution. Exposure to a solution at pSr 5.2 was used to ascertain the predominant troponin C (TnC) isoform present (see refs. 22, 24, 27, 39). Subsequent Western blotting was used to confirm the TnC isoform present in all human, chicken, and toad fibers and some rat fibers.
All treatments were applied in relaxing solution at pH 7.1, except for GSSG treatment, which was applied in relaxing solution at pH 8.5. A 100 mM stock of reduced glutathione (GSH) was made in a potassium HDTA (hexa-methylene-diamine-tetraacetate) solution similar to the relaxing solution but with all EGTA replaced with HDTA; the pH of the stock was readjusted to 7.10 with KOH, and then diluted 20-fold to give 5 mM in the final solution. A 100 mM GSSG stock was made in relaxing solution at pH 8.5 and diluted 10-fold in the final solution. A 100 mM stock solution of 2,2′-dithiodipyridine (DTDP) was made in absolute ethanol and diluted 1,000-fold in the final solution to 100 µM; matching control solutions with the same amount of ethanol (0.1 %) had no noticeably different effect than controls without ethanol. Similarly, NEM was made as a 200 mM or 25 mM stock in ethanol and diluted 1,000-fold in the final solution. Dithiothreitol (DTT) was added to relaxing solution at 10 mM final concentration from a 1 M stock made in double-distilled water. Treatment with S-nitrosoglutathione (GSNO) can cause S-glutathionylation of protein thiols if applied immediately after dissolving the GSNO in solution (i.e., within 1 min) (see refs. 11, 24); this was specifically avoided in the present experiments and GSNO was only applied ~10 min after preparing the solution (called GSNOdel by Dutka et al. (11), which only resulted in S-nitrosylation. SNAP (S-nitroso-N-acetyl-penicillamine, 2 or 10 mM) was added to the final solution shortly before use (within 2 min). Ascorbate was made in a 100 mM stock in relaxing solution, with the pH adjusted to 7.10, and protected from strong light during storage and application. For biotin-labeling of cysteines, following an initial 5 min wash in relaxing solution and then various treatments as specified, individual mechanically skinned EDL fiber segments were treated for 1 min in relaxing solution with 100 µM EZ-Link Biotin-HPDP (Thermo Fisher Scientific), added from a 16 mM stock made in DMSO. All treatments were applied in relaxing solution (pCa >9).
Contractile apparatus experiments and analysis.
The force-Ca2+ relationship was determined in each fiber as previously described (21, 25) by exposing the skinned fiber segment to a sequence of solutions heavily buffered at progressively higher free [Ca2+] (pCa >9 to 4.7, the latter eliciting maximum force), and then the fiber was fully relaxed again in the relaxing solution. This procedure was performed twice before (“control”) and twice after each treatment to verify reproducibility and also gauge any small changes occurring with repeated activation and over time. Force produced at each [Ca2+] within a given sequence was expressed relative to maximum force generated in that same sequence, and analyzed by individually fitting a Hill curve to each sequence, for each fiber segment, using GraphPad Prism 4 software, yielding separate pCa50 value and h values (pCa at half-maximum force and Hill coefficient, respectively) for every case. Maximum force reached in each force-[Ca2+] sequence was expressed relative to the control level before any treatment in the given fiber, after correcting for the small decline occurring with each repetition of the force staircase (typically ~2 to 3% in EDL fibers), as gauged from the initial control repetitions in the given fiber (see also ref. 25). Similarly, there was a very small decrease in Ca2+ sensitivity with each staircase repetition (typically ~0.0015 pCa units); this effect was adjusted for when assessing the change in Ca2+ sensitivity with nitrosylation (GSNO or SNAP) or S-glutathionylation treatments, by averaging the change in pCa50 occurring with the treatment and that occurring when subsequently reversing the effect by DTT exposure.
Dissociation of myosin and myosin light chains.
To clearly distinguish S-glutathionylation of troponin I and myosin light chain 1 (MLC1) by Western blotting, skinned EDL fibers were subjected to S-glutathionylation treatment by a 1 min exposure to 100 µM DTDP followed by a 2 min exposure to 3.5 mM biotinylated glutathione ethyl ester (BioGEE, G36000, Thermo Fisher Scientific), and then any remaining free cysteines blocked by exposure to 5 mM NEM for 5 min (all agents applied in relaxing solution). The fiber was then treated with 1% Triton X-100 in relaxing solution for 5 min, washed for 5 min in relaxing solution, and then placed in 10 µl of relaxing solution with 500 mM KCl for 30 min (“wash solution”), to induce dissociation of myosin and associated myosin light chain proteins out of the skinned fiber (6). The fiber was then transferred to another 10 µl aliquot of the same solution and SDS added to both this solution and the wash solution, and these fiber and wash samples (containing the thin filaments and structural proteins, and myosin and myosin-associated proteins, respectively) were run in adjacent lanes on SDS-PAGE for Western blotting.
Western blotting of skinned fibers was performed on the entire fiber constituents without any fractionation, using nonreducing SDS-PAGE (see ref. 24). Each sample consisted of a single fiber segment (~2 mm in length) or a group of four or five segments (~60 µg wet wt), as specified. Where multiple skinned fiber segments were examined for S-glutathionylation, they were tied together with a silk suture, washed in relaxing solution for at least 5 min, and then transferred successively to the various treatment solutions as required. After the specified treatments, each fiber sample was placed in relaxing solution with 5 mM NEM for 5 min to block free sulfhydryl sites and then placed in nonreducing buffer for SDS-PAGE (final concentration: 125 mM Tris pH 6.8, 10% glycerol, 4% SDS, 0.01% bromophenol blue, 5 mM NEM). Proteins were separated on the specified percentage SDS-PAGE gel (details provided in figure legends, Criterion TGX gels were from Bio-Rad, Hercules, CA) and then wet transferred to nitrocellulose for 60 min at 100 V in a circulating ice-cooled bath with transfer buffer containing 140 mM glycine, 37 mM Tris-base, and 20% methanol. Membranes were then variously probed with anti-GSH (mouse monoclonal, 1 in 1,000, catalog no. 101-A, Virogen, Cincinnati, OH); streptavidin [polyclonal horseradish peroxidase (HRP), 1 in 20,000, catalog no. 21140, Pierce, Thermo Fisher Scientific, Australia); anti-TnI (rabbit polyclonal, 1 in 1,000, catalog no. 4002, Cell Signaling Technology, Danvers, MA), anti-TnC (rabbit polyclonal, 1 in 400, catalog no. sc-20642, Santa Cruz Biotechnology, Santa Cruz, CA), anti-myosin light chain 1 (mouse monoclonal, 1 in 1,000, catalog no. F310, Developmental Studies Hybridoma Bank (DSHB), Iowa City, IA), anti-actin (rabbit affinity isolated, catalog no. A2066, Sigma) all diluted in 1% bovine serum albumin in phosphate-buffered saline with 0.025% Tween. Following exposure to relevant HRP secondary antibodies and a series of washes in Tris-buffered saline with Tween, chemiluminescent substrate (SuperSignal West Femto, Pierce) was applied to membranes and Western blot images taken with ChemiDoc XRS or MP both fitted with a charge-coupled device (CCD) camera using Quantity One software (Bio-Rad). With the membrane position unchanged, the white light source was switched on to obtain an image of the pre-stained molecular weight markers on the membrane.
S-nitrosylation was examined using the biotin-switch technique (12, 17) (e.g., Fig. 8). Each sample consisted of three EDL skinned fiber segments tied together with silk suture. Each sample was subjected to a given sequence of timed treatments by transfer through a set of 0.5 ml Eppendorf tubes containing relaxing solution with specific reagents, with wash periods in relaxing solution between each treatment. All samples were first subjected to strong reducing treatment (100 mM DTT, 10 min) to reverse any resting level of S-glutathionylation or S-nitrosylation or other reversible oxidation states. As indicated in Fig. 8, samples were then S-nitrosylated by GSNO (2 mM, 10 min) or SNAP (2 mM, 10 min), with or without a preceding S-glutathionylation treatment by GSSG (10 mM at pH 8.5 for 15 min) or DTDP-GSH (100 µM DTDP for 5 min, followed by 5 mM GSH for 2 min). Samples were then subjected to NEM blocking treatment (20 mM, 20 min), and then S-nitrosylation detected by a 12-h exposure to relaxing solution with ascorbate (50 mM, pH 7.1) and HPDP-biotin (100 µM), protected from light. For a negative control, other samples were treated with DTT, then immediately blocked with NEM and labeled with HPDP-biotin for 12 h. Two positive controls were performed: one where the fibers were treated with DTT and then immediately labeled with HPDP-biotin, and the other where DTT (10 mM, 10 min) rather than ascorbate was used after the NEM block to reduce all reversible oxidation before labeling with HPDP-biotin (e.g., Fig. 8B). Finally, all samples were equilibrated for 2 min in relaxing solution with 5 mM NEM and then placed in nonreducing buffer for SDS-PAGE (final concentration: 125 mM Tris pH 6.8, 10% glycerol, 4% SDS, 0.01% bromophenol blue, 5 mM NEM). Proteins were then separated on 4–15% Criterion Stain Free gels, wet transferred to nitrocellulose, and probed for biotin with streptavidin, as above.
The troponin exchange experiments were performed on rat soleus type I fibers, as described in a preceding associated study (24). The skinned fibers were first mounted on the transducer, treated with Triton X-100 in relaxing solution (1% vol/vol) for 10 min, and then washed in relaxing solution. Troponin exchange was achieved by bathing the skinned fiber segment for 1 h in a low ionic strength rigor solution with zero Ca2+ and Mg2+ (mM: 2.5 EGTA; 2.5 EDTA; 10 HEPES, pH 7.1 with KOH) with porcine fast troponin (10 mg/ml, Sigma T2275) and 2 mM DTT.
Mass spectroscopy and proteomic analysis.
Proteomic experiments were performed in biological duplicate. Each sample consisted of 10 EDL fiber segments (~25 μg total protein) tied together with silk suture, which was permeabilized with Triton X-100 for 30 min (all treatments and washes in relaxing solution), washed, and treated with 10 mM DTT for 10 min, washed, and either subjected to the standard S-glutathionylation treatment (100 µM DTDP for 5 min followed by 5 mM GSH for 2 min) or left untreated (control). Samples were then blocked with 5 mM NEM for 2 min and placed in nonreducing SDS buffer with 5 mM NEM and separated using electrophoresis (12% nonreducing gel; 1.5M Tris-Cl pH 8.8, 1.5M Tris-Cl pH 6.7, 10% SDS, 30% acrylamide [Acryl:Bis = 37.5:1], 20% APS, TEMED). Individual bands corresponding to ~23 to 27 kDa were excised, destained (50 mM ammonium bicarbonate/acetonitrile), alkylated (50 mM iodoacetic acid for 30 min), and trypsinized [0.2 μg trypsin (Promega Sequencing Grade) for 16 h at 37°C]. Peptides were desalted using reverse-phase C18 StageTips, and eluted in 85% (vol/vol) acetonitrile (ACN) in 0.5% (vol/vol) formic acid (FA). A nanoflow UPLC instrument (Ultimate 3000 RSLCnano, Thermo Fisher Scientific, Scoresby, VIC, Australia) was coupled online to an LTQ Orbitrap Elite mass spectrometer (Thermo Fisher Scientific) with a nanoelectrospray ion source (Thermo Fisher Scientific). Peptides were loaded (Acclaim PepMap100, 5 mm × 300 μm i.d., μ-Precolumn packed with 5 μm C18 beads, Thermo Fisher Scientific) and separated (PepMapRSLC C18, 25 cm, 75 μm inner diameter, 2 μm 100Å, Thermo Fisher Scientific) with a 70-min linear gradient from 0 to 100% (vol/vol) phase B [0.1% (vol/vol) FA in 80% (vol/vol) ACN] at a flow rate of 250 nl/min operated at 45°C.
The mass spectrometer was operated in data-dependent mode where the top 20 most abundant precursor ions in the survey scan (350–1500 Th) were selected for MS/MS fragmentation. Survey scans were acquired at a resolution of 120,000. Unassigned precursor ion charge states and singly charged species were rejected, and peptide match disabled. Maximum injection time was 150 ms, automatic gain control (AGC) target 1 × 106, CID at 35% energy for a maximum injection time of 150 ms with AGC target of 5,000. Dynamic exclusion was activated for 30 s. Data were acquired using Xcalibur software (version 2.1; Thermo Fisher Scientific) and mgf (mascot generic file) files generated by Proteome Discoverer (version 2.1; Thermo Fisher Scientific).
Database searching and protein identification.
Database searches were performed using Mascot 2.4 (Walter and Eliza Hall Institute Mascot Server). Peptide lists were generated from a tryptic digestion with up to two missed cleavages, carbamidomethylation of cysteines as fixed modifications, and as variable modifications glutathione and N-ethylmaleimide. Precursor mass tolerance was 10 ppm, product ions were searched at 0.6 Da tolerances, and minimum peptide length defined at 6 aa.
Cysteine residue numbering.
Note that the reactive cysteine residue on TnIf was originally numbered as Cys133 in Mollica et al. (24) in accord with earlier references (8, 40), but here it is now denoted as Cys134 to fit with more recent numbering (15, 33).
Values are presented as means ± SD, with n denoting the number of fibers examined. Statistical significance (P < 0.05) was determined with Student’s two-tailed paired t-test.
We have previously shown that treating mammalian fast-twitch fibers with the sulfhydryl-specific agent DTDP and then with GSH, or just with oxidized GSSG alone, caused S-glutathionylation of a protein with apparent molecular mass of ~25 kDa, corresponding with TnIf (e.g., see Fig. 10 in Ref. 24). To demonstrate that this signal indeed reflected labeling of TnIf and not MLC1, which runs at a similar apparent molecular weight, these proteins were physically separated by dissociating myosin and the associated light chains with a high [KCl] solution, and then the dissociated proteins and those remaining in the fiber (i.e., the thin filaments and structural proteins) were run in adjacent lanes on SDS-PAGE for Western blotting (see materials and methods). This separation clearly showed that the DTDP-GSH treatment used here, which elicited a large increase in Ca2+ sensitivity, produced marked S-glutathionylation of TnIf and not of MLC1 (Fig. 1). Furthermore, mass spectroscopy of control and treated muscle fiber samples directly demonstrated that the DTDP-GSH treatment produced S-glutathionylation of Cys134 on TnIf (Fig. 2). No S-glutathionylation of Cys134 was found in the control samples, whereas the tryptic fragments containing Cys134 identified in the DTDP-GSH treated samples had all undergone S-glutathionylation, and interestingly were cleaved at Lys130 and not at Lys132 as found for the fragments in the control samples (see discussion).
Effects of S-nitrosylation treatments.
In accord with previous observations (11, 35), treatment of rat skinned fast-twitch (type II) fibers with the S-nitrosylating agents GSNO or SNAP caused a decrease in Ca2+ sensitivity by ~−0.06 to −0.08 pCa units with little or no change in maximum Ca2+-activated force (e.g., Fig. 3) (Tables 1 and 2). Similar effects were seen when treating fibers with a relatively low concentration of GSNO for a prolonged time (200 µM GSNO for 30 min) or with a higher concentration for a shorter time (2 mM for 2 min) (Table 1). The effect on Ca2+ sensitivity was largely or fully reversed by a 10 min exposure to 10 mM DTT, indicating involvement of cysteine residues, and also partially reversed by a 10 min exposure to ascorbate (2 to 50 mM) indicating the effects were due at least in part to S-nitrosylation of the cysteine residues (12); the mean ± SD increase in pCa50 upon ascorbate reversal was +0.044 ± 0.004, n = 4 for GSNO-treated fibers and +0.079 ± 0.014, n = 3 for SNAP-treated fibers, which corresponded to 59 ± 14% and 65 ± 7%, respectively, of the reversal found with DTT treatment in the same fibers (e.g., Fig. 3). Ascorbate treatment had no effect on Ca2+ sensitivity in fibers that had already been reduced by DTT (change in pCa50: −0.009 ± 0.009, n = 4).
As reported by Spencer and Posterino (35), the S-nitrosylation treatments had no significant effect on Ca2+ sensitivity in rat type I (slow-twitch) fibers (Table 1). Similarly, we found that in human vastus lateralis muscle fibers, SNAP treatment caused a decrease in Ca2+ sensitivity in type II (fast-twitch) fibers (change in pCa50: −0.090 ± 0.018, n = 10) and not in type I fibers (−0.009 ± 0.006, n = 3). In contrast, SNAP treatment had no effect on the Ca2+ sensitivity in type II fibers of either chicken or cane toad (Table 1). These findings are highly analogous to the effects of S-glutathionylation treatment, which was found to affect (increase not decrease) Ca2+ sensitivity in mammalian type II fibers and not in mammalian type I fibers nor in chicken or toad type II fibers; those effects were the result of S-glutathionylation of Cys134 on troponin I (24), which is present only in mammalian type II fibers (15, 40). To further explore this parallel, we examined whether the S-nitrosylation and S-glutathionylation treatments displayed competitive actions, which would be suggestive of a common site of action.
Competitive effects of S-nitrosylation and S-glutathionylation treatments.
After a rat type II fiber had been subjected to S-glutathiolynation treatment with GSSG, which increased Ca2+ sensitivity (e.g., Fig. 4), S-nitrosylation treatment with GSNO caused no further change in Ca2+ sensitivity (mean change in pCa50 +0.004 ± 0.004 pCa units, n = 3), but following subsequent reversal of the S-glutathiolynation with DTT, GSNO treatment once again elicited a substantial decrease in Ca2+ sensitivity (−0.073 ± 0.002, n = 3). A similar blocking effect was seen irrespective of whether GSSG or DTDP-GSH treatment was used as the S-glutathionylating pretreatment, or whether GSNO or SNAP treatment was used to elicit S-nitrosylation (data not shown).
Conversely, in rat type II fibers that had undergone S-nitrosylation treatment with SNAP (10 mM, 2 min), subsequent S-glutathionylation treatment with DTDP-GSH resulted in the Ca2+ sensitivity increasing to only a small net amount above the original starting level (final level +0.041 ± 0.001 pCa units, n = 3), which was only ~16% of the increase in Ca2+ sensitivity produced by the same S-glutathionylation treatment when the fibers had not been pretreated with SNAP. This indicates that the S-nitrosylation pretreatment with SNAP was sufficient to block most of the effect on Ca2+ sensitivity of this strong S-glutathionylation treatment (see refs. 21, 24). Furthermore, Western blotting of skinned EDL fibers for anti-GSH directly showed that S-nitrosylating pretreatment with SNAP largely prevented S-gluthionylation of TnIf (e.g., Fig. 5A).
Block by NEM.
The preceding experiments indicated that the S-glutathionylation and S-nitrosylation treatments each largely blocked the action of the other treatment on Ca2+ sensitivity. To further test whether this was because the two treatments both targeted the same cysteine residue(s), we examined the blocking action of N-ethylmaleimide (NEM) on each of the processes. We have previously shown that a 2 min application of a low concentration (25 µM) of the alkylating agent NEM largely prevents S-glutathionylation of TnIf, irreversibly blocking >85% of the normal Ca2+ sensitivity increase (24). We found here that the same NEM treatment (25 µM for 2 min) also blocked ~90% of the Ca2+ sensitivity shift to S-nitrosylation treatment by either GSNO or SNAP (right-hand column for treatments 3 to 6 in Table 2). Importantly also, S-nitrosylation pre-treatment prevented NEM from irreversibly blocking the effects of a subsequent S-glutathionylation treatment (applied after DTT reversal of the effects of the initial S-nitrosylation) (e.g., Fig. 6, A and B), and conversely, pretreatment by S-glutathionylation prevented NEM from blocking the effects of S-nitrosylation (e.g., Fig. 6C), irrespective of which specific S-nitrosylation and S-glutathionylation treatments were used (see summarized data in Table 2).
Furthermore, Western blotting of rat type II fibers with anti-GSH directly showed that NEM blocked S-glutathionylation of TnIf (Fig. 5B and also see ref. 24) and that pretreatment with SNAP prevented such NEM block (Fig. 5B). Moreover, NEM treatment blocked the ability of biotin-HPDP (a biotin-tagged analogue of DTDP) to react with and label TnIf at Cys134, and pretreatment with SNAP prevented this blocking action of NEM (e.g., Fig. 7). These data demonstrate that SNAP and NEM both competitively target the same cysteine residue on TnI that undergoes S-glutathionylation, namely Cys134.
Biotin-switch assay for S-nitrosylation.
The biotin-switch technique (12, 17) was used to directly identify proteins in the rat type II muscle fibers undergoing S-nitrosylation with the GSNO and SNAP treatments. This showed that both treatments caused significant S-nitrosylation of TnIf, as well as S-nitrosylation of the fast isoform of myosin light chain 1 (MLC1) (seen at ~25 kDa) and myosin light chain 3 (MLC3) (seen at ~16 kDa) (see Fig. 8A), and also another protein running at ~140 kDa which was probably myosin binding protein C (MyBPC) (not shown) (see ref. (37). Furthermore, it was found that the 10 min treatment with 2 mM GSNO produced a comparable level of S-nitrosylation of TnIf as that seen with the positive control treatment in which fibers were exposed to the biotin-labeling agent without any pretreatment with the S-nitrosylating and reducing agents (ascorbate or DTT) (e.g., Fig. 8B). This showed that the standard GSNO treatment produced very substantial S-nitrosylation of TnIf. Furthermore, S-glutathionylation pretreatment of the fibers with either GSSG or DTDP-GSH was found to markedly block the ability of the GSNO treatment to produce S-nitrosylation of TnIf (see Fig. 8B).
GSH treatment simply reverses S-nitrosylation of TnIf.
Given that Cys134 on TnIf evidently can undergo either S-glutathionylation or S-nitrosylation, with opposing effects on Ca2+ sensitivity of the contractile apparatus, an important further question was whether exposure of the S-nitrosylated residue (i.e., RSNO) to reduced glutathione (GSH) would result in it being reduced back to a free sulfhydryl (i.e., RSH) or instead becoming S-glutathionylated (i.e., RSSG). This was examined by first S-nitrosylating a rat type II fiber with SNAP or GSNO and then examining the effects of two successive 2 min exposures to GSH (5 mM). The GSNO treatment decreased Ca2+ sensitivity by −0.090 ± 0.016 pCa units (n = 4) and the first GSH exposure resulted in the Ca2+ sensitivity shifting back to its original level (+0.002 ± 0.003 pCa units relative to original control level) and the second GSH exposure caused no further change (+0.005 ± 0.007 pCa units relative to original). The GSH exposures following S-nitrosylation with SNAP had similar effect (final level: +0.004 ± 0.005 and −0.001 ± 0.04 pCa units relative to original level, n = 3). Given that it takes a very prolonged exposure to GSH (>20 min) to even partially reverse S-glutathionylation of TnIf (21), it can be concluded that when Cys134 on TnIf is S-nitrosylated, exposure to GSH simply reduces it back to a free sulfhydryl rather than causes it to undergo S-glutathionylation. It was also verified that Cys134 had been converted back to its reduced state by showing subsequent DTDP-GSH treatment elicited the normal S-glutathionylation effect on Ca2+ sensitivity. In two further cases, fibers were S-nitrosylated with SNAP and then exposed to 10 mM GSSG at normal pH (7.1), which had no effect on the Ca2+ sensitivity, indicating that Cys134 remained in its S-nitrosylated state and GSSG did not cause S-glutathionylation.
Effect of troponin exchange.
Finally, we examined whether exchanging the troponin complex in rat slow-twitch (type 1) muscle fibers with fast-twitch troponin affected the response of the fibers to S-nitrosylation treatment with SNAP. These experiments were done as part of the troponin exchange studies detailed in Mollica et al. (24), where ~35% of the TnI in the soleus fibers was exchanged with TnIf and which resulted in the type I fibers showing increased Ca2+ sensitivity upon S-glutathionylation treatment (~+0.13 pCa units vs. no change before exchange). After such partial exchange with TnIf, S-nitrosylation treatment with SNAP (10 mM, 2 min) resulted in a significant decrease in Ca2+ sensitivity in the three type I fibers examined (pCa50 decreasing by −0.026 ± 0.011 pCa units, compared with −0.001 ± 0.002 pCa units in untreated fibers, Table 1); this sample included the fiber shown in Fig. 7 of Mollica et al. (24) that displayed increased Ca2+ sensitivity to S-glutathionylation treatment after TnIf exchange, as well as a fiber in which SNAP treatment was also tested before the TnIf exchange and found to have no effect on Ca2+ sensitivity (zero change in pCa50).
The findings of this study provide compelling evidence that the action of NO in decreasing Ca2+ sensitivity in skeletal muscle is mediated by S-nitrosylation of Cys134 on TnIf. The evidence for this is as follows: 1) Cys134 on TnIf in mammalian fast-twitch muscle undergoes S-glutathionylation and S-nitrosylation respectively when muscle fibers are subjected to the specific S-glutathionylation treatments (DTDP-GSH or GGSG at pH 8.5) and S-nitrosylation treatments (GSNO or SNAP) used here (see mass spectroscopy results here in Fig. 2 and in Su et al. (37) and Figs. 1 and 8); 2) the decrease in Ca2+ sensitivity with S-nitrosylation treatment is seen only in mammalian fast-twitch (i.e., type II) fibers (e.g., rat, human, and rabbit), and not in mammalian slow-twitch (i.e., type I) fibers nor in toad or chicken type II fibers (see results and Table 1), in accord with the presence of Cys134 on TnIf; 3) the effects of S-nitrosylation treatment can be reversed with ascorbate (Fig. 3), a specific reversal agent (12); 4) S-glutathionylation treatment in mammalian fast-twitch fibers blocks S-nitrosylation of TnIf (Fig. 8) and its effect on Ca2+ sensitivity (Fig. 4), and S-nitrosylation treatment blocks S-glutathionylation of TnIf (Fig. 5A) and its effects on Ca2+ sensitivity (see results); 5) NEM irreversibly blocks both S-glutathionylation and S-nitrosylation with very similar efficacy (~90% block by 20 µM NEM for 2 min) (Table 2 and ref. (24); 6) S-nitrosylation pretreatment prevents NEM from blocking S-glutathionylation of TnIf (Figs. 5B and 7) and its effect on Ca2+ sensitivity (Fig. 6, A and B), and conversely S-glutathionylation pretreatment prevents NEM from blocking the effects of S-nitrosylation (Fig. 6C); and finally, 7) S-nitrosylation treatment decreases Ca2+ sensitivity in rat slow-twitch fibers after exchanging in fast-twitch troponin (see results).
A substantial proportion of cysteine residues in many proteins, including in skeletal muscle (37), can undergo S-nitrosylation, but fewer cysteines residues seemingly can undergo S-glutathionylation, and only a small subset are able to undergo both types of modification (4, 13, 14). The findings here indicate that Cys134 on TnIf is able to be S-glutathionylated or S-nitrosylated, and that these alternate modifications of the one cysteine residue have opposing functional effects, increasing or decreasing the Ca2+ sensitivity of contractile apparatus, respectively (e.g., Fig. 4). Cys134 is in the flexible and highly mobile COOH-terminal domain of TnIf, immediately adjacent to the “switch” region that binds to the hydrophobic pocket in the N-lobe of TnC in the Ca2+-bound state and swings back to be frequently near actin in the absence of Ca2+ (1, 32). S-glutathionylation effectively adds a negative charge at a cysteine residue, which together with the accompanying steric effects is thought to exert an action similar to that occurring with protein phosphorylation (9, 19). It seems that the net effect of S-glutathionylation at Cys134 is to bias the movements of the switch region towards the TnC bound state, so that the interaction between TnI and TnC controlling contractile activation occurs at lower cytoplasmic [Ca2+]. The effect of S-glutathionylation of Cys134 in altering TnIf’s tertiary structure is also directly indicated by the slower migration of the S-glutathionylated protein on SDS-PAGE (24) and the reduced susceptibility to trypsin cleavage of the nearby Lys132 residue, as apparent in our mass spectroscopy data (Fig. 2). S-nitrosylation of Cys134, on the other hand, decreases the Ca2+ sensitivity of contractile activation, indicating that its overall effect presumably is to bias the movements of the switch region of TnIf away from TnC.
In the native troponin complex in situ, Cys134 on TnIf is readily accessible and reactive both in the presence and absence of Ca2+ (8, 15), and hence readily susceptible to either S-glutathionylation or S-nitrosylation. The mass spectroscopy study of Su et al. (37) found that the GSNO treatment in mouse muscle homogenates also resulted in S-nitrosylation of Cys49 and Cys65 on TnIf and of Cys99 on TnCf. However, S-nitrosylation of these residues likely only occurred because the muscle homogenate was treated in a low ionic strength solution in the absence of any Ca2+ and Mg2+, conditions which induce dissociation of the troponin complex (see ref. 24), because all these three cysteine residues are normally inaccessible to modification in the native troponin complex (8, 15) (and note absence of S-nitrosylation of TnC in Fig. 8A).
Given that S-glutathionylation and S-nitrosylation of Cys134 have opposing functional effects, it was important to examine whether the millimolar levels of reduced GSH normally present in rested muscle fibers (18) might be expected to cause any S-nitrosylated residues to undergo S-glutathionylation (i.e., RSNO + GSH → RSSG + HNO) (see refs. 13, 19). It was instead found that the presence of GSH evidently converted S-nitrosylated Cys134 residues back to their reduced state (i.e., RSNO + GSH → RSH + GSNO), simply reversing the decrease in Ca2+ sensitivity rather than inducing a marked increase. It is interesting to relate this to the findings of Andrade et al. (3), where application of the NO donor S-nitroso-N-acetylcysteine (SNAC) to intact fast-twitch fibers of the mouse decreased Ca2+ sensitivity by ~−0.065 pCa units, which was reversed within 1 min simply by washout of the SNAC without application of any specific reducing treatment. In the skinned fibers examined in the present study, sensitivity changes elicited by S-nitrosylation or S-glutathionylation treatments remained unchanged indefinitely until specifically reversed by application of a reducing treatment, such as DTT or GSH. It seems that in the intact fiber experiments of Andrade et al. (3), the presence of extracellular SNAC gave rise to a steady influx of NO, resulting directly or indirectly in S-nitrosylation of Cys134 on TnIf, and that when the NO influx ceased upon washing away the SNAC, the remaining level of reduced GSH present in the fiber was sufficient to quite rapidly reverse the S-nitrosylation of TnIf. (It is possible that the reversal of the S-nitrosylation effects in the experiments of Andrade et al. (3) might also have been aided by hemolysis of the RS-NO bond by the ultraviolet light used for imaging intracellular Ca2+ in those fibers).
The above findings together offer important insight into the possible actions of NO in skeletal muscle fibers, indicating that if NO generation within a fiber were to markedly increase in some situation, it could be expected to readily target Cys134 on TnIf, not only directly decreasing Ca2+ sensitivity but also blocking any Ca2+ sensitivity increase to S-glutathionylation, preventing its beneficial effects in exercising muscle (2, 24). If the increase in NO were relatively small and brief, its inhibitory effects on Ca2+ sensitivity would probably be only transient, being quickly reversed by the normal reducing environment within the muscle fiber, whereas if the increase in NO were very large or prolonged it presumably would also perturb the redox environment of the fiber and the inhibitory effects of the NO may be long-lasting. However, if the increase in NO were preceded by an increase in reactive oxygen species, it is quite likely that Cys134 would have already undergone S-glutathionylation, and the Ca2+ sensitivity would be heightened and remain unaffected by the NO, particularly given that the reversal of S-glutathionylation of Cys134 is relatively slow (21).
The overall importance of NO’s direct effects on muscle force is currently unclear and likely differs considerably in various conditions. NO is generated in normal resting skeletal muscle and its production increases with muscular activity (5, 20, 30). Current data are equivocal as to whether the direct inhibitory effects of NO are a significant factor in normal muscle fatigue (30, 34), but this is a complicated question because NO not only has direct inhibitory effects on contractile function but it also has stimulatory effects on Ca2+ release from the sarcoplasmic reticulum (3, 29) and on the blood supply to the muscle. Furthermore, high or very prolonged levels of NO may not only decrease Ca2+ sensitivity but also reduce maximum force production of the contractile proteins by also acting on the myosin heads (26), including via the production and action of peroxynitrite (10, 38). Such inhibitory actions of NO are believed to play a major role in the decreased muscle function in hypoxia (28, 41) and in sepsis (7, 23). The levels of NO and GSNO applied in the present study were likely very much higher than the levels reached in-vivo, but they were applied for only relatively brief periods, and because these nitrosylating treatments caused little or no decrease in maximum force production it is apparent that they elicited reversible physiological alterations and not the irreversible pathological modifications of the contractile apparatus that occur in certain circumstances.
This study provides evidence that the direct inhibitory effect of nitric oxide on Ca2+ sensitivity in skeletal muscle is due to S-nitrosylation of Cys134 on TnIf. Significantly, this same site can undergo S-glutathionylation in the presence of oxidants and glutathione, which has the opposite functional effect, increasing the Ca2+ sensitivity of muscle contraction. Production of nitric oxide and reactive oxygen species both increase with muscle activation, as well as in hypoxia and in particular pathological conditions. Both S-nitrosylation and S-glutathionylation can have important protective effects, preventing irreversible oxidative damage (e.g., sulfonation) of key cysteine residues. S-glutathionylation of TnIf has been observed to occur in human muscle after prolonged cycling (24), and would be expected to be beneficial to muscle performance by compensating to some extent for the actions of the many metabolic factors that decrease contractile Ca2+ sensitivity with exercise (2). Given its opposing competitive effects, the extent of S-nitrosylation of Cys134 remains an important but unresolved issue, with its relative role likely being greater in specific exercise and disease conditions.
We thank the National Health and Medical Research Council of Australia for financial support (Grant no. 616 1051460).
No conflicts of interest, financial or otherwise, are declared by the author(s).
T.D., J.M., D.G., G.P., R.M., and G.L. conceived and designed the research; T.D., J.M., C.L., V.W., D.G., and R.M. performed experiments; T.D., J.M., C.L., V.W., D.G., R.M., and G.L. analyzed data; T.D., J.M., C.L., V.W., D.G., G.P., R.M., and G.L. interpreted results of experiments; T.D., J.M., V.W., D.G., and R.M. prepared figures; T.D., D.G., R.M., and G.L. edited and revised manuscript; T.D., J.M., C.L., V.W., D.G., G.P., R.M., and G.L. approved final version of manuscript; R.M. and G.L. drafted manuscript.
We thank Maria Cellini, Heidy Flores and Barney Frankish for technical assistance. Proteomics in this study was supported using the Mass Spectrometry and Proteomics facility, La Trobe University. D. W. G. was supported by the LIMS Molecular Biology Stone Fellowship, La Trobe University Research Focus Area Leadership Grant, and La Trobe University Start-up Fund. Proteomics in this study was supported using the Mass Spectrometry and Proteomics facility, La Trobe University.
- Copyright © 2017 the American Physiological Society