DEVELOPMENTAL GENETICS 12:113-122 (1991)
Developmental Protein Synthesis Is Required for the Transcription of Dictyostelium Prespore Genes M.A. BENEDICT, D.A. DESILVER, D.E. PELLETIER, W.H. PENTZ, AND D.I. RATNER Department of Biology, Amherst College, Amherst, Massmhusetts and Lodish, 1983; Chisholm et al., 1984; Mehdy et al., 1983; Oyama and Blumberg, 19861. Cell-specific gene expression is also modulated by the effects of ammonia, “DIF,” and other low molecular weight metabolites still chemically unidentified [reviewed in Williams, 19881. Finally, if normal intercellular contacts are prevented by forcing cells to develop in rapidly shaken suspension, prestalk transcripts accumulate as usual, but (the majority of) prespore genes are not expressed [Chisholm et al., 1984; Mehdy et al., 19831. There is evidence that protein synthesis may be necessary for the transcription of prespore genes and/or for the stabilization of their transcripts. The translational inhibitor cycloheximide prevented the reaccumulation of prespore mRNA species in developing amoebae which had been disaggregated and then stimulated with cyclic AMP [Mehdy et al., 19831.In contrast to the sensitivity of prespore transcripts, prestalk and nonspecific messages (the latter present at equal levels in both cell types) were insensitive to cycloheximide. No conclusion concerning the role of protein synthesis in prespore gene expression could be drawn, however, because of the possibility that the particular inhibitor used had directly affected other cellular processes. Cycloheximide is known to block maturation of ribosomal RNA precursors in Dictyostelium [Iwabuchi et al., 19711, and to disrupt DNA and RNA synthesis in many systems [reviewed in Pestka, 19771. In several cases, the effect of cycloheximide upon DNA and RNA synthesis or upon amino acid uptake could not have been a necessary consequence of its disruption of protein synthesis, since other modes of translational inhibition Key words: Dictyostelium discoideum, cyclodid not produce the same range of deleterious effects heximide, emetine, protein synthesis, mRNA accu[McMahon, 1975; Timberlake and Griffin, 19731. mulation, transcription The first aim of this study was to establish whether our previous observations with cycloheximide reflected a bona fide requirement for protein synthesis in the INTRODUCTION accumulation of prespore transcripts. We have comThe development of spore and stalk cells of Dictyo- pared the effects of cycloheximide with those of three stelium discoideum depends upon the regulated expression of diverse genes in the two cell types. Hybridization studies of the accumulation of individual gene transcripts have provided information about the developmental kinetics and physiological regulators of gene Received for publication July 31, 1990. expression. Cyclic AMP, for example, is necessary for the induction of all the prespore-specific genes studied Address reprint requests to Dr. D.I. Ratner, Department of Biology, to date, and for a major class of prestalk genes [Barklis Amherst College, Amherst, MA 01002. It has been established previABSTRACT ously that the maintenance of expression of prespore-specific genes of Dictyostelium discoideum is prevented by the translational inhibitor cycloheximide. The drug had no effect upon the level of transcripts of the other genes examined, prestalkspecific or cell type-nonspecific. However, the interpretation of this result is open to question, because of possible nonspecific effects of cycloheximide. We have now characterized the cellular specificity and temporal profiles of mRNA accumulation of additional Dictyostelium cDNA clones, and have examined other inhibitors of in vivo protein synthesis. Four structurally and mechanistically distinct translational inhibitors each prevented the reaccumulation of prespore transcripts in cyclic AMP-primed, disaggregated amoebae. These results establish the importance of developmental protein synthesis in the accumulation of prespore gene transcripts. Nuclear run-on transcription assays were used to learn whether protein synthesis is required primarily for mRNA synthesis or transcript stability. A transcriptional time course first demonstrated that the abundance of these cellspecific transcripts during development mirrors their rates of synthesis. Significantly, the protein synthesis requirement of the prespore genes examined also occurs at the level of mRNA transcription, implying the existence of one or more developmentally regulated transcriptional activators.
0 1991 WILEY-LISS, INC.
114
BENEDICT ET AL.
other translational inhibitors, anisomycin, pactamycin, and emetine. For each inhibitor studied, prespore transcript accumulation was invariably dependent upon concomitant protein synthesis. The absence of prespore mRNA in treated cultures could reflect, a priori, a lack of prespore transcription, or a heightened lability of these particular transcripts. The modulation of Dictyostelium gene activity is known to depend upon both of these mechanisms under different circumstances. Cyclic AMP in fact both stimulates the transcription of cell-specific genes [Landfear et al., 1982; Pears and Williams, 1988; Datta and Firtel, 1988; Manrow and Jacobson, 19881 and stabilizes many of these same transcripts [Mangiarotti et al., 1983; Landfear et al., 19821, though the stabilization may be transient [Manrow and Jacobson, 19881. DIF serves as a transcriptional activator of some prestalk genes [Williams et al., 19871 while a t the same time repressing the transcription of at least one prespore gene [Early and Williams, 19881. The dissociation of multicellular aggregates, on the other hand, results in the decay of some prestalk and many prespore transcripts [Barklis and Lodish, 1983; Mehdy et al., 1983; Amara and Lodish, 1987; Mangiarotti et al., 19891. Thus the requirement of prespore genes for developmental protein synthesis could be attributed either to transcriptional andlor post-transcriptional events. To choose between these alternatives, we performed a series of nuclear run-on transcription assays. The results described here show the importance of transcriptional mechanisms in determining both the temporal profile of the genes studied and especially the dependence of prespore mRNA levels upon developmental protein synthesis.
MATERIALS AND METHODS Methods for the growth and development of D . discoideum strain NC-4, for RNA isolation and hybridization, and for the identification of cell-specific cDNA clones have recently been published [Ratner et al., 19891. Procedures for the determination of rates of transcription a r e outlined in the legends to Figures 4-6 and in Table 1; they will be more thoroughly described elsewhere [DeSilver, Benedict, and Ratner, in preparation]. RESULTS Identification of Cell-Specific Genes Our original study of cell-specific gene expression employed a limited set of genes identified initially in terms of their temporal specificity and only later examined individually for cell-type specificity [Mehdy et al., 19831. We have since directly screened a developmental cDNA library for additional prespore- and prestalk-specific clones [Ratner et al., 19891.Colony hybridization to probes reverse transcribed from cell-specific poly-A RNA identified candidate clones corre+
P +Qo
Qy
* C lone 10
80 28 Fig. 1. Cell-type specificity of cloned genes revealed by hybridization to total cellular RNA. Equal quantities of total RNA purified from prespore or prestalk cells [Mehdy et al., 1983; Ratner and Borth, 19831 were fractionated by agarose gel electrophoresis and transferred to a nylon membrane. The purity of the cell populations used as a source of RNA, as judged by immunofluorescence, was 98% for prespore cells and 85% for prestalk. RNA blots were hybridized to the nick-translated plasmids indicated. Transcript sizes were estimated a t DR10, 2.2 kb; DR28, 0.9 kb; DR37, 1.7 kb; DR53, 0.5 kb; DR80, 1.6 kb; and DR82,2.1 kb. (Reprinted with the publisher's permission from Ratner et al. 119891
sponding to prespore or prestalk mRNA species. Dot blot hybridizations to immobilized DNA minipreps of promising clones afforded further indication of cell specificity. However, specificity is best demonstrated by RNA blotting experiments, in which radiolabeled cDNA plasmids are hybridized to size-fractionated total RNA isolated from the two purified cell types. Figure 1 shows the prestalk bias of DR80 and the prespore specificity of clones 10, 37, 53, and 82. The very slight hybridization of the four prespore genes to the prestalk RNA preparation is readily attributable to the measured impurity of the prestalk cell population (see Fig. 1 legend). DR28 clearly is also prespore biased, though the bias appears less than that of the other prespore
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION CLONE
HOURS OF DEVELOPMENT 0
3
6
9
12
15 18 21
10
37
53
82 28
80 0
3
6
9
12
15 18
21
Fig. 2. Developmental profile of mRNA accumulation. Amoebae grown upon bacterial lawns were harvested, washed, and placed upon buffered pads for development. At the times indicated above each lane, amoebae were collected and total RNA extracted; the fractionated and blotted RNA was then hybridized to each of the probes shown. (Reprinted with permission from Ratner et al. [19891
genes and must therefore reflect some production of DR28 transcripts in cells identified as prestalk. Prespore-specific clone DRlO has been described previously a s clone 2-H3 [Mehdy et al., 19831. While it was possible, a priori, that some of the prespore cDNA clones were redundant isolates of a single gene, hybridizations with mixed probes established the unique size of each clone’s transcript (see Fig. 1, legend). Thus, each of the prespore clones described corresponds to a different gene.
Developmental Kinetics of mRNA Accumulation We have determined the developmental profile of mRNA accumulation for each of the genes presented above, by allowing amoebae to develop upon pads for varying times before they were harvested and RNA prepared. Figure 2 reveals the extent of hybridization of the various cDNA probes to their gene transcripts. The four prespore genes, DR 10, 37, 53, and 82, are similar to one another in their times of expression,
115
though subtle differences exist as to the relative extent of accumulation within the active period. Transcripts of these genes were first detected at 9 hours of development, abundant between 12 and 18 hours, and much reduced by 21 hours. These results are generally consistent with the profile of (class 11) prespore genes described previously [Mehdy et al., 1983; Chisholm et al., 19841. Expression of prestalk gene DR80 occurred a t very low levels even in vegetative and early developing cells, and at much higher levels after 6 hours. The appearance of two maxima of accumulation, with the minor peak a t 9 hours and the major peak at 18 hours, is reproducible. Gene DR28 is expressed appreciably a t all stages, but its transcript is considerably enriched between 9 and 15 hours of development. The temporal profiles of expression of clones 80 and 28 do not agree closely with those of the major gene categories described elsewhere. Since both genes are expressed early in development, and even in vegetative cells, the cell type bias evident when the RNA of migrating slug cells is examined (as in Fig. 1) presumably reflects differential transcriptional shut-off andlor transcript lability in the two precursor cell populations [Borth and Ratner, 1983; Jermyn et al., 19871.
Inhibitors of In Vivo Protein Synthesis Cycloheximide is the inhibitor customarily used to disrupt in vivo protein synthesis in Dictyostelium and was the only drug used previously to prevent the accumulation of prespore transcripts [Mehdy et al., 19831.To assess the universality of the cycloheximide effect, we first examined the ability of several other translational inhibitors to block protein synthesis in developing amoebae. In experiments detailed elsewhere [Ratner et al., 19891, we observed that pactamycin, anisomycin, and emetine (but not puromycin) significantly disrupted protein synthesis in developing amoebae. Of these, emetine (500 pg/ml) exhibited the best combination of effective but fully reversible inhibition of amino acid incorporation together with the retention of normal developmental capability (assessed after removal of the drug). Emetine would seem to be a n effective alternative to cycloheximide a s a n in vivo translational inhibitor [cf. also Okamoto, 19791, while emetine, anisomycin, and pactamycin each may be used to confirm results obtained with cycloheximide. Effects of Translational Inhibitors Upon mRNA Levels Cycloheximide is known to prevent the cyclic AMPdependent reaccumulation of prespore, but not prestalk or nonspecific, gene transcripts in disaggregated cells [Mehdy et al., 19831. Our first objective was to confirm the effects of cycloheximide while examining additional clones. Developmental genes were induced by allowing NC4 amoebae to differentiate on the surface of a pad until they reached the ‘‘finger” stage. Cells were then washed from the pads, disaggregated, and
116
BENEDICT ET AL. A
C Y C L O H E XlMlDE 0
3
2;s:
0
3 2 2 8 8
80
53
37
10 0
3
53 3
0
-5 5+ -8 +8
-5
0 3
8
-
8
8
+ - - +
28
80
C
5
ANlSOMYClN
OR PACTAMYCIN
8 8 a n
0
3 5
i i
5 : E a
37
82
53
0
3
5 5 - a
5
p
-
8
a
8
8
D
80
Fig. 3. Effects of translational inhibitors upon the accumulation of particular gene transcripts. Cells developed upon pads to the “finger” stage were disaggregated and suspended in buffer (the 0 lanes); after 3 hours (lanes 3) the culture was divided in half, with cyclic AMP added to each half and inhibitor to one half. Additional cell samples were collected at the 5th and 8th hours of incubation in suspension (lanes 5 and 8). RNA purification, electrophoresis, blotting, and hybridization to the probes indicated were as before. Note that the hours indicated in this figure refer to the time of incubation in suspension
and do not correspond to the simple developmental time course of the previous figure. A mRNA accumulation in an uninhibited culture (“5-” and “8-”) and in one exposed to cycloheximide (“5+” and “8+”).B: mRNA accumulation in a n uninhibited culture (“5-” and “8-”) and in one exposed to emetine (“5+”and ‘%+”). C: mRNA accumulation in an uninhibited culture (“5-” and “8-”), in a culture treated with anisomycin (“5a”and “8a”),and in a culture exposed to pactamycin (“5p” and “ 8 ~ ”(Reprinted ). with permission from Ratner et al. 119891
suspended in buffer in the absence of cyclic AMP. After 3 hours of incubation, during which time most developmental mRNA molecules are turned over [Mangiarotti et al., 19821, cyclic AMP was added to the divided cultures with or without accompanying inhibitors. The reaccumulation of particular gene transcripts over the
next few hours was assessed by hybridization of various probes to blots of total cell RNA. Figure 3 establishes that, as predicted by the earlier study, cycloheximide prevented the reaccumulation of the transcript of prespore gene DR53. The DR53 RNA was present at the finger stage of development (‘‘0”
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION hours of incubation after disaggregation), decayed during suspension in the absence of cyclic AMP (‘‘3’’ hours), and reaccumulated after cyclic AMP addition only in the absence of drug (“5%”and “8-” vs. “ 5 i ” and “8 + ”). Similar results were obtained with another prespore gene, DR37 (data not shown). In contrast, renewed expression of prestalk gene DR80 was unaffected by translational inhibition (evident as hybridization in addition to that due to partial persistance of the older transcripts; contrast “5”and “8” with “0” and “3”). Indeed, DR80 transcripts often appear to be “super-induced” when translation is interrupted, a s seen in the 8 hour samples of panels A, B, and C [cf. also Singleton et al., 19881. If the effect of cycloheximide upon prespore transcript accumulation is truly mediated by its disruption of protein synthesis, then the other inhibitors described above should have similar consequences for mRNA accumulation. Figure 3B shows that emetine also prevented the expression of prespore clones DR10,37, and 53 (as well as DR82, not shown). In contrast, the reaccumulation of prestalkenriched DR80 transcripts was completely resistant to emetine treatment. Messenger RNA of the presporebiased but early gene DR28 appeared to be fairly stable under disaggregation conditions, with little if any additional accumulation discernable after cyclic AMP stimulation. Finally, anisomycin and pactamycin, in separate treatments, likewise interfered with the accumulation of each prespore transcript tested, but not that of prestalk gene DR80 (Fig. 3C). In sum, each of the four translational inhibitors blocked mRNA accumulation from every active prespore gene tested, but none inhibited the expression of the prestalk gene DR80.
117
prestalk-biased genes. Figure 4 gives two examples from a representative time course, contrasting vegetative with late, developing nuclei. The reduction of actin transcription, and the turning on of several cell-specific genes can readily be seen. A thorough and quantitative analysis of such data requires densitometry of the autoradiographs (with appropriately varied times of exposure to achieve a linear film response) as well as repetition of the entire experiment. The averaged results of three independent developing cultures appear in Figure 5. The transcriptional activation of the four prespore-specific clones (DR10, 37, 53, and 82) midway through development is apparent and can be contrasted with the earlier expression of the prespore-biased clone DR28. Prestalk-biased clones (including actin [Mehdy et al., 19831) exhibit diverse profiles, with clone DR71 transcription preceding that of clones 9 and 29, while DR80 synthesis appears bimodal. The relative transcription rates of (six of) these genes should be compared with the temporal profiles of mRNA accumulation in Figure 2. The general agreement of the two figures argues that, for these genes, appropriate developmental expression is established by regulation of the rate of transcript initiation, and not that of message decay (cf. Landfear et al., 1982; and contrast Mangiarotti et al., 1985).
Effects of Translational Inhibitors Upon Transcription Nuclei were also prepared from cells which had been allowed to develop on a substratum to the finger stage, at which point the cells were collected, suspended in buffer, and subsequently primed with cyclic AMP in the presence or absence of cycloheximide. Recall that i t is under such circumstances that the in vivo reaccumuDevelopmental Kinetics of Transcription lation of prespore transcripts is absolutely dependent The above work on gene expression is based entirely upon protein synthesis. Figure 6 shows one represenupon Northern blot estimates of mRNA accumulation; tative comparison of transcriptional rates cyclohexmRNA levels must, a priori, reflect both rates of syn- imide, while Table 1 summarizes the results of 5 indethesis and decay. To gain understanding of the mech- pendent experiments. The results indicate that three of anism(s) responsible for the RNA temporal profiles and the four prespore genes which we have studied (clones protein synthesis dependence, we have undertaken 10, 37, and 82) are transcribed only to a very limited measurements of rates of mRNA synthesis by in vitro extent in the presence of cycloheximide, in comparison nuclear transcription assays. Purified nuclei elongate with prestalk genes. (Cycloheximide reduces prestalk in vitro gene transcripts whose synthesis was initiated transcription 2-5 fold, but prespore transcription is inin vivo, while de novo initiation is minimal [Jacobson hibited 20 fold or considerably more.) Thus the protein et al., 19741. Hybridization of the RNA product of these factorb) whose concomitant synthesis is necessary for reactions with Dictyostelium cDNA plasmids spotted the accumulation of these prespore mRNAs acts a t the onto nitrocellulose filters is proportional to the amount level of gene transcription. Gene 53 would appear to be unusual: despite mRNA of each gene’s transcript that was produced [Landfear et al., 19821. Relative rates of transcription are deter- accumulation kinetics and drug sensitivity equivalent mined by two-dimensional densitometric scanning of to that of the other three prespore genes, DR53 was the autoradiographs and measurement of the inte- insensitive to cycloheximide a t the level of mRNA synthesis. We first surmised that the drugs caused lability grated spot densities. To examine the developmental kinetics of gene tran- of the DR53 message, but experiments directly meascription, nuclei were prepared from amoebae devel- suring transcript decay were, to our surprise, unioped for varying times and their in vitro RNA products formly negative (i.e., DR53 decay rates were actually hybridized to a collection of prespore-specific and decreased by translational inhibition; DeSilver, Bene-
*
118
BENEDICT ET AL.
2 0 hr
0 hr A
9 10 28 2 9 37
41 5 3 71 80 8 2 P
A
9 10 28 2 9 37
41 53 71 80 8 2
P
Fig. 4. Autoradiographic comparison of in vitro nuclear transcripts of vegetative and 20 hour developing cells. NC4 amoebae were harvested from vegetative plates and developed upon pads for up to 26 hours. At 4 hour intervals, cells were harvested and nuclei prepared [Landfear et al., 1982; Alexander and Shinnick, 19851; nuclei were snap-frozen and stored at -70°C. RNA synthesized in vitro [Landfear et al., 1982; Alexander and Shinnick, 19851 was hybridized to triplicate dots of plasmid DNA of the cDNA clones indicated. “ A repre-
sents the plasmid pcDdActinB1 [Bender et al., 19781. “p” represents plasmid pBR322, used as a control to quantitate non-specific hybridization. The figure shows only two representative time points of the developmental time course. For purposes of quantitation, several autoradiographic exposures were analyzed (avoiding the problem of film saturation which affects actin here). Further methodological details will be presented elsewhere [DeSilver, Benedict, and Ratner, in preparation].
dict, and Ratner, in preparation). The resolution of this paradox seems to be subtle differences in developmental timing. Recent experiments indicate that the dependence of DR53 expression upon protein synthesis is lost slightly earlier in development (2-3 hours) than is the case for the other prespore clones. Most importantly, by relying upon split cultures to ensure precise developmental synchrony (half of one culture used for mRNA blotting and half for nuclear transcription), we were able to show that the transcriptional and accumulation requirements of clone DR53 for developmental protein synthesis in fact agree with each other [DeSilver, Benedict, and Ratner, in preparation].
genes examined. In contrast, the accumulation of prestalk clone DR80 mRNA was uninhibited by any of the drugs used (again consistent with the lack of effect
DISCUSSION The earlier observation [Mehdy et al., 19831 that cycloheximide blocked the reaccumulation of several prespore mRNA species has now been extended to three additional inhibitors of protein synthesis. Sensitivity to each drug was observed for all four active prespore
Fig. 5. Developmental profiles of transcription of prespore and prestalk genes. Nuclear transcription assays and RNA hybridization were performed as in Figure 4. The data of that experiment (including time points in addition to the two shown in Fig. 4) were combined with data from two additional time courses (independent cultures). Autoradiographs were scanned by 2-D densitometry and the spot densities integrated. Background hybridization to pBR322 was subtracted separately for each hybridization. The number of nuclei used for transcription was kept constant for every time point within a single experiment, rather than normalizing hybridization as a function of transcriptional activity. For each gene shown, the relative extent of transcription 10-100%) was computed throughout each single time course; finally, the average of the three profiles was obtained and is displayed here. Error bars represent the standard error of the mean. Genes 9, 29, 71, 80, and actin are all prestalk-biased in message accumulation in slugs; clone 28 is prespore-biased, and clones 10, 37, 53, and 82 are prespore-specific [data from Mehdy et al., 1983; this work; and D.R. unpublished[. Details of data analysis will appear elsewhere IDeSilver, Benedict, and Ratner, in preparation].
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION
Pr es por e
Prestalk
*:-.L ACT1 N
I
60 40
20.01
80 ..
80 -.
60 ~.
60 .-
40.-
40 .-
20..
20.0.
t
100
c\?
80 -60 -. 40..
20..
0
8
16
24
0
8
100
DR 80
80 . 60.
0
8
16
24
D e v e l o p m e n t a l T i m e (hrs) Fig. 5.
16
24
119
120
BENEDICT ET AL. TABLE 1. Relative Transcription After Cycloheximide Treatment* Gene Prespore 10 37 53 82 28
(Biased) Prestalk biased Actin 9 29 71 80
1
2
Experiment 3
4
5
Average
0.018 0.000 0.316 0.040 0.093
0.023 0.218 0.046 0.717
0.045 0.000 0.309 0.028 0.293
0.028 0.000 0.299 0.038 0.148
0.038 0.551 0.117 0.300
0.030 (0.011) 0.000 (0.000) 0.339 (0.125) 0.054 (0.036) 0.310 (0.245)
0.108 0.200 0.206 0.249 0.322
0.275 0.591 0.200 0.284 0.329
0.121 0.256 0.259 0.334
0.212 1.042 0.230 0.304 0.410
0.244 0.650 0.306 0.459 0.453
0.192 (0.074) 0.621 (0.345) 0.240 (0.043) 0.311 (0.085) 0.370 (0.059)
-
*Relative transcription rates in translationally inhibited cells. The experimental protocol of Figure 6 was followed in five wholly independent replicates. For each experiment, the ratio of transcription + cycloheximidei- cycloheximide was calculated for each gene. The five ratios and their average are presented (standard deviation in parentheses). Transcription of the prespore-specific genes 10,37, and 82 is roughly 10 times more sensitive to cycloheximide than is that of the prestalk genes. However, transcription of prespore gene 53 is relatively insensitive to inhibition of translation a t this developmental stage. (Residual gene 28 transcription may be cycloheximide resistant, but the transcription of this gene is virtually off by this time in development; cf. Fig. 5 . )
of cycloheximide upon other prestalk genes; Mehdy et al., 1983). These observations do not preclude the existence of other classes of genes that arc sensitive to translational inhibitors. Indeed, others have observed that the early addition of cycloheximide to developing cells prevents the appearance of a specific phosphodiesterase transcript which would otherwise accumulate preferentially in prestalk cells [Franke et al., 19871; the observations here serve to support those authors’ assumption that the cycloheximide target is protein synthesis. Nor is there any conflict between our results and the fact that transcripts of some early developmental genes accumulate preferentially when protein synthesis is blocked [Singleton et al., 19881. The four inhibitors used in the present study differ in structure and mode of action [reviewed extensively in Gale et al., 19811. Variously, they target the small or large ribosomal subunit; inhibit translocation, transpeptidation, or initiation; and either stabilize or deplete polyribosomes [cf. discussion in Ratner et al., 19891. Thus the state of the translational machinery of amoebae treated with the four inhibitors is likely to be different in each case, the obvious common feature being the absence of protein synthesis. A priori, either mRNA synthesis or stabilization could have been the component dependent upon developmental protein synthesis. The data of Table 1 establish that it is transcription of prespore genes which requires concomitant translation. The seeming exception, DR53, also displays transcriptional sensitivity, but t h a t result holds only if protein synthesis is arrested somewhat earlier than in the case of the other three prespore genes studied [DeSilver, Benedict, and Ratner, in preparation]. The simplest interpretation of
these results is that prespore transcription requires absolutely the activity of one or more developmentally regulated polypeptides, i.e. trans-acting transcriptional factors. There is, of course, much precedent for the role of transcriptional activators in the regulated expression of diverse eukaryotic genes [Maniatis et al., 19871. In D. discoideum, there is now biochemical evidence implicating such a protein in the regulation of a prestalkenriched cathepsin [Hjorth et al., 19891. Our experiments suggest that one should expect to find analogous proteins necessary for prespore transcription. Little more can be inferred about the nature of such proteins, except that, a s each of the prespore genes we examined was expressed to some degree a t the time of disaggregation, the critical activator protein must have itself begun to accumulate by that time, and, furthermore, must itself be labile in rapidly shaken suspension culture in the absence of cyclic AMP. That protein could then reaccumulate to effective levels after cyclic AMP stimulation only in the absence of translational inhibitors. Gene 53 becomes transcriptionally resistant to cycloheximide somewhat earlier in development than do the other prespore clones 10, 37, and 82 (Table 1, and Results). The simplest interpretation would be that the essential 53 transcriptional activator is a protein distinct from that needed for transcription of the other genes. However, this is puzzling in so far as, by both Northern blotting (Fig. 2 ) and transcriptional profiles (Fig. 5 ) , the regulated expression of gene 53 seems quite comparable to that of the other three genes. Perhaps the expression of all four genes is synchronized by nonprotein, physiological factors (positive or negative)
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION
Cyc l o h e x i m i d e
Control A
9 10 28 29 37
A 53 71 7 7 80 82
121
P
A
9
10 28 29 37
B 53 71 77 80 82 P
Fig. 6. Effects of translational inhibition upon transcription of prespore and prestalk genes. NC4 cells, allowed to develop on pads to the finger stage, were harvested, disaggregated, and suspended in buffer in the absence of cyclic AMP. After 3 hours, 300 pM cyclic AMP, alone (A) or with 500 pgiml cycloheximide (B), was added and the cells incubated for another 2 hours. Nuclei were then prepared and RNA
transcribed and hybridized to the clones indicated. The designation of clones is as given in the legend to Figure 4. The relative cycloheximide sensitivity (comparing panels A and B) of prespore clones 10, 37, and 82 can be contrasted with the resistance to inhibition of the prestalk clones and of the unusual prespore clone 53 [DeSilver, Benedict, and Ratner, in preparation].
which are unaffected by translational inhibition. Resolution of this and other points requires, as a next step, the isolation of regulatory regions flanking the prespore genes.
Amara JF, Lodish HF (1987): Specific mRNA destabilization in Dictyostelium discoideum requires RNA synthesis. Mol Cell Biol 7: 4585-4588. Barklis E, Lodish HF (1983): Regulation of Dictyostelium discoideum mRNAs specific for prespore or prestalk cells. Cell 32:1139-1148. Bender W, Davidson N, Kindle KL, Taylor WC, Silverman M, Firtel RA (1978):The structure of M6, a recombinant plasmid containing Dictyostelium DNA homologous to actin messenger RNA. Cell 15: 779-788. Borth W, Ratner D (1983): Different synthetic profiles and developmental fates of prespore versus prestalk proteins of Dictyostelium. Differentiation 24:213-219. Chisholm RL, Barklis E, Lodish HF (1984):Mechanism of sequential induction of cell-type specific mRNAs in Dictyostelium differentiation. Nature 310:67-69. Datta S, Firtel RA (1988): An 80-bp cis-acting regulatory region controls CAMPand development regulation of a prestalk gene in Dictyostelium. Genes Dev 2:294-304. Early A, Williams J (1988): A Dictyostelium prespore-specific gene is transcriptionally repressed by DIF in vitro. Development 103:519524. Franke J, Podgorski GJ, Kessin RH (1987): The expression of two transcripts of the phosphodiesterase gene during the development of Dictyostelium drscoideum. Dev Biol 124:504-511. Gale EF, Cundliffe E, Reynolds PE, Richmond MH, Waring MJ (1981):“The Molecular Basis of Antibiotic Action.” New York: John Wiley & Sons. Hjorth-AL, Khanna NC, Firtel RA (1989): A trans-acting factor re-
ACKNOWLEDGMENTS We are grateful to Kenneth Katz for helpful suggestions during the course of these studies. This work was supported by grants DCB-8616136 and DCB-8911049 from the National Science Foundation. NOTE ADDED IN PROOF Recent hybridization studies [D.D. and D.R., unpublished] have shown that clone DR53 corresponds to the gene described previously as D19 [Barklis and Lodish, 1983; Early and Williams, 19881. REFERENCES Alexander S, Shinnick TM (1985):Specific regulation of transcription of the discoidin gene family in Dictyostelium discoideum. Mol Cell Biol 5:984-990.
122
BENEDICT ET AL.
quired for CAMP-induced gene expression in Dictyostelium is regulated developmentally and induced by CAMP.Genes Dev 3:747-759. Iwabuchi M, Mizukami Y, Sameshima M (1971):Synthesis of precursor molecules of ribosomal RNA in the cellular slime mold Dictyostelium discoideum. Biochim Biophys Acta 228:693-700. Jacobson A, Firtel RA, Lodish HF (1974): Synthesis of messenger and ribosomal RNA precursors in isolated nuclei of the cellular slime mold Dictyostelium discoideum. J Mol Biol 82:213-230. Jermyn KA, M, Kay RR, Williams JG (1987): Two distinct classes of prestalk-enriched mRNA sequences in Dictyostelium discoideum. Development 100:745-755. Landfear SM, Lefebvre P, Chung S, Lodish HF (1982):Transcriptional m control ofgene expression during development o f ~ i c t y o s t e ~ i udiscoideum. Mol Cell Biol 2:1417-1426. ~ ~G, ~ ~~ S, ~ l i ~R,f ~~ ~~i ~p,~ d ceccare1li ~ ~~~~ A, d H~~~~ BD (1989):~ ~of specific ~ m~~~ l destabilization ~ dur~ ing Dictyostelium development. Development 106:473-481. Mangiarotti G, Ceccarelli A, Lodish HF (1983): cyclic AMP stabilizes a class of deve~opmental~y regulated Dictyostelium discoideum mRNAs. Nature 301:616-618. Mangiarotti G, Giorda R, Ceccarelli A, Perlo C (1985): mRNA stabilization controls the expression of a class of developmentally regulated genes in Dictyostelium discoideum. Proc Natl Acad Sci USA 82:5786-5790. Mangiarotti G, Lefebvre P, Lodish HF (1982): Differences in the stability of developmentally regulated mRNAs in aggregated and disaggregated Dictyostelium discoideum cells. Dev Biol 89:82-91. Maniatis T, Goodbourn S, Fisher J A (1987): Regulation of inducible and tissue-specific gene expression. Science 236:1237-1245. Manrow R, Jacobson A (1988): mRNA decay rates in late-developing Dictyostelium discoideum cells are heterogeneous, and cyclic AMP does not act directly to stabilize cell-type-specific mRNAs. Mol Cell Biol 8:4088-4097. McMahon D (1975): Cycloheximide is not a specific inhibitor of protein synthesis in vivo. Plant Physiol 558155821.
Mehdy MC, Ratner D, Firtel RA (1983): Induction and modulation of cell-type-specific gene expression in Dictyostelium. Cell 32:763771. Okamoto K (1979):Induction of cyclic AMP phosphodiesterase by disaggregation the multicellular complexes of Dictyostelium discoideum. Eur J Biochem 93:221-227. Oyama M, Blumberg DD (1986): Changes during differentiation in requirements for CAMPfor expression of cell-type-specific mRNAs in the c e h l a r slime mould, Dictyostelium discoideum. Dev Biol 117:550-556. Pears CJ, Williams J G (1988): Multiple copies of a G-rich element upstream ofa CAMP-inducibleDictyostelium gene are necessary but not sufficient for efficient gene expression. Nucleic Acids Res 16: 8467-8486. Inhibitors of protein synthesis. ~iPestka ~ s i(1977): ~ ~ ~ i In Weissbach H, i Pestka ~s (eds): “Molecular Mechanisms of Protein Biosynthesis.” New York: Academic Press, pp 467-553. Ratner D, Borth W (1983): Comparison of differentiating Dictyostelium discoideum types separated by an improved method Of density gradient centrifugation. Exp Cell Res 143:l-13. Ratner DI, Pentz WH, Pelletier DA (1989): Prespore gene expression in Dictyostelium requires protein synthesis, Biochim Biophys Acts 1o08:71-78, Singleton c K , Manning SS, Feng (1988): Effect of protein synthesis inhibition on gene expression during early development of Dictyostelium discoideum, ~~1 cell ~ , 8:10-16, ~ l Timberlake W, Griffin D (1973): Direct inhibition of the uptake of proline by cycloheximide. Biochem Biophys Res Commun 54:216221. Williams J , Ceccarelli A, McRobbie S, Mahjbubani H, Kay RR, Early A, Berks M, Jermyn KA (1987): Direct induction of Dictyostelium prestalk gene expression by DIF provides evidence that DIF is a morphogen. Cell 49:185-192. Williams J G (1988):The role of diffusible molecules in regulating the cellular differentiation of Dictyostelium discoideum. Development 103:l-16.
Developmental Protein Synthesis Is Required for the Transcription of Dictyostelium Prespore Genes M.A. BENEDICT, D.A. DESILVER, D.E. PELLETIER, W.H. PENTZ, AND D.I. RATNER Department of Biology, Amherst College, Amherst, Massmhusetts and Lodish, 1983; Chisholm et al., 1984; Mehdy et al., 1983; Oyama and Blumberg, 19861. Cell-specific gene expression is also modulated by the effects of ammonia, “DIF,” and other low molecular weight metabolites still chemically unidentified [reviewed in Williams, 19881. Finally, if normal intercellular contacts are prevented by forcing cells to develop in rapidly shaken suspension, prestalk transcripts accumulate as usual, but (the majority of) prespore genes are not expressed [Chisholm et al., 1984; Mehdy et al., 19831. There is evidence that protein synthesis may be necessary for the transcription of prespore genes and/or for the stabilization of their transcripts. The translational inhibitor cycloheximide prevented the reaccumulation of prespore mRNA species in developing amoebae which had been disaggregated and then stimulated with cyclic AMP [Mehdy et al., 19831.In contrast to the sensitivity of prespore transcripts, prestalk and nonspecific messages (the latter present at equal levels in both cell types) were insensitive to cycloheximide. No conclusion concerning the role of protein synthesis in prespore gene expression could be drawn, however, because of the possibility that the particular inhibitor used had directly affected other cellular processes. Cycloheximide is known to block maturation of ribosomal RNA precursors in Dictyostelium [Iwabuchi et al., 19711, and to disrupt DNA and RNA synthesis in many systems [reviewed in Pestka, 19771. In several cases, the effect of cycloheximide upon DNA and RNA synthesis or upon amino acid uptake could not have been a necessary consequence of its disruption of protein synthesis, since other modes of translational inhibition Key words: Dictyostelium discoideum, cyclodid not produce the same range of deleterious effects heximide, emetine, protein synthesis, mRNA accu[McMahon, 1975; Timberlake and Griffin, 19731. mulation, transcription The first aim of this study was to establish whether our previous observations with cycloheximide reflected a bona fide requirement for protein synthesis in the INTRODUCTION accumulation of prespore transcripts. We have comThe development of spore and stalk cells of Dictyo- pared the effects of cycloheximide with those of three stelium discoideum depends upon the regulated expression of diverse genes in the two cell types. Hybridization studies of the accumulation of individual gene transcripts have provided information about the developmental kinetics and physiological regulators of gene Received for publication July 31, 1990. expression. Cyclic AMP, for example, is necessary for the induction of all the prespore-specific genes studied Address reprint requests to Dr. D.I. Ratner, Department of Biology, to date, and for a major class of prestalk genes [Barklis Amherst College, Amherst, MA 01002. It has been established previABSTRACT ously that the maintenance of expression of prespore-specific genes of Dictyostelium discoideum is prevented by the translational inhibitor cycloheximide. The drug had no effect upon the level of transcripts of the other genes examined, prestalkspecific or cell type-nonspecific. However, the interpretation of this result is open to question, because of possible nonspecific effects of cycloheximide. We have now characterized the cellular specificity and temporal profiles of mRNA accumulation of additional Dictyostelium cDNA clones, and have examined other inhibitors of in vivo protein synthesis. Four structurally and mechanistically distinct translational inhibitors each prevented the reaccumulation of prespore transcripts in cyclic AMP-primed, disaggregated amoebae. These results establish the importance of developmental protein synthesis in the accumulation of prespore gene transcripts. Nuclear run-on transcription assays were used to learn whether protein synthesis is required primarily for mRNA synthesis or transcript stability. A transcriptional time course first demonstrated that the abundance of these cellspecific transcripts during development mirrors their rates of synthesis. Significantly, the protein synthesis requirement of the prespore genes examined also occurs at the level of mRNA transcription, implying the existence of one or more developmentally regulated transcriptional activators.
0 1991 WILEY-LISS, INC.
114
BENEDICT ET AL.
other translational inhibitors, anisomycin, pactamycin, and emetine. For each inhibitor studied, prespore transcript accumulation was invariably dependent upon concomitant protein synthesis. The absence of prespore mRNA in treated cultures could reflect, a priori, a lack of prespore transcription, or a heightened lability of these particular transcripts. The modulation of Dictyostelium gene activity is known to depend upon both of these mechanisms under different circumstances. Cyclic AMP in fact both stimulates the transcription of cell-specific genes [Landfear et al., 1982; Pears and Williams, 1988; Datta and Firtel, 1988; Manrow and Jacobson, 19881 and stabilizes many of these same transcripts [Mangiarotti et al., 1983; Landfear et al., 19821, though the stabilization may be transient [Manrow and Jacobson, 19881. DIF serves as a transcriptional activator of some prestalk genes [Williams et al., 19871 while a t the same time repressing the transcription of at least one prespore gene [Early and Williams, 19881. The dissociation of multicellular aggregates, on the other hand, results in the decay of some prestalk and many prespore transcripts [Barklis and Lodish, 1983; Mehdy et al., 1983; Amara and Lodish, 1987; Mangiarotti et al., 19891. Thus the requirement of prespore genes for developmental protein synthesis could be attributed either to transcriptional andlor post-transcriptional events. To choose between these alternatives, we performed a series of nuclear run-on transcription assays. The results described here show the importance of transcriptional mechanisms in determining both the temporal profile of the genes studied and especially the dependence of prespore mRNA levels upon developmental protein synthesis.
MATERIALS AND METHODS Methods for the growth and development of D . discoideum strain NC-4, for RNA isolation and hybridization, and for the identification of cell-specific cDNA clones have recently been published [Ratner et al., 19891. Procedures for the determination of rates of transcription a r e outlined in the legends to Figures 4-6 and in Table 1; they will be more thoroughly described elsewhere [DeSilver, Benedict, and Ratner, in preparation]. RESULTS Identification of Cell-Specific Genes Our original study of cell-specific gene expression employed a limited set of genes identified initially in terms of their temporal specificity and only later examined individually for cell-type specificity [Mehdy et al., 19831. We have since directly screened a developmental cDNA library for additional prespore- and prestalk-specific clones [Ratner et al., 19891.Colony hybridization to probes reverse transcribed from cell-specific poly-A RNA identified candidate clones corre+
P +Qo
Qy
* C lone 10
80 28 Fig. 1. Cell-type specificity of cloned genes revealed by hybridization to total cellular RNA. Equal quantities of total RNA purified from prespore or prestalk cells [Mehdy et al., 1983; Ratner and Borth, 19831 were fractionated by agarose gel electrophoresis and transferred to a nylon membrane. The purity of the cell populations used as a source of RNA, as judged by immunofluorescence, was 98% for prespore cells and 85% for prestalk. RNA blots were hybridized to the nick-translated plasmids indicated. Transcript sizes were estimated a t DR10, 2.2 kb; DR28, 0.9 kb; DR37, 1.7 kb; DR53, 0.5 kb; DR80, 1.6 kb; and DR82,2.1 kb. (Reprinted with the publisher's permission from Ratner et al. 119891
sponding to prespore or prestalk mRNA species. Dot blot hybridizations to immobilized DNA minipreps of promising clones afforded further indication of cell specificity. However, specificity is best demonstrated by RNA blotting experiments, in which radiolabeled cDNA plasmids are hybridized to size-fractionated total RNA isolated from the two purified cell types. Figure 1 shows the prestalk bias of DR80 and the prespore specificity of clones 10, 37, 53, and 82. The very slight hybridization of the four prespore genes to the prestalk RNA preparation is readily attributable to the measured impurity of the prestalk cell population (see Fig. 1 legend). DR28 clearly is also prespore biased, though the bias appears less than that of the other prespore
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION CLONE
HOURS OF DEVELOPMENT 0
3
6
9
12
15 18 21
10
37
53
82 28
80 0
3
6
9
12
15 18
21
Fig. 2. Developmental profile of mRNA accumulation. Amoebae grown upon bacterial lawns were harvested, washed, and placed upon buffered pads for development. At the times indicated above each lane, amoebae were collected and total RNA extracted; the fractionated and blotted RNA was then hybridized to each of the probes shown. (Reprinted with permission from Ratner et al. [19891
genes and must therefore reflect some production of DR28 transcripts in cells identified as prestalk. Prespore-specific clone DRlO has been described previously a s clone 2-H3 [Mehdy et al., 19831. While it was possible, a priori, that some of the prespore cDNA clones were redundant isolates of a single gene, hybridizations with mixed probes established the unique size of each clone’s transcript (see Fig. 1, legend). Thus, each of the prespore clones described corresponds to a different gene.
Developmental Kinetics of mRNA Accumulation We have determined the developmental profile of mRNA accumulation for each of the genes presented above, by allowing amoebae to develop upon pads for varying times before they were harvested and RNA prepared. Figure 2 reveals the extent of hybridization of the various cDNA probes to their gene transcripts. The four prespore genes, DR 10, 37, 53, and 82, are similar to one another in their times of expression,
115
though subtle differences exist as to the relative extent of accumulation within the active period. Transcripts of these genes were first detected at 9 hours of development, abundant between 12 and 18 hours, and much reduced by 21 hours. These results are generally consistent with the profile of (class 11) prespore genes described previously [Mehdy et al., 1983; Chisholm et al., 19841. Expression of prestalk gene DR80 occurred a t very low levels even in vegetative and early developing cells, and at much higher levels after 6 hours. The appearance of two maxima of accumulation, with the minor peak a t 9 hours and the major peak at 18 hours, is reproducible. Gene DR28 is expressed appreciably a t all stages, but its transcript is considerably enriched between 9 and 15 hours of development. The temporal profiles of expression of clones 80 and 28 do not agree closely with those of the major gene categories described elsewhere. Since both genes are expressed early in development, and even in vegetative cells, the cell type bias evident when the RNA of migrating slug cells is examined (as in Fig. 1) presumably reflects differential transcriptional shut-off andlor transcript lability in the two precursor cell populations [Borth and Ratner, 1983; Jermyn et al., 19871.
Inhibitors of In Vivo Protein Synthesis Cycloheximide is the inhibitor customarily used to disrupt in vivo protein synthesis in Dictyostelium and was the only drug used previously to prevent the accumulation of prespore transcripts [Mehdy et al., 19831.To assess the universality of the cycloheximide effect, we first examined the ability of several other translational inhibitors to block protein synthesis in developing amoebae. In experiments detailed elsewhere [Ratner et al., 19891, we observed that pactamycin, anisomycin, and emetine (but not puromycin) significantly disrupted protein synthesis in developing amoebae. Of these, emetine (500 pg/ml) exhibited the best combination of effective but fully reversible inhibition of amino acid incorporation together with the retention of normal developmental capability (assessed after removal of the drug). Emetine would seem to be a n effective alternative to cycloheximide a s a n in vivo translational inhibitor [cf. also Okamoto, 19791, while emetine, anisomycin, and pactamycin each may be used to confirm results obtained with cycloheximide. Effects of Translational Inhibitors Upon mRNA Levels Cycloheximide is known to prevent the cyclic AMPdependent reaccumulation of prespore, but not prestalk or nonspecific, gene transcripts in disaggregated cells [Mehdy et al., 19831. Our first objective was to confirm the effects of cycloheximide while examining additional clones. Developmental genes were induced by allowing NC4 amoebae to differentiate on the surface of a pad until they reached the ‘‘finger” stage. Cells were then washed from the pads, disaggregated, and
116
BENEDICT ET AL. A
C Y C L O H E XlMlDE 0
3
2;s:
0
3 2 2 8 8
80
53
37
10 0
3
53 3
0
-5 5+ -8 +8
-5
0 3
8
-
8
8
+ - - +
28
80
C
5
ANlSOMYClN
OR PACTAMYCIN
8 8 a n
0
3 5
i i
5 : E a
37
82
53
0
3
5 5 - a
5
p
-
8
a
8
8
D
80
Fig. 3. Effects of translational inhibitors upon the accumulation of particular gene transcripts. Cells developed upon pads to the “finger” stage were disaggregated and suspended in buffer (the 0 lanes); after 3 hours (lanes 3) the culture was divided in half, with cyclic AMP added to each half and inhibitor to one half. Additional cell samples were collected at the 5th and 8th hours of incubation in suspension (lanes 5 and 8). RNA purification, electrophoresis, blotting, and hybridization to the probes indicated were as before. Note that the hours indicated in this figure refer to the time of incubation in suspension
and do not correspond to the simple developmental time course of the previous figure. A mRNA accumulation in an uninhibited culture (“5-” and “8-”) and in one exposed to cycloheximide (“5+” and “8+”).B: mRNA accumulation in a n uninhibited culture (“5-” and “8-”) and in one exposed to emetine (“5+”and ‘%+”). C: mRNA accumulation in an uninhibited culture (“5-” and “8-”), in a culture treated with anisomycin (“5a”and “8a”),and in a culture exposed to pactamycin (“5p” and “ 8 ~ ”(Reprinted ). with permission from Ratner et al. 119891
suspended in buffer in the absence of cyclic AMP. After 3 hours of incubation, during which time most developmental mRNA molecules are turned over [Mangiarotti et al., 19821, cyclic AMP was added to the divided cultures with or without accompanying inhibitors. The reaccumulation of particular gene transcripts over the
next few hours was assessed by hybridization of various probes to blots of total cell RNA. Figure 3 establishes that, as predicted by the earlier study, cycloheximide prevented the reaccumulation of the transcript of prespore gene DR53. The DR53 RNA was present at the finger stage of development (‘‘0”
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION hours of incubation after disaggregation), decayed during suspension in the absence of cyclic AMP (‘‘3’’ hours), and reaccumulated after cyclic AMP addition only in the absence of drug (“5%”and “8-” vs. “ 5 i ” and “8 + ”). Similar results were obtained with another prespore gene, DR37 (data not shown). In contrast, renewed expression of prestalk gene DR80 was unaffected by translational inhibition (evident as hybridization in addition to that due to partial persistance of the older transcripts; contrast “5”and “8” with “0” and “3”). Indeed, DR80 transcripts often appear to be “super-induced” when translation is interrupted, a s seen in the 8 hour samples of panels A, B, and C [cf. also Singleton et al., 19881. If the effect of cycloheximide upon prespore transcript accumulation is truly mediated by its disruption of protein synthesis, then the other inhibitors described above should have similar consequences for mRNA accumulation. Figure 3B shows that emetine also prevented the expression of prespore clones DR10,37, and 53 (as well as DR82, not shown). In contrast, the reaccumulation of prestalkenriched DR80 transcripts was completely resistant to emetine treatment. Messenger RNA of the presporebiased but early gene DR28 appeared to be fairly stable under disaggregation conditions, with little if any additional accumulation discernable after cyclic AMP stimulation. Finally, anisomycin and pactamycin, in separate treatments, likewise interfered with the accumulation of each prespore transcript tested, but not that of prestalk gene DR80 (Fig. 3C). In sum, each of the four translational inhibitors blocked mRNA accumulation from every active prespore gene tested, but none inhibited the expression of the prestalk gene DR80.
117
prestalk-biased genes. Figure 4 gives two examples from a representative time course, contrasting vegetative with late, developing nuclei. The reduction of actin transcription, and the turning on of several cell-specific genes can readily be seen. A thorough and quantitative analysis of such data requires densitometry of the autoradiographs (with appropriately varied times of exposure to achieve a linear film response) as well as repetition of the entire experiment. The averaged results of three independent developing cultures appear in Figure 5. The transcriptional activation of the four prespore-specific clones (DR10, 37, 53, and 82) midway through development is apparent and can be contrasted with the earlier expression of the prespore-biased clone DR28. Prestalk-biased clones (including actin [Mehdy et al., 19831) exhibit diverse profiles, with clone DR71 transcription preceding that of clones 9 and 29, while DR80 synthesis appears bimodal. The relative transcription rates of (six of) these genes should be compared with the temporal profiles of mRNA accumulation in Figure 2. The general agreement of the two figures argues that, for these genes, appropriate developmental expression is established by regulation of the rate of transcript initiation, and not that of message decay (cf. Landfear et al., 1982; and contrast Mangiarotti et al., 1985).
Effects of Translational Inhibitors Upon Transcription Nuclei were also prepared from cells which had been allowed to develop on a substratum to the finger stage, at which point the cells were collected, suspended in buffer, and subsequently primed with cyclic AMP in the presence or absence of cycloheximide. Recall that i t is under such circumstances that the in vivo reaccumuDevelopmental Kinetics of Transcription lation of prespore transcripts is absolutely dependent The above work on gene expression is based entirely upon protein synthesis. Figure 6 shows one represenupon Northern blot estimates of mRNA accumulation; tative comparison of transcriptional rates cyclohexmRNA levels must, a priori, reflect both rates of syn- imide, while Table 1 summarizes the results of 5 indethesis and decay. To gain understanding of the mech- pendent experiments. The results indicate that three of anism(s) responsible for the RNA temporal profiles and the four prespore genes which we have studied (clones protein synthesis dependence, we have undertaken 10, 37, and 82) are transcribed only to a very limited measurements of rates of mRNA synthesis by in vitro extent in the presence of cycloheximide, in comparison nuclear transcription assays. Purified nuclei elongate with prestalk genes. (Cycloheximide reduces prestalk in vitro gene transcripts whose synthesis was initiated transcription 2-5 fold, but prespore transcription is inin vivo, while de novo initiation is minimal [Jacobson hibited 20 fold or considerably more.) Thus the protein et al., 19741. Hybridization of the RNA product of these factorb) whose concomitant synthesis is necessary for reactions with Dictyostelium cDNA plasmids spotted the accumulation of these prespore mRNAs acts a t the onto nitrocellulose filters is proportional to the amount level of gene transcription. Gene 53 would appear to be unusual: despite mRNA of each gene’s transcript that was produced [Landfear et al., 19821. Relative rates of transcription are deter- accumulation kinetics and drug sensitivity equivalent mined by two-dimensional densitometric scanning of to that of the other three prespore genes, DR53 was the autoradiographs and measurement of the inte- insensitive to cycloheximide a t the level of mRNA synthesis. We first surmised that the drugs caused lability grated spot densities. To examine the developmental kinetics of gene tran- of the DR53 message, but experiments directly meascription, nuclei were prepared from amoebae devel- suring transcript decay were, to our surprise, unioped for varying times and their in vitro RNA products formly negative (i.e., DR53 decay rates were actually hybridized to a collection of prespore-specific and decreased by translational inhibition; DeSilver, Bene-
*
118
BENEDICT ET AL.
2 0 hr
0 hr A
9 10 28 2 9 37
41 5 3 71 80 8 2 P
A
9 10 28 2 9 37
41 53 71 80 8 2
P
Fig. 4. Autoradiographic comparison of in vitro nuclear transcripts of vegetative and 20 hour developing cells. NC4 amoebae were harvested from vegetative plates and developed upon pads for up to 26 hours. At 4 hour intervals, cells were harvested and nuclei prepared [Landfear et al., 1982; Alexander and Shinnick, 19851; nuclei were snap-frozen and stored at -70°C. RNA synthesized in vitro [Landfear et al., 1982; Alexander and Shinnick, 19851 was hybridized to triplicate dots of plasmid DNA of the cDNA clones indicated. “ A repre-
sents the plasmid pcDdActinB1 [Bender et al., 19781. “p” represents plasmid pBR322, used as a control to quantitate non-specific hybridization. The figure shows only two representative time points of the developmental time course. For purposes of quantitation, several autoradiographic exposures were analyzed (avoiding the problem of film saturation which affects actin here). Further methodological details will be presented elsewhere [DeSilver, Benedict, and Ratner, in preparation].
dict, and Ratner, in preparation). The resolution of this paradox seems to be subtle differences in developmental timing. Recent experiments indicate that the dependence of DR53 expression upon protein synthesis is lost slightly earlier in development (2-3 hours) than is the case for the other prespore clones. Most importantly, by relying upon split cultures to ensure precise developmental synchrony (half of one culture used for mRNA blotting and half for nuclear transcription), we were able to show that the transcriptional and accumulation requirements of clone DR53 for developmental protein synthesis in fact agree with each other [DeSilver, Benedict, and Ratner, in preparation].
genes examined. In contrast, the accumulation of prestalk clone DR80 mRNA was uninhibited by any of the drugs used (again consistent with the lack of effect
DISCUSSION The earlier observation [Mehdy et al., 19831 that cycloheximide blocked the reaccumulation of several prespore mRNA species has now been extended to three additional inhibitors of protein synthesis. Sensitivity to each drug was observed for all four active prespore
Fig. 5. Developmental profiles of transcription of prespore and prestalk genes. Nuclear transcription assays and RNA hybridization were performed as in Figure 4. The data of that experiment (including time points in addition to the two shown in Fig. 4) were combined with data from two additional time courses (independent cultures). Autoradiographs were scanned by 2-D densitometry and the spot densities integrated. Background hybridization to pBR322 was subtracted separately for each hybridization. The number of nuclei used for transcription was kept constant for every time point within a single experiment, rather than normalizing hybridization as a function of transcriptional activity. For each gene shown, the relative extent of transcription 10-100%) was computed throughout each single time course; finally, the average of the three profiles was obtained and is displayed here. Error bars represent the standard error of the mean. Genes 9, 29, 71, 80, and actin are all prestalk-biased in message accumulation in slugs; clone 28 is prespore-biased, and clones 10, 37, 53, and 82 are prespore-specific [data from Mehdy et al., 1983; this work; and D.R. unpublished[. Details of data analysis will appear elsewhere IDeSilver, Benedict, and Ratner, in preparation].
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION
Pr es por e
Prestalk
*:-.L ACT1 N
I
60 40
20.01
80 ..
80 -.
60 ~.
60 .-
40.-
40 .-
20..
20.0.
t
100
c\?
80 -60 -. 40..
20..
0
8
16
24
0
8
100
DR 80
80 . 60.
0
8
16
24
D e v e l o p m e n t a l T i m e (hrs) Fig. 5.
16
24
119
120
BENEDICT ET AL. TABLE 1. Relative Transcription After Cycloheximide Treatment* Gene Prespore 10 37 53 82 28
(Biased) Prestalk biased Actin 9 29 71 80
1
2
Experiment 3
4
5
Average
0.018 0.000 0.316 0.040 0.093
0.023 0.218 0.046 0.717
0.045 0.000 0.309 0.028 0.293
0.028 0.000 0.299 0.038 0.148
0.038 0.551 0.117 0.300
0.030 (0.011) 0.000 (0.000) 0.339 (0.125) 0.054 (0.036) 0.310 (0.245)
0.108 0.200 0.206 0.249 0.322
0.275 0.591 0.200 0.284 0.329
0.121 0.256 0.259 0.334
0.212 1.042 0.230 0.304 0.410
0.244 0.650 0.306 0.459 0.453
0.192 (0.074) 0.621 (0.345) 0.240 (0.043) 0.311 (0.085) 0.370 (0.059)
-
*Relative transcription rates in translationally inhibited cells. The experimental protocol of Figure 6 was followed in five wholly independent replicates. For each experiment, the ratio of transcription + cycloheximidei- cycloheximide was calculated for each gene. The five ratios and their average are presented (standard deviation in parentheses). Transcription of the prespore-specific genes 10,37, and 82 is roughly 10 times more sensitive to cycloheximide than is that of the prestalk genes. However, transcription of prespore gene 53 is relatively insensitive to inhibition of translation a t this developmental stage. (Residual gene 28 transcription may be cycloheximide resistant, but the transcription of this gene is virtually off by this time in development; cf. Fig. 5 . )
of cycloheximide upon other prestalk genes; Mehdy et al., 1983). These observations do not preclude the existence of other classes of genes that arc sensitive to translational inhibitors. Indeed, others have observed that the early addition of cycloheximide to developing cells prevents the appearance of a specific phosphodiesterase transcript which would otherwise accumulate preferentially in prestalk cells [Franke et al., 19871; the observations here serve to support those authors’ assumption that the cycloheximide target is protein synthesis. Nor is there any conflict between our results and the fact that transcripts of some early developmental genes accumulate preferentially when protein synthesis is blocked [Singleton et al., 19881. The four inhibitors used in the present study differ in structure and mode of action [reviewed extensively in Gale et al., 19811. Variously, they target the small or large ribosomal subunit; inhibit translocation, transpeptidation, or initiation; and either stabilize or deplete polyribosomes [cf. discussion in Ratner et al., 19891. Thus the state of the translational machinery of amoebae treated with the four inhibitors is likely to be different in each case, the obvious common feature being the absence of protein synthesis. A priori, either mRNA synthesis or stabilization could have been the component dependent upon developmental protein synthesis. The data of Table 1 establish that it is transcription of prespore genes which requires concomitant translation. The seeming exception, DR53, also displays transcriptional sensitivity, but t h a t result holds only if protein synthesis is arrested somewhat earlier than in the case of the other three prespore genes studied [DeSilver, Benedict, and Ratner, in preparation]. The simplest interpretation of
these results is that prespore transcription requires absolutely the activity of one or more developmentally regulated polypeptides, i.e. trans-acting transcriptional factors. There is, of course, much precedent for the role of transcriptional activators in the regulated expression of diverse eukaryotic genes [Maniatis et al., 19871. In D. discoideum, there is now biochemical evidence implicating such a protein in the regulation of a prestalkenriched cathepsin [Hjorth et al., 19891. Our experiments suggest that one should expect to find analogous proteins necessary for prespore transcription. Little more can be inferred about the nature of such proteins, except that, a s each of the prespore genes we examined was expressed to some degree a t the time of disaggregation, the critical activator protein must have itself begun to accumulate by that time, and, furthermore, must itself be labile in rapidly shaken suspension culture in the absence of cyclic AMP. That protein could then reaccumulate to effective levels after cyclic AMP stimulation only in the absence of translational inhibitors. Gene 53 becomes transcriptionally resistant to cycloheximide somewhat earlier in development than do the other prespore clones 10, 37, and 82 (Table 1, and Results). The simplest interpretation would be that the essential 53 transcriptional activator is a protein distinct from that needed for transcription of the other genes. However, this is puzzling in so far as, by both Northern blotting (Fig. 2 ) and transcriptional profiles (Fig. 5 ) , the regulated expression of gene 53 seems quite comparable to that of the other three genes. Perhaps the expression of all four genes is synchronized by nonprotein, physiological factors (positive or negative)
PROTEIN SYNTHESIS REQUIREMENT OF GENE TRANSCRIPTION
Cyc l o h e x i m i d e
Control A
9 10 28 29 37
A 53 71 7 7 80 82
121
P
A
9
10 28 29 37
B 53 71 77 80 82 P
Fig. 6. Effects of translational inhibition upon transcription of prespore and prestalk genes. NC4 cells, allowed to develop on pads to the finger stage, were harvested, disaggregated, and suspended in buffer in the absence of cyclic AMP. After 3 hours, 300 pM cyclic AMP, alone (A) or with 500 pgiml cycloheximide (B), was added and the cells incubated for another 2 hours. Nuclei were then prepared and RNA
transcribed and hybridized to the clones indicated. The designation of clones is as given in the legend to Figure 4. The relative cycloheximide sensitivity (comparing panels A and B) of prespore clones 10, 37, and 82 can be contrasted with the resistance to inhibition of the prestalk clones and of the unusual prespore clone 53 [DeSilver, Benedict, and Ratner, in preparation].
which are unaffected by translational inhibition. Resolution of this and other points requires, as a next step, the isolation of regulatory regions flanking the prespore genes.
Amara JF, Lodish HF (1987): Specific mRNA destabilization in Dictyostelium discoideum requires RNA synthesis. Mol Cell Biol 7: 4585-4588. Barklis E, Lodish HF (1983): Regulation of Dictyostelium discoideum mRNAs specific for prespore or prestalk cells. Cell 32:1139-1148. Bender W, Davidson N, Kindle KL, Taylor WC, Silverman M, Firtel RA (1978):The structure of M6, a recombinant plasmid containing Dictyostelium DNA homologous to actin messenger RNA. Cell 15: 779-788. Borth W, Ratner D (1983): Different synthetic profiles and developmental fates of prespore versus prestalk proteins of Dictyostelium. Differentiation 24:213-219. Chisholm RL, Barklis E, Lodish HF (1984):Mechanism of sequential induction of cell-type specific mRNAs in Dictyostelium differentiation. Nature 310:67-69. Datta S, Firtel RA (1988): An 80-bp cis-acting regulatory region controls CAMPand development regulation of a prestalk gene in Dictyostelium. Genes Dev 2:294-304. Early A, Williams J (1988): A Dictyostelium prespore-specific gene is transcriptionally repressed by DIF in vitro. Development 103:519524. Franke J, Podgorski GJ, Kessin RH (1987): The expression of two transcripts of the phosphodiesterase gene during the development of Dictyostelium drscoideum. Dev Biol 124:504-511. Gale EF, Cundliffe E, Reynolds PE, Richmond MH, Waring MJ (1981):“The Molecular Basis of Antibiotic Action.” New York: John Wiley & Sons. Hjorth-AL, Khanna NC, Firtel RA (1989): A trans-acting factor re-
ACKNOWLEDGMENTS We are grateful to Kenneth Katz for helpful suggestions during the course of these studies. This work was supported by grants DCB-8616136 and DCB-8911049 from the National Science Foundation. NOTE ADDED IN PROOF Recent hybridization studies [D.D. and D.R., unpublished] have shown that clone DR53 corresponds to the gene described previously as D19 [Barklis and Lodish, 1983; Early and Williams, 19881. REFERENCES Alexander S, Shinnick TM (1985):Specific regulation of transcription of the discoidin gene family in Dictyostelium discoideum. Mol Cell Biol 5:984-990.
122
BENEDICT ET AL.
quired for CAMP-induced gene expression in Dictyostelium is regulated developmentally and induced by CAMP.Genes Dev 3:747-759. Iwabuchi M, Mizukami Y, Sameshima M (1971):Synthesis of precursor molecules of ribosomal RNA in the cellular slime mold Dictyostelium discoideum. Biochim Biophys Acta 228:693-700. Jacobson A, Firtel RA, Lodish HF (1974): Synthesis of messenger and ribosomal RNA precursors in isolated nuclei of the cellular slime mold Dictyostelium discoideum. J Mol Biol 82:213-230. Jermyn KA, M, Kay RR, Williams JG (1987): Two distinct classes of prestalk-enriched mRNA sequences in Dictyostelium discoideum. Development 100:745-755. Landfear SM, Lefebvre P, Chung S, Lodish HF (1982):Transcriptional m control ofgene expression during development o f ~ i c t y o s t e ~ i udiscoideum. Mol Cell Biol 2:1417-1426. ~ ~G, ~ ~~ S, ~ l i ~R,f ~~ ~~i ~p,~ d ceccare1li ~ ~~~~ A, d H~~~~ BD (1989):~ ~of specific ~ m~~~ l destabilization ~ dur~ ing Dictyostelium development. Development 106:473-481. Mangiarotti G, Ceccarelli A, Lodish HF (1983): cyclic AMP stabilizes a class of deve~opmental~y regulated Dictyostelium discoideum mRNAs. Nature 301:616-618. Mangiarotti G, Giorda R, Ceccarelli A, Perlo C (1985): mRNA stabilization controls the expression of a class of developmentally regulated genes in Dictyostelium discoideum. Proc Natl Acad Sci USA 82:5786-5790. Mangiarotti G, Lefebvre P, Lodish HF (1982): Differences in the stability of developmentally regulated mRNAs in aggregated and disaggregated Dictyostelium discoideum cells. Dev Biol 89:82-91. Maniatis T, Goodbourn S, Fisher J A (1987): Regulation of inducible and tissue-specific gene expression. Science 236:1237-1245. Manrow R, Jacobson A (1988): mRNA decay rates in late-developing Dictyostelium discoideum cells are heterogeneous, and cyclic AMP does not act directly to stabilize cell-type-specific mRNAs. Mol Cell Biol 8:4088-4097. McMahon D (1975): Cycloheximide is not a specific inhibitor of protein synthesis in vivo. Plant Physiol 558155821.
Mehdy MC, Ratner D, Firtel RA (1983): Induction and modulation of cell-type-specific gene expression in Dictyostelium. Cell 32:763771. Okamoto K (1979):Induction of cyclic AMP phosphodiesterase by disaggregation the multicellular complexes of Dictyostelium discoideum. Eur J Biochem 93:221-227. Oyama M, Blumberg DD (1986): Changes during differentiation in requirements for CAMPfor expression of cell-type-specific mRNAs in the c e h l a r slime mould, Dictyostelium discoideum. Dev Biol 117:550-556. Pears CJ, Williams J G (1988): Multiple copies of a G-rich element upstream ofa CAMP-inducibleDictyostelium gene are necessary but not sufficient for efficient gene expression. Nucleic Acids Res 16: 8467-8486. Inhibitors of protein synthesis. ~iPestka ~ s i(1977): ~ ~ ~ i In Weissbach H, i Pestka ~s (eds): “Molecular Mechanisms of Protein Biosynthesis.” New York: Academic Press, pp 467-553. Ratner D, Borth W (1983): Comparison of differentiating Dictyostelium discoideum types separated by an improved method Of density gradient centrifugation. Exp Cell Res 143:l-13. Ratner DI, Pentz WH, Pelletier DA (1989): Prespore gene expression in Dictyostelium requires protein synthesis, Biochim Biophys Acts 1o08:71-78, Singleton c K , Manning SS, Feng (1988): Effect of protein synthesis inhibition on gene expression during early development of Dictyostelium discoideum, ~~1 cell ~ , 8:10-16, ~ l Timberlake W, Griffin D (1973): Direct inhibition of the uptake of proline by cycloheximide. Biochem Biophys Res Commun 54:216221. Williams J , Ceccarelli A, McRobbie S, Mahjbubani H, Kay RR, Early A, Berks M, Jermyn KA (1987): Direct induction of Dictyostelium prestalk gene expression by DIF provides evidence that DIF is a morphogen. Cell 49:185-192. Williams J G (1988):The role of diffusible molecules in regulating the cellular differentiation of Dictyostelium discoideum. Development 103:l-16.
