Expression of a mutant regulatory subunit of CAMP-dependentprotein kinase in the ~ a c o - 2human colonic carcinoma cell line

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

VERA MIHAJLOVIC Banting and Best Department of Medical Research, University of Toronto, Toronto, Ont., Canada M5G lL6

Banting and Best Department of Medical Research and Department of Pharmacology, University of Toronto, Toronto, Ont., Canada M5G lL6 WOJTEKAUERBACH Research Institute, Hospital for Sick Children, Toronto, Ont., Canada M5G 1x8 MANUELBUCHWALD Department of Molecular and Medical Genetics, University of Toronto, Toronto, Ont., Canada and Research Institute, Hospital for Sick Children, Toronto, Ont., Canada M5G 1x8 AND

Banting and Best Department of Medical Research and Department of Pharmacology, University of Toronto, Toronto, Ont., Canada MSG 1L6 Received February 12, 1992 MIHAJLOVIC, V., KROLCZYK, A. J., AUERBACH, W., BUCHWALD, M., and SCHIMMER, B. P. 1992. Expression of a mutant regulatory subunit of CAMP-dependent protein kinase in the Caco-2 human colonic carcinoma cell line. Biochem. Cell Biol. 70: 1039-1046. Caco-2 human colonic carcinoma cells were transfected with an expression vector encoding a mutant form of RI (regulatory subunit of the type 1 CAMP-dependent protein kinase), driven by the metallothionein 1 promoter. A stable transformant was isolated that expressed the mutant RI gene in a zn2+-inducible manner. The consequences of the RI mutation on CAMP-dependentprotein kinase activity, cell division, and regulation of chloride efflux were examined. When grown in the absence of ZnSO,, protein kinase activity in the transformant was stimulated 2.5-fold by cAMP and approached the levels of CAMP-dependent protein kinase activty seen in parental Caco-2 cells; when treated with ZnSO,, CAMP-dependentprotein kinase activity in the transformant was inhibited by 60%. In the absence of ZnSO, the transformant grew with the same doubling time and to the same saturation density as the untransformed parent. In the presence of ZnSO, the transformant exhibited a CAMP-reversible inhibition of cell division, indicating that a functional CAMP-dependentprotein kinase was required for the growth of these cells in culture. Induction of the mutant RI gene also abolished forskolin-stimulated chloride efflux from these cells, suggesting obligatory roles for cAMP and CAMP-dependentprotein kinase in forskolin's actions on chloride channel activity. We anticipate that this transformant will be useful for further studies on the roles of cAMP and CAMP-dependentprotein kinase in the regulation of intestinal epithelial cells, including regulation of cell proliferation and differentiation, and regulation of chloride channel activity by neurohormones and neurotransmitters. Key words: chloride efflux, cell growth, gene transfer, forskolin. V., KROLCZYK, A. J., AUERBACH, W., BUCHWALD, M., et SCHIMMER, B. P. 1992. Expression of a MIHAJLOVIC, mutant regulatory subunit of CAMP-dependent protein kinase in the Caco-2 human colonic carcinoma cell line. Biochem. Cell Biol. 70 : 1039-1046. Des cellules Caco-2 de carcinome du colon ont CtC transfecttes avec un vecteur d'expression codant une forme mutante de RI (sous-unit6 regulatrice de la prottine kinase AMPc-dtpendante, type 1) et contr6lt par le promoteur de la metallothiontine 1. Un clone stable de cellules transformtes ou I'expression du gene mutant de RI est induite par le z n 2 + a t t t isolt. Les consCquences de la mutation de RI sur l'activitt de la prottine kinase AMPc-dtpendante, sur la division cellulaire et sur la rtgulation de l'efflux du chlore ont ett CtudiCes. Lorsque les cellules croissent en absence de ZnSO,, l'activitt de la prottine kinase dans les cellules transforrntes est stimulte 2,s fois par 1'AMPc et atteint presque le niveau d'activitt de la prottine kinase AMPc-dtpendante observe dans les cellules Caco-2 parentales; lorsque les cellules transformies sont en prtsence de ZnSO,, l'activitt de la protkine kinase AMPc-dtpendante est diminute de 60%. En absence de ZnSO,, les cellules transformtes en culture ont le meme temps de doublement et elles atteignent la m&medensite de saturation que les cellules parentales non transformtes. En prtsence de ZnSO,, la division des cellules transformkes est inhibte et cette inhibition est contrte par I'AMPc, ce qui indique qu'une protkine kinase AMPc-dtpendante fonctionnelle est requise pour la croissance de ces cellules en culture. L'induction du gtne mutant de RI abolit tgalement la stimulation par la forskoline de la sortie du chlore hors des cellules, ce qui suggtre que 1'AMPc

ABBREVIATIONS: RI, regulatory subunit of type 1 CAMP-dependent protein kinase; CFTR, cystic fibrosis membrane transductance regulator; CF, cystic fibrosis; a-MEM, alpha minimal essential medium; MES, a-(N-morpho1ino)ethanesulfonicacid; bp, base pair(s); kb, kilobase@); TPA, 12-decanoylphorbol-13-acetate;MDCK, Madin-Darby canine kidney. ' ~ u t h o rto whom all correspondence should be sent at the following address: Banting and Best Department of Medical Research, University of Toronto, 112 College Street, Toronto, Ont., Canada MSG 1L6. Printed in Canada / lmprime au Canada

1040

BIOCHEM. CELL BIOL. VOL. 70,

1992

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

et la protkine kinase AMPc-dkpendante jouent des r8les essentiels dans le mkcanisme d'action de la forskoline sur l'activitk des canaux du chlore. Nous anticipons que ces cellules transformkes seront utiles pour d'autres etudes sur les r8les de 1'AMPc et de la protkine kinase AMPc-dkpendante dans la rkgulation des cellules kpithkliales de l'intestin, incluant la regulation de leur prolifkration et de leur diffkrenciation, et dans la rkgulation de l'activitk des canaux du chlore par les neurohormones et les neurotransmetteurs. Mots clds : efflux du chlore, croissance cellulaire, transfert du gtne, forskoline. [Traduit par la redaction]

Introduction The fluid and electrolyte compositions of the intestine, exocrine pancreas, and sweat gland are determined in large measure by neurohormone- and neurotransmitter-regulated chloride secretions from apical chloride channels (Frizzell and Halm 1990). The physiological importance of these channels is clearly illustrated in patients with cystic fibrosis. These patients have mutations in the gene encoding CFTR (Riordan et al. 1989), a membrane protein that is responsible for regulated chloride conductance in secretory epithelial cells (Anderson et al. 1991b; Drumm et al. 1990). As a consequence of these mutations, affected patients have intestinal obstructions, repeated airway infections, pancreatic insufficiency, abnormal mucus secretions, and abnormal sweat electrolytes resulting from tissue dehydration. Second messengers implicated in the regulation of chloride flux in intestinal epithelia include CAMP, inositol trisphosphate, diacylglycerol, c a 2 + (Dawson 1991), and possibly cGMP (O'Loughlin et al. 1991). In addition, second messenger independent processes involving activated guanyl nucleotide binding regulatory proteins have been implicated in epithelial chloride channel activation (Tilly et al. 1991), and effects of cyclic nucleotides on ion channel activation independent of protein phosphorylation also may be possible (Gold and Nakamura 1987). Although these various second messenger systems can regulate chloride flux, it is not clear to what extent these different second messengers act via independent signalling pathways. For intestinal epithelial cells, Anderson and Welsh (1991) have suggested that distinct chloride channels may be involved in c a 2 + - and CAMPregulated secretion, implying distinct signalling pathways. On the other hand, the CFTR chloride channel contains phosphorylation domains for CAMP-dependent protein kinase and protein kinase C that contribute to regulated chloride flux (Rich et al. 1991; Riordan et al. 1989; Tabcharani et al. 1991), implying that distinct regulatory pathways converge at a common target. Consistent with the conclusion that the same channel can be regulated by multiple second messenger pathways are the observations that impairment of CAMP-regulated chloride channels in enterocytes from C F patients is accompanied by defects in chloride channel regulation by cGMP and c a 2 + (de Jonge et al. 1989). Other studies suggest even more complex interactions, involving cross-talk among CAMP-, cGMP-, c a 2 + - , and protein-kinase-C-dependent signalling systems (Harper 1988). One approach that has been used effectively t o sort out the relative importance of different second messenger systems in signal transduction involves the use of mutations that disrupt second messenger pathways at specific sites. For example, mutations in RI that behave as dominant inhibitors of CAMP-dependent protein kinase activity have been extremely useful in evaluating the relative importance of cAMP and CAMP-dependent protein kinase in the signal transduction cascade (e.g., Schimmer 1989). The construc-

tion of expression vectors encoding these mutant RI genes has permitted an extension of this genetic approach to a variety of hormone-responsive systems and thus has broadened its general utility (Clegg et al. 1987). In this report, we describe the transformation of the human colonic carcinoma cell line Caco-2 (Rousset 1986) with a mutant RI gene and evaluate the usefulness of this transformant in studies of CAMP-dependent regulation of cell proliferation and chloride efflux.

Materials and methods Cell culture and gene transfer

Caco-2 cells (Rousset 1986) were seeded at a density of lo6 cells/75-cm2 plastic tissue culture flask. The cells were grown in monolayer at 36.S°C in a-MEM supplemented with 10% heatinactivated fetal bovine serum, 200 U penicillin G sodium/mL and 0.27 mg streptomycin sulfate/mL in a humidified atmosphere of 95% air - 5% COz. The culture medium was changed every 3rd or 4th day and cells were subcultured every 10th day using a solution of 0.1% trypsin and 0.5 mM EDTA in phosphate-buffered saline. Cells were transfected by lipofection with pMT-REV(AB)neo, an expression plasmid encoding a dominant, inhibitory mutant form of murine RI regulated by the mouse metallothionein 1 promoter together with a neomycin-resistance gene (Rogers et al. 1990). For transfection, cells were grown in 10-cm plastic tissue culture dishes until they reached 80% confluence. Cells were washed twice with serum-free a-MEM, incubated for 1 h at 37°C in a-MEM (3 mL) plus Lipofectin Reagent (30 pL), and then incubated for an additional 6 h with 25 pg of supercoiled pMT-REV(AB)neo DNA. At the end of the incubation, the DNA was removed, and cells were rinsed in a-MEM plus serum and incubated in fresh medium for 24 h to permit expression of the transfected gene. Cells were subcultured at a 1:3 dilution and grown in culture medium containing G418 (400 pg/mL). G418-resistant colonies were isolated and passaged in medium containing G418 at 200 pg/mL. CAMP-dependentprotein kinase activity

CAMP-dependent protein kinase activity was assayed in 100 000 x g supernatant fractions prepared from cell homogenates ~ [ y - 3 2 to ~ histone ] ~ ~ F2B ~ by measuring the transfer of 3 2 from in the presence of varying concentrations of cAMP essentially as described (Pon et al. 1990). Enzyme activity was assayed using 35 pg of cell extract in 170 pL of a reaction mixture containing 100 pg histone F2B, 150 pg bovine serum albumin, 10 mM MgSO, 10 mM 1,4-dithiothreitol, 0.5 mM 3-isobutyl-l] ~ ~ ~cpm/ methylxanthine, 5 mM NaF, 20 pM [ y - 3 2 ~(100-200 pmol), and 20 mM MES (pH 6.5). Phosphorylated substrate was collected on squares of phosphocellulose paper and washed free of radioactive nucleotides with trichloroacetic acid (Sahal and Fujita-Yamaguchi 1987). RNA-blot analysis

Total cellular RNA was isolated from Caco-2 cells by extraction with guanidine thiocyanate and centrifugation through CsCl (Chirgwin et al. 1979).RNA samples were electrophoresedthrough formaldehyde-agarose gels, blotted onto nylon membranes, and probed with a 770-bp BamHI-EcoRI cDNA fragment from the RI gene in pRevlO, an expression vector encoding a wild-type mouse RI (Clegg et al. 1987).

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

MIHAJLOVIC ET AL.

FIG. 1. CAMP-dependent protein kinase activtiy in Caco-2 and Rev5 cells. CAMP-dependentprotein kinase activity was measured as a function of varying cAMP concentrations in supernatant fractions from Caco-2 (0) and the mutant RI transformant Rev5 (W). Enzyme activity was assayed in cells grown under standard conditions (a) or in cells incubated in growth medium containing 175 pM ZnSO, for 40 h prior to assay (b). Protein kinase activity is expressed as pmol 3 2 transferred.min-'amg ~ protein-'.

Measurement of cell proliferation Cells were plated at a density of lo4 cells/l-cm well in multiwell tissue culture plates and grown for 1-8 days in growth medium. The Rev5 transformant was cultured in the presence of G418 to maintain the neomycin-resistance gene. Proliferation was assayed by measuring increases in DNA content using a fluorometric assay (Kissane and Robins 1958) with salmon sperm DNA as standard. Measurement of chloride efflux Chloride efflux was evaluated using an isotopic iodide efflux assay described by Venglarik et al. (1990). Cells were plated at a density of lo6 cells/35-mm plastic tissue culture dish and cultured for 2 or 3 days before assay. Cells were loaded with radioactive iodide by incubation for 1 h in Hepes-Ringer solution (140 mM NaCI, 3.3 mM KH,PO,, 0.83 mM K,HP04, 1 mM CaSO,, 1 mM MgSO,, 10 mM glucose, 10 mM Hepes, pH 7.4) containing 2.5 pCi/mL of [ 1 2 5 ~ ]and ~ a then ~ washed four times with 2 mL of isotope-free Hepes-Ringer solution to remove unincorporated isotope. The rate of iodide efflux was measured by incubating the labeled cells with 1-mL aliquots of the Hepes-Ringer solution for 60-s intervals over a period of 14 min. Agonists of chloride efflux were added to the incubation solutions after the 3rd min of incubation. The rate of efflux (r) was determined from the percentage of radioactive counts released at each time interval (Venglarik et al. 1990). Rea ents 1% [ I]NaI (16.3 mCi/pg iodine; 1 Ci = 37 GBq) was obtained from Amersham Canada Ltd., Oakville, Ont., and [ y - 3 2 ~ (3000 Ci/mmol) was purchased from DuPont-NEN Canada, Markham, Ont. Forskolin, 8BrcAMP, and salmon sperm DNA were from Sigma Chemical Co., St. Louis, Mo.; histone F2B (a slightly lysine-rich histone fraction prepared from calf thymus) was from US Biochemical Corp., Cleveland, Ohio; tissue culture medium, sera, G418 and Lipofectin Reagent were from GibcoBRL, Canada, Burlington, Ont. ~ ~ t r nylon a n membranes ~ ~ (Schleicher and Schuell, Int., Keene, N.H.) were purchased from Xymotech Biosystems, Mount Royal, Que.

Results CAMP-dependent protein kinase activity in Caco-2 cells transfected with a mutant R I gene The expression vector pMT-REV(AB)neo contains a cDNA encoding a dominant inhibitory form of mouse RI that renders CAMP-dependent protein kinases resistant to activation by CAMP. The plasmid also contains a neomycinresistance gene that facilitates the recovery of transformants by selective growth in the neomycin analog G418 (Rogers et al. 1990). Transfection of Caco-2 cells with the mutant RI expression vector pMT-REV(AB)neo gave rise to eight G418-resistant colonies that were assayed for CAMPdependent protein kinase activity. Dose-response relationships for CAMP-dependent protein kinase activity were determined by the curve-fitting procedure of De Lean et al. (1978). Extracts from one clone, designated Rev5, exhibited a nearly wild-type CAMP-dependent protein kinase activity under normal culture conditions., Protein kinase activity in extracts from parental Caco-2 cells, determined from six separate experiments, was stimulated by cAMP 2.7-fold to a maximum of 300 10 U of activity (1 U equals ~ from [ y - 3 2 ~ to ]histone ~ ~ ~per 1 picomole 3 2 transferred minute per milligram protein). The ED5ovalue (the concentration of agonist producing half-maximal activation) of the enzyme for cAMP was 0.06 + 0.01 pM. A single experiment is illustrated in Fig. 1. Enzyme activity in the Rev5 ] transformant ~ ~ ~ was stimulated by maximally effective concentrations of cAMP approximately 2.5-fold over the basal level, reaching a V,, that was 90% of the value obtained with the Caco-2 parent (Fig. la). The ED,, for CAMP, 0.05 + 0.01 pM, was similar to that found in parental Caco-2 cells. Treatment of parental Caco-2 cells with ZnS04 inhibited the wild-type enzyme approximately 25% (Fig. lb). In contrast, when Rev5 was incubated with ZnS04 to induce the expression of mutant RI from the

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

1042

BIOCHEM. CELL BIOL. VOL. 70, 1992

F I G . 2. RI transcripts in Caco-2 and Rev5 cells. Total RNA was isolated from Caco-2 and Rev5 cells grown in the presence or absence of 175 pM ZnSO, for 40 h. RI transcripts were analyzed by Northern blot hybridization using a nick-translated RI cDNA probe. The endogenous 3.6-kb RI transcripts and the 1.7-kb RI transcript resulting from transfection are indicated by arrows.

mouse metallothionein 1 promoter, CAMP-stimulated protein kinase activity decreased by 60% to a maximum of 120 5 U of activity. The ED50 of the zn2+-induced enzyme for CAMP was only marginally affected (Fig. lb). CAMP-dependent protein kinase activities recovered in extracts from the seven other neomycin-resistant clones ranged from 80 to 100% of the values obtained with the Caco-2 parent and were unaffected by ZnS04 (data not shown). Inasmuch as the levels of CAMP-dependent protein kinase activity in these transformants approached the levels seen in the Caco-2 parent and were insensitive to ZnS04, we conclude that these transformants expressed the neomycin-resistance gene, but did not express the gene encoding the mutant form of RI.

*

RNA blot hybridization analysis of RZ transcripts As shown in Fig. 2, the zn2+-induced inhibition of CAMP-dependent protein kinase activity was accompanied by a zn2+-dependent increase in RI transcripts from the pMT-REV(AB)neo vector. Untransfected Caco-2 cells, grown in the presence or absence of ZnS04 contained a major RI transcript of approximately 3.6 kb. In the Rev5 transformant, two hybridizing bands were seen with the RI probe: a band at 3.6 kb corresponding to the endogenous form of RI and a new band at 1.7 kb corresponding to the transcript size predicted from the RI sequence encoded in the expression vector (Clegg et al. 1988). Treatment of the Rev5 transformant with ZnS04 for 40 h caused a selective, 14-fold increase in the mutant RI transcript at 1.7 kb, as determined by densitometric analysis. The uniform levels of the endogenous RI transcript in each track ensured that equal amounts of mRNA were compared (Fig. 2). Growth regulation in the RevS transformant The recovery of a single transformant in which z n 2 + treatment was required to observe a CAMP-resistantprotein kinase activity, and the failure to observe any Caco-2 transformants with constitutively impaired CAMP-dependent protein kinase, suggested that the RI mutation might have an inhibitory effect on the proliferation of these colonic epithelial cells. Accordingly, the growth parameters of parental Caco-2 cells and the Rev5 transformant were determined by measuring DNA accumulation as a function of time in culture. As shown in Fig. 3, Caco-2 cells and the Rev5 transformant grown in the absence of ZnS04 had similar doubling times of approximately 22 h and reached similar saturation densities. Treatment of Rev5 cells with ZnS04 to induce the expression of mutant RI and inhibit

Days FIG. 3. Effects of the RI mutation on Caco-2 and Rev5 cell growth. Caco-2 ( a , 0 ) and RevS (m, 0 ) cells were grown in the 0 ) of 175 pM ZnSO, and assayed presence ( a , a) or absence (0, for DNA content over 8 days. ZnSO, treatment was initiated 40 h before plating cells and was continued throughout the growth experiment. Results are averaged from three separate experiments each carried out in triplicate and presented as means k SEM.

CAMP-dependent protein kinase activity inhibited cell growth by 60-70% over 6 days in culture. Thereafter, the Rev5 cells seemed to escape from the growth inhibitory effects of ZnS04. The growth inhibitory effects of ZnS04 in Rev5 cells seemed to be directly related to the induction of the mutant RI gene and inhibition of CAMP-dependent protein kinase activity, since this inhibitory effect was not observed in the Caco-2 parent. Furthermore, growth of the Zn2+-treated Rev5 transformant with low doses of 8BrcAMP reversed the inhibition of cell growth. Figure 4 illustrates the most dramatic reversal of the zn2+-induced inhibition of cell growth obtained with 8BrcAMP. In a second experiment, a more modest effect was observed, presumably reflecting variations in the residual low level of CAMP-dependent protein kinase activity available for activation by 8BrcAMP after z n 2 + treatment. The reversal of the growth-inhibitory effect of ZnS04 in the Rev5 transformant by 8BrcAMP further supports our conclusion that the inhibition of cell growth results specifically from the impairment in CAMP-dependent protein kinase activity. The resumption of Rev5 growth after prolonged exposure to z n 2 + (Fig. 3) may reflect the selective outgrowth of a subpopulation of unstable transformants no longer expressing the mutant RI.

Evaluation of chloride channel activity in the Rev5 transformant Chloride channel activity was evaluated by measuring the rate of radioactive iodide efflux in the presence or absence of stimulating agents (Venglarik et al. 1990). CAMPdependent activation of iodide efflux was tested using a stimulating cocktail of 10 pM forskolin plus 1 mM 8BrcAMP as described (Bear and Reyes 1992)to ensure maximum responses. This activation cocktail stimulated the rate of iodide efflux from parental Caco-2 cells approximately

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

MIHAJLOVIC ET AL.

o !

1

0

2

f

4

6

Day

FIG. 4. CAMPreverses the growth inhibitory effects of ZnSO, in Rev5 cells. Rev5 cells were treated for 40 h with 175 pM ZnSO, and then plated and grown for 4 days in the presence of 175 pM ZnSO, with (D) or without (u) 0.1 rnM 8BrcAMP. DNA content was analyzed over the 4-day interval as indicated. Results are averaged from duplicate or triplicate determinations at each point.

1.7-fold. The effect of the CAMP-generating cocktail was transient. It reached a maximum 3-4 min after application of the stimulus and then declined (Fig. 5a). Subsequent experiments demonstrated that forskolin was the active component of the stimulating cocktail, and the cAMP analogs 8BrcAMP, dibutyryl-CAMP, and N6-monobutyryl-CAMP were without effect. Similarly, we were unable to observe stimulating effects of TPA (1 pM) or the c a 2 + ionophores A23187 (1 pM) and ionomycin (2 pM) on the efflux of radioactive iodide (data not shown). The Rev5 transformant, in the absence of ZnS04, responded to the stimulating cocktail with increases in iodide efflux that paralleled the activity seen in the Caco-2 parent (Fig. 5 4 ; however, treatment of Rev5 with 175 pM ZnS04 for 16 h before the assay abolished the stimulation (Fig. 54. As shown previously, exposure of cells to ZnS04 for 12-24 h was sufficient to see the induction of mutant CAMPdependent protein kinase activity from the plasmid (Clegg et al. 1987; Bringhurst et al. 1989). In contrast, parental Caco-2 cells treated with ZnS04 still responded to the stimulating cocktail with a I .7-fold increase in iodide efflux, though basal and stimulated levels were inhibited slightly (approximately 25%; Fig. 5b). Therefore, the selective inhibitory effect of ZnS04 on iodide efflux in Rev5 cells presumably reflected the induction of the mutant RI and inhibition of CAMP-dependent protein kinase activity. Discussion The fluid and electrolyte compositions of the intestine are determined in large measure by CAMP-dependentelectrogenic chloride secretions from crypt epithelial cells and electrogenic sodium reabsorption from surface epithelia (Welsh et al. 1982). Two cell lines that have been used widely to study the regulation of intestinal fluid and electrolyte balance are the T84 and Caco-2 human colonic carcinoma cell lines. These cells exhibit neurohormone and CAMP-dependent chloride efflux (Bear and Reyes 1992; Burnham and Fondacaro 1989; Dharmsathaphorn and Pandol1986; Mandel et al. 1986) and

1043

express the chloride conductance regulator protein CFTR (Buchwald et al. 1991; Riordan et al. 1989), and thus provide useful model systems to study the regulation of intestinal chloride secretion. T84 cells are well-differentiated epithelial cells that carry out chloride transport in response to a variety of secretagogues including vasoactive intestinal polypeptide, prostaglandin E2, forskolin, carbachol, and the c a 2 +-dependent ionophores A23 187 and ionomycin (Dharmsathaphorn and Pandol 1986; Venglarik et al. 1990). Caco-2 cells are of interest because they differentiate in culture, changing from fetal crypt-like cells to villus-like cells over approximately 20 days in culture (Rousset 1986). The induction of specific hydrolases such as sucrase-isomaltase and cytoskeletal elements such as villin (Rousset 1986) provide markers for the functional differentiation of these cells. Caco-2 cells exhibit forskolin-stimulated chloride efflux in logarithmically growing cultures, but lose the regulated efflux when maintained in stationary phase. Responses reach a maximum four-fold stimulation after approximately 10 days in culture as cells enter the stationary phase of growth and then decline to a 1.25-fold increase after approximately 20 days in culture (A. Krolczyk and B. P. Schimmer, unpublished observations). Rogers et al. (1990) transfected T84 cells with an expression vector encoding a dominant mutant form of RI (i.e., pMTREV(AB)neo) and recovered two transfectants in which the cAMPdependent protein kinase was constitutively resistant to activation by CAMP. When the protein kinase defective transformants were examined for regulated chloride transport, they were found to be resistant to the secretory effects of forskolin, vasoactive intestinal polypeptide, and prostaglandin E2,but responsive to the c a 2 + ionophores ionomycin and A23187 (Rogers et al. 1990). These results suggest that cAMP and CAMP-dependent protein kinase were obligatory components of forskolin, vasoactive intestinal polypeptide, and prostaglandin E2 actions, but were not required for the calcium-mediated response. The effects of the protein kinase mutation on the actions of carbachol or phorbol esters were not reported. In this study, the mutant RI expression vector was used to isolate a protein kinase defective clone from the Caco-2 human colonic carcinoma cell line. We were able to isolate a single protein kinase defective transformant from Caco-2 cells in which CAMP-resistant protein kinase activity was observed only after cells were treated with ZnS04 to induce the synthesis of mutant RI. As determined in our subsequent experiments (Figs. 3 and 4), induction of mutant RI resulted in a CAMP-reversibleinhibition of cell replication. These observations indicate that a normal CAMP-dependent protein kinase activity is required for Caco-2 growth and suggest that the loss of a CAMP-responsive protein kinase is lethal to these cells. In a number of cell lines, cAMP inhibits cell growth and mutations affecting CAMP-dependent protein kinase activity seem not to have an adverse effect on cell survival (Schimmer 1989). In addition, a number of cell lines transfected with mutant RI expression vectors yield transformants that readily proliferate with constitutively impaired CAMP-dependent protein kinase activity (see Rogers et al. 1990 for brief review). On the other hand, some cell lines clearly require cAMP for growth. For example, lymphoid and canine MDCK kidney epithelial cells require cAMP at low concentrations to proliferate, whereas higher doses of cAMP inhibit cell division (Devis et a/. 1985; Taub

1044

BIOCHEM. CELL BIOL. VOL. 70, 1992

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

,.

0.05

1

0.25

1

0.05 5

0

10

15

(b) Caco-2 + ZnS04

1 10

5

I)

Tlme, mln

Time, min

15

-

(e) Rev5

0.05

0.05

0

5

10

15

Time, min

'

0

'

'

5

10

15

Time, min

FIG. 5. CAMP-stimulatediodide efflux in Caco-2 and Rev5 cells. Caco-2 and Rev5 cells were treated for 16 h in the presence or absence of 175 pM ZnSO, as indicated and loaded with [ 1 2 5 ~ as ] ~described a~ in Materials and methods. Cells were assayed for release of 1 2 5 at ~ 60-s intervals in the absence (0) or presence (a)of a stimulating cocktail of 10 pM forskolin and 1 mM 8BrcAMP. Stimulus was applied after the 3rd min of assay and rates of efflux (r) were calculated from the equation r = (In Rl - In R2) + (t, - t,), where Rl and R2are the percentages of counts remaining in the cells at times 1 ( f , ) and 2 (t,) (Venglarik et al. 1990). As shown, forskolin + 8BrcAMP stimulated chloride efflux from untransfected Caco-2 cells in the presence or absence of ZnSO, (a and b, respectively). In the transfected Rev5 clone, however, forskolin + 8BrcAMP only stimulated chloride efflux in the absence of ZnSO, (c); in the presence of ZnSO,, chloride efflux was markedly inhibited (6).

et al. 1979; Whitfield et al. 1973). Our results (Figs. 3 and 4) suggest that Caco-2 cells belong to the group that require cAMP and CAMP-dependentprotein kinase activity for cell division. It follows that transformants exhibiting constitutively impaired CAMP-dependent protein kinase activity could not survive without a second-site mutation to compensate for the CAMP-dependent growth requirement. Since the Caco-2 transformant described here exhibits CAMPresistant protein kinase activity only after induction with z n 2 + ,the uninduced transformant provides an endogenous control to exclude phenotypic changes that are independent of the alterations in CAMP-dependent protein kinase activity.

The effects of the mutant RI on CAMP-dependent chloride efflux were evaluated in Caco-2 cells using an isotopic iodide efflux assay. This assay is based on the selectivity of the chloride channel for iodide (Frizzell and Halm 1990) and has been validated for Caco-2 cells (Bear and Reyes 1992). We demonstrated that a CAMP-stimulating cocktail of forskolin plus 8BrcAMP was capable of increasing the rate of chloride efflux in Caco-2 cells; however, the active component of the cocktail was forskolin, not the cAMP analog. The absence of an effect of 8BrcAMP and other cAMP analogs on chloride efflux seemed inconsistent with a role for cAMP in the regulation of chloride efflux in these cells,

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

MIHAJLOV'IC ET

and raised the possibility that forskolin acted in a CAMPindependent manner. Indeed, forskolin has been shown to interact with a glucose transporter, regulate ion channel activity, stimulate cGMP accumulation, inhibit CAMPphosphodiesterase activity, and activate protein kinase C (Anderson et al. 1991a; Laurenza et al. 1989). Our finding that the action of forskolin on chloride efflux is specifically inhibited upon induction of a CAMP-resistantprotein kinase (Fig. 5) argues strongly for the obligatory roles of CAMP and CAMP-dependentprotein kinase in forskolin's actions. Therefore, another explanation is required for the seemingly contradictory results that 8BrcAMP had no effect on chloride efflux while affecting cell growth (Fig. 4). One possibility is that the rate of entry of the cAMP analog into the cell might be a limiting factor, since the effects of 8BrcAMP on growth were measured over periods of days, whereas the effects of 8BrcAMP on chloride efflux were measured in minutes. The lack of effect of c a 2 + ionophores and phorbol esters on chloride efflux in Caco-2 cells also was unexpected. It is possible that the 2.7 mM c a 2 + present in the a-MEM tissue culture medium is sufficient to maximally activate the c a 2 + signalling pathway in these cells or that c a 2 + and phorbol esters are required in combination to produce an increase in chloride efflux. These possibilities are being investigated. The availability of a Caco-2 transformant with a zn2+-inducibledefect in CAMP-dependent protein kinase will permit an investigation into the roles of cAMP and CAMP-dependent protein kinase in the actions of a variety of secretagogues on chloride efflux and the importance of the cAMP signalling system in the differentiation of Caco-2 cells from secretory crypt to reabsorptive villus cells. Acknowledgements This work was supported by research grants from the Canadian Cystic Fibrosis Foundation and the National Cancer Institute of Canada. Vera Mihajlovic received postdoctoral fellowship support from the Canadian Cystic Fibrosis Foundation. We thank G.S. McKnight (Department of Pharmacology, Seattle, Wash.) for generously providing pMT-REV(AB)neo and pRevlO, and C. Bear for her helpful discussion. Anderson, M.P., and Welsh, M.J. 1991. Calcium and cAMP activate different chloride channels in the apical membrane of normal and cystic fibrosis epithelia. Proc. Natl. Acad. Sci. U.S.A. 88: 6003-6007. Anderson, R.J., Breckon, R., and Colston, D. 1991a. Regulation by forskolin of cyclic AMP phosphodiesterase and protein kinase C activity in LLC-PK, cells. Biochem. J. 279: 23-27. Anderson, M.P., Rich, D.P.. Gregory, R.J., Smith, A.E., and Welsh, M.J. 1991b. Generation of CAMP-activatedchloride currents by expression of CFTR. Science (Washington, D.C.), 251: 679-682. Bear, C.E., and Reyes, E.F. 1992. CAMP-activatedC1 conductance in the colonic cell line: Caco-2. Am. J. Physiol. 262: C251-C256. Bringhurst, F.R., Zajac, J.D., Dagget, A.S., Skurat, R.N., and Kronenberg, H.M. 1989. Inhibition of parathyroid hormone responsiveness in clonal osteoblastic cells expressing a mutant form of 3 ' -5 ' -cyclic adenosine monophosphate-dependent protein kinase. Mol. Endocrinol. 3: 60-67. Buchwald, M., Sood, R., and Auerbach, W. 1991. Regulation of expression of CFTR in human intestinal epithelial cells. In The identification of the CF (cystic fibrosis) gene. Edited by L.-C. Tsui,

AL.

1045

G. Romeo, R. Greger, and S. Gorini. Plenum Press, New York. pp. 241-252. Burnham, D.B., and Fondacaro, J .D. 1989. Secretagogue-induced protein phosphorylation and chloride transport in Caco-2 cells. Am. J. Physiol. 256: G808-G816. Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J., and Rutter, W. J. 1979. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry, 18: 5294-5299. Clegg, C.H., Correll, L.A., Cadd, G.G., and McKnight, G.S. 1987. Inhibition of intracellular CAMP-dependentprotein kinase using mutant genes of the regulatory type I subunit. J. Biol. Chem. 262: 13 111 - 13 119. Clegg, C.H., Cadd, G.G., and McKnight, G.S. 1988. Genetic characteristization of a brain-specific form of the type I regulatory subunit of CAMP-dependent protein kinase. Proc. Natl. Acad. Sci. U.S.A. 85: 3703-3707. Dawson, D.C. 1991. Ion channels and colonic salt transport. Annu. Rev. Physiol. 53: 321-339. de Jonge, H.R., van den Berghe, N., Tilly, B.C., Kansen, M., and Bijman. J. 1989. (Dys)regulation of epithelial chloride channels. Biochem. Soc. Trans. 17: 816-818. De Lean, A., Munson, P.J., and Rodbard, D. 1978. Simultaneous analysis of families of sigmoidal curves: application to bioassay, radioligand assay, and physiological dose-response curves. Am. J. Physiol. 235: E97-E102. Devis, P.E., Grohol, S.H., and Taub, M. 1985. Dibutyryl cyclic AMP resistant MDCK cells in serum free medium have reduced cyclic AMP dependent protein kinase activity and a diminished effect of PGE, on differentiated function. J. Cell. Physiol. 125: 23-35. Dharmsathaphorn, K., and Pandol, S.J. 1986. Mechanism of chloride secretion induced by carbachol in a colonic epithelial cell line. J. Clin. Invest. 77: 348-354. Drumm, M.L., Pope, H.A., Cliff, W.H., Rommens, J.M., Marvin, S.A., Tsui, L.-C., Collins, F.S., Frizzell, R.A., and Wilson, J.M. 1990. Correction of the cystic fibrosis defect in vitro by retrovirus-mediated gene transfer. Cell, 62: 1227-1233. Frizzell, R.A., and Halm, D.R. 1990. Chloride channels in epithelial cells. In Channels and noise in epithelial tissues. Vol. 37. Edited by S.I. Helman and W.V. Driessche. Academic Press Inc., San Diego. pp. 247-282. Gold, G.H., and Nakamura, T. 1987. Cyclic nucleotide-gated conductances: a new class of ion channels mediates visual and olfactory transduction. Trends Pharmacol. Sci. 8: 312-316. Harper, J .F. 1988. Stimulus-secretioncoupling: second messengerregulated exocytosis. Adv. Second Messenger Phosphoprotein Res. 22: 193-318. Kissane, J.M., and Robins, E. 1958. The fluorometric measurement of deoxyribonucleic acid in animal tissues with special reference to the central nervous sytem. J. Biol. Chem. 233: 184-188. Laurenza, A., Sutkowski, E.M., and Seamon, K.B. 1989. Forskolin: a specific stimulator of adenylyl cyclase or a diterpene with multiple sites of action? Trends Pharmacol. Sci. 10: 442-447. Mandel, K.G., Dharmsathaphorn, K., and McRoberts, J.A. 1986. Characterization of a cyclic AMP-activated C1- transport pathway in the apical membrane of a human colonic epithelial cell line. J. Biol. Chem. 261: 704-712. O'Loughlin, E.V., Hunt, D.M., Gaskin, K.J., Stil, D., Bruzuszcak, I.M., Martin, H.C.O., Bambach, M.C, and Smith, R. 1991. Abnormal epithelial transport in cystic fibrosis jejunum. Am. J. Physiol. 260: G758-G763. Pon, D.J., Wong, M., Riordan, J.R., and Schimmer, B.P. 1990. CAMP-bindingproteins in epithelial cells cultured from human sweat glands. Am. J. Physiol. 258: C1036-C1043. Rich, D.P., Gregory, R.J., Anderson, M.P., Manavalan, P., Smith, A.E., and Welsh, M.J. 1991..Effect of deleting the R domain on CFTR-generated chloride channels. Science (Washington, D.C.), 253: 205-207. Riordan, J.R., Rommens, J.M., Kerem, B.-S., Alon, N.,

Biochem. Cell Biol. 1992.70:1039-1046. Downloaded from www.nrcresearchpress.com by Calif Dig Lib - Davis on 01/09/15. For personal use only.

1046

BIOCHEM. CELL BIOL. VOL. 70, 1992

Rozmahel, R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J.-L., Drumm, M.L., Iannuzzi, M.C., Collins. F.S., and Tsui, L.-C. 1989. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science (Washington, D.C.), 245: 1066-1073. Rogers, K.V., Goldman, P.S.. Frizzell, R .A., and McKnight, G.S. 1990. Regulation of C1- transport in T84 cell clones expressing a mutant regulatory subunit of CAMP-dependent protein kinase. Proc. Natl. Acad. Sci. U.S.A. 87: 8975-8979. Rousset, M. 1986. The human colon carcinoma cell lines, HT-29 and Caco-2: two in vitro models for the study of intestinal differentiation. Biochimie, 68: 1035-1040. Sahal, D., and Fujita-Yamaguchi, Y. 1987. Protein kinase assay by paper-trichloroacetic acid method: high performance using phosphocellulose paper and washing an ensemble of samples on flat sheets. Anal. Biochem. 167: 23-30. Schimmer, B.P. 1989. Cyclic AMP and other effectors of cyclic AMP-dependent pathways. In Drug resistance in mammalian cells. Vol. 1. Edited by R.S. Gupta. CRC Press Inc., Boca Raton. pp. 185-210. Tabcharani, J.A., Chang, X.-B., Riordan, J.R., and Hanrahan, J.W. 1991. Phosphorylation-regulated C1- channel in CHO

cells stably expressing the cystic fibrosis gene. Nature (London), 352: 628-63 1. Taub, M., Chuman, L., Saier, M.H.J., and Sato, G. 1979. Growth of Madin-Darby canine kidney epithelial cell (MDCK) line in hormone-supplemented, serum-free medium. Proc. Natl. Acad. Sci. U.S.A. 76: 3338-3342. Tilly, B.C., Kansen, M., van Gageldonk, P.G.M., van den Berghe, N., Galjaard, H., Bijman, J., and de Jonge, H.R. 1991. G-proteins mediate intestinal chloride channel activation. J. Biol. Chem. 266: 2036-2040. Venglarik, C.J., Bridges, R.J., and Frizzell, R.A. 1990. A simple assay for agonist-regulated C1 and K conductances in saltsecreting epithelial cells. Am. J. Physiol. 259: C358-C364. Welsh, M.J., Smith, P.L., Fromm, M., and Frizzell, R.A. 1992. Crypts are the site of intestinal fluid and electrolyte secretion. Science (Washington, D.C.), 218: 1219-1221. Whitfield, J.F., Rixon, R.H., MacManus, J.P., and Balk, S.D. 1973. Calcium cyclic adenosine 3 ' ,5 ' -monophosphate and the control of cell proliferation: a review. In Vitro, 8: 257-278.