Activation of Ras and the Mitogen-Activated Protein Kinase Pathway Promotes Protein Degradation in Muscle Cells of Caenorhabditis elegans

To discover and study intracellular signals that regulate proteolysisin muscle, we have employed transgenic strains of Caenorhabditiselegans that produce a soluble LacZ reporter protein limitedto body-wall and vulval muscles. This reporter protein is stablein well-fed wild-type animals, but its degradation is triggeredupon a shift to 25°C in a strain carrying a temperature-sensitiveactivating mutation in the Ras oncogene homologue let-60. Thesemutants are not physiologically starved, inasmuch as growthrates are normal at 25°C. Ras-induced degradation is notprevented by the presence of cycloheximide added at or beforethe temperature shift and thus uses preexisting proteolyticsystems and signaling components. Furthermore, degradation istriggered when adult animals are shifted to conditions of 25°C,confirming that Ras acutely promotes protein degradation inmuscles whose developmental history is normal. Reduction-of-functionmutations in the downstream protein kinase Raf (lin-45), MEK(mek-2), or mitogen-activated protein kinase (MAPK) (mpk-1)prevent Ras-induced protein degradation, whereas activated MPK-1is sufficient to trigger degradation, indicating that this kinasecascade is the principal route by which Ras signaling triggersprotein degradation in muscle. This pathway is activated inhypodermal cells by the LET-23 epidermal growth factor receptorhomologue, but an activating mutation in let-23 does not promoteproteolysis in muscle. Starvation-induced LacZ reporter degradationis unaffected by reduction-of-function mutations in Ras, Raf,MEK, or MAPK, implying that Ras activation and starvation triggerproteolysis by mechanisms that are at least partially independent.This is the first evidence that Ras-Raf-MEK-MAPK signaling activatesprotein degradation in differentiated muscle.

INTRODUCTION

Top

Introduction

The regulation of proteolysis in muscle is a clinically significantproblem because of its relevance to muscle wasting and atrophyin skeletal muscle and pathological hypertrophy and atrophyin cardiac muscle. Muscle proteolysis is promoted by fastingor chronic starvation (18, 29), by disuse or denervation (19,31), in sepsis (10), and in cancer cachexia (53). Catastrophicloss of muscle protein is associated with increased mortalityin patients with cancer, diabetes, sepsis, heart failure, orAIDS (30, 52), and extreme muscle wasting can itself be theproximal cause of death.

The molecular mechanisms of proteolysis in muscle have beenextensively studied (3, 4, 15, 31-33, 44, 45, 51), althoughthere is considerable uncertainty remaining about the specificroles of the proteolytic systems involved. Moreover, severalexternal signals that regulate muscle proteolysis have beenidentified but we know little about how intramuscular signal-transductionmechanisms might couple external stimuli to the regulation ofprotein degradation.

We have been developing a genetically tractable model for studyingthe regulation of proteolysis in muscle, using strains of thenematode Caenorhabditis elegans that express a transgene-codedLacZ reporter protein specifically in body-wall and vulval muscles.This protein is completely stable in well-fed wild-type animals,but its degradation can be triggered by starvation (59) or byloss of cholinergic neural inputs (50). In both instances, theactivation of preexisting proteolytic systems is sufficientto account for the observed protein degradation. The reporterprotein is ubiquitinated in starved animals (59), and inhibitordata indicate that proteasome activity is required for its degradation(50).

Ras signaling is known to be important in a number of diseasestates associated with muscle wasting. We report here that proteolysisin muscle is triggered acutely by activation of the Ras homologueLET-60, that signaling occurs by the Raf-MEK-MAPK cascade ofprotein kinases, and that the resultant protein degradationuses a preexisting proteolytic system. We also show that reduction-of-functionmutations in this pathway do not affect starvation-induced ordenervation-induced protein degradation, implying that the variousmechanisms of stimulating proteolysis in muscle are at leastpartly distinct.

MATERIALS AND METHODS

Top

Materials and Methods

Nematode strains were maintained, grown, and roughly age synchronizedas described previously (59), except that strains containinglet-60(ga89^ts) were routinely kept at 16°C. The geneticmarkers used in this study were as follows: on linkage groupLG I, mek-2(ku114); on LG II, let-23(sa62); on LG III, mpk-1(n2521);on (LG) IV, let-60(ga89^ts), let-60(n1046), let-60(n2021), lin-45(sy96),him-8(e1489), cha-1(p1182^ts), dpy-13(e184), and unc-24(e138);on LG V, ccIs55(unc-54::lacZ); and on LG X, gap-1(n1329), gap-2(pe103),unc-2(e55), and unc-18(e81). A strain containing the integratedtransgene gaIs37 [EF1 ${alpha}$ -D-mek(gf) hs-mpk-1(wt)] IV was obtainedfrom S. Kim; at 25°C, the constitutively active DrosophilaMEK homologue activates wild-type C. elegans MPK-1 expressedfrom a heat shock promoter (28). Strains containing unlinkedmutations and/or the integrated transgene (59) ccIs55(unc-54::lacZ)on linkage group V or the transgene iwIs16(act-4::lacZ) on linkagegroup X (49) were constructed by conventional genetic methods.A double-mutant strain containing let-60(ga89^ts) and lin-45(sy96)on LG IV was obtained by constructing a lin-45/unc-24 let-60heterozygote, selecting (at 25°C) their nonuncoordinatedmultivulva (MuV) progeny (presumed lin-45 let-60/unc-24 let-60recombinants arising by crossing over between let-60 and unc-24),and then selecting their vulvaless (Vul) progeny (at 20°C)to obtain lin-45 homozygotes. The presence of suppressed let-60(ga89^ts)in this strain was confirmed by the production of MuV individualsat 25°C after outcrossing.

Methods for histochemical staining of ß-galactosidasewith 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside(X-Gal) were as described previously (59). Staining times were1 to 2 h at room temperature, governed by visual examinationof stained control animals (wild type or mutant at permissivetemperature) included with every experiment. Stained animalswere photographed on Kodak Ektachrome EH-160 transparency filmunder bright-field illumination with a red suppression filteror, in some later experiments, with a Nikon 990 digital camera.The automatically controlled exposure is dominated by the bright-fieldbackground and depends little on the intensity of staining ofthe animals being photographed and is thus approximately constantfrom experiment to experiment.

Methods for gel electrophoresis, fluorometric assay of ß-galactosidaseactivity, and immunoblotting and quantitation were as describedpreviously (59). Movement rates were measured as previouslydescribed (6), except that rates were measured with animalssuspended in liquid (70 mM potassium phosphate, 70 mM NaCl [pH7]). Optical measurement of worm growth was as described previously(6).

RESULTS

Top

Results

Ras activation promotes proteolysis in muscle. We followed proteolysis in muscle by using an UNC-54::LacZ reporterfusion protein (17) that forms tetramers with ß-galactosidaseactivity (59). The fusion protein contains only part of theN-terminal UNC-54 (myosin heavy chain) ATPase domain (17) andtherefore does not associate with myofibrils but remains solublein the cytosol.

To determine whether Ras signaling regulates muscle proteindegradation, we constructed strains containing the unc-54::lacZtransgene in combination with the let-60 Ras mutant allele ga89^ts.The ga89^ts mutant Ras has temperature-sensitive GTPase activity(16), such that Ras signaling is activated at the nonpermissivetemperature (25°C). Well-fed ga89^ts homozygotes expressedapproximately wild-type levels of LacZ fusion protein when grownat 16°C. To measure protein degradation without regard toany potential effects of Ras on transcription or translationof the unc-54::lacZ transgene, we grew animals to mid-adulthoodat 16°C, such that lacZ expression was essentially completed.Studies have shown previously (59) that the fusion protein iscompletely stable in well-fed wild-type animals (Fig. 1c), sothat any subsequent decline in LacZ levels (as measured by eitherß-galactosidase activity or immunoreactive protein)in a mutant strain must represent net degradation of the preexistingLacZ fusion protein. We found that activated-Ras (ga89^ts) homozygotes,when shifted to 25°C as adults, showed a time-dependentloss of preexisting reporter protein that did not occur in similarlytreated wild-type animals (Fig. 1). In these and all subsequentexperiments below, the amounts of 146-kDa reporter protein (Fig.1b, c, and d) and the levels of ß-galactosidase activity(Fig. 1c and d) declined in parallel.

View larger version (31K):
[in this window]
[in a new window]
FIG. 1. Ras activation promotes proteolysis in muscle. (a) Animals stained for ß-galactosidase activity (blue) after growth at 16°C (left) or 72 h after young adults grown at 16°C were shifted to 25°C (right) either with (+CH) or without (-CH) cycloheximide at 400 µg/ml. (b) Immunoblot analysis of 146-kDa fusion protein in 30-worm lysates. Each band shown from left to right in the two rows of bands is the 146-kDa band from the corresponding animal shown from left to right in panel a. (c) Kinetics of loss of ß-galactosidase activity at 16°C (open circles and squares) or at 25°C (filled circles and squares) in wild-type (circles) or let-60(ga89^ts) (squares) homozygotes. Diamonds show decline of protein in ga89^ts after a shift to 25°C, measured by Western blot analysis similar to that described for panel b. (d) Quantitation of 146-kDa fusion protein (open bars) by integration of the bands shown in panel b for the 72-h group and of ß-galactosidase activity (striped bars) by fluorometric assay of 10-worm lysates. Data are means ± standard deviations of triplicate samples.

In a strain containing activated-Ras (ga89^ts) and ccIs56, anunc-54::lacZ fusion construct identical to ccIs55 but with theaddition of a simian virus 40 nuclear localization signal (17),we observed ß-galactosidase initially in both musclecytosol and nuclei. At 72 h after a shift to 25°C, therewas no histochemically detectable ß-galactosidaseactivity in cytosol but some was still present in the nuclei(R. Harper and L. Jacobson, unpublished data). It was not possibleto tell if any nuclear fusion protein had been degraded, butevidently nuclear fusion protein is at least somewhat protectedfrom degradation relative to the same protein in the musclecytosol.

We have also examined in a more limited way the effect of activatedRas on the stability of an ACT-4::LacZ fusion encoded by theintegrated transgene iwIs16 (49). In contrast to the much largerN-terminal moiety on the UNC-54::LacZ fusion protein, the ACT-4::LacZfusion protein adds only the first 14 amino acid residues ofACT-4 (actin) to the N terminus of Escherichia coli LacZ andis expressed from the act-4 promoter in 95 body-wall musclecells, the eight vulval muscles, and the spermathecae. The ACT-4::LacZfusion was also completely stable in wild-type animals and wasdegraded (in muscles but not the spermathecae) in activated-Ras(ga89^ts) homozygotes after a shift to 25°C, with kineticsapproximately the same as those we observed for the UNC-54::LacZfusion. This suggests that the in vivo stabilities of the fusionproteins are determined primarily by the large LacZ regions(1,015 residues) rather than by the smaller N-terminal fusionregions. Furthermore, we examined a strain expressing soluble,cytosolic green fluorescent protein from an unc-54 promoter(but with no added amino acid residues) and, after activationof ga89^ts Ras at 25°C, observed time-dependent loss of greenfluorescent protein fluorescence in body-wall muscle over 72h. Thus, these various reporter proteins provided us with informationabout a Ras-induced proteolytic process in body-wall musclecells that is not specific for LacZ.

We infer that protein catabolism is provoked by activation ofRas signaling rather than by some other peculiar property ofga89^ts mutant Ras, because protein is also degraded in otherstrains with Ras activation. We observed (Fig. 2) protein degradationin strains homozygous for a constitutively active (47) let-60allele (n1046) and in strains with increased Ras activationas the result of reduction-of-function mutations in the RasGTPase-activating protein gene (20, 21) gap-1 or gap-2. Additionally,the gap-1 reduction-of-function mutation enhanced protein degradationat 25°C in a strain carrying let-60(ga89^ts). It is notablethat the gap-1 and gap-2 single mutants do not show the MuVphenotype (20, 21) characteristic of strong Ras activation (47).

View larger version (24K):
[in this window]
[in a new window]
FIG. 2. Ras activation, but not LET-23 EGFR activation, promotes degradation. (a) Adult animals grown at 20°C were stained for ß-galactosidase activity. (b) ß-Galactosidase activity as a function of age of wild-type and let-23(sa62)-activated EGFR animals. Age-synchronized L1 larvae were grown for 84 h at 16°C and then shifted to 25°C at the 0 time point. Data represent means ± standard deviations of fluorometric assays of triplicate 10-worm samples.

In the course of this and other studies in our laboratory, wehave examined a wide variety of C. elegans mutants for expressionand degradation of this UNC-54::LacZ fusion protein. It waspreviously reported (50) that LacZ degradation (in starved animalsonly) is promoted by mutations affecting the presynaptic synthesisor release of acetylcholine or the action of postsynaptic acetylcholinereceptors; however, many other mutations that affect morphology(e.g., dumpy, blister, roller, etc.) or behavior (e.g., uncoordinated,chemotaxis defective, egg laying defective, etc.) have littleor no effect on LacZ expression or stability as assessed byhistochemical staining.

Ras activation induces changes in gene expression through theaction of various downstream effectors, which in many casesact to promote the transcription of specific genes. To determineif Ras activation stimulates proteolysis by inducing de novogene expression, we followed protein degradation in ga89^ts mutantanimals treated with cycloheximide. Studies have previouslyshown that cycloheximide at this concentration blocks proteinsynthesis within a few minutes at most but does not affect therate of protein degradation in muscle (59). If Ras-induced proteolysisrequired the participation of newly synthesized proteins (e.g.,signaling proteins, proteases, etc.), then cycloheximide shouldblock the Ras effect and the LacZ reporter protein should remainstable after a shift to 25°C. To the contrary, we observedthat onset of Ras-induced protein degradation was not blockedby cycloheximide added at the time of a shift to 25°C (Fig.1), implying that stimulation of proteolysis does not requirethe synthesis of new proteins but is mediated by preexistingsignaling proteins and proteases. Pretreatment with cycloheximidefor 6 h prior to a shift to 25°C did not alter this result(data not shown), mitigating against the possibility that relevantRas-induced gene expression takes place in a (hypothetical)brief period before protein synthesis is blocked by cycloheximide.Cycloheximide-treated wild-type controls showed little or noLacZ degradation over the same time interval (Fig. 1) (59).

The Ras effect is physiological rather than developmental. Ras activation is widely known to affect the developmental fatesof specific cells in C. elegans (47) as well as in other organisms(8, 11, 39). To distinguish developmental effects of Ras onthe muscle cells from acute physiological effects on proteindegradation in those cells, we performed temperature upshiftand downshift experiments with let-60(ga89^ts) homozygotes. Figure3 shows that animals shifted from 25 to 15°C conditionsat or prior to adulthood retained full ß-galactosidaseactivity and that animals shifted from 15 to 25°C conditionswell after adulthood still lost significant activity. In otherwords, early Ras activation at 25°C does not provoke laterprotein degradation if the mutant animals are returned to thepermissive temperature (15°C) prior to reaching adulthood.Conversely, mutant animals raised to adulthood at 15°C,and thus having a near-wild-type experience during development,can still be provoked after reaching adulthood to degrade muscleprotein when Ras is activated by a shift to 25°C. Theseobservations imply that Ras activation promotes protein degradationby acutely stimulating proteolysis rather than by altering somedevelopmental process in the muscles. This is entirely consistentwith the inference from the cycloheximide experiments (Fig.1) that Ras activation causes signal to pass via preexistingproteins to a preexisting proteolytic system.

View larger version (21K):
[in this window]
[in a new window]
FIG. 3. Ras activation effects on ß-galactosidase in muscle are acute rather than developmental. Ras mutant [let-60(ga89^ts)] animals were grown at 15°C or 25°C, and L1-L2 animals at each temperature were then transferred to nine new plates. At 0, 8, 24, 32, 53, 57, 68, 72, and 80 h after transfer, one plate was moved from 15°C to 25°C and one was moved from 25°C to 15°C. At 119 h after transfer, 50 animals from each plate were stained for ß-galactosidase activity and the stained animals were examined to score the fraction showing full-body staining (ordinate). Times of temperature shift were normalized to growth at 20°C (7) to generate the abscissa.

The data in Fig. 3 show that there is some variability betweenindividual animals in the rate or extent of protein degradation,and ga89^ts individuals also show only a moderate incidence ofthe MuV phenotype resulting from ectopic differentiation ofhypodermal cells (16) in response to Ras activation. That is,both the MuV and protein-degradation phenotypes display variableexpressivity and perhaps variable penetrance. However, we foundthat those ga89^ts individuals that are MuV did not degrade theLacZ reporter any more than their non-MuV siblings and, conversely,those ga89^ts individuals with the most pronounced LacZ degradationwere not more likely to be MuV. We interpret this lack of correlationbetween activated-Ras phenotypes in hypodermis and muscle tomean that individuals differ not only in organism-wide levelof Ras activation but also in the efficiency with which Rassignal affects the ultimate phenotypic output. Animals showingthe MuV phenotype differ from each other in their numbers ofectopic differentiation events, which is indicative of cell-to-cellvariability. In contrast, we did not observe much variationbetween individual muscle cells in the same activated-Ras animalin the extent of protein degradation (although there was variationbetween individual animals), suggesting a systemic rather thancell-to-cell variation in downstream execution of events leadingto the phenotype.

The ectopic differentiation of hypodermal cells to produce theMuV phenotype is efficiently signaled by activation of the LET-23receptor (2, 48), a homologue of mammalian epidermal growthfactor receptor (EGFR) which signals via Ras (47). To test whetherLET-23 also signals to Ras to induce muscle proteolysis andto further examine the lack of correlation between the Ras-inducedprotein-degradation and MuV phenotypes, we constructed a straincarrying the unc-54::lacZ transgene and an activating mutationin the extracellular domain of LET-23 (25). This mutant strain[homozygous for let-23(sa62)] showed high penetrance and expressivityof the MuV phenotype (Fig. 2a) but on the basis of histochemicalstaining did not obviously degrade the LacZ reporter. Quantitativefluorometric assaying of ß-galactosidase activity(Fig. 2b) showed that the ß-galactosidase levels inthe activated-LET-23 mutants were about 50 to 70% of those inwild-type animals of similar chronological age; this is approximatelyproportionate to the reduced body size of the mutants (Fig.2a), so the lower activity on a per-animal basis is likely toreflect only generally reduced growth and protein synthesisrather than increased protein degradation as a result of LET-23activation. Furthermore, there was no loss of LacZ reporterprotein as assayed over time by Western blotting and reporteractivity did not disappear in let-23(sa62) animals treated withcycloheximide. We conclude that the LET-23 EGFR does not signalto cause Ras activation in muscle, consistent with the factthat no LET-23 expression has been reported in muscle (23).

Raf, MEK, and MAPK are downstream Ras effectors. Ras effects may be mediated by a wide variety of downstreameffectors (8); the effect on hypodermal development in C. elegansinvolves signal transduction by the Raf-MEK-MAPK cascade ofprotein kinases (47). To determine whether Ras signals proteindegradation in muscle by the same pathway, we constructed strainscontaining the unc-54::lacZ transgene and temperature-activatedlet-60(ga89^ts) in combination with mutant alleles that producereduction of function in the protein kinases Raf [lin-45(sy96)],MEK [mek-2(ku114)], or MAPK [mpk-1(n2521)]. In each case, amutation in Raf, MEK, or MAPK efficiently suppressed Ras-inducedprotein degradation after temperature upshift (Fig. 4). Ras-inducedprotein degradation in ga89^ts mutant animals was also preventedby 1 µM PD-98059, a pharmacological inhibitor of MEK activity(1, 14). Conversely, mutational activation of a MAPK transgene[EF1 ${alpha}$ -D-mek(gf) + hs-mpk-1(wt)] (28) was sufficient to provokeprotein degradation (Fig. 4). Note that in this case there wasan obvious lag between the shift to 25°C and the onset ofreporter protein degradation, as would be expected if degradationwere triggered only after MPK-1 had been expressed de novo fromthe heat shock promoter.

View larger version (41K):
[in this window]
[in a new window]
FIG. 4. Reduction-of-function mutations in downstream effectors suppress Ras-induced protein degradation. (a) Histochemical stain for ß-galactosidase activity in muscle. (b) Immunoblot analysis of 146-kDa fusion protein in 30-worm lysates. Each row of bands shows the 146-kDa lacZ fusion bands from the corresponding double-mutant animal shown in panel a. (c) Quantitation of 146-kDa fusion protein (open bars) by integration of the bands at 72 h (gel in panel b) and of ß-galactosidase activity (filled bars) by fluorometric assay of 10-worm lysates made 72 h after a shift to 25°C. Data represent means ± standard deviations of triplicate samples.

We infer that the Raf-MEK-MAPK cascade is the principal routeby which Ras signaling promotes protein degradation in muscle.The relevant downstream target of MAPK action is presently unknown(the LacZ fusion protein itself contains no recognizable MAPKphosphorylation site), but the cycloheximide data (Fig. 1) implythat the known outputs involving activation of transcriptionfactors (47) are insufficient to explain how Ras induces proteolysis.

Ras activation does not cause starvation or loss of neural function. Five lines of evidence indicate that protein degradation inactivated-Ras mutants is not the result of starvation. First,ga89^ts animals raised at 25°C grew at a rate indistinguishablefrom that of wild-type animals (Fig. 5), whereas starved animalsshow arrested development or markedly decreased growth rate.Second, ga89^ts animals raised at 25°C do not enter the dauerlarva state associated with poor nutritional status (9). Third,ga89^ts animals did not catabolize lipid stored in the intestinalcells after a shift to 25°C (as judged by staining withSudan Black [reference 27 and data not shown]), whereas thesestores are completely depleted after about 8 h of acute starvation(S. J. Barmada and L. A. Jacobson, unpublished data). Fourth,ga89^ts animals degraded LacZ protein in all muscles at 25°C(Fig. 1), whereas in starved wild-type animals (or starved ga89^tsmutants at 16°C) the protein is degraded in the posteriorbody-wall muscles but preserved in head and vulval muscles (50).Fifth, Ras-induced reporter degradation was significantly slower(50% degradation in about 72 h at 25°C; Fig. 1) than isstarvation-induced degradation (50% degradation in 16 h at 20°C[59]). This kinetic discrepancy is large enough to precludeexperiments to determine whether Ras-induced and starvation-inducedprotein degradation would be quantitatively additive.

View larger version (16K):
[in this window]
[in a new window]
FIG. 5. Growth rates of wild-type (filled diamonds) and Ras mutants [let-60(ga89^ts); open squares] at 25°C. Eggs were laid on plates for 2 h; the 0 time point represents the end of this period. Data show means ± standard deviations of results from 6 animals at each time point.

We have so far found no evidence that relates Ras-Raf-MEK-MAPKsignaling to starvation-induced degradation of protein in muscle.Reduction-of-function mutations in Ras [let-60(n2021)], Raf,MEK, or MAPK affected neither the rate nor the anterior-posteriorcellular specificity of starvation-induced LacZ degradation(Fig. 6). Thus, activation of the Ras-MAPK pathway is not necessaryfor starvation-induced protein degradation.

View larger version (11K):
[in this window]
[in a new window]
FIG. 6. Reduction-of-function mutations in mek-2 or mpk-1 do not block starvation- or denervation-induced proteolysis. Histochemical stains for ß-galactosidase activity in muscle are shown. Starvation of adult animals was for 24 h at 20°C as described previously (59). At 25°C the cha-1(p1182^ts) mutation blocks acetylcholine synthesis; these animals were essentially wild-type at 20°C in both movement and protein degradation. The mek-2 mutants appear short because of a dpy-13 mutation in the background.

There was a rough correlation between Ras-induced protein degradationand a progressive loss of mobility (Table 1) in the mutant animalsafter temperature shift. This does not appear to result fromproteolysis of myofilament proteins leading to major structuraldefects, since myofibrillar structure appeared grossly normalafter staining with Oregon Green phalloidin (data not shown).Soluble proteins (including the LacZ fusion) and myofilamentproteins are likely to be degraded by distinct routes (44);thus, the loss of movement may reflect degradation of solublemuscle enzymes needed to power contractions. Alternatively,the mobility defect may be unrelated to protein degradationin muscle.

View this table:
[in this window]
[in a new window]
TABLE 1. Ras activation leads to time-dependent loss of mobility^a

It seems unlikely that Ras activation affects muscle simplyby provoking functional denervation (50), since Ras-inducedproteolysis was not prevented by the nicotinic agonist levamisole,which does prevent denervation-induced proteolysis (50). Conversely,neither mek-2 nor mpk-1 reduction-of-function mutations preventedthe protein degradation induced by disruption of cholinergicsignaling in a cha-1(p1182^ts) mutant at nonpermissive temperature(Fig. 6), implying that MEK-MAPK signaling is not required fordenervation-induced proteolysis. However, these observationsdo not exclude the possibility that signals from denervationand Ras activation converge at some downstream effector.

We presently believe that starvation and Ras activation triggerdifferent proteolytic mechanisms. We have previously shown thatin starved or denervated animals, the LacZ reporter proteinbecomes conjugated to ubiquitin (59) and that its subsequentinactivation is prevented by inhibitors of the proteasome (50).By contrast, the protein degradation provoked by Ras activationwas not prevented by proteasome inhibitors Z-leu₃-CHO (Fig.7) or clasto-lactacystin-ß-lactone (not shown), implyingthat Ras-induced degradation occurs by a mechanism independentof proteasome activity. Furthermore, in starvation-induced degradationit is characteristic that 146-kDa protein (as detected on Westernblots) is lost more rapidly than is ß-galactosidaseactivity, because proteolytic fragments remain noncovalentlyassociated in active tetramers (59). By contrast, our data indicatethat during Ras-induced reporter degradation there was a verygood correlation between the disappearance of 146-kDa proteinand that of activity (Fig. 1), perhaps implying that early proteolyticcleavages result in loss of activity. Taken together, theseobservations thus suggest that Ras-induced reporter degradationoccurs by a mechanism distinct from that provoked by starvationor denervation.

View larger version (18K):
[in this window]
[in a new window]
FIG. 7. Proteasome inhibitor Z-leu₃-CHO does not prevent Ras-induced protein degradation. Mutant strains [cha-1(p1182^ts) or let-60(ga89^ts)] were grown at 16°C and shifted as young adults to 25°C either with or without 200 µM Z-leu₃-CHO (MG-132; BIOMOL Research Laboratories) and then stained with X-Gal 68 h later. The cha-1^ts mutants were unable to synthesize acetylcholine at 25°C and were both genetically denervated and starved because they ceased feeding.

DISCUSSION

Top

Discussion

We have shown that Ras activation acutely provokes the degradationof reporter proteins in muscle by signaling to the Raf-MEK-MAPKcascade and that Ras-induced proteolysis is largely distinctfrom starvation-induced or denervation-induced proteolysis insignal transduction and probably in the proteolytic system(s)that is triggered.

There is certainly no reason to suppose that the degradationof the UNC-54::LacZ fusion protein in any way represents thecatabolism of normal UNC-54 myosin heavy chains. The N-terminalUNC-54 fragment represents only a portion of the myosin ATPasedomain (263 residues of the 1,966 present in full-length UNC-54)and is evidently insufficient to support assembly of the fusionprotein into myofilaments. It was reported (59) that early instarvation-induced degradation, this region may be separatedfrom the LacZ domain by a set of proteolytic cleavages. Ourphalloidin staining results indicate that myofibrils are largelyresistant to the proteolytic processes that degrade the reporterproteins. We believe that these soluble reporter proteins areinstead reporting on the processes that are most likely to affectthe soluble proteins of the muscle cytosol. There are few reliableidentifications of endogenous muscle-specific, soluble cytosolicproteins, and few tools are available for studying them. Experimentsto identify other endogenous substrates among soluble muscleproteins are presently in progress in our laboratory.

There are some caveats to the use of fusion proteins as reporters,one of which is that if misfolded, they may enter a misfolded-proteindegradation pathway (37). However, it is very unlikely thatthe reporter proteins we have studied are degraded simply becausethey are recognized as misfolded. There is no direct way toassess the conformation of the N-terminal UNC-54 fragment, butwe infer that the C-terminal 1,015-residue portion of the UNC-54::LacZreporter protein achieves a native or near-native ß-galactosidaseconformation because the protein-protein subunit contacts areproperly formed to promote association into tetramers (59) andthe assembled tetramers have ß-galactosidase activity.The crucial point is that in the context of well-fed wild-typeanimals, this fusion protein is completely stable (59) and thusdoes not by default enter a misfolded-protein degradation pathwayby virtue of its structure alone. Even after Ras activationits degradation is not rapid, proceeding at a rate four- tofivefold lower than in starved animals.

Neither is it plausible that these fusion proteins are somehowidentified as foreign and consequently singled out for degradationby a proteolytic system that has no endogenous protein substrates.It is extremely difficult to imagine how such discriminationmight be achieved at the molecular level; indeed, even the best-knownexamples of enzymes that selectively degrade foreign molecules,the DNA restriction endonucleases, do not differentially identifyforeign DNA molecules at all but instead ignore resident DNAbecause its recognition sites have been methylated. Thus, thetransgene-coded proteins in our studies are probably reportingon the activity of a protein-degradation system that has naturallyoccurring protein substrates in muscle cells. In many of themost familiar instances of Ras-Raf-MEK-MAPK signaling, downstreameffectors are phosphorylated by MAPK to trigger an increasein transcription leading to either a developmental decision(e.g., the differentiation of hypodermal precursor cells toform the C. elegans vulva [47]) or a more general increase inbiosynthetic activity (e.g., increased cell proliferation aftertransformation). We have consequently become accustomed to thinkingof Ras signaling as having fundamentally anabolic consequences,and it thus seems paradoxical at first that Ras-MAPK signalingshould also promote the catabolism of existing proteins.

There are some cases in mammals where Ras-MAPK signaling hasbeen implicated in promoting proteolysis of specific proteins,such as cell cycle regulators cyclin D1 (42), cyclin-dependentkinase inhibitor p27Kip1 (26), and myc (41). It has also beenreported that matrix metalloprotease (22) and cathepsin (43)activities are upregulated by Ras signaling. However, this isthe first report that Ras-MAPK signaling promotes protein degradationin differentiated muscle.

The body-wall muscle of C. elegans is most nearly similar tothe striated skeletal muscle of vertebrates (55). A signalingprocess in mammalian skeletal muscle analogous to that we haveuncovered in C. elegans muscle is suggested by the observations(38) in cultured mouse myotubes that expression of activatedRaf promotes atrophy and, conversely, that expression of dominant-negativeRaf promotes hypertrophy. It is important to note, however,that these cellular responses were not shown directly to reflectRaf-responsive change in the rate of proteolysis.

Signaling by the Ras-MAPK pathway opposes the conversion ofcommitted mammalian myoblasts into terminally differentiatedskeletal muscle (13, 35, 54), and Ras activation is classicallyassociated with transformation or dedifferentiation (11), althoughit is not clear that the same Ras effectors mediate all theseeffects (56). On the other hand, it is evident that in orderto lose its differentiated phenotype(s) and structures, a cellnot only must synthesize new proteins but also must degradeat least some key proteins characteristic of the differentiatedstate (e.g., myoD in muscle [34, 46]). A similar argument mayapply to remodeling of differentiated tissue. Protein degradationin skeletal muscle is increased by heavy exercise (5, 12, 24,36), which also promotes an increase in MAPK activation (40,57, 58). Thus, our results raise the intriguing possibilitythat the stimulation of proteolysis by Ras-MAPK signaling isan important feature of transformation and/or remodeling inmammalian muscle.

It is also worth considering more generally whether the temporalsequence of stimulation of protein synthesis and of proteindegradation, and/or the balance between these processes, determineswhether Ras signaling results in catabolic, anabolic, or remodelingeffects. It may be significant in this context that Ras-stimulatedprotein degradation appears to involve the activation of a preexistingproteolytic system rather than to depend upon the synthesisof new proteins. This implies that the earliest response toRas activation might be an increase in degradation of existingproteins and that a lag would likely intervene before Ras-promotedstimulation of transcription resulted in the eventual synthesisand accumulation of new proteins. A temporal separation of thiskind would be well suited to facilitating a change in differentiatedstate or significant remodeling of a relatively plastic tissuelike muscle.

ACKNOWLEDGMENTS

This work was supported by NSF grants MCB-9630841 and MCB-0090734.

We are grateful to A. Fire, D. Eisenmann, S. Kim, A. Golden,J. Shaw, and M. Han for generously providing nematode strains,M. Sundaram for useful discussions, and R. Harper for technicalassistance. Many strains were obtained from the CaenorhabditisGenetics Center, which is supported by the NIH National Centerfor Research Resources.

FOOTNOTES

* Corresponding author. Mailing address: Department of Biological Sciences, University of Pittsburgh, Pittsburgh, PA 15260. Phone: (412) 624-4647. Fax: (412) 624-4759. E-mail: LJAC{at}pitt.edu.

REFERENCES

Top