ABSTRACT
Transcriptional regulation of the s38 chorion gene was studied using P element-mediated germline transformation. A 5–27 kb DNA fragment containing the s38 gene and 5′- and 3′-flanking sequences, was tested for its ability to be transcribed with correct developmental specificity. Five single-insert transformed lines were generated by microinjection of this DNA fragment cloned into a marked P element transformation vector. In each line, the transformed gene was transcribed according to the precise developmental pattern followed by the native s38 gene. The 1– 3 kb at the 5′ end of this tested fragment was fused to the E. coli lac z gene. This fragment was also capable of initiating transcription of E. coli lac Z RNA with the developmental profile of the native s38 gene. In vitro deletion studies are underway to determine which sequences in the 1–3 kb fragment are necessary for regulating the developmental expression of the gene.
INTRODUCTION
The regulation of gene expression in eucaryotes has been the focus of much study in the field of developmental biology. Inducible gene systems, such as those regulated by heavy metals or hormones, are attractive because they lend themselves to simple experimental manipulation, and the differences in RNA levels between induced and uninduced states can be dramatic. Genes exhibiting major changes in the level of expression during development in response to less-well-characterized signals are equally interesting to study, and provide the opportunity to teach us about timing or tissue-specific factors or signals controlling their expression. The approaches to understanding gene regulation range from defining hormonal parameters to classical genetics to in vitro mutagenesis and DNA sequencing. In the latter approach, the cloned gene can be altered specifically, and then re-introduced by transfection or injection into an environment where the effect of the mutation can be studied (McKnight, Gavis, Kingsbury & Axel, 1981; McKnight & Kingsbury, 1982; Brinster et al. 1981; Palmiter et al. 1982a; Palmiter, Chen & Brinster, 1982b).
Our laboratory has concentrated its efforts on understanding the regulation of gene expression of a set of genes encoding the chorion or eggshell of Drosophila. The chorion genes exhibit dramatic time- and tissue-specific developmental regulation (Spradling et al. 1981; Griffin-Shea, Thireos & Kafatos, 1982). The P element transformation techniques recently developed in this laboratory have enabled us to ask direct questions about the specific function of the sequences surrounding chorion genes and the importance of the organization of these genes for their proper developmental regulation. In this paper we describe our studies on the controls governing expression of the s38 chorion gene which encodes a major polypeptide of relative molecular mass 38 × 103.
THE CHORION GENE SYSTEM
The chorion of Drosophila is an architecturally complex structure composed of three layers and approximately 20 polypeptides (Waring & Mahowald, 1979). Eggshell protein synthesis and construction take place during the late stages of oocyte development and closely follow the synthesis of the major chorion RNAs (Waring & Mahowald, 1979; Spradling & Mahowald, 1979).
Transcription of the chorion genes occurs in the ovarian follicle cells according to a precisely timed developmental programme. All of the major chorion RNAs are detected in the last stages (11–14) of oogenesis (Parks & Spradling, 1981; Griffin-Shea et al. 1982). Within these stages, however, RNA synthesis from each gene proceeds with an independently timed pattern. The s38 RNA is detected maximally at stage 12, which lasts 1·5 h, and is also present during stages 11 (a 20-min time span) and stage 13 (a 2·5h period). The RNA from the s38 gene is highly abundant, accounting for approximately 10 % of the poly A-I-RNA in the stage-12 egg chambers (unpublished observations).
Genes encoding the most abundant chorion proteins are located in two genomic clusters (Spradling et al. 1980). This organization facilitates an unusual kind of regulation in Drosophila. Prior to transcription, the DNA in each cluster selectively amplifies in the ovarian follicle cells (Spradling & Mahowald, 1980). The chorion genes are then abundantly transcribed, producing large quantities of RNA and protein within the last 5 h of oogenesis.
The s38 gene, located in the X-chromosome chorion gene cluster (position 7F1) is centrally located within a region that amplifies 16-fold (Spradling, 1981). It is surrounded by at least nine other transcription units that are active during the late stages of oogenesis (Parks, Kalfayan & Spradling, 1982 and unpublished observations). We were interested in learning whether the clustering and the amplification of these genes was actually required for properly regulated initiation of transcription. If not, we wished to define the minimum DNA sequences necessary for regulation of both timing and tissue-specific expression of the genes. We found that at most, 1·3 kilobases (kb) of the DNA at and including some of the 5′-portion of the s38 gene regulates the timed expression of this gene.
P ELEMENT-MEDIATED TRANSFORMATION
The regulation of chorion and other genes’ expression can be studied with relative ease in Drosophila now that the technique of P element-mediated germline transformation is possible (Spradling & Rubin, 1982; Rubin & Spradling, 1982). Genetic studies led to the discovery of the transposable P elements and defined two types of elements, complete and defective (Engels, 1981). More recently, both classes of P elements have been cloned and sequenced (O’Hare & Rubin, 1983). The complete elements, capable of catalysing their own transposition in a permissive environment (an M strain), are 2·9 kilobases and are terminated by inverted 31 base pair repeats. The defective elements, ranging in size from 0·5–l·6kb lack sequences from the middle portion of the DNA, but have conserved the 31 base repeat at their ends. They are capable of transposing only in the presence of complete P elements. Thus, the middle portion of the element is thought to encode a trans-acting transposase-like product while the ends are required for the integration.
In our transformation system, cloned P elements are injected directly into the pole plasm of preblastoderm, M-strain embryos. Intact P elements are used to catalyse the transposition of defective element vectors that carry between their ends the cloned gene (such as the s38 gene) to be studied. The transposon integrates into random chromosomal sites in the developing germline cells, and is subsequently stably inherited. An additional marker gene, such as the rosy+ gene is also incorporated into the transposon so that transformation events into rosy-strains can be assayed by examination of the eye colour.
EXPRESSION AND REGULATION OF THE s38 chorion gene
A marked transposon (pl A25) containing the s38 gene was constructed for use in this study (Wakimoto, Kalfayan & Spradling, 1983; DeCicco et al. 1983) and is shown in Fig. 1. The essential features of this transposon are the P element ends, a cloned restriction fragment containing the wild-type rosy gene, and an altered s38 gene. The s38 gene produces a 1·4 kb RNA that is wholly contained within a 4·7 kb Eco R1 restriction fragment (Spradling, 1981; Parks et al. 1982). This fragment has approximately 700 bases flanking the 5′ end of the gene and 2·2 kb past the 3′ end. Since no mutations in the s38 gene have been discovered, the gene was altered by the insertion of a heterologous DNA fragment (derived from the bacteriophage ml 3) into the transcribed portion of the gene. This change was designed to produce a 2·0 kb RNA that could be distinguished from the s38 transcript by size and by hybridization to the heterologous inserted DNA sequence.
Construction of the s38-M13 transposon. A 4·7 kb Eco-Rl fragment containing the s38 gene was modified by the insertion of a 572 bp fragment from M13 into the transcribed portion of the s38 gene. The arrow indicates the approximate position of the insert (open box) near the 5′ end of the message and the direction of transcription. The resulting 5· 27 kb Eco R1 fragment was cloned into the Eco R1 site of the transformation vector PV11 so that the region containing chorion and rosy+ gene sequences was flanked by P element ends (solid blocks). This plasmid was co-injected with the complete P element, p25·1 into rosy-M embryos.
Construction of the s38-M13 transposon. A 4·7 kb Eco-Rl fragment containing the s38 gene was modified by the insertion of a 572 bp fragment from M13 into the transcribed portion of the s38 gene. The arrow indicates the approximate position of the insert (open box) near the 5′ end of the message and the direction of transcription. The resulting 5· 27 kb Eco R1 fragment was cloned into the Eco R1 site of the transformation vector PV11 so that the region containing chorion and rosy+ gene sequences was flanked by P element ends (solid blocks). This plasmid was co-injected with the complete P element, p25·1 into rosy-M embryos.
Co-injection of the transposon, plA25, with the complete cloned P element, p7π25·1, into rosy-M embryos resulted in a number of transformants which were used to establish five different single insert lines. The genetic and cytological data will be presented in detail in a later publication. In addition, two multiple insert lines containing transposons at different sets of sites were obtained.
Examination of restricted genomic DNA from each single insert transformed line showed that the transposons had integrated, unrearranged, into the genome. An example of this work, showing that the transformed Eco R1 fragment containing the s38-ml3 gene is of the expected size, is shown in Fig. 2.
Analysis of transformed DNA in single insert lines. Two to three micrograms of genomic DNA from males (who do not amplify their chorion genes) from rosy-flies-and from transformed lines No. 4 and 5 was restricted with Eco Rl, separated on 1 % agarose gels and transferred to nitrocellulose filters. The DNA blots were hybridized with radiolabelled plasmid DNAs from either pUC 8 (which contains sequences homologous to the inserted M13 DNA and should hybridize specifically to the transformed gene) or p103.47 (which contains the 4·7 kb Eco Rl fragment containing the s38 chorion gene and should hybridize to both the endogenous and the transformed DNA). DNA was digested with Hindlll and Eco Rl. Lanes marked pP are Eco Rl digested cloned plA25 fragments.
Analysis of transformed DNA in single insert lines. Two to three micrograms of genomic DNA from males (who do not amplify their chorion genes) from rosy-flies-and from transformed lines No. 4 and 5 was restricted with Eco Rl, separated on 1 % agarose gels and transferred to nitrocellulose filters. The DNA blots were hybridized with radiolabelled plasmid DNAs from either pUC 8 (which contains sequences homologous to the inserted M13 DNA and should hybridize specifically to the transformed gene) or p103.47 (which contains the 4·7 kb Eco Rl fragment containing the s38 chorion gene and should hybridize to both the endogenous and the transformed DNA). DNA was digested with Hindlll and Eco Rl. Lanes marked pP are Eco Rl digested cloned plA25 fragments.
Transcripts from the transformed genes were detected in each transformed line during the proper developmental time period. Staged poly A4-RNA from each of the late stages of oogenesis as well as from pooled early stages was prepared and screened for homology to the inserted ml3 sequence. The same RNA preparations were then rescreened with DNA homologous to the s38 message. The results in all cases were the same, and a typical example taken from one of the single insert lines is shown in Fig. 3. Clearly, transcription of the transformed gene initiates in parallel with that of the endogenous gene. Differences in the relative amounts of the transformed versus the endogenous RNAs in stages 11–13 are apparent, however, and can be explained in part by the fact that this transformed gene does not amplify, while the endogenous gene is undergoing amplification between stages 11 and 13. In addition the altered message may be less stable than its normal counterpart. No transformed transcripts were detected in different stages of development (data not shown).
Developmental profile of the s38-M13 transcript. PolyA+ RNA was isolated from staged egg chambers of the transformed line No. 4 and assayed by Northern blotting for the presence of the s38-M13 transcript. The RNA was size fractionated on 1 % agarose-formaldehyde gels. Lanes labelled 1–9 contain RNA from the early stages from 5 ovaries. Other lanes contain RNAs from 100 egg chambers at the stage indicated, except for the stage 11 lane, which has only 50 egg chambers of RNA. The same blot was hybridized first with the radiolabelled pUC 8 probe (see legend for Fig. 2), washed free of hybridized material and rehybridized to labelled p103.47.
Developmental profile of the s38-M13 transcript. PolyA+ RNA was isolated from staged egg chambers of the transformed line No. 4 and assayed by Northern blotting for the presence of the s38-M13 transcript. The RNA was size fractionated on 1 % agarose-formaldehyde gels. Lanes labelled 1–9 contain RNA from the early stages from 5 ovaries. Other lanes contain RNAs from 100 egg chambers at the stage indicated, except for the stage 11 lane, which has only 50 egg chambers of RNA. The same blot was hybridized first with the radiolabelled pUC 8 probe (see legend for Fig. 2), washed free of hybridized material and rehybridized to labelled p103.47.
A second transposon, pACSn, was constructed to begin to test the limits of the DNA sequences required for developmental regulation of the s38 gene. In this plasmid 1·3 kb at the 5′ end of the 4-7 kb fragment containing the s38 gene was fused to the lac z gene of E. coli (Fig. 4). The DNA was injected as described above and RNA from the transformed lines was analysed for hybridization to the cloned lac z gene. As witnessed with the larger fragment, the transposed gene was regulated with the proper developmental timing for the s38 gene (Fig. 5).
Construction of the s38-lacz fusion genes in three reading frames. The indicated steps were carried out according to standard procedures. The 1·3 kb Eco Rl-Bam HI fragment at the 5′ end of the s38 gene was ligated into a similarly digested transformation vector, Carnegie 1. A fragment from the lac z gene of E. coli was fused into the s38 DNA at the Bam HI site in all three orientations, using different enzymes in a polylinker. An 8·1 kb Sal fragment containing the rosy+ gene was ligated into the vector between the P element ends. Solid blocks indicate P element ends. The open box represents the pBR322 backbone sequences necessary to propagate the plasmid in bacteria. They are lost in the integration into Drosophila DNA.
Construction of the s38-lacz fusion genes in three reading frames. The indicated steps were carried out according to standard procedures. The 1·3 kb Eco Rl-Bam HI fragment at the 5′ end of the s38 gene was ligated into a similarly digested transformation vector, Carnegie 1. A fragment from the lac z gene of E. coli was fused into the s38 DNA at the Bam HI site in all three orientations, using different enzymes in a polylinker. An 8·1 kb Sal fragment containing the rosy+ gene was ligated into the vector between the P element ends. Solid blocks indicate P element ends. The open box represents the pBR322 backbone sequences necessary to propagate the plasmid in bacteria. They are lost in the integration into Drosophila DNA.
Developmental profile of the s38-lac z transcript. The Northern blot shown here was prepared essentially as described in the legend for Fig. 3 except that RNA from 100 egg chambers was used in each of stages 10–13 and from 200 egg chambers in stage 14. The hybridization probe was a plasmid, pMC1871 containing the lac z gene.
Developmental profile of the s38-lac z transcript. The Northern blot shown here was prepared essentially as described in the legend for Fig. 3 except that RNA from 100 egg chambers was used in each of stages 10–13 and from 200 egg chambers in stage 14. The hybridization probe was a plasmid, pMC1871 containing the lac z gene.
DISCUSSION
We have shown that a 1·3 kb DNA sequence, containing approximately 700 nucleotides of sequences flanking the 5′ end of the s38 gene is capable of directing properly timed synthesis of the s38 RNA. We are currently attempting to narrow this sequence down to a minimum. In addition, the data indicate that the s38 gene does not need to reside in its normal position within the cluster of DNA that amplifies. Furthermore, properly timed transcription can occur in the absence of amplification (data not shown).
Several other investigators have used P element-mediated transformation to study developmental regulation of other Drosophila genes (Scholnick, Morgan & Hirsch, 1983; Goldberg, Posakony & Maniatis, 1983; Hazelrigg, Levis & Rubin, 1984; Spradling & Rubin, 1983). The conclusion that emerges from these studies is that Drosophila genes do not require a great deal of sequence flanking each gene to direct developmentally regulated expression. Our results address directly expression at the level of transcription. We have shown that no more than 700 nucleotides beyond the 5′ end of the s38 gene are required for directing the precisely timed initiation of transcription of this gene.
The chorion genes differ from the other genes that have been studied, however, in their unusual clustered organization and developmentally controlled amplification. It is therefore interesting to note that neither clustering nor amplification are required for proper developmental regulation of RNA synthesis.
Recent transformation studies in this laboratory have shown that small fragments of DNA within central region of the amplified cluster of chorion genes possess the ability to amplify when removed from the cluster, but that the level of amplification is sensitive to chromosomal position (DeCicco & Spradling, manuscript in preparation). The sensitivity to chromosomal position can be buffered in part by surrounding the amplifying sequences with DNA from the amplified cluster. Interestingly, transcription of the transformed s38 gene is much lower than that of the endogenous gene. The levels of transformed RNA vary slightly depending on their chromosomal position, as noted for the rosy+ gene (Spradling & Rubin, 1983), the Ddc+ gene (Scholnick et al. 1983) and the Adh+ gene (Goldberg et al. 1983). However, the overall level of s38 RNA is lower than predicted, even after accounting for the absence of amplification. This is perhaps due to a decreased stability in the altered s38 gene. We are also currently investigating the possibility that, as for amplification, additional surrounding sequences from the cluster increase the level of synthesis from the transformed gene due to distant ‘enhancer-like’ sequences.
Our current strategy for defining the minimum sequences required for timing and tissue-specific expression of the s38 gene involve the construction of a series of deletions from the 5′ and 3′ ends of the T3 kb regulatory fragment. We expect that this approach will tell us whether the timing control is separable from that for tissue-specific regulation of the gene.