Characterization of a cyclic AMP-resistant Chinese hamster ovary cell ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY

.

Vol. 260, No. 26, Issue of November 15,pp. 13927-13933,1985 Printed in U.S.A.

Characterization of a Cyclic AMP-resistant Chinese Hamster Ovary Cell Mutant Containing Both Wild-typeand Mutant Species of Type I Regulatory Subunit of Cyclic AMP-dependentProtein Kinase* (Received for publication, November 14,1984)

Toolsee J. Singhs, JacobHochmanp, Roberto Vernal!, Margaret Chapman, Irene AbrahamJJ, Ira H. Pastan, andMichael M.Gottesman** From the Laboratory of Molecdur Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892

We havecharacterized a cyclic AMP-resistant with theconclusion that thereare at least two different Chinese hamster ovary(CHO) cell mutant in which one species of RI present in CHO cells and that one of these of two major species of type I regulatory subunit (RI) species is altered in the mutant analyzed in this work. of cyclic A ~ P - d e ~ n d e n protein t kinase is altered. Wild-type CHO cell extracts contain two cyclic AMPdependent proteinkinase activities. AS shown by DEAE-cellulose chromatography, there is a peak of Among the many effects cAMPhas on cultured fibroblasts type I protein kinactivity in mutant extracts, but such as Chinese hamster ovary (CHO') cellsare an inhibition the type I1 protein kinase activity is considerably re- of cell growth (1-3), altered amino acid and sugar transport duced even though free type I1regulatory subunit (RII) (4),and decreased agglutinabilityby lectins (5,6). It has been is present, The type I kinase from the mutant has an altered RI(RI*) whoseKD for thebinding of S-NS[~'P] proposed that these different effects of cAMP on cultured cAMP (KD= 1.3 x lov6M) is increased by more than cells are all the result of phosphorylation reactions mediated 200-fold compared to RI from the wild-type enzyme by the CAMP-dependent protein kinases (7, 8). if this is indeed true thenin principle any cell that lacks these enzyme (KD= 5.5 x M). No differences were foundbetween the catalytic subunits from the wild-type and activities should not show any CAMP-mediated effects. Exmutant type I kinases. A large portion of RI inmutant tensive genetic analysis in the S49 lymphoma system where and wild-type extracts is present in the free form. The cAMP is cytotoxic and in other cell systems has confirmed RI* derived from mutanttype I protein kinase shows this conclusion (reviewedin Ref. 9). altered labeling by 8-NS['eP]cAMP (& = 1.3 x lo-' To test this and alternative hypotheses regarding CAMP M) whereas the free RI from the mutant is labeled effects in cultured fibroblasts we have isolated a number of normally by the photo~finity label (& = 7.2 X lo-' independent mutants from CHO cells (10). These m u ~ n t s M), suggesting that theRI* which binds to the catalytic are resistant to the growth-inhibitory effects CAMPhas on subunit is functionally different from the freeform of the wild-type cells. The CAMP-dependent protein kinase acRI. The decreased amount of type 11kinase activityin tivity in a number of these mutants has been characterized the mutant appears be todue to competition of RI*with (11,12). In these mutants, CAMPeffects on morphology (lo), RII forbinding to the catalytic subunit. Translation of nutxient transport (41, and ~ n d u ~ i oofno ~ i t ~ decarboxn e mRNA from wild-type CHO cells results in thesynthe- ylase (13) were blocked confirming that normal CAMP-desis of two different charge forms ofRI, providing pendent protein kinase activity is needed for these effects in biochemicalconfirmation of two differentspecies of RI cultured fibroblasts. in CHO cells. Additional biochemical evidence based Recent findings basedon the use of the photoaffinity labe1 on isoelectric focusing behavior of 8-N3[32P]cAMP-laS-~~["P]cAMP have indicated that some cell types, such as beled RI species and 136S]methionine-label~ RI from mutant and wild-type extracts confirms the charge the neuroblastomacell (14) and the CHO cell (12), have heterogeneity of RI species in CHO cells. These genetic significant quantities of a CAMP-binding protein which is not and biochemical data taken together are consistent associated with either type I or type I1CAMP-dependent protein kinase activities resolved on DEAE-cellulose chro* The costs of publication of this article were defrayed in part by mato~aphy.This binding activity resembles the regulatory the payment of page charges. This article must therefore be hereby subunit of type I CAMP-dependent protein kinase (RI) and marked "advertisement" in accordance with 18 U.S.C. Section 1734 in some cases may be formed by dissociation of this holoenzyme. in this work, however, we report the characterization solely to indicate this fact. $.Present address: Section on Metabolic Regulation, Endocrinology of a CAMP-resistantCHO mutant (10248)in which the affinand Reproduction Branch, National Institute of Child Health and ity of 8-N3[32P]~AMP for the mutant hol~n~me-associated Human Development, National Institutes of Health, Bethesda, MD RI (Ri*) appears to be approximately 2 orders of magnitude 20205. f Present address: Department of Genetics, Institute of Life Sci- less than the affinity for free RI. These data, taken together with biochemicaldata demonstrating the charge heterogeneences, The Hebrew University,Jerusalem, Israel. ity of RI species in CHO cells, indicate that there are at least ll Present address:Universitadegli Studi, Istituto di Patologia Generale, Roma, Italy. 11 Present address:CellBiology Department, The WpjohnCo., Kalamazoo, MI 49007. ** To whom correspondence and reprint requests should be sent. Bethesda, at: National Institutes of Health, Building 37, Room 2E18, MD 20892.

'The abbreviations usedare:CHO,Chinese hamster ovary; 8N3[32P]~AMP, 8-a~ido[~~P]adenosine 3':5'-monophosphate;CAMP, adenosine 3':5'-~onophospha~;RI and RII, regulatory subunits of type I (PKI) and typeI1 (PKII) protein kinases; C, catalytic subunit of type I and type I1 protein kinases; SDS, sodium dodecyl sulfate.

13927

13928

CHO Mutant with a Defective RI

two major functional species of RI-like CAMP-binding proteins in CHO cells. Recent evidence based on the likelihood that there is only one gene coding for RI (accompanying paper; Ref. 15) suggests the p o ~ i b ~ ithat t y thetwo forms of RI in themutant may be the result of a mutation in one of two separate alleles of a single gene for RI.

from [%3]methionine-labeled extracts using S t a ~ h y ~ aurew ~ c c ~ was as previously described (22) except that buffer A contained 0.05% SDS, using a rabbit anti" antibody generouslyprovided byJ. Beavo (University of W a s h i n ~ ~ n ) .

RESULTS

Protein Kinase A c ~ a id P ~ H ~ ~ ~ FActivities - b i ~ ~ ~ in Extra&-Extracts from the mutant and wild-type cells were analyzedfor their CAMP-dependent protein kinase and f3H]cAMP-bmdingactivities. Fig. lA shows the activation of Type IIA histone, the protease inhibitor aprotinin, and all unla- the kinases by different concentrations of CAMP. It can be beled nucleotides were purchased from Sigma. [d"PJATP was from observed that the kinase activation curve from the mutant New England Nuclear and 8-N3[32P]cAMPfrom ICN. Protein assay extract is shifted to the right, requiring greater than 10-fold kits andall reagents for SDS-gel electrophoresis were from Bio-Rad. higher concentration of CAMP for maximal activation. Fig. 1B shows the binding of r3H]cAMP bythe mutant andwildMethods type cell extracts. In contrast to the data obtained for kinase Cell Cultures and Mutant Selection-CHO cells were grown in a- activity (Fig. IA),these two curves are quite similar. This modified minimal essential medium as previously described (10). The finding suggests that most of the RI and RII in the mutantis ~ P E R I M E N T A LPROCEDURES

mutant 10248 wasselected from ethyl methanesulfonate-mutagenized cultures of wild-type (10001) cells.Clone 10248 wasselected because of its ability to grow in medium supplemented with cholera toxin (1 pg/ml) and theophylline (1 mM) as previously described (10). In order to grow cells for preparation of extracts, both mutant and wild-type cells were plated on 100-mm tissue cuiture dishes (Falcon) at 37 "C (5 X lo6ce11s dish) and harvested 3 days later when they were about 90% confluent. Sucrose G ~ d i e n ~ - S u c r ~gradient e centrifugation was carried out according to the method of Martin and Ames (16). Linear gradients (11 ml) of5-15% were prepared in buffer A (5 mM sodium phosphate and 1 mM EDTA (pH 7.6), 1 m ~ v ldithiothreitol, 1% aprotinin).Protein samples were dialyzed overnight against two changes of buffer A and concentrated by u l t r a ~ l t ~ t i oAliquots n. (0.3 ml containing 0.8-1.0 mg of protein) were gently layered on gradients which were then spun for 18 h at 40,000 rpm in a SW 40 Ti rotor at 4 "C. Phosphorylase b, hemoglobin, type I protein kinase, and free RI were used as marker proteinsand were run simultaneously in parallel gradients. After the run, each gradientwas fractionated by collecting 0.32-ml samples from a hole punctured in the bottom of the tube. Fractions were assayed for kinase activity and the incorporation of 8-N,[82P]~AMP. Preparation of Regulatory and Catalytic Subunits-For these experiments the pooled and concentrated typeI and type I1 holoenzymes from DEAE columns were used (12). The wild-type and mutant enzymes were separately incubated with 10 PM CAMP for 30 min at 3 "C to allow for binding equilibrium, then at 30 "C for 5 min to facilitate dissociation into R and C subunits. The dissociated enzyme was then applied to a small DEAE-cellulose column (0.7 X 1.5 cm) equilibrated with buffer A. The C subunit does not bind to D E S cellulose and was collected in the flow-through in a tube containing 100 pg/ml bovine serum albumin. The column was then washed with 1 ml of buffer A containing 10 p~ CAMP. The wash fraction was combined with the void volume of the column. This represents the preparation of C used in this study. The column was further washed with 5 ml of buffer A alone. The free regulatory subunits (RI and RZI) were eluted with 0.2 M NaCI in buffer A. Any residual free C A W was removed by passing RI and RIIthrough a Sephadex G-25 column (0.7 X 10.0 em) equilibra~d buffer in A. Slab Isoelectric Focusirag of P h o t ~ ~ ~EI-Photoaffint y - ~ ~ ity-labeled samples from extracts or DEAE fractions were prepared as previously described (12, 14). Slab isoelectric focusing was as described by Doherty et al. (17) with the following modifications. Samples were acetone precipitated prior to loading (18) and m p h o lines were 1 ml ofpH 4-6 and 1ml of pH 5.7 (LKB). Electrode buffers were 0.5% ethylenediamine (basic) and 0.2% H&304 (acidic). I n Vitro Translation of CHO mRNA and Two-dimensional Gel Analysis of Affinity Purified Translated RI-Total RNA from CHO cells was prepared as previously described (19) and [3SS]methioninelabeled proteins from translation reaction mixes were affinity purified on N6-(2-aminoethyl)-cAMP-agarose gels. Two-dimensional gel electrophoresis was as described previously (18). Other Methods-Protein kinase activity was measured using histone as substrate as described by Corbin and Reimann (ZO), ['HJ CAMP-bindingwas determined as described by Gilman (21), and 8N,['T]cAMP incorporation and quantitation into regulatory subunits was as previously described (12). Immunoprecipitation of RI

I

0

0

9

I

I

8

7

-

-.-

I

I

6 5 Log [CAMPI, M

I

I

4

3

6

C

8

e

a

E" . 8

-! z 4 0

z 3 ~ 0

n

f 4

-F

C-4

2

Wild Type

o"0 Mutant

0

-

L-[cAMPl,

M

FIG. 1. Protein kinaseand CAMP-binding activities in mutant and wild-type extracts. A, cell extract (35 pg of protein) from

wild-type (A) or mutant (m) cells was assayed for kinase activity in the presence of various concentrations of CAMP. B, additional aliquots (210 pg of protein) of the mutant andwild-type extracts were also assayed for I3H]cAMP binding.Symbols are thesame as in A.

CHO Mutant with a Defective RI probably normal in its ability to bind CAMP.As we will show below, RI from the type I kinase holoenzyme in this mutant is defective. Alteredbinding of CAMPis not detected in crude extracts because these extracts contain an excess of RI and RII not associated with holoenzyme. In addition, the Gilman ['HICAMP binding assay (21) used in these initial studies may detect only one of two CAMP-binding sites on RI and hence not given a complete picture of the ability of RI to bind CAMP.As will be shown below, RI* in holoenzyme and free RI have different characteristics in themutant. Fractionation of Extracts on DEAE-cellulose-We have previously shown(11,12) that extracts from wild-typeCHO cells can be fractionated on DEAE-cellulose into type I and type I1 protein kinases. In addition, we used the photoaffinity analogue of CAMP, 8-N3(32P]cAMP, to show that up to 70% of RI is not associated with enzyme activity but is present in wild-type extracts as the free regulatory subunit (12). Fig. 2 shows the DEAE profile of the mutant extract. The type I protein kinase (peak I,) is present in apparently normal amounts when compared to its wild-type counterpart. The type I1 holoenzymeis,however, present in very small amounts. We have previously reported the characterization of a mutant (10215) missing type I1 protein kinase in which the defect resided in the catalytic subunit isolated from the type I kinase (11).In mutant 10248,some of the protein kinase activity does not bind to the DEAE-cellulose and is recovered in thewash (fraction I6 in Fig. 2). This phenomenon is consistently seen for mutant 10248, but since it occasionally occurs in wild-type extracts as well we cannot be certain of its significance. Also shown in Fig. 2 is the incorporation of 8-N3[32P]cAMPinto the different fractions. It can be seen that there is no incorporation of the label into the type I protein kinase. However, a large peak of RI not associated with type I kinase is readily observed. The amount of nonholoenzyme-associatedRI from the mutant is comparable to that found in wild-type extracts (12). Another smaller peak of RI (3% of total RI)binding is observed and co-elutes with the kinase activity that does not bind to thecolumn (peak I b

-

in Fig. 2). Although type I1 prot+n kinase activity is absent, a peak of RII binding is present. This peak is comparable in amount to that found in wild-type extracts (11). Enzyme Activation and 8-N3P2PJcAMPImorporetion into Type I Protein Kinase from Wild-type and Mutant ExtractsThe results of Figs. 1and 2 suggest that the type I kinase in the mutant may be defective, To investigate this possibility more closely we have used the type I holoenzyme partially purified from wild-type and mutant extracts by DEAE-cellulose chromatography.In studying the activation of the purified mutant and wild-type enzymes by different concentrations of CAMP,we have found that the shift to the right in the curve observed with the mutant extract (Fig. 1) is also seen with the partially purified type I holoenzyme (data not shown). Fig. 3A shows the incorporation of increasing amounts of 8-N3[32P]cAMPinto thetype I holoenzyme from mutant andwild-type extracts. It can be seen that appreciable incorporation of the label by the mutant enzyme occurs only at concentrations greater than M. At the highest concentration of the label tested in this experiment (3 X 10" M) incorporation by the mutant RL" is still not saturating (Fig. 3A). In separate experiments (datanot shown) using higher concentrations of 8-N3[32P]cAMP,the mutant RImaximally incorporated approximately the same amount of label as its wild-type counterpart. The concentrations of label required for half-maximal incorporation by the wild-type and mutant enzymes are 5.5 X and 1.3 X M, respectively. It is feasible that CAMP or 8-N3[32P]~AMP may bind less efficiently to RI of the type I holoenzyme from mutant extracts because of an altered interaction between RI and C. Such unfavorable interactions can dccur if either RI or C is defective. Fig.3B shows that thedefective incorporation of 8N,[32P]cAMP into RI* persists even when free RI* derived from the type I holoenzyme is used. Hence, this latter result strongly suggeststhat RI* of the type I holoenzyme from the mutant extract is defective. The Catalytic Subunit of the Type I Kinase Has Normal Phsphotransferase Activity-The presence of a defective RI* I

RI

FIG.2. Fractionation of extract from mutant cells by DEAE-cellulose chromatography. A sample containing 35 mg of protein was applied to (0.7 X 10.5 aDEAE-cellulosecolumn cm) previously equilibratedin buffer A. The column was washed with 20 ml of buffer A, then eluted with a linear gradient of 35 ml of buffer A and 35 ml of buffer A containing 0.4 M NaCl. Fractions of 1ml were collected.Aliquots (35 pl) from fractions were assayed forprotein kinase activity in the absence (A) and presence (A) of 10 b~ CAMP. Another aliquot (54 pl) was assayedfor the incorporation ofS-Na[32P]cAMP (IO4 M)into RI (0)and RII (m). -, [NaCl] determined by conductivity.

16

-

14

-

4.0

t

12 -

10

13929

-

8-

6-

4-

2-

0FRACTION NUMBER

1

Defective RI

a CHO with Mutant

13930

more convincingargument for the nonidentical nature of the two species of RI is supplied by examining their 8-N3[32P] A 3 4CAMP-incorporation curves. Fig. 4 shows the incorporation Wild Type of increasing concentrations of the label by free RI from G 4 Mutant a mutant and wild-type extracts. These curves are similar for mutant and wild-type cells (half-maximal binding at 7.2 x 3P 10- and 7.0 X 10” M ,respectively). However,RI not associated with holoenzyme incorporates 8-N3[32P]cAMPwith much greater affinity than RI* derived fromthe type I kinase 2/I / of themutant (half-maximal binding at 1.8 X M) (compare Figs. 3B and 4). These differences cannot be due to the Y/ fact that RI* in Fig. 3B is preloaded with CAMP since no such 1/ D/’ differences are observed in RI* labeling in Fig. 3A (unloaded) ,’ 0 ’ compared to Fig. 3B (loaded) or between labeling of wild-type ”/---~ RI seen in Fig. 3B (loaded) compared to Fig. 4 (unloaded). , OF t Hence, these data imply the existence of two forms of RI in B Dissociated Type I Kinase this mutantsuggesting that the RIwe find unassociated with 4 holoenzyme is not derived simply by dissociation of holoenI zyme. We havefound that free RI in mutant10248 is able to A A A inhibit the histone kinase activity of beef heart C (data not A O h shown). 3P RII Is Present as the Free Regulatory Subunit in Mutant / / Extracts-Fractionation of mutant extracts on DEAE-cellu/ / lose fails to show the presence of a type I1 protein kinase 2o/ 0 / activity peak although the incorporation of 8-N3[32PplcAMP / to RIIis detected (Fig. 2). We have consideredthe possibility // that even though no type I1 kinase activity is present the 1A’ 0 latter enzyme may bepresent in a catalytically inactive holo0 :/a” enzyme complex. Wehave attempted to resolve this question c ” ” ”””” L by sizing the RII peak from DEAE-cellulose (see Fig. 2) on a 7 6 5 4 sucrose gradient. Fig. 5 shows that theRII binding peak from - Log [8-N,- I’PlcAMPI. M the mutant (Fig. 5C) migrates the same as RI derived from FIG. 3. Concentration-dependent incorporation of 8-NS the type I1 holoenzyme (Fig. 5B). It should be pointed out [3zP]cAMPby wild-type and mutant RI either in the presence that the presence of RI binding in all the profiles shown in or absence of C. A, aliquots (18 pg of protein) of the pooled and Fig. 5 is due to thecontamination of the type I1 kinase peak concentrated type I kinase from wild-type (A) and mutant (U) extracts were assayed for 8-N3[32P]eAMPincorporation in theconcen- by free RI when fractions are pooled from DEAEkellulose (see Fig.2).Weconclude that RII is not associated with . the type I holoenzyme (0.8 mg) from tration range 0.01-30 p ~ B, wild-type and mutant cells was dissociated into R and C subunits as catalytic subunit as atype I1 holoenzyme in the mutant described under “Methods.” The incorporation of 8-N,[32P]cAMP extract. into the free RI was done as for the holoenzyme. Symbols are the Heterogeneityof RI in Mutant andWild-type Extracts-To same as in A. ascertain whether the mutation in CHO strain 10248 produced I

I

A

TypeIKiise

A

4

” ”

I

I

I

I

I

does not exclude the possibility that theC subunit of the type I holoenzyme may also be defective. Characterization of another CHO mutant (10215) from this laboratory with a similar phenotype established that the type I kinase had a defective C. When compared with the type I enzyme from wild-type cells, the mutant C showed an altered ATP curve (11).We examined the effect of increasing concentrations of ATP on mutant and wild-type extracts. We have found that the type I C from wild-type and mutant(10248) extracts have the same K, for ATP, approximately 3 X M. This is so regardless of whether C is in the presence or absence of RI. Furthermore, C fromboth the wild-type and mutant(10248) type I kinases also phosphorylate histone to thesame extent (K, for histone, approximately 0.2 mg/ml). Free RI in Mutant Extracts Is Not Derived from Type I Protein Kinuse-As was shown in Fig. 2, the RIincorporation of 8-N3[32P]cAMPwas primarily represented by a largepeak of RI not associated with type I holoenzyme. One possible hypothesis is that this RI is derived fromtype I protein kinase which has become dissociated during the preparation and fractionation of the extract on DEAE-cellulose. If this is the case then an equal amount of C would be expected to be present. However, using the heat-stable inhibitor (23) of protein kinase we find no free C activity (data not shown). A

-.C

I

se

0

-. W

E

,

I

8

Free RI

Wild Type 4-

D-0 Mutant

0

E

0

P

~”“-n

” ”

A

a 0

”

A

I

I

zOD

L ‘

8

L

7

I

6

5

4

- Log [~-N,-[’PICAMPI,M FIG. 4. Concentration-dependent incorporation of 8-Na [3aP]cAMPby free RI from wild-type and mutant extracts. Extracts were fractionated ona DEAE-cellulose column and thepeak of free RI binding (see Fig. 2) used for this experiment. Aliquots (23 pg) of the pooled and concentrated free RI were assayed for 8-N3[”P] CAMPincorporation in theconcentration range 0.01-30 p ~ A, . wildtype; 17,mutant.

CHO with Mutant

a Defective RI

13931

B A ”

abC-

d-

l

FIG. 5. Sucrose gradient centrifugation of type I1 kinase from wild-type and mutant extracts. Sucrose gradient centrifugation was performed as described under “Methods.” Sedimentation is from right to left. A, type I1 holoenzyme from wild-type extracts. B, RII, prepared from the type I1 kinase from wild-type cells as described under “Methods” was sedimented. C, RII binding peak from mutant extracts (seeFig. 2). The symbols used are: A,A, protein kinase activity assayed in the presence and absence of 1 p~ CAMP, respectively; 0,M, incorporation of 8-N3[32P]cAMPinto RI and RII, respectively.

2

1

2

3 4

FIG. 6. Isoelectric focusing behavior of RI from wild-type and mutant extracts and DEAE fractions of wild-type extracts: effect of alkaline phosphatase treatment.Panel A, extracts of wild-type (lane I ) and mutant (lane 2) cells were labeled with 8-N3[32P]cAMPand run on slab isoelectric focusinggels as described under “Experimental Procedures.” Samples were loaded at the basic end (top)of the gel. The letters are used to identify the four isoelectric forms of RI found in CHO cells, labeled u-d, from the most basic to themost acidic form. Forms b and c are not well-resolved on this gel. Panel B, DEAE fractions from wild-typeextracts containing type I kinase holoenzyme (lane 1 and 3) and free RI (lanes 2 and 4 ) were labeled with 8-N3[32P]cAMPand separated on slab isoelectrofocusing gels without any treatment (lanes 1 and 2) and after treatment for 30 min at 37 “C with Escherichia coli alkaline phosphatase (2 pg/ml) (lanes 3 and 4).

(c and d ) are reduced in intensity when compared with the basic bands (a and b) after treatment with alkaline phosphatase (lanes 1 and 2, compared withlanes 3 and 41, suggesting an alteration in the charge of the RI subunit, we photoaffnitylabeled extracts of mutant 10248 with 8-N3[32P]cAMPand that they are phosphorylated speciesof the two more basically migrating wild-typeRI subunits. Fig. 6B also comparesthe 8compared the labeled products with photoaffinity labeled wild-type extracts. Fig. 6 shows an autoradiogram of a slab N3[32P]cAMP-labelingpattern of RI associated with type I isoelectric focusing gel in which mutant andwild-type extracts kinase (lane l ) , with free RI (lane 2). These patterns are are displayed. Both extracts show four isoelectric species of similar, with all four isoelectric species present. This result RI under conditions in which RI appears as a single band on indicates that these different species of RI are all found in a one-dimensional SDS-polyacrylamide gel (Fig. 6A, lanes 1 both holoenzyme-associatedand free RI species in wild-type and 2). The two most basic bands are increased in mutant CHO extracts. Because 8-N3[32P]cAMP might not be expected to label the 10248relative to the acidic bands, but no new bands appear, mutant RI* subunits in 10248 efficiently, we also metaboliindicating that themutation in mutant 10248 doesnot alter the charge of RI* as detected by 8-N3[32P]cAMP labeling. cally labeled wild-typeand mutant extracts with [35S]methiTreatment of type I kinase and free RI from wild-typeCHO onine and immunoprecipitated RI* and RI from these excells with alkaline phosphatase prior to photoaffinity-labeling tracts. Fig. 7 shows two-dimensional gel analyses of these is shown in Fig. 6B. It can be seen that thetwo acidicbands labeled immunoprecipitates. As can be seen, wild-type ex-

CHO Mutant with a Defective Rl

13932

A

B

C

D

FIG. 7. Immunoprecipitation and two-dimensional gel analysis of RI from wild-type and mutant extracts of CHO cells. 5 X 10‘ cells were labeled with 1 mCi of [%]methionine for 3 h and extracts containing approximately IO7 cpm were prepared and immunoprecipitated with anti-RI antibody as described under “Experimental Procedures.” The first dimension was isoelectric focusing, run from left (basic end) to right (acid end). The vertical dimension was SDS-polyacrylamide gel. Panel A shows the two-dimensional gel of an extract from wild-type cells precipitated with normal rabbit serum; Panel B, wild-type extract with anti-RI serum; Panel C, mutant extract with normal rabbit serum; Panel D, mutant extract with anti-RI serum. The letters identify the four isoelectric species of RI specifically immunoprecipitated by the anti-RI serum.

CAMP-dependentprotein kinase activity, confirming the hypothesis that this enzyme is essential for growth-inhibitory effects of high cAMP levelsin mammalian cells(7). To date, four of these CAMP-dependent protein kinase mutants have been characterized in detail. In one mutant (10215) the type I1 kinase is missingand thetype I enzyme, although present in normal amounts, has a defective C subunit (11).Two other mutants (10265 and 10223) havea missing type I kinase and reduced amounts of the type I1 enzyme (12).A third mutant (10260)has very little type I or type I1 activity (12). In this study we present the detailed characterization of a novel CHO CAMP-resistant mutant (10248).Mutant 10248 is FIG. 8. Two-dimensional electrophoresis of affinity puri- resistant to cAMP effects on growth, morphology (lo), nufied RI translated from mRNA isolated from wild-type CHO cells. The direction of electrophoresis was as described in the legend trient transport (4) and induction of ornithine decarboxylase to Fig. 7. The letters identify the two isoelectric species of RI which activity (13). Based on assaysin crude extracts, this mutant co-migrate with 8-Ns[32P]cAMP-labeledextracts from wild-type cells. was originally thought to have normal levels ofCAMP-deThe additional [36S]methionine-labeledproteins seen on this gel are pendent protein kinase activity and CAMP-binding activity presumed to be other nucleotide-binding proteins with affinity for (10). The studies reported here, however, indicate that there the N‘-(aminoethy1)-CAMP-Sepharose. are at least two defects in this mutant affecting its protein kinase activity. These two defects,(discussed below) are 1) tracts contain four species of RI (Fig. 7B), corresponding to loss of type I1 kinase activity and 2 ) altered affinity of holothe four 8-N3[32P]cAMP-labeled species, whereasthe mutant enzyme-associatedRI for CAMP.As indicated below, we hyextracts have reducedamounts of the most acidic species (Fig.pothesize that these two defects are secondary to a single 70). alteration in one of two speciesof RI in this mutant. Direct support for the hypothesis that wild-type extracts We have shown that extracts of this mutant contain only contain two species of RI comes from translation of mRNA type I protein kinase. The type I1 enzyme is much reduced, from wild-type cells. Fig. 8 shows a fluorogram of a two- although free RII is present. In this respect, this mutant dimensional gel of [3SS]methionine-labeledtranslation prod- seems very similar to another of our mutants (10215) charucts isolated after N6-(aminoethy1)-CAMP-Sepharoseaffinity acterized previously (11).Extracts of the latter also have no chromatography. Two CAMP-binding speciesof differing is- detectabletype I1 kinase with freeRII being present. However, oelectric points corresponding to the two basic forms of RI although the type I kinases from both 10248 and 10215 are shown in Fig. 7 are clearly seen. defective,the defect in the former (10248) appears to be in its regulatory subunit (see below) whereas the catalytic subunit DISCUSSION of mutant 10215has been shownto be altered (11).A similar We have isolated over30 independent CAMP-resistant alteration in the C subunit from mutant 10248 has not been CHO mutants in our laboratory over the last several years found. Evidence presented in this paper that mutant 10248 has a (10,24)? The great majority of these have had defects in their defective RI* subunit may be summarizedas follows. 1)IsoC. Roth and M. M. Gottesman, unpublished data. lated RI* from PKI binds and incorporates 8-N3[32P]cAMP

CHO Mutant witha DefectiveRI with a 200-fold lesser affinity than wild-type RI or free RI from the same mutant (Fig. 3); 2) PKI has a shifted doseresponse to CAMP. The difference in magnitude of this shift compared to the 200-fold reduction in 8-N3cAMP incorporation may reflectthe different affinity of 8-NacAMPand cAMP for their two binding sites on RI (25) or may simply reflect the non-equilibrium nature of these measurements; and 3) RII andC subunits in themutant appear to be normal. The functional alteration in RI in mutant10248 is similar to that seen in a class of mutants of S49 cellswhose resistance to dibutyryl CAMP is due to an altered RI. Many of these well-characterizedS49 mutants havebeenshown to have isoelectric variant RIs (26), as has also recently been demonstrated in 8-Br-CAMP resistant Y1 adrenal cells (27). We have sought direct physical evidence for an altered RI* in mutant 10248without success. As seen in Figs. 6 and 7 there is no alterationin theisoelectric mobilityof mutant RI*, and two-dimensional tryptic peptide maps of8-N3[’2P]cAMPlabeled peptides from mutant RI* and wild-type RI show no differences (data not shown). However, by DNA-mediated gene transfer itis possible to transfer the complete phenotype of mutant 10248 byintroducing the mutantRI gene detected with an RI probe by Southern blotting into wild-type cells (accompanyingpaper, Ref. 15). These additional data provide genetic evidence that thephenotype of mutant 10248is linked to analteration in the RI* gene. How does the defect in RI* in mutant 10248 explain its failure to respond to CAMP? Studies on phosphorylation in intact wild-type and mutant CHO cells indicate that mutant 10248shares with mutant 10215an inability to pho~horylate a 52,000 molecular weightprotein after cAMP treatment (28). Both of these mutants lack type I1 kinase, and both have a type I kinase whose activation requires increased amounts of CAMP,albeit for different reasons. Either of these defects or both together could account for the failure of 10248to respond to C A M P . The reason for the lack of type I1 protein kinase in mutants 10248 and 10215 may relate to the failure of C to dissociate from RI under usual conditions of CAMPstimulation of PKI. By this model, mutant PKI acts to bind most available C, reducing the amount of PKII in the cell at steady state. By the same argument, wild-type RI in the mutant would also not be associated with C a t steady-state. The present mutant (10248) gives us some intriguing information about the origin in crude extracts of RI not associated with holoenzyme. We havenoted the presence of this form of RI in both wild-type and other mutant extractsbefore (12). It has also been observedin cultured neuroblastoma cells (13). A persistent question has been whether this free RI was derived in some way by dissociation of type I kinase., The data in Fig. 6 suggest that all four isoelectric species of RI are found in both type I kinase and free RI fractions from DEAE columns. However,the data from mutant 10248indicate that free RI may under some conditions be different from RI* complexed with C to form type I holoenzyme. These two RIs inthemutant have dramatically different incorporation curves for 8-Ns[32P]~AMP, the RI* fromtype I kinase showing defective incorporation of the nucleotide either in theabsence or presence of C. The data presented in the accompanying paper (15) suggest that there is only one RI gene in CHO

13933

cells. Hence,in mutant 10248 RI* fromtype I kinase and free RI may represent two alleles of this single gene. This idea is supported by the demonst~tionof two different CAMP-binding translation products from CHO mRNA (Fig. 8). It is also possible that free RI in the mutant is encoded by a genenot detected by the bovine RI probe used in the accompanying paper or that both forms of RI are derived from differently processed forms of the same transcript. Nevertheless, the characterization of mutant 10248 reported here confhms the conclusion that there are multiple ways in which protein kinase can be altered to give rise to CAMP-resistant cells. A c k n o w ~ g m e ~ - W ewould Iike to thank J. Siverman and J. Sharrar for help with the typing of the manuscript and C. Roth and J. Castaiio for critical discussions in theearly phases of this work. REFERENCES 1. Johnson, G. S., Friedman, R. M., and Pastan, I. (1971)Proc. Natl. Acad. Sei. U. S. A. 68,425429 2. Johnson, G. S., and Pastan, I. (1972)J. N&l. Cancer Inst. 48, 1377-1387 3. Hsie, A. W., and Puck, T. T. (1971)Proc. Natl. Acmd. Sei. U. S. A. 68,358-361 4. LeCam, A., Gottesman, M. M., and Pastan, I. (1980)J. Biol. Chem. 255,8103-8108 5. VanVeen, J., Roberts, R. M., and Noonan, K. ,D. (1976)J. Cell Biol. 70,204-216 6. WiIlingham, N., and Pastan, I. (1975)J. CeU BioZ. 67,146-159 7. Krebs, E. G. (1972)Curr. Top. Cell. Regul. 5,99-133 8. Kuo, J. F., and Greengard, P. (1969)Proc. Natl. Acad. Sei. U.S. A. 64,1349-1355 9. Gottesman, M. M. (1980)Cell 22,329-330 10. Gottesman, M. M., LeCam, A., Bukowski, M., and Pastan, I. (1980)Somatic CeU Geenet. 6,45-61 11. Evain, D., Gottesman, M.M., Pastan, I., and Anderson, W.B. (1979)J. Bwl. Chem. 254,6931-6937 12. S i g h , T. J., Roth, C., Gottesman, M. M., and Pastan, I. H. (1981) J. Biol. Chem. 256,926-932 13. Lichti, U.,and Gottesman, M.M. (1982)J. Cell. Physiot. 113, 433-439 14. Walter, U., Costa, M. R. C., Breakefield, X. O., and Greengard, P. (1979)Proc. Natl. Acad. Sci. U. S. A. 76,3251-3255 15. Abraham, I., Brill, S., Hyde, J., Fleischmann, R, Chapman, M., and Gottesman, M. M. (1985)J. Biol. Chem. 260,13934-13940 16. Martin, R G., and Ames, B. N.(1961)J. Biol. Chem. 236,13721379 17. Doherty, P. J., Tsao, J., Schimmer, B. P., Mumby, M.C., and Beavo, I. A. (1981)Cold Spring Habor Gong Cell Proliferation, 8,211-225 18. Cabral, F., and Schatz, G. (1979)Methods EnzymoL 56G, 602613 19. Gottesman, M. M., and Sobel, M. E. (1980)CeU 19,449-455 20. Corbin, J. D., and Reimann, E. M. (1974)Methods Enzymol. 38, 287-290 21. Gilman, A. G. (1970)Proc. Natl. Acad. Sci. U. S. A. 67,305-312 22. Gottesman, M. M., and Cabral, F. (1981)Biochemistry 20,16591665 23. Walsh, D. A,, Ashby, C. D., Gonzalez, C., Calkins, D., Fischer, E. H., and Krebs, E. G. (1971)J. Biol. Chem. 246,1977-1985 24. Gottesman, M. M. (1983)Methods Enzyml. 99F,197-206 25. Rannels, S. R., and Corbin, J. D. (1980)J. Biol. Chem. 255, 7085-7088 26. Steinberg, R. A,, O’FarrelI, P.H., Friedrich, U., and Coffino, P. (1977)CeU 10,381-391 27. Doherty, P. J., Tsao, J., Schimmer, B. P., Mumby, M. C., and Beavo, J. A. (1982)J. Biol. Chem. 257,5877-5883 28. LeCam, A., Nicolas, J.-C., S i g h , T. J., Cabral, F., Pastan, I., and Gottesman, M. M. (1981)J. BWZ. Chem. 256,933-941