Proteomic Evidence for Roles for Nucleolin and Poly[ADP-ribosyl] Transferase in Drug Resistance Zongming Fu and Catherine Fenselau* Department of Chemistry and Biochemistry,University of Maryland, College Park, Maryland 20742 Received April 25, 2005
One-hundred twenty-four proteins have been identified in the soluble nuclear protein mixture from MCF-7 human breast cancer cells, of which more than 90% are classically categorized as nuclear proteins. Proteins were also studied from three drug resistant MDF-7 lines, selected previously from the same parent line by exposure to etoposide, to mitoxantrone, or to adriamycin in the presence of verapamil. Both quantitative gel comparisons and stable isotope labeling were used to identify a total of fourteen proteins whose abundances are altered by more than 2-fold in the three resistant lines. Several cytoskeleton proteins, cytokeratin 8, cytokeratin 19, septin 2, and alpha tropomyosin, are decreased in common across the three resistant cell lines. PARP-l (poly[ADP-ribosyl]transefrase or connexion) is found to be less abundant in all three resistant lines. Nucleolin is more abundant in lines resistant to etoposide and mitoxantrone, while the mitotic checkpoint protein BUB 3 is more abundant in the line resistant to adriamycin/verapamil. Keywords: nuclear proteins • drug resistance • metabolic labeling • isotope ratios • 2-D gel electrophoresis • mitoxantrone • etoposide • adriamycin • MCF-7 cancer cells
Introduction Drug resistance accounts for most failures of chemotherapy and has been described as the single most common reason for the discontinuation of a drug. It is important to understand the mechanisms of drug resistance in order to improve clinical treatment and develop new drugs. Efforts have been made to study mechanisms at the molecular and cellular levels over the past two decades, and these mechanisms may be classified as providing increased drug efflux, decreased drug uptake, enhanced intracellular detoxification, up-regulation of the DNA repair system, inadequate drug activation, and inhibition of apoptosis.1 Although it appears that they are multi-factorial, mechanisms of drug resistance are still not well understood. In this study, we have used proteomic techniques to examine changes in abundances of nuclear proteins in drug susceptible and three related drug resistant sub-lines of human breast cancer MCF-7 cells. Many chemotherapeutic agents, including alkylating agents, cross-linking agents, intercalating agents, and topoisomerase inhibitors have been designed to act in the nucleus of the cancer cell.2 The identification of nuclear proteins whose abundances are altered in acquired drug resistance may contribute to understanding more fully the mechanisms of acquired drug resistance. A comparison could also be made of the changes in proteins across the three cell lines selected for resistance to three different drugs, two with a similar mode of action in the nucleus, and one thought to act by a more diverse mechanism. Proteins fractionated by twodimensional gel electrophoresis were quantitatively compared * To whom correspondence should be addressed. E-mail: Fenselau@ wam.umd.edu. 10.1021/pr0501158 CCC: $30.25
2005 American Chemical Society
using comparative densitometry and also with metabolic isotope labeling.3-7 Nuclear proteins were analyzed in four MCF-7 cell lines. The MCF-7 cell line has been widely used to study the effects of anticancer drugs since it was established from a pleural effusion of a patient with metastatic mammary carcinoma in 1973.8 The parental line used in this study is a member of the NCI panel of 60 cell lines developed to screen new antitumor agents.9 The three drug-resistant MCF-7 cell lines studied here are all derived from this single parental cell line. The etoposide (VP-16) resistant MCF-7 line (MCF-7/VP) was developed by Dr. Ken Cowan,10 by selection during culture in increasing concentrations of etoposide. It is reported to be 28-fold resistant to etoposide (VP-16), 21-fold resistant to teniposide and 9-fold resistant to adriamycin.10 Its mode of action involves locking topoisomerase II in a cleavable complex and activating a Ca/Mg endonuclease to trigger apoptosis. P-glycoprotein is not overexpressed in these cells; the levels and sequence of topoisomerase II are little changed. Resistance in this cell line is characterized as multifactorial,10,11 and is thus appropriately investigated with proteomic techniques. The second drug resistant cell line studied is MCF-7/MX, also provided by Dr. Ken Cowan.12 MCF-7/MX cells are approximately 4000-fold resistant to the cytotoxic effects of mitoxantrone, and 50- to 180-fold cross-resistant to analogues of camptothecin, a known topoisomerase I poison. They are also approximately 10-fold cross-resistant to adriamycin and etoposide.12,13 Mitoxantrone is considered to act by inhibiting the strand passing activity of topoisomerase II. Functional assays and Western blot analysis have revealed no difference in activities or abundances of topoisomerase I and II in Journal of Proteome Research 2005, 4, 1583-1591
1583
Published on Web 09/21/2005
research articles mitoxantrone resistant MCF-7 cells.14 Mitoxantrone resistance in MCF-7 cells is not mediated by P-glycoprotein, however a new pump, termed breast cancer resistant protein (BCRP) was found overexpressed in the cell line.15 The mechanisms of drug resistance in this cell line are also considered to be multifactorial.14 The third drug resistant cell line studied here is MCF-7/ AdrVp, developed by Dr. Douglas Ross at the University of Maryland Medical School.16 It was isolated by selecting human breast cancer MCF-7 cells viable in incremental increases of adriamycin (Adr) in the presence of verapamil, a potent inhibitor of the P-glycoprotein (Pgp) membrane pump and P-glycoprotein is not overexpressed in this cell line.16 MCF-7/ AdrVp cells are 900-fold resistant to adriamycin, which is thought to act by intercalation into DNA and interference with topoisomerase II, the generation of free radicals and oxidative stress.16,17 BCRP was found overexpressed in this line,16 however other mechanisms appear to be required to account for the level of resistance observed.15 Finally, as a positive control, we used the drug resistant cell line OVCAR-8, which has been repeatedly selected by exposing MCF-7 cultures to increased concentrations of adriamycin.18 It is 192-fold resistant to adriamycin.18 Both its genetic and proteomic characteristics distinguish it from MCF-7 cells and related sub-lines.19-25
Experimental Section Cell Culture. Four MCF-7 cell lines and the OVCAR line were cultured in MEM (Sigma, St. Louis, MO) with 10% of FBS (Sigma) and 1% penicillin streptomycin. Isotope labeled control MCF-7 cells were cultured in the same MEM, except that the essential amino acids lysine and arginine were replaced with their isotopic 13C6 counterparts (Cambridge Isotope Laboratories, Cambridge, MA). Isotope labeled MCF-7 cells were processed and checked separately to make sure that isotope labeled amino acids were incorporated into the proteins in the nuclei before they were mixed with resistant cells. Every 6 months, the drug resistant cell lines were subjected to a reselection cycle of three passages with culture medium containing increased concentrations of the appropriate drugs. Isolation of Nuclei and Preparation of Nuclear Proteins. Cultured MCF-7 cells were harvested at 95% confluence. A nuclei isolation kit (Sigma, Product No. NUC-201) was used to isolate and purify MCF-7 nuclei, according to user instructions. The nuclei pellets were resuspended in a buffer containing 0.42 M NaCl, 20 mM HEPES, 1.5 mM MgCl2, 0.2 mM EDTA, 25% (v/v) Glycerol, PH7, and then were vortexed vigorously for 15 s every 10 min, for a total of 40 min on ice. Then the suspension was centrifuged at 16 000 × g for 10 min. The supernatant fraction was collected and relative protein concentrations were measured using the Bio-Rad DC Protein Assay (Bio-Rad, Hercules, CA). Bovine serum albumin (BSA) was used to make the standard curve. 2-D Gel Electrophoresis. Three-hundred fifty micrograms of protein were used to conduct each 2-D gel electrophoresis. Before loading the first dimension strip, the protein sample was desalted with Biospin-6 spin columns (Bio-Rad) and added to three and a half micrograms of micrococcal nuclease and incubated at 37 °C for 30 min to eliminate possible DNA or RNA contamination. The sample was then dried by vacuum centrifugation and 320 µL of rehydration buffer was added, which contained 7 M urea, 2 M thiourea, 2% CHAPS, 50 mM 1584
Journal of Proteome Research • Vol. 4, No. 5, 2005
Fu and Fenselau
DTT and 1% of IPG buffer (Pharmacia, Piscataway, NJ). The solution was incubated at room temperature for 1 h. Seventeen cm pH 3-10 NL ReadyStripTM IPG strips (Bio-Rad)) were used for the first dimension isoelectric focusing (12 h rehydration, 60 000 hv), 193 × 183 × 1.0 mm,; 8-16% Tris-HCl IPG COMB precast gels (Bio-Rad) were used for the second dimension SDS polyacrylamide gel electrophoresis (16 mA for 30 min and then 24 mA for 4 h and 40 min). At the end of the run, the gels were removed and fixed in a solution containing acetic acid and methanol. Staining was performed with Bio-Safe colloidal Coomassie blue G-250 (Bio-Rad,) for 2 h. The gels were then rinsed with water until the desired contrast was achieved for densitometric image analysis. Analysis of Gel Images. A GS-800 calibrated densitometer (Bio-Rad,) was used to visualize Coomassie stained gel slabs. After the gels were scanned, the images were exported as TIFF files. The TIFF files of 2-D gel images of control and drug resistant MCF-7 nuclear proteins were compared using Compugen Z3 software (Compugen Ltd., Tel Aviv, Israel). In each gel, the intensity of a spot was measured and recorded. Ratios were calculated between corresponding spots on two plates. The image correspondence generated by the program was checked manually. Spots of interest were also examined using enlarged images. For quantitative comparisons, measurements were made on 3 to 8 separate cell harvests, with 3 to 5 gels from each. Mass Spectrometry Analysis. After analysis of the gel image, the spots of interest were excised and subjected to tryptic ingel digestion as described by Jensen et al.26 The peptides generated were extracted as described and desalted using ZipTip C18 pipet tips (Millipore, Bellerica, MA). Routinely, the desalted samples were first analyzed using an AXIMA-CFR MALDI -TOF (Kratos, Chestnut Ridge, NY) mass spectrometer. The instrument was operated in the reflectron positive ion mode. The laser power was set at 45-50 on the manufacturer’s scale of 0 to 180. The monoisotopic peaks of trypsin autolysis products, m/z 842.51 and 2211.11, were used for internal calibration of each spectrum, or melittin and angiotensin II were used as external calibrants. An aliquot of 0.5 ul of sample solution was put on the plate and covered with 0.5 µL of 50 mM alpha-hydroxycinnamic acid in 0.1% trifluoroacetic acid/70% acetonitrile, then dried and introduced into the instrument. Spectra were recorded by accumulating 50-100 laser shots, depending on the quality of the sample. The data from the spectra were used to identify proteins by peptide mass fingerprinting, using search programs from www.matrixscience.com. The NCBI and Swiss-Prot databases were searched. Identifications were confirmed manually by checking for consistency in the pI and molecular weight values provided by the 2-D gel array. If no conclusive identification could be acquired, then the sample was analyzed using nanospray tandem mass spectrometry to identify proteins by microsequencing on a hybrid quadrupole time-of -flight (TOF) mass spectrometer, QSTAR Pulsar Qq-TOF (Applied Biosystems, Foster City, CA). Typically, 2 ul of the peptide solution in acetic acid/methanol/ water (2:49:49, v/v/v) was loaded into a capillary tip (Protana, Odense, Denmark) and mounted in the nanospray source. Spray voltage was set at 900 V. MS scan range was from 350 to 1200 m/z. Doubly or triply charged peptide ions were selected for sequencing. Collision energy was ramped from 15 to 50 eV. The BioExplore program integrated with the instrument was used to identify proteins using the resulting MS/MS spectra. Other database search programs
Role for Nucleolin and PARP-1 in Drug Resistance
research articles
such as TagIdent (http://www.expasy.ch) and MS-Blast (http://www.dove.embl-heidelberg.de/blast2/msblast.html) were used as necessary. Relative Protein Quantitation by Isotope Labeling. 13C6lysine and 13C6-arginine were introduced into the drug susceptible MCF-7 cells grown through five passages and the extent to which lysine and arginine were replaced by 13C6-lysine and 13 C6-arginine counterparts in the nuclear proteins was evaluated by mass spectrometry as reported by Gehrmann et al.7 Labeled and unlabeled cells were mixed in approximately 1:1 ratios, and peptides from proteins whose abundances were known from 2-D gel images not to be modified were examined to normalize the mixing ratios of cells more precisely. Relative abundance (RA) is defined as the peak area of the monoisotopic peak from the unlabeled peptide [U] divided by the peak area of the monoisotopic peak from the labeled peptide [L]. The RA value was corrected for incomplete labeling according to the following equation, where X is the percent incorporation of the labeled amino acid in protein from the labeled cells: RA )
[U] - (100 - X)/X[L] 100[L]/X
Quantitative comparisons were averaged from 3 separate cell harvests.
Results Typically, (6-8) × 106 cells were harvested from a 150 cm2 flask. Their total wet weight ranged between 0.17 and 0.20 g. One gram of cell pellet provided between 0.17 nad 0.24 g of nuclei, from which 7.5 to 9.0 mg of nuclear proteins could be recovered. Following separation by 2-D gel electrophoresis and in-gel digestion, the peptide products were subjected to mass spectrometry analysis. Typically, each protein identified by peptide mass fingerprinting had sequence coverage above 30%. In order for a protein identification to be considered valid, at least two sequence tags were required. The Mascot scores all exceed the 95% confidence criteria set by the software. For acceptable identifications the expected molecular weights and pI values of each protein fit well with its position in the 2-D gel map. Figure 1 shows an annotated 2-D gel electrophoresis map of nuclear proteins from control MCF-7 cells. Proteins identified are listed in Table 1. Although 161 spots were identified, only 124 different proteins were tallied. Some proteins or isoforms appeared in more than one spot in the gel. For example, spots 134, 135, 136, and 137 contained isoforms of protein hnRNP L. In rare cases, a spot was found to contain more than one protein. For example, two proteins, BAF53 (pI 5.39, MW 47430) and hnRNP F (pI 5.38, MW 45985) were identified in spot 72. Proteins having significantly smaller molecular weights (based on gel migration) than their theoretical values were judged to be truncated. After 5 passages, the cells cultured in growth media containing 13C6- arginine and 13C6-lysine were harvested and nuclear proteins were extracted. The protein sample was subjected to 2-D gel electrophoresis. The gel images were digitized and compared with those from control MCF-7 cells cultured in regular media. The images were found to overlap well. Spectra for all the peptides analyzed showed that the percent of isotope incorporation was 90 ( 2%. The expression profile of etoposide (VP-16) resistant MCF-7 cells was compared with that of control MCF-7 cells, using both
Figure 1. Annotated 2-D gel map of nuclear proteins from control MCF-7 cells. Numbers code the proteins that were identified.
densitometry and metabolic labeling methods. Of the 160 proteins and isoforms studied, abundances of 9 proteins were found to be altered by more than 2-fold in etoposide resistant cells. Nucleolin, HMG 1, 40 s ribosomal protein SA and cyclophilin B are more abundant. Cytokeratin 8, cytokeratin 19, septin 2, PARP-1 and alpha tropomyosin were less abundant. Figure 2 shows an example of gel arrays and mass spectra of two peptides from a protein whose abundance levels are the same in two cell lines. The protein was identified as nucleophosmin. It was subsequently used as a control for cell mixing. Figure 3 shows gel arrays and mass spectra of peptides from nucleolin, which reveal that it is more abundant in MCF-7/VP cells. Figure 4 shows gel arrays and mass spectra of peptides from septin 2, which was less abundant in MCF-7/VP cells. Supporting Information Table 1 (see the Supporting Information) summarizes protein relative abundances in VP-16 resistant cells, determined with both methods. The proteins with significantly altered abundances (more than 2-fold) are summarized in Table 2. The nuclear proteins from mitoxantrone resistant MCF-7 cells were analyzed by both densitometric comparison and metabolic labeling methods, however in this experiment the selection of spots to be analyzed by the metabolic labeling method was guided by the results of densitometric comparisons. Thus the number of spots analyzed by isotope ratios was reduced. Prohibitin, 78K glucose-regulated protein (GRP78), HMG-1, 40S ribosomal protein SA, and nucleolin were found to be significantly more abundant (more than 2-fold). Cytokeratin 19, cytokeratin 8, septin 2, PARP-1, and alpha tropomyosin were found to be less abundant (Table 2 and Supporting Information Table 2). The nuclear proteins from MCF-7/AdrVp were also analyzed by both densitometric comparison and isotope labeling methods. In this line, mitotic checkpoint protein BUB 3 and cyclophilin B were found more abundant, while septin 2, septin 7, cytokeratin 19, cytokeratin 8, and alpha tropomyosin were Journal of Proteome Research • Vol. 4, No. 5, 2005 1585
research articles
Fu and Fenselau
Table 1. Proteins Identified from Nuclear Extraction of Control MCF-7 Cell
1586
spot noa
protein ID
pI
MW
accession no.
1 2-3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29-30 31 32 33 34 35 36 37-38 39 40 41 42 43 44 45 46 47-48 49 50-53 54 55-57 58 59 60 61-63 64 65 66-68 69 70 71 72/1 72/2 73 74-75 76 77 78-83 84 85-86 87 88 89 90 91 92 93-94
Micrococcal Nuclease snRNP-G Signal recognition particle 9Kd Pro P10 protein Ubiquitin U6 snRNA-associated Sm-like protein LSm2 S100 calcium-binding protein A16 S100 calcium-binding protein A13 Enhancer of rudimentary homolog snRNP-F U6 snRNA-associated Sm-like protein LSm7 60S acidic ribosomal protein P2 U6 snRNA-associated Sm-like protein LSm3 NHP-2 like protein 1 Single-stranded DNA-binding protein 40S ribosomal protein S12 Putative RNA-binding protein 3 VAMP-associated protein DIMI protein homolog 60s ribosomal protein L23a Cyclophilin B NIF3L1BP1 protein 40S ribosomal protein S7 HMG 2 HMG-4L FK506-binding protein 3 PSA 2 HMG-1 HSP 27 eIF4E 40 S ribosomal protein SA Prohibitin Breast carcinoma amplified sequence 2 U2 small nuclear ribonucleoprotein A′ Guanine nucleotide-binding protein beta PSA 1 Nuclear protein Hcc-1 Annexin IV Annexin V Splicing factor SC35 EF-1-beta Alpha tropomyosin Guanine Nucleotide -binding protein hnRNP H3 Annexin II hnRNP A2/B1 hnRNP A1 60 S acidic ribosomal protein p0 CapZ alpha-1 eIF-3 beta 2 Nucleophosmin Cytokeratin 19 hnRNP C Set Protein Cytokeratin 8 (truncated) Actin beta Transcription factor NF-AT 45K chain Elongation initiation factor 4A-1 BAF 53 hnRNP F PP1A hnRNP Do hnRNPA3 Mitotic checkpoint protein BUB 3 hnRNP A/B THO complex subunit 3 Actin gamma Calcium-binding transporter Septin 2 Creatine kinase Septin 7 Elongation factor 1-alpha 1 Regulator of chromosome condensation Proliferation-association pro 2G4
9.8 8.98 8.27 7.3 6.56 6.04 6.28 5.91 5.63 4.7 5.1 4.42 4.58 8.72 9.59 6.36 8.86 6.85 5.53 10.44 9.33 5.69 10.09 7.77 8.45 9.29 7.12 5.62 5.98 5.79 4.79 5.57 5.48 8.72 7.6 6.15 6.1 5.85 4.94 11.88 4.5 4.72 5.6 6.37 7.56 8.97 9.26 5.71 5.45 5.38 4.47 5.04 4.95 4.12 4.76 5.15 8.26 5.32 5.39 5.38 5.94 7.62 8.74 6.36 6.49 5.7 5.31 5.31 6.25 8.6 8.85 9.1 7.18 6.13
19299 8490 9974 11064 8560 10828 11794 11464 12422 9776 11595 11658 11707 14165 17249 14728 17160 27211 16775 17684 22785 23039 22113 24059 21125 25161 25865 24747 22768 25082 32833 29786 26115 28398 35055 29822 23656 35729 35783 25560 24617 32856 37307 36903 38677 37464 37464 34252 32903 36479 32746 44079 33667 32084 53515 42408 44897 46352 47430 45985 37488 38581 39947 37131 35945 38747 41766 45790 41689 47007 48756 50451 44941 44101
P43209 Q15357 P49458 P60903 P02248 Q9Y333 Q96FQ6 Q99584 Q14259 P49458 Q9UK45 P05387 Q9Y4Z1 P55769 Q04837 P25398 P98179 O95292 O14834 P29316 P23284 Q6I9Y2 P62081 P26583 Q9UJ13 Q00688 P25787 P09429 P04792 P06730 P08861 P35232 O75934 P09661 P25388 P25786 P82979 P09525 P08758 Q01130 P24534 P09493 P11016 P31942 P07355 P22622 P09651 P05388 P52907 Q13347 P06748 P08727 P07910 Q01105 P05787 P60709 gi|1082855 P04765 O96019 P52597 P62136 Q14103 P51991 O43684 Q99729 Q96J01 P02571 Q9P129 Q15019 P12532 Q16181 P04720 P18754 Q9UQ80
Journal of Proteome Research • Vol. 4, No. 5, 2005
research articles
Role for Nucleolin and PARP-1 in Drug Resistance Table 1. (Continued) 95 96 97 98 99-100 101 102 103 104 105 106 107 108 109 110 111 112-115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134-137 138-139 140-141 142 143-144 145 146-147 148-149 150 151 152-153 154 155-156 157-158 159 160 161 a
Elongation factor 1-gamma Cleavage stimulation factor DEK oncogene RUV-like 2 HLA-B associated transcript-1 hnRNP H Septin 6 Glutamate dehydrogenase 1 ATP synthase alpha chain SRP 54 54 kDa DNA- and RNA-binding Protein ATP synthase beta chain HAT type B subunit CAF-1 subunit Splicing factor 3A subunit 3 hnRNP K (truncated) hnRNP K Copine III T-complex protein 1, beta Nuclear matrix pro 200 RuvB-like 1 Aspartyl-tRNA synthetase Glutamate dehydrogenase 1 U4/U6 sn RNP Phenylanyl-tRNA synthetase alpha chain Probable RNA-dependent helicase p68 T-complex protein 1, alpha Nucleolin (76K) Heat shock cognate 71 kDa HSP 70-1 Stress-70 protein Annexin VI Delta coat protein Stress-induced-phosphoprotein 1 Phenylanyl-tRNA synthetase beta chain hnRNP L hnRNP Q ATP-dependent DNA helicase II, 80kDa subunit licensing factor ATP-dependent DNA helicase II, 70kDa subunit ATP-dependent helicase DDX1 EF-2 C1-THF synthase Splicing factor PQ PARP-1 Structural maintenance of chromosome 3 Structural maintenance of chromosome 1 Bifunctional aminoacyl-tRNA synthetase TER ATPase Nucleolin (100K) hnRNP U GRP 78
6.25 6.12 8.69 5.49 5.44 5.89 6.24 7.66 9.16 8.87 9.01 4.64 4.89 4.74 5.27 5.39 5.39 5.6 6.01 6.14 6.02 6.11 7.66 7.05 7.46 9.06 5.8 4.41 5.37 5.48 5.87 5.42 5.89 6.4 6.4 6.65 8.68 5.55 6.08 6.23 6.81 6.42 6.94 9.45 8.99 6.77 7.51 7.77 5.14 4.41 5.76 5.16
50429 48324 42933 51125 49416 49198 49685 61359 59714 55668 54066 56525 48132 47911 58812 51230 51230 60947 57452 55603 50538 57100 61359 59097 57396 69105 60306 76224 71082 70294 73635 75695 57174 63227 66088 60719 69590 83091 81884 69953 82380 96116 101364 762116 113680 141454 143144 161923 89950 76224 90423 78000
P26641 Q05048 P35659 Q9Y230 Q13838 P31943 Q14141 P00367 P25705 P61011 Q15233 Q14283 Q16576 Q09028 Q12874 Q07244 Q07244 O75131 P78371 Q9UMS4 Q9Y265 P14868 P00367 O43172 Q9Y285 P17844 P17987 P19338 P11142 P08107 P38646 P08133 P48444 P31938 Q9NSD9 P14866 O60506 P13010 P33993 P12956 Q92499 P13639 P11586 P23246 P09874 Q9UQE7 Q14683 P07814 P55072 P19388 Q00839 P11021
The numbers identify spots shown in Figure 1.
found to be less abundant (Table 2 and Supporting Information Table 3). In the positive control experiment, the nuclear proteins from OVCAR-8 were analyzed with comparative densitometry only. As expected, the expression profile of nuclear proteins from these cells was found to be dramatically different from that of control MCF-7 cells. Supporting Information Table 4 lists nuclear proteins for which abundance changes greater than 2-fold were measured in the OVCAR cells. About one-fifth of the nuclear proteins that are detected in parental MCF-7 cells are not detected in 2-D gel arrays of nuclear proteins from OVCAR-8 cells.
Discussion A 2-D gel map of soluble nuclear proteins from MCF-7 cancer cells has been obtained in this research. Our samples
contained highly enriched soluble nuclear proteins. More than 90% of those identified are classically cataloged as nuclear proteins or proteins that may be present in the nucleus. Some of the “nonnuclear” proteins have also been found by other researchers in nuclear protein preparations.27,28 For example, ATP synthase beta chain (spot 104) is classified as a mitochondrial protein, however it has also been found in the proteomic analysis of human nucleolus by Andersen et al.27 Its observation by more than one laboratory may indicate that this protein is closely associated with the nucleus. Changes in protein abundances are reported in this paper. The proteomic methods used do not measure changes in expression or regulation. We used densitometric comparisons and isotope labeling methods to analyze differentially abundant soluble nuclear proteins in drug resistant and susceptible MCF-7 cells. The results obtained by the two methods are Journal of Proteome Research • Vol. 4, No. 5, 2005 1587
research articles
Figure 2. Nucleophosmin has no change in abundance in MCF-7/VP cells. (A) Enlarged images showing nucleophosmin with equal abundance in control (left) and MCF-7/VP (right) cells. (B) Metabolic labeling showing nucleophosmin with equal abundance.
Fu and Fenselau
Figure 4. Septin 2 has a lower abundance in MCF-7/VP cells. (A) Enlarged images showing septin 2 with lower abundance in MCF-7/VP cells (right) than in control MCF-7 cells (left). (B) Metabolic labeling showing septin 2 with lower abundance in MCF-7/VP cells. Table 2. Summary of Abundance Changes in Three Drug Resistant MCF-7 Cell Lines*
Higher abundance v. Lower abundance V. * On the basis of labeling. ** On the basis of densitometric comparison. (All measurements were made on three separate harvests.
Figure 3. Nucleolin has a higher abundance in MCF-7/VP cells. (A) Enlarged images showing nucleolin with higher abundance in MCF-7/VP cells (right) than in control MCF-7 cells (left). (B) Metabolic labeling showing nucleolin with higher abundance in MCF-7/VP cells.
consistent (Supporting Information Tables 1-3) in most cases. Two-dimensional gel electrophoresis remains the technique of choice in many proteomic laboratories. It can separate and provide a global view of many proteins, and importantly, it allows isoforms to be detected and analyzed. After the gel images are digitized, they can be compared conveniently at very low cost. For densitometric comparisons, reproducibility of sample preparations is crucial. The poor quality of isoelectric focusing of extremely basic proteins poses a special problem for determining their relative abundances. For a protein with extremely high abundance, computer software sometimes does not properly recognize the spot as comprising a single protein. This may also cause comparison errors. Densitometric comparisons also cannot deal well with a spot containing more than 1588
Journal of Proteome Research • Vol. 4, No. 5, 2005
one protein. On the other hand, isotope labeling methods can correct for process variations. The relative abundances of several proteins in one spot can be differentiated. However, the high cost of isotopes is a disadvantage, and it is readily applicable only to cultured cells. Table 2 summarizes the changes in abundance observed in nuclear proteins from the three drug resistant cell lines. Some of these proteins may play important roles in conferring drug resistance in MCF-7 cancer cells. Proteins are considered first, whose abundances are found to be decreased by at least 2-fold. These entries in Table 2 mostly comprise cytoskeletal proteins, including alpha tropomyosin (spot 45 in Figure 1), cytokeratin 19 (spot 61-63), cytokeratin 8 (spot 66-68) and septin 2 (spot 88), found to be less abundant in MCF-7/VP, MCF-7/MX, and MCF-7/AdrVp cells than in control MCF-7 cells. Septin 7 was found to be less abundant in MCF-7/AdrVp only. Cytoskeletal protein functions range from cellular architecture to signal transduction and apoptosis.29 Cytokeratins are common contaminants from
Role for Nucleolin and PARP-1 in Drug Resistance
human fingers, however in our experiments the cytokeratins from the control cells were found to carry isotope labels, which ruled out contamination as their source. Several papers have already linked differential expression of these proteins to drug resistance.30-32 Cytokeratins and alpha tropomyosin have been reported to be up- and down-regulated, respectively, in breast cancer cells compared to normal breast cells.33 Tropomyosin has been found to be down-regulated in drug resistant SNU 638 gastric cancer cells compared to their parental cells.32 However, the nucleus is not the major subcellular location of these proteins, and the results observed in this research may indirectly reflect the abundances of cytoskeletal proteins in the whole cell. Analysis of whole cell lysate should give more conclusive evidence on the cause and implications of decreased abundances of these proteins. PARP-1 (spot 151) was found to be significantly less abundant in MCF-7 cells resistant to etoposide (MCF-7/VP) and mitoxantrone (MCF-7/MX) than in drug susceptible MCF-7 cells, with ratios of 0.40 ( 0.06 and 0.46 ( 0.06, respectively (Table 6). It also has a lower abundance in MCF-7/AdrVp (0.65(0.10), which did not quite meet the 2-fold threshold set for “significant” change (Table 2). PARP-1 is a 114 kDa nuclear zinc-finger enzyme, which catalyzes the attachment of ADP ribose units to target proteins.34 It is now recognized that PARP-1 activates a unique signaling pathway leading to apoptosis.35-40 Following genotoxic stimulation. PARP-1 is activated to carry out massive synthesis of poly [ADP-ribose], which subsequently causes NAD+ depletion. Poly [ADP-ribose] formation and NAD+ depletion trigger mitochondrial depolarization and release of a mitochondrial apoptosis-inducing factor (AIF), which promotes programmed cell death through a caspase-independent pathway. Recent studies show that DNA-damaging agents selectively induce cell death through this pathway, independent of p-53 or Bcl-2 family proteins.39,40 In addition, independent research has shown that PARP-1 is upregulated at the early stage of apoptosis induced by UV irradiation.41 Mouse cells lacking PARP-1 have been observed to have increased resistance to anticancer therapy.42 Overproduction of PARP-1 has been implicated in the molecular pathway leading to cell death by energy depletion following stress.43 In the present research, the lower abundance of PARP-1 is closely correlated with resistance to several drugs. The lowered abundance potentially contributes to drug resistance in MCF-7 cells by repressing cell death through the PARP-1 signaling pathway. When proteins in Table 2 are considered whose abundances are increased, it can be seen that changes are more similar in cells that are resistant to the two drugs with similar modes of action, etoposide, and mitoxantrone. Only cyclophilin B is increased in all three of the resistant MCF-7 lines. Among many functions, cyclophilin B (peptidyl-prolyl isomerase) has been shown to interact with the mammary hormone prolactin to potentiate proliferation and cell growth as an intra-nuclear transcriptional inducer.44 Two forms of nucleolin were detected in MCF-7 cells. These can be seen in the 2-D gel map (Figure 1) around 100 KDa and 76 KDa (spots 126 and 159), and peptide mapping indicates that they differ in the lengths of their N-terminal domains, Nucleolin of 76K Da was found to have significantly higher abundances in MCF-7/VP and MCF-(Table 2). Abundances of both forms are only marginally decreased in the MCF-7/AdrVp line. Nucleolin is a multifunctional nuclear protein with high abundance.45 As its contribution to drug resistance in cancer
research articles cells, we focus on its functions in apoptosis and DNA repair. Apoptosis in U937 leukemia cells is accompanied by a lowered level and altered localization of nucleolin within the nucleus.41 It is suggested that apoptosis induced by taxol and okadaic acid treatment in HL-60 cells occurs through a process that involves down-regulation of nucleolin and destabilization of bcl-2 mRNA. Nucleolin was identified as a binding protein for AU-rich elements involved in the stabilization of mRNA of bcl-2, a known apoptosis inhibitor.46 Nucleolin has also been identified as a stress-responsive protein. Nucleolin expression levels were found to be increased in Chinese hamster ovary (CHO) cells after UV or ionizing radiation exposure.47 It has been reported to form a complex with replication protein A after cell stress to prevent initiation of DNA replication and to mobilize DNA repair.48 The C-terminal domain of nucleolin accelerates nucleic acid annealing. Its ability to promote homologous DNA pairing in vitro may indicate a similar function in vivo.49 It has been demonstrated that nucleolin interacts with topoisomerase I. It is not known if nucleolin interacts with topoisomerase II, the target of etoposide and mitoxantrone. Taken together, the literature cited suggests that up-regulation of nucleolin may inhibit apoptosis and enhance DNA repair in drug resistant MCF-7 cells. High mobility group protein 1(HMG-1) (also called HMGB-1) (spots 29, 30 in Figure 1) was found to be more abundant in MCF-7/VP and MCF-7/MX cells. HMG1 is a nonhistone chromosomal protein, which seems to play an architectural role in the assembly of nuclear protein complexes in a variety of biological processes, including DNA repair, but the mechanism of these processes is not clear.50,51 It has been found to inhibit cell death in yeast and mammalian cells and to be abundantly expressed in human breast carcinoma,52 The higher abundance of HMG1 may contribute to drug resistance to VP-16 and mitoxantrone by enhancing DNA repair, or modulating chromatin structure, thus interfering with drug activity. It is widely accepted that glucose regulated protein 78 has anti-apoptotic activity,53,54 and its overexpression has been reported by others to occur in association with drug resistance.55,56 In the present study it is found with increased abundance in mitoxantrone resistant cells (4.91 ( 0.46), though not in the cell line resistant to adriamycin in the presence of verapamil. Its abundance is not changed in MCF-7/VP cells. It is classified as a molecular chaperone localized in the endothelial reticulum,57 and its presence in the nucleus is most likely due to translocation after genotoxic stress.58 Among the cells lines examined here, prohibitin (spot 34 in Figure 1) was found to be more abundant only in the mitoxantrone resistant line. Prohibitin is a 30 KDa protein, evolutionarily conserved in a variety of organisms.59 It is mainly located in the inner membrane of mitochondria, and is recruited into the nucleus in response to genotoxic stress.60 Prohibitin has been shown to be up-regulated in association with cisplatin resistance in a head and neck carcinoma cell line61 and to block apoptosis induced by the topoisomerase I inhibitor camptothecin in a human B cell line.62 It is not clear yet if prohibitin acts as a cell survival or anti-apoptotic factor.63,64 Mitotic checkpoint protein BUB 3 (spot 77) is more abundant only in MCF-7/AdrVp cells (Table 1). Its abundance in MCF-7/VP and MCF-7/MX is unchanged, relative to control MCF-7 cells. This protein plays an important role in the function of the mitotic spindle checkpoint.65,66 BUB 3, together with two other checkpoint proteins BUB1 and BUB2, has been Journal of Proteome Research • Vol. 4, No. 5, 2005 1589
research articles observed to be overexpressed in gastric cancer.67 For MCF-7 cancer cells under genotoxic stress caused by chemotherapeutic agents, the higher abundance of mitotic protein BUB 3 may mean enhancement of this checkpoint in the cell cycle, allowing enough time for DNA repair, cytoskeleton assembly, and correction of defects caused by the drug. Overall, in the present research, we have found that MCF-7/VP, MCF-7/MX, and MCF-7/AdrVp cell lines have some similarities in their patterns of abundance changes in nuclear proteins, and also some important differences. They all have lower abundances of cytoskeletal proteins and increased abundance of cyclophilin. Beyond that, it appears that MCF-7/VP and MCF-7/MX have more in common with each other than with the MCF-7/AdrVp line. For example, changes in nucleolin and PARP-1 would seem to have high potential to deal with genotoxic stress and resist cell death, and similar changes are observed in the two cell lines. On the other hand, the nucleolin level in MCF-7/AdrVp cells is about the same as that in control MCF-7 cells, and the abundance of PARP in MCF-7/AdrVp is lower than that in control MCF-7 cells, but does not exceed the threshold of 2-fold set for “significant” change. As pointed out earlier, etoposide and mitoxantrone are both topoisomerase II poisons, sharing similar mechanisms of action and structures. Only in the third resistant line, MCT-7/ AdrVp, an important cell cycle regulation protein, mitotic checkpoint protein BUB 3 is found to be more abundant. Thus, resistance in that line may derive from other mechanisms. Our objective is to contribute to understanding the mechanisms of acquired drug resistance in cancer cells. Proteomics strategies allow analysis of many proteins at once and focus the list of targets for further investigation. Using a proteomics approach, we have been able to identify some nuclear proteins whose changes of abundance appear to correlate with drug resistance in MCF-7 breast cancer cells. These proteins are actively involved in chromatin structure, DNA repair, cell cycle regulation and protein folding. This report is the first indication that most of them might contribute to drug resistance. Once their roles in drug resistance are confirmed by molecular biological or other methods, these proteins can be exploited as biomarkers for diagnosis of the development of drug resistance, for the design of individualized therapy, and as new drug targets. The results of the studies reported here support the hypothesis that drug resistance in cancer cells is a multi-factorial process. Many proteins are involved directly or indirectly, although one or a few proteins may play a dominant role. Different sets of proteins may be involved to deal with different genotoxic assaults.
Acknowledgment. This work was supported by the NIH Grant No. GM21248. We thank Drs. K. H. Cowan (Eppley Cancer Center, University of Nebraska), D.D. Ross and P. Gutierrez (Greenebaum Cancer Center, University of Maryland) for MCF-7 cell lines. We also thank Dr. Y. Hathout for helpful discussions and Mr. A. Chertov for technical assistance. Supporting Information Available: Supporting Information Table 1: Relative abundance of nuclear proteins from MCF-7/VP cells. Supporting Information Table 2: Relative abundance of nuclear proteins from MCF-7/MX cells. Supporting Information Table 3: Relative abundance of nuclear proteins from MCF-7/AdrVp cells. Supporting Information Table 4: Nuclear proteins having abundance changes in 1590
Journal of Proteome Research • Vol. 4, No. 5, 2005
Fu and Fenselau
OVCAR-8 cells. This material is available free of charge via the Internet at http://pubs.acs.org.
References (1) Gottesman, M. M. Annu. Rev. Med. 2002, 53, 615-627. (2) Gutierrez, P. L.; Desai T. T. BioMedicina 1999, 2, 235-240. (3) Oda, Y.; Huang, K.; Cross, F. R., Cowburn, D.; Chait, B. T. Proc. Natl. Acad. Sci U.S.A. 1999, 96, 6591-6596. (4) Chen, X.; Smith, L. M.; Bradbury, E. M. Anal. Chem. 2000, 72, 1134-1143. (5) Jiang, H.; English, A. M. J. Proteome Res. 2002, 1, 345-350. (6) Ong. S.-E.; Blagoev. B.; Dratchmarove, I.; Kristensen, D. B.; Steen, H.; Pandey, A.; Mann M. Mol. Cell Proteomics 2002, 1, 376-386. (7) Gehrmann, M. L.: Hathout, Y.: Fenselau, C. J. Proteome Res. 2004, 3, 1063-1068. (8) Soule, H. D.; Vazquez, J.; Brennan, M. J. Natl. Cancer Inst. 1973, 51, 1409-1416. (9) Weinstein, J. N.; Buolamwini, J. K. Curr. Pharmaceut. Design 2000, 6, 473-483. (10) Schneider, E.; Horton, J. K.; Yang, C. H.; Nakagawa, M.; Cowan, K. H. Cancer Res. 1994, 54, 152-158. (11) Cory, A. H.; Cory, J. G. Adv. Enzyme Regul. 2001, 41, 177-188. (12) Nakagawa, M.,; Schneider, E.; Dixon, K. H.; Horton, J. K.; Kelley, K.; Morrow, C.; Cowan, K. H. Cancer Res. 1992, 52, 6175-6181. (13) Yang, C. J.; Horton, J. K.; Cowan, K. H.; Schneider, E. Proc. Am. Assoc. Cancer Res. 1996, 37, 308. (14) Yang, C. J.; Cowan, K. H.; Schneider, E. 1995 Cancer Res. 1995, 55, 4004-4009. (15) Doyle, L. A.; Ross, D. D. Oncogene 2003, 22, 7340-7358. (16) Ross, D. D.; Yang, W.; Abruzzo, L. V.; Dalton, W. S.; Schneider, E.; Lage, H.; Dietel, M.; Greenberger, L.; Cole, S. P.; Doyle, L. A. Natl. Cancer Inst. 1999, 91, 429-433. (17) Gewirtz, D. Biochem. Pharmacol. 1999, 57, 727-741. (18) Batist, G.; Tulpule, A.; Sinha, B. K.; Katki, A. G.; Myers, C. E.; Cowan, K. H. J. Biol. Chem. 1986, 261, 15544-15549. (19) Scudiero, D. A.; Monks, A.; Sausville, V. A. J. Natl. Cancer. Inst. 1998, 90, 862. (20) Roschke, A. V.; Tonon, G.; Gehlhaus, K. S.; McTyre, N.; Bussey, K. J.; Lababidi. S.; Scudiero, D. A.; Weinstein, J. N.; Kirsch, I. R. Cancer Res. 2003, 63, 8634-8647. (21) Fairchild, C. R.; Ivy, P. S.; Kao-Shan, C. S.; Whang-Peng, J.; Rosen, N.; Israel, M. A.; Melera, P. W.; Cowan, K. H.; Goldsmith, M. E. Cancer Res. 1987, 47, 5141-5148. (22) Devarajan, E.; Chen, J.; Multani, A. S.; Pathak, S.; Sahin, A. A.; Mehta, K. Int. J. Oncol. 2002, 20, 913-920. (23) Pirinia, F.; Breuleux, M.; Schneider, E.; Hochmeister, M.; Bates, S. E.; Marti, A.; Hotz, M. A.; Betticher, D. C.; Borner, M. M. J. Natl. Cancer. Inst. 2002, 92, 1535-1536. (24) Mehta, K.; Devarajan, E.; Chen, J.; Multani, A.; Pthak, S. J. Natl. Cancer. Inst. 2002, 94, 1652-1654. (25) Hathout, Y.; Gehrmann, M. L.; Chertov, A.; Fenselau, C. Cancer Lett. 2004, 210, 245-253. (26) Jensen, O. N.; Wilm, M.; Shevchenko, A.; Mann, M. Methods in Molecular Biology. Link, A. J., Ed.; Human Press: Totowa, New Jersey, 1999, 112, 513-530. (27) Andersen, J. S.; Lyon, C. E.; Fox, A. H.; Leung, A. K.; Lam, Y. W.; Steen, H.; Mann, M.; Lamond, A. I. Curr. Biol. 2002, 12, 1-11. (28) Scherl, A.; Coute´, Y.; De´on, C.; Aleth Calle´, A.; Kindbeite, K.; Sanchez, J. C.; Greco, A.; Hochstrasser, D.; Diaz, J. J. Mol. Biol. Cell. 2002, 13, 4100-4109. (29) Hanrahan, J.; Snyder, M. Mol. Biol. Cell. 2003, 12, 663-673. (30) Bauman, P. A.; Dalton, W. S.; Anderson, J. M.; Cress, A. E. Proc. Natl. Acad. Sci. USA. 1994, 91, 5311-5314. (31) Bichat, F.; Mouawad, R.; Solis-Recendez, G.; Khayat, D.; Bastian, G. Anticancer Res. 1997, 17, 3393-3401. (32) Yoo, B. C.; Ku, J. L.; Hong, S. H.; Shin, Y. K.; Park, S. Y.; Kim, H. K.; Park, J. G. Int. J. Cancer 2004, 108, 532-539. (33) Hondermarck, H.; Vercoutter-Edouart, A. S.; Re´villion, F.; Lemoine, J.; Belkoura, I. E.; Nurcombe, V.; Peyrat, J. P. Proteomics 2001, 1, 1216-1232. (34) D’Amours, D.; Desnoyers, D.; D’Silva, I.; Poirier, G. G. Biochem. J. 1999, 342, 249-268. (35) Kaufmann, S. H.; Desnoyers, S.; Ottaviano, Y.; Davidson, N. E.; Poirier, G. G. Cancer Res. 1993, 53, 3976-3985. (36) Wasierka-Gadek, J.; Bugajska-Schretter, A.; Lo¨w-Baselli, A.; GraslKraupp, B. Mol. Carcinogen 1999, 24, 263-275. (37) Chiarugi, A.; Moskowitz, W. A. Science 2002, 297, 200-201. (38) Yu, S. W.; Wang, H.; Poitras, M. F.; Coombs, C.; Bowers, W. J.; Federoff, H. J.; Poirier, G. G.; Dawson, T. M.; Dawson, V. L. Science 2002, 297, 259-263.
research articles
Role for Nucleolin and PARP-1 in Drug Resistance (39) Pettitt, A. R.; Sherrington, P. D.; Cawley, J. C. Cancer Res. 2000, 60, 4187-4193. (40) Zong, W. X.; Ditsworth, D.; Bauer, D. E.; Wang, Z. Q.; Thompson, C. B. Genes Dev. 2004, 18, 1272-1282. (41) Mi, Y.C.; Thomas, S. D.; Xu, X. H.; Casson, L. K.; Miller, D. M.; Bates, P. J. J. Biol. Chem. 2003, 278, 8572-8579. (42) Wurzer, G.; Herceg, Z.; and Wsierska-Gadek J. Cancer Res. 2000, 60, 4238-4244. (43) Drazen, D. L.; Bilu, D.; Edwards, N.; Nelson, R. J. Mol. Med. 2001, 7, 761-766. (44) Rycyzyn, M. A.; Clevenger, C. V. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6790-6795. (45) Ginisty, H.; Sicard, H.; Roger, B.; Bouvet, P. J. Cell. Sci. 1999, 112, 761-72. (46) Sengupta, T. K.; Bandyopadhyay, S.; Fernandes, D. J.; Spicer, E. K. J. Biol. Chem. 2004, 279, 10855-10863. (47) Yang, C.; Maiguel, D. A.; Carrier, F. Nucleic Acids Res. 2002, 30, 2251-2260. (48) Daniely, Y.; Dimitrova, D. D.; Borowiec, J. A. Mol. Cell. Biol. 2002, 22, 6014-6022. (49) Thyagarajan, B.; Lundberg, R.; Rafferty, M.; Campbell, C. Somat. Cell. Mol. Genet. 1998, 24, 263-272. (50) Bustin M. Mol. Cell. Biol. 1999, 19, 5237-5246. (51) Brezniceanu, M. L.; Volp, K.; Bosser, S.; Solbach, C.; Lichter, P.; Joos, S.; Zornig, M. FASEB J. 2003, 17, 1295-1297. (52) Kawahara, N.; Tanaka, T.; Yokomizo, A.; Nanri, H.; Ono, M.; Wada, M.; Kohno, K.; Takenaka, K.; K Sugimachi, K.; and Kuwano, M. Cancer Res. 1996, 56, 5330-5333. (53) Gabai, V. L.; Meriin, A. B.; Yaglom, J. A.; Volloch, V. Z.; Sherman, M. Y. FEBS Lett. 1998, 30, 1-4.
(54) Kiang, J. G.; Gist, I. D.; Tsokos, G. C. FASEB J. 1998, 12, 1571-9. (55) Shen, J.; Hughes, C.; Gessner, T.; Subjeck, J. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 3278-3282. (56) Chatterjee, S.; Cheng, M. F.; Berger, R. B.; Berger, S. J.; Berger, N. A. Cancer Res. 1995, 5, 868-873. (57) Haas, I. G. Experientia 1994, 50, 1012-1020. (58) Moreno-Flores, M. T.; Olazabal, U. E.; Kreutzberg, G. W. Exp. Neurol. 1997, 146, 10-6. (59) Sato, T.; Saito, H.; Swensen, J.; Olifant, A.; Wood, C.; Danner, D.; Sakamoto, T.; Takita, K.; Kasumi, F.; Miki, Y. Cancer Res. 1992, 52, 1643-1646. (60) Nijtmans, L. G.; de Jong, L.; Artal Sanz, M.; Coates, P. J.; Berden, J. A.; Back, J. W.; Muijsers, A. O.; van der Spek, H.; Grivell, L. A. EMBO. J. 2000, 19, 2444-2451. (61) Johnsson, A.; Zeelenberg, I.; Min, Y.; Hilinski. J.; Berry, C.; Howell, S. B.; Los, G. Brit. J. Cancer 2000, 83, 1047-1054. (62) Fusaro, G.; Wang, S.; Chellappan, S. Oncogene 2002, 21, 45394548. (63) McClung, J. K.; Jupe, E. R.; Liu, X. T.; Dell’Orco, R. T. Exp. Gerontol. 1995, 30, 99-124. (64) Fraser, M.; Leung, B.; Jahani-Asl, A.; Yan, X. J.; Thompson, W. E.; Tsang, B. K. Reprod. Biolo. Endocrin. 2003, 1, 66-78. (65) Planas-Silva, M. D.; Weinberg, R. A. Curr. Opin. Cell. Biol. 1997, 9, 768-772. (66) Sherr, C. J. Science 1996, 274, 1672-1677. (67) Grabsch, H.; Takeno, S.; Parsons, W. J.; Pomjanski, N.; Boecking, A.; Gabbert, H. E.; Mueller, W. J. Patholo. 2003, 200, 16-22.
PR0501158
Journal of Proteome Research • Vol. 4, No. 5, 2005 1591
