Effects of 6-and 8-substituted Analogs of Adenosine 3’: S-Monophosphate on Phosphoenolpyruvate Carboxykinase and Tyrosine Aminotransf erase in Hepatoma Cell Cultures*

A variety of 6. and 8-substituted analogs of CAMP (cyclic adenosine 3’:5’-monophosphate) have been tested for their ability to increase the activity of tyrosine aminotransferase (EC 2.6.1.5) in cultured Reuber H35 hepatoma cells.

Some analogs, particularly the 8-thio-substituted ones, produced effects approximately equivalent to those generated by NO,@'-dibutyryl CAMP. In contrast, CAMP and its 02'-monobutyryl derivative were relatively ineffective even at very high concentrations, whereas three other analogs actually depressed the activity of the aminotransferase. Changes in enzyme activity generated by the various analogs were paralleled closely by changes in the relative rate of aminotransferase synthesis.
An excellent correlation was found to exist between the ability of any given analog to influence the activity of tyrosine aminotransferase and that of phosphoenolpyruvate carboxykinase (EC 4.1.1.32).
A similar correlation was found to exist between the ability of various analogs to elevate the activity of these enzymes and to inhibit reversibly the growth of H35 cells. Only one of five inhibitors of CAMP phosphodiesterase activity tested produced any increase in aminotransferase activity when added alone. All of the 6-and 8-substituted analogs tested, including noninducers, stimulated fl histone phosphorylation in crude rat liver extracts with approximately equal potencies. On the other hand, dibutyryl CAMP was only a weak activator of protein kinase in vitro, even though it is a potent enzyme inducer.
A possible resolution of this apparent discrepancy has been provided by preliminary analyses of site-specific fl histone phosphorylation in whole cells. Only compounds active as aminotransferase inducers are capable of stimulating phosphorylation of the serine-37 residue of endogenous f1 histone (3-to lo-fold). * This work was supported by Grants AM 16753, AM 53197 and GM 01983 and by a grant from the American Diabetes Association.
A Dortion of this work was presented in preliminary form at the Fifth Annual Miami Winter Symposium, January 1973. This communication is Paner IV of a series in which the preceding paper is Ref. 30. Previous reports from this laboratory have described the stimulatory effect of N6,02'-dibutyryl cyclic adenosine 3': 5'.monophosphate (BtzchMP) on the synthesis of tyrosine aminotransferase (EC 2.6.1.5) and phosphoenolpyruvate carboxykinase (EC 4.1.1.32) in cultured rat hepatoma cells (l-4). The evidence t,o date in this system suggests a post-transcriptional site of action of t'lie cyclic nucleotide, perhaps at the polysomal level (Z-4).
One of the questions we would like to address is whether a CAMP'-dependent protein kinase plays a central role in this process.
As one means of exploring this possibility, we have made use of a variety of new analogs of CAMP for studies with intact cells.
If a protein kinase does mediate the effects of CAMP analogs on enzyme induction,l then one would expect, for example, to find a good correlation between the effects of any given analog on the activity of the aminotransferase and that of the carboxykinase.
In searching for nonbutyrylated analogs of CAMP that might be active as enzyme inducers, we turned to the 6-and 8-substituted derivatives prepared recently by several groups (7-10). A number of these compounds have been found to be at least equipotent to P&cAMP in several biological systems (7-14). Most siguihmt is t,he finding that many of these analogs are capable of act,ivating protein kinase without apparent structural modification (8,9,13,14), in contrast to the dibutyryl derivative (15)(16)(17)(18).
In addition, many of the compounds are resistant to hydrolysis by CAMP phosphodiesterase and several are potent inhibitors of this enzyme (8,9,13,19). It has proven difficult to establish the involvement of protein kinase in more than a small number of CAMP-mediated regulatory processes (20)(21)(22).
Evidence for or against the participation of this enzyme in enzyme induction under clearly physiological conditions, therefore, would be highly desirable.
As will be seen, the use of a variety of CAMP analogs with diverse abilities to 1 The abbreviations used are: CAMP, cyclic adenosine 3':5'monophosphate; BtacAMP, N6,0z'-dibutyryl cyclic adenosine 3':5'-monophosphate; cIMP, cGMP, and cCMP, the cyclic 3':5'-mimic the effects of BtsoAMP has provided a means of addressing the question of protein kinase involvement in enzyme induction. The details of these procedures have been extensively described previously (5, 6). Radioactivity in total soluble protein was determined in 25-pl samples of the supernatant fraction by the filter paper disc procedure of Mans and Novelli (27), which includes incubation of the discs with 5y0 trichloroacetic acid at 85".

Growth Studies
Plastic flasks (25 cmz) were inoculated at a density of 0.3 to 0.5 X lo6 cells.
The analogs were first added 24 hours after subculture and they were added again 48 hours later with a change of medium.
The next day (4 days after subculture, just past the mid-log phase), the cells were harvested as described above and lysed with 0.15 M KCl-1 mM EUTA containing 0.1% Non-idet P40. After centrifugation at 5000 x g for 10 min, the pellets were suspended in 0.1 N NaOH and assayed for their DNA content by the modified diphenylamine procedure (28 1. Time cum-se of effects of cAMP analogs on tyrosine aminotransferase activity in H35 cells. All analogs were added at 0.5 mM and cells were harvested for assay at the times indicated, as described under "Exuerimental Procedures." Each value represents the average of four to six observations, with standard errors of 5 to 10%. The maximal-fold increase (i.e. 100% of the maximal response) in aminotransferase activity observed with each analog is given in brackets below. A : l --;o, BttcAMP trichloracetic acid was used in the fi histone experiments. All other biochemicals used were obtained from Sigma Chemical Co.

Time Course oj E$ects of CAMP Analogs on Tyrosine Amino-
transferuse Activity-We have previously reported that Bt2cAMP causes a rapid, nearly linear increase in tyrosine aminotransferase activity in cultured H35 cells (l-4).
-4s shown in Fig. lA, careful analysis has revealed that a slight lag period actually appears to precede full development of the response to the dibutyryl analogs, which is eliminated when Na-monobutyryl CAMP is employed.
In contrast to these active butyrylated analogs, O2monobutyryl c,4MP is totally ineffective at 0.5 mM, although significant activity can be detected at 2 rnbr (Table 1). Similarly, cAMP itself is only marginally effective and only at very high concentrations. Theophylline, a phosphodiesterase inhibitor (8), does potentiate this effect to a modest extent, suggesting that it may be truly ascribed to a CAMP-mediated phenomenon.
A reduced response is actually obtained with the highest concent'ration of CAMP tested (20 mM). Among other cyclic nucleotides tested, only cIMP had a significant effect (although theophylline did not appreciably enhance its effects).
The other cyclic nucleotides, including cGMP and its 8Br derivat,ive, proved to be essentially inactive at the concentration employed.
Among the new synthetic derivatives of CAMP tested in this system, the 8-thio-substituted analogs are part,icularly effective, as shown in Fig. IB. The kinetics of the response of tyrosine aminotransferase to 8MeS-CAMP and 8EtS-CAMP is essent,ially idemical with that wit,h NG-monobutyryl CAMP.
The response of the aminotransferase after exposure to 8PhCH&cAMP, on the other hand, exhibited a brief lag period similar to that observed with P&CAMP. In contrast to the dibutyryl analog, the effects of the three 8-thio-substituted analogs and Ne-monobutyryl CAMP wore off more rapidly, particularly in the case of 8PhCH&CAMP.
The dose of the iV6-monobutyryl analog used 233 The various additions were made, and 5 hours later the cells were harvested for assays, except those exposed to 10 and 20 mM CAMP and O2 -monobutyryl CAMP, which were harvested at 3 hours.  (2) 35 (2)   0 (7) 40 (6) 10 (5)  0 (7) 13 (5) a The values shown are the average percentage increase in aminotransferase activity observed in one or two experiments, with the number of separate observations in parentheses.
in this study-was suboptimal (see Fig. 2), and with higher concentrations the effect persists for a longer period of time. From previous studies, it is clear t,hat H35 cells are capable of metabolizing both BtzcAMP and Na-monobutyryl CAMP (30), which presumably accounts for the eventual decline in the response. The maximal elevation of enzyme activity with all of the analogs was found to occur at 3 to 4 hours.
This interval is somewhat less than that expected for achievement of the maximal increase in aminotransferase activity if its rate of synthesis had been increased at time zero or shortly thereafter (31), since this enzyme has a half-life of about 15-i hours (12). A true steady state con&ion was not achieved with most of the analogs employed.
A decline in elevated enzyme activity occurred after 3 to 4 hours, for reasons as yet unknown.
Increasing the conccntration of analog has not been found to alter the time of peak enzyme response or its duration when this has been tested.
Dose-Response Relationships- Fig.   2 illustJratjes bhe dose-response relationships among the effective analogs for elevating tyrosine aminotransferase activity (measured at 4 hours). The three 8thio-substituted analogs exhibit'ed identical dose-response curves, which have been combined for simplicity.
Btz-CAMP is comparably potent (l), but, the N6-monobutyryl and 8Rr analogs proved to be considerably less so. On the other hand, BMeS-CAMP possesses a potency only slightly lower than the 8.thio analogs.
It should be noted t,hat minimal effect#s can be observed with the most potent analogs, including B&CAMP (l), at external concentrations of 5 to 10 CIM. The intracellular concentration range expected to result is thus well within that of CAMP generated in rat liver by concentrations of glucagon that produce significant metabolic responses (33 to be 5 times more potent than BtzcAMl' and 8-thio-substitut,ed analogs. Rate of Synthesis Measurements-BtzcAMI' has been shown to produce its effect' on tyrosine aminotransferase activity by increasing the relative rat'e of synthesis ol" this enzyme in organ culture (5), adult rat liver (6), and II35 cells (1). It was of interest, therefore, to determine whether the other active analogs of cAMP produced their effects by t.he same mechanism.
(We have not tested all of t,he analogs in t'his regard, but the results to date leave little doubt that different mechanisms are operative.) An excellent correlation was folmd to exist between the degree of effect of each analog test)ed on enzymatic activity and on the relative rate of enzyme synthesis (Table II).
Two of the analogs t,est,ed a&ally depressed the activity of the enzyme and also depressed incorporation of [3H]leucine into both the aminotransfemsc and total soluble protein.
These results suggest that the active analogs owe their ability to regulate the activity of this enzyme to the same mechanism as that influenced by BtscAMP.
Possible changes in the rate of enzyme degradation produced by by these analogs have not been tested, but 13tzcAhII? has been shown to be wit,hout detectable effect on this process (1). Correlation of Analog E$ects on Tyrosim Aminotransferase and P-Enolpyruvate Carboxykinase-U&CAMP has been shown to induce both tyrosine aminotransferase and I'-enolpyruvate carboxykinase in adult rat Iiver (6, 34) and in H35 cells (1, 2). In addition, hormones which act by way of promoting cAMY synthesis produce similar increases in the activit,y of both enzymes in fetal and adult rat liver (5, 6j.
In an effort to -determine whether induction of these two enzymes is mediated by a common intracellular factor, we were interested in determining whether analogs active in elevating aminotransferase activity are also capable of influencing carboxykinase activity to a comparabie extent.
With the use of a constant conccntrhon and a fixed time in-tervaJ3 it was found that an excellent correlation does exist' be-  and P-enolpyruvate carboxykinase activities in H35 cells. The analogs were all added at 0.5 mM and cells were harvested for assay at, 5 hours, as described under "Experimental Procedures." Each value is the mean of at least three observations (on the average, eight observations), with standard errors of 10 to ZO"/,. The average increase in activity generaled by Bt2cAMP was 2%fold for the aminotransferase and 2.0-fold for the carboxykinase.
The values for Analogs 1, 2, and 3 are given as percentage of decrease in control activity.
tween the ability of any given analog to influence tyrosine aminotransferase activity and that of Y-enolpyruvate carboxykinsse (Fig. 3). Although this relationship has not' been extensively studied under a wide variety of conditions, no deviation from the observed correlation has ever been observed. The carrelation coefficient for the relationship in Fig. 3 is 0.98, with a p value of tO.OO1.
As are seen below, t)he dose-response relationships for increases in the act,ivity of bot,h enzymes generated by BtncAMP and 8MeS-CAMP are also essentially identical. In addition, the other cyclic nucleotides tested (see Table I), which had either no effect or caused only modest increases in the activity of the aminotransferase, produced comparable respouses in carboxykinase activity. Effects of Inhiibilo~s of CAMP Phosphodiesterase-One can envision at least two possible mechanisms by which the 6-and g-substituted analogs of CAMP could exert their effects. Some of these analogs have been reported to inhibit the activity of CAMP phosphodiesterase (8,9,13) and could thus produce an increase in intracellular CAMP which, in turn, could bring about the observed changes in enzyme activities.
It is equally possible that they act directly by activating protein kinase in these cells (or interact with a specific binding protein, if protein kinase is steady state after exposure to CAMP analogs for 5 to 6 hours (1,3). Even though the response of the aminotransferase to some of the analogs begins to drop off by 5 hours (see Fig. lB), this difference does not appreciably alter the observed correlat,ion other than to underestimate slightly the effect of 8PhCh&cAMP on the aminotransferase (see Figs. 1 and 3). Correetion for this small difference serves only to improve the observed correlation. not the mediator of CAMP action on enzyme induction).
As an indirect test of the first possibility, a number of known inhibitors of phosphodiesterase activity have been tested for their ability to elevate tyrusine aminotransferase activity _ Theophylline is capable of modestly elevating the activity of the aminotransferase in organ culture (35) and adult liver (5) and it' pot,ent,iates the effects of hormones and CAMP in these systems (5, 35), but it has little or no effect in H35 cells when added alone (Table III), as reported previously (I).
It is able, however, to potentiate weakly the effects of exogenous CAMP, as seen in Table 1. Among other inhibitors tested, most u-ere totally inactive, but RO-20-1724 (11, 13, 36) had a consistently modest effect on aminotransferase activity. None of these compounds, however, was capable of generating the degree of increase in enzymatic activity observed with any of the act'ive cAMP analogs, even when concent,ra.tions great,er than t,hose indicat.ed were employed.
In addition, preliminary experiments have shown that an active analog (8PhCH&-CAMP) does not perceptibly alter the intracellular concentration of cAMP in H35 cells, which is normally about 0.1 PM. Eflects of CAMP Analogs on Protein Kinuse ilctivity-'The other likely possible mechanism of action of cAMP analogs, namely the direct activation of protein kinase, has been tested initially mrith rat liver extracts for the most part.
It, has been reported that' all of t,he analogs used in these studies do activat,e prot)ein kinases from various tissues in z&o (8,9,13,14). The activity of protein kinase (with calf thymus fl histone as a substrate) in H35 cells extracts is stimulat,ed by CAMP, but the degree of effect is quite variable for reasons that have not been accounted for as yet.
From studies by Hohmann and Langan, and our own preliminary results, it is known that phosphorylation of the serine-37 residue of fl histone in intact H35 cells is stimulated 3-to lo-fold by I3tncAMP. These results demonstrate the presence of functional cAMl-'-dependent, prot,ein kinase activity in whole cells.
The inconsistent stimulation of H35 protein kinase activity by CAMP in titro appears to result from variable dissociation of the holoenzyme upon lysis, thereby increasing basal enzymatic activity and proportionately reducing the CAMP dependence of f, histone phosphorylation. This explanation is supported by the fact that the ability of crude extracts from H35 cells to phosphorylate fl histone is essentially identical with that  Sfter lysis and centrifugation at 20,000 X g for 15 min, aliquots of the soluble extracts were incubated with [-y-32P]ATP (80,000 cpm per nmol; 40 x lo8 cpm total) and 1 mg of purified fl histone from calf thymus with 50 mM Tris (pH 7.5), 5 mM M&12, and 1 rn~ dithiothreitol for 20 min at 37" with or without the indicated nucleotides. The fi histone was then isolated and subjected to digestion with trypsin. The tryptic peptide containing serine-37 was isolated by paper electrophoresis and thin layer chromatography before counting.

Nucleotide added Concentration
[y-a*P]ATP incorporated i n to swine-37 of fl histonea a Each value is the average of two observations zt standard deviation.
of rat liver extracts when CAMP is present.
Furthermore, addition of the inhibitor (37) of caMI'-dependent protein kinase from skeletal muscle is able to depress the basal activity of H35 extracts toward fl histone as a substrate by as much as 90 to 95% when the degree of CAMP dependence is low. In contrast, the basal activity of the crude rat liver kinase is depressed by only 25 to 50%.
With the exception of Bt,cAMP, all of the 6 and 8-substituted analogs, as well as CAMP itself, were equally effective at I PM as activators of rat liver protein kinase with fl histone as substrate (Table IV).
Dose-response studies have shown that the 6-and 8substitutcd analogs tcst,ed are roughly equipotent (average concentration for half-maximal activation, 20 00 40 nM), including ones that are inactive as enzyme inducers (e.g.

8ffzN-CAMP).
Others have reported that CAMP analogs with O2 substituents are much less effective as kinase activators (13,17,18), and we have also found &cAMP and 02'-monobutyryl CAMP (data not shown) to be much less effect#ive in t#his regard. The fact that any response to these two analogs is observed prer;umably results from the presence of esterase activity in these cell-free extracts (17,38).
In one experiment in which a substantial effect of free CAMP on protein kinase activity was detected with H35 cell extracts, both CAMP and EMeS-CAMP were found to stimulate sitespecific fl histone phosphorylation to an cquivalcnt extent (Table V).
In this experiment, the tryptic peptide of fr histone containing t'he serine-37 residue was isolated and shown to be the site of enhanced phosphorylation in titro. Tryptic peptides containing other phosphorylated moieties did not exhibit any change in degree of phospharylation with either cyclic nucleotide. These results are cntircly analogous to those reported in rat liver by Langan (22,26,39,40).
In preliminary experiments we have also found that active aminotransferase inducers stimulate markedly the incorporation of "Pi (from Na332P04) into se&e-37 of fr histone in intact H35 cells.
In contrast, all three noninducing analogs (8H&cAMP, 8HOEtHN-CAMP, and 6HS-CAMP) are totally without effect on this process in whole cells, even thqugh they st,imulate protein kinase activity in cell-free ext,racts.

Correlation
of Analog E$ects on Growth Rate and Enzyme Ac-G&y--We have reported previously that Bt-CAMP, 8HS CAMP and 8MeS-CAMP are all capable of inhibiting the rate of growth of H35 cells in a reversible manner by reducing t,he rate of DNA replication (3,29,30,41). Although the experiments have been performed under different conditions (e.g. serum-free versus serum-containing medium, 3-to B-hour incubation versus several days' incubation, etc.), the dose-response relationships for inhibition of growth rate and elevation of carboxykinase activity (as well as that of the aminotransferase; see Fig. 2) by Bt2cAMP and 8MeScAMP are similar (Fig. 4). The structural requirements for growth inhibition and effects on aminotransferase (or carboxykinase) activity also appear to be similar (Table VI). Although this relationship has not been tested extensively as yet, it, is fair t,o say that, all analogs which are capable of elevating t'he activity of both enzymes are also capable of reversibly inhibiting the rate of growth of H35 cells, with the one exception not,ed below.
Four of the analogs were found to produce irreversibly toxic effects on H35 cells (Table VI).
Removal of these compounds after a a-day exposure did not reverse their killing effects.
Only one of these analogs, BMeS-CAMP, was able to increase the activity of the aminotransferase or the carboxykinase (Figs. 2  and 3). The lethal effects of 8H2N-, 8HOEtHN-, and 6HS-CAMP may well be unrelated to any CAMP-mediated event, since they are also able to kill another hepatoma cell line (HTC) whose growth is unaffected by analogs which reversibly inhibit H35 cell growth (Table VII).
These analogs are also lethal to a cultured human astrocytorna cell line whose growth is reversibly inhibited by BtlcAMP.5

DISCUSSION
The results of these studies strongly suggest that the structural specificities for cyclic nucleotides of the intracellular components involved in the regulation of tyrosine aminotransferase and phosphoenolpyruvate carboxykinase synthesis by CAMP are Fro. 4. Dose-response relationships for the effects of BtzcAMP and 8MeS-CAMP on growth rate and P-enolpyruvate carboxykinase activity in H35 cells. In t,he growt,h experiments the analogs were first added 24 hours after subculture and again 48 hours later with a change of medium.
The cells were harvested 24 hours later and assayed for DNA content, as described under "Experimental Procedures." Each value is the mean of three observations in two separate experiments, with standard errors of less than 10%. For studies of carboxykinase activity, serumdeprived confluent cells were exposed to t,he analogs for 5 hours prior to harvest and assayed as described under "Experimental Procedures." Each value is the mean of four observations in a single experiment, with average standard errors of 10 to 157,. The maximal-fold increase (i.e. 100% of the maximal response) in carboxykinase activity is given in brackets below. A, BtzcAMP   Analogs were first added 24 hours after subculture and re-added on Day 3 with fresh medium.
* Based on the increase in tyrosine aminotransferase activity after exposure of confluent stationary phase cells to 0.5 mM of each analog for 3 hours.
Each value was obtained with four to eight separate flasks in at least two experiments.
* Defined as irreversible cytotoxicity as judged by a continual decline in cell number even after withdrawal of the analog. Cells released into the medium were no longer viable, as judged by their inability to reattach and grow after transfer to medium devoid of analogs.
highly similar, if not identical. A similar, although less well documented, correlation appears to hold for CAMP analog effects on growth regulation and enzyme induction. At least two possible explanations can be envisioned to account for the observed correlations: 237 TABLE VII E$ects of analogs of CAMP on growth of cultured HTC cells Flasks were seeded initially at about 0.5 X 10" cells. After 1 day the analogs were added and again on Day 3 after a change of medium.
On t,he 4th day the cells were harvested for assay of their DNA content. 1. All of the analogs act by inhibiting CAMP phosplrodiesterase activity to various degrees, and the subsequent rise in intracellular CAMP brings about the observed response. Although this possibility is difficult to eliminate totally, a number of facts argue against it. First of all, no rise in intracellular CAMP has been observed in preliminary experiments with H35 cells exposed to an analog active as an inducer.
Second, known inhibitors of CAMP phosphodiesterase have little effect and only weakly poteutiate the effects of exogenous CAMP. As in most other systems (8,12,13), the active analogs produce much greater responses than phosphodiesterase inhibitors in the absence of exogenous CAMP.
Although permeability problems may account in part for the poor cffcctiveness of these inhibitors (42), it is difficult to imagine that they are all subject to the problem at the same time that all of the active analogs are not. Third, an analog t,hat, has been found to be totally inactive as an inhibitor of the partially purified phosphodiesterase from two tissues (12, 13), 8HO-CAMP, is as effective as Ut-CAMP in provoking a rise in aminotransferase activity in H35 cells and markedly enhances glucose production in liver slices (14). Finally, 02'.monobutyryl CAMP has been shown to be a better inhibitor of phosphodiesterase from four tissues than either N6monobutyryl CAMP or &CAMP (17), and yet it is much less effective as an enzyme inducer.
2. The analogs act by stimulating protein kinase directly, and (a) a number of protein kinases exist that exhibit different substrate specificities but possess similar structural specificities for cyclic nucleotide activation; or (b) a single protein kinase (or two or three closely related isozymes) is activated by the various analogs to differing degrees, and the resulting catalytic subunit phosphorylates different protein substrates specific for the particular process being regulated (i.e. induction of the aminokansferase, induction of the carboxykinase, inhibition of growth, etc.), which ultimately produces the observed differing degrees of response.
Thcsc possibilities are also difficult to evaluate with currently available information.
The second possibility is most consistent with the known facts but is supported by circumstantial evidence at best. The preliminary observation that analogs active as enzyme inducers (a total of nine have been tested to date) produce marked stimulation of specific fi histone phosphorylation in whole cells, whereas the three noninducing analogs tested are completely without effect by guest on March 24, 2020 http://www.jbc.org/ Downloaded from on this process, provides the strongest and most direct evidence in support of protein kinase involvement. (It should be pointed out that this analysis is viewed as providing an index of t,hc degree of activation of protein kinase in v&o rather than as irnplying that the process of f1 histone phosphorylation per se is a prerequisite for induction of these two enzymes.) The fact that cAlMI%iependent phosphorylation of f, histone exhibits the same site specificity in H35 cells in viva and in vitro as it does in rat liver suggests that the CAMP-dependent protein kinase present in these cells is very similar to, if not identical with, that in normal liver (20-22, 26, 39, 40, 46-50).
The variably reduced dependence of this enzyme on CAMP in cell-free extracts may well be a technical problem. This suggestion is supported by the fact that the degree of protein kinase activation by CAMP analogs in intact H35 cells (3-to X-fold) is simiilar to that produced by BtzcAMP in rat liver (22,26,39,40). It is possible, however, that these cells do possess a variably reduced quantit,y of t,he regulatory subunit of protein kinase, as suggested for H'I'C cells (51) (53).
The concept of a single protein kinase (or two or three closely related forms) and multiple protein substrates also predicts that cellular variants should be found in which one or more responses to CAMP analogs are alt,ered or missing.
The reduced response of the aminotransferase in MH,C, cells can be explained on this basis (30).
In addition, the facts that the aminotransferase can be partially induced in HTC cells by U&AMP and SMeS-cAMY (30,54) but that the rates of growth or DNA synthesis are not inhibited by these analogs are also explicable on this basis (29,30).
The phosphorylation of rat liver ribosome-associated proteins has been shown to be stimulable by glucagon and Bt'zcAMP in vivo (55) and by CAMP in v&o (50,56,57), but t'here is no evidence that such an event triggers alterations in ribosomal function (58).
Since CAMP analogs do not seem to have any consistent major elfeat on over-all hepat'ic protein synthesis (5, B), it is possible that phosphorylation-induced changes in ribosome function will only be demonstrable when the synthesis of a specific protein is examined. It remains to be seen whether this represents the mechanism by which CAMP analogs regulate specific protein ,syntheais at the translatiorml level. The fact that the three analogs that are lethal to H35 cells also kill HTC cells suggests that this effect can be ascribed to even& that are apparently unrelated to cAMI'-mediated phenomena, &ce the growth of HTC cells is not inhibited by cAMP analogs active as enzyme inducers. &Amino analogs of CAMP are also toxic to nongrowing cells, since they inhibit protein synthesis ia H35 cells deprived of serum for -20 hours (see Table III).
The nature of the mechanism by which 8-amino derivatives exert their toxic effects is not clear.
In the case of GIL%CAMP, this derivative is not resistant to phosphodiesterase attack (9) and in the process is converted to 6.thioinosinic acid, a cytotoxic agent (59). It is of interest that this analog is more effective as an activator of t#he cGMP-dependent protein kinase from lobster tail muscle than it is as an activator of the CAMPdependent protein kinase (9). GMeS-CAMP is the only exception to the observed correlations in that it is a moderate enzyme inducer and yet is irreversibly toxic to H35 cells.
Since the e xperimental conditions are different, it is conceivable that this analog could produce CAMPlike effects in nongrowing cells, leading to increases in aminotransferasc and carboxykinase activities, whereas, in growing cells over an extended period of time, additional non-CAMPmediated effects could occur. This suggestion is supported by the fact that GMeS-CAMP does markedly inhibit bhe rate of growth of HTC cells and produces severe cytotoxic effects at 0.5 mM similar to those in H35 cells. These non-CAMP effects could well result from formation of the corresponding 6MeSadenylafe, which is structurally relsted to the cytutoxk 6-t&+inosinic acid and also inhibits tumor growth (59). The toxic effects of GMeS-CAMP would be expected to predominate even though protein kinase might bc activated, since the latter event apparently leads only to a reversible slowing of the growth rate (3,29,41).
It should be pointed out that, had we relied exclusively upon in vitro assay of protein kinase activation by the various analogs, it would have been concluded long ago that this enzyme was not likely to participate in the response of the cells to cAMP analogs, This is illustrated by the fact that &U&N-CAMP does not induce the aminotransferase but is a good activator of protein kinase i~z z-i&o (see Table II and Refs. 8 and 13 to 14).6 Because of the problems of possible diffcrenccs in uptake or metabolism (or both) of various analogs to compounds which have different abilities to activate protein kinasc or to inhibit phosphodiesterase (as appears to be the case with the 8H&-CAMP and Ut~cAMP derivatives), the cffccts of the analogs on protein kinase and phosphodiesterase activities have been assessed primarily in whole cells, thereby circumventing these problems.
Studies of the uptake and metabolism of various analogs are planned as soon as the appropriate radioactive derivatives are synthesized.
ArG-Monobutyryl CAMP and Et,-CAMP appear to derive their enhanced efficacy in whole cells, relative to CAMP itself, from their resistance to phosphodiesterase attack and possible inhibition of this enzyme in addition (8,9,13,19). The acdivity of N6-monobutyryl cALNIP both in vivo and in vitro suggests that intracellular production of this analog may well account, in most, cases, for the metabolic activity of I!&-CAMP in whole cells, since the latter is much less effective in vitro (15)(16)(17)19). Addition of trjtiated IXzcAMP to H35 cells leads to the intracellular accumulat,ion of both labeled NC-and @'-monobutyryl CAMP and free CAMP (30). Similar results have been reported for other cultured cells (19,60), although at least one exception 6 A rough correlation has been found to exist between the relative potencies of a series of CAMP analogs as activators of the partially purified bovine brain kinase and their ability to increase tyrosine aminotransferase activity in H35 cells (4, 17). Although such a correlation is certainly compatible with the concept of protein kinase involvement in enzyme induction, unfortunately it is dithcult to see how potency studies with a partially purified brain kin&se preparation can adequately assess possible differences in the uptake, metabolism, and inherent ability of the various analogs to act,ivate protein kinase in intact cultured hepatoma cells.
As indicated, the best correlation analysis would seem to be that using f 1 histone phosphorylation in whole cells a.8 an index of protein kinase activation. has been noted (61). The weak effects of 02'-monobutyryl CAMP in H35 cells may be the result of poor cellular penetration or the fact that it is apparently not active until converted to CAMP (which is not resistant to attack by phosphodiesterase), or bot,h (15)(16)(17)19). Preliminary experiments have revealed that CAMP penetrates H35 cells very poorly, if at all. These results presumably account for the very weak ability of CAMP as an enzyme inducer.
Others have also reported that CAMP is poorly t,aken up and rapidly degraded in perfused liver (62) and in other Gssues (18, 19, 60, 63j. From the prcscnt results, and from those of others, it seems evident that some of the newer analogs of CAMP deserve consideration as alternatives to the use of BtncAMP. The availability of a series of CAMP analogs possessing differential abilities to affect cAMPmediated processes should be of invaluable assistance in terms of their use as molecular probes. They also offer some promise as tissue-specific growth-inhibiting agents for use in cancer chemotherapy (29,30,41,51,58,64,65).
Ackno~ledgmenls-The expert technical assistance of Carolyn Bearg is gratefully acknowledged.
We would also like to thank the group at ICN Nucleic acid Research lnstitut,e for their invaluable assistance in providing us with c.4MP analogs. We are especially grateful to Dr. Jon Miller for providing us with informat'ion concerning his studies on the effects of CAMP analogs on tyrosine aminotransferase activity in rat liver prior to publication.