potent acrasin yet reported is adenosine-3',5'-cyclic phosphate (cyclic AMP).* ... spotted on agar plates containing cyclic AMP, the amoebae form characteristic.
THE ACRASIN ACTIVITY OF 3', 51-CYCLIC NUCLEOTIDES BY BRUCE M. CHASSY, L. L. LOVE, AND MICAH I. KRICHEVSKY ENVIRONMENTAL MECHANISMS SECTION, NATIONAL INSTITUTE OF DENTAL RESEARCH, NATIONAL INSTITUTES OF HEALTH, BETHESDA, MARYLAND
Communicated by Marshall Nirenberg, June 25, 1969
Abstract.-3',5'-Cyclic nucleotides are acrasins for the cellular slime mold, Dictyostelium discoideum; they have chemotactic activity on the myxamoebae at very low concentrations. However, not all cyclic nucleotides can evoke the adhesiveness necessary for aggregate formation. 2',3'-Cyclic nucleotides and dibutyrylcyclic adenosine monophosphate are not acrasins, though they enhance the rate of differentiation and morphogenesis. The addition of 3',5'-cyclic nucleotides to agar test plates stimulates the rate of morphogenesis. While the stimulatory effect of cyclic nucleotides is concentration-dependent, it is fairly uniform between pH 5 and 7, with an optimum at pH 6. The acrasins may be metabolized by extracellular phosphodiesterase to 5'-nucelotides which may then stimulate differentiation and morphogenesis. The cellular slime molds (myxamoebae) are capable of vegetative growth.' However, when challenged with certain conditions (e.g., lack of nutrients or low humidity), they aggregate to form a "multicellular organism" which develops into fruiting bodies consisting primarily of stalks and spores. These myxamoebae are one of the simplest systems available for the study of growth and development. For many years biologists have attempted to elucidate the chemical and physical events involved in the aggregation, morphogenesis, and differentiation of these organisms. In particular, the forces involved in attracting unicellular organisms to form a multicellular organism have received considerable study. As early as 1902, both Olive2 and Potts3 suggested the presence of a chemical mediator of aggregation. In 1942, Runyon4 reported that a dialyzable small molecule could orient, or attract, amoebae. The case for chemotaxis was further advanced when Bonner' showed that the amoebae secrete a diffusible chemical which can orient and attract other amoebae. Bonner named the agent acrasin after the genus Acrasiales. The chemical identity of acrasin was obscured by the acrasin-like activity of a number of compounds and extracts. The most potent acrasin yet reported is adenosine-3',5'-cyclic phosphate (cyclic AMP).* Konijn and his collaborators reported on the acrasin activity of cyclic AMIP and noted its occurrence in bacteria, amoebae, and other natural sources which were known to attract amoebae.6 When the amoeba Dictyostelium discoideum is spotted on agar plates containing cyclic AMP, the amoebae form characteristic rings;t cyclic AMP causes the development of the stickiness or adhesiveness necessary for aggregation.7 Cyclic AIMP has relatively little effect on the course of the morphogenetic process, however. Recently, Krichevsky et al. showed that 5'-nucleotides (e.g., adenosine-5'phosphate (5'-AMP)) can stimulate differentiation even though they lack acrasin activity.8 D. discoideum excretes an extracellular enzyme capable of 296
VOL. 64, 1969
297
BIOCHEMISTRY: CHASSY ET AL.
hydrolyzing cyclic Ai'IP to 5'-AMIP.9 Chassy and co-workers reported that this enzyme is not specific for cyclic AMIP but hydrolyzes other cyclic nucleotides. 10 These results suggest a relationship between the factors responsible for aggregation and those responsible for differentiation, possibly mediated by the cyclic nucleotide phosphodiesterase. Materials and Methods. -Strains: D. discoideum NC-4 grown on Escherichia coli B was used in these experiments. Bioassay: Details of the spore count assay procedure for stimulation or inhibition of the rate of morphogenesis were published previously." Various compounds were incorporated directly into purified washed agar, and droplets of amoebae were placed on the plates. Both acrasin activity and stimulation of the rate of differentiation could be followed when the plates were kept at 18°. Acrasin activity was also followed by observing migration of amoebae out of a droplet toward a juxtaposed droplet or streak of the compound being tested. Photography: Micrographs were taken with a Zeiss WL stand equipped with a 4X planapochromat 0.16 NA objective, lOX eyepiece, 0.32 NA bright field condenser, and a Polaroid camera back. Bright-field transmitted illumination was used. Larger areas were recorded macroscopically (> magnification) with a Bausch and Lomb photomicrograph camera system fitted with a 48-mm lens using transmitted light (green filter). Polaroid 3000 black-and-white film was used for all photographs. Chemicals: N6,2'-di-n-butyryl-3',5'-cyclic AMP; guanosine- and uridine-2',3'-cyclic phosphates; and 5'-nucleotides were obtained from Schwarz BioResearch. Adenosineand cytidine-2',3'-cyclic phosphates; cytidine-, uridine-, thymidine(deoxyribofuranosyl)-, and guanosine-3',5'-cyclic phosphate were purchased from Sigma. Inosine3',5'-cyclic phosphate was prepared according to Borden and Smith'2 or by deamination of cyclic AMP, based on the method of Fujisawa."3 N6-butyryl-adeinosine-3',5'-cyclic phosphate was prepared by boiling an aqueous solution of dibutyryl cyclic AMP until hydrolysis, as judged by thin-layer chromatography, was complete (about 10 min). The purity of synthetic and commercial nucleotides and cyclic nucleotides was verified by the ultraviolet adsorption spectrum (Gilford model 2400 recording slectrophotometer) and by their homogeneity in thin-layer chromatography (microcrystalline cellulose plates with isobutyric acid-1 N NH4OH, 100:60 v/v; or ethanol-i 31 ammonium acetate, 7 :3 v/v as the developing solvents). Acceptable purity was taken as less than 2 7 minor spots on thin-layer chromatography and ±5% agreement with reported extinction coefficients. Results. The acrasin activity of cyclic AMIP was demonstrated by the following experiment. Figure 1 shows the results when a droplet of amoebae is a
FIG. 1.-Photographs of agar test plate (3.2 X) at 18 hr after application of amoebae with (a) no additions to the agar, (b) 1 mI AMIP final concentration, and (c) 1 m\M1 cyclic AM1P (black lines delineate portions of separate photographs spliced together).
C~~~~~1
i
* . # 6
2982BIOCHEMISTRY: CHASSY ET AL.
PROC. N. A. S.
placed on the agar plate and allowed to stand overnight. On washed agar (Fig. la), large aggregates of amoebae have formed and are ready to migrate. On agar containing AMP (Fig. lb), pseudoplasmodia (slugs) have already formed and-are migrating, thus are apparently further along in the differentiation process. On agar with cyclic AMP (Fig. ic), a peripheral ring is clearly visible. The diameter of the droplet containing amoebae was approximately half the diameter of the peripheral ring; that is, the amoebae have migrated outward. This outward migration is a good direct test of the acrasin activity of cyclic AMP.7 In addition, the cells in the center show a mixed development; there are more slugs than in Figure la but fewer than in Figure lb. Thus, the incorporation of cyclic A1\P in the agar results in outward migration of amoebae from the edge of the original spot, as well as stimulation of the rate of morphogenesis within the confines of the original spot. Various 3',5'-cyclic nucleotides were tested for acrasin activity (Fig. 2). Figure 2a shows the effects of four concentrations of cyclic UMP. On agar with lowest concentration, the amoebae have migrated outward but have not formed the distinct ring seen with cyclic AMP (Fig. ic). There are numerous small aggregates in the center portion occupied by the majority of amoebae from the original droplet. With increased UMP concentrations, the ring is more pronounced; there are numerous small aggregates in the center region; and fruits begin to appear. Thus, cyclic UMP stimulates differentiation inside the spot. Some slugs have migrated to the outer periphery. At 3 mM\ cyclic UMP, aggregation is inhibited. Three effects of cyclic UMP can be noted: It attracts amoebae at all concentrations tested; it stimulates differentiation at intermediate concentrations; and it inhibits differentiation at higher concentrations. With cyclic CiMIP in the agar test plates (Fig. 2b), the amoebae migrate out of the original droplet, thus demonstrating the acrasin activity of cyclic CI\iP. As with cyclic U\IP, cyclic CMP stimulates spore formation. Although cyclic CMP attracts the amoebae, it does not make them as sticky as does cyclic A1iP (with the possible exception of the highest concentration tested (3.0 mM)). Rings occur only with the highest concentrations of cyclic CAMP; even then, most of the amoebae migrate into an area only slightly larger than the original droplet. Figure 2c shows the effects of cyclic GMP in the agar test plates. At lower concentrations, cyclic GMP stimulates differentiation and has some attractive power; however, the cells do not aggregate outside the original spot. At 0.7-1.0 mMX1 cyclic GMP, the attraction is more pronounced, but the lack of adhesive character is more readily observable. Very high concentrations of cyclic GMP (3.0 mMI) completely inhibit migration and differentiation. In some cases, such inhibitions are overcome with time and the amoebae differentiate to form fruiting bodies; however, if the concentration is sufficiently high, development is completely inhibited. When cyclic IM1P is used (Fig. 2d), the attraction outward is marked, with clear ring formation at higher concentrations. Some stimulation is observed, although the highest concentration (1 mM) is inhibitory. Similar acrasin activity is observed with cyclic TA1IP (not shown); 3',5'-cyclic TMP stimulates at low concentrations and inhibits at higher concentrations. In order to further test the specificity of the chemo-
BIOCHEMISTRY: CHASSY ET AL.
VOL. 64, 1969
299
3
4
(ii
l
2
l
:a -.
3
2£i 0
4
l
.;;
4
FiG. 2.-Photographs (after 18-20 hr) of agar test plates (3X) with various 3',5'-cyclic nmcleotides added to the agar as follows: (a) cyclic UMIP (1, 0.1 mAI; 2, 0.3 mAI; 3, 1.0 mM 4, 3.0 mAI); (b) cyclic CMP (1, 0.05 mMI; 2,0.3 mM; 3, 0.7 mMI; 4, 3.0 mAI); (c) cyclic GAIP (1, 0.3 rM; 2, 0.7 mM; 3, 1.0 mAI; 4, 3.0 mM); (d) cyclic IMP (1, 0.1 mM; 2, 0.3 mAI; 3, 0.7 mAI; 4, 1.0 mM); white lines indicate portions taken from photographs of separate plates).
tactic response, dibutyryl cyclic AMIP and its hydrolysis product, butyryl cyclic AM\IP, were also tested. Although dibutyryl cyclic AMP stimulates differen-
tiation, no rings or acrasin activity was observed with this compound. Butyryl cyclic ASMP had acrasin activity but less stimulatory potency than dibutyryl cyclic AMP. Adenosine-, guanosine-, cytidine-, and uridine-2',3'-cyclic phosphates wera stimulatory to the rate of morphogenesis; however, these compounds showed no acrasin activity. The stimulations produced by mononucleotides and 3',5'-cyclic mononucleotides are illustrated in Table 1. When AMP is included in agar plates, the rate of differentiation is accelerated (that is, there is a greater number of fruits than in a water control at a given time). Cyclic AiMP is as stimulatory as A\MP (Table 1), although it inhibits at higher concentrations (above 2 m'M). At low
BIOCHEMISTRY: CHASSY ET AL.
300
PROC. N. A. S.
TABLE 1. Stimulation of rate of morphogenesis by cyclic nucleotides: dependence on concentration. Compound
Number of Fruits* -(Millimolarity) t
(control) 0
Expt. 1
AMP cyclic AMP
12 12
0.2
0.4
0.6
1.0
2.0
64 60
83 56
98 88
118 174
124 81
(Millimolarity) Expt. 2
(control) 0
cyclic UMP cyclic CMP Dibutyryl cyclic AMP cyclic TMP cyclic IMP cyclic GMP
8 8 18 18 22 22
0.1
0.3
0.7
1.0
3.0
26 6 23 37 32 33
33 35 53 13 80 133
74 126
77 112 28 60 0
0 90 0 13 0 0
57 50 8 113
52
* All plates were counted when water controls began to fruit, usually 15-25 hr.
t Obtained by incorporating aliquots of concentrated stock solutions directly into warm agar prior to pouring plates. For details see reference 11 and under Materials and Methods.
concentrations, cyclic IMP and cyclic GMP both stimulate differentiation with optima at approximately 0.3 mM; at higher concentrations, they inhibit (Table 1). Cyclic UMP has a similar effect (optimum at 0.7-1.0 mM). Stimulation at low concentration and inhibition at higher concentrations are again observed with cyclic TMP and dibutyryl cyclic AMP (Table 1). However, cyclic CMP is different in that it stimulates the entire concentration range tested (Table 1). The pH of test plates is not critical to morphogenesis between pH 5 and pH 7. Table 2 shows that both 5'-AMP and cyclic AMP exhibit pH optima approximately at pH 6, in contrast to the controls containing nonstimulatory low levels of sodium phosphate buffer. When cyclic nucleotides were incorporated into test plates at 3 pH values (Table 2), a variability of stimulation similar to that found with AMP was observed. As mentioned earlier, cyclic nucleotides can inhibit differentiation at high concentrations. Where inhibition is observed, the cells may slowly recover and differentiate, or they may be permanently arrested. A 40-fold magnification of the peripheral ring produced by 1 mM cyclic AMP is shown in Figure 3a. At successive 12-hour intervals (Fig. 3b and c), the amoebae have migrated away from the clumps forming the ring. Although the amoebae in the center of the original droplet have already differentiated and fruited, those in the peripheral area dispersed without further morphogenesis. At lower, but still inhibitory, TABLE 2. Effect of pH on stimulation of rate of morphogenesis by nucleotides. Number of Fruits*
pHt
Additions to test plate
5
6
7
None cyclic AMP (0.2 mM) cyclic AMP (1.0 mM) AMP (1.0 mM)
33 48 9 77
29 126 45 106
26 32 32 68
* All plates counted after standing at 180 for 19 hr. t pH achieved by incorporating 2 mM sodium phosphate into test plates.
BIOCHEMISTRY: CHASSY ET AL.
VOL. 64, 1969
301
I .,
4
t-
.
* '.... i _
do an_ I;,
An,
, B,,s,
4.A
e
4 ,'>.
z
.,
, 4,**%
a.
'4..
.-
FIG. 3.-Photographs of area of peripheral rings (21.2 X) formed on agar plates containing 1 mM cyclic AMP; taken at 24, 36, and 48 hr (a, b, and c, respectively). Arrow points to center of original droplet of amoebae.
levels of cyclic nucleotides, differentiation may be retarded but not permanently blocked. 1\Iyxamoebae located in the peripheral ring formed on cyclic GMIPor cyclic CMP-impregnated plates are shown in Figure 4. The slugs have formed mature fruiting bodies which are normal in all respects. The absence of slime tracks demonstrates that the slugs fruited directly from the peripheral ring aggregate. Another direct assay of chemotactic activity can be performed by placing streaks of amoebae opposite streaks of the test compound. In Figure 5a these streaks can be seen clearly: a streak of amoebae suspension opposite a streak of cyclic AiMP. After 18 hours, a clearly defined line of single amoebae and slugs has migrated toward the streak of cyclic AMP. In a control experiment with 5'-A1\IP substituted for cyclic AMIP (Fig. 5c), there is no such attraction; instead, these cells are beginning to aggregate within the original boundaries of the streak of amoebae. In Figure $id and e, respectively, cyclic GMIP and cyclic INIP were tested in streaks placed opposite amoebae. Single amoebae have
t
c FIG. 4.-Photographs of spores formed in peripheral ring on plates containing 1 mM cyclic GMP (a and c) or 1 mMI cyclic CMP (b). (a) and (b) are 21.2 X and (c) is 42.4 X; in both cases focus is on the agar surface rather than the aerial spore itself.
302
BIOCHEMISTRY: CHASSY ET AL.
PROC. N. A. S.
been attracted toward the cyclic nucleotide area, and the remaining amoebae have aggregated into slugs. Cyclic GMP and cyclic IMP did not confer adhesive properties in the previous assay procedure; rather, they gave rise to diffuse rings outside the boundaries of the original droplets (Fig. 2c and d). In this experiment, the lack of adhesion is demonstrated by the large field of individual amoebae that have migrated toward the cyclic nucleotide area without aggregating. Discussion.-Konijn and his collaborators have shown that cyclic AMP has potent acrasin activity.6' I This report demonstrates that 3',5'-cyclic nucleotides as a group have acrasin activity. Of the cyclic nucleotides tested, only dibutyryl cyclic AMP failed to attract the myxamoeba, D. discoideum. One of the difficulties in the search for the chemical identity of acrasin was that extracts from various sources possessed acrasin activity. Although cyclic AMP is itself widespread in nature, some of the cyclic nucleotides studied here also occur naturally and thus would complicate a search for a single acrasin structure. In addition, since cyclic-2',3'-nucleotides and di'butyryl cyclic AMP have no acrasin activity, it appears that an unhindered 3',5'-cyclic phosphate is required for activity. Though the 3',5'-cyclic phosphodiester bond seems to be a common feature of acrasins, the unique chemical role of nucleotides containing this bond has yet to be elucidated. All the cyclic nucleotides tested had about the same ability to attract amoebae; however, not all FIG. 5.-Photographs of amoe- evoked the adhesive properties necessary for agbae in streak attraction assay (see Materials and Methods). gregation (those that could, still conferred less The arrow points in the direc- adhesiveness than cyclic AMP). Cyclic AMP is tion of the streak of compound being tested; black lines the most effective acrasin because it both attracts scratched in plates delineate well and confers good adhesive properties. Thus, original confines of streaks. (a) Agar streaked with 40 mM attraction and adhesion are distinct properties; a cyclic AMP at initiation of the given compound may possess a potency for either experiment; (b) shows this of the processes without the other. Although atsame plate after 24 hr at 180; (c) pictures the effect of a streak traction of amoebae is caused by nucleotides of 40 mM AMP after 18 hr; bearing a 3',5'-phosphodiester bond, the specific (d) and (e) are the same experiment as above, but with structural requirements for producing the adhe40 mM cyclic GMP and cyclic sion necessary for the formation of a migrating IMP, respectively, streaked on slug require further research. the agar.
XVOL. 64, 1969
BIOCHEMISTRY: CHASSY ET AL.
303
i\Iononucleotides8 and compounds which cause intracellular accumulation of mononucleotides14 15 stimulate morphogenesis; this fact may be related to the acrasin activity of cyclic nucleotides. Cyclic nucleotides are themselves stimulatory to morphogenesis; however, D. discoideum elaborates an extracellular enzyme which can hydrolyze cyclic nucleotides to their corresponding 5'-nucleotides.9 Thus, the cyclic nucleotides per se may not be stimulators; rather, their hydrolysis products, the 5'-nucleotides, may be stimulators. Cells inhibited initially by higher concentrations of cyclic nucleotides may in time overcome the inhibition and differentiate; they may, therefore, possess metabolic means of reversing the effect of the cyclic nucleotide. Chassy et al. have reported that the cyclic nucleotide phosphodiesterase has a broad specificity for cyclic nucleotides.10 This enzyme could provide both a source of 5'-nucleotides for the stimulation of morphogenesis and also a mechanism for the slow reversal of inhibition. The foregoing is consistent with the postulate that cyclic nucleotides attract the cellular slime mold, D. discoideum, and generate the adhesion which gives rise to aggregates. The concentration gradient required for directional information is potentiated by hydrolysis of cyclic nucleotide to 5'-nucleotide by the extracellular phosphodiesterase. The resultant 5'-nucleotide may then stimulate further differentiation and morphogenesis. Such a hypothesis invites direct study of not only the permeability and fate of cyclic nucleotides and 5'-nucleotides in differentiating cultures, but also the nature of the chemotactic and adhesive processes. The hypothesis also requires direct proof for a role of the enzyme as an antiacrasin mediator of morphogenesis. * Cyclic nucleotides are abbreviated as follows: uridine-; thymidine(deoxyribo)-; guanosine-; inosine-; cytidine-; N6,2'-O-di-n-butyryladenosine-; N6-butyryladenosine-3',5'-cyclic phosphate as cyclic UMIP, cyclic TMP, cyclic GMP, cyclic INIP, cyclic CMzIP, dibutyryl cyclic AMP, but cyclic AMP, respectively. t The rings may be explained by decreased cyclic AMP concentration in the area covered by the droplet of amoebae. This could result from dilution by the droplet or from the action of the extracellular phosphodiesterase, or both. The amoebae then migrate outward in an expanding ring toward higher cyclic AMP concentration. I Bonner, J. T., The Cellular Slime Molds (Princeton: Princeton University Press, 1967),
2nd ed., p. 3. 2 Olive, E. W., Proc. Boston Soc. Nat. Hist., 30, 451 (1902). 3Potts, G., Flora, 91, 281 (1902). 4Runyon, E. H., Collecting Net, 17, 88 (1942). 5 Bonner, J. T., J. Exptl. Zool., 106, 1 (1947). 6 Konijn, T. M., J. G. C. van De Meene, J. T. Bonner, and D. S. Barkley, these PROCEEDINGS, 58, 1152 (1967). 7 Konijn, T. A., D. S. Barkley, Y. Y. Chang, and J. T. Bonner, Am. Naturalist, 102, 225
(1968). 8 Krichevsky, M. I., L. L. Love, and B. M. Chassy, J. Gen. Microbiol., in press. 9 Chang, Y. Y., Science, 161, 57 (1968). '0 Chassy, B. M., L. L. Love, and M. I. Krichevsky, Federation Proc., 28, 842 (1969). 1' Krichevsky, M. I., and B. E. Wright, J. Gen. Microbiol., 32, 195 (1963). 12Borden, R. K., and M. Smith, J. Org. Chem., 31, 3247 (1966). 13 Fujisawa, K., J. Vitaminol. (Kyoto), 3, 295 (1957). 4 Krichevsky, MI. I., and L. L. Love, J. Gen. Microbiol., 34, 483 (1964). 15 Ibid., 50, 15 (1968).
