Vol. 3, No. 10

MOLECULAR AND CELLULAR BIOLOGY, OCt. 1983, p. 1724-1729 0270-7306/83/101724-06$02.00/0 Copyright © 1983, American Society for Microbiology

Analysis of DNA Structural Patterns and Sequence Organization at the Larval Cuticle Locus in Drosophila melanogaster JOEL C. EISSENBERG* AND SARAH C. R. ELGIN Department of Biology, Washington University, St. Louis, Missouri 63130 Received 6 June 1983/Accepted 6 July 1983

We examined the pattern of DNA organization at the larval cuticle gene complex 44D of Drosophila melanogaster, using micrococcal nuclease and the 1,10-phenanthroline-cuprous complex. The initial cleavage patterns obtained with both reagents exhibited "gaps" at the positions of each of the genes examined, as well as at a pseudogene sequence contained within the complex. An additional gap for which no gene exists was observed for both patterns. The cleavage pattern obtained with micrococcal nuclease was unaltered, at a level of resolution of ±50 base pairs, in a mutant containing a transposable element. Analysis of the sequence data from this 5.5-kilobase gene cluster indicated that the sequence per se, and not the general base composition, is a dominant factor in determining the patterns observed.

Fine-structure analysis of the eucaryotic chromosome has identified specific features of chromatin architecture associated with gene expression and functional organization. For example, in nearly all cases examined, sites hypersensitive to DNase I have been found within 1 kilobase 5' to active or potentially active genes (4). These DNase I-hypersensitive sites are often tissue specific, corresponding to the developmental specificity of gene expression. Other cases of structural specificity in chromatin have been observed; for example, yeast centromeric chromatin shows a high degree of organization (2), suggesting precise nucleosome positioning in this region. Whereas the molecular basis for these structures has not been established, it seems likely that the underlying DNA sequences play an important role in the local organization of chromatin. Consequently, it is of interest to examine patterns in DNA structure. Recent experiments from this laboratory, using diverse DNA cleavage reagents to digest protein-free DNA substrates (primarily locus 67B1 of D. melanogaster), have revealed a striking pattern; sequences within coding regions show relatively low sensitivity to cleavage, whereas flanking nontranscribed sequences show relatively higher sensitivity (3, 9). The similarity of digestion patterns obtained with such seemingly unrelated reagents as micrococcal nuclease and 1,10-phenanthroline-cuprous complex suggests a common basis in DNA conformation, determined by the underlying base sequence.

We have now used DNA cleavage reagents as probes of DNA structure at a second D. melanogaster gene cluster, the larval cuticle gene complex at 44D. In addition to a set of four developmentally regulated genes, the cuticle gene complex contains a pseudogene sequence (13). We wished to test whether a pseudogene sequence, apparently incapable of encoding a usable transcript, would retain the DNA structural properties of a functional gene as assayed by sensitivity to cleavage. In addition, a mutant with a 7.3-kilobase insertion is available at this locus, permitting one to look at the consequences of such a change in genome organization. The cuticle gene complex has been completely sequenced (13; this paper), allowing a comparison of the cleavage pattern with base composition and a preliminary search for corresponding sequence patterns. MATERIALS AND METHODS Plasmid and phage DNA preparation. Recombinant phages XDmLCP2 and XDm2/3LCP1 and the recombinant plasmids pCPI-11 and pCPIII-9, described in reference 13, were all generous gifts of M. Snyder. Recombinant phage DNA was prepared by polyethylene glycol precipitation of intact phage, followed by phenol extraction and ethanol precipitation of phage DNA. Recombinant plasmids were prepared from cleared lysates by banding supercoiled DNA twice in cesium chloride-ethidium bromide density gradients, extracting with isopropanol-water (9:1), and precipitating three times with ethanol. All operations with recombinant DNA were carried out in accordance with National Institutes of Health guidelines.

1724

DNA STRUCTURE OF D. MELANOGASTER CUTICLE LOCUS

VOL. 3, 1983

a.

H M S Beogle

I

JR

I

A DmLCP2

ADm2/3LCP1

W

1 Ils IIIt9r I~ m p #

I

I

1725

PH~~~~

lpCp"

'pCPI-lll ;

a --

-

-

s_

5 kb

FIG. 1. Schematic map of the larval cuticle gene complex of D. melanogaster, showing the position of the HMS Beagle transposable element in the 2/3 mutant. Lines below the map indicate the extent of the genomic sequences included in the phage clones XDmLCP2 (13) and XDm2/3LCP1 (14). Also indicated are the sequences contained in the plasmid clones pCPI-11 and pCPIII-9 used in the indirect end-labeling analyses (13). Letters above the map represent cleavage sites for the following restriction enzymes: R = EcoRI; H = HindIlI; S = Sall; B = BamHI, K = KnpI; P = HpaI; X = XhoI. Letters with dots above them indicate sites used in the sequencing strategy. Kb, Kilobases. Digestion protocol and indirect end-labeling analysis. Digestions of phage DNA with micrococcal nuclease (Worthington) were performed essentially according to reference 9, except that the reactions were terminated by the addition of EDTA to 2 mM, followed by heating for 10 min at 65°C. Digestions with 1,10phenanthroline-cuprous complex were carried out essentially according to reference 3. Agarose (Seakem) gel electrophoresis, Southern transfers, 32P-labeling of plasmid probes, and dextran sulfate hybridizations were carried out as described in references 9 and 17. Washes were performed in (i) 5x SSC (SSC = 0.15 M NaCI plus 0.015 M sodium citrate)-2x Denhardt solution-0.2% sodium dodecyl sulfate for 1 h at 65°C, (ii) 5x SSC-0.2% sodium dodecyl sulfate for 1 h at 65°C, and (iii) 2x SSC for 1 h at 65°C. Autoradiography was carried out at room temperature, using Kodak XAR-5 X-ray film. Sequencing methodology. The BamHI-BamHI fragment containing the unsequenced portion of the cuticle complex (13) was cloned from XDmLCP2 into the BamHI site of pUC9 (15). Sequencing was then performed on subclones from this plasmid (i) from the HindIll site and (ii) an adjacent HpaI site, rightwards, and (iii) from the KpnI site leftwards. Sequencing was performed essentially according to Maxam and Gilbert (11) as modified by Smith and Calvo (12). All fragments to be sequenced were labeled by filling in 3' ends with reverse transcriptase, fractionated on 4% acrylamide gels, and purified over DE52 cellulose (Whatman).

RESULTS We have used an indirect end-labeling approach to map the micrococcal nuclease cleavage pattern at the larval cuticle gene locus. Purified DNA from a bacteriophage lambda clone, XDmLCP2, containing the entire complex was digested to various extents with micrococcal nuclease and then cut completely with EcoRI. The DNA was then fractionated on an

agarose gel, blotted to nitrocellulose paper, and probed with the plasmid subclone pCPI-11 (see map, Fig. 1). The results of this experiment are shown in Fig. 2. The regular pattern of strong nuclease cleavage sites is interrupted at positions corresponding to genes II and III. In addition, an obvious gap in the pattern occurs at the pseudogene sequence, %lI. By using a different subclone, pCP IlI-9, to probe from the other end of the EcoRI fragment, it is possible to confirm the positions of these gaps (Fig. 3). Mapping from this side clearly shows an additional gap appearing in the position of gene I. An identical pattern is obtained when purified genomic DNA is used (data not shown). The chemical cleavage reagent 1,10-phenanthroline-cuprous complex also shows a strong pattern of preferential cleavage on naked DNA, a pattern which is virtually identical to that generated by micrococcal nuclease at locus 67B1 (3). When XDmLCP2 DNA is cleaved to various extents with this reagent and the same EcoRI fragment is probed with either plasmid probe, a pattern of fragments is obtained which is essentially identical to the one generated by micrococcal nuclease digestion at this locus (Fig. 4). In addition to the gaps in both patterns which correspond to the functional genes and the pseudogene, there is an additional region which might be scored as a gap, mapping between gene I and the pseudogene. No transcript has been detected from this region (M. Snyder, personal communication); however, it seemed possible that another pseudogene might be present. When the position of this gap was mapped and compared with the available sequence data (13), a substantial portion was found to overlap a

1726

EISSENBERG AND ELGIN

A

D~ ~

MOL. CELL. BIOL.

Whereas a consensus sequence for the target of micrococcal nuclease cutting has been proposed (7), sequence level analyses of the cutting preferences for this (3) and other enzymes (1, 6) have shown a significant effect of surrounding sequences upon cutting efficiencies. We have examined the possible effects of neighboring sequences on the preferential cutting pattern seen here by examining a cloned sequence, XDm2/3LCP1, derived from the cuticle gene mutant 2/3 (14). In this case, a 7.3-kilobase transposable element has been inserted at the TATA box site 30 bases 5' to gene III. Micrococcal nuclease digests of both mutant and wild-

B

.U. .~ %. ...S

NU',

I

I

II FIG. 2. Cleavage pattern of micrococcal nuclease at the larval cuticle gene complex, visualized with the plasmid probe pCPI-11. Samples of XDmLCP2 DNA were digested for (A) 3, (B) 4. (C) 5. and (D) 6 min with micrococcal nuclease and then cut to completion with EcoRI. Samples were electrophoresed in 1.25% agarose, blotted to nitrocellulose paper. and probed with 32P-labeled pCPI-11. The map to the right of the autoradiogram indicates the relative positions of genes in the larval cuticle gene complex.

I I

i

region of DNA which was not yet sequenced. We therefore subcloned a BainHI fragment containing this region from XDmLCP2 and sequenced it (Fig. 5), completing a sequence map for the entire complex. Analysis of this sequence revealed no significant open reading frames, and dot matrix comparisons to the other cuticle gene sequences showed insufficient homology to suggest a pseudocuticle gene (by the criteria of eight of nine base matches; data not shown). In addition, the low guanine plus cytosine (G+C) content of this sequence, approximately 30% (Fig. 6), is more typical of noncoding sequences. Taken together, these data seem to rule out another coding region at this position.

lIII

FIG. 3. Cleavage pattern of micrococcal nuclease at the larval cuticle gene

complex, visualized with the plasmid probe pCPIII-9. Samples were treated as in the legend to Fig. 2, except that the blot was probed with 32 P-labeled pCPIII-9.

VOL. 3, 1983

DNA STRUCTURE OF D. MELANOGASTER CUTICLE LOCUS

III

1727

type clones, cut with EcoRI and sized on agarose, were blotted to nitrocellulose paper and

p

probed with the gene I subclone pCPI-11. The results of this side-by-side comparison of micrococcal cleavage products (Fig. 7) reveal no apparent alteration in the fragment patterns throughout the homologous sequences. Interestingly, a gap appears in the position of the 260base pair terminal direct repeat, although the remainder of the transposable element DNA examined exhibits a periodic cleavage pattern characteristic of noncoding DNA.

I

:.;

I

III

A

B

C

D

FIG. 4. Cleavage pattern of 1,10-phenanthroline at the larval cuticle gene complex. Samples of XDmLCP2 DNA were digested with 1,10-phenanthroline-cuprous complex for (A, D) 3 and (B, C) 4 min and then cut to completion with EcoRI. Samples were electrophoresed in 1.25% agarose, blotted to nitrocellulose paper, and probed with (A, B) 32P-labeled pCPI-II or (C, D) 32P-labeled pCPIII-9.

DISCUSSION The results obtained upon micrococcal nuclease cleavage of cuticle gene DNA are consistent with those previously observed at the heat shock locus 67B1 (9). The regular pattern of fragments obtained across the cuticle gene locus is interrupted at the positions of each of the three genes examined. A similar pattern distinguishing gene and spacer has been obtained at a number of other loci transcribed by RNA polymerase II in D. melanogaster (M. A. Keene and S. C. R. Elgin, manuscript in preparation). In addition, the pseudogene sequence in the cuticle gene cluster shows a similar reduced sensitivity to nuclease and absence of a regular 200-base pair periodic cleavage pattern. The chemical cleavage reagent 1,10-phenanthroline-cuprous com-

is 6 38 66 20 96 46 56 76 la CTTATTAACTGTTGTGACAAAATAGAOT TTT*TATCAACTZGGCAATACAATATCTAAAAGTT CTTTCO"GCGCrTTTATGCATTATTTAT

116

126

130

146

156

16

176

166

246

256

2"

276

266

i6

2"

296

36

TGAGGTGCATTTGCAGACAGACAACAACCAACAAATTCCCAGTCTTATTTTAAMAGAGA^ATOCTATACTCAGCTOGAAGTGTGACCGGTATATTT 210

226

236

CAATACTTTTTTAATATCGGTAAAAATTTGGAAACAAATAATAATATGAAAAAAATTATTGAATTGTTCAAAACGTATAACGTAGCAGTTTTAGGCCGT

310 320 336 346 356 36 376 366 39a 480 TTGTGGGCCTCTGGAATCTTCATATCTAATCTTAACTCTCTAGTTTTTG AATCGAGCCTTCATACGGACATACTCTGCTGTTGATCGCTTTTAAAAT

416

420

430

440

456

466

476

460

496

5OO

S16

520

536

546

556

S

576

568

596

666

616

620

630

646

656

ATATATACTTTATATGGTATAAACGCCTCCTTCTACCTGTTGCATAATTTTCCTATATTCCATTTTACTTCGAGTAACGAGTATTATCATTCAGTTATC

ICGTATGTTTTTATCTCATTATCTTCTAGATTATTTTACGGTTCAATCTCTTTACTTTTATOCTGCCTTCTATTGGTTATCACACAATATAATAAGACAA

FIG. 5. Sequence of spacer DNA between gene I and the pseudogene of the larval cuticle complex. This begins at position 1,697 of Fig. 8 of reference 13.

sequence

1728

EISSENBERG AND ELGIN

MOL. CELL. BIOL.

a

100,

I

Ill}

*

'I U

i 82,];t

4, XII

111

IHI0]ltt?ifIIJlItt t flt,ttt!

FIG. 6. (a) Computer-generated plot of the percentage of G+C residues within the EcoRl fragment of XDmLCP2 analyzed. Percentages were calculated for 100 bases around a given position, in single-base increments across the fragment. (b) Micrococcal nuclease cleavage map of the same EcoRI fragment, showing the relative positions of the genes and pseudogene. The map is drawn to the same scale as in part (a) to facilitate comparisons. Arrows below the map indicate the positions of strong micrococcal nuclease cleavages in the DNA. Vertical lines indicate faint cuts within gaps, visible upon long exposure. Short arrows indicate cleavages of intermediate intensity. The star indicates the position of a gap in the micrococcal pattern which does not correspond to a known gene or pseudogene sequence.

plex behaves in an identical fashion to micrococcal nuclease at this locus. The influence of surrounding DNA on nuclease cleavage patterns suggests a strong conformational component to DNA-ligand interactions. Even sequence-specific enzymes such as the restriction enzymes PstI and Hinfl and highly sequence-dependent reagents such as the intercalators actinomycin D and Hoechst 33258 show an apparent sensitivity to regional sequence variations (1, 6, 10, 16). It was therefore of interest to test a naturally occurring sequence rearrangement for its effect on the observed pattern of micrococcal nuclease cleavage in the cuticle gene complex. In a direct comparison of the wild-type and mutant 2/3 patterns at this locus at the level of resolution of agarose gel electrophoresis, no detectable perturbation was observed at the insertion boundary. It would be interesting to examine the very short-range effects of such insertions at the DNA sequence level. One possible common basis for the virtually identical cleavage patterns observed for these mechanistically diverse reagents is suggested by the observation that the gene and pseudogene sequences are relatively G+C rich (about 60%) compared with the noncoding spacer DNA (30 to 40%; see Fig. 6). There are two lines of evidence which seem to argue against the simple hypothesis that the action of these reagents is dominated by adenine-thymine richness alone. First, whereas micrococcal nuclease shows a preference for adenine-thymine-rich sequences, data on the binding preferences of 1,10-phenanthroline and other intercalating compounds indi-

cate some preference for G-C-rich DNA (5, 8, 16). In both cases, however, attempts to identify a consensus recognition site at the sequence level have been confounded by the strong influence of neighboring sequences. Second, in the specific case of the cuticle gene complex, an obvious gap in the cleavage pattern of both reagents occurs in a sequence for which no transcript has been identified and no pseudogene sequence is detectable and which has an overall G+C content similar to that of the more frequently cleaved spacer sequences in this same complex. Thus, it appears that specific base sequences and not general base composition are the principal basis for the observed cleavage patterns. We are attempting now to utilize the extensive sequence information available at this locus to classify a hierarchy of conditions which lead to the establishment of a nuclease-sensitive structure in naked DNA. The maintenance of noncoding eucaryotic DNA in a periodic repeating pattern detectable both by micrococcal nuclease and an intercalating reagent suggests a possible role for such conformational patterns in the organization of chromatin. Sequences leading to these conformations might be subject to selective pressure in eucaryotic DNA, in the absence of overriding protein-coding requirements. In that light, it is interesting to note that the pseudogene sequence retains its genelike nuclease-resistant character despite a number of coding defects which makes its expression extremely unlikely. One prediction that could be made from this argument is that intervening sequences, which are not subject to stringent peptide-encoding sequence con-

DNA STRUCTURE OF D. MELANOGASTER CUTICLE LOCUS

VOL. 3, 1983

wr A

reagents on both DNA and chromatin, at the cuticle gene locus and others.

2/3 B

C

1729

D

ACKNOWLEDGMENTS B

III I II

l

We thank M. Snyder and his colleagues for gifts of recombinant phages XDmLCP2 and XDm2/3LCP1 and recombinant plasmids pCPI-11 and pCPIII-9, M. Schuler for helpful advice on DNA sequencing, and M. Brandenberg for writing the program used to generate Fig. 6. This work was supported by Public Health Service grant GM30273 to S.C.R.E. from the National Institutes of Health.

11 LITERATURE CITED

I

I FIG. 7. Comparisons of the micrococcal nuclease patterns obtained for the cuticle sequences in (A, B) XDmLCP2 and (C, D) XDm2/3LCP1. Samples were treated as in the legend to Fig. 2, and the blot was probed with 32P-labeled pCPI-11. "B" in the righthand map indicates the position of the HMS Beagle transposable element, and the box at the end of the element indicates the position of one of the terminal direct repeat elements.

straints, might resemble intergenic spacers in a nuclease cleavage assay. Whereas the introns of the cuticle genes are too short to make a reliable test of this hypothesis, longer introns do indeed fit this model (Keene and Elgin, in preparation). An important inference we would like to draw from the DNA structural data so far discussed is that conformations recognized by diverse cleavage reagents might also be recognized by chromosomal proteins. In particular, conformational information might be utilized by histones to position nucleosomes in a nonrandom array to accommodate regulatory interactions or transcription in adjacent sequences or both. We are currently testing this inference by comparing the cleavage patterns for a variety of nucleolytic

1. Armstrong, K., and W. R. Bauer. 1982. Preferential sitedependent cleavage by restriction endonuclease Pst I. Nucleic Acids Res. 10:993-1007. 2. Bloom, K. S., and J. Carbon. 1982. Yeast centromere DNA is in a unique and highly ordered structure in chromosomes and small circular minichromosomes. Cell 29:305-317. 3. Cartwright, I. L., and S. C. R. Elgin. 1982. Analysis of chromatin structure and DNA sequence organization: use of the 1,10-phenanthroline-cuprous complex. Nucleic Acids Res. 10:5835-5852. 4. Elgin, S. C. R. 1981. DNase I-hypersensitive sites of chromatin. Cell 27:413-415. 5. Feigon, J., W. Leupin, W. A. Denny, and D. R. Kearns. 1982. Binding of ethidium derivatives to natural DNA: a 300 MHz 'H NMR study. Nucleic Acids Res. 10:749-762. 6. Hofer, B., G. Ruhe, A. Koch, and H. Koster. 1982. Primary and secondary specificity of the cleavage of 'single strand' DNA by endonuclease Hinf I. Nucleic Acids Res. 10:2763-2773. 7. Horz, W., and W. Altenburger. 1981. Sequence specific cleavage of DNA by micrococcal nuclease. Nucleic Acids Res. 9:2643-2658. 8. Howe-Grant, M., and S. J. Lippard. 1979. binding of platinum (II) intercalation reagents to deoxyribonucleic acid. Dependence on base-pair composition, nature of the intercalator, and ionic strength. Biochemistry 18:57625769. 9. Keene, M. A., and S. C. R. Elgin. 1981. Micrococcal nuclease as a probe of DNA sequence organization and chromatin structure. Cell 27:57-64. 10. Martin, R. F., and N. Holmes. 1983. Use of an "2I-labeled DNA ligand to probe DNA structure. Nature (London) 302:452-454. 11. Maxam, A. M., and W. Gilbert. 1980. Sequencing endlabeled DNA with base-specific chemical cleavages. Methods Enzymol. 65:499-560. 12. Smith, D. R., and S. M. Calvo. 1980. Nucleotide sequence of the E. coli gene coding for dihydrofolate reductase. Nucleic Acids Res. 8:2255-2274. 13. Snyder, M., M. Hunkapiller, D. Yuen, D. Silvert, J. Fristrom, and N. Davidson. 1982. Cuticle protein genes of Drosophila: structure, organization and evolution of four clustered genes. Cell 29:1027-1040. 14. Snyder, M. P., D. Kimbreil, M. Hunkapiller, R. Hill, J. Fristrom, and N. Davidson. 1982. A transposable element that splits the promoter region inactivates a Drosophila cuticle protein gene. Proc. NatI. Acad. Sci. U.S.A. 79:7430-7434. 15. Vieira, J., and J. Messing. 1982. The pUC plasmids, an M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19:259-268. 16. Wells, R. D. 1971. The binding of actinomycin D to DNA. Prog. Mol. Subcell. Biol. 2:21-32. 17. Wu, C., P. M. Bingham, K. J. Livak, R. Holmgren, and S. C. R. Elgin. 1979. The chromatin structure of specific genes. 1. Evidence for higher order domains of defined DNA sequence. Cell 16:797-806.