Cell. Vol. 16, 575-566,

March

1979.

Copyright

8 1979 by MfT

Sequence Organization of Two Recombinant Plasmids Containing Genes for the Major Heat Shock-Induced Protein of 0. melanogaster Elizabeth A. Craig, Brian J. McCarthy and Samuel C. Wadsworth* Department of Biochemistry and Biophysics University of California, San Francisco San Francisco, California 94143

Summary We have isolated recombinant DNA clones which include cDNA and chromosomal DNA sequences of the major heat shock-inducible gene of Drosophila. With the cDNA fragments used as specific hybridization probes, DNA:DNA reassociation and in situ hybridization analysis demonstrated that the DNA sequences are repeated approximately 7 times in the haploid Drosophila genome, and that gene sequences are present at both the 87A and 87C loci on the cytological map. The cloned cDNA and homologous cloned chromosomal DNA hybridized to mRNA which translated in vitro into the major 70K heat shock-specific protein. Here we summarize a study of the organization of genes coding for the 70K heat shock-specific protein contained in the two recombinant chromosomal DNA plasmids pG3 and pG5. On the basis of R loop hybridization experiments and restriction enzyme analysis, we conclude that a 14 kb fragment, G3, contains three copies of the gene coding for the 70K protein. A second 9.2 kb fragment, G5, contains one copy of the gene coding for the 70K protein. Hybridization of labeled poly(A)-containing RNA to restriction endonuclease-cleaved DNA indicates that the mRNA coding regions in G3 and G5 are each approximately 2100 bp long. The three tandemly repeated genes of G3 are separated by approximately 1400 bp of spacer DNA. The two internal spacer regions in G3 appear to be identical, whereas differences in restriction enzyme sites indicate that the sequences adjacent to the cluster differ from the internal spacer and from each other. Introduction The heat shock genes of Drosophila comprise an especially interesting inducible system which is amenable to detailed study. Expression of the heat shock genes can be modulated quite easily; gene expression is induced within minutes of changing the culture temperature from 25 to 37°C (Ritossa, 1962; Ashburner, 1970). In D. melanogaster, puffs are induced at at least eight specific loci in polytene chromosomes. Discrete species of RNA are * Present address: Worcester ogy, Shrewsbury, Massachusetts

Foundation 01545.

for Experimental

Biol-

synthesized (Spradling, Penman and Pardue, 1975) and translated into proteins specific for the “heat shock” response (Tissieres, Mitchell and Tracy, 1974; Lewis, Helmsing and Ashburner, 1975; McKenzie, Henikoff and Meselson, 1975; Mirault et al., 1978; Moran et al., 1978). The puffs are clearly sites for transcription of the newly synthesized RNA species since radioactive RNA precursors added at the time of the heat shock are incorporated into RNA at these loci (Ritossa, 1964; Tissieres et al., 1974; Bonner and Pardue, 1976), and radiolabeled mRNA isolated from heat-shocked cells hybridizes to these chromosomal loci as well (McKenzie et al., 1975; Spradling et al., 1975; Bonner and Pardue, 1976; Spradling, Pardue and Penman, 1977). In general, the heat shock loci are not closely linked; puffs are observed on three chromosome arms (Ashburner, 1970). Even the heat shock-specific regions at the complex 87AlC loci are separated by at least lo-12 fine bands, corresponding to at least several hundred kilobases of DNA. Drosophila cells in culture respond to a heat shock by synthesizing a set of polypeptides indistinguishable from those made in excised salivary glands (Moran et al., 1978). Existing polysomes break down upon heat shock and are replaced by polysomes that carry heat shock-specific mRNA as the major class of mRNA molecules (McKenzie et al., 1975). Thus polysomal RNA from heat-shocked tissue culture cells provides a convenient enriched source of mRNA molecules coding for these proteins. Previous studies have indicated that the most abundant of these mRNA species hybridizes to the 87A and 87C loci and encodes the 70K heat shockspecific protein (McKenzie et al., 1975; Spradling et al., 1975). Sucrose gradient sedimentation experiments demonstrated that the major heat shock-induced mRNA sediments at approximately 20s (McKenzie et al., 1975; Spradling et al., 1977; Moran et al., 1978) and translates in vitro to produce the 70K heat shock-induced protein (Moran et al., 1978). The most precise quantitation of the size of this major heat shock-induced mRNA comes from the work of Henikoff and Meselson (1977); electrophoresis in methylmercury agarose gels, which substantially eliminates secondary structure in RNA molecules, indicates a length of approximately 2.7 kb for this RNA species. While the most prominent heat shock-induced RNA species hybridizes to 87A and 87C, a minor temperature-induced RNA species (band A4 of Spradling et al., 1977) is reported to hybridize only to 87C. Lis, Prestidge and Hogness (1978) have isolated a series of recombinant DNA clones by hybridization to heat shock-specific poly(A)-containing cytoplasmic RNAs. One of the fragments,

Cell 576

cDm703, carries a basic 1.5 kb Hind III repeating unit that hybridizes only to 87C and is complementary to RNA species of 4, 3 and 2 kb in length. Our goal was to isolate recombinant DNA clones carrying genes for each of the heat shock-induced proteins and, if feasible, to isolate all copies of any reiterated heat shock genes. Since our long-term objective was to use these cloned genes in the study of the regulation of transcription, we chose a tissue culture line as the source of DNA for transformation. We report the isolation and characterization of recombinant cDNA and chromosomal DNA clones specific for the 87A and 87C heat shock loci. Results Identification of cDNA Clones Specific for the Complex Locus 87Af 87C Since mRNA coding for heat shock proteins represents a large proportion of poly(A)-containing polysomal RNA in cells incubated at 3PC, we sought to use cDNA cloning techniques to isolate recombinant DNA clones for heat shock-specific loci. Characterization of the in vitro translation products of the RNA preparations used as substrates for synthesis of double-stranded cDNA in these experiments demonstrated the presence of mRNA in high abundance for seven heat shock-specific proteins. Prior to insertion into E. coli, the doublestranded DNA was cut with endonuclease Hae III and ligated into plasmid vectors using synthetic restriction endonuclease recognition sequences as described in Experimental Procedures. DNA from four of the first six transformants isolated hybridized with labeled heat shock-specific RNA; two were chosen for further study. These inserts, designated HS4 and HS12, are 320 and 180 bp in length, respectively, and do not cross-anneal. Radiolabeled probes were prepared from HS4 and HS12 DNA and hybridized to squashes of salivary gland polytene chromosomes. As shown in Figure 1, 3H-cRNA or 3H-nick-translated DNA from both HS4 and HS12 hybridize to the 87A/87C heat shock-specific loci in a ratio of 1 :1.5. Hybridization was not detected at any other loci. To determine which of the heat shock-specific proteins is encoded at the 87A/87C loci, we applied the hybridization selection method of J. B. Lewis et al. (1975) to purify mRNA complementary to the HS4 and HS12 DNA (Figure 2). The translation products in lanes (c) and (e) result from mRNA species that hybridized to HS4 and HS12 sequences, respectively; each contains a prominent 70,000 dalton protein as well as several less intense bands migrating slightly more slowly than the 70K protein. The HS4 and HS12 DNAs clearly fail to hybridize with mRNA encoding either the 84K or 68K heat

Figure

I.

In Situ Hybridization

(a) 3H-cRNA synthesized from exposure. (b) 3H-nick-translated 28 day exposure.

of HS4 and HS12 Sequences HS4: 15,000 cpm per slide; 22 day HS12 DNA: 46.000 cpm per slide;

shock proteins, nor does pBR322 DNA alone hybridize with any heat shock mRNA (lane g). The faint bands that appear in lane (g) also appear in lanes (a), (c) and (e), and represent endogenous template activity. The prominent smaller polypeptide seen only in lane (e) is not reproducibly observed in mRNA hybridization selection experiments and is perhaps a degradation product of the 70K protein: this phenomenon has not been investigated further. Preliminary Analysis of 87AI87C Gene Structure Using cDNA Clones DNA reassociation experiments were performed to estimate the number of genes encoding the 70K protein in the Drosophila genome. Isolated cDNA inserts of plasmids HS4 and HS12 were labeled in vitro and reassociated in the presence and absence of Drosophila DNA. The Drosophila DNA increased the rate of reassociation by the factor expected if the genes were repeated about 7 times per haploid genome (data not shown). To analyze the organization of the 70K protein genes, Drosophila DNA was cleaved with the restriction endonucleases Eco RI, Hind III, Pst I and Barn HI, electrophoresed on a 1% agarose gel and transferred to a nitrocellulose filter (Southern, 1975). Hybridization with 32P-nick-translated probes

Sequence 577

Organization

Figure 2. Gel Electrophoresis yapatite-Purified mRNA

in Cloned

Heat

of Translation

Shock

Genes

Products

of Hydrox-

Approximately 0.2 pg of either HS4, HS12 or pBR322 DNA were cleaved with Hind III, denatured and hybridized with heat shock mRNA as described in Experimental Procedures. The in vitro translation products were electrophoresed on a 10% acrylamide gel. Lanes (a-g) show mRNA-dependent reticulocyte translation assays with the following additions: (a) no RNA; (b) RNA not hybridized to HS4 DNA; (c) RNA purified by hybridization to HS4 DNA: (d) RNA not hybridized to HS12 DNA; (e) RNA purified by hybridization to HS12 DNA; (f) RNA not hybridized to pBR322 DNA; (g) RNA purified by hybridization to pBR322 DNA.

synthesized from plasmids HS4 and HSI 2 was used to determine the molecular weight distribution of homologous genome fragments. A complex pattern was obtained (Figure 3). Eco RI and Hind III digests showed similar patterns of hybridization with both probes. Most of the hybridization was to two large Eco RI fragments of approximately 27 and 17 kb, and to two large Hind III fragments of 33 and 14 kb. Estimates of the sizes of these large fragments are imprecise because of poor resolution in the gels in the larger size range. HS4 hybridized strongly to a 1.2 kb Pst I fragment and weakly to 4.7, 4.3 and 2.7 kb fragments. HS12 hybridized strongly to a 2.3 kb Pst I fragment and weakly to a 3.7 kb fragment. The pattern of hybridization to Barn HI-digested DNA is the most complicated. HS12 hybridized predominantly to a 3.5 and a 2.0 kb fragment; hybridization was also detected to 4.6, 4.1 and 2.1 kb fragments. HS4 hybridized to five distinct bands ranging in size from 1 .8 to 4.6 kb. The 3.4 kb fragment always showed the most intense signal of hybridization. Hybridization to several high molecular weight fragments was also detected. Isolation Having

of Genome Clones identified two restriction

endonucleases

Figure 3. Hybridization of HS4 Several Restriction Endonucleases

and

HS12

to DNA

Cleaved

with

5 pg of Schneider’s ceil Drosophila DNA were cleaved with Eco RI, Hind Ill, Pst I or Barn HI, electrophoresed on a 1% agarose gel for 6 hr and transferred to filters as described in Experimental Procedures. Recombinant plasmids HS4 and HS12 were nicktranslated and hybridized (2.5 x lo8 and 5 x lo6 cpm, respectively) to the DNA bound to the filters. Exposure was for 2 days. The numbers on the left indicate the size in kilobases and position of Eco RI-cleaved lambda DNA fragments; those on the right indicate the position of Hind Ill-cleaved adenovirus 2 (Ad2) DNA.

Eco RI and Hind III that generate relatively large Drosophila DNA fragments containing the heat shock genes at 87A/87C, we attempted to isolate chromosomal DNA clones containing these DNA fragments. DNA was cleaved with Eco RI or Hind III and fractionated on sucrose gradients. Material sed imenting faster than 11-l 2 kb, about 10% of the total DNA, was pooled. The pooled DNA was I igated to Eco RI- or Hind Ill-cleaved plasmid and used to transform E. coli RRl. The resulting banks were screened for recombinant plasmids containing the 70K genes using the 32P-labeled insert of HS4. Two plasmids which hybridized with HS4 were selected for further study: G3 contained a 14 kb Hind III fragment and G5 a 9.2 kb Eco RI fragment. To show that the chromosomal DNA clones G3 and G5 also coded for the major heat shock proteins (the 70K group), two types of in vitro protein synthesis experiments were performed: a positive mRNA selection (J. B. Lewis et al., 1975) and a hybrid-arrested translation experiment (HART) (Paterson, Roberts and Kuff, 1977). Examples of these data are shown in Figure 4. It can be seen clearly in the HART experiment that translation of the 70K band and the several faint bands just above it (indicated by the arrows) are blocked by hybridization to G3 or G5 DNAs (lanes a and c). The adjacent

Cell 578

Figure

4. Hybridization-Arrested

Translation

and mRNA

Selection

Analysis

of Chromosomal

Clones

Approximately 1 pg of G3 DNA (cleaved with Hind Ill) or G5 DNA (cleaved with Eco RI) was denatured and hybridized with heat shock mRNA as described in Experimental Procedures. (a) Translation of heat shock mRNA hybridized with G3 DNA. (b) An equal aliquot of sample (a), heat-denatured prior to translation. (c) Translation of heat shock mRNA hybridized with 65 DNA. (d) An equal aliquot of sample (c), heat-denatured prior to translation. (e) Translation of mRNA purified by hybridization to G3 DNA. (f) Translation of mRNA that failed to hybridize with G3 DNA. (g) Translation products of mRNA-dependent reticulocyte system without added RNA. Kc cells and salivary glands were labeled as described in Experimental Procedures. The separate panels are from different experiments; electrophoresis was in 12% acrylamide gels in each case.

lanes (b and d) demonstrate that upon denaturation of the hybrids, the 70K proteins are again a prominent translation product. The converse experiment, mRNA selection, confirms the results of the HART experiment: mRNA that hybridizes to G3 DNA translates into the 70K group of proteins (lane e). mRNA selected by hybridization to either G3 or G5 DNA does not translate into any other heat shock-specific proteins. Thus both types of in vitro translation experiments show that mRNA species for the major 70K protein and the fainter bands above it share sequence homology. These proteins are also evident in Kc cells or explanted salivary glands labeled in culture, as indicated by arrows in Figure 4. Restriction Map of G3 A physical map of G3 was constructed by cleavage with 12 restriction endonucleases (Figure 5). Inspection of the stained gels after certain endonuclease digests indicated that some fragments were present in greater than equimolar amounts; in

particular, the 1.1 and 2.3 kb Pst I fragments, the 2.7 kb Hint II fragment and the 3.4 kb Barn HI fragment exhibited this characteristic (Figure 6). Densitometric analysis reveals the more intense staining of several of the bands containing these fragments relative to larger fragments in the same digest. The sum of the limit products of digestion with these three enzymes is less than the total length of the DNA. The correct length is obtained, however, if we assume that the 3.4 kb Barn HI fragment is present twice in G3 and that the 1.1 and 2.3 kb Pst I fragments and the 2.7 kb Hint II fragment are present 3 times. An extensive analysis of partial digestion products, subcleavage of isolated fragments and cleavage of subclones of the internal Barn HI-Barn HI fragment and Barn HI-Hind III terminal fragments confirmed the observation that blocks of sequences are reiterated 3 times in G3. This arrangement of sequences is most clearly demonstrated by digestion of the subclones of G3 with Hpa I and Sac I (Figure 7). Neither of these enzymes cleaves

Sequence 579

Organization

in Cloned

Heat

Shock

Genes

63

Figure

5. Map of the mRNA

Coding

Regions

of G3 and G5

The distances between restriction endonuclease sites were determined by the relative rates of migration of DNA fragments to Hind IIIcleaved Ad2 and Hae Ill-cleaved @X174 DNA in 1 and 1.5% agarose gels. The placement of sites was determined from analysis of partial digestion products and cleavage of isolated and subcloned fragments with additional enzymes. The boundaries of the mRNA coding regions, designated by the solid lines above the maps, were determined from the results of hybridization of labeled polysomal RNA to the clones as described in the text and in the legend to Figure 6. The brackets indicate the boundaries of subclones BHI, BH15 and B6. The internal Barn HI subclone BE is labeled at two locations. These sequences are repeated twice in G3, and since the 12 restriction endonucleases used thus far detected no differences in sequence, it is unknown which unit 68 represents.

the vector pBR322. Hpa I cleaves G3 twice, yielding a 3.3 and a 1.6 kb fragment. Neither of the two Barn HI-Hind III subclones (BHI, which is the right-hand 4.2 kb fragment, and BH15, which is the left-hand 3.3 kb fragment) is cut by Hpa I (Figure 5). The Barn HI subclone 88 is cut once with Hpa I, yielding a linear molecule of 7.7 kb. Since G3 is cleaved twice with Hpa I, the Barn HI fragment must be repeated twice, and they must be adjacent to yield a 3.3 kb fragment upon cleavage. Sac I does not cleave BHI but does cleave BH15 (the left-hand fragment) once. 88 is also cleaved once, and since this fragment is duplicated, the results indicate that G3 is cleaved 3 times with Sac I, consistent with 3 fold reiteration of some sequences in the insert. Mapping of mRNA to G3 To determine the number and size of genes for the 70K protein contained in these genome DNA fragments, an R loop hybridization experiment was performed. Poly(A)-containing RNA isolated from polysomes of heat-shocked Kc cells was incubated

under R loop conditions with Hind Ill-cleaved G3 and spread for electron microscopy. Hind III releases the Drosophila insert from the plasmid cloning vehicle pBR322, which also served as a size marker. We observed several arrangements of R loops in G3. Some molecules contained a single R loop, in the center of the DNA fragment (Figure 8a). Figure 8c illustrates a second class of molecules in which the center of the molecule lacks an R loop but is flanked by an R loop on either side. Other molecules with two R loops contain the central one and either the left or right “right’‘-hand R loops (Figures 8e and 89). Measurements of all the molecules show clearly that the outside R loops are located at different distances from the ends of the DNA fragment, and we have arbitrarily designated the shorter of these lengths as the “left” end of G3. Considered together, these observed arrangements of R loops indicate that G3 contains three mRNA coding regions. In fact, occasional molecules were observed with three R loops, but in nearly all cases one or more of the R loops was

Cell 580

all three sites is approximately 2000 bp. The average inter-R loop distance is approximately 1550 bp. The G3 molecule itself is nearly 14,500 bp long, consistent with size determinations on agarose gels. We have no evidence for intervening sequences, but cannot exclude the existence of short intervening sequences (