11-Deoxy,16,16-Dimethyl Prostaglandin E2 InducesSpecific Proteins in Association with Its Ability to

Protect Against Oxidative Stress

Kelly M. Towndrow,†,‡ Zhe Jia,‡ Herng-Hsiang Lo, Maria D. Person,Terrence J. Monks, and Serrine S. Lau*

Center for Molecular and Cellular Toxicology, Division of Pharmacology and Toxicology,College of Pharmacy, The University of Texas at Austin, Austin, Texas 78712

Received June 21, 2002

Prostaglandins (PGs) act locally to maintain cellular homeostasis and stimulate stressresponse signaling pathways. These cellular effects are diverse and are tissue-dependent. PGE2,and the synthetic analogue, 11-deoxy,16,16-dimethyl PGE2 (DDM-PGE2), protect renal proximaltubular epithelial (LLC-PK1) cells against cellular injury induced by the potent nephrotoxicand nephrocarcinogenic metabolite of hydroquinone, 2,3,5-tris-(glutathion-S-yl)hydroquinone.Although this cytoprotective response (in LLC-PK1 cells) is mediated through a thromboxaneor thromboxane-like receptor coupled to AP-1 signaling pathways, the mechanism of cyto-protection is unknown. In this study, we utilized HPLC-electrospray ionization tandem massspectrometric (ESI MS/MS) and matrix-assisted laser desorption ionization time-of-flight massspectrometric (MALDI TOF) analysis of proteins isolated from DDM-PGE2-stimulated LLC-PK1 cells to identify candidate cytoprotective proteins. DDM-PGE2 selectively stimulated thesynthesis of several proteins in LLC-PK1 cells. Peptide sequencing by ESI-MS/MS of in-geltryptic protein digests revealed the identity of eight proteins: endothelial actin binding protein,myosin, elongation factor 2 (EF-2), elongation factor 1R-1 (EF-1R), heat shock protein 90â(HSP90â), glucose-regulated protein 78 (GRP 78), membrane-organizing extension spike protein,and actin. Both ESI-MS/MS and MALDI-MS analysis resulted in the same protein identifica-tion. Western analysis confirmed the temporal induction of the majority of these proteins,including EF-2, EF-1R, HSP90â, GRP78, and actin. The collective expression of these proteinssuggests that DDM-PGE2-mediated cytoprotection may involve alterations in cytoskeletalorganization and/or stimulation of an endoplasmic reticulum (ER) stress response. The presentstudies provide insights into potential downstream targets of PG signaling.

IntroductionPGs1 have diverse intercellular and intracellular effects

and are involved in numerous physiological and patho-physiological processes. PGs offer cytoprotection againstvarious cellular stresses, although most studies havefocused on the protective effects of PGs in liver (1) andgastric mucosa (2). The mechanism(s) underlying PG-mediated cytoprotection is unknown, but the ability ofPGs to protect cells in culture against toxic insultsimplies that this mechanism(s) is cellular in nature (3).One potential mechanism of cytoprotection may involvePG-mediated regulation of cytoskeletal organization,

since the combination of PGE2 and a PGI2 analoguefacilitates the repair of ischemic ileum epithelium (4).This cytoprotection is associated with elevations in cAMPand Ca2+ leading to cytoskeletal-mediated tight junctionclosure, likely via actin filament relaxation, which mayconfer cytoprotection by regulating membrane perme-ability (4). The effects of PGE2 on cell shape have beennoted in many systems, including cultured osteoblasts(5). As an example, PGE1 and PGE2 induce morphologicalchanges and the selective breakdown of actin microfila-ments but not microtubules or vimentin filaments (5).

The coupling of bisphosphonates to PGE2 to formsynthetic PGE2-bisphosphonate conjugates may be po-tentially therapeutic in the treatment of osteoporosis (6).Such conjugates may be useful due to their ability toselectively target bone and to be slowly hydrolyzed intotwo therapeutically useful compounds, bisphosphonatesand PGE2, which are potential bone resorption inhibitorsand bone formation stimulators, respectively (6). More-over, the protection of Caco-2 cells against ethanol-induced damage by 16,16-dimethyl PGE2 correlates withincreased PKC activity and Ca2+ efflux and a subsequentstabilization of microtubules (7). Other possible PGcytoprotective mechanisms may include alterations inblood flow, enhanced regeneration, alterations in toxicantmetabolism, and increased membrane stability (1).

* To whom correspondence should be addressed. Tel: 512-471-5190.Fax: 512-471-5002. E-mail: slau@mail.utexas.edu.

† Present address: Department of Investigative Toxicology, LillyResearch Laboratories, Eli Lilly and Company, Greenfield, IN 46140.

‡ These authors contributed equally to this work.1 Abbreviations: acetonitrile, ACN; 11-deoxy,16,16-dimethyl prosta-

glandin E2, DDM-PGE2; 16,16-dimethyl prostaglandin E2, dmPGE2;elongation factor, EF; HPLC-electrospray tandem mass spectrometry,ESI MS/MS; glucose-regulated protein, GRP; heat shock protein, HSP;membrane-organizing extension spike protein, moesin; endoplasmicreticulum, ER; [1S-[1R,2R(Z),3â(1E,3S),4R]]-7-[3-[3-hydroxy-4-(4-iodo-phenoxy)-1-butenyl]-7-oxabicyclo[2.2.1]hept-2-yl]-5-heptenoic acid, IBOP;[1S-[1R,2â(Z),3R,5R]]-7-[3-[[(4-iodophenyl)sulfonyl]amino]-6,6-dimethylbicyclo[3.1.1]hept-2-yl]-5-heptenoic acid, ISAP; matrix-assistedlaser desorption/ionization, MALDI; protein kinase C, PKC; prosta-glandin, PG; 2,3,5-tris-(glutathion-S-yl)hydroquinone, TGHQ; throm-boxane, TX; time-of-flight, TOF; 12-O-tetradecanoyl-phorbol-13-acetate,TPA; unfolded protein response, UPR.

312 Chem. Res. Toxicol. 2003, 16, 312-319

10.1021/tx020048l CCC: $25.00 © 2003 American Chemical SocietyPublished on Web 02/04/2003

PGE2 generally acts via G-protein-coupled cell surfacereceptors, designated EP1-EP4, each of which is coupledto different signal transduction systems. However, thecomplexity of PG signaling mechanisms is illustrated bythe fact that (i) PGs can be transported into cells via aPG transporter (8-9), (ii) EP receptors can also belocalized at the nuclear membrane (10), and (iii) somePGs (i.e., 11-deoxy-∆12, 14-PGJ2) are ligands for otherreceptors, such as the peroxisome proliferator-activatedreceptor family of nuclear receptors (11). Thus, PGsignaling is multifaceted and complicated by autocrineand paracrine effects.

DDM-PGE2 and PGE2 protect renal proximal tubularepithelial (LLC-PK1) cells against TGHQ-mediated cyto-toxicity (12). This cytoprotective response is apparentlynot mediated via known EP receptor subtypes, sinceknown EP agonists do not confer protection. The DDM-PGE2 protective response is mediated through a PKC-coupled pathway since (i) TPA, an activator of PKC, alsoinduces cytoprotection; (ii) DDM-PGE2 increases TREbinding activity; and (iii) cytoprotection is overcome byPKC-inhibiting concentrations of H-89 (12). The cyto-protective response to DDM-PGE2 is mediated by a TXA2 receptor (13), as both U46619 and IBOP, TXA2

receptor agonists, also protect against TGHQ-mediatedcytotoxicity. Furthermore, DDM-PGE2-mediated cyto-protection and TRE and NF-κB binding activity areinhibited by the TX receptor antagonists, SQ29548 andISAP (13). Sulfalazine, a TX A2 synthase inhibitor, alsoblocks cytoprotection and NF-κB binding activity inducedby DDM-PGE2, indicating divergent downstream signal-ing pathways (13). Although it is clear that DDM-PGE2

is cytoprotective through a receptor-mediated pathway,the mechanism(s) underlying this response remains un-known. The studies described herein were conducted toidentify proteins expressed in DDM-PGE2-stimulatedLLC-PK1 cells that may participate in DDM-PGE2-mediated cytoprotection.

Experimental Procedures

Cell Culture. LLC-PK1 cells were purchased from theAmerican Type Culture Collection (CL101) and cultured inDulbecco’s modified Eagle medium with 4.5 g/L glucose (DMEM;GIBCO BRL, Grand Island, NY) supplemented with 10% fetalbovine serum (FBS; Atlanta Biologicals, Atlanta, GA) in a 37°C/5% CO2-humidified incubator.

35S-Methionine Protein Labeling. Postconfluent LLC-PK1cells were exposed to 1 or 2 µM DDM-PGE2 (Caymen Chemicals,Ann Arbor, MI), an ethanol vehicle control, 10 µM U46619(Caymen Chemicals), 10 ng/mL TPA (Calbiochem, San Diego,CA), or 0.1% (v/v) dimethyl sulfoxide (DMSO) vehicle control inDMEM (methionine/cysteine free) with 25 mM HEPES and 10%FBS medium (pH 7.4) containing 35S-methionine (ICN; 0.1 mCi/ml) for 24 h. At the end of the experiment, cells were washedthree times with PBS, scraped, and collected in PBS. Cells werepelleted at 200 g (10 min) and lysed in buffer (80:10:10 stockbuffer [10 mM Tris, 10 mM NaCl, 3 mM MgCl2, 1 mM EDTA,0.1% NP-40, pH 7.4], 10 mg/mL PMSF, 100 mM sodiumorthovanadate; 1 complete mini protease inhibitor cocktail tablet[containing antipain dihydrochloride (50 µg/mL), bestatin (40µg/mL), chymostatin (60 µg/mL), E-64 (10 µg/mL), leupeptin (0.5µg/mL), pepstatin (0.7 µg/mL), phosphoramidon (300 µg/mL),pefabloc SC (1 mg/mL), EDTA disodium salt (0.5 mg/mL), andaprotinin (2 µg/mL)] in 10 mL) for 15-30 min on ice. After lysis,the samples were subjected to one freeze/thaw cycle andcentrifuged at 16 000g (10 min), and the supernatant wascollected. The concentration of 35S-methionine-labeled proteins

was determined by the Bradford method (14), and proteins wereseparated on an 8% denaturing polyacrylamide gel (16 cm ×16 cm SDS-PAGE), stained with Coomassie Blue R, dried, andanalyzed by autoradiography.

In-Gel Protein Tryptic Digestion. Differentially expressedprotein bands were identified by autoradiography and thenselected for in-gel digest after Coomassie Blue staining. 35S-Methionine-labeled proteins were separated by SDS-PAGE andprocessed as described above, except without drying the gel.Typically, pooled individual protein bands originating from acombined total of 300-600 µg of cellular lysate were a sufficientamount for LC-MS/MS analysis. In-gel tryptic digestion wasbased on a modification of the method of Shevchenko et al. (15).Prior to in-gel digest, individual bands were cut into 1 mm piecesand destained in 5% acetic acid and 50% methanol to removethe Coomassie Blue. Gel pieces were dehydrated with ACN, andresidual ACN was evaporated in a SpeedVac. Proteins were thenreduced with 10 mM DTT in 100 mM NH4HCO3 at roomtemperature for 1 h. Residual DTT was removed, and cysteineswere alkylated with 50 mM iodoacetamide (in 100 mM NH4HCO3)for 1 h. After the residual iodoacetamide was removed, gel pieceswere subjected twice to washing (100 mM NH4HCO3 for 10 min)and dehydration (5 min in ACN). Gels were dried for 2-3 minin a SpeedVac and rehydrated on ice with 20 ng/µL sequencinggrade modified trypsin (Promega, Madison, WI; in 50 mMNH4HCO3) for 10-15 min. Excess trypsin was removed, 20 µLof 50 mM NH4HCO3 was added, and gel pieces were digestedovernight at 37 °C. After digestion, peptides were extractedtwice in 75 µL of 5% formic acid/50% ACN.

Peptide Sequencing by LC-ESI-MS/MS. Peptides wereanalyzed with an electrospray ion trap mass spectrometer(ThermoFinnigan LCQ, San Jose, CA) coupled to a microboreHPLC (Magic 2002, Michrom BioResources, Auburn, CA).Samples were injected into a MAGIC MS C18 (5 µm, 200 Å, 0.5mm × 50 mm) column and eluted with a 5-95% mobile phaseB (90:10:0.09:0.02 ACN:H2O:acetic acid:trifluoroacetic acid) over45 min followed by 95% B for 10 min. Mobile phase A consistedof 2:98:0.1:0.02 ACN:H2O:acetic acid:trifluoroacetic acid. A flowrate of 20 µL/min was used, and a mass range of 400-2000 Dawas recorded for the full scan, which was followed by a singleZoom scan and MS/MS of the most intense peak. Individualpeptide sequences were identified with SEQUEST incorporatedinto ThermoFinnigan BIOWORKS software to match MS/MSspectra to amino acid sequences in the National Center forBiotechnology Information (NCBI) or OWL protein database.Samples from several experiments were analyzed, and onlyproteins seen in multiple runs are reported here.

The two high molecular weight bands (bands 1 and 2 inFigure 1) were also analyzed by microspray LC-MS/MS. A resist-ive splitter (Magic Variable Splitter, Michrom BioResources)was used to split the precolumn flow from a 20 µL/min input to0.35 µL/min through the column. A custom-built microsprayinterface, according to the design of Gatlin et al. (16), wasmounted on the LCQ, using a PicoFrit (New Objective, Cam-bridge, MA) 75 µm i.d. x 5 cm column with a 15 µm tip filledwith BioBasic C18 material (5 µm). A 30 min linear gradientfrom 5 to 65% B followed by a 15 min wash at 95% B was usedto elute the digested peptides, with A (0.5% acetic acid, 0.005%trifluoroacetic acid in water) and B (0.5% acetic acid, 0.005%trifluoroacetic acid, 90% ACN, 10% water). The LCQ acquireda single MS over the m/z range 360-2000 followed by two data-dependent MS/MS scans with dynamic exclusion. Peptidesequences were identified using the TurboSEQUEST softwareas described above.

Peptide Mapping by MALDI-MS. Peptides were directlyanalyzed with a MALDI TOF mass spectrometer (PerSeptiveBiosystems Voyager De-Str, Framingham, MA). Spectra wereacquired in positive ion mode using the reflectron detector overa mass range of 700-3600 Da. The sample and R-cyano-4-hydroxycinnamic acid (Agilent Technologies, Palo Alto, CA)matrix were mixed 1:1 and drop-dried on the target. A mixtureof 1 ng/µL cze standards (Bio-Rad, Hercules, CA) and 2.5 ng/µL

DDM-PGE2-Induced Proteins in Renal Cells Chem. Res. Toxicol., Vol. 16, No. 3, 2003 313

adrenocorticotropic hormone was used for external calibrationof MALDI spectra. To minimize salt interference and enhancesensitivity, protein samples were desalted using a C18 ZipTip(Millipore, Bedford, MA) washed and eluted with 0.1% formicacid and 0.1% formic acid/50% ACN, respectively.

The peptide mass list for each sample was entered in theprotein identification program, MS-Fit, in the Protein ProspectorSuite (www.prospector.ucsf.edu). The OWL database wassearched using a 75 ppm peptide mass tolerance for trypticdigest and a maximum of one missed cleavage and carbamido-methylation of the cysteine residues. To enhance search sensi-tivity, a mass tolerance of 50 ppm coupled with an externalcalibration, at a position adjacent to the sample spot on theMALDI target, was utilized with mass searches for proteinsample one.

Western Blot Analysis. Cells were washed three times withice cold PBS, scraped, and lysed in buffer (50 mM Tris-HCL;pH 7.5, 100 mM NaCL, 10 mM sodium fluoride, 5 mM EDTA,1% Triton X-100, 40 mM â-glycerophosphate, 0.5 mM sodiumorthovanadate, 0.25 mM PMSF, 1 complete mini proteaseinhibitor cocktail tablet in 10 mL) for 15-30 min on ice. Thecell lysates were centrifuged at 14 000g for 20 min at 4 °C. Thesupernatant containing the total protein was collected andstored at -80 °C. Protein concentration was measured by theBio-Rad DC protein assay kit (Bio-Rad Laboratories). Pro-teins were separated on a 10% denaturing polyacrylamide geland transferred to a nitrocellulose membrane by electroblotting(wet transfer). Primary antibodies with different dilution factorsand manufacturers are listed as follows: elongation factor 2 (EF-2; 1:250, Santa Crutz Biotechnologies, Santa Cruz, CA); HSP90â(1:4000, Stressgene Biotechnologies, Victoria, BC, Canada);GRP78 (1:40 000, a generous gift from Dr. James Stevens, EliLilly & Co); elongation factor 1R-1 (EF-1R; 1:1000, UpstateBiotechnology, Lake Placid, NY); actin (1:10 000, OncogeneResearch Products, Boston, MA). GAPDH (1:1000, a generousgift from Dr. Kline, University of Texas at Austin) was used asthe housekeeping protein for loading normalization. Secondary

antibodies (1:3000 dilution for all blots except GRP78, whichwas 1:10 000) were purchased from the Santa Cruz company.Protein expression was visualized using enhanced chemilumi-nescence (Amersham, Arlington Heights, IL) according to themanufacturer’s specifications.

Results

DDM-PGE2-Mediated Induction of Protein Syn-thesis. A 24 h pretreatment with DDM-PGE2 (1 or 2 µM)is required to protect LLC-PK1 cells from TGHQ (300µM)-mediated cytotoxicity (12). However, the mecha-nism(s) underlying this cytoprotective response is un-known. 35S-Methionine labeling of newly synthesizedproteins was utilized to identify candidate cytoprotectiveproteins induced during a 24 h exposure of LLC-PK1 cellsto DDM-PGE2. DDM-PGE2 (1 or 2 µM, 24 h) increases(120 ( 6%; mean ( SE) overall protein synthesis in LLC-PK1 cells, and at least eight proteins were selectivelyinduced by DDM-PGE2 (Figure 1) as determined byincreased levels of 35S-methionine incorporation.

Induction of proteins 1, 5, and 8 was evident as earlyas 4 h following DDM-PGE2 treatment and persisted to24 h (Figure 1). Proteins 3, 4, and 7 were only slightlyinduced at 4 h but were clearly elevated by 8, 12, and 24h. A weaker induction of protein 6 was detected at 12and 24 h after exposure of LLC-PK1 cells to DDM-PGE2.A less robust induction was seen with protein 2, whichwas induced by 8 h (Figure 1).

Consistent with the hypothesis that DDM-PGE2 maybe acting through a TX or TX-like receptor (13), U46619(10 µM), a TX A2 receptor agonist, and the phorbol ester,TPA (10 ng/mL), were also found to increase total proteinsynthesis (119 ( 6 and 122 ( 1%; mean ( SE, respec-tively). Furthermore, induction of specific proteins, inparticular, proteins 1, 4, 5, 7, and 8, were similar in LLC-PK1 cells after treatment of DDM-PGE2, U46619, andTPA (Figure 2).

Mass Spectrometric Analysis of DDM-PGE2-Induced Proteins. To determine the identity of proteinselevated by DDM-PGE2, a mass spectrometric strategyinvolving in-gel tryptic digestion of SDS-PAGE sepa-rated 35S-methionine-labeled proteins was employed. Theeight proteins induced during the 24 h pretreatmentwindow (Figure 1) were targeted for tryptic in-gel diges-tion and LC-ESI-MS/MS analysis. Eight peptides wereidentified from an in-gel tryptic digest of protein 1 (Table1). A computer-generated search of the human subset ofthe OWL database identified protein 1 as the 280.7 kDaendothelial actin-binding protein (ABP-280; also callednonmuscle filamin; Table 1). These eight peptides com-prise 3.5% of the amino acid sequence of ABP-280 (i.e.,3.5% amino acid coverage). Six peptides from proteinband 2 were identified to cover 5.4% of the amino acidsequence for myosin. Four peptides (comprising 5.7% ofprotein 3) were identified to be derived from EF-2, witha molecular mass of 95.3 kDa (Table 1). Digests of protein4 produced eight peptides, which were attributed toHSP90â, a 83.2 kDa protein (with an amino acid coverageof 11.8%; Table 1). In one instance, peptides for bothHSP90â and HSP90R were found (data not shown).Database searching of the complete OWL databaserevealed 13 peptides from protein 5 (27.8% amino acidcoverage), which were identified as the 78 kDa GRP(GRP78) (Table 1). Searching the porcine subset of OWLdetermined that six peptides from protein 6 are derived

Figure 1. Time-dependent induction of DDM-PGE2 responsiveproteins. LLC-PK1 cells were exposed to vehicle (ethanol; E) or1 µM DDM-PGE2 (D) for 4, 8, 12, and 24 h in the presence of35S-methionine. Cell lysates (1 × 106 CPM/lane) were separatedby SDS-PAGE and analyzed by autoradiography as describedin the Experimental Procedures section.

314 Chem. Res. Toxicol., Vol. 16, No. 3, 2003 Towndrow et al.

from the 67.5 kDa moesin. Five peptides isolated fromprotein 7 were identified as EF-1R. Protein 8 wasidentified as cytoplasmic actin from 11 peptides compris-ing 27.7% amino acid coverage.

One set of in-gel protein digests was confirmed byMALDI-TOF-MS peptide mapping. Both LC-ESI-MS/MSpeptide sequencing and MALDI-MS peptide mass map-ping techniques resulted in the same protein identifica-tion for all samples (data not shown). The MS/MSfragmentation pattern for the peptide identified asTHINIVVIGHVDSGK from the EF-1R protein band isshown in Figure 3. Several other peptides isolated fromthe EF-1R band gave similar high quality MS/MS spectra,adding further confidence in the protein identification.Figure 4 is a representative MALDI-MS spectrum of thetryptic peptides identified from the EF-1R band. Thetryptic fragment assignments for the EF-1R sequence,trypsin autolysis, and keratin contaminant peaks fromsample handling are labeled. Both ESI-MS/MS andMALDI-TOF show complete agreement on the identifica-tion of the DDM-PGE2-induced proteins, although itremained possible that only the dominant protein(s) ineach band had been identified. Therefore, to validate themass spectral identification, expression of inducibleproteins was subsequently confirmed by western blotanalysis (Figure 5). Elevations in EF-1R and EF-2expression were evident as early as 4 h and weresustained through 24 h of DDM-PGE2 exposure. HSP90âwas maximally induced at 4 h and represented a plateauthrough 24 h of DDM-PGE2 exposure. In addition, adelayed temporal induction of GRP78 and actin, following8-24 h of DDM-PGE2 treatment, was observed.

Discussion

Although DDM-PGE2 induces overall protein synthesis,it also selectively stimulates the synthesis of specificproteins (Figure 1), including filamin, myosin, EF-2,HSP90â, moesin, GRP78, EF-1R, and actin (Table 1). Thecytoprotective effects observed in LLC-PK1 cells followingDDM-PGE2 treatment are coupled to a TX-like receptor(13) linked to the PKC pathway (12). Consistent withthese findings, DDM-PGE2 and U46619, the TX receptoragonists, and TPA, the PKC stimulator, all induce asimilar spectrum of proteins in LLC-PK1 cells (Figure2). HPLC-ESI-MS/MS analyses provided the identity ofeight proteins induced following treatment of LLC-PK1cells with DDM-PGE2, with at least four peptides identi-fied for each protein sample by peptide fragmentationand database searching, ensuring correct identificationof the dominant protein in each gel slice (Table 1).MALDI-MS peptide mass mapping confirmed the identityof proteins identified by multiple HPLC-ESI-MS/MSexperiments.

With available antibodies, western blot analysis fur-ther confirmed the mass spectrometric identification ofEF-1R, EF-2, HSP90â, GRP78, and actin (Figure 5).Unfortunately, currently available commercial antibodiesagainst endothelial actin binding proteins and myosinwere not specific, giving equivocal results. Moreover,although 35S-methionine labeling experiments showedweak induction of protein 6, identified as moesin by massspectral analysis, western analysis failed to show induc-tion (data not shown). It is possible that a low abundant,yet unknown cytoprotective protein(s) induced by DDM-PGE2 comigrated with moesin on gel electrophoresis. Onthe other hand, the weak induction of protein 6 variedfrom experiment to experiment. Whether moesin playsan important role in DDM-PGE2-induced cytoprotectionremains to be determined.

The induction of endothelial actin binding protein ABP-280 (filamin), myosin, EF-2, EF-1R, HSP90â, GRP78,moesin, and actin by DDM-PGE2 (Table 1) providesinsight into the mechanism of DDM-PGE2-mediatedcytoprotection. The known cellular functions and sub-cellular localization of these proteins are listed in Table2. These proteins are interrelated in the following way:(i) ABP-280, EFs (17, 18), and HSP90â (19) are all actinbinding proteins; (ii) ABP-280, myosin, moesin, and actinare associated with the cytoskeleton; (iii) EFs are criticalcomponents of the protein synthetic machinery, whichresides in close association with the cytoskeleton (18, 20);(iv) EF-1R has other nonprotein synthetic functions,including involvement in cytoskeletal organization (21);and (v) GRP78 functions in nascent protein translocationand protein processing in the ER (22). Thus, DDM-PGE2-mediated cytoprotection may involve alterations in cyto-skeletal organization and/or stimulation of an ER stressresponse as discussed below.

Cyclooxygenase products have been implicated in mi-togenic signaling pathways (23-26). Likewise, DDM-PGE2 increases overall protein synthesis but also inducesthe synthesis of specific proteins (Figure 1). Furthermore,DDM-PGE2 induces DNA synthesis in LLC-PK1 cells(data not shown). EFs are critical components of theprotein synthesis machinery (Table 2). EF-2, like EF-1R,is involved in peptide elongation, specifically by trans-location of the nascent protein chain between the A andthe P sites within the ribosome. EF-1R is an actin binding

Figure 2. DDM-PGE2, U46619 (both TX receptor agonists),and TPA produce a similar induction of protein synthesis inLLC-PK1 cells. LLC-PK1 cells were exposed to vehicle (DMSOor ethanol (ETOH)), 1 µM DDM-PGE2, 10 µM U46619, or 10ng/mL TPA for 24 h in the presence of 35S-methionine. Celllysates (1 × 106 CPM/lane) were separated by SDS-PAGE andanalyzed by autoradiography as described in the ExperimentalProcedures section.

DDM-PGE2-Induced Proteins in Renal Cells Chem. Res. Toxicol., Vol. 16, No. 3, 2003 315

protein that facilitates the GTP-dependent transfer ofaminoacyl-tRNA to the ribosome (21). The upregulationof EF-2 and EF-1R (Figures 1 and 5, Table 1) is consistentwith the increased protein synthesis observed followingtreatment of LLC-PK1 cells with DDM-PGE2.

The increase in actin (whether F-actin or G-actincannot be determined) by DDM-PGE2 is likely coupledto the induction of EF-1R, which has a role in cytoskeletalreorganization, including the regulation of microtubulestability and actin filament bundling (21). Thus, EF-1R

and actin are linked via their cytoskeletal and proteinsynthesis roles, respectively. A number of proteins identi-fied from DDM-PGE2-stimulated LLC-PK1 cells exhibitcytoskeletal functions (Table 2). Actin is the main com-ponent of the cytoskeleton, which functions to maintaincell shape and integrity. Banan et al. (27) demonstratedthat dmPGE2 increased the fraction of F-actin (poly-merized) to G-actin (unpolymerized), prevented collapseof the actin cytoskeleton, and protected a rat intestinalcell line from ethanol-mediated toxicity. Similarly,

Table 1: Summary of Results from LC-ESI-MS/MS Analysis

sample mol mass (kDa) peptide sequence A.Aa (%) OWL ref protein

1 280.7 LVSIDSKAIVDGNLKAGVAPLQVKGAGTGGLGLAVEGPSEAKVTVLFAGQHIAKLLGWIQNKDAGEGLLAVQITDPEGKPKLIALLEVLSQK

3.5 ABP2_HUMAN (H)b endothelial actin-bindingprotein (ABP-280; nonmusclefilamin)

2 226.5 HSQAVEELAEQLEQTKRVSHLLGINVTDFTRNLPIYSEEIVEMYKIRELESQISELQEDLESERSGFEPASLKEEVGEEAIVELVENGKKLQQELDDLLVDLDHQR

5.4 MYSN_HUMAN myosin, heavy chain,nonmuscle type A

3 95.3 NPADLPKVFSGLVSTGLKLDSEDKDKEGKPLLKAYLPVNESFGFTADLR

5.7 EF2_HUMAN (H) EF-2

4 83.2 LGIHEDSTNRYESLTDPSKSLVSVTKSIYYITGESKSLTNDWEDHLAVKGVVDSEDLPLNISRALLFIPRHSQFIGYPITLYLEK

11.8 HS9B_HUMAN (H) HSP90â (HSP84)

5 72.3 IQQLVKVLEDSDLKTWNDPSVQQDIKTKPYIQVDIGGGQTKNQLTSNPENTVFDAKSQIFSTASDNQPTVTIKITPSYVAFTPEGERKSDIDEIVLVGGSTRVTHAVVTVPAYFNDAQRIINEPTAAAIAYGLDKRIEWLESHQDADIEDFKAKFEELNMDLFRDNHLLGTFDLTGIPPAPR

27.8 GR78_MESAU 78 KDa glucose-regulatedprotein (GRP78; BiP)

6 67.5 IQVWHEEHRFVIKPIDKKFVIKPIDKESPLLFKEDAVLEYLKLFFLQVK

7.1 MOES_PIG (P) membrane organizing extensionspike protein (moesin)

7 50.1 EVSTYIKQLIVGVNKIGGIGTVPVGRLPIQDVYKTHINIVVIGHVDSGK

10.6 EF11_HUMAN EF-1R-1

8 41.7 ILTERIIAPPERKAGFAGDDAPRLDLAGRIIAPPERQEYDESGPSIVHRGILTLKIWHHTFYNELRAVFPSIVGRPRVAPEEHPVLLTEAPLNPKSYELPDGQVITIGNER

27.7 ACTB_HUMAN (H)ACTG_HUMAN (H)

actin, cytoplasmic 1 (â-actin)actin, cytoplasmic 2 (γ-actin)

a Amino acid coverage (%). b Species OWL database subset; p, porcine; h, human.

316 Chem. Res. Toxicol., Vol. 16, No. 3, 2003 Towndrow et al.

dmPGE2 reversed toxicity in Caco-2 cells by increasingPKC activity and microtubule stability (7). Moreover,carbacyclin (PGI2 analogue) and PGE2 together protectedileum intestinal mucosa from ischemic injury by initiat-ing cytoskeletal-mediated closure of tight junctions,which was proposed to be protective by maintainingproper membrane permeability (4). Interestingly, disrup-tion of the actin cytoskeleton of proximal tubules due toischemic or other injurious agents is well-documented(28-31). Thus, maintenance of cytoskeletal integrity islikely an important component of the cytoprotectiveproperties of DDM-PGE2.

Proximal tubular cells express distinct apical and baso-lateral membrane domains, which can be disorganizedduring injury, resulting in a loss of cell polarity (32).Spectrin and ankrin link cortical actin to basolateralspecific proteins, such as the Na, K-ATPase, thereby

maintaining normal basolateral membrane composition(32). Filamin is a spectrinlike protein that functions tocross-link actin microfilaments to the plasma membrane(32) and is induced in DDM-PGE2-stimulated cells (Table1). Moreover, the apical brush border is maintained byactin cross-linking via actin binding proteins such asezrin (28), which can mediate survival of LLC-PK1 cells(33). In contrast, dissociation of ezrin has been observedfollowing anoxia in rabbit proximal tubules (34). Takentogether, the data suggest that DDM-PGE2 may becytoprotective by upregulating moesin and filamin ex-pression and maintaining proximal tubular cell polarity.

HSP90 is an abundant (1-2% of total protein) cytosolicmolecular chaperone that is involved in the correctfolding of proteins (35-37). HSP90 production is inducedfollowing heat stress, but its specific role is not as well-characterized as that of HSP70 (35-37). The two isoformsof HSP90, R and â, share significant homology but appearto have different cellular functions. HSP90R is an actinand tubulin binding protein (38), and HSP90â is struc-turally related to the human microtubule interactingprotein, MIP-90 (39) that may protect microtubulesduring stress. DDM-PGE2-mediated elevations in HSP90âmay therefore prevent cytoskeletal damage induced byTGHQ through its association with cytoskeletal struc-tures and/or by facilitating repair of damaged proteins.

Toxicants induce ER stress proteins during proximaltubular epithelial cell death (40-42). Prior induction ofER stress (termed ER tolerance) and overexpression ofER chaperones protect proximal tubular epithelial cellsfrom subsequent chemical insult (40-42). Induction ofcritical ER chaperones (termed unfolded protein re-sponse; UPR), such as GRP78 (BiP), GRP94, and cal-reticulin, is essential to this cytoprotective response.Presumably, overexpression of ER chaperones is cyto-protective via increasing the ER’s ability to deal with

Figure 3. LC-ESI-MS/MS spectrum of EF-1R peptide THIN-ILVVIGHVDSGK. The MS/MS spectrum of one peptide fromthe LC-ESI-MS/MS analysis of protein band 7 is shown. Thepeptide was identified as a tryptic peptide from EF-1R, and theb and y fragmentation ions are labeled next to their correspond-ing peaks. The data are collected in the centroid mode.

Figure 4. MALDI-MS spectrum of EF-1R peptides identified from protein band 7. The EF-1R tryptic fragments, trypsin autolysis,and keratin contaminant peptides identified from the MALDI-MS spectrum of the gel digest of band 7 are labeled next to theircorresponding peaks. The inset shows the sequence for EF-1R with the tryptic peptides detected in bold. The human sequence isused because the porcine sequence is not available.

DDM-PGE2-Induced Proteins in Renal Cells Chem. Res. Toxicol., Vol. 16, No. 3, 2003 317

unfolded proteins and Ca2+ depletion following toxicantinsult (43, 44).

PGs induce GRP78 expression, and ∆12-PGJ2 inducesGRP78 gene expression in HeLa cells via the UPRresponse element (45). Induction of GRP78 gene expres-sion was observed following exposure of normal ratkidney cells to a synthetic PG with a cyclopentenonestructure but not a synthetic cyclopentanone PG (46).While the identification of GRP78 from DDM-PGE2-

stimulated LLC-PK1 cells described herein is consistentwith the ability of PGs to alter stress protein expression,the structure-activity relationship for PG-mediated ef-fects on stress protein expression is unclear, since DDM-PGE2 is a cyclopentanone PG.

In summary, DDM-PGE2 stimulates the synthesis ofspecific proteins, each of which is likely involved in theoverall cytoprotective response. Mass spectrometric analy-sis of proteins isolated from DDM-PGE2-stimulated LLC-PK1 cells identified several potentially cytoprotectiveproteins, namely, filamin, myosin, moesin, EF-2, HSP90â,GRP78, EF-1R, and actin. Western analysis confirmed atemporal induction of the latter five proteins. Inductionof these particular proteins suggests potential DDM-PGE2-mediated cytoprotective mechanisms involve cyto-skeletal changes, maintenance of proximal tubular epi-thelial cell polarity, activation of UPR, and possiblyincreased proximal tubular epithelial regenerative capac-ity.

Acknowledgment. This work was supported in partby awards from the National Institute of General MedicalSciences to S.S.L. (GM 56321), the National Institute ofEnvironmental Health Sciences Center Grant (P30-ES07784), and NIEHS Training Grant T32-ES07247 toK.M.T. MALDI spectra were acquired in the MassSpectrometry Facility of the University of California atSan Francisco under the direction of Dr. Al Burlingameand supported by NIH Grant NCRR RR01614.

References

(1) Ruwart, M. J. (1986) Protection of the liver against variousdamaging agents. In Biological Protection with Prostaglandins(Cohen, M. M., Ed.) pp 229-243, CRC Press, New York.

(2) Robert, A. (1979) Cytoprotection by prostaglandins. Gastroenterol-ogy 77, 761-767.

(3) Paller, M. S., Manivel, J. C., Patten, M., and Barry, M. (1992)Prostaglandins protect kidneys against ischemic and toxic injuryby a cellular effect. Kidney Int. 42, 1345-1354.

(4) Blikslager, A. T., Roberts, M. C., Rhoads, J. M., and Argenzio, R.A. (1997) Prostaglandins I2 and E2 have a syngeristic role inrescuing epithelial barrier function in porcine ileum. J. Clin.Invest. 100, 1928-1933.

(5) Yang, R. S., Fu, W. M., Wang, S. M., Lu, K. S., Liu, T. K., andLin-Shiau, S. Y. (1998) Morphological changes induced by prosta-glandin E in cultured osteoblasts. Bone 22, 629-636.

(6) Gil, L., Han, Y., Opas, E. E., Rodan, G. A., Ruel, R., Seedor, J.G., Tyler, P. C., and Young, R. N. (1999) Prostaglandin E2-bisphosphonate conjugates: potential agents for treatment ofosteoporosis. Bioorg. Med. Chem. 7, 901-919.

(7) Banan, A., Smith, G. S., Deshpande, Y., Rieckenberg, C. L.,Kokoska, E. R., and Miller, T. A. (1999) Prostaglandins protect

Table 2: Function and Localization of Proteins Identified from DDM-PGE2-Stimulated LLC-PK1 Cellsa

protein function subcellular localization

endothelial actin-bindingprotein (ABP-280; nonmuscle filamin)

promotes actin filament branchingand links actin filaments tomembrane glycoproteins

peripheral cytoplasm

nonmuscle myosin, heavy chain type A appears to play a role in cytokinesis,cell shape, and specialized functionssuch as secretion and capping

cytoplasmic

EF-2 causes GTP-dependent nascentprotein chain translocation betweenthe A and P sites of the ribosome

cytoplasmic

HSP90â molecular chaperone; refolding ofdenatured proteins

cytoplasmic

78 kDa glucose-regulatedprotein (GRP78)

ER stress protein ER lumen

EF-1R mediates GTP-dependent aa-tRNAtranslocation to the A site of the ribosome

cytoplasmic

actin cytoskeletal cytoplasmica Adapted from http://prowl1.rockefeller.edu/prowl/proteininfo.html.

Figure 5. Time-dependent induction of EF-1R, EF-2, HSP90â,GRP78, and actin following exposure of LLC-PK1 cells to DDM-PGE2. LLC-PK1 cells were exposed to vehicle (ethanol; EtOH)or 2 µM DDM-PGE2 for 4-24 h. (A) Protein lysates wereseparated by SDS-PAGE and blotted, and protein levels weredetermined as described in the Experimental Procedures sec-tion. GAPDH was used as a housekeeping protein for normal-ization of sample loading. (B) Quantification of protein expres-sion by densitometry.

318 Chem. Res. Toxicol., Vol. 16, No. 3, 2003 Towndrow et al.

human intestinal cells against ethanol injury by stabilizingmicrotubules: role of protein kinase C and enhanced calciumefflux. Dig. Dis. Sci. 44, 697-707.

(8) Kanai, N., Lu, R., Satriano, J. A., Bao, Y., Wolkoff, Schuster, V.L. (1995) Identification and characterization of a prostaglandintransporter. Science 268, 866-869.

(9) Lu, R., Kanai, N., Bao, Y., and Schuster, V. L. (1996) Cloning, invitro expression, and tissue distribution of a human prostaglandintransporter cDNA (hPGT). J. Clin. Invest. 98, 1142-1149.

(10) Bhattacharya, M., Peri, K., Ribeiro-da-Silva, A., Almazan, G.,Shichi, H., Hou, X., Varma, D. R., and Chemtob, S. (1999)Localization of functional prostaglandin E2 receptors EP3 and EP4in the nuclear envelope. J. Biol. Chem. 274, 15719-15724.

(11) Forman, B. M., Tontonoz, P., Chen, J., Brun, R. P., Spiegelman,B. M., and Evans, R. M. (1995) 15-Deoxy-δ 12, 14-prostaglandinJ2 is a ligand for the adipocyte determination factor PPARγ. Cell83, 803-812.

(12) Weber, T. J., Monks, T. J., and Lau, S. S. (1997) DDM-PGE2-mediated cytoprotection in renal epithelial cells: evidence for apharmacologically distinct receptor. Am. J. Physiol. 273, F507-F515.

(13) Weber, T. J., Monks, T. J., and Lau, S. S. (2000) DDM-PGE2-mediated cytoprotection in renal epithelial cells by a thromboxaneA2 receptor coupled to NF-κB. Am. J. Physiol. 278, F270-F278.

(14) Bradford, M. (1976) A rapid and sensitive method for thequantitation of microgram quantities of protein utilizing theprinciple dye-binding. Anal. Biochem. 72, 248-254.

(15) Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996) Massspectrometric sequencing of proteins from silver-stained poly-acrylamide gels. Anal. Chem. 68, 850-858.

(16) Gatlin, C. L., Kleemann, G. R., Hays, L. G., Link, A. J., and Yates,J. R. (1998) Protein identification at the low femtomole level fromsilver-stained gels using a new fritless electrospray interface forliquid chromatography-microspray and nanospray mass spec-trometry. Anal. Biochem. 263, 93-101.

(17) Bektas, M., Nurten, R., Gurel, Z., Sayers, Z., and Bermek (1994)Interactions of eukaryotic elongation factor 2 and actin: a possiblelink between protein synthetic machinery and cytoskeleton. FEBSLett. 356, 89-93.

(18) Liu, G., Tang, J., Edmonds, B. T., Murray, J., Levin, S., andCondeelis, J. (1996) F-actin sequesters elongation factor 1R frominteraction with aminoacyl-tRNA in a pH-dependent reaction. J.Cell Biol. 135, 953-963.

(19) Koyasu, S., Nishida, E., Kadowaki, T., Matsuzaki, F., Iida, K.,Harada, F., Kasuga, M., Sakai, H., and Yahara, I. (1986) Twomammalian heat shock proteins, hsp90 and hsp100, are actin-binding proteins. Proc. Natl. Acad. Sci. U.S.A. 83, 8054-8058.

(20) Shestakova, E. A., Motuz, L. P., and Gavrilova, L. P. (1993) Co-localization of components of the protein-synthesizing machinerywith the cytoskeleton in G0-arrested cells. Cell Biol. Int. 17, 417-424.

(21) Negrutskii, B. S., and El’skaya, A. V. (1998) Eukaryotic transla-tion elongation factor 1R: Structure, expression, functions, andpossible role in aminoacyl-tRNA channeling. Prog. Nucleic AcidsRes. Mol. Biol. 60, 47-78.

(22) Gething, M. J., and Sambrook, J. (1992) Protein folding in thecell. Nature 355, 33-45.

(23) Handler, J. A., Danilowicz, R. M., and Eling, T. E. (1990)Mitogenic signaling by epidermal growth factor (EGF), but notplatelet-derived growth factor, requires arachidonic acid metabo-lism in BALB/c 3T3 cells. Modulation of EGF-dependent c-mycexpression by prostaglandins. J. Biol. Chem. 265, 3669-3673.

(24) Glasglow, W. C., Afshari, C. A., Barrett, J. C., and Eling, T. (1992)Modulation of the epidermal growth factor mitogenic responseby metabolites of linoleic and arachidonic acid in Syrian hamsterembryo fibroblasts. Differential effects in tumor suppressor gene(+) and (-) phenotypes. J. Biol. Chem. 267, 10771-10779.

(25) Graves, L. M., Bornfeldt, K. E., Sidhu, J. S., Argast, G. M., Raines.E. W., Ross, R., Leslie, C. C., and Krebs, E. G. (1996) Platelet-derived growth factor stimulates protein kinase A through amitogen-activated protein kinase-dependent pathway in humanarterial smooth muscle cells. J. Biol. Chem. 271, 505-511.

(26) Castano, E., Dalmau, M., Marti, M., Berrocal, F., Bartrons, R.,and Gil, J. (1997) Inhibition of DNA synthesis by aspirin in Swiss3T3 fibroblasts. J. Pharmacol. Exp. Ther. 280, 366-372.

(27) Banan, A., Smith, G. S., Kokoska, E. R., and Miller, T. A. (2000)Role of actin cytoskeleton in prostaglandin-induced protectionagainst ethanol in an intestinal epithelial cell line. J. Surg. Res.88, 104-113.

(28) Edelstein, C. L., Ling, H., and Schrier, R. W. (1997) The natureof renal cell injury. Kidney Int. 51, 1341-1351.

(29) Bonventre, J. V. (1993) Mechanisms of ischemic acute renalfailure. Kidney Int. 43, 1160-1178.

(30) Raman, N., and Atkinson, S. J. (1999) Rho controls actin cyto-skeletal assembly in renal epithelial cells during ATP depletionand recovery. Am. J. Physiol. 276, C1312-C1324.

(31) Kruidering, M., van de Water, B., Zhan, Y., Baelde, J. J., Heer,E., Mulder, G. J., Stevens, J. L., and Nagelkerke, J. F. (1998)Cisplatin effects on F-actin and matrix proteins precede renaltubular cell detachment and apoptosis in vitro. Cell Death Differ.5, 601-614.

(32) Gorlin, J. B., Yamin, R., Egan, S., Stewart, M., Stossel, T. P.,Kwiatkowski, D. J., and Hartwig, J. H. (1990) Human endothelialactin-binding protein (ABP-280, nonmuscle filamin): A molecularleaf spring. J. Cell Biol. 111, 1089-1105.

(33) Gautreau, A., Poullet, P., Louvard, D., and Arpin, M. (1999) Ezrin,a plasma membrane-microfilament linker, signals cell survivalthrough the phosphatidylinositol 3-kinase/Akt pathway. Proc.Natl. Acad. Sci. U.S.A. 96, 7300-7305.

(34) Chen, J., Doctor, R. B., and Mandel, L. J. (1994) Cytoskeletaldissociation of ezrin during renal anoxia: role in microvillarinjury. Am J. Physiol. 267, C784-C795.

(35) Buchner, J. (1999) Hsp90 and Co.- a holding for folding. TrendsBiochem. Sci. 24, 136-141.

(36) Caplan, A. J. (1999) Hsp90’s secrets unfold: new insights fromstructural and functional studies. Trends Cell Biol. 9, 262-268.

(37) Huang, H., Lee, W., Lin, J. H., Jian, S. C., Mao, S. J., Yang, P.C., Huang, T. Y., and Liu, Y. C. (1999) Molecular cloning andcharacterization of porcine cDNA encoding a 90-kDa heat shockprotein and it’s expression following hyperthermia. Gene 226,307-315.

(38) Csermely, P., Schnaider, T., Soti, C., Prohaszka, Z., and Nardai,G. (1998) The 90-kDa molecular chaperone family: Structure,function, and clinical applications: A comprehensive review.Pharmacol. Ther. 79, 129-168.

(39) Cambiazo, V., Gonzalez, M., Isamit, C., and Maccioni, R. B. (1999)The â-isoform of heat shock protein hsp-90 is structurally relatedwith human microtubule-interacting protein Mip-90. FEBS Lett457, 343-347.

(40) Liu, H., Bowes, R. C., III, van de Water, B., Sillence, C.,Nagelkerke, J. F., and Stevens, J. L. (1997) Endoplasmic reticu-lum chaperones GRP78 and calreticulin prevent oxidative stress,Ca2+ disturbances, and cell death in renal epithelial cells. J. Biol.Chem. 272, 21751-21759.

(41) Liu, H., Miller, E., van de Water, B., and Stevens, J. L. (1998)Endoplasmic reticulum stress proteins block oxidant-induced Ca2+

increases, and cell death. J. Biol. Chem. 273, 12858-12862.(42) van de Water, B., Wang, Y., Asmellash, S., Liu, H., Zhan, Y.,

Miller, E., and Stevens, J. L. (1999) Distinct endoplasmic reticu-lum signaling pathways regulate apoptotic and necrotic cell deathfollowing iodoacetamide treatment. Chem. Res. Toxicol. 12, 943-951.

(43) Kaufman, R. J. (1999) Stress signaling from the lumen of theendoplasmic reticulum: coordination of gene transcriptional andtranslational controls. Genes Dev. 13, 1211-1233.

(44) Brostrom, C. O., and Brostrom, M. A. (1998) Regulation oftranslational initiation during cellular response to stress. Prog.Nucleic Acids Res. Mol. Biol. 58, 79-125.

(45) Odani, N., Negishi, Takahashi, S., Kitano, Y., Kozutsumi, Y., andIchikawa, A. (1996) Regulation of BiP gene expression by cyclo-pentenone prostaglandins through unfolded protein response. J.Biol. Chem. 28, 16609-16613.

(46) Takahashi, S., Odani, N., Tomokiyo, K., Furuta, K., Suzuki, M.,Ichikawa, A., and Negisihi, M. (1998) Localization of cyclopen-tenone prostaglandin to the endoplasmic reticulum and inductionof BiP mRNA. Biochem. J. 335, 35-42.

TX020048L

DDM-PGE2-Induced Proteins in Renal Cells Chem. Res. Toxicol., Vol. 16, No. 3, 2003 319