DEVELOPMENTAL GENETICS 12:196-205 (1991)
The Relative Importance of Transcriptional and Post Transcriptional Regulation of Drosophila Chorion Gene Expression During Oogenesis CHARLES P. ROMANO, JUAN CARLOS MARTINEZ-CRUZADO, AND FOTIS C. KAFATOS Department of Cellular and Developmental Biology, The Biological Laboratories, Harvard University, Cambridge (C.P.R., J.C.M.C., F.C.K.) and The Institute of Molecular Biology and Biotechnology, Research Center of Crete, Crete, Greece (F.C.K.) To determine the relative roles ABSTRACT of transcriptional and post-transcriptional events in establishing the temporal pattern of chorion gene expression in Drosophila, we have examined chorion gene transcription, RNA accumulation, and protein synthesis in follicles of selected pre-, early-, and late-choriogenic stages. Chorion gene transcription was assayed in follicle cell nuclei by nuclear run-on reactions. For the s i5, s 16, s18, 536, and s38 chorion genes, the periods of intense transcription are as predicted from the dynamics of RNA accumulation and protein synthesis, indicating that these genes are primarily regulated at the transcriptional level. In contrast, gene s 19 appears subject to post-transcriptional control at stage 14, when transcription rates are substantially higher than predicted from the observed RNA levels. Transcription of regions between the clustered and tandemly oriented chorion genes was also examined. In contrast to many RNA polymerase II transcribed genes, for the s78 and s36 chorion genes run-on transcription appears to terminate within about 100 base pairs downstream of the polyadenylation sites, corroborating previous reports based on electron microscopy of s36 [Osheim et a/., EMBO J 5:3591-3596, 19861.
Key words: Run-on tra nscription, develop men-
ceeds through the remainder of oogenesis [Spradling and Leys, 19881, chorion gene mRNAs accumulate and decay according to a temporal program that is distinct for each gene [Griffin-Shea et al., 1982; Parks et al., 19861. With the exception of the transient prechoriogenic accumulation of untranslated s15 mRNA, chorion proteins are generally synthesized when the corresponding chorion mRNA accumulates [Petri et al., 1976; Waring and Mahowald, 1979; Thireos et al., 1980; Griffin-Shea et al., 1982; Parks et al., 19861. The chorion genes of Drosophila thus serve as a useful model system for the study of temporally regulated gene expression. The fundamental question of the relative regulatory roles of transcriptional and post-transcriptional events in establishing the temporal pattern of chorion gene expression has yet to be addressed directly and systematically. Indirect evidence favoring transcriptional regulation has been obtained in P-element-mediated transformation experiments. Chorion genes, marked with foreign DNA by internal insertion or fusion, produce temporally regulated transcripts when transferred to different chromosomal sites [Kafatos et al., 1985; Wakimoto et al., 19861; this demonstrates that amplification is not involved in temporal programming but leaves open the possibility that transcribed regions may influence that programming by a post-transcriptional mechanism. However, the early s36 and late s15
tally regulated gene expression, follicle cell nuclei
INTRODUCTION During the final stages of Drosophila oogenesis, the terminally differentiated follicle cells surrounding the oocyte construct the proteinaceous eggshell or chorion, which serves to protect the egg and embryo [reviewed by Mahowald and Kambysellis, 19801.The two clusters of major chorion genes are selectively amplified within the already polyploid follicle cells to produce the massive amounts of chorion proteins in the short time available [Spradling, 1981; reviewed by Delidakis et al., 19891. After the onset of amplification, which pro-
0 1991 WILEY-LISS. INC.
Received for publication August 7, 1990; accepted October 8, 1990. Address reprint requests to Dr. Fotis C. Kafatos, Department of Cellular and Developmental Biology, The Biological Laboratories, Harvard University, 16 Divinity Ave., Cambridge, MA 02138. Dr. Romano is now at Monsanto, Plant Gene Expression GroupAASG, 700 Chesterfield Village Parkway, St. Louis, MO 63198. Dr. Martinez-Cruzado is now a t the Departamento de Biologie, Recinto Universiterio de Mayaguez, Universidad de Puerto Rico Mayaguez, PR 00708.
DROSOPHILA CHORION GENE EXPRESSION
197
promoters alone can direct an alcohol dehydrogenase reporter gene to be expressed in a manner temporally indistinguishable from that of the endogenous s36 and s15 chorion genes [Romano et al., 19881. Moreover, the ability to modify the s15 promoter and thereby change the temporal specificity of the resulting transcripts [Mariani et al., 19881 further argues that this chorion gene is transcriptionally regulated. Chorion gene transcription has also been studied by EM analysis of chromatin spreads in which nascent transcripts of the amplified chorion genes can be directly visualized [Osheim and Miller, 19831. Transcriptionally active, amplified DNAs with the characteristics of the X and 3rd chromosome chorion gene clusters are only observed in follicular nuclei of the stage at which the mRNAs corresponding to the respective clusters accumulate [Osheim and Miller, 19831. However, as positive identification of a chorion cluster by this technique is dependent on its active transcription, these results must be interpreted with caution. The transcriptional activity of higher eukaryotic genes has traditionally been measured via nuclear runon assays. In such assays, transcriptionally engaged RNA polymerase molecules synthesize radiolabeled nascent transcripts, which are quantified to reveal the number of active RNA polymerase molecules on a given gene a t the time the nuclei were isolated [reviewed by Darnell, 19821. Additional information can be obtained from such studies, e.g., that c-myc and cmyb are regulated at the level of transcriptional elongation rather than initiation [Nepveu and Marcu, 1986; Bentley and Groudine, 1986; Bender et al., 19871. Nuclear run-on analysis has rarely been employed in the analysis of tissue specific gene expression in organisms such as Drosophila, as it is often difficult to obtain enough nuclei to perform the assay. However, we have been able to use this approach to examine chorion gene transcription because of our ability to isolate significant numbers of follicles (i.e., egg chambers) of a given choriogenic stage, as well as the fact that each follicle contains about 1,200 polyploid follicle cell nuclei in which the chorion genes are amplified. We report here the temporal pattern, boundaries, and levels of chorion gene transcription in follicle cell nuclei as assayed by nuclear run-on transcription. These results are compared and contrasted with the temporal patterns of chorion mRNA accumulation and protein synthesis.
morphological criteria [King, 19701, until a minimum of about 400 were collected. After they settled into a minimal volume of Ringer's (about 30 pl), and equal volume of Solution A (4mM Mg Acetate, 6 mM CaCl,, 20 mM Tris-HC1, pH 8.0,1% N P 4 0 ,l mM PMSF, and 1 mM DTT) was added and the resultant mixture gently disrupted with five strokes of a loosely fitting Dounce homogenizer. The homogenate was layered over 10 ml of Solution B (1 M sucrose, 5 mM Mg Acetate, 10 mM Tris-HC1 pH 8.0, 0.5 mM DTT, and 0.5 mM PMSF) in a 15 ml Corex Tube and spun at 4,000 RPM for 10 min in a Sorvall HB40 rotor. The pelleted nuclei were resuspended in Solution C (25% glycerol, 4.7 mM Mg Acetate, 50 mM Tris-HC1, pH 8.0, 0.1 mM EDTA, and 5 mM DTT) and spun at 4,000 RPM for 10 min in an Eppendorf microcentrifuge. Pelleted nuclei were resuspended in 10 to 15 pl of Solution C and 10 pl of nuclei were added to 10 p1 of 2X transcription mix (22.8% glycerol, 9.4 mM Mg Acetate, 1.8 mM MnCl,, 54 mM Tris-HC1, pH 8.0,160 mM KC1, and 1mM each of ATP, GTP, and CTP with 250 pCi a-32P-UTP a t 800 Ci/ mmol) at 25°C. In some experiments, nuclei in solution C were preincubated a t 4°C for 5 min with or without 2 pg/ml a-amanitin (Sigma Chemical Co., St. Louis, MO) prior to the transcription reaction. When indicated, nuclei in solution C were added to a n equal volume of 2 x transcription mix at 1.2% sarkosyl (N-lauroyl sarcosine sodium salt; ultra-pure molecular biology grade from IBI) for a final sarkosyl concentration of 0.6%. After 20 min a t 25"C, the run-on reaction was terminated by disruption in Urea:SDS (7 M Urea, 2% SDS, 10 mM Tris-HC1, pH 8.0, 350 mM NaCl, and 1 mM EDTA) and total nucleic acid along with 20 pg carrier tRNA was isolated by phenol:CHCl, extraction followed by ethanol precipitation. Genomic DNA was removed by Promega RQ1 RNAse free DNAse treatment and run-on transcripts were purified by phenol:CHCl, extraction followed by ethanol/NH,Acetate precipitation. About 5 x lo6 cpm of incorporated counts were typically obtained.
MATERIALS AND METHODS Isolation of Nuclear Run On Transcripts From Follicles Dissected ovaries from cold anesthetized Drosophila melanogaster ry506were disrupted in Drosophila Ringer's solution at 4°C by passage through a BSA coated pasteur pipet [Petri et al., 19761. This and all steps up until placing the nuclei in the transcription mix were done a t 4°C. Follicles were staged by hand according to
Southern and Slot Blots RNA was heated in water for 1min at 75"C, brought to 1 0 SSC, ~ and slotted to nylon membrane with a suction manifold. DNA was slotted a s previously described [Romano et al., 19881. Nucleic acid was crosslinked to the filters by UV irradiation [Church and Gilbert, 19841. Southern blots of restriction enzyme digested plasmids were done as reported [Delidakis and Kafatos, 19871.
Isolation of DNA and RNA DNA was isolated from nuclei as described [Delidakis and Kafatos, 19871 but was treated with RNAse prior to slot blotting. Total nucleic acid was isolated from follicles as described [Thireos et al., 19801and was treated with Promega RQ 1 RNAse free DNAse to obtain RNA for slot blot analysis.
198
ROMANO ET AL.
Hybridization Conditions RNA and DNA slot blots were hybridized and washed as described [Romano et al., 19881, except that oligo-labeled fragment probes in excess were used instead of RNA or nick translated probes [Feinberg and Vogelstein, 19831. Run-on transcripts were hybridized in 1 ml of solution to Southern or slot-blot filters with 0.5 pg denatured, immobilized DNA probe per slot or band for at least 24 hr, and were washed as above. Under these conditions, 0.5 pg of immobilized probe was shown to be in excess of run-on transcript. Filters probed with run-on transcripts were treated with 5 kgl ml RNAse A in 2~ SSC a t 37°C for 30 min prior to autoradiography . Densitometry To quantitate autoradiographic signals, filters were exposed without screen or preflash to Kodak XAR-5 film to obtain exposures within the linear response range of the film. The autoradiograms were then scanned with a Helena Laboratories R and D densitometer equipped with a Quick Quant I11 integrating computer. Values obtained from filters probed with run-on transcripts were then normalized to the length of complementary sequence present in the DNA immobilized on the filter. When noted, the length normalized runon transcription values were also normalized to the experimentally determined gene copy number. Relative gene copy numbers were determined by quantitating slot blots of genomic DNA from the nuclei used in the run-on experiments as described above (e.g., see Fig. 4,DNA); the densitometric values from the amplified chorion genes were divided by the value for the unamplified tubulin gene to obtain gene copy numbers. Values obtained from hybridizing immobilized RNAs or DNAs with oligolabelled probes in excess were normalized for the specific activity of those probes; specific activities were calculated as described (Promega Prime-a-Gene Labeling System).
lowed by hybridization with double stranded probes (Fig. 4). The quantitative measurements of stage specific transcription and RNA accumulation obtained in two independent but identical experiments were combined and tabulated in Table 2. Only early (st.12) and late (st.14) choriogenic stage follicles were used, as stages 11 and 13 are too transient and rare for run-on transcription assays. In interpreting the results, it should be noted that choriogenesis is an exceedingly rapid process [ca. 5 hr from stage 11to 14; King, 19701. It should also be noted that the developmentally middle chorion genes ( s l 9 and s16) are maximally expressed during st. 13 while the early genes (s36 and s38) are turning off and the late genes (s18 and s15) are turning on, in terms of RNA accumulation and protein synthesis [Petri et al., 1976; Waring and Mahowald, 1979; Griffin-Shea et al., 1982; Fenerjian et al., 1989; see also Fig. 41.
Nuclear Run-on Transcription of the Major Chorion Genes is RNA Polymerase I1 Dependent and Displays Stage Specific Differences To identify the RNA polymerase activity involved in the synthesis of the chorion gene run-on transcripts, nuclei from stage 1 through 14 follicles were briefly exposed to 2 kg/ml a-amanitin prior t o the nuclear runon reaction. While total incorporation decreased by about 50%, run-on transcription of the chorion genes decreased by over 95% (as estimated for s15; Fig. 2, st.1-14 2 amanitin). As the RNA polymerases I and I11 of Drosophila and other insects are not inhibited by low concentrations of a-amanitin [Roeder, 1976; Greenleaf et al., 19761, this result demonstrates that the chorion run-on transcripts are products of RNA polymerase I1 activity. To determine the level of chorion gene transcription during different stages of oogenesis, nuclear run-on analysis of transcripts from pre-, early-, and latechoriogenic follicle nuclei (stages l-lOA, 12, and 14, respectively) was performed. Unique patterns of transcriptional activity are characteristic of each stage. In RESULTS prechoriogenic stages (Fig. 2, st. l-lOa), high levels of Experimental Strategy histone gene (his) run-on transcripts, along with low The organization of the major chorion gene clusters levels of ribosomal protein 49 (rp), al-tubulin (tb), and of Drosophila melanogaster is shown in Figure 1, and yolk protein 1 (ypl) run-on transcripts are observed, the plasmids used in this study are listed in Table 1. To demonstrating that these nuclei are transcriptionally monitor the stage and strand specificities of transcrip- competent; however, with the exception of very low s38 tion (Fig. 2), single-stranded M13 clones (chorion transcription, no chorion gene run-on transcripts are genes) or double-stranded pUC clones (control genes) detected. Transient prechoriogenic expression of s15, were slot blotted and probed with run-on transcripts of which apparently occurs only under certain conditions, staged nuclei. To determine the physical boundaries of was not detected in either the nuclear run-on assays or transcription, double-stranded chorion clones were di- in the slot blot analysis of stable RNA [Thireos et al., gested with restriction enzymes, Southern blotted, and 1980; Figs. 2 and 41. Since the single quiescent oocyte probed with run-on transcripts from follicular nuclei of nucleus within each follicle is unlikely to yield signifall stages (st.1-14; Fig. 3). Levels of accumulated icant amounts of run-on transcript [Mahowald and chorion RNA in the follicles and relative chorion gene Tiefert, 19701, signals from the His, rp49, and copy numbers in the nuclei used for stage specific run- aTub84B genes may be largely due to run-on tranon reactions were also determined by slot blotting fol- scripts from the giant nurse cell nuclei [Mahowald and
-
-
s 37
s38
s 36
I
1
So
DROSOPHZLA CHORION GENE EXPRESSION
Xb’Bs
BO
I Xb
+----
L
b 1
r
1
11
[
1
s 36 .
K I
Pv BS EC
Xh
J
199
’
-1
Ec
Ec
Stage12 strand specific transcriDtion Transcription boundaries
R N A / D N A Probes Stage 12 mRNA level Gene location 1 kb
I
1
So Ps Ac
I
1
AC Hi
r - L 1
I
l
l
1
1
I
1
Bg Xb (Ec) XhSa Xh
I
I
HI Ps Ec
I
I
Sm
Hn
1
t -
+---------
r
1]
s19
s16
I
1
Stage 14 strand specific transcription
Transcription boundaries
R N A / D N A Probes Stage 14 mRNA level Gene location 1 kb
Fig. 1. Organization of the major chorion gene clusters, transcription boundaries, and levels of transcription. At the top and bottom of each diagram the solid boxes extend from the cap to the polyadenylation site of the indicated chorion genes. Approximate positions of the s37, J, and K transcripts are shown [Parks et al., 19861. Abbreviations of restriction sites: Ac = A d , Ba =BamHI, Bg =BglII, Bs =BsrnI, Ec=EcoRI, Hi=HindlII, Hn=HincII, Ps=PstI, Pv=PuuII, Sa=SalI, Xb =XbaI, Xh =XhoI. Just below the restriction maps, levels of copy number corrected strand specific transcription during stage 12 (s36 and s38) and stage 14 (s15, s16, s18, and s19) are shown by the width of the solidly shaded arrows; unshaded areas indicate untranscribed regions, and dashed arrows indicate that transcription was undetectable. These arrows indicate the strand specificity of the stage specific
transcripts and correspond in length to the M13 clones used as probes. Transcription boundaries from stage 1-14 run-on reactions are indicated by the cross hatched bi-directional arrows, which represent restriction fragment probes; the molar levels of transcription (width) in the gene coding regions are arbitrarily set as equal, while those in 3’ flanking regions are set relative to the level of transcription in the preceding gene. Asterisks indicate tested regions which do not overlap with the transcription units, and dashed lines indicate regions with undetectable levels of transcription. Fragments used to generate probes for DNA and RNA quantitation are shown below by thin solid lines. Finally, the experimentally determined RNA levels at stage 12 (s36 and s38) and stage 14 ( s l 5 ,sl6, s f 8 ,and sl9) are indicated by the width of the stippled arrows.
Kambysellis, 1980; Ambrosio and Schedl, 1985; Matthews and Kaufmann, 19891. The traces of s38 signal may be due to production of the overlapping “J” transcript in nurse cell nuclei [Fig. 1and Parks et al., 19861. However, the run-on y p l transcripts presumably
originate from follicle cell nuclei [Barnett et al., 1980; Brennan et al., 19823. Since y p l transcription is detectable in the pre-choriogenic nuclei, the absence of detectable levels of chorion gene transcription at this stage of follicular development is significant.
200
ROMANO ET AL. TABLE 1. Plasmids Gene s15
s16 s18 si8-si9 s19 s3 6 s36 s38 s38 rP49 YPl aTub84B His
Insert 1.16 kb XbdlSalI 1.98 kb EcoRIlHindIII 1.99 kb PstIlBglII 5.65 kb SalIlEcoRI 0.78 kb XhoIlHindIII 0.75 kb XhoIlPvuII 1.99 kb XbdlEcoRI 0.98 kb BamHIIXbd 4.7 kb EcoRI 0.6 kb EcoRIlHindIII 3.8 kb HindIII 0.6 kb PvuII 5.0 kb HindIII
In contrast to prechoriogenic nuclei, the transcription levels in early choriogenic nuclei are high for s36 and s38, low for s16 and s19, even lower for s18, and undetectable for s15 (Fig. 2, st.12; Table 2). While the histone genes are transcribed at low levels, no rp49 or aTub84B run-on transcripts are detectable (presumably due to the degeneration of nurse cell nuclei). Transcription of ypl is also undetectable, in agreement with the known absence of y pl mRNA in stage 12 follicles [Brennan et al., 19821. Finally, transcription in late choriogenic stage nuclei is high for s15 and s19, moderate for s18, and low for s I 6 (Fig. 2, st.14; Table 2). In contrast to the high levels of transcriptional activity observed at stage 12, the s36 and s38 genes are now transcriptionally quiescent; this was confirmed with longer probes (Table 11, bearing the entire transcription unit of s36 or s38 (data not shown). As in the early choriogenic stage nuclei, low levels of histone transcripts and no transcripts from three other control genes are observed, as expected. The effects of the anionic detergent sarkosyl on the nuclear run-on reactions were also examined. Sarkosyl inhibits RNA polymerase I1 initiation in vitro, strips nucleosomes from chromatin, and, in some but not all instances, lifts transcriptional elongation blocks [Green et al., 1975; Hawley and Roeder, 1985; Bentley and Groudine, 1986; Rougvie and Lis, 19881. However, sarkosyl does not inhibit elongation reactions of RNA polymerase I1 in vitro and often stimulates elongation in nuclear run-on reactions [Hawley and Roeder, 1985; Green et al., 1975; Gariglio et al., 19811. In early and late choriogenic stage nuclei, addition of 0.6% sarkosyl resulted in moderate quantitative changes but did not alter the qualitative pattern of chorion gene transcription (data not shown). Consequently, continued transcriptional initiation in vitro and sarkosyl sensitive transcriptional elongation blocks are not responsible for the temporal pattern of run-on chorion gene transcription (Fig. 2). However, sarkosyl addition does yield low but detectable levels of premature s l 6 , s18, and s36 transcription in pre-choriogenic nuclei (data
Wong et al., 1985 D. King, personal communication Wong et al., 1985 Wong et al., 1985 Wong et al., 1985 Spradling et al., 1987 Spradling et al., 1987 Spradling et al., 1987 Parks et al., 1986 Wong et al., 1981 Barnett et al., 1980 Theurkauf et al., 1986 Lifton et al.. 1978
mp18I 19 mp18119 mpl8119 puc9 mp18119 mp18l19 puc19 mp18119 pBR322 pBR322 pBR322 pSP73 pBR322
Stage
1 -14
Stage
t
-1tamanitin
S
I -10A a
12
S
a S
14 a Fig. 2. Run-on transcription in nuclei of staged follicles, and the effect of a-amanitin. The upper panel shows run-on chorion transcription in stage 1-14 follicle nuclei, in the presence and absence of 2 Fgiml a-amanitin. Filter-immobilized plasmids were used to detect the transcripts. Filters were probed with 1.6 X lo7 cpm (-1 or 0.74 x lo7 cpm ( + ) of run-on transcript and exposed for 20 h r with a n intensifying screen. The bottom panels show run-on transcription in stage l-lOa, 12, and 14 follicle nuclei. Chorion gene slots (~1.5, s l 6 , s18, s19, s36, and s38) marked “s”represent sense strand run-on transcription,and those marked “a” represent anti-sense strand transcription: single stranded M13 probes of appropriate polarity were used as probes for these transcripts. All non-chorion gene slots represent both sense and anti-sense run-on transcription, detected with plasmid DNA. Duplicate filters were hybridized with 5 x 10‘ (st.l-lOA), 1.4 X lo6 (st.12),or 2 X 10‘ (st.14) cpm of run-on transcript and exposed for 168 hr with a n intensifying screen. Analysis oftubulin DNA amounts also revealed that the DNA input for the st. 1-10a run-on reaction was about 35%higher than for the st. 12 reaction and about 120% higher than for the st.14 reaction. rp, ribosomal protein 49 gene; ypl, yolk protein 1 gene; tb, atubulin gene; his, histone gene cluster; pBR, pBR322.
DROSOPHILA CHORION GENE EXPRESSION
201
Fig. 3. Mapping of run-on transcription boundaries. Subclones of genomic DNA encoding s36 or s18, s15, and s l 9 genes were digested with the indicated restriction enzymes, Southern blotted, and probed with 3 x l o 7 cpm of stage 1-14 run-on transcript. A map of the clones, the ethidium bromide stained gel, a light autoradiographic exposure (216 hr without intensifying screen), and a dark autoradiographic exposure (168 hr with intensifying screen) are shown from left to right.
Locations of the cap ( + 1) and polyadenylation sites (polyA) of the transcription units are as shown. Note that fragments B and E comigrate on the gel at the right; the signal is due to run-on transcription in E, as no transcription was observed in a separate experiment in a fragment containing exclusively B. Data of these and other experiments are presented schematically in Figure 1. COD, coding; FL, flanking; IG, intergenic.
not shown). Further analysis is necessary to determine if these chorion genes are occupied by “paused” transcription complexes prior to their normal activation [cf. Rougvie and Lis, 19881. The observed transcription of chorion genes is largely strand specific. Synthesis of antisense transcripts is barely detectable for s36, s38, and s18 even when transcription of the normal strand is high (st. 12 for s36 and s38, st. 14 for s18; antisense transcripts are not detectable for the less intensively transcribed sl6 gene). Surprisingly, however, significant levels of antisense signal are evident for gene s l 5 a t stage 14 (from 12 to 25% of the sense strand signal); s19 antisense transcripts were also observed at st. 14 in one experiment (25%), but not in a second. Sequence analysis indicates no significant self-complementarity in these chorion genes [Wong et al., 19851, eliminating the possibility that the antisense signal could be due to sense transcripts. Furthermore, in situ hybridization experiments indicate that stable s15 antisense RNA accumulates in stage 14 follicle cells (M.J. Conboy and B.D. Mariani, personal communication). The promoterb)
which yield the sl5 and s l 9 antisense transcripts have not been identified.
Most Transcription Occurs in the Known Transcription Units To determine the transcribed regions of the chorion clusters, Southern blots of appropriately restricted plasmids were probed with run-on transcripts from st.1-14 nuclei (Figs. 1,3). As expected, the coding regions of the chorion genes were transcribed at high levels. Because of lack of termination, nuclear run-on transcripts of regions 3’ to the polyadenylation sites of most RNA polymerase I1 genes are usually equimolar to transcripts of the adjacent coding regions [Birnstiel et al., 19851. However, no substantial transcription was observed immediately downstream of four chorion genes. For s36, the 556 bp region just 127 bp downstream of the polyadenylation site is transcribed a t less than 2% of the molar level of the coding region (Fig. 3, s36, fragment D), while for s37 the corresponding value is less than 5% (Fig. 1,raw data not shown). Similarly,
202
ROMANO ET AL.
Fig. 4. Analysis of DNA and RNA from staged follicles used in nuclear run-on experiments. DNA from staged (l-lOA, 12,14) nuclei, or RNA from staged follicles, were isolated and analysed on slot blots as described (Materials and Methods). These follicles and nuclei were aliquots of the samples used in the Figure 2 nuclear run-on experiments. Filters were probed with oligonucleotide labelled fragments (see RNAiDNA probes in Fig. 1) a t 1.4 x lo9 to 3.7 x lo9 cpmipg, and were exposed without intensifying screen for either 96 hr (DNA and bottom tubulin RNA panels) or 18 hr (remaining RNA panels). The bottom tubulin (tb) DNA panel is a darker print of the 96 h r exposure.
Note that the aTub84B tubulin probe may also hybridize to other tubulin sequences present in these ovarian samples [Tub85E, aTub84D, and aTub67C for DNA, and aTub67C for RNA; Theurkauf and Wensink, 1986; Matthews et al., 19821. The right-most panel is a more complete set of staged RNAs, analyzed by gel electrophoresis followed by blot hybridization, from [Fenerjian et al., 19891; comparisons of intensities a t different stages are appropriate for the same gene, but not between genes, as the specific activities of the probes and the exposures varied for different genes.
for s l 8 the 800 bp region just 80 bp downstream of the polyadenylation site as well as the 1,128 bp region upstream of the gene are transcribed at less than the detection limit (< 12% of the molar transcription level in the coding region; Fig. 1 and data not shown). A 265 bp fragment extending from 37 bases upstream to 223 bases downstream of the s15 polyadenylation site is only transcribed a t approximately 5% of s15 coding region levels (Fig. 3, s18-sl6-sl9, fragment E). However, substantial transcription does occur further downstream in the s15/s19 intergenic DNA, as well a s downstream of s19, in both cases a t approximately 2530%of the molar levels of transcription within the respective coding regions (Fig. 3, s18-sl5-sl9, fragments F,H); these levels are comparable to the s15 and s l 9 antisense run-on transcription. The origin of these intergenic transcripts is unclear, as strand specific probes were not employed.
DNA and RNA Levels DNA samples from aliquots of the nuclei used in the run-on reactions were immobilized on nylon membranes and probed to determine the relative gene copy numbers, as well as the relative input of nuclei (Fig. 4, DNA). All of the chorion genes were amplified to similar levels within a given cluster, consistent with the results of previous experiments [Spradling, 19811. Therefore, normalization for gene copy number was superfluous. To control for variations in oocyte staging and to permit the direct comparison of chorion gene transcription and mRNA accumulation, the RNA levels were determined in aliquots of the same follicles used for nuclear preparations (Fig. 4, RNA). In prechoriogenic follicles, no chorion RNA could be detected. Early choriogenic follicles had high levels of s36 and s38 RNA, low but detectable levels of sl6 and s19 RNA,
DROSOPHILA CHORION GENE EXPRESSION
203
TABLE 2. Quantitation of Chorion Gene Transcription and mRNA Accumulation* Gene Stage 12 s15 s16 s18 s19 s36 s38 Stage 14 Sl5 s16 s18 s19 s36 s38
Relative transcription level
Relative mRNA level
The Relative Importance of Transcriptional and Post Transcriptional Regulation of Drosophila Chorion Gene Expression During Oogenesis CHARLES P. ROMANO, JUAN CARLOS MARTINEZ-CRUZADO, AND FOTIS C. KAFATOS Department of Cellular and Developmental Biology, The Biological Laboratories, Harvard University, Cambridge (C.P.R., J.C.M.C., F.C.K.) and The Institute of Molecular Biology and Biotechnology, Research Center of Crete, Crete, Greece (F.C.K.) To determine the relative roles ABSTRACT of transcriptional and post-transcriptional events in establishing the temporal pattern of chorion gene expression in Drosophila, we have examined chorion gene transcription, RNA accumulation, and protein synthesis in follicles of selected pre-, early-, and late-choriogenic stages. Chorion gene transcription was assayed in follicle cell nuclei by nuclear run-on reactions. For the s i5, s 16, s18, 536, and s38 chorion genes, the periods of intense transcription are as predicted from the dynamics of RNA accumulation and protein synthesis, indicating that these genes are primarily regulated at the transcriptional level. In contrast, gene s 19 appears subject to post-transcriptional control at stage 14, when transcription rates are substantially higher than predicted from the observed RNA levels. Transcription of regions between the clustered and tandemly oriented chorion genes was also examined. In contrast to many RNA polymerase II transcribed genes, for the s78 and s36 chorion genes run-on transcription appears to terminate within about 100 base pairs downstream of the polyadenylation sites, corroborating previous reports based on electron microscopy of s36 [Osheim et a/., EMBO J 5:3591-3596, 19861.
Key words: Run-on tra nscription, develop men-
ceeds through the remainder of oogenesis [Spradling and Leys, 19881, chorion gene mRNAs accumulate and decay according to a temporal program that is distinct for each gene [Griffin-Shea et al., 1982; Parks et al., 19861. With the exception of the transient prechoriogenic accumulation of untranslated s15 mRNA, chorion proteins are generally synthesized when the corresponding chorion mRNA accumulates [Petri et al., 1976; Waring and Mahowald, 1979; Thireos et al., 1980; Griffin-Shea et al., 1982; Parks et al., 19861. The chorion genes of Drosophila thus serve as a useful model system for the study of temporally regulated gene expression. The fundamental question of the relative regulatory roles of transcriptional and post-transcriptional events in establishing the temporal pattern of chorion gene expression has yet to be addressed directly and systematically. Indirect evidence favoring transcriptional regulation has been obtained in P-element-mediated transformation experiments. Chorion genes, marked with foreign DNA by internal insertion or fusion, produce temporally regulated transcripts when transferred to different chromosomal sites [Kafatos et al., 1985; Wakimoto et al., 19861; this demonstrates that amplification is not involved in temporal programming but leaves open the possibility that transcribed regions may influence that programming by a post-transcriptional mechanism. However, the early s36 and late s15
tally regulated gene expression, follicle cell nuclei
INTRODUCTION During the final stages of Drosophila oogenesis, the terminally differentiated follicle cells surrounding the oocyte construct the proteinaceous eggshell or chorion, which serves to protect the egg and embryo [reviewed by Mahowald and Kambysellis, 19801.The two clusters of major chorion genes are selectively amplified within the already polyploid follicle cells to produce the massive amounts of chorion proteins in the short time available [Spradling, 1981; reviewed by Delidakis et al., 19891. After the onset of amplification, which pro-
0 1991 WILEY-LISS. INC.
Received for publication August 7, 1990; accepted October 8, 1990. Address reprint requests to Dr. Fotis C. Kafatos, Department of Cellular and Developmental Biology, The Biological Laboratories, Harvard University, 16 Divinity Ave., Cambridge, MA 02138. Dr. Romano is now at Monsanto, Plant Gene Expression GroupAASG, 700 Chesterfield Village Parkway, St. Louis, MO 63198. Dr. Martinez-Cruzado is now a t the Departamento de Biologie, Recinto Universiterio de Mayaguez, Universidad de Puerto Rico Mayaguez, PR 00708.
DROSOPHILA CHORION GENE EXPRESSION
197
promoters alone can direct an alcohol dehydrogenase reporter gene to be expressed in a manner temporally indistinguishable from that of the endogenous s36 and s15 chorion genes [Romano et al., 19881. Moreover, the ability to modify the s15 promoter and thereby change the temporal specificity of the resulting transcripts [Mariani et al., 19881 further argues that this chorion gene is transcriptionally regulated. Chorion gene transcription has also been studied by EM analysis of chromatin spreads in which nascent transcripts of the amplified chorion genes can be directly visualized [Osheim and Miller, 19831. Transcriptionally active, amplified DNAs with the characteristics of the X and 3rd chromosome chorion gene clusters are only observed in follicular nuclei of the stage at which the mRNAs corresponding to the respective clusters accumulate [Osheim and Miller, 19831. However, as positive identification of a chorion cluster by this technique is dependent on its active transcription, these results must be interpreted with caution. The transcriptional activity of higher eukaryotic genes has traditionally been measured via nuclear runon assays. In such assays, transcriptionally engaged RNA polymerase molecules synthesize radiolabeled nascent transcripts, which are quantified to reveal the number of active RNA polymerase molecules on a given gene a t the time the nuclei were isolated [reviewed by Darnell, 19821. Additional information can be obtained from such studies, e.g., that c-myc and cmyb are regulated at the level of transcriptional elongation rather than initiation [Nepveu and Marcu, 1986; Bentley and Groudine, 1986; Bender et al., 19871. Nuclear run-on analysis has rarely been employed in the analysis of tissue specific gene expression in organisms such as Drosophila, as it is often difficult to obtain enough nuclei to perform the assay. However, we have been able to use this approach to examine chorion gene transcription because of our ability to isolate significant numbers of follicles (i.e., egg chambers) of a given choriogenic stage, as well as the fact that each follicle contains about 1,200 polyploid follicle cell nuclei in which the chorion genes are amplified. We report here the temporal pattern, boundaries, and levels of chorion gene transcription in follicle cell nuclei as assayed by nuclear run-on transcription. These results are compared and contrasted with the temporal patterns of chorion mRNA accumulation and protein synthesis.
morphological criteria [King, 19701, until a minimum of about 400 were collected. After they settled into a minimal volume of Ringer's (about 30 pl), and equal volume of Solution A (4mM Mg Acetate, 6 mM CaCl,, 20 mM Tris-HC1, pH 8.0,1% N P 4 0 ,l mM PMSF, and 1 mM DTT) was added and the resultant mixture gently disrupted with five strokes of a loosely fitting Dounce homogenizer. The homogenate was layered over 10 ml of Solution B (1 M sucrose, 5 mM Mg Acetate, 10 mM Tris-HC1 pH 8.0, 0.5 mM DTT, and 0.5 mM PMSF) in a 15 ml Corex Tube and spun at 4,000 RPM for 10 min in a Sorvall HB40 rotor. The pelleted nuclei were resuspended in Solution C (25% glycerol, 4.7 mM Mg Acetate, 50 mM Tris-HC1, pH 8.0, 0.1 mM EDTA, and 5 mM DTT) and spun at 4,000 RPM for 10 min in an Eppendorf microcentrifuge. Pelleted nuclei were resuspended in 10 to 15 pl of Solution C and 10 pl of nuclei were added to 10 p1 of 2X transcription mix (22.8% glycerol, 9.4 mM Mg Acetate, 1.8 mM MnCl,, 54 mM Tris-HC1, pH 8.0,160 mM KC1, and 1mM each of ATP, GTP, and CTP with 250 pCi a-32P-UTP a t 800 Ci/ mmol) at 25°C. In some experiments, nuclei in solution C were preincubated a t 4°C for 5 min with or without 2 pg/ml a-amanitin (Sigma Chemical Co., St. Louis, MO) prior to the transcription reaction. When indicated, nuclei in solution C were added to a n equal volume of 2 x transcription mix at 1.2% sarkosyl (N-lauroyl sarcosine sodium salt; ultra-pure molecular biology grade from IBI) for a final sarkosyl concentration of 0.6%. After 20 min a t 25"C, the run-on reaction was terminated by disruption in Urea:SDS (7 M Urea, 2% SDS, 10 mM Tris-HC1, pH 8.0, 350 mM NaCl, and 1 mM EDTA) and total nucleic acid along with 20 pg carrier tRNA was isolated by phenol:CHCl, extraction followed by ethanol precipitation. Genomic DNA was removed by Promega RQ1 RNAse free DNAse treatment and run-on transcripts were purified by phenol:CHCl, extraction followed by ethanol/NH,Acetate precipitation. About 5 x lo6 cpm of incorporated counts were typically obtained.
MATERIALS AND METHODS Isolation of Nuclear Run On Transcripts From Follicles Dissected ovaries from cold anesthetized Drosophila melanogaster ry506were disrupted in Drosophila Ringer's solution at 4°C by passage through a BSA coated pasteur pipet [Petri et al., 19761. This and all steps up until placing the nuclei in the transcription mix were done a t 4°C. Follicles were staged by hand according to
Southern and Slot Blots RNA was heated in water for 1min at 75"C, brought to 1 0 SSC, ~ and slotted to nylon membrane with a suction manifold. DNA was slotted a s previously described [Romano et al., 19881. Nucleic acid was crosslinked to the filters by UV irradiation [Church and Gilbert, 19841. Southern blots of restriction enzyme digested plasmids were done as reported [Delidakis and Kafatos, 19871.
Isolation of DNA and RNA DNA was isolated from nuclei as described [Delidakis and Kafatos, 19871 but was treated with RNAse prior to slot blotting. Total nucleic acid was isolated from follicles as described [Thireos et al., 19801and was treated with Promega RQ 1 RNAse free DNAse to obtain RNA for slot blot analysis.
198
ROMANO ET AL.
Hybridization Conditions RNA and DNA slot blots were hybridized and washed as described [Romano et al., 19881, except that oligo-labeled fragment probes in excess were used instead of RNA or nick translated probes [Feinberg and Vogelstein, 19831. Run-on transcripts were hybridized in 1 ml of solution to Southern or slot-blot filters with 0.5 pg denatured, immobilized DNA probe per slot or band for at least 24 hr, and were washed as above. Under these conditions, 0.5 pg of immobilized probe was shown to be in excess of run-on transcript. Filters probed with run-on transcripts were treated with 5 kgl ml RNAse A in 2~ SSC a t 37°C for 30 min prior to autoradiography . Densitometry To quantitate autoradiographic signals, filters were exposed without screen or preflash to Kodak XAR-5 film to obtain exposures within the linear response range of the film. The autoradiograms were then scanned with a Helena Laboratories R and D densitometer equipped with a Quick Quant I11 integrating computer. Values obtained from filters probed with run-on transcripts were then normalized to the length of complementary sequence present in the DNA immobilized on the filter. When noted, the length normalized runon transcription values were also normalized to the experimentally determined gene copy number. Relative gene copy numbers were determined by quantitating slot blots of genomic DNA from the nuclei used in the run-on experiments as described above (e.g., see Fig. 4,DNA); the densitometric values from the amplified chorion genes were divided by the value for the unamplified tubulin gene to obtain gene copy numbers. Values obtained from hybridizing immobilized RNAs or DNAs with oligolabelled probes in excess were normalized for the specific activity of those probes; specific activities were calculated as described (Promega Prime-a-Gene Labeling System).
lowed by hybridization with double stranded probes (Fig. 4). The quantitative measurements of stage specific transcription and RNA accumulation obtained in two independent but identical experiments were combined and tabulated in Table 2. Only early (st.12) and late (st.14) choriogenic stage follicles were used, as stages 11 and 13 are too transient and rare for run-on transcription assays. In interpreting the results, it should be noted that choriogenesis is an exceedingly rapid process [ca. 5 hr from stage 11to 14; King, 19701. It should also be noted that the developmentally middle chorion genes ( s l 9 and s16) are maximally expressed during st. 13 while the early genes (s36 and s38) are turning off and the late genes (s18 and s15) are turning on, in terms of RNA accumulation and protein synthesis [Petri et al., 1976; Waring and Mahowald, 1979; Griffin-Shea et al., 1982; Fenerjian et al., 1989; see also Fig. 41.
Nuclear Run-on Transcription of the Major Chorion Genes is RNA Polymerase I1 Dependent and Displays Stage Specific Differences To identify the RNA polymerase activity involved in the synthesis of the chorion gene run-on transcripts, nuclei from stage 1 through 14 follicles were briefly exposed to 2 kg/ml a-amanitin prior t o the nuclear runon reaction. While total incorporation decreased by about 50%, run-on transcription of the chorion genes decreased by over 95% (as estimated for s15; Fig. 2, st.1-14 2 amanitin). As the RNA polymerases I and I11 of Drosophila and other insects are not inhibited by low concentrations of a-amanitin [Roeder, 1976; Greenleaf et al., 19761, this result demonstrates that the chorion run-on transcripts are products of RNA polymerase I1 activity. To determine the level of chorion gene transcription during different stages of oogenesis, nuclear run-on analysis of transcripts from pre-, early-, and latechoriogenic follicle nuclei (stages l-lOA, 12, and 14, respectively) was performed. Unique patterns of transcriptional activity are characteristic of each stage. In RESULTS prechoriogenic stages (Fig. 2, st. l-lOa), high levels of Experimental Strategy histone gene (his) run-on transcripts, along with low The organization of the major chorion gene clusters levels of ribosomal protein 49 (rp), al-tubulin (tb), and of Drosophila melanogaster is shown in Figure 1, and yolk protein 1 (ypl) run-on transcripts are observed, the plasmids used in this study are listed in Table 1. To demonstrating that these nuclei are transcriptionally monitor the stage and strand specificities of transcrip- competent; however, with the exception of very low s38 tion (Fig. 2), single-stranded M13 clones (chorion transcription, no chorion gene run-on transcripts are genes) or double-stranded pUC clones (control genes) detected. Transient prechoriogenic expression of s15, were slot blotted and probed with run-on transcripts of which apparently occurs only under certain conditions, staged nuclei. To determine the physical boundaries of was not detected in either the nuclear run-on assays or transcription, double-stranded chorion clones were di- in the slot blot analysis of stable RNA [Thireos et al., gested with restriction enzymes, Southern blotted, and 1980; Figs. 2 and 41. Since the single quiescent oocyte probed with run-on transcripts from follicular nuclei of nucleus within each follicle is unlikely to yield signifall stages (st.1-14; Fig. 3). Levels of accumulated icant amounts of run-on transcript [Mahowald and chorion RNA in the follicles and relative chorion gene Tiefert, 19701, signals from the His, rp49, and copy numbers in the nuclei used for stage specific run- aTub84B genes may be largely due to run-on tranon reactions were also determined by slot blotting fol- scripts from the giant nurse cell nuclei [Mahowald and
-
-
s 37
s38
s 36
I
1
So
DROSOPHZLA CHORION GENE EXPRESSION
Xb’Bs
BO
I Xb
+----
L
b 1
r
1
11
[
1
s 36 .
K I
Pv BS EC
Xh
J
199
’
-1
Ec
Ec
Stage12 strand specific transcriDtion Transcription boundaries
R N A / D N A Probes Stage 12 mRNA level Gene location 1 kb
I
1
So Ps Ac
I
1
AC Hi
r - L 1
I
l
l
1
1
I
1
Bg Xb (Ec) XhSa Xh
I
I
HI Ps Ec
I
I
Sm
Hn
1
t -
+---------
r
1]
s19
s16
I
1
Stage 14 strand specific transcription
Transcription boundaries
R N A / D N A Probes Stage 14 mRNA level Gene location 1 kb
Fig. 1. Organization of the major chorion gene clusters, transcription boundaries, and levels of transcription. At the top and bottom of each diagram the solid boxes extend from the cap to the polyadenylation site of the indicated chorion genes. Approximate positions of the s37, J, and K transcripts are shown [Parks et al., 19861. Abbreviations of restriction sites: Ac = A d , Ba =BamHI, Bg =BglII, Bs =BsrnI, Ec=EcoRI, Hi=HindlII, Hn=HincII, Ps=PstI, Pv=PuuII, Sa=SalI, Xb =XbaI, Xh =XhoI. Just below the restriction maps, levels of copy number corrected strand specific transcription during stage 12 (s36 and s38) and stage 14 (s15, s16, s18, and s19) are shown by the width of the solidly shaded arrows; unshaded areas indicate untranscribed regions, and dashed arrows indicate that transcription was undetectable. These arrows indicate the strand specificity of the stage specific
transcripts and correspond in length to the M13 clones used as probes. Transcription boundaries from stage 1-14 run-on reactions are indicated by the cross hatched bi-directional arrows, which represent restriction fragment probes; the molar levels of transcription (width) in the gene coding regions are arbitrarily set as equal, while those in 3’ flanking regions are set relative to the level of transcription in the preceding gene. Asterisks indicate tested regions which do not overlap with the transcription units, and dashed lines indicate regions with undetectable levels of transcription. Fragments used to generate probes for DNA and RNA quantitation are shown below by thin solid lines. Finally, the experimentally determined RNA levels at stage 12 (s36 and s38) and stage 14 ( s l 5 ,sl6, s f 8 ,and sl9) are indicated by the width of the stippled arrows.
Kambysellis, 1980; Ambrosio and Schedl, 1985; Matthews and Kaufmann, 19891. The traces of s38 signal may be due to production of the overlapping “J” transcript in nurse cell nuclei [Fig. 1and Parks et al., 19861. However, the run-on y p l transcripts presumably
originate from follicle cell nuclei [Barnett et al., 1980; Brennan et al., 19823. Since y p l transcription is detectable in the pre-choriogenic nuclei, the absence of detectable levels of chorion gene transcription at this stage of follicular development is significant.
200
ROMANO ET AL. TABLE 1. Plasmids Gene s15
s16 s18 si8-si9 s19 s3 6 s36 s38 s38 rP49 YPl aTub84B His
Insert 1.16 kb XbdlSalI 1.98 kb EcoRIlHindIII 1.99 kb PstIlBglII 5.65 kb SalIlEcoRI 0.78 kb XhoIlHindIII 0.75 kb XhoIlPvuII 1.99 kb XbdlEcoRI 0.98 kb BamHIIXbd 4.7 kb EcoRI 0.6 kb EcoRIlHindIII 3.8 kb HindIII 0.6 kb PvuII 5.0 kb HindIII
In contrast to prechoriogenic nuclei, the transcription levels in early choriogenic nuclei are high for s36 and s38, low for s16 and s19, even lower for s18, and undetectable for s15 (Fig. 2, st.12; Table 2). While the histone genes are transcribed at low levels, no rp49 or aTub84B run-on transcripts are detectable (presumably due to the degeneration of nurse cell nuclei). Transcription of ypl is also undetectable, in agreement with the known absence of y pl mRNA in stage 12 follicles [Brennan et al., 19821. Finally, transcription in late choriogenic stage nuclei is high for s15 and s19, moderate for s18, and low for s I 6 (Fig. 2, st.14; Table 2). In contrast to the high levels of transcriptional activity observed at stage 12, the s36 and s38 genes are now transcriptionally quiescent; this was confirmed with longer probes (Table 11, bearing the entire transcription unit of s36 or s38 (data not shown). As in the early choriogenic stage nuclei, low levels of histone transcripts and no transcripts from three other control genes are observed, as expected. The effects of the anionic detergent sarkosyl on the nuclear run-on reactions were also examined. Sarkosyl inhibits RNA polymerase I1 initiation in vitro, strips nucleosomes from chromatin, and, in some but not all instances, lifts transcriptional elongation blocks [Green et al., 1975; Hawley and Roeder, 1985; Bentley and Groudine, 1986; Rougvie and Lis, 19881. However, sarkosyl does not inhibit elongation reactions of RNA polymerase I1 in vitro and often stimulates elongation in nuclear run-on reactions [Hawley and Roeder, 1985; Green et al., 1975; Gariglio et al., 19811. In early and late choriogenic stage nuclei, addition of 0.6% sarkosyl resulted in moderate quantitative changes but did not alter the qualitative pattern of chorion gene transcription (data not shown). Consequently, continued transcriptional initiation in vitro and sarkosyl sensitive transcriptional elongation blocks are not responsible for the temporal pattern of run-on chorion gene transcription (Fig. 2). However, sarkosyl addition does yield low but detectable levels of premature s l 6 , s18, and s36 transcription in pre-choriogenic nuclei (data
Wong et al., 1985 D. King, personal communication Wong et al., 1985 Wong et al., 1985 Wong et al., 1985 Spradling et al., 1987 Spradling et al., 1987 Spradling et al., 1987 Parks et al., 1986 Wong et al., 1981 Barnett et al., 1980 Theurkauf et al., 1986 Lifton et al.. 1978
mp18I 19 mp18119 mpl8119 puc9 mp18119 mp18l19 puc19 mp18119 pBR322 pBR322 pBR322 pSP73 pBR322
Stage
1 -14
Stage
t
-1tamanitin
S
I -10A a
12
S
a S
14 a Fig. 2. Run-on transcription in nuclei of staged follicles, and the effect of a-amanitin. The upper panel shows run-on chorion transcription in stage 1-14 follicle nuclei, in the presence and absence of 2 Fgiml a-amanitin. Filter-immobilized plasmids were used to detect the transcripts. Filters were probed with 1.6 X lo7 cpm (-1 or 0.74 x lo7 cpm ( + ) of run-on transcript and exposed for 20 h r with a n intensifying screen. The bottom panels show run-on transcription in stage l-lOa, 12, and 14 follicle nuclei. Chorion gene slots (~1.5, s l 6 , s18, s19, s36, and s38) marked “s”represent sense strand run-on transcription,and those marked “a” represent anti-sense strand transcription: single stranded M13 probes of appropriate polarity were used as probes for these transcripts. All non-chorion gene slots represent both sense and anti-sense run-on transcription, detected with plasmid DNA. Duplicate filters were hybridized with 5 x 10‘ (st.l-lOA), 1.4 X lo6 (st.12),or 2 X 10‘ (st.14) cpm of run-on transcript and exposed for 168 hr with a n intensifying screen. Analysis oftubulin DNA amounts also revealed that the DNA input for the st. 1-10a run-on reaction was about 35%higher than for the st. 12 reaction and about 120% higher than for the st.14 reaction. rp, ribosomal protein 49 gene; ypl, yolk protein 1 gene; tb, atubulin gene; his, histone gene cluster; pBR, pBR322.
DROSOPHILA CHORION GENE EXPRESSION
201
Fig. 3. Mapping of run-on transcription boundaries. Subclones of genomic DNA encoding s36 or s18, s15, and s l 9 genes were digested with the indicated restriction enzymes, Southern blotted, and probed with 3 x l o 7 cpm of stage 1-14 run-on transcript. A map of the clones, the ethidium bromide stained gel, a light autoradiographic exposure (216 hr without intensifying screen), and a dark autoradiographic exposure (168 hr with intensifying screen) are shown from left to right.
Locations of the cap ( + 1) and polyadenylation sites (polyA) of the transcription units are as shown. Note that fragments B and E comigrate on the gel at the right; the signal is due to run-on transcription in E, as no transcription was observed in a separate experiment in a fragment containing exclusively B. Data of these and other experiments are presented schematically in Figure 1. COD, coding; FL, flanking; IG, intergenic.
not shown). Further analysis is necessary to determine if these chorion genes are occupied by “paused” transcription complexes prior to their normal activation [cf. Rougvie and Lis, 19881. The observed transcription of chorion genes is largely strand specific. Synthesis of antisense transcripts is barely detectable for s36, s38, and s18 even when transcription of the normal strand is high (st. 12 for s36 and s38, st. 14 for s18; antisense transcripts are not detectable for the less intensively transcribed sl6 gene). Surprisingly, however, significant levels of antisense signal are evident for gene s l 5 a t stage 14 (from 12 to 25% of the sense strand signal); s19 antisense transcripts were also observed at st. 14 in one experiment (25%), but not in a second. Sequence analysis indicates no significant self-complementarity in these chorion genes [Wong et al., 19851, eliminating the possibility that the antisense signal could be due to sense transcripts. Furthermore, in situ hybridization experiments indicate that stable s15 antisense RNA accumulates in stage 14 follicle cells (M.J. Conboy and B.D. Mariani, personal communication). The promoterb)
which yield the sl5 and s l 9 antisense transcripts have not been identified.
Most Transcription Occurs in the Known Transcription Units To determine the transcribed regions of the chorion clusters, Southern blots of appropriately restricted plasmids were probed with run-on transcripts from st.1-14 nuclei (Figs. 1,3). As expected, the coding regions of the chorion genes were transcribed at high levels. Because of lack of termination, nuclear run-on transcripts of regions 3’ to the polyadenylation sites of most RNA polymerase I1 genes are usually equimolar to transcripts of the adjacent coding regions [Birnstiel et al., 19851. However, no substantial transcription was observed immediately downstream of four chorion genes. For s36, the 556 bp region just 127 bp downstream of the polyadenylation site is transcribed a t less than 2% of the molar level of the coding region (Fig. 3, s36, fragment D), while for s37 the corresponding value is less than 5% (Fig. 1,raw data not shown). Similarly,
202
ROMANO ET AL.
Fig. 4. Analysis of DNA and RNA from staged follicles used in nuclear run-on experiments. DNA from staged (l-lOA, 12,14) nuclei, or RNA from staged follicles, were isolated and analysed on slot blots as described (Materials and Methods). These follicles and nuclei were aliquots of the samples used in the Figure 2 nuclear run-on experiments. Filters were probed with oligonucleotide labelled fragments (see RNAiDNA probes in Fig. 1) a t 1.4 x lo9 to 3.7 x lo9 cpmipg, and were exposed without intensifying screen for either 96 hr (DNA and bottom tubulin RNA panels) or 18 hr (remaining RNA panels). The bottom tubulin (tb) DNA panel is a darker print of the 96 h r exposure.
Note that the aTub84B tubulin probe may also hybridize to other tubulin sequences present in these ovarian samples [Tub85E, aTub84D, and aTub67C for DNA, and aTub67C for RNA; Theurkauf and Wensink, 1986; Matthews et al., 19821. The right-most panel is a more complete set of staged RNAs, analyzed by gel electrophoresis followed by blot hybridization, from [Fenerjian et al., 19891; comparisons of intensities a t different stages are appropriate for the same gene, but not between genes, as the specific activities of the probes and the exposures varied for different genes.
for s l 8 the 800 bp region just 80 bp downstream of the polyadenylation site as well as the 1,128 bp region upstream of the gene are transcribed at less than the detection limit (< 12% of the molar transcription level in the coding region; Fig. 1 and data not shown). A 265 bp fragment extending from 37 bases upstream to 223 bases downstream of the s15 polyadenylation site is only transcribed a t approximately 5% of s15 coding region levels (Fig. 3, s18-sl6-sl9, fragment E). However, substantial transcription does occur further downstream in the s15/s19 intergenic DNA, as well a s downstream of s19, in both cases a t approximately 2530%of the molar levels of transcription within the respective coding regions (Fig. 3, s18-sl5-sl9, fragments F,H); these levels are comparable to the s15 and s l 9 antisense run-on transcription. The origin of these intergenic transcripts is unclear, as strand specific probes were not employed.
DNA and RNA Levels DNA samples from aliquots of the nuclei used in the run-on reactions were immobilized on nylon membranes and probed to determine the relative gene copy numbers, as well as the relative input of nuclei (Fig. 4, DNA). All of the chorion genes were amplified to similar levels within a given cluster, consistent with the results of previous experiments [Spradling, 19811. Therefore, normalization for gene copy number was superfluous. To control for variations in oocyte staging and to permit the direct comparison of chorion gene transcription and mRNA accumulation, the RNA levels were determined in aliquots of the same follicles used for nuclear preparations (Fig. 4, RNA). In prechoriogenic follicles, no chorion RNA could be detected. Early choriogenic follicles had high levels of s36 and s38 RNA, low but detectable levels of sl6 and s19 RNA,
DROSOPHILA CHORION GENE EXPRESSION
203
TABLE 2. Quantitation of Chorion Gene Transcription and mRNA Accumulation* Gene Stage 12 s15 s16 s18 s19 s36 s38 Stage 14 Sl5 s16 s18 s19 s36 s38
Relative transcription level
Relative mRNA level
