Proc. Nati. Acad. Sci. USA Vol. 82, pp. 4018-4022, June 1985 Biochemistry Torsional stress induces an S1 -hypersensitive site within the of the Xenopus laevis oocyte-type 5S RNA: gene : (/transcription factor/DNA-protein interaction) WANDA F. REYNOLDS AND JOEL M. GOTTESFELD Department of , Research Institute of Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, CA 92037 Communicated by James Bonner, February 22, 1985

ABSTRACT The internal promoter of the Xenopus laevis gene promoter adopts an S1 nuclease sensitive conformation oocyte-type 5S RNA gene is preferentially cleaved by S1 and in supercoiled DNA. This altered conformation may be Bal-31 in plasmid DNA. S1 nuclease sensitivity is similarly induced or stabilized by TFIIIA in linear DNA. largely dependent on supercoiling; however, Bal-31 cleaves These findings provide a clear correlation between an S1 within the 5S RNA gene in linear as well as in supercoiled DNA. nuclease-sensitive conformation and a promoter element. The S1 nuclease-hypersensitive site is centered at position +48-52 of the gene at the 5' boundary of the promoter. A METHODS DNase I-hypersensitive site is induced at this position upon DNAs. For DNase I "footprint" analysis, pXlo 3'A+56 binding of the transcription factor, TFIIIA, specific for the 5S was digested with EcoRI and end-labeled with polynucleotide RNA gene. The somatic-type 5S RNA gene promoter is not kinase (Bethesda Research Laboratories) and ['y-32P]ATP. preferentially cleaved by S1 nuclease or Bal-31 nuclease in The DNA was secondarily digested with HindIII and the supercoiled DNA, nor does TFIIIA induce a DNase I site at 530-bp fragment containing the gene was isolated by poly- position +50. This differential promoter response may be acrylamide gel electrophoresis. The fragment was labeled on related to a 4-fold difference in TFIIIA affinity between the the coding strand, 60 bp 3' to the gene. pXlsll was digested oocyte and somatic 5S RNA genes. with Hpa II and end-labeled with the Klenow fragment of DNA polymerase and [a-32P]dCTP. The DNA was secondar- The 5' flanking sequences oftranscriptionally active genes in ily digested with HindIII and the 380-bp fragment containing chromatin-are often hypersensitive to nucleases. This sensi- the gene was isolated from a polyacrylamide gel. The frag- tivity extends not only to DNase I, , and ment was labeled on the coding strand, 107 bp 5' to the gene. restriction enzymes (1) but also to single-strand-specific Digestion of DNA with S1 and Bal-31 Nucleases. DNA (1-2 reagents such as S1 nuclease and bromoacetaldehyde (2, 3). ,ug) was digested with up to 1 unit of S1 nuclease (Bethesda It has been suggested that hypersensitive regions correspond Research Laboratories or P-L Biochemicals) in a reaction to regulatory sequences and may be sites of interaction with volume of 30 p.1 in a buffer containing 30 mM sodium acetate effector proteins. Upstream elements that are hypersensitive at pH 4.8, 50 mM NaCl, and 1 mM ZnCl2. After a 30-min in active chromatin are in some instances S1 nuclease incubation at 37°C, the reaction was stopped with EDTA (20 hypersensitive in supercoiled DNA. The significance of this mM). The DNA was purified, digested with an appropriate correlation is made apparent by recent evidence that active , electrophoresed in 1.2% agarose gels in chromatin is subject to torsional stress (4, 5). 1 x TAE buffer (40 mM Tris acetate, pH 7.8/2 mM EDTA), Examples of upstream elements that are hypersensitive as transferred to nitrocellulose filters, and hybridized with a supercoiled DNA include those of the chicken P-globin gene nick-translated probe abutting the restriction site (15). (2, 6), Drosophila heat shock protein 70 (7) and histone genes Plasmids were digested with Bal-31 (Bethesda Research (8), sea urchin histone genes (9), adenovirus major late region Laboratories) (units, as defined by the supplier, indicated in (10), and the simian virus 40 origin/enhancer region (11). figures) for 1 min at 22°C in a 20-,ul reaction volume in a buffer Although these findings are provocative, there is as yet little containing 20 mM Tris HCl at pH 8.1, 100 mM NaCl, 12 mM correlation between hypersensitive sites (HSSs) in CaC12, 12 mM MgCl2, and 1 mM EDTA. After extraction with and known elements. In phenol and precipitation with ethanol, the DNA was digested supercoiled DNA promoter fact, and analyzed as described above. sequences giving rise to a HSS upstream of the Drosophila Mapping of S1 Nuclease Cutting Sites. pXlo8 (5-10 ,ug per heat shock protein 70 gene can be deleted without signifi- lane) was digested with S1 nuclease at low concentrations cantly affecting promoter function in vivo (12). Only in the (0.1-0.2 unit/30 ,ul) for 30 min at 37°C. This nuclease cases of certain (T)ATA homologies (6, 10) and sequences concentration produced less than 5% linear molecules and within the simian virus 40 origin/enhancer region (11) do such ensured a low level of nicking. The DNA was then digested HSSs correspond to identifiable promoter elements. To help with EcoRI and electrophoresed in a 1.2% agarose gel as clarify the biological significance of S1 nuclease hypersensi- above. Sections ofthe agarose gel containing DNA fragments tive structures, a system is needed in which the promoter of interest (see Fig. 3) were excised and the DNA was sequences are well defined and specific transcription factors electroeluted and digested with HindIII. The DNA was have been identified. The Xenopus SS RNA genes provide end-labeled with the Klenow fragment of DNA polymerase such a system. Deletion mutant analysis has delineated a and [a-32P]dATP, followed by electrophoresis in 6% promoter element internal to the 120-base-pair (bp) gene (13). acrylamide nondenaturing gels in TBE buffer (90 mM Tris A transcription factor, TFIIIA, specific for the 5S RNA gene borate, pH 8.3/2 mM EDTA). Labeled fragments were has been shown to bind this promoter region (14). In this excised from the polyacrylamide gel, crush-eluted and elec- report, we present evidence that the oocyte-type 5S RNA trophoresed in sequencing gels (7.6% acrylamide/0.4% bisacrylamide/8.3 M urea/ix TBE). The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: bp, base pair(s); HSS, hypersensitive site; TFIIIA, in accordance with 18 U.S.C. §1734 solely to indicate this fact. transcription factor specific for 5S RNA gene. 4018 Downloaded by guest on September 26, 2021 Biochemistry: Reynolds and Gottesfeld Proc. Natl. Acad. Sci. USA 82 (1985) 4019

DNase I Footprinting. TFIIIA was isolated from immature electrophoresed in agarose gels, blotted onto nitrocellulose oocytes as 7S particles consisting of TFIIIA in association and hybridized with a nick-translated probe abutting the with 5S RNA or as purified TFIIIA according to published EcoRI site. This indirect end-labeling procedure (15) revealed procedures (16, 17). Identical footprints were obtained with a repeating pattern ofcleavage sites within the 5S insert (Fig. 7S particles and purified TFIIIA. Singly end-labeled DNA 1B). S1 nuclease preferentially cleaves at the approximate fragments were incubated with saturating amounts of 7S center of each gene and pseudogene. The A+T-rich spacer particles in the presence of RNase A (10 gg/ml) for 10 min sequence, which should be the most likely to "breathe," is followed by DNase I digestion (2 pg/ml for 1 min at 220C). not preferentially cleaved. S1 nuclease cleavage within the 5S The DNA fragments were purified and electrophoresed on gene is largely dependent on negative supercoiling. Only at sequencing gels. the highest nuclease concentration employed was cleavage RESULTS observed within linear 5S DNA (Fig. 1C). Furthermore, the cleavage sites in linear DNA are less defined than within S1 Nuclease Cleavage of 5S Plasmid DNA. We examined the supercoiled DNA. A plasmid containing a single copy of the S1 nuclease sensitivity of a plasmid, pXlo8, containing four 5S gene and the 5' flanking sequences but lacking the tandem repeating units of oocyte-type 5S DNA. As shown in pseudogene gives rise to a similar S1 nuclease sensitivity Fig. lA, each 720-bp repeat contains the 5S gene, a pseudo- profile (Fig. 1D). S1 nuclease cleaves at one site central to the gene, and approximately 500 bp of spacer sequences (18). but does not cleave within The pseudogene is a direct duplication of the first 101-bp of gene the 5' or 3' flanking the gene with 10 base substitutions. Though apparently not sequences. There are a number of S1 nuclease cleavage sites transcribed in vivo, the pseudogene retains transcriptional within the prokaryotic vector sequences. We note, however, activity in cell-free extracts (19). Supercoiled pXlo8 was that within the 720-bp oocyte 5S DNA insert, only one site digested with S1 nuclease at concentrations producing 5-50% linear molecules. The DNA was then digested with A 0 0.1 0.05 0.02 0.01 0.1 0 0.1 0.05 EcoRI, NS.D SCD- A EcoRI H 720 H H H H BamrHI

SC- 11073120 -425 Pseudogene 5S gene A+T-rich spacer Si B Supercoiled Linear B Hindlil 0 1 D XIo H 01

1.i Ht

_ -Bam - _=, _ H U H w S....fl m

so H ml

FIG. 1. S1 nuclease cleavage of 5S plasmid DNA. (A) pXlo8 contains four 720-bp repeat units of Xenopus laevis oocyte-type 5S DNA (18) inserted at the HindIII site of pMB9. H, HindIII site. (B) H Supercoiled pXlo8 was digested with the indicated units of S1 nuclease followed by EcoRI. The DNA was electrophoresed on a 1.2% agarose gel, transferred onto nitrocellulose, and hybridized to FIG. 2. Bal-31 cleavage pattern. (A) Ethidium bromide-stained the nick-translated EcoRI/HindIII fragment of pMB9 (see A). An gel showing, in the first six lanes, supercoiled pXlo8 digested with the autoradiograph is shown. The marker lane (HindIII) contains EcoRI- indicated units of Bal-31 (1 min, 22°C). The last three lanes show digested pXlo8 partially cleaved with HindIlI. Diagram at right BamHI-linearized pXlo8 digested with Bal-31. N, nicked monomer; shows relative position of S1 cuts within 5S DNA insert. (C) SCD, supercoiled dimer; L, linear; SC, supercoiled. (B) Supercoiled EcoRI-linearized pXlo8 was digested with the indicated units of S1 pXlo8 (first five lanes) and BamHI-linearized pXlo8 (seventh to ninth nuclease and analyzed as in B. Markings at right indicate S1 cut sites lanes) were digested with the indicated units of Bal-31, giving rise to within the 5S DNA insert. (D) Xlo: S1 nuclease digestion pattern of the percent linear molecules shown in A. The DNA was digested with pXlo 3'A +56 containing 530 bp ofoocyte-type 5S DNA, including the EcoRI, electrophoresed, blotted, and hybridized with the nick- A+T-rich spacer, the gene, and 56 bp 3' to the gene but lacking the translated EcoRI/HindIII fragment of pMB9. The marker lanes (Hd) pseudogene. After S1 digestion, the DNA was cleaved with Pst I, contain EcoRI-digested pXlo8 partially cleaved with HindIII. The electrophoresed, blotted, and hybridized to the nick-translated Pst relative positions of Bal-31 cutting sites within the 5S genes and I/EcoRI fragment of pBR322. pseudogenes are indicated at left. H, HindIII site; Bam, BamHI site. Downloaded by guest on September 26, 2021 4020 Biochemistry: Reynolds and Gottesfeld Proc. Natl. Acad Sci. USA 82 (1985)

(two copies) is recognized and that site is within a regulatory supercoiled pXlo8. These enzymes, like Bal-31, appear to region (as shown below). recognize a broader region than the S1 site in supercoiled Bal-31 Cleavage of 5S Plasmid DNA. Bal-31 nuclease has DNA. The fact that several enzymes recognize a site within endonuclease as well as exonuclease activity. Like S1 the 5S gene in linear DNA suggests a structural feature as nuclease, Bal-31 preferentially cleaves single-stranded DNA. opposed to a mere sequence preference. Unlike S1, Bal-31 is active at physiological pH. Supercoiled Mapping the S1 Nuclease Cutting Site. To map the S1 pXlo8 was digested with Bal-31 at concentrations resulting in nuclease cleavage site precisely, fragments having an Sl-cut 5-70% linear molecules (Fig. 2A), and this digestion was end were electroeluted from the agarose gels, secondarily followed by EcoRI digestion, electrophoresis, blotting, and digested to create smaller fragments, and sized on polyacryl- hybridization as described above. As shown in Fig. 2B, amide gels. As shown in Fig. 3A, we excised that region ofthe Bal-31 preferentially cleaves within the 5S gene and pseudo- agarose gel containing DNA fragments extending from the gene, producing a pattern similar to that obtained with S1 EcoRI site to S1 nuclease cuts within the second 5S repeat nuclease, although the cleavage sites are broader. The unit. The fragments, 1000-1500 bp long, were electroeluted breadth of the Bal-31 cleavage site could be due in part to the and digested with HindIII to create fragments 370 bp long exonucleolytic activity ofthis enzyme. We also examined the (EcoRI to HindIII), 720 bp long (corresponding to the intact Bal-31 cleavage pattern on linear 5S plasmid DNA. To avoid first repeat), and less than 300 bp long (resulting from S1 exonucleolytic digestion at the EcoRI end, pXlo8 was initially nuclease cuts within the gene and pseudogene of the second linearized with BamHI (for restriction map, see Fig. LA), repeat). The fragments were end-labeled and displayed in followed by Bal-31. The DNA was then secondarily digested nondenaturing gels as shown in Fig. 3B. This revealed a major with EcoRI, electrophoresed, blotted, and hybridized to the S1 nuclease cleavage site between positions +35 to +55 of probe abutting the EcoRI end. Surprisingly, Bal-31 cleaves the gene and pseudogene. (There is a second cleavage site the 5S gene and pseudogene with equal efficiencies in linear within the pseudogene between +5 and +15, which presum- and supercoiled DNA. Other enzymes such as Neurospora ably results from sequence changes.) The labeled fragments crassa single-strand nuclease, micrococcal nuclease (data corresponding to the hypersensitive site within the 5S gene not shown), and, to a lesser extent, S1 nuclease, cleave were crush-eluted from the nondenaturing gel and electro- within the 5S gene and pseudogene in linear as well as in phoresed under denaturing conditions. This indicated a

A H H H H H RI w _ ~~~~X,..I

Electroelute HindII

RI H i H 370 720 Si ends FIG. 3. Localization of S1 nuclease cutting site. (A) The autoradiogram shows S1 nuclease-digested pXlo8, analyzed by indirect end-labeling as in Fig. 1B. B 0 0.1 370 M The region of a corresponding agarose gel containing DNA fragments with S1 nuclease cut ends within the gene or pseudogene of the second repeat unit was 404 excised (as indicated by brackets) and the DNA was electroeluted and digested with Hind1lI to give rise to _ 309 smaller fragments as shown. H, HindIII site; RI, EcoRI site. (B) pXlo8 was digested with the indicated units of S1 nuclease and treated as described for A. The resulting fragments were end-labeled and elec- 238242 trophoresed on nondenaturing 6% acrylamide gels. 242- 217 Markers are Hpa II fragments ofpBR322. Positions of 17 the major S1 cuts within the gene and pseudogene are 201 indicated at right. (C) The region of the gel in B (lanes 190 0.2 and 0.3) containing fragments with Si-cut ends 160* . ISO8 within nucleotides +35-55 of the gene (indicated by 147- bracket) was excised and the DNA was crush-eluted 120- and electrophoresed in a denaturing acrylamide gel 110. (corresponding lanes 0.2 and 0.3). pBR322 markers (lanes M) were calibrated with 5S restriction frag- 76 ments. The lane marked 370 refers to the 370-bp 90- EcoRI/HindIII fragment crush-eluted from the gel in 76- 1 B to assay extent of S1 nicking. (D) The region of the 67- 67 gel in B containing fragments with Si-cut ends within oW nucleotides +35-55 of the pseudogene was excised and the DNA was crush-eluted and electrophoresed under denaturing conditions as in C. The first two lanes contain DNA eluted from lanes 0.2 and 0.3 in B. The third lane contains pBR322 markers. At left, the location of cutting sites is indicated (nucleotides +39-55). Downloaded by guest on September 26, 2021 Biochemistry: Reynolds and Gottesfeld Proc. Natl. Acad. Sci. USA 82 (1985) 4021 cluster of minor cutting sites surrounding a distinct site at transcribed only during oogenesis and early development +48-52 (Fig. 3C) at the 5' boundary ofthe internal promoter. (21). The somatic-type genes are not developmentally regu- The sequences surrounding this site exhibit a moderate lated and are transcribed in all active cell types, including degree of purine-pyrimidine asymmetry (Fig. 4B). The oocytes. We examined the response of somatic 5S DNA to noncoding strand sequences extending from +47 to +61 are superhelical stress. pXlsll contains a single 883-bp repeat purine rich (12/15), whereas those 5' of the site are pyrimi- unit of X. laevis somatic SS DNA. The repeat contains one dine rich (11/14). Fragments corresponding to S1 nuclease copy of the gene and 760 bp of G+C-rich spacer sequences cleavage at the 5' boundary ofthe pseudogene promoter were (20). Supercoiled pXlsll was digested with S1 nuclease or similarly crush-eluted and displayed under denaturing condi- Bal-31 and then with Sal I. The DNA was electrophoresed in tions (Fig. 3D). Cleavage sites extend from +39 to 55, agarose gels, blotted onto nitrocellulose, and hybridized with although S1 "nibbling" of breathing ends may create some a probe abutting the Sal I site. As shown in Fig. SA, major S1 length heterogeneity. Also, the pseudogene sequence. differs nuclease sites are situated at approximately -150 and -25 from that ofthe gene at three positions within this region (Fig. from the 5' end of the gene but not within the gene. Similarly, 4B) such that the exact pattern of cleavage may differ. Bal-31 cleaves at or near the -150 site but does not recognize TFIIHA-Induced HSS at +50. As shown by DNase I the -25 region or the internal promoter. The two types of SS footprinting (Fig. 4), TFIIIA binds the internal promoter from RNA genes have radically different upstream sequences. The +45 to +95 (14). Within the coding strand footprint, TFIIIA oocyte SS gene is flanked by A+T-rich sequences, whereas induces a DNase I-HSS at +49-51 within the oocyte-type 5S the somatic 5S gene is flanked by G+C-rich sequences gene (14). This suggests that TFIIIA binding might, like containing several regions of marked purine/pyrimidine superhelical stress, induce an altered conformation at the 5' asymmetry (18, 20). It is possible that the 5' flanking promoter boundary. We note that the +50 region within the sequences could influence the conformation of the 5S pro- noncoding strand footprint is not hypersensitive (14). In- moters. For example, the sites upstream of the somatic 5S stead, enhanced cleavages are located at 60-64 and 74-76. gene could act as a sink for torsional stress, thus preventing Nuclease Sensitivity of Somatic 5S DNA. In Xenopus, there emergence of an intragenic site. Alternatively, the absence of are two major families of 5S RNA genes, the oocyte type, a site within the somatic promoter may result from sequence present in 20,000 copies per haploid , and the somatic changes (Fig. 4B). The oocyte and somatic SS promoters type, present in 400 copies. The oocyte 5S RNA genes are differ in sequence at four positions (+47, 53, 55, and 56) near the 5' boundary. Interestingly, a TFIIIA footprint on the A M 1 2 coding strand of the X. laevis somatic 5S gene does not include an enhanced DNase I cleavage at +50 (Fig. 5B). 160 % 147 DISCUSSION 45 The 5' boundary of the oocyte-type 5S RNA promoter is 122 -15 cleaved preferentially by S1 nuclease. This site centered at 110 W +48-52 is largely dependent on supercoiling. In contrast to S1 nuclease, Bal-31 cleaves within the 5S RNA gene in linear as well as supercoiled DNA. This indicates an unusual DNA 90 % A Bal Si B - - 76 ,__ -95 N L_w 67 ~mm

-5 Lg5

TFIIIA B DNase OR~~~\ 40 w0 _ 50 Xlo G T C TG A T C T C A G A A G C G A T A C A G G G T A5 Si nuclease 872- ,4, _- XIS G T C T G A T C T C G G A A G C C A A G C A GG G T XIo 4/ G T C T G A T C T jG A A G T G A T A C A G G G G 603- OR_ O.- do- FIG. 4. DNase I-hypersensitive site in presence of TFIIIA. (A) _. Al. DNase I footprint analysis of TFIIIA bound to oocyte 5S RNA gene promoter. A singly end-labeled (coding strand) fragment containing FIG. 5. Nuclease sensitivity of somatic 5S DNA. (A) Bal-31 and the oocyte 5S gene was incubated in the absence (lane 1) or presence S1 nuclease sensitivity. pXls11 contains a single HindIII repeat unit (lane 2) of TFIIIA and lightly digested with DNase I. The DNA was (883 bp) of X. laevis somatic 5S DNA inserted at the HindIII site of purified and electrophoresed in a sequencing gel. An autoradiograph pBR322 (20). Supercoiled pXls11 was digested with 0.1 unit of Bal-31 is shown. TFIIIA protects a region extending from nucleotides +45 (1 min 22°C) or 1 unit of S1 nuclease (30 min, 37°C) and then digested to 95 of the gene with an enhanced DNase I site at nucleotide 49-51. with Sal I. The DNA was electrophoresed, blotted, and hybridized The markers (M) are Hpa II-digested pBR322 fragments. The with the nick-translated Sal I/HindIll fragment of pBR322. Markers position of the DNase I-hypersensitive site was furthersorroborated are end-labeled Hae IlI-digested DNA of phage 4X174. The diagram by comparison with a G-only cleavage of the 5S fragment (not indicates the relative position of the 5S DNA insert. (B) DNase I shown). (B) Sequences are shown for nucleotides 37 to 62 of the footprint analysis of TFIIIA bound to somatic 5S DNA. A singly Xenopus laevis oocyte-type 5S gene (Xlo) (18), the somatic-type end-labeled (coding strand) fragment containing the somatic 5S RNA (Xls) (20), and the Xlo pseudogene (Xlo 4O) (18). In Xlo, the arrow gene was incubated in the absence (-) or presence (+) of TFIIIA as above indicates the position (+2) of the TFIIIA-induced DNase I 7S particles and lightly digested with DNase I. The DNA fragments cleavage site, The arrows below indicate the major S1 nuclease were isolated and displayed in sequencing gels. An autoradiograph is cleavage sites in supercoiled DNA. In Xls and Xlo 41, nucleotides that shown. TFIIIA protection extends from approximately +45 to 95. differ from the Xlo sequence are underlined. The +50 region is not hypersensitive. Downloaded by guest on September 26, 2021 4022 Biochemistry: Reynolds and Gottesfeld Proc. Natl. Acad. Sci. USA 82 (1985) secondary structure present in linear molecules that is ac- increased TFIIIA affinity of the somatic gene derives from centuated by torsional stress to become accessible to S1 sequence differences near the 5' promoter boundary (19). nuclease. A TFIIIA-induced DNase I site at +50 in linear These sequence changes (47, 53, 55, and 56) presumably DNA suggests that factor binding might trigger or stabilize a account also for the lack of S1 nuclease and Bal-31 sensitivity similar conformational change. at +50 of the somatic 5S promoter. This raises the possibility The question remains as to the nature of the S1 nuclease- that the nuclease-sensitive conformation within the oocyte 5S sensitive DNA conformation. The S1 site is dependent, for promoter is related to the 4-fold difference in TFIIIA affinity. the most part, on supercoiling. Since DNA unwinding re- These findings suggest that the oocyte 5S promoter under- lieves superhelical stress, unwinding probably accompanies goes a conformational change in response to torsional stress. site formation. DNA configurations that result in unwinding The functional relevance of this response is suggested by and are associated with S1 HSSs include Z-DNA (22), recent evidence linking torsional stress and active chromatin. cruciforms (23) and possibly DNA slippage along direct The transcriptionally active fraction of simian virus 40 repeats (7). The sequences surrounding the 5S DNA HSS are minichromosomes or 5S plasmid DNA assembled in Xenopus not compatible with these structures. The site is, however, oocytes can be relaxed with topoisomerase I or DNase I (4, situated within a region of purine-pyrimidine asymmetry, a 5). Novobiocin, an inhibitor oftopoisomerase II, has a similar characteristic commonly associated with S1 nuclease sensi- relaxing effect, suggesting that negative supercoils are ac- tivity in or near regulatory DNA. Such asymmetry is thought tively introduced by this enzyme (5). These findings imply to cause helical distortion, perhaps resulting in non-B-form that active nucleosomes do not constrain DNA supercoils but DNA (11). are capable of disassembling, allowing the DNA to uncoil. The +50 HSS is situated within a region shown by deletion The energy of negative supercoils would then become avail- analysis to be crucial for transcription initiation. Deletion of able to drive the DNA into alternative conformations. If, like the first 50 bp of the Xenopus borealis somatic 5S RNA gene S1 nuclease, specific DNA binding proteins are able to still resulted in 5S-sized transcripts that initiated within recognize variations in helical geometry as well as nucleotide vector sequences. Deletion to +55, however, allowed little to sequence, a promoter might exhibit optimal factor affinity no transcription (13). This 50-55 region, though required for only when associated with active nucleosomes. transcription, is not essential for TFIIIA binding (24). Those We thank D. D. Brown and co-workers for the generous gifts of5S nucleotides that are essential and that make tight contact with plasmids. We also thank E. Plaza for preparing the manuscript. This the factor are located at the 3' end of the promoter between work was supported in part by grants from the National Institutes of positions +70 and 90 (25). Nevertheless, TFIIIA interaction Health (GM26453 and CA06824). This is publication no. 3476-MB with the 5' region of the promoter is important for gene from the Research Institute of Scripps Clinic. activity. Partial proteolysis of the 40-kDa factor has been shown to result in three domains (17). A 20-kDa domain that 1. Elgin, S. C. R. (1984) Nature (London) 309, 213-214. 2. Larsen, A. & Weintraub, H. (1982) Cell 29, 609-622. binds the 3' promoter region is insufficient for transcription 3. Kohwi-Shigematsu, T., Gelinas, R. & Weintraub, H. (1983) Proc. in systems reconstituted in vitro. An adjacent 10-kDa domain Natl. Acad. Sci. USA 80, 4389-4393. interacts with the 5' region of the promoter and permits a low 4. Luchnik, A. N., Bakayev, V. V., Zbarsky, I. B. & Georgiev, G. P. level of transcription. A third domain of 10 kDa apparently (1982) EMBO J. 1, 1353-1358. does not bind the DNA directly but does give rise to 80% of 5. Ryoji, M. & Worcel, A. (1984) Cell 37, 21-32. the effect of TFIIIA. Interestingly, this latter 6. Schon, E., Evans, T., Welsh, J. & Efstratiadis, A. (1983) Cell 35, stimulatory 837-848. domain also enhances DNase I hypersensitivity at the 5' 7. Mace, H. A. F., Pelham, H. R. B. & Travers, A. A. (1983) Nature boundary of the footprint and within the 5' half of the (London) 304, 555-557. promoter. Thus, the induction of HSSs within this region is 8. Glikin, G. C., Gargiulo, G., Rena-Descalzi, L. & Worcel, A. (1983) associated with the stimulatory activity of TFIIIA. Nature (London) 303, 770-774. Unlike the oocyte-type 5S promoter, the somatic-type does 9. Hentschel, C. C. (1982) Nature (London) 295, 714-716. to stress. 10. Goding, C. R. & Russell, W. C. (1983) Nucleic Acids Res. 11, not give rise to a HSS in response superhelical 21-26. Neither S1 nuclease nor Bal-31 cleaves within the gene. The 11. Evans, T., Schon, E., Gora-Maslak, G., Patterson, J. & oocyte and somatic promoters also respond differently to Efstratiadis, A. (1984) Nucleic Acids Res. 12, 8043-8072. TFIIIA binding. The TFIIIA footprint within the somatic 5S 12. Dudler, R. & Travers, A. A. (1984) Cell 38, 391-398. gene does not include the +50 HSS characteristic of the 13. Sakonju, S., Bogenhagen, D. F. & Brown, D. D. (1980) Cell 19, gene. it appears that a 13-25. oocyte-type Thus, supercoil-induced 14. Engelke, D. R., Ng, S.-Y., Shastry, B. S. & Roeder, R. (1980) Cell S1 nuclease HSS at +50 correlates with a factor-induced 19, 717-728. DNase I HSS at +50. This correlation suggests that the 15. Wu, C. (1980) Nature (London) 286, 854-860. oocyte 5S promoter undergoes an analogous conformational 16. Pelham, H. R. B. & Brown, D. D. (1980) Proc. Natl. Acad. Sci. change in response to torsional stress or TFIIIA binding. USA 77, 4170-4174. There is reason to suspect that TFIIIA could locally stress the 17. Smith, D. R., Jackson, I. J. & Brown, D. D. (1984) Cell 37, unwinds 5S DNA to a small 645-652. DNA in that it oocyte-type 18. Miller, J. R., Cartwright, E. M., Brownlee, G. G., Fedoroff, N. V. degree (26) and has single-stranded binding preference (27). & Brown, D. D. (1978) Cell 13, 717-725. The differential response of the oocyte and somatic pro- 19. Wormington, W. M., Bogenhagen, D. F., Jordan, E. & Brown, moters may be related to their differential expression in vivo. D. D. (1981) Cell 24, 809-817. During oogenesis, the oocyte 5S genes are transcribed at high 20. Peterson, R. C., Doering, J. L. & Brown, D. D. (1980) Cell 20, efficiency. Transcription falls during oocyte maturation and 131-141. 5S 21. Wormington, W. M. & Brown, D. D. (1983) Dev. Biol. 99, 248-257. is not detectable in somatic cells (21). The somatic genes 22. Singleton, C. K., Klysik, J., Stirdivant, S. M. & Wells, R. D. are transcribed in all cell types, including oocytes. The basis (1982) Nature (London) 229, 312-316. of this differential regulation is unclear, since in vitro both 23. Panayotatos, N. & Wells, R. D. (1981) Nature (London) 289, genes apparently require the same transcription factors in 466-470. addition to RNA polymerase III. The somatic promoter does, 24. Sakonju, S., Brown, D. D., Engelke, D., Ng, S., Shastry, B. S. & however, exhibit a 4-fold greater affinity for TFIIIA in Roeder, R. G. (1981) Cell 23, 665-669. vitro This 4-fold 25. Sakonju, S. & Brown, D. D. (1982) Cell 31, 395-405. competition studies in (19). binding prefer- 26. Reynolds, W. & Gottesfeld, J. (1983) Proc. Natl. Acad. Sci. USA ence, in conjunction with decreasing TFIIIA concentrations 80, 1862-1866. during embryogenesis, is thought to play a role in the 27. Hanas, J., Bogenhagen, D. & Wu, C. W. (1984) Nucleic Acids Res. developmental inactivation of the oocyte 5S genes. The 12, 2745-2758. Downloaded by guest on September 26, 2021