Induction of alkaline phosphatase in mouse L cells by overexpression of the catalytic subunit of cAMP-dependent protein kinase.

Mouse L929 cells were used to study the mechanism of cAMP induction of alkaline phosphatase (AP) activity. Following treatment with 200 microM 8-chlorophenylthio-cAMP (CPT-cAMP), alkaline phosphatase enzyme activity was observed to increase 80-fold after 24 h. The CPT-cAMP dose response of the alkaline phosphatase enzyme activity correlated well with the CPT-cAMP activation of cAMP-dependent protein kinase in L cells. A cDNA clone for the alkaline phosphatase was isolated and used to demonstrate a 10-fold increase in alkaline phosphatase mRNA levels after a 24-h treatment of L cells with CPT-cAMP. Increased mRNA levels were first detected 4-6 h, after CPT-cAMP treatment, and the level of alkaline phosphatase mRNA decreased rapidly after removal of CPT-cAMP. In vitro nuclear transcription studies showed that a 3-fold increase in alkaline phosphatase gene transcription was detectable 6 h after CPT treatment, and this increase was blocked by cycloheximide. In order to determine if the catalytic (C) subunit of cAMP-dependent protein kinase was able to mediate the induction of AP, L cells were transfected with expression vectors containing the metallothionein promoter and coding for the C alpha isoform of the catalytic subunit of cAMP-dependent protein kinase or for a catalytic subunit in which lysine 72 had been mutated to methionine (C alpha K72M). Zinc treatment of stably transfected cells expressing the wild-type C subunit showed an increase in protein kinase activity and an increase in AP activity. Zinc treatment of cells containing the mutant C subunit expression vector produced an increase in the amount of a protein which was recognized by C subunit antibodies on Western blots, but these cells showed no increase in protein kinase activity or in AP activity. We conclude that the C subunit is sufficient for transcriptional induction of the AP gene and that the phosphotransferase activity of the C subunit is required for this induction.

Mouse L929 cells were used to study the mechanism of CAMP induction of alkaline phosphatase (AP) activity. Following treatment with 200 WM &chlorophenylthio-CAMP (CPT-CAMP), alkaline phosphatase enzyme activity was observed to increase SO-fold after 24 h. The CPT-CAMP dose response of the alkaline phosphatase enzyme activity correlated well with the CPT-CAMP activation of CAMP-dependent protein kinase in L cells. A cDNA clone for the alkaline phosphatase was isolated and used to demonstrate a lo-fold increase in alkaline phosphatase mRNA levels after a 24-h treatment of L cells with CPT-CAMP. Increased mRNA levels were first detected 4-6 h, after CPT-CAMP treatment, and the level of alkaline phosphatase mRNA decreased rapidly after removal of CPT-CAMP. In vitro nuclear transcription studies showed that a 3fold increase in alkaline phosphatase gene transcription was detectable 6 h after CPT treatment, and this increase was blocked by cycloheximide. In order to determine if the catalytic (C) subunit of CAMP-dependent protein kinase was able to mediate the induction of AP, L cells were transfected with expression vectors containing the metallothionein promoter and coding for the Cal isoform of the catalytic subunit of CAMP-dependent protein kinase or for a catalytic subunit in which lysine 72 had been mutated to methionine (CcuK72M). Zinc treatment of stably transfected cells expressing the wild-type C subunit showed an increase in protein kinase activity and an increase in AP activity. Zinc treatment of cells containing the mutant C subunit expression vector produced an increase in the amount of a protein which was recognized by C subunit antibodies on Western blots, but these cells showed no increase in protein kinase activity or in AP activity. We conclude that the C subunit is sufficient for transcriptional induction of the AP gene and that the phosphotransferase activity of the C subunit is required for this induction.
A preponderance of experimental evidence suggests that the mechanisms by which CAMP regulates gene transcription are distinct in eukaryotic and prokaryotic organisms (Roesler et al., 1988 vator protein, which is itself a DNA-binding protein (Weber et al., 1982). While CAMP is bound to catabolite activator protein, it enhances transcriptional initiation by interacting with specific promoter sequences and RNA polymerase. In eukaryotes, however, the major receptor for CAMP is the regulatory (R)' subunit of CAMP-dependent protein kinase (Beebe and Corbin, 1986;Taylor et al., 1988). CAMP-dependent protein kinase exists as a catalytically inactive tetramer of two regulatory subunits and two catalytic (C) subunits. The binding of two molecules of CAMP to each R subunit results in the dissociation of the holoenzyme and release of catalytically active C subunit. The release of C subunit from the holoenzyme results in phosphorylation of serine and threonine residues in many cellular proteins. Although multiple isoforms of both the R and C subunits have been characterized, their functional significance is currently uncertain (Beebe and Corbin, 1986;Chrivia et al., 1988). The majority of evidence to date would support a model for CAMP regulation of eukaryotic gene transcription where the C subunit released from the inactive holoenzyme is translocated to the nucleus (Boney et al., 1983;Riabowol et al., 1988). There the C subunit is thought to phosphorylate proteins important in the regulation of gene transcription (Montminy and Bilezikjian, 1987;Grove et al., 1987). In most cases these phosphorylated proteins are themselves thought to be DNA-binding proteins (Mitchell and Tjian, 1985). The expression of many eukaryotic genes has been reported to be induced by CAMP, but one of the most dramatic inductions reported to date has been the induction of AP activity in mouse L cells (Firestone and Heath, 1981). Alkaline phosphatase is a membrane-bound glycoprotein that exists in several isoforms that can be experimentally distinguished by differences in antigenicity and inhibitor sensitivity (Gum and Raetz, 1985). Alkaline phosphatase activity in bone and other tissues is thought to play a role in extracellular phosphate metabolism (Stigbrand and Fishman, 1984). Dibutyryl-CAMP treatment of mouse L cells resulted in a 2000-fold increase in the specific activity of AP in cell extracts. Furthermore, in vitro translation of mRNA from cbntrol and CAMP-treated cells suggested that this induction was due to an increased abundance of the mRNA coding for AP (Firestone and Heath, 1981). Although the mechanism of induction of enzyme activity was not demonstrated clearly, this system offers several advantages for the analysis of CAMP-dependent protein kinase function in the regulation of gene expression.  (Uhler and McKnight, 1987) except that the BarnHI/ Sac1 fragment containing the 5' region of the mouse metallothionein promoter (Glanville et al., 1981) was replaced with the BamHI/SacI fragment of the murine mammary tumor virus promoter (Majors and Varmus, 1983). This expression vector was constructed in an attempt to generate a promoter that would respond to both zinc and glucocorticoids.
However, the induction by zinc was unaffected by the change in promoter structure, and the maximal glucocorticoid induction of Cal mRNA was only 3-fold over untreated cells. The Ca protein coding region is identical in this vector and the previously described pColEV (Uhler and McKnight, 1987). The pCwK72M expression vector was constructed by oligonucleotide mutagenesis (Zoller and Smith, 1984) using an oligonucleotide with the sequence GTCGAA-GATCATCATGGCGTA to mutaaenize the sense strand of the Sac11 PvuII fragment of pColEV subcloned into M13mp18. The entire So&j BgmI fragment was sequenced following mutagenesis and subcloned into pCaEV to generate pColK72M, which is identical to pCaEV except at nucleotide 390 of the published Co cDNA sequence where the codon AAG coding for lysine 72 has been changed to ATG coding for methionine, and at nucleotide 397 where the leucine codon has been changed from TTA to TTC to create a new TaoI recognition sequence.
Hence, the only difference in the proteins produced by pCuEV and pCaK72M is that the latter protein has a methionine at position 72 whereas the former retains the lysine at this position as in the wild-type Ca sequence (Uhler et al., 1986a). Alkaline Phosphatase Assay-For quantitation of AP enzyme activity, the medium was aspirated from lo-cm culture plates, and the plates were washed twice with cold Tris-buffered saline (0.01 M Tris (pH 7.3), 0.15 M NaCl).
The cells were scraped off into cold Trisbuffered saline and pelleted at 800 rpm for 5 min. The pellet was then resuspended in cold Tris-buffered saline and sonicated. This cell extract was assayed for AP activity usingp-nitrophenyl phosphate as the substrate as described (Firestone and Heath, 1981). Activity is reported as units/g total protein, where 1 unit represents 1 pmol of p-nitrophenyl phosphate hydrolyzed per min. All protein concentrations were determined spectrophotometrically using a dye binding assay (Bio-Rad).
J&use Assay-Cell pellets were resuspended in homogenization buffer (10 mM NaP04 (pH 7.0), 1 mM EDTA, 1 mM dithiothreitol, 250 mM sucrose) and the protein concentration adjusted to 2 mg/ml. Kinase assays were performed as described previously (Uhler and McKnight, 1987 for isolation of total RNA and oligo(dT)-cellulose for poly(A)RNA purification as described (Uhler et &.,'1986b). A cDNA library of 400,000 independent clones in XgtlO was constructed as described previously (Uhler et al., 1986b) and screened using an oligonucleotide (ATGATCTCACCATTTTTAGTACTGGCCATCG-GCACCTGCCTTACCAAC) corresponding to the first 48 coding nucleotides of the published mouse placental AP cDNA (Terao and Mintz, 1987 Blot Analysis-Cell extracts prepared as above for the kinase assay were boiled for 5 min after adding sample buffer (Uhler and McKnight, 1987). The denatured samples were cooled on ice, electrophoresed on a 10% SDS-PAGE gel (Laemmli, 1970), and transferred to nitrocellulose. Laboratories) were loaded onto a formaldehyde-l% agarose gel for Northern analysis as described previously (Uhler et al., 198613).

Time Course of CPT-CAMP
Induction of AP-We observed a dramatic increase in L cell alkaline phosphatase enzyme activity after treatment with the CAMP analogue CPT-CAMP (Fig. IA). This CAMP analogue was chosen for these studies because it has previously been shown to be more resistant to degradation by phosphodiesterases and is a more potent activator of CAMP-dependent protein kinase than other CAMP analogues such as dibutyryl CAMP and 8-bromo-CAMP (Miller et al., 1975). In addition, its metabolism does not generate butyrate, which has been shown in some systems to induce alkaline phosphatase (Gum et al., 1987 increase in enzyme activity was detected between 6 and 8 h after treatment.
An 80-fold increase was observed at 24 h after CPT-CAMP treatment from 0.14 unit/g total cellular protein to 10.6 units/g.
CPT-CAMP Activation of CAMP-dependent Protein Kinase and Induction of AP Actiuity-To assess the possible correlation between alkaline phosphatase induction and CAMPdependent protein kinase activation, dose-response curves for both alkaline phosphatase activity and CAMP-dependent protein kinase activity were generated after treatment with various concentrations of CPT-CAMP. For kinase activity determinations, L cells were harvested after 1 h of treatment since this time point represented peak activity of the kinase activity (data not shown). For alkaline phosphatase assays, the experiment was performed at 24 h of treatment when alkaline phosphatase activity was still linearly increasing, even at maximal CPT-CAMP concentration.
As shown in Fig. lB, there is a correlation between the degree of kinase activation and the induction of alkaline phosphatase enzyme activity in mouse L cells at higher levels of CPT-CAMP (>30 PM CPT-CAMP). At the lowest CPT-CAMP (10 pM) concentrations, the basal level of kinase activity represented 15% activation of total kinase whereas at the highest concentration (3 mM) more than 90% of the kinase was activated. Neither kinase activation nor alkaline phosphatase induction was saturated at 3 mM CPT-CAMP.
Concentrations of CPT-CAMP greater than 3 mM were difficult to obtain because of the low solubility of CPT-CAMP in DMEM. Down-regulation of CAMP-dependent Protein Kinase by CPT-CAMP-The kinase activities shown in Fig. 1B were determined after 1 h of CPT-CAMP treatment because long term treatment of L cells with CPT-CAMP caused a significant decrease in the total amount of cellular CAMP-dependent protein kinase present. It can be seen (Fig. 1C) that 3 mM CPT-CAMP treatment for 24 h caused a 45% reduction in the total amount of protein kinase activity when compared with the total kinase activity in control L cells. This decrease in kinase activity was dependent on the dose of CPT-CAMP used and was even more pronounced at 48 h than at 24 h. An analogous decrease in total CAMP-dependent protein kinase activity has also been seen in porcine LLC-PKi cells after elevation of cellular CAMP levels (Hemmings, 1986). The loss of activity in LLC-PKI was shown to be due to specific proteolysis of the C subunit and may represent one cellular mechanism for down-regulating the activated kinase. Isolation of an AP cDNA Clone from Mouse L Cells-In order to determine if the increase in alkaline phosphatase enzyme activity was because of an increase in the amount of mRNA coding for alkaline phosphatase, an AP cDNA clone was isolated. Using an oligonucleotide probe based on the first 48 nucleotides of the coding region of the published mouse placental alkaline phosphatase cDNA sequence (Terao and Mintz, 1987), we screened a cDNA library from CPT-CAMP-treated L cells and isolated two alkaline phosphatase cDNA clones after screening 400,000 recombinant phage. Sequencing of these cDNAs showed them to be 98% identical to the published placental cDNA except in the 5'-noncoding region. The first 92 nucleotides of the 5' end of the mouse L cell AP cDNA clone showed no homology to the placental cDNA (Fig. 2). However, within the region beginning at a position 105 bases upstream of the translational initiation codon and continuing through to the poly(A) tail, the mouse L cell and placental cDNAs showed nearly complete identity (data not shown). Sequencing of a large number of rat liver cDNA clones for alkaline phosphatase as well as the rat alkaline phosphatase gene has suggested that two independent promoters within the rat alkaline phosphatase gene are functional in rat liver (Toh et al., 1989). The two promoters are immediately upstream of exons 1 and 2 of the rat bone/liver/ kidney gene. As shown in Fig. 2 the first 92 nucleotides of the 5'-noncoding region in the mouse L cell cDNA which are not highly homologous to the mouse placental cDNA nucleotide sequence (37% identity) show a high degree of homology with the rat exon 2 (89% identity). Conversely, the mouse placental sequence shows a high degree of homology (85% identity) to rat exon 1 in this region. These nucleotide homologies and the fact that two independent L cell alkaline phosphatase cDNA clones contained the sequence homologous to rat exon 2 suggest that in mouse L cells the promoter for exon 2 is inducible with CAMP. The mouse L cell alkaline phosphatase cDNA was used for in the amount of radiolabeled cDNA probe hybridizing to this Northern blot analysis of poly(A) RNA from control and 2.5-kb band after CPT-CAMP treatment, and this level of CPT-CAMP treated mouse L cells (Fig. 3A). The Northern induction was confirmed independently by a solution hybridblot indicates a substantial induction of the 2.5kb alkaline ization assay (data not shown). Further analysis showed that phosphatase mRNA in the poly(A) RNA from CPT-CAMP-the induced level of alkaline phosphatase mRNA was sustreated cells compared with control cells. Densitometric scan-tained for at least 72 h of CPT-CAMP treatment (Fig. 3B), ning of the autoradiogram indicated at least a IO-fold increase that the increased level of alkaline phosphatase mRNA was first detectable 4-6 h after CPT-CAMP treatment (Fig. 3C), and that following removal of CPT-CAMP the level of alkaline phosphatase mRNA rapidly declined with a half-life of approximately 2 h (Fig. 30). Quantitation of amount of AP mRNA present in these RNA samples by a specific solution hybridization assays confirmed the results by Northern blot analysis and showed that even at maximal induction the AP mRNA represented less than 0.001% of the total mRNA present (data not shown).
CPT-CAMP Induction of AP Gene Transcription-The alkaline phosphatase cDNA was also used to determine whether this increase in mRNA was due to increased alkaline phosphatase gene transcription.
Nuclei were isolated from both control L cells and from L cells treated with 3 mM CPT-CAMP. The nascent RNA transcripts were elongated in vitro in the presence of [32P]UTP (McKnight and Palmiter, 1979). The labeled RNA was hybridized to pGEM-4 plasmid containing the alkaline phosphatase cDNA, a mouse y-actin cDNA fragment, or pGEM-4 alone. The mouse y-actin gene has been shown previously to be expressed at a low level in mouse L cells (Tokunaga et al., 1988). As shown in Table I, alkaline phosphatase gene transcription increased from 5.8 ppm in control L cells to 15.5 ppm after 6 h of CPT-CAMP treatment. In the same samples the y-actin transcription rate remained unchanged. In other experiments (data not shown) the basal AP gene transcription rate in L cells was as low as 3.2 ppm, and CPT-CAMP treatment resulted in a 5-fold induction of the AP gene transcription.
In addition, the increase of alkaline phosphatase transcription rate measured at 6 h was almost completely blocked by treatment of the L cells with cycloheximide ( Table I), suggesting that synthesis of an intermediate protein is required for alkaline phosphatase induction.
Stable Transfection of Mouse L cells with C Subunit Expression Vectors-To determine directly if CAMP-dependent protein kinase was involved in the transcriptional response of the AP gene to CAMP, L cells were co-transfected with an inducible expression vector for the Ca subunit of CAMPdependent protein kinase and an expression vector for the selectable marker, neomycin phosphotransferase.
The expression vector for the Co! subunit of CAMP-dependent protein kinase has been demonstrated previously to code for a catalytically active protein in mouse NIH/3T3 cells (Uhler and McKnight, 1987). G-418-resistant clones were isolated and characterized for their expression of the Cal subunit mRNA in the presence and absence of ZnSO+ L cell clones expressing a mutated form of the Ca subunit in which the lysine at position 72 was changed to methionine were also isolated by transfection with an alternate expression vector. This lysine residue has been demonstrated by affinity labeling experiments to be involved in ATP binding by the catalytic subunit (Zoller et al., 1979). In vitro mutations that change the analogous lysine to methionine in other protein kinases have shown to inactivate those protein kinases (Chou et al., 1987;Weinmaster et al., 1986;Snyder et al., 1985;Kamps and Sefton, 1986). In both transfections, clones that showed induced Ca subunit mRNA in response to ZnSO, treatment were characterized with respect to induction of kinase activity and AP activity. Although results for only one or two clones of each class of transfectant, normal or mutant C subunit, are shown, similar results were obtained for at least five clones of each class of transfectant.
Effects of Overexpression of C Subunit mRNA Levels on Kinase Activity-A study of ZnSOl induction of C subunit mRNA and kinase activity was performed by comparing the effect of ZnSOl on wild-type L cells, the CaB-transfected L cell clone containing the CLY subunit expression vector, and the CaK72M13 clone containing the mutant Ca subunit expression vector in which the lysine at position 72 has been changed to methionine.
Representative results are shown in Fig. 4. Fig. 4A shows the level of Ca subunit mRNA in these three different cell types in the absence or presence of ZnSO.+ Wild-type L cells contain approximately 50 molecules of Ccu mRNA per cell, and ZnS04 treatment has no effect on the level of Ca mRNA in these cells. The Ca2 cells, which have been transfected with the Ca expression vector, contain approximately 80 molecules of Ca mRNA per cell in the absence of ZnS04, but this increases to 400 molecules/cell after ZnS04 treatment. In the CaK72M13 clone, harboring the mutant Ca subunit expression vector, the Ca subunit mRNA level increases from 500 molecules/cell under basal conditions to 7000 molecules/cell after ZnS04 treatment. The effect of these Ca subunit mRNA levels on kinase activity in the three different cell types is shown in Fig. 4B. There is no effect of ZnSOl on kinase activity in wild-type L cells where endogenously active kinase activity is 50 units/ mg protein, and total kinase activity is 800 units/mg. ZnS04 treatment of the Ca2 cells, however, increases the endogenously active kinase activity 3-fold from 80 to 240 units/mg. This same treatment increases the total kinase from 900 units/mg in the basal state to 2100 units/mg after ZnSOl induction of transcription from the Ca expression vector. In the CaK72M13 cells, however, the endogenously active kinase activities are 50 units/mg protein in the absence or presence of ZnSO1, and total kinase activities are 450 units/mg protein independent of whether or not the cells have been treated with ZnSO+ Thus, even though the CaK72M13 cells are expressing a 140-fold excess of the mutant Ca mRNA, they show no increase in kinase activity. We conclude from these experiments that the C subunit containing the lysine to methionine change at position 72 possesses little if any kinase activity.
Effect of Increased C Subunit on AP Activity in L Cell Transfectants-The ability of each of these L cell clones to induce AP in response to ZnSOl and CPT-CAMP was tested. As shown in Fig. 5 units/g in the control cells to 4.4 units/g in the CPT-cAMPtreated cells. In contrast to the L cells, the Co2 cells responded to ZnSO., treatment with a 20-fold increase in AP activity from 0.06 to 1.2 units/g. These same cells showed an 80-fold increase in AP activity when cells treated with CPT-CAMP (5.0 units/g) are compared with untreated cells (0.06 units/ g). The CatK72M13 cells did not show an increase in AP activity with ZnSOl treatment (0.28 units/g in control cells and 0.16 units/g in ZnSO1-treated cells) but did respond to CPT-CAMP with a lo-fold increase in AP activity (3.3 units/ g in CPT-CAMP-treated cells). Northern Blot Analysis of AP mRNA Induced by ZnSO, and CAMP-In order to test the possibility that CPT-CAMP and ZnS04 treatment of Ca2 cells caused induction of mRNAs for the different isoforms of AP, Northern blot analysis of these cells was performed using conditions of high stringency. Poly(A) RNA from the C~y2 cells was isolated after no treatment or after ZnSOl or CPT-CAMP treatment and subjected to Northern blot analysis using the AP cDNA as radiolabeled probe. As is shown in Fig. 6, both the ZnS04 and CAMP treatments increased the level of 2.5-kb AP mRNA in these cells although CAMP treatment resulted in 2-3-fold higher levels of AP mRNA than ZnSOI treatment.
Measurement of Alkaline Phosphatase Gene Transcription Rates in Transfected L Cells-The AP gene transcription rates were measured in L cells and in C~2 cells treated with CPT-CAMP or ZnSOl to test whether the increase in AP mRNA levels was due to an increase in AP gene transcription (Fig.  7). Whereas CAMP treatment of both mouse L cells and Ca2 cells resulted in a 2.5-fold increase in gene transcription, ZnSOl treatment increased AP gene transcription only in the Ca2 cells. Thus, overexpression of the C subunit induced by ZnSOl treatment resulted in an increased rate of AP gene transcription.
Furthermore, since the basal transcription rate of the AP gene is low in L cells and near the level of sensitivity for this assay (approximately 0.5-2.0 ppm), the fold induction determined by this assay may be an underestimate of the actual degree of induction. Nuclei were prepared from confluent lo-cm tissue culture dishes that had been cultured in DMEM with 10% fetal calf serum in the presence or absence of 3 mM CPT-CAMP or 110 uM ZnSO1. The nuclei were incubated with ["'PIUTP and radiolabeled RNA transcripts isolated. The RNA was hybridized to nitrocellulose filters adsorbed with pGEM-4 DNA containing AP or no insert. Transcription rates were determined from the amount of "*P RNA specifically hybridized to each filter and expressed in ppm. Transcription rates for mouse L cells (white bars) and Ctv2 transfectants (black bars) are shown.
Western Blot Analysis of C and R Subunits of CAMPdependent Protein Kinase in Transfected L Cell-In order to assess if the CaK72M expression vector was able to produce C subunit protein in the stably transfected cells, cell extracts were prepared from Ca2 and CaK72M18, a cell line expressing levels of mutant Ccu subunit similar to those C~uK72M13 cells. These extracts were subjected to Western blot analysis to study the protein produced by the expression vectors (Fig. 8). ZnS04 treatment of Ca2 cells produced an increased level of C subunit that co-migrates on SDS-PAGE with the endoge- cells. Two hundred micrograms of protein cell extract from Co2 or ColK72M18 was electrophoresed on a 10% SDS-PAGE gel and transferred to nitrocellulose. Twenty nanograms of purified bovine heart C subunit was also run for size comparison. The Western blot was incubated with a polyclonal antibody against the C subunit followed by an AP-coupled second antibody as described under "Materials and Methods." After color development and identification of the C subunit, the blot was further incubated with a polyclonal antibody against the RI subunit.
nous C subunit in L cells. Quantitation of the C subunit present in CCY~ cell extracts using lz51-protein A (Uhler and McKnight, 1987) showed that ZnS04 treatment of Ccu2 cells produced a 3-fold increase in immunoreactive C subunit. There was also an increase in immunoreactivity of a band that co-migrates with the RI subunit. This is consistent with a similar compensatory increase in RI protein in response to C subunit overexpression which was described previously for transfected mouse NIH/3T3 cells (Uhler and McKnight, 1987). ZnSOl treatment of CaK72M13 or CcuK72M18 cells, however, caused an increase in the amount of a protein immunologically related to the C subunit but which migrated slightly faster in SDS-PAGE than endogenous C subunit. Quantitation of the C subunit induction in CaK72M18 cells showed an 8-lo-fold induction of total C subunit protein.
Since identical ZnS04 concentrations produced X-fold higher levels of Ca mRNA in &K72M cells than in Ccv2 cells (Fig.  4A), the CcuK72M protein can be estimated to be 5fold less stable than the wild-type C subunit. Although small increases in RI subunit were observed occasionally after ZnSOl treatment of CaK72M18 cells, the RI subunit compensation was never as large as that seen in Ca2 (Fig. 8).
Thus, cells expressing the CaK72M subunit did show induction of both Ca subunit mRNA and Ccu subunit protein as shown in Figs. 4A and 8, respectively, but this mutant protein lacks kinase activity (Fig. 4B). Furthermore, the mutant C subunit protein in the CaK72M13 cells appears to be incapable of inducing AP activity in response to ZnS04 (Fig. 5) although the endogenous Ca subunit gene product is still able to induce the AP gene in response to CPT-CAMP. DISCUSSION Our results demonstrate that the previously reported induction of alkaline phosphatase by dibutyryl CAMP can also be mimicked by CPT-CAMP. This lends support to the notion that induction of alkaline phosphatase occurs through a CAMP-dependent mechanism and not through the effect of butyrate. The observation that induction of alkaline phosphatase enzyme activity correlates with activation of CAMPdependent protein kinase suggested that effect of CAMP may be mediated by activation of the kinase. It is interesting to note that even at 3 mM CPT-CAMP the induction of alkaline phosphatase activity has not reached a maximum (Fig. 1B). This is in contrast to many other cellular responses to CAMP which are maximally induced at concentrations of CAMP which activate a much smaller fraction of the total cellular CAMP-dependent protein kinase. For example, in perfused liver, phosphorylase is maximally stimulated by epinephrine concentrations that activate only 30-35% of the total CAMPdependent protein kinase (Keely et al., 1975). Maximal stimulation of ACTH release occurs at concentrations of corticotropin releasing factor which activate only 50% of the type I kinase whereas type II kinase is not activated at all by corticotropin releasing factor (Litvin et al., 1984). It is possible that this difference in response to CAMP-dependent protein kinase activation between alkaline phosphatase induction and other cellular responses reflects differences in the kinase isoforms which mediate these various cellular CAMP responses or that these cellular functions occur in different cellular compartments.
The present finding that the cDNA for the alkaline phosphatase induced in mouse L cells is identical to the mouse placental alkaline phosphatase cDNA was not unexpected. Three isoforms of human alkaline phosphatase have been described: an intestinal isoform, a placental isoform, and a bone/liver/kidney isoform. In humans it is the bone/liver/ kidney isoform which is CAMP inducible.
Genetic evidence suggests, however, that only two forms of alkaline phosphatase exist in the mouse (Terao et al., 1988), a placental and an intestinal isoform. Furthermore, the placental cDNA was used to show that the placental isoform is expressed in tissues other than the placenta. Biochemical characterization of the potency of various peptide inhibitors has also suggested that in mouse the placental and L cell alkaline phosphatase are very similar (Gum and Raetz, 1983). Thus, in mouse it appears as if the bone/liver/kidney and placental alkaline phosphatases are encoded by the same gene and that this form of alkaline phosphatase is expressed in L cells.
It was reported previously that the amount of mRNA coding for alkaline phosphatase as determined by in vitro translation increases with dibutyryl CAMP treatment (Firestone and Heath, 1981). This is in complete agreement with present experiments, which determined that alkaline phosphatase mRNA levels increased at least IO-fold as determined by Northern blot hybridization. Furthermore, a modest (3-fold) and transient increase in alkaline phosphatase gene transcription was detected by in vitro nuclear transcription assay. It appears from data presented here that this CAMP induction of alkaline phosphatase gene transcription is a secondary event and that alkaline phosphatase may belong to a small group of genes that are known to be transcriptionally regulated at a secondary level by CAMP. Further study of alkaline phosphatase gene promoter structure and particularly the promoter for exon 2 may help to characterize the mechanism of secondary effects of CAMP on gene transcription.
It is interesting to note, however, that the rat alkaline phosphatase promoter region upstream of exon 2 does not contain a classic CAMP response element (Toh et al., 1989).
Although AP enzyme activity was observed to increase 80fold over 24 h of CPT-CAMP treatment (Fig. IA), the AP mRNA levels were observed to increase only lo-20-fold (Fig.  3A). Some of this discrepancy may be due to the inaccuracy of measuring the basal level of AP mRNA in uninduced L cells. The level of AP mRNA after maximal induction by CPT-CAMP was estimated to be 0.001% of the total mRNA by both RNA solution hybridization assay and the relative abundance of the AP cDNA in the cDNA library. However, it is possible that CAMP exerts some effect at the level of translation of AP mRNA or at the level of the AP protein stabilization.
A much larger discrepancy is seen when comparing the 3-fold induction of AP gene transcription with the IO-20-fold increase in AP mRNA. Here again, the basal level of AP gene transcription was difficult to determine exactly, and the actual fold induction by CAMP may be higher than that determined in our assay. However, in poly(A) RNA from control cells, we consistently observed low amounts of heterogeneous high molecular weight RNAs that hybridized to the AP cDNA (Fig. 3A). Furthermore, CPT-CAMP treatment consistently reduced the amount of hybridization to these high molecular weight RNAs. This observation might suggest that CPT-CAMP increases the formation of mature AP mRNA from unprocessed RNA transcripts.
A precedent for this type of regulation has been reported for the regulation of al-acid glycoprotein gene by glucocorticoids in a rat hepatoma cell line (Vannice et al., 1984). Alternatively, the apparent discrepancy between AP gene transcriptional induction and AP mRNA induction may represent stabilization of the AP mRNA by CPT-CAMP.
However, although other modes of regulation most probably affect AP gene expression, the results of in vitro transcription assays (Table I) demonstrate that CAMP does regulate transcription of the AP gene. Since the induction of overexpression of the CLY subunit in Ca2 cells by ZnSOl treatment was sufficient to induce AP enzyme activity and AP mRNA levels, the major effect of CAMP on AP gene expression appears to be to regulate the kinase activity of the C subunit. Several other Ca overexpressing cell lines were generated during the course of these experiments, and zinc induction of Co mRNA levels had similar effects on alkaline phosphatase activity, but the Ca2 cells were convenient for the studies presented here because of their low basal level of CLY mRNA expression.
Although overexpression of the Ca! subunit does lead to increased levels of R subunit protein, this R subunit results from an increased stability of the R subunit in the holoenzyme complex as compared with the free R subunit (Steinberg and Agard, 1981) and is not due to an increase in the mRNA for the R subunit (Uhler and McKnight, 1987). In addition, the compensating R subunit protein is complexed in an inactive holoenzyme complex with C subunit as determined by kinase activity (Uhler and McKnight, 1987). From the experimental data presented here, induction of alkaline phosphatase activity requires the kinase activity of the CLY subunit. Expression of a mutant Ca subunit in which the lysine residue at position 72 in the protein has been changed to methionine is incapable of inducing alkaline phosphatase activity. An analogous lysine residue has been shown by affinity labeling to be involved in ATP binding to the porcine C subunit (Zoller et al., 1979). This lysine is part of a sequence motif strictly conserved among protein kinases, and mutagenesis of this lysine residue has been shown to abolish kinase activity for other protein kinases including the human insulin receptor (Chou et al., 1987), the fps oncogene (Weinmaster et al., 1986), and v-src (Snyder et al., 1985;Kamps and Sefton, 1986). The present finding that the analogous mutation in the CLY subunit abolishes kinase activity is therefore not surprising.
However, the fact that this protein is unable to induce alkaline phosphatase activity in intact cells suggests that protein phosphorylation by the CCX subunit plays a central role in CAMP regulation of alkaline phosphatase gene expression. Several differences were noted between the CatK72M mutant protein and the wild-type Ca protein.
First, the CcuK72M mutant protein consistently migrated more rapidly than the wild-type protein on SDS-PAGE. Second, the CaK72M protein appeared to be less stable than the wild-type protein in that much higher levels of the &K72M mRNA as compared with wild-type Ca mRNA were required to see similar amounts of the two proteins (see Figs. 4A and 8). Finally, the &K72M mutant was not able to stabilize the RI subunit as well as the wild-type Ca subunit. Since the mutated lysine residue has been shown to play a role in ATP binding (Zoller et al., 1979), it is possible that autophosphorylation of the C subunit has been affected and that lack of autophosphorylation may in turn affect conformation as seen by altered migration in SDS-PAGE, proteolytic degradation, and holoenzyme formation of the Cal subunit. Alternatively, it may be the simple change in charge from positively charged lysine to neutral methionine which alters the rate of migration of the mutant protein.
The conclusion that phosphorylation by the C subunit of CAMP-dependent protein kinase is required for transcriptional regulation of the alkaline phosphatase gene by CAMP is consistent with recently reported results in other systems. An expression vector for the protein kinase inhibitor peptide has been shown to inhibit the CAMP stimulation of transcription of the human enkephalin promoter in a transient expression assay system (Grove et al., 1987), and C subunit expression is able to stimulate CAMP-responsive transcription in a similar transient expression assay system (Mellon et al., 1989). In addition, microinjection of the C subunit has been shown to stimulate transcription from the vasoactive intestinal peptide gene promoter and the c-fos promoter whereas microinjection of the R subunits did not stimulate transcription from these promoters (Riabowol et al., 1988). Although these studies clearly implicate the kinase activity of the C subunit in regulation of gene transcription, in neither of these cases was it possible to assay for CAMP-dependent protein kinase in the cell and directly correlate kinase activity with the transcriptional response. The mouse L cell alkaline phosphatase system presented here possesses some advantages over other gene transcription systems and has facilitated the demonstration that the kinase activity of the C subunit of CAMPdependent protein kinase is sufficient and necessary for transcriptional induction of the mouse L cell alkaline phosphatase gene.