[CANCER RESEARCH 45, 743-750, February 1985]

Selective Protein Phosphorylation in Heterogeneous Subpopulations of HumanColon Carcinoma Cells1

Subhas Chakrabarty,2 Yih Jan, Charles A. Miller, and Michael G. Brattata

Bristol-Baylor Laboratory, Department of Pharmacology, Baylor College of Medicine, Houston, Texas 77030

ABSTRACT

Endogenous membrane and cytosolic and nuclear proteinphosphorylations were compared among three well-character

ized subpopulations of human colonie carcinoma cells that wereoriginally isolated from a single human primary colon tumor.These intratumoral subpopulations of cells were found to differsignificantly in their biological properties. Analysis of phospho-proteins by two-dimensional electrophoresis following 32Pphos-

phorylation of subcellular fractions in a cell-free system or labeling

intact cells in vivo revealed significant differences in the selectivephosphorylation of membrane, cytosol, and nuclear proteins.The two-dimensional membrane, cytosol, and nuclear phospho-

protein profiles distinguished the three subpopulations of coloniecarcinoma cells from each other.

Silver-staining proteins from the three subpopulations werealso compared. The two-dimensional, silver-stained electropho-

retic profiles of nuclear proteins were essentially the same for allthree subpopulations. The silver-stained electrophoretic profile

of membrane and cytosolic proteins revealed only minor differences in the expression of polypeptides. Nevertheless, thesechanges could also distinguish the three subpopulations. Theresults of this study suggest that minor differences in the expression of cytosolic and membrane proteins exist in intratumoralsubpopulations of colonie cells. However, a significantly greaterdegree of heterogeneity was found to be associated with post-

translational modification of proteins by phosphorylation and/ordephosphorylation. These modifications could play an importantrole in determining the expression of different biological properties among subpopulations of malignant cells.

INTRODUCTION

While advances have been made in the treatment of sometypes of cancers in recent years, the cure rates for coloniecarcinoma have remained largely unchanged for the past 2decades (11). One reason for this is the lack of effective chem-

otherapeutic agents against this disease. Heterogeneity of themalignant cells in a single tumor may contribute to the ineffectiveness of chemotherapeutic agents against colonie carcinoma(2, 3). New approaches are needed for the therapy of coloniecarcinoma, and a better understanding of the biological andbiochemical properties of the disease would be helpful in theidentification of potential therapeutic approaches (11,18).

Previously, we reported the identification, isolation, and establishment of cell lines from subpopulations of malignant cells of aprimary culture derived from a surgical specimen of coloniecarcinoma (2). The subpopulations, designated HCT 116, HCT

1Supported by Grant PDT-109 from the American Cancer Society and NIH

Grant CA34432.2To whom requests for reprints should be addressed.

Received January 23, 1984; accepted October 31, 1984.

116a, and HCT 116b, showed significant differences in severalbiological properties generally associated with markers of neo-

plastic transformation, such as tumorigenicity, karyology, morphology, growth in semisolid medium, growth rate, saturationdensity, and sensitivity to mitomycin C both In vitro and in athymicmice (2,17).

Differences in phenotypic expression of biological propertiesmay be related to changes in cellular processes that are involvedin the regulation of cellular functions. Protein kinases have beenknown to play a vital role in the regulation of cellular physiology(1, 12), and recently, Greig ef al. (13) have shown that heterogeneity of phosphorylation of cellular proteins is associated withB-16 melanoma clones of varying metastatic capabilities. In

addition, there is evidence suggesting that the biochemical actions of many regulatory molecules, such as hormones, and viraltransforming genes are also mediated through the action ofspecific protein kinases that are involved in the selective phosphorylation of cellular proteins (5-8,10, 20).

In this paper, we describe the KP phosphorylation of mem

brane, cytosolic, and nuclear proteins from the 3 subpopulations(HCT 116, HCT 116a, and HCT 116b) of human colonie carcinoma cells. Two-dimensional electrophoresis was used in theanalysis of phosphoproteins and silver-staining proteins from

each of the subcellular fractions. While only minor differenceswere found in the silver-staining electrophoretic patterns, signif

icant heterogeneity among the phosphoprotein electrophoreticprofiles of the 3 subpopulations was observed.

MATERIALS AND METHODS

Tissue Culture. Cells were maintained in 25-sq cm plastic flasks inMcCoy's Medium 5A supplemented with antibiotics and 10% fetal bovine

serum as described previously (2). Cultures were subcultured weekly ata 1:50 split with 0.25% trypsin:0.1% EDTA in tissue culture medium.The experiments described below were performed with cells betweenpassages 200 and 250. Tissue culture supplies were obtained fromGrand Island Biological Co. (Grand Island, NY).

Characteristics of HCT 116, HCT 116a, and HCT 116b were reportedpreviously except for a detailed analysis of karyology (2, 17). Briefly,HCT 116a was found to be more tumorigenic than HCT 116 with a 2-fold-higher growth rate in athymic mice. HCT 116b grew poorly in athymic

mice and had a growth rate of about 20% that of HCT 116a. Thetumorigenicities of the 3 subpopulations in athymic mice at the timewhen the phosphorylation experiments were performed are summarizedin Table 1. HCT 116a had a doubling rate which was about 50% of thatof the other 2 subpopulations in tissue culture. HCT 116 grew as anorderly monolayer of polygonal cells in tissue culture, while HCT 116bhad a more elongated morphology. HCT 116a grew as anchorage-indifferent, grape-like clusters. HCT 116a was the most sensitive line to

mitomycin C both in vitro and in xenografts, while HCT 116b was about6-fold more resistant to the drug in vitro, and xenografts did not respond

to mitomycin C. HCT 116 had an intermediate level of sensitivity to thedrug. HCT 116a produced large amounts of carcinoembryonic antigen

CANCER RESEARCH VOL. 45 FEBRUARY 1985

743

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

Table 1

Comparison of tumorigenicity in athymic mice at early and late passage ofmalignant cell subpopulations in tissue culture

Tumor volume was estimated by caliper measurements after s.c. injection ofa x b2

cells by the formula, —-—, where a is the maximum length of the tumor, and b

is the perpendicularto a (17).

Size of xenograft (cumm)SubpopulationsHCT116a

HCT116HCT116bCells

from passage251113±178°

605 ±59241 ±27Cells

from passage2501010

±100650 ±125217±13*

Mean ±S.D.

(about 2000 ng/24 hr in confluent cultures), while the other subpopula-

tions secreted negligible amounts of this glycoprotein.Karyotypes from G-banding were utilized in distinguishing the subpop-

ulations by marker chromosomes. Karyotype analysis showed that HCT116b had 47.XY karyotype with an additional chromosome 8 and additional material on the long arm of chromosome 13. HCT 116 had a 46.XYkaryotype with additional material on the short arm of chromosome 18,while HCT 116a had an apparently normal 46,XY karyotype at thebanding level achieved. Karyology was performed by standard methodsdescribed previously (3).

Preparation of Subcellular Fractions. Confluent cells (1.0 g) wereharvested from the tissue culture flasks by scraping and were washed 3times in cold 0.9% NaCI solution (saline). Washed cells were swollen for10 min at 4°in 10 volumes of 10 mw Tris:10 mM NaCI:8 mw MgCI2.

Swollen cells were pelleted by low-speed centrifugation and disrupted in10 ml of 0.5% Nonidet P-40 in 10 mw Tris:10 HIM NaCI:1.5 mw MgCI2

(disruption buffer) with a Dounce homogenizer (Fisher Scientific). Thisprocedure essentially broke up the cells without disrupting the nuclei.Nuclei were judged to be well dispersed and intact by phase-contrast

microscopy (29). Nuclei were removed by centrifugation at 1,000 x g for15 min, washed twice with cold 0.88 M sucrose containing 5 mM MgCI2,and used in the in vitro phosphorylation experiments described below.Polysemes and mitochondria were removed by centrifugation at 20,000x g for 15 min, and the cytosols were separated from the crudemembranes by centrifugation of the supernatants at 100,000 x g for 16hr. The supernatants obtained following the last centrifugation weredesignated as cytosol fractions which were used in the in vitro phosphorylation experiments. The pelleted crude membranes were resolubilizedby homogenization in disruption buffer with a tight-fitting, Teflon tip

homogenizer. These fractions were designated as membrane fractionswhich were used in the in vitro phosphorylation experiments. Proteinconcentrations were determined by the Bio-Rad dye binding assay (Bio-

Rad, Richmond, CA). Phosphorylation assays were performed immediately following the preparation of subcellular fractions described above.No significant changes in the in w'fro-phosphorylated phosphoprotein

profiles were observed upon storage of the samples in lyophilized format -70° for 1 week.

Phosphorylation Assays. Membrane and cytosol proteins (100 ¿/g)were phosphorylated in a kinase reaction mixture (200 n\) which consisted of 50 mM Tris-HCI (pH 7.2), 10 mw NaF, 10 mM MgCI2, 30 MMcyclic AMP, 5 fiM ethyleneglycol-bis(0-aminoethylether)-N,A/,A/',A/'-tetra-acetic acid, and 50 MCi of [-><-32P]ATP(Amersham; 200 Ci/mmol) (14).The reaction mixtures were incubated at 30°with continuous shaking

for 12 min, and the reactions were terminated by the addition of 200 p\of ice-cold 10% trichloroacetic acid; 5 »\of 1.0% bovine serum albumin

solution were added as carrier for precipitated proteins. Precipitatedproteins were pelleted in a bench-top microfuge, washed 3 times with

cold acetone, and resolubilized in lysis buffer containing 9.5 M urea, 2%Nonidet P-40, 2% Ampholine (pH 3.5 to 10), and 5% /3-mercaptoethanol,as described by O'Farrell (22).

Washed nuclei from 10 flasks (75 sq cm) of cells were similarilyphosphorylated in 0.5 ml of kinase reaction buffer containing 0.15 M NaCI

in the presence of 200 ¿<Ciof [T-^PJATP. At the end of the incubation

period, the nuclei were immediately put on ice, washed twice with coldsaline containing 2 mM phenylmethylsulfonyl fluoride, and homogenizedin lysis buffer. Insoluble chromatin and nucleic acids were removed bycentrifugation in a bench-top centrifuge, and DNase I and RNase (Calbi-

ochem, Los Angeles, CA) were added to the supernatant to a concentration of 50 Mg/ml, respectively.

In Vivo Phosphorylation. The medium from cells grown to 16 hrbefore confluency was replaced with phosphate-free medium (GrandIsland Biological Co.) supplemented with 10% fetal bovine serum:0.01%NaHCO3:0.001 % phenol red and containing 1 to 2 mCi of 32P¡(Centichem)per 75-sq cm flask of cells. The cells were incubated overnight for 16 hr,

and subcellular fractions were harvested as described above with washing and disruption buffers containing 10 mM 0-glycerophosphate, 0.5 mMethyleneglycol-bis(/3-aminoethylether)-A/,A/,A/',A/'-tetraacetic acid, 10 mM

NaF, 0.5 HIM phenylmethylsulfonyl fluoride, leupeptin (5 /¿g/ml),aprotinin(10 ^g/ml), 1.0 mM A/-ethylmaleimide, and 1.0 mM p-hydroxymercuriben-

zoate. This mixture of enzyme inhibitors was used to prevent proteolyticdegradation. Phosphoprotein analysis was performed immediately following the isolation of subcellular fractions. No significant changes in thephosphoprotein profiles were observed upon storage of the samples forup to 2 weeks in lyophilized form at -70°.

Analysis of Phosphorylated Proteins and Silver-stained Proteins.32P-phosphorylated proteins were identified by 2-dimensional gel electro-

phoresis. Two-dimensional gel electrophoresis was performed as described by O'Farrell (22) and Sidman (26) with some modifications.

Isoelectric focusing was performed in 200-^1 micropipets (Fisher Scientific) in 4% polyacrylamide gels with LKB Ampholines (2%) in the pHrange of 3.5 to 10. Forty ng of each subcellular fraction were loadedonto the focusing gels. Isoelectric focusing was performed to equilibriumat 3100 V hr; a stable pH gradient of 4.2 to 9.2 was obtained in the gel.Following isoelectric focusing, the gels were adapted in sample bufferfor 15 min before annealing onto the second-dimensional slab gels which

consisted of 3.5% stacking and 12% separating polyacrylamide gels.The sample buffer consisted of 6.25% 1 M Tris-HCI (pH 6.8), 5% 0-mercaptoethanol, 10% glycerol, 2.3% sodium dodecyl sulfate, and0.002% bromphenol blue in H2O. Visualization of proteins separated onthe 2-dimensional gels was accomplished by a sensitive silver-staining

procedure as described by Sammons ef al. (25). The pH gradients in thefocusing gels were estimated by slicing 0.5-cm segments of the gels,

and the pH was measured after the addition of 200 ¡Aof distilled H2O.Molecular weights were estimated by running molecular weight standards (Bio-Rad Laboratories) in the molecular weight range of 21,000 to200,000 in the second dimension of electrophoresis. ^P-labeled proteins

were visualized by exposing the acrylamide gels to Kodak X-Omat Rfilms with intensifying screen (Dupont) at -70°. The experiments de

scribed in this paper had been performed at least 3 times. Only thosereproducible differences are described in "Results."

RESULTS

Cytosol Proteins. The electrophoretic profilesof silver-stainedproteins obtained from the cytosol fractions of HCT 116b, HCT116, and HCT 116a are shown in Fig. 1, A to C, respectively.The profiles of all 3 subpopulations were quite similar withrespect to both major (heavily stained) and minor proteins.Profiles of HCT 116b and HCT 116 were essentially identical toeach other, except that HCT 116 had a prominent protein (labeledSpot A) at pi 6.9, M, 34,000, which was absent from HCT 116b.HCT 116a also had a prominent protein at pi 6.9, M, 34,000.HCT 116a showed several differences from both of the othersubpopulations. These included a marked reduction of spot 1 (pi8.2, M, 35,000) and the absence of Spot 2 (pi 8.0, M, 35,000).Spot M (pi 5.2, Mr 54,000) was present as a major protein in

CANCER RESEARCH VOL. 45 FEBRUARY 1985

744

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

HCT 116a but was not observed in profiles of HCT 116 or HCT116b.

Two-dimensional electrophoretic analysis of cytosolic phos-

phoproteins from these colonie carcinoma sublines, following anin vitro phosphorylation procedure, revealed significant differences in the selective 32Pphosphorylation of proteins labeled A

to H (Fig. 2, A to C). Broken circles indicate the areas in theelectrophoretic profiles of protein spots that were not phospho-rylated or in which the intensity of 32Plabeling had been greatly

reduced in comparison to equivalent protein spots from theelectrophoretogram of another subpopulation that was 32P la

beled. For example, Proteins A and B were highly phosphorylatedin the HCT 116b subpopulation (Fig. 2A), while only Protein Bwas highly phosphorylated in the HCT 116 subpopulation (Fig.2B), and only Protein A was highly phosphorylated in the HCT116a subpopulation (Fig. 2C). Four of 8 of the cytosolic proteinsdescribed were phosphorylated in HCT 116b cells, 7 of 8 of thecytosolic proteins were phosphorylated in HCT 116 cells, andonly 2 of 8 of the cytosolic proteins were phosphorylated in HCT116a cells.

Significant differences in the selective phosphorylation of cy-

tosol proteins were also observed among the 3 subpopulationsof cells following an in vivo labeling procedure using 32P|.Eleven

cytosol proteins were found to be selectively phosphorylated(Fig. 3, A to C). Proteins E and H corresponded (in terms ofmigration) to Proteins E and H in Fig. 2. Proteins A to D describedin Fig. 2 did not appear to be selectively phosphorylated, andother proteins (1 to 9) were found to be selectively phosphorylated by the in vivo labeling procedure. Four of 11 of the proteinsdescribed were phosphorylated in the HCT 116b cells, 9 of 11of the proteins were phosphorylated in the HCT 116 cells, and 6of 11 of the proteins were phosphorylated in the HCT 116a cells.

Comparison of Figs. 2 and 3 revealed both similarities anddifferences between the results of the in vitro 32Pphosphorylationand the in vivo 32P phosphorylation experiments. Some phos-

phoproteins identified in Fig. 2 appeared to be closely related tothose identified in Fig. 3, and new phosphoproteins were identified in Fig. 3 that were not seen in Fig. 2. Protein C (Fig. 2)appeared to be equivalent to Protein 9 (Fig. 3) that was highlyphosphorylated in the HCT 116 and HCT 116a cells but not inthe HCT 116b cells. Proteins E, H, G, and F in Fig. 2 corresponded to Proteins E, H, 6, and 7 in Fig. 3; however, theselective phosphorylation patterns of these proteins among the3 subpopulations of cells were different between Figs. 2 and 3.Additional cytosolic proteins (1 to 5 and 8) were phosphorylatedby the in vivo 32P phosphorylation procedure but not by the in

vitro procedure.Membrane Proteins. Fig. 4 shows the electrophoretic profiles

of silver-stained proteins from the 3 subpopulations of cells. Thearrangement of 5 silver-stained protein spots in the area of pis

6.9 to 7.8 and molecular weights of 50,000 to 60,000 was foundto be similar between HCT 116b and HCT 116 cells. However,a different arrangement of protein spots in this area was observed for HCT 116a cells. This area is marked by parallelograms,

Two membrane proteins (X and Z) were found to be selectivelyphosphorylated among the subpopulations of cells by the in vitrolabeling procedure (Fig. 5). Protein Z was phosphorylated in HCT116b cells and HCT 116 cells (Fig. 5, A and B), while only ProteinX was phosphorylated in HCT 116a cells (Fig. 5C). Protein X

was phosphorylated in both the HCT 116 and HCT 116a cellsbut not in HCT 116b cells. Eight membrane proteins (N, O, P, Q,R, S, T, and U) were found to be selectively phosphorylated bythe in vivo labeling procedure (Fig. 6, A to C). Five of theseproteins were phosphorylated in the HCT 116b cells, and 3 ofthese proteins were phosphorylated in the HCT 116 and HCT116a cells.

The phosphoprotein profile observed for the in w'fro-labeled

experiments was not well resolved in the regions of the gels ofpi 5.6 and molecular weight of 45,000 (Fig. 5). The phosphoproteins in these areas were clearly resolved in the gels followingthe in wVo-labeling procedure. Proteins P, Q, and 0 were found

to be highly phosphorylated in HCT 116b and HCT 116 cells incomparison to HCT 116a cells (Fig. 6). Protein Z in Fig. 5 probablycorresponded to Protein S in Fig. 6, and the phosphorylationpattern of this protein was found to be different among the 3 cellsubpopulations depending upon the procedures used in the 32P

labeling. The pattern of phosphorylation of Protein X observedin Fig. 5 was not observed in Fig. 6. As in the case for cytosolicproteins, additional proteins, such as R, N, T, and U, were foundto be selectively phosphorylated among the 3 subpopulations ofcells by the in vivo labeling procedure (Fig. 6).

Nuclear Proteins. No significant qualitative or quantitativedifferences in the electrophoretic profiles of silver-stained pro

teins from the 3 subpopulations of cells were found. Therefore,the silver-stained electrophoretic profiles of the nuclear proteins

are not shown. Four proteins (labeled 1 to 4) were found to beselectively phosphorylated by the in vitro phosphorylation procedure (Fig. 7, A to C), and 5 proteins (labeled 5 to 9) were foundto be selectively phosphorylated by the in vivo phosphorylationprocedure (Fig. 8, A to C). Protein 1 in Fig. 7 probably corresponded to Protein 6 in Fig. 8; however, the phosphorylationprofile of this protein among the subpopulations was observedto be different, depending on the method of 32Pphosphorylation

used in the experiments. The differences and or similarities ofthe phosphoprotein patterns observed by the in vitro and in vivolabeling procedures are further discussed below.

DISCUSSION

Previous work had shown that the 3 subpopulations of coloncarcinoma cells utilized in this study possessed significant differences in their tumorigenic properties, drug sensitivities, growthrates, morphology, karyology, and growth in semisolid medium(2,17). The biological properties of HCT 116 are intermediate tothose of the other 2 subpopulations with respect to growth rateof xenografts, mitomycin C sensitivity, and the degree of abnormality in its karyotype. In other respects, HCT 116 appears tobe more closely related to the properties of HCT 116b than tothose of HCT 116a. HCT 116 and HCT 116b have similar growthrates in tissue culture, saturation densities, and anchorage-

independent growth, and both subpopulations showed a reducedexpression of carcinoembryonic antigen relative to HCT 116a.

Though the biological properties of these subpopulations ofcells were initially characterized in 1981 (2), it should be pointedout that their tumorigenic properties, growth rates, cellular morphology, and growth in semisolid medium are routinely monitoredin our laboratory, and the biological properties have been maintained for these subpopulations of cells in these studies.

The 2-dimensional electrophoretic profiles of nuclear proteins

CANCER RESEARCH VOL. 45 FEBRUARY 1985

745

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

stained with silver did not reveal any qualitative or consistentquantitative changes among these subpopulations. The electro-phoretic profiles of silver-stained cytosolic and membrane proteins also revealed extensive homology among the subpopula-

tions. However, a limited number of consistent differences in thesilver-stained electrophoretic profiles have been observed. Theseproteins did not correspond to any of the phosphoproteinsdescribed. The profiles of phosphoproteins appeared to be theresult of selective phosphorylation and/or dephosphorylation ofcellular proteins, which were expressed in all 3 subpopulations.

Selective phosphorylation of cellular proteins was observed inthe 3 subpopulations of colon carcinoma cells regardless ofwhether the proteins were phosphorylated in a cell-free system

or phosphorylated in vivo, in intact cells. However, the differences among the phosphoprotein patterns of subcellular fractions from the 3 subpopulations labeled in vitro did not matchthe differences obtained from the corresponding subcellular fractions labeled in vivo, although in both cases, selective phospho-

rylations among the 3 subpopulations of cells were observed.There are several reasons for the differences between the inw'fro-labeled and the in wVo-labeled phosphoprotein patterns,

(a) The presence of ethyleneglycol-bis(^-aminoethylether)-A/,A/,A/',A/'-tetraacetic acid in these experiments may have re

duced the activities of protein kinases, the activities of whichrequire the presence of Ça2"",Ca2+ plus calmodulin, or Caz+ plus

phospholipid. (ó)The compartmentalization of substrates andenzymes was disrupted by the in vitro procedure and may haveresulted in reduced substrate specificity for the enzymes. Thisis exemplified by the membrane phosphoprotein patterns inwhich only 2 membrane proteins were selectively phosphorylatedby the in vitro procedure in contrast to 8 membrane proteins thatwere selectively phosphorylated by the in vivo procedure. Thephosphorylation of proteins observed by the in vivo procedurewas probably more reflective of the protein kinase and phospha-

tase activities under physiological conditions. It should be pointedout that homologous silver-staining patterns were observedwhen the gels were stained with silver, following autoradiogra-

phy, regardless of whether the fractions had been labeled in vitroor the fractions had been isolated after in vivo labeling. Takentogether, the selective phosphorylation of cellular proteins observed by both the in vitro and in vivo labeling techniquesindicated a qualitative and/or quantitative differences in the protein kinases and/or phosphatases in the 3 subpopulations ofcolonie carcinoma cells.

Extensive homology of membrane proteins, as assessed by2-dimensional electrophoresis, has been reported between 2

sublines of rat prostatic tumor which represented opposite extremes in growth rate and histology (15). The identification oftumor-associated cellular proteins by 2-dimensional electropho

resis that were not observed in normal tissues has also beenreported by a number of investigators (24, 27-29). Differential

patterns of expression of cytokeratins in normal, tumor, andcultured cells have been documented as well (19). At present,the significance of the differences identified in the expression ofmembrane and cytosolic silver-stained proteins among the 3

subpopulations of cells is not known. Nevertheless, the difference in the arrangement of 5 membrane proteins in the area ofpis 6.9 to 7.8 and molecular weights of 50,000 to 60,000 andthe expression of the cytosolic Protein M (pi 5.2, M, 54,000)distinguished the 3 subpopulations of cells. The silver-stained

nuclear protein profiles, on the other hand, were found to behomologous among the 3 subpopulations of cells. The numberof silver-stained subcellular proteins greatly exceeded that of the32P-labeled proteins, yet the major molecular differences among

the subpopulations of cells were found to be in their phosphoprotein electrophoretic patterns.

Selective phosphorylation of cellular proteins may play animportant role in the generation of tumor cell diversity, sincephosphorylation appears to be an important means of growthcontrol and differentiation (9,12, 23). Tyrosine residues did notappear to be phosphorylated in those proteins that were selectively phosphorylated among the subpopulations of cells. Treatment of the gels with 1 M KOH for 2 hr at 55° completelyremoved the [32P]phosphates from these proteins (6). Selective

phosphorylation of cellular proteins by endogenous protein kinases has also been observed between tumor and normal cells(21). Greig ef a/. (13) have observed greater heterogeneity ofphosphorylation than of polypeptide expression among clonesof B16 melanoma with different metastatic capabilities. It shouldbe pointed out that the experiments described here were performed with established cell lines with well-defined biological

properties. Assessment of differences in phosphoprotein patterns among heterogeneous populations of malignant cells fromfresh biopsies would be highly desirable. However, this willrequire the development of the technological capability of determining kinases, phosphatases, and appropriate substrate proteins either indirectly by immunological methods in populationsof unresolved cells or after the resolution of specific subpopula-tions of cells from the biopsies or their primary cultures. Eitherapproach presents formidable technical difficulties at the presenttime. It is not known whether the differences in phosphoproteinexpression we observed are due to the lack of protein kinaseactivities or an increase in cellular phosphatase activities. Different levels of protein kinases have been reported between tu-

morous and normal tissues (4) and between differentiated andundifferentiated cells (16). The isolation and characterization ofthe enzymes that are involved in the generation of selectivephosphorylation among the subpopulations of colonie carcinomacells will be the subject of future studies.

REFERENCES

1. Antoniades, H. N., and Owen, A. J. Growth factors and regulation of cellgrowth. Annu. Rev. Med., 33: 445-463, 1982.

2. Brattain, M. G., Fine, W. D., Khaled, F. M, Thompson, J., and Brattain, D. E.Heterogeneity of malignant cells from a human colonie carcinoma. CancerRes., 41: 1751-1756,1981.

3. Brattain, M. G., Marks, M. E., McCombs, J., Finely, W., and Brattain, D. E.Characterization of human colon carcinoma cell lines isolated from a singleprimary tumor. Br. J. Cancer, 47: 373-381,1983.

4. Chung, L. W. K., and Breitweiser, K. Comparison of soluble protein kinasesfrom normal rat prostates and prostate Dunning tumors. Cancer Res., 83:3297-3304,1983.

5. Cohen, P. The role of protein phosphorylation in neural and hormonal controlof cellular activity. Nature (Lond.), 296: 613-619,1982.

6. Cooper, J. A., and Hunter, T. Changes in protein phosphorylation in Roussarcoma virus-transformed chicken embryo cells. Mol. Cell. Bid., 1:165-178,1981.

7. Cooper, J. A., and Hunter, T. Four different classes of retroviruses inducephosphorylation of tyrosine present in similar cellular proteins. Mol. Cell. Bid.,7:393-407,1981.

8. Ek, B.. and Heldin, C-H. Characterization of a tyrosine-specific kinase activityin human fibroWast membranes stimulated by platelet-derived growth factor.J. Biol. Chem., 257: 10486-10492,1982.

9. Erikson, E., Shealy, D. J., and Erikson, R. L. Evidence that viral transforminggene products and epidermal growth factor stimulate phosphorylation of thesame cellular protein with similar specificity. J. Biol. Chem., 256: 11381-

CANCER RESEARCH VOL. 45 FEBRUARY 1985

746

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

11384,1981.10. Erikson, R. L., Purchio, A. F., Erikson, E., Collett, M. S., and Brugge, J. S.

Molecular events in cells transformed by Rous sarcoma virus. J. Cell Biol., 87:319-325,1980.

11. Freeman, H. J., and Kim, Y. S. Membrane biochemistry and physiology ofhuman colonie cancer. Cancer Bull., 30: 237-242,1978.

12. Greengard, P. Phosphorylated proteins as physiological effectors. Science(Wash. DC), 199: 146-152,1978.

13. Greig, R. G., Caltabiano, L., Reid, R., Jr., Feild, J., and Poste, G. Heterogeneityof protein phosphorylation in metastatic variants of B16 melanoma. CancerRes., 43: 6057-6065,1983.

14. Kobayashi, N., Tanaka, A., and Kaji, A. In vitro phosphorylation of the 36kdprotein in extracts from Rous sarcoma virus-transformed chicken fibroblasts.J. Biol. Chem., 256: 3053-3058,1981.

15. Kozlovskis, P. L., Claflin, A. J., Fletcher, M. A., McKinney, E. C., and Rubin,R. W. Two-dimensional gel electrophoresis of membrane proteins from theR3327 prostate adenocarcinoma. Cancer Res., 42: 2748-2756, 1982.

16. Kraft, A. S., and Anderson, W. B. Characterization of cytosolic calcium-

activated phospholipid dependent protein kinase activity in embyronal carcinoma cells. J. Biol. Chem., 258: 9178-9183,1983.

17. Long, B. H., Willson, J. K. V., Brattain, D. E., Musial, S. T., and Brattain, M.G. The effects of mitomycin C on human colon carcinoma cells. J. Nati. CancerInst, 73: 787-792, 1984.

18. Marks, M. E., Danbury, B. H., Miller, C. A., and Brattain, M. G. Cell surfaceproteins and glycoproteins from biologically different human colon carcinomacell lines J. Nati. Cancer Inst., 71: 663-671, 1983.

19. Moll, R., Franke, W. W., Schiller, D. L., Geiger, B., and Krepier, R. The catalogof human cytokeratins: patterns of expression in normal epithelia, tumors, and

cultured cells. Cell, 3): 11-24,1982.20. Nakamura, K. D., Martinez, R., and Weber, M. J. Tyrosine phosphorylation of

specific proteins after mitogen stimulation of chicken embryo fibroblasts. Mol.Cell. Biol., 3:380-390,1983.

21. Nakaya, K., Shimizu, T., Segawa, Y., Nakajo, S., and Nakamura, Y. Characterization of protein phosphorylation of the cytosol of AH-66 hepatoma asertescells. Cancer Res., 43: 3778-3782,1983.

22. O'Farrell, P. H. High resolution two-dimensional electrophoresis of proteins. J.

Biol. Chem., 250: 4007-4021, 1975.23. Reynolds, F. H., Jr., Todaro, G. J., Fryling, C., and Stephenson, J. R. Human

transforming growth factors induce tyrosine phosphorylation of EGF receptors.Nature (Lond.) 292: 259-262, 1981.

24. Ruoslahti, E., Oh, E., and Jalanko, H. Differences in the nuclear proteins ofnormal and malignant liver cells. Oncodev. Biol. Med., 1:17-26,1980.

25. Sammons, D. W., Adams, L. D., and Nishizawa, E. E. Ultrasensitive silver-based color staining of polypeptides in polyacrylamide gels. Electrophoresis,2:135-141,1981.

26. Sidman, C. Two-dimensional gel electrophoresis. Immunol. Methods, 2: 57-74,1981.

27. Wu, B. C., Spohn, W. H., and Busch, H. Two-dimensional gel electrophoresisof nuclear phosphoproteins of Novikoff hepatoma and regenerating liver.Physiol. Chem. Phys., 72:11-20, 1980.

28. Wu, B. C., Spohn, W. H., and Busch, H. Comparison of nuclear proteins ofseveral human tumors and normal cells by two-dimensional gel electrophoresis.Cancer Res., 41: 336-342,1981.

29. Yeoman, L. C., Jordan, J. J., Busch, R. K., Taylor, C. W., Savage, H. E., andBusch, H. A fetal protein in chromatin of Novikoff hepatoma and Walker 256carcinosarcoma tumors that is absent from normal and regenerating rat liver.Proc. Nati. Acad. Sci. USA, 73: 3258-3262, 1976.

CANCER RESEARCH VOL. 45 FEBRUARY 1985

747

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

OB E"u? 8 If

fifi*s2li<

ysii

'

CANCER RESEARCH VOL. 45 FEBRUARY 1985

748

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

-45

8.6 7.« 6.9 5.6l III

4A

MW XI IO'3

IEFpi 7.8 6.9 5.6

MW x

€fci -31

5A

-66

-45

Fig. 4. Two-dimensional electrophoretic analysis of membrane proteins that were silver stained from HCT 116b (A), HCT 116 (B), and HCT 116a (C) cells.Theelectrophoretic procedures were identical to those described in Fig. 1. Parallelograms mark the areas in which a difference in the arrangement of 5 protein spots wasdetected. Abbreviations as in Fig. 1.

Fig. 5. Two-dimensional electrophoretic analysis of membrane ^P-labeled proteins (in vitro labeled) from HCT 116b (A), HCT 116 (B), and HCT 116a (C) cells. The

procedures of electrophoresis and phosphoprotein detection were identical to those described in Figs. 1 and 2. X and Z represent equivalent proteins from the 3subpopulations of cells in which selective phosphorylation was detected. Abbreviations as in Fig. 1.

CANCER RESEARCH VOL. 45 FEBRUARY 1985

749

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR ALTERATIONS AND TUMOR CELL HETEROGENEITY

IEF pi 7.8 6.9 5.67.8 6.9 5.6

col » '

Ocoj

IEF pi 69 56 4.8 6.9 5.6 48 69 5.6 4.8««——te I- m\nr x

-66

IEF

-45

8A 8B

MWx10'3

-66

-45

8CFig.6. Two-dimensional electrophoretic analysis of membrane ^P-labeled proteins (in vivo labeled) from HCT 116b (/>),HOT 116 (B), and HCT 116a (C) cells. The

procedures of electrophoresis and phosphoprotein detection were identical to those described in Figs. 1 and 2. N to U represent equivalent proteins from the 3subpopulations of cells in which selective phosphorylationwas detected. Abbreviations as in Fig. 1.

Fig.7. Two-dimensional electrophoretic analysis of MP-labelednuclear proteins (in vitro labeled) from HCT 116b (A), HCT 116 (B), and HCT 116a (C) cells. Theprocedures of electrophoresis and phosphoprotein detection were identical to those described in Figs. 1 and 2. Spots 1 to 4 represent equivalent proteins from the 3subpopulations of cells in which selective phosphorylation was detected. Abbreviations as in Fig. 1.

Fig.8. Two-dimensional electrophoretic analysis of ^P-labeled nuclear proteins (in vivo labeled) from HCT 116b (A), HCT 116 (B), and HCT 116a (C) cells. Theprocedures of electrophoresis and phosphoprotein detection were identical to those described in Figs. 1 and 2. Spois 5 to 9 represent equivalent proteins from the 3subpopulations of cells in which selective phosphorylation was detected. Abbreviations as in Fig. 1.

CANCER RESEARCH VOL. 45 FEBRUARY 1985

750

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1985;45:743-750. Cancer Res   Subhas Chakrabarty, Yih Jan, Charles A. Miller, et al.   Subpopulations of Human Colon Carcinoma CellsSelective Protein Phosphorylation in Heterogeneous

  Updated version

  http://cancerres.aacrjournals.org/content/45/2/743

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/45/2/743To request permission to re-use all or part of this article, use this link

on May 2, 2021. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from