Cyclic AMP Responses Are Suppressed in Mammalian Cells Expressing the Yeast Low K,,, CAMP-Phosphodiesterase Gene*

A genomic DNA fragment from Saccharomyces cere- visiae which contains the SRA5 (=PDE2) gene, coding for a low K,,, CAMP-phosphodiesterase, was transfected into Chinese hamster ovary cells. Clones carry- ing the CAMP-phosphodiesterase gene were capable of growth in the presence of cholera toxin, which slows the growth of untransfected cells by elevating their CAMP levels. The cholera toxin-resistant transfected cell lines expressed high levels of CAMP-phosphodies-terase mRNA and CAMP-phosphodiesterase activity. Basal intracellular CAMP levels were not significantly affected by the presence of the yeast CAMP-phospho- diesterase gene, but elevation of CAMP levels in response to cholera toxin or prostaglandin El was sup- pressed. Induction of the CAMP-responsive tyrosine aminotransferase promoter by cholera toxin was also blocked in cell lines carrying the yeast CAMP-phospho-diesterase gene. Cholera toxin-resistant transfected cell

A genomic DNA fragment from Saccharomyces cerevisiae which contains the SRA5 (=PDE2) gene, coding for a low K,,, CAMP-phosphodiesterase, was transfected into Chinese hamster ovary cells. Clones carrying the CAMP-phosphodiesterase gene were capable of growth in the presence of cholera toxin, which slows the growth of untransfected cells by elevating their CAMP levels. The cholera toxin-resistant transfected cell lines expressed high levels of CAMP-phosphodiesterase mRNA and CAMP-phosphodiesterase activity. Basal intracellular CAMP levels were not significantly affected by the presence of the yeast CAMP-phosphodiesterase gene, but elevation of CAMP levels in response to cholera toxin or prostaglandin El was suppressed.
Induction of the CAMP-responsive tyrosine aminotransferase promoter by cholera toxin was also blocked in cell lines carrying the yeast CAMP-phosphodiesterase gene. Cholera toxin-resistant transfected cell lines were sensitive to the growth inhibitory effects of Ne,02'-dibutyryladenosine 3',5'-monophosphate, which can be used to bypass the effects of the yeast CAMP-phosphodiesterase.
Many hormones exert their intracellular effects by binding to cell surface receptors and, through the intermediary actions of GTP-binding proteins, cause an effector enzyme to increase synthesis of a second messenger molecule. In many cases, the effector enzyme is adenylate cyclase and the intracellular second messenger is CAMP. The conclusion that a particular hormone requires an intracellular increase in CAMP to exert its effect is often based on the following correlation: if CAMP concentrations rise after application of the hormone and if a variety of other agents, which include cholera toxin, pertussis toxin, forskolin, or phosphodiesterase inhibitors, cause an effect similar to that caused by the hormone, a case has been made that the intracellular second messenger is CAMP. There are several difficulties with this approach. Forskolin and phosphodiesterase inhibitors are nonspecific; the former affects ion channels (1,2) and the latter, because they are often purine derivatives, may react with adenosine receptors or other intracellular targets. ADP-ribosylating toxins, such as the cholera and pertussis toxins, may react with more than one type of GTP-binding protein, or their GTP-binding pro- tein substrate may couple to different second messenger systems. For these reasons, proof that elevation of CAMP concentrations is a crucial step in a hormonal induction should, in addition to the pharmacological criteria described above, be based on studies of mutants which cannot respond to CAMP. Such mutants exist in Chinese hamster ovary cells, S49, PC12, and a limited number of other cells, but are difficult to isolate (see Ref. 3 for review). We have devised genetic means to suppress the rise in CAMP levels caused by cholera toxin and hormones. The method can be used with virtually any cell and has the further advantage that suppression of the CAMP second messenger response can be overcome by addition of certain activators of the CAMP-dependent protein kinase, thus creating an experimental situation similar to that offered by a conditional mutant. Developmental sequences for which the agonists are not known may also be examined to determine whether an increase in CAMP is essential for cellular differentiation.
Finally, the genetic approach should work in uiuo, where pharmacological approaches are limited.
To control CAMP levels we have introduced an exogenous CAMP-phosphodiesterase gene into mammalian cells. The choice of genes was limited to the eukaryotic phosphodiesterase genes that have been cloned and sequenced. These include genes from Drosophila melanogaster (4), Dictyostelium discoideum (5), Saccharomyces cereuisiae (6)(7)(8), and recently mammalian cells (9)(10)(11). The yeast low K,,, CAMP-phosphodiesterase was the most suitable for introduction into mammalian cells because it has a low K,,,, it is CAMP specific, and its activity is not modulated by calmodulin or cGMP (12, 13). The gene coding for the yeast CAMP-phosphodiesterase was cloned by complementation of a mutation that suppresses the effects of an activated RAS.2 gene and was named PDE2 (6) or SRA5 (7)  which has the yeast SRA5 (=PDE2) gene inserted into the unique BamHI site of pDOP, as shown in Fig. 1 panels C and D were shown by DNA slot blot analysis to contain the yeast CAMP-phosphodiesterase gene (data not shown). Cholera toxin slows the growth of CHO cells enough to distinguish resistant colonies growing on plastic or in soft agar. Growth of strain 10248, which has an altered regulatory subunit of the CAMP-dependent protein kinase was unaffected by cholera toxin. As predicted, the growth of CAMPphosphodiesterase containing clones was unaffected by cholera toxin. The morphological effect of cholera toxin, which causes the cells to become spindle shaped and clustered, is also blocked in the CHO-PDE cell lines as shown in Fig. 3E. Transcription of the Yeast PDE2/SRA5 Gene-The yeast CAMP-phosphodiesterase gene is transcribed in these cholera toxin resistant clones. Fig. 4 shows a Northern blot of 20 pg of total RNA extracted from the untransfected parental cell line CHO 10001, a cell line transfected with the pSV2neo control plasmid (CHO-Cl), and the three cell lines containing the yeast CAMP-phosphodiesterase gene (CHO-PDE4, CHO-PDES, and CHO-PDElO).
The blot was probed with the 0.5kb BamHI-EcoRI fragment from the yeast SRA5 gene in pDOPSRA5.
Only the CHO-PDE cell lines carrying the SRA5 gene produced detectable yeast CAMP-phosphodiesterase transcripts of the predicted size. One transcript extended from an initiation site in the 5' long terminal repeat (LTR) of murine sarcoma virus to the polyadenylation site of the SRA5 gene (approximately 3.4 kb), the other extended from the 5' LTR to the polyadenylation site in the 3' LTR (approximately 6.7 kb). CHO-PDE cell lines 11-16 were also tested and contained the SRAS transcripts (data not shown). Detection of the Yeast Enzyme by Kinetic Analysis-The yeast CAMP-phosphodiesterase is easily detected in enzyme assays because of its 170 nM K,,,. Mammalian phosphodiesterases have a range of K,,, values between 2-fold and 500-fold higher (32). Extracts of CHO cells were assayed at low CAMP concentrations (below the K, of the yeast enzyme) where the yeast enzyme is relatively more efficient than the endogenous and CHO-PDE16 had respectively 8. 2-, I.&, 5.3-, 10.6-, 7.9-, 8%, 10.7-, and 9.9-fold more CAMPphosphodiesterase activity than control CHO-C3 or CHO-Cl cells when measured at a low substrate concentration (45 nM). Furthermore, kinetic analysis using different CAMP concentrations showed that the component of the CAMP-phosphodiesterase activity unique to the extracts from the pDOP-SRA5 transfected cells had an apparent K, of 100-300 nM (Fig. 5). This is close to the 170 nM value reported for the yeast low K,,, CAMP-phosphodiesterase (12). As the CAMP levels are increased while the amount of enzyme is kept constant, the fraction of total CAMP hydrolyzed decreases, as shown in Fig. 5. However, even in the substrate range of l-10 pM, where the contribution of endogenous phosphodiesterases to the CAMP-hydrolyzing activity is increased (due to their higher K,,,), the majority of the hydrolytic activity is due to the yeast CAMP-phosphodiesterase.  CAMP levels in pDOP control transfected (CHO-C3) cells. Cells incubated for 6 h in the presence of 100 rig/ml cholera toxin had a 36-fold increase in CAMP levels ( Table I). The CHO-PDE12 cell line showed no increase in CAMP levels when grown in the presence of cholera toxin for 6 h ( Table I). The reduction of cholera toxin-stimulated CAMP levels is consistent with the cholera toxin-resistant phenotype of the CHO-PDE cell lines and presumably results from the rapid conversion of CAMP to 5'-AMP because of the presence of the yeast low K,,, CAMP-phosphodiesterase activity. Elevation of CAMP levels by prostaglandin E1 were also diminished in the CHO-PDE cell lines (see below). Despite the presence of the yeast low K, CAMP-phosphodiesterase activity in CHO-PDE cell lines, the basal CAMP levels of these cell lines were not significantly different from the control cell lines (Table I). In yeast, CAMP levels are regulated over a large concentration range by a negative feedback loop which includes the catalytic subunit of the CAMP-dependent protein kinase, the CDC25, RASl, and RAS2 gene products, and adenylate cyclase (33). The maintenance of basal CAMP levels in CHO cells expressing large amounts of CAMP-phosphodiesterase might be caused by a similar compensatory mechanism which activates the endogenous adenylate cyclase. To assay adenylate cyclase activity in mammalian cell extracts it is necessary to include a phosphodiesterase inhibitor such as IBMX in the reaction mixture. The yeast enzyme is relatively insensitive to the phosphodiesterase inhibitors IBMX and RO 20-1724 (I(& > 300 pM at 25 nM CAMP; data not shown). Therefore, an excess of nonradioactive CAMP is required in the reaction mixture for the CHO-PDE cell lines to quench the yeast CAMP-phosphodiesterase activity and prevent hydrolysis of the radioactive reaction product. Basal levels of adenylate cyclase or activation of the membrane adenylate cyclase by pretreatment of cells with cholera toxin in vivo, or activation in vitro by NaF, GTP, or GTP+ showed no significant difference between control and SRA5 expressing cell lines (data not shown), indicating that the CHO-PDE cells do not have an activated adenylate cyclase to compensate for the high phosphodiesterase activity.

Control of Cholera
Prostaglundin E1 Effects-One potential use of the technique developed here is to block the CAMP-mediated effects of hormones. PGE, causes a rapid rise in CAMP levels in CHO cells (Fig. 6) on the experiment, peaking around 3 min (Fig. 6A). In the CHO-PDE12 cell line, CAMP levels were stimulated only 2-4-fold by 10 pM PGEi, peaking at about 1 min (Fig. 6A). Integration of the two curves in Fig. 6A revealed that approximately 70% of the CAMP response of the CHO-C3 control cell line was eliminated in the CHO-PDE12 cell line. When CAMP levels were measured 1 min after stimulation with different PGE, concentrations, CHO-C3 cells and CHO-PDE12 cells required equivalent amounts of PGEl (about 0.3 pM) to induce half-maximal activation of CAMP levels (Fig.  6B). This is what one would expect to occur in the presence of large amounts of CAMP-phosphodiesterase, when receptor binding, coupling to G-proteins, and activation of the adenylate cyclase are unaffected.
Reversibility of the Effects of the SRA5 Gene-The method used here blocks the CAMP signal transduction system at a step before CAMP can activate the CAMP-dependent protein kinase. Because this molecule is left intact, it is possible to activate the CAMP-dependent protein kinase in cells that contain the yeast CAMP-phosphodiesterase by adding cellpermeant CAMP analogues that are not hydrolyzed but which bind to and activate the CAMP-dependent protein kinase. N6,02'-dibutyryladenosine 3',5'-monophosphate (N6,02'-db-CAMP) is one such analogue. At 1 mM, N6,02'-db-CAMP inhibits growth of both control CHO-C3 and CHO-PDE12 cells by about 60%. Furthermore, the cell shape changes that are induced by cholera toxin in CHO-C3 cells but not in CHO-PDEl2 cells, can be seen in both cell lines after treatment with 1 mM N6,02'-db-CAMP, as shown in Fig. 3 (compare panels C and F). A detailed pharmacological analysis of the yeast low K,,, CAMP-phosphodiesterase is described in the accompanying paper (39). Effects on a CAMP Inducible Promoter-The 3.0-kb upstream segment of the CAMP-inducible TAT promoter (27,28) was cloned in the correct and reverse orientations into the pSV,UMS-CAT reporter plasmid (pTAT-CAT and pTAT,-CAT, respectively) and transiently transfected into CHO-Cl, CHO-PDElO, and CHO-PDE12 cell lines. Induction of TAT promoter activity by cholera toxin (100 rig/ml) was seen as a 2.0-2.5-fold increase in CAT activity in the control cell line CHO-Cl, which does not express the yeast CAMP-phosphodiesterase activity (Fig. 7). In cell lines CHO-PDElO and CHO-PDE12, in which the cholera toxin-induced increase in CAMP levels is ablated due to expression of the yeast CAMP-phosphodiesterase, there was no increase in TAT promoter activity as assessed by CAT activity (Fig. 7). Transfection of the pTAT,-CAT plasmid containing the TAT promoter in the reverse orientation resulted in no detectable CAT activity in any of the cell lines. Thus, the inability of the cells expressing the yeast CAMP-phosphodiesterase to increase CAMP levels affects gene induction as well as growth and morphology. DISCUSSION CHO cells transfected with the vector pDOPSRA5 produce a substantial amount of active yeast CAMP-phosphodiesterase, which prevents CAMP levels from accumulating upon stimulation of the adenylate cyclase. Expression of the yeast CAMP-phosphodiesterase activity in the CHO-PDE cell lines permitted normal growth and morphology in the presence of cholera toxin. The rise in intracellular CAMP levels that occurs after stimulation with PGEl was also limited. Activation of the CAMP-stimulatable tyrosine aminotransferase promoter by cholera toxin was prevented in CHO-PDE cell lines. The effects of the enzyme can be overcome with CAMP analogues that are not hydrolyzed but which activate the CAMP-dependent protein kinase, creating an experimental situation similar to that provided by a conditional mutant. It is not clear why basal CAMP levels are unaffected by expression of yeast CAMP-phosphodiesterase in CHO cells. CHO cells expressing the yeast CAMP-phosphodiesterase do not seem to compensate for this activity by activating adenylate cyclase. One possibility is that there is a CAMP pool that is sequestered in a compartment not accessible to the CAMP-phosphodiesterase (such as a pool bound to the regulatory subunit of the CAMP-dependent protein kinase). The cholera toxin or PGEi-sensitive pool of CAMP is apparently accessible to the yeast CAMP-phosphodiesterase. These results imply that basal CAMP levels do not reflect phosphodiesterase activity in CHO cells.
The ability to elevate CAMP levels in eukaryotic cells by pharmacological means has aided in understanding the elements of the second messenger cascade and has led to clinical applications, ranging from grafting of cholera toxin-treated skin allografts (34) to the treatment of cardiac failure with phosphodiesterase inhibitors (35). There are experimental situations in which it would be useful to be able to reduce CAMP levels. However, few strategies to reduce CAMP levels or to antagonize the effects of CAMP on the CAMP-dependent protein kinase have been developed. One method that has been used is the introduction of genes coding for mutant regulatory subunits of the CAMP-dependent protein kinase (18,36). The resultant regulatory subunits bind CAMP with greatly reduced affinity but still bind the catalytic subunit. In another approach, sequences of the inhibitor protein of the catalytic subunit of the CAMP-dependent protein kinase have been introduced transiently into cells and the induction of cotransfected CAMP-responsive reporter genes by CAMP analogues has been reduced (37,38). It has not yet been possible to create permanent cell lines which express an excess of protein kinase inhibitor, the effects of which are irreversible (38).
The elevation in CAMP levels by cholera toxin is chronic and is known to block cell growth by activation of the CAMPdependent protein kinase. In those clones that we have analyzed, there is enough phosphodiesterase activity to reduce CAMP levels so that the cells can grow. A mutated CAMPdependent protein kinase would give the same phenotype (as shown in Fig. 2). However, in these transfected lines we know that this enzyme is intact. First, the frequency of the transfectants is too high to be accounted for by mutagenesis. Second, N6-0"-db-CAMP, which is a poor substrate for the yeast enzyme, causes the CHO-PDE cells to assume a morphology that is identical to CHO-C control cells treated with cholera toxin or CAMP analogues, which could only happen if the CAMP-dependent protein kinase can be activated. The CAMP response to PGEl is fast and transient.
Even under these conditions the yeast CAMP-phosphodiesterase expressed in the CHO-PDE12 cells is able to substantially reduce the CAMP response to PGE1. About 70% of the normal CAMP increase of control cells is eliminated.
Whether this is sufficient to completely block the activation of the CAMPdependent protein kinase by PGEl is not known. In CHO cells PGEl produces no effects on growth or morphology, so that these events cannot be monitored at a phenotypic level as they can be with cholera toxin.
The pDOP-SRA5 vector provides a transferable genetic tool to suppress the effects of adenylate cyclase stimulation in eukaryotic cells in a pharmacologically reversible manner. Cells expressing high amounts of CAMP-phosphodiesterase can be used for studying the role of CAMP in the control of cell growth, differentiation, or metastasis in uitro. The gene should continue to function in uiuo if cells are placed into syngeneic or immunologically incompetent animals, offering experimental possibilities currently unavailable with a strictly pharmacological approach. One possibility is that the expression of the yeast CAMP-phosphodiesterase gene can be placed under the control of specific promoters and limited to particular target tissues in transgenic animals.
to Dr. J. E. Schwartzbauer for the pDOP plasmid, and to Dr. R. Kahn for help with the adenylate cyclase assays.