If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Dept. of Molecular and CellularPhysiology, Tufts University School of Medicine, Arnold 809, 136 HarrisonAve., Boston, MA 02111. Tel.: 617-636-0408; Fax: 617-636-0445;
* The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked“advertisement” in accordance with 18 U.S.C. Section 1734solely to indicate this fact.
Chaperone-mediated autophagy (CMA) is a selective lysosomal proteindegradative process that is activated in higher organisms under conditions ofprolonged starvation and in cell culture by the removal of serum. Ketonebodies are comprised of three compounds (β-hydroxybutyrate, acetoacetate,and acetone) that circulate during starvation, especially during prolongedstarvation. Here we have investigated the hypothesis that ketone bodies induceCMA. We found that physiological concentrations of β-hydroxybutyrate(BOH) induced proteolysis in cells maintained in media with serum and withoutserum; however, acetoacetate only induced proteolysis in cells maintained inmedia with serum. Lysosomes isolated from BOH-treated cells displayed anincreased ability to degrade both glyceraldehyde-3-phosphate dehydrogenase andribonuclease A, substrates for CMA. Isolated lysosomes from cells maintainedin media without serum also demonstrated an increased ability to degradeglyceraldehyde-3-phosphate dehydrogenase and ribonuclease A when the reactionwas supplemented with BOH. Such treatment did not affect the levels oflysosome-associated membrane protein 2a or lysosomal heat shock cognateprotein of 70 kDa, two rate-limiting proteins in CMA. However, pretreatment ofglyceraldehyde-3-phosphate and ribonuclease A with BOH increased their rate ofdegradation by isolated lysosomes. Lysosomes pretreated with BOH showed noincrease in proteolysis, suggesting that BOH acts on the substrates toincrease their rates of proteolysis. Using OxyBlot™ analysis to detectcarbonyl formation on proteins, one common marker of protein oxidation, weshowed that treatment of substrates with BOH increased their oxidation.Neither glycerol, another compound that increases in circulation duringprolonged starvation, nor butanol or butanone, compounds closely related toBOH, had an effect on CMA. The induction of CMA by ketone bodies may providean important physiological mechanism for the activation of CMA duringprolonged starvation.
For eukaryotic cells to maintain homeostasis, a balance between proteinsynthesis and proteolysis must exist. This balance is achieved in a variety ofways. Cells have the ability to control protein synthesis by regulatingtranscription and translation(
). Autophagy, literally“self-eating,” is a cellular process that allows cells to removeproteins, organelles, and foreign bodies from the cytosol and deliver them tolysosomes for degradation.
CMA is a process activated during long term starvation in which cellsselectively degrade proteins in order to recycle their amino acids or use themfor energy. During nutrient deprivation, substrates that contain a consensusmotif related to KFERQ (
Ketone bodies are produced by the liver during long term starvation inresponse to rapid lipolysis. Ketone bodies can be utilized by muscle andcontribute to the preservation of muscle mass during prolonged survival; theycan also be used as an energy source for the brain(
). We also show that theincrease in proteolysis observed is at least in part due to the stimulation ofCMA. Lastly, we show that ketone bodies induce CMA by oxidizing substrates,permitting them to be recognized by the CMA machinery and imported into thelysosome more efficiently.
MATERIALS AND METHODS
Cell Culture and Isolation of Lysosomes—Human embryonicfibroblasts (IMR-90) were obtained from Corriel Cell Repositories (Camden,NJ). Cells were maintained in Dulbecco's modified Eagle's medium (Sigma)supplemented with 10% newborn calf serum and 1% antibiotic(penicillin/streptomycin with fungicide; Invitrogen). IMR-90 cells were usedfor experiments at 20–40 population doubling levels (PDLs). To deprivecells of serum, cultures were washed several times with Hanks' balanced salts'solution (Invitrogen) and then placed in media without serum. Lysosomes fromIMR-90 cells were isolated as described previously(
).Lysosomal preparations that contained >10% broken lysosomes, as determinedby β-hexosaminidase latency, were not used.
Chemicals and Antibodies—All chemicals were obtained fromSigma unless otherwise indicated. GAPDH was obtained from Roche AppliedScience, and RNase A was obtained from Worthington Biochemical Company(Lakewood, NJ). Antibodies against lamp2a were obtained from ZymedLaboratories Inc.. Antibodies against GAPDH were obtained from Biodesign(Saco, ME), and antibodies against hsc70 were obtained from MaineBiotechnology Services (Portland, ME). Detection of the formation of carbonylgroups in proteins was performed by following the manufacturer's instructionsusing the OxyBlot™ oxidized protein detection kit supplied by ChemiconInternational (Temecula, CA). [14C]GAPDH and [14C]RNaseA were radiolabeled by using [14C]formaldehyde reductivemethylation as described previously(
Cellular Protein Degradation—Confluent cells were labeledwith [3H]leucine (2 μCi/ml) for 48 h in media containing 10%newborn calf serum. Cells were washed twice with Hanks' balanced saltsolution, and the media were replaced with either complete media or mediawithout serum both containing excess (2.8 mm) unlabeled leucine(
). Aliquots of media (200μl) were taken at the indicated time intervals and precipitated in 20%trichloroacetic acid. Proteolysis was determined by measuring the percent ofacid soluble radioactivity compared with the total radioactivity of celllysates prepared by the addition of 0.1 n NaOH and 0.1% sodiumdeoxycholate (
). Briefly, GAPDH wasincubated with purified lysosomes treated with 100 μmchymostatin A. The GAPDH was incubated with the lysosomes for 30 min and thenwashed and treated with 10 μg of proteinase K and 1 μmCaCl2 for 15 min. The reaction was stopped by incubation with4-(2-aminoethyl)-benzenesulfonyl fluoride. Lysosomes were solubilized andprocessed for SDS-PAGE.
Immunoprecipitation of GAPDH—Cells were placed in conditionsof serum deprivation for 24 h. The cells were washed twice in ice-coldphosphate-buffered saline and incubated in Nonidet P-40 lysis buffer (50mm Tris-HCl (pH 8.0), 150 mm NaCl, 1% Nonidet P-40, 0.5μg/ml phenylmethylsulfonyl fluoride, and 0.01 μg/ml each aprotinin,leupeptin, and pepstatin for 15 min The lysate was collected by centrifugingthe samples at 14,000 × g for 15 min. Aliquots were then takenand precleared with protein A-Sepharose beads (Amersham Biosciences) for 1 h.Antibody (8 μg) was added to the precleared lysate and incubated overnightat 4 °C. The antibody-antigen complex was removed from the lysate bycentrifuging the samples at 10,000 × g for 30 s to pellet theprotein A-Sepharose beads. The supernatant was discarded, and beads washed anadditional three times. The beads were resuspended in Laemmli sample bufferand subjected to SDS-PAGE followed by immunoblot analysis.
General Methods—Protein determination was performed usingthe Lowry method with bovine serum albumin as a standard(
) werevisualized using chemiluminescence detection methods (Western Lightning;PerkinElmer Life Sciences). Radioactivity was determined by counting samplesin a Packard liquid scintillation analyzer with quenching detected by anautomatic external standard (Packard Tri-Carb 2100 TR). Densitometric analysiswas performed by using Adobe Photoshop 7.0. Statistical analyses werepreformed using the two-tailed Student's t test.
In confluent cultures of IMR-90 human fibroblasts, serum withdrawalactivates CMA (
). Using pulse-chaseanalysis, we analyzed the effect of BOH on cellular proteolysis in IMR-90cells maintained in media with or without serum(Fig. 1a). We observedan increase in proteolysis in BOH-treated cells maintained either with orwithout serum as compared with control. In cells maintained in media withserum, proteolysis increased to that of control cells maintained withoutserum. We found similar effects of BOH-induced proteolysis using Chinesehamster ovary cells (data not shown). Treating cells with acetoacetateresulted in an increase in proteolysis only in cells maintained in mediacontaining serum (Fig.1b). Because BOH had a more dramatic activation of CMA,we chose to focus on the BOH induction of CMA. Because CMA is activated byserum withdrawal, all cell cultures were serum-deprived 24 h prior toexperimentation unless noted otherwise.
Treatment with oxidizing agents during prolonged nutrient deprivation maydecrease lysosomal stability(
). To address the issue oflysosomal stability we performed a β-hexosaminidase latency assay.Lysosomes were isolated from cells in the absence or presence of BOH. Theisolated lysosomes were kept at 37 °C, and the latency was measured ateither 0, 15 or 30 min after isolation. The stability of the lysosomes fromtreated cells was indistinguishable from those isolated from treated cells atall time points (Fig. 2). Afterisolation, ∼1% of the total lysosomes were broken in both treated anduntreated lysosomes. After incubation at 37 °C for 15 and 30 min, thepercentage of broken lysosomes rose to 1.5 ± 0.5 and 1.75 ±0.5%, respectively. Therefore, we conclude that treatment with BOH does notaffect the stability of the lysosomes.
To explore the mechanisms for the BOH-induced CMA, we performed invitro CMA degradation assays using isolated lysosomes. In these assays,isolated lysosomes combined with substrates of CMA such as[14C]GAPDH and [14C]RNase A allowed us to determine thepercentage of lysosomal protein degradation by measuring the change oftrichloroacetic acid-precipitable radioactivity into trichloroaceticacid-soluble radioactivity(
The lysosomes isolated from BOH-treated cells showed an increase indegradation of both [14C]GAPDH(Fig. 3a) and[14C]RNase A (Fig.3b) as compared with controls. Degradation of both[14C]GAPDH and [14C]RNase increased with increasedamounts of lysosomal protein (0,14, 28, and 56 μg) At 14 μg of lysosomalprotein there was no significant difference in the degradation of[14C]GAPDH or [14C]RNase A between lysosomes isolatedfrom control versus treated cells; however, at 28 and 56 μg oflysosomal protein from treated cells there was a 2- and 3-fold increase in thedegradation of [14C]GAPDH, respectively, in treated compared withcontrol cells. Similarly, with 28 and 56 μg of lysosomal protein there wasa 3- and 5-fold increase in the degradation of [14C]RNase A,respectively, compared with control. These data indicate that the observedincrease in proteolysis was due, at least in part, to the stimulation ofCMA.
To determine the mechanism of activation of CMA by BOH we isolatedlysosomes from cells and then added varying concentrations of BOH(Fig. 4). The degradation wasenhanced by increasing the concentration of BOH, achieving a maximal responseat 4 mm. The observed increase in the proteolysis of[14C]GAPDH and [14C]RNase A indicated that BOH wasacting on the lysosome and/or substrates to elicit an effect.
We next wanted to determine whether BOH has an effect on the levels of keyproteins involved in CMA, namely lamp2a and lyhsc70. The amount of lamp2a inthe lysosome membrane is rate-limiting for CMA under a variety ofphysiological and pathological conditions(
). We isolated lysosomesfrom control and BOH-treated cells and measured protein levels by immunoblotanalysis. We found that the amount of hsc70 and lamp2a in the membrane andlyhsc70 in the membrane and matrix of lysosomes were equivalent in control andtreated cells (Fig. 5).Therefore, BOH was inducing CMA without altering the levels of these keyproteins.
The proteolysis assays using isolated lysosomes (Figs.3 and4) reflect a combination ofsubstrate binding, transport, and degradation(
). To establish whethersubstrate was entering BOH-treated lysosomes more efficiently than untreatedlysosomes, we performed a proteinase K protection assay. We isolated lysosomesfrom IMR-90 cells and treated them with protease inhibitors. Lysosomes werethen incubated with BOH and GAPDH or GAPDH alone. After 1 h of incubation, wewashed the lysosomes and treated with proteinase K to cleave any remainingsubstrate bound to the lysosomal membrane. We then carried out immunoblots ofthe isolated lysosomes to measure the amount of GAPDH that had beentransported into the lysosomes (Fig.6). The lysosomes incubated with BOH showed a 2.5-fold increase inthe amount of GAPDH that was transported into the lysosome(Fig. 6). These data suggestthat in a reaction mixture containing BOH, GAPDH, and lysosomes, the GAPDHundergoes a more efficient uptake compared with a reaction mixture deprived ofBOH, suggesting that BOH acts on the substrate, the lysosome, or both.
Next, we isolated lysosomes from IMR-90 cells. The lysosomes and/orGAPDH/RNase A were pretreated with 4 mm BOH and then subjected to atransport assay (Fig. 7). Wefound that when substrate was preincubated with BOH there was a 2-foldincrease in its degradation compared with untreated substrates. BOH-pretreatedlysosomes degraded both substrates at the same rate as untreated lysosomes.These data suggest that BOH acts on substrates and causes them to be degradedby CMA more efficiently.
One possibility is that BOH directly or indirectly causes proteins to beoxidized, making the damaged proteins better substrates for CMA. To test thisidea, we looked at the accumulation of oxidized GAPDH in BOH-treated IMR-90cells. We immunoprecipitated GAPDH from either BOH-treated or untreated cellsand then performed an OxyBlot™ assay, which derivatizes carbonyl groupsto a 2,4-dinitrophenylhydrazone (DNP) moiety. The DNP moiety can then bedetected using anti-DNP antibodies and is a method to assay for one form ofoxidative damage to a protein. We then subjected the immunoprecipitated GAPDHto an OxyBlot™ assay followed by SDS-PAGE and immunoblot analysis usingthe anti-DNP antibody as our primary antibody. We found that GAPDH fromBOH-treated cells was oxidized ∼3-fold more than GAPDH from control cells,suggesting one mechanism for the stimulation of CMA by BOH(Fig. 8a,). RNase A isnot an intracellular protein, so we could not follow the same procedures asfor GAPDH. In this case, we incubated RNase A with 4 mm BOH andfound that there was an increase in the appearance of carbonyl groups comparedwith control. RNase A incubated with BOH for 4 or 24 h showed ∼2- and3-fold increases, respectively, in carbonyl groups compared with untreatedRNase A (Fig. 8b).
Finally, to determine whether the stimulation of CMA by BOH was specificfor BOH, we tested compounds structurally similar to BOH, butanone, andbutanol. We treated cells with both butanone and butanol (4 mmeach) (Fig. 9a), butwe did not observe any effect on proteolysis in the absence or presence ofserum. We next asked if another compound whose concentration increases duringstarvation, glycerol, could stimulate CMA. We treated cells with glycerolconcentrations ranging from 0.05 to 5.0 mm and measured theiraffects on CMA (Fig.9b). There was no significant affect of glycerol on thestimulation of proteolysis. Taken together, these data suggest that BOHstimulates CMA in a specific manner.
Five lines of evidence indicate that ketone body formation leads toincreased proteolysis by CMA. First, IMR-90 cells treated with BOH and, to alesser extent, acetoacetate increased proteolysis in cells maintained inserum-supplemented media (Fig. 1,a and b). Acetone could not be tested because ofits volatility.
Second, we found that lysosomes isolated from BOH-treated cells transportedand degraded substrates of CMA at a higher rate than control lysosomes(Fig. 3, a and3b). Third, BOH-treated substrates were degraded bylysosomes at a higher rate than untreated substrates, suggesting that BOH actsthrough substrate modification to stimulate CMA(Fig. 7). Fourth, BOHincubation with RNase A increases the formation of carbonyl groups in RNase A(Fig. 8b). Lastly, wedemonstrated that GAPDH, a substrate for CMA, immunoprecipitated from cellstreated with BOH showing a higher occurrence of oxidative damage as comparedwith control (Fig.8a).
The effects of ketone bodies have been extensively studied on tissues thatutilize them for energy, such as skeletal muscle and brain. The use of ketonebodies as an energy source prevents the catabolism of essential proteins andpreserves amino acid pools within the cell during times of nutritional stress(
).The effects of ketone bodies on cell types that do not utilize ketone bodiesfor energy, such as IMR-90 cells, have not been closely studied even thoughthey survive in an environment in which ketone bodies reach significantly highlevels of 1–12 mm(
). The ability of BOH toincrease protein breakdown in times of nutritional stress may be one moreevolutionary mechanism for cells to survive prolonged nutrient deprivationwithout using ketone bodies for energy.
Previous studies have shown that oxidized proteins make better substratesfor CMA (
), who demonstrated thatoxidized proteins have the ability to bind and translocate into the lysosomemore rapidly than non-oxidized proteins. The mechanism of enhanced substrateuptake due to oxidation is not clear. We do know that substrates must beunfolded prior to translocating across the lysosomal membrane(
). Therefore, the unfoldingof proteins after oxidation may promote recognition by thechaperone-cochaperone complex and increase the rate of delivery and uptake ofoxidized substrates into the lysosome.
It is interesting to note that in Fig.4 the broken lysosomes actually have a decreasing degradativecapacity in the presence of increasing concentrations of BOH. This may beexplained by the fact that increasing concentrations of BOH may be damagingand inactivating the degradative enzymes, thus causing less proteolysis.Future studies could focus on the effect that BOH treatment has on enzymefunction. Also in Fig. 3 weobserved that there was no difference between the amount of degradationbetween 0 and 14 μg of lysosomal protein. One explanation of thisobservation is that our degradation assay is not sensitive enough to detectchanges in degradation with this amount of lysosomal protein. With nolysosomes added we found 1–2% protein degradation. Undoubtedly, oneneeds enough of a signal above 1–2% to reflect CMA. If this were thecase one would expect to see an increase in degradation by adding higherconcentrations of lysosomes, which is in fact what we observed.
We demonstrated that there was an increase in the degradation of[14C]GAPDH by lysosomes isolated from cells treated with 4mm BOH for 24 h (Fig.3). However, we saw no increase in proteolysis when lysosomes werepreincubated with 4 mm BOH for 20 min(Fig. 7). One explanation forthis apparent discrepancy is that isolated lysosomes normally have CMAsubstrates bound. These substrates will compete with [14C]GAPDH or[14C]RNase for uptake. The lysosomes isolated from BOH-treatedcells will have fewer CMA substrates bound because they are more efficientlytranslocated and degraded (Fig.6). Therefore, lysosomes isolated from BOH-treated cells will havea higher activity for added radiolabeled substrates. The lysosomespreincubated with BOH for 20 min may not have had sufficient time to processnonradioactive bound CMA substrates. We are unable to test this idea, becausethe isolated lysosomes are not stable during prolonged incubation times.
Whenever we compared intracellular proteolysis rates, cells were ofequivalent PDLs between 20 and 40. However, proteolytic rates betweenexperiments were somewhat variable (Figs.1, a and band 9, a andb). In 100 independent experiments we obtained half-livesof 48 ± 6 and 28 ± 6 h in cells maintained with or withoutserum, respectively. IMR-90 cells in media without serum have been shown toincrease the half-life of long-lived proteins by 20 min per PDL(
). Therefore, varying PDLsbetween 20 and 40 h will lead to variability in at½ of ±7 h.
There are several possible mechanisms to explain BOH stimulation of CMA.First, BOH likely causes protein oxidation by increasing reactive oxygenspecies, which could lead to an increase in total oxidized proteins(
). These results aresomewhat different from ours, as we could not see a measurable change in thelevels of lamp2a or the luminal chaperone. These differences could beexplained by the fact that Kiffin et al.(
) dosed their rodents twicewith paraquat or hydrogen peroxide to oxidize proteins. Treatment withpro-oxidants of this magnitude may increase the levels of oxidized proteinsand force the cell to up-regulate components of CMA in order to deal with theincreased demand of substrates. BOH, on the other hand, may only oxidize asmall percentage of proteins, permitting the cell to degrade the damagedproteins without increasing components of the pathway.
), and cause the cell toactivate CMA as part of the nutrient stress response. The cells would induceCMA without the need to up-regulate components of the pathway. This does notseem likely, because immunoprecipitated GAPDH from BOH-treated cells had a3-fold increase in the level of oxidation as compared with control. Futurestudies should investigate the effects of oxidants of varying strengths on themodification of substrates and the subsequent activation of CMA.
Third, several studies suggest that ketone bodies have the ability to causeincreased free radical formation and lipid peroxidation in erythrocytes andendothelial cells (
). We did observe anincrease in proteolysis in acetoacetate-treated cells maintained in serum, butnot in cells maintained without serum (Fig.1b). This could be explained by the fact thatacetoacetate has the ability to down-regulate the insulin receptor at the mRNAlevel by 56% (
).Down-regulation of the insulin receptor would stimulate the cell to respond tonutrient deprivation and activate CMA. This is one explanation as to why weonly see an activation of CMA in cells maintained in serum. Future studiesshould look at what other factors contribute to CMA activation.
In these studies we have demonstrated that ketone bodies, more specificallyBOH, stimulate CMA by causing the oxidation of substrates. In addition, duringprolonged starvation CMA is activated because of increased lamp2a in thelysosomal membrane and increased lyhsc70 in the lysosomal lumen(
). Our data indicate thatketone bodies can also stimulate CMA by affecting substrate proteins duringprolonged starvation in vivo. This finding gives us further insightinto the physiological importance of CMA stimulation during times of nutrientdeprivation.
We thank Dr. Ana Maria Cuervo, Dr. Ira Herman, Dr. Laura Liscum, andNicholas T. Mesires, M.S. for valuable discussions and critical evaluation ofthis manuscript. We also thank Dr. Alfred J. Meijer for suggesting to us thatketone bodies may activate CMA.